With Voice and Pen
This page intentionally left blank
With Voice and Pen coming to know medie val song and ho w it wa s made
}
Leo Treitler
1
3
Great Clarendon Street, Oxford, ox2 6dp Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi São Paulo Shanghai Taipei Tokyo Toronto Oxford is a registered trade mark of Oxford University Press in the UK and certain other countries Published in the United States by Oxford University Press Inc., New York
© Leo Treitler 2003 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2003 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press or as expressly permitted by law, or under the terms agreed with the appropriate reprographics rights organizations. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above. You must not circulate the book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging-in-Publication Data Treitler, Leo, 1931– With voice and pen: coming to know medieval song and how it was made / Leo Treitler. p. cm. Includes bibliographical references (p. ) and index. Contents: Medieval improvisation – Written music and oral music : improvisation in medieval performance – The Vatican organum treatise and the organum of Notre Dame of Paris : perspectives on the development of a literate music culture in Europe – Peripheral and central – On the structure of the Alleluia melisma : a Western tendency in Western chant? – Homer and Gregory : the transmission of epic poetry and plainchant – Centonate chant : Übles Flickwerk or e pluribus unus? – Lingering questions about oral literature – The politics of reception: tailoring the present as fulfilment of a desired past – Oral, written, and literate process in the music of the Middle Ages – Observations on the transmission of some Aquitanian tropes – History and ontology of the musical work – The early history of music writing in the West – Reading and singing: on the genesis of Occidental musical writing – Speaking of Jesus – Medieval music and language – The marriage of poetry and music in medieval song. 1. Vocal music – 500–1400 – History and criticism. 2. Music and literature. I. Title. ml1402.t73 2003
782.42'09'02–dc21
2002192579 isbn 0-19-816644-3 1 3 5 7 9 10 8 6 4 2
Typeset by Figaro, Launton, ox26 5dg Printed in Great Britain on acid-free paper by Biddles Ltd., Guildford & King’s Lynn
This book is dedicated to Helmut Hucke and to the memory of Barbara Thornton and Fritz Reckow
This page intentionally left blank
Pre face
vii
Preface When, in 1960, I reached that stage in my graduate studies in musicology at Princeton University when one declares a dissertation topic, I set out to choose, as I now reflect, with a good deal more naivety than is displayed by most dissertation students whom I have since been privileged to advise. I would pursue a medieval topic, I decided, and in particular the question of ‘the origin of the conductus’. That the easy use of the ‘conductus’ label as though to designate a clearly circumscribed item in the modern panorama of ‘medieval music’ is hardly justified by medieval usage or practice, and that this can give a measure of how much invention has been entailed in the construction of that panorama and altogether of the panorama of ‘Middle Ages’, is something of which my awareness was still well in the future at that time. And the decision that the burning question about such a shadowy genre should be just the one about its origin was more a matter of falling into a disciplinary habit of the time than of active decision. When I proposed the topic to Oliver Strunk and asked him to be my adviser he expressed surprise. I had arrived at Princeton as a foreigner in such territory (something that I now think worked to my advantage), having first earned a master’s degree jointly in music composition and history at the University of Chicago, writing a thesis on harmonic procedure in the Fourth String Quartet of Béla Bartók, and then spent a year studying composition in Berlin. Strunk was not much impressed with the music involved, but he allowed that ‘the problems are pretty interesting all right’, and agreed to oversee my work. He seemed to know better than I—and conveyed in that characteristically spare way of his—how deeply the theoretical problems and topics entailed in the study of medieval music would engage me. As one does in developing a dissertation topic we reined mine in, making it more concrete, less ambitious, and more realistic. It would be a study of a tradition of monastic songs with rhythmic verse on sacred topics originally transmitted in Aquitanian manuscripts of the eleventh and twelfth centuries, a rich tradition of art songs made for virtuoso singers whose study did not really require the justification of a role in the investigation of origins that had led me to them. Once into those repertories I came to think otherwise about Strunk’s assessment of their musical qualities, and I have since been struck by the number of modern performing ensembles who have drawn from them in their concerts and recordings. I found there were thirty-three songs that I could read and place in their intended tonal frameworks with some confidence. The first task was to establish texts for them, and with that my learning began. Of the thirty-three songs of my
viii
Pre face
repertory, twenty-five are represented in at least two sources, and no two sources of any song in the collection are identical in their musical texts. In the exceptional case of one song for which notation had been written out for all the strophes (the songs are all strophic and normally the melody was written out just once) no two strophes are identical in the notation even of a melisma that concludes each one, where the differences are not explained by the singing of different words. My training and experience with Western music of later times led me to think that I ought to be able to explain such differences as reflections of different ideas about the structures of the songs, but as far as I could tell the differences did not seem consequential in that way and usually one version seemed to work as well as another. And then there was always the hovering question of the determinacy of the neumes, which, to be sure, did seem to be oriented to a single etched line as a way of conveying information about relative pitch distance between notes, but with which one can never be quite certain that an apparent difference is an intentional one. As long as I reported variants I believed I was serving the material as well as could be done. I first posed this general problem in the original version of Chapter 4, where I suggested that it should be regarded not only as an irritant to editors but also as a potential source of insight into the nature of the early medieval music culture. That is how I regarded it, and it opened up successive waves of questions that have engaged me, on and off, during some forty years, and well beyond the domain of medieval music in many directions: what is the ontological state, the state of being, of music that is represented in such a fluid way? How can we understand its principles of organization and progression and how can we represent what such an item essentially is? What were the traditions and conventions on the basis of which it was composed? How was it transmitted and known to the people who sang it? How firmly did the notators mean just exactly what they wrote down? Did the notations serve them and the singers as prescriptions for the singing, or as descriptions, or as touchstones, or as exemplifications within a range of possibilities for the singing? In any case what were the principles on which the notations functioned and what can we understand about the invention of the notational systems and what prompted it? How was music composed, transmitted, and learnt before their invention? Each question provoked others and as they unfolded they linked together in a dense network rather than any kind of logical sequence or hierarchical structure. The network spread from philology to history, criticism, and philosophy. But then, that is the original, broad sense of philology. One of the most important learning experiences for me over this long span of time was a growing awareness of the handicaps under which we, with mental resources and forms of thought that we have accumulated over centuries after the fact but especially since the nineteenth century, labour in trying to capture the spirit of a musical culture that was alive a millennium and more ago. I mean
Pre face
ix
the resources of language, of aesthetic beliefs, of analytical procedures, of images of the music-historical geography, landscape, and time spans, with their peaks and troughs and central and peripheral phenomena, the influence of feelings of cultural and national, racial, and gender identity and alienation on historical thinking, the modes of representing histories in narratives and panoramas, ideas about how things change, evolve, develop, progress, ideas of what is musical and what is not, ways of thinking about the ontology of music and about the relations between language and music, of our expectations and uses of musical notation. I intend the contradiction implied in the juxtaposition of ‘handicap’ and ‘resources’, which is the dilemma that we share with all historians but which I believe is especially acute for historians of the Middle Ages. That is because ‘The Middle Ages’ and things ‘medieval’ bear, more than other epochs and the creations of their cultures, the burden of the purposes for which they were invented and perpetuated as historiographic concepts by people who were not members of the culture they invented. Brian Stock, a historian of medieval literature, has caught the weight of this burden: ‘The Renaissance invented the Middle Ages in order to define itself, the Enlightenment perpetuated them in order to admire itself; and the Romantics revived them in order to escape from themselves . . . “The Middle Ages” thus constitutes one of the most prevalent cultural myths of the modern world’ (‘The Middle Ages as Subject and Object’). For ‘medieval music’ that means it must be both pejorated in the service of the Renaissance and the Enlightenment (representations of medieval music have been especially constrained by this role, as we see in Ch. 16) and idealized in the service of the Romantics. And we must add a role as foundation and model of European or Western music, which has been highly influential in the characterization and judging of medieval music (see, especially, Ch. 9). But I said ‘dilemma’, for we cannot divest ourselves of our mental resources and replace them with those of our medieval objects, which we can only try to infer with hopes of occasional limited success. (In Ch. 5 I report on one such success achieved by Fritz Reckow, one which would, if allowed, constitute a basis for the revision of much of the modern characterization of medieval song; another is Mary Carruthers’s The Book of Memory: A Study of Memory in Medieval Culture, which could contribute substantially to our conjurings about the role of memory in the transmission and performance of medieval song.) The best we can do, beyond such occasional successes at capturing medieval beliefs and conceptions, is to be alert to representations that sound suspiciously like our own conceptions, projected like the voice of a ventriloquist to the bodies of medieval dummies of the historian-puppet-master’s manufacture. This book is meant to be attentive throughout to the relations between representation and reality that is thus encountered—in Chapters 1 and 2, 6, 7, and 8, 9, 11, 13, 15, and 16, and most explicitly Chapter 5, where I reprint and, hopefully, correct my own early disfigurement
x
Pre face
of a phenomenon of style history within a medieval genre in accordance with an ideological map of the world’s music in the Middle Ages—as egregious an instance of historiographic conditioning of historical discovery as can be found in the literature of this field. This question about ‘Coming to Know’ is as much the subject of this book as are the particular song traditions that are studied in its chapters and the central questions about their composition and transmission—orally through improvisation, working out (precomposition, as some put it), and memory; and scriptually through notation. Indeed it was more than anything else in order to make explicit in the new introductions that I have added the—often tacit—epistemological conditions of study that I took the decision to join these far-flung essays in a book. But that also provided a welcome opportunity to revise them where that seemed necessary and to take account of other work in the field that has appeared since the initial publications. There are references here and there in the literature to an ongoing ‘debate’ about topics discussed in this book that have to do mainly with aspects of the history of Latin chant, particularly its composition and transmission—oral tradition, improvisation, the role of memory, the stability of the repertory, the beginning of written transmission and its role vis-à-vis performance (for example, László Dobszay, ‘The Debate about the Oral and Written Transmission of Chant’). I believe that the disagreements about these matters—they are hardly debates in any formal sense—arise mostly out of different ways of construing the processes concerned and the names for them. They can be sublimated with attention to some of the epistemological matters mentioned just above, and with a bit of linguistic archaeology. Both are attempted in this book (see Chs. 1 and 2 with respect to ‘improvisation’ and Ch. 6, especially the introduction, with respect to ‘memory’). There is, however, one object of contention that has at least the appearance that it might be settled through the analysis of hard evidence. The oldest surviving specimens of musical notation from the medieval West have been dated to the first third or so of the ninth century. The oldest books of chant provided throughout with musical notation have been dated to some time around the turn of the ninth to the tenth century. Kenneth Levy has compiled evidence in support of his hypothesis that a fully neumated chant book was nevertheless available to the court of Charlemagne a century before the writing of the oldest surviving ones, and that it played a key role in the court’s project of replacing local chant traditions with the tradition of Rome. The evidence has been discussed in several publications, and there is no consensus about its conclusiveness (see Ch. 6 for details). But the importance of this question, too, depends on a matter of epistemology. Can it be assumed that the members of chant communities singing since childhood in oral traditions, upon receiving for the first time a text of a tradition different from theirs, would have accepted its authority and turned at once to the
Pre face
xi
project of replacing their tradition with the foreign one represented by the text? This would have entailed not only unlearning their tradition and learning a foreign one, it would have had to entail as well a radical cognitive shift, creating a role for a written text to intercede in the pathway from ear to voice—at least for their leaders. No evidential support for such a scenario has turned up; on the contrary, only anecdotal reports about singers, not books, being sent around to effect such an exchange, and about chant communities blaming one another for the initial failures of the project. But it would seem to be an unspoken premiss for asserting the significance in this sense of any proposed dating of the origin of the written tradition. In a new propaedeutic for chant studies Richard Crocker relegates such questions to what he calls the ‘prehistory’ of Gregorian chant because they concern a time from which we have no written sources, and he wants us to regard speculation about them as ‘romantic fantasy’. He consequently urges a change of focus to the critical study of chant that is preserved in readable sources (see the introduction to Ch. 2 for details). The posing of such alternatives and the recommendation to choose one over the other suggests an isolation from the broad range of historical fields that has much bedeviled music-historical studies in general. Historians of all subjects have ever occupied themselves with reconstructions and hypotheses that come down essentially to informed and reasoned imaginings about how the surviving evidence came to be how it is. I am thinking not only of workers in disciplines that identify themselves explicitly as some kind of ‘history’, but also of palaeontologists, palaeographers, botanists, zoologists, students of evolution, archaeologists, astronomers, linguists. As for the history of chant, unless one assumes that the pieces we have in writing were all composed at the time they were initially written down—something that no one has suggested—they are clues to the practices with which they are continuous, no matter how well or badly we exploit them as such. I doubt that our curiosity about the connections will be dampened, any more than would the curiosities of historians in all those other fields about their ‘prehistories’. We are all guided, consciously or not, by the idea expressed in R. G. Collingwood’s classic formulation (The Idea of History): How, or on what conditions, can the historian know the past? In considering this question, the first point to notice is that the past is never a given fact . . . If the historian has no direct or empirical knowledge of his facts, and no transmitted or testimonial knowledge of them, what kind of knowledge has he? My historical review of the idea of history has resulted in the emergence of an answer to this question: namely that the historian must re-enact the past in his own mind.
xii
Pre face
There should be no disputing, in any case, that performance was the focus of traditions of medieval song, and I could not conceive of making a book about those traditions that would illustrate them only with scores. The accompanying compact disc, I feel, transforms the book. It, as representative of the work of the performers beyond it, makes its own research contribution, in making manifest the musicality and virtuosity of singers in all aspects of performance that must have been assumed by composers and pedagogues—if these were other than singers themselves. That must be a major component of any representation of medieval song traditions, although it has had hardly any place in scholarly representations and is in effect suppressed in the aural representations of modern performance traditions such as the chanting tradition that follows the style created by the choirs of the monastery of St-Pierre of Solesmes and that has been widely disseminated in recordings. The production of the recording was supported with generous grants from the Research Foundation of the City University of New York, where I was a member of the faculty of the Graduate Center during the entire time of the preparation of this book; and from the Weiss–Brown Publication Subvention Award of the Newberry Library in Chicago. I was particularly moved to have the latter support. Howard Brown was my colleague at the University of Chicago very early in my work on this subject. He was the director of the Collegium Musicum there, in which his companion Roger Weiss was a tenor. When I gave a talk on work coming out of my dissertation—my first—at the national meeting of the American Musicological Society in Columbus, Ohio, in 1961, Howard’s ensemble illustrated it. So I feel, in a sense, that the project has come full circle. I have been fortunate, during this work, to be nourished by continuous contact through nearly half a century with a number of mainly European colleagues and friends who share my passion for our subject and from whom I have learnt more and received more stimulation and support than I can possibly recount, contact without which I could not have produced work of any merit in this field. They are Helmut Hucke, the late Fritz Reckow, Hartmut Möller, and Andreas Haug in Germany; László Dobszay in Hungary; Wulf Arlt and Max Haas in Switzerland; Ritva Jacobsson, Gunilla Björkvall, and Gunilla Iversen in Sweden; Susan Rankin in England; the late Nino Pirrotta in the United States and Italy; Edward Nowacki and Charles Atkinson in the United States. Their influence is evident throughout this book, which I offer as a measure of my appreciation and of my respect for them. Just typing their names brings a rush of blood to my head. I have been equally fortunate in being able to recruit for the performances on the compact disc a number of friends whom I count among the great singers of medieval music: Benjamin Bagby and the late Barbara Thornton of the ensemble Sequentia; Dominique Vellard of the Ensemble Gilles Binchois; Katerina
Pre face
xiii
Livljanich of the ensemble Dialogus. Their singing has enlivened my understanding of medieval song beyond what I could have imagined. I am grateful to them for that and for their enthusiastic contributions to this project. The decision to add the Blues performance by Lightnin’ Hopkins came very late. The parallels with certain kinds of medieval performance models that are discussed in Chapter 6 have long been evident to me. I thought I should share that impression. But then, too, I decided to make explicit the exception I take to an attitude that runs just below the surface of some of the literature in the field, that emphasis on oral tradition and, even more, improvisation, is an offence to the dignity and worth of a tradition like that of Gregorian chant. At the same time I signal thereby my regret that the study of oral traditions and of improvisation as a primary mode of creation has in effect been relegated mainly to the field of ethnomusicology, historical musicology accepting as a subject in general only such improvisation as ornaments of written compositions. Historical musicology suffers a loss from such a division. Having my colleague Bonnie Blackburn serve as copy-editor of this book has been a stroke of great good fortune for me. Beyond attending to normal editorial tasks with consummate skill she has rescued me from numerous errors and inconsistencies, provided valuable information bearing on my subject, and followed my reasoning closely, calling my attention where necessary to its unclear or dubious turns. The book is much the better for her collaboration, and I am most grateful for it. I do not, however, mean to imply by this that what has survived this process necessarily has her endorsement. L.T. Lake Hill, New York
xiv
Ac knowledgement s
Acknowledgements The original versions of the chapters comprising this book appeared in the following publications and are included here, as revised, with the permission of the publishers. Revisions have been made in the interest of clarity, to eliminate irrelevancies, and to correct what I now regard as errors. Chapter 1. The World of Music, 3 (1991). Chapter 2. El Códice Calixtino y la música de su tiempo: actas del simposio organizado por la Fundación Pedro Barrie de la Maza en La Coruña y Santiago de Compostela, 20–23 de septiembre de 1999, ed. José López-Calo and Carlos Villanueva (Fundación Pedro Barrie de Maza: La Coruña, 2001). Chapter 3. ‘Der Vatikanische Organumtraktat und das Organum von Notre Dame de Paris: Perspektiven der Entwicklung einer schriftlichen Musikkultur in Europa’, Basler Jahrbuch für historische Musikpraxis, 7 (1983), translated from the German by Leo Treitler. Chapter 4. Roundtable ‘Peripherie und Zentrum in der Musik des Hohen Mittelalters’, Deutsche Musikforschende Gesellschaft, Internationaler Kongress Berlin 1974, Kongressbericht (Kassel, 1980). The segment reprinted here was originally in English. Chapter 5. Harold S. Powers (ed.), Studies in Music History: Essays for Oliver Strunk (Princeton, 1968). Chapter 6. MQ 60 (1974). Chapter 7. JAMS 28 (1975). Chapter 8. Review of John Miles Foley, The Theory of Oral Composition: History and Methodology (Folkloristics; Bloomington: Indiana University Press, 1988) and id. (ed.), Oral-Formulaic Theory: A Folklore Casebook (Garland Folklore Casebooks, 5, and Garland Reference Library Series of the Humanities, 739; New York: Garland Publishing, 1990), in Music Library Association Notes, 50 (1993), 70–6. Chapter 9. Journal of the Royal Musical Association, 116 (1991). Chapter 10. Speculum: Journal of the Medieval Academy of America, 65 (1981). Chapter 11. Forum musicologicum, 3 (1982).
Ac knowledgement s
xv
Chapter 12. JAAC 51 (1993). Chapter 13. JAMS 35 (1982). Chapter 14. EMH 4 (1984). Chapter 15. Münchener Beiträge zur Mediävistik und Renaissance-Forschung, 26: Liturgische Tropen (Munich, 1985). Chapter 16. Studies in the History of Music: Music and Language (New York: Broude Brothers, 1982). Chapter 17. Under the title ‘Once More: Medieval Music and Language’, in Essays in Honor of David G. Hughes (Isham Library Papers, 4; Cambridge: Harvard University Press, 1995). The version in this book has a new beginning.
xvi
Ac knowledgement s
Contents List of Plates List of Figures List of Tables List of Musical Examples Notes on the Compact Disc Abbreviations 1. Medieval Improvisation
xvii xviii xix xx xxiii xxvii 1
2. Written Music and Oral Music: Improvisation in Medieval Performance
39
3. The Vatican Organum Treatise and the Organum of Notre Dame of Paris: Perspectives on the Development of a Literate Music Culture in Europe
68
4. ‘Peripheral’ and ‘Central’
84
5. On the Structure of the Alleluia Melisma: A Western Tendency in Western Chant(?)
103
6. Homer and Gregory: The Transmission of Epic Poetry and Plainchant
131
7. ‘Centonate’ Chant: Übles Flickwerk or e pluribus unus?
186
8. Lingering Questions about ‘Oral Literature’
202
9. The Politics of Reception: Tailoring the Present as Fulfilment of a Desired Past
211
10. Oral, Written, and Literate Process in the Music of the Middle Ages
230
11. Observations on the Transmission of Some Aquitanian Tropes
252
12. History and the Ontology of the Musical Work
298
13. The Early History of Music Writing in the West
317
14. Reading and Singing: On the Genesis of Occidental Music Writing
365
15. Speaking of Jesus
429
16. Medieval Music and Language
435
17. The Marriage of Poetry and Music in Medieval Song
457
Bibliography Index
483 497
List of Music al Example s
xvii
List of Plates Plates are between pp. 162 and 163 I Drawing accompanying the notational representation of the first ecclesiastical mode in the trope MS Paris lat. 1118 II Drawing accompanying the notational representation of the second ecclesiastical mode in the trope MS Paris lat. 1118 III Drawing accompanying the notational representation of the third ecclesiastical mode in the trope MS Paris lat. 1118 IV Drawing accompanying the notational representation of the fourth ecclesiastical mode in the trope MS Paris lat. 1118 V Sculptural representation of the first and second ecclesiastical modes on a capital from the abbey church of Cluny VI Sculptural representation of the third and fourth ecclesiastical modes on a capital from the abbey church of Cluny VII Frontispiece from Paris lat. 1141 VIII Treves, Stadtbibliothek, uncatalogued single leaf IX Frontispiece from St. Gallen 390–91 (Codex Hartker) X Frontispiece from Munich clm 17403 XI The Musical Hall of Fame, a painting distributed in reproduction with the magazine The Etude, December, 1911 XII Cartoon by Eugene Mihaesco XIII Drawing of a lion and a porcupine by Villard de Honnecourt, c.1235 (Paris fr. 19093) XIV Saint-Bénigne de Dijon, Christ in Majesty with the Animals XV Vézelay, Église de la Madeleine, Christ in Majesty
xviii
List of Music al Example s
List of Figures 2.1 Simha Arom’s analysis of the temporal organization of the Aka Pygmy performance ‘Bobangi’ 3.1 Alleluia Hic Martinus from the Vatican Organum Treatise 5.1 (a) Alleluia Domine deus salutis, in St. Gallen 339 (b) Alleluia Ostende nobis, in Laon 239 (c) Alleluia Eripe me, in St. Gallen 339 (d) Alleluia Beatus vir qui suffert, in St. Gallen 339 6.1 Sequence Diem festum de sancto Bartholomeo, in Kremsmünster 309 13.1 An example of cantus planus binatim in Berlin, former Deutsche Staatsbibliothek, 40592 13.2 Prosa notation in Oxford Bodley 775 13.3 Reinforced iconic script in Benevento VI. 34 13.4 Reinforced iconic script in Vercelli 162 13.5 Reinforced iconic script in Padua A 47 13.6 Reinforced iconic script in London Lansdowne 462 13.7 Symbolic script in Vercelli 161 13.8 Signs interpreted as accents in ‘transition’ to neumes in Paris, Bibl. Ste-Geneviève 557 14.1 Musical notation in the Musica enchiriadis 14.2 Notational cues in the tenth-century gospel book from St. Gallen, Geneva lat. 37a
50 79 125
142 322 335 345 347 348 350 351 362 376 395
List of Music al Example s
xix
List of Tables 6.1 11.1 13.1 14.1
The two categories of constraints determining chant types Trope verse in four sources Eight neume characters as written in fourteen neumatic scripts Provenance of question mark types
162 275 336 421
xx
List of Music al Example s
List of Musical Examples 1.1 Transcription of the offertory Factus est dominus from Codex Bodmer, with variants in Vatican Vat. lat. 5319 and San Pietro F 22 1.2 Hypothetical revision of a passage in Ex. 1.1 according to inferred rules 1.3 Hypothetical revision of another passage in Ex. 1.1 according to inferred rules 1.4 Figure 8 from Max Haas, Mündliche Überlieferung und altrömischer Choral 1.5 The gradual Sciant gentes in the Frankish Gregorian transmission; the Old Roman transmission of Codex Bodmer; VaticanVat. lat. 5319; and Vatican San Pietro F 22 1.6 Written sources of different provenance from the twelfth century: (a) final cadence of the organum alleluia Hic Martinus; (b) passage from a MS from the Norman colony of Sicily; (c) passage from an organum of the Parisian Magnus liber organi; (d )–( f ) passages from an Aquitanian source 1.7 Intonation formula and model antiphon for the first mode, from Paris lat. 1121 2.1 The communion antiphon In splendoribus from the Graduale triplex 3.1 Six examples illustrating the eighteenth rule from the Vatican Organum Treatise 3.2 Beginning of an introit antiphon from Vatican Vat. lat. 5319 3.3 Three passages from the organum compositions in the Vatican Organum Treatise 3.4 Five passages from the organum compositions in the Vatican Organum Treatise 3.5 Two passages from the organum compositions in the Vatican Organum Treatise 3.6 Four passages from the organum compositions in the Vatican Organum Treatise 3.7 Transcription of alleluia Hic Martinus from the Vatican Organum Treatise 4.1 Two songs by Bernart de Ventadorn, Ara’m conseillatz and Pos pregatz 4.2 Trope Dulciter agnicole to the introit Introduxit vos 4.3 Versus Catolicorum concio
14 19 20 23
27
35 37 61 73 75 75 75 76 77 80 93 96 98
List of Music al Example s 4.4 5.1 5.2 5.3 5.4 5.5 5.6
Notre Dame organum duplum Judea et Iherusalem Alleluia Dies sanctificatus Alleluia Surrexit Dominus The melisma of the alleluia Beatus vir qui suffert Alleluia and melisma of the alleluia Te martyrum Alleluia and melisma from the alleluia Beatus vir sanctus Martinus Rhythmic prosa from Misset and Aubry, Les Proses d’Adam de Saint Victor 5.7 Two versions of the melisma from the alleluia versus Eripe me 5. App. Seventeen alleluia fragments in four groups 6.1 Melisma on ‘dierum’ in the second verse of the offertory Deus enim firmavit 6.2 The melody of Ex. 6.1 with prosula texts from Bamberg lit. 5; Rome, Bibl. naz. lat. 1343, and Paris lat. 776 6.3 Mode 2 Gregorian tract Deus, Deus meus, comparing fourteen verse melodies and the final phrases of mode 2 tracts De necessitatibus, Qui habitat, and Confitemini 6.4 Initial phrases of the first verses of six Gregorian and five Old Roman mode 2 tracts with schematic representation of the obligatory features of such phrases 6.5 Second phrases of several verses of the mode 2 tract Deus, Deus meus in Gregorian and Old Roman transmission 6.6 The verse melody of the alleluia Dies sanctificatus in standard and special Gregorian and Old Roman transmissions 6.7 Conclusion of the alleluia verse Paratum cor meum in Gregorian transmission from Paris lat. 903, Benevento VI. 34, and London Harl. 4951 6.8 Cadential melismas of the respond and verse of the Gregorian gradual Timete dominum; cadential melismas of the gradual Timete dominum and the alleluia Dies sanctificatus in the Old Roman transmission 9.1 Introit antiphon Rorate caeli in the Old Roman and Gregorian traditions 10.1 Four versions of the trope verse Filius ecce patrem conpellans taliter inquit 10.2 Cadential melismas of three strophes of the versus Ex ade vitio 11.1 Trope Dulciter agnicole 11.2 Trope Hodie Stephanus martir 11.3 Trope Filius ecce patrem 11.4 Trope Adveniente Xristo
xxi 100 109 121 121 121 121 122 126 127 140 141
163
173 174 178
179
180 217 243 245 258 264 267 271
xxii 11.5 11.6 11.7 11.8 12.1 12.2
12.3 12.4 13.1 14.1 14.2 14.3 14.4 14.5 15.1 16.1 16.2 17.1 17.2 17.3 17.4 17.5 17.6
List of Music al Example s Trope In Ihordane Trope Filii karissimi/clarissimi domino Trope Iubilent omnes Trope Discipulis flammas Chopin, Mazurka Op. 7 No. 5 in C major Second (più mosso) section of Chopin, Waltz Op. 64 No. 2 in C sharp minor with transcriptions from recordings by Alfred Cortot (1934), Ignaz Paderewski, and Sergei Rachmaninoff Chopin, Waltz Op. posthumous 70 No. 1 in G flat major, autograph version with variants in the Fontana version Chopin, Nocturne Op. 62 No. 1, bars 53–5: sketch, autograph fair copies, first and later Trope introduction Deus pater filium to the Christmas introit Puer natus est nobis Gregorian gradual Exsultabunt sancti Genealogy settings Responsorial verse tone of mode 1 from Exeter Introit and communion psalmody from St. Gallen 381 Table of positurae Two trope elements from Paris lat. 909 and Apt 17, with the antiphon phrases they introduce, from Paris lat. 776 Antiphon Petrus autem servabatur Two versions of the trope verse Discipulis flammas and two versions of the trope verse Omnigenis linguis Abelard, Dolorum solatium, fourth strophe Versus Radix iesse castitatis lilium Alternative version of Radix iesse Versus Lilium floruit, from Paris lat. 3549 Troubadour song Lan quant li jorn, by Jaufre Rudel Walther von der Vogelweide, Palästinalied
277 282 285 294 303
307 310 313 353 386 396 399 400 411 430 444 446 462 464 469 471 473 479
Notes on the Compact Disc Two emphases throughout this book make the enclosed compact disc indispensable for illustration of the book’s interpretations and, more directly, as conveyer of its meanings. Both can seem so self-evident as to be redundant, but they need stating explicitly. First, medieval song was a singer’s art, an art of belting out through mastery of melody and vocal skills the thoughts, sentiments, and images of scriptural prose and sacred and secular poetry. Second, the composing and broadcasting of medieval song took place in singing; it was an oral tradition, whether or not any of it came to be represented by the ciphers of musical notation. In the absence of singing, the notated musical examples in the book would have presented themselves to readers as objects, held fast in their notational matrices. The emphases of this book, instead, do not only require recorded performances, they ask for singing that celebrates the oneness of verbal and musical expression as it celebrates voice. The performances on the disc answer to those demands, but I must put it the other way around: it is largely from such performances that I have learnt the need to make historical interpretations responsible to the moment of singing out, something that has been too much neglected in the scholarly study and— paradoxically—even in many recorded performances of medieval song (I think especially of the tradition of chant singing owing to the choirs of the Abbey of St-Pierre de Solesmes and their followers, which tends to the suppression of the vocal virtuosity, versatility, and sensuousness implicit in the written record made by notators of the Middle Ages).
content s of the compact disc 1 Eighth mode intonation formula and model antiphon from Paris lat. 1121 (Aquitanian, 11th c.). Eighth mode introit Introduxit vos (Frankish*) with trope verses from Paris lat. 1121 2 First mode introit antiphon Rorate caeli (Frankish) 3 First mode introit antiphon Rorate caeli (Roman; Vatican Vat. lat. 5319, 11th c.) 4 First mode gradual Sciant gentes (Frankish) 5 First mode gradual Sciant gentes (Roman; Codex Bodmer, 11th c.) 6 Second mode alleluia v. Dies sanctificatus (Frankish) 7 Second mode alleluia v. Dies sanctificatus (Roman; Codex Bodmer) 8 Organum: first mode alleluia v. Hic Martinus (Vatican Ottob. lat. 3025, the Vatican Organum Treatise) 9 Second mode tract Deus deus meus (Frankish)
xxiv
Note s on The Compact Disc
10 11 12 13 14 15
Lightnin’ Hopkins, Blues, Goin’ Away Fifth mode offertory Factus est Dominus (Roman; Codex Bodmer) Versus Lilium floruit, Paris 3549 (12th c.) Versus Radix iesse, Paris lat. 1139 (c.1098) Versus Radix iesse, Le Puy A/V/7/009 (c.1588) Jaufre Rudel, troubadour, Lanquand li jorn, Paris fr. 20050 (13th c.). For sources see Chapter 17 n. 17 16 Walther von der Vogelweide, Minnesänger, Palästinalied (13th c.). For sources see Chapter 17 n. 19
* Chant items identified as ‘Frankish’ are based on the neumes and transcriptions of the Graduale triplex, with the exception of Track 9, which is based on Paris lat. 776 (Albi, 11th c.). Attentive listeners will notice differences of pitch content between this performance and the transcription in Chapter 6, Ex. 6.1, which is based on the Vatican Edition. I leave these differences as a tacit commentary on the concept of the ‘fixity’ of the repertory. See the introduction to Chapter 6 on this subject.
Tracks 1–7 and 11 are performed by members of the ensemble Dialogus: Catherine Sergent, Caroline Magalhaes, and Katarina Livljanic (Director). They are joined in track 8 by Lucia Nigohossian. Track 9 is performed by Dominique Vellard, Director of the ensemble Gilles Binchois; track 10 is performed by Lightnin’ Hopkins; tracks 12–16 are performed by members of the ensemble Sequentia: track 12, Benjamin Bagby, Co-Director, and Eric Mentzel; track 13, Benjamin Bagby; track 14, Barbara Thornton, Co-Director and the women’s choir of Sequentia; track 15, Barbara Thornton; track 16, Benjamin Bagby. Track 9 is taken from Chant Grégorien (STIL 2106 S84), with permission of STIL records. Track 10 is taken from Goin’ Away (Prestige/Bluesville OBCCD522-2 [BV-1073]), courtesy of Fantasy, Inc. Track 13 is taken from Shining Light: Music from Aquitania (Deutsche Harmonia Mundi 05472 77370 2), courtesy of the RCA Victor Groups. Track 15 has been digitally remastered from the cassette tape issued with The Union of Words and Music in Medieval Poetry, edited by Rebecca A. Baltzer, Thomas Cable, and James I. Wimsatt (Austin: University of Texas Press, 1991), used with the permission of the University of Texas Press. All four items are used with the permission of their original publishers. The remaining tracks are original to this Compact Disc. Benjamin Bagby supervised the recording of the new items in Paris. Robert Berkovitz remastered tracks 14 and 15 and provided altogether indispensable guidance in the preparation of the recordings. The design for the Compact Disc was made by Leann Davis Alspaugh. Katerina Livljanich has provided the following comment on her ensemble’s performances of chant on this recording.
No te s on the C om pac t Disc
xxv
a note about chant performance¹ All stories are haunted by the ghosts of the stories they might have been (Salman Rushdie, Shame) Like stories, medieval chant manuscripts seem to be haunted by the ghosts of how they might have been written. These performances of Frankish and Roman chant are just one such story: what we can hear on this recording is just one possible image of their sound, for they certainly did sound in different ways in the mouths of medieval cantors throughout Europe. When we perform chant of the Frankish tradition, we are confronted with a multitude of taciturn manuscripts from different places and centuries that transmit very similar or identical images of chant melodies. To which does one give the privilege of being considered ‘good sources for performance’? And what do we mean by ‘good source’? A diastematically precise one? Richer in rhythmical indications? Or just more legible to us than the other manuscripts? Would the same ‘good source’ be considered good by a contemporary singer and a medieval cantor for whom oral tradition was a true performance source and manuscripts were not ‘scores’ in the sense that we know that word? Would the criteria be the same? Through decades of chant research scholars have considered some types of notation as more precise than others with respect to rhythm, giving particular privilege to the earliest neumes of the St. Gallen and Metz families. Yet this knowledge does not encourage us to translate them into a set of tables and recipes for a precise performance of each neume. The same text can reveal different truths to different readers. The discipline known as Gregorian semiology brought a new perspective to the understanding of the earliest neumes, but as singers we are aware that each performance style is just one possible reading, an interpretation of an interpretation. In this particular performance we try to take into account all the precision of the rich indications for rhythm and neume grouping provided by St. Gallen manuscripts for the Frankish pieces. Besides careful reference to rhythmic nuances in the neumatic script, there are many different levels we may consider when we want to incarnate these signs in sound. There is the rhetorical function of each piece, its modal identity, ornamental richness, a profile crystallized during centuries of oral transmission. All these elements influence our performance decisions. Yet we shall never be able to know precisely what constituted long and short durations or fast and slow movement for St. Gallen cantors and scribes, how these values related to one another, and how flexible they were in their symbiosis with ¹ The CD accompanying this book contains recordings of several medieval chants and songs, interpreted by various performers. This note is not a musicological commentary about all of them, but an individual performer’s point of view concerning the following chants: introit Rorate caeli, alleluia Dies sanctificatus, gradual Sciant gentes, all in Frankish and Roman traditions, offertory Factus est dominus in the Roman tradition, the troped introit Introduxit vos, and an organum from the Vatican Organum Treatise.
xxvi
N o te s on the C om pac t Disc
the words of a chant. Yet medieval chant did not survive only through the mirror of St. Gallen neumes, and if we want to perform chant repertory from other sources we should not be trapped by a St. Gallen myopia or apply parameters from one notation to another. It seems that the ultimate help and guidance in the performance of neumes comes from the words that we are singing, the sense of the story or the sentiment we are conveying. Only in connection with the words do neumes reveal their inner logic. This matter becomes much more complex when we look at Roman chant. Should we then say ‘As we do not know how to interpret the rhythm of this type of notation it is useless to sing this repertory’? Some layers in chants of the Roman repertory are more or less comparable to the melodies of the Frankish tradition. If we take the example of the Christmas alleluia Dies sanctificatus we shall gain a very significant amount of information from similarities in the ornamental character and phrasing of that melody between Roman and St. Gallen manuscripts. If, on the other hand, we wish to apply an analysis of the same kind to the offertory Factus est dominus we shall be confronted by less similar versions. For the highly repetitive Roman melody there will be different problems in the performance, presented by its inner construction and by the role of musical formulas in the declamation of the text that cannot be reduced to an isolated analysis of neumes. Then what do we really want to do when we make this music sound? Do we want it to sound as it did in a medieval liturgy? (But then, in which century, at which place? in which acoustics?) Or do we simply want it to sound ‘well’? ‘Well’ for which audience? Are we not conditioned by our musicological visions and convictions, by our musical taste, by a tendency even to tailor messages from the manuscripts sometimes in order to make them suit our own theories? From the same text performers will always read slightly different musical realities and each of these realities will be haunted by the ghosts of songs they could have been. (The way of singing three solo verses in Factus est dominus shows this plurality, and we purposely stress the different vocal and personal identities of the three singers who render them.) The attempt to reduce the song to neumes always reminds me of forbidden literature by dissident writers. When their books finally became available and openly published, one realized that many writers had been screaming for decades about the impossibility of expressing themselves in their language, but when the time came to speak, in the hysterical joy over their rediscovered tongue they forgot what they actually intended to say. In our joy over discovering neumes we might be in danger of reducing music to its script, and forgetting that we can sing.
A b b re v i at i on s
xxvii
Abbreviations man uscrip t sigla Apt Arras Autun Bamberg Benevento Berlin Bologna Brussels Cambridge Chartres Codex Bodmer Codex Las Huelgas Cologne Cortona Douai Einsiedeln F Geneva Heidelberg Karlsruhe Kassel Kremsmünster Laon Le Puy London Lucca Madrid Milan Modena Montecassino Montpellier Munich
Archives de la Basilique Sainte-Anne Bibliothèque municipale Bibliothèque municipale Staatsbibliothek Biblioteca capitolare Staatsbibliothek zu Berlin — Preußischer Kulturbesitz Biblioteca universitaria Bibliothèque royale Corpus Christi College Bibliothèque municipale Private collection Martin Bodmer (Cologny, Switzerland), Codex 74 Burgos, Monasterio de Las Huelgas, Codex Las Huelgas Dombibliothek Biblioteca comunale Bibliothèque municipale Stiftsbibliothek Florence, Biblioteca Mediceo-Laurenziana, Pluteo 29, I Bibliothèque universitaire et publique Universitätsbibliothek Landesbibliothek Murhardsche Bibliothek Stiftsbibliothek Bibliothèque municipale Le Puy en Velay, Bibliothèque municipale British Library Biblioteca capitolare Biblioteca nacional Biblioteca Ambrosiana Biblioteca capitolare Monumento nazionale di Montecassino, Biblioteca Faculté de Médecine Bayerische Staatsbibliothek
xxviii Münster Naples Oxford Padua Paris Paris, Bibl. Ste-Gen. Pistoia Rome, Bibl. Ang. Rome, Bibl. Cas. Rome, Bibl. naz. Rome, Bibl. Vall. St. Gallen St Petersburg Treves Trier, Dombibl. Trier, Stadtbibl. Troyes Turin Valenciennes Vatican Vatican Organum Treatise Vercelli Vic Vienna Wolfenbüttel W1 W2 Worcester Zwettl
A b b re v i at i on s Staatsarchiv Biblioteca nazionale Vittorio Emanuele III Bodleian Library Biblioteca capitolare Bibliothèque nationale de France Bibliothèque Ste-Genevieve Biblioteca capitolare Biblioteca Angelica Biblioteca Casanatense Biblioteca nazionale Vittorio Emanuele II Biblioteca Vallicelliana Stiftsbibliothek Rossiyskaya Natsional'naya Biblioteka Stadtbibliothek Dombibliothek Stadtbibliothek Bibliothèque de la ville Biblioteca nazionale universitaria Bibliothèque municipale Biblioteca Apostolica Vaticana Biblioteca Apostolica Vaticana, Ottob. lat. 3025 Biblioteca capitolare Biblioteca museo episcopal Österreichische Nationalbibliothek Herzog August Bibliothek Wolfenbüttel, Herzog August Bibliothek Helmstedt 628 Wolfenbüttel, Herzog August Bibliothek Helmstedt 1099 Cathedral Library Stiftsbibliothek
bibliographical AfMw AM BAV BNF Cantus Planus
CLA
Archiv für Musikwissenschaft Acta musicologica Biblioteca Apostolica Vaticana Bibliothèque Nationale de France International Musicological Society Study Group Cantus Planus: Papers Read at the Third Meeting, Tihany, Hungary, 19–24 September 1988 (Budapest, 1990) E. A. Lowe, Codices Latini antiquiores, 11 vols. plus Supplement (Oxford, 1934–71)
A b b re v i at i on s CT
xxix
Corpus Troporum, ed. Ritva Jonsson, Gunilla Björkvall, and Gunilla Iversen (Acta universitatis Stockholmiensis, Studia Latina Stockholmiensia; Stockholm and Uppsala, 1975– ) EG Études grégoriennes EMH Early Music History IMS Berkeley International Musicological Society: Report of the Twelfth Congress, ed. Daniel Heartz and Bonnie Wade (Kassel, 1981) Hucbald, Guido, and John Hucbald, Guido, and John on Music, trans. Warren Babb, ed. Claude V. Palisca (New Haven, 1978) Hughes FS Graeme M. Boone (ed.), Essays on Medieval Music in Honor of David G. Hughes (Isham Library Papers, 4; Cambridge, Mass., 1995) JAAC Journal of Aesthetics and Art Criticism JAMS Journal of the American Musicological Society JM Journal of Musicology KmJb Kirchenmusikalisches Jahrbuch Mf Die Musikforschung MGG Die Musik in Geschichte und Gegenwart, ed. Friedrich Blume, 16 vols. (Kassel, 1949–79) Die Musik in Geschichte und Gegenwart, 2nd edn., ed. MGG2 Ludwig Finscher (Kassel, 1994– ) MGH Monumenta Germaniae historia MMMA Monumenta monodica medii aevi MQ Musical Quarterly NG The New Grove Dictionary of Music and Musicians, ed. Stanley Sadie, 20 vols. (London, 1980) NG II The New Grove Dictionary of Music and Musicians, 2nd edn., ed. Stanley Sadie, 29 vols. (London, 2001) NLH New Literary History Palaeographie der Musik Palaeographie der Musik, i: Die einstimmige Musik des Mittelalters, ed. Wulf Arlt (Cologne, 1973–9) [in 5 pts]; pt 2: Max Haas, Byzantinische und slavische Notationen (1973); pt 3: Solange Corbin, Die Neumen (1977); pt 4: Ewald Jammers: Aufzeichnungsweisen der einstimmigen ausserliturgischen Musik des Mittelalters (1975)
xxx PM Strunk’s Source Readings
WM
A b b re v i at i on s Paléographie musicale (Solesmes, 1889– ) Oliver Strunk (ed.), Strunk’s Source Readings in Music History, rev. edn., Leo Treitler, gen. ed. (New York, 1998) The World of Music, 3 (1991)
Medie val Impr ovisat i on
chapter 1
Medieval Improvisation
} Music historians are favoured with a rich tradition of descriptive language for the practices that we investigate. But for that favour we are obliged to pay the price of circumspection about the historical appropriateness with which we use that language in particular cases. A familiar example is the often-discussed case of the interpretation of the symphonies, sonatas, and chamber works of Haydn, Mozart, and Beethoven and their contemporaries in terms of the concept of ‘sonata form’ that was fully articulated only in composition manuals of the second quarter of the nineteenth century. What we must evaluate is the degree to which this helps to elucidate those works, as against the possibility that in interpreting them under a conception that was given expression after the fact, we somewhat distort their composers’ conceptions of them. Without entering into the complexities of the issue over intentionality that this case can raise, we can probably agree that the matter is pretty well settled with the acknowledgement that those nineteenth-century formulations are themselves based on the works of those composers as models. There is little serious rejection of them as anachronistic and distorting. But this is no more than a particular case of the task constantly faced by historians of every subject: elucidating some part of the past with the language and concepts of the historian’s present, while seeking at the same time to capture the language and concepts of that past and breathe life into them so as to grasp them as the tools of that past’s own present. The task is complicated even further by the history that the language and concepts of the historian’s present bear, history that is often not in the historian’s consciousness, and which may be alien to the historian’s, as well as to the historical object’s, present. The history of language can in such stealthy reversal capture and imprison thought. Must we resign ourselves to the distorting effects of that, yielding, with Nietzsche, that ‘We cease to think when we refuse to do so under the constraint of language?’¹ Or can we oppose such effects through work in the archaeology of language? Improvisation—word and concept woven as leading thread into the texture of ¹ The Will to Power, trans. Walter Kaufmann and R. J. Hollingdale, ed. Walter Kaufmann (New York, 1964), 283.
1
2
Medieval Impro visat i on discourse about medieval music-making—presents this dilemma in shrill voice. The first two chapters in this book are concerned to demonstrate that, and to exemplify the benefits towards an unfettered historical view from such archaeological work. The subject of this chapter—‘improvisation’, considered as word, concept, and practice—presents a particularly vexing case. The word (along with its verb, adjective, and adverb forms) is modern in all European languages (details are given in Ch. 2). That alone is not reason enough to avoid its use in describing medieval practices. But in its etymological roots and in its wide usage as of the nineteenth and twentieth centuries it can have negative and even pejorative connotations when it is held up against the value-norm of through-composed, well-structured, unified music. These are described at the beginning of the following chapter. Bruno Nettl has reflected on the cultural meaning of those connotations and I summarize his reflections further on in this introduction. Because such connotations come to the fore in the context of scholarship on chant—more prominently by far than in the context of the study of any other Western art music—until the publication of the original version of this chapter in 1991 I avoided the word ‘improvisation’ and its variants in writing about oral traditions of chant and related music in order not to give the impression that I would place those under the rubric of traditions carried on by improvident musicians. Nonetheless, as it happened, even my talk of oral traditions was translated by some readers as language about improvisation, with all the negative connotations that have been impressed on that word during its modern history. As Nettl wrote, as editor, in the preface to the collection of essays in which this one first appeared, ‘Improvisation was linked with oral tradition [in the Western tradition of music scholarship], and it was sometimes assumed that all music not notated must be in some way improvised.’² So when Nettl invited me to contribute to that collection I welcomed the opportunity to write about the subject in the company of scholars for whom ‘improvisation’ is not negatively valued. Their subjects are improvisation by Arab musicians, by North American jazz musicians, and by children singing while at play in an Australian aboriginal culture.³ Their perspectives are alike in four fundamental respects: first, they understand ‘improvisation’ neutrally, descriptively, as ‘creativity in the context of performance’, as Racy puts it (p. 8);⁴ second and third, they do not conceive a boundary between improvisation and composition so sharp as to make them categorical opposites, and so they do not construe the relationship to be a hierarchy in which durable works produced through composition are more highvalued than the ephemera that emerge in improvisational activity; fourth, they refer ² ‘New Perspectives on Improvisation’, WM 3 (1991), 3–6. ³ Ali Jihad Racy, ‘Creativity and Ambience: An Ecstatic Feedback Model from Arab Music’, ibid. 7–28; Gregory E. Smith, ‘In Quest of a New Perspective on Improvised Jazz: A View from the Balkans’, 29–52; Margaret J. Kartomi, ‘Musical Improvisations by Children at Play’, 53–66. ⁴ Nettl, in his article ‘Improvisation’ in The New Harvard Dictionary of Music, ed. Don Randel (Cambridge, Mass., 1986), 392, puts it: ‘The creation of music in the course of performance.’ The last phrase served him for the title of his subsequent anthology, In the Course of Performance: Studies in the World of Musical Improvisation, ed. Nettl with Melinda Russell (Chicago, 1998).
Medie val Impr ovisat i on to entirely different sets of values for the musical creations of the cultures about which they write from those on which we judge the masterworks of Western classical music. Kartomi writes, for example: ‘We need to recognize that these expressions are artistic. Child art, which has come to be highly appreciated by adults for its unique styles and outlook, normally displays a high measure of spontaneous creativity as opposed to the laborious workings behind most adult art works’ (p. 56). Racy concludes his account this way: ‘The ecstatic feedback model of creativity . . . places prime emphasis on the process of music making. Music is seen as a participatory phenomenon that involves direct emotional exchange between performers and listeners. Tarab artistry is directly and organically intertwined with the ecstatic and interactive dynamics of the performance event’ (p. 22). Above all, for none of these writers—any more than for ethnomusicologists and jazz scholars in general—is ‘improvisation’ a value term one way or the other. And so I took this opportunity to introduce the term in the context of medieval studies, cleared of the dark tones with which it has been stained in the literature of chant studies especially, in order to suggest the varying ways and various contexts in which music-making in the European Middle Ages was in general like music-making in the cultures that these authors describe. I believe that is what Nettl expected of me, for it was his project in that publication, as in the anthology of 1998, to redress the neglect of ‘the various orders of creativity’ that lead to finished works in favor of the focus on works themselves that has characterized ‘the history of musicology’, as Nettl put it in the introduction to his anthology (p. 1). Putting it positively, he has been working to establish a place in musical studies for learning about improvisation, a place that is commensurate with the prominence and the range of its practices over a very extensive historical and geographical terrain, and that requires the neutralization of the concept that is attempted in the 1998 anthology. But even there we find signs of how difficult a task that is and how much attention it requires. In Stephen Blum’s illuminating contribution to the book, ‘Recognizing Improvisation’, there is a citation of Aristotle’s famous speculations on the origins of drama in the Poetics. In the translation of Stephen Halliwell⁵ we read of ‘those who . . . brought poetry into being by their improvisations [autoschediasmaton]’, and of ‘tragedy, having come into being from an improvisational origin [arches autoschediastike].’ Blum provides a translation from the Arabic of Ibn Sina’s commentary on this passage, in which the word ‘extemporize’ in its several forms appears in place of ‘improvise’.⁶ That Dahiyat’s rendering is more proximate to the sense of the original is suggested by references in the Iliad and the Odyssey to fighting ‘at close quarters’ (autoschedios; Blum elucidates ‘i.e. with minimal opportunity to plan one’s moves before making them’). This concords with dictionary definitions of schediasma as whim or caprice and of schediazo as acting off-hand, giving free play, or pejoratively acting with insufficient care. Translating these concepts with the word ‘improvisation’ presupposes its prior negative loading, which would pre-empt the kind of neutrality for which Nettl’s authors aim. ⁵ The Poetics of Aristotle:Translation and Commentary (Chapel Hill, NC, 1987) ⁶ Avicenna’s Commentary on the Poetics of Aristotle, trans. Ismael M. Dahiyat (Leiden, 1974).
3
4
Medieval Impro visat i on It is worth trying to understand how it happens that suggestions of improvisation in the background of medieval chant tradition, and even implications of improvisation read into talk about the oral tradition of chant practice, have been so much understood in the sense of this negative loading. To do so we must first take account of the paradoxical fact that what is negative in the connotations of the word was actually celebrated when the word came into general use in European languages in the nineteenth century. The signs are in words incorporated in the titles of many pieces that were composed at the time: caprice or capriccio, fantasy, impromptu, moment musicale, rhapsody, all of them suggesting a musical expression that is a whim of the moment, an impulse, unplanned as to its occurrence and its moment-to-moment progression, guided more by passion and intuition than by the rational faculty, emphasizing that rather than elegant design. Pieces with such words in their titles were not necessarily in fact such unplanned productions (one thinks of the wealth of analytical literature about Chopin’s Polonaise Fantasie), but they were meant to give that impression, to have that aura, exactly in prominent and perhaps titillating contrast with works—often by the same composers—that were carefully worked out. It is precisely the implication of that aura of the impulsive and the unplanned that has been taken as an offence to the stature, dignity, and aesthetic grandiloquence of chant when improvisation is mentioned or implied in reference to it. (As I write this we in the USA are confronted with a dramatic demonstration of how unsettling and disturbing the experience of an improvisatory process can be. I mean the course of resolution of the presidential election of 2000, which was unpredictable at every turn.) I think of three reasons why this should be so: first, Gregorian chant has been the music of the Western Christian Church for over a millennium; second, it is regarded as the foundational music of Western culture, the counterpart for music of the visual and literary arts of Greek Antiquity; third, there is consequently strong motivation to see embodied in it what have been regarded since the nineteenth century as the essential aesthetic principles and values of Western classical music. This subject is discussed at length in the introduction to Chapter 5 and in Chapter 9, the second part of whose title hints at this function. Clearly the connotations of ‘improvisation’ fly in the face of this value system. Nettl attributes the neglect of improvisation in musical studies substantially to this conflict, putting it rather sharply: The musical establishment to which the profession of musicology belongs . . . connects improvisation as a musical practice, but even more as a concept, with a kind of third world of music. Jazz, the music of non-Western cultures, folk music, and all music in oral tradition are somehow included here . . . In the conception of the art music world, improvisation embodies the absence of precise planning and discipline. The most common contrast in the American public conception . . . is between composed art music and improvised jazz, in which art music is correlated with discipline, art for art’s sake, reliability, and predictability, while the correlation between classical musicians as middle
Medie val Impr ovisat i on class and conventionally moral, as against the old stereotype of jazz musicians as unreliable, with unconventional dress and sexual mores, excessive use of alcohol and drugs, and more. Improvisation as the music of people who don’t plan ahead . . .⁷ However closely this catches ‘the American public conception’, a stereotyped contrast between the ‘discipline, reliability, and predictability’ of Western classical art music and the ‘absence of precise planning and discipline’ of other, especially improvised, musics, is certainly in evidence in the modern literature on medieval chant. It is what has made it so difficult to conceive of chant as an improvisatory art. In the face of that stereotype I invite a comparison between the Gregorian tract Deus, Deus meus (CD track 9), which is discussed in Chapter 6, and the modern American blues song Goin’ Away (CD track 10), composed and performed by Lightnin’ Hopkins. The two performances have parallel expressive vocations, each in its own social setting: the intonation of a lengthy personal lament, each piling on its woeful plaints in the manner of a litany, that is, in a series. The expressive impact is thus delivered in large measure by the almost numbing repetitiveness of the singing. The words are unfolded in a series of verses, each intoned to a model fixed by the same musical pattern, a pattern that is thus heard over and over. In both cases it is a matter of four phrases in which each verse is declaimed. What fixes the pattern in the tract—which is of course melody alone—are the goal tones of the four phrases: d, c, f,d, reinforced by formulaic approaches to those tones (a detailed analysis is given in Ch. 6). In the blues the pattern is also fixed by a sequence of focal points for the four phrases in which each verse is likewise declaimed, but there they are the tones of the tonal-harmonic hierarchy, given out by the guitar: tonic, subdominant, tonic– dominant–tonic (this last progression often made via the subdominant). In both cases the melodic phrases and sub-phrases are adapted to the sense-units of the words, which they, in effect, punctuate (this principle is articulated in the present chapter). We would find the same parallels between any second-mode Gregorian tract and any blues. Floating the words on this sort of musical patterning is a procedure well adapted to improvisation, and its use in these traditions of cultures so distant from one another can be understood against the background of the fact that both indeed began as improvising traditions. The similarities illustrate the principle that biologists call convergent evolution, producing homologues. What do we gain from the comparison? I shall try a paradoxical formulation. On the one hand we exploit the familiar experience of the blues performance as a kind of hermeneutic bridge to the distant chant—it helps us to think ‘this is like something we know’. On the other hand we approach the tract liberated from the insulating genre classification that has been a principal dimension of the organization of Western music history, and we are released from the habit of apprehending it, like all items of chant, as a piece of fixed, absolute music. That habit is, after all, a premiss of the interpretation of chants as structurally unified works. In that sense ⁷ Nettl, In the Course of Performance, 6–7.
5
6
Medieval Impro visat i on we de-familiarize the tract. The comparison enables us to transcend the category distinction of art music and popular music, which is especially dubious for the Middle Ages. And in identifying a fundamental musical universal, it transcends the cultural opposition that has been another premiss of the analysis of plainchant (discussed in the introduction to Ch. 5 and in Ch. 9). The last phrase of the passage from Nettl’s introduction puts in a nutshell a central point about the improvisation concept that runs afoul of one of the most sanctified axioms about the transmission of chant both before and after the precipitation of the repertory into notated books. Whether it is in the Romantic conception that valorized all those pseudo-extempore compositions, or in the pejorative reading of ‘improvisation’ that we find in the literature on chant, or in the criteria for successful improvisation in music cultures other than the Western art-music culture, there is intimation—true or not—that improvised music is characterized by variation; the rendering of an improvised item is different from one time to the next: intentionally, if this is regarded as a value, inevitably when it is the opposite, usually called ‘stability’, that is valued. It is the latter conception that prevails in the literature on chant history, and that is why the possibility of improvisation in the establishment of the chant tradition has been thought to be ruled out by the homogeneity of the written transmission in the earliest notated books. The problematics of the place of the stability concept in the discourse about chant history are discussed at length in the introduction to Chapter 6. In this chapter the analysis of the Roman offertory Factus est dominus is intended as a counter-example in the domain of chant to the axiom of the necessary variability of improvised music. The analysis is given in terms of rules or constraints according to which the succession of discrete melodic elements is determined by the semantic, syntactic, and phonetic parameters of the words to be vocalized. The rules are of such a density and precision as to control nearly every note. In a comparison of the three existing sources for the chant a small number of variants emerge, sufficient, perhaps, to suggest that the near replication is not simply a result of copying (in each case the variant can be understood as the product of an alternative rule), and showing the small degree of play in the determination of the chant. I suggest that a singer who had memorized the melodic elements and the rules was prepared to vocalize the words, and that the outcome could be virtually identical from one performance to another. The conclusion is that variability is not inevitable, but depends on the density of the constraints controlling the performance. Another cross-cultural comparison, drawn in Chapter 2, reinforces this point. There it concerns the highly rule-bound offertory discussed in Chapter 1 and an example of the entrancing polyphonic singing of Central African pygmies. This is a highly complex oral performance practice that is made possible by a rigorous organization in accordance with extremely strict rules which, here, too, bring about performances that are nearly identical from one time to the next. While the original version of this chapter was the occasion for the abandonment of my own long-guarded scruple against any appeal to the label ‘improvisation’ for characterizing medieval performance—in particular of chant—it was by no means
Medie val Impr ovisat i on the first occasion on which that tradition has been so characterized. On the contrary, as Helmut Hucke wrote in the opening sentence of his essay ‘Improvisation im Gregorianischen Gesang’, ‘That there was improvising in the practice of Gregorian chant is a supposition that is often expressed.’⁸ For an example he cited a passage from an article by Walter Lipphardt: Germinal melodies of the time around 600 lay at the root of Roman cantorial practice, but these were, in characteristic southern fashion, forever varied, ornamented, reconfigured by the singers, for as long as the melodies were maintained on Italian soil. It was in the special nature of the Frankish melodic tradition that it gave up the improvisational art of the South and became fixed, but by virtue of that and at the same time it became ‘sacral’ and unified.⁹ At the root of this brief account, we might say, lies a germinal theme of historical chant scholarship, with these motifs: improvisation is a practice of southern peoples, particularly Italians; its essence is perpetual variation; when the Roman melodic tradition was taken over by the people of the North (the Franks) it became fixed—fixity is as much an expression of the northern character as is variability an expression of the southern character; with fixity goes unity: fixity is a necessary condition of unity, as it is a necessary condition of authenticity; and it does not lie very far below the surface that both unity and authenticity are to be preferred to their absence—their attainment constitutes an advance. What I have translated as ‘ornamented’ in the first sentence of the passage from Lipphardt is the German verziert, and the association of that word with ‘improvisation’ is a clue to the context of meaning that the latter carried along when it was introduced into the domain of medieval chant composition and performance. Verzierung referred to the practice by virtuoso performers in the eighteenth and nineteenth centuries—mainly singers—of ornamenting the composed and written solos of the works they performed in concert. The practice was highly valued by audiences, who were, however, impatient with performers whose improvisations they had heard before. It was from such a tradition that the improvisation concept, first applied to music in the eighteenth century, acquired its connotation of perpetual variation, which runs counter, of course, to the investment in the theme of the chant’s fixity. And it is one of the two contexts evoked by the phrase ‘improvisatory flights’, cited in the following chapter as an ironic characterization meant to cast doubt on the suggestion of improvisation in chant tradition. The other, clearly, is the eighteenth-century fashion for the ‘free fantasy’, as described and exemplified by C. P. E. Bach, with its qualities of flowing spontaneity, eccentricity, abruptness, and unpredictability.¹⁰ But these could not be further removed from the ideals and methods of chant practice, before or after the age of music writing. What is very clear in Lipphardt’s language is the a priori nature of the narrative. ⁸ KmJb 38 (1954), 5–8. ⁹ ‘Gregor der Große und sein Anteil am römischen Antiphonar’, Atti del congresso di musica sacra 1950 (Tournai, 1952), 248 ff. ¹⁰ See the article ‘Improvisation’ in MGG² iv, §V. 18. und 19. Jahrhundert.
7
8
Medieval Impro visat i on Each of its motifs expresses a value—now aesthetic, now cultural or even national, a prejudice in the literal sense—that has priority in the interpretation of evidence. And we might well wonder whether such a narrative could ever actually be the clear outcome of an open engagement with the historical material, were it not already pre-interpreted through the elements of the narrative itself. We must recognize that historical accounts, especially when they are cast in narrative form, inevitably have a fictive aspect, but we do want them to be true and realistic accounts of their part of the historical world in flux, as far as we can make them so. On close reading Lipphardt’s unidimensional and unidirectional account must strike us, not so much as false, but as incapable of being confirmed as true and lacking any semblance of an appearance of reality. This fundamental historiographic issue will come up repeatedly throughout the discussions in this book. Hucke continued, in the second sentence of the cited essay, ‘But until now our knowledge hardly goes beyond this supposition [i.e. that Gregorian chanting entailed improvisation]’. He motivated the undertaking to push our knowledge further with the simple but profound admonition that ‘one must take into account that, after all, this all happens in the absence of writing’.¹¹ And he directed it with the question whether the melodies transmitted in writing show ‘improvisatory markings’, whether we can regard them as ‘inscribed improvisations’, to be put to a population in the field that he had defined on the grounds of an intuitive sense that answers would be found there—Gradual verses of the fifth mode, in the Old Roman transmission (‘Improvisation im Gregorianischen Gesang’). There he thought he did indeed find the ‘marks of a regulated improvisational practice, making use of rigorous stylization principles, formal schemes, and fixed melodic elements’ (p. 5). What he did not explain, in the latter essay or anywhere else, is why these should be regarded as ‘improvisatory markings’, or what he really meant by that. At the conclusion of the earlier of the two essays he had drawn a contrast between the making of chants through the ‘free’ stringing together of fixed elements or lines— he called that ‘improvisatory’—and their composition following a ‘model’ or ‘configurational principle’. I can see the difference, but why one should be called ‘improvisatory’ and the other not, is not so clear. Much later, for example in ‘Toward a New Historical View of Gregorian Chant’,¹² Hucke seemed to be satisfied to show how attention to these elements and other factors allow us to understand, through re-enactment of the composition of individual chants, how such chants could have been composed orally, without any reference to ‘improvisation’, and without any effort at systematic classification of techniques. His late analyses show that such re-enactment is really the ultimate critical approach to the chants we have before us in written form. With this shift of focus the fixity–variability topos ceased to be such an issue for him. In ‘Improvisation im Gregorianischen Gesang’, however, he was still under the sway of the paradigm sketched by Lipphardt and its priorities. He concluded: ‘At the latest about the middle of the ¹¹ ‘Musikalische Formen der Officiumsantiphonen’, KmJb 37 (1953), 7–33. ¹² JAMS 33 (1980), 437–67.
Medie val Impr ovisat i on eighth century the melodies were set down in writing. While the Old Roman repertory was in part further developed, the Frankish transmission, arising through consistent revision of the Old Roman transmission, shows signs of a practice that remained note-for-note faithful to its tradition.’¹³ Hucke subsequently revised his views about the time of the writing down, the relation between the Old Roman and Frankish transmissions, the degree of uniformity in the early written Frankish transmission and what it shows (or does not show) about fixity in the oral tradition, and altogether about the possibility of fleshing out a narrative as unidimensional and unidirectional as that sketched by Lipphardt. We are undoubtedly influenced in our thoughts about ‘improvisation’ by our feelings about the sorts of life strategies—or absence of them—that are often so labelled. That must interfere with the effort to ground—as I am trying to do—a sense of improvisation as a practice or behaviour that can result in orderliness and stability in any domain. Mary Catherine Bateson begins her book Composing a Life with a description of this tension: This is a book about life as an improvisatory art, about the ways we combine familiar and unfamiliar components in response to new situations, following an underlying grammar and an evolving aesthetic. It started from a disgruntled reflection on my own life as a sort of desperate improvisation in which I was constantly trying to make something coherent from conflicting elements to fit rapidly changing settings. Improvisation can be either a last resort or an established way of evoking creativity.¹⁴
af terword I note that I have used the word ‘tradition’ nineteen times in this introduction. It is a term that deserves more careful reflection than we are accustomed to affording it in our literature. Two pieces of writing are especially recommended to the reader for their encouragement of such reflection: Brian Stock, ‘Tradition and Modernity: Modes from the Past’,¹⁵ and Ruth Finnegan, ‘Tradition, But what Tradition, and for Whom’?¹⁶
}
¹³ ‘Improvisation im Gregorianischen Gesang’, 8. ¹⁴ Composing a Life (New York, 1989), 1. My emphasis. ¹⁵ Listening for the Text: On the Uses of the Past (Baltimore, 1990), 159–71. ¹⁶ The Milman Parry Lecture on Oral Tradition for 1989–90, in the journal Oral Tradition, 6 (1991), 104–25.
9
10
Medieval Impro visat i on
I My title for this chapter smooths over what will appear with a little reflection to be a paradox, but what on further reflection can show a way into the subject. ‘Improvise’ conveys a negative, from the Latin improvisus (‘unforeseen’). Without the negative, Latin has left us ‘provide’, which, looking ahead, is what we should like to do for our children. And so dictionaries give us such definitions as ‘to perform without preparation’. They lead us to another negative that conveys the sense of an unreflective and sudden character about such performance: ‘im-promptu’, they tell us, is how we act on the spur of the moment. The prompter guides the actor; an impromptu performance goes forward without guidance. Impromptus by nineteenth-century composers are intended to suggest the result of ‘sudden inspiration’, making a romantic virtue of an adversity.¹⁷ The spur of a moment is its crest, from which we are impelled to move, as though by the force of gravity, impulsively, out of the normal pulse or rhythm of things; ex tempore, out of time. It is no wonder, then, that The Concise Oxford Dictionary of Music gives us ‘improvisation: a performance according to the inventive whim of the moment’¹⁸ or that the author of a recent essay on orality in Gregorian chant would summon such language as ‘improvisatory license’ and the ‘freedom’ and ‘vagaries’ of ‘improvisatory flights’.¹⁹ In this negative frisson about ‘improvisation’ there is a hint that it is something special, the exception to something normal, more grounded, something with all the attributes that improvisation lacks: preparation, guidance, planning ahead, proceeding apace deliberately. That something, of course, is ‘composition’. In 1854 Eduard Hanslick, who can still be read as a spokesman for values that tacitly determine our use of language in the description of music, made a central point of the aspect of deliberate planning entailed in that concept: ‘The composer works slowly and intermittently, . . . forming the musical artwork for posterity.’²⁰ Bruno Nettl, the author of the Harvard Dictionary article, concluded wisely that ‘It seems most appropriate to reserve the term improvisation for cultures and repertories in which a distinction from non-improvised and precomposed forms can be recognized’ (p. 392). The wisdom is in the implicit recognition that the very concept of ‘improvisation’ as we have seen it anchored in language is a product of cultures that have valorized its opposite—composition—as a norm, whether or not as a higher form; but also in the implicit recognition that no culture is likely to ¹⁷ New Harvard Dictionary of Music, 391. ¹⁸ Concise Oxford Dictionary of Music (Oxford, 1980), 315. ¹⁹ Kenneth Levy, ‘On Gregorian Orality’, JAMS 42 (1990), 185–227, repr. in his Gregorian Chant and the Carolingians (Princeton, 1998), ch. 6. ²⁰ On the Musically Beautiful: A Contribution toward the Revision of the Aesthetic of Music, trans. and ed. Geoffrey Payzant (Indianapolis, Ind., 1986), 49.
Medie val Impr ovisat i on
11
thrive alone on caprice in the making of music. A music culture is not likely to produce works entitled ‘Caprice Viennoise’ or ‘Fantasie-Impromptu’ or the like, except against the backdrop of works entitled ‘Hammerklavier Sonata’ or ‘Jupiter Symphony’ or the like. That points the way into the paradox. We cannot identify such a contrast in the musics of the medieval European culture. But that is not, as might be thought, because ‘composition’, both oral and written, is not well established in that culture; it is because we cannot identify improvised music, in the negative sense that is stamped on current usage. According to the implicit sense of its title, this chapter threatens to have no subject, and no proper place in this collection. What might seem especially paradoxical about this is that music in the medieval culture was to a very great extent produced and presented in unwritten traditions. As Nino Pirrotta put it, ‘The music from which we make history, the written tradition of music, may be likened to the visible tip of an iceberg, most of which is submerged and invisible.’²¹ He reckoned the submerged part to amount to seveneighths, and called it ‘the vast uncharted domain of unwritten music’.²² If this may be said of music throughout European history, then all the more so of music in the medieval culture, which first devised and began to use a technology of musical notation sometime after the turn of the ninth century, but even then for highly specialized purposes (the illustration of pedagogical handbooks and the recording of the repertories of the liturgical music of the Church). The production of music on a broad front, on principles not dependent on writing or dependent on it in only a collateral way—as in the composition of a counterpoint to a melody— continued otherwise unabated. The production of music as the actualization of both written and unwritten composition is a premiss for the understanding of medieval music cultures. What we cannot conclude, however, is that music that was not performed from score or from some aural counterpart of photographic memory was ‘improvised’, if we mean by that to emphasize the absence of plan. On the contrary, musicmaking in unwritten traditions—just because they are unwritten—depends on systems and methods of procedure that entail design in the sense of both praxis and poiesis. One had to have a procedural plan, even if one did not have a pre-vision about how it would turn out. With this recognition the appearance of paradox fades, and my subject becomes clear: generative contexts for music-making in the Middle Ages. This formulation circumvents the idea of an opposition between ‘composition’ and ‘improvisation’, considered as opposites. The interest is in systems that functioned in unwritten traditions, but on principles that could and did continue to guide composition when it became normal to write its products down. ²¹ ‘The Oral and Written Traditions of Music’, in Music and Culture in Italy from the Middle Ages to the Baroque: A Collection of Essays (Cambridge, Mass., 1984), 72–9. ²² ‘New Glimpses of an Unwritten Tradition’, ibid. 51–71 at 54.
12
Medieval Impro visat i on
Categorical differences between the compositional processes of written and unwritten traditions are no more characteristic of the greater part of the Middle Ages than are differences between music that has been pre-composed and music that has been improvised. Indeed, singer-notators must surely sometimes have improvised as they were writing down. We take advantage of this continuity, in fact, to learn about improvisatory traditions from what has been written down. Even in the late Middle Ages composing and performing could be thought of as a single act. A music master of the thirteenth century writing a handbook instructing singers in the invention of counterpoints to be sung along with traditional chant melodies states his purpose as teaching how ‘to compose and perform discant ex improviso’—the last two words of that can be rendered ‘extemporaneously’, ‘spontaneously’, or ‘in an impromptu manner’.²³ Whether or not this writer intends the implication that there can also be composition that is worked out in advance of performance, there is abundant musical evidence of that throughout the Middle Ages—for the composition of chant before the age of music writing, as we shall see in the introduction to Chapter 2, and for polyphony that has come down in writing on which its performance was probably dependent. I am thinking of music in two, three, and four parts, made around the turn of the thirteenth century for the liturgy of the cathedral of Notre Dame in Paris. The music is organized in regular metrical, harmonic, and voiceleading phases, and the coordination of the three in performance depends entirely on the durational patterns in which each of the voices sing. Contemporary awareness of that, and of a separation between composition and performance, is signalled implicitly by an increasing specificity in the indication of temporal duration by the notation of that music through the written transmission of the same works through successive sources.²⁴ But it is very likely that even in the same musical establishment, at the same time, singers were making polyphonic settings of liturgical chant extemporaneously by following the rules given in such manuals as the one just mentioned. Evidence has been brought recently of improvised music in the very same liturgy of Notre Dame as late as the sixteenth century.²⁵ Pirrotta has passed on to us a proud remark of Poliziano in a letter about his pupil Piero de’ Medici, that he could sing from musical notation or to the accompaniment of a ‘cithara’. This was in the fifteenth century, but, Pirrotta goes on, ‘if we go back to the fourteenth century, we find a similar dichotomy between the written tradition of the so-called Italian Ars nova and the widespread custom of sung poetry, which was not only cultivated without the help of any kind of notation, but seems to a certain extent to have intentionally avoided the strictures of notation’.²⁶ ²³ ‘Componere et proferre discantum ex improviso’; Anonymous II, in the section ‘Sequitur de discantu’, in Scriptores de musica medii aevi, ed. E. de Coussemaker, 4 vols. (Paris, 1864–76), i. 311. ²⁴ Details are given in my essay ‘Regarding Meter and Rhythm in the Ars Antiqua’, MQ 65 (1979), 524–58. ²⁵ e.g. Craig Wright, Music and Ceremony at Notre Dame of Paris 500–1550 (Cambridge, 1989), 159. ²⁶ ‘Ars Nova and Stil Novo’, in Music and Culture in Italy from the Middle Ages to the Baroque, 27.
Medie val Impr ovisat i on
13
It is the music of the Ars nova that gets our attention, inevitably, because we can still perform it and hear it and study it. But that gives us a view of history that is in a way distorted, especially for this period of the late Middle Ages when written music was rather special and exceptional. Pirrotta writes: ‘The composers of the Ars nova were all monks, priests, canons, or church organists . . . I have come to consider their activity as a private hobby, appreciated only by a few connoisseurs.’²⁷ The very term Ars nova referred more in its day to a technology of music writing than to a musical style. But that in its day there had come to be a clear and reflected distinction between music whose composition coincided with its performance, and music whose composition was separated from its performance by notation— and necessarily so—of that there can be no doubt.
II Example 1.1 is a transcription of an offertory chant for the mass on the fourth Sunday of Lent, Factus est dominus, made from Codex Bodmer 74, a service book with musical notation for the mass compiled for the cathedral of Santa Cecilia in Trastevere, Rome, in 1071.²⁸ This codex is the oldest surviving fully notated service book of Roman origin, and it seems very near to the earliest use of musical notation altogether in Rome.²⁹ But as the liturgical tradition represented by the codex has been identified in Rome as early as the eighth century,³⁰ the music for that tradition must have been an unwritten one for at least three centuries. Whether that musical tradition is what we now see recorded in Codex Bodmer and the several other books like it that were compiled during the eleventh to thirteenth centuries is something we cannot of course know with confidence.³¹ Putting it most cautiously, the textual status of the melody of Ex. 1.1 may be anything from that of a direct transcription from the unwritten tradition to that of a new composition immediately written down, and it may represent the original writing down itself or a copy at the remove of one or more generations going back ²⁷ ‘Ars Nova and Stil Novo’, in Music and Culture in Italy from the Middle Ages to the Baroque, 29. ²⁸ Das Graduale von Santa Cecilia in Trastevere (Cod. Bodmer 74), ed. Max Lütolf (Cologny and Geneva, 1987), fo. 62v. ²⁹ This has seemed to be the case, judging only from the absence of older sources. But recent analysis of related manuscript material has produced no evidence that the written transmission of the Roman tradition represented by the Bodmer Codex began earlier than the 11th c., approximately two centuries after the beginning of the Frankish written tradition. The evidence is presented by John Boe in two articles: ‘Chant Notation in Eleventh-Century Roman Manuscripts’, in Hughes FS, 43–57; and ‘Music Notation in Archivio San Pietro C 105 and in the Farfa Breviary, Chigi C.VI.177’, EMH 18 (1999), 1–45. ³⁰ See Michel Huglo, ‘Le Chant “vieux-romain”: liste des manuscrits et témoins indirects’, Sacris erudiri, 4 (1996), 96–124. ³¹ See Helmut Hucke’s survey of the sources for this tradition, known in the current scholarly literature as Old Roman, in ‘Gregorian and Old Roman Chant’, NG vii. 693–7.
14
Medieval Impro visat i on
Ex. 1.1. Transcription of the offertory Factus est dominus from Codex Bodmer, fo. 62v, with variants in Vatican Vat. lat. 5319 and San Pietro F 22. CD track 11 Codex Bodmer, fo. 62 v
Fac
tus
est
Off.
do
mi
nus
(i) 5319 fo. 65 v
* fir ma
men
K
T
tum
me
P
S
um
et sal
vum
P (ii) F 22 fo. 39 v
K
T
me
fe
P
M
S
S
S
S
cit
ab in
i
mi
cis
me
P (ii) F 22
is
K'
T
po
ten ti
P
M
bus
et
ab
his
P (ii) F 22
K'
T'
Q
runt
me.
5319 8
o
de
T' (iii) F 22 * The figure
signifies a liquescent neume
Q
V1
Fac
tus
qui
15
Medie val Impr ovisat i on Ex. 1.1. cont.
est
do
mi
nus
fir ma
men
(i) 5319
K
T
P
tum
me
um
K
et re
T
P
M
me
um
et li (Repetenda)
P
S
K
us
spe ra
bo in
be
ra
fu
gi
um
K
T
tor
me
T'
Q
X
e
umV2 Per
T'' 5319
se T'''
P
K
in
quar T
P
S
S
S
me
os
et
com
pre
P
S
los
et
S
don
ec
de
T''
hen dam
non con
fi
cos
K
K
K'
i mi
il
T''
P
ver
tar
T
P
ci
ant
M
et
li
16
Medieval Impro visat i on
Ex. 1.1. cont. X V3 Pre cin
xi
(v) 5319
sti
K
T''
P
ad
bel
lum
et
me
Vir tu
tem
M
S
S
S
sup
plan
tas
ti
ab
S
S
S
S
in
i
mi
cos
P
M
me
et
S
S
dis
ti
S
me
in
os
K
T'''
sub
tus
S
S
S
S
i
mi
co
rum
K
mi chi
S
S
S
S
S
K
en
tes
me
dis
per
di
S
S
S
S
me
o
rum
de
T'''
P
M
dor
sum
et T'
S
ho
di Q
sti (iv) 5319
Medie val Impr ovisat i on
17
as far as the early part of the century (but no earlier). There is no reason to suspect the extreme case of brand new compositions in connection with the writing down in such a traditional practice, and I shall shortly show why such a hypothesis is virtually ruled out. (This question is taken up again in Ch. 2 and its introduction.) On the other hand, it seems inevitable that the very process of writing down would have affected its product editorially. In any case we may be confident that the notations of Codex Bodmer are witnesses in some sense to the unwritten tradition that preceded it (and perhaps continued around it), and they are as close in time to such a tradition as it is possible to come with surviving sources that can be read. A medieval liturgical chant is first and foremost to be understood as the presentation in singing of an ecclesiastical verbal text. The primary task of melody in chant is the presentation of language in such a way as to project its sound structure and meaning with maximum distinctness, while at the same time being faithful to principles of melodic syntax and grammar that are required to preserve coherence of idiom and genre and to the constraints of performance practice in the liturgical situation (essentially a matter of who is to sing what, and in what order). The analysis of chant begins with the description of these tasks and obligations. The opening song of the offertory (marked ‘off’, as in the manuscript) was sung by the choir, and was followed by several verses—here three—for the cantor. Usually the end of each verse is cued in the sources to a final segment of the offertory, which was sung again by the choir as a refrain (called repetenda). This always ends on the final of the mode of the chant, so that in the many instances when the verse ends on some other tone, the singing of the repetenda brought about tonal closure. In Ex. 1.1, however, there is no such cue at the end of the first verse, while the end of the second verse is cued to the last line of the first, et liberator . . . This unusual circumstance certainly has some connection with another that emerges in a comparison of this chant with its liturgical counterpart in the Frankish Gregorian tradition. It is an offertory with two verses, with words that correspond as follows to the words of the Roman counterpart: Roman
Frankish
Offertory verse 1 verse 2 verse 3
Offertory verse 1 verse 2
The situation of the repetenda cues and the near identity of the words and melody of the Roman offertory with its first verse suggest that some ancestor of the
18
Medieval Impro visat i on
Roman chant comprised an offertory and two verses, like the Frankish counterpart,³² and that someone later, for whatever reason, fashioned a new offertory for it, based in words and melody on what was originally the offertory and became the first verse.³³ This is of great interest for my examination here precisely because it provides strong evidence against the possibility that the Roman chant was newly composed at the time of its writing down in the Bodmer Codex. We may confidently proceed to read it as a witness to its unwritten tradition, and in fact as fresh a witness as it is possible to find in the transmission of medieval chant. The following, then, will be an analysis aimed at identifying the principles on which such a chant could have been made in that tradition. The primary melodic units of Ex. 1.1 are the settings of the individual syllables of the text. Few syllables are declaimed by a single note, most by groups of two to nine notes. There are two occurrences of a very long melisma (marked X), and there is a rising two-note group (S) that is repeated in straight recitation (cf., especially, the third verse). The chant comprises the settings of 148 syllables. Given the narrow range of the melody (leaving aside the melismas X, e–b) the number of groups of different melodic contents is bound to be quite limited. I count about sixteen. But giving an exact count is difficult, because it depends on what is to count as ‘same’ and what as ‘different’. So in Ex. 1.1 I have labelled with the letter T three configurations that are different, when it comes down to the last detail. But they all turn around the note g—moving up to a and down to f—and more important they all have the same function within the melodic discourse: each one is the middle figure in a sequence of three figures—KTP or KTQ—that always and only articulate the sense-units in which the words are grouped: ‘The Lord is my firmament [KTP] and he delivered me [KTP] from my strongest enemies [KTP] and from them that hated me [KTQ].’ The clausula KTQ provides a melodic full close on the tonic f for the completion of sentences, which means completion of whole sections of the chant (offertory or verse); KTP makes a half-close: a clear melodic caesura, but ending on the unstable tone g, with implication of continuation and eventual close on f. The clausula KTP or KTQ (including their variants) is written out sixteen times. In twelve of those instances the clausula sings three syllables, with stress on the penultimate syllable. In the remaining four instances the stress is on the ante³² This would be most readily understandable on the hypothesis advanced by Hucke (in the article cited in the previous note), that the two traditions that have come down as ‘Gregorian’ and ‘Old Roman’ are really Frankish and Roman branches of the original Gregorian tradition. The new offertory for the Roman version would then have been fashioned after the separation of the two branches, something for which the watershed is the writing down of the Gregorian melodic tradition sometime around the turn of the 9th c. ³³ Joseph Dyer has offered a different interpretation of the relationship between the two versions of this offertory. See ‘Tropis semper variantibus: Compositional Strategies in the Offertories of Old Roman Chant’, EMH 17 (1998), 1–60; the discussion to which I refer is on pp. 25–9. This difference does not affect my analysis of the Roman version and the conclusions I draw from it.
19
Medie val Impr ovisat i on Ex. 1.2. Hypothetical revision of a passage in Ex. 1.1 according to inferred rules S
S
S
S
(V3) in
i
mi
cos
K'
me os
sub
T''
P
tus
me
penultimate syllable (off. poténtibus, off. hodérunt me, V2 defíciant, V3 súbtus me). In all but the last of these, the figure K carries the unaccented syllable before the accented one, as elsewhere, but it is turned upward to the tone a rather than down to f, and the accented syllable is given the extra descending figure g–f before the continuation with TP or TQ. On these principles the treatment of subtus me in the third verse is anomalous. Following the accommodation to the antepenultimate accent in the three other instances, the setting would be as shown in Ex. 1.2. I cannot, in fact, see why that would not have been the preferred setting. (The setting of potentibus might appear to be an exception also. But the g on ten is written as a liquescent neume, consistent with the liquid n, and would most likely have been sung with a portamento descending towards f.) This sort of critical engagement with the written material, in terms of the application of principles that it bespeaks, has the value that through it we test our understanding of the principles themselves. They are principles that experienced singers would have internalized and called upon, once faced with an ecclesiastical text and the definition of a liturgical moment and its appropriate performance practice. And they must have constituted a semiotic code in the projection of the chant to the liturgical community that would have clarified for them the articulation and sequence of sense-units in the—mostly biblical—language that was being unfolded for them.³⁴ The formulae would have been apprehended as markers, signalling the location of any moment in the chant, hinting at what is coming next, identifying moments in the language as beginnings, pauses, or full stops. My invasive correction of the musical text betrays my approach to it as an inscription, more of a code than of a closed and autonomous work. After the clausula KTP there is a continuation, and in almost every instance it is with the conjunction et.³⁵ The melodic figure mainly used for continuation after the clausula is M, and following that there is commonly recitation on the repeated ³⁴ This interpretation is different from the one that has been more conventionally offered in the literature regarding the use of stereotyped and recurrent figures or formulae in chant melodies. The practice has been labelled ‘centonization’, and it has been interpreted in the sense of the composition, i.e. putting together, of melodies from formulae in the manner of a mosaic or patchwork, a technique that is said to bestow on the melodies an aesthetic quality of unity. The difference is crucial, for I see the formulae as essential elements in the improvisatory system in the unwritten tradition, whereas the conventional interpretation incorporates them in a conception of composition as if in a written tradition. ³⁵ This paratactic stringing out of the biblical text has itself been interpreted in the light of the unwritten tradition in the background of the Bible. Walter Ong, in his book Orality and Literacy (New York, 1982), 37, calls it an oral residue in a written text.
20
Medieval Impro visat i on
Ex. 1.3. Hypothetical revision of another passage in Ex. 1.1 according to inferred rules T (V2) me
P
os
et
S
S
com
pre
K
hen dam
figure S, which continues until the next clausula (cf. V3, et inimicorum meorum dedisti michi dorsum). The recitation is usually articulated with the following clausula by way of a rise to a from which there is a descent to the f that begins the figure K of the clausula. The treatment of the recitation in the second verse at et comprehendam is anomalous on the criterion of the general practice, with the series of S figures following directly after the P of the clausula. I have written a suggested alternative version as Ex. 1.3. The length of the recitation between M and the clausula depends, of course, on the number of syllables in the sense-unit. The longest is the one in the third verse, beginning et inimicorum: sixteen syllables, comprising two sung as the continuing figure M, ten recited to figure S, one on the articulating note a and three sung as the clausula. At the other extreme is the phrase et ab his qui oderunt me in off.: eight syllables: two as the continuation, M, four as the clausula K'T'Q. In the light of the general pattern we can understand the settings of the remaining two syllables: his is sung to just one recitation figure, qui rises to the a, as in all recitations. Now we can interpret et salvum (off.) and sperabo (V1): the initial a of the continuing figure M, followed by one instance of the reciting S figure and the articulating tone a. Et non (V2), one syllable shorter, makes do with one S figure and the tone a. Finally I turn to beginnings, first those of the second and third verses: the melisma X. This is a standard melody for offertories in F in this tradition. It usually begins a verse, and aims towards a clausula. Differences are a function of differences in the textual situation. The sustained vowel through the melisma is that of an accented syllable. In verse 2 that syllable is preceded by two others in the same word, in verse 3 by two; hence the requirement for the additional g at the beginning of the latter. The exit from the melisma can be understood through comparison with a passage like that in the offertory, with text (me-)um / et salvum me fecit. In verse 2 the melisma is finished off with the P figure on the syllable (perse)quar. The next three syllables (in-i-mi-[cos]) are given the abbreviated M figure followed by the articulating a that prepares the clausula KTP. In verse 3, what was the P figure in verse 2 is broken up to accommodate two syllables (-sti me), and the continuation on virtutem is like the corresponding continuation in the second verse.
Medie val Impr ovisat i on
21
The melisma X, the reciting figure S, and the sequence KTP are the most welldefined and stereotyped features of the melodic idiom of offertories in F in this tradition. All else can be thought of as ways of getting into and out of those, or getting from one to another. Thus the beginning phase of the offertory section of Ex. 1.1 is not in itself stereotyped, but is a loosely strung exploration of the two overlapping modules f–g–a and g–a–b that are then shaped together in the stereotyped configuration of figure K. We have this chant in two other sources for the Roman tradition: the eleventhor twelfth-century gradual Vatican Vat. lat. 5319 for the Lateran Basilica, and the thirteenth-century gradual Vatican San Pietro F22 for the St Peter’s Basilica (this version has no verses, for by the time of the compilation of F22 these had been dropped from the performance of the offertory throughout the Western church). I have shown, beneath the transcription from the Bodmer Codex in Ex. 1.1, a few variants from these sources that are instructive from the point of view of this study: i. This figure is written for the fourth syllable in both offertory and V1 in MS 5319. ii. This version of figure P is written consistently throughout MS F22. It retains its identity as a constituent formula in the P position of the idiom as that was construed by the singing community of St Peter’s, just as its counterpart does at Santa Cecilia in Trastevere. iii. In F22 the figure Q is obviously the simplest adaptation of P for ending on the tonic. This is a different way of deriving a Q figure than the way that is represented in the Bodmer Codex. The point is that each tradition has developed its own internally consistent way of carrying out the function at this place in the chant, within the idiom that they share. Thus function is primary in setting the formal constraints, and formula is a particular concretization of a function. iv. Here the versions of Codex Bodmer and 5319 seem to be responding to two different cues, each correctly in its way. Codex Bodmer responds to the fact that the text has come to the end of a sentence, and writes the clausula accordingly. 5319 responds, presumably, to the coming repetenda, and so writes the half-close. But it does not, in fact, cue the repetenda here (it does so at the end of the second verse only). These variants are not random corruptions of an original; they show each of the local ecclesiastic-musical communities using the generative system of this idiom as they know it with consistency and deliberateness, and thereby they show us the boundaries of the system as such. Variant v is of quite a different sort. It is most likely a random corruption, but one that has come about in the written tradition through writing. MS 5319 writes the melisma X one step higher than Bodmer.
22
Medieval Impro visat i on
They come back together at the articulating tone a just before the clausula. It might be thought that, as this occurs at the beginning of the third verse, the melisma was sung a degree higher than in the second verse for heightened effect. But this sort of effect, which we know from commercial popular music of today and as a nineteenth-century operatic effect, is not in the medieval aesthetic. In any case, of nineteen occurrences of the melisma in MS 5319, only this one is written at that pitch level. It is most likely a writing error, either in copying or in writing down from memory or dictation. What I have described is an interwoven texture of materials and procedures for making a melodic presentation of an ecclesiastical text suited for a liturgical occasion of a particular kind. This context of themes and principles is an idiom of composition, written or unwritten. Because of the proximity of the Bodmer Codex to the unwritten tradition in Rome, and because of the virtual certainty that the Roman Factus est dominus is a chant of long tradition, we can read this inscription of it with a fair degree of confidence as representative of the unwritten tradition. Chants of the same genre, composed in the same idiom with different ecclesiastical texts, are related as members of a family.³⁶ The description has been fearsomely verbose and complex—so many factors in so many different parameters, so many rules. But it is different to think of such idioms being internalized non-verbally by singers who practise them daily and have been doing so since childhood. Perhaps most important and difficult is for us to think that the functions of concrete melodic formulae would have been known as immanent qualities, as much so as their actual note contents—that the figure g–f–e–f–g, for example, would have been called up from the singer’s memory only as he was coming to sing the last syllable of a clause that is syntactically complete but not the end of a sentence. We know the phrase ‘good night’ not only for its phonetic and semantic contents but also as a phrase we utter during a certain portion of the day and as the sign of a certain action (taking leave or going to sleep). Uttered in a different context it would have a different illocutionary meaning.³⁷ (Think of ‘Good night!’ as an expression of disgust.) The most helpful exposition of the relationship between the melodic contents and the functions of formulae is presented by Max Haas as a central conception for his recent book Mündliche Überlieferung und altrömischer Choral: Historische und analytische computergestützte Untersuchungen.³⁸ Haas’s exposition accom³⁶ Here I have reference to the concept of ‘tune family’ from the domain of folklore studies. A fine exposition of that concept and its history, and of its relevance to chant studies, is presented in Robert Labaree’s ‘ “Finding” Troubadour Song: Melodic Variability and Melodic Idiom in Three Monophonic Traditions’ (Ph.D. diss., Wesleyan University, 1989). ³⁷ Regarding ‘illocutionary meaning’ in music see my article ‘Methods, Style, Analysis’ for the round table ‘Current Methods of Stylistic Analysis’, in Report of the Eleventh [IMS] Congress Copenhagen 1972 (Copenhagen, 1974), 61–70 and Ch. 13 n. 9. ³⁸ (Berne, 1997), 48–51.
23
Medie val Impr ovisat i on Ex. 1.4. Figure 8 from Max Haas, Mündliche Überlieferung und altrömischer Choral, 51
8
Ab
oc
cul
tis
me
is
plishes two essential purposes at once: it shows what the point of formulaicism was, and it shows how we can reliably recognize and communicate about formulae when we see, or hear them. Because of its great value I interrupt my own exposition in order to summarize it. My summary follows Haas’s language closely, in my translation. Every chant in a monophonic liturgical repertory has a definite function: chant x for function y. This is not just a matter of classification, it means that singers of a particular chant community use the chant as the chant for that specific occasion. Thus a chant in a manuscript stands for something, it is accordingly a sign or an index (Indicator). What is more, it is a conventional, not a natural, index, which can be used correctly according to the convention, or incorrectly. Its function is its meaning. It either has that meaning or it does not. A ‘chant community’ is defined by its agreement about the meanings of its signs in this sense. Now Haas brings the concept of ‘conventional index’ down to the level of tone segments within chant melodies. His analyses show an association between specific segments and their functions in specific contexts. (The interested reader is directed to Haas’s book for the demonstration of how segments are identified.) Example 1.4 shows the simplest of his examples. Haas reports that in 219 responds or verses of Old Roman graduals the note sequence g–b–a occurs 296 times. Of those occurrences seventeen fall on a single syllable. In all seventeen instances the syllable is accented and is intoned by the sequence g–b–a as the first characteristic segment of a model melody. It is not that the sequence g–b–a marks an accented syllable in such a position somehow by its very nature or essence, but simply that this sequence was conventionally used by that chant community in that sense. That makes it an index for the convention, for which it is unimportant whether there has been an explicit agreement about it. ‘Music steps in in order to make the syntactic, semantic, and phonetic conditions and interrelations of the words audible . . . The concept “formula” referring to music will be used henceforward in this work only to refer to the multiple occurrence of segments that have a specific meaning and only that meaning’ (p. 51). (Under such a definition the formula tabulations and labellings found in numerous modern works of chant analysis would be meaningless.) The composition of Ex. 1.1 is so closely determined by the rules and constraints in various categories that I have described, affecting virtually every moment of the performance, that repeated performances under the same network of constraints
24
Medieval Impro visat i on
are bound to be very much alike and might easily be identical, although we could say that they were performed as improvisations by singers who knew the rules very well. It is an extreme case in that respect, we might say a boundary case. Then again, singers could only have learnt the rules by committing the chant, or others like it, to memory. This is just to illustrate how the concepts of improvising and singing from memory flow into one another in this environment. Put it in terms of what the singer has learnt and kept in mind as the basis for performance: whether it is the chant as a whole that he recalls more or less exactly, or some concatenation of principles of selection and order and actual melodic bits or formulae—call it an improvisatory system—that allows him to make it up as he sings. The answer may lie with one of these hypothetical formulations or the other, or anywhere along the continuum of which they are the boundary paradigms. If the singer produces the same chant time after time, that can be accounted for by the model at either extreme. If we are asked to give directions for getting from one place to another we might do so by performing an internalized verbal script that we have repeated many times, or we might refer to a mental map that we have of the region and make up our directions as we go, being reinforced in our performance by the familiarity of critical turning points as we arrive at them. Neither method is in itself bound to produce more nearly identical directions from one time to the next than is the other. If we are the recipients of such directions we tend to do better in following them if we can convert the script to a mental map than if we just repeat the script verbatim. The differentiation that I am suggesting here is interestingly reflected in a differentiation of cognitive faculties that we make when we speak of performing music without the use of a score: ‘by heart’ and ‘by ear’. The former is easily translated as ‘from memory’. The sense of ‘playing by ear’ is more elusive. We can take a clue from a common metaphorical use of the phrase: someone says in setting off on a trip, for example, that she means to ‘play it by ear’, that is to proceed without a detailed itinerary, in a way by improvising. In the context of musical performance we usually speak of the ability to ‘play by ear’ as the ability to perform something from an internalized sense of how it goes (e.g. someone who can play an evening of songs by George Gershwin without having learnt them from score or practised them), or to repeat something right after hearing it, based not only on a good memory but also on a good knowledge of the style or idiom. I shall call such a basis for a performance an aural paradigm. The skill of performing by heart was explicitly required and cultivated in the Middle Ages; the phrase itself has come down to us from that time.³⁹ I shall try to ³⁹ This is explicit in Aurelian’s Musica disciplina and Notker’s preface to his sequence collection. The passages are cited in the introduction to Ch. 6.
