Oral and Written Transmission in Chant (Music in Medieval Europe) [1 ed.] 9780754626268, 9781315090283, 0754626261

Citation preview

Oral and Written Transmission in Chant

Series Editor: Thomas Forrest Kelly Titles in the Series: Chant and its

Origins Kelly

Thomas Forrest

Oral and Written Transmission in Chant Thomas Forrest

Kelly

Embellishing the Liturgy Alejandro Enrique Planchart Poets and Singers Elizabeth Aubrey Ars

antiqua

Edward H. Roesner Ars

nova

John L. Nádas and Michael Scott Cuthbert Instruments and their Music in the Middle

Ages Timothy J.

McGee

Music in Medieval Europe

Oral and Written Transmission in Chant

Edited by

Thomas Farrest Kelly Harvard University, USA

I~ ~~o~!!!~~~up LONDON AND NEW YORK

First published 2009 by Ashgate Publishing Published 2017 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OXl 4 4RN 711 Third Avenue, New York, NY 10017, USA Routledge is an imprint ofthe Taylor & Francis Group, an informa business

Copyright © Thomas Forrest Kelly 2009. For copyright of individual articles please refer to the Acknowledgements. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data

Oral and written transmission in chant. - (Music in medieval Europe) 1. Gregorian chants - History and criticism 2. Musical notation - History -To 1500 3. Transmission of texts I. Kelly, Thomas Forrest 782.3'222'0902 Library of Congress Control Number: 2007943185

ISBN 13: 978-0-7546-2626-8 (hbk)

DOI: 10.4324/9781315090283

Contents Acknowledgementsvii Series Preface ix Introduction xi

PART I 1 2

MUSIC-WRITING

Leo Treitler (1982), ‘The Early History of Music Writing in the West’, Journal of the American Musicological Society, 35, pp. 237-79.3 Charles M. Atkinson (1995), ‘De Accentibus Toni Oritur Nota Quae Dicitur Neuma: Prosodic Accents, the Accent Theory and the Paleofrankish Script’, in Graeme M. Boone (ed.), Essays on Medieval Music in Honor of David G. Hughes, Cambridge, MA: Harvard University Department of Music, pp. 17-42.47

NOTATION II PART AND PERFORMANCE 3 4

5

The Restoration of the Chant and Seventy-Five Years of Recording’, Early Music, 7, pp. 197-217.75 Nino Albarosa (1983), ‘The Pontificio Istituto di Musica Sacra in Rome and the Semiological School of Dom Eugène Cardine’, Journal of the Plainsong and Medieval Music Society, 6, pp. 26-33.91 Lance W. Brunner (1982), ‘The Performance of Plainchant: Some Preliminary Observations of the New Era’, Early Music, 10, pp. 316-28.9

Mary Berry (1979), ‘Gregorian Chant:

PART III ORAL AND WRITTEN TRANSMISSION 6 7 8

Leo Treitler

‘Homer and Gregory: The Transmission of Epic Poetry and Musical The Quarterly, 60, pp. 333-72.1 5 Plainchant’, Leo Treitler (1975), ‘“Centonate” Chant: Übles Flickwerk or E pluribus unus?', Journal of the American Musicological Society, 28, pp. 1-23.15 David G. Hughes (1987), ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’, Journal of the American Musicological Society, 40, pp. 377-404.

(1974),

179

9

Kenneth Levy (1987), ‘Charlemagne’s Archetype of Gregorian Chant’, Journal of the American Musicological Society, 40, pp. 1-30.207 10 Leo Treitler (1988), ‘Communications’, concerning Levy and Hughes, above, Journal of the American Musicological Society, 41, pp. 566-75. Levy’s response, pp. 575-78;

Hughes’

response, pp. 578-79.237

11

12 13

14

15

László Dobszay

(1993), ‘The Debate about the Oral and Written Transmission of International Chant’, Musicological Society Congress Report XV: Madrid 1992 de [Revista Musicologia, 16], pp. 706-29.251 Kenneth Levy (1990), On Gregorian Orality ’, Journal of the American Musicological Society, 43, pp. 185-227.275 Emma Hornby (2004), ‘The Transmission of Western Chant in the 8th and 9th Centuries: Evaluating Kenneth Levy’s Reading of the Evidence’, The Journal of Musicology, 21, pp. 418-57.319 Edward Nowacki (1998), ‘Chant Research at the Turn of the Century and the Analytical Programme of Helmut Hucke’, Plainsong and Medieval Music, 7, pp. 47-71.359 Susan Rankin (1995), ‘Ways of Telling Stories’, in Graeme M. Boone (ed.), ‘

Essays on Medieval Music in Honor of David G. Hughes, Cambridge, MA: Harvard University Department of Music, pp. 371-94.385 16 Theodore Karp (1990), ‘Interrelationships among Gregorian Chants: An Alternative View of Creativity in Early Chant’, in Eugene K. Wolf and Edward H. Roesner (eds), Studies in Musical Sources and Style: Essays in Honor of Jan LaRue, Madison, WI: A-R Editions, pp. 1—40. 409 Index 449

Acknowledgements The editor and material.

publishers

wish to thank the

following

for

permission

to use

copyright

Inc. for the essay: Theodore Karp (1990), ‘Interrelationships among Gregorian Chants: An Alternative View of Creativity in Early Chant’, in Eugene K. Wolf and Edward H. Roesner (eds), Studies in Musical Sources and Style: Essays in Honor ofJan LaRue, Madison, WI: A-R Editions, pp. 1-40. All rights reserved. Used with permission.

AR-Editions,

Press for the essays: Nino Albarosa (1983), ‘The Pontificio Istituto de Musica Sacra in Rome and the Semiological School of Dom Eugène Cardine’, Journal of the Plainsong and Medieval Music Society, 6, pp. 26-33; Edward Nowacki (1998), ‘Chant Research at the Turn of the Century and the Analytical Programme of Helmut Hucke’, Plainsong and Medieval Music, 7, pp. 47-71.

Cambridge University

Harvard University Department of Music for the essays: Charles M. Atkinson (1995), ‘De Accentibus Toni Oritur Nota Quae Dicitur Neuma: Prosodic Accents, the Accent Theory and the Paleofrankish Script’, in Graeme M. Boone (ed.), Essays on Medieval Music in Honor of David G. Hughes, Cambridge, MA: Harvard University Department of Music, pp. 17-42; Susan Rankin (1995), ‘Ways of Telling Stories’, in Graeme M. Boone (ed.), Essays on Medieval Music in Honor of David G. Hughes, Cambridge, MA: Harvard University Department of Music, pp. 371-94. László Dobszay for the essay: László Dobszay (1993), ‘The Debate about the Oral and Written Transmission of Chant’, International Musicological Society Congress Report XV: Madrid 1992 [Revista de Musicologia, 16], pp. 706-29. Oxford University Press for the essays: Mary Berry (1979), ‘Gregorian Chant: The Restoration of the Chant and Seventy-Five Years of Recording’, Early Music, 7, pp. 197-217; Lance W. Brunner (1982), ‘The Performance of Plainchant: Some Preliminary Observations of the New Era’, Early Music, 10, pp. 316-28; Leo Treitler (1974), ‘Homer and Gregory: The Transmission of Epic Poetry and Plainchant’, The Musical Quarterly, 60, pp. 333-72. Copyright ©1974 by G.

Schirmer, Inc.

University of California Press for the essays: Leo Treitler (1982), ‘The Early History of Music Writing in the West’, Journal of the American Musicological Society, 35, pp. 237-79. Published by University of California Press on behalf of the American Musicological Society; Leo Treitler

(1975), ‘“Centonate” Chant: Übles Flickwerk or E pluribus unus?', Journal of the American Musicological Society, 28, pp. 1-23. Published by University of California Press on behalf of the American Musicological Society; David G. Hughes (1987), ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’, Journal of the American

Oral and Written Transmission in Chant

Musicological Society, 40, pp. 377-404; Kenneth Levy (1987), ‘Charlemagne’s Archetype of Gregorian Chant’, Journal ofthe American Musicological Society, 40, pp. 1-30. Published by University of California Press on behalf of the American Musicological Society; Leo Treitler (1988), ‘Communications’, concerning Levy and Hughes, above, Journal of the American Musicological Society, 41, pp. 566-75. Levy’s response, pp. 575-78; Hughes’ response, pp. 578-79. Published by University of California Press on behalf of the American Musicological Society; Kenneth Levy (1990), ‘On Gregorian Orality ’, Journal ofthe American Musicological Society, 43, pp. 185-227. Published by University of California Press on behalf of the American Musicological Society; Emma Hornby (2004), ‘The Transmission of Western Chant in the 8th and 9th Centuries: Evaluating Kenneth Levy’s Reading of the Evidence’, The Journal of Musicology, 21, pp. 418-57. Copyright © 2005 by the Regents of the University of California. All rights reserved. Images The

Abbey ‘Triplex’.

of Salesmes for the

image:

The introit Polulus Sion

it will appear in the

as

new

Abbey of Salsmes for the manuscript examples in Figures 1 2 3 and 5 from Emma Hornby (2004), ‘The Transmission of Western Chant in the 8th and 9th Centuries: Evaluating Kenneth Levy’s Reading of the Evidence’, The Journal of Musicology, 21, pp. 418-57. The

,

,

Heinrich Heine Universität, Düsseldorf for the image: Ad te levavi in Paleoffankish script, Die Hansdschrift ist Leihgabe der Stadt Düsseldorf an die Universitäts- und Landesbibliothek

Düsseldorf. St Gall Stiftsbibliothek for Plates 1, 2 and 3 from Susan Rankin (1995), ‘Ways of Telling Stories’, in Graeme M. Boone (ed.), Essays on Medieval Music in Honor ofDavid G. Hughes, Cambridge, MA: Harvard University Department of Music, pp. 371-94.

Every effort has been made to trace all the copyright holders, but if any have been inadvertently overlooked the

opportunity.

publishers

will be

pleased

to make the necessary arrangement at the first

Series Preface

This series of volumes provides an overview of the best current scholarship in the study of medieval music. Each volume is edited by a ranking expert, and each presents a selection of writings, mostly in English which, taken together, sketch a picture of the shape of the field and of the nature of current inquiry. The volumes are organized in such a way that readers may go directly to an area that interests them, or they may provide themselves a substantial introduction to the wider field by reading through the entire volume. There is of course no such thing as the Middle Ages, at least with respect to the history of music. The Middle Ages if they are plural at all get their name as the temporal space between the decline of classical antiquity and its rediscovery in the Renaissance. Such a definition might once have been useful in literature and the fine arts, but it makes little sense in music. The history of Western music begins, not with the music of Greece and Rome (about which we know far too little) but with the music of the Latin Christian church. The body of music known as Gregorian chant, and other similar repertories, are the first music that survives to us in Western culture, and is the foundation on which much later music is built, and the basis for describing music in its time and forever after. We continue to use the term ‘medieval’ for this music, even though it is the beginning of it all; there is some convenience in this, because historians in other fields continue to find the term useful; what musicians are doing in the twelfth century, however non-medieval it appears to us, is likely to be considered medieval by colleagues in other fields. The chronological period in question is far from being a single thing. If we consider the Middle Ages as extending from the fall of the Roman Empire, perhaps in 476 when Odoacer deposed Romulus Augustus, into the fifteenth century, we have defined a period of about a millenium, far longer than all subsequent style-periods (‘Renaissance’, ‘Baroque’, ‘Classical’, ‘Romantic’ etc.) put together; and yet we tend to think of it as one thing. This is the fallacy of historical parallax, and it owes its existence to two facts; first that things that are nearer to us appear to be larger, so that the history of the twentieth century looms enormous while the distant Middle Ages appear comparatively insignificant. Second, the progressive loss of historical materials over time means that more information survives from recent periods than from more distant ones, leading to the temptation to gauge importance –

–

by sheer

volume. There may be those who would have organized these volumes in other ways. One could have presented geographical volumes, for example: Medieval Music in the British Isles, in France, and so on. Or there might have been volumes focused on particular source materials, or individuals. Such materials can be found within some of these volumes, but our organization here is based on the way in which scholars seem in the main to organize and conceptualize the surviving materials. The approach here is largely chronological, with an admixture of stylistic considerations. The result is that changing styles of composition result in volumes focused on different genres– in time, or discontinuous in

tropes, polyphony, lyric

style

and usage. There

–

are

that are not of course also volumes

–

entirely separate notably those on chant

Oral and Written Transmission in Chant

and on instrumental music that focus on certain aspects of music through the whole period. Instrumental music, of which very little survives from the Middle Ages, is often neglected in favour of music that does survive for very good reason; but we do wish to consider what we can know about instruments and their music. And liturgical chant, especially the repertory known as Gregorian chant, is present throughout our period, and indeed is the only music in Western culture to have been in continuous use from the beginnings of Western music (indeed it could be said to define its beginnings) right through until the present. The seven volumes collected here, then, have the challenge of introducing readers to an –

–

history and style, and of presenting the best of recent musical taken that, scholarship. together, they will increase access to this rich body of music, and provide scholars and students with an authoritative guide to the best of current thinking about the music of the Middle Ages. enormous

swathe of musical We trust

THOMAS FORREST KELLY Series Editor

Introduction

The writing down of music is one of the triumphant technologies of the West. Without writing, the performance of music involves some combination of memory and improvisation. Isidore of Seville famously wrote that ‘unless sounds are remembered by man, they perish, for they cannot be written down’ (from Isidore’s Etymologies, translated in Strunk, 1998, p. 149). Whatever is the pre-history of chant, we can know it only indirectly, through the surviving artifacts and those artifacts all are written documents. It is a remarkable fact that the early manuscripts of chant, all of them from Frankish lands, transmit a repertory that is fully formed, –

essentially complete and melodically complex and the variants are surprisingly few, it might seem, for a repertory transmitted orally for some considerable time. Until the second half of the twentieth century, chant scholars spent time considering the body of chant as a sort of grand work of art, and applied to it the techniques of classical philology, in which a scientific attempt is made to recover a putative authorial original. Much –

excellent work has been done, and many useful results obtained (see Part III in Chant and its Origins, the companion volume to this one). More recently, however, owing largely to the pioneering work of Helmut Hucke (see his ‘Towards a New Historical View of Gregorian Chant’, Chapter 16 in the companion volume) and Leo Treitler (two important contributions are reprinted here), the consideration of chant as a repertory developed and transmitted in oral tradition has made it necessary to rethink earlier assumptions, and to apply new techniques and processes to the study of how chant develops and travels. Stylistic analysis can give some idea of whether a melody is fixed or variable; formulaic construction can give a hint of an improvisational past. It is an important step that scholars of medieval music have taken up the ideas and techniques of scholars of folklore, of oral transmission, of ethnomusicology; for the music we study is in fact an ancient music transmitted for a time in oral culture; and we study a culture not our own, whose informants are not people but manuscripts. All depends,

ironically,

on

deducing oral issues from written documents. writing of music takes on a primary importance,

not only because writing is what preserves medieval music, but also because if we know the origins and purpose of music-writing we will know a great deal about the music itself. The musical situation of the ninth and tenth centuries, when the earliest surviving notated chant books appeared, has prompted a great deal of writing. By this time the liturgy is essentially fixed, so far as we can tell; and yet it is the first moment at which we can truly say something about the melodies to which the texts of the liturgy were sung. From chant books, as from other liturgical books, historical facts can be deduced, and an attempt to discover historical layers and developments has engaged liturgists for generations. But

And

so

the

questions are different, because we can only assume that the melody we see written in unheightened neumes in the earliest manuscripts is the same as the one whose pitches we can read in later manuscripts, and that it has been sung for some length of time, longer or shorter, before its earliest surviving writing. musical

Oral and Written Transmission in Chant

Since the nineteenth century, the origin of neumes has been of concern to scholars. The as received view in the twentieth century has been that neumes arise in one of two ways a series of points, each representing a note, or as a series of signs, reducible to two (dot and stroke), and all derived ultimately from accent-signs used in the writing of language. –

Music-writing Leo Treitler’s ‘The

Early History of Music Writing in the West’ ( Chapter 1 ) gives an overview previous thinking, and he classifies early types of neumatic writing. His conspectus of neumatic writing styles provides an introduction to the look of the various styles of neume, and his division into two groups, iconic and symbolic, follows the semiotic terminology of the philosopher Charles Peirce. A semiotic approach to music-writing has the advantage, for Treitler, of considering the purpose and function of the technology of writing, rather than simply its origin and evolution. Moreover, it raises issues of the use of notation in a musical culture that is or can be at least in part independent of written records. 1 Chapter 2 Charles Atkinson’s ‘De Accentibus’, deals with the question of the origins of music-writing. Although Levy, Treitler and others seem to have dismissed the ‘accent-theory’ of the origin of musical signs as too simple to explain the complexity of Western music-writing, Atkinson suggests that prosodic accents, as expressed in ancient treatises on grammar, formed the basis for the music script known as Paleofrankish. Earlier versions of such a theory had held that the acute accent became the medieval virga, the grave accent the punctum and the circumflex accent the clivis. Atkinson demonstrates that the prosodic signs make sense only in the context of the special Paleofrankish script, and that there is an unquestionable link between the prosodic signs and the rise of music-writing. of

,

Notation and Performance The

and significance of musical notation are the subject of much current will be presented in later sections of this volume. But it must be and more discussion, remembered that the measure of value of musical notation for most modem musicians is the extent to which it enables the making of music. Gregorian chant has continued to be sung, in some places even until today, despite the changes wrought in the practice of the liturgy by the

origins, history

Second Vatican Council. The performance of chant in the twentieth and twenty-first centuries falls naturally into two areas: chant as part of religious devotion and chant as music for concert or recording. The two are not mutually exclusive, of course, since many monastic choirs make recordings, and many performing groups are fully aware of the religious significance of the music they sing. Nevertheless, the distinction is one worth making, since the listener who is

1 with

an

a

worshipper

in

a

monastery, and who hears each

piece

of chant in its context of

Treitler’s essay, along with his other essays presented here and a number of others, is reprinted introduction in Treitler (2003) ; readers may wish to consult this volume not only as a handy

compendium of much of Treitler’s best work, but also for the subsequent reception.

about his studies and their

author’s

own

thoughts and

second

thoughts

Oral and Written Transmission in Chant

lections, psalmody, prayer and

so on, is different from the listener whose recording most of chant that might never actually be sung together in the liturgy. likely juxtaposes pieces Part II considers the performance of chant in our own time, and the various ways in which

chant notations are interpreted. In Chapter 3 Mary Berry, a distinguished scholar and performer, reviews seventy-five years of chant performance, and makes us aware that the prevailing style of performance for the early years of recordings is that of the monastery of Solesmes. This is

surprise, since Solesmes is the editor of the chant books of the Roman Catholic Church, and many of its books incorporate aspects of the rhythmic theory developed by Dom André Mocquereau and beautifully exemplified in the many recordings issued under the direction of Dom Joseph Gajard and his successor Dom Jean Claire. It is a beautifully smooth style some

no

–

would say Debussian predicated on the equality of notes, and on gradually shifting rhythmic emphases based on small musical units of two or three notes (what Dom Mocquereau called the ‘fundamental rhythmic cell’). –

Berry is also concerned to recognize, however, the enormously important contribution of the Solesmes monk Dom Eugène Cardine to our knowledge of chant performance. Cardine’s ideas about Gregorian rhythm, developed over many decades of study and teaching at the Pontifical Institute of Sacred Music in Rome, derive from his scrutiny of a small number of early manuscripts that contain particularly nuanced versions of neumatic notation, in which alternative ways of writing notes, and of grouping and disaggregating them, reveal information about the relative lengths and weights of notes. Much of Cardine’s theory is summarized in his His teaching in Rome produced a series of devoted students, and there is now what many call a ‘Cardine school’, or a ‘Semiological school’, of scholars and performers, many of them in Germany, who are committed to research and performance along the lines traced out by Cardine; the journal Beiträge zur Gregorianik serves as a sort of house organ. In Chapter 4 Nino Albarosa, one of Cardine’s most passionate students, describes

Sémiologie grégorienne (1970).

the effect of Cardine, and the studies of many of his students, in his account of Cardine’s work at the Pontificio Istituto de Musica Sacra. These and other, wider issues concerning chant performance are usefully summarized by Lance Brunner in ‘The Performance of Plainchant’ ( Chapter 5 ). Much has happened since Brunner’s survey, but he is able to point out to the reader many issues that are of concern to historians, scholars and listeners today and will continue to be so in the future.

Oral and Written Transmission Back to the early neumes: we have more than once alluded to the inevitable question of how chant came to be what it is. As to what it is, we can only know it after it is written down. But what musical practices preceded the writing, in the tenth century, of complete chant books? Is it possible that the chant was already centuries old when it was written, and that it had

changed very little in those centuries? How else can we explain the remarkable consistency in the written tradition once we find it? And why was it written down then, or at all, if it could be held in memory for centuries? And what is the ‘it’ that is held in memory? In his now-famous article ‘Homer and Gregory’ (Chapter 6 ) Leo Treitler attempts a view of making and performing chant without notation.

Taking as his starting point thinking about the techniques of oral composition and re-composition in performance, as described and theorized by Millman Parry and Albert Lord,

with respect to Homeric tradition on the one hand and living epic singers on the other, Treitler considers Gregory the Great in the sense of Homer: an authority figure who stands for a large and living tradition. How ecclesiastical singers produced and reproduced chants is his real subject, and a simple explanation of a fixed repertory in a fixed form is not the only, or the best, explanation, in view of the way in which Treitler sees the operation of human memory and tradition. More

importantly, it has to do with how music comes into existence –as an authorial archetype constantly reinvented composition. This is why the nature of writing is so – not important only because it gives us our earliest music, but because how we interpret the act of writing itself may tell us something about the music it seeks to represent. Certainly by the time of writing, the repertory of Gregorian chant was a body of individual pieces, even though the careful consideration of the surviving music may well give evidence of formulaic structure suggesting a past that involved improvisational variability or evidence of gradually changing oral tradition. In Chapter 7 his trilingually-titled ‘“Centonate” Chant: Übles Flickwerk or E pluribus unus?', Treitler addresses this issue, considering as he does so some of the major writers on Gregorian form and esthetics: Peter Wagner’s Gregorianische Formenlehre (1921), Paolo Ferretti’s Estetica gregoriana (1934), and François-Auguste Gevaert, whose study of the Gregorian office antiphons, La melopée antique dans le chant de l ’église latine (1895), identified what we now call melody-types, indicating that antiphons could be created on the basis of generally understood melodic procedures rather than as deliberately and self-consciously independent works of art. This leads Treitler to a consideration of the phenomenon of recurrent passages of music in certain genres of chant, what Ferretti called ‘centonization’, the assembly of a ‘patchwork’ composition out of pre-existent melodic units or formulae. Ferretti’s conception of assembling a piece from constituent elements allowed for an aesthetic in which individual pieces are nevertheless individual works. Treitler’s essay likens this conception of art to the difficulties that arose in the late eighteenth century when the epic poetry of Homer was recognized as being created perhaps along similar lines, challenging the work-concept of great individual creation. Chant is perhaps a set of procedures, adopted anew at each occasion, whose written form reflects a frozen version of what might have gone somewhat differently but always in a well-formed and shapely way on a different or as a

,

–

–

occasion. The matter of oral transmission is an area of animated discussion and varying opinion, and the discussion has not subsided, as the reader will see. David Hughes, in Chapter 8 ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’, argues that there is no evidence for any sort of transmission other that what we might think of as normal: a written or oral passing on of fixed melodies from one person to another. His consideration of the variants in chant transmission, and the remarkable fixity of the chant as we have it, makes any sort of continuing re-improvisation unlikely, however much the chant may have originated in oral tradition in the distant, for the moment unrecoverable, past. That the chant is transmitted faithfully in our surviving documents is a remarkable thing; it would be not so remarkable, however, if the practice of writing the chant had originated not with our earliest chant books of about 900, but with a neumed archetype a hundred years earlier. This latter scenario is suggested by Kenneth Levy in Chapter 9 ‘Charlemagne’s Archetype of Gregorian Chant’. The notion that musical notation in neumes began at about the time of ,

,

our a

earliest

sources

short time earlier

might seem self-evident, especially as the surviving sources of chant from are without such notation. But Levy suggests that, based on a variety of

interesting ‘indices’, a complete neumed book of mass-chants must have existed at the time Charlemagne. The absence of physical evidence is countered by Levy with suggestions about distribution and about similarity of neumations in distant places. The issue has been heatedly debated, as the reader will soon discover. We reprint as Chapter 10 from the Journal of the American Musicological Society, a flurry of communications from Treitler, Hughes and Levy on many of the matters arising from these issues and their specific contributions: Treitler reacts to Levy’s idea of a neumed archetype about 900; Levy reacts to that; and Hughes reacts to Treitler’s communication. 2 That such varying views are held by such distinguished scholars says much about the difficulties inherent in the study of music so distant from ourselves. Some further discussions of oral tradition, writing and history broaden the scope of the discussion. László Dobszay, in Chapter 11 ‘The Debate about the Oral and Written Transmission of Chant’, surveys the issue and brings to bear the substantial experience of Hungarian scholarship in the collecting and study of folk-song and other folk traditions. Kenneth Levy weighs in again with ‘On Gregorian Orality’ ( Chapter 12 ), in which he uses multiple writings of an offertory to suggest a means of viewing and understanding the variations that arise in oral composition and transmission before the age of writing. And Emma Hornby, a scholar known for her careful and detailed analyses of Old-Roman repertories, in Chapter 13 reviews Levy’s work with respect to the early writing of chant, and in a carefully reasoned study, reviewing much evidence and scholarship, considers that Levy’s case for an early neumed archetype is not the only possible explanation for the situation as we find it, and argues that of

,

,

oral transmission cannot be ruled out. Edward Nowacki’s Chapter 14 ‘Chant Research at the Turn of the Century’, is a translation and commentary of an important but underread essay by the eminent scholar Helmut ,

Hucke. Most of Hucke’s work is in German, but his influence on chant scholarship in the twentieth century was significant. In this essay Nowacki translates and comments on Hucke’s consideration of the musical aspects of Gregorian and Old-Roman chant. This might be read

historically in conjunction with materials in the companion volume on the relationship of those two repertories, but Hucke’s concern here is with musical material and how it comes to be, and the role that oral transmission might play in the formation of melody. Susan Rankin’s ‘Ways of Telling Stories’ ( Chapter 15) approaches the beginnings of writing by the detailed study of a single place in this case the important monastery of St. Gall. By a careful study of the earliest musical notations at the monastery, she is able to show the –

‘consistent and competent integration of musical notation into [the] writing techniques’ (p. 408) of its scriptorium in the last quarter of the ninth century. In Chapter 16 Theodore Karp, who has worked for decades on relationships within Gregorian melodies, considers how it may be that melodic elements, normally thought to be characteristic of a mode and a genre, can sometimes be found in a wider range of melodies. Karp attributes this to ‘crossing’, the phenomenon whereby a singer is reminded by similarity of a parallel passage in another place, resulting in a sort of mixture that combines elements. Karp’s examples make clear that this is a feature of the chant before writing, but also a practice

Much of Kenneth Levy’s interesting work on chant (including the essays in this volume companion volume) has been reprinted, with revisions and commentaries by Professor Levy, in his Gregorian Chant and the Carolingians (1998), allowing the reader to see the full extent of Levy’s considerable contribution to the discourse on the early history of chant. 2

and its

that continues into the literate phase of transmission. He posits an early stage of chant in which categories of mode and genre were far from fixed. Karp’s essay here is related to his larger work, Aspects of Orality and Formularity in Gregorian Chant (1998), which repays careful study. Even from this short introduction the reader will be able to discern the paradoxes of the

study of medieval chant. The sources are too many for one scholar to comprehend and yet too few to answer our questions. The early history of chant is a history of orality, of transmission by mouth to ear, and yet we can study it only through the use of written documents. The chant is said to originate in Rome, and yet the sources originate from Gaul. There is much wisdom and much information available to the reader in these two volumes on chant, but the wisest scholars are those who learn, as the reader will learn, that the amount of work remaining before us is much greater than what lies behind. References Cardine , E.

‘

( 1970 ), Sémiologie grégorienne ’, Études grégoriennes, 11 , pp. 1 158 (English trans., Robert M. Fowels, Gregorian Semiology, Solesmes, 1982 ). Ferretti , Paolo ( 1934 ), Estetica gregoriana: trattato delle forme musicali del canto gregoriano , Rome -

.

Gevaert ,

François-Auguste ( 1895 ), La melopée antique dans le chant de I’église latine Ghent Karp Theodore ( 1998 ), Aspects of Orality and Formularity in Gregorian Chant Evanston, IL : Northwestern University Press Levy Kenneth ( 1998 ), Gregorian Chant and the Carolingians Princeton, NJ : Princeton University .

,

,

,

.

,

,

Press

.

Strunk , W. Oliver ( 1998 ), Source Norton

Readings

in Music

History

,

rev.

edn, ed. Leo Treitler New York : ,

.

Treitler, Leo ( 2003 ), With Voice and Pen: Oxford : Oxford University Press

Coming

to Know Medieval

.

Wagner Peter ( 1921 ), Gregorianische Formenlehre Leipzig ,

,

.

Song

and How it

was

Made ,

Part I

Music

Writing

DOI: 10.4324/9781315090283-1

[1] The

Early History

of Music

the West

*

Writing

in

By LEO TREITLER

PERSPECTIVE on

the

early history of Western music has

been

OUR changing somewhat lately in consequence of the serious attention being given to the fact that the Gregorian Chant tradition was, in its early centuries, an oral performance practice. In this essay I wish to focus on the other side of that picture: the nature, origin, and evolution of the early practice of writing music. was translated after the ninth century into from a performance practice represented in But the evolution writing. to a tradition of writing, composing, transmission, and reading, took over a span of centuries. Written sources originating during that place period 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 1 more as a model for performance than as a blueprint. In the framework of these interpretations, old questions about the nature, function, and early history of music writing present themselves

The oral tradition

in If we

a new are to

light.

do

more

than to hold these ideas about the uses of early sources of individual songs as intriguing

music books and of the

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? *

This paper is an expansion of a lecture first presented at Brandeis University in 1981. Its final version was written while I was a fellow at the Humanities Research Centre of the Australian National University. 1 Full expositions of these interpretations, and citation of further literature, are given in Helmut Hucke, “Toward a New Historical View of Gregorian Chant,” this JOURNAL, XXXIII (1980) , 437-67, and Leo Treitler, “Oral, Written, and Literate Process in the Transmission of Medieval Music,” Speculum, LVI (1981) , 471-91.

April,

DOI: 10.4324/9781315090283-2

Oral and Written Transmission in Chant

changes in the nature and function of notation correspond to long-range changes in the relations of transmission? These questions must lead to the study of notations in the light of

What

these

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?

their

There from

is

thus

a

paleography

shift of focus to

the

in

investigation of early notations paleography, which concerns

semiotics. Musical

itself with the classification of

signs

and the identification of their

periods places of use, has been a mature discipline since early in the century. In the semiotics of musical notation, which would and

concern

itself with the functional relationships between sign systems they signify while taking into account the situation of the

and what

person(s)

to whom

they signify, virtually everything

remains to be

done. 2 I we think about the way that musical notation functions I altogether? propose to approach that question from the vantagepoint of our modern pitch notation. The purpose will be to set out

How do

some

general

The first

concepts and terminology. of our system is the

metaphor that the pitch pitches of higher and lower frequencies occupy higher and lower positions, respectively, along the vertical dimension, and that the magnitude of an interval between two points of the spectrum is the length of the segment 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. premise

spectrum has vertical extension, that

This is concretized as the staff or system of staves, which also has horizontal extension in order to provide for the representation of

pitch-succession through time. The staff pitch-spectrum extended through time. 2

is in effect

a

sign

for the

“Semiotics” is the English term for the study of signs and signification, American philosopher Charles S. Peirce (Collected Papers [Cambridge, Mass.], 1931-58). “Sémiologie” is the French term, following the Swiss linguist Ferdinand de Saussure (Cours de linguistique générale [Paris, 1915]). Despite its title, Eugène Cardine’s “Sémiologie grégorienne,” Études grégoriennes, XI (1970) , 1—158, does not take up the problem that is posed here. Its interest is restricted to the investigation of the notational expressions of given melodic passages in selected sources. While that could be the beginning of a semiotic investigation, it does not yet constitute one.

following the

Oral and Written Transmission in Chant

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. 3 What I have just described The

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. A symbol is a sign that represents by virtue of

habitual association between the sign and its referent. There need be anything inherent about the sign itself that suggests what is represented in order for it to function in the symbolic mode, only the

an

not

consistent habit of

having

it stand for that

thing.

A clear

example

would be the male and female signs, ♂ 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 4 component in our system of pitch notation. In placing “notes” on the staff we are in effect

picking out positions 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 in the two-dimensional field determined

3 Peirce, Collected Papers. See also Roman Jakobson’s further articulation of Peirce’s ideas, in “Language in Relation to Other Communication Systems,” Selected Writings, II (The Hague, 1971), 697-708. A summary is given by Terrence Hawkes, in Structuralism and Semiotics (Berkeley and Los Angeles, 1977), pp. 123-50. 4 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 an habitual one between form and meaning. But some will recognize in the sign a vestigial letter “F,” the reference tone for the clef. This demonstrates a most important point that I shall want 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 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.

which the

phrase is a metaphor succession of

“melodic line” provides a clue. That that captures our sense of melody as more than a

expression

common

We express that sense also by is, we conceive of melody

pitches.

“melodic motion.” That

speaking of a single,

as

thing, and we read the notational representation of melody graph or picture of melody, in which our mind’s eye or ear

continuous as a

dots.

connects the

the notation is is and that the second of Peirce’s modes iconic mode, functioning of representation. A sign functions in the iconic mode when it In

representing melody by drawing a picture of it, in

an

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” is symbolic. 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 represented can be expected to cross the road in that vicinity. If we have not learned 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 represents

both the active involvement of the beholder in the process of 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 account

representation,

whole and

duplicate

it in

our

performance.

We

can

do that

more

the melodic patterns of the idiom are conventionalized readily, and familiar to us. For most Western readers of modem Western the

more

notation the iconic

score

of

a

mode, whereas

tonal

melody

in the

score

will function

of

primarily

in the

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 this would obviously not be true of specialists in the performance of non-tonal a

music. It is

a

special feature, and it has proven to be an advantage, of our pitch notation that it has this versatility with respect to its

system of

potential for

use as a more symbolic or more iconic representational From our own experience with it, it seems reasonable to system. 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

signs

are

are

symbolic—the pitches

symbols

of those

names

the medieval West)—and do

reading. They

lack the

are

(like

not lend

isomorphism

named, and the notational

some

early

letter notations in

themselves

with musical

to

such swift

configurations

is the underlying principle of the Western system. Such notations be very precise, but they are not made for fluent reading. 5 Peirce identified

that can

representation, which he called by returning to our bring two traffic and the animal-crossing sign. They signs, the one-way sign not only identify, they call for an action or a behavior: not to enter the a

third mode of

the indexical mode. We

that into view

can

or to proceed with caution so as not to collide with 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 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

one-way street, creatures

warning.” And it can function,

in the last case, as an imperative—a this typewriter as I approach the end

as

road sign, the bell that rings on of each line, punctuation marks. Musical notation has an indexical component in this imperative sense. It is most apparent in tablatures, which tell the player where to his place 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

signs, and the exact hierarchical arrangement will depend the way the system is used, the nature of the referent field, and the 6 competencies of the readers. system of

on

5

Walter Kaufmann, Musical Notations of the Orient (Bloomington, 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 tone 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,” The New Grove Dictionary [London, 1980], 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 modem notation can also be considered an “analog” 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 “analog” indicators. 6 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

There is yet

one

other

important differentiation

to be

made in

analyzing the nature and function of musical notations, which is best approached through a parallel differentiation in language-writing. “Spoken words are the symbols of mental experience and written words are the symbols of spoken words.” 7 John of Salisbury, scholar and Bishop of Chartres in the twelfth century, wrote: “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.” 8 The poet Francisco Gómez de Quevedo (1580-1645) wrote: Aristotle wrote:

deserts with few, but wise books bound I live in conversation with the departed, and listen with my

“Hiding in the peace of these

together

the dead.” 9

Evidently for all three, meaning was signified by spoken, not the written word. The path of representation went from graphic sign through sound. This concords with evidence that to

eyes the for

a

long time

before silent

reading

became

normal,

the contents of

through the ear, that is, through reading aloud. On the other hand I. J. Gelb, in his classic A Study of Writing,

books

were

received

“Fully developed writing became a device for expressing linguistic elements by means of visible marks.” 10 The difference reflects

wrote:

two

different

sorts of historical situation—a norm for oral communication

dominating

for

a

considerable time after the introduction of

writing, and an advanced state of literacy in which silent reading is the norm and the referent of writing is the linguistic structure itself. These may be, but are not necessarily successive stages. They may be conditions that coexist simultaneously with respect to different languages, e.g., Latin and the vernacular languages in the Middle Ages. Such a distinction must be recognized in the uses and communications of music writing. When we read a score for study we do not necessarily reproduce in our minds the sounds of music. What the score

can

represent

to

us

is the structure of the music. 11 The

decline in the use of letter notations in the West, is attributable to the general unsuitability of symbolic notations for fluent reading. 7 See Ignace J. Gelb, A Study of Writing (Chicago, 1952), p. 13. 8 See M. T. Clanchy, From Memory to Written Record (London, 1979), p. 204. 9 “Retirado en la paz de estos desiertos, Con pocos, pero doctos, libros juntos, Vivo en conversación con los defuntos, Y escucho con mis ojos a los muertos.” Francisco Gómez de Quevedo, Poemas morales, No. 131, Obra poética, I, ed. José Manuel Blecua (Madrid, 1969), 253. 10 Gelb, A Study of Writing, p. 13. Emphasis added. 11 This conception of the relationship between a score and the work it represents has been presented by Benjamin Boretz (“Nelson Goodman’s Languages of Art from a Musical Point of View,” in Perspectives on Contemporary Music Theory, ed. Benjamin

of reading scores in this way reflects an advanced state of musical literacy in the culture, and it does not enter into the early history of music writing at all. Drawing all of 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 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 paleographic one. A conception of the historical development, if it is to be more

possibility

than

a

only

from the

description of the forms in chronological sequence,

standpoint.

analysis

can

of individual notations from such

a

proceed semiotic

12

II How did the was

early writers and readers of music conceive of what it signified? To begin with, what we call

that their notations

as an attribute of in as spoken language, particular scriptural language, 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.” 13 The earliest formal analyses that have come down to us derive the divisions of the melodies from the divisions of the texts they declaim into sense units. And the analyses are couched in the

melody

seems

earliest to have been understood

Boretz and Edward T. Cone [New York, 1972], pp. 31-44) and Kendall Z. Walton (“The Presentation and Portrayal of Sound Patterns,” In Theory Only, II/11-12 [1977],

3-16). 12

I wish

to

make

explicit

that this

approach

the

to

study

of notations entails

a

rejection of the analysis of Nelson Goodman in Languages of Art (Indianapolis, 1968) on virtually every possible ground: 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. 13 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”). .