Medie val Impr ovisat i on
25
show in the last section that the skill of performing by ear was likewise cultivated, even though it was not as prominent in the conceptual vocabulary.⁴⁰ Example 1.1 affords us a rather close look at the elements of what we may consider a hypothetical improvisational system for the composition of plainchant in one genre and one tradition. Whether it manifests a system that was in operation as such a system at the time of the writing down, and how accurate a record it is of what was sung in its liturgical place in the unwritten tradition, we cannot know— hence its hypothetical status. But we can be certain that mastery of such principles would have been—and still would be—sufficient basis for chanting such a liturgical text. And we know, too, how typical it is of the Roman tradition. The high density of constraints and the restricted scope of the formulaic repertory are interrelated characteristic features of the Roman tradition, and are in clear contrast with the Frankish tradition. I suggest in Chapter 6 that this difference is a consequence of the fact that the Roman tradition continued as an unwritten one during two centuries after the time when the Gregorian tradition became established in books. On this interpretation the more prolix formulaic treasure of the Frankish tradition is a consequence of the earlier arrest of the stereotyping and reduction that is characteristic of oral traditions through the processes of canon formation and writing down. This could seem paradoxical from the perspective on improvisation as impromptu and unplanned performance. But the point is that in a highly regulated oral tradition, in a context where uniformity and consistency of practice are highly valued, the evolutionary pressure is likely to favour the conservation and reduction of means. I believe this situation exemplifies precisely Nettl’s generalization that ‘improvisatory systems . . . tend to be more compact stylistically than those in which improvisation is absent’.⁴¹ That is why the a priori conception of improvisation as the exercise of a momentary whim is of so little use in regard to medieval music. This brings me to a point of aesthetics. The Frankish tradition of Gregorian chant has been much praised in the scholarly literature for its qualities of unity, consistency, coherence, logic, rationality, balance, and tightness of construction. It has more than once been likened on that score to the music of Beethoven (see Chs. 7 and 9). We can find qualities answering to such language in Ex. 1.1—some might say in excess. But those qualities arose from the exigencies of a regulated generative system in a situation where the demand for consistency of practice was great, whether or not they were recognized as aesthetic values (there is no indication in ⁴⁰ I am aware of one early source in which there is explicit mention of learning by ear, the oft-cited passage in De harmonica institutione, ch. 5, by Hucbald of St-Amand (c.900) in which he compared neumes with alphabetic notation. Since neumes do not provide information about the interval contents of a melody, he writes, one cannot perceive its melodic content ‘unless one gets it by ear from someone’ (Hucbald, Guido, and John, 36). ⁴¹ ‘Improvisation’, in The New Harvard Dictionary of Music, 392.
26
Medieval Impro visat i on
medieval writing about chant that they were; see the introduction to Ch. 5). The critical categories identified in that language are the projections of aesthetic values that have prevailed during the period since the turn of the twentieth century in which medieval chant became the object of scholarly study and revival in modern editions. In investing the chant with those formal qualities scholars have given historical weight to their importance as aesthetic values, just because of the position that chant has assumed as the founding tradition of Western music.⁴² Herein lies another side of the seeming paradox that grows out of the embedded negative connotation of ‘improvisation’. Just as there would be, by its lights, an inherent contradiction in the idea of an improvisatory tradition that produces a textual canon, so also there would be a contradiction in the idea of an impromptu practice producing melodies whose leading attributes are balance, cohesion, rationality, etc., all those qualities of good design, in the sense of good intentions. That is why, again, the improvisation concept must be neutralized if it is to be of any use in describing medieval musical traditions. But then it is indispensable. We must be able to use it, liberated from the negative connotations with which it has been stigmatized in some recent literature, and in full recognition that we cannot sharply circumscribe its referents.
III In each of the Old Roman sources for the gradual Sciant gentes (Ex. 1.5, CD tracks 4 and 5) there is a tendency for melodic figures to recur without change (e.g. the settings of (De)us meus and ut rotam in the Verse).⁴³ This is a characteristic of the Old Roman tradition in general that we see in the extreme in Factus est dominus. We see it in the Gregorian tradition, too, although to a lesser degree. It is this repetitiveness and the corollary economy of formulaic contents of the Old Roman melodies that give them ‘the appearance of an oral tradition that has only recently been written down’.⁴⁴ Behind this remark is the idea that, whether or not performers in oral traditions tend towards the elaboration of florid, ornamental features, they tend in any case towards an economy in the number of constituent elements that need to be managed, such as the melodic formulae of chant. And as the Old Roman tradition was a strictly oral one for about two centuries longer than the Gregorian tradition, it displays this latter characteristic in substantially greater measure. ⁴² There is further discussion of this in Ch. 9. ⁴³ See Ruth Steiner’s discussion of the Old Roman transmission of this gradual in Neues Handbuch der Musikwissenschaft, ii: Die Musik des Mittelalters, ed. Hartmut Möller and Rudolf Stefan (Laaber, 1991), 45–9. ⁴⁴ Thomas Connolly, ‘The Graduale of S. Cecilia in Trastevere and the Old Roman Tradition’, JAMS 28 (1975), 413–58 at 434.
27
Medie val Impr ovisat i on
Ex. 1.5. The gradual Sciant gentes: (1) in the Frankish Gregorian transmission, Vatican edition; (2) in the Old Roman transmission of Codex Bodmer, fo. 35v; (3) in Vatican Vat. lat. 5319, fo. 36r–v; (4); Vatican San Pietro F 22, fo. 16r Sci
ant
gen
tes
1 8
2 8
3 8
4 8
quo
ni
am
no
men ti
bi de
1 8
2 8
3 8
4 8
tu so 1 8
2 8
3 8
4 8
lus
al
tis
si
mus
us
28
Medieval Impro visat i on
Ex. 1.5. cont. su per
om
nem
1 8
2 8
3 8
4 8
ter
ram
1 8
2 8
3 8
4 8
V De
us
me
us
1 8
V De 2 8
3 8
4 8
us
29
Medie val Impr ovisat i on Ex. 1.5. cont. po
ne
il
los
ut ro
tam
1 8
2 8
3 8
4 8
(tam)
et sic
ut
sic
ut
sti
1 8
2 8
vel ud 3 8
ud 4 8
pu 1 8
2 8
3 8
4 8
lam
an
te fa
ci em
30
Medieval Impro visat i on
Ex. 1.5. cont. ven
ti.
1 8
2 8
3 8
4 8
Comparison of the versions of Sciant gentes that are inscribed in the three surviving sources of the Old Roman tradition for the mass (the version from the Bodmer Codex is on CD track 5), first with the Frankish Gregorian version (CD track 4), then with one another, will direct us to still other aspects of compositional methods in the unwritten practice. In both traditions the melody is composed on the same pattern of melodic range-modules: centred on the pentachord d–a, extending downwards to the lower c and upwards to the upper c' (with a single upward extension to f ' in the Gregorian version), with d as the final and a as the central pivotal tone. (In the Gregorian tradition it is classified as first mode. There is no modal classification in the Roman tradition.) This is one basis of the composition. The other is the design of the melody for the clear projection and appropriate emphasis of the liturgical text which was itself edited for singing. The principal decision in accomplishing that task was the division of the liturgical text into sense-units, and we can see that all the versions reflect the same reading.⁴⁵ The sense-units are articulated in that the melodies mark the end of each with some cadential figure, and begin the next with an opening figure. In this aspect the melody plays a role parallel to punctuation in a text to be recited. But then within the setting of each sense-unit, the declamatory pace of the text is the same in all versions as well, as controlled by the brevity or length of the melodic figure carrying each syllable. Thus all opening syllables are relatively short, but the syllables ro-(tam), sti-( pulam) and fa-(ciem) in the Verse are carried by long melismas in all versions. The division of the text into sense-units is, of course, a response to the imma⁴⁵ Mary Carruthers writes: ‘remembering material sententialiter [according to the sense units], following a definition of Isidore of Seville, would mean to remember it in chunks, . . . by its constituent “ideas” ’. The Book of Memory: A Study of Memory in Medieval Culture (Cambridge, 1990), 90.
Medie val Impr ovisat i on
31
nent syntactical properties of the text itself, and it would have been possible for singers in the same and different traditions to arrive at the same divisions independently. The choice of modal type can also have been a coincidence—there are not so many from which to choose, after all—and that choice can account for many of the similarities with respect to range and principal tones. But such things cannot be said as readily about decisions regarding the declamatory pace. The similarity of all versions with respect to them is the most compelling sign of a continuing tradition for the chant, a tradition that goes back to a time before the separation of the Frankish Gregorian and Old Roman traditions, and that was maintained through unwritten tradition at least until the Carolingian era in the Frankish regions, and at least until the eleventh century in Rome. This is a constant at a deep level that is far more significant for such a tradition than the details of the melodic outline itself, a fact that tends to be obscured by the habit of comparing the melodies of chants as complete and fixed configurations of tones. Among the three Old Roman versions there are variants of different kinds. For one, there are the signs of fluidity in the tradition, for example the number of descents f–d in the final melisma, the details of the descent from a to d on the very last syllable of the Verse, or of the melismas on both syllables of the last word (terram) of the Respond. These have the appearance of an ad libitum component in what is on the whole a rather consistent transmission of the melody. We do not mainly find such a component in the written transmission of the Frankish Gregorian tradition, whatever the degree of its textual variation.⁴⁶ It may be tempting to attribute this difference to the continuing unwritten transmission in the Old Roman tradition and the older Frankish Gregorian textual tradition, but again that would be only the exercise of an ungrounded bias. In the tradition of the tropes, a tradition for which only a written transmission can be documented, a fluidity greater than that of the Old Roman tradition is characteristic (see Ch. 11). A related bias would associate the fluidity in the transmission with the florid nature of the melodies, and the two together with the phenomenon of orality. But the Frankish Gregorian tradition abounds in stereotyped melismas that are transmitted in stable form from source to source and even from melody to melody, and that were sufficiently well known as to require only cuing in written sources much of the time.⁴⁷ I identified another type of variant first in the discussion of the Old Roman offertory Factus est dominus, the writing out of the melisma X in the third verse, in the version of MS 5319, at a level one degree higher than in the other sources. I ⁴⁶ Helmut Hucke has reasoned from the extent of textual variation between early notated sources such as Laon 239 and St. Gallen 359 (20–40% between those two) that such sources cannot have been accomplished simply through copying of a musical text, and that an Urtext of the Gregorian gradual cannot be retrieved through text-critical methods. See ‘Gregorianische Fragen’, Mf 41 (1988), 304–30 at 327. ⁴⁷ See the article ‘Neuma’ in NG xiii. 23–5. The bias referred to here has undoubted connections with the tendency to characterize the Old Roman style as Mediterranean. See Ch. 9.
32
Medieval Impro visat i on
simply called it an error in that discussion, but there are other things like it in the transmission of Sciant gentes that suggest a better context of meaning. In the melodic setting of (quoni-)am no-(men) in the Respond, the version of the Bodmer Codex is inscribed first two degrees (-am), then one degree higher than that of MS 5319. MS F22 begins in agreement with Bodmer, and continues in agreement with 5319. In the second and third neumes of the melisma on (altissimus) in the Respond, there are several disagreements about pitch height. In none of these can any one version be singled out as an error, as can the displaced melisma X. All such variants in these sources are clues to a difficulty that the notators must have had in writing down their melodies (if not these notators, then the notators from whose sources these were copied). Remembering that we confront here what is more or less the border between oral and written transmission in the Old Roman tradition, it is a difficulty in setting the melodies as the notators knew them into the notational matrix for the first time. This must have been a problem already for notators in Frankish regions earlier, translating the turns and figures of melodies into non-diastematic neumes. But the first notators of Old Roman chant would have had to add to their tasks the exact translation of melodic interval distance into the fine differentiation of vertical distance that was the operating principle of a notational system that did not evolve with the experience of their musical tradition in writing, but lay ready for their use as novices. Unlike other chant communities, the Romans did not develop their own notational idiolect but adopted the music script of the neighbouring Beneventan community. That task is second nature to us but must have been totally strange to them. So here variants were created through written transmission that would not likely have occurred in unwritten transmission.⁴⁸
IV ‘To compose and perform discant extemporaneously’: that describes a very substantial part of learned musical performance in the Middle Ages. ‘Discantus’ is a term that had been in use since the time of the Carolingian mania for producing pedagogical books of all sorts, including aspects of musical theory and practice. It referred to the performance of plainchant in two vocal parts: the chant itself and an extemporized voice that—taking account of the whole of the medieval period—could range from a strict counterpoint to an obbligato vocalise. ln the ⁴⁸ Charles Atkinson has written about problems of chant transmission and classification that arise only when notators are confronted with the task of ‘capturing [their] precise intervallic structures in diastematic notation, with its dependence on a diatonic tone system’. See ‘From “Vitium” to “Tonus acquisitus”: On the Evolution of the Notational Matrix of Medieval Chant’, in Cantus Planus, 181–97. David G. Hughes, in ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’, makes reference to the problems that must have been posed by ‘the reduction to written symbols’ (JAMS 40 (1987), 377–404 at 394).
Medie val Impr ovisat i on
33
twelfth century it was joined by the term ‘organum’, which at first could refer to the same range of practice as ‘discant’. From the thirteenth century onwards the two words became specialized, ‘discantus’ for counterpoint in the narrow sense, ‘organum’ for florid accompaniment. The plainchant melodies were written down from the beginning of the practice, and sometimes counter-melodies that were invented to accompany them were also notated here and there, even whole repertories of them, some of which were widely transmitted. But extemporizing on the chant—‘singing on the book’, as it was sometimes called—continued to be practised long after the introduction of music writing.⁴⁹ No handbook that has survived from that time has a sharper orientation to this practice as a performer’s art than one written in the early thirteenth century under the rubric ‘[Here] begins the art of organum’, generally known today as the ‘Vatican Organum Treatise’.⁵⁰ It begins ‘An organum is a song that follows a preceding song, for the cantor ought to precede while the maker of the organum ought to follow, and the cantor ought to finish first . . .’.⁵¹ Organum is conceived here as something accomplished in the act of singing.⁵² Addressing medieval improvisation in terms of what it was rather than what it was not will bring primary attention to performance, not just as medium for the presentation of works, but as primary musical fact. That is why I have chosen to focus on the Vatican Organum Treatise for introducing this aspect of medieval improvisation. In order to appreciate something that stands out about this handbook within its genre, we must briefly consider some general features of the genre. From the beginning in the ninth century pedagogical treatises on music were modelled on grammatical treatises. They usually comprised normative precepts about a current practice, with illustrative examples drawn from the practice itself. In grammatical treatises the examples are drawn from language texts. In the handbooks that described and classified plainchant the examples are drawn from the canon of existing chant. But the authors of discant and organum handbooks, which are in ⁴⁹ A very great part of the history of European musical practice is subsumed under this topic. It has left many musical, theoretical, and pedagogical primary sources, and it has been the subject of an enormous secondary literature. The reader is referred to the succinct articles ‘Discant’ and ‘Organum’ in the New Harvard Dictionary of Music and the articles ‘Discant’, ‘Organum’, and ‘Organum and Discant: Bibliography’, in NG, to the chapter ‘Early Polyphony’ in The New Oxford History of Music, ii, 2nd edn. (Oxford, 1990), 485–553, and the article ‘Organum’ in MGG², Sachteil vii. 854–81. ⁵⁰ Vatican, BAV Ottob. lat. 3024, fos. 46v–50v. Facsimile and translation in Irving Godt and Benito Rivera, ‘The Vatican Organum Treatise—A Colour Reproduction, Transcription, and Translation’, in Luther Dittmer (ed.), Gordon Athol Anderson (1929–1981) In Memoriam (Henryville, Pa., 1984), 264–345. ⁵¹ ‘Organum est cantus subsequens precedentem quia cantor debet precedere organizator vero sequi, et cantor debet primitus finire’ (ibid. 293). ⁵² How unfortunate, then, that the new translation of the treatise begins ‘An organum is a vocal line [my emphasis] . . .’, electing for cantus a phrase from modern analytical discourse that tends in its reference to an object, something in a score, and thereby instantly throwing the reader off from the focus on the act of singing that is in the original (Godt and Rivera, p. 293).
34
Medieval Impro visat i on
effect composition manuals, had to make up their own illustrations by applying the rules that were embodied in the precepts. And in the Vatican treatise the author exceeded by far the bare illustration of his rules, writing passages of organum that get to be quite dazzling sometimes. Examples are shown in Chapter 3, and a performance of one very long composition is included on the accompanying CD (track 8). That notation of this performer’s art could be incorporated into music of the written tradition on a broad front can be easily seen in a comparison of a passage from that organum composition (Ex. 1.6(a), final cadence of alleluia Hic Martinus) with passages from written collections of the same period but of widely separated provenance: a passage from a twelfth-century manuscript from the Norman colony of Sicily (Ex. 1.6(b)), a passage from an organum of the Parisian Magnus liber organi, also of the twelfth century (Ex. 1.6(c)), and three passages from a twelfth-century Aquitanian source (Ex. 1.6(d–f )). The Aquitanian examples—which look as though they were from compositions for two organal voices—are of especial interest: two virtuosic vocal parts, linked in regulated integrative sequential patterns. The point of departure is contrary motion, but there are little imitations, crab motions, and voice exchanges—even retrograde inversions, if one likes to put it that way. But putting it that way should not encourage us to assume that such intricacies are matters for written tradition alone. Polyphonic games based on such techniques were much involved in the regulated improvisation of the Middle Ages.⁵³
V The idea of ‘aural paradigm’ that has come out here belongs among the basic supports for our understanding of medieval improvisation, for in one of its senses improvisation is composing by ear, as I suggested in the first section. I shall conclude with a brief account of a central phenomenon in medieval music practice that functioned in the sense of aural paradigms, and that demonstrates the continuity, ⁵³ An elaborate instance of such a regulated vocal game is that of the famous Summer Canon, as interpreted by Shai Burstyn in ‘Gerald of Wales and the Sumer Canon’, JM 2 (1983), 135–50. (The word ‘game’ is mine, not his.) Burstyn demonstrates an improvising polyphonic practice entailing short melodic formulae in voice exchanges and ‘unassuming rounds’ which produces a pattern of chordal sonorities that he interprets as itself a controlling factor in the improvisation. In Ex. 1.6(d–f ) above the descending triadic sonorities that are implied by the two voices descending in contrary motion must also have been an aural paradigm in getting the patterns right. What is suggested here, then, is the idea of a chordal ostinato pattern as control for an improvisational practice. There is undoubtedly more to be learnt about such patterns as another factor in medieval traditions of performance-composition. Lawrence Gushee has hinted at something like it in his suggestion of a background of minstrel practice for the polyphonic compositions of Guillaume de Machaut; see ‘Two Central Places: Paris and the French Court in the Early Fourteenth Century’, in Bericht über den Internationalen Musikwissenschaftlichen Kongress Berlin 1974 (Kassel, 1980), 135–50.
Medie val Impr ovisat i on
35
Ex. 1.6. Written sources of different provenance from the twelfth century: (a) final cadence of the organum alleluia Hic Martinus; (b) passage from a MS from the Norman colony of Sicily (Madrid 19241, fo. 11r); (c) passage from an organum of the Parisian Magnus liber organi (W1, fo. 25r); (d )–(f ) passages from an Aquitanian source (London Add. 36881, fos. 3/3v/2v) (a)
(i)
(b)
8
(c)
8
runt
8
(d )
8
8
(e)
8
8
36
Medieval Impro visat i on
Ex. 1.6. cont. (f)
8
8
not only of improvisation with composition in the Middle Ages, but also the continuity of both with performance from memory (‘by heart’) and from the score. The fundamental fact of these continuities as a condition of medieval music practice is a theme of this essay. The phenomenon to which I have reference is one of several that are known under the name ‘tonary’. First and foremost the term refers to a kind of thematic index of Gregorian chants listed according to a practical principle that has to do with practice for the performance of the antiphons. These were sung as refrains in the singing of psalms; the psalm verses themselves were sung to psalm tones, which were differentiated according to the modes, and each of which was provided with a variety of terminations. The antiphons were classified according, first, to the mode of the psalm tone with which each was to be alternated in performance, and then according to the termination of that psalm tone that made what was regarded as the best melodic link to the beginning of the antiphon. Knowing which antiphon was to be sung according to the liturgical occasion, the singers would consult the index to fix on the psalm tone and termination—given by the category in which the antiphon is listed—that they should choose for the psalm verses. In addition to such an index, tonaries include one or two sets of melodies, each set comprising one melody for each mode. The melodies of one such set are called ‘intonation formulae’ in the modern literature; they are vocalises on strings of vowel-rich, highly singable nonsense syllables such as noanoeane or noneoeane (a different string of syllables is provided for each mode). The melodies of the other set have been called ‘model antiphons’. Each is an antiphon-like setting of a biblical passage that begins with the number of the mode represented by the particular melody, presumably as an aide-memoire. For example, the first-mode melody sets the phrase Primum quaerite regnum Dei (‘Consider first the kingdom of God’) from Matthew 6: 33. Examples 1.7(a) and (b) are transcriptions of the intonation formula and model antiphon for the first mode as they are notated in Paris lat. 1121 (early eleventh century). These ‘model antiphons’ are not, in fact, antiphons used in the service. They must be paradigms, in some sense, but for
Medie val Impr ovisat i on
37
Ex. 1.7. Intonation formula (a) and model antiphon (b) for the first mode, from Paris lat. 1121 (a)
8
No
a
no
e
a
ne
(b)
8
Pri
mum que ri te
reg num De i
what? Michel Huglo calls them ‘touchstones’ for aiding singers in recognizing the modes, something they had to be able to do in order to sing the liturgy.⁵⁴ But this practical explanation does not cover all the circumstances about the tonaries. For one, the indexes include chants other than antiphons that did not need to be matched with psalm tones (e.g. responsories and graduals). For another, the intonation formulae and model antiphons are redundant on that explanation. For a third, tonaries comprising only those paradigms, no index, were often written at the beginnings of tropers. As to the first circumstance, we could speculate that the tonaries set in motion a drive to classification, which continued beyond the antiphons. As to the second, the listing of antiphons is never complete, and singers might have used the paradigms for determining the modes of the unlisted ones. For an explanation of the third circumstance we have to look for a different sort of purpose. When it comes to the tropers, we must take account of the facts that (1) the melodic transmissions in their notations are not nearly as consistent on the whole as are the notations of the traditional liturgical chants that are their hosts; their repertories are not canonical in the same way; and (2) their notations are rarely so precise that they can be read off by someone who does not know them. That may also be true of the notations of the chants themselves, but their uniformity in transmission allowed the notations to be read rather more in the sense of emblematic mnemonics. But in any case in the age of staffless neumes—even those we call ‘diastematic’—‘reading music’ cannot have been conducted with the fluency that is now our standard. The singer who held the melodic paradigms of the tonaries in his musical consciousness must have depended on them as referents and for orientation in responding to the cues of the notational signs. Knowledge of normative modal patterns can be a sufficient basis for interpreting the notational signs. Putting it the other way round, the patterns of the neumes can provide clues to the ⁵⁴ ‘Tonary’, NG II, xxv. 694–8.
38
Medieval Impro visat i on
mode; they can trigger the mental presentation of the whole modal configuration that then guides the reading of the neumes. When I speak of the mental presentation of the whole modal configuration I think of the singer calling upon a kind of tacit knowledge. And the most common way that we have of referring to such tacit knowledge when it is knowledge of music is to say that we have it, or know it, ‘by ear’. These modal paradigms constituted archetypes that were the basis of medieval melodic tradition for the liturgy. Singers had to have them mentally present—‘in their ears’, we might say—for the specifically practical purpose that I have described, for their reading of staffless neumes, and altogether to function in the melodic tradition. That helps us to understand the inclusion of tonaries in virtually every kind of liturgical book: graduals and antiphoners, tropers and sequentiaries, psalters, not just in books whose contents had to be matched with psalm tones. When it came to traditions like those of the tropes, the ‘reading’ of music was not a simple decoding of something fixed, it was a reconstitution on the basis of both the aural paradigms for the modes and the cues of notation. These aural paradigms were bridges connecting acts that we tend too often to regard as separate: reading, remembering, improvising. Among medieval tonaries some are illustrated with human figures dancing, juggling, playing musical instruments. They have been interpreted as jongleurs and as King David with his attending musicians (Pls. I–IV, from Paris lat. 1118). Figures like them in function have been carved on columns in the cathedrals at Cluny (Pls. V–VIII, on capitals of the abbey church at Cluny, now in the Musée Ochier in Cluny), Autun, and Vézelay, more or less contemporary with the manuscript illustrations. In the cathedrals of course they do not illustrate tonaries, but they are associated with inscriptions that identify them as representations of the modes, like the words laid under the model antiphon melodies in the tonaries. Their presence transcends all reference to systematic considerations such as the classification of antiphons and their linkage to psalm tones. They are transformations into the visual realm of the concrete universals that are the modal paradigms. They are emblems of the presence of music and, as Tilman Seebass has interpreted it, in their display of instrumental music and dance they are emblematic, not of the abstract discipline of musica but of music based on action.⁵⁵ ⁵⁵ Musikdarstellung und Psalterillustration im frühen Mittelalter (Berne, 1973).
W r i t t e n M u s i c a n d O ra l M u s i c
chapter 2
Written Music and Oral Music: Improvisation in Medieval Performance
} Current talk about how music was made, preserved, and disseminated in the medieval world tends to turn rather obsessively on a small vocabulary of words that play in a system of variously configured linkages and oppositions: words like oral,literate (or written), composed,improvised, stored in memory or retrieved from it, variable (or fluid), stable (or fixed). There is greater consensus that these are the essential terms in any representation of medieval music-making than there is about the exact nature of the musical states, phenomena, or processes that the words are supposed to stand for. They derive meaning, rather, from the way they are thought to be related to one another, and such meaning can be quite rigid, narrow, and of an a priori nature. Improvised and variable are commonly placed in a reciprocal relationship—if a piece is a product of improvisation it is expected to be variable in subsequent performances. If it appears in variant versions that raises suspicion, at least, that it might have been improvised; memorized and written are linked reciprocally in similar fashion with stable or fixed, as is composed, which in turn is treated as a virtual synonym of written, while oral is linked with variable, improvised is used as the opposite of both composed and memorized, and so on. This system supports such assertions as the following, delivered in a discussion of troubadour–trouvère tradition: ‘Where oral tradition prevails, no performer need feel obliged to repeat the same song identically at each performance.’¹ This is stated with unusual directness and explicitness here, but the rule it promulgates is usually working somewhere beneath the surface whenever it comes to representing a musical tradition that functions in a (musically) scriptless culture. When that tradition is subject to a belief in its canonical status, however, like the belief that is visible in discourse since the Carolingian era about the status of the Gregorian chant tradition, this principle works at cross-purposes with an equally active belief in the stability of the repertory through its oral transmission. In the case of chant the tension is discernible in the wariness with which its acknowledged origins in oral tradition has been approached in modern historical writing, espe¹ Giulio Cattin, Music of the Middle Ages (Cambridge, 1984), 134.
39
40
W r i t t e n M u s i c a n d O ra l M u s i c
cially given the tendency to collapse oral tradition and improvisation into a single category. This is where the concept of memorization comes into play, as the preferred alternative to improvisation because it can account for the stabilization of the tradition before its initial writing down that seems to be suggested by the uniformity found in the earliest written transmissions of the chant. But assigning this role to memorization locks it into a conception of photographic, or note-for-note memorization that cannot be demonstrated for the period in question, and that is dubious on most general accounts of the functioning of memory (see the introduction to Ch. 6). Throughout this system the meanings of the key terms are generated from their assigned functions within the system rather than from their aptness as descriptive terms for what has actually been learnt about aspects of music-making in the medieval world. In short the system is a self-referential language game, sealed off from the happenings that it is meant to describe. Its workings as such are lubricated, however, by the familiarity of each of its key terms, and the concept that it stands for, from the context of modern practice and description in which we have learnt them: a concept of ‘improvisation’ under which each performance of something is different from every other one; a concept of memorization as the recording and retention of whole songs note for note; a concept of fixity or stability that presupposes clear and shared criteria of what counts as ‘the same’ from one performance to another, to which must be added equally clear and shared standards for the relationship between what is written and what is performed. But we have no independent idea of what the fit of any of these conceptions is to the medieval music culture, and in this chapter I mean to cast doubt on the assumption that the system they constitute functioned in the Middle Ages. So it is no wonder that the field of historical studies in medieval music has been unable to bring the elements of this system into balance with one another and to arrive at agreement about a coherent picture of music-making in the medieval world. As far as the history of chant is concerned that leaves in place what László Dobszay has aptly called ‘The Mystery of the Gregorianum’, in an article of that title.² The title plays on a heading in the first volume of Peter Wagner’s Einführung in die gregorianischen Melodien, ‘Die gregorianische Frage’.³ The question, as Wagner formulated it, is whether the ‘codification of the mass and prayer service, and the chant that belongs with it, which superseded all other versions from the seventh century onwards, and which in all essential respects is still current today, goes back to Gregory I, who was pope during 590–604’. Dobszay asks a string of questions that are forced out by that one: ‘What does it mean to create or codify the Gregorian repertory at the beginning of the 7th century? . . . How could the material have been recorded by the author or codifier? How could it come to those who used it? How could its survival and integrity have been guaranteed?’ Those, for Dobszay, add up to a mystery, transported from the agency of Gregory I (which has ² In Musica antiqua Europa orientalis (Bydgosz, 1994), 483–7. ³ Einführung in die gregorianischen Melodien: Ein Handbuch der Choralwissenschaft, i: Ursprung und Entwicklung der liturgischen Gesangsformen bis zum Ausgange des Mittelalters (Fribourg, 1895; 3rd edn., Leipzig, 1911, repr. 1962; Engl. trans., 1907), 191.
W r i t t e n M u s i c a n d O ra l M u s i c come increasingly into question) to the liturgical-musical tradition itself, the ‘Gregorianum’. Dobszay shows his awareness of the functioning of the system I have described, when he writes ‘Earlier research . . . attributed 19th-century methods of “composition,” transmission and performance retrospectively to the age of Gregory. These ideas presuppose . . . the use of music notation. If this means of transmission were lacking one could not, it seemed, . . . speak of “fixing the melody,” “the propagation of melodies,” “the reception of the repertory over a wide area,” upholding standards of musical accuracy, etc.’ That is the mystery. In a subsequent paper, ‘Two Paradigms of Orality: The Office and the Mass’, Dobszay focuses the mystery on the role of memory, which, on his view, must carry the burden of explanation when musical notation is not available.⁴ But he raises the concept to a level of sophistication and differentiation that has nowhere else been seen: Most scholars agree that the music put in writing in the course of the ninth century was a music that had earlier been fixed in the living practice. In other words: by means of oral tradition the memory was able to retain the repertory during the decades between the Frankish reception and the first notated manuscripts . . . What was this memory like? Was it reconstructive or did it preserve music like a photocopy? The question is too simple, as Dobszay shows; it depends upon the nature of the material to be remembered. Formulaic chants such as graduals, typical chants such as alleluias, individual chants such as introits, representing different stylistic and chronological layers, all call upon different aspects of the memory faculty. ‘Already the first manuscripts contain a complex repertory . . . Thus the memory which preceded written recording had layers of different depth, it went back different distances in time and combined different types of remembering.’ This perspective on memory is not only more differentiated but, most important, it is informed by the condition of the music itself, not by what is required of the memory concept in an explanatory narrative. The mystery does not fade with the beginning of music writing. As Dobszay writes in a third essay, we do not know ‘what the music scribe recorded by this device [of musical notation] and to what extent his recorded material is a Vorschrift or Nachschrift. Is the substantial identity of the material as written down in the 10th–11th century sources more important or the differences found therein?’⁵ This important question requires fleshing out. Behind it is the observation I have already made, that we do not know what the relationship was between what was written and what was sung. ‘The substantial identity of the material as written down’ can be as much a phenomenon of a written tradition—taking that phrase quite literally— as of the performance tradition. I interpret Dobszay’s remark in the sense that the ⁴ Address for the roundtable ‘Schrift und Verschriftlichung’, during the conference ‘Musik der Karolingerzeit’, Paderborn, 23–6 Oct. 1999. ⁵ ‘The Debate about the Oral and Written Transmission of Chant’, Revista de musicología, 16 (1993), 706–29.
41
42
W r i t t e n M u s i c a n d O ra l M u s i c differences might be quite significant in having persisted despite the apparent intention of producing a uniform tradition of both performance and written material. In any case, Dobszay writes, ‘The differences cannot be ignored.’ Richard Crocker’s recent long and deep reflection on the past and future of chant studies shares something of Dobszay’s sense of mystery.⁶ Crocker writes that, having read and admired David Hiley’s handbook, Western Plainchant,⁷ he was left, nevertheless, with ‘a feeling that Gregorian chant was still remote, mysterious, unattainable . . . [T]he sense of mystery resists our musical understanding, the sense of remoteness our historical understanding’ (p. 33). It is especially about what he calls the ‘prehistory’ of Gregorian chant—its history before the ninth century—that Crocker writes in this mode. The boundary is set by the beginning of music writing in the ninth century, and the language reveals the tacit and remarkable bias that ‘history’ begins only with the existence of written sources. Note the contrast with Dobszay’s blunt assertion that ‘any description of chant history and transmission that makes a sharp dividing line at the point of the introduction of neumatic notation is false’.⁸ Crocker continues: Before the ninth century . . . if we feel we need to make statements it must be not with confidence but with imagination and a sense of adventure . . . We have always assumed . . . that the repertory recorded after the ninth century was formed by its past . . . [But I]n order to study how Gregorian chant was informed by its musical past, that past must be manufactured (pp. 34–5). . . . The ‘prehistory’ of Gregorian chant consists of the centuries before 900 . . . Before that time we only guess at the music, and that truth needs repeating at every opportunity (p. 39). . . . I can caution that any ideas we derive from the Gregorian itself about how Western Latin Chant might have been before ca. 650 are fantasy (p. 41) . . . I would read it as a Romantic fantasy, a fascination with remote origins (p. 85) . . . [W]e cannot say why the chant is the way it is on the basis of how it got that way, simply because we do not know . . . We have to take the Gregorian repertory the way we find it in the sources. And to the degree that these sources . . . show us the repertory in varying use at various times and places, they too are not much use in saying how the chant came to be. (p. 42) The challenge of this bold reassessment of the limits of knowability in our field—following Dobszay I shall call it ‘agnosticism’—deserves to be engaged, especially as it aims ultimately at a reassessment of what questions are worth asking. That is brought out clearly in Edward Nowacki’s summary: Richard Crocker has been especially cogent in recommending the abandonment of lost causes and the adoption of new or neglected research methods. ⁶ ‘Gregorian Studies in the Twenty-First Century’, Plainsong and Medieval Music, 4 (1995), 33–86. ⁷ Western Plainchant: A Handbook (Oxford, 1993). ⁸ Unpublished paper, ‘Two Paradigms of Orality: The Office and the Mass’.
W r i t t e n M u s i c a n d O ra l M u s i c The attempt to read the oldest melodic texts of plainchant as the trace of the sumptuous oral tradition that preceded them . . . is ultimately unsatisfying in his view, perhaps even vain. In its place he urges us to cultivate a critical, evaluative understanding of chants composed in the era of musical notation, that is, as authoritative compositions of the time when they were copied rather than as witnesses to an obsolescent oral tradition. Putting this advice into practice would require more listening, performing, and remembering of chants as they are and less speculation about how they got that way.⁹ To be sure, Crocker’s programme is a reaction particularly to the intense preoccupation during the second half of the twentieth century with questions about the history of Gregorian chant prior to the time of its redaction, but it seems, more than that, a challenge to virtually the whole tradition of historical study in the field. He writes ‘For the last hundred years . . . Gregorian studies have been preoccupied with the prehistory of the repertory.’ This challenge arises, at least in part, out of well-founded distrust of a number of long-accepted verities, the ‘lost causes’ mentioned by Nowacki: for one, the assumption of a stable liturgy before 900 as a basis for theories about the prehistory of chant; recent research persuades Crocker that ‘the early history of liturgy has little demonstrable relationship with the Gregorian repertory as we know it (p. 41); for another, the ‘myth of Roman origins’, and with it the central place that has been afforded the question of the relationship between Roman and Frankish chant (the creation of that myth is described in Ch. 6); for yet another, a belief in the fixity of the chant in the oral tradition through the instrumentality of memory, unsupported, however, by any consideration of the feasibility of that idea, or indeed of its meaningfulness. Crocker asks ‘What did it mean, in the absence of written records, to believe that a chant was the same in successive performances . . . And what is the meaning of asking such questions in the absence of any documents on which to base an answer?’ There can be various reactions to this challenge, and I shall confess my own. What Crocker writes about the overemphasis of the question of origins, not only in this case but in historical studies of music altogether; the preoccupation with the idea of fixity in the prehistory of chant when we do not even know what that would have meant and how it would have been measured at the time; the construction of historical scenarios entirely out of invented elements; all that resonates with much that I have thought and expressed in this book (the last point especially in the introduction to Ch. 6). On the other hand, Crocker’s agnostic and, in a way, new-critical stance seems like a renunciation of knowledge that we could have and that might be worth having and that, in its form, may be rooted in the way we configure Western music history. There are two general considerations. The inquiry after origins. A recent visit to the Rose Center for Earth and Science ⁹ ‘Chant Research at the Turn of the Century and the Analytical Programme of Helmut Hucke’, Plainsong and Medieval Music, 7 (1998), 47–71.
43
44
W r i t t e n M u s i c a n d O ra l M u s i c at the American Museum of Natural History in New York City gave food for thought about this topic. One space in the centre is dedicated to a demonstration of the ‘big bang’ theory of the origin of the universe. The space is located along the pathway that visitors must follow. The theory is just that, by no means regarded as established fact or as ever susceptible of confirmation. But the very investment in this demonstration and its strategic placement in the Center testifies to the essential place of the theme of origin in the configuration of the field of inquiry concerned with the history of the physical universe. This is confirmed in the representation of the universe in perpetual change that is projected onto the ceiling of the centre’s main theatre. The concepts of origin and development are immanent in the image of the universe that is presented. This invites comparison with the fields of evolutionary biology and geology, which are likewise organized by a perspective from the viewpoint of origin and development. Under the influence of such sciences historical representations of music, as indeed of virtually all aspects of culture, were for a long time cast in the same form. Over the past few decades, however, historians of culture have come to assert a greater interest, on the whole, in the ontology and phenomenology of our objects and in their engagement with the cultures in which they were created, than in their origins. It will always seem natural to hold in mind the form of a galaxy or an organism or a rock formation in a single cognition together with the history of its development or evolution from the point of its origin. But it no longer seems so natural to invest the question of the developmental background of something like a painting by Hieronymus Bosch with such importance, especially when that distracts attention from an analysis of its qualities and their meaning in the artist’s world. Crocker’s call for a turn away from the preoccupation with the origins and prehistory of chant and towards a critical, evaluative study of the chant that has come down in writing catches this current historiographic temper. But that was more easily said than it is likely to be carried out, because of the special investment in the theme of origins in the particular case of chant history. After all, in writing the history of Western music historians have, since the beginning of the practice in the eighteenth century, tried to begin at the beginning, or point of origin and that role they have consistently vested in the chant of the medieval Latin church. In that sense the study of chant has been the contribution from the music-historical side to the ongoing and widespread search for the culture’s origins, the culture’s creation myth. It is hard to see, then, how we would give up the enquiry into the origins of this originating music without altogether giving up that configuration for music history and the desire to learn about the origins of our culture. It does seem to me, at least, that just such a radical reconfiguration of the historiography of Western music is entailed in Crocker’s current position. That is quite a reversal for the author of A History of Western Musical Style (New York, 1966).¹⁰ As for ‘the myth of Roman origins’, everyone knows that it belongs to the creation myth of Gregorian chant that was invented in the ninth century (see Ch. 6 ¹⁰ See my discussion of the historiographic style of that book in Music and the Historical Imagination (Cambridge, Mass., 1989), 135–41.
W r i t t e n M u s i c a n d O ra l M u s i c here), but even if we could agree that it is difficult to confirm, or perhaps even that it is probably false, there might be some resistance to renouncing enquiry after the purposes that were to be served by its invention and after the core of reality on which it embroidered. Interest in the history of Western chant before the written tradition has not been aimed only at the identification of origins. Without so much as a sidelong glance at the question of origins, efforts have been under way for decades to understand aspects of the style of chants that have come down in writing as accommodations on the part of composers to the exigencies of composition and transmission in the oral tradition. Properties of written chant serve as the basis of extrapolations to the unwritten tradition. But their interpretation as such gives us a handle on them precisely for the project of the critical evaluation that Crocker wants us to undertake. Gregorian chants do display properties that would have facilitated both their composition and their recall in performance in the absence of music writing (and their reception by the religious communities for which they were intended). We cannot ignore this functional aspect in proceeding with our critical evaluations, any more than we would ignore the basso continuo in analysing a Baroque sonata. This is said, of course, under the widely shared belief in a continuity of the repertory from the oral tradition into the written one, not as any sort of proof that such chants were composed within the oral tradition. The alternative to this belief is that the repertory that is transmitted in the earliest notated books was composed at the time of the writing down. No one has made such a claim, and I believe that it is counterintuitive. There is in any case an abundance of evidence and reasoning against it—all the indications of fixity in the unwritten tradition summarized in the introduction to Chapter 6, and Dobszay’s conclusion, cited earlier, that ‘the memory which preceded written recording had layers of different depth, it went back different distances in time and combined different types of remembering’. Consequently a hypothesis that the Gregorian repertory was composed at the time of the initial writing down would have to entail either the belief that the composers intended, and were able, to make it appear like an old tradition, or the assumption that they composed in imitation of such a tradition. Either would make the posing of the hypothesis pointless. Prime among the stylistic properties visible in the written sources that can be seen as aids to composition and recall in the oral tradition are the systematic use of melody models and formulae in some types of chant and the very intricate ways in which aspects of language are determinants of the melodic process in virtually all chants. Crocker does not speak of the first two of these, and he has reserved his greatest scepticism, even scorn, for the suggestion of the third. This calls for some discussion. He writes that the melody of chant ‘cannot be said to communicate either the sound or the sense of the words with anything like the effectiveness claimed or to be expected for a style allegedly derived from words’ (p. 44). That claim is difficult to evaluate as it is formulated. I suppose what is meant is that neither the sound- nor the sense-contents of the words of specific chants can be said to correlate in specific
45
46
W r i t t e n M u s i c a n d O ra l M u s i c ways with the properties of their melodies to the degree that would be ‘expected for a style allegedly derived from words’. In order to get an impression about whether this seems to be so we need to have it put in terms of links that may or may not be observable between specific aspects of language and specific aspects of melody, and what the nature of those links is. That will come down to two general questions. With respect to the relationship between melody and the sounds of words, it will be whether melodic process seems to correlate with the phonetic aspects of language—accent, rhyme, assonance, word length, syllable length, and syllable count. With respect to the relationship between melody and the sense of words, it will be whether melodic process seems to correlate with the syntactic aspect of language—the grouping and distribution of words in functional sense-units. If the answers to those two questions are affirmative we can say that those aspects of the words are factors in decisions leading to corresponding aspects of the melodies. Whether that should count as ‘communicating the sound or sense of the words’ is an awkward question; I suppose that it should. But it would decidedly be a different kind of ‘communicating’ than occurs when melody imitates the sense of the words with madrigalisms or word-painting, something that hardly plays any role at all in medieval chant, as everyone knows (I wonder whether that is not what Crocker was denying). In the final two chapters of this book links of both kinds between words and music are demonstrated in extra-liturgical songs, as well as links between poetic and melodic form and the projection of rhetorical voice through melody. The analysis of the Old Roman offertory in Chapter 1 gives a positive answer to both questions. Hucke’s analyses in ‘Toward a New Historical View of Gregorian Chant’¹¹ demonstrate the match of sense groups with melodic phrases, something that had already been asserted a millennium earlier in the De musica of Johannes.¹² Edward Nowacki has demonstrated the link between word accent and melodic formation in eighth-mode tracts of the Old Roman tradition.¹³ In a likewise empirical mode and with a simple example Max Haas has provided both another demonstration of the link between melodic process and text declamation, and an important ¹¹ In one of the most important sentences in this essay Hucke writes ‘The basic principle of composition in Gregorian chant is the division of the text into units defined by sense; the melodic phrases correspond to these sense units’ (p. 452). ¹² ‘Just as in prose three kinds of distinctiones are recognized, which can also be called “pauses”— namely, the colon, that is, member; the comma or incisio; and the period, clausula or circuitus—so also it is in chant. In prose, where one makes a pause in reading aloud, this is called a colon; when the sentence is divided by an appropriate punctuation mark, it is called a comma; when the sentence is brought to an end, it is a period . . . Likewise, when a chant makes a pause by dwelling on the fourth or fifth note above the final, there is a colon; when in mid-course it returns to the final, there is a comma; when it arrives at the final at the end, there is a period.’ Johannes, ‘De Musica’, in Hucbald, Guido, and John, 116. Harold Powers called attention to this passage and its significance as a key to its author’s conception of the relationship between language and music in ‘Language Models and Musical Analysis’, Ethnomusicology, 24 (1980), 1–60. ¹³ ‘Text Declamation as a Determinant of Melodic Form in the Old Roman Eighth-Mode Tracts’, EMH 6 (1986), 193–226. The author announces his intention as follows: ‘Particular attention will be paid to the regular and, in some cases, predictable ways in which the accentuation, phrasing and syntax of the text determine melodic form.’
W r i t t e n M u s i c a n d O ra l M u s i c insight into the significance of such a linkage (see the discussion of Ex. 1.4 in the preceding chapter). It belongs to the grammar of this singing tradition that the melodic property g–b–a and the phonetic property ‘accented syllable’ are linked in the position ‘first characteristic segment of a model melody’.¹⁴ All these demonstrations seem to be overlooked or overridden by Crocker’s claim, and in trying to understand that I return to Nowacki’s essay cited just above. He writes: The essay will take up the subject of text expression in its classic formulation by Dominicus Johner¹⁵ and show how he was prevented from arriving at a satisfactory result by the very terms in which he framed the question (p. 193) . . . The assumptions shared by all chant scholars of his day effectively limited the study of text expression to the search for those rare and debatable instances where the melody might be regarded as forming an analogue to the semantic reference of the words (p. 226) . . . Lacking our structuralist perspective, he took no notice of the indispensable role of structure in the expression of meaning. Just as it is the formal properties of sentences that enable words to form meaningful utterances, so it is precisely the routine formal properties of plainchant that give it the capacity to project the poetic and syntactic structures of its texts in a way that makes them plain to the understanding and vivid to the imagination. (pp. 225–6) Emphasis on the relationships between poetry and melody in medieval song in general is not meant here in the spirit of encomium—praising the musicians for their efforts towards the unity of words and music; on the contrary, all indications are that they would probably not have started out with an idea of words and music as separate expressive media that one could choose to unify or not. The purpose is to identify one of the chief compositional means at their disposal and in particular an aid to oral composition, accommodating melody to the properties of words. Crocker’s assertion to the contrary, like his agnosticism with respect to the chances of learning about the condition of chant prior to the writing down, is contrary to his own salutary call for the critical study of the chant that has come down to us in writing and that we can sing and hear and analyse. In his book The Advent Project: The Later-Seventh-Century Creation of the Roman Mass Proper (Berkeley, 2002) James McKinnon shows us that the reason books containing chants for the celebration of mass have come down to us from the Middle Ages is that the Roman schola cantorum, working under the patronage of the pope, conceived the idea that there should be a cycle of chants of each of the liturgical genres that comprise the mass proper, and set out to compose them. They adapted scriptural texts for chanting, arranged them in narrative sequence, and assigned them to feasts according to their theological meanings. And it is only reasonable to conclude that they devised for each a way of intoning it musically. It was ¹⁴ Mündliche Überlieferung und altrömischer Choral, 50–1. ¹⁵ Wort und Ton im Choral (2nd edn., Leipzig, 1953).
47
48
W r i t t e n M u s i c a n d O ra l M u s i c in each of these aspects an ambitious act of composition. On the musical side it had to be an act of oral composition, as they had no way of representing melodies in writing. What that might have entailed we can only begin to infer through the close comparative analysis of chants that have come down in writing since the tenth century, but we can begin to infer it. We would do well simply to speak, with McKinnon, of the invention of medieval song in general as ‘composition’, written or oral, and to acknowledge, with Crocker, that there is no point in worrying about the boundary, when it is the oral, between pre-composition and performance from memory on one side, and improvisation on the other.
} The title of this chapter was shrewdly assigned to me by Padre José López-Calo of Santiago de Compostela in Spain, organizer of the conference for which it was originally written, and I thank him for provoking me to think again about concepts that I thought I had understood. The title projects one of the main dimensions in which we have tried to describe medieval music practice, with connections to every aspect of medieval music history. But the words of which it is composed are understood in different ways, reflecting different interests and inherited outlooks on musical tradition and culture, and different ways of thinking. So many differences can create the appearance of disagreement about the subject among scholars who devote themselves to the study of it, a situation that challenges us to differentiate truly substantive disagreements from those that result from the control that language has over our thinking. For a striking example consider the seemingly contradictory characterizations of the music culture of Paris centred in the cathedral of Notre Dame around 1200 offered by two students of the subject. Craig Wright presents evidence for the oral learning of chant by rote, for the chanting of the Divine Office from memory (by heart) rather than from the notated books that were made in increasing numbers, and for the extempore singing of organum, even in three and four parts.¹⁶ David Hiley, writing about improvisation in the Middle Ages, labels that music culture a ‘script culture’.¹⁷ How can there be so sharp and fundamental a disagreement between two such knowledgeable scholars in such authoritative publications about a subject that has been so much investigated on the basis of such a wealth of evidence? The answer is that this is not so much a question about the right interpretation of evidence as it is about the use of language and concepts that are the paraphernalia of our understanding. That is the subject that, for me, is raised by the title that was assigned to me and that I now take as my subject. ¹⁶ Music and Ceremony at Notre Dame of Paris, 325 ff. ¹⁷ MGG², Sachteil iv. 550.
W r i t t e n M u s i c a n d O ra l M u s i c
49
For a first approach I cite an essay by the ethnomusicologist Simha Arom, who has studied the indigenous music of Central Africa for many years.¹⁸ Its interest for my subject is in what it shows about how highly disciplined, structured, and consistent oral performance can be, and in Arom’s observations about the concepts with which we struggle in trying to describe medieval practice. His purpose in the description of African polyphony is to show the extreme rigor with which its performances are conducted. ‘It is’, he writes, ‘one of the most complex oral musics that one can find.’ His study, and the practice it describes, can serve us as a caveat against a priori judgements about what is and is not possible in the oral production of music, and the a priori characterization of improvisation as a capricious, random, uncontrolled way of making music. For his melodic orientation the pygmy singer disposes, for each piece, of an extremely simple and most often implicit model which he knows perfectly . . . In the swarming multiplicity in which the most complete freedom seems to reign, a rigorous organization is manifested. Each member of the community knows, at every point of the duration of each song, and in conformity with the function of that point, what variations he can execute. Improvisation, as far-ranging as it can appear to be, is thus limited to the interior of this framework. . . . It appears that all these pieces are, from the point of view of form, put together according to extremely strict principles. Nothing is left to chance. Each musician knows exactly where and how to place his particular sound in the overall structure, and the scope for individual improvisation is very limited. (p. 135)
Fig. 2.1 shows Arom’s analytical diagram of a piece, Bobangi, in which the total duration of seventy-two time units (the concept of ‘time unit’ here is something like the medieval concept of integer valor) is articulated in eight different patterns. In order to make this analysis Arom had to devise a method for isolating each participant’s performance, whereas each was able to perform only together with the rest. His ingenious solution was to have each participant perform his or her part while listening to the performance of the whole through earphones as Arom recorded the performance. This [performance] technique is the fruit of a long apprenticeship in the course of which children progressively accumulate a repertory of models and formulas that they will later utilize and transmit . . . All the members of the population dispose in effect of a virtually unlimited array of melodic modules proper to each piece and usable at the appropriate moment of the piece’s duration. Just as in speaking, the speaker of a language can, at diverse points of the phrase that he enunciates, choose from among several terms of ¹⁸ Simha Arom, ‘Une parenté inattendue: polyphonies médiévales et polyphonies africaines’, in C. Meyer (ed.), Polyphonies de tradition orale: histoire et traditions vivantes. Actes du Colloque de Royaument 1990 (Paris, 1993), 133–45. Example 2.1 is cited from an Italian version: ‘Su alcune impreviste parentele fra le polifonie medievali e africane’, in Maurizio Agamennone (ed.), Polifonie: procedimenti, tassonomie e forme: una riflessione ‘a più voci’ (Venice, 1996), 163–80. The performance that is analysed in Example 1 can be heard on the following compact disc: ‘Aka Pygmy Music’, Music and Musicians of the World; UNESCO Collection D 8054, track 5.
50
W r i t t e n M u s i c a n d O ra l M u s i c
Fig. 2.1. Simha Arom’s analysis of the temporal organization of the Aka Pygmy performance ‘Bobangi’ (from ‘Su alcune impreviste parentele fra le polifonie medievali e africane’) similar sense, in the same way each pygmy chooses intuitively, at the appropriate moment, those melodic formulas that he prefers to enunciate at this point of the song. (p. 135)
I stress the prominence given to the phrase ‘the appropriate moment’ in this account. It corresponds to the essential aspect of the function that each melodic segment concretizes in the unfolding oral production of medieval music, a concept to which I return further on. In conclusion Arom writes: In a culture of oral tradition, the perpetuation of the music can be assured only by the presence in the musician of a mental reference, a model. The model is a representation in sound, at once global and simplified, of a musical entity; it manifests its singularity in that it condenses in a sketch the ensemble of its essential features and only those. The model amounts to the most bare realization of a piece which can be identified as such by its listeners. Every realization of a polyphony or a polyrhythm is then nothing other than the incarnation of its model. (p. 137)
I have used the expressions ‘aural paradigms’ and ‘concrete universals’ in the oral production of medieval music in just this sense (see Ch. 1). Arom continues: In a certain way, the model is to the musics of oral tradition what the written object is to Western learned music: both are a support of memory. Writing confers to the work a finished character, achieved, and to which nothing essential can be added, while the
W r i t t e n M u s i c a n d O ra l M u s i c
51
model is a sketch of which nothing can be curtailed (which cannot be abridged). One could summarize this opposition thus: the written work is maximal, the model which structures a traditional practice is minimal.
I cannot help recalling at this point Walter Benjamin’s haunting remark that the finished work is the death mask of its intuition,¹⁹ and Feruccio Busoni’s longer, more optimistic declaration about what happens when a musical inspiration is written down: Every notation is already the transcription of an abstract idea. At the moment that the pen overpowers it the thought loses its original shape. The intent to write it down forces the choosing of metre and key. Form and sound medium, on which the composer must decide, set the course and boundaries still more . . . Even if something original that is irrepressible survives of the idea . . ., from the moment of the decision it is reduced to the type of a class. The idea becomes a sonata or a concerto . . . That is an arrangement of the original . . . The performance of a work is also a transcription and, as free as it might be, this, too, can never drive the original from the world. For the musical art work, before it begins to sound and after its sound has faded, remains fully present and undisturbed. It is both in and beyond time, and its essence is that it can give us a palpable representation of the otherwise ungraspable concept of the ideality of time.²⁰
The sharp distinction that Arom has made in comparing the musics of oral tradition with the written music of learned Western traditions is extremely telling as regards the written music of modern learned Western traditions. When it comes to the written music of medieval traditions things are not so straightforward. Written music of the Middle Ages may convey models, it may not be finished, it may not be maximal, in all those ways being like oral music. It may also be the opposite, as we shall see: finished, maximal, conveying every aspect of a performance. The epithet ‘written music’ received the ontological sense in which Arom invoked it in the early nineteenth century, for example when Johann Friedrich Herbart wrote ‘In strict composition music does not depend on forte and piano. The tones need simply to be heard, indeed only to be read’ (my emphasis).²¹ This is as much as to say that the music—whose inner essence, Herbart argues, is form— is located in the writing: the notational text is ‘written music’. But it does not follow that with the advent of music writing music was at once deposited into this ontological state, a tacit assumption that lies behind much thinking about this subject as regards medieval music and that has been one source of poor communication in the discussion. What is neglected is the different sense in ¹⁹ Walter Benjamin, ‘Die Technik des Schriftstellers in dreizehn Thesen’, in Gesammelte Schriften, iv/1, ed. Tillman Rexroth (Frankfurt, 1981), 107. ²⁰ Feruccio Busoni, Entwurf einer neuen Ästhetik der Tonkunst (Leipzig, 1916; new edn. Frankfurt, 1974), 29–30. The optimistic turn at the conclusion of the passage should be seen in the light of Busoni’s prolific activity as a transcriber. ²¹ Cited in Edward Lippman, A History of Western Musical Aesthetics (Lincoln, Nebr., 1992), 296 from Enzyklopädie der Philosophie (Halle, 1831).
52
W r i t t e n M u s i c a n d O ra l M u s i c
which the notational text in the Middle Ages is ‘written music’, reflecting quite a different music concept. If nineteenth-century theorists conceived that the reference of musical scores is to musical structures—a conception that has held in some circles to the present²²—the evidence from medieval writing about music and from the technology of most medieval systems of music writing suggests a conception among medieval theorists and notators that the reference of musical scores—from the advent of music writing—is to the sounding voice in motion. As this semiotic relationship is immanent in the technology of the writing systems themselves, music writing and written music have the same referents. And that is a fundamental difference between the medieval and the modern situations. For in the modern situation the passage from music writing to written music is made through a step of abstraction; the structures that are conceived to be the referents of written music are not immanent in the music writing. To speak of ‘the advent of music writing’ is to raise the first clearly debatable issue of substance to come up here, about which one would think there ought to be a definitive answer. It is the question whether musical notation was invented or developed out of older continuous writing practices. There is no question that neumes made use of older sign systems related to the performance of texts— punctuation marks of syntactical articulation—and that the rising and falling movement of the voice that was at the core of the melody concept—called ‘accentus’ by medieval writers—was once denoted by a still-older sign system— the prosodic accents. Neumes combine these functions. But the question remains whether either system had musical signification before the appearance of the neumatic scripts. The choice is fundamental: is neumatic notation a stage in a continuous evolution of the graphic representation of singing in which one cannot mark a beginning—that is, has there always been notation, or at least protonotation? Or does music writing constitute, as Wulf Arlt has put it, ‘a qualitative leap with the step into script’?²³ Here a remark by Henry John Chaytor is very much to the point : ‘The essence of criticism is the ability to understand states very different from those in which we live.’²⁴
²² In a subtle analysis that cannot be adequately summarized here, Benjamin Boretz writes ‘Sounds . . . are not part of music, however essential they are to its transmission. And neither are paint, pigment and canvas part of painting . . . Sounds, in fact, are not even what musical notation specifies . . . What scores do specify is information about music-structural components. . . . [I]t is sound-successions rather than notations that are the real symbolic languages of music; and notes require prior music-structural interpretation to be regarded as music-determinately symbolic of sounds.’ ‘Nelson Goodman’s Languages of Art from a Musical Point of View’, in Benjamin Boretz and Edward T. Cone (eds.), Perspectives on Contemporary Music Theory (New York, 1972), 31–44, esp. 34–5. ²³ ‘Anschaulichkeit und analytischer Charakter: Kriterien der Beschreibung und Analyse früher Neumenschriften’, in Michel Huglo (ed.), Musicologie médiévale: notations et séquences (Paris, 1987), 29–57, esp. 29. ²⁴ From Script to Print: An Introduction to Medieval Vernacular Literature (Cambridge, 1950), 1.
W r i t t e n M u s i c a n d O ra l M u s i c
53
The title assigned to me brings next the concept of improvisation, the most troublesome concept of all because of the tyranny of the present-day usage of the term. Rudolf Frisius writes in the general introduction to the article ‘Improvisation’ in MGG, second edition, ‘In general language usage “improvisation” means uncalculated, unprepared; in the Latin sense of the word, unforeseen . . . In the domain of music the aspect of the unforeseen refers to the absence of a written Vorlage . . . The unpredictability [of the improvised item] is tied to the unrepeatability (“Einmaligkeit”) of its sounding’ (my emphasis).²⁵ By way of contrast, consider Imogene Horsley’s far simpler definition in New Grove: ‘The creation of a musical work, or the final form of a musical work, as it is being performed.’²⁶ All the negatives are absent, leaving only the act of composition in performance. (The phrases ‘musical work’ and especially ‘the final form of a musical work’ [my emphases] are habitual and complicating, but the problems of aesthetics and ontology they raise are beyond my concerns here.) How much Frisius’ emphasis is a nineteenthcentury one can be seen from remarks of Stendhal about the composition of his novel The Charterhouse of Parma, a text of 493 pages in the most recent English version, which he dictated in fifty-two successive days in 1838: ‘I was improvising as I dictated. I never knew while dictating one chapter what would happen in the next one’; and ‘When I was dictating this, I had no idea what the next chapter would contain. Improvisation.’²⁷ In efforts to sketch a picture of ways of making music in the twelfth and thirteenth centuries, attention sooner or later falls on the ‘Vatican Organum Treatise’, which Craig Wright has characterized as ‘undoubtedly the most important, and at the same time most problematic, monument of early organum from the Ile de France’.²⁸ What is problematic is to understand the status of the examples that follow each rule and far exceed its stipulations. Again, modern scholars have offered quite different interpretations—though not on the basis of different evidence—that make a difference for the image of the practice to which this important document is witness. According to Wright, in these examples ‘the singer is provided with a fund of stock phrases that can be invoked according to the melodic movement of the plainsong tenor. [This brings to mind the formulae learnt by the Pygmy singers of Central Africa.] Once these musical formulae were part of the performer’s aural memory, virtually any chant could be sung extempore in two-voice organum without recourse to a written polyphonic manuscript.’ He reinforces this view of the way things were done by citing the identity of ‘several of ²⁵ MGG², Sachteil iv. 538–9. ²⁶ ‘Improvisation’, NG ix. 31. ²⁷ Stendhal, La Chartreuse de Parme, ed. Henri Martineau (Paris, 1950), 526 and 582. The latest English version is The Charterhouse of Parma, trans. Richard Howard (New York, 1999). ²⁸ Music and Ceremony at Notre Dame of Paris, 336.
54
W r i t t e n M u s i c a n d O ra l M u s i c
the melodic formulae of the Vatican Treatise . . . note for note [with melodies] in the Magnus liber organi of Leoninus . . .²⁹ These concordances’, he writes, ‘are too numerous and too lengthy to be mere coincidences, but it is not apparent whether the tradition of the treatise influenced the compositional process of Leoninus or whether it reflects a later distillation of formulae extracted from his Magnus liber organi’. Edward Roesner writes in a recent article: Whatever their origin and their intended function in the Vatican treatise itself, these examples afford the modern analyst a glimpse into how the musicians of Notre Dame would have thought about and reacted to the duplum material in the written documents on which their performances of the organa directly or indirectly depended. Whatever their circumstances of origin and function, the Vatican examples suggest that the singer had a variety of approaches to the individual ‘composed’ melodic gesture or phrase available to him. It could have been expanded, contracted, closed off with a cadence, led seamlessly into the following phrase, or shaped in any of the other sorts of ways recorded in the organum manuscripts. Or, if the material were well enough known, or stable enough for some other reason, it could have been performed ‘straight’—as written, as ‘originally conceived’.³⁰
Here I must enter a brief excursus on the concepts of ‘stock phrases’ and ‘performing from memory’. Whatever conception of ‘memorized’ we have in mind, whether a notion of the note-by-note storage of musical material and its note-bynote retrieval, or the reconstruction of the same material on the basis of salient features and Gestalten, one question is always skipped over, and that is the question raised by Crocker in the essay cited in the introduction: what would have counted as ‘the same’ in the musical culture we are talking about? The question is apropos because there is a tacit premiss of ‘sameness’ between score and performance and between one performance of something and another underlying all this talk of ‘stock phrases’ and ‘memorized’ material. The discussions have proceeded in confidence that medieval musicians held such a concept, and that its criteria are manifest. But it is an unearned confidence that falls back ultimately on the standard for sameness of our own textual music culture. There is no evidence for that from the Middle Ages. On the contrary it seems that for the Middle Ages our criteria for ‘sameness’ as far as music is concerned can be recognized consistently only at the level of what Arom calls the ‘model’. Otherwise we are dependent on the criteria of ritual function and/or language text.³¹ We can be surprised at how different the understanding of ‘sameness’ could be from our own standards as recently as the early twentieth century. I have been studying recordings of Chopin’s music made during the first three decades of the century, and I find that performances of the same pieces, even by the same ²⁹ See Stephen Immel, ‘The Vatican Organum Treatise Re-examined’, EMH 20 (2001), 121–72. ³⁰ ‘Who “Made” the Magnus liber?’, EMH 20 (2001), 227–66. ³¹ This is demonstrated in Ch. 11.
W r i t t e n M u s i c a n d O ra l M u s i c
55
performers, can vary widely—in regard not only to what we call ‘interpretation’ but to the notes played and to the overall formal arrangement of sections. Sometimes we have the testimony of the performer that he is playing the score exactly as written, or proclaiming the performer’s obligation to be altogether faithful to the score. I interpret this, not as hypocrisy, but as evidence of a different notion of what such faithfulness entails. Chopin himself would surprise us; he was capable of dispatching on the same day fair copies of the same piece (same title, opus number, key) to each of his publishers, copies that we would say do not entirely agree with one another. Yet Chopin was a notoriously fastidious man (see Ch. 12). Hiley supposes that the many ‘revisions’ of organa dupla in the Notre Dame sources are preferred versions in a script culture. Still, he says, the repeated use of figures favoured by each source is the product of an oral improvisation practice. On the other hand, he insists, ‘improvisation in the domain of three- and fourvoiced Parisian organa is completely ruled out’.³² Wright does not rule it out, citing rubrics in twelfth- and thirteenth-century chant manuscripts for Notre Dame that call for organum by three and four singers. Incidentally, this evidence suggests that we might wish to reconsider the certainty with which we have regarded the famous edict of Bishop Odo of Sully in 1198 ordering the singing of the gradual for the Feast of the Circumcision in two, three-, or four-part organum as a sort of commission for Perotin’s four-part setting that is in the principal manuscripts of Notre Dame polyphony.³³ Of course that risks the loss of one of the few dates on which we have pinned the lifespan of this composer. Sarah Fuller, in her New Oxford History of Music chapter ‘Early Polyphony’, cites two of the Vatican rules and observes ‘The writers of such instructions no longer teach how to generate entire phrases or pieces. Instead they enumerate a collection of stock formulas which appear in vacuo, with no guidance on how they might appropriately be fitted together in composition.’³⁴ This is true enough, but it is hard to imagine how verbal instructions for fitting together the stock formulae might have been formulated, especially in a music culture where rote learning by imitation was the norm. And perhaps it would not have suited the author’s musical ³² ‘Improvisation’, MGG ², Sachteil iv. 553. ³³ Wright, Music and Ceremony, prints a translation of the entire edict on p. 239. On p. 289 he makes the usual assumption linking Odo’s order, the four-part Viderunt omnes attributed to Perotin by Anonymous IV, and the four-part setting of that text that is notated in the principal manuscripts, despite his own documentation of the normal practice of extempore organum singing—in as many as four parts—in that environment. If he is right about the latter, then there is no particular reason to presume that just the text that is in those manuscripts corresponds to the performance that would have been made in response to Odo’s instruction. To presume that amounts to another projection of the relations of the modern musical economy onto a medieval practice, another failure to follow the evidence to the representation of a music culture that worked differently from ours so far as the relation of scores to performances are concerned. ³⁴ The New Oxford History of Music, ii: The Early Middle Ages to 1300, ed. Richard Crocker and David Hiley (2nd edn., Oxford, 1990), 485–556 at 524.
56
W r i t t e n M u s i c a n d O ra l M u s i c
temper to specify just one way of fitting them together. The difference that seems to stand out is one between the emphases on ‘extempore singing’ and ‘composition’. Fuller consistently writes of ‘composition’ throughout her chapter and I believe she is right to do so, just as McKinnon is right to speak of the composition of chant before the age of writing. Claude Palisca, in his discussion of the Vatican Treatise in his New Grove article ‘Theory, Theorists’, does not hesitate to write that what the examples exemplify are ‘improvised melismas’.³⁵ Hiley, comparing the organum purum of the Magnus liber organi with the examples in the Vatican, writes ‘It is imaginable that both forms of polyphony were improvised in that one set in standard figures that keep recurring in these repertories and that undoubtedly come from a reservoir of well-practised turns.’ He then cites Coussemaker’s thirteenth-century Anonymous II treatise, whose subject matter, in the author’s words, is ‘componere et proferre discantum ex improviso’.³⁶ In this, which Hiley counts as one of the earliest appearances of the ‘improvisation’ concept in the medieval literature, we can find a hint about how to take in what seems like a bewildering array of different opinions about the state of musicmaking in the twelfth and thirteenth centuries that emerges from this attention to the Vatican Organum Treatise. The Anonymous II author does not seem to regard composing and improvising as alternatives, and he does not regard composing as an act separate from or prior to performing. The hint is that we might do better to follow his example, and that we ought to include performing from memory and extemporizing in this group of concepts that are fully compatible for the time and not separated by sharp boundaries. If we do consider following his example, we may well reconsider whether he was talking about what we call ‘improvisation’ at all. Then the apparent differences among the modern authors whom I have been citing could melt away. They just show us struggling to measure and describe the musical practices of a foreign culture with the measuring instruments and descriptive language of our own. The disagreements are symptoms of the fact that these do not fit those practices very well. I need to attend to one more issue that has been thrown up in this discussion and that I have so far skipped over. It is raised by Wright’s assertion that the ‘concordances’ between the Vatican Organum Treatise and the Magnus liber organi are too numerous and too long to be ‘coincidental’. I am interested in the question what is meant by ‘coincidence?’ Recently, during the course of a dinner-table conversation with friends and colleagues in Paris, it emerged that two of us have the same Personal Identification Number attached to our bank accounts—in Paris and New York, respectively, and my Parisian friend confided that a third colleague who lives in Cambridge, ³⁵ xviii. 748. ³⁶ MGG², Sachteil iv. 550. The ‘Anonymous II’ text is in Scriptores de musica medii aevi, ed. Coussemaker, i. 311.
W r i t t e n M u s i c a n d O ra l M u s i c
57
Massachusetts, also has that number. What an astonishing coincidence, we said. But the number is the shelf number of a manuscript in a large national library. So, not as much of a coincidence as it would be if I shared that number with my dentist and my automobile repairman. But still, there are many libraries with many music manuscripts in their holdings. Still, again, all of us have shared an interest in that one in particular. What I am doing here is trying to gauge just how much of a ‘coincidence’ this curious phenomenon is. We say ‘coincidence’ with two emphases: two or more things happening at the same time or in the same way, and happening at the same time or in the same way without any apparent cause, that is by chance or accident. German has two different words, gleichzeitig and zufällig. Obviously Wright intended the second emphasis. The two sources could not have come to have the same formulae by chance, and there is a cause: the formulae in one are copies—direct or indirect—of the formulae in the other. ‘Coincidence’ suggests randomness, absence of control, lack of planning, accident. The fear of that is a shadow lurking over talk about ‘improvisation’. Its flavour can be sensed in language that has been used to stigmatize suggestions about oral tradition in discussions of the transmission of chant: ‘improvisatory maneuver’, ‘improvisatory license’, ‘the freedom and vagaries of improvisatory flights’, lacking ‘predetermination and coordination’.³⁷ These characterizations resonate with a definition of ‘Improvisieren’ that we find, for example, in Gassner’s Universal-Lexikon der Tonkunst (Stuttgart, 1849) as ‘das sogenannte freie phantasieren’, which allows itself, ‘in wahrhaft ungebundener Freiheit, die mannigfachsten poetischen Licenzen’. Corollary to this attitude is the expectation that the products of an improvisation will differ from one instance to another. This is implicit in Frisius’ remark in the MGG article that the unpredictability of improvised music is a consequence of the uniqueness of the conditions of the first sounding. But that follows only on the supposition that the improvisation is unconstrained, in other words, that it is random (think once again of Stendhal’s remarks). On this point it is worth noting that Simha Arom showed with respect to the oral traditions of West African polyphony that even the exact replication of a performance is not only a possible but a normal consequence of proceeding under a clear set of constraints, if the constraints are sufficiently dense. That is for the analyst to determine, as I have tried to do in Chapter 1. Addressing the question about sameness and difference in the transmission of medieval epic poetry the literary scholar Franz Bäuml distinguished between two kinds of constants through transmission (what Max Haas calls ‘Überlieferungstreue’ in his MGG² article, ‘St. Martial’—we might translate ‘transmission constant’³⁸): ‘stofflich’ and ‘wörtlich’. Wörtlich can be translated, for music, as ³⁷ Levy, Gregorian Chant and the Carolingians, ch. 6. ³⁸ viii. 843.
58
W r i t t e n M u s i c a n d O ra l M u s i c
notengetreu, note-for-note identity. Constants that are stofflich correspond to Arom’s models. Bäuml also called them Schablone, templates.³⁹ Many comparisons have been published of versions of medieval compositions that reveal the sameness at the level of the stoffliche and difference at the level of the wörtliche. For example, Roesner, in the article cited earlier, compares four clausulae and speaks of ‘points of tenor/duplum congruence’ yielding the same ‘contrapuntal skeleton, different workings out of the same melodic-harmonic strategy . . . different solutions’ to the same musical task. The generation of music on the basis of such stylization principles—whether at the moment of performance or before, whether in writing or in performing— played at least as large a role in the Middle Ages as direct replication and deserves at least as prominent a place in our repertory of interpretative concepts. The ghost of randomness scared up by the thought of improvisation has been a red herring. ‘It seems most appropriate to reserve the term improvisation for cultures and repertories in which a distinction from non-improvised and precomposed forms can be recognized.’⁴⁰ This aperçu of the ethnomusicologist Bruno Nettl, which is worth reiterating in the context of this chapter, amounts to an implicit recognition that the very concept of improvisation, as it is anchored in modern languages, is a product of cultures that have valorized its opposite—composition—as a norm, even as a higher form. It is also an implicit recognition that no culture is likely to thrive alone on caprice in the making of music. We cannot identify such a binary system in the music of the Middle Ages, and the famous passage cited from the Anonymous II treatise shows that its author would not have recognized such a differentiation. Recall the title of the treatise: ‘How to compose and bring forth (perform) discant ex improviso’. All Latin dictionaries and glossaries understand the forms improvisus, de improviso, and ex improviso in the sense of ‘sudden’, ‘unexpected’. There is no verb in Latin that would correspond to our word ‘improvise’. The title of the treatise is not an early appearance of the ‘improvisation’ concept at all. Such a word seems to have appeared first in Italian in 1547, improvvisare, to speak, compose, or write ‘all’improvviso, senza preparazione’.⁴¹ In French improviser, with the sense of ‘chanter sans préparation’, is known since 1660.⁴² De improviso appears in Castilian in 1570, with the same meaning as the Latin phrase (suddenly, unexpected). Improvisar is found in nineteenth-century sources as a derivative of the French improviser.⁴³ The Oxford English Dictionary reports the use of Improviso ³⁹ ‘Der Übergang mündlicher zur artes-bestimmten Literatur des MA’, in N. Voorwinden and M. de Haan (eds.), Oral Poetry: Das Problem der Mündlichkeit mittelalterlicher Dichtung (Darmstadt, 1979), 238–50. Haas carried this distinction into his book Mündliche Überlieferung und altrömischer Choral, 27. ⁴⁰ ‘Improvisation’, New Harvard Dictionary of Music, 392. ⁴¹ Dizionario etimologico della lingua italiana, ed. Manlio Cortelazzo and Paolo Zolli (Bologna, 1979). ⁴² Dictionnaire historique de la langue française, ed. Alain Ray (Paris, 1998). ⁴³ Diccionario crítico etimológico de la lengua castellana, ed. J. Corominas (Berne, 1954).
W r i t t e n M u s i c a n d O ra l M u s i c
59
in English in 1786, with the meaning ‘improvised, extempore’. The earliest appearance of the verb ‘to improvise’—‘to sing or say extempore, to compose (verse, music, etc.) on the spur of the moment (impromptu)’—is reported for 1826. The Etymologisches Wörterbuch des Deutschen reports Improvisieren as an eighteenthcentury term with the sense of ‘Aus dem Stegreif, ohne Vorbereitung gestalten, abgeleitet von italienisch improvviso’.⁴⁴ This association with the Italian is writ large in nineteenth-century music dictionaries, where ‘improvisation’ is commonly identified as an Italian practice. It is important to see not only that medieval Latin used no word corresponding to our ‘improvise’ but that it knew no antonym for componere. There is no evidence of a European conception corresponding to our conception of ‘improvisation’ as opposed to ‘composition’ before the sixteenth century. This is not to say that there were not medieval practices that we would call ‘improvisation’, but then obviously we must take care not to burden our representation of and attitudes towards such practices with the properties connoted by that word when it came into usage. Even when it entered modern European languages, it was not as an antonym for the descendants of componere. It took the invention of a conception of composition that entailed the ideas of deep structure and unity and monumentality and fixity and lasting durability—a conception elaborated by Eduard Hanslick in the middle of the nineteenth century⁴⁵ and by numerous theorists of music since then—to create at the same time a place for a conception of improvisation as its opposite. Perhaps we can see an early expression of this conception in Nicolaus Listenius’ distinction of 1537 between ‘practical music’ (performance) and ‘poetic music’ [from poeiesis] (‘which consists in making or constructing, that is in such labor that even after the artificer is dead, leaves a perfect and absolute work’).⁴⁶ And ⁴⁴ Etymologisches Wörterbuch des Deutschen, ed. Wolfgang Pfeiffer (Berlin, 1998). The pursuit of the phrase ‘aus dem Stegreif ’ leads down the most fascinating and peculiar byway in the metaphorical language of this subject. ‘Stegreif ’ occurs in modern German only in that phrase, never by itself. Native speakers who read, hear, and speak the word freely rarely have any idea about its meaning, except in the company of its partners in the phrase. It derives from ‘Stegareif ’, known since the 11th c. A clue to its specific meaning is its Old English cognate ‘stigrap’, which evolved as the modern ‘stirrup’. Another clue is the current Swedish ‘stiga’ for ‘climb’ and ‘steg’ for ‘step’. In the 17th c. ‘Stegreif ’ gave way to ‘Steigbügel’, which remains in use; both share the descriptive meaning of a frame or rim used for stepping or climbing up. But the former, which no longer ever occurs alone, survives in the phrase ‘aus dem Stegreif ’, which came into use in the 18th c. for a statement or proclamation made from the stirrups, that is without dismounting from one’s horse, as it were on the spur of the moment. (The virtual identity of metaphorical connotation of these two phrases with equine association is apparently pure coincidence. ‘On the spur of the moment’ is an extension of ‘on the spur’, at full speed or with the utmost haste. The coincidence does, however, point to the prominence of the horse in the lives of people creating this language.) ⁴⁵ Eduard Hanslick, Vom Musikalisch-Schönen: Ein Beitrag zur Revision der Ästhetik der Tonkunst (Leipzig, 1854) ⁴⁶ Musica: ab authore denuo recognita multisque novis regulis et exemplis adaucta (1549) (Rochester, 1954). The translation is by Lydia Goehr, in The Imaginary Museum of Musical Works (Oxford, 1992), 116. Goehr disputes the interpretation by several writers of Listenius’ phrase opus perfectum et absolutum to ‘indicate the presence of the work concept’. She could have disarmed this claim by observing that it is based, at least in part, on the readiness to read perfectum et absolutum in the sense of their seeming modern German or Eng-
60
W r i t t e n M u s i c a n d O ra l M u s i c
perhaps that distinction has an earlier counterpart in Johannes Tinctoris’ distinction between counterpoint (mainly realized in the act of performance), and what he called res facta (from chose faite)—polyphonic music that, because all voices were coordinated with one another, had to be (pre)composed and, presumably, written.⁴⁷ A third pair of terms that seems to correspond somewhat to these two, known from the same period, is sortisatio, ‘the improvised joining of various melodies to some chant’, and composition, ‘the premeditated combining of melodies related to each other but not necessarily with any reference to a cantus firmus’.⁴⁸ Clearly this system of meanings as a whole is foreign to the Middle Ages. But it is invoked, whether we intend it or not, when we speak of improvisation in medieval music, and even, because of the habit of linking the oral–written duality with the improvisation–composition antimony, when we speak of oral tradition in medieval music, and that situation gives rise to disagreements that are, to put it bluntly, absurd. Language is like musical notation in this regard, that both have reference to, and derive their meaning from, an inhabited world of practices and relations and language usage. Lifting terms such as ‘improvisation’ and ‘composition’ from the world of nineteenth-century music practice and aesthetic beliefs in order to describe through them medieval practices was bound to produce the kinds of seeming disagreements we have seen, which may not be disagreements at all. Nettl’s suggestion turns out to be correct from the point of view of the history of language, as do McKinnon’s and Fuller’s consistent use of the word ‘composition’ for the practices they describe. If we speak of ‘improvisation’ in medieval music practice, as I do in Chapter 1, let it be with reference to those aspects of the practice that made composition ex improviso possible and that are visible in the products that have come down in writing, not with reference to some sort of opposite of composition in a binary system. In an essay of 1992 that adumbrates The Advent Project James McKinnon presented material that allows us to see the difference and its importance for historical interpretation.⁴⁹ The essay gives us a glimpse of a substantial act of composition in the eighth century, before the age of music writing. Its product is the set of ten Gregorian communions for Advent and Christmas Day (McKinnon eventually extended his analysis beyond this group but I limit myself to it for purposes of this review). Clear signs of what he called ‘compositional planning’ are evident in the texts of the communions, drawn from the Prophets or the Psalter, lish cognates—perfect and absolute—when what both words (in their context) say is ‘complete’ or ‘finished’. There is no disputing that Listenius was making a clear ontological distinction between performance as a process of making, which is ongoing, and what is left behind—in written form, he suggests— when the process is finished, even after the maker is dead. To be sure, this is not the whole of the work concept that was constructed during the 18th–19th cc. ⁴⁷ Bonnie J. Blackburn, ‘Res facta’, MGG ². ⁴⁸ Howard M. Brown, ‘Sortisatio’, NG. ⁴⁹ ‘The Eighth-Century Frankish-Roman Communion’, JAMS 45 (1992), 179–227.
61
W r i t t e n M u s i c a n d O ra l M u s i c Ex. 2.1 The communion antiphon In splendoribus from the Graduale triplex
In
an
splendo
te
ri
lu
bus
ci
sanc to
fe
rum
rum,*
ex
ge
u
te
nu
ro
i
te.
which ‘have a carefully worked out thematic relationship to the season’—for example, on the eve of Christmas Isaiah predicts ‘All flesh shall see the salvation of the Lord’; on Christmas Day David announces that ‘All the ends of the earth have seen the salvation of our God’. The Advent-to-Christmas Day group, McKinnon demonstrated, forms a compositional unit, a feature that, he suggested, extends to the music itself. When it came to choosing melodies for this group of texts the composer(s), in seven of the ten cases, called on one of the strong Gestalten or tune models in the Gregorian tradition. As McKinnon described them, they are ‘short and lovely chants of great tonal coherence that center about the interval d–f and that make, at some point in their brief existence, an intensifying gesture toward upper c'’ (see Ex. 2.1). The seven melodies came to be assigned to modes 1 or 6, but other melodies of the type, of other genres of the mass proper, were assigned to modes 2, 4, and 5 as well. In the case of longer texts that were set to melodies of this family the move to C may be put off for quite a while, but it comes eventually. Melodies of this family may easily be identified by looking first in the books of the Vatican Edition at chants with a C clef on the top line of the staff. Not all will be melodies of this type, but many will be. What I report as the use of a single melody-type, or model, for seven of the ten chants of this category, is a characteristic phenomenon in chant tradition that I describe in Chapter 6 as ‘thrift’, and as that word suggests, it served a practical purpose in the oral tradition. McKinnon interpreted it in the sense of an aesthetic intention that ‘enhance[s] the impression that the Advent-Christmas day group forms a compositional unit’. He attributed the melodies on the grounds of this unity to a single composer and on the grounds of their beauty to a masterful one who, he speculated, would most likely have been a member of the Roman schola cantorum. Referring to the focus of the melodies on the tones d and f he dubbed that composer ‘The Master of the Re-Fa Advent Lyrics’. The intention of this type of appellation—borrowed from the language of art history⁵⁰—is to identify a creator of individual works that share an original, consistent, and unique style that ⁵⁰ I have recently seen, for example, paintings in the National Gallery of Art in Washington, DC attributed to the ‘Master of the Life of St. John the Baptist’, whose activity is traced back to Rimini in the second quarter of the 14th c.
62
W r i t t e n M u s i c a n d O ra l M u s i c
can be recognized by connoisseurs despite the artist’s anonymity. This presumes a degree of individuality in the work and a level of connoisseurship that allow experts to say ‘this is a work of that master, that is not’. Such precise attributions in the case of Gregorian communion chants are out of the question, not because we lack the connoisseurship but because model melodies of such simple and traditional types lack sufficiently differentiating features. Individual differences among them can, for the most part, be understood as accommodations to the different syntactical and phonetic properties of their language texts. (Helmut Hucke has attempted hypothetical accounts of compositional choices made in the case of longer, more complex chants of the same family such as graduals, where texts are not alone sufficient to account for differences. See ‘Toward a New Historical View of Gregorian Chant’.) The implication of a modern sense of ‘composition’ underlying McKinnon’s identification of a composer for the communions is made explicit in his conclusion that ‘the communion temporale . . . is a remarkable creation of epic proportions, like some great symphony’. Another complication about McKinnon’s account of those chants as ‘compositions’ in this sense is raised by their temporal and cultural distance from his source for the melodies that are the objects of his encomium. He transcribed his specimen chant from the securely diastematic (alphabetic) notation of the eleventhcentury manuscript Montpellier H.159, copied in Dijon in northern France, and from that he intends us to apprehend the artistic qualities of what one composer in eighth-century Rome achieved. Of course we always have to contend with this kind of insecurity when we try to reach back behind the veil of obscurity that is lifted only partly with the invention of music-writing systems and fully with the normalized use of diastematic notations (here Crocker’s reservations come to mind again). But especially because of McKinnon’s emphasis on the individuality of this composer’s work we have to wonder whether that is audible through the noise of three centuries of—mainly oral—transmission. Max Haas has suggested a way of viewing this difference of conceptions of medieval composition and its products through attention to an aspect of the language-like nature of chant idioms. Recall Noam Chomsky’s classic formulation of what it is that linguistics aims to understand about language: how it is that speakers of a language can utter sentences they have never spoken before and understand sentences they have never heard before. Haas writes ‘A piece of chant has virtually “more” repertoire in itself than, for example, a piano sonata of Beethoven. Whoever occupies herself long enough with a chant corpus seems to be able quite simply to transform a Latin text that has never been used for a gradual into a pretended gradual. But someone who has assiduously studied the thirty-two Beethoven piano sonatas will still not be able to write the thirty-third sonata.’⁵¹ ⁵¹ Mündliche Überlieferung und altrömischer Choral, 178.
W r i t t e n M u s i c a n d O ra l M u s i c
63
The remark should not be simply dismissed with reference to the greater complexity of Beethoven’s piano sonatas. The point is that each member of a chant family embodies the generative principles and surface features of the whole family, something that Beethoven surely avoided in writing his piano sonatas. This bears comparison with Simha Arom’s aperçu that the written work is maximal, the model that structures a traditional practice is minimal. It seems appropriate at this point to ask whether medieval usage seems to lean towards either of these interpretations. An answer is at hand in Markus Bandur’s article ‘Compositio’ in the Handwörterbuch der musikalischen Terminologie.⁵² It is that medieval usage would not suggest either of these interpretations. The history of the compositio concept runs quite contrary to the implications of ‘The Master of the Re-Fa Lyrics’ for the eighth century, while it is neutral vis-à-vis the interpretation from the viewpoint of melodic tradition. Compositio and componere, as well as their relatives conficio and commodo and even simply facere, are verbal commonplaces in discussions of the creation of monody and organum (or discant) in the music literature of the Middle Ages from the beginning (the ninth century), and Bandur observes that their usage shows them to be continually understandable by virtue of their lexical effectiveness, hence not needing definition. Their reference is always to an act of putting together or making a melody or adapting a melody to a given text or an organal voice to a given cantus (as in Ad organum faciendum), or to the property of having been put together (being composite). There is not the least suggestion that this act necessarily entails writing, a linkage, writes Bandur, that arose only out of the definition of the composer as ‘musicus poeticus’ in the seventeenth century, became definitive only with the differentiation of composition and performance in the early eighteenth century, and has become so habitual in our time that ‘writing’ and ‘composing’ are treated as synonyms. Writers of the Middle Ages make no distinction between written and oral composition, and no differentiation between the act of composing and its product as a work. Bandur comments that, in view of the independence of the ‘composition’ concept from writing in the Middle Ages, it was well suited to practices that he calls the improvisational production of new melodies. I believe he meant nothing more than the oral production of melodies, not the full modern sense of ‘improvisation’. But he is perfectly right to subvert the modern notion of an opposition between composition and improvisation in the Middle Ages. What I have itemized here are associations and oppositions that have accrued around the concept of composition more or less during the sixteenth to the eighteenth centuries and that have been conveyed into our conceptual language by nineteenth-century writers on music theory and aesthetics.
⁵² Ed. Hans Heinrich Eggebrecht (Wiesbaden, 1972– ).
64
W r i t t e n M u s i c a n d O ra l M u s i c
The question about reading an eleventh-century notation as a token of an eighthcentury compositional act hints at a different kind of spectre that haunts these deliberations and that needs to be taken more seriously. It is our habit of thinking that in looking at a notational image we really look through it at the music behind it as though through a window. And that amounts to a belief in an identity relationship between notational image and music, two autonomous media that are normally linked through correspondence and reference, sometimes through resemblance or isomorphism, but not ever really through identity. This habit is strong enough to resist contradiction by phenomena of music writing throughout history—for example Chopin’s autographs and annotations, mentioned above and discussed at length in Chapter 12, which give the impression of a composer with definite mental images of his music, struggling to find adequate notational expressions for them—perhaps because it seems to be reinforced by the practice of other composers, for example Beethoven and Schoenberg. When it is projected onto the music of the Middle Ages it can be a source of confusion, disagreement, and irresolvable paradox. Nino Pirrotta and Susan Rankin have unintentionally illustrated the paradoxical fix in which the scholar can land in interpreting medieval notations. Pirrotta, writing about a polyphonic fragment at Foligno (Italy), declared ‘The clash of G against F . . . is a joy to my musicologist’s eye, if not to my ear, as evidence that the writing down of the piece had nothing to do with the requirements of its performance.’⁵³ Rankin, likewise writing about a specimen of early notated Italian polyphony, asserted that ‘the music as notated would not make sense in a performance’.⁵⁴ Both these comments were intended to undermine the assumption of the equivalence of text and sound in those particular cases, but if they are to succeed in that they have to be read with just the assumption they aim to subvert. For it is only when the notational signs are read ‘literally’ that they give evidence that they probably were not read ‘literally’. These cases are extreme because taken literally they would record musical nonsense that can be easily rejected. But medieval music manuscripts abound in cases that present the paradox in acute but more subtle form: multiple transmissions of tropes, versus, polyphonic compositions, displaying so much variation as to raise the suspicion that singers would not have rendered a composition from a given source each time in the same way (once again the question about the meaning of ‘same’). But in order to entertain that suspicion the scholar must presume that the source corresponds to just one reading, not to a range of possible readings—he or she must read it ‘literally’. It is an irresolvable paradox that would have delighted Italo Calvino, whose very apropos musings ⁵³ ‘Church Polyphony a propos of a New Fragment at Foligno’, in Music and Culture in Italy, 124. ⁵⁴ ‘Between Oral and Written: Twelfth-Century Italian Sources of Polyphony’, in Giulio Cattin and F. Alberto Gallo (eds.), Un millennio di polifonia liturgica tra oralità e scrittura (Quaderni di Musica e storia, 3; 2002), 75–98.
W r i t t e n M u s i c a n d O ra l M u s i c
65
about the written and the unwritten in Se una notte d’inverno un viaggiatore are cited by Rankin (I quote from the English translation): At times I think of the subject matter of the book to be written as of something that already exists: thoughts already thought, dialogue already spoken, stories already happened, places and settings seen; the book should be simply the equivalent of the unwritten world translated into writing. At other times [and here is what we have not sufficiently allowed about music writing] on the contrary, I seem to understand that between the book to be written and things that already exist there can only be a kind of complementary relationship: the book should be the written counterpart of the unwritten world; its subject should be what does not exist and cannot exist except when written, but whose absence is obscurely felt by that which exists, in its own incompleteness. I see that one way or another I keep circling around the idea of an interdependence between the unwritten world and the book I should write. This is why writing presents itself to me as an operation of such weight that I remain crushed by it.⁵⁵
The paradox is circumvented if we allow ourselves to think of notations as being in a sense opaque rather than transparent, autonomous rather than subordinate to the sounding music even while corresponding to it and referring to it in the very formation of its signs, equal in status with it, subject to its own values and aspirations, and performing more and different functions than just the prescriptive one vis-à-vis performance. That impression is given especially by notations that are drawn with great care and precision and with an abundance of detail, properties that we might be tempted to think were aimed at giving the singer the most specific and detailed prescriptions for their performance, if we did not take into account what we know from indications that singers were trained by rote from childhood and were expected to sing from memory long after the advent of music writing.⁵⁶ Those indications make it difficult to imagine that singers would have been dependent in practising their traditions on such precise notations, in the way that today’s singers would be dependent for their performance of Luigi Nono’s Il canto sospeso on his highly detailed and precise notation. Such precise notations of the Middle Ages would be the language of ‘written music’ that would answer the demand for completeness and correctness and exemplary status, as much so as voiced music. Out of such reflection an analogy suggests itself in which written music was to sounding music as written language was to sounding language, in the interpretation of that relationship during the Carolingian era that has been presented by David Ganz.⁵⁷ Ganz writes freely of ‘written language’, noting that ‘classical grammarians had subordinated written language to speech’,
⁵⁵ If on a Winter’s Night a Traveler . . . (New York, 1981), 171. ⁵⁶ An exemplary report on such a notation is given by Wulf Arlt in Stiftsbibliothek Sankt Gallen: Codices 484 & 381, ed. Wulf Arlt and Susan Rankin (Zürich, 1996), i, Kommentar, ⌺ als Notator, 59–77. ⁵⁷ ‘The Preconditions for Caroline Minuscule’, Viator, 18 (1987), 23–43, esp. 40–2.
66
W r i t t e n M u s i c a n d O ra l M u s i c
citing Augustine’s De dialectica: ‘every word is a sound, for when it is written it is not a word but the sign of a word . . . For written letters indicate to the eyes something other than themselves and indicate to the mind utterances beyond themselves.’ ‘But Isidore and Alcuin reject this teaching’, writes Ganz. ‘Isidore is not concerned with letters as a conversation with those who are absent’, as Aristotle had put it. ‘Alcuin is concerned with individual letters as a way to help the reader to read correctly.’ In the light of this shifting of the hierarchy, Ganz concludes, ‘the written norm could alter the structure of the spoken norm, and because of its higher value written language might have a specific, and exemplary, status’. Although the relationships of writing and singing cannot have been exactly like those of writing and speaking, there is no doubt that music writing began in circumstances conditioned by the insistence of imperial authority on the importance of written texts that must be complete, correct, and legible. At the very least, then, this characterization of the situation of language writing is food for thought about the situation of music writing. The title of Edward Roesner’s essay ‘Who “Made” the Magnus liber?’ is exactly consonant with medieval usage. Through the study of the music in its transmission the essay brings us closer to an understanding of twelfth–thirteenth century compositional and performance practice in the conceptions of that time than we have been before. With the word ‘made’ Roesner cites the famous account of Anonymous IV, ‘Master Leoninus . . . so it has been said . . . made the great book of organum on the Gradual and Antiphonary to enrich the Divine Service.’⁵⁸ Roesner’s emphasis of this word recognizes its interchangeability in medieval usage with ‘composed’, and that is a clue for us to disengage medieval references to ‘composing’ from the sense of this word in modern times, especially since the nineteenth century and especially from its use as antonym of ‘improvising’. The title used by Anonymous IV—Magnus liber—would trip off for us the implication of a stable corpus. But in fact the transmission displays a high degree of variation. It also displays passages that recur with virtually no change. This situation suggests to Roesner that ‘the act of writing the music down, of copying it, may have been tantamount to the act of composition itself ’ (p. 234). I make similar observations in Chapter 11 about the transmission of earlier traditions, such as those of the tropes and versus. But at the same time ‘performance traditions played a significant role in shaping the substance of the music in the Magnus liber’ (p. 234). This somewhat tempers the reputation of the Parisian repertory as representing the beginning of polyphonic composition in the modern sense. Thus the ‘relationship between the ⁵⁸ Translation by Roesner, ‘Who “Made” the Magnus liber?’, 227, from Fritz Reckow, Der Musiktraktat des Anonymus 4, i: Edition, ed. Fritz Reckow (Wiesbaden, 1967), 46.
W r i t t e n M u s i c a n d O ra l M u s i c
67
composer, the work, and the performer, between text and transmission, may have been a less linear one—or a less compartmentalized one—than we are often wont to suppose’ (p. 234). ‘Compartmentalized’ is important. These words mark roles that were not vested in different individuals, as they are now, when the combination of roles in one person is noted as something special. The written transmission shows notators determining details of a musical text at the moment of writing. Contrary to the assumption—commonplace for us— that it is the music that controls or dictates the notation, here the notation controls the music as well. In that sense writing down is a counterpart of performing. So Roesner writes ‘Some of the variations in these organa may be . . . the work of a copyist with a particular notion of how the music should go, or perhaps the product of the imposition of a certain manuscript “house style” ’. As in the early writing down of chant, there is much more to the relationship between notation and music in the Middle Ages than that of equivalence. Roesner concludes: ‘In the kind of creative milieu I am hypothesizing, composition and improvisation are intermingled. [I wonder whether it is even necessary to say this. Judging from what I have reported here, what we call improvisation was folded into what they called composition. It is the tyranny of the language, the names of separate categories, with which we operate that forces us to draw up a representation of a compound process, put together from strands that may not have been separate in the first place.] Predetermination and ad hoc decision making merge, and there is no clear line separating the written from the oral creating and transmission of musical texts’ (p. 264) . . . ‘Who produced the versions of the magnus liber that we know from the surviving manuscripts? The organista, the singer, the scribe—it would seem that all of them had a hand in it, and that each of them played a role that can be deemed “creative”’ (p. 266). For my own conclusion I can only indicate by way of reiteration the direction in which it seems to me these reflections send us. If ‘improvisation’ with its modern overtones is not properly medieval, if instead medieval music was produced through methods of composition that do not differentiate ‘written’ and ‘oral’ music—or writing and oral process—which do not differentiate methods of composition or characteristics of the product, then we must also try out a different conception of music writing, as a contingent, not a constituent aspect of the music-making complex, an aspect with some autonomy, with its own aims, values, purposes, prestige, reciprocating with its own influence on music, not as something ancillary and merely utilitarian. The Calvino paradox already suggests this.
68
Th e Vatic an Organ um Treatise
chapter 3
The Vatican Organum Treatise and the Organum of Notre Dame of Paris: Perspectives on the Development of a Literate Music Culture in Europe
} It is the attitude throughout this book that if we wish to understand its subject as nearly as possible from the vantage point of its own time, in its own life, we must approach ‘medieval song’ as practices rather than as repertories of closed objects, as processes rather than their frozen products. It is the making of songs—whether composing, improvising, writing, but in any case singing—that is the focus, not the structure of what has come down to us in writing. Singing is at the centre, and that is what we have tacitly taken for granted or simply ignored in our studies. I aim to redress that neglect in some part with this chapter, which takes as its principal subject a pedagogical handbook addressed to singers. Reading with that in mind will force our attention to singing, and singing of a very spectacular kind that must have been widespread. The chapter was written originally as a lecture delivered at the Schola Cantorum Basiliensis in Basle for an audience of students of the performance of early music. I have been privileged to be associated with performers through most of the years during which I pursued the studies whose reports are collected in this book, and those associations have been most influential in the formation of my own understanding of my subject. That has been especially so with the singers of medieval song who are represented on the accompanying CD.
} The tradition of music pedagogy in the European Middle Ages reaches back to the Carolingian era. It began in an atmosphere that was sparked by the educa-
Th e Vatic an Organ um Treatise
69
tional initiatives of Charlemagne’s court and was driven by a true writing rage. Pedagogical books were written about the most diverse subjects—from cookery books to music treatises. The earliest music-pedagogical books that have come down to us from the Middle Ages are the Musica disciplina by Aurelian of Réôme¹ and the anonymous Musica enchiriadis.² Both were written around the middle of the ninth century, and both aimed to comprehend an active current practice—the tradition of liturgical chant performance—in a set of norms. As model for the pedagogical structure of the Ars musica that was being newly created the writers elected the ancient Ars grammatica. From that model they borrowed the format in which each aspect of the practice was to be explained, codified, and legitimated through pedagogical principles and citation of phenomena that fall under universal laws. They also borrowed from the Ars grammatica the idea that the pedagogical principles were to be illustrated with examples drawn from the practice being taught. This created a role in the ars musica for a music script, and indeed these two treatises are among the oldest documents of medieval music writing. There are also several neumations of texts among the oldest documents that stand more or less closely to the liturgy. My focus here for the moment will be on the role of notation in the pedagogical books, in comparison with its role in the practical sources. The most striking difference lies in the fact that the notations in the pedagogical books were conceived primarily for the designation of pitch height, while those in the oldest practical sources served as indications of aspects of the elocution of the words: of voice quality (quilisma, gutturalis, tristropha), of the sounding out of the syllables and the consequences for what could be sung on them (liquescent and non-liquescent neumes; a liquescent neume, for example, would have signalled ‘do not sing a melisma on this syllable’), and of the matching of the syllables to melodic turns. The exact fixing of the pitches was not the main task of the oldest neumations in the practical sources. This difference makes sense if we consider that the melodic traditions of the ninth century were known through oral transmission, while the pedagogical treatises were preoccupied with matters that depended on the distinctions among notes and the systems in which they functioned. Aurelian took on the task of differentiating chant melodies according to their tonal types. He proceeded from complete melodies, and took it for granted that the singers to whom he addressed his treatise had the repertory by heart—indeed ¹ Aurelian of Réôme, Musica disciplina, ed. Lawrence Gushee (Corpus scriptorum de musica, 21; Rome, 1975). For the latest investigation of the date and provenance of the treatise see Barbara Haggh, ‘Traktat “Musica Disciplina” Aureliana Reomensis: Proweniencja i Datowanie’, Muzyka, 2 (2000), 25–78. ² Musica et Scolica enchiriadis, una cum aliquibus tractatulis adiunctis. Recensio nova post Gerbertinam altera ad fidem omnium manuscriptorum, ed. Hans Schmid (Bayerische Akademie der Wissenschaften, Veröffentlichungen der musikhistorischen Kommission, 3; Munich, 1981).
70
Th e Vatic an Organ um Treatise
he made clear that he regarded that as a criterion of professional competence. When he wanted to indicate something about a particular melodic passage he could do so by citing the words that it intoned; there are just a few isolated examples in the treatise with neumes. The author of the Musica enchiriadis, on the other hand, faced the task of building up a tone system. For that he provided many illustrations that he called descriptiones (something like ‘diagrams’), using a notation that exactly specified the positions of individual notes in the diatonic series. These descriptiones served not only as examples to illustrate his verbal descriptions in visible form. The author also used them as arguments, in that he made his explanations dependent on the descriptiones, rather in the manner of diagrams in scientific treatises. The neumes in the earliest practical sources are in no way to be regarded as binding prescriptions. They could not be, for in so far as they were intended to direct singers—something not at all to be taken for granted³—they could serve an oral performance tradition only as guidelines. Nor should the greater precision of its pitch information persuade us that the notation of the Musica enchiriadis was a prescriptive notation for performers. And while its type can be found in some practical sources, they are relatively few and early. Thinking of musical notation as a medium for the instruction of performers—something that is second nature to us—seems on the whole to have been foreign to musicians of the ninth century and not among the original motives for the invention of notations. Only around the end of the century do we sense implicit clues to the formation of such an idea. Hucbald of St-Amand, in his treatise De harmonica institutione, complained about the inadequacy of ‘the signs which custom has handed down to us and which in various regions are given no less various shapes’ (the neumes) for the transmission of melodies.⁴ He suggested as remedy the introduction of an alphabetic notation in combination with the neumes, which would make possible the precise indication of interval distance while retaining the information about performance nuances and the matching of note groups with syllables encoded in the neumes. Although this suggestion did not lead to a widespread use of letter notations, the principle underlying them was carried forward with the writing of notes on staff lines. In that the staff has an unambiguous representational character vis-à-vis interval distance, thereby making letter notations redundant, it made it possible to ³ Indication of a different kind of use is found, for example, in the 11th-c. Casus monestarii sancti galli by Ekkehard IV of St. Gallen, who reports on a copy of the ‘authentic Roman antiphoner’ that was placed on an altar in the church of the monastery. Whenever anything ‘in the chant disagreed [with it] all errors of this type were corrected, as in a mirror.’ The implication seems to be that the book was kept as a check on what was sung, and if there were disagreements the singing would be corrected with the book as standard. But it seems fairly clear that the singing was not done from the book. (Monumenta Germaniae historica, Scriptores rerum sangallensium: Annales, chronica et historiae aevi carolini, ed. Georg Heinrich Perz (1829), ch. 47, ll. 13–14.) ⁴ Martin Gerbert, Scriptores ecclesiastici de musica sacra potissimum, i (St. Blasien, 1784), 104–52. Translated as Melodic Instruction in Hucbald, Guido, and John, 3–46.
Th e Vatic an Organ um Treatise
71
combine the information conveyed by both types of notation—letters and neumes—in a single system. This development enabled young singers to learn the chants from a book without the help of a teacher, something that Hucbald foresaw, and about which a music pedagogue of the eleventh century known today only as ‘Pseudo-Odo’ boasted.⁵ That this conception had already been fully formed in some cases can be seen in practical sources of the eleventh century. In music-pedagogical writing we first encounter this prescriptive character of notation in the Vatican Organum Treatise, and that is what marks the historical significance of that pedagogical text.⁶ It is the first of a new type of music treatise, and at the same time musical notation is given a new function. The notation in the treatise is prescriptive, but in a special sense. It is precisely readable—that is also the case with the notation of the Musica enchiriadis—–but the reader is meant to internalize from it models for proceeding without notation. That is consistent with the arrangement of the treatise. In contrast to the Musica disciplina and Musica enchiriadis, it does not transmit a descriptive or explanatory doctrine: it gives instruction for practising singers. The musical examples, however, are more than illustrations of the verbal instructions. Their information about the practice far exceeds the information given in the rules. The notation itself transmits a prescriptive doctrine. Since the practical examples—again in contrast to those of the ninth-century treatises—do not have the character of instantiating a fixed practice but are new embodiments of principles, some of which are not even explained, the treatise is to be viewed as a kind of composition manual: the examples are presented as models for composition. The Vatican Organum Treatise transmits the earliest composition pedagogy of the Western tradition known to us. The objection could be raised against this assertion that the treatise teaches improvisation, not composition. Such an objection would rest on a false dichotomy and an erroneous viewpoint: that improvisation and composition are mutually exclusive practices and that the former is associated with oral, the latter with written expression. But when it comes to medieval music we have no reliable criteria for differentiating between pieces that were orally conceived and later written down and those that were immediately written down as fixed composition. Such distinctions further overlook the fact that for a long time ‘written composition’ in many genres and traditions was a matter of writing down music that was conceived according to the principles of oral tradition. Medieval music culture remained long after the introduction of musical notation an essentially oral culture. It makes far more sense to differentiate ‘oral’ and ‘written’ composition in the ⁵ ‘With the passage of only a few . . . days they were singing at first sight, extempore, and without a mistake any written music’; Strunk’s Source Readings, 199. ⁶ Godt and Rivera, ‘The Vatican Organum Treatise’.
72
Th e Vatic an Organ um Treatise
Middle Ages than ‘improvisation’ and ‘composition’. (See the discussions of this general subject in Chs. 1 and 2.) The Vatican Organum Treatise comes down to us in just one source written on five leaves in the first quarter of the thirteenth century. These are currently bound into the manuscript Vatican Ottob. lat. 3025, as folios 46–50. The question of the treatise’s integration into this manuscript is discussed in the edition by Godt and Rivera but has little bearing on the interpretations that I offer here. On palaeographic grounds a northern French or English provenance has been suggested.⁷ Zaminer, in his commentary, called attention to the similarities of style between the organa of the Vatican treatise and the early Parisian organum, and in interpreting them he initiated continuing efforts to link the two genetically (see Ch. 2). But this follows an old historiographic habit by which stylistic similarities are unreflectively taken as evidence of linear descent.⁸ Under a less constrained type of interpretation, the Vatican treatise teaches a widespread vernacular art that is a counterpart of the tradition out of which the Parisian composers refined a highly specialized and elite compositional style. If we suppose that singers would have memorized the examples as models, that would account both for the impression that we see in them a widespread practice, and for the resemblance of many of them to passages in Notre Dame organum that was noted in Chapter 2, without our having to posit a direct genetic connection. The text of the treatise begins with a definition of organum that focuses on performance. I cite just the beginning of the definition: ‘Organum is a song (cantus) that follows a preceding song, for the cantor ought to precede and the organizator ought to follow and the cantor should finish first. For an organum by itself is worth nothing unless some song goes together with it.’⁹ This passage leads to a treatment of consonances and a lengthy discussion about problems encountered in the singing of B and B(, and upon that follows the main business of the treatise, the enunciation of thirty-one rules for interval progressions. These follow a uniform format, which I exemplify with the eighteenth rule: if the chant ascends by a second and the organum begins at the octave, let the organum descend by a seventh, and it will arrive at the unison, as follows.¹⁰ There follow examples of the rule in which the distance between the first and last notes in the organum is indeed a seventh, and in which the two voices indeed begin at the octave and end at the unison; but these provide only the framework for what proves to be a rather extravagant coloratura; see Ex. 3.1. ⁷ Frieder Zaminer, Der Vatikanische Organum-Traktat (Münchner Veröffentlichungen zur Musikgeschichte, 2; Tutzing, 1959), 22–4. ⁸ ‘The task of music history is the investigation and the setting forth of the developmental paths of music’; Guido Adler, Methode der Musikgeschichte (Leipzig, 1919), 9. ⁹ Godt and Rivera, ‘The Vatican Organum Treatise, 293. For the passage, see above, Ch. 1 n. 51. ¹⁰ Ibid. 299.
73
Th e Vatic an Organ um Treatise
Ex. 3.1. Six examples illustrating the eighteenth rule from the Vatican Organum Treatise (a)
8
fa
re.
fa
re.
fa
re.
(b)
8
(c)
8
(d )
8
fa
re.
sol
mi.
sol
mi.
(e)
8
(f)
8
The treatise contains 343 examples, of which 251 directly exemplify the rules and ninety-two constitute an addendum with no direct connection to the rules. It concludes with three full liturgical organum compositions, an alleluia and two
74
Th e Vatic an Organ um Treatise
responds. (Fig. 3.1 below is a facsimile of the first of these, Ex. 3.7 a transcription, and a performance is on track 8 of the CD.) It is a practical manual that is oriented entirely to the generative principles of ad hoc organum compositions, and that draws only to a very limited extent on theoretical explanations. In the examples the intervals of the upper voice are connected through passing notes and through diminutions that progress in their own characteristic melodic patterns. In the organum compositions this technique is used to a much wider extent. Thus the teaching about interval progressions becomes a teaching about melismatic organum. But about that teaching the rules have absolutely nothing to say. To repeat, the examples are not descriptiones in the sense of the Musica enchiriadis. The expression ‘melismatic organum’ identifies the formal class of two-voiced elaborations of medieval chant. One element of that class that must have been of great importance in that time, and to which the Vatican Organum Treatise draws attention, is not given expression in that phrase. I now focus on that element, first in the domain of monophonic chant. In recitation passages of the Old Roman tradition the words are commonly sung not on a single note, as in the Frankish tradition, but on a repeated flourish of two or three notes; see Ex. 3.2. This exemplifies a differentiating quality of the two traditions of Gregorian chant that has often been described: the Roman tradition manifests an ornamental melismatic style, whereas the Frankish tradition is characterized by a relatively tight melos and a clearly arranged and discernible form. The priority of ornament in the Roman tradition is to be understood not only as a taste for a decorative style; it also manifests a performance-involved impetus— a valorization of singing itself and a virtuoso capacity for it. This gives substance to the anecdote about Frankish singers sent to Rome by Charlemagne to learn the Roman manner of singing, who were then declared by Romans to be too barbaric, with voices too rough, to be able to master that style (see Ch. 6). If we compare the teaching of the Vatican treatise, as illustrated in the examples and especially in the organum compositions, with the Parisian organum, then the aspect of a virtuoso singer’s art stands in the foreground. Particular features of both suggest that the composers were singers. The examples and the compositions evoke an impression similar to that made by the examples in Giulio Caccini’s Le nuove musiche.¹¹ This can be highlighted with a few examples. First a counterpart of the recitation style in the Roman transmission of Gregorian chant (see Ex. 3.3). As Caccini has been mentioned, I show some Caccini-like properties in the treatise. Often a note can be written from four to eight times in immediate succession, something like Caccini’s trillo. This can occur in two sorts of contexts: 1. As an intonation at the beginning of a piece or section. Such intonations begin either directly on the repeated note (Ex. 3.4(a)) or on the note just below ¹¹ Ready access in Strunk’s Source Readings, 607–16.
75
Th e Vatic an Organ um Treatise Ex. 3.2.
Beginning of an introit antiphon from Vatican Vat. lat. 5319, fo. 53r *
8
O
cu
li me
i
sem
per
ad
do
mi
num
Ex. 3.3. Three passages from the organum compositions in the Vatican Organum Treatise, fo. 50r (a)
8
(b)
8
(c)
8
Ex. 3.4. Five passages from the organum compositions in the Vatican Organum Treatise: (a) fo. 50r; (b) fo. 49v; (c) fo. 50v; (d ) fo. 49v (a)
8
(b)
8
(c)
8
(d )
8
(e)
8
76
Th e Vatic an Organ um Treatise
Ex. 3.5. Two passages from the organum compositions in the Vatican Organum Treatise: (a) fo. 49r; (b) fo. 49v (a)
8
(b)
8
that to be repeated (Ex. 3.4(b)). Once an intonation begins on the note above (Ex. 3.4(c)). After the note repetition the melody is led into a melismatic flourish that either returns to the initial reciting note (Ex. 3.4(d )) or aims towards a different note (Ex. 3.4(e)). 2. At the bottom or the top of a melismatic arch (Exs. 3.5(a–b)). As drawn out as intonation phrases may be, cadential phrases can be far more drawn out. They are striking not only in their length, but also in their range, their large leaps, and the cascading of the voice. The reader is urged to try singing the cadential passages in Ex. 3.6, and then to imagine the effect of a virtuoso singer performing them with complete mastery. The ornamental melos displayed here, which must evoke the aural image of great virtuosity in performance, represents a very widespread vernacular style throughout Europe in this age. It is far more extravagant than what we find in the Parisian organa—especially the later ones. Zaminer, in his edition, made a very interesting observation about the notation in the treatise. He counted 110 different notational signs used to represent only thirty different melodic figures. That means that every melodic figure is represented by an average of 3.66 different signs. The same thirty melodic figures occur in the Parisian organa, but they are represented by far fewer signs. The correspondence between script and music is noticeably closer in the Parisian organa than in the organum of the Vatican treatise. The author (or notator) of the treatise wrote in each case one of several possible signs in order to make visible the compositional possibilities that he aimed to teach. The reader was to assimilate the principles embodied in the examples and then to forget their concrete written exemplifications. In contrast, the writers of the Parisian sources, especially the later ones, used a script that was logical and consistent, because the written score represented the key for the performance. The concept ‘prescription’ must therefore be understood as weighted in two different ways: on the one hand as point of departure or jumping-off point for a performance whose contents the singer shapes in some measure, on the other as detailed directions for the performance. Notation serves in different ways and degrees as guideline for the performance.
Th e Vatic an Organ um Treatise
77
Ex. 3.6. Four passages from the organum compositions in the Vatican Organum Treatise: (a) and (b) fo. 49v; (c) and (d ) fo. 49r (a)
8
(b)
8
(c)
8
8
(d )
8
8
We know that the Parisian organum manuscripts transmit late and in some sense edited versions. For the history of written practice they are of particular interest because they transmit not only the last, most refined phase of the Notre Dame tradition, but also something of its transformation up to the time of the writing down, with respect to elements of musical style as well as to the notation. As for the musical style, the latest versions, in contrast to the Vatican organa, display a kind of economy or efficiency, but one might also say squareness, of form (avoiding the implication of ‘progress’; see the introduction to Ch. 5). At the same time there is a sharpening and greater consistency in the notation of the later sources. That is not a coincidence; the two phenomena seem to be manifestations of the same mentality, a desire for economy and orderliness. It has been tempting to think that the formal symmetry and clarity of the later style and the interest in consistency of transmission require a more precise and consistent notation. But such symmetries could have been in support of memory in an oral tradition. The possibility should also be considered that the causal relation could have run in the opposite direction, that the notation could have influenced compositional practice, especially as the domain of notation came to take a central
78
Th e Vatic an Organ um Treatise
place in the music theory of the time. In any case, the changes undergone in notational techniques should be seen in the context of the altered role that notation was to play, a change that resulted from the mutations in the constellation of relations among composition, writing, and performance. The Vatican Organum Treatise offers two important insights on the topic of this chapter. The first concerns the domain of non-literate and literate medieval music tradition. Attention to this domain is an indispensable condition for understanding medieval music history. Right alongside that, however, the passion for singing, of vocalizing, must be taken into account as another dimension. For it is just this dimension that prohibits us from postulating an all-too-orthogenic development from scriptless to scripted practice. The second insight concerns the transformation to which the phenomenon of visible exemplification in music pedagogy was subject: in that we have observed this transformation and interrogated it, two central growth processes within medieval music history have presented themselves to us: the coming-to-be of scriptuality and the rise of a pre-scriptive composition.
Th e Vatic an Organ um Treatise
Fig. 3.1. Alleluia Hic Martinus from the Vatican Organum Treatise, fo. 49r
79
80
Th e Vatic an Organ um Treatise
Ex. 3.7. Transcription of alleluia Hic Martinus from the Vatican Organum Treatise, fo. 49r
8
Al
8
8
le
8
8
8
lu
ya.
81
Th e Vatic an Organ um Treatise Ex. 3.7. cont.
8
Hic
8
Mar
8
[ti ]
8
8
8
nus pau
per
et
82
Th e Vatic an Organ um Treatise
Ex. 3.7. cont.
8
mo
di
8
8
cus
di
ce
ves
lum
in
gre
8
8
8
di
tur
him
83
Th e Vatic an Organ um Treatise Ex. 3.7. cont.
8
8
nis
8
ce
le
8
8
8
sti
bus
84
‘ Per ipheral’ and ‘Central’
chapter 4
‘Peripheral’ and ‘Central’
} The original version of this chapter was written for a symposium with a task of a kind that was not then commonly undertaken in the discipline of music-historical studies. It was to conduct a critical examination, on the grounds of musical analysis, of the full historiographic system through which the material that has come down from a particular period had been evaluated, categorized, and ordered—in this case the music of the twelfth to the fourteenth century. The other participants were Rudolf Flotzinger, Lawrence Gushee, Max Haas, David G. Hughes, Michel Huglo, and Fritz Reckow. Organizer and moderator of the Symposium was Wulf Arlt. In his thoughtful and provocative introduction to the published proceedings of the symposium Arlt explained his choice for discussion of the concept-pair ‘Peripherie/Zentrum’ as the linchpin of the system through which the whole of Western music composed and performed during those three centuries had been given an historical accounting. The challenge to music historians uncovering sources from that period during the early decades of the twentieth century came from their discovery of material that was fragmentary, crude or primitive in appearance, backward-looking, along with music that they could evaluate as sophisticated and advanced: how could it all be folded into an orderly historical account, especially one that would be faithful to the a priori commitment to the task that had been set for music history by Guido Adler, to ‘investigate and mark out the developmental paths of music’? It could be done by separating the material into the categories ‘central’ and ‘peripheral’, which reinforced the implication brought out in the reception of Adler’s dictum that a main line of development is, by definition, one that passes through centres. This historiographic housekeeping implicates, one after another, a whole network of interdependent conceptions: the dependence of the music conception and music theory on the distinction of quadrivial and trivial arts; the very concept of artes and its opposite number in another duality, usus, that is, the duality of theory and practice which has, in turn, been linked habitually with ‘composition’ and ‘improvisation’—there is a bias towards regarding a place where music is ‘composed’ as a centre but not a place where music is ‘improvised’ (this suggests another reason why the idea of improvised chant has seemed so intoler-
‘ Per ipheral’ and ‘Central’
85
able: it threatens to consign chant—the originary European music—to peripheral status). Arlt cites the alternative conception of Jacques Handschin, who wondered whether this ‘peripheral’ style should not be regarded as the real foundation of polyphonic composition in the twelfth and thirteenth centuries rather than as some degenerated provincial practice, and whether the Notre Dame style, by contrast, does not appear in such a light as an ‘excessive special development’.¹ Nino Pirrotta expressed a similar interpretation more directly: the kind of polyphony often called ‘archaic’ or ‘peripheral’ (although found in relatively recent and central sources) belongs to the normal practice of polyphony in most churches, large and small, of the Western world. On the other hand, the kind of artistic polyphony we have become used to considering standard for fourteenth- and fifteenth-century sacred music is but the valuable expression of a minority, the advancing standard of a special elite.² Such style designations as ‘archaic’ and ‘peripheral’ and their opposites ‘advanced’ and ‘central’, which are themselves central to this general historical picture, arose in the interest of preserving an overall conception of the history of art as a continuous progression to higher stages of development. As assessments of chronological and geographical placement they are often more virtual than real. And this raises the question whether the view of the Western or European music culture as being characterized by an always active drive to the new, which feeds this historical picture, is one that was held by medieval musicians themselves.³
} Any time we wish to prepare a medieval piece for performance or study we encounter the problem of establishing a musical text, a problem that is raised by variation in the transmission of musical items from one source to another. Usually those problems have been discussed in peripheral compartments of our publications: footnotes, prefaces, appendices, etc. My purpose here is to move the issues entailed into the centre, and to ask ‘What does it mean that the editorial problem exists?’ To consider this question requires that we go behind methods of text criticism, to the conceptions on which the idea of text criticism in the domain of medieval music has been based: that the pieces with which we deal are the products of an act of composition (in the modern sense) or redaction; that it is the understanding of ¹ Citations on pp. 23–4 n. 2 of Arlt’s introduction to the Roundtable ‘Peripherie und Zentrum in der Musik des Hohen Mittelalters’, Bericht über den Internationalen Musikwissenschaftlichen Kongress Berlin 1974 (Kassel, 1980). ² ‘Church Polyphony apropos of a New Fragment at Foligno’, in Music and Culture in Italy, 124–5. ³ See the literature cited by Arlt on p. 24 nn. 7–9 of his introduction.
86
‘ Per ipheral’ and ‘Central’
composers that they should normally remain fixed; and that variants arise either through revision or through error (corruption). These conceptions refer to a certain relation between musical production and transmission that has prevailed for a long time in the Western tradition. But the assumption that it was established with the beginning of musical notation is supported only by habit, not by the evidence of the sources. What is not self-evident is that composers produced unique, fully determined works, that musical notation uniquely represented these, that musical sources were from the first copied from other written sources, and that they served from the first as prescriptions for performance. That is, we cannot say that the processes of production and transmission became separate from the moment that music first came to be written down. The suggestion to which I have been leading is that phenomena that have been perceived mainly as problems for editorial technique can be placed in the centre of attention as evidence bearing on the nature and relations of musical production and transmission and on the ontology of music in the Middle Ages. With reference to a few examples, I wish to try first to rearticulate the idea of ‘transmission’. What ways do we know for describing and explaining the family resemblances and the differences between the versions of a piece? Most commonly we try to relate versions to one another either directly, through the concept of ‘variant’, or indirectly through the concept of ‘archetype’. I have learnt a different way of approaching the task in which the focus is shifted for the time being away from the search for archetypes and for the direct genetic relations between versions, to the study of each version as a new construction of the piece based on the same underlying model; in so far as we are speaking of the same piece I would speak, then, of its repeated reconstruction. Now this is not meant to cast away for ever the concept of ‘archetype’ or to deny the possibility of direct collateral relationships. Nor is it meant to propose the radical position that all variants must be regarded as equally valid. It is proposed as a shift of focus, as a deliberate research strategy in order to set up a different framework that has, I think, its own advantages. This shift brings into view an altered construction of the tasks involved in the study of transmissions: identification of the forms and schemata underlying the transmission of genres, families, and individual pieces; study of the relationships among such forms and schemata; investigation of the extent and manner of interdependence among schemata in the melodic, rhythmic, harmonic, and textual domains; differentiation of the levels within repertories and individual pieces on which these schemata operate; study of the particular reconstruction, that is, search for the factors that enter into it and that determine family resemblances and individual differences. The image here is that of the performer, or the notator, not reproducing the piece in performance exactly from score or from memory, not copying it directly
‘ Per ipheral’ and ‘Central’
87
from a written antecedent, but reconstructing it each time on the basis of a model, and under the influence of local and individual principles guiding the reconstruction. I think of such models also as archetypes, but in a sense different from the hypothetical archetype of a manuscript stemma. My suggestion is not that this idea uniquely describes the medieval musical situation, but rather that an understanding of the history of musical production and transmission in the Middle Ages requires this image as well as the traditional one. These ideas can be concretized with reference to a number of specimen problems in the study of transmission. They are presented in section II of this chapter (Exs. 4.1–3) and should be consulted in parallel with this general discussion. In the analysis of a transmission it is useful to distinguish among the following layers. The concept of ‘layers’ is to be understood strictly as a construct; other layers might be identified, and above all there is a continuity between layers, as the examples show. 1. The données of a musical system—for example, the diatonic system; modal conventions, either those associated with psalmody or those associated with the octave species; melodic formations based on third-chains, etc. 2. The particular formal-tonal characteristics of a melodic family, for example, of the alleluias of the ‘Dies sanctificatus’ family, or the graduals of the ‘Justus ut palma’ family. 3. The concrete formulae at the surface that are transmitted along with underlying schemata, yet that are short of fully determining the musical surface. 4. The surface of the piece as it is represented in each notated version. The study of a transmission in these terms is a matter of identifying the levels at which the piece remains invariable through transmission, and of identifying the peculiarities of each reconstruction and the determinants of those peculiarities. This approach opens out to a reconsideration of the role of notation in relation to performance. For us the role of notation is primarily prescriptive: instructions from the composer to the performer. But at the same time a notated score is for us a signification of the piece, spelt out in as many details and parameters as are covered by the notational system. For medieval music we cannot take that for granted. What are the alternatives? Perhaps the most important is that the notation provides a system of signals or cues that do not directly instruct the performer in detail, but that put him on course in his own process of decision making about the performance. In that sense all notation is indexical (see Ch. 13). Adiastematic neumes are of that sort, as is modal rhythmic notation.⁴ Does this mean that once pitch notations indicated intervals precisely they left no decisions about pitch to the performer? The degree of pitch variation in the written transmission of many repertories makes that unlikely. Of particular interest in this respect is the ⁴ See my essay ‘Regarding Meter and Rhythm’.
88
‘ Per ipheral’ and ‘Central’
comparison of adiastematic and diastematic notations of the same repertories, where the degree of variation is significantly greater with diastematic notation. This suggests that the notator is now making decisions that had been left before to the performer—decisions that adiastematic notations could not record. It puts the notator in the position of performer. From another perspective the notation of a musical item can have the status of a protocol of a performance of it. The notator has heard it (or indeed has asked for it to be performed), and he translates it into writing. This way of putting it brings out the possibility that the act of writing down does not necessarily have as its immediate goal a performance. What other objectives can there have been? Perhaps to preserve the piece, to record its existence and symbolize knowledge and possession of it. We must admit that the purposes and meanings of early notations, and the modes of their signification, are mysteries that we are far from having solved. It is more than a question of how to transcribe them. But whatever the reasons for writing down a piece, once a notator has undertaken to do so it must be written out all the way. There were no notations corresponding to our analytical abstractions of the essential features of a piece. The only way to record or show or transmit a piece was to perform it or write out a concrete version of it. We have no idea how the notator or the performer would have responded to the sorts of variants we find. It is for us to determine in particular cases at what level the notation means to be precise, from the point of view of our analytical and editorial purposes; to determine the notation’s level of signification in the particular musical context in which it is employed. This way of thinking about musical production and transmission cuts across the categories ‘oral’ and ‘written’ transmission. Reconstruction as the primary way of musical production is not something that came to an end with the introduction of notation. It was several centuries after that event before the relations of transmission were so altered as to make this characterization inappropriate. I am inclined to downgrade the significance of the oral–written dichotomy as a primary style discriminator, and I would regard the introduction of music writing as a necessary but not a sufficient cause of fundamental changes of style. Using the framework that I have sketched here, I wish to try a characterization of a major alteration in the relations of musical transmission in the high Middle Ages. The central fact is the separation of production and transmission. Its major symptom is the investment in the notated source of a significantly larger share of control over the performance. In terms of my earlier formulation about the role of notation, I would put it that the signification of the notation is at the level of the musical surface. That means that decisions about how the piece is to go are not left to the performer, but are communicated to him in detail through the notation. This remark points in two directions: (1) what it means in terms of the history of style that this should have happened, and (2) changes in the requirements for notation in consequence of these stylistic matters.
‘ Per ipheral’ and ‘Central’
89
The point with respect to the history of style is that notation became prescriptive when the stylistic circumstances demanded that pieces be the same from one performance to another. Three aspects of the stylistic situation are of primary importance from this point of view: 1. The rise of a rhythmic conception—by the late twelfth century in practice, by the late thirteenth century in theory—in which the basic organizing principle is the perfection, a periodic division of time in which the beginning of each period is accented and the movement of music is reckoned with respect to that position, in which rhythm is created by the grouping of notes with respect to the perfections, and in which form is a matter of the architectonic, subalternating organization of rhythmic groups and/or perfections. The perfection becomes the common currency or standard measure for the coordination of all durations and all metrical units. This is, of course, the system of modal rhythm, and it amounts to the establishment of our modern system of rhythm and metre. 2. The potent idea that the alternation of consonance and dissonance should be controlled by this articulation of time—the idea that dissonance resolution should occur normally on the downbeat. That turned the control of harmony from the realization of a sound ideal to an instrument of musical form, and it opened the avenue of association between rhythm and voice-leading. 3. Entering into this development is a powerful tendency to periodic and subalternating melodic phrase structure, which is expressed in polyphony through all sorts of symmetrical relations between voices. This relates to the prevalence of accentual, rhymed verse, and to the clarification of a harmonic modal system as a system of octave species. We really do not know enough to say what the causal dynamics are here. What we do know is that all these factors are fully operative by the time of the transmission of the Parisian Magnus liber (see Ex. 4.4 below and its discussion). In these circumstances duration cannot any longer be something like a contingent ornament in the province of the performer’s art, as was suggested by the old rules for duration and consonance in organum. These were clearly addressed to performers, whereas the rule for duration and consonance in discant that was promulgated by Franco of Cologne was clearly addressed to composers.⁵ That is, duration was built into the relationship between the voices, along with consonance. It was not any longer contingent. Then it must become the objective for notation to give unequivocal signification of both pitch and duration. That requirement was fully realized in Franconian notation, but the ideal for such a situation is visible in thirteenth-century writing on mensurable music in general, and it is reflected in the pre-Franconian modal sources of the thirteenth century. ⁵ Franco of Cologne, Ars cantus mensurabilis; English version in Strunk’s Source Readings, 226–44.
90
‘ Per ipheral’ and ‘Central’
How does this rather long anacrusis prepare us for thinking about the conceptual dichotomy, ‘periphery–centre?’ Where production and transmission are both matters of reconstruction and essentially continuous, it is hard to see how there can be a distinction between centres and peripheries on any other basis than that of the level of activity and the interest in collecting. Centres are places where there is a high level of cultural and/or liturgical activity, peripheries places where activity is less dense. But it is harder in such circumstances to see a distinction in the sense of centres of production and distribution as against receivers and consumers. For the structure of production and transmission is such that every place is equipped for production, and music is produced locally for local consumption. On the other hand, with the alteration in the relations of production and transmission that I have characterized the ‘centre–periphery’ duality becomes descriptively plausible as a measure of the musical geography. And it leads to the expectation of a transmission that is, in principle, stable. That expectation can be measured against particular cases. One of the great open questions regarding the production and transmission of music in the high Middle Ages is the question about the status of St-Martial de Limoges as a centre. The question is open because the old idea of a ‘school’ of St-Martial as antecedent to the School of Notre Dame has lost its plausibility. We have come far enough in our study of the sources to understand that St-Martial was more a collecting than a producing centre. But I have the impression that our main accommodation to such evidence has been to replace the expression ‘St-Martial’ with the expression ‘Aquitanian’, and then to continue talking as before. The question that remains open is, how does this decentralization affect our understanding of the nature of Aquitanian musical production and transmission, of the historical position of Aquitanian music, and of the musical map of Europe from the tenth to the twelfth centuries? I propose that we address these questions by focusing on the repertories of tropes of the tenth and eleventh centuries, and the repertories of polyphony of the eleventh and twelfth centuries. For both repertories I propose an approach from the viewpoints that I have developed so far, with particular attention to these questions: How coherent are the collections of sources and the repertories? What does the study of ‘variants’ show about the relation between production and transmission? How far are the processes of production and transmission separable in our understanding of the music? How does the transmission within Aquitanian families of sources compare with transmission between Aquitanian and nonAquitanian sources? That is to say, how far can we talk about an export business, and on the other hand, how far can we recognize in different places similar principles of style as basis for the local production of music? This is to ask, how specific are the styles in question? Finally, in view of the nature of musical production and transmission in these repertories, what is the role that notation plays in them?
‘ Per ipheral’ and ‘Central’
91
Regarding the trope repertories, there are these questions: 1. Evans,⁶ Hughes,⁷ and Weiss,⁸ grouping the Aquitanian tropers of the tenth and eleventh centuries according to their repertories, have produced approximately the same results, while using different criteria: repertory concordances and notation (Evans); repertory concordances alone (Hughes); melodic traditions (Weiss). That suggests confidence in the results, and raises the question of significance in the terms that I have posed here. 2. In that direction, Evans speaks of ‘standard’ repertory, Hughes of a ‘central core’ of the repertory, represented in the most closely knit of the manuscript groups. As that central manuscript group includes all the tropers that were drawn up for use at St-Martial or at least in the area of Limoges, one can speak of a central St-Martial group. But by the same reasoning, St-Martial is not a centre of distribution, for this central manuscript group has a low rate of concordance with manuscripts outside the group. It is rather to be viewed in the context of a wider area of distribution, in which it is distinguished by local growth and by the standardization of local melodic traditions. 3. Hughes asserts that no one of the extant tropers can have served directly as source for any of the others, and he doubts that this can be accounted for simply through the disappearance of all the intermediate sources. Reflecting on this conclusion, he considers whether it suggests that the tropes might have been transmitted orally, but then backs away, because the variants are ‘too few’ for an oral transmission. This is puzzling all around. 4. Only Weiss has based his conclusions on the study of melodic transmission. (That is important because, of course, the comparison of repertories alone, i.e. of the feasts troped and the texts of the tropes, cannot tell us anything about the nature of melody transmission.) His method is based on the phenomenon of alternative melodic settings of the same trope texts. These double transmissions especially pose a challenge to our music-critical methods. The question is, what criteria do we have for determining when and how two melodies are the same? And behind that lies the question, what is our understanding of the nature of melodic transmission? Example 4.2 below is addressed to these questions. Turning now to Aquitanian polyphony of the eleventh and twelfth centuries, we have long understood that the four sources traditionally identified as the sources for St-Martial polyphony do not show a high degree of coherence, either individually or as a group. Now Fuller has shown that they represent what were originally nine sources.⁹ They include just over sixty polyphonic pieces, and nearly ⁶ Paul Evans, The Early Trope Repertory of St. Martial de Limoges (Princeton, 1970). ⁷ David G. Hughes, ‘Further Notes on the Grouping of the Aquitanian Tropers’, JAMS 19 (1966), 3–12. ⁸ Günther Weiss, ‘Zum Problem der Gruppierung südfranzösischer Tropare’, AfMw 21 (1964), 163–71. ⁹ Sarah Fuller, ‘Aquitanian Polyphony of the Eleventh and Twelfth Centuries’ (Ph.D. diss., University of California at Berkeley, 1969).
92
‘ Per ipheral’ and ‘Central’
two-thirds of them are unica. There is no possibility of speaking of a common ancestor or common origin, nor much hope of establishing a filiation. They are a decentralized group, without a core repertory. The concordance situation that looked thin when we counted four sources looks even thinner when we count nine. As with the tropes, this much alone suggests local production for local use, some circulation but not much, and it pushes us to search for other ways of thinking about production and transmission than those to which we have been accustomed. In the polyphony, is there a difference between the direct transmission of an upper-voice melody and the reconstruction of one? To answer we need to learn the principles on which added voices are made in relation to the constraints imposed by the given voice. How far can variant ‘versions’ be understood as different applications of the same principles in the presence of the same given voice, and how far should they be regarded as transformations of some other version? To what extent can we recognize different objectives or different principles for the making of organum as the causes of variant ‘versions’? How do the relationships among Aquitanian versions compare with the relationships between Aquitanian and non-Aquitanian versions? That is, how specific is the style of Aquitanian polyphony? Example 4.3 is addressed to these questions. Example 4.1: troubadour songs by Bernart de Ventadorn (a) Ara’m conseillatz seignor (Appel no. 6)¹⁰ Rhyme scheme: AB Melodic phrases Source G: MN Source R: ST
AB
CC
M'N' OO' ST UU
DD PQ S' T'
Melodic correspondences between the two versions: M-S, N-T, O-U. The variant M-M' in G is an expansion. R is, by contrast, literal about reflecting repetitions in the rhyme scheme through melodic phrase repetitions (ST ST). The same literalness is shown by S'T'. Regarding tonality, G projects a tonal configuration frequently encountered in medieval monody but not strictly accountable in terms of the modal system. We can understand it as a C tonality with close on the cofinal. Despite the close on g the melody is not heard as G plagal because the c is too strongly tonicized to be heard as an accessory. Because of the greater consistency about matching phrases in the version of R, e becomes pivotal and, ultimately, the final. ¹⁰ Karl Appel, Die Singweisen Bernarts von Ventadorn (Halle, 1934).
93
‘ Per ipheral’ and ‘Central’
Ex. 4.1. Two songs by Bernart de Ventadorn: (a) Ara’m conseillatz; (b) Pos pregatz, from Milan R 71 (G) and Paris fr. 22543 (R) (a)
8
Source G M
N
A ra’mcon se
illatz
se
ignor
Source R S 8
8
Ar m’a co sse
lhatz
se
nhors
C’u na don na’m R S
det
U na do na’m
det
8
Mas e R U
s’a
mor
8
sen
Vos c’a vatz va
lor
e
sen
C’ai a ma da
lon
ga
men
lon
ia
men
mic
pri
vat
pri
vat
greus no’m
fo.
greu no’m
fo.
T s’a
mor
C’ay a ma da
O'
ra
Mas e ras
sai
de
ver
tat
Qu’il a al tr’a U
say
per
ver
tat
G P 8
e
N'
G O 8
ber
T
G M' 8
Vos c’a veç sa
Que fay au tr’a
mic
Q
Et anc de null R S'
con
Et anc de nulh
com
pa
gnon
Com pa gna tan T'
pa
nho
Com pa nha tan
94
‘ Per ipheral’ and ‘Central’
Ex. 4.1. cont. (b)
8
Source G M Pos
pre
N gaz
mi,
se
ignor,
Qu’ieu chant, eu
Source R S 8
Pus
mi
pre
iatz,
se
nhor,
Qu’ieu chant, yeu
Qant cuit chan
E
tar,
eu
plor
A
8
rai
co
eu
sai.
Man tas
ves,
q’ieu
o
sai.
Ben chant, qan
mal
li
vai.
T cant cug chan
tar,
plor
Greu vei
O ree
chan
ta
dor,
Greu vei
V retz chan
ta
dor,
G P 8
ta
ra,
R U 8
chan
l’o
G P 8
rai.
O
R S 8
te
T
G M' 8
chan
Vai mi donc mal d’a mor?
Ben chan,
O
can
mal
l’es
Q
Anz meilz qe non fez mai. E donc, per qe
R U
V
A mi del mal d’a mor.
Va mielhs que no fez
m’es mai?
W may. E doncx, per que’mn’es mai?
(b) Pos pregaz mi, seignor (Appel no. 36) Rhyme scheme: AB Melodic phrases Source G: MN Source R: ST
tai.
AB
AB
AB
B
M'O PO ST UV
PO UV
Q W
‘ Per ipheral’ and ‘Central’
95
Melodic correspondences between the two versions: M-S, N-T. Beyond that, the different conceptions of the piece in the two sources take control. M' is an expansion, like G’s setting of the second A line in no. 6. Is O a mistake (a tone too low)? Perhaps, but it can be understood as deliberate. For it takes a new pitch level that is picked up in P and that prepares the recurrence of O in the sixth line. Then the repetition of PO for the seventh and eighth lines closes the circle. It is a more subtle arrangement than that of R but just as coherent and purposeful. The differences in the settings of the ninth line are significant. R consistently projects an F tonality. The endings of U and V surround the final, and W makes the expected, conclusive close. G begins in a g pentachord and slips down to one on f. The repeated sequence P–O makes a clear antecedent–consequent effect, and the ending of Q leaves the whole melody open. That makes sense in view of the text: the difficulty of singing well, the open question at the end. But only the first strophe ends with a question. For the rest, the ending of the R version is more fitting. What can one conclude? In the case of both songs there are correspondences between the sources with respect to overall form and the course of some phrases. Beyond that the versions are quite different. Each is coherent in terms of its own conception, and what is suggested is a certain base sense of how the song goes, and then individual reconstructions informed by that sense. There is a consistency about the differences between R and G: R tends to be more literal, smoother, simpler, more consistent (therefore ‘better’, writes Appel). G seems more subtle, more differentiated, more individualized. Both sources, R and G, were prepared for musical notation, but the staves for substantial numbers of the songs in both remain empty. That is, writing down the music was not a part of the same editorial process as assembling the texts. The sources for them must not have provided all the melodies. Both sources are dated to the fourteenth century, nearly two centuries after Bernart. There are no other sources for either melody. The very idea of an authentic or original version in such a situation seems chimerical. Example 4.2: Trope Dulciter agnicole to the introit Introduxit vos¹¹ The material comprises the trope in the versions of Paris lat. 909 (A) and Apt 17 (B) (Ex. 4.2). This is one of Weiss’s double transmissions, that is, he interprets the two settings of the trope text as different melodies. The discussion is addressed to that question, and to the larger question of the nature of transmission in this genre. The trope text is the same in both versions, and its setting falls in the same places with respect to the phrases of the chant. In both versions trope line I falls in ¹¹ Transcriptions from: Introitus Tropen, i: Das Repertoire der südfranzösischen Tropare des 10. und 11. Jahrhunderts, ed. Günther Weiss (MMMA 3; Kassel, 1970), 332–5.
96
‘ Per ipheral’ and ‘Central’
Ex. 4.2. Trope Dulciter agnicole to the introit Introduxit vos from Paris lat. 909 and Apt 17 I
8
Dul
ci
ter
a
gni co
le
fes
ti
vum du
ci te pa
scha,
gni co
lae
fes
ti
vum du
ci te pa
scha,
* 8
Dul
ci
ter
a
x
8
iam qui
a
8
Iam qui
a
Jor
Jor
da
nis me ru x
is
tis
fon
te
re
da
nis me ru
is
tis
fon
te
re
nas
ci
nas
ci
II
8
Ci vi
bus ae
8
Ci vi
bus ae
the re
the
re
8
Dix
e
rat hoc Mo y
y
8
Dix
e
rat hoc Mo
i
8
8
ma
is ma
nan
tem
man
na
de cho ris,
nan
tes
man
na
de co
ris,
y
III
IV
is
ses
ses
quon dam,
Xpis
tus iu
bet
at
nunc
quon dam,
Xpis
tus iu
bet
at
nunc
z
Ae ther, hu mus, pariter z
pe la
gus
u
ni
ta
vo
ce
de
can tant:
Hae ter, hu mus,
pe la
gus
u
ni
ta
vo
ce
re
sul tent:
* Such figures signify liquescent neumes
‘ Per ipheral’ and ‘Central’
97
two sections (Ia and Ib), with caesura on g. Trope line II corresponds melodically to trope line Ia in both cases, and line III corresponds to Ib. And in both cases line IV very closely corresponds to line III. This is so whether or not there is any detailed melodic similarity between the two settings of the respective lines. The similarities are in the internal correspondences within the two versions. Something like the formal system of this trope is transmitted. There are, however, some correspondences of melodic detail at the surface, indicated by the horizontal brackets x, y, and z. So it cannot be so easily said that these are ‘different’ melodies. If we incorporate its formal system as well as its surface details into our understanding of what the melody is, it seems that it is the same melody and that we are confronted with two different reconstructions of it. The formal and material similarities that we find between the two versions represent the basis of its transmission. The question whether that is an oral or a written transmission does not seem to be to the point at all. Example 4.3: Aquitanian polyphony (Catolicorum concio) The material comprises six organal settings of a single cantus melody: A and B are Aquitanian, CI–III are the three strophes of a setting in the Codex Las Huelgas,¹² and D is a setting from western Switzerland.¹³ Arabic numerals refer to the notes of the cantus. In the main the three versions of C can be understood as one organum presented three times, but there are these differences: I1 is a more elaborate opening for the initial strophe. II1 and III1 begin identically in the first group, but descend to d in independent ways. II and III5–18 are identical, but differ in detail from I15–18. In 16–18 this is a matter of different elaborations of the same consonant notes. Such variants are understandable from the viewpoint of a repeated remaking of the organum on essentially the same principles; their differences do not seem to be otherwise purposive. Comparing D with C, there are identities or close similarities with CI at the opening, and with all strophes of C over 2, 7–8, and 11–14. But these do not force the conclusion of a direct transmission. They can be understood in the light of the independent production of organal settings of the same melody on the basis of similar principles of counterpoint and a single general manner of melodic elaboration. A shows detailed melodic similarities with C and D only over 13, but the manner of melodic elaboration is in general similar in all three settings. We can find similar figures in all the organal voices, but in different places with respect to the ¹² Higini Anglès, El còdex musical de Las Huelgas (Barcelona, 1931). ¹³ Published by Arnold Geering in Die Organa und mehrstimmigen Conductus in den Handschriften des deutschen Sprachgebietes vom 13. bis 16. Jahrhundert (Berne, 1952).
98
‘ Per ipheral’ and ‘Central’
Ex. 4.3. Versus Catolicorum concio from (a) Paris lat. 1139, fo. 44v; (b) London Add. 36881, fo. 13v; (c) Codex Las Huelgas, fo. 22; and (d) Oxford Lat. liturg. d 5 (a) 8
(b) 8
(c) 8
8
I II III
8
(d) 8
cantus variants
8
Cantus: A, B, C 1 8
Ca
2
3
4
5
6
7
8
to
li
co
rum
con
ci
o
cantus. 6–8 and 16–19 are the two cadential passages of the cantus melody. In the organal voices of A, B, and C the notes consonant with the cantus in those passages are c'–b–d', that is, the underlying cadential counterpoint in those versions is the same. That reinforces the notion of a production of organum on similar principles of counterpoint and melodic style, and it also shows that making an organum was a goal-directed process that produced the greatest similarities at cadences. The version of B is throughout closest to an underlying note-againstnote counterpoint, hence the dependence of the organal voice on the cantus is most direct. The versions of A and CI are at the opposite extreme from that, and their organal voices are determined largely independently by melodic principles. Summing up, these versions exemplify a situation in which organum is produced on the basis of underlying principles of style that are similar over a wide geographical area and over a long period of time. It is the cantus and the principles of making organum that are transmitted; the organum is made locally. The example does not suggest a specifically Aquitanian manner. But this case represents one extreme. Certainly the direct transmission of whole organum melodies from one source to another and from one region to another has been demonstrated (for example by Wolf Arlt, in his contribution to the symposium for which this chapter was originally written).
99
‘ Per ipheral’ and ‘Central’ Ex. 4.3. cont. (a) 8
(b) 8
(c) 8
I II
8
III
8
(d) 8
cantus variants
8
Cantus: A, B, C 9 10 8
sum
mo,
11
12
13
14
15
16
17
18
sum
mo,
sum
mo
cum
gau
di
o
Example 4.4: Magnus liber organum duplum: Judea et Iherusalem The example shows the setting of ‘Judea’ in the versions of W1, F, and W2. The fifth stave in each system shows a reduction of the organal voice, designed to bring out its voice leading relationships in terms of the elaboration of the principal consonant notes accompanying the tenor. Those notes are written as half notes. Upward- and downward-directed stems identify the two registers in which the voice moves. Slurs indicate long-range voice-leading connections. Arabic figures identify ordines, lower-case roman figures identify notational groups. The discussion here will concern in detail only the passage 1–5. The passage has a clear overall progression comprising the opening sonority, sustained ad libitum, and three phrases that I take to be eight perfections in length each: 2, 3, and 4–5 (the third phrase comes out to eight perfections on the plausible supposition that 5 would be sustained for two perfections). This exemplifies the association of a clear modal micro-rhythm with a balanced macro-structure of phrases matched in duration. (Reckow has drawn attention to this association in his discussions of copula,¹⁴ but it obtains as well in symmetrical structures of the ¹⁴ Fritz Reckow, Die Copula: Über einige Zusammenhänge zwischen Setzweise, Formbildung, Rhythmus und Vortragsstil in der Mehrstimmigkeit von Notre Dame (Akademie der Wissenschaften und der Literatur. Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse, 13, Jahrgang 1972; Mainz, 1972); id., ‘Das
100
‘ Per ipheral’ and ‘Central’
Ex. 4.4. Notre Dame organum duplum Judea et Iherusalem from W1, W2, and F 1
2
3
W1 8
i
F 8
i
ii
iii
W2 8
Ju
8
4a
4
5
6
W1 8
u
F 8
v
iv
W2 8
(Ju )
de
a
8
7
7a
8
9
W1 8
F 8
W2 8
a
8
a
‘ Per ipheral’ and ‘Central’
101
sort shown here. It is in general a mark of the establishment of modal rhythm and its notation in organum duplum.) The organal voice establishes itself in two registers at the outset, the octave (1) and the fifth (2) of the tenor’s f. Its action through 6 can be understood as double voice-leading in those two registers, elaborating those two notes. The critical fact for this interpretation is that the organal voice moves by step within each of the registers, and moves by skip between them. The essential lines of motion can be summarized thus: in 1–3 the duplum makes a linear descent from f ' to c' in the upper register, and a linear descent from c' to f in the lower register. But in the second descent in 3 the f ' is left by skip, and that keeps it hanging, so to speak, as an upper ceiling that is reclaimed in 8–9. In 4–8 there is a linear return in the upper register from c' to f ', and in the lower register from f to c'. By virtue of these underlying lines the entire passage from 1 to 9 is unified. The movement of the tenor is worked into the total action, rather than the upper voice being simply a series of melismas above the notes of the tenor. How that is done is exemplified by the preparation for the first tenor movement to d. In 4 the two registers of the duplum are collapsed and brought down to f, from whence they make a strong thrust to a above d in 5. But we must focus on the progression of the duplum melody from another point of view. It moves in two- and three-note groups that make the detailed contours of the line, and that are articulated through notational groupings. Of course these correspond to the modal-rhythmic grouping of the line that is expressed through the notation. And the rhythm supports the articulation of the voice-leading. Thus in 2 the accented beginnings of the odd-numbered perfections are always occupied by the c'; that is where the dissonances are most consistently resolved. In 3, where the main descent takes place, the stressed position of the odd numbered perfections is occupied first by f ', then again by f ', then by c', and finally by f. I am proposing, with this interpretation, the following reading of 3, in the more explicit notation of W2:
Rhythm, in this example, is not a surface element imposed from without, but an integrated compositional element. The transmission of this piece through the three sources is uniform. The variants that occur are mainly notational, and they can be understood in the light of a clarification of the essential processes that have been observed to this point. In 3 the extra c' of F and W2 (i) makes for an easy continuation of the first-mode rhythmic pattern. In W2 the binaria (ii) and the breve–long (iii) further clarify that Organum’, in Wulf Arlt, Ernst Lichtenhahn, Hans Oesch, and Max Haas (eds.), Gattungen der Musik in Einzeldarstellungen: Gedenkschrift Leo Schrade (Berne, 1973), 434–96.
102
‘ Per ipheral’ and ‘Central’
pattern. In 4 the breve–long of W2 (iv) makes clear a second-mode rhythmic pattern, and the final note with plica in F and W2 (v) in place of the binaria of W1 suggests a reversal to instead of . That is consistent with the tendency to stress the consonant note. But it may be that it simply makes explicit what a performer would have done in any case. That would be to suggest that the notation of F and W2 is more prescriptive than that of W1. And in turn perhaps that would be a more precise way of putting what I mean in speaking of a ‘more explicit’ modal notation.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
103
chapter 5
On the Structure of Alleluia Melisma: A Western Tendency in Western Chant(?)
} We try to reproduce the reality, but the harder we try, the more we find the pictures that make up the stock-in-trade of the spectacle of history forcing themselves upon us . . . Our concern with history . . . is a concern with preformed images already imprinted on our brains, images at which we keep staring while the truth lies elsewhere, away from it all, somewhere as yet undiscovered. W. G. Sebald, Austerlitz (New York, 2001), 71–2 All the essays that are the originals of the chapters in this book are, in one way or another, efforts at representing a distant reality, the lyric practices of the European early Middle Ages. I include this earliest one because, reconsidered today, it leads straight to the serious reflection that is needed about the terms and conditions, as well as the hazards, under which we attempt any such representation. It does so for me, quite frankly, because of so much in it that seems to me wrong, even offensive, now. I could have chosen simply to bypass it for that reason. But that would be a missed opportunity. The differences with respect to musical configuration that are reported in it still seem to me to be fundamental; it is the configuration of history in which I confidently embedded that report that now strikes me as being quite mistaken. I have taken this as a lesson in the way that our apprehension of the music of the past is coloured by our historiographic preconceptions—historiographic in the broadest sense. It is a lesson that we cannot learn too many times, and that is why I include this badly flawed early essay here, preceded by reflections on the fundamental problem itself. The title of this chapter is the one I gave to the original essay. The question mark has been added as a token of my present questioning. The first hazard to be acknowledged is the chance that under the present-day postmodern creed any claim to represent a reality in a piece of historical writing— a reality that is separate from the writing itself, and which the writing has as its object—will be suspect. I refer to the way that caveats about the limits of objectivity in historical writing, and acknowledgement of its necessarily fictive aspect, can
104
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
nowadays be run all the way to the denial of any historical ‘reality’ at all, to the dismissal of a concept of historical truth as the means to bring such ‘reality’ into view, and to the doctrine that history is through and through linguistic.¹ For example Tzvetan Todorov in 1970: ‘Nowadays we can no longer believe in an external immutable reality, nor in a literature which would merely be the transcription of this reality. Words have gained an autonomy which things have lost.’² I do not share this extreme view, although I have more than once expressed reservations about the objectivist standard, and have urged the acknowledgement and cultivation of the imaginative component of historical writing.³ I do not doubt that in the past real people lived and breathed and acted in intentional and instinctual ways in response to needs, challenges, and impulses that arose in their lives, and that such action included the making of music and musical cultures. And I believe that those lives and intentions and acts—as well as their products—are what we aim to represent in historical writing, no matter how precariously. The worries about our ability to do so leads properly not to the frantic denial that there is anything to represent but to a sensitivity and vigilance about the commitments, interests, obligations, ideologies, and habits—conscious or unconscious, fresh or stale— that influence our representations. The need for that is illustrated most sharply in Chapter 9. Such sensitivity, when it comes to the subject of this book, might well begin with the awareness that the very labels Middle Ages and ‘medieval’ carry with them an idea of distance—psychological, cultural, aesthetic, stylistic—that was embodied in the feelings of the people of the Renaissance who devised them to express a consciousness of their own fundamental difference from a cultural epoch that they regarded as being well behind them. This perception of the Middle Ages as distant and other continues to be manifested directly in the connotations of the word ‘medieval’, which can go to old-fashioned, primitive, irrational, superstitious, cruel, etc. And it is manifested indirectly when it is said that the cultural era that extended until sometime in the mid-twentieth century began with the Renaissance (or that it is the Renaissance that has just ended).⁴ Now that era has been renamed ‘modern’, through a paradoxical reversal that results from the juxtaposition of that word with ‘postmodern’ in yet another historiographic duality. These labels, too, signify a perceived (or desired) rupture from a past that is declared to be fundamentally at odds with present sensibilities. ‘Modern’ has come to mean old-fashioned. The comparison is telling, for in the case of the ‘modern’/‘postmodern’ duality we—at least those of us whose lifetime has spanned the shift—have seen radical transformations in the apprehension or conceptualization of cultural traits and intellectual habits when they are viewed from without after having been experienced from within. Nowadays we may experience such transformations when conversing with another person, as uncertainty about whether we are even talking about the same thing. ¹ ² ³ ⁴
See my essay ‘Postmodern Signs in Recent Musical Studies’, JM 13 (1995), 3–17. ‘The Fantastic in Fiction’, 20th Century Studies, 3 (May 1970), 76–92 at 91–2. In my book, Music and the Historical Imagination. See Leonard B. Meyer, ‘The End of the Renaissance?’, Hudson Review, 16 (1963), 169–86.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
105
Both ‘medieval’ and ‘modern’ (in the transformed sense of old-fashioned) are labels that result from historiographical rather than historical reflection. They impose heavy conditions on historical description in both cases even before it begins, and it is an exquisite irony that the modern era was conceived as a new beginning after the rejected Middle Ages and that it has expired, evidently cast off just like that era. But such one-sided pejoration of the Middle Ages has somewhat retreated when confronted by other cultural streams that were fed by European pride in self, most prominently the Romantic revival of the medieval, the discovery of two medieval Renaissances—those of the Carolingian epoch and of the twelfth century—and the idea that Europe, as historical, cultural, and geographic entity, is a creation of the Middle Ages.⁵ The result has been an ambivalence—one might even say neurosis—in the attitudes of historians of the Middle Ages to their subject: at once as source and as cultural other. Such a confusion of attitudes must influence the kinds of representations that historians make, and on a more general level the ambivalence must play out altogether as a dialectic between representation and reality. This is especially acute in the case of European music history, which has been understood since the Romantic era to begin with the Middle Ages, unlike the history of art and literature, which claim ancient Greek patrimony (see Ch. 9). The Latin chant of the medieval Roman Church—Gregorian chant, specifically—is the oldest European music that we know, and by virtue of that priority it has been granted the presumptive right to be regarded as the progenitor of European music. So, sure enough, scholars have long engaged in the identification of quintessentially European principles in that music, even to the extent of identifying them with principles that dominate the music of Beethoven, the all-time paragon of European music. This works both ways, as I observe in Chapter 9: the definition of what is European is informed by the traits of the chant that has come down as Gregorian, and the traits of European music that have been identified as such since the nineteenth century operate tacitly as teleological ideals for the recognition of those traits in the chant. But this circle has not held against waves of discovery forcing recognition that medieval chant as a whole does not consistently display those traits that have, by broad consent, been held to be Western. A response to this challenge has been a play of appropriation and alienation: identifying music of a cultural self and of a cultural other, sometimes called Western and Oriental, respectively, all within the heritage of Latin chant, and the overall view that the former separated off from the latter and became the foundation of European music.⁶ The chapter that I am introducing here was conceived in the context of such exercises, in the 1960s, in locating the present and the medieval past in the grand ⁵ For example, see Denys Hay, Europe: The Emergence of an Idea (Edinburgh, 1957). ⁶ Another instance in which research was channelled by an a priori historiographic duality—with absurd results—is discussed in Ch. 4. There it concerns the duality ‘periphery and centre’, which begins as a geographic distinction and becomes a distinction of styles, and in the end forces the characterization of Parisian musical practices in the 12th and 13th cc. as central and peripheral styles both located at once in the centre.
106
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
sweep of Western cultural history, something that has itself become a thing of the past (a fact that is looked upon with rejoicing in some quarters, where meta-narratives are scorned, and with nostalgia in others, where the abandonment of large theses and bold interpretations in favour of detailed description is regretted).⁷ And what I now doubt is that, in recruiting the music that I interpreted in the essay for the service of these grand themes, I paid sufficient heed to an admonition that reaches us from a venerable source, and that seems beyond questioning: ‘Every civilization and every period has its own possibilities of aesthetic perfection . . . The works of art of the different peoples and periods, as well as their general forms of life, must be understood as products of variable individual conditions, and have to be judged each by its own development, not by absolute rules of beauty and ugliness.’⁸ The germ of this chapter is embedded in an earlier essay of mine that is not reproduced here, one in which I began the historiographic reflections that are carried further here.⁹ They were provoked by the assault upon musical tradition [that] is said to come from two directions at once: from composers of totally organized music . . . [in which] the relationships are so complex as to elude the senses, and from composers whose aim it is to check the listener’s efforts to hear organized structures—either by deriving the order of musical events from . . . random sources, or by leaving the order for the performer to determine, usually in spur-of-the-moment decisions during performance. Both seem to deny necessities that we have long taken to be essentials of the nature of music.¹⁰ What those essentials come down to, I thought, are the principles of movement towards goals, symmetry, closure, and a superordination of parts, or hierarchical structure. And my aim in the 1964 article was to pursue our notion of musical order into its early history. It was my thesis that these principles could be observed in medieval melody earliest in settings of rhymed accentual Latin poetry, which are non-ritual art songs, and that they reflect there the rhythmic regularities and symmetries of that poetry. Consequently, I speculated, it seems reasonable to suggest that these procedures are not indigenous to the language of the chant, but rather that they entered that language under the influence of the development in view here, of a melodic style that matches the formal properties of the poetry. The chapter here introduced was presented as a kind of negative affirmation of that tentative conclusion, through demonstration of the rarity of such structures in the one chant genre that might be expected to manifest them, the wordless melismas of the alleluia, and of the late date of the few instances that do show them, especially when those result from apparent revisions of earlier versions. Whatever else I may now ⁷ See ‘What Kind of Story is History?’ in my Music and the Historical Imagination. ⁸ Erich Auerbach, ‘Vico and Aesthetic Historicism’, JAAC 8 (1949), 110–18 at 110. ⁹ ‘Musical Syntax in the Middle Ages: Background to an Aesthetic Problem’, Perspectives of New Music, 4 (1965), 75–85. ¹⁰ Ibid. 75–6. See also Meyer, ‘The End of the Renaissance?’
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
107
question about the essay, I believe that conclusion still has merit and constitutes an important fact about the history of Western music. There was a second explicit historiographic purpose: to locate the origin of our notion of musical order, not in the Renaissance, but in the high Middle Ages. In making this argument I was contributing, on the musical side, to a different myth of the Middle Ages—not the myth of the primitive, barbaric dark ages, but the Middle Ages of Charles Homer Haskins’s The Renaissance of the Twelfth Century (Cambridge, 1927) in which Europe emerges triumphantly from localism, superstition, and oral chaos (about the coupling of the last two words see Ch. 1) into rationality, law and order, and strong central government—in other words a Europe worthy of being the fountainhead of our culture.¹¹ Underlying the whole project, therefore, was a third historiographic purpose, tacit except in the title of this chapter: to present a concept of Western music as a standard for the evaluation of all music. Rationality, order, and the unity of integrated form that is the counterpart of strong central government, become the defining properties of Western music, and any music that lacks these virtues risks being perceived not only as foreign, but as deficient and not worthy of close analysis. It risks being treated as though it were hardly music. Peter Wagner characterized the contrasting cultural styles that he thought to recognize within the tradition of Gregorian chant in the paragraph that I quote immediately after this introduction: on one side regularity, plan, clear disposition in groups, transparent structure (‘structure’ is my translation of Wagner’s Gliederung and Aufbau, and I shall return to the concept presently), symmetrical configuration, clear construction cast in a manifest order, a product of aesthetic deliberation. On the other side undulation in unregulated fashion, organic (he meant this in the sense of unconstrained growth, not in the sense of the nineteenth-century aesthetic doctrine), without regularity or plan, without clear disposition in groups, without symmetrical arrangement of parts. This opposition is reflected in the way that Wagner’s discussion of the alleluia Dies sanctificatus and my discussion that is based on it are focused on properties that the melody lacks—specifically formal properties. Hence the first item of business in this reconsideration must be a new account of that chant, focused on the materials, procedures, conventions and constraints, and aims that seem to inform it, unconstrained by the biases that guided my interpretation then. To speak of Gliederung is to speak of the articulation of something in its constituent members or parts. Our arms and legs are Glieder. Members of an organization are its Mitglieder. The concept accommodates both the part–whole relationship and the joints, so to speak. To speak of Aufbau is to speak of building up, construction, accumulation. The concept accommodates both the putting together and the resulting construction. In translating both as ‘structure’—there is little choice—we remove the trace of the dynamic that is in both originals. ‘Structure’ stops the clock. Invoking, as it does, architecture and engineering, it makes one think at once of buildings and bridges or other things that are built, and of the ¹¹ See Brian Stock, ‘The Middle Ages as Subject and Object’, NLH 5 (1974), 527–48.
108
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
way their members fit together and support one another. It directs attention to the thing as a whole, seen at once (synchronically) in its entirety. It suggests a system of connections and dependencies, and since the age of structuralism, at least, it suggests depth and relationships through depth such that details at any one level result from the transformation or interpretation of structures at deeper levels. In respect to music such an orientation has of course not just seeped in from structuralist linguistics or anthropology, it is a product of Western compositional history and is a part of every music student’s training in that music. It is a special conception, which has arisen out of a particular—late medieval—musical practice and aesthetic to whose description it can contribute very substantially. If we universalize it we are bound sometimes to misrepresent it. It is a conception that frames research, and can frame the researcher within it. In returning now to the same alleluia melody I do not ask about the forms or the formal categories in which it is encased, and certainly not, as before, whether it has any form at all, but about the directed continuities and movements that define its coherence as it unfolds in time. And I shall take a narrative perspective, asking as I follow the melody what linkages and internal references become active in the immediate succession of notes (and in the relation of note groups to units of language) but also across the melody, following larger dynamics and patterns that emerge through accumulation. The narrative, that is, need not follow the melody from note to note, but it must be about temporal process. To borrow a phrase from Heinrich Schenker, it will be a narrative of ‘the life of notes’.¹² I shall refer to notes and note-groups or figures (corresponding to neumes) that are numbered in the transcription. See Ex. 5.1 (track 6 on the CD; an Old Roman version is included on track 7 for comparison). 1–6: a rising motion, reiterating the second note (1–2), that much suggesting the second note as a first point of arrival, a quality that is enhanced rhythmically by the fact that it is presented as a single note. 3–4 seems at first like an echo of 1–2 a fourth higher, but 4 dispels that as it falls back to the first note of 3. 5 reiterates the falling movement of 4, but falls through a larger interval. Its goal is a new note (e) , filling in the skip from 2 to 3. This reference back to the d–f skip is kept active in 6–7, which reverses the skip. That, in turn, has the effect of closing off this segment of the melody as a phrase that sets the first complete word. That this is done with a return to the note d reinforces the initial impression of that note as a goal, and one that is treated as relatively stable. The opening c, on the other hand, is relatively unstable, and the opening dynamic entails a pull that d has on c or, to put it the other way about, a push that c has towards d. The d, when it is reiterated in 2, is the only note in the phrase that stands alone. The phrase opens in tension and closes in resolution. The effect is not unlike that of the resolution of an accented dissonance. 7–17: the upward sweep of the melody at the very beginning is halted and deflected downwards in 4, where it had reached the highest tone to that point. 7 ¹² Erläuterungsausgabe der letzten fünf Klaviersonaten Beethovens, 4 vols. (Vienna, 1913–21), Opus 109, pp. 56–7. See my discussion in ‘Beethoven’s “Expressive” Markings’, Beethoven Forum, 7 (1999), 89–106.
109
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a Ex. 5.1. Alleluia Dies sanctificatus, transcribed from Trier, Dombibl. 153
V
1
2 3
4
5
Al
le lu
ia.
18
19
20 21
6
7
8 9
22
23 24 25 26
es 39
12 13 14
15 16 17
ij.
Di 38
10 11
40 41
42
27 28 29 30
31 32 33
34 35 36 37
sanc ti fi ca
tus il lux
it no
43 44 45
46 47
48
49
50
bis: ve 51 52 53 54
55 56 57 58 59 60 61 62
63 64
ni
te gen tes et a do ra te
Do mi
69
70
di 87
88
71 72 73 74 75 76
77 78
e
de scen dit lux ma
89
90
91 92 93
gna su
per
79 80 81 82 83 84
94
95 96
ter
ram.
97 98
65
66 67 68
num, qui a ho 85
86
99
takes up 3 again, as a platform to push one note still higher (a), the highest note that the melody will reach altogether. The moment is given weight again by being presented as a single note (8), and as the pivot for another downward turn (9), this time a much sharper turn. Looking ahead for a moment we can see how sparingly this note is sounded. There is a kind of elasticity about the melody altogether, as it is deflected downwards whenever it reaches the a (21,38,69,84), except at 95–8, where it persists in holding the a against this elasticity. The persistence makes of this the moment of greatest tension, the climactic moment, and after it the melody ends quickly with a single figure that links to 6,17, 26, 64–5, 74–6—i.e. all the instances in which the melody rests at the point of its origin and resolution. This is very much a matter of pacing and planning, displaying an exquisite sense of time. Returning to 9, now, it is the deflection from the ceiling note, first by step, then by the largest leap in the entire melody (a leap of a fifth is virtually the largest leap that occurs in this style, and its occurrence is always a marked signifying moment).
110
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
Its target is the first recurrence of the c since the beginning, and as then the movement is straight to d and its drawn-out tilting with f —drawn out, really, until the end of the phrase (10–17). It seems like a settling down after the downward leap (9–10). The f itself is reiterated four times during this drawn-out stabilizing process. The stability of f in this melodic matrix is even more strikingly exploited in the verse, where it is reiterated eleven times in succession with just one intervening deflection to g (51–61). In this passage it is the note for reciting a good bit of the verse, and that gives it a role parallel to d (see 27–35). The notes c and f play quite different roles in the dynamic of the melody. While f is something of a partner to d, sharing its dynamic roles, linked to it in cadential figures, pivotal in the turning of the melody down to d or up to a, c is a counterweight to d, its opposite, as unstable as d is stable. Hearing c we anticipate d, but certainly not the other way around. C is thus a factor for activating and reactivating the melody, a potential that is exploited, for example, in the continuation from 43 to 44. 43 concludes a phrase, coinciding with the end of a syntactical group in the verse: ‘This sacred day enlightens us.’ In modern editions a colon is set at that point, corresponding in function to the c. Both point forwards to the beginning of the next phrase. (This correspondence of function between melody and punctuation in articulating the sense of language is absolutely characteristic, and is surely a factor in the early history of musical notation. It is discussed fully in Chs. 13 and 14.) In the continuation from 89 to 90, however, the c is not followed by d, but by its partner in stability, f. The effect is a delay in the arrival of the anticipated d, in what becomes the climactic phrase of the whole melody. The postponement of the d enhances the effect of holding out against the deflection from a. It is, again, a matter of great sensitivity in the pacing of the dynamic events of the melody, very far from ‘undulation in unregulated fashion, without plan’. The sequence d–f–a constitutes a kind of melodic potential that supports the melody. It is most explicitly actualized in the figures 38 and 84 and the sequence 69–76. Pulling against it is the sequence anchored in c,c–e–g. This is most explicitly actualized in 39 and 85, each time playing against a figure based on the principal sequence d–f–a (38 and 84). I call it principal because of its anchorage in the principal tone, d. Figure 98 almost combines both sequences. It lacks only the d that is skipped, creating a need for the return to it that comes in the final figure. Once accustomed to these potentialities, we can discern clearly the parallel ways in which 7–17 and 91–9 conclude the responsory and verse, respectively: opening from f, reaching up to a, shifting from the sonority of the primary third-sequence (d–f–a) to that of the contrasting one (c–e–g) in descending to c, the c forcing the cadence. Both times the opening move from f to a sounds the upper half of the d–f–a sonority, and the cadence complements that with the lower half, to make the close. But what a difference there is in the pacing of the closures: in the first instance a drawnout cadential process—f tilting alternately with d and c (10–17)—and in the second a close like the springing of a trap (99). There is one downward leap of a fourth, from d to A (48). This links itself to the high a as its octave and it completes the full identification of the sound-space of
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a the melody as a whole. The occurrence of this moment of completion so far into the melody is itself a form-defining element. (This sound-space, with the as as outer boundaries and d as central source note, is, of course, the space that came to be defined as a dimension of the first plagal mode when the modal system was imposed on the melodic practice around 800.) Following the action of the chant only this far should render it sufficiently clear that Wagner’s negative characterization of the melody as melismatic undulation without regulation or plan, without symmetry or clear disposition of parts, without the direction of ‘aesthetic deliberation’, is not so much inaccurate as irrelevant and inappropriate. Imagine describing the opening of Debussy’s Prélude à l’après-midi d’un faune as melismatic undulation without regulation or plan, without symmetry or clear disposition of parts, failing to come up, in short, to the standard of Western music. My error in adopting Wagner’s judgement of the alleluia was precisely in failing to heed Auerbach’s precept that ‘The works of art of the different peoples and periods, as well as their general forms of life, . . . have to be judged each by its own development, not by absolute rules of beauty and ugliness’ (p. 9), and then elaborating Wagner’s construction of a teleological history in which the absolute rules signalled by the phrase ‘transparent structure’ constitute the necessary and sufficient criteria of ‘aesthetic deliberation’ and, really, coherence. And the crowning error is displayed in the arrogance of my title: the identification of ‘transparent structure’ as an essential, and by implication, defining property of ‘Western’, and only ‘Western’ music. In the original essay this ideology is articulated in the context of a narrative of incremental change in the form-building process towards the goal of ‘transparent structure’: three melismas . . . ‘reflect a further stride in the direction of change’; another is ‘even more modern’; still another shows ‘a reduction of the line, with a resultant sharpening of the form’; in another the melody had ‘not yet assumed its final form’. This rests on two assumptions, neither of them self-evident: that the variants compared in the essay should be regarded as revisions, and that they should be aligned in the sense of a progression towards a goal. As to the first, the variants can also be regarded simply as different remakings in a tradition that entails remaking in the process of transmission. As to the second, the apparent stylistic change could just as easily elicit disappointment over a decline in place of the triumphalist rejoicing over the rise of ‘Western music’: a loss of the sense of suppleness and subtlety of movement compensated only by a form-building procedure that is at once more mechanical and more dependent on the juxtaposition of large blocks or the sweep of large-scale linear processes. The turn to such regularities—which can enhance memorability—might reflect an effort to shore up a declining oral tradition in which spontaneity was a value and a skill. The category of ‘revision’ will come up again later. I have found in this reconsideration a different spirit in the alleluia Dies sanctificatus: pacing, timing, the making of form through accumulation, linkage, association, anticipation, postponement, quickening and retarding movement, changing tensions—everything to do with the flow of the melody in time. I think
111
112
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
of it as the product of a highly sensitized, subtle musicianship, trained in an oral tradition in the art of continuity, not as a hierarchically integrated structure. Is it the product of ‘aesthetic deliberation’? Certainly, as long as we do not read ‘deliberation’ with the historically acquired connotation of the composer working in the sketchbook. I find that my apprehension of this melody upon reconsideration resonates strongly with the results of an enquiry conducted by Fritz Reckow and reported in his essay ‘Processus und Structura: Über Gattungstradition und Formverständnis im Mittelalter’.¹³ Reckow’s enquiry pursues the question ‘what did medieval authors articulate thoroughly in regard to musical form or at least bother to mention?’ He asks this question mainly in the context of the study of conductus and motet and we can therefore only tentatively regard his observations as relevant to the time of the alleluia Dies sanctificatus through backward projection. I shall return to this question after summarizing Reckow’s observations, but at the very least what he has found will suggest second thoughts about the widespread characterization of Gregorian chants according to the criterion of formal unity, for that criterion would not likely have had meaning for medieval musicians engaged in the practice of chant. I cite Reckow’s essay at some length in order to bring out the sharpness of the contrast between the conception of medieval writers and the values that we modern music historians have superimposed onto the music. From the early Middle Ages onwards, music was described as movement, as a process, a ‘course’ from a beginning towards an end . . . The descriptive vocabulary of music theory is characterized throughout by an abundance of expressions that in clear nuances refer to various kinds and aspects of movement down to the point of rest—running, walking, jumping, straying, retarding, stopping, resting, and waiting . . . The image of music is obviously determined by the thought of a concrete audible performance executing sounds in movement. The concept of motion is so obviously present that it apparently does not need a proper theoretical exposition or justification. It therefore marked musical understanding of form all the more forcefully . . . The ‘dynamic’ concept of music as a ‘course’, manifested in such a vocabulary, touches not only apperception of form in general but also composition in particular. The fitting together of parts is seen as a sequel in time, as a ‘path’ that the composer—and after him everyone who performs or hears the composition—‘walks’ from a beginning to an end. This is not only in the obvious sense that all music is an ‘event in time’, or an experience that ‘contains time’, but also in the sense—not so self-evident—that a composition can be consciously conceived and executed as a flow in time, with consequences for its form(ation), for the relation of parts to one another, as for the specific kind of connection and cohesion of the whole. . . . A primary concern is not the ‘closure’ of the whole but apparently the contrast between parts following one another . . . To the conception of form that comprehends the whole as a processus corresponds a particular interest in the diversitas of parts. And in consistency ¹³ Musiktheorie, 1 (1986), 5–29.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
113
with that, music theory thematizes less the creation of unity than the avoidence of monotony.¹⁴ It is in discussions of the motet around 1300 that Reckow first finds a concept of structura, spelt out in terms of tonal unity and periodic regularity. The criterion of diversitas is superseded by that of similitudo. Coming back to the question of the relevance of this overview to the music conception underlying Gregorian chant, it is clear that the processus concept investigated by Reckow is a derivative of the earlier medieval, and ultimately ancient, conception of melody as a movement of the voice (see Ch. 13, esp. n. 34) whereas the structura conception that is found first in the fourteenth century is the first sign of Wagner’s ‘Aufbau’ and ‘Gliederung’ that underlies the dubious characterizations of Gregorian chant in writing discussed especially in Chapters 7 and 9. There is no evidence of such a music conception holding sway in the Middle Ages from any time between the earliest theoretical writing in the ninth century and c.1300, that is, no evidence that anyone would have thought of Gregorian chant in that way during the time of its creation and early transmission. The historiographic problem that is posed here is inscribed broadly across this book, as it must be if the book is to be an honest representation of the course of coming to know a music culture so distant from ours, yet present in ours in so many ways that we can too easily be inclined to describe that culture in terms that have grown up with ours. Just now we have encountered the tension between the habit of interpreting and evaluating a medieval tradition on the criterion of a modern value that we have long held as second nature. Measuring music on the criteria of structure, unity, and closure is confronted by indications that just those criteria were foreign to the culture under study. If we were to react by feeling compelled to revise the interpretation, bringing it into consonance with that evidence, it would be out of a belief that the historical text should be a window that yields an unclouded view of some past, and testimony from that past must be granted the highest authority. But we may hesitate, acknowledging that the historical text can never be altogether transparent, that it inescapably mediates between past and present, can only represent the past, cannot just reveal it. Nothing is quite what it was after we’ve formed a clear picture of it.¹⁵ If the medieval participants in the Gregorian chant community had no conception of or interest in the structure of chants, we may not want to let that deflect our interest in them from that point of view in the present. We may believe that point of view to open on to a more illuminating picture of the nature and history of Gregorian chants. Many scholars seem to have acted on that belief (but none, I think, as a conscious choice against the possibility that their belief is anachronistic). That opens the way to the incursion of modern language, concepts, and values. The ¹⁴ Translation by Siegmund Levarie. ¹⁵ Stephen Dunne, ‘Burying the Cat’, in Different Hours: Poems (New York, 2000), 111.
114
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
investment, and consequently the risk, is greater still if the historical significance and aesthetic value of Gregorian chants is rated on a sort of structural-unity scale, according to which such chants as the alleluia Dies sanctificatus are judged primitive, not even the product of ‘aesthetic deliberation’. The issue has been simmering throughout this introduction and will boil up repeatedly throughout this book. Historians have long cleaved to one or the other of those two standards for the status of historical texts, which is to say those two beliefs about the nature of their task, whether or not they have made their choice in consciousness of them as alternatives. The textbook model for this polarity has been the contrast between the histories of Thucydides and Herodotus. Lord Macaulay, writing in the early nineteenth century, a time of intense discussion about this question, characterized it thus: ‘The history of Thucydides differs from that of Herodotus as a portrait differs from the representation of an imaginary scene.’¹⁶ It is my impression that throughout its history, and even today, historians working in the domain of this book have not done so out of that sort of consciousness. The field is marked by a contradiction: it is committed to the goal of producing true displays of the past anchored in hard evidence and objective reasoning, whereas its literature abounds in theories and narratives that can hardly be confirmed or disconfirmed because they are informed at least as much by belief systems, cultural myths, historiographic habit, and the demands of the narrative form itself, as they are by evidence and reason, and which may indeed be founded on very little evidence at all (see the introduction to Ch. 6 for a particularly interesting case in the current literature). It is one of the chief aims of this book to encourage a more open and explicit engagement of this dimension of our work. We can make out a distinct polarity in the beginnings of modern historiography around 1800, a time when historians wrote explicitly about what the choices were and where they placed themselves within them. This polarity framed, for example, the reputation that was instantly generated by James Boswell’s Life of Samuel Johnson, LL.D., which came out in 1791. It was the judgement of contemporaries that ‘Boswell didn’t exactly write the Life, he merely took it down—a faithful but mindless stenographer; the virtues of the book simply reflected the virtues of its subject. . . . Boswell was a “great fool,” Macaulay wrote in 1831, and while conceding the merits of the Life, he added that it was a book only a fool could have written.’¹⁷ This should be read in conjunction with Macaulay’s dictum that ‘some of the best writers of fiction have been among the worst writers of history’ (he might have been thinking of Sir Walter Scott). In 1821 Wilhelm von Humboldt began a lecture on ‘The Historian’s Task’ with this statement: ‘Die Aufgabe des Geschichtschreibers ist die Darstellung des Geschehenen’ (‘The task of the writer of history is the representation of what hap¹⁶ ‘History and Literature’, in Fritz Stern (ed.), The Varieties of History: From Voltaire to the Present (New York, 1956), 74–5. ¹⁷ Cited from Charles McGrath, review of Boswell’s Presumptuous Task (New York, 2000) in The New York Times Book Review, 19 Aug. 2001, p. 12.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
115
pened’).¹⁸ Three years later Leopold von Ranke declared it to be the aim of history ‘merely to show how it [the past] eigentlich was’.¹⁹ I leave eigentlich untranslated because it has many different English equivalents that can lend the statement different nuances of meaning in translation. It may be actually, truly, really, essentially, intrinsically, properly, strictly speaking, after all, exactly, originally, authentically. In any case, the statement has appropriately been called ‘the most famous statement in all of historiography’,²⁰ and vouchsafed mythic status in the discourse of historiography.²¹ Ranke’s statement is indeed much cited, more often than not far too simply as token of a positivist doctrine of historiography. Bann’s essay, circumventing the arguments about the statement’s lexical meaning and focusing instead on the context in which it was made, thus what Ranke meant in making it, brings better understanding. First there is the statement’s place in Ranke’s text, as conclusion to this sentence: ‘History has had assigned to it the task of judging the past, of instructing the present for the benefit of ages to come [by Voltaire, for example]. The present study does not assume such a high office; it wants merely to show . . .’. The crucial word here is ‘merely’. This is a matter of history’s rank and authority and of the social, political, and rhetorical force of the historical text. It has its own rhetorical force, taking history down from its pretensions to philosophy and disavowing a role for it as cultural and national myth (see Ch. 9). Then there is the juxtaposition of Ranke’s zeigen (show) with Humboldt’s darstellen (represent). This is a matter of the poetics of the historical text and of the epistemological presuppositions from which it arises. And the broadest context is that of the general discussions of the aims of history and the shift that was taking place through those discussions in beliefs about the historian’s task. Humboldt craftily underscored the contrast of historiographic modes in his continuation. ‘The simple representation [of what took place] is at once the primary, indispensable condition of [the historian’s] work and the highest achievement he will be able to attain. Regarded from this vantage-point he appears to be merely passively receptive and reproductive rather than actively inventive’²² — appears to be but isn’t really, is the implication: representing, creating illusions, simulating, imitating (mimesis) is the artful literary work of the historian, all in the interest of verisimilitude, but achieved through the working of a text. ‘Verisimilitude’ suggests ‘lifelike’, but that expression can have two quite contrary connotations: just the way it is or was in real life, but then to make it lifelike the historian must breathe life into it. That is the classicist aesthetic for representation, and it is what Ranke aimed, through his own craftiness, to disavow with the two simple words ‘[er will] bloss zeigen’. ¹⁸ Wilhelm von Humboldt, Werke, i (Darmstadt, 1960), 585. ¹⁹ Geschichte der romanischen und germanischen Völker von 1494 bis 1514 (2nd edn., Leipzig, 1874), p. vii. ²⁰ Leonard Krieger, Ranke: The Meaning of History (Chicago, 1977), 4. ²¹ For an analysis of this influential bit of reception history see Stephen Bann, ‘The Historian as Taxidermist: Ranke, Barante, Waterton’, in The Clothing of Clio (Cambridge, 1989). ²² In Stern (ed.), The Varieties of History, 57.
116
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
For that he gained his reputation as the progenitor of a new professional historiography of the nineteenth century that was to be entirely outside the province of literature, a historiography in which the function of the linguistic sign was to be transparent to the historical signified. Lord Acton, in his ‘Inaugural Lecture on the Study of History’, pronounced Ranke ‘the representative of the age which instituted the modern study of History. He taught it to be critical, to be colourless [for ‘colour’ read ‘personal style’], and to be new.’²³ The pretension ‘to be colourless’, however, like that to ‘merely show’ how things were in the past, betrays itself. The French encyclopedist Jean-François Marmontel wrote ‘Someone has said that for the historian the best style is one that resembles limpid water. But if it has no colour in itself, it will naturally take on that of its subject, as the stream takes on the tint of the sand which forms its bed.’²⁴ The analogy takes me back to the situation that touched off this excursion into historiography, the tradition of chant criticism and history that takes unity of form as its point of focus, confronted by evidence that in the world of that music this was hardly a relevant or even a known concept. We writers in that field have no doubt believed our texts to be transparent to their objects, but in fact our texts have taken on the tint of the bed in which the stream of their tradition flows—in this case the language, concepts, and values of our scholarly tradition. That this is a tension that springs up whenever accounts of human events are put before a public in whatever medium, was brought home to me just as I was writing this passage by a sentence in a recently published journalistic film review. The review, by Maria Margaronis, concerns the cinematic version of Louis de Bernière’s novel about the Nazi occupation of a Greek island, Captain Corelli’s Mandolin.²⁵ The reviewer describes an early scene in which a group of villagers crowd around a list of casualties posted in the public square. Until this point we have seen only minstrel-show peasants, conversing in broken English and executing perfectly choreographed folk dances. But now an elderly man collapses with grief and cries out in his own language ‘Oh god, oh god, my boy. My boy has been killed’. The moment feels real; while it lasts, contact is made with history. Then we are back in the fantasy world of director John Madden’s Anglo-Hollywood confection, where all the women are brave and the children above average, the Italians love pasta and opera, and the Greeks are good-hearted and proud. The aspect of ‘aesthetic deliberation’ in the paradigm under which Wagner wrote and which I had adopted in 1968 refers tacitly to the idea of the scriptuality of music that is taken as an immanent property linked to its coherence and that is implicit in the concept of structure.²⁶ Scriptuality, in this conception, is a property that music ²³ Lectures on Modern History, ed. John Neville Figgis and Reginald Vere (London, 1969) 32. ²⁴ Éléments de littérature, iv (Œuvres complètes, viii; Paris, 1787), 113. ²⁵ The Nation, 20/27 Aug. 2001, p. 36. ²⁶ Helmut Hucke recognized this in his essay ‘Gregorianische Fragen’: ‘It was taken as self-evident that Gregorian Chant was scriptual (schriftliche) music, just as it was taken as self-evident that Western music was scriptual music’ (p. 305).
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
117
can have immanently, not simply a matter of the technology of transmission. It is the other side of a duality in which being ‘unregulated and unplanned’ is a condition of improvised music. It is under this conception of music as inherently scriptual in so far as it is coherent, that I had unreflectively interpreted the different versions of the same melodies that I showed in the essay in the sense of revisions.²⁷ It lies not far below the surface of these associations that scriptuality and the coherence with which it is intertwined are factors in the advanced or superior characteristics of Western music, and as is so often the case, that conception received explicit articulation before it entered the historical unconscious. Johann Nicolaus Forkel, reacting in his Allgemeine Geschichte der Musik to reports about the music of ‘wild and uncultivated people’, wrote that the melodic disconnectedness of the notes of this music was so extreme that it could not be apprehended by travellers and written down with European notation. The notes of their music, he said, were melodically unconnected to one another (‘unzusammenhängend; nicht mit einander verbunden’)—literally incoherent, unlike the music for which European notation was devised.²⁸ Jean-Jacques Rousseau, in his Essai sur l’origine des langues, wrote ‘All people who have string instruments are obliged to tune them in consonances, but those who lack them have inflections in their songs that we call false because they do not follow our system and we cannot write them down. This is what one has noticed about the songs of the savages of America.’²⁹ The other connotation of ‘aesthetic deliberation’ that affects analytical methodology is not so readily documented, although it is equally potent. It is the assumption that the schemata that we produce in analysis represent the compositional process and through that representation constitute a sound basis for a historical account. This is particularly troublesome when the analysis is conducted from the vantage point of a culture so far removed from that of the artefact being analysed as mid-twentieth-century European-American culture is removed from eighth- or ninth-century European culture. What I presented in the essay on which this chapter is based may be perfectly sound analyses of the scores, but that is not to say they reveal how the pieces were composed, or that they highlight what were for the composers and listeners of the time the values that guided them or what they would have regarded as the music’s effects or achievements. What they highlight are the principles that I believed generate coherent music. At the time the essay was published the probable response to this from the point of view of critical theory might have been that it violates the scruple against the ²⁷ The concept of ‘revision’ is like the concept of ‘centonization’ in that both are generated out of a mindset that tacitly assumes the ‘scriptuality’ of music. ²⁸ (Leipzig, 1788), 5. ²⁹ ‘Tous les peuples qui ont des instrumens à cordes sont forcés de les accorder par des consonances, mais ceux qui n’en ont pas ont dans leurs chants des inflexions que nous nommons fausses parce qu’elles n’entrent pas dans notre sistème et que nous ne pouvons les noter. C’est ce qu’on à remarque sur les chants des sauvages de l’Amérique’; ed. Charles Porset (Bordeaux, 1970), 181. Regarding both Forkel and Rousseau see Wulf Arlt, ‘Der Beitrag des 18. Jahrhunderts zum Verständnis der Tonschrift’, in Carl Dahlhaus and Hans Oesch (eds.), Festschrift für Arno Volk (Cologne, 1974), 47–90.
118
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
intentional fallacy. From the point of view of the paradigm of style history³⁰ it might have been that my questions about values, intentions, and effects at the time are irrelevant to the narrative of orthogenic and teleological development in which I embedded the analyses. It is the narrative that has priority: the story that claims to recognize in those early melodies the potential of what was actualized in music of later epochs, and that story does not register a claim to realistic representation of the musical culture; on the contrary its interest in that culture is the narcissistic hope of being able to make out in it the embryonic image of our own culture. But instead it succeeds, through that ambition, in projecting a full-blown cultural selfimage, and therein lies the irony (see Ch. 9). But the dogma that one cannot know the intentions of other minds and should not aim to act on such knowledge if one could attain it, no longer strikes fear the way it once did.³¹ And the priority once granted to the narrative has since come to be shared, at least, by the principle of aesthetic historicism long ago enunciated by Vico and reasserted more recently by Auerbach and Stock.³² The conclusion for historiography is a caveat about the control that narrative can exercise in the interpretation of historical evidence, not a denial of the aims of large-scale narrative history or of synthesizing interpretative hypotheses in general. It is a call, rather, to alertness to the ways in which narratives in historical studies can take charge and direct the selection and interpretation of evidence, especially when they act out ideological agendas that may be tacit, and even more so when they are formed about dualities of self and other. There is a story to be told about the creative consolidation of a musical culture in Carolingian Europe, and that story does indeed entail new interactions of music and language that resulted in a different idea of musical form from what is at work in the old chant tradition; that aspect of the reasoning in this chapter can, I think, be left standing. The story also entails a number of profoundly influential inventions: musical notation, the idea of a corpus of music that can be complete, authentic, authoritative, enclosed within a book, subject to a comprehensive system of classification, the product of an author, taught, subject to systematic analysis and speculative theorizing. The story would also be about the profound effect that this consolidation had on the history of European—and not only European—music. It is a story with political and cultural plot-lines, not just musical ones. And to tell it would call for the same close attention to detail as well as to large outlines, as the close analysis of a chant melody.
} ³⁰ See my Music and the Historical Imagination, 85. ³¹ There is no need to review the reasons for this change. The issue is discussed in Music and the Historical Imagination, 67–78. A sufficient way of reassuring ourselves about this was once offered me by the philosopher Peter Kivy, who pointed out that when he drives between his home and university via the New Jersey Turnpike he is constantly forming hypotheses about the intentions of other minds (those of the other drivers) and acting on them. ³² See nn. 8 and 12.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
119
It has been a central concern of the literature on the Gregorian alleluia to pursue that genre from its early Judaeo-Christian beginnings to its Byzantine prototypes, and thence through a uniquely long period of composition that was marked both by major changes of style and by a steady growth of the repertory. Peter Wagner read meaning into those style changes for the growth of an indigenous Western—as opposed to an inherited Oriental—compositional tradition. Contrasting the melismas of the alleluias with those of the mass responsories and the tracts, he wrote: We have recognized the organic course of the series of notes as the most prominent characteristic of melisma structure in the mass responsories and in the tracts. Less often do they constitute a symmetrical pattern, effected principally through the repetition of note groups. The situation in the alleluia chants, however, is quite the reverse. In only a very few instances do the melismas proceed without regularity or plan, without disposition in groups that are immediately and clearly recognizable. Most of the others are models of transparent structure, with all the parts organized in a symmetrical configuration. In this the former group shows itself to be the older, the latter as more recent and as the product of aesthetic deliberation. . . . The older group forgoes any symmetrical arrangement of its parts; in the more recent alleluia chants the melismas are always clearly constructed and of a manifest order. The former are reminiscent of the melismatic melody of the Orient, which still today undulates in unregulated fashion. The latter display Latin, Roman traits in their clear structure.³³
To represent the older group, Wagner quoted the melody shown in Ex. 5.1. It grows out of the repeated intonation of a limited number of short, characteristic melodic turns and a frequent pausing on the two notes of choice, f and especially d. These circumstances can indeed result in the near identity of extended passages, as in the first and third phrases of the verse. But it seems fundamentally different from the invention of a longer line that is repeated as a matter of structural principle. Such is clearly the case in Ex. 5.2, the alleluia Surrexit Dominus, whose melody is quoted as an extreme case of the repetition scheme by Friedrich Gennrich.³⁴ This chant incorporates all three of the features by which Wagner characterized his later group: the repetition of the alleluia melisma after the verse to give an overall ABA form, the adaptation of the alleluia melody for the verse itself, and the repetition of melodic groups within the verse.³⁵ ³³ Gregorianische Formenlehre (Leipzig, 1921), 398. ³⁴ In Grundriss einer Formenlehre des mittelalterlichen Liedes (Halle, 1932), 108. ³⁵ Wagner proposed this stylistic contrast as an internal criterion for chronology among the alleluias. Specifically, he proposed the year 608 as a watershed separating the types represented above, on the following argument: the alleluia Adorabo (Graduale Romanum, 1731) shows a tendency to symmetry in the verse but it does not repeat the alleluia melisma after the verse; hence it is to be considered transitional. The melody is unique to the Adorabo text, so it must be original to that text and its ritual function—the consecration of a church. The first known occasion of that ritual—the dedication of the Roman Pantheon in 608—fixes that year as the approximate time of separation between the two alleluia types. Wagner executed this series of deductions in a speculative way, and we must still regard it with some circumspection, for more is assumed than is demonstrated about the transitional role, and especially the age, of the Adorabo melody.