.

terminology

of

language

structure. It would not

be

overstating the

case to say that melody was conceived of and functioned extension of the prosody of language. 14

as

an

The elemental melodic phenomenon, in that conception, was the the voice in declaiming a syllable of text; as it was also

movement of

put, it

was

the movement of

melody itself

in that act. The ninth-century

music scholar Aurelian of Réôme referred to such inflections with the term “accentus.” 15 This interpretation is reinforced in the description by the tenth-century poet Notker of the relation between words and music in the

“Every

sequence:

movement

16

syllable.” That describes sequence. But in general inflections were conceived

of the

melody

should have

a

single

call the syllabic setting of the it suggests that the syllables and their what

as

the

we

primary units of melody. syllables that are the referents

It is those melodic inflections of the

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 17 signs themselves. To be sure, the figures passed through individual and that is reflected in the formation of the neumes. In notes, alphabetic notations the aspect of pitch sequence is most to the fore. But in the overall picture these are quite exceptional (and even they are written so that the letters cluster in groups corresponding to neumes). 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 18 accompanying the declaiming of syllables of text. On this subject cf. Ritva Jonsson and Leo Treitler, “Medieval Music and in Studies in the History of Music, I: Music and Language (in press). 15 Aurelian of Réôme, Musica disciplina, ed. Lawrence Gushee, Corpus Scriptorum de Musica, XXI (n.p., 1975), II, 76. 1424-77. It has sometimes been mistakenly thought that Aurelian refers 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 Jacques Handschin, “Eine alte Neumenschrift,” Acta musicologica, XXII (1950), 69-97, esp. 69-73, and Solange Corbin, Die Neumen, Palaeographie der Musik, I/3 (Cologne, 1977), 18. 14

Language,”

16

Jacques Froger, ‘L’Épître

de Notker sur les lettres significatives,” Études V (1962), 23-72. Cf. Michel Huglo, “Les Noms des neumes et leur origine,” Études grégoriennes, I

grégoriennes, 17

(1954), 53-67. 18 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 analog-digital distinction. It is again

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. 19 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 (Appendix, Figure I ). 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 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

important

to

convey. III

Now to the signs themselves. Table I gives a selective overview of neumes in use in Europe and England between the ninth and twelfth centuries. It is introduced here is

no

claim for its

as a

point of departure only, and

there

completeness. 20

apparent in the notation of duration, in the contrast between modal and mensural notation, and again in the contrast between 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. 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 accomodate 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 treated at length in my essay “Homer and Gregory: The Transmission of Epic Poetry and Plainchant,” Tbe Musical Quarterly, LX (1974), 333-72, section II. 20 The Table is basically that of Corbin (Die Neumen, Anhang). 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 19

TABLE I

A few definitions will facilitate the sorting out of the paleographic material as it is briefly summarized in the Table. Character. The class of representations of a particular melodic inflection. In Table I the characters each

one

headed

that

began

the

are

shown

as

vertical

columns,

it in tables of essentially this form

by given 21 produced in the eleventh century. The characters had become standard by the time of their representation in the tables of the eleventh century and after, but we do not know how long before then. That means that there is something anachronistic about speaking of, say, “Punctum” and “Virga” in sources anterior to the neume

to

tables,

name

be

and

we

will do well to hold that in mind.

inscription of a character. Each entry in the to distinguish neume from character because there can be two quite different graphic signs that are both called “Clivis,” for example, and 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 Neume. A conventional

Table is

a neume.

We need

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. (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. (f) In the row marked “Benevento” the first two forms of the Climacus have been

added. Two cautionary remarks are required. First, the appearance of a neume under the column in all the rows except that marked “Paleofrankish” 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, alternate 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 that in the first two cases alternate forms are in complementary distribution, meaning that they do not occur in the same context. In the third case alternate forms are in random distribution. The semiotic task requires us to determine wherever possible which of these relations obtains. 21 Huglo, “Les Noms des neumes.”

Virga

identifies this script as “Messine.” Solange Corbin argues persuasively that her designation more accurately fits the facts of geographic distribution.) Of the three class-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, 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 scriptcategories. Still, what is shown does correspond to paleographic realities and it 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 compound neumes.

(3)

The Punctum and

Virga

descend from Latin

grammatical

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” in respect to their formation out of dots (cf. Table I ). 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 notes, respectively. The “point neumes” function on the basis of the relative vertical positions of their component dots. The peculiarity of the classification scheme lies really 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 result from a later branching-off from the accent-neumes. Ewald Jammers subsequently posited a separate origin for the point-neumes. 22 Jacques Handschin proposed a different interpretation of the distinction that 22

See Corbin, Die Neumen, pp. is untenable.

Jammers’s theory

17-18

for

a

summary and for

a

demonstration that

opened the way to 23 shortly suggest. The

accent

a

substantial advance in

theory of origin,

as

understanding,

I shall call

it, goes back

as

to

I shall

Edmond

subsequently been transmitted and developed various contributors to the Paléographie musicale, 25 by Peter by Wagner, Johannes Wolf, Henry Bannister, Gregory Suñol, Willi 26 Apel, and Solange Corbin. To my knowledge only Handschin has publicly recognized any of the serious flaws in the theory. Certainly its widespread acceptance

de Coussemaker. 24 It has

has been premature, for it opens onto issues of great complexity touching upon every outstanding problem about the early history of notation. Nevertheless there is

our

of the

theory,

and I shall

proceed

some

from

kernel of truth in each aspect with points (1)

those, starting

(2).

and

In the Scandicus column of Table 1

compound

of two Puncta and

a

,

the

most common

Virga. (Only script

8

form

is a

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

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 common to all scripts and thus constitutes a premise of the system as a whole. I shall call it the principle of directionality. 27 such

a

Handschin, “Eine alte Neumenschrift.” Edmond de Coussemaker, Histoire de I'harmotiie au moyen age (Paris, 1852; repr. Hildesheim, 1966), Chap. II, “Origines des neumes.” 25 Paléographie musicale (Solesmes, 1889- ). 26 Citations in Corbin, Die Neumen, p. 19, to which must be added Willi Apel, Gregorian Chant (Bloomington, 1958), pp. 108-109. 23

24

27

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 premise of the spatial metaphor that I identified

at the beginning, are distinct in directionality as a principle in the

their

exploitation of that premise. I mean (1) formation of individual neumes; (2) diastematy, which is the representation of melodic interval-size through vertical distance on the writing surface; 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

reflect the

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 are 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 that is 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 separately, because of certain idiosyncracies they

rather discuss them

And in two of those scripts, 7 and 10, the Virga is rarely or 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

display). not at

only the top of an ascending group, high note absolutely (7, 9-10).

or

the motion to the top, not

a

Now consider the three-note groups with change of direction and Porrectus). In scripts 7b and 10 their components are

(Torculus

quite evident:

in

7b the Torculus comprises

three Puncta and 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. Covering all three phenomena with the term “diastematy” distorts the understanding of individual scripts and of the historical relations among them (e.g., by characterizing instances of the first and third phenomena as deficient instances of the second).

Porrectus Punctum-Punctum-Virga. That is also how the Porrectus is made in io, 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 groups therefore confirms the given in the preceding paragraph.

characterization of that class Can

we

make inferences about the components of the Torculus they are not in plain view? It will be instructive

and Porrectus where to

try. The Torculus in all such

Punctum-Virga-Punctum Punctum-Punctum-Punctum. by extension of the reasoning

cases

(or Pes-Punctum),

or

would be read

as

might be preferred, just the Pes. But there is no independent reason for preferring regarding 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 or scripts would be a compound of Virga-Punctum-Virga j j 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

The former

presumably again

one or

that in the

represents

only

the other of those

scripts

the top of

configurations.

of the B class the

This confirms

Virga,

when

used,

an ascent. 28

28 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,” Acta musicologica, XXXV [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 really been

With this There

Porrectus in mind we return to the Clivis. cursive forms: that in the A-class scripts beginning, like

analysis of the

are two

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. the

If we take the Clivis and Porrectus of the B scripts to be compound the initial horizontal element would not be the most direct

neumes,

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 scripts accomplish more effectively is

and Porrectus of the later B 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,

compounds,

represents high This differentiation establishes a

origin lower

or

a

and the Punctum

a

alone

lower

or

in

one.

29

symbolic mode of representation. The

entirely arbitrary, and whatever may have been its nothing innate about the signs that would suggest the

convention is there is

note

higher position

of the note that is

represented.

scripts are based on the principle of directionality: the vertical position of the note-signs vis-á-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, The B

their

neumes

constitute

a

graphic

imitation of the

contours

of the

melodic line. 30 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. 29 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 defining 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 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 Punctum. Such a rule has in fact been 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. 30 Handschin, “Eine alte Neumenschrift,” effectively recognized these two scriptclasses 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 I'harmonie) had already glimpsed

something

of the distinction. There will be

more to

say about Handschin’s paper.

Because of the

sharpness of their differences

the versions of the Clivis and Porrectus marks of the two of a

script-classes.

In

general

in form and

be taken

as

function,

identifying

if the Clivis and Porrectus

less these

the script can script configurations: J be expected in most cases to have all the other characteristics of A scripts. If they are in more or less these configurations: J the script can be expected in most cases to have all the other characteristics of the B scripts. Henceforward I shall refer to the A scripts as symbolic because the symbolic mode dominates in them. (The iconic mode does still function in them, 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. are

in

can

more or

(This is not to forget that the very basis on which the iconic mode can function, the spatial metaphor, is itself a symbolic principle.) They are scripts 7-10. No scripts are primarily indexical, although the indexical mode functions in scripts of both major classes. 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 historical development in the relationship between symbolic and iconic scripts, what

can

that suggest about

larger patterns of cultural development? IV

Is an historical development vis-á-vis symbolic and iconic writing discernible? In the present state of our knowledge we cannot give

chronological priority do

to one or

the other notational mode.

Specifically,

know whether the Paleofrankish and early Aquitanian on the one hand, or the Germanic ones, on the other, were the scripts, earliest ones in use, and we do not know what position should be given to the Breton script vis-á-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 be best approached first by way of an we

analysis

not

of the Beneventan

script,

which I have

so

far put aside.

Table I shows, for that

both the iconic and symbolic forms manuscripts those forms are in distribution. For both characters the symbolic

script,

of the Clivis and Porrectus. But in the strict

complementary

form ( ) is used when its first note the preceding note. The iconic form (J

J

of the preceding note that the “symbolic” form is 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 ha/e been analyzed. 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

represents

descent from

represents an ascent from ) is used when its first

a

(Appendix, Figure ). 2

But that

or

repetition

means

note, it begins with an upstroke. When it repeats the is below it, it begins with a Punctum (Appendix,

or

preceding note Figure 2 ). 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 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, i.e., its shape represents the direction of the melodic line from the preceding note to the note that

(see Appendix, Figure 2 ). In the manuscript represented there, Benevento, Archivio cap., VI.34, only the form with downward slant is used. The Beneventan manuscripts Rome, Bibl. Vat., lat. 10673 and Benevento, Archivio cap., VI.35, 38, and 40 use both

it represents

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

have identified elsewhere. Its greater

we

a matter

of the extension of the

shown in

Figure

2

is

explicitness

is in effect

of directionality. (The

principle quite precisely diastematic

well, but separately here.) as

script

that is

a

separate matter, and it will be discussed In one important regard the Beneventan script differs from most of the iconic scripts. Single notes are represented by both Punctum and

In the late

Virga.

eleventh/early twelfth-century gradual Benevento,

Archivio cap., 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

just the sort of operational rule for the Virga that cannot be inferred for the of the notations North. 31 There are exceptions in early symbolic Archivio Benevento, cap., VI.34, but not nearly so many as in the earlier Beneventan manuscript, Rome, Bibl. 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 of aspect the sharpening in precision of the script, along with exact 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 precise than

following

note.

Now this is

differentiation of Punctum and

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

“the

signs

representing them, then we would expect to find a strict signs in the earliest sources and a more one as notations became more clearly diastematic. But the

for

differentiation between the casual

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 the

question

writing

adaptation

of

once-symbolic

arises whether this

suggests

a

forms for iconic

general tendency in music

toward the dominance of the iconic mode.

cannot answer that

question

there is evidence for such

a

with respect

tendency

function,

to

Although

we

the very earliest stage,

after the first hundred years. 32

n. 29. Even in the St. Gall script there are inconsistencies within the 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 Virga where another has Punctum in the same passage. Examples of this may be seen in Corbin, Die Neumen, pp. 48-52. 31

See

same

32 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

I have identified reinforced iconic of the twelfth century and after. 33

scripts

in the

following sources

French

Paris, Bibl. Ste. Geneviève, 547 (Solange Corbin, Répertoire de manuscrits médiévaux, I [Paris, 1965], Plate 6). 2. Troyes, Bibl. de la ville, 1047 (Paléographiemusicale [hereafter PM] III, Plate 195). 3. Paris, Bibl. nat., n.a. lat. 1235 (PM III, Plate 195b). 4. Paris, Bibl. nat., n.a. lat. 10511 (PM III, Plate 198a). Italian 1.

5- Vercelli, Bibl. cap., 162 (Appendix to this article, Figure 3 ). 6. Lucca, Bibl. cap., 601 (the entire MS in PM IX). 7. Lucca, Bibl. cap., 602. 8. Lucca, Bibl. cap., 603 (PM II, Plate 34). 9. Lucca, Bibl. cap., 609 (PM II, Plate 35). 10. Pistoia, Bibl. cap., 121 (Bruno Stäblein, Schriftbild der einstimmigen Musik, Musikgeschichte in Bildem, III/4 [Leipzig, 1975], Plate 24). 11. Pistoia, Bibl. cap., 119 (Stäblein, Schriftbild, Plate 25). 12. Padua, Bibl. cap., 47 (Appendix to this article, Figure 4 ). 13. Verona, Bibl. cap., C.V. (PM II, Plate 14a). 14. Monte Cassino, Bibl. dell’abbazia, NN. 339 (PM II, Plate 19). 15. Monte Cassino, Bibl. dell’abbazia, NN. 540 (PM II, Plate 21a). 16. Monte Cassino, Bibl. dell’abbazia, NN. 546 (PM II, Plate 22). 17. Rome, Bibl. Vat., Barb. XII, 12 (PM II, Plate 23). 18. Rome, Bibl. Vat., lat. 5319 (PM II, Plate 28). 19. Rome, Bibl. Vat., lat. 6078 (Stäblein, Schriftbild, Plate 23). 20. Rome, Bibl. casanatense, C. IV. I (PM II, Plate 29). 21. Naples, Bibl. naz., VI. G. 38 (PM II, Plate 30). 22. Paris, Bibl. nat., n.a. lat. 1669 (PM II, Plate 31). 23. Rome, Bibl. vallicelliana, C. 32 (PM II, Plate 33a). 24. Cortona, Bibl. com. 12 (PM II, Plate 33b). 25. Modena, Bibl. cap., O. I. 16 (PM II, Plate 36). 26. Modena, Bibl. cap., O. I. 7 (PM II, Plate 37a). 27. Turin, Bibl. naz., F. IV. 18 (PM II, Plate 37b). 28. Bologna, Bibl. univ., 2748 (Stäblein, Schriftbild, Plate 19). Gothic 29. 30. 31.

London, Brit. Lib., Add. 26884 (PM III, Plate 135). Zwettl, Stiftsbibl., 199 (Stäblein, Schriftbild, Plate 11). Oxford, Bodl. Lib., Rawl. C. 892 (Stäblein, Schriftbild, Plate 12). English

32. London, Brit. Lib., Add. 12194 (PM III, Plate 201). 33. London, Brit. Lib., Lansd. 462 (Appendix to this article, Figure 5 ). and follow the direction of

greater detail. 33 The location of

development,

published

parentheses following each

both

geographical

facsimiles of notational

source.

and institutional, in

specimens

is indicated in

There

are

earlier

and 32,

2, 3, 5,

sources

at

least,

of the

same areas

in which

a

of provenance

as

Numbers is used:

strictly symbolic script

Troyes, Bibl. de la ville, 894 (Corbin, Die Neumen, Plate 25). 3. Paris, Bibl. nat., lat. 9449 (Corbin, Die Neumen, Plate 29). 5. Vercelli, Bibl. cap., 161 (Appendix to this article, Figure 6 ). 32. All English sources of the eleventh century, with the exception of the organum notation of the Winchester Troper (see below). 2.

This is strong evidence for a widespread tendency to change over symbolic to iconic notations. A hint of the sense of such

from

transformation is

provided by a variant of the rule for the forms of the by the scribe of the Sarum Gradual. He

Clivis and Porrectus used writes

J

where the

neume

whenever the first note represents a descent except This begins a new syllable. Then he writes J

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, Archivio cap., VI.34 that begin with Clivis, i.e., where there is no question about the preceding note, forty-seven are written J and four are writtenJ Only five pieces begin with Porrectus; three are writtenJ and one Perhaps that a sense about these forms as an betrays symbolic underlying or J

background norm. A change of representational function is evident also in the other of the two scripts that I had put aside. Table I 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 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 ). Both devices remind us of the primary syllable of text ( J function of melody as inflection of language. In the Nonantola script the once-“symbolic” Clivis and Porrectus J are used above the text because they provide attachment lines to the syllables. In melismatic passages, where there is

no text

for the

to, the normal rules for reinforced iconic example, Stäblein, Schriftbild, Plate 16). There

writing English.

are

indications of

a

tendency

toward

iconic

in two other notational traditions: the The

manuscript Paris,

neumes to

attach

scripts apply (cf., for modes of

Aquitanian

music

and the

Bibl. nat., lat. 1240, from the tenth

surviving source of liturgical music from Its notation shows essential features of the later script that Aquitania. 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 ( J ), a form of the Pes that is otherwise found in northern French scripts (J ), and these signs for single notes: J

century,

(there

is the oldest

is

nothing resembling the Virga of the later Aquitanian sources, compound neumes). It has been observed repeatedly that Paris 1240 is more closely

except in

related with regard to its contents to Northern sources, especially the Nevers troper Paris 9449, than to later Aquitanian sources. 34 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 I shows a single 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 1121 the single notes are all Puncta, identical in form. In Paris 1240 they are differentiated as J for high notes and J for low notes (J is always and only the lowest note, a is not quite so consistent about its context and function). And what is written a and v in Paris 1240, is usually written Virga and Punctum, respectively, in Paris 9449, where those signs represent high and low notes, respectively. 35 34 Paul Evans, “Northern French Elements in an Early Aquitainian Troper,” in Speculum musicae artis: Festgabe für Heinrich Husmann, ed. Hans Becker and Reinhard Gerlach (Munich, 1970), pp. 103-10, and Ellen Ryer, “The Introit Trope Repertory at Nevers: MSS Pa. Bn. Lat. 9449 and Pa. n.a. lat. 1235” (Ph.D. diss., Univ. of

California, Berkeley, 1981), pp. 39-72, 74-213. 35

Here I must

about a methodological trap that can be entailed in comparison presupposes that we are seeing alternate things; otherwise it tells us nothing. So it behooves us to

enter a caveat

making such comparisons.

The

ways of writing the same maximize the chances that the assumption is right. The best way to do that is to choose melodies and, even better, formulas, that we know to be stable, and sources that we know to be closely related. The melody chosen for Example 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 behavior 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. 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, p. 95) compares the inscriptions of the Trisagion in

Example

1

Trope introduction to the Christmas Introit Puer natus (Ritva Jonsson, Corpus troporum [Stockholm 1975], #34)

1.

fol.

2.

Pa 1235, Pa 9449, 3. Pa 1240, 4. Pa 1118,

fol. fol. fol.

184r 7V 9r 9V

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 J is a symbol for the lowest note of a descent. In Benevento VI.34 the same sign a note lower than the preceding. The Beneventan script allows 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 J and J look like acute and grave accents, respectively, and their counterparts in the

indicates

J

North

Virga and Punctum, respectively. But the relationships do quite right from the point of view of the theory. For the would lead us to expect the J and J of Paris 1240 to be replaced theory not

are

work out

Paris 1240, Paris 903 (the eleventh-century 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 premise 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 therefore seriously flawed by their weakness.)

in subsequent Aquitanian sources. But that different script altogether; in Aquitania those neumes are Puncta. And in fact it would not really be correct to say

by Virga and Punctum in

happens replaced by that

they

a

were

replaced by Virga and Punctum in the

North. The

one

speculation 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 of Paris 1240 were adapted from a Northern script, for which there are no surviving sources, and in which similar 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 we

Winchester

Troper, the eleventh-century manuscript Cambridge, Corpus College 473. In the main, English neumatic writing was symbolic, as Table I shows. But in the notation of the new organum parts provided for some of the traditional chants in Corpus Christi 473, the repertory of forms in the script was enlarged by the Christi

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. 36 The situation of this

manuscript contributes to our understanding 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 toward more informative notations must have been motivated by the need to represent non-traditional matter, and also by the need 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 I 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 to the present of the

context

.

day

would

surely

show that those written in

symbolic scripts

I mention this only briefly here. The differences in the forms and placement of 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, who is preparing a detailed report on the whole matter. 36

neumes

outnumber

by far those

written in iconic

scripts.

It

the

was

symbolic

too, whose vestigial Virgae had become meaningless, that were adapted for the modal and then the mensural system for representing relative duration. The iconic forms in fact died out. If we ask why, a

forms,

Archivio cap., VI.34 and the Sarum Gradual and No. 32 above) will suggest a perfectly (Appendix, Figure satisfactory explanation. The iconic system was made redundant and obsolete by the universal establishment of diastematy, ultimately in look at

Benevento,

2

the form of the staff. The

symbolic-iconic

the

system

was

development of diastematy

shift. In

written with strict

some

sources

diastematy,

ran

parallel

to

the reinforced iconic

even on

the staff. That the

principle succeeded over the iconic one in the evolutionary seen in late sources of this type. The latest known to me is is history Brit. London, Lib., Lansd. 462, from the Sarum use in the fifteenth

diastematic

century (No. 33 in the list above; Appendix, Figure 5 ). In this manuscript both types of Clivis and Porrectus are used, and in the majority of cases their use follows the usual principles of 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 unclarity 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

Peter

Wagner,

with the cautious

concurrence

of

Handschin, 37

asserted that the

principle of diastematy was at work in neumatic from the beginning. That is not only because of the writing virtually of of the early diastematic scripts, but also some possible priority because the principle of directionality plays a role in the formation of at

least

scripts. This is in contrast Byzantine notation, for example.

some neumes

symbolic

nature of

in all

Handschin’s discussion of these

matters

is,

so

far

to the

as

I

strictly

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? 37

Peter

(Leipzig,

Wagner, Einführung

in die

1912; repr. Hildesheim, 1962), pp. 84-86.

Melodien, II: Neumenkunde Handschin, “Eine alte Neumenschrift,"

Gregorianischen Chap.

12.

It was his paper that suggested that standpoint to me, and I am indebted to it, despite my somewhat different interpretations. Handschin called the Paleofrankish 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 positions, 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 because all scripts can be more-or-less diastematic, and all of them have an element of directionality in their compound neumes. But there are still problems about 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 Paleofrankish

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. Gall. For suppose that a Porrectus were to follow a higher tone: the upstroke cannot represent the distance to the first tone of the neume, because that would be reached

by

a

motion from above.

despite these problems of characterization, Handschin produced virtually the same classification as that based on the iconicsymbolic principle. 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 But

usualis).

The criterion would be the

distance

to

interval

distance,

proportionality of vertical decipherability without of way saying what degree of

exact

which

means

Otherwise there is no condition suffices in order for a notation to be approximation considered diastematic. Paris 1240 and 9449 are generally said to have diastema tic and adiastematic notations, respectively, but they both show about the same degree of directionality (see Example 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

foreknowledge.

to that

according to which is that

(Paris 903, but not 1118). But Paris 1235, 9449 in Nevers, is also diastematic according to that by the same token one could say that the

that criterion

a successor to

criterion,

so

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 here.

First,

it is

spared

generally

handicap). There are two problems good practice to say that a thing has the

this

not

characteristics of its successors, but has them in deficient measure (it would be like saying that the music of Dufay exemplifies triadic but

tonality,

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,

suited their purpose. 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 that may be

of historical

development—a scriptorium converting diastematy (e.g., presumably, 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: cases

a matter

to

Symbolic Scripts

Iconic Scripts

Adiastematic

St. Gall, Bologna etc., Paris 9449 (Nevers)

Paris 1240 and 1118, Chartres, Bibl. mun., 47 (Breton), Rome, Bibl. Vat., lat. 10673

Diastematic

The Liber usualis and most medieval staff notations

Paris 903, Benevento VI. 34, Paris 1235 (Nevers), Sarum Gradual

(Beneventan)

The first

scripts

that

were

diastematically Symbolic scripts

written

Aquitanian, Beneventan).

(Lothringian, diastematically only when they were put on the

are were

iconic writen

staff. Diastematic and iconic writing surely arise out of the same conception about the task of notation and how it should be carried out. It is hardly surprising that they came together earlier. Despite the fact that iconic scripts as such became obsolete, the principle of representation of which they were the first embodiment was carried forward in the diastematy that became universal. There is, however, one essential difference between the iconic principle for the formation of neumes and the diastematic principle for their disposition on the page. The focus of representation in the former is the direction of the melodic line. It retains the essential nature and purpose of neumatic scripts, the representation of the inflection of the syllables of language. The focus of diastematic writing—especially with the introduction of the staff—is the representation of relative pitch. The invention of diastematy must be associated with the shift in the nature of Western notation, from a system of signs for the inflection of language to a pitch notation. Of course the concomitants and consequences of that were boundless. VI

From this perspective on the early evolution of music writing, it will be of interest to return to the comparison with the history of language writing. There are engaging parallels, and as they concern different media, different cultures, and different historical periods and time-spans, they have something to contribute to the history of 38

human

cognition. The earliest writing depended on iconic representation: pictures that expressed meaning directly, without intervening linguistic forms. At the next clearly definable stage the signs had become symbols through a process of abstraction, and they had come to represent words through a process of phoneticization (the attachment of sound values to the symbols). Such a system is called logographic. But no completely logographic system, with a different symbol for each word in the vocabulary, has ever developed. Short-circuiting the sort of cumbersome system that would have

38

The

Writing.

following

remarks

on

been,

language writing

are

word-signs based

on

came to

be

Gelb, A Study of

used as syllable signs, and that constituted a first stage in the evolution of analytical systems of writing. The application of that principle

writing systems produced syllabaries, of which Elamite, West Semitic, Cypro-Minoan, and Japanese. The last stage, alphabetic writing, was essentially a matter of making explicit the vowels of the syllables, and separating them off

throughout Gelb

entire

counts four:

from the

consonants.

symbolic and iconic components of the representing language and music, the histories of the two media seem to run in opposite directions: iconic to symbolic in language writing, symbolic to iconic in music writing. But the result is nevertheless similar in respect to two essential points: (1) The two systems became increasingly analytical. (2) They developed into systems based, not on the association of a unique signal with each referent, but on a principle that allows a limited number of signs to represent an infinite number of referents, through combinations that With respect to the

systems for

the reader need the

never

have

seen

before, given only that

he has learned

principle. With

respect to music writing we

have evidence of

a

contemporary

advantage system. Around the year 1000 Odo of St. Maur wrote, “I taught certain actual boys and youths by means of this art [notation] so that some after three days, others after four days, and one after a single week of training in it, were able to learn several antiphons and in a short time to sing them without hesitation, not hearing them sung by anyone, but contenting themselves

awareness

of such

of the

with

simply

a

copy written

a

according

to

the rules.” 39

By contrast, the Byzantine notational system always remained a symbolic one: arbitrary signs for melodic intervals. And there are hints that as late as the seventeenth century had not become established in the East. In

a

genuine literate tradition

1647 a French Dominican of Jacques description singing in the Byzantine “The Greeks have music books, but they rarely look at them liturgy: while singing The Greeks seldom sing from a book at the pulpit, and even more rarely do they conduct or teach singing with written notation at hand.” 40

named

Goar

.

.

.

wrote a

.

39

Oliver Strunk, ed., Source

40

Euchoiogion,

Hucke, “Toward

Readings in

Music

History (New York, 1950),

In ordinem Sacri Ministerii notae, a New Historical View,” p. 448.

21

pp.

103-104.

(Paris, 1647), p. 30. Cited by

VII

It was a further service of Handschin’s article to sort out some of the evidence bearing on the accent theory of origins. In particular he provided an answer to the question, To what did Aurelian refer when he used the expressions “acutus accentus” and “circumflexio” in the nineteenth chapter of his Musica disciplina? I quote his conclusion: Es ergibt sich eine klare Präsumption für die Annahme, dass mit “acutus accentus” nicht das Zeichen für den Hochton, sondem der Podatus, und mit “circumflexio” nicht der Podatus oder die Clivis, sondem der Torculus gemeint ist. 41 .

.

.

Handschin inferred this from Aurelian’s text, and he supported it script, one of whose main is in the the musical text oldest source for Aurelian’s exemplars treatise. In that script a slanted up-stroke ( ) represents not a single with reference to the Paleofrankish

J

high

note,

as

the

theory

would

require,

but

a two-note

ascent; and

a

semicircle (J) represents, not a two-note descent (according to the theory, Circumflex becomes Clivis) but a three-note group, rising and

falling. This

was

very

damaging

to

the

of the unshakable

theory, of course. (It is a knowledge in our field that by subsequent writers who,

accent

tenacity of sign Handschin’s paper is dutifully cited nevertheless, continue to transmit the theory he attacked.) What Handschin missed, and Corbin caught, 42 is that in his use of the word “accent” Aurelian was not really referring to neumes at all, but to melodic figures. The neumes are there, and they allow us to read the melodic figures, but it is the latter to which Aurelian refers. In a way that is an even harder blow to the theory, but in a way it helps, too. The accent theory of origin is, if nothing else, too specific and too simple. It is hard to imagine what sort of evidence could confirm it. In fact it is hard to say what the theory really means. Paolo Ferretti gave a graphic interpretation in his text for the thirteenth volume of the Paléographie musicale, an imagined representation of the 43 evolution of the Punctum from the grave accent: J The picture shows

41 42 43

exactly why the theory

is too

specific

Handschin, “Eine alte Neumenschrift,” p. 71. Corbin, Die Neumen, p. 18. PM XIII, 65.

and

too

simple.

There is

not, and could not be, evidence for such an evolution. But the passage from Aurelian does show that, in the early attempts to conceptualize about

melody,

the concept of “accent”

could refer to both

an

came to

inflection and the

sign

mind. And

for it

(the

as

“accent”

latter in other

sources), the concept and phenomenon of accent evidently do swim in the murky waters from which we shall have to fish out the components of whatever understanding we are ever going to have about the very beginnings of music writing. The task for this section of my essay is

no more

than

to

reasoned identification of factors that may inhabit those

practices, performance practices, and conventions in the similar to those of

melody.

pursuing the accent theory (1) Handschin showed

Virga, That

language

notes as

waters—manuscript

domain that have functions

can

that the Paleofrankish a

a

historical considerations,

reach all the relevant factors still further. We

and that what looked like

is,

try for

Virga

script

used

or acute accent was a

by no

Pes.

of the oldest scripts does not differentiate signs for single Punctum and Virga; nor do most of the other iconic scripts.

one

symbolic scripts that differentiation is not consistent, as one would expect it to be were it truly the basis of the notational practice in the first instance. The oldest script with a reliably Even among the

far as I know, is that of Benevento VI. 34, after the beginning. This is to be contrasted with a notational picture like that of Figure 1 where most syllables are marked by a Virga-like form. The contrast casts doubt on

consistent differentiation, written

some

so

two centuries

,

the

that the Punctum-Virga distinction is inherent in the This sort of marking of syllables constitutes one of the system. that phenomena require systematic study. How common is the what is its geographical and temporal extent, in what practice,

proposition

contexts does it occur?

(2) Post-Classical

inflected language, i.e., it was Hence the talk about the accents marking rising and falling inflections of speech has no real referent. (3) The neume tables, in which we have the earliest evidence of a systematization of the writing system, are no older than the eleventh century. They are the first evidence of systematic thought about notation from a theoretical point of view. Music writing was invented not

spoken

with

a

Latin

pitch

was not an

accent.

by practitioners, and systematized after a long period of practice. The import of the accent theory of origin has to be that musicians faced with the practical task of writing down their melodies turned to an ancient theory (the theory of accents transmitted by the Roman

grammarians on the model of the Greek grammarians) that did not correspond to the practice of their time. That is not easy to believe. (4) Corbin writes: “Not only is it possible to recognize [a] transition [from

marks

accent

to

notational 44

it

signs]

in the sources, but indeed

It is not clear what her referents

were happens quite frequently.” but are the in texts numerous sources which here, presumably they are marked occasionally with signs that may have been intended as accent marks (Appendix, Figure 7 ). But uncertainty about whether a thing belongs to one category or another is not evidence that it marks the transition between them. Sources containing such signs require systematic study, with the same questions asked as were put about syllable markers. (5) Quite apart from the problematic claim of the theory about the pitch significance of the accent marks, there is also the claim that, by

the time of the

writing down of the first neume tables in the eleventh century, their names had been changed to “Virgula” or “Virga” and “Punctum.” Now “Punctum” is known since the fourth century as a punctuation mark, indeed the prototypical punctuation mark. “Virgula" is the name given to the mark of the acute accent by Martianus Capella in the fifth century. In the Middle Ages it, too, was the name of a punctuation mark (which survives in the modem French “virgule," for “comma”). Was there a connection between accentuation and

punctuation

that would have

suggested

this

exchange

of names?

And if so, did melody and its notation play into that connection? Medieval punctuation was essentially an aid to the reader of Latin, a way of helping him bring out the sense of a text as he read it aloud. Cassiodorus (sixth century) wrote that the “positurae” or “puncta” “are, as it were, paths of meaning (‘sensus’) and lanterns to words, as instructive to readers as the best commentaries.”45 It did its work by

marking comma,

off the

colon,

sense

and

units of the text, conceived

as

the

hierarchy

periodus.

just this hierarchy of sense units in language that melody to project. Here is a writer of ca. 1100 (known only as Johannes), drawing a parallel between what amounts to the syntax of Now it is

was

understood

language

and music:

in prose three kinds of distinctiones are recognized, which can also be called “pauses”—namely, the colon, that is, member; the comma or

Just as 44 45

Corbin, Die Neumen, p. 18. Quoted in M. B. Parkes, “Medieval Punctuation,

Medieval

Eloquence,

ed.

J. J. Murphy (Berkeley

and Los

or

Pause and Effect,” in

Angeles, 1978),

pp. 127-42.

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

incisio;

.

.

.

.

Likewise, when a chant makes a pause by dwelling on the fourth or fifth 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 note

period.

46

Then in the example that he provides, the author makes a point of it that the “distinctions” of the melody exactly match those of the text it declaims. The cadences that mark off the distinctions of the melody correspond to the pauses and lowering of the voice that set off the distinctions of the text. Notational signs and punctuation marks therefore play a similar role in guiding the singer/reader in bringing out the sense of a text. 47

Now I quote

again

from Parkes:

In the ninth century Hildemar [possibly a monk from Corbie], writing to Bishop Ursus of Benevento about the art of reading, emphasized the relationship between punctuation and accentuation. He said that prose is split up by three points, adding “Do not be amazed that I have placed a sign of acute accent in the middle of the sense, since, as I have learned from learned men, to these three points three accents are appropriate: the grave as far as the middle of the sensus of the whole sententia, the acute only in the middle of the sensus, and then the circumflex up to the full sensus." Hildemar also incorporated this letter into his commentary on the Rule of St. Benedict, where it forms part of the exposition of and regulates reading aloud to the community Chapter 38 which prescribes in the refectory. 48 .