120
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
I show now three alleluia melismas that reflect still a further stride in the direction of change that Wagner observed between the types of Exs. 5.1 and 5.2 (see Exs. 5.3–5). Like the melody of Ex. 5.2, these involve repetitions of one sort or another, but unlike the latter, the repetitions are not literal. The differentiation of the matched phrases is at the cadence, and in each case it is a differentiation of endings on the confinalis and finalis, respectively. This has the effect of polarizing each of the passages, of imposing on them a direction on the phrase-to-phrase— over and above the note-to-note—level. A serial construction of literal repetitions is potentially open-ended, for it involves no process that is directed towards a goal. In contrast, repetitions like those of Exs. 5.3–5 function in closed, integrated musical configurations. The difference is critical, for it is just towards structures of the sort that these examples represent in miniature that much of the subsequent musical practice of the West has aimed. In the Middle Ages, directed musical structures of this sort, depending then principally on matched but differentiated phrases, were the counterparts in music of the symmetries of rhymed, accentual Latin verse. Earliest in the sacred, nonliturgical songs,³⁶ later in the tropes and proses, phrases of equal length matched isosyllabic lines, masculine and feminine cadences matched the alternation of full with catalectic lines, and rhyme schemes were reflected in range and cadence differentiations that depended on the structure of the modal octave. For a single representative specimen I offer the Ex. 5.6.³⁷ This manner of organizing music was widely adopted in the later Middle Ages and after: in monody and polyphony, in vocal and instrumental music, in sacred and secular music. On the face of it, and especially on Wagner’s criteria, it would appear to be a rather sharp departure from the core of plainsong practice. If it was to count at all in chant composition, it is reasonable that it should have been substantially confined to melismatic writing, for this sort of structure applied to a prose text would constitute a crude misalliance of words and music. Such a case is melody no. 14 in the Appendix to this chapter. But that it should occur at all in chant poses some interesting questions regarding the bearing of chant composition and non-liturgical practice upon one another. One wonders whether the structuring techniques developed to support chant melismas suggested ways of setting the new poetry without obscuring its form; or whether the composers of new chant imitated the song writers in the only place where it was feasible for them to do so, in the melisma; or, indeed, whether both were not pursuing the same ideals of symmetry—but with the song writer, like the instrumental musician, working in a field more fertile for their realization. Very ³⁶ Songs, for example, by the St. Gallen poets Ratpert (c.890), Hartmann the Younger (925), and Waldram (active c.900), preserved in such sources as St. Gallen 360 and 381. The real vogue of musical settings of accentual verse waited until the 11th and especially the 12th c. ³⁷ This review of the characteristics of the musical settings of accentual verse is necessarily brief. For a detailed study I refer the reader to my ‘Musical Syntax in the Middle Ages’.
121
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
Ex. 5.2. Alleluia Surrexit Dominus, reproduced from Gennrich, Grundriss einer Formenlehre des mittelalterlichen Liedes, 108 * *
Solo: Al Choir: Al Solo: V Sur et Choir: Al
le le re ap pa le
*
lu lu xit Do ru it Pe tro lu
*
ia. ia. mi nus ia.
ij. (f )
ve ve
ij.
re, re,
* These notes are sung in the verse only
Ex. 5.3. The melisma of the alleluia Beatus vir qui suffert (Graduale romanum, [43]) A 8
ij. A'
8
Ex. 5.4. Alleluia and melisma of the alleluia Te martyrum (Graduale romanum, 30) B
A 8
Al
le
lu
ia. ij.
A' 8
Ex. 5.5. Alleluia and melisma from the alleluia Beatus vir sanctus Martinus (Graduale romanum,652) A 8
Al le Bb
8
A' 8
A
lu
ia. b
ij. b'
122
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
Ex. 5.6. Rhythmic prosa from Misset and Aubry, Les Proses d’Adam de Saint Victor (Paris, 1900), 238
8
8
Chris tum Si cut
fes pig
ti nus
te se
mo me
nens nex
per Ver
rem tris
a bum
ges bra
mo Pa
tat chi
rem tris
lu is
be vo
ne tis
men cor
iux strin
ce po
ta xit
re re
um; um.
likely this is a chicken and egg question, and it makes little sense to hope to choose one answer over another. But it is a matter worth considering in any case, for it offers a way of holding liturgical and non-liturgical music up to the same light. Specifically, one would like to know how widely represented those structures are in the chant, and when in the history of chant they entered the repertory. I propose to make an approach to these questions through the alleluias of the Graduale romanum. As a beginning I should like to refer back to Exs. 5.3–5 and recapitulate the essential features of those forms that distinguish them from the simple repetition schemes represented by Ex. 5.2. The elements of the form are phrases matched minimally in length and contour. They are arranged in a closed configuration, either contiguously or with an intervening phrase or phrases, and closure is a matter of the differentiation of phrases on the basis of the structure of the modal octave. Usually the antecedent phrase ends on some note other than the final, while the consequent ends on the final. On occasion, as we shall see, both may end on the final, but they lie in the tetrachord and pentachord, respectively, of the modal octave (see no. 14 in the Appendix). The effect in either case is a structure that combines the quality of symmetry—shared with the simple repetition schemes—with that of direction, something which the latter lack. It is a matter of dynamic, as opposed to static, form. The Graduale romanum contains 296 alleluia texts set to a number of melodies which—taking into account the multiple use of some melodies—is something less than that. If we project onto the repertory in the Graduale the ratio of texts to melodies that Apel reported for the Liber usualis (1967)³⁸—3:2—we can expect approximately 200 melodies, more or less. Among those, seventeen contain melismas that answer to the criteria I have outlined above. In this simple statistic, crude as it is, there is the suggestion of an answer to one of our questions: the sort of order that became common in non-liturgical music of all kinds is to be found in ³⁸ Gregorian Chant (Bloomington, Ind., 1958), 383.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
123
the melismas of fewer than 10 per cent of the alleluia melodies. The clarity and symmetry of the remainder of that majority of the alleluias of which Wagner wrote is a matter of simple repetition schemes. There remains the question, when in its history did this notion of order enter the chant? Two sorts of evidence will be relevant here: that bearing on the age of the seventeen melodies, and variant versions of some of the melismas that suggest change in the direction of the form they ultimately assumed. Of the seventeen melodies of this study, only five are associated with texts that occur in the graduals of the eighth and ninth centuries included in Hesbert’s Antiphonale missarum sextuplex (Rome, 1967). It is reasonable, then, to judge the approximate age of the remainder according to their earliest appearance in subsequent graduals. Lacking, still, the concordances for the gradual, we shall have to rely on sources for which inventories are already available—apart from the Sextuplex sources, principally the publications of the Paléographie musicale and, to a limited extent, some isolated single publications. The obvious hazard here is that we may miss an earlier source for this or that melody. How decisive that hazard is, we may judge after we have considered the evidence. The seventeen melodies reproduced in the Appendix, most of them melismas, are cast in four groups according to the earliest occurrence of a text with which each is associated in the Graduale romanum.³⁹ Those groups are constituted as follows: Group I: earliest occurrence in a source with texts only—the Sextuplex sources, eighth/ninth centuries. Group II: earliest occurrence in a musical source showing the direction, but not the magnitude, of melodic intervals, ninth/tenth centuries. Group III: earliest occurrence in a musical source showing both direction and magnitude of melodic intervals, eleventh/thirteenth centuries. Group IV: the melodies reported here do not occur in any of the published sources and are presumed to be of later origin than those of I, II, and III. With regard to formal principle, the melodies constitute five types. Three of these are founded on the antecedent–consequent principle: the simple AA' (nos. 1–3, 5–8, 11, 12), the rounded ABA' (no. 4), and the bar-like AAA' or AA'A' (nos. 10 and 13). The fourth is the principle of the melodic sequence (no. 9). The last three melodies (nos. 15–17) embody a principle that is perhaps even more modern. Each involves an embellished linear motion expressed in an arrangement of short motifs that is directed towards the final. In no. 17 the motion is first to the confinalis, then to the finalis. I have underlaid all three melodies with reductions showing their outlines. While these melodies are not built in matched phrases, they share ³⁹ This cautious wording merely reflects the obvious: we cannot presume to have found the oldest source for a melody if we can read only the words.
124
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
with the others a forward motion towards a goal and a sure sense of the tonality. The examples suggest a tentative answer to our second question, regarding the location of the seventeen melodies, and the way of organizing music they manifest, in the long history of the alleluia. Five melodies are represented in sources of the eighth–ninth centuries (but only the first of these in the very oldest sources). Two melodies were introduced into the repertory in the ninth–tenth centuries, and three additional ones made their earliest appearance in the eleventh–thirteenth centuries. The remaining seven melodies evidently had their origins, in any case, after the eleventh century. That this is not too conservative an impression is confirmed by another circumstance. While the melodies of Group II cannot be read precisely with regard to the magnitude of their intervals, their parallel construction would be reflected in parallel neumatic groups, as in Fig. 5.1(a), a facsimile of a source for melody no. 5. However, early sources for melodies 1, 6, and 7 (Fig. 5.1(b–d )) show neumatic patterns that suggest something other than the clearly parallel construction of the Graduale romanum versions.⁴⁰ In all three cases a reduction of the line, with a resulting sharpening of the form, seems to have taken place sometime after these early versions were recorded. In one case, melody no. 6, we can see the difference quite clearly, for the melody had not yet assumed its final form when it was included in the eleventh-century Gradual of St Yrieix. Example 5.7 compares that version with the final one as it appears, for instance, in the thirteenth-century Sarum Gradual.⁴¹ We must presume from this that in the period of the Group II sources the structures of the Appendix had even less currency than the Appendix itself would suggest. And of course it underlines, too, the caution I have already expressed, that we cannot assume that the Group I melodies were sung in their final form at the time of the Sextuplex sources. To our conclusions regarding the relatively minor role played by goal-oriented structures among the alleluias, we must add the high probability that they entered the repertory gradually, and relatively late in its history. It seems fair to say, then, that composers who employed them widely elsewhere did not in any case take their models from plainsong. But perhaps one can read even further meaning from this apparently tentative introduction of a new way of organizing melodies in an era just preceding the invention of diastematic notations. Notations that represent the direction but not the magnitude of melodic intervals function largely as mnemonics. As such they are especially well suited to music in which standard melodic turns and patterns, already known to the singer, ⁴⁰ The reproductions are from the following published sources: a, c, and d, Paléographie musicale, ser. 1, i (Solesmes, 1889); b, PM, ser. 1, 10 (Tournai, 1909). ⁴¹ Paris lat. 903, published in PM 13 (Tournai, 1925), 242; Graduale sarisburiensis, ed. W. H. Frere (London, 1894), 147.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
125
Fig. 5.1. Four alleluia melodies: (a) alleluia Domine deus salutis, in St. Gallen 339, p. 128; (b) alleluia Ostende nobis, in Laon 239, p. 126; (c) alleluia Eripe me, in St. Gallen 339, p. 127; (d) alleluia Beatus vir qui suffert, in St. Gallen 339, p. 130
play a significant role. But as composers came to favour constructions founded in the modal scales and their pitch hierarchies, those notations would have lost their effectiveness. A pattern of neumes repeated very nearly to the last detail is really a poor representation for a phrase structure that depends for its functioning upon cadential variations. The taste for such constructions may well have constituted one of the evolutionary pressures that resulted in notations that represented the magnitude of intervals precisely. Underlying, and coinciding with this fundamental difference in the manner of composing melodies, is a sharp turn in the Western conception of mode, from Aurelian’s collections of formulae to the ordered pitch systems earliest discussed in the Alia musica. The change in the aims of notation of which I have spoken followed just on the heels of the new modal doctrine.
126
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
Ex. 5.7. The melisma from the alleluia versus Eripe me transcribed from the graduals (a) St-Yrieix (PM 13, p. 242) and (b) Sarum (Graduale sarisburiense, 147) (a)
(b)
In short, melodic style, its theoretical wellspring, and its graphic representation appear to have undergone fundamental modification in closely coordinated fashion.
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a
127
appendix Seventeen Alleluia Melody Fragments from the Graduale romanum Each fragment is accompanied by the following information: its location in the Graduale romanum (GR), the oldest sources, the incipit of a verse text with which it is associated in the latter, and the mode assignment. In general I have quoted only that portion of the melody that is involved in the formal procedures to which I have reference in the text. In some instances, however, I have thought it instructive to show the context in which that portion falls. Thus in no. 10, I mean primarily to show the bar-like arrangement of the phrases bbb', but also to show the order of the full alleluia section in which this fits. Group I 1.
A
GR 286, closing melisma of the verse. Graduals of Monza, Rheinau, Mont-Blandin, Compiègne, Corbie, Senlis. V Dominus in Sina. Mode 8.
8
A' 8
2.
A
GR [49] , beginning of the alleluia melisma. Graduals of Compiègne, Corbie, Senlis. V Justus ut palma. Mode 1.
8
A' 8
3.
A
GR 378, beginning of the alleluia melisma. Graduals of Compiègne, Corbie, Senlis. V Paratum cor meum. Mode 3.
8
A' 8
4.
8
A
Al B
8
ij. A'
8
le
lu
ia.
GR [30]. Graduals of Compiègne and Corbie. V Te martyrum. Mode 5.
128 5.
O n t h e St ruc t u re of A l l e l u i a M e l i s m a A
GR 351, end of the alleluia melisma. Graduals of Compiègne, Corbie, Senlis. V Domine Deus salutis. Mode 3
8
A' 8
Group II 6.
A
GR 343. St. Gallen 359 and 339, Laon 239, Chartres 47. V Eripe me. Mode 2.
ij. A'
7.
A
GR [43]. St. Gallen 359 and 339, Einsiedeln 121, Laon 239. V Beatus vir qui suffert. Mode 1.
8
A' 8
Group III 8.
8
A et sic
ut
GR [19], within the verse. Montpellier H 159. V Sancti tui Domine florebunt. Mode 8.
o A'
8
dor
bal sa
mi
9.
8
Al le A
lu
ia.
8
A1 8
A'
ij. A'
GR 340. Gradual of St Yrieix. V Magnus Dominus. Mode 7. (This melody has been cited by Bruno Stäblein as a fully developed specimen of Wagner’s second category, in the article ‘Alleluia’, MGG, i. 345.)
O n t h e S t ruc t u re of A l l e l u i a M e l i s m a 10.
8
A
Al le b
lu
GR 652. Sarum Gradual. V Beatus vir sanctus Martinus. Mode 5.
ia. ij. b
129
b'
8
A 8
Group IV 11.
8
A
GR 39**. V Gloriosus Deus. Mode 4.
ij. A
8
12.
A
8
ex cor A'
8
ut
de
om
13.
8
con
nemnar ra
jun
ti o nem De
A
et
re
re:
i.
GR [146], within the verse. V In Deo speravit. Mode 1.
flo A
8
ge
GR 528, within the verse. V In multitudine. Mode 8.
ru it
ca
A'
8
14.
8
A V Sel
GR [139], within the verse. V vum me
A' 8
(fac)
De
us
fac,
130
O n t h e St ruc t u re of A l l e l u i a M e l i s m a
15.
Salvum me fac. Mode 3 [sic]. GR 510, the end of the alleluia melisma. V Videbitis. Mode 8.
8
8
16.
8
e van ge
li
zan
ti
bus
GR 59**, within the verse. V Dominus dabit. Mode 6.
8
17.
8
8
8
8
8
8
Al le
lu
ia,
ij.
GR 609. V Sancte Michael Archangele. Mode 8.
Homer and Gre go ry
131
chapter 6
Homer and Gregory: The Transmission of Epic Poetry and Plainchant
} The chanting of the medieval Latin church—ritually ordered intonations of scriptural texts that had been edited as librettos for chanting—was a kind of traditional practice specific to local liturgical communities but under pressure since the Carolingian era to cleave to the practice of Rome—as the touchstone of authenticity and correctness—and thereby to constitute a uniform, universal practice with a unity that would correspond to the political and cultural unity of the Holy Roman Empire. That practice, represented by language- and music-writing in books that have come down since the tenth century and after, came to be known in the ninth century as ‘Gregorian Chant’, on the grounds of a legend regarding Pope Gregory the Great (590–604) and the origin of the body of chant named after him. This chapter confronts the implications of that legend with an attempt to work out a detailed account of how such a prolix body of traditional song can have come into existence and have been disseminated in an oral culture. Like the practitioners of any traditional music the singers of chant composed their melodies following overall formal models and patterns and calling on specific formulaic units of melody at appropriate points, all of these having emerged from the practice itself, and all as native and natural to the singers as their mother tongues. Their melodies can be somewhat unique and individual, or somewhat less so regarding individual chants, depending on the genre of the chant. In any case uniqueness and originality seems not to have been singled out by them as a special value, as it has been in other cultures. Many of the melodies cluster in types marked by common features: opening and closing patterns and turns of phrases; ways of creating hierarchies within the medieval pitch array—the diatonic articulation of the pitch spectrum; ways of creating a dynamic of tension and resolution, climaxes and cadences; the shapes of individual phrases and the overall progression of phrases; and stereotyped formulae, which are marked by their melodic details and by the place in the overall dynamic of the melodies in which they are found—that is, by their function. What this comes down to is likeness in configuration and principles of coherence, and likeness in the melodic stuff, two categories that, however,
132
Homer and Gre g ory
flow into one another. All these features have evolved from within, through the practice itself. The clusters or families of ritual song tend to retain an association with their ritual use—their place in order of liturgical service of the mass or the office. Differences between individual melodies of a type can very often be recognized as responses to the details of the words they were made to intone. The makers of chant followed quite specific principles in adapting melodies to words, paying attention to all parameters of language—semantic, syntactic, and phonetic. The words were thus among the principal determinants of the overall progression as well as the fine details of the melody. At the same time the chant composers’ adaptation of the melodic conventions and materials of the tradition to the words was a principal arena for the exercise of their inventive artistry. This chapter reports my effort to develop a plausible representation of the way that chants might have been made that would at the same time be a way into a critical account of them as sung language. Its objective is genetic criticism, an objective well suited to the study of a tradition whose products are not appropriately regarded as autonomous works. That is more nearly possible through the study of such song families than of more individual songs because of the opportunity for comparative study, which is the principal method of analysis employed in this book. This choice is not meant, however, to encourage the impression that Gregorian chant composition was entirely a matter of the realization of such sets of patterns, principles, and formulae, or that such a representation would account for all chants that have come down to us in writing. The language of this statement of purpose is carefully chosen to make clear the hypothetical status of what is proposed. I believe it cannot move out of this status into certainty, but I do regard it as a hypothesis superior to the principal alternative view, which I discuss in Chapter 7. My representation had to satisfy another criterion. The tradition of Gregorian chant was not initially dependent on any kind of musical notation. It thrived for the first several centuries in cultures that knew no musical notation of any kind. There is general consensus about this among scholars, and although there are insinuations and implications here and there in the modern literature about a use of notation, or perhaps proto-notation, from the beginning, no one has lately supported such an idea explicitly. My representation had to be plausible in the sense that it had to represent a way of making chant that could function without notation, and so I speak of ‘oral composition’. I believe that it is plausible in that sense, and that therein lies one of its principal advantages over the main alternative view, which has been formulated as though for the practice of a writing and reading community. We cannot say, and do not need to say, that the properties that are highlighted in my analyses are proof of an oral practice, only that they can plausibly be interpreted as the modalities and the impress of an oral practice. There is a perplexing question about the extent to which the tradition became dependent on notation after its introduction in the ninth century, together with its corollary, whether and to what extent chant performance and transmission
Homer and Gre go ry
133
remained in any sense orally based once singers had access to notated books. No uniform answer can ever be given to these questions. On one side there is testimony from the ninth century and after from musicians who saw notation as a more efficient alternative to oral tradition—the imitation of a chanting-master— for teaching and transmitting melodies; on the other there is testimony about continued reliance on memory from the same ninth century, when the technology of notation was already known, on through the sixteenth century, when church singers could still be prohibited from using their books while chanting. In a summary essay included in a pedagogical book of survey essays, cautious therefore in avoiding controversial claims, James McKinnon writes: The Gregorian chant was also transmitted orally to the north, and it continued, in a sense, to be orally transmitted long after the advent of notated chant books, whether these existed from about 800 or 900. It is generally agreed that for a number of centuries chant continued to be sung at liturgical services without benefit of notated chant books, while the books—those that survive are too small to have been of any use in actual performance—must have existed somehow as reference works, perhaps to be consulted as a memory aid before the service.¹ Perplexing as they are, these questions are marginal to the main purpose of this chapter, which is to offer a representation of the making and performing of chant without the use of notation, whenever that may have been. The reception of this essay and of subsequent articulations of its central idea—in publications of mine and also of Helmut Hucke’s—has been marked by some uneasiness, and at the core of the discomfort is the implicit suggestion that some have read in our writings that chant performance continued to be a matter of oral composition or improvisation after the Frankish adaptation of Roman chant and even after the appearance of notated chant books, and that consequently performance would have continued to be marked by contingency and variability. That implication does not square with the widely shared belief that the chant had become stabilized by this time, with the variability smoothed off either gradually through practice or removed in a concentrated compositional project. But as McKinnon reports, I have ‘suggested rather narrow limits of chronology and liturgical genre for the application of the theory’.² I imagine that two factors contributing to the contrary reading by some are the collapsing of ‘oral composition’ and ‘improvisation’ into a single concept, which in turn implies an ‘unstable’ transmission, and my subsequent writing about trope transmission, where stability seems mainly not to have been a primary value and ‘reconstruction’ seems the appropriate way to characterize the transmission. The first factor, based on a modern idea about ‘improvisation’, is discussed in Chapters 1 ¹ ‘The Emergence of Gregorian Chant in the Carolingian Era’, in James McKinnon (ed.), Antiquity and the Middle Ages: From Ancient Greece to the 15th Century (Music and Society; Englewood Cliffs, NJ, 1990), 111–12. ² Ibid. 111.
134
Homer and Gre g ory
and 2, the second in Chapter 11. As to the second, the modes of transmission of chants and of tropes or songs of other paraliturgical genres are incomparably different—politically, ecclesiastically, and musically, as I show in Chapter 11. But the distance between the two viewpoints at play here diminishes with the recognition that the medieval techniques of oral composition and of chanting from memory were far more alike—even continuous—than has been imagined from the vantage point of the modern conception of ‘memorization’. We cannot draw a sharp line between the two as cognitive processes and then use that line to mark off stages in the history of chant tradition. This can be put simply: the same melodic properties that would have facilitated composing the chants would have facilitated remembering them. It was in order to encourage recognition of this that I included in the original essay a summary of the classic theory of memory developed by the social psychologist Charles Bartlett that led the fundamental change in thinking about memory from the simple and commonplace notion that experiences or poems or songs or images are fully recorded and stored in a fixed state in a memory bank and later retrieved in that state on demand in remembering, towards the general idea that remembering is an active process of construction or reconstruction on the basis of formal schemes, salient details, and cues. Bartlett’s general conception was widely shared when I wrote the original of this chapter, and it has since received additional reinforcement and refinement from investigations in three areas: experimental cognitive psychology, surveys of oral traditions focusing on the role of memory, and the study of medieval writing on memory. I cite brief representative formulations from those areas. First, some conclusions from Daniel Schacter’s Searching for Memory: The Brain, the Mind, and the Past (New York, 1996): Scientists agree that the brain does not operate like a camera or a copying machine . . . The cognitive psychologist Ulric Neisser called the idea that faithful copies of experience are kept in the mind, only to reappear again at some later time pretty much in their original form, the ‘reappearance hypothesis’. Neisser proposed instead that only bits and pieces of incoming data are represented in memory. These retained fragments of experience in turn provide a basis for reconstructing a past event, much as a paleontologist is able to reconstruct a dinosaur from fragments of bone. ‘Out of a few stored bone chips’, reflected Neisser, ‘we remember a Dinosaur.’³ Further from Schacter: How are we to think about what is retrieved when we recall a past experience? Does the act of retrieval simply serve to activate, or bring into conscious awareness, a dormant memory? . . . I have already suggested an alternative possibility, rooted in Neisser’s analogy that retrieving a memory is like reconstructing a dinosaur from fragments of bone (p. 69) . . . for the rememberer the engram (the stored fragments of an episode) and the memory (the subjective ³ Ulric Neisser, Cognitive Psychology (New York, 1967), 285; the passage from Schacter’s book is on p. 40.
Homer and Gre go ry
135
experience of recollecting a past event) are not the same thing. The stored fragments contribute to the conscious experience of remembering, but they are only part of it (p. 70) . . . The idea that there is a one-to-one correspondence between a bit of information stored away somewhere in our brain and the conscious experience of a memory that results from activating this bit of information is so intuitively compelling that it seems almost nonsensical to question it. Yet scientists who study memory and theorize about it are increasingly skeptical of this idea . . . When we remember, we complete a pattern with the best match available in memory; we do not shine a spotlight on a stored picture . . . We must leave behind our familiar preconceptions if we are to understand how we convert the fragmentary remains of experience into the autobiographical narratives that endure over time and constitute the stories of our lives. (p. 71) Similar scepticism about the ‘intuitively compelling’ commonplace notion of memorization and recall is expressed near the outset of David C. Rubin’s book Memory in Oral Traditions: The Cognitive Psychology of Epic, Ballads, and Countingout Rhymes (Oxford, 1985): Drawing from the literature on oral traditions as a whole, transmission cannot be viewed as verbatim recall of the fixed texts of literate readers. Remembering a piece from an oral tradition does not require the recall of the exact words: recall of the general meaning and form is sufficient (p. 6) . . . In many situations, oral traditions provide a more appropriate model of everyday human behavior than do psychological experiments on memory. At times, people do have to remember what exactly happened on one particular occasion, as in eyewitness testimony. It is much more common, however, to abstract and remember the structure from many similar events, no one of which by itself is the best version. (p. 7) Previewing his unfolding account, Rubin writes: Following Parry and Lord, oral traditions are viewed as human behavior, not as reified texts . . . Oral traditions depend on human memory for their preservation (p. 9) . . . Oral traditions must, therefore, have developed forms of organization (i.e., rules, redundancies, constraints and strategies to decrease the changes that human memory imposes on the more casual transmission of verbal material) (p. 10) . . . The cognitive factors discussed so far have focused on the structure of pieces in oral traditions, but the process of transmission also needs to be considered . . . Recall of a piece in an oral tradition is serial. It starts at the first word and proceeds sequentially to the end . . . Thus what is sung cues what remains to be sung, and the various cues that unfold as the song is sung help distinguish an item from all others in memory . . . This local, implicit, word-by-word or phrase-by-phrase cueing, which is the dominant form of cueing in oral traditions, is easiest to discuss as a network of associations . . . (p. 12)
136
Homer and Gre g ory
The recent research that has the most direct bearing on our subject is Mary Carruthers’s detailed study of the uses, functioning, and conception of memory in A Book of Memory (Cambridge, 1990). Carruthers reports what she finds throughout medieval writing on memory: testimony to its absorbing, integrating, transforming, and organizing action. It is regarded not as an alternative to creativity but the route to it. It is the basis of extempore performance, but not in the sense of parroting what has been learnt by rote. Such performance is fluent but orderly. Carruthers writes ‘how greatly we misunderstand when we reduce ancient and medieval memoria to our word “memorization”’. Max Haas throws a different light on the subject, inviting us to test the concept behind this word with a mental experiment.⁴ Imagine the task of ‘memorizing’ the Manhattan telephone directory. Our protest that we can’t do it—which he takes for granted—is an indication that we do not in practice fix in memory sequential arrangements of signs that do not relate to one another except by some arbitrary ordering system (e.g. the alphabet). Memory depends on the sense that the elements make in relation to one another and in the context of the whole. There isn’t an independent faculty of memory sweeping over its material and sucking it up like a vacuum cleaner, there is an interdependence between memorization and the stuff that is to be memorized and with the social context in which the learning by heart takes place.⁵ Haas uses the expression auswendig lernen, which is the German equivalent of ‘to learn by heart’. Comparing the two expressions, which seem so different at first glance, is illuminating. Aurelian of Réôme spoke in the ninth century of the obligation of every cantor to have ‘ingrafted by memory in the sheath of his heart the inflection of all the verses through all the modes, and the difference between the tones for the verses of the antiphons, introits, and of the responses’.⁶ Notker of St. Gallen (d. 912) wrote in the preface to his collection of sequences, ‘When I was still young and very long melodies repeatedly entrusted to memory escaped my unsteady little heart, I began to reason with myself in what way I could bind them fast.’⁷ The heart has long been thought of as the seat of understanding as well as of feeling. The expressions of Aurelian and Notker are witness to the long-held belief that knowledge in memory is deep-seated knowledge, and whether we are now conscious of that or not, it is a sense that survives in our expression to know something ‘by heart’. The German expression auswendig lernen is a truncation of the expression ‘to know something inwendig und auswendig’, or as we say, ‘inside and out’—to know it thoroughly. In that case too, then, an expression for thorough, deep-seated knowledge has come through to modern usage as an explicit expression for knowing from memory. ⁴ Mündliche Überlieferung und altrömischer Choral, 37–9. ⁵ This observation has been made by Dobszay with specific reference to the genres of Gregorian chant. See the introduction to Ch. 2. ⁶ Joseph Perry Ponte, III, ‘Aureliani Reomensis Musica Disciplina: A Revised Text, Translation, and Commentary’ (Ph.D. diss., Brandeis University, 1961), ch. 19. ⁷ Translation by Andreas Haug, in ‘The Functions of Musical Notation in the Early Middle Ages’,
Homer and Gre go ry
137
Haas draws attention to the way that the modern concept of an independent memorizing faculty, as it entails the idea of putting contents in storage, implies a storage capacity of finite proportions. He cites Kenneth Levy’s calculation that the Gregorian mass proper and office chants would require some eighty hours of performance time, which Levy says is comparable to the time required for the symphonies, quartets, and sonatas of Beethoven, or the operas of Wagner from Lohengrin to Parsifal. This exceeds human capacity, asserts Levy (‘Gregorian Orality’). Haas then cites Rubin’s calculation that the repertory of a certain Serbian epic bard would run to about 600 hours of performance time (p. 139). But this is not in order to correct Levy’s estimate of the limit of human memory capacity; it is to suggest that these calculations just do not make any sense, for no more than memory is demonstrably an independent faculty, does the storage capacity of memory have an independently fixed limit as though it were something like a grain silo. How much is remembered depends, again, on the interdependence between the stuff that is to be learned from memory and the techniques by which that learning is accomplished. In pursuit of that subject Haas takes two things for granted: the stuff of memorized material in the oral tradition was grasped in a way that was instrumental for the spreading of the tradition; and the eventual use of writing did not have as a consequence the abandonment of the mnemotechnical apparatus. This second point has a wider meaning than appears at first glance. For the features that had mnemotechnical utility were not marked as such. That is they were the very stuff itself, not some attached handholds for the memory that could be dropped as superfluous because of writing. If a particular corpus in oral tradition has an arsenal of formulae that arose in response to mnemotechnical needs, that says something about their genesis. But as sayable or singable groups they must have had a sensible function within the context of an epic or song that went hand in hand with their mnemotechnical function. When such material is written down its mnemotechnical markers are retained but tend to fade from view if they are no longer in a context in which they are understood to serve that function. In any case we have only written clues for memory achievements, clues that are not marked for us as such at first glance; we must newly make them visible in that light. Thought about the role of memory on the commonplace notion of ‘memorize’ and ‘memory bank’ ⁸ functioning in the oral tradition rests on the presupposition of the stabilization of the melodies before they came to be written down, but that notion has also been recruited in the service of this historical theme. The relationship of interdependence between the two ideas has given both an a priori aspect, providing mutual support where each seems beyond the reach of confirmation and even comprehension. To speak of a stabilized corpus of melodies in an oral unpublished paper for the round table ‘Schrift und Verschriftung’, during the conference ‘Karl der Große und Papst Leo III. in Paderborn’, Paderborn, 1999. ⁸ Kenneth Levy, in ‘Charlemagne’s Archetype of Gregorian Chant’, JAMS 40 (1987), 1–30 (repr. in Gregorian Chant and the Carolingians, ch. 4), writes that Aurelian’s descriptions of chants were ‘unlikely to be a scroll through a memory-bank’ (p. 10).
138
Homer and Gre g ory
tradition is to say, more or less, that every performance of each melody was substantially the same as every other one according to some criterion. But how would we know that such was the case, and as I ask in Chapter 2, how would we know what would have counted as ‘the same’ to singers 1,000 to 1,500 years ago, or what the criterion of ‘sameness’ would have been? We have difficulty enough judging about that when there are written scores, as I show in Chapter 11. I now believe that I was too ready to adopt the stabilization theme in interpreting the evidence that is presented in section VI of this chapter, and that the two questions—that of the role of memory and that of the stabilization of the repertory before the writing down— should be addressed separately, each on its own evidence and reasoning. Having revisited the theory of memory in the foregoing, I turn now to the topic of the stabilization, or as I prefer to say, fixity, of the Gregorian melodies attained in the oral tradition.⁹ While we cannot say in any concrete way what that would have entailed in the details of practice, we can recognize indications of various kinds that participants in the pre-literate musical culture behaved as though they were dealing with a practice that they regarded as being in some respects fixed, at least ideally. They are only indications, none conclusive. 1. Many chant texts, as I have observed, are adaptations of biblical passages, evidently made for singing.¹⁰ It is a reasonable inference that these were prepared as part of a compositional act that included music. Putting it the other way around, we might ask what would have guided the adapters of the texts in their editing, if not what they had in mind about how they would be sung? 2. Hucke has considered the oldest tonaries from the point of view of this question. The very oldest, from St Riquier, c.800, is without notation, lists chants by mode, but gives only a handful of examples for each genre in each mode. Still, it is a reasonable inference that the existence of the rest and the ability of cantors to match the rest to the examples was a precondition for their ability to function as examples. The next oldest, that from Metz, c.870, is also without notation, but as it classifies the melodies by mode and within modal categories by the psalm-tone cadences from which they continue, it gives reason to think that they refer to melodies with more or less fixed beginnings, at least. Aurelian’s tonary of about the same time provides very sparse notation, but since it describes many details of individual chant melodies it refers to individual, citeable pieces with musical identities, and his descriptions correspond to notations of those pieces in later sources.¹¹ 3. Aurelian and Notker, in the passages cited above, bear witness to the value placed on the fixing of chants in memory, but we do not know how far that corresponds to our notion of fixity. 4. The inference under (1) is encouraged by Hucke’s demonstration (in ⁹ The reason for my preference: if the talk is about ‘stability’ there is a certain bias about that term that inevitably leaks through. Stability is valued; nothing unstable—the foundations of houses, personal relationships, personalities, the stock market—is good. ¹⁰ Kenneth Levy has dubbed them ‘libretti’ (‘Toledo, Rome, and the Legacy of Gaul’, EMH 4 (1984), 49–99), repr. in Gregorian Chant and the Carolingians, 31–81. ¹¹ Hucke, ‘Gregorianische Fragen’.
Homer and Gre go ry
139
‘Gregorianische Fragen’) that the semantic, phonetic, and syntactic properties of the prepared chant texts were determining factors in the provision of melodic settings. Whether or not that was part of the same compositional act, the application of such principles would itself have been a factor for fixity. 5. In section VI of this chapter as well as section VI of Hucke’s ‘Gregorianische Fragen’ we encounter anomalies resulting from the mixture of recognizably different melody types in the same chant or from radically drifting formulae, which were nevertheless accepted into the standard repertory and came into the written transmission as individual melodies. But they are also witness to a practice of chant performance through invention and re-invention with traditional material and principles, a practice that made the anomalies possible. It is important to note that these phenomena, as well as those identified under (4) are indications of both the fixity of individual chants and the techniques of oral composition through which they were produced. The two are not contradictory. 6. A rather special kind of evidence of fixity was presented by Karlheinz Schlager, who showed that poets who applied words (one syllable to each note) to melismas transmitted in the earliest sources rather painstakingly preserved in their poetry the melodic elements that are denoted by note groupings in the neumes of those melismatic sources. This they contrived through the determination of word boundaries in their texts: word boundaries coincide with what are neume boundaries in the melismatic originals, especially striking when that is the case in different texts fitted to the same melisma. (See Ex. 6.1, in which such a melisma is shown as it appears in three neumatic sources and transcribed from the alphabetic notation of a fourth, and Ex. 6.2, in which four texts are supplied to the same transcribed melody.) Vertical lines in the transcription (Ex. 6.1) indicate agreement about neume boundaries among the four sources; vertical lines in the texts of Ex. 6.2 indicate agreement about word boundaries despite difference of word contents. This allows us to see the correspondence of neume boundaries with word boundaries. That suggests that the poets knew those melodic groupings as the constituent elements of the melodies, fixed as such prior to their writing down. That would mean that the earliest notations of such melodies transmitted the immanent groupings within the melodies by means of the neumes. Of course this practice has something important to teach us about what it was that early neumatic notations were designed to convey. The syllabic setting necessitated the resolution of the neumes, which were replaced by single-note neumes. That reduced the information about the melodic contour that was carried by the melismatic neumes. But because of the scruple observed by musical practitioners regarding neume boundaries and word boundaries, the aggregative property of neumes with respect to their constituent notes was shifted, in a way, to words. The words themselves thus could have conveyed the information about note groupings that would have supported the singer’s memory in performance. As Schlager suggested, the words could have functioned as a kind of musical notation. Another way of putting this is that what was represented of the melody or its singing was evidently to include the immanent property of note
140
Homer and Gre g ory
Ex. 6.1. Melisma on ‘dierum’ in the second verse of the offertory Deus enim firmavit, in Montpellier H 159, fo. 148; Einsiedeln 121, fo. 29; Rome, Bibl. Ang. 123, fo. 29r; and Paris lat. 776, fo. 13r
grouping that was shown by the neumes. When these were forced to break up into their constituent notes by the syllabic declamation of the added words, their function was taken over by the words themselves.¹² 7. Andreas Haug has reported on a German sequence collection with neumes, dated to about 1100, that contains pieces of local distribution. In the margins, next to the beginnings of such pieces, in arrow-shaped boxes that point to those beginnings, the scribe has written, also with neumes, the beginnings of sequences that circulated internationally (see Fig. 6.1). The neumes of these incipits correspond to those of the beginnings of the pieces in the main text, and the scansion of the poetry corresponds exactly. The interpretation is clear: each such marginal entry is an instruction to the singer to sing this text to the well-known tune of such-and-such. It puts him on the track of the melody, and his remembering of the melody guides him in the reading of the notation, which is also an act of reconstruction. But at the same time the notation supports him in his remembering of the melody. This ¹² See Karlheinz Schlager, ‘Die Neumenschrift im Licht der Melismentextierung’, AfMw 38 (1981), 296–316. Three other arguments to the same conclusion that cannot easily be summarized here because of space limitations are presented by David Hughes, in ‘Evidence for the Traditional View of Gregorian Chant’ (see esp. p. 394), Kenneth Levy in ‘On Gregorian Orality’, repr. in Gregorian Chant and the Carolingians, 141–77 (where this conclusion is the main thesis), and Ch. 5 of this book.
Homer and Gre go ry
141
The melody of Ex. 6.1 with prosula texts from Bamberg lit. 5, fo. 4r (A); Rome,
Ex. 6.2. Bibl. naz. lat. 1343, fo. 19v (B); Paris lat. 776, fo. 13r (C and D), after Schlager, ‘Die Neumenschrift’
142
Homer and Gre g ory
Fig. 6.1. Sequence Diem festum de sancto Bartholomeo, in Kremsmünster 309, fo. 214v
Homer and Gre go ry
143
interesting phenomenon of written transmission points to what must have been a fundamental relation of mutual support between reading and remembering in the performance of music after the entry of music writing into the picture.¹³ To this point I have tried to register indications of fixity that step forwards and impress themselves on us here and there out of the material that comes down from the daily practice of and commentary about chanting. It requires close observation to recognize them, but not extraordinary ingenuity to force them out. They just do give the inescapable impression that by the time an entire corpus of mass proper chants for the ecclesiastical year was put into writing, the practice of chanting had resulted in the stereotyping of aspects of the practice. That is hardly surprising: people who engage in repetitive ritual practices form habits. But this is not to go so far as to say that the writing down was just a matter of transcribing a corpus of melodies (the texts already existed in writing) that was fixed as a whole from the outset or that had in any case become fixed in some comprehensive way before it was written down. That is the purport of three theories of different forms that I believe exceed the limits of knowability in this, as any, historical investigation. I must step back to consider the problem. To begin, here is a question to be put to any such theory: what, in practical terms, can the claim that the repertory had attained fixity by the time of its initial writing down mean? Presumably that for every chant, every performance would have been like every other performance with respect to the properties that are thought to have become fixed (usually the interval contents of the melody). How would we know that? Perhaps from the comparison of notational representations of the same items in different early notated sources. But that works only if we believe those represen¹³ Sequence Diem festum de sancto Bartholomeo in Kremsmünster 309, fo. 214v. Andreas Haug, ‘Zum Wechselspiel von Schrift und Gedächtnis im Zeitalter der Neumen’, Cantus Planus, 33–47. Haug reports on similar practices in other sources on p. 36 n. 10. A similar situation in the Old Roman antiphoner London Add. 29987 has been reported by Edward Nowacki in ‘The Gregorian Office Antiphons and the Comparative Method’, JM 4 (1985–6), 243–75, esp. 273–4. It concerns a series of antiphons for the Easter season on only the word Alleluia, repeated. Each antiphon is preceded on the page by the beginning of a fully-texted antiphon assigned to another feast (completely written out in another part of the manuscript), whose melody it resembles very closely but not fully. It might seem that here, too, the incipit was an indication to the singer to follow the model of its melody. But the differences yield further insights into the nature of the composition–performance process in such circumstances. Nowacki was able to show that the similarity between alleluia antiphons of the same melodic type is often closer than that between any alleluia antiphon and its fully texted model, but also that one fully texted antiphon may resemble all the alleluia antiphons of its type more closely than any other; i.e. it is the best representative of that melodic type. That means that, despite the fact that the scribe wrote out the incipits of different melodies, it was a melodic archetype more than the particular melody of the incipit that guided him in writing out the corresponding alleluia antiphon. And we can only guess what would have guided the singer using that book in the performance of any alleluia antiphon: the melody cited in the incipit, the notated melody, or some internalized sense of that melodic archetype. What we cannot assume is that it must have been all one or another of these. Remembering, reading, and remaking must have been continuous with one another, and, probably, indistinguishable to the singer. Here again the conclusions from the musical evidence resonate closely with the theory and practice of memory in the Middle Ages. Carruthers (The Book of Memory) writes, ‘In none of the evidence I have discovered is . . . writing itself a supplanter of memory. . . . Books are themselves memorial cues and aids’ (p. 16).
144
Homer and Gre g ory
tations to be snapshots of performances made just before the writing down in the corresponding establishments. That would be easy to believe if we knew those performances to have been already uniform. The circularity is manifest. Skipping over it nevertheless, what, as I have already asked, should be the criterion of likeness that we would need in order to interpret the claim? Hopefully it would be one that the practitioners would have recognized. Can we hope to corner it by using our own judgement about what are significant and insignificant differences? Oliver Strunk recognized this as ‘the recurrent central problem of early Christian music’: ‘How can we control the evidence of our earliest manuscripts? To what extent does their melodic tradition reflect that of earlier times? . . . Throughout the early Christian world an impenetrable barrier of oral tradition lies between all but the latest melodies and the earliest attempts to reduce them to writing.’¹⁴ Despite this famous remark of Strunk’s, a great deal of research activity—perhaps even the main focus of historical chant research—has focused precisely on this time of unwritten tradition. Efforts to reach behind the notated manuscripts have depended largely on a principle derived essentially from the adaptation to music of the dictum formulated by the liturgist Walter Frere: ‘Fixity means antiquity, that is to say that if the same formulary appears in many sources it must be relatively old. It is described, moreover, as having “stability”, whereas “instability” exists where various texts appear from source to source for a particular liturgical item.’¹⁵ McKinnon translated the principle this way for chant transmission: ‘If one compares the same chants as they appear in a generously representative selection of early notated manuscripts and finds them to be nearly identical, then one must conclude that the chants existed in substantially the same form for a considerable period before the manuscripts were written’ (p. 99). McKinnon reaffirmed the Frere principle in his last book as the main support on which his portrayal of the musical aspect of the late seventh-century creation of the Roman mass proper rests.¹⁶ Without its support he could hardly have represented Roman chants of the late seventh century with specimens drawn from notated eleventh-century Gregorian sources. Two other principles are invoked in the book in support of this possibility: McKinnon’s ‘opinion that the intention of the Roman singers was to create fixed and stable chants’ (p. 393; he cautiously adds ‘whether they were able to realize this intention [is] open to question’); and the principle that when chants were assigned to specific feasts it was both text and music that was provided, a reasonable enough supposition, but perhaps not secure enough to assure us that an eleventh-century source transmits melodies that were composed four centuries earlier. McKinnon’s is the first of the three comprehensive theories about the fixity of the repertory before the writing down to which I referred. As influential as the ¹⁴ ‘St. Gregory of Nazianzus and the Proper Hymns for Easter’, in Oliver Strunk, Essays on Music in the Medieval World (New York, 1977), 60–1. ¹⁵ As formulated by McKinnon, ‘The Emergence of Gregorian Chant in the Carolingian Era’, 105–6, based on W. H. Frere, The Sarum Gradual and the Gregorian Antiphonale Missarum (London, 1895). ¹⁶ The Advent Project, 154–6.
Homer and Gre go ry
145
Frere axiom on which it mainly rests has been, it is far from foolproof. To rely on it we would have to rule out other ways of accounting for the ‘near identity’ of written versions of the same chants, for example that it results from the wave of manuscript copying that must have washed over Europe after the invention of notation. A concrete instance of such an alternative hypothesis has been put forward by Kenneth Levy (although he has not offered it as an alternative to the Frere principle, which he also endorses). Levy has been arguing for ‘the existence of an authoritative [earlier in the same passage he writes ‘definitive’] neumed recension of the Gregorian propers ca. 800’—in a word: an Urtext. He calls this Charlemagne’s Archetype of Gregorian Chant, the title of the essay in which he first published the hypothesis (p. 1). And he calls the surviving neumed manuscripts of the late ninth and tenth centuries its ‘early descendants’ (p. 4). If the versions of chants found in those manuscripts are alike, Levy’s hypothesis would direct us to attribute that to their stemmatic derivation from the archetype as the simplest explanation. We could not at the same time take that likeness as evidence for their greater antiquity than the archetype, yet McKinnon affirmed both the Frere principle and the persuasiveness of Levy’s hypothesis.¹⁷ Whatever the merits of Levy’s evidence and reasoning, there is a more general question that makes them moot and that has not, to my knowledge, been explicitly addressed: what role are we to imagine this hypothetical book playing in the historical circumstances in which it has been set? Levy writes: In my view the neumatic notation is likely to have been employed during the later eighth century in effecting the changeover from Gallican to Gregorian musical repertoires, and the Gregorian propers, whose descendents we know in the Editio Vaticana, [presumably via ‘the surviving neumed manuscripts of the late ninth and tenth centuries’] would be a fruit of Charlemagne’s Carolingian Renaissance. (p. 25) What kind of neumatic notation would this have been? If we are to place it in the context of the early history of chant notation as far as we know it, we would probably say adiastematic neumes. But it is widely understood that scores written in such neumes were serviceable for ‘reading’ only by singers who already knew the melodic traditions to which they referred . How can such neumes have been instrumental in indoctrinating singers in a foreign melodic tradition? How can a book written with such a notation have been ‘definitive’ and ‘authoritative’, as though it had been the Editio Vaticana itself? Levy did apparently recognize the problem that ¹⁷ Serious questions have been raised about the interpretation of evidence along the path to this hypothesis. See my ‘Communication’ in JAMS 41 (1988), 566–75; Hucke, ‘Gregorianische Fragen’; David Hughes, ‘The Implications of Variants for Chant Transmission’, in Peter Cahn and AnnKatrin Heimer (eds.), De Musica et Cantu: Studien zur Geschichte der Kirchenmusik und der Oper. Helmut Hucke zum 60. Geburtstag (Hildesheim, 1993), 65–73; and A. Marcel J. Zijlstra’s communication ‘On “Abbott Helisachar’s Antiphoner” by Kenneth Levy’, JAMS 50 (1997), 238–42. I leave them aside here in order to focus on the larger historical problems about the hypothesis that arise even if those interpretations are granted.
146
Homer and Gre g ory
this poses for his hypothesis, and to solve it he extended the hypothesis into the domain of the history of music writing, marshalling the conjectured ancestors of the dispersed field of early specimens of neume writing in an orderly succession of ‘scenarios’. A ‘graphic’ notation of the type of palaeofrankish neumes, dated to the 760s or 770s but with origins perhaps reaching back to the time of Gregory the Great, exact enough as to pitch position to be adopted by a conjectured ‘commission charged with the responsibility of producing a pilot neumation’, served for an ‘initial transfer of chants from oral delivery to written record’, to a ‘scholarly or scientific text’. Underlying this narrative must be the presumption of a fixed corpus of chants that is being thus ‘transferred’. Levy’s is the second of the comprehensive theories of the fixity of the chant corpus as a whole before the initial writing down. In a second stage the ancestors of the remaining types of the earliest neumatic scripts, ‘gestural notations’, were created—this ‘could have been accomplished in weeks’—for the ‘promulgation of the authoritative Frankish-Gregorian neumed antiphoner’, a ‘performing edition’ that Levy dates to the 770s–790s.¹⁸ This seems to subvert the explanatory purpose of the hypothesis; most students of this subject share the view that ‘performing editions’ is just what the earliest existing notated chant books, written at least a century after the 790s, were not, and as their neumes are adiastematic the problem to which the hypothesis seems tacitly directed remains. The ‘scientific text’, with its graphic neumes, would have been better suited to the purpose. Perhaps it was ruled out because the story also requires that the earliest existing chant books, all written with (adiastematic) ‘gestural’ neumes, should count as the descendants of Charlemagne’s archetype. One problem about this narrative is that it lacks the richness and complexity of real life: it spins out a single plot line, following the stages of a single development. What we know about the ‘graphic’ palaeofrankish notation and the ‘gestural’ adiastematic neume scripts is that they manifest fundamentally different ideas about how to represent melodic turns on parchment. We know nothing about their chronological ordering and have no reason to arrange them in a historical continuum (see the introduction to Ch. 13, especially the reference to Susan Rankin’s observations on this subject). There is something striking about the state of the historiography of the field in all this. Levy’s main hypothesis, regarding Charlemagne’s ‘archetype’, is presented on a structure of evidence that is palpable, concrete, and known to have existed, all of it newly brought together to bear on the question of the time when music writing entered the history of chant transmission. We shall form our impressions of it according to the way we evaluate his interpretations of the evidence. The corollary hypothesis, on the other hand, is a narrative about the history of neumatic notation, about whose elements (the commission, the neumatic script they invented, the script into which this was transformed, the scientific edition, the performing edition) there is not the shadow of a hint that they existed or occurred at all. They are all creations of the author’s imagination, and there is no way at all of assessing the historical truth-value of the narrative. Its value rests, not on the quality of its ¹⁸ ‘On the Origin of Neumes’; ch. 5 of Gregorian Chant and the Carolingians.
Homer and Gre go ry
147
deductive reasoning from hard evidence, but rather on its satisfaction of the need for a coherent story with which the historian seeks to account for the condition of the sources once they appear. For this, however, there is still a missing element: the story needs to persuade its readers that in the [musically] scriptless culture in which it is placed, where music had always been passed from teacher to pupil or from one singer to another; musical notation, immediately upon its invention, would have functioned in the prescriptive mode that we take for granted today; and the authority of a written text would at once have been accepted and texts adopted as the preferred medium of transmission. It seems that these ideas have been so inculcated into the habitus of the field of musical practice as it is today that Levy did not see fit to mention them. Perhaps readers will find a negative judgement in this account, because the narrative seems clearly to violate the basic standard for historical knowledge that has guided the field of historical musicology and especially the sub-field with which we are concerned here, that it should be limited to realistic reconstructions through the logical treatment of concrete evidence. But we could instead welcome this narrative as a bellwether in a movement away from the confinement of the knowable in this sub-field to the narrow space allowed by that standard, especially considering how much of early medieval music history took place beyond its boundaries. We could embrace for this sub-field the assertion of Hayden White that ‘Historical narratives are verbal fictions, the contents of which are as much invented as found and the forms of which have more in common with their counterparts in literature than they have with those in the sciences.’¹⁹ Where better than in the historical study of early medieval music do we recognize the temptation—indeed the need—to follow the movement of postmodern historiography towards the primacy of narrative and the boundary region of history and fictions. Richard Crocker, in the assessment and prognostication cited in the introduction to Chapter 2, has acknowledged—albeit without fully embracing—this temptation: Before the ninth century . . . we cannot confidently make statements about how the music went; I will conclude that if we feel we need to make statements (and I do), it must be not with confidence but with imagination and a sense of adventure . . . (‘Gregorian Studies in the Twenty-First Century’, 34). In order to study how Gregorian chant was informed by its musical past, that past must be manufactured (p. 35). Any ideas we derive from Gregorian itself about how Western Latin chant might have been before ca. 650 is fantasy . . . I would read it as a Romantic fantasy, a fascination with remote origins. (pp. 84–5) My enthusiasm for the boldness of Levy’s move to narrative is dampened only by the fact that the tropes of his discourse are such as we would ordinarily use to characterize modern music culture. That the narrative is given in such detail is all to the good. But the details are disturbingly familiar, disturbing because, whatever our ¹⁹ ‘The Historical Text as Literary Artifact’, in Tropics of Discourse (Baltimore, 1978), 82.
148
Homer and Gre g ory
historiographical persuasions may be, we do after all expect historical constructions or reconstructions to bring out the uniqueness of the historical moment. If we are going to depend so heavily on our historical imagination it behoves us to extend its reach beyond the familiar surrounding world of our present. Otherwise we are at risk of giving our histories a circular character. Such circularity is the subject of Chapter 9. The third of the comprehensive theories of the fixity of the chant corpus as a whole prior to the initial writing down is the one most directly and exclusively reliant on the Frere principle as transposed to the music of chant, and it is the one that most clearly exposes the difficulty of imagining what, in concrete terms, can really be meant by such theories. It is the version presented by David Hughes in his essay ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’. Hughes formulated his claim in several different ways: The evidence of the manuscripts clearly shows the chant to have acquired fixed form well before the appearance of the earliest surviving notated manuscripts’ (p. 377); ‘One must conclude that by the time of the earliest surviving notated manuscripts the traditional Gregorian chant had acquired a fixed form throughout the Latin west . . .’ (p. 394); ‘The extreme stability of the basic tradition is obvious enough’ (p. 398); ‘The tradition of chant was one in which every singer attempted to reproduce exactly what he had been taught, in which accuracy was of the highest importance’ (p. 400). These claims ask for belief about three matters: that uniformity in an ancient written musical tradition is easily recognized; that the notations into which the melodies were initially transcribed from oral tradition were capable of encoding the information about which the theory makes its claims, that is, their pitch contents; and that ‘the earliest surviving notated manuscripts’ were in fact achieved through transcription from oral performance or memory. When Frere wrote of ‘the same [liturgical] formulary’ he did not implicate any problems about criteria of sameness. It is not difficult to decide whether or not two liturgical formularies are the same. But it is quite otherwise when one speaks of versions of chant being ‘nearly’ identical, and even less so when one speaks of ‘slight’ or ‘negligible’ variations that do not ‘significantly alter the identity of the melody’ (McKinnon, ‘The Emergence of Gregorian Chant’, 100). Hughes writes of ‘minor’ and ‘trivial’ variants. But we lack clearly formulated criteria for our own judgements about what makes variants ‘slight’, ‘negligible’, ‘insignificant’, ‘minor’, and ‘trivial’, and for determining the boundaries of a chant’s ‘identity’; we have no way of knowing whether any intuitive criteria that do tacitly guide our judgements about such matters would have been at all sensible to medieval singers. A further risk is that when comparisons leading to judgements about antiquity are made at the level of surface detail in the notes, as they are in Hughes’s study, rather than of structural aspects that relate types to one another, they may point more to recent contacts than to deep common roots. An example of this is the comparison in Chapter 1 of the Frankish and Old Roman transmissions of the gradual Sciant gentes from the
149
Homer and Gre go ry
point of view of the declamatory pace of the words—where it is syllabic, where neumatic, where melismas fall. This represents, as Helmut Hucke has observed, the most fundamental, and probably the first, decision made by the originator of the chant, and it carried through the transmission to the written descendants.²⁰ Consistency at that level gives greater confidence that we see something of the original, regardless of the degree of similarity or dissimilarity in the note patterns at the surface. In his 1988 essay ‘Gregorianische Fragen’ Hucke wrote: A written transmission expresses nothing more than the information that it contains . . . Staffless neumes manifestly contain much more information than one might think at first. But about the ‘pitch content’ that Hughes considers to be assured, they inform us badly. And we have no assurance that the transmission had ever been uniform with respect to pitch content . . . A cantor’s handbook is witness only to what the cantor could look up in it. How he transformed such information into practice is something about which the degree of correspondence among all cantor’s handbooks can say very little. (p. 326) For a ready approximation of the degree of such correspondence Hucke refers us to the neumations copied into the Graduale triplex from the manuscripts Laon 239, St. Gallen 359, and Einsiedeln 121. Comparing neumations of corresponding syllables in the first with either the second or the third, he reports deviations of some 20–40 per cent.²¹ So much regarding the first and second beliefs required by the ‘stability’ theory. As for the third, what we have in view is the representation of a singing tradition by means of a text, not a transformation or conversion from one to the other. Can we actually imagine the achievement of this through countless independent acts of dictation and transcription throughout Europe into staffless, probably adiastematic neumes, acts that resulted directly in a uniform written tradition that should be read as reflecting the uniformity of those dispersed moments of dictation? Or does the uniformity of the written tradition result from the circulation of approved texts as models for copying, in a cultural and political atmosphere in which the highest value is placed on uniform texts? Signs of a condition of and belief in fixity in aspects of the oral tradition do indeed greet us in various kinds of testimony and in surviving musical texts themselves, as we have seen. The efforts to impose this theme uniformly across our historical view from above, however, as a sort of article of faith, seem rather to overreach. Our unwillingness to accept Strunk’s pessimistic outlook has not resulted from the discovery or invention of new methods undreamt of by him, but from our need to fill in that long lacuna in our knowledge. But that has meant a greater willingness to entertain fictive accounts than was displayed by that master of research who, intuitive though he was, was primarily moved by the goal of ‘controlling the ²⁰ ‘Toward a New Historical View of Gregorian Chant’, 452.
²¹ See Ch. 1 n. 46.
150
Homer and Gre g ory
evidence’. But then we should not delude ourselves; our accounts these days are forged more under the influence of our general outlooks and of our ideas about what is a good story than by positivistic criteria that guide us in judging what is confirmable and disconfirmable. Levy’s is not the only narrative of the early history of chant that shows a tendency to approach the boundary between history and fiction. We are like today’s performers of medieval music, trying to be as faithful as they can be to the sources, but in the end moulding that evidence according to their ideas of beauty. On the whole the battles over ‘authenticity’ are over. Nothing so much crystallized my intuitive sense that the nature of the chant as it has come down to us in written form since the tenth century must reflect the ways in which it was made and transmitted in the absence of writing, as my reading of Albert Lord’s book The Singer of Tales (1960). That book, and the research programme that it represents, served me as both example and model: example in showing that it was possible to develop on the basis of a living practice an account of the technique through which epic bards could carry out the prodigious feats of remembering and composing entailed in their performances that lasted for hours; model in combining analytical rigour with critical appreciation and imagination. There is no intention of identifying twentieth-century Balkan epic bards with medieval ecclesiastical chanters, or of identifying epic poems with Gregorian chants. The similarity that is illuminating is formal: in the terms and methods of the analysis of their productions and of their ways of producing them. As Haas observes, the citing of the context of epic poetry is more a metaphor than an analogy, following I. A. Richards’s characterization of metaphor as ‘a reciprocal borrowing and trafficking between thoughts, a transaction between different contexts’.²² Witness to the heuristic value of the formal paradigm developed by Parry and Lord is the way that some scholars have turned it, by way of my work on chant, to the study of jazz improvisation—specifically the improvisations of Lester Young and Bill Evans.²³ Of course there is no suggestion of a similarity between Gregorian chant and the jazz performances or compositions of Lester Young or Bill Evans. What we learn through the fit of the same paradigm to such widely divergent traditions is that oral traditions have things in common in the way they work. One respect in which this book is fundamentally at odds with The Singer of Tales is Albert Lord’s dogmatic insistence on a defining opposition between oral composition and performance from memory. Lord thought it was essential for him to maintain that when one epic bard was able after one hearing to repeat the lengthy performance of another virtually verbatim, it was not a superhuman feat of instant ²² Haas, Mündliche Überlieferung und altrömischer Choral, 32. Richards, The Philosophy of Rhetoric (Oxford, 1936), 93. ²³ Lawrence Gushee, ‘Lester Young’s “Shoeshine Boy”’, in IMS Berkeley, 151–69; Gregory Smith, ‘Homer and Gregory and Bill Evans? The Theory of Formulaic Composition in the Context of Jazz Piano Improvisation’ (Ph.D. diss., Harvard University, 1984); id., ‘In Quest of a New Perspective on Improvised Jazz’.
Homer and Gre go ry
151
memorization but the normal performance of a trained singer who knew the rules and constraints, and had only to hear the subject and the story outline to produce what seemed like a replica of what he had heard. His denial that this was memory working is only a consequence of his tacit assumption of the literal-recording-andreproduction concept of memory. It would hardly have been possible for him to accept an objection that he was providing an instance of a virtuoso exercise of memory at work, from a different point of view. But he somewhat moderated his view about this towards the end of his life, while maintaining that he had not done so, in his essay ‘Memory, Fixity, and Genre in Oral Traditional Poetries’.²⁴ There he leaves us with a useful thought: ‘fixity’ cannot really be a property of an oral poem. It is fixed by being written down. Until then it may be sung repeatedly the same way, through whatever mnemotechnical devices of which the singer disposes. But what is called ‘fixity’ is a property of the singer’s performance. A comment is required regarding the adaptation to chant composition of Lord’s concept of ‘thrift’ that is developed in section VII of this chapter. I have thought to recognize such a phenomenon in the tendency to sing again what one has just sung in the same chant, in the same office or mass, in antiphons for the same day, in the same cycle of feasts, and I have interpreted it in the light of the exigencies of oral performance.²⁵ In his book The Advent Project James McKinnon has interpreted such clustering of melody-types within the same feast cycle as the musical counterpart, in a large-scale compositional project, of the narrative continuity—and therefore unity—of the liturgical texts. The effect of the two interpretations is perhaps not so different, since the intention of such an overall design, if that is what it was, would in any case make both text and music memorable. But whether we would interpret instances of melodic clustering in shorter time spans—the offices of the same day, chants of the same office or mass—as products of an intention to unify is not so clear. I am left to wonder whether McKinnon’s musical interpretations have not been made from a modern perspective that might not have been matched by a medieval one (as in his identification of a ‘Master of the Re-Fa Lyrics’, discussed in Chapter 2). There are really two kinds of stories that are stirred up by the question posed at the beginning of this chapter: the kind that gives a direct answer to the question—any account of the methods by which ecclesiastical singers invented and transmitted plainchants from the beginning of the tradition and onwards—and the kind that follows the history of such accounts: the historiography of the question. The chapter begins with a story of the second kind, or at least the beginning of such a story (it is continued in Ch. 7). This is motivated, first, by the fact that for the centuries prior to the first production of notated chant books the first kind of story can be told only as one or another set of suppositions, whereas the beginning of the second kind of ²⁴ In John Miles Foley (ed.), Oral Traditional Literature: A Festschrift for Albert Bates Lord (Columbus, Ohio, 1981), 451–61. ²⁵ See Nowacki’s discussion of the concept in ‘The Gregorian Office Antiphons’, 243.
152
Homer and Gre g ory
story—the invention of the genre of ‘Gregorian chant’ through the creation of the Gregorian legend as a literary and iconographic tradition—is visible to us with a clarity that is rarely matched in historiography. The second motivation is my conviction that historical understanding is compounded of insights into ‘what happened’ or ‘how it was’, and insights into the history of the way those things have been understood. Wrapped up in the story of how ‘Gregorian chant’ got its name are the main questions that would usually be regarded as more specifically historical and scientific: the origins of the tradition, the ontology of the chant, the manner of its production and transmission, the nature of its forms. The tradition of beliefs about the nature and origin and authority of Gregorian chant that had its beginning in the ninth century remains a leading component in the field of concepts and beliefs that confront modern scholars trying to make sense of the practice that tradition has encapsulated in its interpretations, and in order to evaluate it we must be able to understand the interests and needs that it has served and continues to serve. What we can understand about the past resides in the confluence of what tradition and research tell us.
} There is a difficult question about the history of plainchant that is often given simple answers: before the age of music writing, how was it possible to invent plainchants, to preserve them, to learn and perform them? In brief, what was the manner of plainchant transmission prior to the use of musical notation? A generation after Gregory the Great, Isidore of Seville wrote ‘The sound [of music], since it is a sensible thing, flows by into the past and is impressed upon the memory . . . Unless sounds are retained in memory by man they are lost, because they cannot be written.’²⁶ The idea that melodies were retained through memory has come down to the present day as the principal answer to the question about oral transmission. It has lately been joined by the suggestion that plainchant performance in those days had something of improvising in it. But what have we really understood about the mechanisms of oral transmission, once we have said ‘memorized’ and ‘improvised’? What follows here is an essay into the possibility of developing an approach to the problem that will meet two conditions. First, it must recognize that the way the music has turned out is a consequence of the constraints of an oral tradition, ²⁶ Translation by James McKinnon in Strunk’s Source Readings, 149. McKinnon notes that ‘the last sentence of this passage has been taken by many as evidence that musical notation was not known in Isidore’s time’. There is, indeed, no reason to believe that musical notation was known then, but Isidore’s remark really does not yield evidence one way or the other about the question. Sounds cannot, of course, be written, ever. It is signs for attributes of sounds that are written as musical notation. This is not a verbal quibble, and it is worth mentioning, for it exemplifies the risk of adapting our reading of sources of another time to our prior beliefs about their contents and then using them to confirm those beliefs.
Homer and Gre go ry
153
hence that learning to understand the music and learning to understand how it was transmitted is a single task. Second, it must be realistic from the point of view of the human cognitive processes that are invoked. The main evidence is the written transmission. The objective is to learn to frame questions that must be put to the written transmission in order that it may give testimony about the oral transmission that lies behind it. In the few specimen analyses that are offered the focus will be on the Frankish Gregorian transmission. The citation of parallel Old Roman transmissions is intended mainly to exploit the rich opportunity for learning about the nature of oral transmission from the comparison of two traditions that have branched from one sometime before their commitment to writing. It is meant only incidentally to contribute to the discussion about the historical relations between the two. The theory of oral composition that has emerged from the study of epic poetry is taken here as a paradigm—a concrete research model that specifies the problems to be investigated, identifies categories of evidence and methods of investigation, and establishes criteria for satisfactory answers. There is no suggestion of similarity between the two practices and their products other than in the heuristic value of the method of enquiry that follows that paradigm.
I The problem will be entered near the end, in the Carolingian era. That was a time when many things happened in liturgical music that have reverberated ever since, and well beyond the domain of liturgical music: the invention of notation and the writing down of the repertory, the ordering and editing of the repertory in the effort to normalize and facilitate the practice, to that end the installation of the eight-mode system, the identification of an author, and more. I begin with reflections on the last, a phenomenon that is more symptomatic than instrumental: the identification of Gregory the Great as the author of the plainchant. There are really two questions about Gregory’s place in music history: first, the question of the interpretation that we can put upon Gregory’s activities during the fourteen years of his papacy vis-à-vis the singing of the Divine Office; second, the question of the way we can understand the Carolingian view of his activities. The evidence with respect to the first question is not adequate to form a clear impression. But there is material with respect to the second question that can significantly illuminate the musical circumstances in the ninth century. There is a family of medieval illustrations in which Gregory is shown seated, with a book that he either holds in his hands or lets rest on a lectern (Plate VII). A dove, whose beak is pointed towards Gregory’s ear, is perched on his shoulder or hovers by his head. Lower down and off to the sides are two monks, separated
154
Homer and Gre g ory
from Gregory by a cloth screen. One is inscribing something in a book, the other is peeking at Gregory and the dove from the other side of the screen, in an attitude of astonishment. Sometimes just one monk is shown playing both roles: peeking, but holding the stylus that shows he has interrupted his writing (Plate VIII). The tradition of these illustrations can be traced back to the third quarter of the ninth century.²⁷ But there is a description of exactly this scene in a biography of Gregory written in the first half of the century by Paul Warnefried, a monk of Monte Cassino and deacon of the Church in Rome.²⁸ It is explained that the monk has become curious about the pauses in the dictation. When he peeks through the screen he sees that the dove’s beak is at the Pope’s ear during the silent moments and turned away when Gregory dictates. The gesture of astonishment is his reaction to the sight of this miracle. Sometime between the writing of this biography and the compilation of the tenth-century Hartker Codex the scene took on musical content; Paul has it that Gregory is dictating his commentary on Ezekiel, under the inspiration of the dove as embodiment of the Holy Spirit. An illustration in the Hartker Codex shows the scribe’s writing desk turned sufficiently to reveal the St. Gallen neumes in which he records the melodies that Gregory dictates (Plate IX). About 875 John Hymmonides, also a Roman deacon and Gregory biographer, credited the pope as follows: The exceedingly diligent Gregory, motivated by the compunction of musical sweetness, compiled a centonate antiphoner of chants, a task of great usefulness. He also founded the schola cantorum, which still sings in the holy church of Rome according to its original instructions. And he built two dwellings for the schola, with the proceeds from some plots of land: one near the steps of the basilica of St. Peter the Apostle, and another near the lodgings of the Lateran Palace, where even today are preserved with fitting reverence the bed on which Gregory lay while singing, the switch with which he threatened the boys, as well as the authentic antiphoner.²⁹
It appears from these documents that the legend of Gregory the musician had crystallized in the late ninth century, but not before.³⁰ Behind the legend is a piece ²⁷ See J. Croquison, ‘Les Origines de l’iconographie grégorienne’, Cahiers archéologiques, 12 (1962), 249–62. Concerning the illustration in Paris lat. 1141, see Florentin Mütherich’s introduction to the facs. edn.: Sakramentar von Metz, Fragment (Graz, 1972). Iconographic studies of these illustrations have heretofore taken no account of their relation to the history of Gregorian chant, although that relationship has much to do with their changing form. It will be an incidental contribution of this essay to bring two subjects together. ²⁸ ‘S. Gregorii magni vita’, Patrologia latina, ed. J. P. Migne, lxxv (Paris, 1849), cols. 57–8. ²⁹ Translation by James McKinnon in Strunk’s Source Readings, 179. ³⁰ This should not be said without mentioning certain documents from which an origin in the 8th c. might be inferred. Around the year 750 Bishop Egbert of York referred to Gregory as the author of an antiphoner. The gradual of Mont Blandin (late 8th or early 9th c.) is inscribed as the work of Gregory. In a prologue to the gradual of Monza it is claimed that ‘this book of musical art of the Schola cantorum’ was assembled by Gregory (the book is without musical notation). The evidence bearing on this question has been reviewed by Helmut Hucke, whose conclusion it is that nothing more is suggested than a belief in
Homer and Gre go ry
155
of history that is at once political, ecclesiastical, and musical. The essential fact is Charlemagne’s ambition for the reform and unification of the Frankish church according to the practice in Rome. This served the end of political unification, but it was also motivated by a deep religiosity and a genuine feeling of reverence for Rome. The project quite explicitly included an attempt to impose everywhere a uniform tradition of plainchant. Contemporary reports about those attempts will remind us of the legend of Rashomon. John the Deacon wrote thus: Of the various European peoples it was the Germans and the Gauls who were especially able to learn and repeatedly to relearn the suavity of the schola’s song, but they were by no means able to maintain it without distortion, as much because of their carelessness . . . as because of their native brutishness. For Alpine bodies, which make an incredible din with the thundering of their voices, do not properly echo the elegance of the received melody, because the barbaric savagery of a drunken gullet, when it attempts to sing the gentle cantilena with its inflections and repercussions, emits, by a kind of innate cracking, rough tones with a confused sound like a cart upon steps. And so it disquiets the spirits of those listeners that it should have mollified, irritating and disturbing them instead. Hence it is that in the time of this Gregory, when Augustine went to Britain, cantors of the Roman school were dispersed throughout the West and instructed the barbarians with distinction. After they died the Western churches so corrupted the received body of chant that a certain John, a Roman cantor (together with Theodore, a Roman citizen yet also archbishop of York), was sent by bishop Vitalian to Britain by way of Gaul; and John recalled the children of the churches in every place to the pristine sweetness of the chant, and preserved for many years, as much by himself as through his disciples, the rule of Roman doctrine. But our patrician Charles, the king of the Franks, disturbed when at Rome by the discrepancy between the Roman and the Gallican chant, is said to have asked—when the impudence of the Gauls argued that the chant was corrupted by certain tunes of ours, while on the contrary our melodies demonstrably represented the authentic antiphoner—whether the stream or the fountain is liable to preserve the clearer water. When they replied that it was the fountain, he wisely added: ‘therefore it is necessary that we, who have up to now drunk the tainted water of the stream, return to the flowing source of the perennial fountain’. Shortly afterward, then, he left two of his diligent clergymen with Hadrian, a bishop at the time, and, after they Gregory’s work in the liturgical arena. See his ‘Die Entstehung der Überlieferung von einer musikalischen Tätigkeit Gregors des Großen’, Mf 8 (1955), 259–64. The prologue to the gradual of Monza belongs to a tradition about which Bruno Stäblein has written (see ‘Gregorius Praesul, der Prolog zum römischen Antiphonale’, in Musik und Verlag: Karl Vötterle zum 65. Geburtstag (Kassel, 1968), 537–61). He argues that the prologue originally referred to Gregory II (715–31) and that the attribution changed meaning in the late 9th c. That is parallel to the suggestion I am making about the imposition of new meaning on the Gregory illustrations about the same time. Van Dijk has argued that Bishop Egbert, too, referred to Gregory II, who propagated a sacramentary and gradual through the empire, which were adopted at the Council of Cloveshoe in 747 (see his ‘Urban and Papal Rites in 7th- and 8th-Century Rome’, Sacris erudiri, 12 (1961), 411–87). Solange Corbin, L’Église à la conquète de sa musique (Paris, 1960), cites still other documents of the late 8th c. which, by their silence on the matter, support the inference that the Gregory legend had not yet been created.
156
Homer and Gre g ory
had been schooled with the necessary refinement, he employed them to recall the province of Metz to the sweetness of the original chant, and through her, to correct his entire region of Gaul. But when after a considerable time, with those who had been educated at Rome now dead, that most sage of kings had observed that the chant of the other Gallican churches differed from that of Metz, and had heard someone boasting that one chant had been corrupted by the other; ‘Again, he said, let us return to the source. Then Pope Hadrian [I], moved by the pleas of the king . . . sent two cantors to Gaul, by whose counsel the king recognized that all indeed had corrupted the suavity of the Roman chant by a sort of carelessness, and saw that Metz, in fact, differed by just a little, and only because of native savagery.³¹
Now here is the other side of the story, from a biography of Charlemagne originating about the same time at St. Gallen and sometimes attributed to Notker: Charles, . . . sad that all the provinces, regions, and cities differed from one another in the divine praises, that is, in the melodies of the chant, took care to request from Stephen [II], pope of blessed memory [Stephen had died eleven years before Charlemagne took the throne] that he send additional clerics who were greatly skilled in the divine chant. Stephen . . . gave assent and dispatched from the Apostolic See to Charles in Francia twelve clerics who were greatly learned in the chant, according to the number of the twelve apostles. By Francia, incidentally, which I have just mentioned, I mean all the provinces beyond the Alps. When the above-mentioned clerics departed from Rome, they plotted among themselves (since all Greeks and Romans are ever consumed with envy of Frankish glory) how they could so alter the chant that its unity and harmony might never be enjoyed in a realm and province other than their own. So they came to Charles and were received with honor and dispersed to the most prestigious locations. And, in these various localities, everyone of them strove to sing, and to teach others to sing, as differently and corruptly as they could possibly contrive. But the exceedingly clever Charles celebrated the feasts of Christmas and the Epiphany one year at Trier and Metz and very alertly and sharply comprehended the quality of the chants, indeed penetrated to their very essence, and then in the next year he followed the same festivals at Paris and Tours and heard nothing of that sound which he had experienced the year before in the above-mentioned places. Thus he discovered in the course of time how those he had sent to different places had come to differ from one another, and he conveyed the matter to Pope Leo [III] of blessed memory, the successor to Stephen [there were four popes and four decades between Stephen and Leo]. Leo, after recalling the cantors to Rome and condemning them to exile or to lifelong confinement, said to the illustrious Charles: ‘If I send others to you, they, blinded by envy like those before them, will not neglect to deceive you. Rather I will attempt to satisfy your wishes in this manner: give me two very intelligent clerics of your own, in such a way as not to alert my clergy that they belong to you, and they shall acquire, God willing, the total proficiency in this skill that you seek.’ It was done in this way, and after a reasonable length of time Leo returned the clerics to Charles perfectly instructed.³² ³¹ The translation is by McKinnon, in Strunk’s Source Readings, 179–80. ³² The translation is by McKinnon, ibid. 181–2.
Homer and Gre go ry
157
We are not likely to learn who was more nearly telling the truth about the points of difference that divide these narratives. But we can learn quite a lot about the events of a generation earlier from those things about which they are in agreement in their explanations for the outcome of those events: that there was a deliberate effort to impose on the Franks and the Gauls a uniform way of singing based on the different tradition of Rome, that this was Charlemagne’s project, not the papacy’s, that it encountered substantial obstacles, that the medium of transmission was human singers, not books, and that whatever changes occurred in the manner of singing were understood to have human causes—the competencies (or lack of such) or deliberate actions of the singers; they were not natural processes such as ‘decay’ or ‘evolution’. There is another report about Charlemagne’s activities that gives substance to the metaphor about the source and the stream. He commissioned Paul the Deacon to obtain from Hadrian a ‘pure Gregorian’ sacramentary. After some two years Hadrian sent up a sacramentary, along with a letter: ‘Regarding the sacramentary of our holy predecessor, Pope Gregory, thus: since the grammarian Paul has asked it of us for you for a long time, and especially one that is unmixed and handed down by the holy Church, we have sent one to your Royal Majesty with John, monk and abbot of the city of Ravenna.’³³ Why the delay? The quality of the ‘Gregorian’ sacramentary that was eventually submitted suggests a kind of answer. It was so deficient in its provisions as to be unusable by itself, and although it was adopted as the official text, it was regularly supplemented with the Frankish Gelasian sacramentary. What is more, it failed to provide formularies for some feasts that were celebrated in Gregory’s time while providing for others that are of a later origin.³⁴ The papal court, in short, seems not to have been prepared in either the conceptual or the practical sense for Charlemagne’s request. The identification of the ‘Hadrianum’—the resultant sacramentary that combined Roman and Frankish elements—as ‘Gregorian’ was, like the Gregorian legend itself, a product of the Carolingian urge to see the creation of a uniform and Holy Roman Empire as work with divine ordination. On the musical side this movement reflects a new way of thinking. One had to conceive of melodies as being potentially stable from one performance to another, as being fixed from the moment of their invention. One could be intolerant of the variations of local performance practice, and one could interpret them as corruptions of an original. Putting together all of these reports, we see a Carolingian view of things emerging that will look quite familiar to us. A corpus of music has originated in a certain time and place and through the agency of a particular person. At the moment of ³³ See Theodor Klauser, ‘Die liturgischen Austauschbeziehungen zwischen der römischen und der fränkisch-deutschen Kirche vom achten bis zum elften Jahrhundert’, Historisches Jahrbuch, 53 (1933), 169–89. ³⁴ Jean Deshusses, ‘Le “Supplément” au sacramentaire grégorien: Alcuin ou saint Benoît d’Aniane?’, Archiv für Liturgiewissenschaft, 9/1 (1965), 48–71.
158
Homer and Gre g ory
its origin it is written down. The repertory has spread through a transmission that has left a multiplicity of versions. After some time it is observed that the versions do not agree with one another. This is interpreted as the consequence of a process of corruption, and an editorial enterprise is undertaken to restore and re-establish the original. As standard for the enterprise a source closest to the point of origin, an original version, is sought. The restoration called for in the story was the first such project in the history of European musicology. For the most part its objectives and premisses are the same as those for the Gregorian chant restoration that has been under way since the turn of the twentieth century. The difference is one of method—the agreement of many sources rather than the authority of a single source. It is interesting to see that our view of musical transmission goes back so far. I believe there are no earlier signs of it, that it was in fact a new way of thinking, not only about transmission but about music in general. That impression is reinforced by the fact that it is just at this time that the project of writing down the repertory was launched. If we can associate the invention of the Gregory legend with the facts of political-ecclesiastical history, can we also associate the writing down with those circumstances? It is tempting to see noted service books as instruments of the campaign to bring about uniformity. But the matter must surely be more complicated than that. More than a century after the earliest writing down Odo³⁵ and Guido³⁶ spoke of a notation that referred directly to a sound-space articulated as discrete steps. That meant melodies could be represented in terms of scale degrees that were precisely measured with a monochord, and they could be read off from the page though they might be totally unfamiliar to the singer. Both men boasted of the novelty and the advantages of that. (The pictorial tradition of the Gregory legend followed this conceptual and practical innovation. See Pl. X.) Had there been such a notation with the background of such a pitch system, it would certainly have been a factor in the musical disputations of Carolingian times (indeed it was still in the service of the elusive Gregorian reform that Guido saw the principal utility of his system). But the oldest notations, referring as they do to familiar melodic patterns rather than to pitches, were not well suited to be didactic instruments. It is easier to believe that they supported the performance of what was already known and accepted.³⁷ That interpretation is consonant with the implications of the invention of the Gregory legend for the degree of fixity of the repertory. But it raises the question why the performance of the melodies would have required the sort of support that the writing down afforded. Can we be ³⁵ Pseudo-Odo of Cluny, ‘Dialogue on Music’, Strunk’s Source Readings, 198–210. ³⁶ Guido of Arezzo, ‘Prologue to his Antiphoner’, ibid. 211–13. ³⁷ The transcription of the oral tradition would be the role assigned by Kenneth Levy to the Palaeofrankish neumes in the theory discussed in the introduction to this essay.
Homer and Gre go ry
159
concrete about what sorts of changes in the way one knew melodies would have prepared them for writing down, and would have made singers dependent on notation? Is there evidence of such a change? These questions take us back to the need to achieve some understanding of the nature of oral transmission.
II In the absence of scores the medium of transmission was performance. That is not something we need to demonstrate; but we do need to think about it. Is performance without scores tantamount to performance from memory? As our scholarly habits have been conditioned by the study of texts, our recourse in their absence has been to the concept of memory as a medium of storage comparable to a score: things are committed to memory whole, and there they lie fixed and lifeless until they are retrieved whole. We say that the singer has memorized a melody as though we might be saying that he had swallowed a score. Modern psychology tells us that this is an unrealistic view of the process of remembering. In a classic of the field Frederic C. Bartlett has written that remembering is a process not of reproduction but of reconstruction.³⁸ I shall try to describe the important points of Bartlett’s theory, especially as they might illuminate the problem about oral transmission. 1. The theory of remembering depends on the theory of perception, for the way we recall experiences depends on how we grasp them in the first place. 2. Perceiving is not passive reception, it is active organizing. We strive to assimilate newly presented material into the setting of patterns and schemata left from the encounter with past experience. But that always results in the reorganization of those patterns. Thus perceiving, and indeed conscious life in general, is a continual process of adjusting our own records of our past. 3. In perceiving we draw out certain salient features of the matter presented that are for us especially prominent. These serve as signposts for the process of assimilating and reorganizing. 4. Those signposts play a central role in remembering. As perceiving is not simply a matter of the reception of stimuli, so remembering is not simply the storage of stimuli strung together and their later reproduction. Rather it is an active ³⁸ Remembering: A Study in Experimental and Social Psychology (Cambridge, 1932; repr. 1972). The interested reader is referred to the stunning confirmations and articulations of Bartlett’s fundamental insights that have been made by Jean Piaget and Barbel Inhelder. See Inhelder, ‘Memory and Intelligence in the Child’, in D. Elkind and J. H. Flavel (eds.), Studies in Cognitive Development: Essays in Honor of Jean Piaget (Oxford, 1969); and Piaget and Inhelder, Memory and Intelligence (trans. from the French Mémoire et Intelligence, Paris, 1968; English version, London, 1973). The central thesis there is that memory is a reconstructive faculty and is consequently subject to improvement due to progress in intelligence.
160
Homer and Gre g ory
process of grouping appropriate details about such salient features. It is a process of construction, not reproduction. 5. In remembering, therefore, we activate and reorganize the patterns of past experience. This has two important corollaries: (a) each recall is based not on some fixed model outside ourselves, but on our own assimilated version of the matter recalled—not on the ‘original’ but on our most recent rendering; and (b) recall must be in conformity with the existing schemata in which our mind is organized. What does not conform will tend to be corrected or eliminated. 6. The latter tendency, together with the role of persistent detail in remembering, leads to stereotyped forms. This is especially important in the conventionalization of the forms of cultural expression. 7. In the recall of narratives, beginnings and ends especially provide those stand-out, persistent features that serve as the focal points of the reconstruction. Consequently it is beginnings and ends that tend to become most stereotyped in repeated recall. 8. Form, as well as salient detail, is persistent and is therefore an important factor in what makes remembering possible. 9. A salient detail may be common to two or more themes or streams of interest, and it may serve as a crossing-point between them. In that way the theme originally presented may be left and another entered. 10. Remembering and imaginative construction are on a single continuum. They differ from one another in degree, but not in kind. When we have understood ‘memory’ in the sense of these ideas, the notion of transmission through memory becomes quite plausible. But now we must approach the question from the other direction. In the absence of scores, the medium of composition is performance. Again, that does not require demonstration; still it deserves some reflection. Two sorts of ideas have been advanced about composition in performance, both ideas of improvisation in plainchant performance. The first is a rather uncritical notion about improvisation as an unplanned and unregulated performance, smacking somehow of Oriental ornateness and South European licentiousness, a stereotyped idea that has been invoked in an effort to account for a certain prolixity and surface variability in the transmission of the Old Roman chant. Passing over the questionable nature of this conception in all respects, what is really most misleading about it is its treatment of improvisation as a special practice, invoked to account for the special characteristics of a special transmission; whereas we require an understanding of oral transmission as a normal practice whose object and effect is to preserve traditions, not play loose with them. That need has been more nearly met by the conception of an improvisational practice that carries the sense of elaboration and variation upon a Grundgestalt (Peter Wagner did not speak of improvisation; nevertheless one thinks of his char-
Homer and Gre go ry
161
acterization of the tracts as ‘psalmody-like variations’). The plainchant transmission does suggest that sort of thing in a general way, but the idea can be misleading in two particulars if it is narrowly applied: if we hope to identify a fixed Grundgestalt for each melodic group, it will always elude us; and if we carry out the idea to its implications for the techniques of elaboration, we shall run aground with the concept of diminution. Any account of the oral invention of plainchant, to be realistic, must look to the practical, recognizing that in composition through performance the primary, pervasive, and controlling condition is the continuity of the performance. The singer does not make sketches, does not consult a catalogue of formulae and deliberate about which ones he will string together, does not have before him a skeleton outline of the melody that is to be elaborated, and does not go back and make revisions. She or he will have planned before beginning, and will have paused at moments of articulation, thinking what should come next while scanning the next phrase of the text. A realistic account of plainchant composition must consider that any of the general features of the melodies may represent accommodations to this condition.
III I focus now on a single melodic group, considered from the viewpoint of oral transmission. The classification of the plainchant repertory is based upon the intersection of two factors: the melody’s place in the order of service and its modal group. (The assignment within the liturgical calendar does not, in the main, have musical import.) Taking these factors as coordinates, we can construct a table in which we can locate any individual chant (see Table 6.1.). In general the positions in this table represent the distinct musical types into which the repertory falls. But as the chart stands it is misleadingly simple and systematic. Not all positions are occupied (e.g., tracts fall only in modes 2 and 8); on the other hand, there can be more than one melodic type with the same modal-liturgical designation. This simply reflects the fact that the melodic families of the repertory had formed before the eight-mode system was imposed on them. What is meant by ‘melodic type’ will become apparent from the study of such a family. It will be convenient to choose the verses of a single tract, a collection of discrete melodies of whose uniform history we may be confident (Ex. 6.3, CD track 9).³⁹ The tracts were ways of singing extended series of psalm verses, by the soloist and without any refrains. We can understand much about the course run by ³⁹ Gregorian chant transmissions in this and the following examples are taken from the Graduale Romanum except as noted. Old Roman transmissions are transcribed from Vatican Vat. lat. 5319.
162
Homer and Gre g ory
Table 6.1. The two categories of constraints determining chant types Modes:
1
2
3
4
5
6
7
8
office antiphons office responds introits graduals alleluias tracts offertories communions
the melody for each verse in the light of its role as carrier for those texts. As the majority of psalm verses comprise paired lines, the main articulation in the melody comes between those lines. Normally the melody comes to rest once within each half-verse as well, so that most verses are declaimed as four phrases (marked a, b, c, d in Ex. 6.3). In the tracts the setting of the individual verse is the melodic whole. Its constituent units are the phrases (a–d), and in certain places their initial and cadential formulae (I use the expression ‘formula’ in a precise sense that will shortly become apparent). The tract as a whole is marked off as a musically closed entity only by the special music for the opening of the first verse and the close of the last. (The firstverse a phrases of six Gregorian tracts of mode 2 can be seen in Ex. 6.4 below. The last-verse d phrases of three of the same tracts are shown beneath that of Deus, Deus meus in Ex. 6.3.) The aim of the following description is to develop an account of the family of verse melodies with respect to their common characteristics and with respect to the musical (including verbal-musical) factors that may be responsible for both variance and invariance from one to another. The centre of a musical declamation of a psalm verse is a recitation on a single note, the tenor, prepared by an intonation (a tuning up) and discharged through a fall to the final. There is a dynamic tension between the tenor and the final, the tenor being the more highly charged and the final the more relaxed and stable tone. In such elaborate forms of psalmody as the tracts and graduals this relationship plays a great role in the formation of melodies. The a phrase of verse 10 will demonstrate the point most directly. It is without intonation (there is no need for the singer to tune up to the recitation note in any internal verse, for he has already
163
Homer and Gre go ry
Ex. 6.3. Mode 2 Gregorian tract Deus, Deus meus, comparing fourteen verse melodies and the final phrases of mode 2 tracts De necessitatibus, Qui habitat, and Confitemini a V 1 De us, a (continuation)
a 2 Lon
ge
a 3 De
us
me
us
cla ma
bo
per di
em
a 4 Tu
au
tem
a 5 In
te
spe
ra
ve
runt pa tres
a 6 Ad
te
cla ma ve
runt
a 7 E go au
tem
a 8 Om
nes
qui
vi
de
bant
me
a 9 Spe
ra
vit
in
Do
mi
no
si de ra
ve
tis
Do
mi
num
tur
Do
mi
no
na
ce
a 10 Ip si ve
ro
con
runt
et
a 11 Li
be
ra
me
a 12 Qui ti me a 13 An nun ti
a
bi
a 14 Po pu
lo
qui
tur
164
Homer and Gre g ory
Ex. 6.3. cont. b 1 De us
me us,
re
spi
ce
in
me:
b 2 a
sa
lu
ta
me
a
b 3 nec ex au
di
es:
bi
tas,
b 4 in san
cto
ha
b 5 no
stri:
b 6 et sal
vi
fa
cti
sunt:
b 7 sum
ver
mis,
a
sper
na
ri
pi
et non
ho
mo:
b 8 ban
tur me:
b 9 e
at
e
i:
b 10 con spe
xe
runt me:
b 11 de
o
re
le
o
nis:
te
e
um:
b 12 lau
da
b 13 ge 14
ne ra ti
o
ven tu
ra:
165
Homer and Gre go ry Ex. 6.3. cont. c 1 qua re
me
c 2 c
ver
ba
6
5
4
3
2
in
no
cte
et
non
3 c 4 laus
Is
ra
c 5 spe
ra
ve
runt,
ra
ve
runt,
c 6 in
te
spe
op
pro
bri
um
ho
mi num,
lo
cu
ti
sunt
la
bi
salvum fa
ci
at
e
um,
di vi se
runt
si
bi
o
rum
c 7 c 8 is
c 9 c 10 c 11 et a cor
ni
bus
u
ni cor
nu
c 12 u ni ver
sum se
men
et an nun
ti a
bunt
Ja
cob,
c 13
14
cae
li
1
166
Homer and Gre g ory
Ex. 6.3. cont. d 1 de
d
ri
li
qui
sti?
2 d
de li
cto
rum
me
o
rum.
3 ad in
d
si
pi
en
ti am
mi
hi.
4 el.
d 5 et
d
li
be
ra
sti
e
os.
6 et
d
non
sunt
con
fu
si.
7 d
et
ab jec ti
o
ple
bis.
et
mo ve
runt
ca
put.
am
vult
e
um.
ta
me
8 d 9 quo ni
d1 10
ve
d2
sti men
et
d
su per
ve stem me
am mi
se
a
runt
sor
tem.
11 d
hu
mi
li
ta
tem
me
am.
ca
te
e
um.
12 d
ma
gni
ju
sti
fi
13 d
ti
am
e
jus.
14 quem fa cit De necessitatibus, final phrase d
Do
mi
nus.
fa ci Qui habitat, final phrase d
en
tes va
na.
sa lu Confitemini, final phrase d
ta
re me
um.
ta
ri tu
o.
in sa
lu
Homer and Gre go ry
167
been there), hence the phrase is reduced to a tilting back and forth between tenor and final. The relation between those two notes is brought home by the stress situation of the words. The strongly stressed syllables 1, 3, and 9 are up on the tenor, whereas the final provides a point of relaxation for the second syllable and especially for the anacrustic syllables 5–8. The a phrase in verse 5 is made much like that in verse 10, but it is given an intonation that allows a longer anacrusis to the first main stress, which falls only on the fifth syllable. The a phrases of verses 12 and 13 are again sung about the f–d core, with intonation and extended anacrusis to the agogic accent on (Do) mi (no). No doubt the beginning on e in the thirteenth verse has to do with the larger number of syllables. The reader is urged to study other a phrases while keeping the same considerations in mind. Attention is directed in particular to verses 8, 4, 7, and 2, in that order. The verbal text was, of course, one of the données with which the singer had to work, and his task was to exploit the resources of melody in rendering his understanding of its proper declamation and elocution. He did so, in this instance, within the constraints of that complex of melodic functions and associations that I have described in terms of intonation, recitation, cadence, tenor, and final. And then there is a third sort of donnée that we can isolate by focusing upon the opening a phrases shown in Ex. 6.4: these eventually get around to the sort of movement between tenor and final that goes on in the other a phrases. But they begin in a special way that is unique to opening verses, and in their beginnings they are all but identical in detail. The differences are understandable, again, in terms of the stress situations of the words. These phrases, then, would be appropriately described in terms of a specialized and concrete opening formula that stands out of a more generalized melodic procedure. Taken together those are the resources with which the singer could declaim the first word or words of the tract. The relationship between special formula and general melodic procedure is one that I shall try to sharpen further, in this and other contexts. But first it will be necessary to step back and consider some consistencies of melodic progression from one verse melody to another. First, there is an order of cadence tones for the four phrases that is the same in all verses: d–c–f–d. This constitutes an elementary tonal action in which each phrase plays its role; it conditions the details of melodic movement in each phrase. The cadence to c brings the melody to rest at mid-point on a tone that is most sharply contrasted with the final, and the short third phrase provides a pivot for the return to the final in the fourth phrase. The possibility for the third phrase to play this pivotal role rests on second-mode associations of its cadence note, f, in two directions. On one side it is the goal of a direct intonational movement from c; on the other it is the point of melodic tension that is discharged through the fall to the final. This is to say, the
168
Homer and Gre g ory
tonal action through which all the verse melodies are directed has the same fundamental basis as the second-mode psalm tone.⁴⁰ Second, there are consistencies from verse to verse with respect to the details of movement within each phrase. To begin with, this is a consequence of the general constraints already identified: the centrality of the recitation note in all phrases, and the fixed order of the cadential notes. In b phrases, for example, the result is a direct motion from the tenor through e and d to c. This motion has become stylized as a melismatic cadence that is uniform for something like half of each b phrase. With rare exceptions the c phrases reverse this motion, picking up the cadential c of the preceding phrase, and moving up to the tenor and cadence note f. This, too, has become stereotyped, and in most verses the phrase is fully taken up with a uniform cadence of six neumatic elements (marked 6, 5 . . . 1 above phrase c of verse 3 in Ex. 6.3). Exceptions are verses 2 and 11 with only the last three elements, verse 1 with the last two, and verses 4 and 14. In both the b and c verses, then, we have something like what we observed in the a phrases of first verses: general constraints produce general uniformities, and in such contexts there are highly particularized standard passages. It is important to observe at once that the level of detail to which such passages have become standard goes beyond what is determined by the general constraints. Passages somewhat different in detail would have worked under the same constraints. That will be demonstrated in particular cases. I shall refer to the system of constraints of a melody or a phrase as its ‘formulaic system’, and to the standard passages as ‘formulae’. Formulae have become stereotyped through practice under the control of formulaic systems, and this has resulted from and been instrumental for a tradition of oral performancecomposition. Formulaic system and formula, as concepts for the study of the transmission of plainchant, stand for a kind of knowledge that the singer had in knowing a melodic type. It was held along with knowledge of the overall process for melodies of that type (for the verse melodies of tracts of the second mode, the four-phrase form and the role of each phrase). This is said as an alternative to the notion that the singer had ‘memorized’ each melody, individually and as a whole. But there is no difficulty about locating these ideas in the theory of memory that was described above. In that theory scheme, form, order, arrangement are persistent and are instrumental in remembering. That corresponds to what I have suggested about overall form and the formulaic systems of phrases. But in incorporating the controlling role of such factors as ⁴⁰ This statement is supported by the evidence; but not so the assertion, often made, that the tract melodies are elaborations of the psalm tone, and what follows from that, that the four phrases of the former are intonation, mediation, flex, and termination respectively. Further to this subject, see below, n. 53.