.

.

This passage, therefore, associates accent signs with punctuation marks in the function of separating sense units rather than in indicating a raising or lowering of pitch (except indirectly, as articulation of phrases). And that suggests a different sort of connection between accent and melody, or accent and the notation of melody. Whereas the claimed connection between melody or its notation and pitch accent flies in the face of the known facts about the accentuation of Latin, the connection to accent conceived in a role similar to that of

46 Johannes, “De Musica,” trans. Warren Babb, in Hucbald, Guido, and John Music, ed. Claude Palisca (New Haven, 1978), p. 116. 47 Cf. Jonsson and Treitler, “Medieval Music and Language.” 48 Parkes, “Medieval Punctuation,” p. 128.

on

punctuation is consistent with the contemporary testimony. It makes plausible the otherwise puzzling remark by the anonymous author of the treatise Quid est cantus?, quoted by Michel Huglo: “De accentibus toni oritur nota quae dicitur neuma.” 49

The

systematic study of early sources with attention to the relationship of accent-like markings to the syntax of the texts on which they are placed (see, for example, Figure 7 in the Appendix), as well as a study of the relationship of medieval punctuation to the early history of music

are research tasks that will be essential in the 50 of the investigation origins of music writing. the time of the earliest neume During systems it was not only the Punctum ( J ) and Virgula (J) and their combination ( J ) that were used as both neumes and punctuation marks for pauses of various degrees. It was also the Trigon ( ) and the Quilisma ( J ), both used to mark a question. In the eleventh century a figure like the iconic Clivis ( J ) is found, either alone or in combination with a Punctum ( J ), to mark a pause of low or intermediate hierarchichal level. 51 Whether these signs and their functions were developed for one practice and borrowed for the other, and if so, what the direction of borrowing was, are questions we cannot yet answer. The main point for the moment is to establish that there is this common aspect to musical notation and punctuation. In fact this was recognized some time ago, and has been forgotten since. P. Bohn concluded that neumes and punctuation signs have the same origins and significance and are in fact the same signs. 52 Jean Baptiste Thibaut went further to give priority to the punctuation marks, from which he claimed the neumes were derived. 53 Two years later Elias Avery Lowe comment-

writing,

J

49

Huglo, “Les Noms des neumes,” p. 62. The idea of an overlap in function between punctuation and musical notation suggests a context for an observation that was communicated to me several years ago by Professor Helmut Hucke, and that I have since been able to confirm. Psalters that were written at St. Gall before the beginning of music writing there are provided with punctuation. Those that were written afterwards are not so provided. 51 Bernhard Bischoff, Paläographie des römischen Altertums und des abendländiscben Mittelalters (Berlin, 1979), pp. 214-19, provides a convenient survey and bibliography. An extensive survey of medieval punctuation is under way by Professor M. B. Parkes of Oxford University. I wish to thank Professor Parkes for his generosity in me access to a giving preliminary draft of his survey. 52 P. Bohn, “Das liturgische Rezitativ und dessen Bezeichnung in den liturgischen Biichem des Mittelalters,” Monatshefte für Musikgeschichte, XIX (1887), 29-36, 45-52, 50

61-68, 78-80. 53

Jean Baptiste Thibaut, Monuments de Petersburg, 1912), Chaps. 1 and 2.

la notation

ekphonétique

de

l'église

latine (St.

ed,

“It

seems

claim rested

improbable

to

dates

on erroneous

(5) Thibaut’s

title

me,”

and in any

assigned

brings up

the

case

to two

one

other

showed that the 54

key manuscripts. practice from which

Latin neumatic notation has been

thought at one time or another to be derived, ecphonetic notation. These are systems of lectionary signs that were added to Hebrew-Masoretic and Byzantine liturgical texts at points of syntactical division. They represented melodic formulas that served to articulate the

while the more or

beginnings and ends of text-clauses (senseunit), intervening stretches of text would have been recited

less in monotone.

The

principal recent proponent of a derivation of Latin neumatic writing from ecphonetic notation is Eric Werner. 55 Werner’s reasoning on this subject is a mix of appropriate observations and questions, baseless conjecture, and argument from premises that take their conclusions for granted. He asks, “1) To what extent were these notations genuinely intended to carry musical phrases? 2) Wherein lies the distinction between general grammatic-phonetic accents, punctuation, and real signs of musical cadences? 3) How did these accents that served the purpose of music, grammar, mnemonics, and punctuation simultaneously, originate in the first place?” He observes that “The first and most important object” of the ecphonetic signs was

project “the exact syntactic punctuations of the text.” At the same time he makes assertions of such vast scope that it is hard to know either what they mean or how they could be substantiated. “Since the to

neumes as

the

of the Roman Church

Byzantine signs,

Greek,

and

it is

early Roman

stem

logical

ultimately

from the

compare the neumes.”

to

accents or

same source

Hebrew, Medieval

Werner claims that ecphonesis passed from the Hebrew, to the Byzantine, and thence to the Western tradition. Now it must first be stated plainly that there is no documentation of an ecphonetic notational practice as such in the West. Characteristics of Latin chant that could be taken as residuals of such a practice are the more stereotypically formulaic structures at the beginnings and ends of phrases, and the coordination of formulaic melodic cadences with fixed clausula types in the texts. But it does not require the positing of an ecphonetic notational practice to account for those characteristics.

They are

normal features of melodies that

are

the

far

emerged,

an

oral

seems

that

products

of

tradition.

Judging 54 55

from the evidence that has

so

it

Elias Avery Lowe, then Loew, The Beneventan Script (Oxford, 1914), p. 237, Eric Werner, The Sacred Bridge (New York, 1959), Chaps. 3 and 4.

n.

r.

bringing the ecphonetic practice into the picture gains us nothing. The general idea behind such a practice—a role for melody and hence for musical

in

signs

bringing

out

the

syntactical

structure

of laguage—is

something we have already recognized as a factor in the early history of Latin neumatic writing, without having to make claims about The

an

ecphonetic

notation.

focused, but than what has been handed down so far. true-to-life, perhaps we shall a never have detailed and accurate Probably portrayal of the of not a as those that as one “beginning notation”—especially simple have been proposed—any more than we shall have such a portrayal of “the beginning of music” or “the beginning of life.” But given the

picture

that remains after this review is less

more

systematic and comparative study of documented practices and surviving sources, we have a good chance of learning much more about the factors involved, functional aspects. The State

University

especially

of New York

at

if

Stony

we

concentrate on

Brook

Appendix

Figure

1.

Oxford, Bodleian Library, MS 775, fol. 127v

the

Figure

2.

Benevento, Archivio

capitolare,

MS

VI.34,

fol. 27v

Figure

3. Vercelli, Biblioteca

capitolare,

MS 162, fols. 10v—11r

Figure

4. Padua, Biblioteca

capitolare,

MS

A47,

fol. 3V

Figure

5. London, British

Library, MS Lansdowne 462, fol. 98v

Figure

6.

Vercelli,

Biblioteca

capitolare,

MS 161, fols.

58v-59r

Figure

7.

Paris, Bibliothèque

Ste.

Geneviève,

MS 557, fol. 18v

]This essay appears in a revised form and with the addition of an introduction in Leo Treitler (2003/2007), With Voice and Pen: Coming to Know Medieval Song and How it was Made, Oxford and New York: Oxford

University Press.

De Accentibus Toni Oritur Nota [2]

Quae Dicitur Neuma: Prosodic Accents, the Accent

Theory, and the Paleofrankish Script

CHARLES M. ATKINSON

An anonymous Frankish author

probably writing begins his treatment of chant as follows:

in the tenth century

What is chant? It is skill in the musical art, the inflection and modulation of the voice.

Why is

musical art

or

‘chanting,’ that is, from skill in the origins and its structure are from the feet of syllables. Indeed, it is

it called ‘chant’? From

in the modulation of the voice. Its

revealed in the accents of tone

or

founded in acute, grave, and circumflex accents of tone.

The author then makes the

statement

quoted in the

title of this paper:

De accentibus toni oritur nota quae dicitur neuma (From the accents of tone 1 sign which is called neume).

arises the notational

This passage is drawn from the treatise Quid est cantus, sometimes referred to as the Vatican Anonymous, first edited by Peter Wagner in the peritia musicae artis, inflexio vocis et modulatio. Quare dicitur cantus? peritia musicae artis vel vocis modulatione. Ortus quoque suus atque compositio ex accentibus toni vel ex pedibus syllabarum ostenditur. Ex accentibus vero toni demonstratur in acuto et gravi et circumflexo. Ex pedibus denique syllabarum ostenditur in brevi et longa. De accentibus toni oritur nota [a later hand has added above this ‘figura’] quae dicitur neuma.” Biblioteca Vaticana, Cod. lat. Palat. 235, fol. 38v, ed. Peter Wagner, “Un piccolo trattato sul canto ecclesiastico in un manoscritto del secolo X-XI,” Rassegna gregoriana 3 (1904): 481-84. A fascimile of the passage in question may be found in Enrico Marriott 1

a

“Quid est

cantus?

canendo, idest,

a

DOI: 10.4324/9781315090283-3

Oral and Written Transmission in Chant

third volume of one

of the

only

Rassegna gregoriana medieval

I

know of

temporal proximity to the

quote it here because it is concerning the origins of

reporting and given the clarity of its statement about the basis of the neumes, one is surprised to find that if mentioned at all, it has been treated only in passing in several recent articles on the beginnings of music-writing in the West. Moreover, given the Vatican Anonymous’s unequivocal statement that the neumes were derived from prosodic accents, it is a bit surprising to find a large portion of one of those articles devoted to a demonstration that “Carolingian notation could not possibly have been derived from the accents.” neumes.

Given its

in 1904.

statements I

events it is

2

3

The apparent conflict between what medieval and modern scholars have to say about accent and its relation to the origins of neumes is the subject

had

of the present paper. It will not be my purpose here to attempt a new “Universale Neumenkunde,” or to outline a new theory of the origins of neumes, or even to resurrect an

old

one.

Instead, I should like

to

examine

the concept of accent at its source, in the grammatical treatises of antiquity and within the educational milieu of the Carolingian court, so that we may better understand what the

theory

of such

we mean

accents as

when

they

we

talk about

prosodic accents and practical musical

relate to the advent of

notation in the West. 4 Our first task in this endeavor

must

be

to

ascertain whether and how

Carolingian musicians might have acquired a knowledge of prosodic accents. To approach this question we must examine at least briefly the educational system of the Frankish court. Upon his ascent to the Frankish throne in 771, Charlemagne became the ruler of an empire that was increasing in size and importance, and would become

even

larger and

of administration

were

more influential during his reign. New techniques required, and indeed were forthcoming. One of the

Bannister, Monumenti Vaticani di Paleografia musicale latina (Leipzig: Harrassowitz, 1913), 1. It is given in parallel Latin and German in Peter Wagner, Einführung in die gregorianischen Melodien: Ein Handbuch der Choralwissenschaft 2: Neumenkunde (Leipzig: Breitkopf & Härtel, 1911-12), 355-56. 2 See, for example, Leo Treitler, “The Early History of Music Writing in the West,” Journal of the American Musicological Society 35 (1982): 237-79, and “Reading and Singing: On the Genesis of Occidental Music-Writing,” Early Music History 4 (1984): 135-208; Helmut Hucke, “Die Anfänge der abendländischen Notenschrift,” in Festschrift Rudolf Elvers zum 60. Geburtstag, ed. Ernst Hettrich and Hans Schneider (Tutzing: Hans Schneider, 1985), 271-88, and “Gregorianische Fragen,” Die Musikforschung 41 (1988): 304-30; and Kenneth Levy, “On the Origin of Neumes,” Early Music History 7 (1987): 59-90. no.

3

Treitler, “Reading and Singing,” 184.

4

Although he

does not treat this topic expressis verbis, Mathias Bielitz offers much information useful for its pursuit in his Musik und Grammatik: Studien zur mittelalterlichen Musiktheorie, Beiträge zur Musikforschung 4 (Munich-Salzburg: Katzbichler, 1977). See in particular 114-20.

Oral and Written Transmission in Chant

administrative

techniques

that

Charlemagne adopted

is of

special

interest

here. Under his rule the administration of both secular and ecclesiastical affairs became

increasingly dependent upon written documents, especially in the period after 800. As Ganshof has pointed out, “Whatever the field, Charlemagne attached great importance to setting things down in writing.” His administrators obviously needed to know how to read and write—and they needed to know how to do those things well. One Carolingian author reports that Charles “devoted just as much ardor to abolishing imperfections of language as he devoted to vanquishing his enemies on the field 5

6

of battle.”7 As Ganshof and others have shown, the ability to read and write was a was not in great supply in the Frankish kingdom in the mid- to

skill that

8 century. Thus, Charlemagne had to create an educational system that could provide trained administrators in both sacred and secular

late

eighth

realms. In order men

of the

do that he

brought to his court some of the most gifted day—Paulinus of Aquileia, Peter of Pisa, and Theodulf of to

Orléans among them—and convinced Alcuin of York to take over the direction of the palace school and to assist him in implementing a program of educational reform. 9

Unfortunately, we have little direct information about Alcuin’s organization of the palace school, but we can reconstruct its outlines from various

5

François Louis Ganshof, “The Use of the Written Word in Charlemagne’s Administration,” Carolingians and the Frankish Monarchy, trans. Janet Sondheimer (Ithaca, N.Y.: Cornell

in The

University Press, 1971), 125-42. Ganshof remarks that from about the middle of the seventh century until the time of Pippin III “the only use for written records seems to have been to furnish proof of individual rights, or to assist in such proof.” He continues: “So far as we can judge from the sources, the

use

of the written word for administrative purposes started to revive under

When we come to the reign of Pippin III, though only to a very modest extent. Charlemagne we find a change in the situation. The number of sources to enlighten us about administrative records becomes more plentiful; and although more numerous for the period after the imperial coronation, they are spread over the entire reign. This abundance of documentation is novel and revealing” (125-26). .

6

.

.

Ganshof, “The Use of the Written Word,” 126. MGH, Poetae latini aevi carolini I, 89-90. Cited after Pierre Riché, Écoles et enseignement dans le Haut Moyen Age: Fin du Vesiècle—milieu du xie siècle (Paris: Picard, 1989), 112. 8 Ganshof, “The Use of the Written Word,” 135. See also Max Ludwig Wolfram Laistner, Thought and Letters in Western Europe, A.D. 500 to 900 (Ithaca, N.Y.: Cornell University Press, 1931/1966), 191-92, and Joseph Dyer, “The Monastic Origins of Western Music Theory,” in Laszlo Dobszay et al., eds., International Musicological Society Study Group “Cantus Planus": Papers Read at the Third Meeting, Tihany, Hungary, 19-24 September 1988 (Budapest: Hungarian Academy of Sciences, Institute for Musicology, 1990), 199-225, particularly 210-11. 9 On this see Franz Brunhölzl, “Der Bildungsauftrag der Hofschule,” in Karl der Groβe: Lebenswerk und Nachleben 2: Das geistige Leben, ed. Bernhard Bischoff (Düsseldorf: Schwann, 1965), 28-41, and Laistner, Thought and Letters, 192-99. 7

sources,

school

such

at

as

the program he later established at St. Martin of Tours. The one division for Bible study, a second one for the liberal

Tours had

arts

(among which

and

a

discipline of astronomy received special treatment), specifically to grammar. From this outline and from Alcuin’s Disputatio de vera philosophia, the preface to his treatise on grammar, one can deduce that the ultimate and highest goal of instruction was the study of the Bible, but one would progress to that through study of the liberal arts. The first of these—in both a literal and a figurative sense—was grammar. Beyond the most fundamental types of elementary instruction, such as the learning of reading and writing and the memorization of certain liturgical texts, the discipline of grammar was the one liberal art whose study we can reasonably assume to have been expected of everyone. As the testimony of Alcuin’s own treatise on grammar, along with ninth-century library catalogues as well as editions and commentaries by Paul the Deacon, Peter of Pisa, Smaragdous, Walafrid Strabo, Murethach, Sedulius Scotus, and Remigius of Auxerre tell us, the chief authority for grammar was the

third devoted

10

11

12

13

10

Brunhölzl, “Der Bildungsauftrag,"30. On this, see Alc. epist. 121 in MGH, Epist. IV 176, 32ff. In another instance an unknown bishop tells Alcuin that he should oversee instruction, and names grammar, reading, and study of the Bible as subjects (Alc. epist. 161; MGH, Epist. IV 260, 13ff.). In Brunhölzl’s view, such witnesses tell us that Alcuin’s poem 26 also depicts the court school itself. 11 Brunhölzl, “Der Bildungsauftrag,” 33-34. For the text of the Disputatio, see Migne PL 101, 849-54. In the Disputatio, Alcuin is asked by his students to explain the stages leading up to philosophy. Alcuin responds by quoting the Bible: “Wisdom hath builded her house, she hath hewn out her seven pillars” (Prov. 9:1). The seven pillars are to be understood not just as the seven Gifts of the Holy Spirit, but also as the seven liberal arts. Cf. PL 101, 853-54. On Alcuin as a teacher, see Eleanor Shipley Duckett, Alcuin, Friend of Charlemagne: His World and His Work (Hamden, Conn.: Archon Books, 1965), 109-17. 12 The texts in question are the Pater noster, the Credo, and the Psalms. Cf. Bernhard Bischoff, “Elementarunterricht und Probationes Pennae in der ersten Hälfte des Mittelalters,” in Classical and Medieval Studies in Honor of Edward Kennard Rand, ed. Leslie Webber Jones (New York: the editor, 1938), 9-20, the expanded form of this article in Bischoff, Mittelalterliche Studien: Ausgewählte Aufsätze zur Schriftkunde und Literaturgeschichte 1 (Stuttgart: Hiersemann, 1966), 74-87, and Brunhölzl, “Der Bildungsauftrag,” 39. 13 Cf. Riché, Écoles et enseignement, 111-12, and Brunhölzl, Geschichte der lateinischen Literatur des Mittelalters 1: Von Cassiodor bis zum Ausklang der karolingischen Erneuerung (Munich: Fink, 1975), 246-47. Brunhölzl characterizes grammar as the “Grundlage aller Wissenschaften,” and goes on to say (247): “Es ist dabei von sekundärer Bedeutung, daβ man in der Praxis von den artes liberales zunächst fast nur die Fächer des Triviums—Grammatik, Dialektik, Rhetorik—pflegte, unter denen wiederum die Grammatik den weitaus gröβten Raum einnahm, während die Einbeziehung der rechnenden Disziplinen des Quadriviums—Arithmetik, Geometrie, Musik und Astronomie—in nennenswertem Umfang erst etwa seit der Mitte des 9. Jahrhunderts und auch da allmählich und an den verschiedenen Schulen in sehr verschiedenem Maβe erfolgte.”

Donatus. 14 The

saying that “every schoolboy had

his Donatus”

was no

less

in the Carolingian period than it was in late antiquity. In our quest to learn what the Carolingians would have learned about

true

prosodic

accents,

with the sections that

were

our attention to Donatus, along in two of the three manuals of the liberal arts grammar

we on

must

therefore turn

the “standardized textbooks” for the

Isidore of Seville and Martianus of

course,

Cassiodorus, but the

Carolingians—namely, Capella. (The third of those textbooks is, chapter on grammar in his Institutiones is 15

16

simply a two-page précis of Donatus’s Ars maior.) Excerpt 1 presents the opening sentences of each of the Donatus’s Ars maior, Book I, drawn from the Louis Holtz. 17

new

six

chapters of by

critical edition

Excerpt 1

from Donatus, Ars maior, Book I 1. De Voce

ictus, sensibilis auditu, quantum in ipso

est. Omnis vox aut

articulata est aut confusa. Articulata est, quae litteris confusa, quae scribi non potest.

conprehendi potest;

Vox est aer

2. De Littera

Littera est pars minima uocis articulatae. Litterarum aliae sunt uocales, aliae semiuocales, aliae mutae. Vocales

consonantes. Consonantium aliae sunt

sunt, quae per se proferuntur et per i o u.

se

syllabam faciunt. Sunt autem numero

quinque a e

14 Cf. Alcuin, Grammatica, PL 101, cols. 854-902. Donatus is cited early in the treatise .”) and several times subsequently. Cf. Brunhölzl, “Der (855C: “Ut reor, in Donato legimus Bildungsauftrag,” 39. For listings of Donatus in ninth-century library catalogues, see Paul Lehmann, Mittelalterliche Bibliothekskataloge Deutschlands und der Schweiz 1 (Munich: Beck, 1918), in particular the catalogues from St. Gall and Reichenau. For a listing of the most important .

extant

manuscripts,

as

well

.

as a

discussion of medieval editions of and commentaries

upon Donatus, see Louis Holtz, Donat et la tradition del'enseignement grammatical: Étude sur l’Ars Donati et sa diffusion (IVe–IXe siècle) et édition critique, Documents, études et répertoires

publiés par l’Institut de recherche et d’histoire des textes (Paris: Centre national de la recherche scientifique, 1981), 354-421 (manuscripts) and 438-41 (Carolingian editions and commentaries). 15

On this

point as it pertains to instruction in music, see Michel Huglo, “Le développement carolingienne,” Latomus 34 (1975): 131-51. See Cassiodori Senatoris Institutions, ed. R. A. B. Mynors (Oxford: Clarendon, 1937), 94-96.

du vocabulaire de l’Ars musica à l’époque 16

17

Holtz, Donat et la tradition, 603-12.

3. De Syllaba Syllaba est conprehensio litterarum uel unius uocalis enuntiatio temporum capax. Syllabarum aliae sunt breues, aliae longae, aliae communes. 4. De Pedibus Pes est syllabarum et temporum certa dinumeratio. Accidunt uni cuique pedi arsis et thesis, numerus syllabarum, tempus, resolutio, figura, metrum. Pedes disyllabi sunt quattuor, trisyllabi octo, duplices sedecim. 5. De Tonis

Tonos alii accentus, alii tenores nominant. Toni

igitur

tres sunt, acutus,

grauis, circumflexus. 6. De Posituris Tres sunt [omnino]

positurae

uel distinctiones quas θέσεΙς

Graeci uocant,

distinctio, subdistinctio, media distinctio.

Following the opening chapter on articulate versus inarticulate speech—with inarticulate being defined as that which cannot be written down using letters—the five remaining chapters provide a syntactic description of Latin, proceeding from the smallest element, the letter, up to the sentence or period. (Larger divisions, such as paragraphs or sections of a speech, would be covered in rhetoric and logic.) Two chapters, De pedibus and De tonis, are devoted to the treatment of whole words, while the final chapter 18

treats

the

sentence

But the

and its subdivisions of colon and

chapters

in

tonis and De posituris.

which

Excerpt

we are

most

comma.

interested

are

the final two, De

2 presents sections of these two

chapters.

19

Excerpt 2

from Donatus, Ars maior, Book I, chapters 5 and 6 5. De Tonis

[A] Tonos alii accentus, alii

tenores nominant. Toni

grauis, circumflexus. Acutus

igitur tres

sunt, acutus,

in Graecis dictionibus tria loca teneat, et antepaenultimum, apud Latinos paenultimum et cum

ultimum, paenultimum antepaenultimum tenet, ultimum numquam. Circumflexus autem, quotlibet syllabarum sit dictio, non tenebit nisi paenultimum locum. Grauis poni in eadem dictione uel commune cum

18

cum

acuto uel cum circumflexo

potest,

et hoc illi non est

ceteris.

Holtz, Donat et la tradition, 58-62, points out that the organization of this book of the Ars on principles set forth in Melampous, a commentator on the grammar of Dionysios of Thrace (see edition cited in n. 20 below), and is quite close to theτὲχνη περὶ φωνῆς of Diogenes of Babylonia. 19 Holtz, Donat et la tradition. maior is based

[B] Acutus

grauis

accentus est nota per

obliquum ascendens

nota a summo in dexteram

in dexteram partem /, � 5C;, circumflexus nota

partem descendens

de acuto et

graui facta ^, longus linea a sinistra in dexteram partem aequaliter uirgula similiter iacens, sed panda atque contractior̆, hyfen uirgula subiecta uersui: hac nota subter posita duo uerba, cum ita res exigit, copulamus, ut ante̮tulit gressum et Turnus ut ante̮uolans. Huic contraria est diastole, dextera pars [quaedam] circuli ad imam litteram adposita: hac nota male cohaerentia discernuntur, ut est ereptae, uirginis ira et uiridique in litore conspicitur, sus. Apostrophos circuli item pars dextera, sed ad summam litteram adposita: hac nota deesse ostendimus parti ducta —, breuis

orationis ultimam uocalem, cuius crimine

dignum

consonans

Duxisti? Ceterum dasian

littera uocali addita uel detracta

⊦

remanet, ut est tanton'

et

psilen

⫞

apud

me

Latinos H

significat.

6. De Positvris Graeci [C] Tres sunt [omnino] positurae uel distinctiones quas θέσεΙς uocant, distinctio, subdistinctio, media distinctio. Distinctio est, ubi finitur sententia:

plena

huius

Subdistinctio est, ubi

punctum

ad

summam

litteram

ponimus.

multum superest de sententia, quod tamen inferendum sit; huius punctum ad imam litteram

non

necessario separatum mox ponimus. Media distinctio est, ubi fere tantum de sententia superest, quantum iam diximus, cum tamen respirandum sit: huius punctum ad mediam litteram ponimus. In lectione tota sententia periodos dicitur, cuius partes sunt cola et commata.

A close connection with music is established

chapter

by the

5 : “Tonos alii accentus, alii tenores nominant”

others tenors”).

ultimately

it echoes

very first sentence of

(“Some call

tones accents,

Although passage in Quintilian, this definition goes back to the Greek grammarians, for whom tonos a

(from the verb τείνω) is the standard term for vocal inflection and the

important element ofπροσῳδία.

20

most

Greek may be seen in the fact that two of its terms, tonus and tenor, have Greek cognates, and that the term accentus is a translation of the Greek προσῳδία. 21 After

listing

Its direct

lineage from the

the three types of accents—acute, grave, and circumflex— stating, for example, that the

Donatus contrasts Greek usage with Latin,

20 See Marcus Fabius Quintilianus, Institutio oratoria, ed. L. Radermacher (Leipzig: Teubner, 1907), I.v.22. For examples of the Greek usage, see Dionysii Thracis Ars grammatica et Scholia in Dionysii Thracis Artem grammaticam, ed. Gustav Uhlig, Grammatici Graeci, 1/1 and 3

(Leipzig: Teubner, 1883/1901). 21

On this

see

the excerpts from Isidore and Martianus below, as well

as

the article “Tenor”

by Dagmar Hoffmann-Axthelm in Handwörterbuch der musikalischen Terminologie.

22 any of three positions within a word in Greek, but in in Latin. He also mentions that the grave accent can be

acute can appear in

only two positions placed in the same word with either an acute or circumflex, but that this trait does not pertain to the other two—that is, an acute and a circumflex may not appear together in the same word. Following a set of examples demonstrating the placement of accents in words of varying numbers of syllables and quantities of vowel sounds, Donatus finally tells us what the accents actually are (see section B of excerpt):2 “An acute accent is a nota [a graphic sign] ascending toward the right in oblique motion; a grave accent is a nota descending toward the right from above; a circumflex accent is a nota made from acute and grave.” He goes on to describe the notae used for long and short syllables, the hyphen, diastole, and apostrophe, and closes the chapter with examples of the dasia and the psile, signifying the addition or absence of the letter H in Latin. (Musicologists will recognize the dasia as the stem for most of the notational signs in the Musica enchiriadis.) The final chapter of Ars maior, Book I (section C in excerpt 2 ) discusses the positurae or punctuation signs: a high point placed at the end of a sentence or distinction; a low-level point placed at the end of a subdistinction or comma, where not much of the sentence is left; and a mid-level point for a medial distinction or colon, where enough of the sentence is left so as to require a new breath. To provide examples of these distinctions, Murethach of Auxerre, an early ninth-century commentator on Donatus, used the text 23

for the Introit Iubilate Deo omnis

terra:

“‘Iubilate Deo omnis terra,’ here is

medial distinction; ‘psalmum dicite nomini eius,’ here is ‘date gloriam laudi eius,’ here a distinction.” 24

22

a

a

subdistinction;

Holtz, Donat et la tradition, 62, states that Donatus’s word dictio is his equivalent ofλέξΙς

in

Diogenes, meaning.

and is used to

signify

a

word

as a

sounding

and written

entity,

apart from its

23

Cf. Musica et Scolica enchiriadis, una cum aliquibus tractatulis adiunctis, ed. Hans Schmid, Bayerische Akademie der Wissenschaften: Veröffentlichungen der Musikhistorischen Kommission 3 (Munich: Verlag der Bayerischen Akademie der Wissenschaften, 1981). The treatise is also available in Martin Gerbert, Scriptores ecclesiastici de musica sacra potissimum 1 (St. Blasien: Typis San-Blasianis, 1784; repr. Milan: Bollettino bibliografico musicale, 1931), 152-73. 24

The passage with its lemma is as follows: “HUIUS PUNCTUM AD MEDIAM LITTERAM PONIMUS recte ad mediam litteram, quia medius sensus est ibi. Cuius rei exemplum est

(372, 21). Et

'Iubilate Deo omnis terra,’ ecce media distinctio; ‘psalmum dicite nomini eius,’ ecce subdistinctio; ‘date gloriam laudi eius,’ ecce distinctio.” Murethach (Muridac), In Donati Artem maiorem, ed. Louis Holtz, Corpus Christianorum: Continuatio mediaevalis 40 (Turnhout: Brepols, 1977), 45. According to Holtz, Donat et la tradition, 439, Murethach taught at Auxerre before 840. “Iubilate Deo omnis terra” is the text of the Introit for the third Sunday after Easter.

In

keeping with the progressive, systematic

character of the six

chapters

of Book I of the Ars maior, Donatus’s treatment of the accents and positurae in the last two of these chapters provides a complete set of graphic signs to convey the inflection and articulation of a text, from individual syllables to phrases, to the totality of a complete sentence. With regard to the accent signs themselves, I believe it is significant that Donatus does not define them in terms of their pitch inflections, but rather in terms of their appearance as graphic figures. His use of the term nota for these signs is completely in keeping with the use of that word in late antiquity as a notational sign for various linguistic purposes and even for music: Boethius uses the same term when he presents the Alypian notation in his De musica. Even though Donatus himself does not describe the accents in terms of their pitch inflections, ninth-century commentators such as Murethach and Sedulius Scotus do. But they derive their own descriptions from two intervening sources—Isidore of Seville and Martianus Capella. As Louis Holtz has pointed out, the sections on grammar in Isidore’s Etymologiarum incorporate and expand upon Donatus’s work. This is readily apparent from Isidore’s treatment of accentus and positura, which appears in slightly abbreviated form in excerpt 3. 25

26

27

Excerpt 3 from Isidore of Seville, Etymologiarum, I [A] De Accentibus. Accentus, qui Graece prosodia dicitur [ex Graeco nomen Latine “ad,” ᾠδή Graece, Latine “cantus” est. Hoc accepit]. Nam Graece πρóς nomen de verbo ad verbum expressum est. Latini autem habent et alia nomina. Nam accentus et tonos et tenores dicunt, quia ibi sonus crescit et

enim

quod iuxta cantum sit, sicut adverbium quia quod acuat et erigat syllabam, gravis, quod deprimat et deponat. Est enim contrarius acuto. Circumflexus, quia de acuto et gravi constat. Incipiens enim ab acuto in gravem desinit, desinit. Accentus autem dictus,

iuxta verbum est. Acutus accentus dictus,

atque ita dum ascendit et descendit, circumflexus efficitur. [B] De Figvris Accentvvm. Figurae accentuum decem sunt, quae a grammaticis pro verborum distinctionibus adponuntur. ’Oξει̂α, id est acutus accentus, linea a sinistra parte in dexteram partem sursum ducta, fit

25

Boethius, De institutione musica, IV, 3. Holtz, Donat et la tradition, 250 and 432. 27 Isidore of Seville, Etymologiarum, I, xviii-xx, excerpts, from Isidori Hispalensis Episcopi Etymologiarum sive Originum Libri XX, ed. W. M. Lindsay, Scriptorum Classicorum Bibliotheca Oxoniensis (Oxford: Clarendon, 1911). 26

56

Oral and Written Transmission in Chant

26

Charles M. Atkinson

ita: / . Ba.peia, id est gravis, linea a summo sinistrae in dexteram deposita, fit ita: \ . Tieptcr1twµEVTJ , id est circumflexus, '!inea de acuto et gravi facta, exprimitur ita: ". Ma.Kp63

Ratisbon

11

Benevento

11-

Benevento

12

Besangon

11

Brescia Montecassino

11

Brittany

10

Cluny

11

12

Near Auxerre

12

Saint-Denis

11

Dijon

11

?Noyon Near Dusseldorf St. Gall

12-

St. Gall

11

St. Gall Monte Amiate N. Italy PPavia

11

Klosterneuburg

12

10

'3 11

Near Laon ? Bologna

Limoges

Monza Monza Chelles PEinsiedeln Narbonne Nevers Paris Poitiers Arne (Champagne) Rheinau

11-12 11 11 12 11

■3 14 12

11

Salzburg Seckau Sens

!3

F-CA 75

Vaa i Vec i Vec 2 Vec 3 Yrx

Arras Vercelli Vercelli Vercelli St. Yrieix

I-VCd 161 I-VCd 146 IVCd 162 F-Pn lat. 003 (Paleographic Mttsicak

11 11 12 12 11

13)

B. Modern Editions and Secondary Literature Chant Bloomington 1958 Apel Willi Gregorian “ Balducci Sanzio L’interpretazione dei gruppi strofici .” Études grégoriennes 18 ( 1979 ): 5 97 Cain Elizabeth English Chant Tradition in the Late Middle Ages: the Introits and Graduals of the Temporale in the Sarum Gradual .” Ph.D. diss., Harvard University 1982 Cardine Eugène Sémiologie grégorienne .” Études grégoriennes 11 ( 1970): .

,

.

.

,

.

,

-

.

“

.

,

.

,

“

.

,

1

-

158

.

Cattin Giulio Music

of the Middle Ages I trans. Steven Botterill Cambridge (originally published as Il Medioevo I, Turin, 1979). .

,

1984

.

,

,

centons des Alleluia anciens .” Études grégoriennes unnumbered pages]. Fassler Margot The Office of the Cantor in Early Western Monastic Rules and Customaries: a Preliminary Investigation .” Early Music History 5 ( 1985 ): 29 51 Froger Jacques Les prétendus quarts de ton dans le chant grégorien et les symboles du MS H. 159 de Montpellier .” Études grégoriennes 17 ( 1978 ): 145 79 Review of Grammar of Gregorian Tonality by Finn Egeland Hansen Études 19 (1980 ): 223 25 Gmelch Joseph Die Vierteltonstufen im Messtonale von Montpellier Veröffentlichungen der gregorianischen Akademie zu Freiburg, no. 6. Eichstatt 1911 Graduale 1957 Le Graduel Romain, Édition critique par les moines de Solesmes Vol. 2 Les sources Solesmes 1957 Graduale 1960 Le Graduel Romain, Édition critique par les moines de Solesmes Vol. 4 Le texte neumatique, part 1, Le groupement des sources Solesmes

Claire

20

Jean

,

“

.

Les formules

( 1981 ): 3 4 -

“

[plus

12

.

,

-

.

“

.

,

-

.

—.

.

,

grégoriennes

-

.

.

,

.

.

,

.

.

.

,

.

,

.

.

.

,

1960 Hansen

,

.

Finn Egeland H159: Tonary of Copenhagen 1974 .

,

St.

Benigne of Dijon

.

2

vols.

.

,

“

Hucke Helmut Towards a New Historical View of Gregorian Chant .” This JOURNAL 33 ( 1980 ): 437 67 Hughes David G. Variants in Antiphon Families: Notation and Tradition." In International Musicological Society, La musique et le rite sacré et profane. Actes du XIIIe Congrès de la Société Internationale de Musicologie, Strasbourg 29 août—3 septembre 1982 ed. Marc Honegger and Paul Prévost 2 : 29 47 Strasbourg 1986 “ Levy Kenneth A Gregorian Processional Antiphon .” SchweizerJahrbuchfür Musikwissenschaft 2 ( 1982 ): 91 102 Charlemagne’s Archetype of Gregorian Chant .” This JOURNAL 40 ( 1987 ): 1 30 .

,

-

.

"

,

,

-

,

.

,

.

,

-

.

“

—.

-

.

.

André ed. Antiphonarium Tonale Missarum XIe siècle: Codex H. 159 de la Bibliothèque de L'École de Médecine de Montpellier. Paléographie Musicale nos. 7 and 8. Solesmes 1901 5 Palisca Claude ed. Hucbald, Guido, and John on Musk trans. Warren Babb New Haven 1978 Schlager Karlheinz ed. Alleluia-Melodien I. Monumenta monodica medii aevi no. 7 Kassel 1968 Smits van Waesberghe Joseph ed. Johannis Afflighemensis de musica cum tonario Corpus Scriptorum de Musica, no. 1 [Rome], 1950 ed. Guidonis Aretini Micrologus. Corpus Scriptorum de Musica no. 4 [ Rome], 1955 Thompson Claude La traduction mélodique du trigon dans les pièces authentiques du Graduale Romanum .” Études grégoriennes 10 ( 1969 ): 29 65 Treitler Leo Homer and Gregory: The Transmission of Epic Poetry and Chant .” Musical Quarterly 60 ( 1974): 333 72 ‘Centonate’ Chant: Übles Flickwerk or E Pluribus Onus ?” This JOURNAL 28 ( 1975 ): 1 23 Transmission and the Study of Music History .” In International Musicological Society, Report of the Twelfth Congress, Berkeley 1977 ed. Daniel Heartz and Bonnie Wade 202 11 Kassel 1981 Observations on the Transmission of some Aquitanian Tropes .” In Aktuelle Fragen der musikbezogenen Mittelalterforschung. Forum Musicologicum 3, ed. Wulf Arlt and Hans Oesch 11 60 Winterthur 1982 Reading and Singing: on the Genesis of Occidental Music-Writing Early Music History 4 ( 1984): 135 208 Van der Werf Hendrik The Emergence of Gregorian Chant Vol. 1 part 2. Rochester 1983 Wagner Peter Einführung in die Gregorianischen Melodien Vol. 1 Ursprung und Entwicklung der liturgischen Gesangsformen 3rd ed., Leipzig 1911 Vol. 2 Neumenkunde, 2nd ed., Leipzig, 1912

Mocquereau

,

,

-

,

.

,

,

,

,

,

.

,

.

,

,

,

.

.

.

—

.

,

.

,

.

,

,

.

“

.

,

-

.

“

.

,

-

.

“

—.

-

.

“

—.

,

-

.

,

.

,

“

—.

-

,

.

.

,

"

."

—.

-

.

.

.

,

.

.

,

,

.

,

.

,

,

,

.

.

,

ABSTRACT The numerous musical variants in the manuscripts of classical Gregorian chant are mostly trivial, having little effect on the melodies. They are more readily understood as mishearings than as misreadings. The few substantive variants are always regional. Many variants involving the replacement of E or B by F or C, or using special neumes such as the salicus or trigon suggest that at one time microtones were used in the chant—a hypothesis supported by some theoretical evidence. The variants taken together show that the chant was fully fixed with respect to pitch before its dissemination throughout the Carolingian Empire and beyond, and hence that any period of improvisational or recreative performance must have occurred prior to that time.

[9] Charlemagne’s

Archetype Gregorian Chant By KENNETH LEVY For Michel Huglo

WAS

on

his

of

65th birthday.

“GREGORIAN CHANT” FIRST WRITTEN DOWN? When were

the propers of the Roman Mass and Office, which WHEN in

an

unbroken line from the later

Carolingians

to

their definitive neumed forms? If current wisdom

we can trace

Solesmes, given is believed, the

neumatic notation was not devised until the first half of the ninth century; it arose in the service of novel and ancillary chants like tropes, sequences, genealogies, celebrant’s and diaconal ekphoneses,

theorist’s illustrations, and polyphony. The central repertory of Gregorian proper chants would remain consigned to professional memories and improvisational maneuver during the early generations of the neumes’ availability; its systematic neumation would not be undertaken until ca. 900. This scenario, whose origin we owe largely to

has won wide acceptance. 1 I believe it is wrong. is aim here to assemble witnesses to the existence of an authoritative

Solange Corbin,

My

neumed recension of the sooner

than is

To

begin,

Gregorian propers ca. 800,

a

century

presently supposed. we

Gregorian chant,

must

deal with two related

of

“archetypes”

containing the verbal texts alone, the other consisting of the same core of liturgy and text plus the supplement of neumes. Inasmuch as early traditions for the Mass propers are better one

documented than those for the Office, my focus will be

on

the

Mass,

1 Corbin 1952 , esp. 226-28, Corbin 1960 , 690-94, Corbin 1977 , 22-42. It has been embraced by Hucke 1980a, 445, and Treitler: “The earliest practical notations served primarily a cueing function for celebrants reciting ecclesiastical readings and The notation of antiphons, responsories, and Mass-Proper items for the prayers. cantor and schola did not begin until the tenth century” (1984, 176). For the Rutgers Symposium of April 4-5, 1986 at which the original version of the present paper was delivered, Dr. Hucke’s prospectus read: “Written tradition of Western music and of chant did not exist at the time of St. Gregory the Great, and not even when Roman Chant was introduced into the Frankish Empire. It did not begin until ca. 900. Before chant was written down around 900, it was transmitted orally. To study the history of chant up to 900 is to study an oral tradition.” .

.

.

.

.

.

DOI: 10.4324/9781315090283-12

Oral and Written Transmission in Chant

the historical situation should be closely parallel for the Office. The oldest full witnesses of the verbal texts of the Mass propers are the half-dozen documents that receive a masterful edition in Hesbert

although

1935 ; they are the first six items in Table All are of apparently north-European

1

.

with dates ranging and later ninth centuries. A substantial

between the later

origin,

eighth points to a standard text-recension circulating in Frankish regions by ca. 800, a recension whose lost original I shall call the Carolingian Text Archetype. In one of the oldest sources, consensus

Bland

among them

(Table

No.

1,

2,

dating

the

ca.

there is

800),

an

annotation

with those of Roman

provisions contrasting manuscript’s 2 antiphoners known to its compiler. Thus there were Italian roots reaching farther back than the Frankish sources, but no substantial relic of an eighth-century Roman text is preserved. As for the

own

Gregorian

propers with

neumes, the

earliest

surviving

witnesses date from ca. 900, a century after those with the text alone. 3 Some of the more important of them are listed in Table 2 These too are in the main northern European, and there is a .

substantial

of consensus among them which I shall call the Carolingian Neumed

degree

pointing to an archetype

Archetype. Their dates It this distribution of the is around 900. surviving sources, with begin text witnesses beginning ca. 800 and neumed witnesses ca. 900, which lends support to the theory that neumes were not supplied to the full Gregorian repertory until a century after the text tradition was established. In my view, the two traditions were closely linked in date and function, with both circulating around the end of the eighth century. Both would represent Charlemagne’s politics of liturgical renewal; both would implement the changes that were set in motion by Charlemagne’s father Pepin at the time of Pope Stephen II’s visit France in 754 (Vogel 1965a). Let me address the view that the Gregorian musical collection began later (“ca. 900”) than the text collection (“ca. 800”), by examining first what I see as its flawed rationale. It is true enough that a century-long gap separates the earliest preserved witnesses of text from those with music, and what neumations there are from the to

middle and later ninth century

chants—tropes, 2

sequences,

are

to

hymns, lections,

new and ancillary Celebrants’ chants,

Hesbert 1935 , No. 179 (the Seventh Sunday after Pentecost): “Ista ebdomata in antefonarios romanos.” The situation is studied in Hesbert 1932-33. Both types are surveyed in Jeffery 1983

non est 3

given

.

TABLE 1 ABLE I. 1.

Carolingian

Text

Archetype:

Early Descendents (Mass-books) No.

Name

Date

Type Antiphonale

Bland[iniensis]

late

8/early

9

Ant. miss. Cantatorium

late late

8/early 8/early

9 9,

Ant. miss.

late 9

Senlis Corbie

Ant. miss. Ant. miss.

betw.

Lucca fragment “Monza Schwesterhandschrift”

List of chants (Advent) Cantatorium

Monza Cantatorium

Compiegne [Compendiensis]

missarum

Presentation

northern France or Switzerland northern Europe northeastern France Soissons?

deluxe

877-882 “shortly after 853”

St. Denis/Senlis? Corbie

elegant elegant

late 8 late 8/early 9,

Lucca? northeast France

routine

or

or

middle 9

middle 9

elegant elegant

purpureus; uncials

purpureus; uncials

(Zurich, Zentralbibliothek, Rheinau 30): see Hesbert 1935, no. 2; Lowe 1934-66, vol. 7, no. 1019; Camber 1968, nos. 802, 1325; Graduale (“8e-9e siecle”); Jeffery 1983, 319. Bland!iniemis](Brussels, Bibl. Royale, MS. 10127-44): see Hesbert 1935, no. 3; Lowe 1934-66, vol. 10, no. 1548; Gamber 1968, nos. 856, 1320; Graduale 1957, 37 (“fin 8e/d£but 9e s.”); Jeffery 1983, 319. Monza (Monza, Tesoro della Basilica S. Giovanni, Cod. CIX): see Hesbert 1935, no. 1 (“8th c.”); Gamber 1968, no. 1310 (Bischoff: “2nd third of 9th c.”); Graduale 1957, 77 (“debut 9e”); Jeffery 1983, 320. Compiigne (Compendiensis: Paris, Bibl. nat., MS. lat. 17436): see Hesbert 193J, no. 4; Gamber 1968, no. 1330; Graduale 1957, 109; Froger 1980 (“Soissons, late 9th century”); Jeffery 1983, 319. Senlis (Paris, Bibl. Ste.-Genev., MS. lat. 111): see Hesbert 1935, no. 6; Gamber 1968, nos. 745, 1322; Graduale 1957, ii3;Jeffery 1983, 319. Corbie (Paris, Bibl. nat., MS. lat. 12050): see Hesbert 1935, no. j; Gamber 1968, nos. 745, 1335; Graduale 1957, ioj; Jeffery 1983, 319. Lucca Fragment (incipits list: Lucca, Bibl. capit., 490, fol. 30-31): see Hesbert 193J, no. 7; Lowe 1934-66, vol. 3, no. 303; Gamber 1968, no. 1302; Graduale 1957, 65 (“end of 8th c.’’); Huglo 1951; Froger 1979; Jeffery 1983, 320. “Monza Schwesterhandschrift" (Berlin, Cleveland, [and Trier]): see Gamber 1968, no. 1311 (Bischoff: “2nd third of 9th c.”); Graduale 1957, 143 (“beg. 9th c.”); Sifffin 1950; Jeffery 1983, 320. Rheinau

1957, J55

Oral and Written Transmission in Chant

Rheinau

Origin

Table

2.

Carolingian Neumed Archetype: Early Descendents (Mass-books)

Laon 239

Date

Type

Name

No.

Antiphonale

ca.

Ant. miss. Cantatorium Cantatorium Ant. miss.

end 9th c. early 10th end

Neume type Lorraine

900-930

missarum

2.

345-

Chartres 47 Saint Gall 359 Laon fragment Valenciennes

fragment

ca.

Breton Saint Gall Lorraine Breton

c.

900

9th c.

6.

Monza excerpt

Ant. miss.

early

10th

c.

proto-

7-

Albi excerpt

Ant. miss. (Process-

early

10th

c.

proto-

ionale)

Nonantolan?

Aquitaine

Laon 239 (Laon, Bibl. mun., MS. 239): see Gamber 1968, no. 1350; Graduale 1957, 57 (“vers 930”); Mocquereau 1909-12 (facs. ed.). Chartres 47 (destroyed): see Gamber 1968, no. 1351; Graduale 1957, 43 (“ioe s.”); Huglo 1979 (“end of 9th c.”); Mocquereau 1912-14 (facs. ed.). Saint-Gall 379 (St. Gallen, Stiftsbibliothek, Cod. 359): see Gamber 1968, no. 1315; Graduale 1957, 132 (“debut du ioe s.”); Mocquereau 1924 (facs. ed.). Laon fragment (Laon, Bibl. mun., 266): see Gamber 1968, no. 1313; Jeffery 1982; Jeffery 1983, 320. Valenciennes fragment

(Valenciennes,

Bibl. mun.,

1957, 148; Jeffery 1983, 320. Monza excerpt (Monza, Duomof-1/101): 20-23, tav. VII. Albi excerpt (Albi, Bibl. mun., 44): see

see

407):

see

Gamber 1968,

Gamber

nos.

1968,

no.

I304d; Graduale

801, 1250, 1336; Dalmonte 1969,

Huglo 1982, 253-68, pi.

V.

etc.—rather than to the Gregorian propers. Yet this fails the plausibilities of survival. With representatives as rare as

there is

an

obligation

to

consider

of

Mass-antiphoners

between the

eighth and tenth century, they to consider what may have failed to survive as are

well as what did. Table 1 has shown the chief early descendents of the Carolingian Text Archetype. Of the eight items listed there, only three— Rheinau, Bland, and Lucca—go back to the late eighth or early ninth century, to the period before we have actual evidence of neumes. The Monza Cantatorium and its Trier-Berlin-Cleveland sister-fragments may date from that remote time, or they may (as Bischoff indicates)

only from the second third of the ninth century (Gamber 1968 1311). As for Compiegne, Senlis, and Corbie, they are of later date, originating at a time when notations are already available. Thus only Rheinau, Bland, and Lucca, are surely earlier than our first

date

,

no.

sources with neumes: only these three survive as representatives of the text-antiphoner during the three-quarters of a century from the late 8th through middle ninth century.

It is hazardous Yet for

to assess the

text-antiphoners

Rheinau, Bland,

and Lucca

we

survival of the

rates

of

Carolingian

manuscripts.

“early” period represented by

have indications that

they cannot have

been altogether rare. A number of documents indicate the obligation of priests to know the content of the Antiphoner. 4 If each fair-size or monastic house had just a single text-antiphoner, the number of copies should have mounted into the hundreds during the first century of the Carolingian ecclesiastical reform. At certain houses there were multiple copies. An inventory of 831 for Centula (St. Riquier) lists six volumes of “Antiphonarii,” of which none still exists. 5 If we measure the survivors against the numbers that are apparently lost, the disappearance-rate is so high that the extant text-antiphoners can not really be taken as statistical indices. They are accidents, all of which might have disappeared. They may tell us nothing about the original situation. As for the neumed antiphoners (Table 2 ), they were doubtless fewer in number than the text antiphoners (Table 1 ). The early

church

neumes were too

small to be read

by choral singers during

a

service,

4 a. An edict, perhaps from the last decade of Charlemagne’s reign, “Haec sunt ,” lists the Creeds, Lord’s Prayer, quae iussa sunt discere omnes ecclesiasticos. contents of the Sacramentary, etc., and then: “9. Cantum Romanorum in nocte. 10. Et ad missa [sic] similiter; 11. Evangelium intellegere, seu lectiones libri comitis [the Gospels and Epistles]; 12. Omelias dominicis diebus...; 15. Scribere cartas et .

.

epistulas (Boretius 1883 235). b. An episcopal(?) edict to priests at a diocesan synod, probably of the early ninth century: Ammonere vos cupio, fratres et filioli mei, ut ista pauca capitula quae hinc scripta sunt intentius audiatis. 1. Imprimis, ut sacerdos Dei de divina scriptura doctus sit.... 2. Ut totum psalterium memoriter teneat. 3. Ut signaculum [the Creeds] et baptisterium [words and prayers of the baptismal service] memoriter teneat. 4. Ut de canonibus doctus sit et suum penitentiale bene sciat. 3. Ut cantum et compotum ,

[calendar matters] sciat. 6. Ut nullus sacerdos feminas

secum habitare permittat.... 7. Ut presbyteri in tabernis bibere non praesumant," etc. (Boretius 1883 , 236-37). These are fundamental concerns for all priests, prominent among them the memorization of the Psalter and Creed, and the knowledge of chant. c. Such prescriptions are elaborated in a capitulary of Bishop Haito of Basel (807-23) whose sixth paragraph lists the Antiphoner as a necessary volume: Quae ipsis sacerdotibus necessaria sunt ad discendum, id est sacramentarium, lectionarius, antifonarius, baptisterium, compotus, canon penitentialis, psalterium, homeliae per circulum anni dominicis diebus et singulis festivitatibus aptae. Ex quibus omnibus si unum defuerit, sacerdotis nomen vix in eo constabit" (Boretius 1883 , 363). d. At the Council of Rispach in 798: Paragraph VIII. Episcopus autem unusquisque in civitate sua scolam constituat et sapientem doctorem, qui secundum traditionem Romanorum possit instruere et lectionibus vacare et inde debitum discere, ut per canonicas horas cursus in aecclesia debeat canere unicuique secundum congruum tempus vel dispositas festivitates, qualiter ille cantus adomet aecclesiam Dei et audientes aedificentur (Werminghoff 1906 , 199). 5 Stäblein 1979, 78*, n. 381; the note lists further instances.

and

they were unlikely to be consulted during performance by ninth-century soloists and choirmasters who still consigned much

about the repertory to their memories. In all, there were fewer musicians who used fully-neumed proper collections than there were priests who used missals, and fewer well-heeled individuals would have kept a noted book for the pure pleasure of ownership. The chance of survival was accordingly smaller for the neumed collections than for the text collections—whose chances we have seen were

vanishingly

small. Other factors should have increased the text collections’ odds of survival. The preserved descendents of the Carolingian Text Archetype

tend

to

be attractive

examples of the

book-creator’s

Sextuplex manuscripts they range from

the

Rheinau and Bland

to

the

purpurei

the

art. Among simple elegance of

of Monza and Berlin-Cleveland

column). For manuscripts of the Carolingian Neumed Archetype (Table 2 ), the already poor prospect of survival was further dimmed by the fact that even when nicely executed, the (See

Table

1

,

last

pages bearing neumes rarely have the tidy attractiveness of those with alone. And where a text-antiphoner might remain useful indefinitely,

text

requiring little change in order to be kept current, the neumed antiphoners were rendered obsolete by notational innovations of the tenth and eleventh centuries. The emergence of staff lines and clefs meant that new books were substituted, and the older ones with prediastematic neumes had little further purpose. All things considered, it is remarkable how few text-antiphoners survive from the later eighth and ninth centuries. It should therefore be no surprise if there are none at

all with

In the

neumes.

the

evaluating ninth-century

plausibilities of survival,

one must

also consider

neumations of music other than Gregorian chant. As a group, these are earlier than the Gregorian neumations, which has fostered the theory that the neumes were invented to serve other repertories than the Gregorian. Yet in most instances the survival of such strays is not attributable to the music itself but to the nature and content of the host manuscript. Certain of the miscellaneous early neumations are for celebrant’s chants (Preface, Exultet), lections, etc., where survival is due to the texts, and to host Sacramentaries and

Lectionaries that

are

exceptionally

fine. 6 In other cases, such

as

the

6 The twenty-one items in the “Table of Extant Examples of 9th-Century Notation” in Hiley 1980 are the best current inventory of neumatic incunabula; some further possibilities appear in the inventory in Corbin 1977 , 21-41, among them the Sélestat Lectionary (Corbin, Taf. 2), and the neumed Exultet in Arsenal 227 (Corbin, Taf. 5).

“oldest dated neumes,” set to the prosula Psalle modulamina, or various relics of the Pentecost Greek Mass (missa graeca), the neumed entries are additions to manuscripts whose principal content is not related to the music. 7 To sum up, the theory of the neumed antiphoners’ origin a century after the text antiphoners’ may be commended for its cautious reliance on the surviving evidence, but it does not stand up to scrutiny. When closely examined, there is no aspect that appears soundly based. The absence of early noted antiphoners does not

validate the “late” scenario of Corbin, scenario to which I

does it

preclude the “early” Speculations have, of course, long existence of noted recensions going back even as nor

now turn.

circulated about the far as Gregory the Great, but no plausible case has yet been made. 8 1 offer seven indices of the existence of an “early” Gregorian neumation, all pointing well back of 900, some pointing back of 800. The first index of an

“early” date for the Gregorian neumation is paleographic common sense. Around the year 900, when the first substantial witnesses of the noted Gregorian propers appear (see Table 2 ), there are already marked differences in regional ductus: the distinctive neumes of Lorraine, Saint Gall, Brittany, etc. The brilliant work of our generation’s Gregorian semiologists, the “école a matter

of

has affirmed the long-held premise—going back to Volumes and 3 of the Paléographic musicale—that a common neumatic archetype lies behind the diverse regional manifestations. Dom Cardine calls this the “archetype d’ecriture” (1977, 174). I prefer a terminological

Cardine,” 2

distinction between the Carolingian Neumed Archetype and its neumeless counterpart, the Carolingian Text Archetype. Yet whatever the name, the varieties of ductus ca. 900, in the main

descendents of an authoritative archetypal neumation, render awkward the claim that “the different regional paleographic styles go back to the very beginning of neume notation” (Hucke 1980a, 445). It is more likely that a period of development lay between the neumed its and first archetype preserved descendents. Allowing for paleographic one should change, suppose at least an intervening half-century, and perhaps much longer. 7 Hiley 1980 , “Table of Extant Examples of 9th-Century Notation,” 7th, 10th, and 12th items. 8 Recent advocates of an eighth century or earlier date have included Anglés 1954, 106-8, arguing for the time of Gregory, and Froger 1978. Froger correctly—as I see it—dates the archetype:". we aim to restore the Gradual to the state in which it was diffused in the Carolingian Empire from the last quarter of the 8th century” (p. 82), but he attempts no justification of the dating. .

.

A second index of an “early” Gregorian neumation can be drawn politics of the Frankish Empire, whose subdivisions were

from the

beginning

before the death of

even

Charlemagne

in

814.

The most

considerable of these came in 843 with the Treaty of Verdun, which formalized the growing split between a French West and German East. Musical consequences are recognizable in the differing states of the East- and West-Frankish sequence repertories. Among the Gregorian neumed propers, there are discrepancies reflecting similar causes but these are small while the agreements are large. Huglo has observed that this points to a noted Gregorian archetype before the middle ninth century (Huglo 1975 ). The third index of an “early” date for the Carolingian Neumed Archetype is the reception of the Gregorian repertory in south Italy.

Charlemagne

took

over

the old

kingdom

of the Lombards after the

Pavia in 774, but in practice he was limited to the capitulation northern Duchy of Spoleto, and it was only ca. 787 that the southern at

Duchy of Benevento came under effective Carolingian control. South Italy is unlikely to receive a Frankish-Gregorian transmission before that time (Gay 1904 25-48). At the later end, Dom Hesbert has placed the arrival of Gregorian chant at Benevento “before ca. 808” (1936, 45off.). I pointed out some time ago that Hesbert’s reasoning was sound but his date represented a faulty reading which has to be changed to “before ca. 838” (Levy 1970, 221, n. 100). Nevertheless, a Gregorian musical transmission would arrive at Benevento between ,

ca.

787

and

ca.

838,

and since the neumatic details of the Beneventan

readings agree with those of northern Europe, the indication is again of an archetypal Gregorian recension before the middle ninth century. This will be amplified in my seventh index, below. The fourth index of an “early” noted archetype depends on the missa graeca, the composite of Byzantine, quasi-Byzantine, and Latin musical elements which was evidently assembled for the celebration of some Frankish imperial Pentecost of the late eighth or early ninth century. Charles Atkinson has tentatively proposed the years 827-35 for the compilation (1982, 144f.). I would prefer to keep it during the last decades of Charlemagne’s reign, in particular between ca. 797 and 814 (Levy 1963 36). Yet for present purposes either dating will suffice. What matters is that there are some six dozen manuscripts of the ninth through twelfth centuries with traces of this Pentecost Greek Mass. 9 They represent nearly every region: France, Germany, Lowlands, England (by way of France). The exception is Italy, which ,

9 Atkinson

1982

,

120-25,

provides

an

exhaustive

inventory.

has

at most

relation to

five

sources

Pentecost, and

with “Greek”

chants, none showing any originating in the Beneventan zone. Frankish-Gregorian tradition that came to none

Now it may be that the Benevento left the north at

a time so late that the Pentecost Greek mattered longer enough for inclusion. Yet the Beneventan transmission occurred “before ca. 838.” Thus a likelier assumption is that it left the north before the elements of the Pentecost Greek Mass were annexed to the Gradual and Troper, hence at the latest by

Mass

no

827-835,

and

perhaps during

The fifth index

the

reign of Charlemagne.

from the Musica disciplina of Aurelian of 10 ca. Reome, composed 850. Lawrence Gushee has demonstrated that Aurelian knew the neumatic notation. 11 Yet to judge from the nature and considerable abundance of his musical citations, Aurelian may not have supplied the treatise with noted illustrations. He may have expected his specialist-readers to have neumed antiphoners for consultation. There is an indication of this in Cap. X, 27-29, where Aurelian cites an unusual passage in two Responsory verses of the first comes

mode: “There is in this mode [the First Authentic] a certain phrase (divisio) in the Noctums that I do not remember finding elsewhere in the whole repertory (latitudinem) of the Responsory verses, except only in the verses of these two Responsories: V Domitte, ne in ira tua; t Timor et tremor, and V Peccantem me cotidie; y Deus in nomine tuo. Which verses, although they could have the same arrangement as the others, nevertheless, because this remained so among the ancients, so also must it remain 12 among us, in their memory.”

Responsories of the

,

10

For a modern edition, see Gushee 1975 Gushee 1963 , 215:“. it is precisely here that we meet with phrases positively indicating the use of musical notation. For instance, we cannot take the phrase ‘by the as metaphorical: '. since, indeed, there are eye’ (oculo) in the following sentence some tones that have, in their inflection, very nearly the same arrangement of their verses, the tenor of the one tone will change to the other, unless they are surveyed by the eye [Gushee’s italics] with a cautious inspection or examination at the middle or the end. This the diligent singer can easily recognize in these two tones, namely, the first ” plagal and the fourth plagal.’ (“. quoniam quidem sunt nonnulli toni qui prope uno eodemque modo ordine versuum in suamet retinent inflexione, et nisi aut in medio aut in fine provida inspectione aut perspicatione antea circumvallentur oculo, unius toni tenor in alterius permutabitur. Quod studiosus cantor in his duobus tonis, videlicet plagis proti et plagis tetrardi otius valet agnoscere,” Gushee 1975 , 118.) I would add that in various instances Aurelian makes detailed melodic comparisons, singling out individual syllables of particular chants of the Gregorian repertory with such directions as “on the 15 th syllable of. ....” This too indicates neumations. 12 “Est in hoc tono quaedam divisio in noctumalibus responsoriis quam non uspiam memini me in latitudine totius responsoriorum versibus repperisse, nisi solummodo in versibus istorum duorum responsorium. Hi autem sunt: Resp. 11

.

.

.

...

.

.

.

.

The survey of the verse repertory that Aurelian describes—“quam non uspiam memini me in latitudinem totius responsoriorum versibus

repperisse”—is unlikely to have been a but rather

a

point

to

point comparison

antiphoner.

And when Aurelian

because it

was

mansit”),

he is

used

likely

“among to

opts

scroll

through a memory bank,

of neumed chants in

to

a

reference

preserve the musical

the ancients”

anomaly (“apud antiquos ita

have found the former tradition in

a

noted

antiphoner. Cap. XIII, the reader is told of a musical passage in two verses that is “not found elsewhere in the prolixity Gradual Gregorian of the whole Antiphonale” (“memini me non alicubi repperisse in 13 prolixitate totius antiphonarii”). One must again conclude that Aurelian is not referring to a singer’s well-stocked memory but to the Then in

noted

“prolixity” of a fully My sixth index

Mass-antiphoner. “early” date comes from the famous or Admonitio generalis, addressed by Charlemagne to the capitulary Frankish clergy on 23 March 789, setting forth the guidelines of his ecclesiastical reforms. In its eightieth chapter there is the injunction to all clergy (“omni clero”) that they fully learn the Roman chant which of the

Pepin ordered substituted for the Gallican chant. In its seventy-second chapter, addressed to “priests” (“sacerdotibus”), there 14

his father is the

injunction,

that there be schools for teaching boys to read. Be sure to emend carefully in every monastery and bishop’s house the psalms [psalmos], notes [notas], chants [cantus], calendar material [computus], grammar[s] [grammaticam], and the Epistles and Gospels [libros catholicos]. For often ". ..

ne in ira tua; y Timor et tremor, et item: Resp. Peccantem me cotidie; f Deus in nomine tuo, qui, cum ordinem possidere queant ceterorum, tamen [quia] apud antiquos ita mansit, apud nos quoque ob eorum memoriam necesse est permanere" (Gushee 1975 , 88-89). I cannot say how how rare this phrase is among the First Mode Responsory verses, but the two verses end identically in the readings of the Worcester Antiphoner (Mocquereau 1922-25, 60 and 137), suggesting that the chants known to Aurelian in 850 were like the ones we know. 13 Gushee 1975 , 99. I am indebted to Professor Gushee for confirming this reading of the Valenciennes manuscript. Aurelian here cites identical verse-endings of the Graduals Tollite portas and Haec dies; our later neumed traditions (Graduate Triplex, 25 and 212) once more indicate that he is talking about the chants we know. 14 “Ut cantum Romanum pleniter discant, et ordinabiliter per nocturnale vel gradale officium peragatur, secundum quod beatae memoriae genitor noster Pippinus rex decertavit ut fieret quando Gallicanum tulit ob unanimitatem apostolicae sedis et sanctae Dei aeclesiae pacificam concordiam” (Boretius 1883 , 61).

Domine,

.

.

enough there are those poor

who

[inemendatos libros]

texts

call upon God do it poorly.” 15

want to

they

well,

but because of

The passage is often cited as a possible indication of neumes, though so far without firm claim (Hiley 1980, 334). Within my present historical framework, Charlemagne’s “notas” may perhaps be taken at face value. At issue are both the word notas and its context: psalmos, notas, cantos, computus, grammaticam, and libros catholicos. These are texts to be scrupulously emended for ecclesiastical establishments. We await adequate guides to Merovingian-Carolingian usage, but a sifting of available glossaries has produced no support for the view that Charlemagne’s “notas” refers to notaries’ shorthand signs (“Tironian notes”) or schoolboys’ jottings about texts under study, rather than to the neumes of plainchant—which is what the context

suggests:

“.

.

.

psalmos,

notas, cantus.

.

.

.” 16

My seventh index of an “early” date for the Carolingian Neumed Archetype points again to 800 and the reign of Charlemagne. It depends in part on eleventh-century evidence and hence cannot, any more than my previous indices, produce firm conclusions about an eighth-century Gregorian neumation. Yet it comes as close to proof as I think

anything can. Despite Charlemagne’s characterization of his father’s thoroughness in suppressing the old “Gallican” chant, 17 it appears that the

15

“Et ut scolae legentium puerorum fiant. Psalmos, notas, cantus, compotum, grammaticam per singula monasteria vel episcopia et libros catholicos bene emendate; quia saepe, dum bene aliqui Deum rogare cupiunt, sed per inemendatos libros male rogant” (Boretius 1883 60). 16 Franz Blatt's Novum Glossarium mediae latinitatis (Copenhagen, 1957- ), the best compilation, begins only with the ninth century; its extensive entries under neuma (pp. 1231-34) and nota (pp. 1391-96), assembled by Anne-Marie Bautier-Regnier, are excerpted in Bautier-Regnier 1964. The comprehensive Mittellateiniscbes Wörterbucb (1959- ) of the Bavarian and Berlin Academies does not reach this portion of the alphabet. I have consulted Maigne d’Amis, Lexicon manuale (Paris, 1860); Forcellini et al., Totius latinitatis lexicon, (Padova 1864-98); Diefenbach, Novum glossarium (Frankfurt, 1867); Du Cange-Favre, Glossarium mediae et infimae latinitatis (Niort, 1883-87); A. Bartal, Glossarium mediae et infimae latinitatis regni bungariae (Leipzig, 1901); F. Arnaldi, Latinitatis italicae medii aevii (Brussels, 1939); A. Souter, A Glossary of Later Latin to 600 A.D. (Oxford, 1949); R. E. Latham, Revised Medieval and Latin Word List from British and Irish Sources (Oxford, 1965); J. F. Niermeyer, Mediae latinitatis lexicon minus (Leiden, 1976); and P. G. W. Glare, Oxford Latin Dictionary (1982). 17 ". quod beatae memoriae genitor noster Pippinus rex decertavit ut fieret ."(Boretius 1883 61). quando Gallicanum tulit ,

.

.

.

.

,

process by which Gregorian chant supplanted Gallican chant in the Frankish homeland was not one of outright exchange of an existing corpus for a wholly different one. Only the final stages can be observed in surviving documents, but from an amount of apparently Gallican material that lingers in the Sextuplex and other early chant sources it would seem there were preliminary editions

liturgical-musical

a while perpetuated locally-esteemed Gallican matter that later became obsolete. In a recent article, I outlined one such class of Gallican survivals: “non-psalmic” Offertories, whose texts are “centonate librettos,” composed of fragments drawn from biblical or post-biblical narrative rather than the verbatim Psalter. Certain

which for

Offertories of this class (among them, Sanctificavit, Erit hie vobis, Oravi Precatus est Moyses, and Vir erat) found permanent places in the

Deum,

Frankish-Gregorian stations such

as

canon,

the final

assigned

Sundays

to later

after

and lesser calendar

Pentecost,

or

the

formerly

aliturgical Thursdays Lent; presently they were absorbed even into the Urban Roman liturgy. For others, there was a briefer regional or in

local survival. Some of these “Gallican

apocrypha,”

attached

to

the

fringes Gregorian traditions, open extraordinary perspectives on early history of the rite (Levy 1984). One such chant is the Pentecost Offertory, Factus est repente. As with other Offertories of this type, its text is a non-psalmic libretto, of

the

centonized from passages in the second chapter of Acts. The chant has already served as the centerpiece of a discussion by Dom Hesbert, who concluded that it was both “old” and “Roman” (1963, 68). I return to Factus est repente now, not because I view it as “Gallican” rather than “Roman” in origin, but because evidence unknown to Hesbert gives it a significant role in the history of the neumed

archetype. Hesbert knew Factus est repente in two recensions, one northern European, the other Italo-Beneventan. In the Frankish north—its presumable homeland—he found it only in Bland, one of the two

oldest un-neumed sources of the Gregorian Antiphonale missarum (Table 1 No. 2), a codex for which there is every indication of late eighth or early ninth century origin. Factus appears in Bland as an ,

alternative to the standard Gregorian Pentecost hoc Deus, whose psalmic text it follows:

Offertory, Confirma

OFF. Confirma hoc Deus quod operatus es in nobis, f. I. Cantate Domino psalmum dicite nomini ejus. V II. In aecclesiis benedicite Domino Deum de fontibus Israhel. f [III.] Regna terre cantate Deo. Item Off. Factus est repente de caelo sonus tamquam advenientis in spiritu vehementis & replevit totam domum hubi erant sedentes alleluia. .

.

y I. Et repleti sunt omnes Spiritu Sancto alleluia alleluia (Hesbert 1935 124, no. 106).

loquentes magnalia

.

Dei

,

the

Among earliest and

Sextuplex traditions, then,

narrowest of

circulations. It

Factus est repente has the already to be obsolescent

seems

ca. 800. For the centuries that followed, Hesbert’s search of than six hundred Graduals and Missals uncovered no further sources from northern Europe. But it did turn up seven from Italy, all from the eleventh and twelfth centuries. Six of these are from dating

by

more

the Beneventan zone, and the seventh is from the Abruzzi, abutting the northern limit of the Beneventan zone. In these Italian versions the Offertory refrain has the same text as Bland, but the verse has an

ampler beginning

and

an

altered conclusion:

BLAND VERSE. Et repleti sunt omnes Spiritu Sancto loquentes magnalia Dei alleluia alleluia. ITALIAN VERSE. Et apparuerunt Apostoli dispertitae linguae tamquam

ignis, seditque Sancto,

et

supra singulos eorum; et repleti coeperunt loqui magnalia Dei. Alleluia.

sunt omnes

Spiritu

The six Beneventan versions are noted, and their agreements indicate a common neumed source. Hesbert’s transcription of the Beneventan Factus est repente, an “antique composition qui, bien exécutée devait être fort belle,” is based upon the ensemble of Beneventan .

.

.

it is shown in Example i (Hesbert 1063, 60). There are two main differences between the “northern” text-tradition of Bland and the composite “southern” tradition of Benevento and the Abruzzi (Ex. 1). One is the discrepancy in

readings;

verse-texts,

something

to which I

will return. The other is the

liturgical assignment. prescribes Factus est repente as an alternate for Pentecost Sunday, appended to the standard Gregorian-Roman Confirma hoc Deus, while the Italian sources have Factus only at Thursday in Pentecost week. Hesbert found a satisfactory explanation: Bland

Factus was

fixture

Pentecost

Sunday but was obliged to give Gregorian provision of Confirma hoc; it was in the Beneventan-Abruzzese tradition by transfer to a preserved that occasion lacked chants (Pentecost Thursday neighboring proper once a

on

way to the newer

long remained aliturgical). 18 Having reclaimed Factus leave

to this extent,

it, the available documents

opening

18 Hesbert 1963, 68f.; Hesbert’s reasoning is doublets” described in Huglo 1971 296. ,

Hesbert was obliged to further avenues. His

no

an

application

of the “loi des

Example

i

Offertory,

Factus est repente (Beneventan versions,

conclusion that it

was

“An Old

Offertory

Hesbert).

for Pentecost” will stand.

of “Roman” origin should have been supposition in when there were indications that the even 1963 suspect with which recensions Factus est repente circulated were “Gregorian” 19 north. in the Frankish The chant has no analogue in promulgated

But his

that it

transcr.

was

the Old-Roman musical repertory

(Cutter 1979 ). And the class of that it represents now centonate-libretto” Offertories “non-psalmic, to be of Gallican ( appears origin Levy 1984 ).