Homer and Gre go ry
169
reciting note and cadence note, ‘formulaic system’ also touches on the concept of ‘salient features’. ‘Formula’ has a double association as well. It, too, may function as ‘salient feature’, in that standard formulae were points of orientation in the reconstruction of chants. But standard formulae are the outcome of a process of stereotyping. This corresponds with particular closeness to the conclusions from experiments in remembering for, as we shall see, standard formulae are especially phenomena of beginnings and ends. I shall return one more time to the relationship between formula and formulaic system as basis for the study of a group of melodies. But first it will be necessary to attempt a description of the context from which those concepts have been borrowed: the theory of oral epic.
IV In the 1930s Milman Parry published the investigations into the formulaic composition of the Homeric epics that gave the principal impetus to a transformation of his field. The ultimate consequence was the theory of oral composition, developed by Parry and his student Albert Lord.⁴¹ From beginnings in the field of Homeric studies, it has become the basis of intense activity in the study of medieval literature in Old French, German, and English, and it touched off the systematic study of the Yugoslavian epic tradition that was its laboratory. As we learn the circumstances of its development, interesting parallels in the field of chant studies must come to mind. The hypothesis was addressed to the Homeric Question, which, like the question raised at the beginning of this chapter, asked about methods of composition and transmission. The Iliad and the Odyssey had early been regarded as the written literary productions of a single author, handed down from the beginning through written sources. In the seventeenth and eighteenth centuries that tradition came to be questioned simultaneously from critical aperceptions about the forms and language of the poems and from reflection about the likely fact that they were composed and communicated without the aid of letters. During the era of scientific philology in the nineteenth century those earlier aperçus became the basis of systematic analysis that aimed to show the poems as layers of accretion upon archetypes, expanded and reworked through successive ⁴¹ The following description is based principally on Milman Parry, The Making of Homeric Verse: The Collected Papers of Milman Parry, ed. Adam Parry (Oxford, 1971), and Albert Lord, The Singer of Tales (Cambridge, Mass., 1960; repr. New York, 1968). In some ways the articulation of the theory due to students of medieval epic is more pertinent to the study of plainchant, as I shall suggest. Specific references to that side of the subject are based principally on J. D. Niles’s dissertation, ‘The Family of Oral Epic’ (University of California, Berkeley, 1972). Extensive bibliographies may be found in all three works.
170
Homer and Gre g ory
stages of editing. Those analyses remained tacitly bound to the assumption of texts originally written or in any case fixed and transmitted in the way of written texts. Their objective was the identification of archetypes and of the chronology of their transformation. But they produced only eternal and irresolvable disagreements. Parry circumvented those disagreements and once again placed the aesthetic question at the centre. The power of his work derived from his direct insight into the formulaic construction of the poems. He put aside the reading of that characteristic as a sign of patchwork construction and interpreted the formulae instead as language invented to express particular essential ideas on the model of persistent sound patterns and metres. He understood that in the light of such an interpretation the formulae must be regarded as traditional, that is, that they can only have been produced by many singers, over several generations. Finally, he understood that such a tradition is by nature an oral one, and he was able to demonstrate that through his study of the still-living tradition of oral epic in Yugoslavia. It is important to observe that Parry never proved—and never attempted to prove— that the surviving texts of the Homeric poems are themselves records of oral performances. His thesis is that we can understand the poems only if we see them as belonging in the domain of oral literature, something that is fundamentally different from written literature. Parry recast the questions of the making of the poems and of their critical analysis so that they became two sides of the same question. Oral composition is composition done in the act of performing. The basis of an oral composition is a framework that is described in terms of two kinds of elements: themes and formulae. ‘Themes’ refers to the full hierarchy of subject matters of the narrative: the subject of the song as a whole (the Odyssey is a return song about the return of Odysseus from the Trojan War); the major episodes (the journey to Hades, the return to Ithaca); and particular scenes (arming for battle, recognitions). The themes at every level are traditional. The Odyssey is parallel in form to other return songs: the arrival at home and the battle to reclaim wife and domain are common to such songs; arming scenes and rescue scenes become standard in form and content, and they are common to songs on different subjects. Themes on this lower level may, indeed, be repeated in detail from one song to another. Each scene or episode is developed in verse, with lines that are regulated according to some metrical-syntactical-semantic pattern. The choice of language is under the control of these factors at several levels: the metre and its articulation (e.g. a hexameter with feminine caesura); the general sense of the sentence (e.g. ‘and X replied’); and particular ways in which the segments of the line are filled to give the particular sense wanted (e.g. a predicate phrase going to the caesura, a proper name with epithet occupying the second half of the line). Each of the component phrases, or the line as a whole, can be standard and repeated many times in
Homer and Gre go ry
171
precisely the same wording, but they do not need to be. It is the pattern that counts most heavily for the oral composition. It requires some care to locate the concept of formula in such a scheme. Parry first defined formula as ‘a group of words which is regularly employed under the same metrical conditions to express a given essential idea’. Ultimately he loosened that conception, de-emphasizing the exact recurrence of word series and emphasizing instead their patterns. That emphasis has been of special importance to Lord and the students of medieval epic. The singer’s competence rests on knowledge of the metrical pattern, the grammatical construction, and the semantic content required for the line. This may bring about actual formulaic identities, but that is not the objective, and it is not essential. The set of constraints that determines the line is its formulaic system. Oral poetry may be closely constrained by formulaic systems—that is, it may be very strictly formulaic—although the rate of recurrence of formulae may not be very high. If the singer has accumulated a repertory of standard formulae, each serves him when his knowledge of theme and formulaic system calls for a phrase of its characteristics. They belong to the complex of habits and associations that enable the singer to compose at high speed. Lord has put it that ‘they emerge like trained reflexes’. But this is not to say that the technique of oral composition depends on the singer’s retention of a stock of standard formulae which he or she strings together. The formulaic analysis of an oral poem is a matter not of making a count of recurrent phrases, but of identifying the formulaic systems that regulate the verses of the poem.⁴² Now for the Homeric epics this can come to very nearly the same thing because of a characteristic that Parry called Thrift: ‘For a given essential idea and a given metrical slot there is one formula and one formula only.’ But in the medieval epics, where the verses are equally under the control of strict formulaic systems, the rate of formula recurrence is much lower. There one speaks rather of Variation than of Thrift, and hence the emphasis of the medievalists on the pattern over the formula. In developing the component themes of a song, the singer may be guided more by the accustomed way of developing the theme than by any uniform idea of the way the theme goes in the song. Themes enter the song as independent units, and because of this there may be inconsistencies between them. There may not be a strong enough idea of the consistency of the song to prevent that. Under the pressure of oral composition the singer may lose track, and can, so to speak, make a ⁴² The parallel issue for the study of plainchant is raised, of course, by the concept of ‘centonization’ and the sort of formula tabulation that is presented as analysis in connection with that concept. It should be apparent that the point of view being developed here is in sharp opposition to that approach. The argument, which would carry this essay too far afield, is made in Ch. 7.
172
Homer and Gre g ory
wrong turn. If a poem shows such inconsistencies, if its verses are under the control of formulaic systems, if it shows recurrent formulae, if it is built up of standard themes, these are taken as prima facie evidence that it is the product of an oral composition. Lord has written, ‘The singer’s mode of composition is dictated by the demands of performance at high speed, and he depends upon inculcated habit and association of sounds, words, phrases, and lines.’ How these habits and associations are formed, how the singer learns—not songs but how to make them—is a matter of the greatest importance for the theory of oral composition. What the singer knows has been assimilated from an early age. Learning is one of the mechanisms of oral transmission. (In plainchant tradition before the age of writing we cannot think of transmission simply as the passage of whole melodies through adult singers. From the moment he entered training as a boy, the singer was assimilating melodic principles and patterns according to type and liturgical moment. That, too, must have been an aspect of transmission.) The oral composition hypothesis finds such resonance in the theory of memory that I believe we may regard the former in all essential respects as a special case of the latter. Rather than standing in mutual opposition, they confirm one another in striking ways.
V Examples 6.4 and 6.5 represent attempts to show the transmission of a phrase in terms of formulaic system and formula. ‘Transmission’ is taken in two senses: the transmission of a melody through different word sequences, and its transmission with the same word sequence through different sources. In these examples the latter sort of comparison is made only between the Gregorian and Old Roman transmissions, but a full analysis in these terms should compare different sources within the same traditions. The upper stave in each example shows those features that every phrase for that position incorporates. It is a representation of what I shall call the formulaic system for that phrase position. There are optional features (shown in parentheses), or there may be choices between two or more alternatives. This can be an important aspect of the formulaic system and must be shown in the analysis. The staves beneath the representation of the formulaic system show a sampling of the melodies that have been composed for the respective positions in the melody. The recurrent formulae in the melodies are identified by horizontal brackets. A cautionary note about these examples is necessary. What is shown in the upper stave in each instance is not a reduction. It is range modules, pitch goals and centres, and surface melodic figures that constitute the constraints of that
173
Homer and Gre go ry
Ex. 6.4. Initial phrases of the first verses of six Gregorian and five Old Roman mode 2 tracts with schematic representation of the obligatory features of such phrases
Gregorian
3
Tu es
vas 3 Gau de
1
3
Qui ha
bi
tat
1 De
3
2 us,
1
3
2
Con fi te
mi
ni
Do
mi
no
3 De Old Roman
ne
ces
si
ne
ces
si
1 De
ti
ta
ti
bus
me
is
3
1 Do
ta
3 4
2 mi
1
bus
ne 2
Do mi
3
ex
au
di
au
di
vi
4
ne 4 3
De us, 1 Qui ha
bi
De us 3
2 tat
me
us
4 in
ad iu to
ri
o
particular phrase type. Melodic details around and between these elements can sometimes be interpreted as prolongations and diminutions, but not consistently or systematically so. The exact alignment of the specimen phrases under the schematic representation of the formulaic system should therefore not be regarded as exactly fixed. Thus, with reference to Ex. 6.4 I would not wish to say that formula 2 is an elaboration of the notes c–d, but rather that it is a concrete way of moving through those notes. Then I would say that the representation of those notes in the formulaic system stands for the fact that in every a phrase there is some sort of movement through them. It is a representation of what is obligatory. Nothing mysterious is intended by these remarks; the point is simply that ‘formulaic system’ is a construct, a way of referring to the singer’s assimilated sense of the pattern
174
Homer and Gre g ory
Ex. 6.5. Second phrases of several verses of the mode 2 tract Deus, Deus meus in Gregorian and Old Roman transmission
Gregorian V1
De us me
V12, (13)
V11
lau
de o
V7, (8) sum ver
us
re
spi ce
in me:
da
te
e
um:
re
le
o
nis:
mis,
et
non ho
mo
Old Roman V2, (9, 12) (a)
V3, (11, 13)
sa
cla ma
V5, (7, 8) pa tres
lu
te me
a
bo
per di
em
no
stri
of a melodic phrase as that can be inferred from the study of numerous instances of phrases like it. In section III there are extensive observations about a phrases in general, and I go on from those for the discussion of Ex. 6.4. What sets off a phrases of first verses is the opening downwards from the final to the low confinal. In this lower tetrachord the norm for melodic movement is the frame d–c–A. The tetrachord is shown vertically in the upper stave of Ex. 6.4 because the formulaic system allows a beginning in either direction through that frame. But eventually the movement is downwards, and in most cases this takes the form of standard formula 1. From the tetrachord there is a turn back to the final, and formula 2 is a standard version of that turn. It is less standard than the opening formula, and that is important. From the final the melody heads for the recitation note through the figure c–d–f, and once there it is guided by the principles regulating a phrases in general. For a further clarification of the distinction formula–formulaic system I have identified with dotted brackets a hypothetical formula 3 at the upward turn to the tenor. The point is that I would not wish to identify those passages as a recurrent
Homer and Gre go ry
175
formula comparable to formulae 1 and 2. Because of the range of variation in those passages, to identify them as such would be to do no more than to show that they have in common the core c–d–f. But that has already been observed with the identification of the formulaic system, hence the very distinction would be undermined. Formulae 1 and 2 are virtually identical in each occurrence, and their description must be quite different from the description of the formulaic system at those points. This might be perceived as a pedantic refinement for the analysis of what is on paper. But as my purpose is to attempt a description of oral composition the distinction is quite important. It refers to different kinds or levels of knowledge with which the singer approaches the performance: knowledge of the general way the melody must go at each turn, and knowledge of particular melodic details. The latter sort of knowledge is reflected in the written transmission by standard formulae, which are by and large standard in their notation as well. Passages like those under bracket 3 do not answer to Lord’s description of formulae that emerge ‘like trained reflexes’; those under brackets 1 and 2 do. Some observations on the a phrases in the Old Roman transmission are in order. The two versions of the phrase from De necessitatibus (Ex. 6.4) are quite similar to one another, and it is tempting to think that one has been derived from the other or both from a common original. In some sense, near or distant, that is always so, for the alternative is to suppose that the tracts of mode 2 were invented in the first instance as formulaic systems. That would distort the concept altogether and it would be contrary to good sense. A formulaic system can be transmitted only through melodies, but that is not to say that the singer can assimilate it only as a melody. He learns one melody and he imitates its pattern in inventing another like it. At some point his inventions do not refer back to the models of concrete melodies but are based on his internalized sense of pattern. That is more or less what it would mean to speak of the transmission of a formulaic system. It suggests a way to develop a systematic comparison between the Old Roman and Gregorian traditions without becoming entrapped at once in perpetual disputations about the directions and priorities of the relationship. To elaborate, we may note first that in Ex. 6.4 the two versions of the phrase in Qui habitat do not show the same degree of surface resemblance as those of De necessitatibus; indeed they are no more alike than are, say, the Gregorian Qui habitat and the Old Roman De necessitatibus. So there is not the same temptation to argue for the direct transmission of a fixed melody–text complex Qui habitat in one direction or another. There may have been such a transmission, but we have no evidence for it; we can be certain only that the phrases are under the control of the same formulaic system. In fact all that the evidence allows us to say with confidence is that in both traditions one sang a tract text Qui habitat; that in singing it one followed in both places the same melodic type, defined by the same formal
176
Homer and Gre g ory
arrangement of four phrases, each governed in both places by the same formulaic system. Our reflections on oral transmission enable us to think that this can have sufficed for the independent production in both traditions of the tracts Qui habitat. There is no need to reach beyond the evidence and speculate how one version of ‘the melody’ might have been transformed into the other. To close the circle on the argument, we note a close resemblance between the Gregorian Confitemini Domino and the Old Roman Domine exaudi. That is largely a matter of the near identity in both of the standard formulae 1 and 2, which appear to be involved in the transmission while other standard formulae are isolated in one or the other tradition (e.g. Old Roman formula 4). To return to the versions of De necessitatibus, the resemblance there suggests a transmission on the level of the formula as well as of the formulaic system. Is that substantially different from speaking of the transmission of the melody as a whole? Not in this instance, as it happens. But it has the advantage of allowing us to make comparisons throughout the whole complex on the basis of a consistent set of concepts. If we can move through both traditions, making systematic comparisons in these terms, locating the levels of similarities and dissimilarities, we shall greatly improve our position for making statements about the relationship between them, both ontological and historical. The b phrases shown in Ex. 6.5 operate under a very much simpler system: optional intonation, recitation on f supported by d, a shift to e, then the cadence through d to c. There is a variant of the system, in which the recitation note gives way to a strong displacement by g (verses 7 and 8 of both the Gregorian and Old Roman versions; also Gregorian verse 11 and Old Roman verse 5). In these instances the f is regained and then displaced by e as in the main system. Standard formulae have developed in both the Gregorian and Old Roman transmissions from the cadence into the phrase to the point of displacement of the recitation note by e. But the formulae are not the same, which rather strengthens the case for putting it that a single formulaic system is operating but that different formulae have become standard, rather than speculating at this point about the transmission of whole melodies.⁴³
VI In the discussion about the tracts of the second mode I took note of the common ground shared by the verse melodies and the psalm tone of the second mode. This reflects an important general fact: the formulaic systems of individual phrases do not necessarily belong to one melodic type only. They may regulate phrases in ⁴³ Of course this statement is not intended to apply at all to the case where Gregorian melodies were taken over whole into the Old Roman repertory long after the separation.
Homer and Gre go ry
177
melodies of different types belonging either to the same or to different modes. The identity of the melody-type is established by the overall form of the melodies (incorporating the formulaic systems), and by its standard formulae. When a formulaic system is shared by two melody-types we may wonder what would have prevented the singer from giving out a standard formula belonging to one while performing a melody of the other type. That can be predicted from the experiments on remembering, and it would be something like the phenomenon of inconsistency in the oral epic. In fact the written transmission of chant shows many instances of this, and it is strong evidence for the role of formula in the oral composition of plainchant. I show a single specimen case. The Gregorian transmission of the verse melody of the second-mode alleluia type Dies sanctificatus follows a sequence of four phrases, each strictly regulated by a formulaic system. Example 6.6 displays the alleluia with the verse Dies sanctificatus in the Gregorian and Old Roman transmissions (CD tracks 6 and 7 respectively). The form of the verse melody is bipartite and parallel, with each half usually comprising two phrases: ||AB||A'C||. The B phrase, when it occurs, is usually a standard cadential formula. That is the formula that I wish to isolate for this demonstration, and I shall identify it as Formula Q. While I believe it is most likely to have originated with this very old melody-type, that is something of which we cannot ever be certain, and it is not at all necessary to establish it for the purposes of the argument to be developed here. What is essential is to recognize its formulaic role and context. Phrase A begins with an ascent to the confinal and a descent to the final, where the melody may recite for a time. This arch is repeated, and there follows a contrasting descent to c in preparation for the cadence on that tone. The B phrase (and of course the C phrase as well) resolves the c to the final d, using f in a pivotal role similar to the role of f in the second-mode tracts. In the Gregorian Formula Q the phrase first descends into the lower tetrachord. But that detail, striking as it is, seems not to be essential to the function of the phrase in the course of the whole. This is an important observation about the formula–formulaic relationship: the most sharply etched details of a standard formula may be quite incidental to the requirements of the formulaic system under which it has developed. There is corroborating evidence for this contention in phrases that occupy the position of Formula Q and discharge its principal functional obligation without making the characteristic descent into the tetrachord. First, Ex. 6.6(c) shows the B phrase as it is in a correlate alleluia verse transmitted in two French sources of the eleventh century. One sees that it amounts essentially to the final phrase of the melody (see Ex. 6.6(a), phrase C). This makes the melody completely parallel, and opens the possibility that originally the second phrase was identical to the fourth and that the B phrase became specialized as Formula Q sometime later. Second, the Old Roman melody (Ex. 6.6(b)) is organized like the Gregorian, and its B
178
Ex. 6.6. The verse melody of the alleluia Dies sanctificatus in standard and special Gregorian and Old Roman transmissions. The special Gregorian transmission of the second phrase is transcribed from London Harl. 4951 and Paris lat. 780 A. a. Gregorian (standard) 8
V Di
es
san
cti fi
ca tus il
lu xit no
il
lu xit no
bis:
b. Old Roman V Di
8
es san cti fi
ca
tus
Formula Q
a. Gregorian (standard) 8
ve
8
ve
ni
te
gen tes
et
a
do
ra
te
Do
mi
num
b. Old Roman ni
te gen tes et a do ra te Do mi Alleluia V Benedictus sit Jesu, London, BL Harl. 4951, fo. 230 v ; Paris, BNF lat. 780, fo. 97
c. Gregorian (special)
num
qui
a
(= phrase C) est
8
quem
ro
co
nat
Do
mi nus
A'. a. Gregorian (standard) 8
qui a
ho
di
e
de
scen dit lux
ma
gna
b. Old Roman 8
ho
C.
di
e
de
a. Gregorian (standard) 8
su
8
su
per
ter
ram.
b. Old Roman per ter
ram.
scen
dit
lux
ma
gna
Homer and Gre g ory
B.
bis:
Homer and Gre go ry
179
Ex. 6.7. Conclusion of the alleluia verse Paratum cor meum in Gregorian transmission from Paris lat. 903, Benevento VI. 34, and London Harl. 4951 V Paratum cor meum. Paris, BNF lat. 903, fo. 253 1 (a)
(= Alleluia and jubilus beginning) 8
(glo)
ri
a
me
a.
V Paratum cor meum. Benevento, Archivio cap. VI. 34, fo. 259 1 Formula Q (b) 8
(glo)
ri
a
me
a.
V Paratum cor meum. London, BL Harl. 4951, fo. 260 (c) 8
(glo)
ri
a
me
a.
phrase has a function similar to that of its Gregorian counterpart. But the Old Roman B phrase is not Formula Q. Formula Q is one of several solutions taken for the problem of how to end the third-mode alleluias of the type Paratum cor meum, a most irregular melody of very unstable tradition (three endings are shown in Ex. 6.7). The uncertainties shown by this melody are characteristic of pieces in E. Throughout much of its extent it behaves as though d were its final and f its reciting note. In all versions of the melody the final melisma is prepared by a cadence on d. In each case the melisma begins with a move up to f. In Ex. 6.7(a) that is taken as the opportunity to begin the Alleluia and jubilus again, and close on the ‘correct’ final, e. (This ending, which in Paris 903 goes only so far as is shown in the transcription, may be compared with the beginning of the alleluia in the Vatican Edition. The latter, it will be noted, takes the Beneventan ending, Ex. 6.7(b).) Example 6.7(b) ends with Formula Q. Example 6.7(c) begins like 6.7(b), but ends with the final phrase of the alleluia Dies sanctificatus (Ex. 6.6(a), phrase C). The graduals classified under mode 1 are better characterized as a loose collection than as a melody-type. There are points of identity between them, especially in their verses. These, however, are not sufficient to give the impression of a strong underlying Gestalt for the group. In this respect they are more like the alleluias of mode 3 than the alleluias and tracts of mode 2. Among the melodies of this group are some that incorporate Formula Q , as final melisma of the verse or of both the respond and the verse. Example 6.8(a) shows the latter case, with the cadential melismas of respond and verse. Note that the music leading up to the cadential melisma of the verse is identical to that in the alleluia Paratum (Ex. 6.7(b)). Example 6.8(b) is the Old Roman version of the end of the verse from the gradual Timete. The cadence here is not Formula Q , but a formula that was sung in the Old Roman tradition at the conclusion of the alleluias in D, such as the alleluia Dies santificatus (Ex. 6.8(c)). (The Old Roman
180
Homer and Gre g ory
Ex. 6.8. (a) Cadential melismas of the respond and verse of the Gregorian gradual Timete dominum; (b) cadential melismas of the gradual Timete dominum and the alleluia Dies sanctificatus in the Old Roman transmission Gregorian gradual Timete Dominum
R ti men
(a)
V bo
Formula Q
ti
bus
e
um. Formula Q
no.
Old Roman gradual Timete Dominum (b) 8
om
ni
bo
no.
Old Roman alleluia, V Dies sanctificatus (c) 8
su
per ter
ram.
alleluia and gradual repertories in D are both much more uniform than their Gregorian counterparts, and that reflects a general characteristic of the former tradition that will presently be discussed in connection with the concept of Thrift.) This rather brings us about full circle. An overview may be helpful, using the following designations: Formula Q , Formula Y (C phrase, Gregorian alleluia Dies sanctificatus), Formula Z (C phrase, Old Roman alleluia Dies sanctificatus). Gregorian Formulae:
Q all. Dies all. Paratum grad. Timete
Old Roman Y all. Dies all. Paratum ..........
Z all. Dies . . . . . . . . . ⁴⁴ grad. Timete
Taken together, the similarity of contexts where the formulae recur and the general weakness of Gestalt in the third-mode alleluia and first-mode gradual groups suggest a way of understanding these borrowings. I believe we can interpret the phenomenon in the light of Bartlett’s idea of crossing. But the question whether we can interpret it in the light of the inconsistencies in oral epic has two sides, and we must address them separately. On one side these examples support the notion of a distinction between the integrity of the melodic Gestalt and the integrity of its component themes or for⁴⁴ The Paratum melody is not in the Old Roman repertory.
Homer and Gre go ry
181
mulae—the weakness of the first-mode gradual and third-mode alleluia as against the persistence and independence of Formula Q.⁴⁵ But can we interpret this phenomenon in the sense of a slip—great Gregory nodding—and, even more important, shall we understand the resulting melodies as inconsistencies? Not, certainly, when the borrowing melodies have such an inconsistent transmission themselves. But that is not always the case. There are strong melodic types that have, through the course of their transmission, incorporated whole phrases of other clearly defined types.⁴⁶ I believe that such mutants complement the more consistent transmission of melody-types in supporting the view of chant performance before the age of notation as a process of making or remaking, based on the complex of patterns and associations and concrete formulae that the singer has assimilated. The examples that have been cited in this section bring up two further questions about transmission that are intriguing and fundamental. They cannot be explored here, but they must at least be mentioned. First, what does it say about the relationship between the Gregorian and Old Roman transmissions that each version of the gradual Timete takes its cadential formula from the local version of the alleluia Dies sanctificatus and that the two formulae are not the same? There are other instances of this sort of relationship, and their investigation would seem to hold some promise. At the very least I conjecture that these are further instances of relationships at a deep-structural level that point more surely to early history than do similarities of surface detail. Second, what does it suggest about the state of the transmission that really striking instances of drifting formulae and crossbreeds of sharply different melodic types could be written down and admitted into the standard repertory? Here I venture a single speculation. It may suggest that the incongruities were not recognized or that, if they were recognized, their tradition was nevertheless respected. In either case it would suggest that by the time of writing these melodies were being transmitted more as individual melodies than as concrete instances of melodic types; that a degree of standardization and individuation in the repertory had taken place; that the model for the performance was a particular plainchant, not the principles of a melodic type. This would be to suggest, finally, that the tradition of oral composition had changed its nature by the time of writing down to a condition in which one could conceive of representing it in writing, ⁴⁵ In this connection it seems important that Formula Q is frequently cued in the Gregorian manuscripts, not fully written out, and this in its appearances in all three melodic groups. ⁴⁶ Some examples: the fifth-mode graduals In Deo speravit and Protector noster, beginning in the manner of first-mode graduals; the fifth-mode gradual Bonum est confiteri Domino, beginning also as first mode, continuing as fifth mode, then taking up a second-mode gradual melody in the verse at et veritatem; the fifth-mode gradual Ego dixi, following the same course as the preceding, the second-mode melody beginning in the verse at in die mala; the seventh-mode gradual Benedictus Dominus, beginning as the eighthmode tract. We have no idea how much of this sort of thing there is in the repertory. A systematic study of the phenomenon would no doubt tell a great deal.
182
Homer and Gre g ory
and perhaps, too, that writing down was something of a rescue operation in circumstances that were a natural consequence of the history of transmission itself. Of course this is a hypothesis about which I speculated earlier for other reasons.
VII The last subject to be introduced here provides both an important link in the sequence of reflections about oral composition and the occasion for raising an important question about the degree to which we can interpret any of the written transmission in the light of oral composition. It is the evidence for a counterpart in the transmission of plainchant of that conservation of means, that reiterative characteristic that Parry called Thrift. 1. In the tracts of mode 2 certain phrase melodies recur in their entirety in one or more verses. More often than not by far, those verses succeed one another directly. Thus in Deus, Deus meus: Gregorian transmission
Old Roman transmission
Phrase
Verses
Verses
a b c d
12 and 13 7 and 8; 12 and 13 5–10 7–10; 11 and 12
5 and 6; 12 and 13 7 and 8 5 and 6; 7 and 8 7–13
(
The phrases shown in the column to the left are identical in the pairs or series of verses shown under the two transmissions. In each case the singer repeated what he had sung in the corresponding place just immediately before. 2. There are several hundred Gregorian alleluias, compared with fifty-four in the Old Roman tradition. Of the fifty-four only eight have unique melodies and the remaining forty-six were sung to a limited number of melodic types. For a single example we may again consider the melodic type for the alleluia Dies sanctificatus. Of sixteen Old Roman alleluias in D, twelve were sung to that type. The jubilus melodies in all twelve are identical, and the verse melodies comprise from two to four phrases, each of which is essentially like its counterpart in the other alleluias. The forms found are ABAC, ABAC (with the second and third phrases elided), ABC, and AC. (Robert Snow has given an overview of the situation for all the Old Roman alleluias.)⁴⁷
⁴⁷ See Snow’s chapter on Old Roman chant in Apel, Gregorian Chant, 496–9.
Homer and Gre go ry
183
3. Paul Cutter, comparing the Gregorian and Old Roman responds of the second mode, showed that in effect the melodies are under the control of a single formulaic system in both transmissions. A major difference is in the far greater proliferation of formulae in the Gregorian transmission, or a higher concentration of current formulae in the Old Roman transmission.⁴⁸ Snow has reported a similar relationship for the graduals of modes 3 and 5 (he might well have added those of mode 2).⁴⁹ 4. Thomas Connolly, studying the introits of both traditions from the point of view of their formulaic construction, reports that in the Old Roman tradition ‘fewer but more rigidly applied formulas are used than in the Gregorian tradition’; and that for the Old Roman introits ‘there are, in fact, only four basic cadences’.⁵⁰ Indeed, once alerted to this phenomenon, we shall notice throughout the transmission all manner of signs of a tendency to be thrifty in the use of musical material; a tendency to sing again what one has just sung in the same plainchant, in the same office or mass, in the same cycle of feasts.⁵¹ There is a strong temptation— and for good reason—to interpret all this in the light of Parry’s concept of Thrift, and to see in it a kind of accommodation to the exigencies of an oral tradition in the face of an ever-growing need for liturgical music. It would be, in this view, another indication of the gathering crisis for which writing down was the solution. Is the case for oral composition stronger, the more evidence there is of thrift? If continued study of the Old Roman transmission sustains the impressions that have been recorded so far, these phenomena are more pronounced there. Then what would it mean to say that oral practice shows more vividly through the Old Roman written transmission than through the Gregorian? Three answers present themselves: first, that the former was a more oral—i.e. more ‘improvisatory’— tradition; second, that the signs of the oral tradition have been more nearly removed from the Gregorian written transmission; third, that the Old Roman tradition was oral for a longer time, that it could develop more of the signs of an oral tradition. The evidence we have in view here is not that melodic ornateness and prolixity that some interpret as signs of improvisation, but rather an economy of means. That rules out the first answer, and it weakens the second as well, for ‘removing the signs’ of oral composition would have amounted to the deliberate proliferation and loosening of formulaic practice. The third answer is the most coherent and the most plausible. ⁴⁸ ‘Die altrömischen und gregorianischen Responsorien im zweiten Modus’, KmJb 54 (1970), 33–40. ⁴⁹ In Apel, Gregorian Chant, 492–4. ⁵⁰ ‘Introits and Archetypes: Some Archaisms of the Old Roman Chant’, JAMS 25 (1972), 157–74. ⁵¹ In the office, the use of the same melodic type for different antiphon texts on the same day; in the mass, similar melodic setting for the same text when it is sung as both introit and gradual, for example (see Connolly, ‘Introits and Archetypes’); the concentration of melody types in feast cycles, e.g. the secondmode gradual in Advent and Easter Week.
184
Homer and Gre g ory
But there is reason to temper these suspicions and speculations. The evidence that I have associated with the idea of Thrift, taken neutrally, comes down to a matter of regularities and uniformities. So, indeed, does formulaic construction. The caveat we must sound is that any uniformity we find in the written transmission might be due, entirely or in part, to a process of redaction that went along with the process of writing down. Such a process surely affected the Gregorian transmission and probably the Old Roman transmission as well. This spectre confronts every inference that is made from the written transmission about what preceded it. That includes the oral-composition hypothesis, but also an assumption that has been made for a very long time in the field: that agreement among versions is a sign of antiquity and points the way for the discovery of archetypes.⁵² We are left with three fundamentally different ways of assessing the meaning of any demonstrated uniformity of transmission: (1) the uniformity is a product of the Carolingian recension of the Franco-Roman chant, (2) it speaks for the antiquity of the transmission and points the way to archetypes; (3) it is both a product and an instrument of the tradition of oral composition. The extent to which the first interpretation applies will have to be determined in each instance through systematic and comparative study of the transmission, organized according to melodic types and sources, with criteria that can be stated and tested for what is and is not a significant similarity. But the choice between the second and third interpretations can never be settled on the grounds of the concrete evidence alone, for they do not really produce conflicting statements about what is and is not true. It comes down, rather, to differences in objectives and outlooks. If they exclude one another, it is by virtue of their preoccupation with different questions. The study of the special characteristics of form in an oral tradition, in view of its transmission, is not meant to replace for ever the search for archetypes. On the contrary, as it brings with it an approach to the comparative study of melodies that is systematic and tangible, it is liable to clarify the question of archetypes and to set the pursuit of archetypes on more secure ground. That is the promise of an opening just now towards the investigation of oral form and oral transmission.⁵³ ⁵² This is an extreme statement of the danger. Many uniformities and systematic practices may have become more uniform or systematic through writing down, and perhaps some originated at that time. But it is doubtful that this consideration will seriously undermine the basic premiss that the written transmission has descended from an older, traditional practice and show instead that all that has been observed here was accomplished with one stroke on parchment. And once again the attribution of oral form is a critical, as well as a historical, assertion. ⁵³ I must express as a footnote a thought that springs directly from the last paragraph but that is too diverting for the main text. If we once begin to speak of a common ancestor for the tract melodies and psalm tone of the second mode, then we must enlarge the family to take in other types classified as second mode, for they show the same family characteristics: introits, alleluias, office responds (graduals, perhaps, but there we must be cautious). Then we are on the trail not of an archetypal melody, but of an old tonality that was the basis for the invention of melodies. The melodies became differentiated into the several melody-types, a process in which the development of formal strategies (e.g. those of our tract and alleluia) and of standard formulae played a role. The tonality became normalized, first as one of the octoechos, then as one of the octave species.
Homer and Gre go ry
185
One question will surely have arisen in the minds of thoughtful readers of this chapter: all of my inferences have been about ‘the singer’, whereas we think of the performance of the antiphonal and responsorial chants of the mass and office as involving choirs. Does this not make the hypothesis of oral composition inapplicable to all but the small minority of chants for soloists, such as the tracts on which I have mainly based the main analysis? Not at all. I shall briefly mention the most important considerations: (1) The hypothesis is meant to apply only through the eighth century at the latest for the Gregorian tradition; for the Old Roman, speculation can be of only the most tentative sort, pending more systematic study; (2) the first testimony we have of antiphonal choral singing is given by Amalarius of Metz, who wrote in the ninth century; earlier than that there is anecdotal evidence of refrain singing by the congregation, and of responsorial refrain singing by a choir; (3) it is not self-evident that refrain singing by a professional choir is in conflict with the practice of oral composition. It may be that refrains were worked out orally by the choir leader and taught to the choir. This question reminds us in general of how very much in the dark we remain about the place of the choir in the early history of plainchant. See Helmut Hucke, ‘Zur Entwicklung des christlichen Kultgesangs zum gregorianischen Gesang’,⁵⁴ and C. Gindele, ‘Doppelchor und Psalm-Vortrag im Frühmittelalter’.⁵⁵ ⁵⁴ Römische Quartalschrift, 48 (1953), 147–94. ⁵⁵ Mf 6 (1953), 296–300.
186
‘ C e n t onat e ’ C h a n t
chapter 7
‘Centonate’ Chant: Übles Flickwerk or e pluribus unus?
} Considering the circumstances and motives under which the chant of the Latin Church came to be called ‘Gregorian’, that label is laden with implications (but only implications) about the aesthetic character of the practice it identifies—implications about its originality, its unique style, its consistency and coherence, these because it flows from the creative impulse of a single author—and about its authority, given the standing of that author among the Church Fathers. The implicit aesthetic character insinuated on the chant by events of the ninth century was drawn out and made explicit by theorists of the twentieth century working, in fact, to develop a ‘Gregorian aesthetic’ of which the elements of that implicit medieval aesthetic character were the core—the properties of originality, formal unity, consistency, and coherence stemming from the creative impulse and genius of an author. In short, what those theorists projected upon Gregorian chant was a work aesthetic and an artist concept that had come to be formulated for European music since the late eighteenth century. This project coincided with the project of restoring the chant to its ‘original’, ‘authentic’ state on the basis of manuscript studies and newly representing it in authoritative editions and a performance practice for the modern Roman Catholic Church, prescribed in those same editions and in phonograph recordings. The two projects joined together in defining the dimensions—aesthetic, archaeological, and editorial-performative—of a newly invented musical tradition. Through the intensive detailed engagement with the chants that was entailed in all aspects of this project, however, theorists were confronted with an apparent anomaly and a challenge. The formulaic character of many chants—the use and reuse of fixed melodic units in many different chants—could be thought to run counter to key aspects of the Gregorian work aesthetic: unity of form, internal coherence, originality, and uniqueness of authorship. The challenge was to find an interpretation of the method of composition with formulae that would be consistent with, and would even reinforce, the aesthetic doctrine. Such an interpretation was found in the concept of ‘centonization’, under which a chant melody is
‘ C e n t onat e ’ C h a n t
187
regarded as a kind of mosaic. By emphasizing that its unity is a unity in diversity in the crafting of which the Gregorian composers demonstrated their artistry, theorists were able to make a virtue of seeming adversity. With this ingenious idea, which has been one of the main interpretative concepts for the analysis of chant, theorists achieved a success that was denied their literary counterparts. At the end of the eighteenth century the close analysis of the form and content of the Homeric epics produced a challenge to the work aesthetic along the same lines, basically, as that just described. For there, too, the recognition of formulaic composition threatened the reception of the poems as works of individual genius. But this challenge initiated a controversy that has not subsided (see Ch. 8), and for which no compromise formulation such as the theory of centonization has been proposed. But for all its ingenuity, the theory of centonization rather badly misrepresents both the practice that it is meant to explain and the criteria of value on which its products would have been judged in its time. It presents a static view of melodies and their component segments as objects, conceived as though fixed in writing, whereas they had to be made or remembered and apprehended in the flow of performance. Conceiving formulae as though they were stones in a mosaic, the theory cannot take account of their functions in the progression of melodic formulae.That constitutes not only a misrepresentation, it marks an inadequacy of the theory as an analytical theory. And in its practice of labelling and classifying formulae it falls short, too, for it raises, but fails to resolve, the problem of how formulae are to be recognized, identified, and compared. This chapter reinforces the demonstration of the preceding one that historical understanding must come from the analysis both of the evidence about the past and of the history of interpretation into the present.
} If the criticism of art aims to demonstrate the individuality, the unity, and the coherence of art works and aims to show that these spring from the creative impulse, what are critics to do in the face of works that appear to recompose old stuff and that bear the mark of many, or of no, hands? In 1795 Friedrich August Wolf published a study about the Homeric epics that was meant to be purely philological, but which in the end stirred up exactly these issues of aesthetics; it caused a great sensation.¹ For Wolf made powerful arguments to show that the works that stood as paragons of the genius of European literature had in fact been handed down from diverse origins through a process of successive transmission and revision; his evidence suggested that great Homer was a chimera. ¹ Prolegomena ad Homerum, sive de operum Homericorum prisca et genuina forma variisque mutationibus et probabili ratione emendandi (Halle, 1795).
188
‘ C e n t onat e ’ C h a n t
With this work Wolf appeared to pull down the Homeric poems from a stature we can measure in this verse, written by Goethe in 1774: Artist’s Morning Song I step up to the altar, And read, as is fitting, My devotional prayers From sacred Homer. Questions about originality and about unity and diversity in art occupied Goethe’s mind throughout his creative life, and often they were focused on the Homeric problem.² His attitudes towards Wolf ’s arguments shifted back and forth: the arguments were irrelevant to an aesthetic appreciation of the poems, which is a matter of feeling, not knowledge; the arguments were sound and showed the poems to be the products of nature; they showed that, despite their diverse origins, the poems manifest the powerful tendency of poetry towards unity, etc. Goethe spoke of these shifts as the ‘systole and diastole’ of his mind. In 1821, his seventy-second year, he addressed a poem to Wolf in which he recanted his earlier conversion to the scholar’s arguments: Homer, wieder Homer— e pluribus unus Homerus Scharfsinnig habt ihr, wie ihr seid, Von aller Verehrung uns befreit, Und wir bekannten überfrei, Daß Ilias nur ein Flickwerk sei.
Sharp-witted as you are, You freed us from all reverence, And we granted too quickly That the Iliad is but a patchwork.
Mög unser Abfall niemand kränken; Denn Jugend weiß uns zu entzünden Daß wir ihn lieber als Ganzes denken, Als Ganzes freudig ihn empfinden.³
May our apostasy trouble no one, As youth has ways of inflaming us, For we think of it rather as whole; In its wholeness do we joyfully experience it.
In 1916 the classicist Wilamowitz picked up on Goethe’s fourth line and wrote: ‘Whoever . . . believes that the Iliad has sprung through one sudden act from individual lays . . . renounces historical understanding. He considers it, as it is, a wretched patchwork (ein übles Flickwerk) and is concerned with it only to separate the good bits from the collector’s worthless mortar.’⁴ The epithet stuck, and as these things happen, it came to be known in the literature as the expression of
² See Joachim Wohlleben, ‘Goethe and the Homeric Question’, Germanic Review, 42 (1967), 251–75. ³ Gedenkausgabe der Werke, Briefe, und Gespräche, 28. August 1949, ed. Ernst Beutler, 24 vols. (Zürich, 1950), i. 561. ⁴ U. von Wilamowitz-Moellendorff, Die ‘Ilias’ und Homer (Berlin, 1916), 32. The translation is from William M. Calder III, ‘Ein übles Flickwerk’, Classical Philology, 64 (1969), 35–6.
‘ C e n t onat e ’ C h a n t
189
Wilamowitz’s attitude to Homer—the very attitude against which he had polemicized.⁵ The field of plainchant studies has had its problems about the origins of its objects, too; and the same questions about unity and diversity—in the objects and in their origins—have been of central concern. The questions have somewhat come to rest, and the way they have settled—to the extent that they have—has very much influenced the shape of the field. Around the beginning of the twentieth century two quite separate traditions converged into critical and scholarly attitudes that informed the Gregorian reform movement. One is the tradition of belief about the origins of medieval Western chant. The other is the tradition of belief about the coherence and indivisibility of art works as a consequence of the unity of artistic creation. In Carolingian times ideological, political, religious, and ecclesiastical factors combined with the facts of musical performance and transmission in the creation of the doctrine of an authentic, original body of plainchant called ‘Gregorian’. The sense in which one has understood Gregory’s role as its creator has varied from the idealized, metaphorical connotation transmitted in the pictorial tradition showing Gregory receiving the chant from the Holy Spirit in the form of a dove, to the more realistic sense suggested by the remark of John the Deacon (c.870) that Gregory compiled an antiphonarius cento, that is, an antiphoner put together from diverse sources.⁶ What matters most about this doctrine is not the variety of its modes of expression but its central idea, which runs through all of them: that Gregorian chant, as it is transmitted in the sources of the Middle Ages, is a unified body of work, of uniform origin. The idea of a moment of origin does not exclude a redaction as the creative act. Gregorian chant was a canon, authentic, uniform, inspired, as surely as were the books of the Bible.⁷ The objective of both the medieval and the ⁵ See Calder, ‘Ein übles Flickwerk’. The issue over the question of patchwork or unity in the Homeric epics has never been resolved. But in some circles it has been set aside, and the Homeric Question reformulated. For a summary and bibliographical citations see Ch. 6. The issue is still very much alive in the German-speaking countries; see Harald Patzer, Dichterische Kunst, und poetisches Handwerk im Homerischen Epos (Wiesbaden, 1972). Patzer shows that the work of the Parry school regarding formulaic composition and the oral transmission of the epic has met with considerable resistance in those countries, in large measure because of its threat to the conception of creative unity that the Homeric epics have been taken to exemplify. ⁶ See Ch. 6 for a report on this tradition. ⁷ It is a matter of side interest here that there is a parallel, older pictorial tradition in which St Jerome is shown receiving the Vulgate from the dove (see Croquison, ‘Les Origines de l’iconographie grégorienne’). In 1912 Dom Raphael Andoyer characterized Gregory’s work as one of ‘organization and revision rather than of composition, strictly speaking’. Gregory’s sources, in Andoyer’s hypothesis, were the melodies transmitted as what is now mainly known as ‘Old Roman’ chant. He describes the process of transformation as one of ‘fixation, of definitive purification, of absolute perfection of the genre. This apogee is occupied by the Gregorian oeuvre. St Gregory, as the tradition of the high Middle Ages affirms, is the greatest musical genius of Latin Christian antiquity, and the pre-Gregorian chant, in that it permits us to understand the creative activity of the great Pope, is, for that reason, more interesting to us’ (‘Le Chant romain antegrégorien’, Revue du chant grégorien, 20 (1912), 69–75, 107–14). Bruno Stäblein has long been arguing for
190
‘ C e n t onat e ’ C h a n t
modem Gregorian reform movements was and is the restoration of that canon as such.⁸ During the time when the modern plainchant restoration was getting under way, aesthetic ideals that prevailed among music historians of all interests added an explicit emphasis to the tradition about uniform origin. The Gregorian canon was regarded as the product of a uniform creative act, of a controlling artistic will. This emphasis provides the underpinning for the doctrine that Gregorian chant is uniform in its principles of musical style, that each chant is informed and made coherent by the same overall principles of artistic design, and that the consistency of those principles is demonstrable through the systematic analysis of the melodies. The discipline that is comprised by that study is what Peter Wagner called ‘gregorianische Formenlehre’ and what Dom Paolo Ferretti called ‘estetica gregoriana’.⁹ That the disciplines they describe come to the same thing, and that they rest on the same premisses, we shall soon see. As their two books, whose titles provided these designations for the discipline, have had the greatest continuing influence on analytical-historical studies in plainchant down to the present time, I shall focus the discussion for a moment on the question of their objectives and premisses, through specific citations of their texts. In the introduction to his Formenlehre, Wagner took it as his objective to set forth the forms of the Gregorian melodies according to historical and aesthetic criteria—to penetrate to the essence of Gregorian art through the study of form. There were two requirements: access to the old medieval melodies, not the corrupted editions in use since the seventeenth century, and an approach that combined stylistic analysis with aesthetic examination. What is meant by aesthetic examination is a search for the sources of beauty in the melodies. That is no idealistic or subjective thing; it is open to systematic, objective analysis. It is a matter of laying bare the principles of musical logic, the aesthetic norms, the künstlerische Gestaltungskraft of Gregorian chant. The principles of musical logic are such as these: The Periodengesetz, a norm for the melodic phrase as an arch that stretches to a point of maximum tension and returns to the point of departure; principles of symmetry, relating antecedent phrases to consequent ones; the principle of architectonic organization, building on the first two principles; the relationship between an underlying architectonic Gestalt and its melodic elaboration. essentially the same assessment of Gregorian chant, most recently in ‘Die Entstehung des gregorianischen Chorals’, Mf 27 (1974), 5–17; he places the time of the redaction about a half century later, however. ⁸ This is still reflected in the new critical edition of the Gradual. Dom Gajard has transmitted and endorsed Dom Gueranger’s remark: ‘When manuscripts of different epochs and countries agree on a version [of a phrase], we may be certain that we have retrieved the Gregorian phrase’ (Le Gradual romain: édition critique, ii: Les Sources (Solesmes, 1957), 11). ⁹ Peter Wagner, Gregorianische Formenlehre; Paolo Ferretti, Estetica gregoriana, ossia Trattato delle forme musicali del canto gregoriano (Rome, 1934) .
‘ C e n t onat e ’ C h a n t
191
In his discussion of the alleluias (pt. iii, §II, ch. 6), Wagner distinguished an old layer whose melodies are ‘regel- und planlos’ from the majority of transmitted alleluia melodies, which are models of symmetrical organization. The later melodies are the work of ‘aesthetic deliberation’, and they differ from the earlier ones as a ‘skilfully laid-out flower bed’ differs from a ‘luxuriantly proliferating growth’.¹⁰ (What this alleges for the earlier alleluias is, in principle, a curious proposition about any traditional art—that it is, in effect, incoherent. That this seems not to have troubled Wagner may be read as a sign that his assessments of earlier—and also of ‘Oriental’ chant are hardly more than a backdrop for his characterizations of composed Latin chant since the Gregorian era. I shall return to this issue when it comes time to speak of ‘centonized’ chant.) The idea of ‘composition’ served Wagner not only for the distinction between early and late chants, but also for the basic ontological classification of the repertory and hence for the basic plan of his book. The repertory falls into ‘bound’ forms—recitation formulae and such that are more or less fixed and whose selection is determined by the text to be intoned—and ‘unbound’ or ‘free’ forms, whose melodies are freely composed within the constraints of form called forth by the liturgical moment. Ferretti’s title is also explained in the introduction to his book: Musical aesthetics is the science or the philosophy of the beautiful applied to the art of sounds. It promotes . . . the understanding of the artistic work itself. . . . Musical aesthetics occupies itself only with the architectonic structure or construction and with all the constructive elements of a musical composition. [It is] the theory of musical form, [for] form constitutes the essence of the work itself, and [is] the primary source of all the properties and qualities of beauty, objectively considered (pp. vii–viii).¹¹
Ferretti alludes throughout to the composers of Gregorian chants, even to ‘Gregorian melody writers’, as ‘artists’, creating under the inspiration of their ‘personal genius’ and under the guidance of the ‘logic’ of musical and musicaltextual principles. Those principles he sets forth with a characteristic zeal for ¹⁰ Wagner erected a watershed for the separation of the two layers on foundations that have always seemed insecure: he saw a transition to the new style in the alleluia Adorabo, assigned to the celebration of the mass for the dedication of a church. Inasmuch as the earliest such celebration took place in 608 on the occasion of the consecration of the Parthenon in Rome as a Christian church, Wagner took that date for the beginning of the new style. Risky as that conclusion may be on evidential grounds, the period it suggests is about right from the viewpoint of the tradition concerning the origins of the Gregorian style. Perhaps that argued in its favour. There is discussion of Wagner’s chronological layering of styles in Ch. 5. ¹¹ Here, as later, Ferretti reflects a viewpoint of Benedetto Croce, whose Estetica come scienza dell’espressione e linguistica generale: teoria e storia of 1900 (11th edn., Bari, 1965) was widely read in the first decades of this century. Croce writes in the first chapter: ‘The aesthetic phenomenon resides in form and is nothing other than form . . . Every [artistic expression] is a unique expression. [Artistic] activity consists of the smelting of impressions into an organic whole. That is what one has always meant to say in introducing the idea that the art work requires unity.’ (Note the explicit emphasis on the activity as the source of artistic unity. This is apparent in the formulations of Wagner and Ferretti as well. Further on this point, see n. 12.)
192
‘ C e n t onat e ’ C h a n t
classification. Their application by the Gregorian composers resulted in chants that are ‘organic’, ‘harmonious’, and ‘homogeneous’. ‘The aesthetics of the musical form of Gregorian chant was born and developed gradually with the rebirth of the old liturgical melodies, restored to their genuine and traditional purity’ (p. viii). Ferretti, like Wagner, conceived that the restoration of the Gregorian canon, based on the premiss of original creation, was a precondition for the study of Gregorian form. Then, with the old melodies in hand, they both aimed to confirm that premiss by demonstrating through systematic analysis the artistic integrity of the individual melodies and the stylistic uniformity of the repertory. Gregorianische Formenlehre and Estetica gregoriana, on the one hand, and the Gregorian reform movement, on the other, were interdependent from the start. If Gregorian chant embodies the aesthetic ideal of a unity that springs from the creative act, how does that square with the facts of its transmission—especially with the fact of its use and reuse of standard melodic material? How, as Wagner asked, can a satisfactory art work result from the setting together of standard melodic passages? Ferretti found a general way to assimilate that fact in an interpretation of plainchant composition that had been published in 1881 by François-Auguste Gevaert: Whereas in the modern epoch the composer’s first objective is to be original, to invent his own motifs with their harmony and instrumentation, the composers . . . of liturgical chant worked in general with traditional themes, from which, through a process of amplification, they produced new chants. . . . In music, as in architecture, invention consists in constructing new works with the aid of material taken from the common domain.¹²
It is of primary importance for Gevaert that the difference is only in the source of the material, not in the creative process of composition or in the artistic value of the product. The comparison with architecture is pointed. It makes no more difference, so far as the creative process and its products are concerned, whether the composer invents his tunes from scratch or bases his work on traditional tunes, than it does whether the architect finds his material in the quarry or in a ruin. The reuse of traditional material in the plainchant had to be reconciled with the artist concept and with the doctrine of original creation in that domain. The genius of Gregorian composers would be manifested in the skill with which they transformed the traditional material. That is an interpretation of the artist concept that ¹² Ferretti, Estetica gregoriana, 89; François-Auguste Gevaert, Histoire et théorie de la musique de l’antiquité (Ghent, 1881), 315–16. From Gevaert’s later publication, La Melopée antique dans le chant de l’église (Ghent, 1895), we learn what he had in mind: the role of whole melodic models—of thèmes—in the transmission of the office antiphons, which he compared with the roles of the ancient Greek nomoi and the Indian ragas.
‘ C e n t onat e ’ C h a n t
193
corresponds exactly to the broadening of the doctrine of original creation to include revision and redaction, whether by Gregory the Great, by his successors in the papacy, or by anonymous papal or imperial composers.¹³ When Wagner posed the question of whether an amalgam of standard melodic passages could qualify as art on his terms, he was rhetorically imagining the reaction of a hypothetical reader to his discussion of ‘wandering melismas’—a concept he had just introduced. With this term Wagner referred to a phenomenon in the transmission of the mass graduals: the recurrence of standard melismas throughout modal groups of gradual chants. (I shall turn in a moment to his answer.) Similar things about the transmission of the office responsories had been reported by Walter Howard Frere.¹⁴ Ferretti called attention to the spread of this phenomenon through the Gregorian repertory and dubbed it ‘centonization’. In his theory of Gregorian aesthetics, he established ‘centonization’ as one of the principal techniques of composition and ‘centonate’ chants as one of the main categories of its formal types. As far as Wagner and Ferretti are concerned, the phenomenon was understood, instantly upon its recognition, in the light of the aesthetic doctrine that I have summarized here. There was nothing forced about that.¹⁵ Their descriptions of the phenomenon read like new expressions of the same doctrine. The idea of centonization falls right out of the theoretical context in which it was first articulated. This will be clear from a few passages of Wagner’s and Ferretti’s texts. (Wagner did not say ‘centonization’, and, on another level of which I shall come to speak, that is an important fact. But his conceptual filing of the phenomenon is exactly parallel to that of Ferretti.) Ferretti: One takes from the fund of traditional music a certain number of formulae belonging to a certain modal and melodic type, and one welds them, one chains them together in such a way as to make an organic whole, homogeneous and logical. At first glance, the melody thus gained will appear to have been created in a single impulse. One will call it original, ¹³ Croce afforded positive theoretical support for this interpretation. The passage cited in n. 10 continues thus: ‘that is to say [the work requires] unity in diversity. [Artistic] expression is a synthesis of multiplicity and diversity in unity. . . . Whoever has the idea for a tragedy, throws a large number of impressions into a melting pot; expressions that are long familiar are smelted with new impressions into a single mass, just as one might place in a furnace unformed brass lumps together with the most refined statues. In order to create a new statue, the old refined ones must be reduced in the same way as the formless lumps.’ This image makes clear the real meaning of Croce’s precept that the aesthetic moment is in the form (see n. 10), for the substance of art, as opposed to its form, is comprised of impressions, which are natural and given. The creative act is, therefore, an act of assimilation and synthesis, the act of transforming impressions into expressions. This also makes clear the sense in which it is the creative activity that is the cause of unity in art. ¹⁴ Antiphonale Sarisburiense (London, 1901–5), introductory volume. ¹⁵ Historians of science have lately been quite seriously interested in the way that the background of theoretical beliefs against which scientific observations are made condition the way in which phenomena are reported. See, in particular, Norwood R. Hanson, Patterns of Discovery (Cambridge, Mass., 1958), and Thomas Kuhn, The Structure of Scientific Revolutions (Chicago, 1962) .
194
‘ C e n t onat e ’ C h a n t
as it is natural and harmonious in all its parts [pp. 95–6]. The Gregorian melody writers . . . [‘le melografi Gregoriani’ (sic)—Gevaert had spoken of the ‘melographes’ of Antiquity; this is revealing of an attitude of which I shall come to speak shortly] composed chants that are true centos, in which musical formulae, taken from here and there, were skilfully bound together. (p. 117)
Wagner (in response to his own rhetorical question): True, one can gather together various stones that, considered in themselves, impress the connoisseur for their polish, their sparkle, and their rareness. But if a valued piece of jewellery is to be produced from them, then they must be given a rich setting, they must be tastefully grouped, and they must be artfully joined together. Are the old melismas chained to one another in this way? The answer is entirely in the affirmative. . . . In connecting such passages together, the old Choralisten [translation is dangerous here] showed a mastery that has just now been uncovered again. (p. 395)
Willi Apel, at the conclusion of his review of the graduals, associated himself with these views of Wagner, but went one better: ‘These melodies, fascinating in their analytical detail, are equally admirable for their synthetic quality, for their cohesion and union. In fact, the perception of their structural properties greatly enhances their significance as unified works of art, no less so than in the case of a sonata by Beethoven.’¹⁶ These passages make two kinds of assertions about centonized chants: they present a theory about their composition and a statement of their value. Why do the assessments have such an ideological ring? What was the need that brought forth these encomiums that border on mania? The music is wreathed around with laurel as it emerges from obscurity into instant monumentality. This writing gives the strong impression that the theory of centonization is functioning to maintain the music on the pedestal to which it has been raised, and that the exercise of gregorianische Formenlehre reaches its consummation in these panegyrics. In the conceptual world in which these authors operated, the only alternative could be the assessment of the melodies as wretched patchworks.¹⁷ Before, I suggested that the establishment of a gregorianische Formenlehre and the Gregorian restoration were interdependent. A more accurate portrayal would be that the restoration had two aspects that are separable only in abstraction: an archaeological side and a critical side. The theory of centonization belongs on the critical side. For the remainder of this chapter, I wish to focus on that, in its aspects ¹⁶ Gregorian Chant, 362. ¹⁷ There is something reminiscent in this about the way that certain other canons of art have been ushered into a revival: the architecture and sculpture of classical antiquity in the 18th c., J. S. Bach and, later, Palestrina, in the 19th. In all four instances, the revival was more than a matter of archaeological restoration. The art works in view were reborn already clad in baptismal dress that bore the marks of the strong values of their midwives. Their very appearance—dress and all—served to confirm those values. But History—i.e. written, construed History—has this background.
‘ C e n t onat e ’ C h a n t
195
as both compositional theory and analytical theory. Having considered its importance for the Gregorian reform movement, I wish to reflect on its adequacy as a theoretical account of the plainchant transmission. I begin with a very general description of the phenomenon that the theory is meant to explain. My description is in three parts: 1. The repertories of Gregorian chant are transmitted in melodic families that are consistent with respect to liturgical category, to modal designation, and to certain musical features, comprising principally melodic strategies or form and recurrent melodic material. The members of a melodic family are related through those features, and they are distinguished from one another in that they have different texts. That will clarify what is meant by ‘family’ and what by ‘transmission’. 2. In the transmission of melodic families, it is the more melismatic passages that are most stable from one melody to another. A certain number of recurrent melismatic passages, or formulae, provide one of the defining characteristics of melodic families of certain types. 3. These are primarily the psalmodic chants of the mass and office that come down from a time before the practice of musical notation. Underlying every psalmodic chant is a stereotyped melodic procedure with very sharply defined points of departure (intonations and initial formulae) and arrival (medial and final cadences) that articulate the melody. It is at these points of articulation that the melodies show the most stereotyped formulae and that they show them most often. Standard formulae tend to occur with decreasing frequency as the melody moves away from the opening and with increasing frequency as it moves towards the cadence. That is, a melodic family may have a standard opening formula that occurs with high frequency at openings. There may be a formula that immediately follows the opening formula, and the tendency is that it may occur in fewer melodies than the opening formula, but not in more melodies, and so on.¹⁸ For each such family, then, a certain number of formulae may be identified, and individual melodies of the family will show these in greater or lesser degree, separated by passages that are more variable. The principal genres that are thus characterized are certain groups of office antiphons and responsories, certain groups of introits and graduals of the mass, and all the tracts. The phenomenon of stereotyped openings and ¹⁸ In other words, it is not normally the case that a non-stereotyped opening is immediately followed by a stereotyped formula, or that a stereotyped formula is followed by a non-stereotyped cadence. The formulae seem to attach to one another with initial and final formulae as anchor points. This general aspect of the phenomenon of formulaic construction is cautiously described here as a tendency. It is very clear in particular cases (several are described in Ch. 6), but it cannot be stated as a general law for all groups of psalmodic chants in the absence of further systematic studies of transmissions from this point of view. Stated in the most general terms, however, the point is hardly surprising. The melodies move from points of departure and seek goals, and the points of departure and goal points are the most hardened aspects of the piece. That seems musically sound, and in an oral tradition it would have special utility. I shall have more to say along these lines below.
196
‘ C e n t onat e ’ C h a n t
cadences is not restricted to these categories. But of all the chants, the psalmodic ones proceed with the highest degree of regularity through phrases marked by cadences and re-initiations, and they show the phenomenon of formulaic stereotyping in the highest degree.¹⁹ Focusing on this last observation in a slightly different way, the identification of a chant melody as ‘centonate’ is far more than a descriptive statement. It embodies an assertion about how the melody was composed and an assertion about how it works as a melody. The phenomenon that it thus interprets is not restricted to the usual categories of ‘centonate’ chant but is a central fact of plainchant transmission. That is why the theory requires review. I have reviewed the answer that the theory of centonization provides to the question of how the chants were composed, how they came to have the appearance they have. I wish to try a brief formulation of an alternative account as background for further reflections on the theory.²⁰ The point of departure is a reminder of something that we know. Most of the music we are talking about had its origin and was widely transmitted before the age of musical notation. The composition of the chant—whatever that may mean—and its transmission took place through the medium of performance. We have the music we are attempting to analyse from written sources, but it came down first through oral tradition and then underwent redaction. Unless we believe that the redaction removed all the characteristics that the music bore through its oral composition and transmission, any account of it must pay heed to that background. When transmission takes place through performance, the controlling condition is the need to keep the performance going, without notation, without a catalogue of formulae, without a given line on which to improvise. There was not much time to deliberate from phrase to phrase—or if the initial composition was worked out phrase by phrase, to deliberate within phrases—and there was no opportunity to revise. That imposes on us a necessity to consider that any of the general or particular features of a plainchant melody of which we wish to give an account may represent an accommodation to this condition. In place of the paradigm in which one presumes an act of composition that produces a piece which, in the absence of writing, is submitted to memory and then repeatedly reproduced in performance, we might think of a repeated process of ¹⁹ Concerning the association between psalmody and standard formulae, see Helmut Hucke, ‘Psalmodie als melodisches Gestaltungsprinzip’, Musik und Altar, 5 (1952–3), 38–47. For a detailed study of a single family of psalmodic chants, the tracts of mode 2, see Ch. 6. Apel (Gregorian Chant, 323–30) offers an analysis of the same family according to the theory of centonization. His book provides a survey of the phenomenon of formulaic composition in the Gregorian repertory as a whole. The subject has been treated most recently in Theodore Karp’s book, Aspects of Orality and Formularity in Gregorian Chant (Evanston, Ill., 1998). ²⁰ What follows is a brief summary of the account that I have proposed in Ch. 6.
‘ C e n t onat e ’ C h a n t
197
performance-composition—something between the reproduction of a fixed, memorized melody and the extempore invention of a new one. I would call it a reconstruction; the performer had to think how the piece was to go and then actively reconstruct it according to what he remembered. In order to do that he would have proceeded from fixed beginnings and sung towards fixed goals, following paths about which he may have needed only a general, configurational sense, being successively reinforced as he went along and recognized the places he had sung correctly. Different places in the melody would have been fixed in different degrees in his mind; there would have been some places where it would have been most helpful to him to have a note-for-note sense of exactly how it went and others where he could go by this way or that, making certain only that he passed through particular notes or note groups of importance and that eventually he arrived at the goal that he had before his mind’s ear, so to speak. But there was always a tendency for paths to be worn smoother the more he sang the melody. And this might have been so also for the tradition in which the singer worked, including the other singers of his generation and those of other generations.²¹ In place of the idea that the formulae preceded the composition of chants and that composition consisted of putting them together, we might presume just the opposite—that the formulae are actually consequent to the composition of the chants. It may well be that they became stereotyped through precisely such a practice as the one I have just described. The positions of the formulae within the melodies, that is, played a crucial role in the process of the oral reconstruction of chants, one that brought about their classification as formulae. As corollary to this interpretation there is an alternative account of the vaunted unity and organic coherence of formulaic chants. It is not that this quality is a consequence of the judicious selection and skilful combining of formulae by the composers but rather that the formulae of a melodic family have grown up within the same matrix, and as each has developed in a particular functional position in the melody, each will naturally give the impression of being very well suited for its position. And since it is the case, as Wagner observed, that the melodies do follow a logical progression, the progression of formulae also gives the impression of being logical.²² There is no need, however, to account for the uniformity of the ²¹ The central thought here is that the transmission of the music is a matter of both reconstruction and reproduction, hence that its production and transmission are not sharply distinguishable. This idea is meant quite deliberately to cut across the traditional concepts of ‘improvisation’ and ‘composition’. And, although the context here necessarily involves us in reflection about oral transmission, there is no intention of implying that such a situation is uniquely characteristic of oral traditions. The idea of transmission through reconstruction cuts across the categories of ‘oral’ and ‘written’ as well. There is a general exposition of this idea in its applicability to a wide range of medieval practices in Ch. 4. ²² It is hard to believe in the possibility of the continued stable, oral transmission of musical utterances that are not logical and coherent on the basis of some set of principles; a stable transmission would not seem to be possible without them. We do not require the premiss of artistic intent to account for them, and we can hardly count their very existence as the special feature of a style. It is hard to think that Wagner had anything
198
‘ C e n t onat e ’ C h a n t
melodies despite their formulaic nature, nor to count that quality a special triumph over adversity. We would not exclaim how miraculous it is that our arms are attached just where they are to our bodies, where they are most serviceable. The other side of the compositional theory of centonization is its analytical theory. The beginning of an analysis according to that theory is the comparison of all the members of a melodic family and the identification of the recurrent melodic formulae that belong to it. These are usually classified as to their normal functional position within the melodies of the family and provided with a system of labels that take account of the functions of the formulae and also of the variants. (Apel’s system of labelling also takes account of the pitch to which each formula goes, so his analyses give, at once, some sense of a tonal outline of the melody.) Passages that cannot be identified with standard formulae are usually identified as ‘free’ passages, although some analysts label such passages as formulae in any case. The result of this general procedure is an analysis of the melody in terms of its formulaic content and its formal arrangement.²³ Trying to compare these two sides of the centonization theory with the alternative interpretation of the formula phenomenon I have suggested is difficult. Attempting the comparison leaves one with a sensation not unlike that experienced when confronting the sort of drawing Gestalt psychologists like to show: there is only one set of stimuli on the paper, but one sees either a duck or a rabbit— never both at once. Perhaps the crucial point on which the two approaches diverge is that the centonization theory presents a static view of the process of composition, whereas it would seem that an account of a process of oral composition through performance should emphasize the dynamic nature of the process, stressing, as I have already suggested, the inexorable forward movement of the process.²⁴ That accounts for the failure of centonization theory to raise any questions about the utility of the formulae for the composers. Why should one have used traditional material to such an extent? Purely as a matter of cultural or aesthetic value? That also accounts for the failure of the theory to raise any questions about the patterns and orders in the occurrence of formulae within melodies. What sense does it make that all the tracts of mode 8 begin with the same formula? All Beethoven’s sonatas in G do not have the same beginning. The static conception of composition incorporates an attitude towards the material elements as fixed, as res facta. The elements are there, and one composes particular in mind when he spoke of the early alleluia melodies as being unplanned and unregulated other than to provide a contrast for the clarity and order of the new chants. ²³ For a concise illustration of the method, see Apel’s summary analysis of the tracts of mode 8 in Gregorian Chant, 319. ²⁴ In the only published criticism of the centonization theory known to me at the time I wrote the original version of this chapter, Helmut Hucke wrote, ‘Ferrettis Centotheorie ist eine bezeichnende Fehlleistung unlebendiger Überanalyse’. ‘Zu einigen Problemen der Choralforschung’, Mf 11 (1958), 385–
‘ C e n t onat e ’ C h a n t
199
with them in that one performs certain manipulations with them and on them. Ferretti laid out an entire classification system for the operations that composers performed on the formulae in putting them together into compositions (suppression, addition, contraction, etc.). Apel speaks of ‘insertions’, ‘reductions’, and ‘omissions’. Omissions from what? Reductions of what? This way of thinking ties directly into Ferretti’s designation of the composers as melody writers. If that was a slip, it betrays a conceptual structure by which Ferretti was bound. The idea of centonate composition is hardly conceivable apart from the context of writing. It suggests a composer at his work table. The only alternative is to think of it as functioning through memory. But that requires a notion of transmission through memory in which the faculty of memory acts like a sheet of parchment on which signs are deposited.²⁵ These problems point to serious inadequacies of centonization theory as an account of the production of plainchants. The heart of the difficulty with the theory, considered as analytic theory, is best approached through a close look at the ‘cento’ concept itself. The Latin word means ‘patchwork’, and it has been taken into the musical vocabulary from the field of late classical literature. There it refers to a poem or some other composition that has been put together out of lines of text from diverse pre-existing sources. Their juxtaposition pointedly creates new meaning that is not communicated by them individually. Often that effect has rhetorical intent. So in early Christian times one recombined lines from Virgil in centos that revealed the Christian thought that had lain hidden in those lines. Fernand Cabrol and, following him, Ferretti demonstrated that the texts of plainchants are frequently centos of lines or expressions from scriptural sources that may be quite diverse.²⁶ The art of juxtaposing texts of diverse scriptural origin was brought to a high point by the writers of tropes, and there the rhetorical purpose seems quite evident.²⁷ Ferretti clearly intended such an interpretation of formulaic chants in characterizing them as centos. For he wrote that, like the liturgical writers, the melody writers took their formulae from here and there (‘formule musicali prese qua e la’, p. 117), and their special merit is that they forged such organic unity out of such diversity. Are there, in fact, such things in music? The last of J. S. Bach’s Goldberg Variations incorporates a tune whose text begins ‘Ich bin so lang nicht bei dir g’west’ and concludes ‘Kraut und Rüben haben mich vertrieben’. If one doesn’t know the references, it all sounds perfectly coherent; if one does, he has the added experience of coherence and diversity and is amused in the bargain. Sometimes, various late medieval practices in which melodic material from here and there is combined 414. ²⁵ The place of memory in theories of plainchant transmission is discussed in Ch. 6. ²⁶ Fernand Cabrol, Dictionnaire d’archéologie chrétienne et de liturgie (Paris, 1910), ii, pt 2, cols. 3255–9; Ferretti, Estetica gregoriana, 110–11. ²⁷ For example, Paul Evans has shown how Old Testament passages were given concrete Christian
200
‘ C e n t onat e ’ C h a n t
(the motet enté, for example), have been referred to under the heading of cento. The propriety of that is not clear, for what we do not know is whether the quotations were intended as, and would have been recognized as, citations. A modern example comes to mind, and, because it is so arch in this context, it can jolt us into consciousness of what is being said with the centonization theory. In the murder scene of Berg’s opera Wozzeck, the instant after Wozzeck has plunged the knife into Marie’s throat, and as she lies dying, a succession of motifs representing associations in Marie’s life flashes through the orchestra. All the language about centonate chant is appropriate to this case: the composer has judiciously selected from here and there (from several earlier scenes) fixed motifs and he has skilfully combined them into a unified whole. In the original sense of ‘cento’, a special point of meaning derives from the juxtaposition that is not in any of the constituent parts individually (Marie’s life passes through her fading consciousness.) There are such things in music. But ‘centonate’ chant is not such a thing. The constituent formulae of such chants are not taken from here and there, they are by nature of the one context in which they arose and with which they remain normally associated.²⁸ Their juxtaposition is a natural outcome of their development, and they are in no way citations. As David Hiley writes in his Western Plainchant: A Handbook: ‘The use of conventional turns in singing these texts seems rather different, a procedure for which the term “cento” is not really appropriate. These are neither pre-composed scraps to be sewn into place, nor quotations of an already existing composition.’²⁹ Behind the versions of plainchant that have come down in writing is a complex history of transmission. The principal factors are the procedures of oral composition and the processes of redaction and rationalization associated with the writing meaning by being brought into juxtaposition with New Testament texts. See The Early Trope Repertory, ch. 4. ²⁸ In saying that formulae are normally uniquely associated with a single context I do not mean to overlook or contradict the well-known fact that formulae have been transferred from one context to another. There are abundant anomalous cases in which this occurs, but there is also a deliberate and widening practice of melisma transfer. The concept of neuma refers to such a practice (as in the case of the ‘triple neuma’ reported by Amalarius—see Apel, Gregorian Chant, 343). But these are phenomena that belong to a tradition later than the period of the oral composition of the bulk of the repertory with which we are concerned here, and they usually involve longer, formally closed passages. Charlotte Roederer has published an instance that nicely demonstrates the point, a transcription of a second verse for the Christmas alleluia Dies sanctificatus, from the Aquitanian gradual Paris lat. 776. The melody is identical to that for the first verse except for the setting of the opening phrase, ‘Ortus est sicut’, principally an extended melisma on est. Because the melisma is not a part of the melody for the first verse, and because the second verse does not occur at all in Gregorian sources, Roederer concludes that the melisma is an indigenous piece of Aquitanian melody. But it is, in fact, the melody for the opening phrase of the verse Libera me de ore leonis (variously the 10th or 11th verse) of the Gregorian tract Deus, Deus meus for Palm Sunday. As this is one of the oldest Gregorian tracts, and the melisma is recognizable in the oldest Gregorian sources, we may presume that it has been transferred to the Aquitanian alleluia, replacing the opening melisma of Dies sanctificatus. Here is a case where all of the language of centonate analysis and composition theory would apply, and that is instructive, for it is a fundamentally different phenomenon from that to which that language is normally applied. See Roederer’s essay ‘Can We Identify an Aquitanian Chant Style?’, JAMS 27 (1974), 75–99.
‘ C e n t onat e ’ C h a n t
201
down. The theory of centonization, whose deep-lying premiss is that the chants as we find them are the products of uniform acts of composition, tends willy-nilly to confound those factors and to obscure the history. It is time for a brief peroration. The answer to the question posed in my title is: neither; the question is wrong.³⁰ ²⁹ Western Plainchant, 75. ³⁰ The original version of this essay continued at this point with a detailed critical commentary on the analysis of an alleluia chant according to the cento theory published by Kenneth Levy, focusing on difficulties that such an analysis encounters. These come down mainly to (1) the arbitrariness of formula identification; (2) the equally arbitrary identification of ‘variants’ of formulae, together with its misleading implication that an ‘original’ has been identified; (3) the controlling and misleading influence of the system of labelling formulae, and, most disabling for the analysis, (4) the exclusive focus on the contents of formulae, to the total neglect of their functions; that is what gives all such analyses their static character. The analysis is in Levy’s ‘The Italian Neophytes’ Chants’, JAMS 23 (1970), 181–227.
202
L i n g e r i n g Qu e s t i on s a b o u t ‘Ora l L i t e rature ’
chapter 8
Lingering Questions about ‘Oral Literature’
} I have written more than once since the publication of the original version of Chapter 6 that the work of Milman Parry and Albert Lord on the Homeric epics was a springboard for me in this subject, but that I did not continue to refer to it after that initial essay. Yet the ongoing discussions about the Homeric Question and about orality and literacy in the language arts do resonate to such a degree with parallel questions as regards the history of music that I think it can be of value for our field to place a brief account of those discussions into our literature by way of a review of two recent books about the subject: The Theory of Oral Composition: History and Methodology by John Miles Foley (Bloomington, Ind., 1988) and OralFormulaic Theory: A Folklore Casebook, edited by John Miles Foley (Garland Folklore Casebooks, 5; Garland Reference Library of the Humanities, 739; New York, 1990). I include the review with some small revisions reflecting parallels that had not occurred to me when I first wrote it.
} Behind Goethe’s late poem on Homer (see Ch. 7) lies the story of a long-lasting crisis of belief for Goethe, of a scholarly controversy in classical philology, and of the testing of an aesthetic ideology and a cultural self-image. The poem’s addressee was Friedrich August Wolf, who twenty-six years earlier had published a study of the Homeric epics in which he argued that they had been handed down from diverse origins through a process of successive transmission and revision. He concluded that writing could not have been available to Homer, that the texts that had come down to his time were assembled by later redactors, and that before then they had been resident in ‘folk memory’ and transmitted orally.¹ Goethe referred to his youthful enthusiasm for Wolf ’s reasoning, which he could accommodate by regarding the epics as products of nature rather than of ¹ Prolegomena ad Homerum (Halle, 1795; English trans. by A. Grafton, G. W. Most, and J. E. G. Zetzel, as Prolegomena to Homer, 1795 (Princeton, 1985)).
L i n g e ri n g Q u e st i on s a b o u t ‘Ora l L i t e rature ’
203
genius. But he struggled with his aesthetic intuition of the unity of the epics, and he tried the compromise that they manifest the powerful tendency of poetry towards unity—no matter their diverse origins. He spoke of these shifting attitudes as the systole and diastole of his mind. The poem shows how the matter had settled for him, declaring his final belief in unity as the overarching quality of these quintessential works of European literature, and in the absolute dependence of unity on an Author, that is, on the creative spark that only an individual of genius (like himself ) can give off. That these issues are still alive is evident from a review recently published in the New York Review of Books by Sir Hugh Lloyd-Jones, Regius Professor emeritus of Greek at Oxford University, of three books: Albert Lord’s Epic Singers and Oral Tradition (Ithaca, 1991), Barry Powell’s Homer and the Origin of the Greek Alphabet (Cambridge, 1991), and Martin Bernal’s Cadmean Letters: The Transmission of the Alphabet to the Aegean and Further West before 1400 BC (Winona Lake, Ind., 1990), considered together with two strongly dissenting responses in a subsequent issue of the same journal (‘Becoming Homer’ by Andrew Rigby and ‘“Becoming Homer”: An Exchange’ by Ronelle Alexander).² It is evident that what LloydJones calls ‘the old battle between analysts and unitarians’ (review, p. 52) is ongoing, and he leaves no doubt that he is on the side of the latter, especially when that battle is focused as the dispute over whether Homer was a literate or an oral poet. His interest in the two books on the history of the alphabet owes to their considerable advancement, with his approval, of the time in which the alphabet was invented—the first to the eleventh century BC, the second to some time before 1400 BC. Until the 1970s it was commonly held that the alphabet was invented only in the first half of the eighth century BC, and that would have been consistent with the belief that Homer, who is commonly dated to c.800 BC, was an oral poet (Lloyd-Jones seems to regard it as self-evident that had the Greeks used letters in Homer’s day, he would have employed them for writing down his epics, notwithstanding the fact that Plato, four centuries later, still cautioned about the undesirable consequences of writing [in the Phaedrus]). In the ‘Exchange’, Ronelle Alexander writes: ‘Whether or not the alphabet existed at that time is not the central point. Rather, it is Sir Hugh’s fundamental opposition both to the general idea of Homer’s “orality” and to the idea that Parry and Lord could ever have thought of comparing Homeric verse with that of illiterate Yugoslav singers [I shall address this point presently]. He seems to view the former as inconceivable and the latter as repugnant’ (pp. 51–2). And Andrew Rigby adds that ‘A more up-to-date approach supports the notion that the Iliad and the Odyssey were composed in performance without significant back reference or editing. . . . It is far from settled how the fact of the composition-and-performance of the Iliad and the Odyssey should affect their criticism; it certainly cannot be ignored’ (p. 57). ² 5 Mar. 1992, pp. 52–7 and 14 May 1992, pp. 51–2.
204
L i n g e r i n g Qu e s t i on s a b o u t ‘Ora l L i t e rature ’
The idea of Homer as an oral poet was not new in 1795 when Wolf wrote. Robert Wood, an amateur classicist, made a strong case for Homer as an oral bard in his An Essay on the Original Genius of Homer of 1765 (full bibliographic citations not supplied here can be found in the excellent bibliography of The Theory of Oral Composition, hereafter Theory) and addressed ‘the difficulty of conceiving how Homer could acquire, retain, and communicate all he knew without the aid of letters’ (Theory, 3). But the historian Flavius Josephus (born ad 37/38) wrote in his last work, Contra Apionem, ‘Throughout the whole range of Greek literature no undisputed work is found more ancient than the poetry of Homer. His date, however, is clearly later than the Trojan war; and even he, they say, did not leave his poems in writing’ (Theory, 2). Building on this assertion, François Hedelin, the Abbé d’Aubignac, wrote in his Conjectures académiques ou Dissertation sur l’Iliade (1715): ‘since this Homer left in writing no works that bear his name, one must conclude that he never composed them, and if he never composed them one must conclude that he did not exist’ (Theory, 3). Homer was a chimera. And in 1714 the philologist Richard Bentley wrote the following in his Remarks upon a Late Discourse of Free-Thinking in a Letter to F. H. D. D. by Phileleutherus Lipsenius: Take my word for it, poor Homer in those circumstances and early times had never had such aspiring thoughts [as carefully to design his poems for the ages, as someone had suggested]. He wrote a sequel of Songs and Rhapsodies, to be sung by himself for small earnings and good cheer, at Festivals and other days of Merriment; the Illias he made for the Men, and the Odysseis for the other Sex. These loose Songs were not collected together in the form of an Epic Poem, until Pisastratus’s time about 500 years after. (Theory,3)
What is interesting is that these judgements did not unleash the kind of controversy that we see beginning with the reaction to Wolf ’s treatise and continuing into the present; and that is, I think, not because of Wolf ’s more rigorous analysis, but because his thesis ran afoul of an emerging work aesthetic on one side, and a renewed imaging of European culture as a Greek patrimony on the other. Such continued hostility to the thesis as is evident in Lloyd-Jones’s review seems clearly motivated by the former. Lloyd-Jones calls on another eighteenth-century cultural character trait to account for the heating up of ‘The Homeric Question’, as it is called, at that time. He observes that the Iliad and the Odyssey ‘have certain features which the taste of many later ages finds perplexing, if not repugnant. They contain numerous repetitions, including the regular use of certain standard epithets to describe particular persons’ (p. 52). And he continues that ‘the eminently rational eighteenth century asked some awkward questions about them . . . and the most notable attempt to answer them came from Wolf ’. This repetitiveness, and the use of such standard phrases, have long since come to be discussed as the formulaic style of the poems.
L i n g e ri n g Q u e st i on s a b o u t ‘Ora l L i t e rature ’
205
A characteristic statement about this was published in 1923 by Antoine Meillet, in Les Origines indo-européennes des metres grecs: The Homeric epic is made up entirely of formulas which are transmitted by the poets. If one takes a sample passage, he can quickly recognize that it is composed of verses or parts of verses that are found again in the same textual form in one or more other passages. Even those verses whose constituents one does not find in another passage also have the character of formulas, and no doubt it is only by chance that they are not preserved elsewhere. (Theory, 9)
Now if it were said about a new novel or film that there is nothing in it that is not formulaic, that would be taken as a negative comment (or at least a comment relegating the work to a lower order of genre productions—say a Kung Fu film), suggesting that the work lacks originality, that it was turned out (one might even say cranked out) in a hurry, that it lacks subtlety, depth, and conceptual sophistication. That can help us to imagine the perplexity into which these poems—whose greatness was a given—cast scholars from the eighteenth century onwards. This is only compounded by the suggestion that the poems were composed and transmitted without the use of writing, which cannot be squared with the a priori and deeply personal conviction that ‘intricacy of design and subtleties of soul [are] wholly alien to the oral technique of composition’³ and that ‘the existence of subtle links between the different parts of the poems’ shows that they ‘cannot be the production of simultaneous composition-in-performance’ (Lloyd-Jones, ‘Exchange’, 52). At this juncture certain parallels with the reception of Latin liturgical chant come into view: 1. Gregorian chant has occupied a position in constructions of the history of Western music somewhat like that of the Homeric epics in constructions of the history of Western literature—not only as font but also as paragon and paradigm. Music as diverse as sonatas and symphonies by Beethoven and the compositions of Anton Webern have been certified according to its standard. 2. Close analytical study of Gregorian chant and related traditions reveals a formulaic style in the principles that determine the course of melodies from one moment to the next, and in the use of recurrent, stereotyped melodic materials. This observation has challenged scholars who profess a work aesthetic in which the individuality, unity, and coherence accruing to the work through the creative activity of an author are the central musical virtues. The history of chant analysis and criticism has largely been determined by the efforts of scholars to square that fundamental fact of chant style with this fundamental aesthetic ideology, efforts that have not reached any conclusive end point. The general idea, however, has been precisely what Goethe proclaimed in the title of his poem: e pluribus unus, ³ Denys Page, The Homeric Odyssey (Oxford, 1955), 14.
206
L i n g e r i n g Qu e s t i on s a b o u t ‘Ora l L i t e rature ’
unity in diversity (see Ch. 7). Ironically, that slogan would now suit the beliefs of musical scholars more than those of literary ones. 3. The suggestion that the formulaic character of the chant may be interpreted in the sense that what has come down to us is the redaction of a traditional and oral art of liturgical singing (and consequently that Gregory the Great, as author of the chant, is a chimera), simultaneously raises the question of the nature, origins, and history of the writing system in which the chant was ultimately recorded. But here there is a major difference. No one seriously disputes that the invention of music-writing systems and their use for the transmission of liturgical chant considerably postdates the formation of those repertories, but 4. the following sentence in Ronelle Alexander’s response to the Lloyd-Jones review may well be keyed to the state of chant scholarship: ‘The very concept of orality, of the possibility that singers can create and pass on to subsequent generations such highly complex verse [read chant repertories] without the aid of writing, is an idea so powerful that its full implications are only beginning to dawn on us’ (‘Exchange’, 52). It is because of these parallels, and not because of any suggestion of similarities between Latin liturgical chant and the Homeric epics or any of the traditions of heroic verse with which they have been compared that the books under review are of interest to students of music history. Just in case readers should wonder whether those volumes are indeed being reviewed, virtually everything that has been recounted here so far (save the remarks on Goethe and the Lloyd-Jones review and its respondents) can be read in John Miles Foley’s The Theory of Oral Composition. This is a detailed report on the theory that had its origins in these enquiries about the Homeric question, and that became a ‘discipline’, Foley calls it, as a methodology was developed and as something of that methodology was tried, in turn, by students of the Bible, Anglo-Saxon poetry (e.g. Beowulf ), the chanson de geste (e.g. the Chanson de Roland), medieval Hispanic poetry (e.g. Poema del mio Cid), medieval German poetry, Byzantine and modern Greek poetry, Arabic folk epic, African poetry, the folk sermon in the southern and south-western United States—Foley counts ‘over one hundred ancient, medieval, and modern traditions’ (p. 36). (Other marks of the disciplinary status of the field and Foley’s place in it: he is director of the Center for Studies in Oral Tradition, University of Missouri at Columbia, and editor of the journal Oral Tradition.) This can be considered the triumph of a method, but it also has aspects of the spread of a fashion. The initiating move for this astonishing evolution was certainly Milman Parry’s decision to test his hypothesis, based, as he said, only on logic, that the diction of Greek heroic poetry was both traditional and oral, by analogy with the diction of South Slavic heroic poetry, which could be studied in performance. That resulted in a comparative technique—first manifested in a publication of 1933—and gradually in a field of what Foley calls ‘comparative oral traditions’,
L i n g e ri n g Q u e st i on s a b o u t ‘Ora l L i t e rature ’
207
something that Parry cannot have anticipated. Foley marks the establishment of that field with the publication, in 1960, of The Singer of Tales by Parry’s student, Albert B. Lord. He characterizes it as ‘The bible of Oral Theory’, and associates it with Parry’s own plan to write a book with the same title, but with which he was unable to proceed because of his early, accidental death. Not all scholars in the field see such a close link between Lord’s work and Parry’s as Foley here implies; Lloyd-Jones is one who does not, and he refers to Parry’s own son, Adam Parry, and to Adam’s wife, Anne Amory Parry—both of whom were Greek scholars of some accomplishment until their early, accidental deaths—who aimed to distance their own work and that of Milman Parry from Lord’s more dogmatic positions. Indeed, if the commanding role of The Singer of Tales has defined the field, it has also limited it in some ways—ways that, as it happens, limit the interest of the work for music historians. For one thing, there has been little self-reflective attention to the culture-historical and aesthetic-ideological context in which I tried to place the Homeric question at the outset, and that has, I believe, influenced the conclusions of researchers in both the literary and musical fields in no small measure (for the musical side, see Ch. 9). There is not a hint of this in Foley’s report. For another limiting factor, the ethnological-anthropological perspective from which scholars before Parry investigated living traditions of oral poetry, and on whom Foley does report, has been closed out of The Singer of Tales. There is something ironic about this, as though this field study of what is, after all, a folk art, need not be considered in terms of cultural practice because it was undertaken to test a philological theory about ancient texts. Third, Lord’s dogmatic insistence on an absolute difference between oral and literate poetry can be challenged on empirical-critical grounds and on the grounds of its questionable, ultimately ethnocentric implications about oral and literate mentalities and capacities. It also discourages attention to the interaction of oral and literate processes that should be of special interest to music historians as well as to students of literary history. Fourth, his equally dogmatic insistence on an absolute difference between oral composition and various processes of private composition and public performance from memory is implicated in the neglect that the nature and role of memory in preliterate and literate cultural practice has suffered; and that, too, is of special interest to music scholars. Lord somewhat softened his position on this point near the end of his life, in an essay contributed to his own Festschrift ,‘Memory, Fixity, and Genre in Oral Traditional Poetries’. Music historians interested in this aspect of the subject will be greatly enlightened by Mary Carruthers’s The Book of Memory: A Study of Memory in Medieval Culture (Cambridge, 1990). And last, as the recognition of formulaic style became the opening to the whole field, the major part of its research activity has been the formulaic analysis of texts, along with debates over the questions ‘what constitutes a formula?’ and ‘what level of
208
L i n g e r i n g Qu e s t i on s a b o u t ‘Ora l L i t e rature ’
formula density is sufficient to identify a text as “oral”?’ Anyone not quite inside the ‘discipline’ reading Foley’s reports on these activities and debates may form the impression that this came to be an activity for its own sake, that the field had driven itself into a blind alley. For music historians, especially, there is not much transfer value here, for, at least so far as medieval chant is concerned, no one questions that the tradition originated as an oral one; the questions are all about the implications of that fact in numerous directions. Because of these questions and reservations, interested readers would do well to study The Theory of Oral Composition in tandem with Foley’s rich anthology, OralFormulaic Theory: A Folklore Casebook. I can briefly touch on only a few of its more pertinent items. The most important critic of the field has been the English social anthropologist Ruth Finnegan, and Foley has wisely included her 1976 essay ‘What is Oral Literature Anyway? Comments in the Light of Some African and Other Comparative Material’, in which some of these very issues are raised. Finnegan brings the excluded anthropological perspective to the subject as well as the common-sense and anti-dogmatic reasoning that are her trademarks. She denies that there is a single, definable phenomenon we can call ‘oral literature’, that a clear boundary can always and everywhere be drawn between oral and written forms of verbal art, that oral composition is a single universal process, that there are identifiable ‘oral’ and ‘literate’ styles. She argues that ‘formula’ is too restrictive a concept to apply across the international spectrum of oral forms and genres and cautions against ‘all the detailed analysis’ (p. 252) associated with it. She argues that oral and literate modes have interacted and overlapped for centuries in extensive areas of the world, and she cautions against the notion of an ‘oral mentality’ or ‘psychology’, which smacks of the idea of the ‘primitive mentality’ of non-literate people (p. 259). Her caveats have the aim, not of undermining studies in the oral verbal arts, but rather of both loosening and broadening them, and of introducing a more pluralistic approach. An essay of very great interest to music scholars—both theorists and historians—is Michael Nagler’s ‘The Traditional Phrase: Theory of Production’, from 1974.⁴ Nagler develops a structuralist generative model for the production of phrases that locates traditional meaning in a network of pre-verbal Gestalten and regards the actual phrase as one instance of a traditional potential. His notion of the underlying Gestalt is that of an ‘abstract template that generates phrases in an open-ended family whose members bear a family resemblance to one another’— he refers explicitly to Ludwig Wittgenstein’s conception. In concluding, he suggests ‘it is time to experiment with a synchronic view of oral composition, on the rough analogy of a generative grammar of a natural language’ (pp. 294–6). There is nothing in what he writes that would restrict such a model exclusively to oral composition, and indeed the virtue of his generative theory is that it provides a ⁴ Foley’s Oral Formulaic Theory, 283–312.
L i n g e ri n g Q u e st i on s a b o u t ‘Ora l L i t e rature ’
209
way of thinking about the continuity of oral compositional process into the age of writing. It is a pity that there is so little communication between our fields; just such an experiment as Nagler calls for has been published by Edward Nowacki: ‘The Syntactical Analysis of Plainchant’.⁵ Joseph Duggan, in ‘Formulaic Language and Mode of Creation’, writes about The Chanson de Roland, the greatest of the chansons de geste, presenting his reasons for believing that ‘the Roland we possess must be a very nearly unadulterated product of oral tradition, little changed, except for its orthography, from the form in which it was first taken down from the lips of a singer or written down by a singer who had acquired literacy’.⁶ This notion that the written source represents the recension of what had begun as an oral composition precisely parallels a way of understanding the ontological state of at least some portions of the liturgical chant repertories that have come down in writing from the Middle Ages. Duggan concludes with a defence, not of the oralformulaic theory, but of the potential quality of formulaic composition itself. ‘It is time for medievalists to cease treating formulas like the repetitions or clichés of the written language. . . . Further progress in medieval French epic criticism depends upon recognition that the assumption of the inferiority of oral poetry to written poetry is invalid’ (p. 102). Walter J. Ong represents a school of communication theorists working in the wake of Marshall McLuhan’s fundamental insight about the inseparability of meaning from medium. He articulates a view of orality and literacy as determining not only the style of verbal communications and the management of knowledge, but in turn the very structure of consciousness; he speaks of an ‘oral’ as opposed to a ‘literate economy of thought and expression’.⁷ His books have been most thought-provoking for students of culture interested in the oral-literate dimension. Ong’s fascinating essay ‘African Talking Drums and Oral Noetics’ is included in Oral-Formulaic Theory. In it he describes the functioning of the subSaharan system of communication across great distances by drums, a system that he calls an ‘acoustic speech-surrogate’ (p. 111) in that it both imitates aspects of the sound of speech and follows the spoken language’s grouping of speech-units in the process of what he calls the ‘disambiguation’ of meaning. It is because of the latter especially that he believes the structure of drum communication can be taken as a paradigm of ‘oral noetics’ (the term can be simplified to ‘oral thinking’). Readers who are just being introduced to this intriguing work should be advised that Ong’s general approach has been criticized for its reductionist and essentialist tendencies, which can lead to the dismissal of the specific cultural practices of particular societies. One such criticism that can be recommended for ⁵ IMS Berkeley, 191–201. ⁶ Foley’s Oral Formulaic Theory, 83–108 at 101. ⁷ Ibid. 109–36 at 113.
210
L i n g e r i n g Qu e s t i on s a b o u t ‘Ora l L i t e rature ’
its direct bearing on the hazards of such a notion of ‘orality’ as a state of mind with respect to music is Steven Feld’s ‘Orality and Consciousness’, in The Oral and the Literate in Music.⁸ It will be fitting to conclude with a passage from the introduction to a study on the oral background of the New Testament by a biblical scholar whose work is cited, but not included in this anthology. He is Werner H. Kelber, and the study is ‘Sayings Collection and Sayings Gospel: A Study in the Clustering Management of Knowledge’: In the Western tradition concepts of language and theories of interpretation are heavily based on the pervasive experience of written records. Whether language functions mimetically or referentially, expressively or emotively, it is commonly thought of in terms of a visually accessible medium. Literary criticism, hermeneutics and phenomenology, structuralism and postmodernism tend to operate with a view of language which is explicitly or implicitly tied to the sensibilities of a typographic culture. Modern critical theory is in great measure text-based and text-referential. Signs are not wanting, however, of a growing interest in matters of oral composition and performance, speech and rhetoric, discourse and dialogue, and in manifoldly tangled interactions of speech and writing.⁹
If this observation, coming from a student of the Bible—The Book—were reflected upon musical studies, it might help us to imagine letting go a lingering tendency to treat ‘orality’ as though it were a disease or a bad odour carried by music at the edges of culture, and to allow orality and performance as factors in music criticism. ⁸ Ed. Yoshihiko Tokumaru and Osamu Yamaguchi (Tokyo, 1986), 18–28. ⁹ Language and Communication: An Interdisciplinary Journal, 9 (1989), 213–31 at 213.
Th e Po litics of Recep tion
211
chapter 9
The Politics of Reception: Tailoring the Present as Fulfilment of a Desired Past
} Much of what has been written here thus far reveals that coming to know medieval song has been more a matter of its appropriation to the familiar—placing our distant objects under the control of modern language, concepts, and values—than is consistent with the goal of hearing them in their own voices. In this chapter we pursue the history of a seemingly opposite strategy of knowing that nevertheless works a similarly constraining effect on our objects: conjuring an ancient unfamiliar, an other, with which they are contrasted to show their match to what we regard as ours, and to reveal in them the roots of what we value in our own music. It is a strategy whose paradoxical character—justifying the past with the present and the present with the past—is meant to be reflected in the chapter’s title. The possibility of opening up a gulf between ‘theirs’ and ‘ours’ depends on the premiss of recognizable cultural, national, racial, even gender identities, to which are assigned musical traits, among others. It is the premiss of essentialism. On that premiss such traits are held to be inevitable by virtue—in this case—of the music’s production by members of a certain nation, culture, race, or gender. Such beliefs are ideological in general, and they manifest specific ideologies of nationalism, racism, and gender bias. The history of chant reception by historians, especially from the eighteenth to the twentieth centuries, has this background, and that is something of which we who trade in chant criticism should be aware. There are hints of it in the earlier chapters of this book. It is brought to the foreground in this one.
} The discourse of history can appear as a medium of proud self-portraiture, as the ritual of a culture in narcissistic self-contemplation, glorying in its uniqueness and superiority and in its descent from revered ancestors. This thought catches history
212
Th e Po litics of Recep tion
as a kind of myth and opens it to anthropological, as well as to historiographical, description. I come to this perspective by way of my reading in the modern reception history of medieval Latin ecclesiastical chant, in which I am struck by the attitudes, needs, and ambitions by which that history has been conditioned, and by the feverish ideological pitch of much of the literature. Just as the story that has been fashioned about medieval chant in the modern literature reads as an allegory about the idea of ‘Western culture’, so I mean my story about that process of fashioning to be an allegory about history and about the cultural self-image that we have inherited. At the moment in the modern era when the chant entered a process of conscious reception and became a subject of history it was positioned at the headwaters of the mainstream of European music. To demonstrate its right to that position, historians promptly undertook to display in it just those qualities that were to count as the condition of value and greatness in the culture that validated them in this way, and to claim for it a descent from the music of the culture that was to be a model for Western culture: Greek antiquity. At the same time the chant, as the oldest known European musical tradition, validated those qualities as the essential touchstone for recognizing what was European in music—and, perhaps of greater importance, what was not non-European. The very prominence of chant as a historical subject from the early youth of the discipline of music history in the late eighteenth century is due in large part to its role in assuring that we have a subject for the discipline at all, that is, ‘Western music’, and in fashioning our own cultural self-image. Here is François-Auguste Gevaert, writing in his book Les Origines du chant liturgique de l’église latine: The chant of the Christian church appeared precisely in an epoch when the intellectual activity of Graeco-Roman society was in decline, and it reached the high point of its development at the moment when the aesthetic and literary sense of the West was in a sleep destined to last a long time. . . . As for the musical interest which the old melodies offered to the artist . . . that is not a matter of pure curiosity, as would be the music of exotic people—Chinese or Hindu—where we encounter bizarre melodies and rhythms, piquant perhaps but at bottom strange to our manner of feeling. The ecclesiastical songs . . . have passed into our blood and have contributed to form what we can call our musical temperament. (p. 6)
As simply as that Gevaert identifies at once the absolute divide between what is ‘ours’ and some (bizarre, strange) Other that will be a central theme in my story. The earliest signs of a historical consciousness about chant reach us from the Carolingian era. The story of chant that we read in eighteenth-century literature takes as its narrative core reports from the ninth century about interactions among
Th e Po litics of Recep tion
213
different chant traditions, and those reports reveal an entanglement in issues of cultural identity and conflict from the very beginning. I begin my account with a report on a short article entitled ‘Plainchant’ in the second volume of Jean-Jacques Rousseau’s Dictionnaire de musique, published in Paris in 1768. Plainchant, wrote Rousseau, is a noble relic, very much disfigured, but very precious, of Greek music, which having passed through the hands of barbarians has not, however, been able to lose all its primitive beauties. There remains yet enough of it to render it much preferable, even in its present condition, and for the use to which it is intended, to those effeminate and theatrical pieces which in some churches are substituted in its place.¹
I underscore a few points. The source for the plainchant that had come down to Rousseau’s time is the music of Greek antiquity. That should be understood, indeed, as a claim—or, better, an assumption— about historical origins and continuity, but also as a statement about the music’s qualities and about the idea of Greek antiquity as the source of artistic standards. ‘Barbarism’ has a negative connotation, but ‘primitive’ does not. It enfolds within its meaning the ideas of closeness to origins, purity, directness, naturalness (the opposite of affectation and artificiality). Charles Burney, in his General History of Music (London, 1776), knits some of this together when he writes of the ‘simplicity’ of plainchant singing, which ‘precludes levity in the composition and licentiousness in the performance’ (p. 423). (That ‘licentiousness’ carried then the sense of sexual wantonness that it now does seems likely from the context we are about to enter. And then, after all, Don Giovanni is identified in the dramatis personae as ‘Giovane Cavaliere estremamente licenzioso’.) The vocabulary of these early histories is bathed in a feeling of nostalgia for a golden age undebased by corruption, but an age, we must remember, about whose music little or nothing was known. But that is just the point about a golden age. It has meaning only as a transcendent idea whose main attribute is the unspoilt quality that attracts nostalgia. To concretize it is already to spoil it. The penumbra of a golden age has always tended to surround stories about the origins of chant, and continues to do so. The eighteenth-century archaeologist and art historian Johann Joachim Winckelmann coined the slogan ‘noble simplicity and quiet grandeur’ for such a desired past. E. H. Gombrich has characterized the document in which it first appeared—Winckelmann’s Gedanken über die Nachahmung der griechischen Werke in der Malerei und Bildhauerkunst (Dresden, 1745)—as the ‘manifesto of classicism’.² Winckelmann provides the meaning of Rousseau’s word ‘effeminate’ ¹ All quotations from Rousseau are taken from the English translation by William Waring (London, 1779). ² The Ideas of Progress and their Impact on Art (New York, 1971), 12–32. All quotations from Winckelmann are taken from these pages of Gombrich’s book, in his translations.
214
Th e Po litics of Recep tion
in this context. It is virtually synonymous with ‘affected’, ‘artificial’, ‘corrupted’, ‘licentious’—the opposite of the qualities of ‘noble simplicity and quiet grandeur’, which are to be understood as manly qualities, as much so as are the sharp outlines and emphatic marking of detail with which Winckelmann characterized the work of Michelangelo, made, he believed, in imitation of Greek sculptures. (Here is his illustration of ‘noble simplicity and quiet grandeur’: ‘Behold the fleet-footed Red Indian running after a stag, how his humours quicken and his muscles and tendons become agile and supple . . . that is how Homer describes his heroes.’) In the effort to avoid harshness and to soften their work, wrote Winckelmann, artists sacrificed significance for charm. ‘We know this process from literature, and from music which also became effeminate.’ The theme of purity and the dread of corruption fills the greater part of Rousseau’s historical sketch about chant: ‘Ambrosius, archbishop of Milan, is said to have been the inventor of the plain-chant. . . . Pope Gregory perfected it, and gave it the form which it preserves at present in Rome and the other churches in which the Roman melody is practiced. The Gallican church allows the Gregorian chant but in part, with much ill-will and almost under duress.’ Then he relates a characteristic anecdote that came to be repeated with ritual regularity, and that we recognize from Chapter 6: The pious Charlemagne being returned to celebrate the festival of Easter with our apostolic Lord at Rome, there arose, during the feast, a quarrel between the Frankish and Italian choristers. The Franks pretended to sing better and more agreeably than the Italians. The latter declaring themselves superior in the ecclesiastic music . . . accused the Franks of corrupting and disfiguring the true melody . . . on account of the natural and barbarous roughness of their throats. This dispute being laid before his majesty, the pious monarch said to his choristers, ‘Tell us, which is the purest and most excellent water, that drawn from the fountain head, or that of the streams which flow at extensive distance from it.’ All consented in pronouncing that of the fountain head the most pure, and that of the rivulets so much inferior as the distance whence it flowed. ‘Have recourse, then’, replied Charlemagne, ‘to the fountain of St. Gregory, whose music you have undoubtedly corrupted.’
It is a linkage—at the very beginning of the story as our eighteenth-century sources know it—between aesthetic value and political purpose. Sir John Hawkins brings in another element in which the story is seen from the Frankish side (Rousseau’s story is based on Roman sources): The account given of this matter by another ancient writer, a monk of St. Gaul [the famous Notker] is that the pope sent to France, at the request of the emperor Charlemagne, twelve excellent singers whose instructions were to reform the music of the French churches and regulate the service so that there might be an uniformity in this respect throughout the kingdom; but that these men, jealous of the glory of France, in their way thither plotted to corrupt and diversify the plain-chant in such a manner as to
Th e Po litics of Recep tion
215
increase the confusion in which it was involved, and thereby render the people for ever incapable of performing it correctly.³
Rousseau’s neoclassicism did not thrive in the climate of national and racial sentiment that settled over the nineteenth century. His idea of chant as a Greek patrimony seems first to have been disputed directly in a general book of music history published in 1834 (the first of the genre ‘Histories of Western Music’) by the amateur music historian R. G. Kiesewetter, a Viennese nobleman, civil servant, and musical man-about-town (he turns up repeatedly between 1820 and 1826 in Beethoven’s conversation books).⁴ ‘It is a preconceived opinion’, he wrote, ‘as widely spread as it is deeply rooted, that modem music was modelled on that of ancient Greece, of which it is in fact merely a continuation’.⁵ His attack on this notion, however, is of no less an a priori, speculative, and ideological nature than is Rousseau’s assertion itself. We get a hint of what was at stake from the German title: Geschichte der europäisch-abendländischen oder unsrer heutigen Musik. I emphasize the last clause: our music of today. He wrote: Modern music flourished only in proportion as it began to separate and withdraw itself from the system laid down and enforced by the Greeks . . . It reached . . . perfection only when it succeeded in completely emancipating itself from the last remnant of the ancient Grecian. . . . That Grecian or Hebrew melodies should have found their way into the assemblies of Christians seems altogether incredible. . . . Their natural horror of everything connected with the heathens . . . was too great for the admission among them of such melodies as had been common to the pagan temples or theatres; whilst they evinced an equal anxiety also to separate themselves from the Jews. (pp. 1–4)
Kiesewetter’s disagreement with Rousseau over the question of the Greek origins of European music seems ironic in a way. His argument against it lays the ground for a nationalist and even racist narrative that was spun out later in the nineteenth century, and especially in the twentieth, but alongside a transformation in European ideas about the character and origins of ancient Greek culture that could have obviated those very arguments by assuaging the otherness that Kiesewetter could not square with his sense of ‘our music’. I shall have to return to all that. The old story about quarrels between the Romans and the Franks has been placed in the centre of modern chant scholarship by a discovery that seems to fly in the face of the belief that is shared by the stories on both sides. It is the belief that the chant tradition that settled into the Western church was disseminated from Rome. The discovery is that the oldest chant books of Roman provenance (from the eleventh and twelfth centuries) transmit melodies that are consistently ³ A General History of the Science and Practice of Music (London, 1776), 138. ⁴ Beethoven: Konversationshefte, ed. Karl-Heinz Köhler and Grita Herre with Günther Brosche (Berlin, 1968– ) i. 372; ii. 319; vi. 175; ix. 204, 216, 291. ⁵ History of the Modern Music of Western Europe, trans. Robert Muller (London, 1848), 1.
216
Th e Po litics of Recep tion
different in style from the Gregorian melodies that are their liturgical counterparts and that are transmitted in Frankish chant books of the tenth century and after. The Roman melodies have come generally to be called ‘Old Roman’, even though there is no agreement about the theory that they are descended from an old pre-Gregorian chant tradition of Roman origin. The oldest books with Gregorian chant known to have been in Rome were imported from France in the twelfth century. This situation is troublesome for the belief that the Gregorian tradition was founded in Rome and disseminated from there throughout Europe. The question remains unresolved, but in the course of the discussion the narrative has been enhanced with explicit and implicit categories of race, nationality, and gender that had become highly articulated during the nineteenth and early twentieth centuries in non-musical humanistic and sociological fields.⁶ We must now look at a pair of melodies from the two traditions in order to get some direct experience of what has touched off these stirrings, the introit Rorate caeli in the Old Roman and Gregorian traditions (see Ex. 9.1, CD tracks 3 and 2, respectively). They are settings of the same words and were sung on the same ritual occasion: the introit of the mass on the fourth Sunday of Advent. The differences of melodic style that are manifest here are typical of the two traditions. The Old Roman melody is more drawn out with melismas, more redundant, more ornate. The outlines of the Gregorian melody are sharper, more direct. Despite these differences the two melodies display a similarity of phrasing and of tonal outline that exemplifies a family resemblance which is common between corresponding chants of the two traditions, sufficiently to send the mind off in thoughts about one being a descendant of the other, or both being descendants of a common ancestor. These questions of genealogy have led to successively broader questions: about the origins of plainchant and its stylistic history, about the origins and nature of European music, and about the character and sources of European culture. One of the scholars most directly responsible for that is Bruno Stäblein (who coined the term ‘Old Roman’). His descriptions of Old Roman and Gregorian chant resonate strongly with the language that we encounter in the eighteenthcentury literature. That is no coincidence; it is, rather, an instance of the recycling of eighteenth-century classicizing values as modernist standards—critical neoclassicism, we might call it.⁷ The Old Roman melodies are described thus (Stäblein’s reference is exactly to the melodies of Ex. 9.1): ‘endless streams of melody that overflow the boundaries of textual divisions . . . like a chain of pearls or a voluptuous gown . . . soft, elegant, charming, and graceful, without sharp edges or corners’.⁸ ⁶ An overview of the source situation and of the theories that have been advanced to account for it is given by Helmut Hucke in ‘Gregorian and Old Roman Chant’, NG vii. 693–7. ⁷ Stäblein’s remarks are quoted from two sources: Die Gesänge des altrömischen Graduale Vat. lat. 5319, ed. Stäblein (MMMA 2; Kassel, 1970), and ‘Die Entstehung des gregorianischen Chorals’. ⁸ ‘Die Enstehung des gregorianischen Chorals’, 11.
217
Th e Po litics of Recep tion
Ex. 9.1. Introit antiphon Rorate caeli in the Old Roman and Gregorian traditions, transcribed from Vatican Vat. lat. 5319 and the Vatican Edition respectively Old Roman 8
8
Gregorian
Ro ra te cae Let the heavens precipitate from above
li
de
su
per
8
8
et
nu
bes
plu
ant ius
tum
and let the clouds rain down the Just One
8
8
a
pe ri
a
tur ter
ra
Let the earth open
8
8
et
ger
mi
net
sal
va
to
rem
and sprout the Saviour
To Rousseau and Winckelmann this would have been the description of an art in a decadent phase. Stäblein takes a patronizing view of the style as ‘naive, youthfully fresh, blossom-like, the expression of a general Italic, folk-like feeling’.⁹ The Gregorian melodies, by contrast, are ‘disciplined and ordered, a product of thorough rational thinking’.¹⁰ They are ‘clear, sculpted configurations, systematically chiselled’.¹¹ They are of a ‘more perfect quality—perfectior scientia’, he writes, quoting the ninth-century writer Walahfrid Strabo.¹² Other descriptive terms that he plucks from medieval sources are vis (strength), virtus (manliness), vigor (vigour), potestas (power), ratio (reason).¹³ ⁹ Die Gesänge des altrömischen Graduale, 38. ¹⁰ Ibid. 38. ¹¹ Ibid. ¹² ‘Die Entstehung des gregorianischen Chorals’, 17. ¹³ Ibid. 14.
218
Th e Po litics of Recep tion
Stäblein regards the Old Roman tradition as the older one, from which the Gregorian tradition was achieved by ‘stripping down and reshaping the unrestrained coloratura of the originals’.¹⁴ This happened in Rome, whence the new melodies were carried to the major ecclesiastical and political centres of the Frankish North during Carolingian times. Through this transformation the provincial Roman melodies were raised to a higher, super-regional level as ‘melodies for a world power’.¹⁵ ‘Rome’ thus has two cultural faces: as the provincial Italic home of a prolix and undisciplined Mediterranean singing practice, and as the place where the efficient and economical model was fashioned for a European melodic style—the old centre of the Roman Empire and the inheritor of Greek culture. The belief in the classical origins of chant as a matter of factual reality has not survived for lack of evidence, but the feeling of it still pervades the literature. The most enigmatic of Rousseau’s dualities is that of the Gregorian and Gallican. This is a recurring topic in the modern literature, but there is little common agreement or understanding about what ‘Gallican chant’ is or was. According to the usually judicious New Harvard Dictionary of Music it was ‘the Latin chant of the churches in Gaul before the importation of Gregorian chant under Pepin and Charlemagne’, a process that Rousseau thought was not very successful. The Harvard Dictionary article concludes that, on the whole, ‘it is difficult to distinguish true Gallican survivals from local French compositions originating after the suppression of Gallican chant’ (p. 332). This did not, however, restrain Stäblein from giving quite a detailed characterization of that tradition in an article on ‘Gallikanische Liturgie’ that was published even before some of the most likely identifications of such chants had been made: In contrast to the classical, unified, concise, sober style of the Roman liturgy, the Gallican, in its most developed blossoming in the eighth century, displays trends that are rather to be called Romantic. The bright, colour-happy richness, a proclivity to fantastic prolixity characterize it, while with the Roman the leading principles are disciplined formulation and restraint. This character of the Gallican liturgy brings it close in some respects to the oriental character.¹⁶
The virtual identity of the characterization of this duality with that of the Old Roman–Gregorian is unmistakable, right up to the last two words, ‘oriental character’, which I shall take up in a moment. As the implied but counter-intuitive claim of an identity of style between Gaul and Italic Rome is not based on the comparison of actual melodies or liturgies, we have to suspect that we are in the presence of an archetype for the Other, functioning to define the Gregorian by contrast with the oriental as basis for ‘our modern Western music’, as Kiesewetter had put it. ¹⁴ ‘Die Entstehung des gregorianischen Chorals’
¹⁵ Ibid.
¹⁶ MGG iv. 1299–1325.
Th e Po litics of Recep tion
219
Next I return to the paradigmatic books for modern chant studies that were discussed in Chapter 7, those of Wagner, Gregorianische Formenlehre, and Ferretti, Estetica gregoriana. Wagner characterized the Gregorian melodies as ‘models of clear formal structure and symmetrical organization, the work of aesthetic deliberation’ (p. 398). Ferretti called them ‘organic’, ‘harmonious’, ‘homogeneous’, and ‘logical’ (pp. vii–viii). Wagner’s language was aimed at a contrast with an older layer of chants whose melodies were ‘unregulated and without plan’. The difference is like that between a ‘skilfully laid-out flower bed’ and a ‘luxuriantly proliferating growth’ (p. 403 ). Those older melodies remind him of the ‘unregulated undulation of the melismatics of the Orient’, whereas the later ones display ‘Latin, Roman traits’. Perhaps that passage was lingering in Stäblein’s mind as he described the Gallican tradition. The most extravagant-sounding encomium in the whole story, again, is the passage in Willi Apel’s 1958 book surveying Gregorian chant. Concluding an analysis of a group of gradual chants, Apel wrote that ‘the perception of their structural properties greatly enhances their significance as unified works of art, no less so than in the case of a sonata by Beethoven’.¹⁷ The task is to validate Gregorian chants as the source of European music by displaying in them just those qualities that count as European on the authority of Beethoven, who epitomizes them as no one else (see Pl. XI). This power of Gregorian chant as the earliest embodiment of Western musical principles under Greek inspiration could then be reflected back in legitimation of later music. Anton Webern, in his apologia Der Weg zur neuen Musik, described the form of a Gregorian alleluia melody and exclaimed that it has ‘already the full structure of the large symphonic forms, expressed exactly as in the symphonies of Beethoven’.¹⁸ This time it is the ‘new music’ that is legitimated by showing it to be in direct line from that ancient music. The symmetry here catches the game of mirrors that I mean to suggest with my title. If there is a single word that can express what is for the modern period the essential attribute of ‘Western music’ throughout its assigned history, that word is ‘form’, preceded by all its qualifiers (rational, logical, unified, concise, symmetrical, organic, etc.). Nowhere is that plainer than in the titles of the two paradigmatic books on chant already cited: Gregorianische Formenlehre and Estetica gregoriana.¹⁹ Wagner’s task, he wrote, was ‘penetration to the essence of Gregorian art through the study of form’ (p. 3). Ferretti declared that ‘form constitutes the essence of the ¹⁷ Gregorian Chant, 362. ¹⁸ Der Weg zur neuen Musik (Vienna, 1960), 23. ¹⁹ That a treatise on musical forms would constitute the extension of the title Estetica gregoriana is as significant as the fact that a theory of form, as in Wagner’s title, would constitute a science of style. Regarding the influence of these aesthetic issues on the analysis of chant, see Ch. 7.
220
Th e Po litics of Recep tion
work itself, and is the primary source of all the properties and qualities of beauty’ (p. vii). In the background is Benedetto Croce’s summation of much nineteenthcentury thought in his Estetica of 1905: ‘The aesthetic fact is form and nothing but form.’²⁰ One generation further in the background is Eduard Hanslick’s famous treatise on musical aesthetics, Vom Musikalisch-Schönen of 1854, where the language is very similar to Wagner’s. In aesthetic analysis, he wrote, ‘we . . . penetrate to the inner nature of the works and try, from the principles of their own structure, to account for the unique efficacy of the impressions we receive from them’.²¹ But the appeal to form is equivocal. Croce’s conception, like Hanslick’s before him, is a conception of inner form that has the sense of ‘musical idea worked out in notes’. Far different, on the other hand, is the meaning we must read in the appropriation of Croce’s slogan by generations of chant scholars. Here form becomes the more pedestrian conception of the enumeration, arrangement, and coherence of parts on such principles as repetition, variation, development, contrast, return, etc. (A clue is the plural in Ferretti’s title: Treatise on the Musical Forms of Gregorian Chant.) This conception embodies a value-gradient according to the principles of closure, symmetry, unity, and the idea that every note is necessary to the whole and no note is superfluous to it. This is the sense of form on which the identification of the essential traits of Western music has rested, the sense that is claimed to be exemplified first in Gregorian chant. It is form as order—the historical state of equilibrium between the disorder that precedes (Peter Wagner’s ‘luxuriantly proliferating growth’, reminding him of the ‘unregulated undulation of the melismatics of the Orient’) and the decadence that follows (Rousseau’s and Winckelmann’s conception of the ‘effeminate’). Form is the hallmark of the work as monument. (‘The composer works slowly and intermittently . . . forming the musical artwork . . . for posterity’.)²² Form is what is lacking in the constructions of the musical Other that are identified in the story I have been tracing. It is no coincidence that those constructions are products of the same early nineteenth-century ideological milieu in which this sense of form came to be explicitly articulated.²³ But why would it have come to mind to characterize chant styles that deviate from these principles as feminine or oriental? The answer must be found in the ²⁰ Aesthetic as Science of Expression and General Linguistic, trans. Douglas Ainslie (New York, 1924), 16. ²¹ On the Musically Beautiful, 6. ²² Ibid. 49. ²³ I was myself once persuaded of the attractiveness of this way of thinking, and have regrettably been a contributor to the mythology based on it. The very title of Ch. 5 conveys the a priori idea that closure and unity of melodic structure and coherence of melodic syntax are essentially Western, as opposed to oriental, features. And the interpretations of melody within the essay I now regard as too restrictive about what constitutes unity and coherence, even within the ‘Western’ tradition. I would rather have my current understanding of that question represented by the analysis of the Old Roman offertory in Ch. 1 and the new analysis of the Gregorian alleluia in Ch. 5. I offer those analyses, too, as a counter-example to the characterizations of Old Roman and Gregorian chant that I have cited in the preceding. On its own terms the Old Roman melody shown there is as coherent and unified as any Gregorian melody. It seems we are for ever apt to allow ideology to command analytical methods which then, of course, produce the accounts we desire.
Th e Po litics of Recep tion
221
history of European identity formation. With respect to the category of the ‘oriental’, the history has been traced by Edward Said, in his book Orientalism (New York, 1978). ‘Orientalism’, writes Said in his introduction, is a way of coming to terms with the Orient [the Near East] that is based on the Orient’s special place in European Western experience. The Orient is not only adjacent to Europe; it is also the place of Europe’s greatest and richest and oldest colonies, the source of its civilizations and languages, its cultural contestant, and one of its deepest and most recurring images of the Other. . . . The Orient has helped to define Europe (or the West) as its contrasting image, idea, personality, experience. [It is] a surrogate, underground self for the West. (pp. 1–2)
I wish to draw out the implications in this last sentence of a psychological aspect of the process that is reflected in the history I have been describing. Through it our culture has refined its sense of identity by viewing itself in the context of a succession of oppositions with other cultural entities that are contrasted in essential ways with ours. But what we regard as the opposite traits of the Other show through as the traits of a surrogate, underground—we might as well say unconscious—Self. The Other is, in effect, a projection of the Self, or rather of an unacknowledged aspect of the Self that is suppressed as unacceptable to that identity that is the speaker for the Self. Because the duality of Self and Other is created as a context for the definition of Self, that role must determine the attributes we ascribe to the Other, as much so as the attributes that the Other has ‘in itself ’. For that reason, and because whatever is declared ‘Other’ must submit to being made other, it is already born into a relationship of power—a relationship without reciprocity—as the submissive and dominated partner. By its very nature, the dualistic epistemology establishes a political relationship between the knower and the known, not only the known that is cast off as Other, but also the known that is claimed for the Self. This comes at a high price for the Self in the commerce of knowing—a price, unfortunately, that the Self cannot easily acknowledge. For by virtue of this distancing, the Self denies itself understanding of the Other. We think: ‘That is something different; it has nothing to do with us’. An ironic consequence of such dualistic strategies is the disqualification of the scholar from judgement about what is cast off as ‘other’. It guarantees that the scholarly discourse will be a monologue, or at best the dialogue of a ventriloquist with his dummy. To put it bluntly, the scholar is set up from the start to make false judgements, as we can sometimes tell when we make an effort to hear the Other in its own voice. I shall cite a case later on. The situation is nuanced by an erotic tension. The Other must be held within the range of a voyeuristic gaze, one whose effect would be destroyed by too much familiarity. This is evident in the descriptive language we have read about Old Roman chant (‘voluptuous gown’, ‘chain of pearls’, ‘soft’).
222
Th e Po litics of Recep tion
The way the dualities of the masculine and the feminine and of the European and the oriental are treated as interchangeable expressions of the same categories of critical interpretation and historical narrative is an insistent clue to a fact of outlook and power. Europe and the Orient are assigned masculine and feminine character respectively, and in both cases it is a relationship of the dominant and the dominated. That is a political reality that was rationalized in theory in the late eighteenth century in essentialist theories of gender character that were developed and expressed both as science—Herbert Spencer, for example, asserted in his 1867 book Principles of Biology that ‘deficiency in reproductive power . . . can be reasonably attributed to the overtaxing of [women’s] brains’ (pp. 512–13)—and in historical explanation: Hanslick, for example, explained ‘why women . . . have not amounted to much as composers’: ‘The cause of this lies . . . in the plastic aspect of musical composing, which demands renunciation of subjectivity . . . while women are by nature preeminently dependent upon feeling . . . it is not feeling which composes music, but the specifically musical, artistically trained talent.²⁴ At the same time there was a parallel development of essentialist theories of racial character, described by George Stocking in his book Race, Culture, and Evolution (1968). What I have just described as a political relationship between the knower and the known reflects real-world politics. It could hardly be otherwise. This linkage between the European–oriental and the male–female dualities is tightened by the connection of both to yet another deep-lying dimension of Western thought. Said cites a book entitled Modern Egypt (New York, 1908) by a certain Lord Cromer: ‘The European is a close reasoner, a natural logician. . . . The Oriental is irrational . . . [his] mind . . . like his picturesque streets, is eminently wanting in symmetry. His reasoning is slipshod. He is singularly deficient in the logical faculty.’²⁵ Plate XII is a political cartoon in which the artist has visualized this stereotyped idea about the contrast between the Western and the oriental minds as an icon of the failure of diplomacy in the Persian Gulf. But it is an idea with a long and distinguished history. In 1801, for example, Friedrich Hölderlin wrote that Homer had been ‘soulful enough to appropriate earthy-Western rationality (abendländische junonische Nüchternheit) for his Apollonian domain’.²⁶ The idea of rationality as a uniquely Western trait depends on a view of a divided human consciousness and cognition that is in itself a part of the Western heritage that we like to follow back to Plato: it is the idea of an opposition between reason on one side and passion and desire led by the senses on the other. This ²⁴ On the Musically Beautiful, 46. ²⁵ See Said, Orientalism, 46. ²⁶ Letter to Casimir Bölendorff, 1801, in Friedrich Hölderlin: Sämtliche Werke und Briefe (3rd edn., Munich, 1981), ii. 926. I am grateful to Patrizia Hucke of the Freie Universität, Berlin, for calling my attention to this letter.
Th e Po litics of Recep tion
223
opposition has always been projected onto the dimension of gender, but the thought network that allowed it to cross naturally to the constructions of nationalities and race was a creation of the nineteenth century. That, it seems, required a clarification which was, in a sense, the other side of the invention of the Orient: the revelation that the culture of Greek antiquity which served as source and model for European culture was a counter-culture to the Orient. Martin Bernal has described this in his book Black Athena.²⁷ He writes of two models of Greek history: one viewing Greece as essentially European or Aryan, the other seeing it as Levantine, on the periphery of the Egyptian and Semitic cultural area. I call them the ‘Aryan’ and the ‘Ancient’ models. . . . According to the Ancient Model, Greek culture had arisen as the result of colonization, around 1500 BC, by Egyptians and Phoenicians who had civilized the native inhabitants. (p. 1)
But under the paradigm of ‘races’ that was applied to all human studies at the end of the eighteenth century: ‘it became increasingly intolerable for Greece, which was seen not only as the epitome of Europe but also its pure childhood, to have been the result of the mixture of native Europeans and colonizing Africans and Semites’ (p. 29). According to ‘the Aryan Model, which most of us have been taught to believe, [and which] developed only during the first half of the nineteenth century, there had been an invasion from the north which had overwhelmed the local “Aegean” or “Pre-Hellenic” culture. Greek civilization is seen as the result of the mixture of the Indo-European-speaking Hellenes and their indigenous subjects’ (p. 30). This is the transformation that came too late to head off Kiesewetter’s vigorous rejection of the idea of Greek patrilineage for Western music. It accomplished what might be expressed by reversing my sub-title: ‘The Tailoring of the Past as Progenitor of a Preferred Idea of the Present’. There are clear signs of its adoption as a general notion into music history, but that did not prompt a restoration of the theory of Greek origins of chant. This may be because there was no good evidence for it. But, more to the point, the question was rather superseded by yet another duality that was assigned a higher priority for obvious ideological reasons as a narrative dimension: the Germanic/Nordic versus the Semitic/oriental. We can read about it in a book published nearly 100 years after Kiesewetter’s, Richard Eichenauer’s 1932 book Musik und Rasse. This is a kind of ‘History of Western Music’, too, but one in which race provides the main taxonomic and explanatory categories. It belongs in the company of books published around the same time in Germany with such titles as Kunst und Rasse, Rasse und Stil, Rasse und Seele, Rassenkunde Europas, Die Rasse in den Geisteswissenschaften, Rassengeschichte des hellenischen und des römischen Volkes, Rassenkunde des jüdischen ²⁷ Black Athena: The Afroasiatic Roots of Classical Civilization, i: The Fabrication of Ancient Greece 1785–1985 (New Brunswick, NJ, 1987).
224
Th e Po litics of Recep tion
Volkes.²⁸ The research field was called ‘Rassenforschung’, heritage of a Romantic preoccupation with its roots in the late eighteenth century. Eichenauer’s idea about Greek culture was a version of the ‘Aryan theory’, which he understood to be the product of recent scholarship: ‘Race research has demonstrated that Greek behavioural patterns as a whole show the picture of a great ascent under Nordic influence followed by a decline brought about by Entnordung [which he also calls Semitisierung]’ (p. 37). Greek musical style is Apollonian, which in turn is identified as Nordic. The history of Gregorian chant, on the other hand, begins in the Dionysian (i.e. oriental/Semitic) domain, with the Jewish chant of the Near East. But with the spread of Christianity the chant ‘wandered into the heartland of the Nordic race’ where ‘the Germanic musical feeling expressed itself ever more strongly’ (p. 67). The outcome, then, would have been a chant tradition with Nordic character, like the Greek, but with no claim that one had descended from the other. Eichenauer did not fail to pick up Peter Wagner’s oriental theme: ‘We recall that one of the strongest orientally flavoured characteristics of ecclesiastical chant were the long melismas. Wagner thus distinguishes an older, still purely oriental group from a younger group with Roman traits. He finds the difference in the turn from unregulated . . . meandering [of the former] to the clear structure [of the latter]’ (p. 89). Then Eichenauer confronts the difficult problem about what is meant by ‘Roman’. He quotes Peter Wagner again: ‘Systematic and design-wise melodic process was ever the spiritual task of a strong side of the Roman genius’; and he quotes Heinrich Besseler, author of the influential handbook Musik des Mittelalters und der Renaissance: ‘In [the Gregorian chants] rational control and the establishment of musical form are felt most strongly as against the voluptuously proliferating oriental melismatic style.’ In all this Eichenauer sees a ‘racial influence. [In the Gregorian melodies] . . . a “new-Roman”, that is Germanic, genius is speaking. . . . [It is] the Nordic-strict constructive spirit that strives for the clear working-through of form’ (p. 89). Something like this dialectical epic has an important role in the musical historical ideology of the more famous Wagner, Richard. It is the conflict between the Germanic folk myth and Christianity that is described in Oper und Drama. Christianity uprooted the mythos, robbing it of its virility and rendering it incapable of procreation. Christianity in effect emasculated the German myth.²⁹ A sexual threat to the idea of the masculine underlies every form of the dread of corruption. This is hinted at in the theme of the effeminate in the eighteenth century and is explicit in Wagner’s essay. ²⁸ Bibliographical details for these books can be found in Eichenauer, Musik und Rasse. ²⁹ Richard Wagner’s Prose Works, ed. and trans. William Ashton Ellis (London, 1893–9; repr. New York, 1966), ii. 39–41.
Th e Po litics of Recep tion
225
Wagner’s mythic image of conflict was in turn given a specific music-historical content in the story of a struggle between the melodic modes of Christian chant and the major–minor tonalities, which were assigned Germanic character. This version of the story was first presented by Hans Joachim Moser in a 1914 article, ‘Die Entstehung des Dur-Gedankens: Ein kulturgeschichtliches Problem’.³⁰ According to Moser the major–minor scales and the church modes represent a priori such radically different ideas that a development of one principle out of the other is ‘ontologically unthinkable’. As such, he wrote, they reflect the opposition between the Germanic and the Mediterranean peoples. As his authority on this subject he cites The Foundations of the Nineteenth Century, a book by Richard Wagner’s son-in-law, the notorious ideologue Houston Stewart Chamberlain, who came to occupy a prominent place in the National Socialist Pantheon.³¹ This book was unsavoury enough for even Moser to disclaim identification with its conclusions for ‘political, religious, and artistic life’, while still accepting its characterization of the racial make-up of the greater European land mass. He wrote: ‘It will be possible to show that this battle between music that is like the church modes and folk-like harmonic music was nothing other than the battle between internationalism and nationalism, between communism and individualism, between “Völkerchaos” and “Germantum”.’ These last two terms require decoding in terms of Chamberlain’s historical ideology: the Mediterranean peoples were always divergent in character—hence ‘Völkerchaos’—while the Germanic peoples were united by a common character, ‘Germantum’. This high valuation of unity as a Germanic characteristic—which has been discussed by Erik Erikson in Childhood and Society (1963)—transferred easily to the canon of values for art, and has done so since the Carolingian era, as we have seen. The smoke of battle has cleared, but we can still read in the most up-to-date and apparently non-ideological surveys of Western music history how under northern influence Gregorian melodies took on a more triadic character, that is, they became more like melodies in the major–minor tonalities.³² As a reminder that this theme of the Germanic–Mediterranean polarity transcends particular musical genres or traditions, we may recall Hanslick’s claim that ‘the tyranny of the upper vocal part among the Italians has one main cause in the mental indolence of those people for whom the sustained penetration with which the northerner likes to follow an ingenious work of harmonic contrapuntal activity [is impossible]’.³³ ³⁰ Sammelbände der Internationalen Musikgesellschaft, 15 (1913–14), 270–95. ³¹ Die Grundlagen des 19. Jahrhunderts (Munich, 1900). ³² ‘A Northern influence on plainchant was that the melodic line became modified through the introduction of more skips, especially by the interval of a third. . . . The tendency of Northern melody was toward organization by third.’ Donald Jay Grout and Claude Palisca, A History of Western Music (4th edn., New York, 1988), 68. ³³ On the Musically Beautiful, 64.
226
Th e Po litics of Recep tion
I may say that the claim about the increasing triadic character of chant under northern influence does not stand up at all well under the comparative study of actual melodies,³⁴ and that is typical of the sort of dissonance that is set off when we can manage to hear medieval voices against the voices that have been imposed on medieval material by imperious historians. The most striking case has already been reported in the Introduction to Chapter 5, bearing on the central role of form in the story I have been telling. It is Fritz Reckow’s account of early medieval writing on what we might call musical form. What he found were descriptions of music through a rich vocabulary that portrays the most finely nuanced and differentiated aspects and modes of movement, but nothing at all in which to recognize a conception of the closed, unified structure of autonomous works.³⁵ A parallel situation has arisen in the recent scholarly literature on jazz, where there is a tendency to fold the music into the Western canon on the grounds of its formal properties. An exemplary case is built in Gunther Schuller’s book Early Jazz: Its Roots and Musical Development (1968). Relevant passages are easily found by looking up ‘form’ and ‘Beethoven’ in the index. Beginning with one, we are led directly to the other: ‘[Duke] Ellington is a strict classicist [Schuller’s emphasis], perhaps only surpassed by Jelly Roll Morton. Ellington’s forms are more concise and symmetrical than those of any number of 19th century Romantic composers. In fact, when compared to the great formal achievements of a Beethoven . . . Ellington’s form seems almost hackneyed and naive in its restraint’ (p. 348). As early as 1935 Percy Grainger, while teaching at New York University, brought Ellington there, and in his introduction called jazz the most classical of popular music and compared it to the music of Beethoven and Brahms.³⁶ In the discussion of jazz’s ancestry, Schuller writes of a ‘misconception that African musical forms are “primitive”. . . . African music—whether in large formal designs or small structural units within these designs—is replete with highly “civilized” concepts’ (p. 27). The language of approbation is strikingly reminiscent of that for chant. A Louis Armstrong solo is characterized as ‘sober and balanced’, a model of ‘structural logic’ (Schuller, p. 103). Language for describing the excluded Other can also be virtually identical to its counterparts regarding chant. In what could be a parody of François-Auguste Gevaert’s talk about exotic musical ‘curiosities’, Dizzy Gillespie quotes an outburst from Cab Calloway about Bebop: ‘Man, listen; will you please don’t be playing all that Chinese music up there?’³⁷ There is plenty of talk about authentic and corrupt jazz traditions; Lincoln Center stages ³⁴ See Charlotte D. Roederer, ‘Can We Identify an Aquitanian Chant Style?’ ³⁵ ‘Prozessus und Structura’. ³⁶ I am grateful to Prof. Krin Gabbard of the State University of New York at Stony Brook for this information, and for other helpful suggestions and information regarding this parallel. ³⁷ Dizzy Gillespie with Al Frazer, To Be or Not . . . to Bop (New York, 1948), 111.
Th e Po litics of Recep tion
227
a ‘Jazz Classics’ series; and now there is talk of a neoclassical restoration, an idea that one critic has called ‘wishful thinking, . . . a biased and unsubstantiated claim that seeks to impose a historicist vision of jazz onto a conflictive and highly unstable . . . movement’.³⁸ In a 1970 review of the Schuller book Lawrence Gushee commented on the ‘constant tendency for a given performance to be treated as a work of art rather than a sample of a process’.³⁹ In modern jazz studies it seems that the more the emphasis is on the description of the form of autonomous works, the less attention is given to the process of making and transmission and to the contexts in which these occur. This is a clue to the interpretation of the preoccupation with form as the outcome of an ideological programme, an outcome that has the effect of falsifying itself. For in the unblinking focus on the interior of autonomous works the investigation of form has always been seized as an alternative to, and therefore an evasion of, the investigation of the world of ethics and politics in which music and our paradigms for the aesthetic have been fashioned. It produces as often as we require it the result that form is what our music has and other music has not, as a security against the fear of disorder and decadence that has characterized the attitude towards the Other. But by its very persistence this attitude gives witness to the politics that maintains it but does not want to be seen. A brilliant film by Maurizio Nichetti, Ladri di saponette (released in the USA under the title Icicle Thief ), is about a film within a film. The inner film, a blackand-white piece of cinema verità about an impoverished young couple and their two small children, with clear reference to De Sica’s Bicycle Thief, is a television production being watched by a young middle-class couch-potato family of the same size in their sterile technicolour apartment. Things begin to go haywire during a commercial, when a slim, near-naked blond English super-mod model jumps into a swimming pool and winds up instead drowning in a wooded pond in the gray world of the inner film, just as the man of the family is bicycling home from his first day of work after a long period of unemployment. He rescues her, wipes the technicolour off her, and takes her home. That derails the film from its already completed course. In a desperate effort to put it back on track the director enters the world of the inner film. ‘Stop improvising!’ he screams at the characters. Bruno, the 6-year-old elder son of the young couple, has become pals with the model, but the director sets him straight about what is to happen. The father’s bicycle is to be struck by a lorry that day on his way home, leaving him incapacitated and again out of work. The mother’s last resort is to be prostitution, and the children are to go to an orphanage. ‘Nothing doing’, says Bruno. ‘You put your ³⁸ Ronald M. Radano, ‘Jazz Neoclassicism and the Mask of Consensus’, in Abstracts of Papers Read at the Joint Meetings of the American Musicological Society, Society for Ethnomusicology, Society for Music Theory, November 7 through November 11, 1990, Oakland, Calif., ed. Bruno Nettl (Urbana, 1990), 47–8. ³⁹ ‘Musicology Rhymes with Ideology’, Arts in Society, 7/2 (1970), 230–6.
228
Th e Po litics of Recep tion
children in an orphanage.’ In the end the whole family leaves, and the director is left trapped in the television set, his face and hands pressed against the screen. He screams to be let out, but the housewife of the technicolour family switches off the set, choking off his voice with it. This offers itself as an allegory for such histories as I have reviewed here, prompting these questions: from whose perspective and in whose interest is the story told? How can we reconcile the needs of historian and audience for myth with the needs of the characters in the story to act and speak out of their own situations and in their own voices? If they escape the confinement and manipulations of their story will we not cheer their freedom just as we cheer Bruno’s? But then there would be no story, only a script. A story like this could drive one into the extreme solipsism of deconstruction. This is one of the dilemmas of historians, which I have neither discovered nor resolved. In plain language it is a dilemma created by contradictions that do arise between a standard of objectivity to which we wish to attain as a methodological and moral responsibility, and the social demands that shape histories. It is as deep and inescapable as another dilemma with which it is inextricably linked, between the need to achieve identity through separation and the distortion and destruction that can result when separation becomes an end in itself. Just as those dilemmas are linked, so are the possibilities of finding ways around them that are not so ridden with contradiction. We cannot really attain the measure of objectivity that we have always craved without gambling on the apparently paradoxical: on an acknowledgement of difference that does not lead to opposition and exclusion; on a sense of identity that does not depend absolutely on the principles of unity and consistency; on a range of responses to the Other that is expanded beyond the stark alternatives of appropriation and alienation; on dialogue with our objects, in which we speak for ourselves and allow them to speak for themselves. In the domain of historical scholarship one often hears such things said nowadays, for example by the medievalist Johann Drumbl, declaring it a responsibility of the interpreting historian ‘to regard what is strange in a text in such a way that the strangeness is retained and not sucked up by the categories of the modern observer’.⁴⁰ We could take this as a direct response to the passage from François-Auguste Gevaert with which we began, about the music of ‘exotic people . . . at bottom strange to our manner of feeling’. Well and good, but it will not be easy to heed this appeal if it is taken simply as a call for the reform of a scholarly bad habit—laziness or lack of imagination. If we are to develop the human capability to transcend the ‘us and them’ syndrome, that will require recognition of the strong need for personal, sexual, cultural, national, racial identity—or, to put it more sharply, recognition of the strong need ⁴⁰ ‘Fremde Texte’, Materiali universitari, lettere, 49 (Milan, 1984), 87–99 at 91.
Th e Po litics of Recep tion
229
to allay anxieties about such matters of identity—that has been served by it. Motivation to do so comes not only from our desires with respect to the interpretation of texts, it ought to grow out of the certain knowledge that such a dualistic construction of texts, but more urgently of the world surrounding our texts, is no longer tenable.
230
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
chapter 10
Oral, Written, and Literate Process in the Music of the Middle Ages
} What is the match of the concepts ‘oral and written transmission’—they are, after all, modern concepts—to what can be garnered about medieval practices? And what really lies under the cover of those labels? This chapter’s title signals my belief that if we are to understand the place of orality in the history of medieval song— oral composition, memorizing and remembering, singing out or writing down without written models—we must try at the same time to understand the place in the same history of music writing itself: the initial writing down, copying, reading, the use of written scores in the preparation for and act of performing, and the significance of written scores in the historical process of creating a musical tradition in a particular culture and under particular social and political circumstances. Now in saying this I may appear to be setting up a framework in which oral tradition on one side is sharply contrasted with written tradition on the other. This would be a great oversimplification, and it is one of my purposes in this chapter to probe both the differentiations and the continuities that we must recognize within the relationship between the oral and written aspects of the making and transmission of medieval song. The chapter’s title also signals a differentiation that I believe is needed with regard to the use of musical notation: possession of a technology of music writing is one thing; its use in laying down a textual tradition—as in the systematic copying of books with notation—is something beyond that; and the establishment of a prevailingly literate culture that depends on writing and reading for its functioning (e.g. the culture of Western art music in the present) is something yet beyond that. This calls for caution about regarding signs of the first as symptoms of the others— for example assuming that neumes in the St. Gallen trope manuscripts 381 and 484, because they are so richly provided with signs for nuances of duration and pitch, played a prescriptive role in the performances of a literate liturgical community (see Ch. 2). In this chapter I posit again the continued functioning of oral processes in performance well beyond the deposition of music in books, and I identify reasons not
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
231
to assume that singers valued them exclusively for their prescriptive role in performance. We can consider a number of different, but not mutually exclusive, ways of thinking about how the oral tradition of chant might have come to be deposited in writing: 1. In the course of its normal oral practice the chant community accumulated a corpus of chants—whether through an evolutionary process or in periods of concentrated compositional or editorial activity—that was committed to and performed from memory, and during the Carolingian spread of writing in all domains this corpus was represented as a whole in musical notation. The closeness with which written images of corresponding chants in the earliest sources resemble one another, from this point of view, reflects the fixed state into which the corpus had settled in the oral tradition. The signs of such a process are reviewed in the introduction to Chapter 6. 2. Early notators who were called upon to provide chant books for their ecclesiastical institutions would have sought models from which to copy, and they would have imitated their graphic properties closely, whether or not these corresponded as closely to the performance practice of the institutions to which those notators belonged. In this way a written tradition (not necessarily a literate one) is established. From this point of view the theme of the fixity of the corpus in the oral tradition is not compelling. The uniform appearance of the corpus in writing is a phenomenon of the writing down—and by implication the assumption that the written texts of an establishment accurately represent its oral performance practice falls into question. 3. The first notators (and successive generations of notators as well) performed like singers in the oral tradition, calling forth chants by following out the rules, forms, and melodic formulae of the several chant types or families, but writing them down rather than singing them out. From this viewpoint the fixity would be at the level of such sets of rules, forms, and formulae—what may be called ‘generative systems’ or ‘performance models’—rather than at the level of the individual chant. That would account for the resemblances throughout such families in the early written sources, as well as those between versions of individual chants in different sources Such general theories about great historical movements tend to be floated high above the din and bustle of the daily events that drive history. In this case it is hard to evaluate the theories for their power to explain those events, for there is so little evidence on the ground that can serve as basis for discriminating among them. The best we can do is to try to stage scenarios in our minds that concretize each of these theories, asking ourselves ‘Just exactly how would this have worked in detail?’ The very possibility of spelling out an answer lends value to a theory, whether or not there is evidence for the verisimilitude of the scenario. Beyond that we can only judge whether the theory and the scenario it suggests ring true—intuitively, on the basis of our experience, on our general evaluation of the kind of theory that it is.
232
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
In James McKinnon’s richly documented scenario in The Advent Project, for example, the Roman schola cantorum—well-supported, highly trained professional cantors habituated to collaboration—working without notation, composed substantial portions of the mass ordinary in coordinated cycles, which would have been committed to memory and dispersed and transmitted that way before they were written down in books. The fact that the words of the chants to which he refers have been demonstrably revised from their scriptural originals in order to suit them for chanting adds to the plausibility of the hypothesis of a deliberate compositional activity. The analyses of individual chants that Helmut Hucke presented in ‘Toward a New Historical View of Gregorian Chant’, for another example, are framed as re-enactments of acts of extempore composition by the ‘notator’, in which compositional decisions are interpreted in terms of the choices confronting him at every turn by the conventions of liturgical genre, mode, model, and the liturgical text. There is no fundamental difference for the product, from this viewpoint, between recording such acts of composition in writing and giving them out viva voce. The analysis of the Old Roman offertory presented in Chapter 1 is of a similar sort. The interpretation offered is that the performance model or generative system that I describe would have constituted a sufficient basis for the iteration and more or less identical reiteration of this chant or for the composition of other chants of the same genre in the same mode articulating different texts, either in singing or in writing. The analysis is not offered as proof that the offertory is the product of an oral tradition but as the portrayal of a compositional process that could function independently of writing, and that probably in the oral tradition. There is no warrant for suspicion that it might have been invented with the initial writing down. Just as plausible is the hypothesis that the close resemblance in so many instances of the versions of the same chant appearing in different written sources is—perhaps only in part— a phenomenon of the writing itself, resulting from copying in the early stages of the writing down, or from an attitude that the score is an emblem or token of the chant, without necessarily being a prescriptive instruction for its performance. I say ‘only in part’ because the many studies that have been made comparing Frankish Gregorian chants with their Old Roman counterparts—for example in Chapter 1—show deep structural similarities that assure us that aspects of what we see in writing belonged to common originals before the separation of the two traditions before the age of writing. To put it in metaphorical terms, there has been an inclination to regard the early sources as photo albums of snapshots of performances, but there is no more reason to assume that than that they are collections of sketches that have come to resemble one another through a process of working out. Equally plausible, again, is a process combining the latter two, the notator generating the chant in his mind, with a model in front of him as he writes. On the other hand it is hard to think that the theory that is most often voiced— the corpus of chants settled gradually into a stable state through normal practice— takes us very far. It is hard to know what it means; it suggests no scenario beyond
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
233
the image of repeated chanting of the item leading gradually to its firming up—like concrete, or a pudding. For the mass proper chants ‘repetition’ would have occurred at one-year intervals. That seems hardly the circumstance for gradual settling. In any case the idea of gradual change itself is modelled on a misconstrued notion of evolutionary change, as gradual and uniform—the neck of the species ‘giraffe’ as a whole gradually lengthening as individuals reach ever higher in the tree for choice leaves. Imagine offering such an explanation if giraffes were to feed only once a year. Natural though it might seem, this idea of the gradual stabilization of the chant in the oral tradition is vague and, when one presses it, lacks explanatory force. The very habit of speaking of the ‘stabilization’ of the corpus deserves some attention. Presumably it is to speak to the tendency that every performance of a chant of the corpus was like every other performance (one usually has reference to the interval contents of the melody). But we must ask again, how would we know that? And what would be the criterion of sameness in the pre-Carolingian chant communities? The answer to both questions, presumably, would lie in the comparison of notational representations of the same items in different sources. But that presumes, again, that those representations are like snapshots of performances in the corresponding establishments. How would we know that? The circularity is manifest. What I have called the generative system or performance model underlying the Old Roman offertory that is analysed in the first chapter depends for its generative role on two properties: its provision of a set of conventional melodic units and, most important, the regulation of the occurrence of those units by a syntax—a systematic pattern of contexts in which the occurrence of each unit is fixed in the sequence of melodic units and by the syntactic function of the language unit to which it responds. This syntactical property is at the heart of formulaic chant, its essential compositional principle, whether the composition is oral or written. Syntactical function is the operative criterion on which formulae can be identified. This emphasis was already brought out in the first chapter, and it is central to analyses in the cited work of Helmut Hucke and Max Haas. Its neglect is the principal weakness of the theory of ‘centonization’ discussed in Chapter 7. The choice of a chant from the oldest of the Old Roman books as an object of analysis in Chapter 1 was made on the grounds of the ontological state of that book’s contents—as close to an oral tradition as any we know. This chapter introduces the study of trope transmissions into the investigation also because of what can be learnt from the ontological state of the compositions that are in the sources. Tropes were composed in the stylistic world of the chant, even if several centuries later, because they were intended as elaborations or extensions of the chants to which they were attached. We learn a number of fundamental things about our subject from their study. They were transmitted in writing from the beginning of the practice of troping, as far as we can tell. Unlike the chants themselves, their written transmission is not the product of the transcription of an oral tradition in musical notation. There is no reason to think that their substantial variability in transmission is the result of an oral practice, other than the a priori but unwarranted
234
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
bias that oral transmission naturally and necessarily produces variation and variation is attributable to oral transmission.¹ We learn from the example of the transmission of a single trope verse, then, that writing was not a guarantor of fixity. This is reinforced in the next chapter. Writing fixes one text, the one that is written; it does not fix the song. Fixity of transmission was not a natural musical value, and when we find it we must search for the particular reasons for it (e.g. the emphasis during the Carolingian era on correctness in accordance with the doctrine of the Roman origin and authority of the chant). The example of the trope transmission has something important to teach about the concept of generative system or performance model. Despite the considerable degree of variation in the transmission it is possible to identify an underlying pattern that is constant throughout it and that is reasonably regarded as the basis for precipitating the trope verse in either writing or singing. The result of this structuralist approach extends the impressions formed on the basis of the analyses of chant transmissions reported to this point to the recognition that composition on the basis of such models was a practice in written as well as oral tradition. Beyond that it shows that composition on such models is not specifically a phenomenon of formulaic chant alone. That is an important conclusion, for a very significant portion of the corpus of chants is not formulaic. A third example in the chapter bears particularly on the relation between notation and performance. It presses harder a question that occurs in the light of the variation in the written transmission of tropes: were the singers who used those scores for their performance (if they did) intended to sing exactly what is written? The example presses this question harder because it comprises three versions of the same passage notated in three consecutive strophes of the same composition by the same notator, but no two are identical. The differences, however, are so small to the eye that one has to wonder whether they would have been observed exactly in performance. And that, in turn, raises the question whether the notation fully and exactly represents the piece. But our suspicions about this can be raised only by reading the notation literally. This is the paradox that was observed in the second chapter, and it haunts our efforts to understand the uses of early medieval notations. Yet we have hardly taken notice of it.
} ¹ Thus David Hughes: ‘Evidently transmission took place either orally (although the variants occurring seem rather few for oral transmission) or from informal working manuscripts’. This is to explain the fluid transmission of tropes, in ‘Further Notes on the Grouping of the Aquitanian Tropers’, 4 n. 4. And Alejandro Planchart: ‘In seeking to explain the surprising uniformity of the early Gregorian tradition in the context of oral transmission’ and ‘tropes imported from elsewhere could be tinkered with and adapted to local taste and traditions . . . These adaptations often show an interaction of oral tradition, in the form of the local chant dialects, with a written-down object, the imported piece’; ‘On the Nature of Transmission and Change in Trope Repertories’, JAMS 41 (1988) 215–49 at 216 and 245 (my emphases).
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
235
The topic of musical transmission in the Middle Ages cannot be discussed without reference to the categories ‘oral’ and ‘written tradition’, and it is my intention to pose the problem in that light explicitly. In doing so I shall refer to the literature about oral poetry and the history of writing, in which these matters have received more attention than they have in the study of music. On the other hand I hope to bring to the attention of students of literature the relevance of evidence from musical studies to the oral–written question, and to begin to identify the parallels between our subjects that can constitute the grounds for a general theory of orality and literacy. I begin with a report that will enable me to develop in a fresh and accessible context some ways of thinking about musical transmission that our habits in dealing with medieval musical traditions have tended so far to exclude. Some years ago I purchased at auction in Vermont the music library of a German emigré musician named Bauer, who had settled near Boston around 1840. It included samples from every level of musical art, from handwritten copies of Haydn symphonies to horse quadrilles for the circus. Somewhere between those extremes was a print of a short piece bearing the title Le Desir: Favorite Waltz by Beethoven, published in 1842 in Boston. The piece is actually more or less the second of Schubert’s Op. 9 Waltzes, commonly known in that incarnation as the Sehnsuchtswalzer.² The tune is the same, but in the pseudo-Beethoven version it is dressed in a flashy triplet figuration that is not Schubert’s, whereas some of the characteristic bittersweet harmonic details that no doubt earned the original its nickname have been washed over by pedestrian alternatives. Schubert’s waltz is in two short sections; a third section has been added to the new version, making it long enough to sell as a separate piece. There is a context for this bit of piracy. Before 1870 at least fifteen editions of Le Desir had been published in the United States, and this was only one of some sixty ‘Beethoven’ waltzes circulating here around that time. A Beethoven boom and waltz fever were simultaneously sweeping the country. Eventually Le Desir reached Brazil. An edition was published in São Paulo around 1950 under the double title Le Desir/Sehnsuchtswalzer. Beethoven is given as the composer in the heading, but a footnote confides that the piece is really by Schubert. There is nothing contradictory about that, as long as we understand that Beethoven’s name had in effect become part of the title. ² The transmission of titles for this piece is dizzying. Schubert’s waltz first appeared under the title Trauerwalzer in 1821, but it came to be known as Sehnsuchtswalzer. Sehnsuchtswalzer may have originated as the title of the first pseudo-Beethoven publication, in 1828. While Beethoven inherited the waltz, Schubert inherited the title. But by 1832 or 1833 Schumann had associated it with Schubert’s waltz, for his Op. 4 piano pieces includes Sehnsuchtswalzer Variationen: scènes musicales sur un thème connu, republished as Scènes musicales sur un thème connu de Fr. Schubert. The theme is identified as ‘Franz Schubert: Sehnsuchtswalzer’, and it is indeed the second of the Op. 9 waltzes. The subtitle, ‘Favorite Waltz’, may be a trace of F. H. Himmel’s Favoritwalzer, on which the third section is based. See Otto Kinkeldey, ‘Beginnings of Beethoven in America’, MQ 13 (1927), 217–48 at 245, and G. Kinsky, Das Werk Beethovens (Munich, 1955), 727.
236
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
There are many variants from one edition to another, and it is not easy to distinguish between notational ineptness, typographical error, misunderstanding of a musical detail, and deliberate revision. None gets the Schubertian harmonies quite right, but the Brazilian edition does better than most, and since it does not have the triplet figuration, it gives the impression of closer contact with the original. But it does have the third section, so we know that the scribe was somehow in touch with the pseudo-Beethoven version as well. Near the end of that section he suffered a minor catastrophe that is as revealing as anything in the whole complex. What would have been the fifth bar from the end has been skipped over (the sixth and fourth bars from the end are exactly as in the other editions). Instead of going right on the scribe evidently decided to compose a new ending that would make sense of the strange lurch that he had inadvertently produced. The result is clumsy and reveals itself as an effort to make the best of a bad situation. The skipped bar suggests an act of copying, whereas the recovery was an act of composing. So was the original invention of the triplet figuration. On the other hand the fading out of harmonic detail from Schubert’s original, and the range of insignificant variation among editions, show that the transmission of Le Desir must in some measure have been a matter of writing it out from memory. The precipitation of the piece into a particular printed edition entailed composing, remembering (and forgetting), and copying, and evidently all at once. While there are variants among some editions, others are identical in every detail and seem even to have been produced from the same plates. In that case a written (printed) tradition has been established for a particular version. The range of variation apparent in the printed editions must have had its counterpart in performance. If we imagine Bauer playing Le Desir at a party, we don’t know that he would have played it exactly as it appears in the score just because he happened to own a printed score that belongs to a written tradition. If we imagine someone rushing home after the party to write out the waltz in the hope of peddling it to another publisher, there is no assurance that he would have written it out just as Bauer played it. That is, performing and writing down have in common that both may entail remembering and composing; writing down may entail copying as well, but not necessarily. In summary, an original composition, circulated through print and through performance, was transformed by way of ‘oral’ (or in this case we might say ‘aural’) process into a new version—perhaps gradually, perhaps all at once. The new version then entered its own written and oral channels. These two channels were not sharply distinct, nor were the traditions of the two versions, which seem to have crossed at least once. Each new edition or performance arose as a realization of a model that would be difficult or impossible to specify exactly. Certainly one cannot assume these models to be identical with either the original composition or its earliest transformation. The realization process entailed copying, remembering,
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
237
and composing, separately or all at once, in various mixes. Even the editions that are associated together in a written tradition based on copying alone have no special claim to authenticity or proximity to an original. And one cannot assume, just because of their mutual identity, that they would have given rise to identical performances. Indeed no edition can be assumed to specify a performance uniquely. The transmission of Le Desir has much of importance in common with the transmission of music in the Middle Ages. It brings out the topic of ‘oral’ (or ‘aural’) and ‘written’ traditions, but in a way that shows the relation between the two to be more complex and more subtle than a straightforward dichotomy. It casts doubt on the notion that there are oral or literate styles. And it has the advantage of leading us quickly to think our way clear of what I call the ‘paradigm of literacy’. Medieval music has mainly been approached by way of this paradigm, by which I mean the model of the composer producing finished works that circulate in fixed state through closed written channels in scores that are blueprints for the performers, who must be able to read them in order to produce performances that are essentially identical from one time to the next. There are signs that a state of literacy answering more or less to that paradigm had evolved in at least one musical tradition by the thirteenth century. But I shall be mainly interested here in the possibilities of describing musical traditions in which that relatively clear condition evidently did not obtain. Such a description will have much in common with descriptions of the evolution of language literacy that have been offered by Clanchy, Goody, and Havelock,³ and it will touch on issues about medieval oral poetry that are still alive.⁴ The first point in the description is the flat assertion that before the introduction of musical notation in Europe in the ninth century, all European musical tradition was exclusively oral. This may be doubted on the grounds of the suspicion, natural enough, that music writing was an older practice but that earlier traces have not survived. The arguments against this supposition are nevertheless persuasive: 1. The oldest music books are books of Gregorian chant, and the oldest chant books with neumes are from the tenth century. (Notational specimens from the ninth century occur scattered individually, in books not mainly having to do with the divine office.)⁵ ³ M. T. Clanchy, From Memory to Written Record: England, 1066–1307 (London, 1979); J. R. Goody, Literacy in Traditional Societies (Cambridge, 1968); and Eric Havelock, Preface to Plato (Cambridge, Mass., 1963); ‘Pre-Literacy and the Pre-Socratics’ (Institute of Classical Studies, University of London, Bulletin no. 13 (1966)); ‘Prologue to Greek Literacy’, lecture given at the University of Cincinnati, 1971 (printed by the University); Origins of Western Literacy (The Ontario Institute for Studies in Education, Monograph Series, 14; Toronto, 1974); ‘The Pre-Literacy of the Greeks’, New Literary History, 8 (1977), 369–91. ⁴ See Ruth Finnegan, Oral Poetry (Cambridge, 1977). ⁵ The details are given by Michel Huglo, ‘Tradition orale et tradition écrite dans la transmission des mélodies grégoriennes’, in H. H. Eggebrecht and M. Lütolf (eds.), Studien zur Tradition in der Musik: Kurt von Fischer zum 60. Geburtstag (Munich, 1973), 31–42. Huglo presents a view similar to the one adopted here
238
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
2. There are complete chant books (some of them identified as ‘books of musical art’) from the ninth and early tenth centuries that lack neumes, containing only texts or text incipits. Some of these assign the texts to ecclesiastical modes.⁶ 3. Analysis of the melodic corpus transmitted in writing in later periods shows that liturgical designation and modal assignment jointly determine the genres of chant, and that given a liturgical text and the genre in which to intone it, the singing of the chant without neumes would have constituted no problem. (There is a detailed exemplification of that in the first chapter.) 4. Lists of chant incipits with their modal assignments, that is, lists giving precisely that information, can be dated back to the eighth century. Chants circulated in an oral tradition at least between the eighth and the tenth centuries, and the supposition of a lost earlier written tradition, even if there were evidence for it, would not alter the certainty of that; but there is no evidence for it. The indirect evidence for the oral tradition reaches up to the point where the direct evidence for the written transmission begins. There is no reason not to read all the evidence at face value. Before the tenth century the tradition of Western art music was an oral tradition. Acknowledgement of this fact is a step of great importance, the necessary first step in every approach to early medieval music. By itself, however, the assertion is neutral. It says only that the channels of transmission were oral, not written; that music was spread through performance, not through books. It says nothing about how the oral tradition worked, or what the features of its products were. Those are things to be learnt from the evidence of the sources; they do not follow from the fact of orality. I wish to maintain that neutrality in what I develop here—that is, I wish to steer clear of the idea of oral and literate styles or mentalities. It should be noted that I distinguish between ‘literate’ and ‘written’ tradition, reserving the former designation for traditions in which writing and reading are the sine qua non of performance. A literate tradition certainly entails writing but that any particular written transmission reflects a literate tradition has to be demonstrated, for it cannot be taken for granted. The indications of an oral tradition that I have cited suggest a continuity of practice from the oral to the written stage. That is confirmed alone by the correspondence between the words and their modal assignments in the neumeless sources of the ninth century and the words and their written musical settings in the books of the tenth. Music writing was introduced into an oral tradition, and it must have been used initially in support of that tradition. This relationship seems characteristic of all communications media that began as oral traditions and regarding the broad outlines of the history of transmission, but he has not developed a similar view about the nature of the oral tradition. My views on both matters approximate those of Hucke, ‘Toward a New Historical View’. ⁶ For details see Helmut Hucke, ‘Der Übergang von mündlicher zu schriftlicher Musiküberlieferung im Mittelalter’, in IMS Berkeley, 180–91 at 188.
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
239
developed written forms.⁷ Recognition of the continuity from oral to written tradition is the second step in an understanding of early medieval music. As for the nature of that continuity, two general views could be taken. The first is that the oral tradition produced a corpus of chants that was retained in memory, and that the written transmission represents a precipitation of that corpus, a translation of it into writing. The second is a little more subtle: that the oral tradition of the chant was a performance tradition in which the singers intoned the words of the Divine Office according to traditional principles that they had assimilated. From this point of view, early notators of music would themselves have been like singers, calling forth the chants from memory by following the principles of the performance practice; instead of singing out in the service, they wrote down in the book. This would be to regard writing down as a kind of performance and to understand the notation as descriptive. There are indications for both interpretations. For the first, there are signs of some degree of fixity in the chant tradition before the writing down. For the second, there are signs of fluidity in the written transmissions of many melodies—of chant and other traditions—and other signs that oral processes continued to operate after the onset of writing.⁸ This is not to suggest that the evidence is contradictory, but rather that the description of early written transmission requires a broader paradigm in which the two models that I have characterized are continuous. The need for such a paradigm becomes plain if one considers a theoretical issue that has been argued in discussions of ‘oral poetry’, namely, the relationship of memorization to oral composition. Albert Lord has always drawn a sharp line between them, recognizing a role for memory only in the unconscious retention of the formulae that according to his paradigm are automatically retrieved during composition.⁹ Ruth Finnegan accepts the distinction, but she argues that the oral transmission of poetry can be a matter of memorizing, and then remembering, whole poems. There is a need on both sides, I believe, to attend to the evidence from cognitive psychology that the distinction is not realistic and that the controversy is therefore moot.¹⁰ From the viewpoint of psychology there would seem to be little question but that oral composition is a kind of remembering, for remembering is always an act of reconstruction on the basis of patterns that have been internalized. In a regulated tradition what looks like the reproduction from ⁷ It constitutes an essential point in the theories of both Clanchy and Havelock. ⁸ Signs of both kinds are described in the introduction to Ch. 6. ⁹ Lord’s article ‘Oral Poetry’ in Princeton Encyclopedia of Poetry and Poetics, ed. Alex Preminger (Princeton, 1974), 591–3, presents the sharpest formulation of his position, which he somewhat softened in a late essay, ‘Memory, Fixity, and Genre in Oral Traditional Poetry’. See the introduction to Ch. 6 for details. ¹⁰ For a survey of research in experimental psychology that bears on the subject of oral transmission, see David Rubin, ‘A Contribution to the Roundtable on Transmission and Form in Oral Traditions’, IMS Berkeley, 173–80. For citations of more recent publications see the introduction to Ch. 6.
240
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
memory of a fixed poem or song is different only in degree from the invention of one that is ‘new’. Written transmission that documents a performance practice is exemplified by the transmission of the Old Roman offertory that is analysed in the first essay.¹¹ What is described there is a generative system for offertories in F in the Roman tradition. Given a text, a knowledge of the principles inferred in the description should be sufficient for the making of such a chant. It also constitutes the description of a genre. The identification of generative systems for the genres of plainchant answers the question about the nature of the oral tradition, for neither the transmission and retention of the principles of such a generative system nor its realization as a particular chant demands or even suggests writing. For that reason alone I call the making of chants on the basis of such systems an oral process. But then there is no reason to think that it is a fundamentally different process if the maker has written the chant down in a book rather than singing it out in the service. That is what it means to suggest that the book may be understood as the documentation of a performance practice, either at first hand or by way of copies. This situation has parallels in the study of literature. Parry, Lord, and Magoun had claimed that formulaic style was an ‘indisputable sign’ of oral composition, whereas Larry Benson demonstrated that formulaic style is ‘characteristic not just of the “oral” epic of Beowulf but also of some written composition in Old English, including Old English translations from Latin originals.’ Benson concluded: ‘to prove that an Old English poem is formulaic is only to prove that it is an Old English poem, and to show that such a work has a high or low percentage of formulas reveals nothing about whether or not it is a literate composition.’¹² It strikes me that the wrong question has been asked of the evidence. Formulaic style, like the generative system described for the Roman offertory, is what Walter Ong has called a ‘residual’ of oral tradition.¹³ Both are understandable in the light of a continuity from oral to written practice. That seems a fundamentally more important historical matter than the question of whether this or that production should be classified as ‘oral’ or ‘literate’. With respect to the ‘residuals’ of oral tradition in written practice, it is necessary to recognize yet another distinction, between those residuals that are still active in the production of poetry and music in the written medium, and those that are literally tags of oral production, transcribed and brought down in writing through a generation or more of copies. Only through control of the evidence is it possible to make judgements about what is involved in particular cases. In Old English literature, as Benson recognized, translation from the Latin is an instance of the first kind. ¹¹ It is recommended to the reader to review that analysis at this point. ¹² Summarized from Finnegan, Oral Poetry, 69–70. ¹³ ‘African Talking Drums and Oral Noetics’, NLH 8 (1977), 411–30.
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
241
The Roman tradition represented by the offertory Factus est dominus is of special interest because of the evidence it presents with respect to these questions. The oldest chant books referred to above are Frankish. The Frankish tradition has been the basis of the official chant of the Western Church ever since Charlemagne, and there is a tradition going back to his time for tracing the origins of that chant—known since then as ‘Gregorian chant’—to Rome. But the oldest surviving chant books of Roman origin are three graduals and two antiphoners, the oldest of them dated to 1071 and the others to the twelfth and thirteenth centuries. Their tradition is distinct from the Frankish one in style and partly so in content, but sufficiently close to it in liturgical and melodic content to suggest a common ancestry. The written tradition of Frankish chant is more than a century older than that of Roman chant. Furthermore, the Frankish tradition continued, in some sense, into modern times, while the written tradition of Roman chant ended with the establishment of the Frankish tradition in Rome in the thirteenth century. By means of indirect liturgical evidence, however, it has been possible to trace the indigenous Roman tradition back to the eighth century.¹⁴ That means at least three centuries of oral tradition, followed by at most one century of written transmission. None of the Roman books can be very far removed, relatively speaking, from the oral tradition in Rome. Analysis of the written transmission reinforces this interpretation and also shows that the Roman tradition had a longer period to develop the characteristic economies of an oral tradition.¹⁵ In fact the sources show that during the century or less of writing, the Roman community did not really develop a written tradition in the strict sense at all. The Romans did not even develop a notational idiolect, as did every other region with its own written tradition; they simply used the system of nearby Benevento. And among the Roman office and mass books there is disagreement on matters of every sort, disagreement that has earned the Roman tradition a reputation for unreliability (on the dubious premiss, of course, that a written tradition must be uniform in order to be reliable, that is, it must reflect consistently the transmitted material in its ‘original’ or ‘authentic’ form). It seems that the individual books present more or less independent precipitations of the oral tradition. Accordingly they are close witnesses to the performance practice and primary sources for the study of early written process in an oral tradition. The realization of the Roman offertory Factus est dominus out of its generative system was an oral process. There are two indications that written process entered into the transmission at the same time. The first is the evidence of a Frankish source as a model, which seems to have interfered with the Roman adaptor’s correct application of the principles of his own tradition. The second is the melisma ¹⁴ Huglo, ‘Le Chant “vieux-romain”’. ¹⁵ This hypothesis is developed in Ch. 6.
242
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
at the wrong pitch level, which offers especially good insight into the nature of writing in pre-literate circumstances. The melisma, we must presume, was fixed as a resource of the performance practice for that genre. The notator wrote it out for the third verse, either from memory or by copying it, perhaps from the second verse; it doesn’t really matter which. Either way, there was not the feedback of content that we take for granted as an aspect of writing in a literate tradition. And so when the notator ended too high at the conclusion of the melisma he was taken by surprise and had to improvise, so to speak. To be sure such things happen often to copyists in traditions that we do not hesitate to call literate. But the point here is that this episode is an aspect of a transmission that also entails oral processes. This notator was not a copyist. Like the Brazilian editor of Le Desir, he was copying and remembering and composing, all at once. The written transmission of the Roman offertory Factus est dominus illustrates with unusual clarity the phenomenon of writing down, with previously written models as reference, in the midst of an oral tradition. The following example, on the other hand, shows the marks of oral process in a tradition that was evidently transmitted in writing from the beginning—the tradition of the liturgical tropes. Their written transmission began more or less contemporaneously with that of the Frankish chant, and it spread over the same regions. If their transmission nevertheless shows signs of oral process, we shall have indication from the other side that oral and written processes functioned quite consistently together in the transmission of music between the introduction of the technology of writing and the establishment of literacy. Example 10.1 is a trope verse for the introduction of the second-mode introit Mihi autem nimis, transcribed from two Aquitanian sources of the eleventh century (a and b), a northern French source of the twelfth century (c), and a Beneventan source of the twelfth century (d ). Verses (e) and ( f ) are taken from tropes for a different introit of the same mode. Their role in the analysis will become clear presently. The first phrase (M ) exemplifies a perplexing feature of the sources of trope melodies—the often heterogeneous nature of the transmission. For some trope melodies, however, there is a high degree of uniformity; there can be a dozen sources for a trope melody that are virtual replicas of one another. The tendency has been to take the uniformities as the norm, reading them in the light of the paradigm of literacy and sweeping the rest under the carpet. One could well do the opposite: read the replicas in the light of the production of multiple editions of Le Desir from the same plates.¹⁶ In the end a satisfactory theory of transmission must take account of both sorts of phenomena. Often settings of a trope text that appear at first to be melodically independent of one another turn out on closer examination to have common features of ¹⁶ The reasons for doing so are developed in Ch. 11.
243
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
Ex. 10.1. Four versions of the trope verse Filius ecce patrem conpellans taliter inquit from (a) Paris lat. 1871; (b) Paris lat. 887; (c) Paris lat. 1235; (d ) Benevento VI. 34; (e) and ( f ), for comparison, two trope verses of the same melodic type from Paris 1871 M
N
(a) 8
Fi
li us
[A]
pos to
ec ce
pa
trem
con
pel
li
cis
ve
ne
ran
dre
as
su
per
as
(b) 8
(c) 8
(d) 8
(e) 8
(f) 8
[C]ul men a
pos to li cum An
N cont.
tra be a tus in
sig
nis.
O
(a) 8
lans
ta
me los
clan ga
li ter in
quit
(b) 8
(c) 8
(d) 8
(e) 8
dis
mus
ho
no ri bus
(f) 8
re
ti net, nos in de ca na
mus
om
nes
structure and content at some underlying level. In Ex. 10.1 the syllabic version of M in Paris lat. 1871 has done the analysis for us, so to speak, by showing the simplest sort of structure, a circling motion around the tonic, presented without any embellishment. The other inscriptions of the phrase are all versions of that (which
244
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
is not to imply that the version of Paris 1871 is the original one; we can say only that the priority that it has is structural). This structure, together with the rules for its elaboration, constitutes the simplest sort of generative system.¹⁷ Certainly there can be more complex generative systems for tropes, but they would be realized in an equally wide range of versions. Just as in the case of the Roman offertory Factus est dominus, the production of trope melodies from such generative systems is an oral process. Now that two examples have been given of oral process in the transmission of medieval music, an attempt at a definition of that process is called for. It is a process in which music is received and coded through hearing, retained schematically in memory, and performed or transferred to writing from some mental idea of it. Of course this is a definition in which place has deliberately been left for the interaction of oral with written processes. Thus ‘hearing’ should be broad enough to include the situation in which a cantor, compiling a book for his own use, sings— aloud or in his mind—from written exemplars and writes down a version of what he has assimilated from his singing. There is a risk that this definition is so broad as to exclude little. But it seems hard to justify distinguishing in any really fundamental way between the situation just described and one in which a notator simply writes down what he has heard somewhere. The intercession of oral process into what has conventionally been regarded as straightforward transcription from one source to another must be one characteristic of pre-literate written transmission. The underlying structure identified for the phrase M in Ex. 10.1 is not unique to this trope, as can be seen from (e) and ( f ). In fact this structure is just a common melodic turn of the mode. One might therefore conclude that the versions of M are independent of one another, and that the resemblances are simply a consequence of their having been composed for the same introit chant, were it not for the close similarity in versions (b), (c), and (d ) of phrase (a). This raises a difficult question, but it is one that must be confronted repeatedly in particular cases. Fundamental to the understanding of transmission is the distinction between what is there by the constraints of the tradition and what by the design of the individual musician, between features common to the musical language and those characteristic of the particular piece. Given a transmission like that of Ex. 10.1, two further questions suggest themselves. First, why was the music written out at all? Second, can one assume that the users of the written sources would necessarily have sung the phrases exactly as written? These questions can be pressed a bit harder with an example from another genre, a cadential melisma from a versus of the Aquitanian tradition, transcribed from a single source (see Ex. 10.2). The piece is a strophic song with ¹⁷ See Nowacki, ‘The Syntactical Analysis of Plainchant’, for an example of how such rules can be formulated rigorously.
245
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
Ex. 10.2. Cadential melismas of three strophes of the versus Ex ade vitio from Paris lat. 1139
8
tra xit pri
8
8
mor
di
a.
fa cta con cor
di
a.
fe
tri am
ne tram vi
Latin text, and the notator has done us the unusual service of writing out the melody for all three strophes (unusual because it was normal to write out the melody just once). One recognizes the same general pattern, but the details are never quite the same twice. In this case the answers to the two questions are: first, the notator seems to have written out exemplifications of the pattern of the melisma—here, too, the notation documents a performance art; second, nothing compels us to assume that the user of the manuscript would have taken care to sing each variant exactly as written, other than the scruples we have learnt in our own training about performing from scores. We do not know that those scruples held then, and no analysis of the melodic passage is likely to reveal something obligatory about just those ways of writing it out. This interpretation gives concrete meaning to something that is intuitively plausible and that has been generally understood by historians of writing: that writing everywhere has been introduced into oral traditions, that it has not initially displaced those traditions but assumed a role within them, and that a fundamentally different condition of literacy, affecting not only the modes of transmission but the contents as well, evolved long after the beginning of writing. Havelock reckons some four centuries, between Homer and Plato, from the adoption of alphabetical writing by the Greeks to the spread of literacy. Clanchy writes that two and a half centuries elapsed, between the Battle of Hastings and Edward I, before writing, which was known throughout that time in England, became the normal basis for government and property transactions. I believe that a musical tradition corresponding to the paradigm of literacy is not demonstrable in Europe before the thirteenth century, four centuries after the earliest music writing.¹⁸ ¹⁸ I refer to the tradition of the so-called School of Notre Dame in Paris. The tradition goes back to the last third of the 12th c., but the written transmission begins in the 13th—hence my hedging between these two centuries at various points in this paper. The grounds for considering this tradition a literate one are given in my article ‘Regarding Meter and Rhythm’. Reckow, Der Musiktraktat des Anonymous 4, ii: Interpretation der Organum Purum-Lehre (Wiesbaden, 1967), offers a similar interpretation of the Notre Dame tradition. His emphasis is on composition as against extempore practice rather than on the difference in transmission,
246
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
If this is so then there must be other ways of understanding the function of written scores during that long period of overlap than as blueprints for performance and closed texts representing fixed works. They served to bestow canonical status on certain traditions, and to bestow authority on their possessors. They served as touchstones for the regulation of oral traditions—as a ‘court of last appeal’, in Havelock’s felicitous phrase.¹⁹ The replication of written documents can be understood in the light of such symbolic, exemplary, and regulatory roles, without reading such replication as a sign of the fixed state of the contents as they circulated through oral channels. For musical documents in particular the relation to performance may be less direct than what we are accustomed to in literate traditions. If the written score was achieved through realization of the principles of a traditional generative system, then perhaps its service to the performer was to exemplify the system rather than to provide note-by-note instructions. The performer would study the book and make a piece like the one represented, rather than reading out from the score. This may be suggested just by the small postcard and even cigarette-package size of the earliest books with musical notation.²⁰ The examples so far noted suggest that the written musical sources before the thirteenth century are documentations of an oral—that is a performance—practice. One other category of evidence allows the relation between performance and writing to be narrowed down still further. When I was a child we thought that the first line of the hymn ‘Gladly the Cross I’d bear’ signalled that it was about a crosseyed bear named Gladly. Of course this can happen only in oral transmission, and if we had made a written version of what we heard it would be a sure sign that oral transmission had occurred. The faulty text could only be the transcription—direct or through copies—of something heard. Such things turn up in the written transmission of tropes, and they constitute direct evidence of an oral process in the transmission.²¹ Here is a clear case. In a number of sources of the eleventh century there is a troped Agnus Dei text given in both transliterated Greek and Latin. In this citation I shall reverse the order, showing first the Latin, then an idealized, composite, ‘correct’ Greek text, and beneath that selected specimens of the Greek text as it appears in some of the sources:²² but these are two sides of the same phenomenon. Edward Roesner has made a substantial new contribution to our understanding of the Notre Dame tradition vis-à-vis orality and literacy in ‘Who “Made” the Magnus Liber?’. ¹⁹ ‘The Pre-Literacy of the Greeks’, 374. ²⁰ See Hucke, ‘Toward a New View’, 447. ²¹ This is something like the ‘Rebus Principle’ described by I. J. Gelb in A Study of Writing (Chicago, 1952), 252. ²² The example is taken from Tropes de l’Agnus Dei, ed. Gunilla Iversen (CT 4; Stockholm, 1980), 59–60. The manuscripts cited are: (a) Paris lat. 1119; (b) Paris lat. 909; (c) Paris lat. 1120; (d) Paris lat. 1834; (e) Paris lat. 1871; (f ) Paris lat. 1084; (g) Vienna 1888; (h) St. Gallen 376; (i) St. Gallen 378; (j) St. Gallen 380; (k) Paris lat. 9449; (l) Oxford Selden supra 27.
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
247
Agnus dei, filius patris O AMNOS TU THEU O HYOS TU PATROS annos oio oyo qui tollis peccata mundi miserere nobis O ERON TIN AMARTIAN TU COSMU ELEYSON YMAS o erontin elyso nymas o eruntin o erontas amarthias o eruntas oe rontas amartias amartas There are many instances of this sort of thing even with Latin texts, and they are all parallel to the case of the cross-eyed bear. It is possible to identify musical counterparts, although these do not have the same immediacy because what greets the eye is not nonsense in the way that the texts are nonsense. The phenomenon may be described in general terms as a sequence of sounds that is realizable in different configurations. In the absence of controls, the configurations are randomly determined. In the case of the cross-eyed bear the control that would come from an understanding of the theological propriety of the text is lacking. With the Greek text the errors result from the absence of controls that would have been exercised by anyone who knew the correct word boundaries of the language. In the case of melody as sound sequence, the configurations are notes grouped together as carriers of single syllables. When the same note-string is linked differently with the same syllable string in different sources, the situation is parallel, and it is again a mark of an oral process. The text is given and the note sequence has been mentally retained. But there are no controls operating to fix the latter upon the former in persistent configurations. What those controls might be are rules for the declamation of the accented syllable in a cadential figure, or stereotyping of an entire cadential formula, as in the case of the Roman offertory Factus est dominus, or the stereotyping or mechanical copying of the entire song. Example 10.1 illustrates the point. The ending of the trope line entails an ascent through the notes c–d–f (at taliter) and an ornamented return through e to d (at inquit). All the versions do that, but Paris 1871 begins one syllable later than the others. This category of variant is very common. It gives the impression that the notator of each source knows how the tune goes with which he is to declaim his text, and the manuscript documents his own process of doing it. In the transmission of the Greek Agnus text, the transformation of the Greek Amnos to the Latin Annos illustrates a different but related phenomenon of oral
248
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
transmission that I shall call the ‘Marylebone’ or ‘Pettycoat-Tail’ phenomenon. ‘Marylebone’—traditionally pronounced ‘Meribon’—is a London railway station named after a district once called Marie la bonne. ‘Pettycoat Tails’ are a variety of Scottish shortbread, adapted from a French confection called Petit gâteau taillé.²³ Around New York City there is a version of the Pater noster with the line ‘And lead us not into Penn Station’. Each time an unfamiliar string of sounds is assimilated into a familiar code. In medieval music that is apparent whenever a song of one tradition has been assimilated into another, taking on the characteristics of the host tradition in the process. In such cases the foreign material, even if it arrived by way of written books, has been transformed by local habits of hearing and performing and then newly transcribed in writing. It is again a matter of the intercession of an oral process in written transmission. Written processes came to dominate in the transmission of some traditions of medieval music. There would be two general ways to understand that intuitively, and both are well substantiated by the evidence. One approach is to see this result as inherent in the nature of the notational system itself. The key was suggested by Charles Seeger: ‘The occidental notational system is par excellence a control system.’²⁴ The history of notation in the Middle Ages can be understood as the incremental institution of controls—controls, that is, as against the habits and constraints that the performer brought to the performance. Nearly all the oldest notations, which did not encode pitches or intervals but only pitch contours, were fully explicit about just one musical parameter, the coordination of groups of notes with syllables of text. This must be taken as their primary role, reflecting in turn the primacy of melody for the articulation and explication of language.²⁵ The encoding of pitch intervals in Western notation was accomplished through the introduction of principles or devices, all of which can be understood as controls: diastematic neumes, the horizontal line as pitch referent and the extension of that idea in the stave, the use of pitch letters, signs of equivalence between notes at different levels, custodes (single notes) at the end of a line to indicate the level of the first note of the next line, and clefs. The increase of controls in the pitch domain, a phenomenon mainly of the eleventh century, was followed and closely paralleled about a century and a half later by the institution of controls in the rhythmic domain. As with pitch, the earliest rhythmic notation (‘modal notation’) was not a blueprint for performance, but set the performer under way in his own construction of the rhythmic patterns. Into that notational system were introduced signs with explicit durational meaning, which functioned as controls. These developments make an especially strong ²³ I learnt of this from Ruth Wallace Wijkman. ²⁴ ‘Oral Tradition in Music’, in Funk and Wagnall’s Standard Dictionary of Folklore, Mythology, and Legend (New York, 1972), 825–9. ²⁵ See Chs. 15–17.
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
249
impression when the sources for a single piece transmit essentially identical contents, save that the later sources substitute more explicit signs. Then there has been, in effect, a notational revision, a purely written process that we can read as a sign of literacy. Any system of notation that functions as a system of controls is based on the assumption that the performer will approach the book with some notion about what he is to perform. The notation gets him started, perhaps, and keeps him on the track. The idea of a control system is just the right way to think about the role of notation in an oral performing tradition. Yet there is a point beyond which controls operate so closely and so constantly that the notation becomes in actuality a system of direct representation rather than of controls. When it functions that way, notation has realized an ideal that was expressed by writers on music during the entire period under discussion here, from Carolingian times to the thirteenth century—for explicit notations that could be read off by performers coming to them cold, for prescriptive rather than descriptive notations.²⁶ Thus Hucbald of Saint-Amand (c.840–930): As the sounds and differences of words are recognized by letters in writing in such a way that the reader is not led into doubt, musical signs were devised so that every melody notated by their means, once these signs have been learned, can be sung even without a teacher. But this can scarcely happen using the signs which custom has handed down to us and which in various regions are given no less various shapes, although they are of some help as an aid to one’s memory.²⁷
Guido of Arezzo (c.1000–50) claimed that by means of the staff notation for which he is known ‘boys who until now have been beaten for their gross ignorance of the Psalms and vulgar letters’ will be able to sing antiphons ‘correctly by themselves’ and ‘without a master, something which, with God’s help, any intelligent and studious person will be able to do if he tries to understand with what great care we have arranged the neumes’.²⁸ And the English writer of the thirteenth century known to us as ‘Anonymous IV’, noted: ‘In the old books the notation was equivocal because the simple signs were all alike. The singers proceeded only by their intellect, saying “I think this one is a long, that one a breve”. Thus they worked a long time before they learnt well something which now everyone can easily learn by means of these explanations, so that today every student will achieve more in one hour than formerly in seven.’²⁹
²⁶ ²⁷ ²⁸ ²⁹
See Charles Seeger, ‘Prescriptive and Descriptive Music-Writing’, MQ 44 (1958), 184–97. Melodic Instruction, in Hucbald, Guido, and John, 36. ‘Prologue to his Antiphoner’, in Strunk’s Source Readings, 212. Reckow, Musiktraktat, i. 49–50.
250
Ora l , W r i t t e n , a n d L i t e rat e P ro ce ss
The one condition that could make possible the realization of this long-desired ideal for musical notation is a conception of writing that is fundamentally different from that of controls functioning in an oral tradition. For the writing down of speech John of Salisbury articulated the new idea in the twelfth century: ‘Fundamentally letters are shapes indicating sounds. Hence they represent things which they bring to mind through the windows of the eyes. Frequently they speak voicelessly the utterances of the absent.’³⁰ An exact counterpart for music writing is this definition by Jehan des Murs (c.1319): ‘The musical note is a quadrilateral figure arbitrarily representative of numbered sound measured by time’, which gives a very precise sense to his claim that ‘whatever can be sung can be written down’.³¹ When writing is thus conceived in the light of a complete equivalence between uttered sound and written sign it can function as an autonomous mode of direct communication, and no longer just as an aid to memory, as Hucbald put it. It means, furthermore, that writing is to be read; reading is the correlate of writing. These two conditions, an autonomous script and the parity of reading and writing, are defining characteristics of a literate tradition.³² This line of reasoning was introduced with the suggestion that there are two approaches to an understanding of the eventual predominance of the written process in the transmission of medieval music. The first proceeds from the viewpoint of the nature of notation. The second proceeds from the viewpoint of the contents that are transmitted. If content is represented with increasing specificity, were there changes in the nature of the contents that called for, or at least exploited, a more explicit notation? In the passage quoted from Jehan des Murs there is a hint of what was in fact a change of the most radical sort in what musical notation had to represent. His attention there is not to pitch, but to time. This reflects the major new preoccupation of composers and theorists of music in the twelfth and thirteenth centuries: the coordination of time, a newly recognized dimension for musical ordering, with pitch. The concern with the duration of notes was not prompted only by a new interest in organizing melodic sequences in rhythmic patterns as a matter of stylistic preference. Because most of the new music of the period involved two or more voices, it was the coordination of voices that forced attention to duration; more specifically and essentially, it was the control of the sequence of consonances and dissonances formed by the voices. The regulation of voices in respect of both pitch and time created a new level of complexity. It called for decisions to be made in advance and communicated, through notation, to performers. The effect of this ³⁰ Cited by Clanchy, From Memory to Written Record, 207. ³¹ Ars novae musicae, in Strunk’s Source Readings, 265, 268. ³² Havelock points to the individuality of written creations as a mark of literate tradition (‘The PreLiteracy of the Greeks’). It hardly needs to be said that an autonomous script is a precondition of such individualism.
Ora l , W ri t t e n , a n d L i t e rat e P r o ce ss
251
was a tendency to fix music, both conceptually and notationally—not for canonical reasons, or initially for aesthetic ones, but primarily for practical reasons.³³ Musical literacy is the transmissional side of that radically new situation. ³³ This point is fully developed in my essay ‘Regarding Meter and Rhythm’.
252
Th e Transmission of Some Aq uitanian Tr ope s
chapter 11
Observations on the Transmission of Some Aquitanian Tropes
} The question that prompted the following study flowed quite naturally from the one posed at the beginning of Chapter 6, which prompted the entire project represented by this book: if the written transmission of Gregorian chant bears the marks of its earlier oral composition and transmission, how does that compare with the features of the early medieval tropes in relation to their modes of composition and transmission? The trope tradition seems natural for the comparison for four reasons: (1) the melodies were composed in the stylistic environment of the chants they staged; (2) there is no indication of any intention on the part of state or ecclesiastical authorities to impose on the repertory of tropes a standard of uniformity or correctness in accordance with some designated authentic tradition as there was in the case of the chant. That is reflected in a fluidity in the written transmission of tropes, in contrast to that of the chants, both in the notated versions of the same tropeverse melodies and in the different selections of trope verses made in different locations for the same chant within the same liturgical moment;¹ (3) the earliest written transmission of the tropes is essentially contemporary with the earliest written transmission chants; and (4) unlike the situation of the chant, there is no independent indication of an unwritten tradition of trope composition and transmission antecedent to the written transmission. For those reasons this comparison has proven to be as important a source of understanding regarding our general subject as the comparison between the Frankish and (old) Roman transmissions of Gregorian chant. I must comment briefly on the assertion I made in stating the fourth point for the comparison, for there has been a tendency to regard the fluidity in the written transmission of tropes as just the sort of indication of an oral tradition that I have claimed does not exist in the case of the tropes.² The tendency rests on the a priori ¹ Something that is easily overlooked with all the attention to the uniformity of the chant tradition is that as a consequence of this fluidity in the trope transmission the actual performed troped mass on any one day in the church calendar would have varied widely from place to place. ² See Ch. 10 n. 1.
Th e Transmission of Some Aq uitanian Tr ope s
253
assumption coupling oral and written tradition with variability and stability respectively: given the fluidity in the written transmission of the tropes, a background of oral tradition is suspected. The interpretation is clearly circular on grounds of logic alone, but there is abundant evidence from cultures around the globe that those couplings are by no means necessarily operative.³ Severing the presumed coupling of writing with fixity is one of the two main outcomes of this study. That is why the word ‘independent’ is emphasized in the formulation of point (4). And it reminds us again of the mysteriousness of early music-writing that has been teasing us throughout this book and that will be addressed head-on in Chs. 13 and 14. The other main outcome is a palpable demonstration of a musical tradition entailing local production for local use, which is nevertheless based on underlying models and patterns that are shared from place to place and transmitted by voice and pen through the sung and written compositions themselves. The more negative philological way of putting this is that all Aquitanian tropers are terminal, that they cannot be fitted on a stemma that will tell anything about their genetic relationships.⁴ This seems to imply some failure in the manuscripts or their makers, or in the methodology of the search. But more likely the problem lies in the expectation that classical philological methods should be able to take the measure of such a tradition of composition, performance, and transmission. A consequence of this outcome is that critical interpretations of this music are questionable if they are based on the premiss of a work concept.
} This study of seven introit tropes or trope groups is addressed at once to a methodological task and a historical question. The methodological task is to find ways of analysing tropes that will take account of their transmission. It is a premiss of this book that in the study of music that pre-dates the time when transmission became demonstrably reproductive, analysis necessarily entails an account of transmission and the study of a transmission is necessarily a task of musical analysis.⁵ With respect to the tropes, that task will come down to a search for concrete musical parameters in terms of which melodic units can be compared, and for criteria by which we can consistently draw the line between ‘versions’ of a single piece and ‘different’ pieces; or to put it more succinctly, for what it means to say ‘same’, ‘different’, ‘version’, and ‘variant’ with reference to this repertory. ³ For references to unwritten performance traditions that can produce music and poetry that is virtually identical from one performance to another see Ch. 1 and Ruth Finnegan’s book Oral Poetry. ⁴ See Hughes, ‘Further Notes on the Grouping of Aquitanian Tropers’. ⁵ The argument for this premiss, and the terms in which it is couched, are elaborated in Ch. 6. My task here is to articulate the concept ‘matrix of transmission’ with reference to the tropes. Certain aspects of the discussion here will also touch on an argument developed at length in Ch. 7.
254
Th e Transmission of Some Aq uitanian Tr ope s
In the end this comes down to the question ‘what is transmitted?’ That question is the access to the historical question to which this chapter is addressed: what is the state of musical production and transmission that is represented by the notations of Aquitanian tropes of the tenth and eleventh centuries, and what is the role that the notation plays in the transmission? And in the intersection of the methodological question, how can we control the musical elements for systematic comparison?, and the historical one—how can we understand the transmission of the repertory?—we may locate the concept of style.