Despite the thoroughness of Hesberts search, two important witnesses of Factus est repente eluded him. One was a noted version 19

Hucke 1954 ,

Huglo

1954 ,

Apel 1956 Gajard ,

1959.

from the Beneventan zone, the earliest yet discovered, datable around the middle eleventh century. The other was a noted version from the Frankish homeland, datable around 1000, hence earlier than any Beneventan neumation, and a unique witness of the northern melodic practice. Singly and together, they expand the evidence so considerably that I would now venture the following four propositions: 1. that the Gallo-Gregorian Offertory Factus est repente reached south Italy from the Frankish north by ca. 800; 2. that Foetus arrived as a component in a full Gregorian musical

recension; 3. that the music of Factus made its transmission but fully neumed;

journey,

not in an

4. that the whole Gregorian recension with which Factus Italy ca. 800 was itself fully neumed.

oral

came to

Hesbert could have encountered the new Beneventan witness only with difficulty since it was in an American collection whose resources were not included in the Solesmes documentation upon which he relied. 20 MS. W.6 of the Walters Art Gallery in Baltimore is a missale plenum of the middle eleventh century, written for Canosa, near Monte Gargano, in the southeast of the Italian peninsula. The text is executed in the “Bari” type of Beneventan script, and is embellished with handsome zoomorphic initials. Small in outer dimensions (19× 12 cm.), the manuscript contains the prayers, lections, and chants for the principal feasts of the Temporale and Sanctorale, but it lacks the bulky provisions for the numbered Sundays and seasonal weekdays. It may have served a prosperous cleric for portable use. 21 The version of Factus in the Canosa Missal makes three important contributions to our dossier. First, it is the only Beneventan version in staffless neumes, hence it is the “earliest” witness of the Italian recension (see Figure 1 ). Second, it is assigned, not as in all other Beneventan manuscripts, to Pentecost Thursday, but to Pentecost Sunday itself. This appearance in the southeastern Beneventan zone, which tends to be more conservative in liturgical usage than the western regions, confirms the Offertory’s archaic assignment to Pentecost—which otherwise is found only in the eighth- to ninth-century Bland. Third, the Canosa missal supplies this confirmation in most striking fashion, for it does not prescribe Factus est repente merely as an alternative to the standard

20

It

was

listed in

Huglo

listing in Monte Cassino 21

197

,

194,

n.

2, and is also found as a Beneventan

318. Full edition and commentary in Rehle 1972.

tonary

Figure

1.

Canosa Missal, f. 152.

Gregorian Confirma hoc (which is stands here

as

the sole

These three

points

the situation

Offertory are

already in Bland); Factus

for Pentecost.

the basis of my first

proposition—that

Factus est repente reached South Italy ca. 800. In brief, the Canosa Missal, with Factus as its sole Pentecost Offertory, would represent an earlier stage of the

Gregorian tradition than any

other known:

a

stage

before the psalmic Offertory Confirma hoc became fixed at Pentecost, hence a stage even “before Bland,” where Confirma hoc is installed as the Pentecost Offertory, and the non-psalmic Factus est repente is relegated to alternate status. Bland, as we know, dates from ca. 800. My second proposition is that the full Gregorian repertory reached Italy at the same time as Factus est repente. It seems unlikely that this obsolescent “Gallican” Offertory journeyed to Italy as an isolated transmission of an individual chant; more plausible is that it came installed

the

as

Mass-antiphoner.

Pentecost

in

Offertory

Thus whatever

can

a

full

be said about the

Gregorian “early” date of

Factus in the Canosa Missal should apply as well to the whole transmission with which it came. We shall soon see decisive support for this.

My

other

new

witness of Factus

est

is the

unique northern Prüm, Paris, lat. 9448,

repente

neumation found in the

Gradual-Troper Figure 2 ). This manuscript has long been recognized as one of the best sources of the Gradual and Troper, yet in recent inventories it has been It was omitted from the Solesmes census of curiously slighted. Graduals and plenary Missals for the critical edition of the Roman Gradual, evidently because it was classed as a Troper ( Graduale 1957 ). It was omitted from Husmann’s census of Tropers for opposite copied

around

1000

of

(see

reasons. 22

Nevertheless Prüm was described by Gautier in 1886 and it has often been used since. 23 The failure of its neumation for Factus est

repente

to draw Hesbert’s

attention

must be

attributed

to a

simple

oversight. The Benedictine

Prüm

just 45 miles south of Aachen. Founded in 710, it capital benefices and visits from the Frankish monarchs, and it enjoyed many retained its prominence for centuries. During the 890s its abbot was abbey

Charlemagne’s birthplace

at

and

was

located

at

Regino, whose writings are among the era’s most informative on theory; and it was from Prüm in 1008 that the monk Bernhard, himself a commentator on liturgy and music, was called to the abbacy of Reichenau by the Emperor Henry II. The abbey’s cultural pretensions around 1000 can be gauged from this “magnifique manuscrit” (Gautier 1886), in which splendid full-page illuminations embellish a text that is rich in liturgical and musical archaisms. music

22 23

study.

".

Husmann 1964 , 110 calls it dem berühmten Prümer Graduale.” Gautier 1886 , 123. Lagerlöf 1983 , 125-78, is an amply illustrated art-historical .

.

Figure

2.

Priim

Gradual-Troper,

f. 51.

Among the latter is its Offertory Factus est repents. Since Prüm represents the same “northern” region as Bland, it comes as no surprise to find Bland’s provisions reproduced. Prüm’s Pentecost Offertory is the standard Gregorian Confirma hoc without verses. Immediately following, and again designated as an alternate (“Item Of[f]”), there is the “apocryphal” Factus est repente with verse Et repleti sunt omnes, both texts just as in Bland. Priim also has Factus cued at the Pentecost Octave, another mark of its former prominence. At the outset, Prüm answers an important question. I have spoken of the apocryphon Factus est repente as a Frankish import to Benevento, basing this on the “Gallican” style of its “non-psalmic, centonate-libretto" text. Yet only the Italian melody has been known, and there was no real way of telling whether the material that travelled south from Gaul to Italy included music as well as text, or whether the in the Beneventan sources was just a local and later composition. The Priim manuscript sets doubt to rest. A comparison of its

melody

(Fig. 2 ) with that of Canosa (Fig. 1 ), or with Hesbert’s Beneventan generic transcription (Ex. 1 ), shows that for the refrain both the text and music are the same. I will shortly return to this melody and subject its northern and southern readings to close scrutiny. But first I must deal with an objection to my theory that the transmission of Factus to Benevento neumation

was

(the

“pre-Bland” Canosa

or

Inasmuch

the oldest Italian witness what assurance is south during the later ninth, tenth,

“early.”

as

missal) dates from the eleventh century,

there that the music did not come or early eleventh century? There can be no assurance. At best I can refine my theory in the light of the Priim exemplar. In the Frankish north, the “Gallican”-style Factus est repente was obsolescent as the Pentecost Offertory by ca. 800, barely surviving among the Sextuplex sources, with only Bland transmitting it as a hanger-on. Prüm confirms the narrowness of northern survival by mirroring the provisions of Bland. Thus few local traditions used Factus est repente even ca. 800, and presumably still fewer kept it through the 9th and 10th centuries. One of these

was

the

source

for

Benevento,

whose

differs from that of Bland-Prüm. Yet the Frankish transmission Benevento is likelier to have occurred in the late 8th or early ninth

verse to

century (the time of Bland refrains and

verses was

or

earlier),

relatively

when the repertory of Offertory holdovers, rather

rich in Gallican

than later on, when the entrenchment of the standard Gregorian canon had narrowed such options. In addition, if one supposes a “late” (ninth to eleventh century) transmission of Factus to Benevento, one must

explain why

the standard

Offertory Confirma

hoc should be

at a time when there is no longer an apparent liturgical for any other Pentecost Offertory than Confirma hoc. Thus a “late” transmission of Factus to Benevento remains a possibility, but is less likely than an “early” transmission, at a time “before Bland,” when Factus would still flourish as an interim Pentecost Offertory in the Gallo-Frankish north. The third of my four propositions above was that when the music of the Factus refrain reached Benevento “ca. 800,” it arrived not in an oral transmission, but neumed. Under normal conditions of evidence there should be no way for documents of the eleventh century to tell us how singers of the eighth to ninth centuries practiced their craft, to assure us that they relied on neumations rather than the fashioning of

displaced warrant

their chants skills. With

the exercise of memory and improvisational an extraordinary constellation of evidence points

through Factus,

my earlier statement that the Priim and Beneventan melodies for the Offertory refrain are the “same,” and inquire how far this sameness extends. An answer can be taken from Figure 3 where the melisma et replevit of the refrain is shown in the readings of Prüm ( Fig. 3a), Benevento VI. 34 (Fig. 3c ), noted tradition

to a

ca.

800. Let

me return to

,

(Fig. 3d ), transcription (Fig. 3b ). Benevento VI. 39

The neumations source. neumes.

There

are

and Hesbert’s

composite

Beneventan

close

enough to indicate a common written agreements of pitch-groupings into compound are

There is the

uncommon neume,

pes

stratus

(a podatus plus

oriscus), which is considered a symptom of a Gallican or “imported” chant. 24 This appears three times in the Prum version of the melisma; in Benevento VI. 34 and VI. 39, each of the three instances is resolved a

podatus plus simplex—replacing

the final oriscus with

punctum-shapes; (which abbreviates its et replevit melisma, retaining only the beginning and end; see Figure 1 ), the first of the pes stratus is nevertheless translated with an oriscus; and as

but in the earlier Canosa missal

of an oriscus in other Beneventan neumations are reflected in Hesbert’s transcription. These agreements in detail indicate that Prum and the Beneventan manuscripts descend from the same noted

vestiges

formulation. Their

proposition—that

common

the

melody

neumatic origin supports my third of Factus arrived in Italy neumed.

24 Cardine 1970 , 131. But the pes stratus, in addition to its appearances in “Gallican” chants and Frankish sequences (see Huglo 1972 , 228 and 238), is common enough in early English and central-Italian collections of the Gregorian propers. In a future paper I will consider the possibility that this neume was applied to the full Gregorian repertory in early west Frankish traditions.

Figure 3. Melisma et Benevento VI. 39.

replevin,

a) Priim; b) transcription; c) Benevento VI. 34; d)

There is further support for this in my fourth proposition—that melody of Factus reached Benevento in a full Gregorian neumed recension. I have already suggested that Factus is likelier to have the

journeyed

chant. Thanks

suggestion

complete repertory than as an isolated to a sharp-eyed observation by Dom Hesbert, this be given solidity, and extended to show that the

from the north in can

transmission entailed the

of

neumes.

analysis newly-restored Pentecost Offertory disclosed that was almost melody entirely independent of the rest of the

Hesbert’s a

use

a

of his

Gregorian repertory. But at one small passage in the refrain—at the repente—Hesbert identified the musical fabric of Factus est repente with that of the Gregorian Offertory Angelas Domini at the words de celo (Hesbert 1963, 64). The parallel is shown in Figure 4 25 The situation is not unusual: a centonate formula appearing in different contexts of text and music, and at points far removed in the Gregorian Mass-book. There is an underlying logic, since the two Offertories are based on non-psalmic, centonate-libretto texts (Angelus word

.

Domini draws

on

Matt.

28),

and both represent the

G-plagal

mode of

this “Gallican” liturgical-historical type. However, there is a significant difference. Where Foetus est repente is an apocryphon of restricted preservation, Angelus Domini found a regular place in the Gregorian

appears in five of the

Sextuplex manuscripts, assigned

to

26 Monday, and the Easter Octave; it settled Holy Saturday, carries for Easter This the standard Gregorian Offertory Monday.

as

canon; it

Easter

historical potential. For if Factus and Angelus related in the neumation of their centonate repentelde celo

uncommon

closely

an

should be

formula,

then what

we

have learned about the

early history

of Factus

repente should apply as well to the history of Angelus Domini. And what we learn about the neumation of Angelus should in turn apply to the full Gregorian recension with which it circulated. Figure 5 compares the neumation of our two melismas: four

est

selected regional neumations of the de celo melisma in Figure 5a (the left-hand column), and the two regional neumations of the repente melisma in Figure 5b (the right-hand column). 27 In Figure 5a1 (de celo) the readings of Lorraine and Saint Gall accompany the square-note version of the Editio vaticana as reproduced in the Graduate Triplex. To judge from the quilisma in the opening figure, from the oriscus as penultimate pitch, and from the general agreements in neumatic disposition and melodic detail, it appears that Lorraine and Saint Gall descend from the same written archetype. But so does Prum (Fig. 5a2 ), whose neumation is close to that of Saint Gall. And so do the twelfth-century south Italian 4a: Graduate sacrosanctae romanae ecclesiae (Desclée, No. 696; Paris, Tournai, see Ex. 1, above. 26 Hesbert 1935 , Nos. 79b, 81a, 87. In light of its paschal assignments and the concordance of its centonate libretto text with a musically related Offertory of the Milanese rite for Easter Sunday (Antiphonale 1935, 210), Angelas Domini may descend from a “Gallican” Easter Offertory, just as Factus would from a “Gallican” Pentecost 25

Fig.

Rome, 1952), 246. Fig. 4b:

Offertory. 27 Fig. ja1:

Graduate triplex 1979 218; 5a2: Paris 9448, 36; 5a3: Benevento VI. 34, 5a4: Benevento VI.33, 84. Fig. 5b1: Hesbert 1963, 62; 5b2: Paris 9448, 51 (Fig. 3); 5b3: Benevento VI.34, 192; 5b4: Baltimore, Walters 6, 152 (Fig. 1). 132;

,

Figure

4. “Centonate” contexts of de celo and

repente.

Figure

5. Neumations of de celo and repente.

neumations of Benevento VI. 34 (Fig. 5a3 ) and its eleventh-century forerunner, Benevento VI. 3 3 (Fig. 5a4 ). In VI. 34 a scandtcus is substituted for the ornamental opening quilisma figure of Lorraine, Saint Gall, and Prúm; but the quilisma survives in Benevento VI.33. Thus there is a common written source behind the neumations of Angelas Domini in tenth- to eleventh-century traditions of the Frankish north and Benevento. As for the melisma repente of the Offertory Factus est,

a

comparison

of the

surviving

neumations

(Prüm

and

Benevento) in Figure 5b confirms what Figure 3 leads us to expect: that behind these two recensions there is also a common neumation. Now do the

themselves

Priim,

archetypal

turn out to

neumations for these two melismas

be the same?

the neumations of de celo

Looking first at the readings of 5a2 ) and repente (Fig. 5b2 ) are

(Fig.

identical. And if minor variants between the Prüm neumations leave doubt about their underlying identity, they are

substantially

of

Benevento, and the Canosa Missal (Figs. 5a3

removed

by

the

readings

VI. 33, of de celo and repente

,

are

precisely

for in Benevento VI. 34, 5b3 , 5a4 5b4), the neumations

the

,

same.

proposition receives substantial explain the exact neumatic good way support. in Prüm and of two contextually unrelated Benevento correspondence That

being

the case, my fourth

There is

no

to

occupying isolated comers of the Gregorian Mass-book, by supposing that what brought their two Offertories Domini and Foetus est repente) from the Frankish north to (Angelas southern Italy was a precisely-neumed, editorially homogenized recension of the full “Gregorian” repertory: in short, a neumed archetype whose existence can now be placed with some assurance

melismas

other than

around the year 800.

aim has been double: to

give greater substance to the concept Carolingian Archetype of Gregorian chant, and to make a hundred-year adjustment in its date, setting this back by about a century, from ca. 900, where it has settled in recent opinion, to shortly before 800, in the middle of Charlemagne’s reign. In my view, My

of the

Neumed

likely to have been employed during the later in eighth century effecting the changeover from Gallican to Gregorian musical repertories, and the authoritative neumation of the Gregorian propers, whose descendents we know in the Editio vaticatia, would be a fruit of Charlemagne’s Carolingian Renaissance. I have no firm proofs. Short of a dated early neumation or a dated description of early neumatic practice, it is difficult to imagine what form such proof could take. Instead, I have a variety of indices which point to the middle 9th and late 8th centuries as times when the noted Gregorian edition existed. Each index has some shortcoming. Concerning the varieties of neumatic ductus ca. 900:1 cannot prove that the regional neumations began in a common ductus and then evolved differently, but that is likelier than for them to have differed from the start. 28 Concerning the divisions of the Empire during the earlier ninth century, or the arrival of Gregorian chant at Benevento “before 838,” or the compilation of the missa graeca during the last decades of neither the Charlemagne’s reign: specific dates nor the connections the neumatic notation is

with the

Gregorian repertory are firm. Concerning the discussions by

Aurelian of Réôme (ca. 850) and the reference to “notas” in Charlemagne’s Admonitio generalis (789): the odds are that these describe neumed Proper collections, but neither instance is concluI will consider certain merits of the “diverse paper titled, “On the Origin of Neumes.”

28

beginnings” theory

in

a

forthcoming

sive. As for my final index, the “apocryphal,” “Gallican” Pentecost Offertory, Factus est repente and its “centonate” relationship with the

Offertory Angelas Domini: this should remove many doubts concerning the “early” circulation of the Carolingian Neumed Archetype; yet the possibility remains that Benevento received its melody Paschal

for Factus est repente after the date, ca. 800, which the early obsolescence of the Factus Offertory has led me to infer. Thus each of my indices has weaknesses; yet each also has strengths. Taken together,

considerable case for the neumed archetype—the ancestor of our later Gregorian recensions—as a product of the same fertile decades around the turn of the ninth century that saw the revised Carolingian editions of the sacramentary, homiliary, lectionaries,

they make

a

29

tonary, etc. I have gone

length in this exercise because the challenge is marshalling arguments for a difficult proof. The consequences go beyond a mere century’s revision in a medieval date. than

more

The

new

one

to some

of

perspectives that emerge from my revised chronology three central issues in the development of Gregorian

historical

bear

on

Concerning the origin of neume-species, my framework offers paleographers a sounder basis for projecting the pre-history of tenth-to eleventh-century neume-species than they have had before. Concerning

chant.

the interrelations of oral and written practice, if my supposition that the Gregorian propers were crystallized in neumes ca. 800 is correct, then various assertions about the continuing effects of oral and improvisational techniques on Gregorian melodies during the 9th 30 century and later will need fundamental review. Concerning the relationship between the “Old-Roman” and “Frankish” musical repertories, if my claim holds that the neumes were employed in the process of shaping the “Gregorian” recension during the later eighth century, then it may be asked whether some of the musical content of the “antefonarios romanos” known to the compiler of Bland ca. 800 was not also at that time cast in neumatic form. That is, how much of the

“Gregorian” musical substance that advocates of the “frankische Überlieferung” have been explaining as an essentially northern stylistic overlay (Hucke 1980b, 696-97) actually represents the Frankish melodic footprint? May the bulk of the Gregorian repertory not be 29 There is no adequate current survey of this liturgical reform (see Vogel 1965 a and 1965b, and Patzelt 1965). For the sacramentary, the situation is amply covered in the editions of the “Gregorian” and Gellone sacramentaries by J. Deshusses (1971 —79 and 1981). For the tonary, we have the masterful Huglo 1971 and also Lipphardt

1965. 30

Most

recent are

Hucke 1980 a and Treitler 1981 , 474.

attributable instead to large-scale appropriations of Roman melodies, with but minor northern retouchings and supplementations, the latter chiefly among the Alleluias and Offertories? On each of these three issues—the early history of neumes, the symbiosis of oral and written practice, and the genesis of the “Old-Roman” and “Gregorian” melodies—much remains to be said. I will return to them in further papers of this series devoted to the emergence of Gregorian Chant. Princeton List

of

University

Works Cited

“

Gregorian Chant .” New Oxford History of Music 2 : 92 127 London 1954 Antiphonale 1935 Antiphonak missarum mediolanensis Rome 1935 Apel Willi The Central Problem of Gregorian Chant .” This JOURNAL 9 Anglès Higini

-

.

,

.

,

.

,

.

.

.

,

“

.

,

( 1956): 118 27 Atkinson Charles Zur Entstehung und Überlieferung der ‘Missa graeca.’ Archiv für Musikwissenscbaft 39 ( 1982 ): 113 45 Bautier-Regnier Anne-Marie A propos des sens de neuma et de nota en latin mediéval .” Revue beige de muskolcgie 18 ( 1964 ): 1 9 Boretius Alfred ed. Capitularia Regum Francorum vol. 1 Monumenta Germaniae Historica Legum Sectio 2. Hannover 1883 Cardine Eugène Sémiologie grégorienne .” Études grégoriennes 11 ( 1970): -

.

“

”

.

,

-

.

“

.

,

-

.

,

,

.

,

,

.

,

“

.

,

1

-

158

.

“

_____ .

Vue d’ensemble

( 1977): 173

le chant

sur

gregorien .” Études grégoriennes

16

-

92

.

“

Les notations neumatiques en France à l’époque carolingienne." Revue d’bistoire de I'église en France 38 ( 1952 ): 225 32 Les neumes .” Histoire de la musique 1 : 689 98 Encyclopédic de la Pléiade, no. 9 Paris 1960 Die Neumen. Vol. 1, fasc. 3 of Paleograpbie der Musik nacb den Plänen Leo Schrades Cologne 1977 Cutter Paul F. The Musical Sources of the Old-Roman Mass Musicological Studies and Documents, no. 36. Neuhausen-Stuttgart 1979 Dalmonte Rossana Catalogo musicale del Duomo di Monza Bologna 1969 Deshusses Jean Le sacramentaire grégorien 3 vols. Spicilegium Friburgense Corbin

,

Solange

.

-

.

“

-

_____ .

.

,

.

.

,

. _____

.

.

,

.

,

,

.

,

.

.

,

16, 24,

.

.

,

.

28

Freiburg

,

1971 , 1979 1982 Liber sacramentorum gellonensis 2 vols. Corpus christianorum, Series latina, nos. 159 159A. Tournai 1981 Froger Jacques The critical edition of the Roman Gradual by the monks of Solesmes .” Journal of the Plainsong and Mediaeval Music Society 1 ( 1978 ): 81 97 Le fragment de Lucques (fin du Vile siècle) Études grégoriennes 18 nos.

.

,

.

,

.

. _____

,

.

,

“

.

,

-

.

“

. _____

.

( i979): 145

-

55

.

”

“

Le lieu de destination et de provenance du ‘Compendiense.’ In Ut concordet voci: Festschrift Eugène Cardine ed. Johannes Berchmans Göschl 338 53 St. Ottilien 1980

. _____

mens

,

-

,

.

,

.

Gajard Joseph ,

( 1959):

7 26

“

‘Vieux-romain’

.

.

,

1968

Léon

,

3

.

Gamber Klaus Codices Gautier 1886

”

‘Grégorien.’ Études grégoriennes

et

-

latini

liturgici

antiquiores

2d ed.

.

Freiburg

Switzerland,

,

.

Histoire de la

.

poésie liturgique

au

moyen age: Les

Tropes

.

Paris

,

.

Gay Jules L'Italie méridionale et I’empire byzantin Bibliotheque françaises d’Athénes et de Rome, no. 90 Paris 1904 .

,

.

.

des écoles

.

,

Graduale 1952 Graduate sacrosanctae romanae ecclesiae Tournai 1952. Graduale 1957 Le Graduel Romain, Édition critique par les moines de Solesmes. Vol 2, Les sources Solesmes [1957]. Graduale triplex Solesmes 1979 Gushee Lawrence A. The Musica disciplina of Aurelian of Réôme Ph.D. diss., Yale Univ. 1963 ed. Aureliani reomensis musica disciplina Corpus scriptorum de musica, no. 21 [ Rome], 1975 Hesbert René-Jean La messe ‘Omnes gentes’ du VIHe dimanche aprés la Pentecôte et l’Antiphonale missarum romain .” Revue grégorienne 17 ( 1932 ): 81 89 170 79 ; 18 ( 1933 ): 1 14 ed. Antiphonale missarum sextuplex Brussels 1935 La tradition Bénéventaine .” Paléographic musicale 14 ( 1936 ): 60 465 Un antique offertoire de la Pentecôte: ‘Factus est repente.’ In voces: Festschrift van Smits Amsterdam Organicae Joseph Waesberghe 59 69 .

.

,

.

.

,

.

.

,

“

.

,

.

,

.

_____,

.

.

“

.

,

-

—

-

.

,

.

,

_____

.

,

“

-

. _____

.

”

“

. _____

-

.

,

I963

,

-

“

David Notation ,” § III 1 2 New Grove Dictionary London 1980 Hucke Helmut Die Einführung des gregorianische Gesangs im Frankenreich .” Romische Quartalschrift 44 ( 1954): 172 87 Toward a New Historical View of Gregorian Chant .” This

Hiley

—

.

,

,

.

.

.

,

“

.

,

-

.

“

_____ .

JOURNAL 33 ( 1980 ): 437 67 (1980a). Gregorian and Old Roman Chant .” New Grove Dictionary London 1980 (1980b). Huglo Michel Le chant ‘vieux-romain’: liste des manuscrits et témoins -

“

.

_____ .

,

“

.

,

indirects .” Sacris erudiri 6 ( 1954): 96 123 Les tonaires: inventaire, analyse, comparison Paris 1971 Altgallikanische Liturgie ”; Romisch-fränkische Liturgie .” In Geschichte der katholischen Kirchenmusik ed. Karl-Gustav Fellerer 219 44 Kassel 1972 De monodiska handskrifternas fordelning i två grupper öst och väst .” Helsingin yliopiston käytännöllisen teologian laitos: Käytännöllisen -

.

. _____

.

.

,

“

_____ .

-

,

.

,

.

,

“

_____ .

teologian julkaisuja 3 ( 1975 ): 47 65 On the origins of the troper-proser .” Journal of the Plainsong and Medieval Music Society 2 ( 1979): 11 18 La tradition musicale aquitaine. Répertoire et notation .” In Liturgie et musique (IXe-XIVe s.) Cahiers de Fanjeaux, no. 17 Toulouse 1982 Husman Heinrich Tropen- und Sequenzenhandschriften. Ser. B, vol. 5:1 of Répertoire international des sources musicales Munich 1964 Jeffery Peter An Early Cantatorium Fragment Related to MS. Laon 239 .” Scriptorium 36 ( 1982 ): 245 52 -

.

“

_____ .

-

.

“

_____ .

.

.

.

“

,

,

.

,

.

-

.

,

.

.

“ . _____

MSS

The Oldest Sources of the Graduate: A Preliminary Checklist of Copied before about 900 AD .” Journal of Musicology 2 ( 1983 ):

316 21 Lagerlöf Martha Rossholm A Book of Songs Placed upon the Altar of the Saviour Giving Praise to the Virgin Mary and Homage to the Emperor .” In Research on Tropes (Konferenser 8. Kungl. Vitterhets Historic och Antikvitets -

.

“

.

,

Akademien) ed. Gunilla Iverson 125 78 Stockholm 1983 -

,

.

,

,

.

“

Kenneth The Byzantine Sanctus and its Modal Tradition in East and West .” Annales musicologiques 6 ( 1963 ): 7 67 The Italian Neophytes’ Chants This Journal 23 ( 1970): 181 227 Rome, and the Legacy of Gaul .” Early Music History 4 _____ . Toledo,

Levy

.

,

-

.

“

_____ .

-

.

.

“

( 1984 ): Lipphardt 1965

-

49 99 ,

.

Walther Der .

Tonar

karolingische

von

Metz

Münster

.

,

Westfalen

,

.

Lowe Elias ,

Avery

.

Codices latini antiquiores ed. Le Codex 239 de la

Mocquereau André ,

,

.

vols. Oxford 1934 66

11

-

,

.

Bibliotbeque de Laon. Paléographie

musicale 10 ( 1909 12 ). ed. Le Codex 47 de la Bibliotbéque de Chartres. Paleographie musicale ( 1912 14). ed. Worcester antiphoner. Paléographie musicale 12 ( 1922 25 ). ed. Cantatorium de Saint-Gall. Paléographie Musicale ser. 2, vol. -

_____ ,

11

-

-

_____ ,

_____ ,

2

( 1924 ). Patzelt Erna Die karolingische Renaissance Graz 1965 Rehle Sieghild ed. Missale beneventanum von Canosa Textus .

.

,

,

liturgici, Siffrin

.

,

.

,

no.

Petrus

,

9 “

.

.

Regensburg

,

1972

Eine Schwesterhandschrift des Graduale

Ephemerides liturgicae 64 ( 1950):

patristici

et

.

von

Monza .”

53 80 Stablein Bruno Monumenta monodica medii aevi. Vol. 2, Die Gesänge des altrömische Graduale Kassel 1979 Treitler Leo Oral, Written, and Literate Process in the Transmission of Medieval Music .” Speculum 56 ( 1981 ): 471 91 .“ Reading and Singing: On the Genesis of Occidental Music-Writing Early Music History 4 ( 1984 ): 135 208 Vogel Cyrille La reforme cultuelle sous Pépin le Bref et sous Charlemagne Graz -

.

.

,

.

,

.

“

.

,

-

.

."

_____

-

.

.

,

.

,

1965 (1965a). “

La reforme liturgique sous Charlemagne .” In Karl der Grosse: Lebenswerk undNachleben ed. Wolfgang Braunfels 2 : 217 32 Düsseldorf

. _____

-

.

,

,

,

1965 (1965b).

Werminghoff

,

Albert ed. Concilia aevi karolini vol. 3. Hannover 1906 ,

,

germaniae historica, Legum

,

1,

part

1.

Monumenta

.

ABSTRACT

Argues that the “Gregorian” repertory of Mass propers was fully neumed under Charlemagne, a century sooner than is generally supposed. The chief witness is an “apocryphal” Offertory, Factus est repente for Pentecost. Affected are widely-held views concerning: (1) the origin of neumes; (2) the impact of

oral-improvisational techniques on Gregorian chant; and (3) relationship of Gregorian and Old Roman chant styles.

the

origin

and

[10]

Communications Leo Treitler

To the Editor

of the JOURNAL:

In his imposing paper “Charlemagne’s Archetype of Gregorian Chant” Kenneth Levy has addressed the very big question about the beginning of music writing in the Western tradition (this JOURNAL 40 [1987]: 1-30). I would like to offer your readers a commentary about Levy’s conclusion, but even more about the nature of his argument and the principles on which it rests, for these raise issues that are profounder and broader than the conclusion itself. First to the conclusion: “The ‘Gregorian' repertory of Mass propers was fully neumed under Charlemagne [ca. 800 is Levy’s approximate date], a century sooner than is generally supposed” (p. 29). As I am among those who have supposed the later date I must briefly say why. It comes down to two sorts of reasons. (The evidence and reasoning are presented in my paper “Reading and Singing: On the Genesis of Occidental Music Writing,” in Early Music History 4 [1984], a paper that Levy cites, but without taking any account of its contents one way or the other.) First, it looks to me as if the early music-writing technology itself is so far dependent, in semiotic principle and paleographic form, on the Carolingian system of punctuation, that the institution of the latter in centers of Carolingian writing toward the end of the 8th century would mark some sort of terminus ante quem non for the invention of the neumes. And it also seems to me that the Mass proper chants, executed by professional cantors and choirs, were probably neumated in a time later than the writing down of the tones given to priests and deacons. To put it simply, if, as I believe, the neumes have been adapted from the system of punctuation that we know from about 780, and if the chants for the cantor and choir were not written down immediately upon the invention of the notational system, it is hard to see how a fully neumated Mass proper would have been in circulation as early as 800. Still, it is clear that Charlemagne and company were pressing very hard to get things of all kinds written down, and it may well be that after the rapid development of the punctuation system and the Carolingian Minuscule hands attention fell to the music-writing system, which, too, could have been accomplished in short order. The second sort of reason for my supposition of the later date is a whole series of indications that the performance of the Mass propers during the 9th century was based on an oral tradition, and that these indications are consistent with the fact that neumated Mass propers have not survived from that century. But there would be nothing so strange about an oral performance practice that continued after people began to write things down— indeed that is the most likely case. Oral and written transmission are not the

DOI: 10.4324/9781315090283-13

Oral and Written Transmission in Chant

two poles of a binary opposition, as Levy seems to believe. From the point of view of what I find historically most interesting in all of this—the evolution of a literate musical culture in the Middle Ages—Levy’s earlier date would not be so very unsettling, even if it does make some of the evidence to which I have referred look strange. Perhaps we could agree to split the difference by adopting Michel Huglo’s suggestion of a terminal date of mid-ninth century (Levy, p. 8). Now Levy writes as though the case for the later date had rested mainly on the lack of surviving sources from an earlier time, and he charges its proponents with failing to consider “the probabilities of survival” (pp. 4-7). There are two things to be said about this charge: (1) it isn’t so, and (2) it’s quite strange that Levy would make it, considering the heavy dependence of his reasoning on the assumption that the pattern of surviving sources is decisive. One section (VI) of my paper of 1984 is addressed to the question whether we should presume that the earliest notational specimens are later survivals and that substantially earlier specimens have all been lost. I cited a number of “positive, direct reasons for accepting that the earliest specimens of music-writing represent more-or-less the beginning of the practice, and they reinforce the negative evidence of the absence of earlier sources” (p. 195). There is similar reasoning in my paper “Oral, Written, and Literate Process in the Transmission of Medieval Music” (Speculum 56 [1981]: 474-75), a paper which Levy also cites without addressing its contents. In 1980 Hucke made clear that the surviving sources define the historical task in a way that does not depend on the “probabilities of survival.” “Even if one wishes, despite this evidence [the evidence of a performance practice in the ninth century that did not depend on notated books], to suppose that there were in certain localities chant books with neumes as early as the ninth century, chant books without neumes were written at least until the tenth century. We must be able to explain the beginning of chant transmission in the Frankish Empire without assuming the use of neumes. the propagation of Gregorian chant in the Empire and the distribution of chant books with neumes are not the same phenomenon: they represent two different stages in the spread of the chant.” (“Toward a New Historical View of Gregorian Chant,” this JOURNAL 33 [1980]: 437-67; Levy cites this paper, too, but again without taking any account of what is in it). .

It

.

reasoning that depends on the presumption that the in the most precise way. In most cases the thought of a single lost deviant source collapses the argument. Two kinds of reasoning are affected: (1) A terminal date for the transmission of a tradition from one place to another is posited on the basis of the absence of an item from the surviving sources of either the place of origin or the place of destination. This is the argument of Levy’s fourth “index,” regarding the missa graeca, compiled 797-814 according to Levy (pp.8-9; the latest reckoning of Charles Atkinson in a contribution to the forthcoming Festschrift for Helmut Hucke gives the date 827-835). The missa graeca is nearly everywhere, but not in Benevento. So Levy concludes that the Gregorian tradition was transmitted to Benevento before its compilation. But the oldest surviving notated sources from Benevento are dated to the mid-11th century. Levy’s “index” depends is in

surviving

fact

.