I The Examples I have drawn selectively from the sources for the examples, and I list here only those sources that I have quoted. I have not reported minor variations of spelling, grammar, or wording in the texts, as these do not seem to me to be identifiable as factors in the musical transmission. The most complete concordance now available for tropes will be found in the volumes of the Corpus Troporum. 1. Introit Introduxit vos, mode 8; trope Dulciter agnicole Sources: Paris lat. 1120, fo. 24r; 909, fo. 25v; 1121, fo. 14v; 1119, fo. 26v; 1084, fo. 1r and fo. 131v; 1118, fo. 48v; 887, fo. 21v (the ‘variant’ in this source published by Weiss⁶ is not accurate according to the MS); Apt 17 (5), fo. 211r. Evans,⁷ 165; Weiss, no. 316. 2. Introit Etenim sederunt, mode 1; trope Hodie Stephanus martyr Sources: 1240, fo. 19v; Paris nouv. acq. lat. 1871 (hereafter 1871), fo. 6r; Bologna 2824, fo. 26v; Modena 1.7, fo. 116r. Evans, 137; Planchart, no. 55; Weiss, no. 24. For comparison: trope Iubare corusco, Paris 1119, fo. 9v; trope O quam pretiosa est, Apt 17, fo. 57 r. 3. Introit Mihi autem nimis, mode 2; trope Filius ecce patrem Sources: 1871, fo. 40r; 887, fo. 44r; Paris nouv. acq. lat. 1235 (hereafter 1235), fo. 242r; Benevento VI 34, fo. 235r. Planchart, no. 105; Weiss, no. 136. For comparison: trope Apostolicis venerandis melos, 1871, fo. 37v; trope Culmen apostolicum Andreas, 1871, fo. 37v. 4. Introit Ecce advenit, mode 2; trope Adveniente Christo Sources: 1121, fo. 9v; 909, fo. 18v; 1119, fo. 14r; 1084, fo. 128v; 1118, fo. 30v; 887, fo. 16v 1871, fo. 10r. Evans, 149; Planchart,⁸ no. 45; Weiss, no. 108. ⁶ The reference is to Introitus-Tropen, i, ed. Weiss. Although this is presented as a critical edition it is quite incomplete in its reporting of variants, which the editor gives the impression of regarding as irritants in the quest for a central or authentic tradition. ⁷ Evans, The Early Trope Repertory. ⁸ Alejandro Enrique Planchart, The Repertory of Tropes at Winchester (Princeton, 1977).
Th e Transmission of Some Aq uitanian Tr ope s
255
5. Introit Ecce advenit, mode 2; trope Haec est preclara dies, line III Sources: 1121, fo. 8v; 909, fo. 17v; 1119, fo. 13v; 1084, fo. 60r; 1118, fo. 30r; 887, fo. 17r; 903, fo. 150r; 1871, fo. 9v; Apt 17, fo. 88r. Evans, 147; Planchart, no. 50; Weiss, no. 113. For comparison: trope Per quam Christicole, 1871, fo. 38v (fourth line of the trope Hic trina sonat for the introit Terribilis est). 6. Introit Ex ore infantium, mode 2; trope Filii carissimi Sources: 909, fo. 16v; 1119, fo. 12r; 887, fo. 15r; 1871, fo. 8v. Evans, 144; Planchart, no. 60; Weiss, no. 123. 7. Introit Os justi, mode 6; two tropes: (a) Iubilent omnes, (b) Psallite omnes Sources: (a) 1121, fo. 33v; (b) 1121, fo. 33r. For the purposes of this chapter I am concerned with the kinds of variation and consistency that are evident, and I have not attempted to record the frequency of occurrence and patterns of distribution of particular readings. Those factors are ultimately relevant to an analysis of transmission, but a detailed report on them would not alter the formulation of the problem that will be developed here. When it comes to a discussion of variant readings, we must be armed with an understanding about the signs on the paper that we are about to read and their relation to the sounds they must be meant, in some sense, to represent. That subject is brought into view by the difficulties that some of these sources present to the transcriber. The difficulties are not merely practical matters to be dealt with as matters of editorial technique; they have significance for the larger task at hand. What I am referring to, of course, is the inconsistency of the vertical placement of the neumes in some of the sources. Generally that is taken as a reflection on the quality and value of those sources. But such an attitude presupposes a role for the notation of the sources that it may never have been expected to play, and it may obscure what we might otherwise learn from the state of the notation about the nature of the transmission. Evans writes: ‘The musical notation [of the sources] is usually very accurate.’⁹ I do not share that impression, and in any case I do not agree with that way of putting it. We could describe the notation of the following sources as ‘accurate’: 1121, 909, 1118, 903, 779,1871; but not the notation of these: 1240, 1120, 1119,1084, 887. Now I should like to consider what might be meant by ‘accurate’ with reference to two concrete examples.
⁹ Evans, The Early Trope Repertory, 33.
256
Th e Transmission of Some Aq uitanian Tr ope s
1. In the 1119 version of Ex. 11.5, line IV, the word ‘hodie’ was at first omitted by the scribe and then entered, together with its neumes, in the margin, thus:
From the more ‘accurate’ source 1871, the passage can be transcribed thus:
8
ho
di
e
2. In the 1120 version of Ex. 11.7(b), line II, the setting of ‘legionibus’ is notated thus:
From the ‘accurate’ source 909, the passage can be transcribed thus:
8
le
gi
o
ni
bus
If we assume in each case that the scribes of the two versions meant to notate the ‘same’ passage, then in the first instance the same d falls at two different levels in the notational field, while the c is notated at a higher level than the first d. In the second instance three successive fs are placed at three different levels in the notational field, whereas an a is written at a level below that of a preceding f, etc. Of course we can try taking the notation of 1120 at face value (on the assumption that it really means to represent intervals through the diastematic placement of the neumes). Then we would get something like this reading of the second passage in 909:
8
le
gi
o
ni
bus
But on the grounds of our knowledge of the musical syntax of the style, gained from notated repertories and from the models notated in tonaries, we would reject that reading, especially in view of the cadential position of the passage. That can mean only that our assumption about the intention to represent intervals accurately is not warranted. Or we might reason that the intention was there, but its execution was prevented by carelessness or incompetence. Now it is easy enough to understand how a notator, working without a horizontal guideline, could lose his place and produce a meandering and—to us—unreliable score. Under such circumstances skill and care could make a difference. But when a notator with a fluent hand represents three successive notes of the same pitch with signs at three different levels in the notational field, it is simply not to be believed that the norms and expectations of his circumstances called for him to place the
Th e Transmission of Some Aq uitanian Tr ope s
257
three signs at the same level, but that he was careless or incompetent. We must accept both that complete consistency of vertical placement was not normative for him, and that the score that he produced was sufficient for his purposes. This last point is reinforced by the first example. The fact that the omitted passage ‘hodie’ was inserted into the margin, with neumes, shows that it was required for whatever purposes the manuscript was to serve. The fact that it was inserted with the neumes in the—to us—unreadable form that we find, shows that they were sufficient to those purposes in that form. What can we say about the state of the notation and its role in the transmission when these are the circumstances? Here another remark of Evans is to the point: ‘There is no evidence before the eleventh century of scribes especially qualified as music copyists, and it seems far more likely that, at least for the earlier tropers, a member of the choir—and probably the trope singer himself—notated the book for his own use.’¹⁰ The import of that remark, considered together with what has just been observed about the notation, is that in a substantial portion of the total sphere of trope performance in the tenth and eleventh centuries, even Aquitanian notation served only to support the singer in performances for which he had to rely on competences other than the ability to ‘read’ the notation (in the modern sense)—whether that means his memory, his ability to extemporize, or what is most likely, some fusion of the two. The conclusion is that such competences must have played a role in the performance, hence in the transmission, of tropes. This general conclusion, based on reflection about the condition of the notation, will be given concrete substance through the detailed analysis of the examples.¹¹ Example 11.1. I begin with a description of the correspondences between the versions of the Paris sources on one hand and that of Apt on the other. Both have the same trope text. In both, line I falls into two sections (Ia and Ib) with caesura on g, corresponding to the articulation of the text. Line II corresponds melodically to Ia in both cases, and line III corresponds to Ib. And in both cases line IV corresponds very closely to line III. This is so whether or not there is detailed melodic identity between the two settings of the respective lines. The similarities between the two versions lie in the internal correspondences within them. Something like the formal system of the trope has been transmitted. In addition, however, there are some correspondences of melodic detail at the surface, indicated by the horizontal brackets 1 and 2. So we can look at the transmission of this trope at the level of its formal arrangements and at the level of surface detail. This way of understanding the matter can be deepened by considering the formal effect of the troped chant as a whole. A two-phrase introduction (Ia–b) has ¹⁰ Evans, The Early Trope Repertory, n. 9. ¹¹ The examples do not necessarily show the entire trope as it is in each source represented. They are meant to show mainly those passages that are relevant to the discussion. The upper-case letters at the ends of the staves denote the lines of the Introit that follow (as in Weiss).
258
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.1. Trope Dulciter agnicole from Paris lat. 909, 1120, 1121, 1084, 1118, and Apt 17 Ia 909/1120/1121
8
1084
Dul
ci
ter
a
gni co
le
fes
ti
vum
dul
ci te
pas
cha
8
1118
8
Apt 17
8
Ib 909/1120/1121/1118 AB 8
1
iam
qui
a
Jor
da
nis me ru
i
stis
fon
te
re
nas ci
1084
8
887
8
Apt 17 A 1
8
II 909/1120/1121/1118 C 8
Ci vi
bus ae
the
re
is ma
nan
tes
man
na
de
cho
ris
887
8
1084
8
Apt 17 B 8
Vi vi
bus
Th e Transmission of Some Aq uitanian Tr ope s
259
Ex. 11.1. cont. III 909/1120/1121/1118/1084/887 D 8
2 Di
xe
rat
hoc Mo
y
ses
quon dam
Xpis
tus
iu
bet
at
nunc
887
8
Apt 17 CD 2
8
IV 909/1120/1121/1118 E 8
Ae ther, hu mus, pa ri
ter
pe la
gus
u
ni
ta
vo
ce
de
can tant
887 E 8
1084 E 8
Apt 17 E 8
Hae ter, hu mus, pe la
gus
u
ni
ta
vo
ce
re
sul tent
Introit Introduxit vos, transcribed from Paris 903, fo. 77 v A
8
In
tro
in
ter
et
ut
in
(h)o
al
le
du
xit
vos
Do
mi
nus
B
8
ram flu
en
tem lac et
mel
al
ni
sem
per
sit
le
lu
ia
C
8
lex
do
mi
D
8
re
ves
tro
E
8
lu
ia
al
le
lu
ia
260
Th e Transmission of Some Aq uitanian Tr ope s
the effect of an antecedent–consequent relationship. That is a matter mainly of the proportions or weight shifts in the two phrases as they balance on g and tilt down to d. In Ia the melody touches down on d near the end of the phrase, after an ascent. In Ib it tilts to d early, from a direct descent in the Aquitanian versions and after a moment of balancing on g in Apt. This is followed by the first chant line(s) which, especially in Paris, come(s) like a re-beginning. Then comes trope line II, another re-beginning in that it is modelled on Ia, followed by a phrase (III) corresponding to the consequent phrase of the introduction (Ib). From this point on the consequent phrase has the effect of a refrain, alternating with chant lines. That enhances the effect of its consequential quality, and it also makes some sense of the faithfulness of the transmission of melodic detail under the brackets 1 and 2. That is, I associate the stability in the transmission of this passage with the sharpness of its formal role. This impression is reinforced by an interesting detail. For Ib, 887 has a clivis over the second syllable which it does not repeat in III and IV, transmitting instead the emphatic syllabic opening that is in all the versions. The descent to d in Ib, III, and IV provides the strongest contrast to the lines of the introit, which are focused on g and its upper and lower neighbours. That is an important aspect of the formal idea of the piece, and it is transmitted in both Aquitania and Apt. There is an important difference between the two versions after the descent to d in the refrain. The Aquitanian version returns to a, the Apt version to g. This difference of detail reflects a more fundamental difference between the two versions. The trope lines in Aquitania articulate a d-pentachord with b as upper neighbour. In the form of the whole this plays into the contrast with the introit lines, whose g-plagal is articulated as two conjunct tetrachords pivoting on g. At the level of the tonal configuration, the trope lines of Apt are closer to the introit lines, and the contrast between alternating trope and introit lines is weaker than it is in the Aquitanian versions. Returning to the details at the surface, this difference is reflected in the fact that in the refrain line the Apt version opens with repeated gs (ornamented in Ib), while the Aquitanian versions consistently open b–a–g. Associated with that difference is the fact that the Aquitanian versions never rise above b, whereas the ceiling in Apt is c'. We may therefore speak of a single underlying idea about the way that the introit Introduxit vos is to be troped, but with a fundamental difference in the formal-tonal realm that determines differences in detail throughout. I suggest the concept of ‘homologous settings’ for such cases. The two settings are homologous with respect to their formal arrangements, and that states the relationship in the most general and unexceptionable way. We must then refine the statement of the relationship by describing the differences in the ways the overall movement is worked out, and the differences of surface detail that are contingent on the structural differences as well as those that are independent of them. I believe this is
261
Th e Transmission of Some Aq uitanian Tr ope s
preferable to forming simple judgements about whether the settings are ‘the same’ or ‘different’, for those terms cannot sufficiently describe musical transmission at this stage in the history of music. Consider the differences of tonal configuration in the versions of Aquitania and Apt in the light of modal affinities. (In the following I shall take office antiphons, given their modal assignments in medieval sources, as representatives of modal types.) The Aquitanian version of the trope begins in a way that is characteristic of first-mode melodies:
8
Dul
ci ter
a
gui
co
le
fes
ti
vum dul ci te pas
tu
am
cha
2. 8
po ten ti am
First-mode antiphons: 1. Per arma justitie; 2. Excita Domine, third and fourth words; 3. Excita Domine, beginning.
But the introit Introduxit vos, like others of its type, is a typical eighth-mode melody:
8
In
tro
du
xit
vox
Do
mi
nus
The Apt version of the trope also begins like a characteristic eighth-mode melody, and we may say that its consistency with the introit is a modal consistency:
8
dul
ci
te
pas
cha
Are we then to understand that the Aquitanian version entails a modal inconsistency, or even a deliberate modal contrast, between trope and chant? ‘Modal inconsistency’ would suggest a lack of schooling; ‘modal contrast’ would suggest a rather sophisticated idea, perhaps like a contrast of tonalities. Or is it that melodic turns and patterns characteristic of plainchants that were classified in different modes could be heard as musically consistent within the background of a single, more general, melodic idiom? This second interpretation would entail the idea of a melodic matrix defined by range, pitch collection, and preferred melodic interval progressions, within which more than one modal configuration could be differentiated. It would entail the understanding of strict modal classification as primarily a utilitarian matter, serving the linkages between psalm verses and refrains in the
262
Th e Transmission of Some Aq uitanian Tr ope s
several liturgical genres of plainchant. Modal differentiation depends on two quite different sorts of melodic detail: tag formulae—especially at beginnings and ends—and procedures for establishing the tonicity of a mode—through range articulation and cadences to the final. Such procedures develop two aspects of tonic function: centrism (the tonic as pivot of melodic balance) and final resolution. The means of modal differentiation partly control the progress of melodies, but not entirely. So that while modal identification is possible, common features remain between melodies of different modes. The idea that the first-mode character of the trope Dulciter agnicole in the Aquitanian version could be assimilated along with the eighth-mode introit Introduxit vos into a single coherent composition becomes more plausible if we consider as a kind of medium between them the first-mode antiphon Cantantibus organis (this is suggested only as a way of thinking ourselves into the sort of knowledge of melodic idiom that a tropist might have had; there is no implication intended that this antiphon had any concrete connection with either the trope or the introit).
8
Can
tan
ti
bus
or
ga
nis
can
ta
bat
Like the introit, this antiphon spans the range of a seventh, from d to c'. The d is a sort of root-tone or tone of origin in both cases, and its role is extended through a chain of thirds (d–f, f–a, a–c). Those thirds are the principal skips in both melodies. But in the antiphon it is the d that is tonicized at cadences, whereas in the introit it is the g. That is the sense in which two different modal configurations are differentiated with a common melodic idiom. It may be that in the composition of tropes the requirement of melodic coherence that we must assume could be satisfied at the level of such an underlying melodic idiom. In the case at hand, the Aquitanian version of the troped introit would be consistent as a whole, in that sense, while at the same time it progresses, by way of a formal principle, through an alternation of emphasis on d (itself and through a) and g. Such would be the meaning, under this second interpretation, of what might otherwise appear to be a modal inconsistency. But to think of this at all in terms of mode would be to refer it to a system that may have little relevance for the relation between a plainchant and its trope. In the Apt version the tropist appears to have pursued a greater modal consistency, or again we may say simply that he remained more narrowly in the melodic track of the introit. In any case the result is a composition that is less dynamic, and less shaped by an idea of the whole; it is more uniform, but less unified. The question that comes to mind is whether this sort of difference between the two versions is displayed consistently throughout the respective collections of tropes. That question calls for systematic study of
Th e Transmission of Some Aq uitanian Tr ope s
263
whole collections from such points of view. The pursuit of such a study promises to enhance our understanding of the background of melodic resources and systems against which tropes (and music in other new genres) were composed and transmitted. As for the modal system, the study of trope transmission from the viewpoint of modal affinities promises to enhance our understanding of modality in that time as a compositional parameter beyond its narrow regulating role for the plainchant. Example 11.2. This set comprises versions with more or less closely similar texts and melodies from Aquitanian and from two Italian sources (b–c), and two tropes of the same introit but with entirely different texts from an Aquitanian and a Provençal source (d and e). a–c comprise four trope lines, I–IV. I have identified three phases of I as P, Q , and R. Comparing versions a–c of I, what is most immediately striking is the divergence in the transmission of text and music. b and c have nearly identical texts, but it is versions a and b that have nearly identical melodies. This is so not only with respect to melodic detail, but especially with respect to critical points of phrase organization. This will be clear from a comparison of versions b and c of I P. b, like a, articulates as three parallel phrases (see the transcription), c articulates as two. In a and b the motion to the peak, a, is reserved for the middle of I Q , while in c the peak is reached in the second phrase of P. In all these respects version b is like a, although its text is identical with that of c. This fact must serve as a counterweight to the observation made in connection with Ex 11.1, that phrase organization reflects text articulation. Here we learn that text articulation is not the unique determinant of melodic process, which can be transmitted independently of text. At the same time, this example reinforces the conclusion from Ex. 11.1, that the question whether two melodies are ‘different’ or whether they are versions of a single melody must be answered through comparison of the underlying movement, as well as the surface features, of the melodies. Now something like this is suggested by a comparison of the a, b, and c versions of Ex. 11.1, on the one hand, with the d and e melodies, on the other. The three phases (PQR) are evident in both groups despite the different texts. And in both, the formal-tonal process is similar, with P rising from and returning to d (or its lower neighbour c); Q , the peak phase, rising from f to a (or its upper neighbour); and the cadential phase R, falling back to d. In d and e the Q phase, in response to the greater length of the text, is reiterative. But its beginning from the pitch f and its climactic role in the melodic procedure are nevertheless clear. This evidently points to a melodic type that transcends individual texts and that is a factor in the transmission. The identification of a special melodic type here would not be sensible if all tropes of Etenim sederunt followed it. But of fourteen tropes of that introit, only the five shown here are of this type.
264
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.2. Trope Hodie Stephanus martir from Paris lat. 1871, 1240, 1119, Apt 17, Bologna 2824, and Modena 1.7 I a1 1871
8
a2 1240
Ho
di
e
Ste
pha
nus
mar
tir
ce los
as
cen
dit
(?) 8
b Bologna
8 Ho c Modena
di
e
in
cli
tus
8 Iu e Apt 17
ba
re
co
rus
8
O
quam pre ti
mar
tir
Ste
pha
nus
8
d 1119
co
o
sa
est
a1 Q
R AB
8
a2
quem pro
phe
ta
du
dum
in
tu
ens
e
ius
vo
ce
di
ce
bat AB
8
b AB 8
c
pa
ra
di
sum lau
re
a
tum
as
cen
dit AB
8
d
8
(d)
pa
ra
di
su
bli
mes
sum
le
le
vi
ta
Ste
8
be
e
8
(e)
8
tus
per
fec
ta
glo
de
quo
ri
a
o
pha
a
as
nus
cen
ad
he
tum ho
qua co
ro
na
lim pro phe ta
ce
vit
ci
dit
the ra
fer
te
re
to
nan
te
Ste
pha num
nit
di cens
265
Th e Transmission of Some Aq uitanian Tr ope s Ex. 11.2. cont. II a1
8
b
In
sur
re
xe
runt
con
tra
me
8
(a1 ) C 8
(b)
iu
de
o
rum
po
pu
li
i
ni
qui C
8
III a1 D 8
b
In
vi
di
o
se la
pi
di
bus op pres
se
runt
me. D
8
(A)
(A')
(B)
IV a1 EF 8
b
Su sci pe
me
um
in
pa
ce
spi
ri
tum. EF
8
In view of the differences in the underlying syntax of I P in versions a and b on the one hand and c on the other (see the discussion of those passages above), and in view of the relationship just noted between versions a–c and d–e of I as a whole, these conclusions seem warranted: 1. I a–e represent a single general melodic type. 2. Ia and b are homologous with respect to syntax and of such a similarity in their surface features that nothing stands in the way of our regarding them as versions of the same melody. Id and e, on the other hand, must be regarded as independent executions of the same melodic type. As for Ic, despite its belonging to the text group of a and b, its divergence from those two in the underlying structure of I P suggests that it is an independent production, like d and e, rather than a version of the a–b group (thus we might have aligned the phrase over ‘Stephanus’ in I Pc with the beginning of I Qd and e). This is an important conclusion for our
266
Th e Transmission of Some Aq uitanian Tr ope s
concept of transmission, for it opens the possibility that similarities between two melodies may be accounted for on the grounds of similarities of text and common melodic types, rather than being understood in the sense of the derivation of one melody from the other. We require a reordering in our understanding of the nature of family relationships in such traditions. I note that at the opening of Q , which is perhaps the strongest moment in the rhetoric of the form, the transmission throughout is most consistent at the level of melodic detail. Like the passages under brackets in Ex. 11.1, it is a bit of concrete formulaic material at a formally critical point, transmitted nearly intact. III tends to proceed in a sort of bar form: AA' (antecedent–consequent) B. This is sharper in the versions 779, 903, and Bologna because of the identical openings of A and A' there. A similar situation obtains in IV. It comprises two phrases, and in Bologna those are integrated by virtue of the fact that the second is an extended variant of the first. This suggests a tendency in Bologna towards symmetrical and integrated forms—something that will require confirmation by further study— and it recalls the suggestion of a special tonal orientation reflected in Apt. The notion of stylization principles characteristic of particular collections and constituting local vector forces in the determination of transmissions deserves full exploration. In this instance, as in the first example, the stylization functions at the level of the underlying structure of the piece and influences the appearance at the surface. There remains one last observation to be made about this example. I, as transmitted in 1240, exemplifies the tendency for that source to set somewhat fewer notes per syllable, although that tendency is nowhere so extreme that we would characterize the music of the later sources as florid by comparison. Evans, reasoning from this general feature of 1240 and from its earlier date, concludes that it is the bearer of bare archetypes of which the later sources present elaborations.¹² But alternatively, it might be that relative bareness of line represents a principle of stylization for 1240. Example 11.3. This group comprises two Aquitanian versions of a trope with more or less similar text and melody (a and b), a northern French version (c), and an Italian version (d). For purposes of comparison I have included Aquitanian versions of two tropes of the same introit, with different texts (e and f ). There are two trope lines (I and II), and I have marked three phases of line I: M, N, and O. Consider I. At a quick glance it will strike everyone that in a–d we have essentially the same melody. What is difficult is to find ways of describing the similarities and differences that a closer examination will bring into focus. That is because this line is unlike the ones in Exs. 11.1 and 2 in that it lacks any obvious and ¹² Paul Evans, ‘Northern French Elements in an Early Aquitanian Troper’, in Speculum musicae artis: Festgabe für Heinrich Husmann zum 60. Geburtstag am 16. Dezember 1968 (Munich, 1970), 103–10.
267
Th e Transmission of Some Aq uitanian Tr ope s Ex. 11.3. Trope Filius ecce patrem from Paris lat. 1871, 887, 1235, and Benevento VI. 34 I a 1871, fo. 40 v M 8
b 887
Fi
li
N
us
ec ce
pa
trem
con
pel
li
cis
ve
ne
ran
dre
am
su
per
as
8
c 1235
8
d Benevento
8
e 1871, fo. 37 v
8
[A] pos to f 1871, fo. 37 v
8
[C]ul men a
a
pos
to
li cum An
tra be
a tus in
sig
nis
O A-C
8
lans
b
ta
li ter in
quit A-C
8
c A-C 8
d A-C 8
e AB 8
dis
f
me los
clan ga
mus
ho
no ri bus AB
8
re
ti net, nos in de ca na
mus
om
nes
268
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.3. cont. II a DE 8
b
Af
fec tu
que
pi o
re pe
tens dic
tan
do
mi ni
strat DE
8
c DE 8
d DE 8
e C-E 8
f
A
mic ti
ho
Mi
ro lu
stran
no re
pre da
to[s] C
8
tes va ri
o
que
vi bra
mi ne cos
mum
sharply defined form. Instead it continues throughout to be balanced on d, almost in the manner of an embellished recitative. In this it resembles the introit itself. Now given just that much of a constraint, and given that it is always the same text that is set and in the same mode, it would be difficult to avoid having any number of versions resemble one another as closely as these do. But in fact there is quite a lot of variation, given the narrow range of possibilities. We might look at the melodies in this example in terms of a number of note groups or modules that we can crystallize from them. I believe we can identify five. I say ‘groups’ or ‘modules’ because there is no fixed way in which the melody moves through their pitches (in this sense it would not be good to call them ‘melodic formulae’). Of course in the actual context of a melody there are no sharp boundaries dividing these groups. They are merely a way of focusing the melodic features of the melodies for purposes of comparison. Here are the proposed groups: 1. The final with its upper and lower neighbours:
8
2. An upward thrust from the final to g, then a descent to f and the final:
8
Th e Transmission of Some Aq uitanian Tr ope s
269
3. A motion to the high a, then a descent through g or f to the final:
8
4. A thrust from c to f, then a descent to the final:
8
5. An upward motion through the lower tetrachord, from A through c to the final:
8
Within the very narrow range of possibilities afforded by the idiom of this trope, the differences between these modules constitute as much contrast as is available. Now to locate them in the melodies: IMa exhibits module 1. The same phrase in b is mainly an elaboration of 4. In c the same phrase exhibits first 1, then 2, then 4. Version d of this phrase elaborates first 2, then 4. INa exhibits 2, while versions b–d of that phrase exhibit 3. 10 exhibits first 5, then 4 in a, 4 with an embedded 3 in b and c, and 4 alone in d. The first phrase of II is an elaboration of 2 in a; it involves both 5 and 4 in b, and 4 alone in d. On close inspection versions that appear at first to be all alike turn out to exhibit the limit of divergence, given the constraints of the idiom. What, then, is the nature of the transmission here? My suggestion is that it is not a transmission of the fixed trope Filius ecce patrem as a whole, but a reconstruction of it based upon the following set of données: the articulation of the text and the consequent articulation of its musical setting; the modular basis of the melody; and certain syntactical principles, that final phrases must ultimately be dominated by 3, that 1 is reserved for beginning phrases, that 5 is reserved for positions at or near the end of the trope line, etc. Once the variation among versions a–d is recognized we can see that tropes e and f operate under the same set of constraints as these, and that the texts are sufficient to account for their greater divergence from a–d than the variation within that group. To account for the transmission of the trope Filius ecce patrem, the conception that its versions are all actualizations of a matrix defined by the constraints that have been described here is preferable to the conception that the versions are related as variants of one another or of some hypothetical archetype. The latter conception suffers from three major weaknesses. First, it has not been possible to identify archetypes or gradients of variance within the existing versions of tropes, on either musical or philological grounds. Second, under this conception it is necessary to posit missing sources in such abundance as to raise suspicion. Third,
270
Th e Transmission of Some Aq uitanian Tr ope s
there is no explanation for such a wide range of variance as is exemplified by the transmission of Filius ecce patrem, whereas the premiss that a fixed melody has been transmitted would seem to call for such an explanation (the more so as the transmission is written). The conception that the versions are all actualizations of an underlying model has four advantages. First, it is of a greater generality. Second, it provides a more adequate account of the observable facts of the transmission (or to put it the other way around, it does not require us to posit masses of irretrievable evidence). Third, it is consonant with a general conception that is proving to be more satisfactory to account for the transmission of other repertories of medieval melody and polyphony.¹³ Fourth, it offers a category for assimilating the facts of variation, other than the stemma that no one has yet been able to reconstruct. Example 11.4. The transmission here involves two widely differing settings of the same trope text (1871 and 887), and several versions closely resembling that of 1871. 1240 corresponds closely to 887. I shall take up first the question of the levels and degrees of difference between 1871 and 887, then I shall consider some of the variants in the 1871 group. I show passages from the other sources in that group only where there are observations to be made about them. Consider I (phrases are marked M–R). The text articulation and the phrase division of the versions of 1871 and 887 are identical through IO. They diverge at P, in that 887 makes no further phrase or text articulation to the end of the line. But with the text articulation and phrase division through IO the musical similarity between the two versions ends, for their intonations of the text take on quite different formal-tonal designs. These can be schematically described thus: 1871
A (M) B (N) C (O) C (PQ ) D (R)
an arch from d to f and back to d an arch from f to c and back to f the climactic phrase, rising to a and falling to d essentially a repetition of the preceding phrase the cadential phrase
887
A (M) B (N) B'(O)
an arch from d to g and down to c motion from f to a the motion of the preceding, then falling to d. (That creates between B and B' an antecedent–consequent relationship. The essential melodic outline is like that of 1871 O, but the role of the phrase in the form is different in each version.) an arch from d to a and down to c. This associates with the first phrase, as a sort of double statement of it. It remains tonally open, and it is closed only by the line ‘quia . . . illi’,
A' (P)
¹³ I refer to the conception developed in Chs. 4, 6, and 7.
271
Th e Transmission of Some Aq uitanian Tr ope s Ex. 11.4. Trope Adveniente Xristo from Paris lat. 1871, 1118, 887, 1119, 909, and 1121 I 1871
8
M
N
Ad
ve ni
en
te
Xri
a
do
sto
stel
la
mag
na
vi
sa
est
1118
8
887
8
1871
8
O et
a
ma
gis
ra tus
est.
887
8
1871
P
Q
R AB
8
Ho 1118
di
e
in
lor
da
ne
bap ti
za
tus est.
Qui
a
AB 8
1119 AB 8
909 AB 8
1121 AB 8
887
8
887 AB 8
Can ta te
et psal li te di cen
tes
il
li
272
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.4. cont. II 1871 C 8
Rex re
gum i
pse reg nans in
tro
no
pa
tris
su
i
su
i
1119
8
1084 C 8
(tro )
no
pa
tris
887 C 8
pa
tris
su
i
III 1871 DE 8
Quod
per
ma
net
nunc
et
in ae ter
num.
1084* DE 8
1121 DE 8
887 DE 8
Po
tes
tas e
ius a
ma ri us
que ad ma re
et ad fi
nes or
bis
ter re.
* The vertical placement of the neumes here is especially vague.
which we may designate in the form as C. This passage is internal to the form in 887 (i.e. the form requires it for closure), not so in 1871. The two versions of lines II and III, respectively, exhibit differences on a similar scale, and we may take this pair as exemplifying a single text in two independent settings. (Of course in line III 887 has an altogether different text.) With respect to the transmission of the version represented in 1871 and its homologues there are a few places of particular interest. The setting of ‘baptizatus est’ at the conclusion of line I shows nearly identical neume patterns in all versions, but apparently not always identical pitches. Of course that may reflect a problem of diastematy; however, 1871, 1118, 909, and 1121 are generally quite con-
Th e Transmission of Some Aq uitanian Tr ope s
273
sistent about diastematy; and the apparent reading of 1119 is consistent with that of 909. A similar situation obtains between the versions of 1871 and 1121 in the setting of ‘in aeternum’ (line III). Such uniformity of neume pattern in the face of a variability of pitch suggests something about the nature of writing down. It is that the primary task for the notator was to set down a pattern of neumes to go with a sequence of text syllables. We have a different sort of reinforcement of that suggestion when two sources exhibit more or less identical neume patterns for the same text passage, but are out of alignment by one syllable (examples are phrases IN in Ex. 11.4, the versions of 1871 and 1118, and phrase IIa in Ex. 11.6 below, the versions of 1871 and 909). I shall cite further evidence for this idea in the discussion of Ex. 11.5, and I shall return to it again at the conclusion, in connection with the question of the role of notation in the transmission of the tropes. This sort of pitch variation seems to be of a different sort than variation at the structural level. At the beginning of line III in Ex. 11.4 the settings of ‘Quod permanet nunc . . .’ in the versions of 1871 and 1121 exhibit differences that reach down to the tonal structure of the melody. 1871 opens with the chain d–f–g–a, 1121 with the chain a–c–d–f. The cadences are parallel, but a fifth apart. All versions come back together at the melisma on ‘et . . .’. Return to the cadence on ‘in aeternum’ at the end of line III. Evans calls attention to the melodic identity between the 1121 version of that cadence and the cadence of the phrase ‘in manu ejus’ of the introit. He calls this a ‘direct quotation’, and interprets it in the light of what he regards as a late tendency in the history of trope composition to quote in the lines of the tropes passages from the chant that is being troped.¹⁴ I must pause over this matter, for it raises questions not only about the interpretation of such a passage but also about the nature of trope composition, and even broader questions about the posing of such questions in the study of this subject. Notice has already been taken that 1121 and 1871 are at variance over this cadence. We could associate the version of 1871 with the cadence of the introit phrase ‘et imperium’ and say that in this case the chant cadence is anticipated. But one begins to get a sense of the arbitrariness of such identifications. Along the same lines, Evans interprets the cadence of trope line II (‘sui’) as a quotation of the cadence of the introit phrase ‘(Dominator do)minus’. To substantiate his interpretation that such quotation is a late development he shows that cadence in the version of 1240:
8
su
i
¹⁴ Evans, The Early Trope Repertory, 92.
274
Th e Transmission of Some Aq uitanian Tr ope s
But the 1240 setting of this trope text corresponds to that of 887, and has nothing whatever to do with the version of 1121. There is no reason to think that the latter is a recomposition of the 1240 cadence. They simply belong to two independent melodies. In the ‘quia’ passage at the conclusion of line I¹⁵ Evans finds another ‘quotation’, this time from the formula for the second mode that is given in the tonary of 1121:
8
No
e
a
gis
It is his interpretation that the passage, being external to the form of the trope line (see above), has been added as an intonation for the beginning of the chant, and that for this purpose a modal formula was used. He appeals once more for substantiation to the appearance of the passage in 1240, which he reads thus:
8
qui
a
He regards the version of 1121 as an alteration of this, hence an addition from outside. But again the version of 1240 corresponds to an entirely different tradition, and there is no reason to think of the ‘standard’ Aquitanian version (1121, 1871, etc.) as representing a later stage of the same thing. What is more, we note that in 887 the ‘quia’ passage is internal to the form of the trope line, and that is the version to which 1240 corresponds. If the passage belonged in the earliest version, there is no reason to think that it was added later. It is even possible to think of it as internal to the form of the ‘standard’ version, for it is quite similar to the opening of trope line I (the ‘quia’ passage is in any case not an exact duplicate of the modal formula). Then the form becomes ABCCDA'. The point about this is that, instead of ‘quotation’, we may see here simply the use of a traditional resource of the mode. Here are the opening and cadence of the first line of ‘Psallite’, in Ex. 11.7:
8
and the cadence of ‘Iubilent’, line I, in the same example:
8
The modal formulae were paradigms for composition, and we shall find their flavour in melodies of corresponding modes without end. But to adapt a model is ¹⁵ Introitus-Tropen, i, ed. Weiss, has not reported the transmission of this passage accurately. The following have the word ‘quia’ with melody as transcribed from 1871: 1871, 1084, 1119, 909, 887, 1118. 1240 has ‘baptizatus est ipsi gloria quia Ecce.’ 887 has ‘baptizatus est cantate et psallite dicentes illi Ecce’ with melody as shown in the transcription.
Th e Transmission of Some Aq uitanian Tr ope s
275
TABLE 11.1. Trope verse in four sources Paris BNF lat.
Gau- de-
a- mus ho-
di- e qui-
a
De-
us
de- scen- dit
1. 1240
2. 13252 3. n.a. 1235
4. 1121
not at all the same thing as to quote a source. (Another instance is the relation between Ex. 11.2, line III and the mode 1 formula in the tonary of 1121.) This issue comes down to a fundamental question about the nature of composition in this time and in this genre. I wonder whether ‘quotation’ is an appropriate concept for this material, whether what we are seeing is not the use and reuse of conventional melodic resources where there is no reason to think that the occurrence in one place signals a reference to the occurrence in another. A concept associated with ‘quotation’ under this same general viewpoint is that of ‘elaboration’. Evans cites what he interprets as a later elaboration of an earlier, simpler cadence (see Table 11.1).¹⁶ Comparing the settings of ‘hodie’ in 1240 and 1235 on one hand and 1121 on the other, Evans interprets the latter as ‘a more ornate version of the same cadential formula’. The difficulty with this interpretation is that even in isolation it would be stretching a point to read this:
8
as a more ornate version of this:
8
In context there is really no reason to do so at all. The passage as a whole in 1121 does not correspond either to that in 1235 or to that in 1240, and there is no reason to regard one as a transformation of the other. The more general question is whether such concepts as ‘elaboration’ and ‘quotation’ are at all apt for such a tradition. The problem of the propriety of those concepts for this tradition is closely ¹⁶ ‘Northern French Elements’, 108.
276
Th e Transmission of Some Aq uitanian Tr ope s
parallel to the problem of the ‘cento’ concept for plainchant composition and transmission (see Ch. 7). Example 11.5. In the transmission of the third trope line we are confronted again by the question, how can we compare the several versions? In what terms, with reference to what features, can we describe similarities and differences? As an approach to these questions I have marked five phrases (a–e) and made syllabic reductions of the versions of 1871 and 1119: 1119
8
1871
b
a
In Ihor da
c
a
ne
Iho
an
ne
bap
ti
za
tus
pa
ter
na
vo
ce
8
1119
8
1871
d
fi
e
li
us
est
pro
tes
ta
tus
cu
ius
ho
nor.
d
8
Decisions about where to place the phrase markers and what notes to include in the reductions are somewhat arbitrary, but in the main I believe the range of choices is not wide, especially when we look at the melodies in terms of their essential underlying motions and modal components. The reductions show a basic sequence common to both versions, but such a degree of difference in particulars as to suggest two independent actualizations of that structure. The common features are these: an opening phrase, a, articulating the lower tetrachord; a second phrase, b, rising to f and descending to linger about d; a third, peak phrase c, rising from d to a; a fourth phrase, d, to which I shall return in a moment; and a fifth, cadential phrase e, elaborating the tones d–f–e–d. At the level of the underlying structure the fourth phrase, d, has an ill-defined function—some sort of extension to accommodate the text before the cadential phrase—and the two d phrases have no melodic features in common. Because of the difference with respect to the point in the text declamation at which the climactic a is reached, the articulation between the c and d phrases comes at different moments. At the surface the two melodies share very little detail. The trope line ‘Per quam Christicole . . .’ of 1871, folio 38v shows the overall motion that is common to both versions of ‘In Ihordane . . .’. That suggests that at this level there is a melodic type at the root of the transmission, whose options have been exercised in the several ways described. And since the same melodic
Th e Transmission of Some Aq uitanian Tr ope s
277
Ex. 11.5. Trope In Ihordane, from Paris lat. 1119, 1121,1084,909, 887,1118,903,1871,779, and Apt 17 1119
8
1121
a
In
b
Ihor
da
ne
a
Iho
an
ne
bap
ti
za
tus
za
to
a
Io
han
ne
idem 8
1084 idem 8
909
8
887
8
1118
8
903
8
bap
1871, fo. 9 v
ti
8
779
8
Apt 17
8
1871, fo. 38 v
8
Per
quam
Xpisti
co
le
type is chosen to intone the tropes of different introits of the second mode, we can see the characteristics of the introit itself as a factor influencing the trope melody at the deep level of the type more significantly than in the details nearer the surface. This is the more so in view of the degree of variation to be found among the sources for ‘In Ihordane . . .’.
278
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.5. cont. 1119
8
c
pa
d
ter
na
vo
ce
fi
li
us
est
pro
tes
ta
de
i
est
ho di
tus
1121
8
1084
8
909
8
887
8
1118
8
903
8
1871
8
779
8
Apt 17
8
1871, fo. 38 v
8
tri
um
phan
tes
su
per
e
a
The sources for ‘In Ihordane . . .’ cluster in two groups, following the two versions revealed by the reductions. This is so in a general sense, with exceptions to be noted separately. The clusters are these: 1119, 1121, 1084, 909, 887, 1118, 903; 1871,779, Apt 17. In discussing the exceptions I should like to reserve phrases d and e for a moment. The b phrase in 1118 and 903 is independent of both versions, but its function (rising to f and descending to linger around d ) is like that of both versions. The c phrase in 1118 begins like the others in its cluster but ends f instead
279
Th e Transmission of Some Aq uitanian Tr ope s Ex. 11.5. cont. 1119
8
1121
e
cu
Introit:
ius
ho
cla
ma
nor. De us
ju
di
ci
um
tu
um Re
gi
da
8
1084
8
909
8
887
8
1118
8
903
8
1871
8
779
8
Apt 17
8
ad 1871, fo. 38 v
8
scan
tus.
dunt.
of d, and that makes a striking structural difference. The c phrase in 903 is quite unique. It begins somewhere between the two basic versions, and takes the peak tone, a, on a weak syllable, thereby weakening the climax and obscuring the articulation between the c and d phrases. This is a difference in the way that is chosen to declaim the text, and the result is a difference in the structure and articulation of phrases. In the 1871 cluster, at the level of the reduction 779 and Apt 17 oppose the pitches
280
Th e Transmission of Some Aq uitanian Tr ope s
d–e–d–c to the e–d–d–c of 1871 in the setting of ‘baptizatus’ in phrase b. In Apt 17 phrase c is the beginning of a departure that carries through phrase d, where there appears to be either a scribal aberration or the protocol of some rather fancy singing. The setting of ‘filius’, at the level of the reduction, opposes a–f–e–d to the a–g–f of the other sources, and phrase d, with its unique text, takes off in a vocalise that covers a range of a tenth and suggests a leap of a seventh between phrases d and e. The transmission of the d phrase in the 1119 group affords an insight, again, into the nature of the notators’ task—in particular the primacy for that task of setting down a pattern of neumes to go with a sequence of syllables (see above). As for the pitches, the transmission shows not that they were less important, but that they were less fixed than the sequence of syllables and the pattern of neumes. Of course the apparent differences in the graphic signs with respect to pitch may reflect a vagueness in the notation (i.e. in the diastematy) as much as it reflects differences in scribal intentions, or in the versions of melodies that are meant to be represented. But we have no way of confidently assessing these alternatives, and that fact alone points to some fundamental characteristics of this notation as a system of representation for this music that I shall try to define carefully at the conclusion. In any case, the transmission of this phrase suggests that if we regard the notation as a pitch notation then it is too precise for the state of the music. It is as though a sharp pen were used to outline what is intended to be a charcoal drawing. In these circumstances what significance can we attribute to the apparent identity of pitch in the versions of 1119, 887, and 1118? Following philological tradition we might say that they represent a central tradition, of which the other versions are variants. But given the nature of the transmission, and in view of the fact that it is not possible to establish a filiation for these sources, that interpretation has nothing but habit to support it. It may be that the three sources are related through copying, but it may also be simply that phrase d was thrice written down with apparently the same pitches. That seems all the more likely because there is no identity among the three sources in the other four phrases. The transmission of the e phrase in the first six versions of the example again demonstrates the primacy of an association between a sequence of syllables and a pattern of neumes—whatever the pitches. Taking pitch into account we see three plausible ways to make the cadence: that of 1119, 1121, and 909; that of 887; and that of 1118. (Note that the clustering of the first three is different from the clustering of the d phrase, and that again makes copying unlikely as the cause of the similarity.) The version apparently shown by 1084 is doubtful; it expresses the mode poorly and it makes a poor connection with the next line of the introit (shown after the version of 1119). It is presumably a scribal error, or a consequence of a vague notational practice. But if 1084 does not mean what it seems to indicate notationally, what does it mean? Something like 1119, with the pitch level slipped down? Or something like 887, with a b instead of an a in the penultimate neume, and a
Th e Transmission of Some Aq uitanian Tr ope s
281
reversal in the final neume? We cannot tell. And that fact itself suggests again that the notators were not setting down a piece they knew as finished; rather, ‘finishing’ the melody required an act of setting it down, or of performing it. These observations will be useful in formulating a general view of the nature and role of notation later on. Example 11.6. The discussion focuses on trope line III. (Lines I and II are included in the example without comment here.) Weiss prints two versions (887 and 1871), and indeed for the greater part of this line there are essentially two versions. I show, however, also the last phrase of 909, which otherwise concords with the version of 1871. The text is articulated in five phrases in all versions (marked a–e). The corresponding phrases of melody are all versions of four basic gestures, all from the idiom of the plagal d-mode: the opening motion from the low c to the tonic d; the expansion of that motion, reaching up to f and descending again; the climactic move from f to the peak, a, and the return to f ; and the cadential descent from f. A close comparison of the three versions will be instructive. 887 moves at once to f in the first phrase, and consequently to a in the second. The other versions proceed more gradually, making the melodic climax only in the third phrase at ‘stolam/ florem’. 887, having reached a climax in the second phrase, has little else to do but begin again in the third, and then yet again along with the others, in the fourth phrase. For 1871/909 that re-beginning makes a rounded form of the whole; in 887 it has become rather repetitious. The fifth phrase is heard as a kind of ‘Nachsatz’ in all versions. Of special interest in phrase d is the identity of the opening in all versions, and then the separation of the three at ‘celesti’. 1871 and 909 follow exactly the same course of notes to the end of the phrase, but in a different distribution in relation to the syllables of the text. In comparing 887 with 909, we might think that an identical melodic course is intended but that there was a problem about the diastematy. I have, however, transcribed both as they appear in the manuscripts, and both are plausible as they appear. And in any case a comparison of the two versions as a whole will not support the idea of an intended identical copy. We have here an irresolvable uncertainty whether the difference between the two versions is a product of writing or whether it represents a real difference in the object. There is no question but that the e phrases are entirely independent of one another. Each of the three goal tones that they choose makes a satisfactory linkage with the beginning of the next line of the introit antiphon, so there need be no suspicions of error here. These differences are understandable only in the sense of a transmission that involves repeated reconstruction on the basis of common underlying ideas about what the piece entails. There are common ideas about the material and about the course of the melody, but those ideas are carried out, in a sense, on a different time plan. One could be tempted to say, simply, that 1871 and 909 are virtually identical
282
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.6. Trope Filii karissimi/clarissimi domino, from Paris lat. 1871, 909, 887, and 1119 I 1871
8
Fi
li
i
ka
ris
si
mi
do
mi
no
cla
ris
si
mi
do
mi
na
ka
ris
si
mi
cla
ris
si
mi
me
los
909
8
887
8
1119
8
1871 AB 8
pan
gi
te
u
na
vo
ce
di
cen
tes
909 AB 8
887 AB 8
1119 AB 8
un
gi
te
cla
man
tes
II 1871 C 8
Fe
ci
sti
lau
da
re
no
men
tu
um
909 C 8
887 C 8
1119
8
Th e Transmission of Some Aq uitanian Tr ope s
283
Ex. 11.6. cont. III 1871
8
a
b
Tri um
phan tes
de
c
os
te
im
pe
re
o
sto
lam
flo
rem
flo
rem
ae
ter
nae
909
8
887
8
ho
te
1871
8
d
Vir
gi
ni
ta
tis
e
os
in
ce
le
sti
glo
ri
a
909
8
887
8
1871
Introit:
e D
8
sunt
se
cu
ti
prop ter
in
i
mi
cos
tu
os.
909 D 8
su
sce
pi
sti
887 D 8
su
sce
pi
sti
and 887 is a different setting altogether. But the difference between 1871 and 909 in phrase d is, on a smaller scale, of quite the same sort as the difference between those two and 887 overall: the same melodic course, run on a different time scale. Example 11.7. This pair is introduced as a sort of control. In the main I have been following multiple settings, or versions of settings, of the same trope text. I believe there has emerged very clearly the fact of a determining role for the text articulation and, corresponding to that, the formal-tonal structure of the melody in the
284
Th e Transmission of Some Aq uitanian Tr ope s
transmission of tropes. But in the discussion of Ex. 11.2 it was suggested that melodic types or models function in the transmission of tropes, sometimes independently of text. This is demonstrated most persuasively by the settings of lines III of the tropes Iubilent and Psallite, both for the same introit and the same feast. The consistencies are not only at the melodic surface but, despite the text differences, also in the phrase articulation and syntax. Evans has called attention to the parallels and similarities between the two melodies.¹⁷ I shall use a shorthand to summarize all of the correspondences that he finds. Lower-case letters refer to Iubilent, arabic numerals to Psallite. I shall use Evans’s language in order to bring out his aperception of the similarities: a ‘appears three times in “Psallite omnes” ’ (1–3). Despite the similarity in the settings of ‘ideo’, ‘“Iubilent” has adapted the melody to agree with the chant cadence, [he refers to ‘judicium’ at the end of the third line of the introit]
8
whereas in “Psallite” it retains its original form [a-3].’ Correspondence is claimed, further, between b and 4, and between c on the one hand and 5 and 6 on the other: The whole of line III can be considered as a paraphrase of the third line of ‘Iubilent’: . . . ‘omnium’ is set to the melody of ‘ipse vero’, ‘corda’ corresponds to ‘illi reddens’, and ‘sequens precepta ideo’ quotes literally the melody of ‘sua promissa ideo’. . . . The melodic material of one trope could be transferred to another trope in the same mode.
This language, especially the expressions that I have italicized, raises again the questions about the status of such surface similarities and what they show about the nature of the transmission. Evans’s language reflects the viewpoint in which a similarity is evidence of a literal quotation from a fixed original, and by implication, a citation of that original. And if it is not a literal quotation, it is nevertheless a deliberate transformation of that fixed original, as in the case of the ‘adaptation’ of the cadence at the end of line III of Iubilent to match the chant cadence—i.e. to ‘quote’ a different original—or as in the cases of ‘ornamentation’ cited earlier. Some questions can be raised about such interpretations. First, the material said to be adapted from one trope to the other is not unique to those two tropes. a-1, b-4, c-5 and 6 can be found in other sixth-mode tropes as well as in sixthmode compositions of other genres. Second, there is something curious about the claim that one trope has been quoted from another in such random fashion. How would it happen that a composer would take for the second phrase of the first line of one trope (Iubilent) the third and seventh phrases of the second line of another (Psallite; i.e. c-5 and 6)? And what evidence is there about which is the original? ¹⁷ Evans, The Early Trope Repertory, 112 ff.
285
Th e Transmission of Some Aq uitanian Tr ope s Ex. 11.7. Trope Iubilent omnes, from Paris lat. 1121 I 1121
8
c Iu
bi
lent
om
nes
fi
de
les
ca
a
ter
va
tim
na
rum
II A
8
de
pro
mant
psal
len
tes
ci
So
us
su
per
b 8
vir
tu tum
ef
fe
ctus
at
que
III d
e
an
ge li cis cho ris
BC 8
con
iun
ctus
est
Tri
sti
se
quens
ve
sti
gi
a
g DE 8
ip
I 1121
8
se
ve
ro
il
li
red dens su
a pro mis
sa
i
de
1
Psal
o
2
li
te
om nes
hu ius
san cti
II
lau dam cum
4
A 8
psal
mi
sta
di
Ho di
cen tes
e
con
ci
vis
5 BC 8
ef
III
7
fe ctus est
cae li
8
co
la 9
rum le 10
gi
o
ni
bus
3 DE
8
Scru ta
to rem om
ni um
cor da se
quens pre ce pta i
de
o
(One of Evans’s suggestions does seem plausible: the correspondence between ‘sequens precepta ideo’ and ‘sua promisse ideo’ in the third lines of the two tropes. We know instances in plainchant—some reported in Chapter 6—where the identity of a single word has served as a switch into a different melodic context.) But otherwise what significance would there be in the citation of one trope in another? And how can a citation be claimed if the melodic turn in question is common to the modal idiom? Of course the simpler alternative would be to see these similarities in the light of the exploitation under similar circumstances of common
286
Th e Transmission of Some Aq uitanian Tr ope s
resources. What those resources are is clear enough. They are the melodic vocabulary of an elaborate psalmody of the sixth mode, the same style in which the Introit to which both melodies belong is made. There is no more reason to think that one trope cites the other than that either cites a responsory of the sixth mode.
II It is time to step back and reflect in general terms on the sorts of difference of interpretation that have arisen during the course of these observations. Example 11.1 brings out the question whether two settings of the same text are ‘different’ entities or ‘versions’ of the same entity. They are ‘different’ if one thinks of tropes as original, fixed compositions. They are ‘versions’ if one thinks of the notated score of a trope as some sort of concretization of a well-formed general idea about the melodic setting of a particular text adapted to a particular liturgical occasion, and of the versions as homologues. The difference comes down to viewpoints about what ‘composition’ and ‘transmission’ were, and these viewpoints control the interpretation of particular evidence. Example 11.7 brings out the question whether melodically similar or identical turns in different settings may be understood as quotations. Answers reflect the same difference of viewpoints about the nature of transmission: the transmission of fixed melodies going back to composed originals, or the repeated actualization or reconstruction of a trope based on a matrix of organizing principles, melodic details that may be shared by different models, and ideas about articulating a text. Example 11.4 brings out the association between the concept of ‘quotation’, reflecting the former viewpoint about composition, and the hypothesis about chant quotation as a late tendency. This hypothesis has as part of its theoretical background the paradigmatic precept that the objective of the research is to ascertain the line of style development in a tradition of tropes regarded as closed, fixed works. That historiographic imperative has had great influence over the interpretation of sources in this field. That Weiss, in what is intended to be a critical edition, fails to report on much of the evidence in the transmission of his material, exemplifies the power that this view of composition has in determining what evidence is regarded as being important. The concept of transmission that I am proposing defines as tasks for research the description of the transmissional matrix and the identification of local stylization factors. Those tasks produce an appetite for variants that can be satisfied only by the total transmission of every item to which one has access. But Weiss can overlook variants as irrelevant to the ‘authentic’ tradition that he seeks to identify for each melody. Similarly, the search for quotations and paraphrases has held attention on sur-
Th e Transmission of Some Aq uitanian Tr ope s
287
face melodic similarities as evidence at the expense of the evidence offered by similarities of syntax, similarities that become evidence as soon as one poses the question, what is the basis of the transmission? Most fundamentally, the differences encountered here reflect differences over the questions that are to be asked. It is a difference between the search for archetypes, on the assumption that the mechanism of transmission is known, and an enquiry about the mechanism of transmission. On one side the leading question is about the development of trope style, on the other it is about the ontology of trope style, with the premiss that style and transmission are inseparable. For a long time the study of tropes was dominated by the debate on the question whether tropes originated as melismas or as new compositions in both words and music. That debate was never much informed by comparative analytical studies that took into account the details of transmission. Clearly such study favours the second of these alternatives in general—that a trope recorded in musical notation represents the actualization of a melodic model under local conditions of stylization and under the constraints of an articulated text—even if in particular cases a trope may have been traced to a pre-existent melisma. Ultimately these questions come down to this one: is a tradition like that of the early medieval tropes appropriately investigated as though they were fixed compositions such as we encounter from the late Middle Ages to the present?
III The transmission of tropes could occur in the act of performance and in the act of writing down, which was a kind of vicarious performance. This is an alternative to the assumption that what is written down must be either a protocol of an act of composition or a copy of something that has been written down before. If writing down is to be regarded as a form of musical performance, then the written score is not necessarily to be regarded in the sense of detailed instructions to the performer. To regard a notated source in that light we ought to require evidence in two categories: (1) evidence that multiple sources for the same music arise through literal copying rather than through repeated independent notation; (2) evidence that the musical circumstances require a narrowing of the range of options to the point of fixing individual compositions. The circumstances may be of two kinds, external factors that establish a canon, as in the case of the chant itself, or a musical syntax that determines the compositional details to a high degree through the close interdependence of the musical parameters. Evidence in the first category alone is not sufficient to support the conclusion of a prescriptive role for notation. Notation may become an artefact or metalanguage, and notational consistency between sources may be a value independent
288
Th e Transmission of Some Aq uitanian Tr ope s
of consistency in performance (more to that point in a moment). Evidence in the second category can generally be expected to be accompanied by evidence in the first category. If the state of the music brings about virtual identity from one performance to the next, we would expect the notation to be explicit and consistent from one source to the next. In fact, as I have suggested elsewhere, we can pinpoint in the sources of the late-medieval corpus of music for the cathedral of Notre Dame of Paris a new interdependence of melodic, rhythmic, and contrapuntal parameters, and at the same time we can demonstrate the copying of scores with revisions made in favour of greater explicitness in the notation.¹⁸ If we lack evidence in both categories we have no reason to regard the written score in the sense of detailed instructions to the performer. The alternative is to understand it as a model for the performance, an exemplar of the piece to be performed. It is as though the notator, in writing down the piece, were saying ‘look, this is how it goes’, just as a performer might say that before performing the piece. That is what it means to suggest that the act of writing down may be regarded as a kind of vicarious performance. It can be interesting to pose these alternatives in terms of the question, what is it that the score represents? Further on in the book I shall suggest reasons for thinking that at the beginning notations more nearly represented singing than works. And where the circumstances of written transmissions suggest—on the grounds of the two criteria just proposed—that the score is prescriptive of the performance in detail, that suggests a shift in the ontology of the music itself and the score seems to represent a work, as well as the singing of it. The idea of transmission as the actualization of a matrix brings together under a single interpretive concept several concepts for describing the making and transmission of medieval music that have usually been regarded as distinct: ‘composing’, ‘writing down’, ‘performing’ (with notation or without, and if without, from ‘memory’ or through ‘improvisation’). It cuts across the traditional dualities in which those concepts have been opposed. For the understanding of musical notation it suggests a new distinction between writing down as an act of actualization and writing down as copying. That in turn raises new questions about the nature of notation as the representation of music. I shall try now to pose those questions.¹⁹
¹⁸ See Ch. 4, the discussion of the organum Iudea et Ierusalem. Further discussion in my article ‘Regarding Meter and Rhythm’. ¹⁹ The discussion initiated here is continued in Ch. 16.
Th e Transmission of Some Aq uitanian Tr ope s
289
IV The hypothesis that I have offered about how the music of the tropes was transmitted corresponds to the concept of oral composition that I have proposed to account for the early transmission of plainchant (see Ch. 6). But while plainchant was a traditional music transmitted for centuries without notation, the transmission of tropes entailed the use of notation in some way from the very beginning. What is more, the tropes do not belong to an ancient tradition. Their composition spans some two centuries, and they are transmitted in a relatively small number of sources, compared with the chant sources. The grounds of their composition and transmission are essentially like the grounds of the oral transmission of such traditional music as the plainchant. What, then, can have been the role that notation played in the transmission of the tropes? The question can be sharpened. A notation is a system of symbols with a capacity for representing the sound combinations and relations of music—for the subject of this book, of sung music. It has been the common understanding that in principle the symbols of Aquitanian notation represent the constituent intervals—i.e. the surface pitch details—of melodies. But if we construe the performance and the notation of an Aquitanian trope as an act of actualizing an underlying model, how could the singer have been served by a notation whose signs represent a level of detail that emerged only as a product of the very act of singing? How could the tacit, wholistic knowledge of a model and the seriatim representation of surface details both have been the point of departure for a performance? Is there some other way of understanding the role of notation than as a string of instructions about what notes to sing? Begin by asking what it means to say that notation in general and Aquitanian notation in particular ‘represents’ sounds and sound relations.²⁰ To say that a symbol ‘represents’ an object is to say that it denotes or stands for that object. In order to do that it must be able to pick that object out of all the objects in its reference field, that is to distinguish it from all the others. To understand the nature of denotation in modern Western pitch notation, it is helpful to think of the lines and spaces of the stave—together with the clefs and sharps and flats—as representing the total field of reference, the system of discrete pitches. The note figures placed on the stave pick out particular positions. Following Goodman, I shall identify the positions on the stave as the characters of the notational system, and each note placed on the stave as an inscription of one of the ²⁰ This discussion has been much informed by Nelson Goodman’s book Languages of Art (Indianapolis, Ind., 1976). I have found Goodman’s formulation of the theory of symbolic representation to be of great help. There are, however, important ways in which his discussion of musical notation suffers from a lack of familiarity with the history of notation and the phenomenology of performance.
290
Th e Transmission of Some Aq uitanian Tr ope s
characters. To us the second space from the bottom, with treble clef, is the character that denotes the pitch a' (440 Hz), the position in our pitch system that corresponds to
A note placed in that position is an inscription of that character. It picks out the pitch a' (440 Hz) from the range of discrete pitches:
The efficacy of the system depends on the distinctness of the characters and their referents, and on a stable linkage between them. Else it would lack the capacity to meet the primary function of a system of symbolic representation, that is the capacity to pick out particular objects from the reference field. The requirements can be stated in terms of two operational propositions: 1. All inscriptions of a character are interchangeable, but the inscription of one character is not interchangeable with the inscription of another. For example all instances of
are interchangeable with one another, but none is interchangeable with any instance of
2. All inscriptions of a character pick out the same pitch from the reference field. (Following Goodman, I shall call the relationship between symbol and referent ‘compliance’; the referent is in compliance with the symbol, or is its compliant.) Then all inscriptions of a character have the same compliant. These conditions are basic to the notation’s usefulness to composers who expect that performers will always follow their intentions in the same way, and to performers who expect the notation to convey in an explicit and detailed way the composer’s intentions. The notated score, then, serves as a bona fide of the identity of the musical work. That it can do so reflects a shared way of thinking about music—in terms of works that are fixed and repeatable. The characteristics of the symbol system reflect something fundamental about the apprehension of the objects represented. If we now try to understand in similar terms the nature of representation in the system of Aquitanian notation, things are not so simple or direct. Several points must be set out. 1. In an idealized view of the system it might be said that the notational field (the vertical space allotted to the neumes between lines of text) corresponds to the
Th e Transmission of Some Aq uitanian Tr ope s
291
stave, and the potential positions within it are the implicit characters whose referents are the discrete pitches of the range exploited by the melodies. After all, one can even superimpose a stave on the notational field of some Aquitanian sources where the vertical placement has been very consistently observed. Then the neumes would be inscriptions of those characters, corresponding to the notes of the modern system. But there are two problems about such a view. First, it approaches the notation from the viewpoint of a modern transcriber, and with that viewpoint we shall not make progress in understanding the utility of the notation to a singer of that time. Even where the vertical placement is most consistent, the relative pitches must be determined by comparing the positions of the neumes and their parts in succession. When there is no custos (e.g. Paris 1118) the process is disrupted with each new line. The neumes are not inscriptions of fixed characters that are potentially present at all times. Second, the observations reported in the discussion of notation near the beginning of this chapter mean that consistent vertical placement was not a necessary condition of notation in the Aquitanian system. 2. Then we must regard as the characters of the system, not the spatial positions but the neume classes themselves (punctum, virga, pes, clivis, torculus, porrectus, etc.). That is the reason for the nomenclature adopted in Chapter 13. Their compliants cannot be the range of pitches, but the classes of pitch groupings denoted by the neume classes (a torculus denotes a three-pitch group in which the second is higher than both the others, etc.). The representation of such pitchgroup classes in association with syllables of text is the common role of all instances of Aquitanian notation (and not Aquitanian notation alone), and we must take that to be the primary role of the notational system as a whole. Whether or not interval distance within and between pitch groups is represented depends on the introduction of contingent devices that I shall call ‘controls’: consistent vertical placement, clefs, custodes, equaliter signs, pitch letters, neume forms that denote particular intervals or positions in the scale, and horizontal clef lines, scratched or inked. Of course eventually some of these controls became constitutive rather than contingent aspects of the system, and that altered the nature of the representation to that of the modern system. But for the system of Aquitanian notation during the period of transmission of the tropes, taken as a whole, that was not the case. 3. We can test this interpretation against the criteria of distinctness of the characters and the stability of the linkage between the characters and their compliants. In view of what was reported near the beginning, we cannot guarantee that the inscription exemplifies a character that is distinct from a different unless we can verify that character that is exemplified by the inscription the score in question is under the control of strict diastematy. But we require no controls whatever to guarantee that and both exemplify a character-type that
292
Th e Transmission of Some Aq uitanian Tr ope s
is distinct from . Further: it appears from what was reported near the beginning , now . Then we canthat the neume might represent, now not guarantee the stability of the linkage between the neume and any one of a number of possible pitch configurations as its compliant. But we can guarantee a stable linkage between the neume class exemplified by and the class of fourpitch groups with an ascent in the first three pitches and a descent from the third to the fourth. We can now interpret the transmission of the d phrase of Ex. 11.5 in this light. With the neume classes as the characters of the system and the pitch-group classes as their compliants, the two conditions that we saw to be necessary for the functioning of the modern system also obtain in the Aquitanian system: 1. All the inscriptions of a character are interchangeable whereas the inscriptions of one character are not interchangeable with the inscriptions of another. All instances of the torculus are interchangeable with one another, whereas none is interchangeable with any instance of the porrectus. This is to say that if a porrectus is inscribed in place of a torculus, it will be recognized as a different character with a different referent. But we cannot say before the fact that if one torculus is replaced by another with different degrees of vertical distance between its elements, the two will be recognized as distinct characters with different referents. That will depend on whether the particular score in view is under the control of strict diastematy, something that cannot be determined for the Aquitanian script in general. 2. Any inscription of a given character will pick out a member of the same class of pitch groupings. Any inscription of the podatus will pick out a two-note group, of which the second note is higher than the first. Such a notation cannot be useful if it is expected that the score should convey to the performer the intentions of a composer (or notator) with respect to all details of pitch; nor can it be useful if the performer is expected always to follow out those intentions in the same way. If the notated score of a trope is to guarantee the identity of the trope, that can only be at some level short of the finished melody that is achieved in performance. But the study of trope transmissions has shown that it is at some such level that the identity of tropes can be found. So here, too, the nature of the symbolic representation reflects something fundamental about the ontological state of the objects represented in the apprehension of its users. What was the utility of such a notation? It provided a basis for the singer’s actualization of a melody, and it identified (‘picked out’) the piece by giving an instance of it. It is in the context of such a conception that I have suggested the notion of ‘control’—guides to (eventually constraints upon) the performer in the act of synthesizing. If this conception of the theoretical status of a score and of its role as Vorlage
Th e Transmission of Some Aq uitanian Tr ope s
293
(the German word is best here) seems mysterious, then perhaps I can suggest a simple analogy as a way of sending the reader’s thoughts in the direction that I have been pursuing. Under my particular habits of work in writing articles such as this, I usually keep a record of my thoughts in handwritten notes and sketches, and eventually prepare a manuscript text. Finally I transcribe from that text onto a typewriter or computer. What I want to reflect about here is the role of the manuscript as Vorlage for the transcription, and its status as representation of the paper. As I transcribe I may copy passages of the text word for word from the manuscript. But at other times I may treat the manuscript as just one concretization of the ideas I want to express. In that case the manuscript serves only to re-present those ideas to me, and I may give them quite different expression. In one instance the Vorlage presents a set of signals for the operation of my fingers on the computer keyboard; in the other it serves as the point of departure for a reconstruction. Note that it may serve in that capacity, even though its text is fully fleshed out. in the second instance the manuscript presents the text at the level of the model. The score, from this viewpoint, may be thought of as a representation of the melody as a whole object. But the melody has a variety of possible expressions. We must now consider whether this viewpoint must be altered if the score takes on a standard form, that is, if all scores for a melody are identical.
V I have concentrated on what might be called the shadowy side of Aquitanian trope transmission. The questions I have raised and the interpretations I have offered might not ever have come up, had the attention been instead on what, by the same token, has passed as the sunnier side, that is on stable transmissions of melodies in notations with sufficient controls to leave little doubt about their contents. Weiss’s tabular summary of the transmission of the trope verse Discipulis flammas (Ex. 11.8) shows the impression that such a transmission would make in a modern representation. Would such evidence contradict my claims about melodies not fixed at the surface? Or would it suggest the coexistence within the Aquitanian trope tradition of two such opposite patterns with respect to transmission, notational type, and the relation between notation and performance? Perhaps the opposition would be more apparent than real. There is a way of interpreting such a situation as is exemplified by Ex. 11.8 that does not contradict the interpretations I have offered so far. It begins with further reflection on the nature of symbolic representation—of which musical notation is a special case— but this time with reference to the pictorial representation of the physical world. There is a view of pictorial representation shared by Nelson Goodman and
294
Th e Transmission of Some Aq uitanian Tr ope s
Ex. 11.8. Trope Discipulis flammas in twelve Aquitanian MSS, as transcribed by Günther Weiss in Introitus-Tropen, i, p. vii
E. H. Gombrich in which the commonplace assumption is denied that representing an object is making a copy or imitation of it.²¹ The fundamental belief they share is that the object of the representation cannot be specified before it has been represented. It is not simply ‘out there’, waiting to be copied on the canvas. As Goodman puts it, the object is ‘achieved’ in the act of representing it, just as I have suggested that the trope melody is achieved in the act of performing it or writing it down. No account of representation is adequate that does not pay heed to its inventive component. ²¹ E. H. Gombrich, Art and Illusion (2nd rev. edn., New York, 1961; repr. Princeton, 1969).
Th e Transmission of Some Aq uitanian Tr ope s
295
Pictorial representation is another kind of denotation. A picture identifies its object, picking it out as one kind of thing from among all the possible kinds in its field of reference. It classifies its objects by means of graphic symbols or figures (Gombrich calls them ‘schemata’). The figures are in effect labels, with which the artist indicates that a such-and-such is being represented. In view of their function in identifying objects graphic labels are conventional, and they may become stereotyped and repeated or copied from one representation to another. They become a system of signs onto themselves. But then the duplication of signs from one representation to another by no means denotes—nor is it necessarily intended to denote—an identity of appearance between the objects being represented. Systems of signs can become conventions in themselves. On the contrary, a single object may be represented through two or more quite different systems of signs, according to the convention in which it is represented. The agreement between a representation and its referent is a matter not of imitation but of habit, a habit in which both the artist and the viewer participate. Gombrich provides a striking illustration for this general view. It is a picture of a lion, drawn c.1235 by Villard de Honnecourt (Pl. XIII). Gombrich writes (p. 76): ‘To us it looks like an ornamental or heraldic image [i.e. like a conventional design], but Villard’s caption tells us that he regarded it in a different light: . . . “Know well that it is drawn from life” . . . He can have meant only that he had drawn his schema in the presence of a real lion. How much of his visual observation he allowed to enter into the formula is a different matter.’ It is not that Villard ‘copied’ the lion in whose presence he worked; he made a representation of it that sufficiently identified it as a lion by means of a conventional schema that his viewers, as we must assume, could read. If that same schema appears in another drawing we can be assured that it, too, represents a lion. But the duplication of the schemata by no means assures the perfect resemblance of two real-life lions to those representations, or to one another. Reflecting these thoughts back onto Ex. 11.8, can we be certain that the identity of the notational representations—assuming they are accurate—guarantees that all performances of the trope Discipulis flammas associated with the sources represented in Ex. 11.8 were identical at the level of detail covered by the notation? I believe we cannot, for there is no guarantee that the notational uniformity is not again a case of stereotyping in the system of signs itself. It may be a case of stereotyping in the neumes (e.g. those in the MSS lat. 1240 and lat. 1120, which have been reproduced from the MSS), and it is certainly a case of stereotyping in the notation of the modern transcriptions from the ten remaining sources. Example 11.8 implies that the neumes of those twelve sources are all diastematic to such a degree of clarity and consistency as to yield unambiguous readings (transcriptions) that are identical down to the last note. But that is wishful thinking and not at all the case. Perhaps the neumes of MSS lat. 903 and 1121 can be
296
Th e Transmission of Some Aq uitanian Tr ope s
characterized that way with some confidence. The neumes of MSS lat. 1084, 887, and 1118 cannot. The table’s picture of the identity of this item throughout the south French transmission—and its generalization by virtue of the presentation of the table as representative—have been supplied by the editor from a dogma about melodic fixity that he has adapted from the literature of chant studies (see the introduction to Ch. 6). Example 11.8 rests on the very premiss that it is meant to establish as a conclusion. If we imagine, as a thought exercise, that we ask each of ten adepts in this field to transcribe the melody of Discipulis flammas into modern notation from one of those ten sources without consulting one another or the other nine sources and if we wonder ‘What is the likelihood that they would produce ten identical transcriptions?’ the answer would surely be ‘very, very low’. But my central aim here is not merely to discredit Ex. 11.8 as evidence. The conception about notational representation that I have been pursuing here would not be subverted even if the table were credible. For what I posit comes down in this case to a view that the intonation of the trope verse Discipulis flammas was conventionally represented by patterns of neumes matched to a sequence of syllables articulated in words, phrases, and sentences. As the verbal side of this complex was invariable, it is as natural to think of notators executing that task by imitating written models, as it is to assume that they did so by imitating local performances in writing or by setting down their mental representation of it—whatever that might mean (this raises questions of ontology that are confronted in Ch. 12). We must allow the first of these possibilities, because we have no evidence whatever that an identity relationship between notation and either performance or some sense of the song as a fixed work—which is a premiss for the second and third possibilities—would have been held as a norm by the participants in that tradition, as it is by participants in some modern traditions of music-making. On the other hand, what has been presented in this chapter makes the assumption of such a norm for that tradition somewhat dubious. But we must also allow that if the notational representation of a song has become stereotyped, for whatever reason, that may in return influence the standardization of performance. The problem of transmission is critical for the understanding of the musical situation in the central Middle Ages. And the tropes are critical for the study of transmission; for their time coincides with the first epoch of music writing, and their sources provide us with substantial evidence about the way they were apprehended in the communities in which they were current. What I hope to have uncovered in this chapter is something of the great complexity of the problem of trope transmission. We require a taxonomic approach to the tropes that takes in all ‘variants’ and that is based not on the impression of surface resemblance but on the analysis of the co-variant options presented by the musical idiom in general, by the stylization principles exhibited by particular col-
Th e Transmission of Some Aq uitanian Tr ope s
297
lections, by the chants with which the tropes are associated, and by the shared aspects that define the identities of individual tropes through their several exemplifications in different sources. At the same time we require an evaluation of the written transmission through a concentrated effort to understand in each instance what is really denoted by the written score (not what ought to be denoted), how, and to what purpose. In this we must be prepared to recognize that, on the one hand, writing is only one aspect of transmission, and on the other hand that it has its own dynamic and develops its own traditions that can impact on the musical transmission itself.
postscrip t Subsequent to the completion of the original version of this chapter, the first volume of the Corpus Troporum appeared in print. The very complexity of the transmission of tropes that I have shown here led the editor and her colleagues to the decision to edit the material as individual elements (verses) first, and only then to show how they are combined in each manuscript appearance. In her editorial commentary she made explicit that ‘Such a method . . . leads us to dismember an organic unity and to present an image of the state of the texts such as has never existed in any manuscript.’²² She defended the solution chosen as the only practical one, for ‘there are in fact very few stable tropes with the same parts repeated in the same order in the several manuscripts’. But perhaps the premiss of organic unity is yet another of those concepts on which we have depended that fits much of our music but not the music of the tradition we have before us here, and is consequently more misleading than helpful in the approach to this material. (Here one could develop an excursus on the history of the ideal of organic unity, which seems to have entered European intellectual history prominently in the late eighteenth century—more or less concurrently with the establishment of text criticism. But that would be another project.) The trope composer seems to have moved through a vast associative field in realizing trope complexes. We may approach the medieval musical understanding if we think of a fluid situation in which the specific merges with and emerges from the generic. Ritva Jacobsson has put it more poetically: ‘The liturgical authors act as bees, according to an image by a medieval author, flying from one flower to another and sucking the sweet honey where they can find it.’²³
²² Tropes du propre de la messe, i: Cycle de Noël, ed. Ritva Jonsson (CT 1; Uppsala, 1975), 21. ²³ Ritva Jacobsson and Leo Treitler, ‘Tropes and the Concept of Genre’, in Pax et Sapientia: Essays on Medieval Music and Texts in Memory of Gordon Anderson (Acta Universitatis Stockholmiensis, Studia Latina Stockholmiensia; Stockholm, 1986), 59–89.
298
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
chapter 12
History and the Ontology of the Musical Work
} If the Mona Lisa is in the Louvre, where is the Toccata and Fugue in D Minor? This question, which is not easy to answer, strikes up the question of the ontological state of music, or, as Roman Ingarden put it in the title of a famous book on the subject, The Work of Music and the Problem of its Identity (Berkeley, 1986). Searching for a single answer corresponding to ‘the Louvre’ is one way of becoming aware of the elusiveness of the question vis-à-vis music—the Cologne Cathedral, the Neue Bach Ausgabe, the hearts and minds of countless music lovers, the CD set Deutsche Grammophon DG 463 016-2, Walt Disney’s film Fantasia. We can complicate the conundrum by trying to decide whether the performance of Leopold Stokowski’s transcription of the Toccata and Fugue in D Minor for symphony orchestra in Fantasia should really be counted an instance of that work, which, after all, was composed for the organ; or a rendition by the Swingle Singers or an arrangement for electronic synthesizer by Wendy Carlos. Still more provocative would be to ask whether a performance that brings a virtuoso obbligato violin part into the fugue should be counted an instance of it. The question expands: not only where is the work located, but what are the margins of its identity—altogether, what is the state of its being? Musical audiences and scholars do not on the whole worry about such questions when it comes to works like the Toccata and Fugue in D Minor. Philosophers of music are an exception, as I report near the beginning of the chapter. Why should we? Because such questions are inescapable in performing, studying, and appreciating medieval music, and they pulse through this book, as the reader must already have noticed. Because of their centrality to my subject , and because of the importance of avoiding too-easy answers imported from the realm of post-Renaissance music I include a paper that I prepared for a gathering of philosophers for a colloquium on the subject ‘Philosophy and the Histories of the Arts’, sponsored by the National Endowment for the Humanities.¹ ¹ National Endowment for the Humanities workshop, ‘Philosophy and the Histories of the Arts’, San Francisco State University, summer 1991.
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
299
What I aimed to contribute was the insistence that the question about the ontological state of music cannot be answered satisfactorily in the abstract and once for all. It must be addressed taking into account the particular historical situation of the music in question, in its particular culture. It is both a historical and a philosophical question. My interest in including the paper in this book is not only to argue that we must not project what we regard as self-evident answers onto medieval music from our experience with modern music, but also to show that those answers are in fact not always so self-evident even with respect to modern music.
} Which is of primary musical importance, musical works, i.e., symphonies, songs, sonatas, etc., or performances of musical works? Are works, or performances, the basic objects of musical attention, musical appreciation, and musical judgment? . . . Does it ever matter what work we say something is a performance of? . . . Should a performance of a fourhand piano transcription of a Brahms symphony count as a performance of that symphony?²
‘What is the work of art that Messiaen composed and which he entitled Et exspecto resurrectionem mortuorum?’³ ‘What exactly did Beethoven compose? . . . what sort of thing is it, this quintet [in E flat for piano and winds Op. 16] which was the outcome of Beethoven’s creative activity? What does it consist in or of?’⁴ ‘How Does a Musical Work Exist?’⁵ Why do philosophers ask such questions? I have an idea about why we musical scholars would do well to ask such questions—and to try to formulate answers with the care that philosophers give to the task. Because the way we conduct our studies is actually informed by answers that we have adopted, but tacitly and not necessarily as a matter of conscious choice. I mean studies in the analysis and criticism, the editing, and the performance of music (especially since all the to-do about ‘historical performance’), and the pursuit of particular historical problems (e.g. the history of musical notation). The best understanding that I have obtained about the reason for the philosopher’s interest in these questions came from my reading of a very brief exposition by Karl Popper.⁶ He writes: ‘A musical composition has a very strange sort of
² Kendall L. Walton, ‘The Presentation and Portrayal of Sound Patterns’, in J. Dancy, J. Moravcsik and C. C. W. Taylor (eds.), Human Agency: Language, Duty and Value (Stanford, Calif., 1988), 237–57. ³ Nicholas Wolterstorff, Works and Worlds of Art (Oxford, 1980), 32. ⁴ Jerrold Levinson, ‘What a Musical Work Is’, Journal of Philosophy, 77 (1980), 5–27 at 5. ⁵ Ingarden, The Work of Music, sixth chapter heading, p. 116. ⁶ In Karl R. Popper and John C. Eccles, The Self and its Brain (Boston, 1983), ‘Dialogue III’, 449–50.
300
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
existence’, and proceeds at once to illustrate with an example, ‘say, a symphony’.⁷ The symphony he happens to choose is Mozart’s Jupiter, which is neither the score [Mozart] wrote . . . nor is it the sum total of the imagined acoustic experiences Mozart had while writing the symphony. Nor is it any of the performances. Nor is it all performances together, nor the class of all performances . . . [It is] a real ideal object which exists, but exists nowhere, and whose existence is somehow the potentiality of its being reinterpreted by human minds. So it is first the work of a human mind or of human minds, the product of human minds; and secondly it is endowed with the potentiality of its being recaptured, perhaps only partly, by human minds again.
Popper has come close to Roman Ingarden’s characterization of the musical work as an intentional object, with a mode of existence like that of a unicorn (which exists in people’s minds and in the contact between minds). The musical work interests him as the clearest exemplification of objects that constitute what he calls ‘World 3, a world which exists nowhere but which does have an existence and which does interact, especially, with human minds-on the basis . . . of human activity’. The World 3 population also includes works of literature, theories, and problems. In describing the scope of human engagement, says Popper, it is necessary to postulate a World of such objects (separate from World 1, the world of cave paintings, decorated instruments, decorated tools, boats, etc.). One would hardly find fault with Popper’s assertion that we do need to identify a domain of objects that we know and with which we engage in that paradoxical mode of existence that is best exemplified by works of music. Music therefore offers itself as a guide—by way of the investigation of its ontology—into a distinct mode of cognition, communication, and interaction that constitutes an essential element among the possibilities of human engagement. That is at least one way of understanding the seemingly unlimited fascination that such questions hold for workers in the field of aesthetics and beyond. The discussion seems to have reference to a non-historical world. It seeks out intrinsic and unchanging roles and relationships that are not conditional upon individual choice and changing historical circumstance. And it abounds with fantastic examples: a sixth symphony by Ludwig van Marthoven (a Martian composer),⁸ or two sound-alike versions of Darius Milhaud’s Le Bœuf sur le toit, The Bull on the Roof and The Cow on the Roof, Debussy’s Quartet arranged for kazoos and guitars,⁹ and other repertory that you might expect to hear in a concert staged by Dr Seuss. Such inventions belong to a research methodology like that of controlled laboratory investigations, which also take place in a non-historical world. This seems fitting enough in a discourse in which a piece of music is referred to as ⁷ I shall want to comment on this posture—typical and illusory, I shall suggest—of fishing around for an example in a random sort of way, as though one would be prepared to work with whatever came up. ⁸ Walton, ‘The Presentation and Portrayal of Sound Patterns’. ⁹ Paul Ziff, Antiaesthetics: An Appreciation of the Cow with the Subtile Nose (Boston, 1984), 44.
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
301
‘an auditory pattern’,¹⁰ a performance as a ‘physical realization’,¹¹ and a score as a ‘conventional and arbitrarily coded statement’ of a musical composition.¹² What is less obvious is that this inquiry has, paradoxically, been directed from a very particular historical stance. And what it conveys is not the existential condition of musical compositions under any circumstance by virtue of their being musical compositions, but rather beliefs about the condition of a limited number of high-art compositions of the European tradition, composed approximately from the time of Bach to the end of the nineteenth century—music of the ‘common practice’ period, but by no means all such music. I hinted at this in my comment on Popper’s choice of an example ‘say, Mozart’s Jupiter Symphony’. Why did he not write ‘say, Rossini’s Tancredi ’ (of which three different versions were produced in 1813), or ‘Stockhausen’s Klavierstück XI’ (a composition comprising eleven segments that are played in a sequence that is determined by the performer only at the moment of performance, and with dynamics and tempi for the segments that rotate independently of them), or ‘the raga Kirwani as performed by Ali Akbar Khan last week at Carnegie Hall’? Because they might be less familiar to his readers, to be sure; but also because they are not works that have been cast once and for all by a composer in a final form that is represented in a score and presented essentially unchanged in performances. (Everything depends, of course, on what lies beneath that word, ‘essentially’.) Or to put it differently, because they leave so much to performers’ choices or to the circumstances of performance, that it would be difficult for the reader to know just what is being pointed to unless it were a specific performance. This is a presumption, to be sure, but I think not an unfair one. The conditions that make Mozart’s Jupiter Symphony so unambiguous as an example belong to a work concept that has informed music-aesthetic, -theoretical, and -historical studies since about 1800.¹³ When authors in this discourse reach out for examples—I ¹⁰ Paul Ziff, Antiaesthetics, 45. ¹¹ Ibid. ¹² Popper and Eccles, The Self and its Brain, 447. Walton does immediately gloss his question about the relative importance of works and performances with this note: ‘I should emphasize that this question, as I construe it, is a question about our cultural institution of music. I am asking what roles pieces and performances have in our institution, how they are regarded and treated by participants in the institution: composers, performers, and appreciators. Neither works nor performances are intrinsically primary, apart from their place in some cultural institution.’ But he is the inventor of Ludwig van Marthoven’s sixth symphony, which he compares to Ludwig van Beethoven’s work of the same title without any concern for the vast (to put it mildly) cultural distance between the two. Ziff, in asking the familiar question about how much can you change—this time in respect to instrumentation—and have it still be the same work (‘double up on the oboes . . . How about three? . . . Five? . . . How about fifty oboes? . . . ten thousand?’), says ‘it all depends on the work in question’. Popper distinguishes his World 3 from Plato’s world of ideas by virtue of its having a history. But none of this alters the fact that this is ‘philosophico-aesthetic research’, as Ziff puts it, aimed at formulating in as nearly unexceptionable a way as possible the ontological status of the musical work in relation to scores and performances; it is not an empirical or historical investigation into the ways that works, scores, and performances have interacted in any actual musical culture. ¹³ See Lydia Goehr, ‘Being True to the Work’, JAAC 47 (1989), 55–67, and The Imaginary Museum of Musical Works.
302
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
don’t mean their invention of counter-examples—these are the conditions that guide their hands. Hence their conclusions regarding the mode of existence of musical works can be somewhat redundant, and the enquiry itself can be somewhat circular.¹⁴ Kendall Walton, perhaps wishing to interrupt the circle, moves to eliminate one of the factors that creates it. He considers first that ‘a work is the pattern specified by its score’, but then notes that ‘some works do not have scores, and in oral musical traditions there are not even conventions for producing scores’. So he sidesteps the score and proposes to define ‘a work as the pattern determined by the sound properties a performance must have to be a correct or flawless performance of it’. But this only tightens the circle; for it nevertheless posits as a condition for the identification of a work some criterion (unspecified) for ‘flawless performance’. In effect the difference between an oral tradition and a literate one with respect to the ontology of the musical works they produce would then turn only on the existence or non-existence of the physical artefact of a score. The condition of musical works in the two situations would be the same. While that might be so in particular cases, it surely cannot be assumed in setting up the enquiry. In some oral traditions, at least, works come into being only in performance and at the instant of performance, and the idea of a ‘flawless performance’ as counterpart to a flawless execution of a score would be a chimera. But we shall soon see that it can be a chimera even in a tradition. Walton’s adjustment tightens the circle, too, in its implication that exceptions to the rules operating under the nineteenth-century work concept must be sought outside the tradition of Western art music during the period of common practice. The remainder of this chapter will be given over to a demonstration that this is not so, with specific reference to works by Frédéric Chopin and their scores and performances. My first example is Chopin’s Mazurka Op. 7 No. 5, which I should like to consider first from the point of view of the two general kinds of roles that are commonly attributed to scores: an instructing role and a designating or identifying role (Ex. 12.1). The score concludes with an instruction that cannot be carried out: Dal segno senza Fine ‘From the sign without end’. The ‘sign’ is the one placed after four introductory measures () and at the double bar in bar 20. The instruction is ¹⁴ I have the impression that, by an unwritten law, the composer whose music is most often selected in investigations about the ontology of musical works is Beethoven. This reflects Beethoven’s long-standing position in our culture as the paradigm (and paragon) of what constitutes serious music altogether. Anytime anyone makes a claim for some musical practice that it belongs in the canon of Western high art music— whether it is Gregorian chant, the jazz compositions of Duke Ellington, or the dodecaphonic compositions of Anton Webern, they sooner or later come to the point of showing its basic similarity with the music of Beethoven. The leading criterion in such comparisons is always closed, unified formal structure, one of the recurrent conditions for work status. See Ch. 9 for examples.
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
303
Ex. 12.1. Chopin, Mazurka Op. 7 No. 5 in C major, from Frédéric Chopin, Mazurken, ed. Zimmermann (Munich, 1975)
304
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
‘[when you get through the last notated measure] go back to the [first] sign [and play through to the second, and keep on doing that]’. The formula is a deviant version of the more common Dal segno al fine (from the sign to the end), and normally the word Fine is written wherever the piece is to end (e.g. the Mazurka Op. 17 No. 3). Even if a performer wants to follow the instruction in Op. 7 No. 5 he or she will sooner or later be forced by circumstances beyond his or her control to give up. As a result the score has multiple realizations, depending on where the performer chooses to stop. Modern technology makes a ‘correct’ performance possible: put it on a tape loop. But who will listen to the whole piece? What, for that matter, is the whole piece? That question takes us into the domain of the identifying function of the score. If we are guided in our idea of what constitutes a work by principles about wholeness and unity, about form and schema, about characteristic structure, then we will be stumped about just what it is that this score identifies. We might say that it is a piece with an open form, with a material but not a formal unity, with a characteristic structure that is circular. I don’t know what to do about wholeness—an idea that has been associated with the work concept for a long time (but not forever) both as a condition of the art object and as a psychological condition for the apprehension of beauty. As a historian I regard the idea that music is exemplified in works considered as structures neither as the description of a universal condition of music nor as selfevident, but as an idea with a history. We first hear of it in the sixteenth century with the concept of the ‘complete and absolute (i.e. autonomous) work’.¹⁵ But that is not to say that every musical item thereafter is governed by it. The interpretation of any particular item of music as governed by the principles embodied in the modern idea of ‘structure’ I regard as a hypothesis, not as a universal whose particular exemplification is the goal of each analysis. And the priority given to that conception in musicology, music theory, and musical aesthetics—regardless of the aesthetic premisses under which the musical object was created and first heard—I regard as the projection of a value that has itself a narrow historicality. I think that Chopin’s score identifies a piece without an ending, and that is the most striking thing about this piece; it is in all other regards rather uninteresting. But in just this one regard it reminds me of Beethoven’s Ninth Symphony, which I ¹⁵ See Carl Dahlhaus, Esthetics of Music (Cambridge, 1982), 10, but also the discussion in Ch. 2 n. 46 of what has been identified as possibly the earliest articulation of the work concept (by Nicolaus Listenius). Arnold Whitall has given a characteristic discussion of what is entailed when one speaks of structure nowadays. He quotes Wallace Berry: ‘Musical structure may be said to be the punctuated shaping of time and “space” into lines of growth, decline, and stasis hierarchically ordered.’ He writes himself that ‘it is only relatively recently that the possibility of musical compositions displaying structure as opposed to form has been envisaged. With the development of structuralism . . . the argument is advanced [now he quotes from Piaget’s book, Structuralism] that “structure in the technical sense of a self-regulating system of transformations is not coincident with form.” The opposite of a true structure is an aggregate, a “mere collection of elements,” which may indeed possess certain formal characteristics, but which lacks the true wholeness and unity of a genuine structure.’ (See the article ‘Form’ in NG II, ix. 92–4.)
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
305
think of as a piece without a beginning. (I once put it that the opening is as though it had always been going on, but someone has just turned up the volume.)¹⁶ This comparison might lead me into the pursuit of the very interesting history of pieces composed precisely against the idea of closure and wholeness since the early nineteenth century. It would be a thread in the history of music that is obscured when we take that idea for granted. I would also associate this feature of the Mazurka with Chopin’s general proclivity for making pieces circular in different ways—directing them back into their beginnings when it seems time to end (what Schumann parodied in his ‘Chopin’ piece in Carnaval), setting up chromatic harmonic pathways through which the piece slips down, down, down until it arrives ineluctably at its point of departure. There are different ways to interpret this—a Romantic fascination with the unfinished, or with music as expression of the infinite, Chopin’s own deep pessimism. Yet another avenue of interpretation is entered by noticing that Chopin has here composed a sort of impression or portrayal of a Polish folk dance of the sort that he heard when he spent every summer in the countryside until he left for Paris. It, too, is music that stops but does not conclude. The longer I go on thinking about it in this way, the less this piece seems a curio or the more its curious aspect opens onto matters of wider historical significance. And the longer I pursue this line of thought the more I think that the score can be perceived to identify not only a characteristic structure but also a certain character. And I think, too, that to the roles of identifying and instructing that are commonly ascribed to the score vis-à-vis the work and the performance, respectively, we must add the possibility of an exemplifying role.¹⁷ The performer of the Mazurka cannot comply with the last instruction in the score. Early recorded performances of Chopin’s music give evidence of a practice of intentional non-conformity with provisions of the score by pre-eminent pianists of the time. Chopin marked the second section of his popular Waltz in C sharp minor Op. 64 No. 2 più mosso, more moved or agitated, and he put a hairpin closing to the ¹⁶ Treitler, Music and the Historical Imagination, 20. ¹⁷ A score exemplifies a work when the community of practitioners to which it is addressed makes performances based on it within a range wider than is circumscribed by notations of the score. We may conclude that is the case from the evidence of actual performances made from known scores. We can infer it, without the same degree of certainty, about traditions no longer active when it is typical for multiple scores of the same work to differ, again in musical parameters that are explicitly specified. There are many such situations in the sources of European medieval music, and the inference is strongest when such multiple scores occur in closely related sources, or even in the same source. (I shall shortly come to instances of both kinds.) Then the work occupies that entire range, and each score exemplifies the range indirectly by denoting one specific realization of it. It can only do that; it is how the notation works. (Composers in the 20th c. have found ways to notate works as multiplicities of realization possibilities, as in the case of Stockhausen’s Klavierstück XI.) But scores can function in the sense of exemplification only for a community of practitioners who know the tradition.
306
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
right under each of the repeated figures. Example 12.2 shows (a) a published edition of this section, and (b) details transcribed from recorded performances by Alfred Cortot, 1934 (C.), Ignaz Paderewski, 1917 (P.), and Sergei Rachmaninoff, 1927 (R.).¹⁸ Cortot leaves off the third beat of the bass in each bar and when he comes to the climactic four bars of the phrase he compresses—he doesn’t give them their full count of twelve beats, only about nine. The important thing is the effect of rushing. Transcriptions of all the autograph sources for this passage—two sketches, the copy Chopin gave to his Paris publisher for the publication, and three presentation copies—and three contemporary published editions, have been published by Jeffrey Kallberg.¹⁹ For the left hand there is not complete agreement between any two of these nine sources. Interestingly enough, Cortot’s omission of the third beat of each bar is like the revised sketch. Kallberg writes at the beginning of his study, ‘Sketches, autographs, manuscript copies, “simultaneous” first editions, . . . annotated editions—when we examine any of these sources we are almost certain to encounter alternate readings. Variants are pervasive in Chopin’s music; indeed it is tempting to assert that they are essential to its aesthetic mode of existence’,²⁰ that is, to its ontological status. Paderewski, when he runs up the scale at the end, does so through a chromatic scale, not the diatonic scale that Chopin wrote. Both his and Cortot’s variants enhance the effect that is suggested by the instruction ‘più mosso’. Rachmaninoff does something that seems at first as though it would have the opposite effect: he calls attention to the descending line of the lowest note in each of the figures at the end of each bar, making the figure as a whole just ornamental, and in the end achieving the same sort of effect. But he does this by violating the score, which says ‘play the last note of each bar most softly’. What are we to think about these performances? They confront us with two successive choice points. First, do we regard them simply as incorrect and thus irrelevant to the discussion, or, somehow respecting their authority, as challenges to our understanding of what works are and how they are denoted by scores and presented by performances? In any case, I think we should avoid the trap of trying to establish their authority by way of their linkage through the generation of the pre-eminent pedagogues Theodor Leschetizky and Anton Rubenstein to contemporaries of Chopin, for that would imply that this is the only way of establishing their authority. Second, if we accept their authority, the choice would be ¹⁸ The performances can be heard on the following recordings: the Cortot recordings of the two waltzes on Alfred Cortot: Frederic Chopin (Hunt CD 510); the Paderewski recording of the Waltz in C sharp minor on The Art of Paderewski (RCA Camden CAL-310); the Rachmaninoff recording of the two waltzes on Rachmaninoff Plays Chopin (RCA VIC 1534). ¹⁹ ‘Are Variants a Problem? “Composer Intentions” in Editing Chopin’, Chopin Studies, 3 (Cambridge, 1990), 257–76. ²⁰ Ibid. 257.
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
307
Ex. 12.2. Second (più mosso) section of Chopin, Waltz Op. 64 No. 2 in C sharp minor: (a) from the Henle edition, and (b) transcribed from recordings by Alfred Cortot (1934) (C.), Ignaz Paderewski (1917) (P.), and Sergei Rachmaninoff (1927) (R.) (a)
Più mosso
3 5 4
4
❊ 38
4 5 4
3
❊
4
5 3
3
4
3
1
❊
4
2 3
4
1
1
5
❊ 4
3
4
2
4 1
4
❊
❊
❊
❊ 43
3
4
3
1
❊
❊ 4
1
2 1
1 2
2 1
4 1
❊
❊
❊
❊
4
❊
❊
❊
49
❊
❊
❊
❊
55
❊ 60
❊
❊ 1
❊
2 1
1
❊ 4
8
2 1
❊
❊
1
❊
❊
308
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
Ex. 12.2. cont. (b)
R. 33
etc.
C.
45
8
3
P.
3
3
3
3
3
between taking a position like Ingarden’s (and Hanslick’s) about the separation of the work from the performance on one hand, and the view that the work is realized only in performance on the other. On the latter view our conception of what the work is must include all three performances, as well as one that would be in exact compliance with the score. As to the former, one could wonder whether Ingarden would have included among the variables of performance that he identified—‘tonal colorings, tempi, dynamic detail, the perspicuity of specific subjects’²¹—such variation in note contents as is revealed in those performances. I doubt it. But if not Ingarden, then perhaps some structuralist theorists, who would identify the work with its underlying structure, in relation to which these variants are superficial details no more subversive of the identity of the work than would be differences of intonation by performers of a work for string instruments. It is not my purpose to argue for one side or the other on any of these questions, but only to argue for the relevance of the evidence of actual practice and belief for the basic question that is my subject. If the work is to be identified with its structure, what is one to say when the structure is itself subject to variation? Chopin’s waltzes, like his mazurkas, are pieces put together of a number of sections or, we might call them, blocks. And one of the things that seems to be potentially fluid about them—from the process of composition right into the performance tradition—is the way these blocks are put together: what sequence, how many repetitions, how to end, etc. Sketches show Chopin playing with just such arrangements.²² The next example shows ²¹ The Work of Music, ch. 1: ‘Performance’, 13–14. ²² Reported in Jeffrey Kallberg, ‘The Problem of Repetition and Return in Chopin’s Mazurkas’, in Jim Samson (ed.), Chopin Studies (Cambridge, 1988), 1–24.
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
309
performers doing the same. The Waltz in G flat major Op. posthumous 70 No. 1 is one of a number of Chopin’s works that were published after his death by his longtime friend and associate Julian Fontana. The performance tradition of the Waltz had always been based on this publication, as had its editorial transmission. Only the critical edition undertaken by Zimmerman in 1978 and published by Henle is based on the autograph score that became available after the Second World War.²³ Example 12.3 shows the four sections of which the piece is made, in the edition based on the autograph, with some rhythmic variants in Fontana’s version. The two versions differ by a matter of one repetition, as follows: Autograph: AA Fontana: AA
BB BB
CC CC
DC DC
DC
A(A) AA
In his recording of 1921 Rachmaninoff happens to play the autograph version. Suspicion that he may have known that version somehow will be allayed by comparing the rhythmic variants in section A. In every case he plays the version of the published Fontana edition. It is imaginable that there was nevertheless an unwritten tradition for Chopin’s repetition scheme, going back to his own performances. But then it would be surprising that Fontana didn’t know it. Cortot, in his 1934 recording, plays the Fontana version through, but then runs through B and A once each before stopping (thus AA BB CC DC DC AA B A). Obviously for both pianists the score both instructs and exemplifies, that is, it allows other exemplifications as well. That may seem hard to swallow from the standpoint of the doctrine that, by virtue of its purpose and function of the score to identify the work and instruct performers, it has first and foremost to be unambiguous. I say it nevertheless on the evidence of such performances, and in the belief that apart from a performance tradition no notation has a final meaning. From this point of view the reference of scores, in one direction, is to works, all right, but in another it is to performance traditions. I was first brought to such a general view about works, scores, and performances from my studies in medieval music traditions. Example 11.3 in Chapter 11 is my transcription of several scores of a work from such a tradition, a trope from the eleventh century. Tropes are a kind of song sung in the mass of the Latin church from the tenth to the twelfth centuries, in the following kind of relationship with Gregorian chant: the chant is a musical setting of a scriptural text, for example, ‘We have seen his star in the East, and are come with gifts to adore the Lord.’ The trope would be a newly composed text introducing the scriptural one, and sung in a style consistent with the style of the chant, for example, ‘With the star pointing the way the kings come from the East, seeking the newborn King of kings, and ²³ Frederic Chopin, Walzer, ed. Ewald Zimmermann (Munich, 1963). See Kallberg, ‘Are Variants a Problem?’
310
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
Ex. 12.3. Chopin, Waltz Op. posthumous 70 No. 1 in G flat major, autograph version A
1
5 3
4
8 3
3
3
3
3
1
2 4
8
5
8
3
B
3
Fine 1
9
8
4
4
4
2
1 3
3
3
etc. 8
13
1
2 2
saying, “We have seen his star . . .” ’, etc. The sort of thing I wish to show about this is what brought me to these problems in the first place. The trope verse shown in the example was sung as an introduction to the traditional chant. Others were sung between sections of the chant. So the performance of the entire chant could entail several trope verses. But from one time to the next and from one place to another the same chant could be sung with a large variety of different introductory and interpolated tropes. Consequently what constituted the whole of a chant in performance with all of its trope verses—say the introit
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
311
Ex. 12.3. cont. 17
4
C
2 4
dolce
1
22
2 4 3
2
D
3 2
3
cresc.
4
26
4
15
31
5
2
4 2 12
36
Da Capo al fine
chant for the Christmas mass—was open and variable. There would have been no conception of such a ‘whole’. The tradition of the tropes was fluid also in the notated presentation of the melodies. My transcriptions show four notated versions of an introit trope verse: two from Aquitania and one each from northern France and southern Italy. I say
312
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
they are versions of one thing rather than different things because they have the same ritual function for the same feast day and the same literary text. The scores of the melodies themselves challenge one’s sense of what counts as ‘the same melody’. My general approach to this problem is to consider them members of the same tune family, like versions of ‘Barbara Allen’. The lower two melodies are there for comparison; they belong to the same tune family as the others but they were sung with different texts and on different ritual occasions. Obviously we can’t read any one of these scores as unique identifier of the work, or denotation of the sound pattern that must be executed to produce a flawless performance of the work. Each score can do no more than exemplify the work. In some traditions exemplification is the main task of scores. In such a tradition identification and instruction are accomplished only indirectly, through exemplification. The score in such a tradition is functionally and ontologically parallel, at the same theoretical level, with the performance. The questions ‘what is this piece’ and ‘how does it go’ could both be answered either by writing out a score or by making a performance. But the tradition of Gregorian chant itself was under institutional constraints from the beginning of its written transmission sometime in the ninth century— constraints imposed more by secular than ecclesiastical authorities—to be practised with the utmost uniformity. And that is reflected in a written transmission of considerable uniformity, compared with the transmission of the tropes. For the chant tradition we must infer a very different ontological state of the work, and a very different role of the score in relation to work and performance. In giving place to the historical dimension of the study of these relationships, then, we need to recognize that there may be cultural, institutional, ideological, and political pressures favouring or requiring one kind of relationship over another. My last example will focus attention on the relations among sketches, scores, and work in the case of Chopin’s Nocturne in B major Op. 62 No. 1. Example 12.4 comprises only three bars (53–5), but the number and variety of sources for these is very instructive. They come at a very crucial moment in the piece, a harmonically enriched reprise in the second section.²⁴ Once established in Paris, Chopin had relations with publishers there, in Germany, and in England. His usual practice was to send manuscripts for pieces as he completed them to all three. He made all the copies himself. And where we are able to compare copies for the same piece they are often not true copies. He usually read proofs for the Paris publications, and there are many instances in which these editions do not correspond exactly to the corresponding autographs, the alterations having presumably been made at the proof-reading stage. Chopin made presentation copies of works and these, too, can disagree with published ²⁴ The example is derived from Kallberg, ‘Are Variants a Problem?’
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
313
Ex. 12.4. Chopin, Nocturne Op. 62 No. 1, bars 53–5: (a) sketch, autograph fair copies, and first editions; (b) later editions: (1) Paderewski Edition (1949) = Brandus Edition; (2) Henle Edition (1966), right hand = Brandus, left hand in bars 53–4 = Brandus; (3) Universal Edition (1980) = Conflation of Brandus and Breitkopf editions (a) Sketch
[] drugi [second]
[ ]
[ ]
Brandus MS Version 3
Version 2 Version 1
Version 2
Breitkopf MS Breitkopf Edition Wessel Edition
[Autograph]
[]
Brandus Edition
314
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
Ex. 12.4. cont. (b) 53 (1)
54
55
(2)
(3)
editions. He earned his livelihood in large part as a fashionable piano teacher, and it was mainly his music that pupils were to learn. When they brought their published scores to their lessons, he sometimes entered changes in them. There is no aspect of the transmission of his music that does not reveal the fluidity of its ontological condition. Evidently for Chopin the score was not the ultimate touchstone for the work. He did not always copy it faithfully when he needed copies. He evidently did not mind having different versions in circulation. By the same token, he did not behave as though he believed that the process of composition had a terminal point (e.g. the writing down). In that light the fluidity in the performing tradition from which I showed examples before is hardly surprising. Here are the details about the sources for the B major Nocturne. We can compare a sketch, the autograph scores that Chopin prepared for the French (Brandus) and German publishers (Breitkopf & Härtel; the autograph prepared for his London publisher, Wessel, has not survived), all three first editions, and the modern editions that have come down from these. There are three versions of the passage we are looking at in the Brandus autograph. The Breitkopf edition differs in bar 54 from its autograph model. The Paderewski edition is identical with the Brandus,
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
315
the Henle edition combines the third version of the Brandus autograph in the right hand with the first version in the left hand. The Universal edition is a conflation of Breitkopf and Brandus. It is evident from his correspondence that Chopin sent the autographs off to his three publishers on the same day. However the differences among them arose, Chopin had the opportunity to eliminate them but did not do so. We would be hard put to order this material in accordance with the paradigm of sketches and versions progressing towards a final and perfected work that has emerged more than anywhere else from the study of Beethoven’s compositional process. (It remains for someone to investigate the mutual influence that this paradigm and the nineteenth-century work concept have had on one another.) The sketch displays a case of composition through the particular working out of a basic and commonplace schema of the musical style—really like the working out of a figure in a drawing from a basic schema. The material, taken all together, shows Chopin working it out in different ways. In that way it is similar to the transmission of the trope. The schema is a sequential descending progression. We might consider it a structural member that underlies all versions. The details of the syncopated left hand and chromatic descent in the inner voice are already in the sketch. But I want to focus on the right hand. In the sketch it plays basically a gradual diatonic descent from c'' to f ', then a quick return in eighth notes past c'' to d('' . This is then articulated or elaborated in different ways in the several finished states. The greatest change takes place in the Brandus autograph, for example version 2, bar 54, making the descent from a(' to f ' a chromatic one. And then especially the Mahlerian touch in version 3, bars 54 and 55, leaping from a(' to e' at the end of 54, and repeating the e' on the downbeat of the third bar and holding it for two long beats, dissonant against every note in the left hand, before the resolution to f ' on the third beat. Evidently in reading proofs for the Brandus edition Chopin changed the right hand at the end of bar 54 to make a smoother descent. For the Breitkopf and Wessel editions this passage is toned down, with the e' resolving to f ' on the second beat. And so it is through the descendants of the two versions. Altogether the Chopin examples show, I think, a fluidity from the processes of composing and inscribing right through to performing. They tend to loosen the image of a hierarchical order from work to score to performance, and they suggest the possibility of a shared ontological level for all three—the possibility that the work may be realized in the performance as well as in the writing down. These things are said not just in the sense that one needs a performance or a score to make the work present or concrete. If you want to make an edition of the trope Filius ecce patrem or Chopin’s B major Nocturne that identifies the work, not just exemplifies it, you must include all the versions. If you want to analyse Chopin’s G
316
H i st ory a n d t h e O n t ol o g y of t h e M u s i c a l Wo rk
flat major Waltz, you ought to take cognizance that the work you are analysing has a fluidity about the form in the written and performance traditions. At the conclusion of the original presentation of this paper to an audience of philosophers I was asked ‘what are the limits? How much can a piece be tinkered with in writing it out or in performance and still retain its identity? Suppose . . .’ and there came another invented example. In response I declared that I decline to participate in any procedures aimed at a claimed unequivocal answer. And I urged that answers be sought, not only by testing the limits through introspection about invented examples, but also by taking into account the actual practice and beliefs of practitioners. That evidence suggests not only that answers vary with the compositional tradition, the composer, and even the individual work, but also that answers with reference to the same works or compositional traditions can change over time and under different historical circumstances. Pianists today do not depart from the published scores of Chopin’s works, and they are usually unaware of the extent to which their predecessors in earlier generations did so.²⁵ But I do not mean to undermine the enquiry. Such differences about the conception of the musical work and its relations to scores and performances are markers of music history. They constitute one of the major dimensions in which music even has a history. So the subject should be of great interest to music historians, just as the music historian’s evidence should be of great interest to the philosopher. ²⁵ This suggests a paradoxical question. Does the modern pianist’s complete faithfulness to Chopin’s scores amount to a disregard of his intentions, or an aspect of inauthenticity?
Th e E a r ly H i st ory of M u s i c W r i t i n g
317
chapter 13
The Early History of Music Writing in the West
} I The focus in this chapter and the next shifts to an aspect of musical transmission that has been hovering around our subject from the beginning, intruding itself forcefully into the foreground from time to time. I mean the invention and early uses of technologies for the representation of the melodic aspect of song in writing. Whoever thinks about that achievement, whether as author or reader, must do so through a veil of strangeness that is as daunting as any that the music historian confronts. We—who are accustomed to musical scores as the signs in which composers encode works, as the instructions that guide performers, and as the objects of analysis and comparison for scholars and students—must strain to understand how a distant and complex musical culture that we view as the progenitor of our own thrived without the use of scores. Then we must try to imagine how, in the course of its history, systems of musical notation were invented and came to mesh with the age-old systems for making, remembering, and disseminating music through performance alone, and at the same time to influence the transformation of those systems. This is as much as to say that we must try to think ourselves outside our own habits of musical thought and practice—no small task. The impediment is our continued dependence on the construction of the institution of music formed around the idea of works as texts given notational specification in scores and given acoustical explication in performance events that comply with the specifications of the score.¹ Despite the highly specific historical context and limited reference of this scheme it tends now to function in all fields of musical scholarship except for ¹ Regarding the constellation of concepts surrounding the work concept and the history of that constellation see Goehr, ‘Being True to the Work’ and The Imaginary Museum of Musical Works. Carl Dahlhaus has written about the work concept as a premiss of our historical view of the same 19th-c. music culture that produced it, in the introduction to his book Die Musik des 19. Jahrhunderts (Neues Handbuch für Musikwissenschaft, 6; Wiesbaden, 1980).
318
Th e E a r ly H i s t ory of M u s i c W ri t i n g
ethnomusicology—that is, in music history, theory/analysis, and music aesthetics—as the representation of the way music is. Subordinating medieval material to these conceptions and expectations has meant evaluating early medieval notational material according to the measure in which it meets the standard of precision in a prescriptive denotation and envisaging performance in a scriptless musical world in terms of two satellite conceptions of the work concept: either as the reproduction of an autonomous text that has been deposited in all its completeness in memory, as a written text would be deposited in a book, or as an improvisation produced through the free exercise of the musical fantasy, speaking with ease of the replacement of one chant tradition practised in a vast area of Europe by another as though that were a matter of trading in old books for new ones, without considering that in a scriptless culture that would have meant not only learning and remembering a complex and prolix new musical system and its repertory, but purging the habits of an old one; making a priori judgements about the nature and limits of the human mind’s creative and retentive capacities without the support of writing; and measuring medieval melody by the critical standards associated with the work concept—closure, unity and not least autonomy, particularly vis-à-vis language. The effort to avoid the use of such templates as constraints on medieval materials, though far from easy, will yield a more objective portrayal, and it will at the same time serve as a reminder of the narrow historicality of the work concept and its satellites. The topics of transmission and notation are nowadays regarded as being naturally interrelated in the most intimate way, but that has not always been so. As early as the mid-nineteenth century the question of the origins of the neumes was among the most prominent topics of discussion in medieval music studies, consistent with the orientation of music history in general to questions of origins, but especially given the regard in which medieval chant was held as itself the original of a European musical tradition, and given that the chant was known and discussed only in the context of its written transmission. But, as Wulf Arlt has written, The discussion about the ‘origin of the neumes’ ran aground because it was fixated—owing to a disposition towards developmental history—on the idea of pushing the new mode of inscription back to an older phenomenon of signification: from lesson signs and other directions for text performance, through cheironomy, to the notation systems of the Eastern churches. In no case could a derivation be established with compelling force and to the exclusion of other phenomena; yet the question of derivation stood so prominently in the foreground that the possibility of a qualitative leap into writing, in which the most widely differing factors played a role without any one of them being identifiable as the single most relevant precondition, was never considered.² ² ‘Anschaulichkeit und analytischer Charakter’, 29. For reviews of the principal theories of origin see Solange Corbin, Die Neumen, Palaeographie der Musik, i/3: 11–21; Jacques Hourlier, ‘L’Origine des neumes’, in J. B. Göschl (ed.), Ut mens concordet voci: Festschrift Eugène Cardine zum 75. Geburtstag (St. Ottilien, 1980), 354–60; and Michel Huglo, ‘Bilan de 50 années de recherches (1939–1989) sur les notations musicales de 850 à 1300’, AM 62 (1990), 224–59. Regarding the ‘cheironomy’ theory of
Th e E a r ly H i st ory of M u s i c W r i t i n g
319
Renewed reflection on the early history of neumes has become an integral and necessary part of the problem complex that surrounds the central question about which there is widespread discussion, over the degree to which the melodic tradition of Gregorian chant had become fixed prior to the inscription of the melodies in musical notation. We now recognize the following as distinguishable but integrally related topics: the composition and dissemination of an unwritten tradition of Gregorian chant in both Frankish and Roman transmissions; the invention of systems of musical notation and their development as they were adapted for the creation and dissemination of a written tradition of the chant; ‘unwritten’ and ‘written transmission’ in performance and redaction I cite this conventional wording not as an endorsement but as a reminder of the ways in which our modern tools of language and concept, because of the presumptions that they embody, stubbornly interfere with the efforts to gain access to the medieval material. The formulation implies both a parallelism—that written and unwritten transmission are both processes that do the same sort of thing, that is, transmit something—and an opposition: that they do so in different, mutually exclusive ways (as one might speak of conveying a message by ‘voice-mail’ or ‘fax’). But both of these implications will be misleading at least some of the time. ‘Written transmission’ presupposes an object, something transmittable. Our most familiar image for this idea is that of someone copying a text from a model, and that is reflected in the habit of saying that a manuscript was copied at such and such a time and place, even if our evidence allows us to say only that it was written then and there. The idea of something transmittable is still more problematic in the case of ‘unwritten transmission’. If it is a well-known hymn that is being transmitted note for note through a stable performance tradition, then the formulation seems apt. But if it is a trope or an organum melody that we have every reason to think was extemporized in performance, then the object has not been transmitted through performance, it has been realized in performance. It comes down to the ontological status of the musical item, and with respect to that the implied parallelism in the binary formulation is not a parallelism at all. What is more, the phrase ‘unwritten transmission’ tends from the first to suggest the idea of performing from a mental repository ( = ‘memory’) of fixed melodies, as the only respectable alternative to an idea of undisciplined improvisation in the duality to which I referred at the beginning. The implied opposition is no more reliable in actual practice. From the very beginning of a written tradition reading, remembering, and extemporizing were continuous acts; they were mutually supportive and interdependent.³
origin see Helmut Hucke, ‘Die Cheironomie und die Entstehung der Neumenschrift’, Mf 32 (1979), 1–16. ³ This attitude is fully and persuasively developed with respect to language texts in Mary J. Carruthers’s The Book of Memory.
320
Th e E a r ly H i s t ory of M u s i c W ri t i n g
II This chapter has four objectives: (1) to distinguish between the analysis of neumatic music writing from the standpoints of palaeography, on the one side, and semiotics, on the other; (2) to suggest what the characteristics of a semiotic analysis of music writing would be; (3) to argue for the importance of a semiotic approach, if we are to integrate the scientific study of musical notation into the historical understanding of the musical cultures that it served, and if we are to have a plausible view of the evolution of musical notation; (4) to sketch a provisional outline of such an evolutionary view, with due care for the sense of ‘evolution’ that can be applied to this material. Some brief remarks about terminology are necessary. As I am writing in English I use the word ‘semiotics’, following the American philosopher Charles Sanders Peirce, whose publications on this subject date from 1867 and after.⁴ It corresponds to the French word sémiologie, introduced by Ferdinand de Saussure in his Cours de linguistique générale (1916). Both terms refer to the discipline that concerns itself with the functional relationship between sign systems and what they signify, in the particular communication contexts in which they function. As Saussure defined it, semiology is ‘a science that studies the life of signs within societies’. It would ‘show what constitutes signs, and what laws govern them’.⁵ Within the general domain that both Peirce and Saussure identified they had somewhat different interests and their work reflects that in differences of emphasis. What I concentrate on here, however, is the approach that they defined in common. In contrast to that approach, the main concern of palaeography is the classification of the signs of past writing systems according to their morphological features and the chronology and provenance of their use. Of course the two concerns overlap, but there are fundamental differences of emphasis and ultimate objective, and it is those that interest me here. In 1963 Dom J. Hourlier proposed the application of the term sémiologie in the analysis of chant sources, to distinguish the interpretation of notational signs, especially in the sense of their rhythmic and expressive signification, from the description and classification of signs in the traditional sense of palaeography.⁶ This proposal was unfortunately restrictive and laid the groundwork for a terminological confusion. Dom Cardine took up the proposal, and as one sees in his Sémiologie grégorienne⁷ the result, however great its value, remains an extension of the domain of Gregorian palaeography. Both publications seem strangely oblivious to the enterprise for which Saussure coined the term, and Sémiologie grégorienne is in no ⁴ Published in Collected Papers of Charles Sanders Peirce, ed. Charles Hartshorne and Paul Weiss (Cambridge, Mass., 1931–58). See especially A Second Trichotemy of Signs, i, pt. 2, 142–3. ⁵ The translation is from the English-language edition, Course in General Linguistics (New York, 1966), 15. ⁶ Sémiologie musicale = EG 6 (1963), 153–7. ⁷ EG 11 (1970), 1–158.
Th e E a r ly H i st ory of M u s i c W r i t i n g
321
way the counterpart of the semiology that is now being vigorously pursued in such fields as anthropology, linguistics, and the study of literature. Be that as it may, nothing would be gained by further discussions of terminology alone. My purpose is to press the case for a study of musical notation in the sense common to Peirce’s ‘semiotics’ and Saussure’s sémiologie. Although the latter term has been pre-empted by Sémiologie grégorienne, the former remains available to us, and in several languages: sémiotique, Semiotik, semiotica. I can introduce the approach through the consideration of a specimen that is not an example of neumatic writing at all. It will allow me to point in an exaggerated way to the differences in levels of meaning that are accessible to palaeographic and semiotic investigation. Figure 13.1 is taken from an Italian manuscript of the fifteenth century originating in Cividale del Friuli.⁸ It is a notation of a widely disseminated sacred song, given here in two parts on facing pages. As a palaeographic item it presents certain interesting and anomalous features. First, there are the designations of the two parts as Tenore and Supranus. This is common terminology for the fifteenth century, and it is consistent with the style of the letters in which those words are written down. But considered in their meanings, those words do not give good descriptions of the vocal parts for which they are labels, for both parts occupy the same vocal range and cross frequently. Second, in the Tenore part, especially, there are vertical strokes that resemble the marks for rests in mensural notation. But if they are read as such they will produce distortions in the declamation of the words, not to mention strange rhythmic effects and the impossibility of coordinating the voices. Even more, the Supranus part has note shapes corresponding to the longs, breves, and semibreves of mensural notation. But the notes of the Tenore are all square, and the setting is note against note. So it cannot be that the notational figures of both parts were meant to be read as the mensural figures they resemble. It is more than likely that the notation only mimics the graphic forms of the mensural system, without really indicating what those forms were designed to convey to performers. In fact the example is a piece of cantus planus binatim (literally, ‘binary plainchant’, in which plain means unmeasured ) notated in fancy dress. It is a palaeographic anomaly. That raises the question, what does it mean that such written artefacts were produced? What is signified by the fact of their production?⁹ We are used to concentrating our attention on the contents of written messages as the source of their ⁸ Berlin, former Deutsche Staatsbibliothek, Ms. mus. 40592, fos. 179v–180r. Reproduced from G. Vecchi and A. Gallo, I più antichi monumenti sacri italiani (Bologna, 1968). ⁹ This question is posed from the vantage point of semiotics. It is at the same time the sort of question that is asked about language utterances from the vantage point of the speech-act theory of language usage. A brief example: if I say to someone in my household on a hot and muggy summer evening ‘the door is open’, it is not merely with the intention of conveying that obvious piece of information, but more with the intention of inducing her to close the door in order to prevent the room from filling with mosquitoes. See Kent Bach and Robert M. Harnish, Linguistic Communication and Speech Acts (Cambridge, Mass., 1979). Just as musical notation could serve as the non-linguistic model for semiotic theory, it could well serve as the non-linguistic model for speech-act theory.
322 Th e E a r ly H i s t ory of M u s i c W ri t i n g
Fig. 13.1. An example of cantus planus binatim in Berlin, former Deutsche Staatsbibliothek, 40592, fos. 179v–180r. Reproduced from Vecchi and Gallo, I più antichi monumenti sacri italiani
Th e E a r ly H i st ory of M u s i c W r i t i n g
323
meaning. But here we cannot even believe the written signs in what they seem to say about aspects of their contents. There seems to be a false relation between sign and signified, whereas, on the other hand, the reference that the signs make to other signs (i.e. to the signs of the mensural system) seems itself to be a source of meaning, without any regard for the musical contents of the notated score. In their mimicry of the forms of notation used in centres of high musical culture, they seem to say something like the following: ‘The sacred chant contained in this book is presented in the lofty forms of the great musical centres, of which it is worthy, and the possession of it brings honour to this establishment.’ It is not only music that may be conveyed through the medium of musical notation, as this case shows us. The false relation between sign and signified suggests that the book functioned more as an emblem for the music that it ‘contains’ than as a guide to its performance. And as to performance the notation tells us that singers would precisely not have relied on it but rather on their oral tradition.The mimicry of mensural notation in the transmission of music of an oral tradition is, in turn, a clue to a musical culture that functioned on two levels: one a literate, progressive culture cultivated in centres of wealth and power, the other a widespread, conservative, and vernacular culture. So while we cannot admire this example greatly for the accuracy with which it conveys the contents of a musical signal, we can nonetheless value it for what we can learn from it about the musical culture in which it and others like it participated. If we were simply to dismiss the example and others like it as defective it would be because of an expectation that the primary or exclusive function of musical notation is to serve as a non-distorting window on the music with which it is associated. That is an expectation that we have after centuries of practice, and it would be unhistorical for us to assume that it was bred into musicians from the time of their first encounter with notation. The possibility that ‘it is not only music that may be conveyed by musical notation’ has been brought home most sharply in a recent study of a rather remarkable phenomenon, the existence of some ninety manuscripts dating from the ninth to the twelfth centuries in which substantial passages of classical Latin poetic texts were provided with neumes.¹⁰ Without aiming necessarily to undermine the leap from this to the supposition that those texts were sung, the author of the cited study identifies reasons to hesitate, pointing out that neumes were used in medieval books not only to mediate information about music but also to highlight accent or stress in pronunciation, syntactic function or word order in grammar, and emphasis in rhetoric, to guide readers to marginal glosses or to corrections in the margin, and even to stand for letters of the alphabet. Adding to the enigmatic character of the phenomenon is the fact that most of the neumations of classical texts are nondiastematic, which would imply an established oral tradition of which, however, there seems to be no other indication. And then most of the neumations are added to written texts that were not prepared for them, being squeezed in between lines of text or even written in margins. That could be taken as an indication that the text ¹⁰ Jan Zielkowski, ‘Nota Bene: Why the Classics were Neumed in the Middle Ages’, Journal of Medieval Latin, 10 (2000), 74–114.
324
Th e E a r ly H i s t ory of M u s i c W ri t i n g
scribe assumed an oral melodic tradition, or, on the other hand, that neumators entered the neumes for some other purpose. At this moment the question remains open, and that is just the point. There are reasons not to leap to assimilate this phenomenon to the medieval song traditions of sequence, versus, and conductus and restrict the study of it to the control of the questions we are accustomed to raising about those traditions. By keeping the question open we may exploit this relatively neglected but prominent phenomenon to learn unexpected things about the early history of musical notation.
III The essential lesson of the example from Cividale is this: in a communication system in which messages are conveyed through conventional signs, neither the analysis of the intrinsic properties of the signs, nor the analysis of the contents that they convey, or even the analysis of those two factors jointly, necessarily exhausts the meaning that may lie in an item. This lesson has been set out systematically in Roman Jakobson’s classic article ‘Linguistics and Poetics’.¹¹ (I preface my brief summary of the essential idea of that article with the observation that, while semiotics developed initially with reference to language, it has been explicitly aimed towards the investigation of all systems of communication through signs, with language as model. Musical notation could well serve as the non-linguistic model.) Jakobson observed that there are numerous contributing factors at work in any communication situation. For speech events he identifies six such factors. Speaker and listener (more generally, addresser and addressee) require no comment. A third factor is the context of the speech event. This volume, the general domain of the topics that it encompasses, its history, the scholarly community that it addresses— all are aspects of the context in which the words I write here have a certain meaning. In a semiotic analysis of musical notation questions of context assume great importance. They will be questions about the competence of the addressees vis-à-vis the music, about musical traditions represented by the notation, about how music is conceived by both addressers and addressees, about the kinds of books in which notations are written or printed, the kinds of institutions in which they are used, and the functions they serve in them, etc. A fourth factor is the code used in the communication (language in speech events, a notational system in written musical communications). A fifth factor is what Jakobson calls a ‘contact’, which we emphasize, for example, when we take a microphone and say ‘testing 1 2 3 4’, or when we remark about the weather to an attractive stranger in the hope of opening a conversation about something that we are really interested in. In a way the introductions to these chapters, which are meant to draw the reader’s attention to aspects and issues in the chapters them¹¹ Published in Thomas A. Sebeok (ed.), Style in Language (Cambridge, 1969); repr. in S. Richard and Fernande de George (eds.), The Structuralists from Marx to Lévi-Strauss (New York, 1972), 85–122.
Th e E a r ly H i st ory of M u s i c W r i t i n g
325
selves, are contacts. The features of the musical score from Cividale play such a role. And finally there is the message itself, that is its contents. Each of these factors determines a different function of the language of a communication system and, as Jakobson writes, ‘Language must be investigated in all the variety of its functions’. We can readily see that the same six factors operate in a musical score as well, but that, as Jakobson observed about the study of language, it has generally been presumed that the message is of primary, even exclusive interest. This amounts to the presumption that the communication event or item has just one function, what Jakobson calls the ‘cognitive’ function, conveying the contents of the message. But according to the various relationships that may be stressed among the factors in the particular case, different communication functions may be identified. Emphasis on the relation between the addresser and the message brings out an expressive function, on which we often focus in the analysis of art. Emphasis on the relation between the message and the addressee brings out a connative (or we might say a vocative) function (e.g. in prayers of supplication). In our fifteenth-century musical example there is an emphasis on the code itself (the notational system) in relation to the context. Usually we try to read through written signs to the contents of what they convey, without paying much attention to the signs for themselves. But here the addressee stops at the signs to read meaning out of their very morphological features. It is similar to the use of musical notation in Flemish paintings representing musical performances. The viewer is not primarily interested in the contents of the musical score represented. What is emphasized in both cases is what Jakobson calls the ‘metalingual’ function. It is the recognition of the metalingual function in our example that provokes further questions about the meaning of the notation. Why did the scribe mimic his mensural models? It is rare that a communication event involves just one of these functions. The several functions operate jointly, in a hierarchy of importance. It is the task of analysis to identify all the functions at work in a communication event or item, and the relations among them. This broadening of attention from the central concern with the signs and contents of messages to the entire array of communication functions and their hierarchical relationships is the aim of semiotics. In a subsequent article, ‘Language in Relation to Other Communication Systems’,¹² Jakobson identified as the ‘pragmatic’ approach to language a study of the cardinal functions of language and their hierarchical relations in the diverse types of communication. ‘This pragmatic approach’, he wrote, ‘must lead mutatis mutandis to an analogous study of other semiotic systems: with which of these or other functions are they endowed, in what combinations, and in what hierarchical order?’ The answers to these questions will be the form of understanding about the multifarious and variable relations between signs and what they signify in particular semiotic systems. It is the principal task of semiotics to investigate those relations, rather than the properties of the signs themselves—that is, to investigate how they are able to function as signs. For the ¹² In Selected Writings, ii (The Hague, 1971), 697–708.
326
Th e E a r ly H i s t ory of M u s i c W ri t i n g
study of musical notation that would be the basic difference between a palaeographic and a semiotic approach. Saussure proceeded on the premiss that, in so far as human action or productions function as signs to convey meaning, there must be an underlying system of conventions and distinctions that makes this meaning possible. Semiological analysis, for him, is a study of systems, not of isolated signs. The strategy of semiology entailed most basically a shift in focus from objects to relations, or from objects to structures—hence Saussure’s role in the founding of structuralism. This chapter exemplifies Jakobson’s and Saussure’s emphasis on the systematic aspect of sign systems with specific reference to Western neumatic writing. I report a number of observations aimed at identifying the neumatic scripts that have long been recognized by musical palaeography, as systems of consistent differentiation, and then I pose the semiotic task as the investigation of the principles that underlie their systematic differentiation.
IV I have omitted from the revised version of this essay section VI of the original version, in which the early history of music writing was compared with the history of language writing, because that comparison was based on a number of premisses that I now regard as false and because it has misleading implications for my subject. First, I had adopted a widely held view of the history of language writing as a stepwise evolution in the signs themselves from pictographs to letters and words. Associated with this view was the assumption that at all stages of this evolution writing had represented speech (that premiss is conveyed in a part of the first section of the original essay that I have also omitted here). The linguist Roy Harris has demonstrated the problematic nature of both premisses on evidentiary grounds, while also emphasizing the distorting nature of such evolutionary accounts as historiographic models.¹³ The palaeographer David Ganz has shown that Carolingian scribes in particular—their attitudes would be most relevant to the understanding of my subject here—did not conceive of writing as a transcription of speech.¹⁴ In taking that evolutionary account of the history of language writing as a standard of comparison, I tacitly projected the implication of an evolution in the sense of a gradual transformation in the signs of music writing—the neumes themselves—from one type to another, whereas the survey on which I report merely shows the increasing use in particular places of one type of neumatic representation over another. It is a matter of particular changes made in particular places, rather than a gradual and uniform transformation of ‘neumatic writing’ writ large. The issue is quite the same as that of the evolution of species in biology.¹⁵ The representation of the history of both ¹³ Roy Harris, The Origin of Writing (LaSalle, Ill., 1986), ch. 3, ‘The Evolutionary Fallacy’. ¹⁴ ‘The Preconditions for Caroline Minuscule’, 40. ¹⁵ ‘Change in a species is not a matter of the uniform transformation of each of its members—the Harvard class of 1950 growing grayer as the years pass, the legs of the flamingo growing longer as the
Th e E a r ly H i st ory of M u s i c W r i t i n g
327
language and music writing in terms of such a transformational evolution has the misleading implication that writing has always served the same task throughout its history. That is not the case with respect to the history of music writing. Finally, there is a real question whether the systems of writing language and music, considered as communication technologies, are at all comparable in any heuristically useful way. For example, it was tempting to see a correspondence between pictographs and neumes that can be seen as representing a melodic line by shadowing its contours—regarding both, that is, as iconic representations. But that was too hastily done. Pictographs depict objects or actions or scenes, not the words with which they would be described. Only if they did that would it be useful to see them as analogous to neumes that represent the movements of the voice in tracing melodies.
V One of the purposes of these introductions is to register the bearing of more recent work in the field on the interpretations presented in the essays. To that end I need to draw attention to the very great relevance of some astute observations and original thinking by Susan Rankin on the subject of this chapter. Its immediate relevance can be seen by focusing on Table 13.1 below, which is the ground for the systematic analysis offered in the chapter. First I emphasize the word systematic; the table’s entries are neither comprehensive nor historically differentiated. They are there to demonstrate what I call the complementary distribution of the aggregate of neume forms into two natural classes that suggest different approaches to the task of representing melody through the use of dots, lines, and curves drawn on parchment. (‘Complementary distribution’, put simply, means that the classes are mutually exclusive.) There is something potentially misleading about the table as it is. Nearly every one of the regional notational styles that I call ‘scripts’ operates with a fixed set of signs that designate standard melodic gestures. So in this chapter I define ‘script’ as ‘a system of standard neumes used by a community of practitioners defined by a common liturgical practice and usually by geographic contiguity’. Usually the names given to the scripts in modern literature refer to that factor of place. Rankin has observed, however, that the script called Palaeofrankish (the name given it by Jacques Handschin in an article described in the essay) does not answer to this definition, especially in its early manifestations. In a very brief discussion she writes waters in which such fowl feed get deeper, the upper voices of organa growing more mobile as... Change is a result of individual variation. That means it need not occur at uniform rates. Such a view of human practices accommodates creative invention, accident, and collective behavior as sources of variation. From such a view the outcome of variation is neither inevitable nor foreseeable. It requires analysis of the particular circumstances under which particular changes occur.’ Treitler, Music and the Historical Imagination, 166. This has application to every subject of music history that has been addressed from the point of view of ‘evolution’—‘the symphony’, ‘organum’, ‘the string quartet’, ‘mensural notation’, etc.
328
Th e E a r ly H i s t ory of M u s i c W ri t i n g
that this script ‘has no such fundamental signs, but instead the principle of joining positions in space which represent notes. . . . Symptomatic of this most extreme iconic . . . approach is this notation’s lack of consciousness of a sign system. In the earliest example . . . the notator used six different ways of expressing the melodic pattern within the space of twelve short lines of notation.’¹⁶ Other scripts use one or two figures to represent that pattern. The impression is that the writer of Palaeofrankish neumes just used whatever way that occurred to him to connect three positions in the vertical dimension. The script had set procedures but not a set repertory of signs, at least in its early manifestations. Handschin, in naming it, implied that it was a kind of proto-notation. That is a supposition that so far seems supported more by the habit of always searching for an originating form (and the assumption that there must be one) than by hard evidence. In any case the Palaeofrankish is not a script in the sense that the others are scripts. It is worth quoting Rankin’s concluding paragraph to the short passage I have cited in order to suggest the wider relevance to the subject of this chapter. The paleofrankish neumes belonged in the very heart of the Carolingian realm, associated above all with St. Amand, but found also on the eastern side at Corvey. That they could co-exist alongside the other developing scripts is only another mark of the Carolingians’ inventive, highly experimental attitude to musical notation. Any attempt to refine a chronological model of the emergence of musical notations beyond that which we already have (which is unable to link specific notational techniques with specific periods before 900) will have to admit heterogeneous chaos as a fundamental characteristic of the situation throughout the period of Carolingian rule.¹⁷
} Our perspective on the early history of Western music has been changing somewhat lately in consequence of the serious attention being given to the fact that the earliest music that has come down to us—the Gregorian chant tradition—was, in its early centuries, an oral performance practice. In this chapter I turn to the other side of that picture: the nature, origin, and evolution of the early practice of writing music, which, in medieval Europe, began in the ninth century.¹⁸ The writing of music evolved over a span of centuries from the graphic representation of performance practices and the illustration of theoretical and peda¹⁶ ‘Carolingian Music’, in Rosamund McKitterick (ed.), Carolingian Culture: Emulation and Innovation (Cambridge, 1994), 274–316 at 302. ¹⁷ Ibid. 303. ¹⁸ I have not been persuaded by Kenneth Levy’s insistence on the existence of a fully developed notational system by 800. See Ch. 6 n. 17, where I cite literature regarding Levy’s interpretation of his evidence and question the significance that he invests in such a date. The burden of this chapter, however, would not be affected if we were to accept it.
Th e E a r ly H i st ory of M u s i c W r i t i n g
329
gogical treatises to the enabling technology for composing, transmitting, and performing music. Written sources originating during the early centuries show many signs of the continuing activities of oral processes along with written ones in the transmission of music, confirmation that writing was for a long time in support of, rather than in competition with, the oral performance tradition. The act of writing was thus a kind of performance analogous to singing out, and the written score served as an exemplification of the song, to be taken more as a model for performance than as a blueprint.¹⁹ In the framework of such interpretations, old questions about the nature, function, and early history of music writing present themselves in a new light. If we are to do more than to hold these ideas as intriguing speculations, we must focus on the question, how did systems of musical notation function? On what principles does the graphic representation of melody—or of the performance of melody— rest? What changes in the nature and function of notation correspond to these long-range changes in the relations of transmission? Such questions must lead to the study of notations in the light of their use in particular conditions: who uses them (that is, who writes and reads them), and for what purposes? What are their modes of representation? What was the conception of the musical objects that they represent? What sort of knowledge and competence did the reader require in order to be able to use them successfully? As we pose such questions the focus in the investigation of early notations must switch from palaeography to semiotics. Musical palaeography, which concerns itself with the classification of signs and the identification of their periods and places of use, has been a mature discipline since early in the last century. In the semiotics of musical notation, which would concern itself with the functional relationships between sign systems and what they signify while taking into account the situation of the person(s) to whom they signify, virtually everything remains to be done.
I How do we think about the way musical notation functions? I propose to approach the question from the vantage point of our commonplace pitch notation. The purpose will be to set out some general concepts and terminology. The first premiss of our system is the metaphor that the pitch spectrum has vertical extension, that pitches of higher and lower frequencies occupy higher and lower positions respectively along the vertical dimension, and that the magnitude of intervals between two points of the spectrum is proportional to the lengths of ¹⁹ Full expositions of these interpretations, and citation of further literature, are given in Hucke, ‘Toward a New Historical View’, and Ch. 10 of this book.
330
Th e E a r ly H i s t ory of M u s i c W ri t i n g
the segments of the vertical subtended by those points. By mapping the vertical dimension onto the writing surface we make a visual analogue of the pitch spectrum. This is concretized as the stave or system of staves, which also has horizontal extension in order to provide for the representation of pitch succession through time. The stave is in effect a sign for the pitch spectrum extended through time. The philosopher Charles Peirce provided a useful and influential conceptual apparatus for the analysis of the functioning of signs in terms of their modes of representation.²⁰ What I have just described about the modern system for representing the extension of the pitch spectrum in time would be identified, in Peirce’s terms, as a symbolic representational mode, representation by virtue of a habitual association between the sign and its referent. There need not be anything inherent about the sign itself that suggests what is represented in order for it to function in the symbolic mode, only the consistent habit of having it stand for that thing. An example would be the male and female signs, ª and ¬. No one who had not been informed about that convention could be expected to guess at the meaning of those signs. The conception of the pitch spectrum as an array along the vertical dimension and its mapping onto a writing surface are arbitrary and conventional, and they constitute a symbolic component in our system of pitch notation.²¹ In placing ‘notes’ on the stave we are in effect picking out positions in the twodimensional field determined by pitch and time. A single note identifies a position in the pitch spectrum. The representation is still symbolic. A succession of notes, representing what we call a ‘melodic line’, identifies a sequence of such positions and represents those pitches, conceived or sounding in succession. But such successions—of signs and sounds—are transformed into something else, to which the common expression ‘melodic line’ provides a clue. That phrase is a metaphor ²⁰ Peirce, Collected Papers. See also Roman Jakobson’s further articulation of Peirce’s idea in ‘Language in Relation to Other Communication Systems’. A summary is given by Terrence Hawkes in Structuralism and Semiotics (Berkeley and Los Angeles, 1977), 123–50. ²¹ Objections can be raised against two assertions in this paragraph. The first is the claim that there is nothing about the forms of the male and female signs that directly suggests their symbolic meanings. Those who know will recognize them as astrological signs representing Mars and Venus, and may see in the arrow of the masculine sign, for example, an aggressive, war-like image that transfers easily to the idea of masculinity; for them the sign is not arbitrary. This is something like reading the sign of the bass clef. For most readers there will be no link other than a habitual one between form and meaning. But some will recognize in the sign a vestigial letter ‘F’, the reference note for the clef. This demonstrates a most important point that I shall wish to make again later: the representational mode of a sign or sign system depends on both its form and the experience of the reader. Sign systems do not function in a single, invariable mode because they are communication media, and what they say depends in part on the condition of the receiver. The other contestable claim is that the designation of pitches as ‘high’ and ‘low’ and their representation by means of signs in ‘high’ and ‘low’ positions on the page is conventional. To some, the spatial representation of pitch will seem natural, based perhaps on the feeling that high pitches lie (or resonate) in high parts of the body and low pitches in the lower parts. But what is certainly not ‘natural’ is the association between what is called ‘high’ in respect to the pitch spectrum—for whatever reasons—and what is called ‘high’ in respect to greater or lesser distance from the observer on the horizontal plane of the page. That association is conventional.
Th e E a r ly H i st ory of M u s i c W r i t i n g
331
that captures our sense of melody as more than a succession of pitches. We express that sense also by speaking of ‘melodic motion’, a conception that resonates with the medieval conception of melody as the voice in motion. That is, we conceive of melody as a single, continuous thing, and we read the notational representation of melody as a graph or picture of melody, in which our mind’s eye or ear connects the dots. In representing melody by drawing a picture of it, the notation is functioning in an iconic mode, and that is the second of Peirce’s modes of representation. A sign functions in the iconic mode when it represents, by virtue of a resemblance that it bears to the thing represented, an isomorphism of some sort between sign and referent. The international traffic sign for ‘One-Way Street’ represents purely symbolically. Signs warning of a cattle- or deer- or kangaroo-crossing have a clear iconic component. One recognizes the referent without having to learn it. What requires learning is that creatures of the type pictured can be expected to cross the road in that vicinity, and we would do well to drive in such a way as to avoid colliding with them. If we have not learnt that, we cannot really tell what the sign is saying to us. The mode in which a sign does its job of representing is not a function exclusively of the characteristics of the sign itself. No analysis of signs as signifiers can be adequate that does not take into account both the active involvement of the beholder in the process of representation and the nature of the referent. When we read our pitch notation in performance, the iconic aspect facilitates the reading. Rather than translating each sign back to its referent in the pitch spectrum, we follow the contour of the line as a whole and duplicate it in our performance. We can do that more readily, the more the melodic patterns of the idiom are conventionalized and familiar to us. For most Western readers of modern Western notation the score of a tonal melody will function primarily in the iconic mode, whereas in the score of a dodecaphonic melody the symbolic component may be more prominent for most such readers, who will more nearly work out the score note by note. But for specialists in the performance of non-tonal music that may not be so. It is a special feature, and it has proven to be an advantage, of our system of pitch notation that it has this versatility with respect to the information conveyed by both its symbolic and iconic aspects. From our own experience with it, it seems reasonable to attribute to the iconic aspect the notation’s potential for fluent reading, which lies in the fact that it can be grasped in configurations of simultaneity and succession. By contrast, most Oriental pitch notations function primarily in the symbolic mode—the pitches are names, and the notational signs are symbols of those names (like some early letter notations in the medieval West)—and they do not lend themselves to such swift reading. They lack the isomorphism with
332
Th e E a r ly H i s t ory of M u s i c W ri t i n g
musical configurations that is the underlying principle of the Western system. Such notations can be very precise, but they are not made for fluent reading.²² Peirce identified a third mode of representation, which he called the indexical mode. We can bring that into view by returning to our two traffic signs, the oneway sign and the animal-crossing sign. They not only identify, they call for an action or a behaviour: not to enter the one-way street, or to proceed with caution so as not to collide with creature of the type shown in case they should be crossing. An index is a sign that represents by virtue of a sequential (and usually causal) link between the sign and what it signifies. It can function descriptively—wet streets as a sign that there has been rain, smoke as a sign of fire, or the implied prediction in ‘red sky in the morning, sailors take warning’. And it can function, as in the last case, as an imperative—a road sign, the bell that once rang on our typewriters as we approached the end of each line, punctuation marks. Musical notation can have an indexical component in this sense. It is most apparent in tablatures, which tell the player where to place his fingers on the instrument. But in varying degrees people read even staff notations that way, translating the signs directly into finger movements. Fingerings in scores increase that tendency. The indexical mode is the third member of the triad of representational modes that are hierarchically arranged in the functioning of every system of signs, and the exact hierarchical arrangement will depend on the way the system is used, the nature of the referent field, and the competencies of the readers.²³ Drawing all these observations together, the understanding of a musical notation requires that we see in it a system of signs working through a hierarchy of modes of representation whose composition will be a function of the use that is made of the notation, the characteristics of the music to which it refers, the relationship between that music and its practitioners, and the types and degrees of ²² Walter Kaufmann, Musical Notations of the Orient (Indianapolis, 1967). Byzantine melodic notation can be similarly characterized. Signs provided at the beginning of the chant indicate the mode and the place of the initial note within the mode. Then a sign is given for each succeeding note, indicating a number of degrees above or below the preceding note. Kenneth Levy (‘Byzantine rite, music of the’, NG iii. 554) calls this a ‘digital’ notation. The modern system, considered as a notation for discrete pitches, the Oriental systems that name pitches, and the medieval Latin letter notations are all similarly digital systems. In its graphic function of representing the contours of melodic lines, the modern notation can also be considered an ‘analogue’ system. What Peirce sought to capture with the dualism of ‘symbolic’ and ‘iconic’ can also be grasped from another side through the dualism of ‘digital’ and ‘analogue’ indicators. The sense that ‘iconic’ and ‘analogue’ have in common is also shared by Wulf Arlt’s concept of ‘Anschaulichkeit’. We are all three trying to catch an approach to the use of signs that depends on isomorphisms between the signs and what they represent. (See Arlt’s ‘Anschaulichkeit und analytischer Charakter’.) ²³ Among Oriental notations it is tablatures that have in the main been used in recording repertories for performance. Pitch notations have more often served in theoretical writing, for purposes of demonstration. This differentiation, as well as the decline in the use of letter notations in the West, is attributable to the general unsuitability of symbolic notations for fluent reading. Exceptions to my generalization about the symbolic nature of Oriental notations—undulating lines depicting vocal or melodic contours—are the Buddhist notational traditions of Tibet and Japan. See Ter Ellingson, ‘Buddhist Musical Notations’, in Tokumaru and Yamaguchi (eds.), The Oral and the Literate in Music, 306–41.
Th e E a r ly H i st ory of M u s i c W r i t i n g
333
competence of the practitioners. These variables can interact in different ways in a single historical situation, and certainly in different historical situations. It is this attention to the context and function of notations that distinguishes a semiotic approach from a palaeographic one. A conception of the historical development, if it is to be more than a description of the forms in a chronological sequence, can proceed only from the analysis of individual notations from such a semiotic standpoint.²⁴
II How did the early writers and readers of music conceive of what it was that their notations signified? To begin with, what we call melody seems earliest to have been understood as an attribute of spoken language, in particular scriptural language, as it was declaimed in the ritual of the Church. That is suggested even by the practice of referring to the performance of chant as ‘speaking’, as well as ‘singing’.²⁵ The earliest formal analyses that have come down to us derive the divisions of the melodies from the divisions of the words they declaim into sense-units. And the analyses are couched in the terminology of language structure. It would not be overstating the case to say that melody was conceived of and functioned as a projection of the prosody of language.²⁶ The elemental melodic phenomenon, in that conception, was the movement of the voice in declaiming a syllable of text; as it was also put, it was the movement of melody itself in that act. It was to such inflections that Aurelian of Réôme referred with the term ‘accentus’.²⁷ This interpretation is reinforced in the description by Notker of St. Gallen of the relation between words and music in the sequence: ‘Every movement of the melody should have a single syllable.’²⁸ That describes what we call the syllabic ²⁴ This approach to the study of notations reveals the inadequacy of the analysis of Nelson Goodman in Languages of Art, owing to his inexperience with the phenomenology and history of music writing and reading and with the structure of music. Thus his attempt to establish a single set of conditions for the functioning of a notation, with the consequence that there can be but one mode of notational representation under any and all circumstances; his failure to regard the users of notations and the purposes for which they are used as factors; his presumption that music is apprehended as aggregates of individual pitches and that is all one has to think about in order to comprehend notation; and his disregard of history. Nevertheless the general terms of his analysis are heuristically valuable if pursued with the sort of awareness and background that he seems to lack. ²⁵ e.g. in the liturgical trope Discipulis flamma for the introit of Pentecost there is the line ‘Ipsi perspicuas dicamus vocibus odas’ (‘Let us speak clear songs to him with our voices’). ²⁶ On this relationship see Chs. 2, 15, and 16. ²⁷ Aurelian of Réôme, Musica disciplina, ed. Gushee. Jacques Handschin thought that Aurelian referred to notational signs in his use of ‘accentus’, but it is quite plain from the context that his referent is the melodic inflection of a syllable. See Handschin, ‘Eine alte Neumenschrift’, AM 22 (1950), 69–97, esp. 69–73, and Solange Corbin, Die Neumen, 18. ²⁸ Jacques Froger, ‘L’Épître de Notker sur les lettres significatives’, EG 5 (1962), 23–72.
334
Th e E a r ly H i s t ory of M u s i c W ri t i n g
setting of the sequence. But in general it shows that the syllables and their inflections were conceived as the primary units of melody. It is those melodic inflections of the syllables that are the referents of the notational signs in the earliest Western systems of music writing. Another word for them was ‘neuma’, and the sign for a ‘neuma’ was a ‘signum neumarum’. By the eleventh century the word ‘neuma’ had passed over into use as the term for the notational signs themselves.²⁹ To be sure, the figures passed through individual notes, and that is reflected in the formation of the neumes. In alphabetic notations the aspect of pitch sequence is most to the fore. But in the overall picture these are quite exceptional (and in some cases even they are written so that the letters cluster in groups under slur-like marks, groups that match the neumes of corresponding neumations). All the evidence taken together converges on a conception of melody for which the first notations were invented not so much as a sequence of pitches, but as a succession of vocal or melodic movements accompanying the declaiming of syllables of text.³⁰ The information encoded in the early neumes was of two kinds: the coordination of melodic inflections with syllables of text (this follows from the conception of melody); and the directions of melodic movement within the inflections represented by the neumes (but not necessarily between them). The neumes did not at first encode information about the distance of movement, that is, the size of the intervals. That information, or better, the competence to supply it, would have come from the singer’s mastery of the grammar of the modes and melodic types.³¹ Given the conception of melody at that time, such a notation would have been adequate and appropriate. These interpretations of the melodic conception and the notation are therefore in mutual support of one another. It is worth showing a case where only the first category of information is conveyed consistently (Fig. 13.2). It shows simply that with each movement of the melody there is a new syllable. The neumes are not much more than syllable markers, and the example answers to Notker’s description of sequences. It is important for us, as we try to understand the sense of the early notation, to recognize that this information alone would have been thought to be worth conveying.
²⁹ Cf. Michel Huglo, ‘Les Noms des neumes et leur origine’, EG 1 (1954), 53–67. ³⁰ The distinction between a notation that emphasizes the configuration even while incorporating signs for individual notes and one that emphasizes the discreteness of individual notes is another aspect of the analogue–digital distinction. It is apparent in the notation of duration, in the contrast between modal and mensural notation, and again in the contrast between the proportional notation and the efforts (in the ‘Ars subtilior’) to specify exact values through notational forms. The interpretation of durational notations from this viewpoint merits a study of its own. ³¹ It is sometimes said that early scores served as mnemonics, implying that the singers had memorized fixed melodies. I prefer my way of putting it because it will accommodate the range of possibilities, from reproducing a fixed melody that has been memorized in that sense, to reconstructing one on the grounds of the singer’s knowledge of an idiom and the constraints of a text. This issue is addressed in Ch. 6.
Th e E a r ly H i st ory of M u s i c W r i t i n g
335
Fig. 13.2. Prosa notation in Oxford Bodley 775, fo. 127v
III Now to the signs themselves. Table 13.1 gives a selective overview of neumes in use in Europe and England between the ninth and twelfth centuries. It is introduced here as a point of departure only, and there is no claim for its completeness.³² A ³² The table is adapted from Corbin, Die Neumen. For purposes of the discussion here the last four columns of Corbin’s table have not been included. The following emendations have been necessary: (a) In the row marked ‘Bologna’ the last neumes in the Torculus and Scandicus columns have been reversed. (b) In the row marked ‘Benevento’ the two forms of the clivis have been reversed in order to make them consistent with the order of the two forms of the porrectus so far as their functional differentiation is concerned. This is to be explained in the text. In the same row the first two forms of the climacus have been added. (c) In the row marked ‘Nonantola’ the first of the neumes under porrectus has been added. (d) In the row marked ‘England’ the last neumes under clivis and porrectus have been added. These forms are used exclusively for the organa in the Winchester Troper. Their significance will be discussed in the text. (e) The scandicus forms have been provided for Nevers and Catalonia. Three cautionary remarks are required. First, the appearance of a neume under the virga column can be misleading. In several notations the virga is not used alone for single notes; only the punctum is used. What appears in the virga column has simply been isolated from compound forms that appear in the other columns. Second, alternative forms appearing in a single column/row can conceal three different possibilities with respect to their use: that they were used at different stages in the history of the notation, that they have different representational functions, or that they were used, so far as we know, indiscriminately. I shall say in the first two cases that alternative forms are in complementary distribution, meaning that they do not occur
336
Th e E a r ly H i s t ory of M u s i c W ri t i n g
Table 13.1. Eight neume characters as written in fourteen neumatic scripts Script
Class
1. St Gallen
A
2. England
A (B)
3. St-Benigne
A
4. Chartres
A
5. Nevers
A
6. Normandy
A
7. Lothringian (Messine)
B
8. Palaeofrankish
B
9. Breton
A/B
10. Aquitanian
B
11. Catalan
A
12. Bologna
A
13. Benevento
B
14. Nonantola
B
Punctum
Virga
Pes
in the same context. In the third case alternative forms are in random distribution. The semiotic task requires us to determine wherever possible which of these relations obtains. Third, since the original publication of this essay Susan Rankin has pointed out, with good reason, that what is shown in the table as ‘Palaeofrankish’ are selections from a much larger array of inscriptions which, rather than constituting conventional signs in a settled, systematic script like the others in the table, manifest, as she puts it, ‘the principle of joining positions in space which represent notes’. I have summarized her view on this subject in the introduction to this chapter. See also Wulf Arlt, catalogue entry for item XI.35: ‘Sakramentar mit “paläofränkischen” Neumen’, in Kunst und Kultur der Karolingerzeit: Karl der Große und Papst Leo III in Paderborn, ii: Katalog der Ausstellung, ed. Christoph Stiegemann and Matthias Wenikoff (Mainz, 1999), 838–9.
Th e E a r ly H i st ory of M u s i c W r i t i n g
Clivis
Torculus
Porrectus
Scandicus
337
Climacus
few definitions will facilitate the sorting out of the palaeographic material as it is briefly summarized in the table. Character. The class of representations of a particular melodic inflection. In Table 13.1 the characters are shown as vertical columns, each one headed by the name given it in tables of essentially this form that were first produced in the eleventh century.³³ The characters had become standard by that time, but we do not ³³ Huglo, ‘Les Noms des neumes’.
338
Th e E a r ly H i s t ory of M u s i c W ri t i n g
know how long before then. That means there is something anachronistic about speaking of, say, ‘punctum’ and ‘virga’ in sources anterior to the neume tables, and we shall do well to hold that in mind. Neume. A conventional inscription of a character. Each entry in the table is a neume. We need to distinguish neume from character because there can, for example, be two quite different graphic signs that are both called ‘clivis’, both of which represent a movement from a higher to a lower tone. Script. A system of standard neumes used by a community of practitioners defined by a common liturgical practice and often, but not always, by geographic contiguity (the contemporaneous use in Normandy and Sicily of the Norman script is the most obvious sort of exception). In the table the scripts are presented in horizontal rows, with their commonly accepted names down the left side. (Only the designation ‘Lothringian’ is a departure from common usage, which identifies this script as ‘Messine’. Solange Corbin argues persuasively that her designation more accurately fits the facts of geographic distribution.) Of the three just defined categories, that of script, as it is presented in the table, constitutes the greatest over-simplification. Some of the scripts shown are more distinct than others, and some would have to be subdivided; others could not be easily subdivided, but one would have to show variants of some of the neumes, and there would still be sources that do not readily fall into any one of the script categories. Still, what is shown does correspond to palaeographic realities and will serve quite well as a point of departure. The most widely circulated theory about the formation, origin, and meaning of the neumes can be stated in three parts: 1. The basic neumes are the punctum and virga, and all other neumes are compounds of these. 2. The referents of the punctum and virga are single low and high notes, respectively. They retain those meanings in composed neumes. 3. The punctum and virga descend from Latin prosodic accents: accentus gravis and accentus acutus, respectively. These originally indicated a downward and upward inflection respectively, hence their neumatic descendants came to signify a low and a high pitch respectively. On the grounds of this theory of their origin, the neumes that have punctum and virga as their basic components are known as ‘accent neumes’. These are contrasted with another class, called ‘point neumes’ with respect to their formation out of dots (cf. Table 13.1). This would appear at once to be an incoherent classification system that poses one class, identified on the basis of a principle of origin, against another, identified on the basis of a principle of morphology. But there is an underlying distinction of function that is more relevant. The ‘accent neumes’ function on the basis of the signification of punctum and virga as low and high
Th e E a r ly H i st ory of M u s i c W r i t i n g
339
notes respectively. The ‘point neumes’ function on the basis of the relative vertical positions of their component dots. The peculiarity of the classification scheme in fact lies only in the unfortunate choice of names for the classes, especially the expression ‘accent neumes’. According to the earliest formulation of this general theory (see below) the point neumes resulted from a branching off from the accent neumes. Ewald Jammers subsequently posited a separate origin for the point neumes.³⁴ Jacques Handschin proposed a different interpretation of the distinction that opened the way to a substantial advance in understanding, as I shall shortly suggest.³⁵ The accent theory of origin, as I shall call it, goes back to Edmond de Coussemaker.³⁶ It was subsequently transmitted and developed by various contributors to the series Paléographie musicale,³⁷ and by Peter Wagner, Johannes Wolf, Henry Bannister, Gregorio Suñol, Willi Apel, and Solange Corbin.³⁸ Most recently Charles Atkinson has offered a re-examination of the theory in which he separates its wheat from its chaff.³⁹ To my knowledge, up to the time of the publication of the original version of this chapter in 1982 only Handschin had publicly recognized any of the serious flaws in the theory. Certainly its widespread acceptance has been premature, for it opens onto issues of great complexity touching upon every outstanding problem about the early history of our notation. Nevertheless there is some kernel of truth in each aspect of the theory, and I shall proceed from those, starting with points (1) and (2). In the scandicus column of Table 13.1, the most common form is a compound of two puncta and a virga. (Only script 8 compounds three puncta, but it uses no virga at all, either alone or in compounds. The cursive forms of scripts 11 and 13 can be interpreted as a connection of two puncta and a virga, but as with all cursive forms such a reading can only be an a posteriori inference, which may be supported if compounding is otherwise established. Script 4 shows a compound of pes and virga. The pes could, in turn, be interpreted as a compound of punctum and virga, but only on indirect evidence; see below.) Three principles are demonstrated by this: (1) that in some instances, at least, neumes are clearly compounded of punctum and virga; (2) that in such cases the virga represents the highest note in the sequence; (3) that within the neume, at least, the upward direction on the page corresponds to an upward direction in the melodic figure. This principle is ³⁴ See Corbin, Die Neumen, 17–18 for a summary and for a demonstration that Jammers’s theory is untenable. ³⁵ Handschin, ‘Eine alte Neumenschrift’. ³⁶ E. de Coussemaker, Histoire de l’harmonie au moyen âge (Paris, 1852; repr. Hildesheim, 1966), ch. 2, ‘Origines des neumes’. ³⁷ The series Paléographie musicale (Solesmes, 1889– ). Hereafter PM. ³⁸ Citations in Corbin, Die Neumen, 19, to which must be added Apel, Gregorian Chant, 108–9. ³⁹ ‘De accentibus toni oritur nota quae dicitur neuma: Prosodic Accents, the Accent Theory, and the Paleofrankish Script’, in Hughes FS, 17–42.
340
Th e E a r ly H i s t ory of M u s i c W ri t i n g
common to all scripts and thus constitutes a premiss of the system as a whole. I shall call it the principle of directionality.⁴⁰ The climacus column also demonstrates the principles of compounding and directionality. But in three of the scripts where the scandicus rises to a virga, the climacus, at least in some forms, does not descend from one (7, 9, 10). In the other cases (1–6, 11–12) the virga is the high note in both directions. Comparison of the rising and falling two-note figures from this point of view gives somewhat less clear-cut results, but they tend in the same general direction. Four scripts have a pes that is a compound of punctum and virga (2, 7, 9–10). For the others it is possible, but not necessary, to think of the pes as such a compound with the elements connected. This is encouraged by the fact that in scripts 4, 6–7, and 9 the hook or thickening at the top of the virga is also made at the top of the cursive pes. Whether the cursive clivis forms in scripts 1–6 and 11–12 should be regarded as virga–punctum compounds is more difficult to say. But in those scripts where the climacus does not descend from a virga the clivis does not do so either (7, 9, 10; from this point on I shall not take 13 and 14 into the general surveys but rather discuss them separately, because of certain idiosyncrasies they display). And in two of those scripts, 7 and 10, the virga is rarely or not at all used for single notes. On the basis of this initial survey we can identify a class of scripts—for now I shall call it the A class, comprising scripts 1–6 and 11–12—in which the virga represents a high note under all circumstances, and the formation of the compound neumes that we have so far considered is based on the differentiation of virga and punctum as high and low notes respectively. The remaining scripts (still not including 13–14) I shall identify as the B class (7–10). In these scripts the virga is either not used at all (8), or it is used only in compound neumes (7 and 10), and when used in compound neumes it represents only the top of an ascending group, or the motion to the top, not a high note absolutely (7, 9–10). Now consider the three-note groups with change of direction (torculus and porrectus). In scripts 7b and 10 their components are quite evident: in 7b the torculus comprises three puncta and the porrectus comprises punctum–punctum– ⁴⁰ In doing so I deliberately avoid the more common designation of this phenomenon as ‘diastematy’. That term has been used to identify three phenomena that, while they have in common the premiss of the spatial metaphor that I identified at the beginning, are distinct in their exploitation of that premiss. I mean (1) directionality as a principle in the formation of individual neumes; (2) diastematy, which is the representation of melodic intervals with information on their magnitude; and (3) a general tendency to reflect the contours of a melody in the overall contours of the line of neumes but without sufficient precision to denote the actual intervals of the melody. In such situations one cannot tell whether there was a representational intent or an unconscious response to the melody as it was being written down. Directionality is common to all Western scripts; diastematy is a specific technique used in certain scripts only; contour-reflection can be found in scripts, of which some became diastematic and others did not. The Greek diastema means ‘interval’. The modern term ‘diastematic’ should accordingly be reserved for interval notations, plain and simple. Non-interval notations are not poor interval notations, and to characterize them as such is to make a nonsense of the whole subject. ‘Poor diastematy’ is no more diastematic than ‘fair weather friends’ are friends.
Th e E a r ly H i st ory of M u s i c W r i t i n g
341
virga. That is also how the porrectus is made in 10, whereas the torculus there comprises punctum–virga–punctum. But despite that difference, the virga is used, when at all, only as the top of an ascent. (That the virga in the torculus of 10 is the goal of the ascent and not the beginning of the descent is confirmed by the clivis in 10.) Scripts 7 and 10, recall, belong to the B class, and the analysis of the three-note group therefore confirms the characterization of that class given in the preceding paragraph. Can we make inferences about the components of the torculus and porrectus where they are not in plain view? It is worth trying. The torculus in all such cases would be read as punctum–virga–punctum (or pes–punctum), or punctum–punctum–punctum. The former might be preferred, just by extension of the reasoning regarding the pes. But there is no independent reason for preferring either interpretation, or even for reading them as compounds. The forms of the porrectus are something else again. If they are compounds, the porrectus of the A class cannot be thought to begin with punctum, for then the initial motion that they represent would be an ascending one, whereas the melodic figure represented begins with a descent. The alternative, then, is to interpret the ascending stroke as a virga, and that means, again, that in the A-class scripts the virga represents a high note absolutely. This is really the only available interpretation of the fact that the porrectus in these scripts begins by tracing an ascending line, and that tends to support the understanding of these neumes as compounds. The porrectus of the A-class scripts would be a compound of virga–punctum– virga → or virga–pes → . This is to be contrasted with the porrectus in the B-class scripts, which does not begin with such an ascent. Where that neume is not cursive (7b and 10) it begins with punctum and continues punctum–punctum or punctum–virga. Where it is cursive its components are presumably one or the other of those configurations. This confirms again that in the scripts of the B class the virga, when used, represents only the top of an ascent.⁴¹ With this analysis of the porrectus in mind we return to the clivis. There are two cursive forms: that in the A-class scripts beginning, like the corresponding porrectus, with the upstroke that betrays the virga component, and that of the B-class scripts beginning, again like the corresponding porrectus, with a horizontal element. ⁴¹ The porrectus of script 9 seems half-way between those of the A and B classes. In that respect, and also in its use of an independent virga, this script seems to straddle the boundary between the A and B classes. It is the only script to do so, as far as I know. Further research on this script is called for. Michel Huglo (‘Le Domaine de la notation bretonne’, AM 35 (1963), 54–84) showed that it is both old and widespread. It has generally been characterized as a mixture of elements of ‘accent’ and ‘point’ scripts. But that assumes both the priority of those two types and the aptness of that way of designating them. No decisive evidence has in fact been adduced to eliminate the opposite possibility, that the Breton script represents a prototype from which the A and B types descended. I am not putting forward such an hypothesis, but at this time we are not in a position to choose between such alternatives.
342
Th e E a r ly H i s t ory of M u s i c W ri t i n g
If we take the clivis and porrectus of the B scripts to be compound neumes, the initial horizontal element would not be the most direct way to connect two puncta. The most direct way is shown by the porrectus and the first of the clivis forms in script 8. Script 8 is the oldest of the B scripts. What the clivis and porrectus of the later B scripts accomplish more effectively is to make more immediately apparent to the eye the starting level of the neume on the page. And that takes us to the essential point of difference between the scripts of the two classes. The A scripts are based on the convention that a virga, alone or in compounds, represents a high note and the punctum a lower one.⁴² This differentiation establishes a symbolic mode of representation. The convention is entirely arbitrary, and whatever may have been its origin there is nothing innate about the signs that would suggest the lower or higher position of the note that is represented. The B scripts are based on the principle of directionality: the vertical position of the note signs vis-à-vis one another and the use of the virga-like stroke to represent the motion to a high note. They function primarily in the iconic mode. Given the spatial metaphor, their neumes constitute a graphic imitation of the contours of the melodic line.⁴³ Because of the sharpness of their differences in form and function, the versions of the clivis and porrectus can be taken as identifying marks of the two script classes. In general if the clivis and porrectus of a script are in more or less the configuration , the scripts can be expected in most cases to have all the other characteristics of A scripts. If they are in more or less these configurations: , the script can be expected in most cases to have all the other characteristics of the B scripts. Henceforth I shall refer to the A scripts as symbolic because the symbolic mode dominates in them. (They do still embody an iconic aspect, especially in their forms of the scandicus and climacus.) To reiterate, they are scripts 1–6 and 11–12. And I shall refer to the B scripts as iconic because that mode dominates in them. (This is not to forget that the very basis on which the iconic mode can function, the spatial metaphor, is itself a symbolic, i.e. conventional, principle.) They are scripts 7–10. No scripts are primarily indexical, although there is an indexical action in all scripts. ⁴² Within a neume this means that the virga represents the highest note. But where there is a succession of neumes including those representing single notes, there is a problem about identifying the boundaries of melodic segments within which the virga is to represent the highest note. A strict complementary distribution of virga and punctum would entail some such rule as this: write a virga for a single note higher than the preceding, and higher than or at the same pitch level as the following note; for all other single notes write a punctum. Such a rule can in fact be inferred from an actual chant book, to be identified presently. But most scripts do not observe any such strict rule. This has a bearing on the question of origins, as I shall shortly suggest. ⁴³ Handschin, ‘Eine alte Neumenschrift’, effectively recognized these two script classes and the functional distinction between them. He identified the scripts of the B class as ‘Tonortschriften’ because of their use of the placement of the neumes on the page. We can now see that Coussemaker (Histoire de l’harmonie) had already glimpsed something of the distinction. There will be more to say about Handschin’s article.
Th e E a r ly H i st ory of M u s i c W r i t i n g
343
Before continuing in terms of this terminology, which may seem somewhat jargon-like, I wish to demystify it from the vantage point of the insight of Susan Rankin cited in the introduction. What the distinction of iconic and symbolic notation types comes down to can be thought of this way: singer/scribes confronted with the task of representing melody or singing graphically took two different approaches initially that, as it turned out, had sufficient overlap so that eventually they could seem to converge—at least from our vantage point—into a single system with distinct executions such as is represented in Table 13.1. One approach was to imitate the contours of melody with line segments connecting points in the two-dimensional space that represented the sound-space, like the notators of the Buddhist traditions of Tibet and Japan.⁴⁴ That would be the approach of the inventors of the ‘Palaeofrankish’ notation. The other approach was to devise a fixed sign for each type of melodic element or progression. That would be the approach underlying all the other scripts, with their more or less imitative ‘inscriptions’ of the distinct notational ‘characters’, as defined above. The consistency with which the characteristic features of symbolic and iconic music writing cluster suggests that the two reflect fundamentally different cognitive principles. Naturally that provokes questions about the historical and cultural significance of this differentiation. I can pursue only the first of these questions here, the question of historical significance. The pursuit of the second will be a major undertaking of its own. The basic question will be, what contrasting features are concomitant with the symbolic–iconic distinction in the institutions in which those two modes of representation were predominantly used? And if there is a discernible trend through time with respect to the symbolic and iconic scripts, what can that suggest about larger patterns of cultural history?
IV Is such a trend discernible? In the present state of our knowledge we cannot give chronological priority to one or the other notational mode. Specifically, we do not know whether the Palaeofrankish and early Aquitanian scripts, on the one hand, or the Germanic ones, on the other, were the earliest ones in use, and we do not know what position should be given to the Breton script vis-à-vis the others. The fact is that all three types were in use in the ninth or the early tenth century. But we can profitably pursue the subsequent history, in which a clear trend is visible. This will best be approached first by way of an analysis of the Beneventan script, which I have so far put aside.
⁴⁴ See above, n. 23.
344
Th e E a r ly H i s t ory of M u s i c W ri t i n g
Table 13.1 shows, for that script, both the iconic and symbolic forms of the clivis and porrectus. But in the manuscripts those forms are in strict complementary distribution. For both characters the symbolic forms and are used when the neume’s first note is higher than the preceding note. The iconic forms and are used when the neume’s first note represents a descent from or repetition of the preceding note (see Fig. 13.3). But that means that the morphologically ‘symbolic’ forms are not functioning symbolically here but rather iconically. The upstroke at the beginning of the neume is not a symbol for a high note: it imitates graphically the ascent to the first note of the neume, exactly like the virga in the iconic scripts that have been analysed. That this is so is clear from a comparison with the clivis and porrectus of the symbolic scripts, which begin with the upstroke regardless of the position of their initial note relative to the preceding note. The same principle is at work in the climacus of the Beneventan script. When its initial note represents an ascent from the preceding note, it begins with an upstroke. When it repeats the preceding note or is below it, it begins with a punctum (Fig. 13.3). The intention is especially clear here. The upstroke to the climacus in the former case is a finer line than the virga that is made in this script, so that we have no doubt about reading it as a representation of the ascent to the first note, which is represented by a punctum. The iconic principle determines the form of the punctum as well. It is slanted in the direction of the preceding note, that is, its shape represents the direction of the melodic line from the preceding note to the note that it represents (see Fig. 13.3). In the manuscript represented there, Benevento VI. 34, only the form with downward slant is used. The Beneventan manuscripts Vatican Vat. lat. 10673 and Benevento VI. 35, 38, and 40 use both the downward- and upward-slanting forms. The Beneventan script observes these principles from its earliest manifestations in the eleventh century. It is eclectic. The forms of the iconic scripts are incorporated, unchanged in principle (I do not take into account here the idiosyncratic style of the script); those of the symbolic scripts are assimilated into a new iconic role. And new elements, such as the puncta and the upstroke to the climacus, are added. The product is a more prescriptive kind of iconic writing than those we have identified elsewhere. Its greater explicitness is in effect a matter of the extension of the principle of directionality. (The script shown in Fig. 13.3 is quite precisely diastematic as well, but that is a separate matter, and it will be discussed separately.) In one important regard the Beneventan script differs from most of the iconic scripts. Single notes are represented by both punctum and virga. In the late eleventh/early twelfth-century gradual Benevento VI. 34, the virga is generally used for a single note that is higher than the preceding and higher than or at the same level as the following note. Now this is just the sort of operational rule for the
Th e E a r ly H i st ory of M u s i c W r i t i n g Fig. 13.3. Reinforced iconic script in Benevento VI. 34, fo. 27v
345
346
Th e E a r ly H i s t ory of M u s i c W ri t i n g
differentiation of punctum and virga that cannot be inferred for the early symbolic notations of the North.⁴⁵ There are exceptions in Benevento VI. 34, but not nearly so many as in the earlier Beneventan manuscript, Vatican Vat. lat. 10673. There a virga may be preceded or followed by a higher note, and there may be two virgae in succession. That is, the rule for the use of the virga was not a constituent of the Beneventan script from the beginning. It came in, we must presume, as an aspect of the sharpening in precision of the script, along with diastematy. That has important implications for the early history of neumatic writing. Evidently the virga rule was not incorporated into whatever writing systems the Beneventan notators took as their models. On the contrary, we find the same types and frequency of exceptions to it in the early Beneventan sources as in the Northern symbolic scripts, exceptions that make it impossible to say anything more than ‘the virga represents a high note’. This has some bearing on our thinking about the question of origins. If the rising and falling accents provided the idea for distinguishing high and low notes and even the signs for representing them, then we would expect to find a strict differentiation between the signs in the earliest sources and a more casual one as notations became more clearly diastematic. But the situation in Benevento, at least, seems to be the reverse. Because the extended use of the principle of directionality in the Beneventan script results in a more explicitly graphic representation of the melodic line, I shall refer to it as a reinforced iconic script. Since this entails the adaptation of oncesymbolic forms for iconic function, the question arises whether this suggests a general tendency in music writing to favour the iconic mode. Although we cannot answer that question with respect to the very earliest stage, there is evidence for such a tendency after the first hundred years.⁴⁶ I have identified reinforced iconic scripts in the following sources of the twelfth century and after.⁴⁷
french 1. Paris, Bibl. Ste-Gen. 547 (Répertoire de manuscrits médiévaux, ed. Solange Corbin, i (Paris, 1965), pl. 6) 2. Troyes 1047 (PM 3, pl. 195) 3. Paris nouv. acq. lat. 1235 (PM 3, pl. 195b) 4. Paris nouv. acq. lat. 10511 (PM 3, pl. 198a)
⁴⁵ See n. 42. Even in the St. Gallen script there are inconsistencies within the same manuscripts about writing the virga or punctum for a note higher than the preceding but lower than the following. The uncertainty about such places is apparent when one manuscript has a virga where another has a punctum in the same passage. Examples of this may be seen in Corbin, Die Neumen, 48–52. ⁴⁶ What follows is a preliminary report on evidence that I have identified so far. In my view it is sufficient to support the idea of such an evolution as a working hypothesis. With the study of more sources it will be possible to refine the hypothesis and follow the direction of development, both geographical and institutional, in greater detail. ⁴⁷ The location of published facsimiles of notational specimens is indicated in parentheses following each source.
Th e E a r ly H i st ory of M u s i c W r i t i n g 347
Fig. 13.4. Reinforced iconic script in Vercelli 162, fos. 10v–11
348
Th e E a r ly H i s t ory of M u s i c W ri t i n g
Fig. 13.5. Reinforced iconic script in Padua A 47, fo. 3v
italian 5. 6. 7. 8.
Vercelli 162 (see Fig. 13.4) Lucca 601 (the entire manuscript in PM 9) Lucca 602 Lucca 603 (PM 2, pl. 34)
Th e E a r ly H i st ory of M u s i c W r i t i n g
349
9. Lucca 609 (PM 2, pl. 35) 10. Pistoia 121 (Bruno Stäblein, Schriftbild der einstimmigen Musik (Musikgeschichte in Bildern, 3, pt. 4; Leipzig, 1975), pl. 24) 11. Pistoia 119 (Stäblein, Schriftbild, pl. 25) 12. Padua A 47 (see Fig. 13.5) 13. Verona C.V. (PM 2, pl. 14a) 14. Montecassino NN. 339 (PM 2, pl. 19) 15. Montecassino NN. 540 (PM 2, pl. 21a) 16. Montecassino NN. 546 (PM 2, pl. 22) 17. Vatican Barb. XII. 12 (PM 2, pl. 23) 18. Vatican Vat. lat. 5319 (PM 2, pl. 28) 19. Vatican Vat. lat. 6078 (Stäblein, Schriftbild, pl. 23) 20. Rome, Bibl. Cas. C. IV. I (PM 2, pl. 29) 21. Naples, Bibl. naz. VI. G. 38 (PM 2, pl. 30) 22. Paris nouv. acq. lat. 1669 (PM 2, pl. 31) 23. Rome, Bibl. Vall. C. 32 (PM 2, pl. 33a) 24. Cortona 12 (PM 2, pl. 33b) 25. Modena O. I. 16 (PM 2, pl. 36) 26. Modena O. I. 7 (PM 2, pl. 37a) 27. Turin F. IV. 18 (PM 2, pl. 37b) 28. Bologna 2748 (Stäblein, Schriftbild, pl. 19)
gothic 29. London Add. 26884 (PM 3, pl. 135) 30. Zwettl 199 (Stäblein, Schriftbild, pl. 11) 31. Oxford Rawl. C. 892 (Stäblein, Schriftbild, pl. 12)
english 32. London Add. 12194 (PM 3, pl. 201) 33. London Lansd. 462 (see Fig. 13.6) There are earlier sources of the same areas of provenance as numbers 2, 3, 5, and 32, at least, in which a strictly symbolic script is used: 2. 3. 5. 32.
Troyes 894 (Corbin, Die Neumen, pl. 25) Paris lat. 9449 (Corbin, Die Neumen, pl. 29) Vercelli 161 (see Fig. 13.7) All English sources of the eleventh century, with the exception of the organum notation of the Winchester Troper (see below)
This shows a widespread tendency to adopt more iconic modes of notation. A hint of what this may have involved is provided by a variant of the rule for the forms of the clivis and porrectus used by the scribe of the Sarum Gradual. He
350
Th e E a r ly H i s t ory of M u s i c W ri t i n g
Fig. 13.6. Reinforced iconic script in London Lansdowne 462, fo. 98v
Th e E a r ly H i st ory of M u s i c W r i t i n g 351
Fig. 13.7. Symbolic script in Vercelli 161, fos. 58v–59r
352
Th e E a r ly H i s t ory of M u s i c W ri t i n g
writes and whenever the first note represents a descent except where the neume begins a new syllable. Then he writes or . That suggests a survival of the understanding of these latter forms as the norm. There seems to be something of this in the Beneventan script too. Of fifty-one pieces in Benevento VI. 34 that begin with a clivis, that is, where there is no question about the preceding note, forty-seven are written and four are written . Only five pieces begin with a porrectus; three are written and one . Perhaps that betrays a sense about these symbolic forms as an underlying or background norm. A change of representational function is evident also in the other of the two scripts that I had put aside. Table 13.1 shows for the Nonantola script both forms of the clivis and porrectus. The unique practice in Nonantola of attaching the neumes to the vowels of the respective syllables that they declaim—as though they had sprouted from their syllables—is well known. Although this has always seemed a curious and exotic practice, it is so in its form only. In the writing of countless medieval music books care was taken to show the precise association of neumes with syllables of text, but usually it was by means of diagonal lines ). grouping together the neumes that belonged with a syllable of text ( Both devices remind us of the primary function of melody as inflection of language. In the Nonantola script the once-‘symbolic’ clivis and porrectus ( and ) are used above the text because they provide attachment lines to the syllables. In melismatic passages, where there is no text for the neumes to attach to, the normal rules for reinforced iconic scripts apply (cf. e.g. Stäblein, Schriftbild, pl. 16). There are indications of a tendency towards iconic modes of music writing in two other notational traditions: the Aquitanian and the English. The manuscript Paris lat. 1240, from the tenth century, is the oldest surviving source of liturgical music from Aquitania. Its notation shows essential features of the later script that has been identified as the Aquitanian script. It also shows features that were not carried on in that script: the symbolic forms of the clivis and porrectus ( and ), a form of the pes that is otherwise found in northern French scripts ( ), and these (there is nothing resembling the virga of the later signs for single notes: Aquitanian sources, except in compound neumes). It has been observed repeatedly that Paris 1240 is more closely related with regard to its contents to northern sources, especially the Nevers troper Paris 9449, than to later Aquitanian sources.⁴⁸ Judging from these few details about the script of Paris 1240 we may infer that music writing as well as repertory in Aquitania was early under the influence of the north. By the eleventh century the Aquitanian script was consistently iconic. The signs for single notes are of special interest. Example 13.1 shows a single ⁴⁸ Evans, ‘Northern French Elements’, and Ellen Ryer, ‘The Introit Trope Repertory at Nevers: MSS Paris, Bn. Lat. 9449 and Paris, n.a. lat. 1235’ (Ph.D. diss., University of California at Berkeley, 1981), 39–72, 74–213.
Th e E a r ly H i st ory of M u s i c W r i t i n g
353
Ex. 13.1. Trope introduction Deus pater filium to the Christmas introit Puer natus est nobis (CT i, no. 34) with versions in (a) Paris lat. 1235, fo. 184r; (b) Paris lat. 9449, fo. 7v; (c) Paris lat. 1240, fo. 9r; (d) Paris lat. 1118 , fo. 9v
(a) (b) (c) (d ) passage in four versions: Paris 1240, Paris 1121 (an eleventh-century Aquitanian source), Paris 9449 (an eleventh-century source from Nevers), and Paris 1235 (a twelfth-century source from Nevers). In Paris 1118 the single notes, with the exception of the liquescent neume on ‘in’, are all puncta, identical in form. In Paris 1240 they are differentiated as for high notes and for low notes ( ) is always and only the lowest note, ( is not quite so consistent about its context and function). And what is written and in Paris 1240 is usually written virga and punctum respectively in Paris 9449, where those signs represent high and low notes respectively.⁴⁹ Like the clivis and porrectus, the slanted single-note figures of Paris 1240 function in the symbolic mode. It is instructive to compare them with the slanted puncta in the Beneventan script, which function in the iconic mode. In Paris 1240 the sign is a symbol for the lowest note of a descent. In Benevento VI. 34 the same sign indicates a note lower than the preceding. The Beneventan script allows ⁴⁹ Here I must enter a caveat about a methodological trap that can be entailed in making such comparisons. The comparison presupposes that we are seeing alternative ways of writing the same things; otherwise it tells us nothing. So it behoves us to maximize the chances that the assumption is right. The best way to do that is to choose melodies and, even better, formulae, that we know to be standard, and sources that we know to be closely related. The melody chosen for Ex. 13.1 is the beginning of a trope to the Christmas introit Puer natus, which is like the highly standard opening of the introit itself. But even after taking such precautions we should not expect mechanical regularity in the translation of the signs of one script into those of another. Each script will exhibit its own patterns of behaviour with respect to such matters as the differentiation of punctum and virga. If, for example, a peak tone is written twice in succession, one script may write its sign for a high note twice, while another (like Paris 9449 in Ex. 13.1) uses its corresponding sign for only the first of the two. Comparisons of scripts in which such factors are not taken into account cannot tell us much. Corbin (Die Neumen, 95) compares the inscriptions of the Trisagion in Paris 1240, Paris 903 (the 11th-c. St-Yrieix Gradual), and the modern Vatican Edition. The first two are not closely enough related in general to encourage confidence that their scribes intended exactly the same melodies. The third gives a version of unknown authority, and its use as a basis for comparison is questionable, to say the least. Only on the troublesome old assumption that the Gregorian melodies were fixed before their writing down and that the written transmission is exclusively the product of copying could this sort of comparison have any value. But even a quick check of the version of Paris 903, which is easily read, against the Vatican Edition reveals a large number of cases in which the two do not correspond, so the example defeats the premiss on which it is supposed to function as an example. (Comparisons of this sort constitute an essential basis for Corbin’s exposition throughout her book, which is flawed by their weakness.)
354
Th e E a r ly H i s t ory of M u s i c W ri t i n g
the sequence ; the script of Paris 1240 does not. That is why we can attribute an imitative function to the sign in Benevento VI. 34 but not in Paris 1240. The use of those neumes in Paris 1240 may appear to have some bearing on the accent theory of origin. The neumes and look like acute and grave accents, respectively, and their counterparts in the north are virga and punctum respectively. But the relationships do not work out quite right from the point of view of the theory. For the theory would lead us to expect the and of Paris 1240 to be replaced by virga and punctum in subsequent Aquitanian sources. But that happens in a different script altogether; in Aquitania those neumes are replaced by puncta. And in fact it would not really be correct to say that they were replaced by virga and punctum in the north. The one speculation we could make in order to fit the case to the theory is that, if the notation really did spread from north to south, the single-note neumes were replaced by virga and punctum. That is not inherently implausible, but there is no evidence for it. A turn to the iconic principle is evident also in the second Winchester Troper, the eleventh-century manuscript Cambridge 473. In the main, English neumatic writing was symbolic, as Table 13.1 shows. But in the notation of the new organum parts provided for some of the traditional chants in Cambridge 473, the repertory of forms in the script was enlarged by the addition of the iconic clivis and porrectus, and there was a discernible effort at diastematy. The chants themselves are written in the prevailing English symbolic script.⁵⁰ The situation of the manuscript contributes to our understanding of the context of this general evolution in notational modes. A symbolic notation that does not encode information about the interval contents of melodic lines is nevertheless adequate in the context of a regulated oral tradition of long standing. But the organum melodies were new, and it was desirable to have the notation convey as much information as possible. Iconic signs were a means to that end. The trend towards more informative notations may have been motivated by the need to represent non-traditional matter, and also by the desire to represent even traditional matter for singers who were not as well versed in the tradition. The iconic scripts were not the only response to those needs, and they were never adopted universally. The easiest way to see that is by looking at the square neumes that head the character columns in Table 13.1. They are the neumes of a symbolic notation, and in their square forms they have been standard from the thirteenth century to the present. A survey of all chant books from the beginning ⁵⁰ I mention this only briefly here. The differences in the forms and placement of neumes between the chant and organum notations have long been known. That the difference is functional as well as morphological was first recognized by my student at Stony Brook George Khouri. A similar differentiation in the tropers St. Gallen 381 and 484 has been noted by Wulf Arlt in Stiftsbibliothek Sankt Gallen: Codices 484 & 381, ed. Arlt and Rankin, 60. A detailed explication of this subject is given by Susan Rankin in ‘Winchester Polyphony: The Early Theory and Practice of Organum’, in Music in the Medieval English Liturgy: Essays for the Centenary of the Plainsong and Medieval Music Society (Oxford, 1992), 59–99.
Th e E a r ly H i st ory of M u s i c W r i t i n g
355
to the present day would surely show that those written in symbolic scripts outnumber by far those written in iconic scripts. It was the symbolic forms, too, whose vestigial virgae had become meaningless, that were adapted for the modal and then the mensural system for representing relative duration. The use of the iconic forms in fact declined and the tendency to adopt them stopped. If we ask why, a look at Benevento VI. 34 and the Sarum Gradual (Fig. 13.3 and no. 32 above) will suggest a perfectly satisfactory explanation. The iconic system of neumes— though not the iconic principle—was made redundant and obsolete by the universal establishment of diastematy, ultimately in the form of the stave. The development of diastematy ran parallel to the symbolic–iconic shift. In some sources the reinforced iconic system was written with strict diastematy, even on the stave. That the diastematic principle succeeded over the iconic one in the evolutionary history is seen in late sources of this type. The latest known to me is London Lansd. 462, from the Sarum use in the fifteenth century (no. 33 in the list above; Fig. 13.6). In this manuscript both types of clivis and porrectus are used, and in the reinforced iconic system. The majority of the exceptions entail an inconsistency about which form is to be written when the first note of the neume repeats the preceding pitch. That is, the two forms of clivis and porrectus are not in random distribution. The inconsistencies reflect an uncertainty over a single issue, in what is a vestigial reinforced iconic system. Diastematy succeeded, presumably, because it was a simpler principle that could at the same time absorb the function of the iconic principle.
V I believe that this observation of a large-scale historical shift is a new and important result. To be sure it requires refinement from many sides. But even so it could emerge only as a result of the classification of notational systems on functional principles, that is, on semiotic, over and above palaeographic, principles. It is one of the most important advantages of a semiotic classification that the very principles that underlie the classification are the focal points for a view of the historical development. There is a parallel here with the understanding of biological evolution. Species classification had long been conducted on the basis of morphological description. During the nineteenth century, and especially under Darwin, progress in the understanding of the origin and development of species depended on the replacement of a morphological description with an analysis of biological systems that takes account of ecological, geographical, and genetic factors. The parallel with the semiotic analysis of notational systems is clear, and it is equally clear that the history of notation is an evolutionary one in exactly the Darwinian sense.
356
Th e E a r ly H i s t ory of M u s i c W ri t i n g
What both sorts of analysis have in common is that they aim to identify what biologists call ‘natural classes’, classes which, as Plato put it, ‘carve nature by the joints’, and which, because of the fundamental character of their principles of differentiation, yield theoretical understanding. To emphasize this point, we can contrast the proposed classification of neumatic scripts as iconic and symbolic with their now common classification in terms of point and accent neumes. The iconic–symbolic classification is based on a principle of function, the point–accent classification on a mixture of morphological description and a theory of origin, with one type identified by its appearance and the other by a theory about its derivation. It is as though one would divide letters of the alphabet into two classes: those that are square in form and those that are thought to have originated as signs for phases of the moon. The iconic–symbolic classification identifies actual classes with some overlap, whereas the point–accent classification identifies categories to which no scripts belong as a whole (with the possible exception of the Palaeofrankish). All other scripts are mixtures of the two classes, which must then be thought of as ideal categories. And consequently any account of the historical development of notation that is based on the point–accent classification must be of an a priori and idealist nature; whereas the iconic–symbolic classification yields a pragmatic historical view that is well supported by the evidence. This contrast speaks, therefore, to the advantages of a turn, just now, towards the analysis of notational systems from the point of view of semiotics.
VI Peter Wagner, with the cautious concurrence of Handschin,⁵¹ asserted that the principle of diastematy was at work in neumatic writing virtually from the beginning. That is not only because of the possible priority of some of the early diastematic scripts, but also because the principle of directionality plays a role in the formation of at least some neumes in all scripts. This is in contrast to the entirely symbolic, non-iconic nature of Byzantine notation, for example. Handschin’s discussion of these matters is, so far as I am aware, the only one in print in which the analysis is undertaken from the standpoint of the question, how did the systems of writing represent? It was his article that suggested that standpoint to me, and I am indebted to it, despite my somewhat different interpretations. Handschin called the Palaeofrankish script ‘diastematic’, pointing out that one can transcribe from it without great difficulty. He posited the spatial metaphor as its controlling principle, observing that cursive lines are used only to connect posi⁵¹ Peter Wagner, Einführung in die gregorianischen Melodien, ii: Neumenkunde: Paläographie des liturgischen Gesanges (Leipzig, 1912, repr. 1962), ch. 12. Handschin, ‘Eine alte Neumenschrift’, 84–6.
Th e E a r ly H i st ory of M u s i c W r i t i n g
357
tions, not as signs for notes. And so he characterized the script as a ‘Tonortschrift’, along with the Aquitanian, Messine, and Nonantola scripts (he recognized the function of the strokes in the latter, and that it was incorrect to designate it an ‘accent script’, as had been conventional). Against these he contrasted the German, the north Italian, north French, Norman, and English scripts, whose signs mark not the positions of the notes, but the distance (‘Strecke’) between them. He recognized the inadequacy of the two prevailing classifications of notations— accent vs. point neumes and diastematic vs. adiastematic scripts—the first because it ignores the functional difference between the two groups that he had identified, the second—as it had been employed—because all scripts have an element of directionality in their compound neumes and can reflect something of the melodic distances between notes without being diastematic in the strict sense. But there are still problems with Handschin’s characterizations of the two types. The ‘Tonort’ principle corresponds to what I have called directionality. Until it is given exact diastematic proportions the position placement of neumes conveys exact information only about direction. Our ability to transcribe, or sing, from the Palaeofrankish script depends on the same competence that the contemporary singer would have required to read the notation, a knowledge of the grammar of the melodies represented. No one, then or now, who lacked that competence could read the notation. On the other hand, the idea that the other type of script is based on the representation of the distance to a pitch by means of the stroke leading up to it can apply to the reinforced iconic scripts, like the Beneventan, but not to the St. Gallen. For suppose that a porrectus were to follow a higher note: the upstroke cannot represent the distance to the first note of the neume, because that would be reached by a motion from above. Yet despite these problems of characterization, Handschin produced virtually the same classification as that based on the iconic–symbolic distinction. It is tempting to think that that was the principle he was really after. A literal translation of ‘diastematic’ would be ‘intervallic’; a ‘diastematic notation’ would be an ‘interval notation’. It is not empty pedantry to insist on reserving that expression for such notations (e.g. Paris 903 and 1235, Benevento VI. 34, the Liber usualis). The criterion would be the exact proportionality of vertical distance to interval distance, which means decipherability without pre-knowledge. Otherwise there is no way of saying what degree of approximation to that condition suffices in order for a notation to be considered diastematic. Paris 1240 and 9449 are generally said to have diastematic and adiastematic notations respectively, but they both show about the same degree of directionality (see Ex. 13.1), something short of the criterion I have just stated. The difference is that some of the successors to Paris 1240 in Aquitania are diastematic according to that criterion (Paris 903, but not 1118). But Paris 1235, which is a successor to 9449 in Nevers, is also diastematic according to that criterion, so that by the same token one could say
358
Th e E a r ly H i s t ory of M u s i c W ri t i n g
that the notation of 9449 is also diastematic. With respect to the Aquitanian script, it is generally said that the diastematy in Paris 1240 and 1118 is poor (9449 has been spared this handicap). There are two problems here. First, it is generally not good practice to say that a thing has the characteristics of its successors, but has them in deficient measure (it would be like saying that the music of Dufay exemplifies triadic tonality, but not very well). Second, it is not to be believed that the notators of Paris 1240 and 1118 intended to write an interval notation, but failed. We have to believe that the notations of those, and indeed of other Aquitanian sources that are not diastematic according to our criterion, were good enough for them. The error behind all these problems is in regarding diastematy as a constitutive property of scripts. It is, instead, a way of writing them. The same scripts can be both diastematic and adiastematic. In some cases that may be a matter of historical development—a scriptorium converting to diastematy (e.g., evidently, Benevento). In other cases a script has been written both ways by contemporaries, and we have to speculate about a functional differentiation (e.g. Paris 903 and 1118, contemporaries in Aquitania; I suggest that one is the work of a scribe, writing a book that is perhaps destined for teaching purposes, the other the work of a cantor, writing a book for his own use). Accordingly, we require a classification scheme something like this: Symbolic scripts
Iconic scripts
Adiastematic
St. Gallen, Bologna, etc., Paris 9449 (Nevers)
Paris 1240 and 1118, Chartres 47 (Breton), Vatican Vat. lat. 10673 (Beneventan)
Diastematic