Levy's

sources are

representative

Oral and Written Transmission in Chant

on the assumption that not a single Beneventan source containing the missa graeca and written during the intervening two centuries has been lost. This sort of reasoning is entailed also in the seventh “index,” regarding the offertory Factus est repente (pp. 11 ff.). The text of this chant appears (“hanging on,” he writes, “barely surviving”) in only one of the six early sources of Hesbert’s Antiphonale missarum sextuplex, and is thus presumed to have been “obsolescent” in the North by ca. 800. And as it appears in several Beneventan sources it is presumed to have been transmitted there by 800— since it was no longer in use in the North after that time. This part of the argument depends, again, on the presumption that not a single northern source containing Factus and written between ca. 800 and the mid-eleventh century has been lost. (In fact Levy has turned up one such source, the Prüm gradual-troper, written ca. 1000. [He writes “in recent inventories it has been curiously slighted,” thereby slighting the most important recent inventory, the Corpus troporum, where it is duly reported.] But he introduces this source only after “establishing” a terminus of ca. 800 for the transmission of Factus to Benevento, using it only for his further arguments that it was transmitted as part of a fully neumed gradual, but not allowing it to re-enter as counter-example for his reasoning for the terminus of ca. 800. I remain mystified by this strategy. Nor does it allay the mystery that Levy characterizes the Prüm entry of Factus as a “musical archaism;” that just closes the circle on the reasoning.) (2) The one principle that is invoked most often in Levy’s paper is the law that detailed agreement among the written sources of a text points to a common textual archetype. I count seven invocations of this law, all quite automatic and unquestioning, and all but the first referring to claimed agreement of neumation as a sign of a neumated archetype. I shall not comment at all on the intangible generality of such a formulation as “the neumatic details of the Beneventan readings agree with those of northern Europe” (p. 8, my italics; this is said in reference to the whole Gregorian repertory), and I shall put on the shelf for a moment the question whether this law can have the same force in the domain of musical texts as it can in the domain of language texts, from which it has been unquestioningly taken over. For now I simply want to call attention to the fact that each such claim depends first of all on the assumption that deviant sources have not been lost. Having made this observation with respect to both types of argument, I want to make clear that I don’t blame Levy for operating on the—always provisional—hypothesis that the patterns displayed by surviving sources represent the original situation. The alternative would be paralysis by a sort of demon theory that nothing is really as it appears. Now I would like to turn to some of the issues that I think are of broader importance, and first of all to take off the shelf the law connecting parallel neumations with the inference of a common archetype. This law has implications for the way that we think about how a musical text is generated, and it is those implications that I think need questioning. For Levy this law is an a priori idea in both of its aspects—that is, in the model of transmission according to which all sources are ultimately the progeny of an archetype, and in the idea that parallel neumation points to the archetype. It can have the force of law only if there are not plausible alternatives. I want to suggest that there is a plausible alternative—an alternative that opens onto a

different, and I would like to say richer, way of thinking about what a musical text is in relation to what it denotes, and about the various ways in which musical texts might have been generated. I shall approach the question by way of an imagined parallel in the transmission of a language text. Let us imagine a number of sources for an early medieval Latin text, all identical in lexical content, and all written out in a manner that was called per cola et commata (phrases comprising sense units are written in separate lines, as a way of guiding the reading-out of the text, exactly in parallel with the purpose of punctuation at the time). And let us imagine that the lineation in all sources is more or less identical. That can be explained by the inference of a common archetype for all sources, but it can also be explained on the hypothesis that the scribes understood the sense of the contents in the same way and translated their understanding into lineations that are more or less identical. And, last-not-least, it can be explained by any number of combinations of these two models. Given the range of possibilities, we would not be entitled automatically to infer an archetypical lineation and take that inference as fact, on which we would then build chronological structures. The neumation of a melody has this in common with the lineation of a language text: both correspond to immanent properties of the thing denoted that would be projected in a performance of it and of which the scribe or notator might take cognizance while writing, and by the same token two scribes or notators might take cognizance of those properties independently of one another. For the language text that property is the sense-grouping of the words. For the melody text that property is the sequence of short figures through which the melody flows, the little turns and directional units in which the melody recognizably articulates (Floskeln is the German word, neumae has something of that idea in the Latin). Neumatic writing is a graphic transcription of that articulation—indeed in its earliest history it is far better understood in that sense than as a denotation of a pitch sequence. I say that the scribe or notator might take cognizance of such properties in writing, and that is also to say that he might not, that he might copy mechanically. To me it is both intuitively and evidentially clear that writing down, especially in the early times that we are talking about here, can have been a mix of copying and putting down what was in the scribe’s head. Levy’s law premises a model in which there was one transcription from the oral-aural state, and from then on it was transmission by copying through closed written channels. That is an a priori idea, and I do not see that we are bound to assume it for all musical genres and circumstances and times. That a melody with well-defined contours would be given parallel neumations by different notators can be understood without requiring the inference of a single textual archetype. I can think of no better example than Levy’s Figures 4 and 5. If your readers will return to those I think they will see that the neumation projects the constituent units of the melody, and that a widely variant neumation would be highly unlikely, given the way that this notational system was adapted to its function. (Karlheinz Schlager did once give persuasive evidence of the stable reflection of the articulation of an offertory melisma through its neumation and the adaptation of several different prosula texts to it. But there is nothing surprising about that. [See “Die Neumenschrift im Lichte der Melismentextierung,” Archiv für

Musikwissenschaft 38 (1981): 300-16.]) All this is no more than to sound a Levy’s claim that “there is no good way to explain the exact correspondence in Prüm and Benevento” other than by way of a notated archetype. In addition to this matter of parallel grouping of notes, Levy singles out the parallel use of the quilisma and oriscus in both versions of the melody shown in his Figures 4 and 5 as a particularly strong indication for a notated archetype, but I think that it is a particularly strong indication for the possibility of independent neumation. We have ninth-century characterizations (from Aurelian and Hucbald) of the performative denotation of the quilisma: a figure sung “with tremulous voice.” If that performative aspect was a constituent of the melody as it was known, then of course it would have been written with the figure that the notational system provided for denoting that aspect, whether by way of copying or not. (There is discussion of this caveat about

in my paper of 1984.) As for the oriscus, its use in the denotation of a cadential figure like the one in Levy’s Figure 5 is rule-bound: for a note preceded by the same pitch and followed by a descent on the next syllable. A notator who knew the melody and knew the rule would write oriscus for that melodic situation. It is strange to claim that its use in the same position in parallel transmissions of the same melodic passage can be explained only by the inference of a written archetype. Levy writes as though neumes were mute forms that had no relation to the sense or the qualities of what they denote, and that were put down only in acts of mechanical copying, not in reflective acts. There are two conclusions: (1) Since there is a plausible alternative explanation for parallel neumations to Levy’s law of archetypes, that law cannot be automatically invoked, and every one of Levy’s “indices” that depends on it is questionable for that reason alone. (2) It follows that when there is talk of the transmission of a melody from one tradition to another during the period from which we have no written sources, we have to allow for the possibility that the transmission was oral and that there was an independent neumation in the host tradition afterwards. That the neumations in the two traditions are parallel does not gainsay that possibility. It is even possible that an initial neumation in Benevento, now lost, approximated the tradition of the Prüm version less closely, and that the surviving written tradition of Benevento is based on a northern exemplar that was introduced in some later time. It is just because of the low probabilities of manuscript survival that we are all whistling in the dark here, and no one is in a position to claim that he has got the tune exactly right. Levy’s attitude toward neumes as inert paleographic items informs the reasoning behind the “first index,” which he identifies as a “matter of paleographic common sense” (p. 7). What it is that is so characterized is the conclusion that “a common neumatic archetype lies behind the diverse regional manifestations” of neumatic writing that are already apparent in the earliest surviving notated Gregorian propers ca. 900, and that “a period of development lay between the neumed archetype and its first preserved descendents. Allowing for paleographic change, one should suppose at least an intervening half-century, and perhaps much longer.” What is it that is actually meant with the phrase “paleographic change” that would have required at least fifty years? Levy doesn’t say, but I think we need to try to

imagine it, not just let it slip by as an a priori idea. Again, I shall concretize the question with a single example. There has been some idea in the literature that the prototypical neumatic script may have been the one called “Paleofrankish” (e.g. Jacques Handschin, “Eine alte Neumenschrift,” Acta musicologica 25 [1953]). Without necessarily endorsing that hypothesis, let us consider a character in that script, the porrectus: a. V And let us compare it to its counterpart in another early script, that of St. Gall: b.N The difference here is representative of the “diverse regional manifestations" of neumatic writing in its earliest known forms, and so it can serve us in the effort to imagine what can be entailed in “paleographic change” or “development.” Handschin saw how fundamental this difference is in that he saw it as a difference in the way that the respective scripts represent melodic He characterized the Paleofrankish script as not in their forms. just figures, a “Tonortschrift” because, as in figure a. above, the sign traces a movement from one position to another in a space that corresponds to the tone space. The porrectus of the St. Gall script, by contrast, is based on the opposition of virga and punctum as symbols of higher and lower notes, respectively. It is in effect a compound of virga-punctum-virga. (Handschin did not put it in quite this way. But he did see that any understanding of the historical relationships among different neumatic script types presupposes an understanding of the principles on which they function in representing melody. See my paper “The Early History of Music Writing in the West,” this JOURNAL 35 [1982]: 263.) For our purposes here it doesn’t matter which of these forms is the earlier If we want to believe that they both derive from a script archetype, we have to be able to imagine a change from one to the other, and that means not just in the sense of a change of forms, but a change from one idea about how neumes represent to the other. Now I cannot escape the impression from the way Levy writes that he has in mind an evolutionary change that amounts to a gradual transformation: figure a. gradually develops an up-stroke at the beginning, or figure b. gradually loses its upstroke (something like Dom Ferretti’s representation of the punctum from the grave accent: n —.♦■ [Paleographie musicale 13, 65]). Such a transformation could take fifty years or But what seems just as likely to me, given the fundamental difference longer. in functional principle, is that someone says, “no, let’s not represent musical figures that way, let’s do it this way,” or that two different starts were made, perhaps one triggered by the other, but based on different principles. Either way the time interval could be fifty years, but also five years or five months. Levy’s “paleographic common sense” comes down to a belief in paleographic development as a gradual transformation of paleographic forms. That is hardly the only view we can take of paleographic change, and it is certainly not privileged as “common sense.” It does not, therefore, constitute very solid ground for establishing chronologies. There is a similar conception of change in straight lines and in one direction behind the liturgical arguments as well, and in particular it is a supporting thread in the complex web of the “seventh index” (pp. 11 ff.). There Levy proceeds from Michel Huglo’s identification of the offertory Factus est repente as Gallican, taking that as evidence of its antiquity (Huglo 1972 in Levy’s bibliography, p. 226). Putting that together with the solo one.

appearance of Factus in but

one of the Sextuplex sources (as a “survivor,” of two offertories provided for Pentecost, as the second on”), “barely hanging

he concludes it was “obsolescent” in the North by ca. 800. On that ground he regards the Canosa gradual (mid-11th century) as representing an “earlier” state of the Gregorian gradual than any of the Sextuplex sources. And in reporting the presence of Factus in the Prüm gradual he regards it as an “archaism.” A simpler, more strightforward report would have it that the Sextuplex sources and the Canosa and Prüm graduals all represent the liturgical conditions in the time and place in and for which each was written, and that they incorporate elements of greater and lesser age. Somewhere in the background of Levy’s particular type of interpretation is a normative idea of liturgical progression, and when particular items are not in step with that progression they are explained with such expressions as “archaic,” “earlier states,” etc. This is no less dubious than those familiar accounts of music in the 18th century in which one trait or composition or composer points to the future whereas another is retrospective. The pluralist character of the historical situation is ironed out in the interest of a smooth, rectilinear historical narrative. In the same publication in which Huglo identified Factus as Gallican, he also identified the offertory Elegerunt apostoli as Gallican, and on the same grounds (p. 226). This offertory, too, is entered in only one of the Sextuplex sources (Senlis) where it is assigned to St. Stephen. On Levy’s reasoning those two facts would mark it, too, as a survivor, barely hanging on. But we find it still as the offertory for St. Stephan in the Graduale romanum (1938). We could regard it as an archaism there. But Huglo gives the simpler explanation that it has replaced the Gregorian offertory In virtute, which is everywhere else in the Sextuplex sources. That suggests that the exchange between the Gallican and Gregorian traditions—at least as far as the offertories are concerned—could be a two-way affair. The principle about the displacement of one liturgy (e.g., the Gallican) by another (e.g., the Gregorian) is a generalization and approximation of what happened in the long run, a kind of sketch. It does not provide so detailed a picture of the liturgical situation at any particular time that inferences can be made from it about the chronology of individual items. (The Elegerunt story is not reported in Levy’s paper.) A similar image of historical change underlies the discussion about “the probabilities of survival,” with particular reference to the history of notation: the neumed antiphoners were rendered obsolete by notational innovations of the tenth and eleventh centuries. The emergence of staff lines and clefs meant that new books were substituted, and the older ones with prediastematic neumes had little further purpose” (p. 6). But now and then one encounters a manuscript that shows that the story of notation cannot be told in a narrative of such a straightforward progressive form. I’ll just mention one that I happen to have run across recently: Munich, Bayerische Staatsbibliothek, Clm 17025, a calendar, gradual, proser and sacramentary from Scheftlarn, written in the 13th century. The notation is reported in The New Grove as “small quadratic notation on four red lines” (17:629). But there are pages where the notator changes from that notation to non-diastematic neumes, perhaps to save space, and possibly to record more familiar matter such as hymns (folios 5, 44, 76v, 101v). We encounter something similar in .

".

.

both continental and British manuscripts with organum, in which the new organal voices are recorded in a more nearly diastematic notation than the old chants. In such situations different modes of notation, functioning on different principles of representation, were simultaneously available and could be applied to needs that differed according to the type of melody that was to be represented and the different competencies of the singers. That suggests a somewhat more complex picture than that of successive notational types strung out on a thin time-line. What becomes more and more evident through this review is how Levy’s hypothesis rests, not so much on his “witnesses,” as it does on these few historiological principles that are the basis for the selection and interpretation of evidence. The paradox is that the more we unravel the intimidatingly complex surface of the argument, the more drastically oversimplified is the representation of the texture of the historical world that comes into view. How it all works is vividly illustrated by the interpretation of Aurelian, Levy’s fifth “index” (pp. 9-10) and the last of the ones on which I shall comment here. Levy offers a series of interpretations of remarks by this enigmatic author, without ever telling us why these should be privileged above other possible interpretations: In referring his “specialist readers” [more likely, his listeners] to particular chants Aurelian “may have expected [them] to have neumed antiphoners for consultation.” “Aurelian makes detached melodic comparisons, singling out individual syllables of particular chants... with such directions as ‘on the fifteenth syllable of .’ This too indicates neumation.” “The survey of the verse repertory that Aurelian describes is unlikely to have been a scroll through a memory bank but a point to point comparison of neumed chants in a reference antiphoner.” “When Aurelian opts to preserve [a] musical anomaly ‘because it was used among the ancients’ he is likely to have found the former tradition in a noted antiphoner.” “The reader is told of a musical passage in two Gregorian gradual verses that is ‘not found elsewhere in the prolixity of the whole antiphoner.’ One must again conclude that Aurelian is not referring to a singer’s well stocked memory but to the ‘prolixity’ of a fully neumed .

.

.

.

.

antiphoner [my italics].”

Each of these passages can be understood just as well in the sense that Aurelian did indeed expect his audience to know from memory the verse repertory that he described, especially in the light of two passages of the same treatise that Levy does not cite: “Although anyone may be called by the name of singer, nevertheless he cannot be perfect unless he has implanted by memory in the sheath of his heart the melody of all the verses through all the modes, and all the differences both of the modes and of the verses of the antiphons, introits, and of responses.” (The passage, beginning ‘Porro autem’ and ending ‘in teca cordis memoriter insitum habuerit’ can be found in Gushee’s edition, p. 118; see Levy’s bibliography for details.) The second passage harks back to Isidore of Seville: ‘The muses, from whom [music] took its name and by whom it was reported to have been discovered, were declared to be the daughters of Jupiter and were said to minister to the memory, for this art, unless it is impressed on the memory, is not retained. non retineatur’; Gushee, p. 61. Both passages (‘Dicebatur autem musae are cited in my 1984 paper, p. 161, where the Latin can also be found.) I wouldn’t read these two passages against the possibility of Levy’s hypothesis .

.

.

of a date of ca. 800 for a notated archetype, for surely there was sometime in the early history of all this when notated books existed but practice relied mainly on oral tradition or memory (those are not different); it is the allowance for pluralistic situations of just this sort that is missing from Levy’s paper. But those passages do speak directly against Levy’s forced use of Aurelian as a witness for the necessity of his hypothesis (“One must again conclude...”). Indeed, only a prior conviction about the conclusion could produce these selections and interpretations of evidence. That goes for the paper as a whole. But there is something more about this circularity. The movement of music through tight scriptual channels defines a condition of musical transmission at one extreme through the history of the high art-music of the West. That condition sets the frame in which this story has been thought out, as is apparent even in the choice of language: writing of a transmission as a “recension” (p. 15); reporting when a manuscript was compiled by saying it was “copied around 1000” (p. 17); reporting differences in the neumation of an item as “replacement,” “resolution,” “translation” (p. 20), always without any effort to back up the special connotations of those words. But the story did not unfold entirely in that frame. Its unfolding is itself the beginning of the erection of that frame, but the reasoning is as though the frame were eternal and universal. David Hughes ("Evidence for the Traditional View of the Transmission of Gregorian Chant,” this JOURNAL 40 (1987): 377-404) allows that “there was perhaps a time at which this was the state of affairs,” referring to his formulation of what amounts to an oral component in the transmission of chant (p. 377). 1 But he writes that that time belongs to a ’’probably irrecoverable phase of chant history.“ I cannot agree with that, and not only because of my belief that a stable written tradition is not necessarily incompatible with a continuing measure of orality in the practice. Even granting that it were, the oral origin is visible through the written surfaces I

Hughes has misrepresented my position in his lead paragraph, especially in the belief that “each performance of a specific chant was in part an and hence to at least some extent different from all others [my italics].” I do not wish to burden your readers now with an extensive commentary on this paper as well. But I do beg to remind them, and Professor Hughes, of a conclusion in my first publication on this subject:". by the time of writing the melodies were being transmitted as individual melodies, not as concrete instances of melodic types; that a degree of standardization and individuation of the repertory had taken place; that the model for each performance was a particular plainchant, not the principles of a melodic type—something more nearly like the ordinary notion of memorization than oral composition. This would be to suggest, finally, that the tradition of oral (“Homer and Gregory: The composition had declined by the time of writing down Transmission of Epic Poetry and Plainchant,” Musical Quarterly 50 (1974): 333-72; the passage cited is on pp. 367—68.) This would, of couse, be entirely consistent with the results that Hughes has presented. And I would like to recall also a passage on p. 346 of the same paper: "... we require an understanding of oral transmission as a normal practice whose object and effect is to preserve traditions, not play loose with them.” With all due respect to Giulio Cattin, his dictum “Where oral tradition prevails, no performer need feel obliged to repeat the same song identically at each performance" cannot be read as authoritative (Hughes, note 34). Hughes has read into my work on this subject an idea that all chant performance during the time spanned by his manuscripts was at least partially improvisatory, and to that extent it was casual and unstable. This misreading on both scores has, I regret to say, constituted a disruptive noise in the channels of communication. 1

am

afraid that

attributing to improvisation, .

me

.

.

.

.

that are its progeny, no matter how many generations of sources and whatever measure of redaction separate them. That is as surely so for the plainchant traditions as it is that the oral-compositional nature of a blues performance would show through generations of pressings of a recording, or of editions of a transcription of it. How the written transmission was related to the performance practice in the communities for which it was written is a separate question that has still to be pursued. But it will surely not yield a uniform answer for all traditions and all communities. The theory of oral composition with respect to plainchant and related traditions was not ever intended to account for variation in transmission as its primary task; variation was but one clue, and not the principal one. The primary task has been to show that the generative systems of the oral tradition—which were carried into the era of composition with the aid of writing—informed the music that was produced, and that they are therefore relevant to the understanding of that music and of the history of its transmission. No matter how uniform the written transmission is, and even if we regard it exclusively as a product of copying faithfully from one source to another, it is transparent to the oral tradition that was its ultimate source. What is striking is the resistance to a serious confrontation with the reality that at some time in history, no matter how far back one wants to push it, the Western musical heritage goes back to an oral tradition that left its mark and that is never entirely out of the picture as a factor in musical practice. This is something for music historians to deal with, as well as ethnomusicologists. I should like to join my colleagues Professors Hughes and Levy in dedicating these remarks to Michel Huglo on his 65th birthday. LEO TREITLER The Graduate Center

City University

To the Editor

of New York

of the JOURNAL:

IN “CHARLEMAGNE’s ARCHETYPE” I proposed an early date—by ca. 800— for an authoritative full neumation of the Gregorian repertory. That is a century before the recent opinions of Corbin, Hucke, and Treitler have put it; it is earlier than any previous qualified opinion. Specialists have long looked to Charlemagne, and on Charlemagne’s word to Pippin before him, as the movers in stabilizing the Gregorian melodic repertory. But no one thus far has undertaken to demonstrate that the repertory so stabilized was cast in neumes. Professor Treitler finds my early date “unsettling.” The reason, as he makes clear, is that it calls into question certain theories he has been promoting for some time, and much of his communication is given over to the restatement of those theories. There is the theory of “re-improvisation” (to use the term of van der Werf, Emergence of Gregorian Chant [1983]); this supposes that our early noted versions of Gregorian chant do not represent fixed melodic shapes, but instead remained susceptible to improvisatory impulse even into the 11th and 12th centuries. There is the related theory of “frozen improvisation” (which Treitler shares with Hucke: this JOURNAL, 33

which supposes that the early Gregorian neumations are “transparent" windows on prior oral-improvisatory practice. And there is the theory of the origin of neumes, which, in a complement to his view of the Carolingian musicians’ improvisatory stance, he identifies in “the Carolingian system of punctuation” (see his “Reading and Singing”: Early Music History 4, [1984]). Treitler’s principal reproaches to me are that I ignore the substance and consequence of these related theories. The length of my paper limited discussion to the most pertinent matters concerning date, but since the other theories have now come up, let me briefly say why I do not think they are

[1980]),

very good. The theory of “re-improvisation” addresses the nature of the neumed recension. Treitler speaks of “an oral performance practice that continued after people began to write things down.” He interprets variants in Gregorian recensions of the 10th and later centuries as indications that the melos remained open to oral, improvisatory input; he even speaks of “a whole series of indications that the performance of the Mass propers during the 9th century was based on an oral tradition”; these are indications one should want to review. For my part, I see a written archetype that reflected a fixed, concretized melos: a model neumation of the late 8th century that provided a detailed record of a crystallized melodic text that excluded improvisation. The unheighted neumes described certain aspects of the melody in considerable detail; they showed relative lengths and some niceties of delivery, like liquescences, oriscus, and quilisma (to give them their later neume-names). That neumatic specificity, present even in the archaic Paleofrank notation, is a good indication of the fixity of the melos. What the early neumes dealt with least well was pitch. For pitch, the neumes offered only silhouettes of ups, downs, and repetitions; they lacked the particulars on levels and interval widths. Yet of all the melodic factors, pitch was most safely left to the memory that served as backup to the neumes. Between Treitler’s views and mine, the essential difference is the nature of the oral input that supplemented the neumes. Where he sees a continuing license to improvise, I see missing particulars that were supplied from verbatim memory. When neumation began, the Gregorian melos seemed secure in professional memories. But soon the interaction between an incomplete, evolving notational system and an attenuating memory resulted in small changes that appear in 10th through 12th century copies. These are copies removed from my proposed “early” neumed model by at least a century. With "frozen Treitler moves the theory of improvisation," to the terrain of prehistory. He observes that backward “re-improvisation” much of plainchant originated in improvisatory maneuver, and he adds that the preserved Gregorian neumations are good windows on oral-improvisational practice. The first observation is safe. When Wagner (Gregorianische Formenlehre, 1921) and Ferretti (Estetica gregoriana, 1934) characterized florid Gregorian chants like the Tracts as realizations of “psalmodic” frameworks, the improvisatory factor was basic to their conception; when Cardine (Congresso Internazionale di musica sacra, [Rome, 1950], 187-91) declared there was an unwritten stage behind the first neumations, that factor became explicit. Treitler’s statement that “the oral origin is visible through the written surfaces that are its progeny no matter how many generations of sources and whatever measure of redaction separate them”

essentially affirms what has long been evident. It is his second observation that raises an issue. He writes, “No matter how uniform the written tradition is, and even if we regard it exclusively as a product of copying faithfully from one source to another, it is transparent to the oral tradition that was its ultimate source [italics mine].” Actually, how good a window on orality does the Gregorian recension supply? We have the one late, remarkably uniform recension. Is it a “transparent” record of oral deliveries, or has the melos on its way to this ultimate recension undergone enough in the way of written compositional and editorial tinkering to distort our rearward vision? It will take more than the arguments Treitler discovers in comparative literary theory (“Homer and Gregory,” Musical Quarterly 1974) to convince me that the solo deliveries of Balkan epic bards illumine the choral ritual of the Gregorian schola cantorum, or that the Gregorian recension we have stands as a “transparent” mirror of improvisational orality, rather than the endproduct of a considered editorial process. The largest part of Treitler’s communication is taken up with the early history of neumes. Here there are two major questions. One is ultimate origins. The other, what Froger (Le graduel romain, IV, 2 [1962]:92) described as the zone brumeuse: that is, supposing there was an “original,” noted archetype, what kind of neumation could it have that would produce the different neume-species found in Lorraine, St. Gall, Brittany, etc. ca. 900? No one has answered these questions, and Treitler complains in effect that my proposal of the “early” written model does not do so. In “Charlemagne’s Archetype” I remarked that a complementary paper “On the Origin of Neumes” was on the way. It appeared in the same year (Early Music History 7 [1987]: 59-90), offering a review of current theories (among them Treitler’s of 1982 and 1984) and some fresh formulations. My aim was to place “Charlemagne’s Archetype” in its notational-historical context. Concerning the zone brumeuse, there are two proposals. First, the model neumation ca. 800 (Froger’s “original”) had a cursive, “conjunct” ductus (as found in the majority of 10th-11th century copies from all over), not a “disjunct” ductus (as in the neumations of Brittany, Aquitaine, Nonantola, and to some extent Lorraine). Second, contrary to what is generally supposed, the model neumation was not “nuance-rich” (like St. Gall 359, Einsiedeln 121, or Laon 236), but “nuance-poor,” as in the majority of surviving early sources. During the 9th and 10th centuries there were clarifications and supplements (episemata, “Romanian” letters, coupures, etc.) that put onto written record some of the detail that was previously left to memory. This was done differently at different places, and it contributed to the variety of the “nuance-rich” notations. My other concern in that paper was the ultimate origin of neumatic notation. My proposal is that the “original” neumes represented a “graphic” method: they amounted to a plotting of relative pitch positions, along with some details of duration and nuanced delivery. Vestiges of this method survive in the notations called “Paleofrank.” Some time later, in “Charlemagne’s Archetype” and its descendants, the “graphic” was replaced by a second method that I describe as “gestural” or “cheironomic.” It is the same method that has long been suggested as an ultimate “origin” of neumes. I propose it only as a later stage, without connection to origins. However I see the implementation of the “gestural” rationale as a major factor contributing to the differentiation of local neume-species.

In sum, my proposals put chronology, memory, and notational etiology into a single frame where certain of the major problems that have long vexed musical Gregorianists may be resolved. I attribute to the generation approaching 800 an activity that other undertakings of Charlemagne’s circle (the reforms of texts, script, education, liturgy, clerical life) might suggest was theirs: the promulgation of a canonical musical repertory in an authoritative, written, neumatic formulation that was designed to facilitate teaching and performance. Earlier stages in that musical development—and perhaps the notational development itself—may, as Charlemagne says, be laid at the doorstep of his father Pippin. I claim no decisive proofs; my interpretations of date and notational method may turn out wrong, and at the least I expect clarifications and adjustments to emerge. But for the moment I believe I have gathered together the main surviving elements of the puzzle, and for the first time have made them fit into a coherent whole. In Professor Treitler’s communication I can discover nothing that shakes my view. Quite the contrary: he finds my date for the neumed model “very unsettling,” and he proposes that we “split the difference by adopting Michel Huglo’s suggestion of a terminal date of mid-ninth century.” That of course brings it near the likely date of Aurelian’s treatise—the 840s. But I think that is not enough. I think my model of an early neumatic promulgation (by ca. 800) of a musically concrete, editorially processed, stabilized melos stands much closer to the reality of the Gregorian musical situation than the late neumation and “frozen-improvisational,” then “re-improvisational,” model that Treitler has endeavored to defend. KENNETH LEVY Princeton

To the Editor

University

of the JOURNAL:

LET ME TAKE care of one small matter first. I cited Cattin’s description of the troubadour-trouvère oral tradition only with reference to that repertoire, and that is made clear in my paper. Central to my thesis is the solidity of the Gregorian tradition (or traditions, if we want to separate the oral from the written in this context): the tradition of secular music is quite different, and I referred to it only by way of contrast. As to misrepresentation, I am inclined to plead slightly guilty with extenuating circumstances: to sum up the complex work of Professors Treitler and Hucke in a sentence or two is no easy matter. I tried to give both authors full credit for diversity and subtlety; but in Professor Treitler’s “Homer and Gregory,” surely the whole issue revolves around the use of formula—and hence around the question of oral composition. “[The singer] will have planned before beginning quickly thinking what should come next as he scans the next phrase of text” (pp. 346-47); “I believe that such sports [standard formulas in unusual positions] complement the more consistent transmission of melody types in supporting the view of chant performance before the age of notation as a process of construction or .” (p. 367). Our difference in this respect lies in our reconstruction .

.

.

.

.

different inferences from the sources as to what had been the status earlier. That freedom declined as writing grew more important would be normal enough. I do not believe that I have ever referred to any aspect of chant performance (or transmission) as “casual and unstable.” For the rest, I am not sure where the disagreement lies, lhat oral and written traditions can and did coexist I have myself argued (“Variants in Antiphon Families: Notation and Tradition,” in La Musique et le rite sacré et profane: Actes du XIIIe Congrès de la Société Internationale de Musicologie, Strasbourg, 29 août—3 septembre 1982, ed. Marc Honegger and Paul Prévost, 2 vols. [Strasbourg, 1986], 2:29-47). That the oral origin of the chant “is visible through the written surfaces that are its progeny” is something with which I would not take, and have not taken issue. And as to resisting the reality of oral tradition, the penultimate paragraph of “Evidence for the Traditional View” sketches a conjectured—but entirely oral—origin for the chant as a whole. There remains "the probably irrecoverable phase of chant history." My contention is that the surviving sources do not supply direct evidence for a transmission other than that hitherto thought of as normal—i.e. one in which younger singers learn the chants either by learning them in rote fashion from an experienced singer, or by reading them from a notated book (the latter necessarily a later, and perhaps overall an exceptional way). If the sources we have are in fact amenable to this sort of normal explanation, then any other sort of transmission must have occurred earlier, and the precise shape of the melodies in that earlier time is hence “probably irrecoverable,” simply because no sources preserving that shape have survived. Of course, other methods of research may effect the recovery, but it is not easy to see what these might be. Since most of the rest of Professor Treitler’s communication is addressed to Professor Levy’s article, I might stop here, but I must add a few words on another subject brought up by Professor Treitler, since it affects my arguments as well as those of Professor Levy. Professor Treitler alleges that the neuming of a melody corresponds to “immanent properties of the things denoted.” It seems to me, rather, that this is something that must be (if possible) established: it is not an axiom. The substantial agreement of the sources in the neuming of melismas may be the result either of the recognition of such “immanent properties,” or (as Professor Levy would have it) of copying from a written exemplar, or conceivably from neither of these causes (e.g. from scribal habits—a process that can conveniently be seen in the Graduale Triplex, where the scribe of Laon clearly favors larger neumeunits than the scribe of Einsiedeln). It is obviously a matter of some importance, and deserves more attention than it has recently received. DAVID G. HUGHES Harvard University

[11] The Debate about the Oral and Written Transmission of Chant

László DOBSZAY

STUDIES by Helmut Hucke and Leo Treitler concerning the oral and written transmission of Gregorian chant,1 followed by responses on the part 3 2 Hughes, Kenneth Levy, Hendrik van der Werf4 and Peter 5 Jeffery as well as by reactions from Treitler,6 have opened a new phase in chant scholarship. In spite of some markedly controversial statements of David

the debate has in many respects succeeded in bringing opinions closer revealing that the disagreements frequently only concerned the

and in

1

H. Hucke, «Toward a New Historical View of Gregorian Chant,» JAMS 33 (1980), 437-467. Idem, «Gregorianische Fragen,» Musikforschung 41 (1988), 304-330. L. Treitler, «Homer and Gregory: The Transmission of Epic Poetry and Plainchant,» MQu 60 (1974), 333-372. Idem, «Centonate Chant: Übles Flickwerk or E pluribus unus?» JAMS 28 (1975), 1-23. Idem, «Oral, Written and Literate Process in the Transmission of Medieval Music,» Speculum 56 (1981), 471-491. Idem, «The Early History of Music Writing in the West,» JAMS 35 (1982), 237-279. Idem, «Orality and Literacy in the Music of the Middle Ages,» Paragon: Bulletin of the Australian and New Zealand Association for Medieval and Renaissance Studies 2 (1984), 143-174. 2 D. G. Hughes, «Evidence for the Traditional View of the Transmission of Gregorian Chant,» JAMS 40 (1987), 377-404. Cf. Communications in JAMS 41 (1988), 578-579 and JAMS 42 (1989), 434-437. 3 K. Levy, «On Gregorian Orality,» JAMS 44 (1990), 185-227. Cf. Communications in JAMS 41 (1988), 575-578 and 44 (1991), 524-525. 4 H. van der Werf, The Emergence of Gregorian Chant. I-II. Rochester, New York, 1983. Cf. Communications in JAMS 42 (1989), 432-434 and 44 (1991), 517-524. 5P. Jeffery, Re-Envisioning Past Musical Cultures Ethnomusicology in the Study of Gregorian Chant, Chicago, London, 1992. 6 Communications in JAMS 41 (1988), 566-575 and 44 (1991), 513-517. -

DOI: 10.4324/9781315090283-14

of agreed facts. After scrutiny of the facts, problems and for solution from several angles, the earlier writings on the possibilities of Gregorian chant now seem a little naive. At the same emergence

interpretation

time, in the light of the debate we are now better able to appreciate earlier initiatives which at the time they expressed were hard to fit into the overall view or appeared to be marginal. As an example I might mention the articles by Benjamin Rajecky written back in the 1960s and '70s which, though taking another field as their point of departure, touched 7 upon the issues now under discussion. The reason why his ideas did not attract the attention of

Gregorian scholars being mostly presented in the literature on

may lie in their such as the Handbuch folk music, in those

days

des Volksliedes in 1975.8 The present paper is not intended to launch sensational discoveries, does it strive to settle disputed questions. All it seeks to achieve is to bring together the problems raised, to point out those areas where the nor

outlines of a new consensus are apparent, and also to indicate any inconsistency or imprecision in either the answers or the questions themselves. Scholars agree that prior to its first recording Gregorian —or let us say pre-Gregorian— chant was handed down by oral transmission. Hucke and Treitler have tried to describe what it might have been like in this orally transmitted state and how the musical material reached the stage of being written down. A statement of secondary importance proved to be

the

challenging, according to which even the written forms should as improvisations (using the term not in the «romantic» sense criticized by Treitler himself but with the meaning he suggested9 ). most

be considered

7 B. Rajeczky, «Népdal-történet és gregorián-kutatás» (=«The Ethnomusicology and Chant Research»), Festschrift für Zoltán Kodály, Budapest, 1943, 308-312. Idem, «Le Chant Grégorien est-il mesuré?,» Études Grégoriennes 8 (1967), 21-28. Idem, «Über die Melodie Nr 773 der Monuments Monodica Medii Aevi I,» Festschrift B. Stäblein, Kassel, 1967, 191-194. Idem, «Zür Frage der Verzierung im Choral,» StudMus 11 (1969), 349-353. Idem, «Gregorián, népének, népdal,» (=«Gregorian Chant, Folk Hymn, Folk Song»), Magyar Zenetörténeti Tanulmányok 1969, 45-64. Idem, «Europäische Volksmusik und Musik des Mittelalters.» StudMus 15 (1973), 201-204. Idem, «Choralforschung und Volksmusik des Mittelarlters?» Acta Musicologica 46 (1974), 181-192. Idem, «Gregorianische Gesänge in der ungarischen Volkstradition,» StudMus 27 (1985), 5-22. Idem. «Ungarn: Choral im Volksmunde.» Festschrift für Ernst Klusen zum 75. Geburtstag, Bonn, 1984, 375-388. Idem, «Gregorian Plainsong and Folksong,» Hungarian Music Quarterly II/2 (1990), 9-16. 8 B. Rajeczky, «Gregorianik und Volksgesang,» Handbuch des Volksliedes. Bd. II.

München, 1975, 393-405. 9 JAMS 44 (1991), 514.

analyses by Hucke and Treitler can perhaps partly be blamed for distracting attention from the main issue (how to imagine the path of development of Gregorian up to the emergence of its written form) and turning it into a secondary issue (whether the music scribe was codifying an inspiration). I propose to raise questions in the inverse direction, starting from the notated codices and heading back towards the beginnings. I shall not, however, touch on the debate about the date of origin and archetype or 10 archetypes of the musical notation. The actual questions are: what the music scribe recorded by this device and to what extent his recorded material is a Vorschrift and to what extent a Nachschrift? Is the substantial identity of the material as written down in the 10th/11th-century sources more important or the differences found therein? An overemphasis on the former gives credence to the notion of a single archetype. The abundance of codices is considered to be the result of a visual copying process, and differences are reckoned to be caused by scribal error. On the other hand, when the variability is accorded greater significance, The

musical and consciousness-related motives in the process of notation assume greater importance. The music scribes seeks to record the form of melody he knows either as the tradition of the community he is familiar with or as his own realization of its possibilities. Each of these processes may have existed, mixed, within one single act of notation. David Hughes stresses the uniform nature of the available melodies and declares the majority of the differences to be trivial. 11 Thus he takes the fixation of the melodies for granted, regardless of whether this fixity was achieved already during the age of oral tradition or only upon the introduction of writing. Nevertheless, the differences between the sources cannot be ignored. They provide abundant examples of tonal rearrangements, changes of final tone, restructuring of the intervallic system even when the tonic remains the same, differences between pentatonic and diatonic dialects, and strings of variants apparent as adjustments of detail but tending collectively to modify the melodic meaning according to a logical musical sense and taste. 12 ( Ex. 1 ).

10K. Levy,«Charlemagne's Archetype of Gregorian Chant,» JAMS 40 (1987), 1-30. 11 JAMS 40 (1987), 380 et sequ. 12 Sources; a) Paris BN cod. 903 (=PalMus XIII), f 252. b) Benevento Bibl. Cap. VI. 34. (=PalMus XV), f 258. c) Missale Notatum Strigoniense (= Musicalia Danubiana 1), f 193v. d) Chartreuse Graduale, Grenoble B. Mun. 84. e) Nevers Graduale, Paris BN n.a.l. 1235, f 115v. f) Quedlinburg Graduale, Berlin Staatsbibl. 40078 f 119v. g) Klosterneuburg Graduale, Graz Univ. Bibl. 807 (= PalMus XIX), f 169v.

Example

1

Though the possibility of copying or of musical error on the part of the scribe cannot be ruled out in every instance, this spectrum of variants must be mote than just infidelity to an archetype, more than a corruption of the tradition. Three additional factors also argue against the notion of corruption. The first is the utmost

insecurity resulting from the non-diastematic the fact that prior to the first notations Roman chant had long been cultivated throughout Europe. We should liturgical remember, for example, that in large areas of what is now Germany church institutions and liturgical singing had become firmly established through the work of Irish missionaries long before the Carolingian notation. The second is

were attempts at centralization from time to time, adherence to local custom must have been at least as strong, particularly if it was not so much a matter of the rite or repertory —so important for

reform. Even if there

the authorities of that age— but of melodic variants of secondary importance. The third point is that the liturgical chant sung in many hundred church institutions all over Europe could not be uniform down to the tiniest detail. We are concerned not only with large and famous monasteries. The emergence of diocesan organization aided the dissemination of liturgical chant over a large number of collegiate and parish churches and schools from the very dawn of the Middle Ages. It would be completely anachronistic to think of ecclesiastical conditions then as resembling those after the later Cistercian or Dominican reforms. The state of the church can only be imagined in a context of, on the one hand, centralistic, authoritative or doctrinaire initiatives, cropping up over and over again, and on the other hand, coexisting with these initiatives, a continuing, living practice actually much richer than anything which could be detailed in reform programmes. In such a situation it is by no means certain that we shall be able to distinguish clearly between the variability which survived the Carolingian reforms and the variability which grew up locally and regionally after the period of reform. Although I do not wish to mix up the nature of events in the 9th century with those in the 12th, I should like to mention a relatively late event which illustrates what I believe must have been typical of the history of music writing. The variant of the Gregorian tradition in Esztergom was first notated on the staff in the middle of the 12th century. 13 The melodic

13 Magyarország Zenetörténete I. Középkor. (The Music History of Hungary. I. Middle Ages.) Ed. by B. Rajeczky. Budapest, 1988, 254-264, 417-420. L. Dobszay, «Plainchant in medieval Hungary,» Journal of the Plainsong & Mediaeval Music Society 13 (1990) 49-78: 50-51, 53-54.

variants, which expressed easily distinguishable from,

a tradition essentially identical with, yet the common European practice, had already

been in use for decades at the time of the exact recording. To record these variants a new manner of notation was developed by revising the earlier adiastematic notation, much practised but inadequate to the new task, along contemporary foreign lines. 14 By means of this notation a material had to be written in staff notation. 15 ( Ex. 2 ).

highly specific melodic

recorded,

nowhere

previously

Example 2

14 Magyarország Zenetörténete I. Középkor, 1988: 174-178, J. Szendrei; «Die Geschichte der Graner Choralnotation,» StudMus 30 (1988), 5-234: 34-48. 15 Sources: a) Klosterneuburg Graduale, Graz Univ. Bibl. 807, f 2v. b) 13th cent. Missale Notatum from Hungary, Güssing, Franciscan Monastery, 1/43, f 16v.

The earlier musical documents of this tradition, without staff notation, had transmitted the outlines of the melodies, but copying from them would hardly have produced a satisfactory result in the new notation. The only source for recording the local variant tradition must have been the living practice of the community itself, while the non-diastematic notations could at best only have been used as a partial control. This

example

draws attention to the fact that the

precise copying

of

a

fixed

repertory and the ad hoc notation of a freely available store of melody represent two extremes, while the actual relationship of the writing and the living practice must be imagined somewhere between them. The main

question is thus not whether the variants retain remnants of improvisation or constitute scribal errors. The explanation of the emergence of Gregorian music does not depend on this interpretation. If, moreover, we consider notation as a frequently accidental or accidentally surviving witness of an actual, rich, colourful, living process, the question of «authenticity» also appears in another light. Let

us

go back in time. However much

we

emphasize

the abundance

of variants in the Gregorian sources, we cannot deny the similarity of items, particularly if they are compared with the corresponding Old Roman items. Because of the essential identity an earlier unifying revision must be taken for granted, the only disputed issue being, at the most, the supposed manner of the revision. It remains to be decided what degree of unity can be imagined and what was the role of writing and of other means toward achieving this unification. The non-diastematic notation is from the beginning unsuited to writing a form of melody fixed from note to note. Nevertheless, if in spite of this we can still

identity between the versions transmitted, we ought orally transmitted cultures did possess the means of recording and memorizing prior to writing. Could such a large corpus of melodies have been retained in the memory? Though the repertory was at the possible time of fixation not yet as large as in the late Middle Ages, it already included groups of items of an individual character or in a transitional style progressing beyond the classical types of melody: I am thinking of such chants as offertories, the cycle of responsories belonging to the scriptura occurrens, and the proprium of many recent

speak

of substantial

to consider whether the

feasts. It is very tempting to digress from the problem by talking about the memory of the «singer», someone with more than average talent and experience, who could recall a melody without the constraint of having to go through it note by note beforehand. There is no denying that solo singing had a more decisive role in early liturgical chant than it has for

example in present Gregorian practice,

and that some genres later sung

the choir had formerly been chanted by soloists. Nevertheless, the greater part of the repertory —particularly the genres containing individually shaped items— had already been sung by the schola, or at least a by group of solo singers, prior the their being put down in writing; their singing would have been disturbed by any substantial variation or improvisational freedom (at least within the given community). Our knowledge of folk music and of Jewish and Eastern Christian liturgical

by

of the illiterate ages had memorizing techniques than we larger have today. Notker’s well-known remark about the difficulty or impossibility of memorizing long melismas 16 actually compels us to draw an

practice justifies

us

in

better memories and

a

assuming

that the

people

arsenal of

inference complementary to the usual

one: a

singer living before Notker's

to be able to memorize the melodies even without the

expected age aid afforded by added texts. In some Eastern rites soloists, schola-like groups, or even entire communities sing together large quantities of text without notation (incidentally with minor variants which do not disturb the overall effect). Christians of the Orthodox rite in Hungary have only recently begun was

17 their melodies. For that matter, the communities or congregations have only the text in their hands (or not even that) while singing melodies in prosaic structure, that is, complicated melodies unfolding in

notating

free forms.

Ethnomusicologists

have attended summer

courses

among

Coptic Christians of Egypt in modern times, where the only task to be mastered was the precise rendition of a huge chant repertory without written music, imprinting it on the mind through continual repetition. It is well known how methodically the ritual melodic material is handed over to each member of the Orthodox Jewish communities, beginning at

the

18

the age of five.

Let

us

consider

a

further aspect of the

matter.

The

majority of scholars

agree that some kind of fixation must have taken place before the first written evidence in the 9th century. Why could we not accept the same premiss for any period of oral tradition? Is it necessary to postulate that the whole repertory had been in a kind of flexible state for three to

16 P.

Wagner, Einführung in die Gregorianischen Melodien. III, Leipzig, 1911, 251-252. Br. Stäblein, Schriftbild der einstimmigen Musik, Leipzig, 1975, 47. 17 Cf. D. Stefanović, «The Phenomenon of Oral Tradition in Transmission of Orthodox Liturgical Chant,» Cantus Planus: Papers read at the Fourth Meeting. Budapest, 1992, 303-306. 18 Report by Ilona Borsai in the Musicological Institut Budapest (1973). Cf. I. Borsai, «Die musikhistorische Bedeutung der orientalischen christlichen Riten,» StudMus 16 (1974), 3-14.

four centuries, then, at a given point in time, became fixed as a whole, finally, still later, put into notation? The alternation between freedom and fixation which is characteristic of the overall history of Latin

and

liturgical chant might well have operated, without the whole repertory necessarily having to fall into one or the other of the opposed categories. If there was a means in the 7th or 8th century of fixing a melody without writing, why not also for a greater or lesser part of the repertory back in the 4th or 5th century? At the beginning of his article David Hughes stated his intention of dealing with Gregorian issues, thus setting aside what could be learned from other repertories. 19 But can they be ignored, if we want to understand Gregorian chant? Irrespective of our position regarding the historical relationship between the Old Roman and the Gregorian repertory, it is striking that such an immense structural similarity could exist between them, notwithstanding the considerable differences. If there be any doubt about the justification of using such modern concepts as ‘underlying structures’ or ‘melodic skeleton', 20 a comparison of these two repertories must surely convince us that the early singers really did have a sense of such structures. Since they are mostly hidden abstractions, which for us have to be revealed by twentieth-century analytical methods, the way in which they must have served as a vital, active force centuries before is indeed a source of wonder. Sed contra factum non est disputatio. Ethnomusicologists have gained insights into the mind of the singer of former times. The Hungarian lament, for instance, demonstrates strikingly the many ways in which an abstract melodic concept be realized, let us say in the manner of an improvisation. 21 (Ex. 3 ). may And it is a gift of fortune if an ethnomusicologist has the luck —as Janka Szendrei once had— to observe a ‘primitive' singer going over in her mind the tonal reference points, basic intervals and structural motives of the improvisation she is about There

are two

to commence. 22

lessons to be learned from this. One

is

that the singer accorded

instinctively capable of such an immerse abstraction should be

19 JAMS 40 (1987), 377-379. 20 P. Jeffery, op. cit p. 22-31, 93-98. 21 L. Kiss B. Rajeczky, Corpus Musicae Popularis Hungaricae. V. Laments, Budapest, 1966. Nrs. 134, 136, 138, 143, 69, 190. For easy comparison I transcribed the excerpts into a notation similar to the chant examples and I marked the elements with letters. 22 J. Szendrei, «Ami egy gyüjtésböl kimaradt» (=«Unrecorded experiences in .

-

collecting laments,») Magyar Zene 28 (1987) 21-22.

Example 3

much greater respect than is customary. In these instances spontaneity is greater achievement than consciousness. The second lesson is that, whatever the historical relationship between the two chant repertories may have been, the hidden presence of the melodic concept rules out the supposition that the Old Roman variant is a kind of subsequent ornamentation of a basic melody (as Apel supposed for Gregorian melismatic material in its entirety23 ). We may plead that we are no longer familiar with the parent melodies of the Gregorian and Old Roman chants. But is a

it not

precisely through

what is

common to

the two

repertories

that we

understand their individual character better?

With that

we

have

come

back

to the

first stage in the history of Latin

chant.But what do we mean by Gregorian chant? The main why the debate has been a heated one is, I believe, that we have

liturgical reason

always identified

not

chant

can

be

thought

the sense in which the word is used. Gregorian of as a repertory, a collection of pieces fixed note

Opusmusik. If fully justified: we

for note,

as

it be

regarded that way

then

Apel's agnosticism

be able to find out whether the notated melodies existed at all before being put into writing. 24 If we leave the memorizing devices of the oral culture out of consideration we is

even

may

risk the

shall

never

statement that the melismatic melodies of the 4th century

years later».25 We have no way of any 11th- or 12th-century melody note by note back to the 4th century. However, the «piece» in this culture is not only a concrete «were

probably forgotten fifty

tracing

sequence of notes but also a melodic concept realized in individual melodies. The Old Roman variants prove most strikingly that the notion of ‘melodic ( Ex. 4 ).

concept’

is very far from

being

a mere

20th-century

fiction. 26

only mean the extension of the multiple Hughes. In the case of melodies with identical

The Old Roman variants

series

investigated by David

texts there

copies adapted

is always the possibility of a direct borrowing prevailing style. Of more significance are

to the

where melodies with different texts

essentially

23 W.

are set

within the

the same musical structures; even

more

or of cases

repertory to significant is the greasame

Apel, Gregorian Chant, Bloomington, 1958, 507-508, 510. W. Apel, «The Central Problem of Gregorian Chant,» JAMS 9 (1956), 118-127. 25 W. Apel, Gregorian Chant, p. 508. 26 a) and c): Graduale Romanum, Solesmes, 1974, 76, 77. b) Old-Roman version from Vat. lat. 5319 (= Monumenta Monodica Medii Aevi II, 1970, 334). 24

Example 4

ter number of items which have different texts but which still obey similar rules: these occur regularly in the Old Roman, Ambrosian and Gregorian 27 repertories. (Ex. 5 ). One can quite understand Peter Jeffery’s doubt as to whether these rules can be defined at all. 28 But we are all well aware of the possibility of definition, even if we have failed so far to carry it out or have done so only through an insufficiently wide spectrum. One has only to think of Edward Nowacki's antiphon studies, 29 Terence Bailey's analysis of the Ambrosian alleluias, 30 or the insights 31 gained during the analysis of the antiphons in Hungarian sources.

We may go even further. In a music which

is not

Opusmusik the style

itself is the primary artistic object, the primary material that is handed down. However great the differences between the Old Roman, Ambrosian and Gregorian repertories may be, they seem very similar (or identical)

compared with any other music. The features that clearly link these so-called styles —or rather sub-styles— are the treatment of text, the melodic motion, the treatment of tonality, the handling of the various intervals, the freedom of rhythm, the role of melodic skeletons, the conjunct when

motion and the function of ornaments and melismas. It may well be that scholarship has not been sufficiently conscious of these stylistic features, but that does not eliminate the fact of the identity of style. This style is the

primary subject-matter of our investigations. In his approach to the antecedents of the 9th-century records the scholar can, of course, plead ignoramus et ignorabimus, if they are taken note by note, but there is no reason to deny the continuity of the tradition in respect of vital musical facts, and consequently we cannot escape the duty to undertake further research. If parallels between Latin liturgical chant and folk music are found in greater number (especially the folk music of peoples not in direct contact with Gregorian chant), and if these parallels affect not

27 a) Old-Roman responsory Vat. lat. B 79 f 87. b) Ambrosian responsory. London, BL Cod. Add. 34209. (= PalMus V), p. 56. c) Gregorian responsory. Worcester Cathedral Library Cod. F. 160 (= PalMus XII), p. 53. 28 Op. cit p. 22-50. 29 E. Ch. Nowacki, Studies on the Office Antiphons of the Old Roman Manuscripts, Dissertation, 1980. Idem: «The Gregorian Office Antiphons and the Comparative Method,» Journal of Musicology 4 (1985), 243-275. 30 T. Bailey, The Ambrosian Alleluias. The Plainsong & Mediaeval Music Society, 1983. 31 Melodiarium Hungariae Medii Aevi II. Antiphons. (Forthcoming) Cf. L. Dobszay, «Experiences in the Musical Classification of Antiphons,» Cantus Planus: Papers read at the Third Meeting. Budapest, 1990, 143-156. .

Example 5

only individual items,

but whole stylistic layers, 32 additional evidence which could illuminate the issues of continuity and chronology gained otherwise so difficult to engage. ( Ex. 6 ). On a previous occasion I called attention to the fact that in the Rule of St Benedict all musical genres of the Office are listed as distinct categories. If we add to that the presence of these genres in the three large may be

we can claim that the difference in their appearance serves additional proof that the same thing is being talked about in both the 6th and the 9th-10th centuries. What Kodály said of folk music may prove valid here as well: it is the existence of variants which proves evidence of the longevity of types, of a memory extending over centuries.33 In the emergence of Gregorian chant three epochs can in my opinion be differentiated, which of course cannot be sharply separated. It makes

repertories, as

sense to

concern.

raise

questions only if they are appropriate to the epoch they Through the chance encounter of texts and musical stereotypes

establishing style,

the first

epoch brought

forth what is most essential in

Gregorian chant, even if we are unable to assign to it one concrete, single «piece». Through the consolidation of texts and the music assigned to them (at least in a given community or region) the second epoch offered better chances for memorization, especially at the scholae stabilized. The spread of Roman

time when the role of Christianity, the energy and the segregation of large

with which musical repertories expanded, dialects, left their mark not only on the mutual

large surviving repertories but also example in the relationship between

on the late

the

a

relationship

of the three

Gregorian tradition,

pentatonic

for

and diatonic dialects.

The third

epoch includes the fixation and subsequent dissemination of Gregorian tradition: not exclusively and not primarily by means of writing, and thus not with the exclusiveness of authentic melodic forms in the sense of Opusmusik. Whether the sets of variants coming before

the

32 a) V. Sztoin, Chants Populates de la partie Centrale de la Bulgarie du Nord, Sofia, 1931, Nrs. 616, 757, B. Bartók, Cintece popolare Romanest din Comitatul Bihor, Bucuresti, 1913, Nr. 100. B. Bartók, Slovenskél'udavéPiesne. I-II, Bratislava, 1959, 1970. Nr. 13/b. —Antiphonale Paulinorum (Zagreb Bibl. Univ. MR 8). —Graduale Romanum 1974, 540. b) V. Sztoin, Chants Populaires Bulgares du Timok á la Vita, Sofia, 1928, Nr. 1229. (Here transcribed in chant notation). —Graduale Romanum 1974, 838. Antiphonale Strigoniense (Bratislava/Pozsony, Knauz 2), f 176v, 75v, Cf. L. Dobszay, A siratóstílus dallamköre zenetörténetünkben és népzenénkben. (= The Melodic Sphere of the Lament Style), Budapest, 1983, 54-82. 33 Z. Kodály, «Néprajz és zenetörténet,» («Ethnography and Music History») Visszatekintés, Budapest, 1974, II. 75, 231. —

Example 6

and after this turning point in the 8th-10th centuries are related to each other remains a mystery. These issues being as difficult to resolve as they are, no one can expect to achieve the precision of the natural sciences. If the social sciences were to strive for such exactness, or with a gesture of agnosticism to give up studying anything that falls short of it, they would not only put up an obstacle to the natural course of human interest but militate against truth as well. For, although it is hard to trace the threads back from the notated codices, to cut off these threads is undoubtedly an historical error. In the course of the investigation several ideas, methods and auxiliary sciences can stimulate lines of inquiry. The scholar can venture freely into this terrain, sure in the knowledge that his colleagues will correct his errors and use his results and that others too have gained.

through

his act of courage

[12] On

Gregorian Orality By KENNETH LEVY

ONE LOOK BACK AT THE BEGINNINGS OF PLAINCHANT

CANunwritten melos behind the musical substances that

and

see an

profiled early neumes? Viewing orality by way of the written record has long had an appeal. When Peter Wagner and Paolo Ferretti described florid Gregorian Tracts as realizations of “psalmodic” frameworks, the oral, improvisatory element was already basic to their conception of melodic origins. For the nineteenth century’s pioneers in plainchant history, this was not yet so. The Mass and Office Propers in preserved neumations were taken to be faithful copies of authoritative written models that were established under Gregory the Great. This view was encouraged by John the Deacon’s Life of Gregory, which hinted at an exemplar of Gregorian times preserved at the Lateran during the later ninth century, and by other Carolingian mythmakers who alluded to a “Gregorian” antecedent for the chanted propers in trope-like are

in

1

2

Peter Wagner, Formenlebre (Leipzig: 1

Vachein.

.

.

Einfiibrung in die Gregorianiscben Melodien, vol. 3, Gregorianiscbe Breitkopf und Hä1921), x: “Ihr Stil: freiere psalmodieahnliPaolo Ferretti, Estetica gregoriana, ossia Trattato delle forme

musicali del cantogregoriano, vol.

1

(Rome: Pontificio Istituto di Musica Sacra, 1934),

132-39. 1 The text (26) does not imply neumes: “Lateranensis ubi usque hodie lectus ejus, in quo recubans modulabatur, et flagellum ipsius, quo pueris minabatur, veneratione congrua cum authentico Antiphonario reservatur Jacques Paul"; Series Latina, vol. 75 (Paris: Gamier, 1862), Migne, Patrologiae Cursus Completus rn. 90. J. L. F. Danjou, who first drew attention to the Gradual-Tonary of Faculte de Medecine, MS H.159 described it Saint-Benigne of Dijon (Montpeller, as “un des Antiphonaires notes au commencement du IXe siecleúar un des clercs que Charleavait fait etudier a Rome, ou par un des chantres queá Adrien avail envoyes en France, lesquels avaient sous les yet l’exeáe note de la main meme de saint Gregoire, exemplaire qu’on víencore au Xe Sainterie SiJane-de-Ltran"; Jean-Ĺran”; cited in Paleographic musicale, vol. 7 (èes: 10 after Revuedela musique liturgique, December 1847. The Belgian JesuitPierre,Ltte, who in 1831 edited a diplomatic facsimile of the Cantatorium, Saint Gall 359, believed “dans ce monument unique, nous possedons veritablement l’oeuvre rium,ngoire”; he supposed that particular manuscript was de Saint Gregoire”; íonaire de copied ca., after Áiphoblementt Saint Gregoire. Facsimile de manuscrit de Saint-Gall, 2nd ed. (Brussels: CH.-J.-A. .

.

.

.

.

Greuse, 1867),

.

.

.

10.

DOI: 10.4324/9781315090283-15

Oral and Written Transmission in Chant

encomia that preface some early musical collections. 3 The association of “Gregory” with neumed plainchant needed revision as it became clear that the surviving noted documents were much later than the sixth-seventh centuries. In place of the written “Gregorian” original, the notion took hold of an unwritten transmission between the melodies’ conception and their writing-down. This was not readily acknowledged, in part because it raised knotty historical and theological problems, but in part because specialists for a long while had enough else to occupy them. Oral-improvisational issues were set aside as they analyzed the received Gregorian versions for their modal behavior and melodic structure, pondered the origin of the neumes, and debated the nature of rhythmic and microtonal nuances. Issues of orality reached the surface at the International Congress of Sacred Music held at Rome in 1950. EugèneCardine reported there on the projected Benedictine édition critique of the Gregorian Mass Propers, and he declared that between the music’s origin and its conversion to neumes there lay a stage of oral transmission. 4 Walter Lipphardt spoke of “Kemmelodien,” and a “lebendige Improvisations-kunst," which he identified with the “Italian South,” contrasting it with the more sculptural shapes he saw as the products of a northern “frankischen Tradition.” 5 Lipphardt’s notions of improvisatory art and Frankish tradition were stimulated by Bruno Stablein’s paper delivered at that same Congress, reviewing the relationship between the Old Roman and Gregorian repertories. 6 With this, the link between melodic origins and musical notation was cut, and in its place stood the questions concerning oral transmission that have since been high on specialists’ agendas. Contributions came from Hucke (begin-

3 “Gregorius praesul conposuit hunc libellum musicae artis scolae qui cantorum”; in the Monza Cantatorium, fol. 2, ed. René-Jean Hesbert, Antipbonale Missarum Sextuplex (Brussels: Vromant, 1935), 2; Bruno Stäblein, ‘“Gregorius praesul,' der Prolog zum römischen Antiphonale,” in Musik und Verlag: Karl Vötterk zum 65. Geburtstag (Kassel: Bärenreiter, 1968), 537-61; Michel Huglo, Les livres de chant liturgique, Typologie des sources du moyen âge occidental. Directeur: L. Genicot. Fasc. 52 (Turnhout: Brepols, 1988), 81, 101-21. 4 Eugène Cardine, “De l’édition critique du Graduel,” in Higini Anglès, ed., Atti del Congresso Internazionale di Musica Sacra (Rome, 1950; reprint, Tournai: Desclée, 1952), 190: “la notation musicale ayant fait son apparition vers la fin du IXe siècle sous forme de neumes nous n’aurons pas là le livre primitif sous sa forme matérielle, puisque la notation musicale était alors inconnue.” 5 Walter Lipphardt, “Gregor der Grosse und sein Anteil am römischen Antiphonar," in Atti del Congresso Internazionale di Musica Sacra, 248-54. 6 Stäblein, “Zur Frühgeschichte des römischen Chorals,” in Atti del Congresso Internazionale di Musica Sacra, 271-75. .

.

.

.

.

.

.

.

.

.

.

.

Oral and Written Transmission in Chant

8

ning in 1953), 7 Apel (1956) Levy (1963), 9 Cutter (1967) and Connolly (1972) 11 Already in 1957, Ferand added the perspective of Then Nettl in 1974 examined the a specialist in improvisation. nature and scope of improvisatory practice with an ethnomusicologist's rigor. At one end of Nettl’s spectrum lay free, ecstatic deliveries of a sort imaginable at ultimate beginnings: music that is the outcome of “unpremeditated, spur of the moment decisions.” At the other end lay a melos that is essentially fixed and memorized, with only a narrow margin for embellishment: the singers are “performing a version of something.” 13 Much of plainchant doubtless began as freely improvised oral delivery, and much of it may have remained near that freer end of the improvisatory spectrum throughout its oral transmission. It may only have been the availability of neumes that rendered the fixing of melodic shapes a practicality. Yet it may also be that some chants during the purely oral stage went from the freely improvisatory toward more calculated methods of melodic production. Events during a stage prior to notation are by their nature IO

,

,

.

12

Helmut Hucke, “Musikalische Formen des Offiziumsantiphonen,” Kircbenmusikaliscbes Jabrbucb 37 (1933): 7-33; idem, “Zur Entwicklung des christlichen Kultgesangs zum Gregorianischen Gesang,” Römiscbe Quartalscbrift 48 (1953): 172-85; idem, “Improvisation im Gregorianischen Gesang,” Kircbenmusikaliscbes Jabrbucb 38 7

5-8. (1954): 8

Willi Apel, “The Central Problem of Gregorian Chant,” this JOURNAL 9 (1956): 118-27; idem, Gregorian Cbant (Bloomington: Indiana University Press, 1958), 324

and 511, dealing with the second-mode tracts and “skeleton melodies.” 9 Kenneth Levy, “The Byzantine Sanctus and its Modal Tradition in East and West,” Annales musicologiques 6 (1958-63): 52-59; idem, “The Trisagion in Byzantium and the West,” Report of the Eleventh Congress of the International Musicological Society, Copenhagen, 1972, ed. Henrik Glahn, Søren Sørenson, and Peter Ryom (Copenhagen: Wilhelm Hansen, 1974) 761-65. 10 F. Cutter, “The Old-Roman Chant Tradition: Oral or Written?,” this Paul JOURNAL 20 (1967): 167-81; idem, “Oral Transmission of the Old-Roman Responsories?." The Musical Quarterly 62 (1976): 182-94. ,

"

Thomas H. Connolly, “Introits and Archetypes: Some Archaisms of the Old Roman Chant,” this JOURNAL 25 (1972): 157-74. 12 Ernest Ferand, “Improvisation,” Die Musik in Gescbicbte und Gegenwart, vol. 6 (Kassel: Bärenreiter, 1957), cols. 1096-1100. “Das melismatische Singen, später in den responsorischen Formen des greg. Gsg. bevorzugt, weist improvisatorische Züge in besonders reichem Masse auf Spuren der Improvisationspraxis sind in den .” (col. meisten Formen des greg. Gsg. nachweisbar oder zumindest zu vermuten .

.

.

.

1098).

.

13 Bruno Nettl, “Thoughts on Improvisation: A Comparative Approach,” The Musical Quarterly 60 (1974): 1—19; cf. 3; also Nettl’s “Types of Tradition and Transmission,” in Cross-Cultural Perspectives on Music, ed. Robert Falck and Timothy Rice (Toronto: University of Toronto Press, 1982), 3-19.

“probably irrecoverable,” follows I will

venture

as

such

a

David

Hughes

has put it,

14

but in what

recovery.

plainchant orality continued to be dealt with by a narrow circle of specialists until a series of papers by Leo Treitler and Helmut Hucke, begun in 1974, brought them to broader musicological attention. 15 Treitler’s and Hucke’s proposals are concerned with four subject-areas. One is the date of the Gregorian corpus’s conversion to neumes; in agreement with Solange Corbin, they put this 16 about 900. Two is the rationale of the early neuming technique; find this they largely in Carolingian punctuation practice. 17 Three is the nature of melodic production after the introduction of neumes; along with Ferand and van der Werf, they envisage oral-improvisational practices that were “reconstructive” or “reimprovisational,” continuing during the first two or three centuries of the noted 18 Questions of

transmission.

“4

this

My

own

views

on

each of these have been rather

“Evidence for the Traditional View of the Transmission of Gregorian Chant,”

JOURNAL

(1987): 377. positions appear in the “Communication” by Treitler in this JOURNAL 41 (1988): 566-75, and in Hucke’s “Gregorianische Fragen,” DieMusikforscbung 41 (1988): 15

40

Recent

304-30, particularly 326-30; these contain responses to criticisms that were launched in my “Charlemagne’s Archetype of Gregorian Chant,” this JOURNAL 40 (1987): 1-30, and by David Hughes in “Evidence for the Traditional View of the Transmission of Gregorian Chant.” Hughes has further remarks in two “Communications”: this JOURNAL 41 (1988): 578-79, and 42 (1989): 435-37. 16 Solange Corbin, “Les notations neumatiques en France à l’époque carolingienne," Revue d'bistoire de I'église en France 38 (1952): 226-28; idem, “Les neumes,” in Roland Manuel, ed., Histoire de la musique, Encyclopédic de la Pléiade, no. 9 (Paris: Gallimard, 1960), vol. 1, 690-94; idem, Die Neumen, Palaeographie der Musik nach den Plänen Leo Schrades, vol. 1, sect. 3 (Cologne: Arno Volk-Verlag, 1977), 22-42; Treitler, “Reading and Singing: On the Genesis of Occidental Music-Writing,” Early Music History 4 (1984): 176; Hucke, “Toward a New Historical View of Gregorian Chant,” this JOURNAL 33 (1980): 445. In “Oral, Written, and Literate Process in the Transmission of Medieval Music,” Speculum 56 (1981): 474-75, Treitler states, “Before the tenth century the tradition of Western art music was an oral tradition.” '7 Treitler, “The Early History of Music Writing in the West,” this JOURNAL 35 (1982): 237-79; idem, “Reading and Singing: On the Genesis of Occidental Music Writing,” 135-208; idem, “Die Entstehung der abendländischen Notenschrift,” Die Musikforscbung 37 (1984): 259-67; Hucke, “Die Anfänge der abendländischen Notenschrift," Festschrift Rudolf Elvers, ed. Ernst Herttrich and Hans Schneider (Tutzing: H. Schneider Verlag, 1985), 271-88. 18 Ferand, “Improvisation,” col. 1099: “Die Unbestimmtheit der Neumenschrift, zumindest in ihren Anfängen, ist in dem Umstand begründet, dass die so aufgezeichneten zeichneten Gesänge selbst keine eindeutigen festen Gebilde waren, sondern lediglich mehr oder weniger in Umrissen gegebene Melodieskelette, die von den Sangern in vielfach veränderter Gestalt improvisatorisch immer wider neugefasst wurden.” Treitler speaks of Gregorian chant in a similar vein: “The oral tradition was translated after the ninth century into writing. But the evolution from a performance practice represented in writing, to a tradition of composing, transmission, and reading, took

in earlier papers. My focus now is on the fourth area, concerning the nature of the melos before neumes were introduced, and the relation between the vanished oral and the written versions that we have. This is the most significant of the areas, and the most challenging, in that it addresses the substance of the prehistoric melos and the basis for forming opinions about it. In Treitler’s and Hucke’s conceptions, good access to the oral states is had by way of the preserved written states, and what these reveal are improvisational practices that were relatively free. 20The Gregorian melodies arose from “the continuity of witnesses of “a performance performance.” The preserved versions are 21 in practice represented writing,” “documentations of a 22 “frozen improvisations”; 23 they offer “transperformance practice,”

different, and they

are

already spelled

out to some extent

19

The act of writing was thus a kind of performance span of centuries. 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”; “The Early History of Music Writing,” 237. He supplies this process with the name of “reconstruction”: “a repeated process of 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 “‘Centonate’ Chant: übles Flickwerk or E pluribus unus?,” this JOURNAL 28 (1975): 11. Hendrik van der Werf gives this the name “reimprovisation”; The Emergence of Gregorian Chant: A Comparative Study of Ambrosian, Roman, and Gregorian Chant (Rochester: the Author, 1983), vol. 1, part 1, A Study of Modes and Melodies, 164-66; “the tradition according to which a given melody was reimprovised for every occasion, including even the act of copying it... ; the persistence of the unheighted neumes in places like Sankt Gallen is testimony to the continued interest in reimprovisation. It made no sense to notate precisely melodies which one knew to vary from one year to another and from one singer to the next”; ibid ., 165. 19 On the nature of the melos after neumes (“reimprovisation”), in my “Communication" to this JOURNAL 41 (1988): 576; on the date of the Gregorian neumation, in “Charlemagne’s Archetype of Gregorian Chant”; on neumatic etiology, in “On the of Neumes,” Early Music History 7 (1987): 59-90. Origin 20 Treitler, “Homer and Gregory: The Transmission of Epic Poetry and Chant,” The Musical Quarterly 60 (1974): 346. 21 Treitler, “The Early History of Music Writing in the West,” 237. 22 Hucke, “Toward a New Historical View of Gregorian Chant,” 453-54; the refrains of two Gregorian responsories “are not different, individual melodies in a strict sense. They are ‘documentations of a performance practice.’” In Hucke’s “Die Anfä"nge der Bearbeitung”: “Das Stammrepertoire des Gregorianischen Gesangs ist Aufeeichnung aus \l=m"\undlicher\l=u"\iberlieferung”; in Bearbeitung in der Musik. Colloquium Kurt von Fischer zum 70. Geburtstag, ed. Dorothea Baumann, Scbweizer Jbhrbuch \l=f"\ur Musikwissenscbaft, Neue Folge, 3 (1983): 16. 23 Hucke, “Derüibergang von \l=m"\undlicher zu schriftlicher Musiüiberlieferung im Mittelalter,” in Report of the Twelfth Congress of the International Musicological Society, Berkeley, 1977, ed. Daniel Heartz and Bonnie Wade (Basel: \l=B"\arenreiter, 1981),

place over a analogous

to

.

.

.

parent” images of bygone oral deliveries. 24 In light of earlier discussions sions of orality, reaching back to Cardine’s and Lipphardt’s declarations of 1950, these assertions are not altogether novel, so that an of discoverer’s fervor that is evident in them is a bit Treitler’s recent “Communication” has it: “what is striking surprising. is the resistance to a serious confrontation with the reality that at some time in history, no matter how far back one wants to push it, the Western musical heritage goes back to an oral tradition that left its mark and that is never entirely out of the picture as a factor in musical practice.” 25 That protests too much. The question has long been, not whether oral elements survive, but how much oral there is, and how one goes about making its identification. Some of the Gregorian neumed versions may represent faithful recordings of bygone oralimprovisational deliveries; they may be “transparent” witnesses of the oral past. Others may represent only the last stages in a multi-stage process where compositional and editorial additions that entered during prior written stages now obstruct the view. What promised to be a major support for conjectures about oral melodic behavior was the analogy, introduced by Treitler in 1974, between Gregorian chant and the works of Balkan epic bards. Plainchant stood to be illuminated by the practice of ancient Homer and of the modem Serbian-Macedonian “singers of tales” whose productions were analyzed in the comparative literary studies of Parry and Lord. 26 The link of “Homer and Gregory” catches the fancy, but apart from both of them once being “oral,” there is little basis for crossover. The improvisatory flights of Gregorian plainchant took in the of medium those of epic bards in the medium of music; place words; and what Parry and Lord dealt with were the verbal texts. For plainchant, the texts come verbatim from Scripture (or sometimes Patristic literature), and are not results of improvisatory elaboration. The improvisations that produced the bardic texts also differ in

occasional

tone

180-90. “In der schriftlichen überlieferung von Solostücken ist, beispielsweise in den Tractus und in den Gradualienversen, eine Vortragspraxis eingefroren worden” (p. 180).

24 Treitler, “Communication” in this JOURNAL 41 (1988): 575: “No matter how uniform the written transmission is, and even if we regard it exclusively as a product of copying faithfully from one source to another, it is transparent to the oral tradition that was its ultimate source.”

25

Treitler, “Communication,” 575. Albert B. Lord, The Singer of Tales (Cambridge: Harvard University Press, 1960); also Lord’s summary, “Oral Poetry,” in Alex Preminger, et al., Princeton Encyclopedia ofPoetry and Poetics (Princeton: Princeton University Press, 1965), 591-93. 26

character from those of the plainchant melos. For the epic, the verbal basis is the stichic one-liner; in Homer, the dactyllic hexameter. Epic continuities are built up in a succession of such short repeated formations. In modem epic there is a tendency for the music to conform to the shapes of the repeating one-liners; 27 a similar conformity is projected for Homeric music, operating under hexametric 28 Freer melismatic, “improvisational” elaborations of the constraints. sort that may have contributed to the shaping of florid Gregorian chants have no counterpart in epic at all. Then too, the improvisers of epic were soloists. There are solo Gregorian chants that display a marked improvisational thrust, as was recognized in Wagner’s description of Tracts as “psalmodie-ahnliche Variationen.” 29 Yet in Gregorian responsorial chants, the solo passages are often linked in style and substance to the choral passages. For the choral deliveries of a Gregorian schola cantorum, considerable predetermination and coordination must be supposed. Such improvisational flights as are generated in a “process of oral composition through performance” 30 would find little place. 31 27 James A. Notopoulos, Modern Greek Heroic Poetry (Ethnic Folkways Library Album FE 4468, New York, 1960) discussed examples of the techniques; see also Stathis Gauntlett, Rebetika Carmina Graeciae Recentioris: A Contribution to the Definition of the Term and the Genre rebetiko tragoudi through Detailed Analysis of its Verses and of the Evolution of its Performance (D. Phil, diss., Oxford University, 1978; Athens: Denise Harvey, 1985), 61: “the processes of improvised oral composition appear to have been expedited by the large degree of uniformity which exists in the metrical and strophic This widespread uniformity would have secured specifications of these verses. considerable flexibility in combination of couplets from different cycles and scope for the interchange of melodies between cycles.” 28 Giovanni Comotti, Music in Greek and Roman Culture (Baltimore: Johns Hopkins University Press, 1989), 14: “the hexameter poetiy of the rhapsodes, recited rather than sung, such as the great Homeric epics (the Iliad and the Odyssey)." For the lyrics of Sappho, “the melodic sequences on which these songs were modulated had evidently a very simple structure that conformed to the metrical figures of the poetic .

text”; ibid 29

,

.

.

21.

Wagner, Einfübrung in die Gregorianiscben Melodien,

vol. 3,

Gregorianiscbe Formenlebre,

353.

Treitler, “‘Centonate’ Chant,” 12. This difficulty is acknowledged by Treitler in the ultimate footnote to his initial paper. His apparent response is to question the antiquity of the Gregorian antiphonal and responsorial Propers (“Homer and Gregory,” 371-72, n. 22): “One question will surely have arisen in the minds of thoughtful readers of the present essay: 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 foregoing analysis? Not at all. I shall briefly mention the most important considerations. (a) The hypothesis is meant to apply only through the eighth century at the 30 31

Apart from the questionable fit between “Homer” and “Gregory,” the more fundamental trouble with such projections about orality is that they are based on what amounts to a single witness. Gregorian chant reaches us in a remarkably uniform state. Hundreds of manuscripts of the ninth through thirteenth centuries represent a tradition that was established by Carolingian musicians under a mandate to impose a single authoritative usage in their realm. This unity of tradition has been underscored by David Hughes. 32 What matters here is that other musical versions are generally lost. The Gregorian text that comes down is a unique source which is not “oral” and “early” but “written” and “late.” Efforts to identify oral elements through analysis of its fabric are bound to remain, as they were with the “psalmodic” designations of Wagner and Ferretti, exercises in personal intuition.

indications of an oral state one wants better evidence. like this may be obtained from cases of parallel readings or Something two or more plainchants (generally representatives Where “multiples.” of different liturgical families) share the same texts and ritual functions, and also some amount of modal-melodic substance, their music in common may represent an earlier “oral” formulation. Dom Mocquereau made comparisons of this sort a century ago, paralleling the Gregorian, Milanese, and Old-Roman versions of the Introit For

surer

latest for the Gregorian tradition; for the Old Roman, speculation can be of only the most tentative sort, pending more systematic study, (b) 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 singing by a choir, (c) In principle, refrain singing is not as such in conflict with the practice of oral composition. 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.” To deny the existence of the eighth-century Proper repertory (the repertory in Hesbert’s Sextuplex) this way seems less than convincing, as does Hucke’s response to the same question in “Der übergang von mündlicher zu schriftlicher Musiküberlieferung im Mittelalter,” 180-90; cf. 182. 32 “Evidence for the Traditional View of the Transmission of Gregorian Chant,” 398: “The extreme stability of the basic tradition is obvious enough. An ideal example of this is the transmission of the Alleluia Dies sanctificatus ; here a complex melismatic tune recurs virtually unchanged in source after source and texting after texting. For an equally convincing and more accessible instance of stability one may leaf through the pages of the Graduale Triplex (Solesmes: Abbaye Saint-Pierre, 1979) noting how rarely the readings of Lan and Mur 3 disagree.” .

.

.

Resurrexi and the Gradual A sunrno caelo. 33 Stäblein made use of similar comparisons in 1950 in launching the current round of speculations about Old Roman chant. 34 Recently, Dom Claire has used multiples to seek out archaic modality and the “corde-mère” of certain modes; 35 Alberto Turco, to consider the “marche vers l’Octoéchos” of certain antiphons; 36 Terence Bailey, the links among Ambrosian, Old Beneventan, and Roman repertories; 37 van der Werf, the existence of “parent melodies.” 38 I have used multiples to address archaic states of the anaphoral chants, of the Easter-vigil mass, and the Trisagion. 39 In 1984, I pointed to multiples found in Old-Hispanic, Gregorian, and Ambrosian traditions as indications that certain Proper chants that are first preserved in neumes of the tenth through twelfth centuries had assumed much of their eventual melodic shape by about 700. 40 For multiples to be of use in viewing oral states, they must meet certain conditions. The neumations should be relatively early, as close as possible to the time of oral transmission. They should be different enough to be seen as the fruits of independent acts of writing, and not as outgrowths of a common written model. Yet their melodic substances should be alike enough to suggest some of the musical physiognomy of an ultimate oral source. These conditions are not easily met. The written versions tend to be late: a Gregorian witness may go back to the later ninth or tenth century, but the counterpart Old Roman, Milanese, Beneventan, or Ravennate witnesses may begin only with the eleventh or twelfth century. The influence of one written version upon another cannot be discounted: the neumed Gregorian version may have been the model for the others. Then too,

33

Paléographic musicale, vol. 2 (Solesmes: Imprimerie Saint-Pierre, 1891), 6-9. Stäblein, “Zur Frühgeschichte des römischen Chorals,” 271-75. A recent contribution is Stefan Klöckner’s “Analytische Untersuchungen an 16. Introiten im I. Ton des altrömischen und des frankisch-gregorianischen Repertoires hinsichtlich einer bewussten melodischen Abhängigkeit,” Beiträge zur Gregorianik 5 (1988): 3-95. Jean Claire, “L’Evolution modale et les répertoires liturgiques occidentaux,” 34

35

Revue

grégorienne 40 (1962): 196-210, 229-45; idem, “La psalmodie responsoriale Revue grégorienne 41 (1963): 7-29; idem, “Les répertoires liturgiques latins l’Octoéchos," Études grégoriennes 15 (1975): 1-181 and supplement.

antique,” avant 36

“Les répertoires liturgiques latins,” 177-221; also his “Melodie-tipo e timbri modali nell’ Antiphonale romanum,” in Studi Gregoriani 3 (1897): 191-241. 37 Terence Bailey, The Ambrosian Alleluias (London: The Plainsong and Mediaeval Music Society, 1983), 53-60. 38 van der Werf, The Emergence of Gregorian Chant, vol. 1, part 1, no. 39 Levy, “The Byzantine Sanctus and its Modal Tradition in East and West,” 7-67; idem, “The Italian Neophytes’ Chants,” this JOURNAL 23 (1970): 181-227; idem “The Trisagion in Byzantium and the West,” 761-65. 40 “Toledo, Rome and the Legacy of Gaul,” Early Music History 4 (1984): 49-99.

when plainchants travel they tend to take on the stylistic features of the regions where they settle. The first things obscured by a new stylization are the surface details of what went before. Some years ago, I examined a situation that met these conditions unusually well. It was a case of what I now propose calling “close multiples.” For the processional Antiphon Deprecamur te Domine, there is a “Gregorian” pedigree perhaps better than that for any other chant. Through citations in Bede and the Vita of St. Augustine of Canterbury it can be traced to the Roman liturgical environment under Gregory the Great. 41 There are four medieval versions of its music. All are in the protos with finals on D. The Old Roman and Milanese versions are independent in their melodic details. Of the other two, one is a widely-circulated “Carolingian” version, and one appears only in the Beneventan zone. They have musical substances that are often identical although their neumations may be independent. This offers a possible window on an oral stage. The closer their agreement, the sharper the view through that window. This notion was advanced tentatively for Deprecamur. Now there may be a stronger case of close multiples with the Offertory Elegerunt apostoli. This chant for St. Stephen Protomartyr has often been cited as a Gallican relic surviving in a Gregorian liturgical environment. 42 I have lent support by observing that Elegerunt's text is not a verbatim excerpt from the Psalter (as is the case with most Gregorian offertories), but a patchwork of “non-psalmic” phrases drawn from Acts 6 and 7. Texts of this “centonate-libretto” sort are the norm among the florid Hispanic Offertories. Inasmuch as Elegerunt serves as a basis for florid Offertories in both the Mozarabic and Gregorian rites, it represents a class of “international” chants for which a Gallican antecedent can be supposed. 43 Among the earliest Gregorian documents, Elegerunt finds only a marginal place. In Hesbert’s collection of old unnoted Mass Propers it 41

“A Gregorian Processional Neue Folge, 2 (1982): 91-102.

Levy,

Antiphon,”

Scbweizer Jabrbucb für Musikwissenscbaft,

Paléograpbie musicale, vol. 15 (Solesmes: Imprimerie Saint-Pierre, 1937-53), 165: “d’origine gallicane”; Michel Huglo, “Altgallikanische Liturgie,” in Karl Gustav Fellerer, Gescbicbte der katboliscben Kircbenmusik (Kassel: Bärenreiter, 1972), vol. 1, 226; Huglo, “Gallican Rite, Music,” in The New Grove Dictionary of Music and Musicians (London: Macmillan, 1980), vol. 7, 116-17; Giacomo Bonifacio Baroffio and Ruth Steiner, “Offertory,” in ibid vol. 13, 515. 43 Levy, “Toledo, Rome and the Legacy of Gaul,” 80-87; Hucke’s “Die Texte der Offertorien,” in Speculum Musicae Artis: Festgabe für Heinrich Husmatnn, ed. Heinz Becker and Reinhard Gerlach (Munich: W. Fink, 1970), 193-203, has a useful 42

,

source-identification and classification of Gregorian

offertory

texts.

only in a usage of St. Denis, assigned to a rare Stephen feast 44 September 9 that is of likely Gallican origin. Most of the regional Italian usages, including the oldest tradition of Rome, have as their Stephen-Offertory a psalmic text (generally In virtute tua) for which an Italianate origin seems likely. 45 By the middle ninth century, however, Elegerunt was included in a neumed recension of the Gregorian Propers that received wide European circulation. Example 1 shows its opening refrain in a reading that is common to French, German, Italian, and English sources. The reading here is that of the Gradual-Tonary of Dijon, compiled ca. 1025, whose “bilingual” notation—coupling neumes with pitch-letters—makes it the earliest

appears on

Propers. 46 many medieval copies of this version, which found its

pitch-specific witness for

the Mass

There are 47 way into the modern Vatican Gradual. That all the copies descend from a single campo aperto neumation is indicated by the melismas on (i) “Elegerunt,” (3) “plenum,” and (6) “lapidaverunt,” whose fourfold pitch-repercussions tend to be launched by a pes stratus (marked with asterisks in Example 1 ). Vestiges of this distinctive neume appear in the Abruzzi, Emilia, Ravenna, the Piedmont, Normandy, Sarum, Einsiedeln-St. Gall, and perhaps Schaffhausen. 48 Such agreements in

44

Hesbert, Antipbonale Missarum Sextuplex, CVI

“Gallican Rite,” 116.

and

no.

148bis;

see

Huglo,

45 In virtute tua is the Stephen Offertory at Rome; Stäblein, Monumenta Monodica Medii Aevi, vol. 2 (Kassel: Bäarenreiter, 1970), 631; Paul F. Cutter, Musical Sources of tbe Old-Roman Mass, Musical Studies and Documents, vol. 39 (American Institute of

1972), 412. Musicology, 46 Paléograpbie musicale, vol. 8 (Solesmes: Imprimerie Saint-Pierre, 1901-5), 267; the B-flats indicated by the manuscript’s alphabetic notation are placed above the staff in the transcription; on the manuscript, see Michel Huglo, “Le tonaire de St. Bénigne de Dijon,” Annales Musicologiques 4 (1956): 7-18; Finn Egeland Hansen, ed., H159: Montpellier: Tonary of St. Benigne ofDijon, (Copenhagen: Dan Fog Musikforlag, 1974); Huglo, review of Hansen, Tbe Grammar of Gregorian Tonality, this JOURNAL 37 (1984): 416-24.

47 The Vatican edition is reprinted along with the Saint Gall neumation in the Graduate Triplex, 634; Elegerunt appears with its complement of verses in the Offertoriale edited by Carolus Ott (Tournai: Desclée, 1935), 161-63, and in an amplified reprint of the latter by Rupert Fischer, Offertoires neumés avec leurs versets d’aprés les manuscrits Laon 239 et Einsiedcln 121 (Solesmes: Abbaye Saint-Pierre, 1978), 161-63. 48 Abruzzi: Biblioteca Vaticana, lat. 4770, fol. j; Emilia: Rome, Angelica 123, fol. 33 (Paléograpbie musicale, vol. 18 [Solesmes: Imprimerie Saint-Pierre, 1969]); Ravenna: Baltimore, Walters Art Gallery, W. 11, fol. 18; the Piedmont: Novalesa, Oxford, Bodleian Library, Douce 222, fol. 109; Normandy: Paris, Bibliothèque Nationale, lat. 10508, addition to fol. 3; Sarum: (W. H. Frere, ed., Graduate Sarisburiense [London: B. Quaritch, 1894], pi. 16); Einsiedeln?: Einsiedeln 121 (Paléograpbie musicale, vol. 4, [Solesmes: Imprimerie Saint-Pierre, 1894], pi. 35); Schaffhausen?: (Paléograpbie musicale, vol. 3, [Solesmes: Imprimerie Saint-Pierre, 1892], pi. 130).

Example

I

“Vulgate” Elegerunt

wide spread of East and West Frankish sources speak for an original neumation that circulated before the divisio imperii of the 840s. 49 Turning to Elegerunt's musical substance, there are noteworthy features in the three upward sweeps of an octave from C to C, which are apparent highlightings for the three key words (I)“Elegerunt,” a

49

See

division.

Huglo, Les Livres de chant liturgiques, 84, on the musical consequences of this

(6)“lapidaverunt,” and “spiritum.” Another response to text-meaning may be seen in the intonation-like figure given to the hortatory “Domine” in line eight. There are some exact melodic correspondences, the most extensive of them involving the pattern KVg, which appears in lines three and six on “plenum” and “lapidaverunt.” This contains two formulaic elements, a and b, which appear elsewhere in different melodic contexts. 50 Element a, with the four-fold pitchrepercussions launched by the pes stratus, is also found in line one on and a vestige of it is perhaps seen in line nine on Ele-ge-runt, b is also found in line eight on “Domi-ne.” Element (a1). “spi-ri-tum” Then there is the cadential element mv, which closes lines six and nine. Such correspondences give this expansive G-mode melody a particular “form.” At the time the authoritative Carolingian-Gregorian repertory was neumed, 51 the pressures for musical-liturgical uniformity were so great that the Gregorian Mass Propers generally reach us in a single melodic version. Elegerunt is unusual in having perhaps half a dozen different musical versions spread among Frankish sources of the tenth through twelfth centuries. The version in Example I enjoyed the widest circulation and longest life, and for that reason it will be called here the “Vulgate” Elegerunt (V). The others will be called “minority” versions. There is an important minority version in sources from the Aquitaine (version A); it is the only Elegerunt found in that region. Example 2 shows it in the reading of an eleventh-century Gradual of Toulouse whose notation is the characteristic “disjunct” neumespecies (“notations à points superposées”) of the French southwest. 52 other Some Aquitainian manuscripts transmit this same reading with minor variants. All apparently descend from a single neumatic

archetype. 53

50 On uses of formulaic, “centonate” materials, see my introduction to “Plainchant” in The New Grove Dictionary of Music and Musicians, vol. 14, 804-5. 51 On the question of date, see my “Charlemagne’s Archetype of Gregorian

Chant,” 52

1-30.

London, British Library, Harleian 4951, fol. 136. On “disjunct” and “conjunct" styles of neumation, see my “On the Origin of Neumes,” 79-81. 53 The reading of Paris, Bibliothèque Nationale, lat. 903 can be seen in Paléographie musicale, vol. 13 (Solesmes: Imprimerie Saint-Pierre, 1925), pl. 23; there are related readings in Paris, Bibliothèque Nationale, lat. 776, fol. 15V; Paris, Bibliothèque Nationale, lat. 780, fol. 10; Paris, Bibliothèque Nationale, nouv. acq. lat. 1177, fol. 9V. On the Aquitanian family of Graduals, see Huglo, “Gradual,” The New Grove Dictionary of Music and Musicians, vol. 7, 605-6.

Example

2

“Aquitanian” Elegerunt

A comparison of the melodies in Examples 1 and 2 shows that a basic musical stock is common to both. Each chant, however, exhibits particular turns of phrase, and these are distinctive and systematic enough to qualify each one as a separate stylization. 54 The Vulgate’s 54

Charlotte Roederer has pointed to a similar case with the “Aquitanian” “Gregorian” processional Antiphon, Stetit/Stabat angelus ad sepulcrum; “Can Identify an Aquitanian Chant Style,” this JOURNAL 27 (1974): 75-99.

and We

from the low C have Aquitanian counterparts, but on (I) “Elegerunt,” Aquitaine rises only to pitch b, while on (6) “lapidaverunt" and (9) “spiritum” it reaches the d, then turns back with a direct descending fifth skip to g that has no Vulgate counterpart. Each chant also has a different scheme of internal repetitions and melodic references. Where the Vulgate’s (6) “lapidaverunt” reproduces its music on (3) “plenum” (Pattern K), and where at both those points the Vulgate has fourfold repercussions launched by a pes stratus (element a, which also appears on (1) “Elegerunt”), the Aquitanian version ignores the Vulgate’s melodic identities, but offers other identities of its own. The Aquitaine version of Pattern K (KAq) in line six reappears at the end of line nine, and the concluding element mAq (in lines six and nine) also appears by itself at the end of line eight. In the Vulgate, these points are only roughly related. There is also a parallel involving element nAq in the first and third of Aquitaine’s lines for which the Vulgate has no real counterpart. These differences in style and layout mark Aquitaine as a musical formulation that is somehow independent of the Vulgate. 55 Further light on the Aquitaine version is cast by a twelfth-century version from Auxerre, to the north of Burgundy. Unlike the “disjunct" neumatic ductus of Aquitaine, the Auxerre notation is “conjunct," and its neumes are disposed on a cleffed staff. 56 But the melodies are practically the same. Example 3 is the transcription of Auxerre, with the manuscript original shown in Figure I 57 It is not clear whether Aquitaine’s disjunct neumes and Auxerre’s conjunct neumes reflect independent neumations, or whether they both go back to a common neumation employing a conjunct ductus, with their differences attributable in part to Aquitaine’s early turn (during the ninth century) toward disjunct neumes, and in part to variants that entered Auxerre’s tradition during its lengthier passage from campo aperto neumes to the staff. For present purposes it will suffice to consider Aquitaine and Auxerre as representatives of a single neumatic formulation, while giving preference to Aquitaine because of its greater age and number of witnesses. Still another melody for Elegerunt is found in the Beneventan zone of south Italy (version B). It is the only Elegerunt that appears there,

three

ascents

.

55 The Alleluias that complete the refrains are not considered here. They tend to represent musical traditions separate from the rest, differing among minority versions as well as among branches of the Vulgate tradition; along with the verses that follow the Elegerunt refrain, they offer substantial avenues for further study. 56 See my “On the Origin of Neumes,” 76-86, for the designations “conjunct” and “disjunct” as substitutes for “accent” and “point” neume-species. 57 Paris, Bibliothèque Nationale, lat. 10511, fol. 154.

Example “

3

Auxerre” Elegerunt

and all copies descend from a single campo aperto neumation whose earliest witnesses are Benevento VI. 33 and Walters 6 (a Missal of Canosa), both written during the eleventh century. The Canosa neumation is shown in Figure 2 58 The chant itself is transcribed in .

58

Benevento, Biblioteca Capitolare, VI. 33, fol. 4; facsimiles 20 (Solesmes: Imprimerie Saint-Pierre, 1983).

musicale, vol.

in

Paléograpbie

Figure

I.

Auxerre Alleluia Video celos and

Offertory Elegerunt.

Figure

2.

Beneventan Offertory

Elegerunt.

twelfth-century staffed version of Benevento VI. 34. This Beneventan Elegerunt shares many of its modal and melodic features with the Vulgate and Aquitaine versions, and it is often quite close in musical particulars to one or the other of them. But it is not really identical with either. There are the same upward sweeps on (1) “Elegerunt,” (6) “lapidaverunt,” and (9) “spiritum,” but the departure-point for “spiritum” is not the low C as in the others but Example

4 after the

59

the d above it. Benevento’s use of pattern KB mirrors the exact melodic correspondences of the Vulgate’s pattern KVg on (3) “plenum fide” and (6) “lapidaverunt Iudei.” Yet it ignores the Aquitaine-Auxerre correspondences between the pattern Ks in lines six and nine. Its melodic substance for this pattern also differs from that of the Vulgate, although it is almost identical with that of Aquitaine59

Paléograpbie

musicale, vol. 15, fol.

22.

Example

4

“Beneventan” Elegerunt

Auxerre. Benevento’s three n-figures in lines one, two, and three reflect the placement of element nAq in Aquitaine, but the Aquitanian music is the same in all three instances, while the Beneventan music is not. Benevento corresponds with the Vulgate in one aspect of its larger “form,” the parallel pattern Ks in lines three and six, yet it maintains its own stylistic profile. In other formal and melodic aspects it is close to Aquitaine, but again somewhat separate. In the end, the written Beneventan version cannot be seen as either the model or the

derivative of the Vulgate or Aquitaine. It is found only in south Italy, but in view of the melodic substance that it shares with the Vulgate and Aquitaine, its ultimate origin should like theirs be sought north of the Alps, on the native ground of its centonate-libretto, “Gallican”-style text. The Beneventan version may have reached south Italy in an early recension of the Carolingian60 Mass Propers, arriving there by the beginning of the ninth century. A further perspective on this comes from England. In much the way that Auxerre expands the picture of the Aquitanian version, so that of Benevento is expanded by a version written in the eleventhcentury staffless neumes of southwest England which appears in both of the Winchester tropers as well as the Kentish Gradual-Troper Cotton Caligula A. 14. 61 I know of no staffed example. The three English readings have a common neumatic model. The neumation of the younger Winchester troper (dating from about 1050?) is seen in 62 A musical relation to the Beneventan version seems Figure 3 possible, closer at the start, with differences increasing toward the end. Benevento and Kent may be outgrowths of a version that once circulated in central or northern France. It probably got to south Italy by about 800; concerning the “Anglo-Saxon” version one can only say that it was at Winchester by about 1000. 63 Now there are three (and perhaps as many as five) different musical traditions for Elegerunt: the Vulgate, Aquitaine-Auxerre, and Benevento-Kent. They share a good deal of melodic substance but differ in their neumation and in aspects of their musical style and .

60

Levy, “Charlemagne’s Archetype of Gregorian Chant,” 11-27; the introduction Gregorian chant to south Italy is discussed in Thomas Forrest Kelly, The Beneventan Chant (Cambridge: Cambridge University Press, 1989), 18-25. of

61 ambridge, Corpus Christi College, 473, fol. 14V; Oxford, Bodleian 775, fol. 116v; London, British Library, Cotton Caligula A. 14, fol. 5 (Paléograpbie musicale, vol. 3, pi. 180). Heinrich Husmann, Tropen undSequenzenenhandschnften, Répertoire international des sources musicales, vol. B V1 (Munich: G. Henle Verlag, 1964), 154—55 suggests Canterbury as the origin of the Caligula troper; Alejandro Enrique

Planchart,

The Repertory of Tropes at Winchester (Princeton: Princeton University Press, vol. 2, 21, describes it as a fragment of an Anglo-Saxon troper from ca. 1050. Oxford, Bodleian 775, fol. 116v. 63 Susan Rankin gives indications that Corbie was a source for English neumed traditions; “Neumatic Notations in Anglo-Saxon England,” in Huglo, ed., Musicologie médiévale: Notations et séquences. Actes de la Table Ronde du C. N. R. S. á I'lnstitut de Recherche et d'Histoire des Textes: 6-7 septembre 1982 (Paris: Centre Nationale de la Recherche Scientifique, 1987), 130-31. Elegerunt is the sole Offertory to appear in the Bodleian copy with a full noted refrain and verse; all other Offertories have a cue for the refrain followed by the neumed verses; this should make it an addition, perhaps to the Offertory cycle once it was in England, or perhaps to the parent French cycle before it went to England.

1977), 62

Figure

3.

Anglo-Saxon Offertory Elegerunt.

form. Some other minority Elegerunts survive only in campo aperto neumes, so their melodic substance remains out of reach. The most Paris significant is a “Northern” version, found in eleventh-century 64 The and Arras of Saint tradition (Cambrai Denis) 75). reading (the of Arras is shown in Figure 4 .

64 Paris, Bibliothèque Bibliothèque Municipale,

Nationale, lat. 9436, fol. 72 (St. Denis usage); Cambrai, (76), fol. 35V.

75

Figure

4. “Northern” Alleluia Video celos and

Offertory Elegerunt.

A similar and perhaps identical version (the neumes are not clear) is found in the archaic “Paleofrankish” notation of Paris MS lat. 17305, dating from the later ninth or tenth century. 65 Of all the Elegerunts, this one, representing a region extending from Paris through the Franco-Belgian north, stands closest to the Carolingian heartland. It is 65 Paris, Bibliothèque Nationale, lat. 17305, fol. 14V; on the “graphic” rationale of the “Paleofrank” notation, see my “On the Origin of Neumes,” 70-79.

in its short melismatic expansion on the word (7)“orantem,” adding the word “Christe” after (8)“Domine Jesu.” Its staffless neumes indicate none of the internal melodic correspondences that serve to relate the Vulgate, Aquitaine, and Beneventan versions. Traces of other minority Elegerunts appear on a leaf of Arsenal MS 1198, where fragments of the text are entered in an eleventh-century 66 French hand; the first three entries bear neumes. The first may be the Vulgate; the second and third do not seem to agree with other known versions. Finally, there is an Old-Hispanic Elegerunt, representing the mixed Gallican-Mozarabic liturgical background from which this Offertory must have sprung. It survives only in staffless neumes of the Hispanic “northern” notational style: as a Sacrificium or Mass Offertory in the tenth-century Antiphoner of Leon, and as a Sono of the Office in the eleventh-century “Liber misticus” of San Millán de la Cogolla, now at Madrid. 67 An apparent parallel between its sixth and ninth lines may reflect the melodic correspondences found between those same lines in the Frankish versions, but no further relationships

unique and in

are

apparent.

With this, the dossier of Elegerunts is complete. In addition to the neumed version that remained in isolation on the Iberian peninsula, there were no fewer than four “Gallo-Gregorian” versions in Carolingian liturgical regions during the ninth through twelfth centuries. If the cognate Elegerunts at Aquitaine and Auxerre, at Benevento and Anglo-Saxon England, and at Paris-Arras and the “Paleofrank” North, are reckoned as separate, there were as many as seven. Whether four or seven, the number is large for a text of the Gregorian Proper when for most texts only a single chant survives. Now the question is, why does Elegerunt have so many? 66

de 1’Arsenal, ms. 1198, fol. 14V; Solange Corbin (with Répertoire de manuscrits méediévaux contenant des notations musicales, vol. 3, Bibliotbèques parisiennes (Paris: Centre National de la Recherche Scientifique, 1974), 38 and pl. VII. 67 Facsimile in Antifonario visigótico mozárabe de la Catedral de Leòn, vol. I, Monumenta Hispaniae Sacra, Serie litúrgica, vol. 5 (Barcelona and Madrid: Consejo Superior de Investigaciones Cientificas, 1959), fol. 74V; Madrid, Academia de la Historia, MS 30, fol. 129; Don M. Randel, An Index to the Chant of the Mozarabic Rite (Princeton: Princeton University Press, 1973); Levy, “Old-Hispanic Chant In Its European Context,” in Españ~~a en laúMdsica de Occidente (Madrid: Ministerio de Cultura,

Paris,

Bibliothèque

Madeleine Bernard),

1987),

vol. 1, 3-14.

simplest explanation would be that despite their manifest differences, the minority melodies are not independent formulations, but instead are decayed readings of the written Vulgate. Erosions of melodic integrity are not uncommon among Gregorian readings of the tenth century and later. They result from a weakening bond between The

the campo aperto numes, in which the chants were first recorded, and the memories of singers upon which the early, incomplete neumatic records depended for essential information, mainly about pitch, that they left out. In a centuries-long development, singers’ memories went their own way and differing opinions about pitch-levels and details of melodic contour became encrusted in local written traditions. The differences between the Vulgate and minority Elegerunts be attributable to such decay. Yet if they are, one should expect may to find similar traces of decay among other Offertories in the same regional recensions. But that does not happen. The differences elsewhere do not approach those among the Elegerunts. In the Aquitaine tradition, nearly all other Offertories agree quite well with the mainline Gregorian readings that are represented in early, carefully neumed witnesses like Laon 239, Chartres 109, and St. Gall 339. That is so with Offertories based on psalmic texts, like Ad te Domine, 68 levavi, representing a putative Roman type where stricter musical conformity might be expected; but it is also so with Angelus Domini, 69 whose text is a “non-psalmic libretto” of the same “Gallican” type as Elegerunt, making it a likelier candidate for comparable melodic discrepancy. The same also applies to the Offertories in the Auxerre tradition, where neither Ad te levavi nor Angelus Domini departs significantly from the standard Gregorian neumation. What differences there are represent the time and place, as can be judged from the proximity of Auxerre’s reading of Ad te levavi to that of a contemporary one from nearby Nevers, as published by van der Werf. 70 The minor differences are of an order that would seem a normal consequence of the centuries of regional transmission that intervened between the promulgation of the Carolingian neumed recension and these descendents. Much the same can be said of the Offertories in the Beneventan recension, where both Ad te levavi and Angelus Domini are in the main line of campo aperto neumations, 71 and so also of the 68

Paléograpbie

musicale, vol. 13, fol. 2. vol. 13, fol. 155. 70 Paris, Bibliothèque Nationale, nouv. acq. lat., 1235, fol. 9V; van der Werf, The Emergence of Gregorian Cbant, vol. 1, part 2, 166-78. 71 Ad te levavi, Benevento VI. 34 (Paléograpbie musicale, vol. 15), fol. Iv; Angelus Domini, Benevento VI.34 (Paléograpbie musicale, vol. 15), fol. 132. 69

Paléograpbie musicale,

Anglo-Saxon

and “Northern” (Paris-Arras-“Paleofrank”) recensions. decay does not account for the differences in

In sum, casual melodic

Elegerunt's multiples. If not decay, then perhaps there was a more purposeful process of “reimprovisation” or “reconstruction,” of the kind recently proposed by van der Werf, Treitler, and Hucke as explanation for differences 72 among Gregorian readings? The notion is that the early neumations did not represent fixed melodic entities. Instead they were points of reference and departure for scribes and singers who during some early centuries of the written transmission continued to exercise age-old improvisatory freedoms. The variant Elegerunts would result from taking one of the neumed versions–perhaps the Vulgate–as the point of departure for elaboration. Yet the melodic variants in mainline Gregorian chants that have been interpreted as the outcome of active reimprovisatory endeavor seem better explained by the passive processes of notational and melodic “decay” just described. The differences between readings are of a minor order, turning on a particular choice of pitch or reciting tone, or on the altered contour of a small melodic group. They show up in versions noted on the staff, and so are at a considerable remove in time and notational technique from the campo aperto neumes that lay at the root of the tradition. Concerning the general run of Gregorian variants, it seems to me that David Hughes was quite correct in observing that “the inference to be drawn is not that the melody was composed anew by improvisation at each performance, but rather that certain kinds of details were somewhat flexible.” 73 With the Elegerunts, the variants are more substantial than elsewhere, so that they more than any others might be identified as the results of reimprovisatory operations. Yet they differ enough among themselves in neumation, style, and form (different text-syllables paired with different melodic substances) that it is difficult to

see

any one of them as the melodic generator of any of are likelier to represent three independent written

the others. They formulations. Two further assumptions in the theories of “decay” and “reimprovisation" also need review: one, that the Vulgate was the oldest Elegerunt; the other, that it began as the most authoritative version. Neither of these can be sustained. The Vulgate became the majority

72

See note 18. “Evidence for the Traditional View of the Transmission of Gregorian Chant,” 398; instead of considering the melodic details “somewhat flexible,” however, I would prefer to say that local memories turned “somewhat fallible.” 73

us as the authoritative version. In the agreement of its neumations in the Carolingian East and West, there is also the mark of its attachment to a mainline Gregorian recension that circulated before Charlemagne’s empire underwent its major political division during the 840s. Yet each of the minority traditions (Aquitaine, Benevento, “Northern,” and Old Hispanic) has a claim to considerable antiquity. Each goes back to a formulation in staffless neumes, and the archaic “Paleofrankish” neumation for Paris-Arras“North” bids to be more ancient than any of the rest. In this company, the Vulgate need not be the oldest, nor the one that at the outset stood first among equals. What may be a better explanation of Elegerunt s multiples would begin with their identification as the musical favorites of different Gallican regions. The geography is clear for Aquitaine/Auxerre, which appears in adjacent zones of the French southwest and Burgundy. It is similarly clear for Paris-Arras-“North,” which has a habitat closer to the Carolingian heartland. Inasmuch as the cognate versions in outlying Benevento and Kent have the same Gallican-style (centonate-libretto) text as all the other Elegerunts, and also the same basic modal-melodic substance as Aquitaine-Auxerre and the Vulgate, they should represent still another Gallo-Carolingian region. And that should also be the case for the Vulgate. In light of the generalities of musical style that link the Vulgate version with the main line of standard Gregorian Offertories, to narrow its origin by means of the other regional Elegerunts should help to focus the search for the pocket of editorial activity—still a prime mystery—where the authoritative Carolingian musical corpus received its definitive melodic and neumatic shaping. Now if the Elegerunts were the musical favorites of different Carolingian regions, the question can be asked, whether they were long-established in those regions, or creations that came relatively late? There are reasons for seeing them as the relics of old and entrenched Gallican traditions. The parallel Eastern and Western neumations for the Vulgate Elegerunt have indicated that this version was circulated by the middle of the ninth century. For the others, a number of hypothetical situations would explain their survival. An authoritative Gregorian repertory might arrive without an Elegerunt at a place that had its own favorite version, which was then annexed to the arriving repertory. Or an authoritative Gregorian repertory might arrive with the Vulgate Elegerunt at a place that had its own favorite, and the import was rejected and supplanted by the local one. Or an authorized Gregorian repertory might arrive with the Vulgate at a

version, and it reaches

that did not want any Elegerunt, and the import was ignored, local use continuing (or taking up) with some other Stephen Offertory like In virtute tua or Posuisti. Each of these situations gives the local versions a certain age and autonomy. The most interesting case would be where the local favorite replaces the arriving Vulgate; this is something to which I will return. Yet it must already be apparent that the Elegerunts promise better than most other cases of multiples. Their differences in neumation, style, and form speak for writings-down that took place independently. Since three of them (the Vulgate, Aquitaine-Auxerre, and Benevento-Kent) share modal-melodic materials, the substance that they share must have a common source, and if their neumations were originally as separate and independent as they seem to be, then that source itself was likely to be oral. There is a remote chance that some archaic written formulation, perhaps cast in a Hellenistic-derived alphabetic or symbolic notation, transmitted a melodic shape that was the common source. But there is no indication that such notations were used to record large amounts of Latin plainchant before neumes, and there is no apparent reason for Elegerunt to receive special

place

treatment.

Now the

Elegerunts'

status as

regional

favorites

notion of their ultimate oral existence. If they

creations, of the later ninth

were

helps support the latter-day melodic

eleventh centuries, this would times when the Gallican traditions whose regional flowerings they seem to represent were overshadowed by the authoritative Carolingian-Gregorian tradition. Once the Gregorian was established in a region, the chances for fresh musical essays with Elegerunt's Gallican-style text greatly diminished. What the multiples suggest instead are earlier times of continuing Gallican vigor, times when the conversions to neumes reflect a climate where local rites still flourish. All four of the main Carolingian traditions (Vulgate, Aquitanian, Beneventan, Northern) are therefore likely to exist by the later eighth century, perhaps in written form already at that time, and in a purely oral form perhaps much farther back. What began as the possibility that the Elegerunts had oral antecedents has now turned into what seems a likelihood. In that case, there is yet another set of questions to address. They concern the relationship between the oral and written melodic states: in brief, to what extent do the neumed versions of these sometime Gallican favorites perpetuate the substances of oral deliveries that went before? For the three Elegerunts that survived long enough to reach pitch-specific notations, an answer comes from their relationships as “close multicome at

through

Example

5

Elegerunts V, A,

and B

and the Vulgate are exceptional in the of melodic detail they share. Their musical differences have been considered above in connection with Examples 1 4 Each was seen as having a distinctive melodic style, a particular set of internal parallels, and an independent neumatic disposition. Now it is their proximities that are important, and these can be judged in Example 5 There are ample agreements of detail on corresponding syllables, words, and phrases. They run throughout, and they are particularly

ples.” Aquitaine, Benevento, amounts

-

.

.

notable in the segments of fabric that

parallels, “accipe.”

as

in the

agreements

on

tied into

form-building “quern,” and (9) for converting a minority version are not

(2)“levitam,” (5)

Earlier I gave as a reason the desire to sustain a local favorite in the face of the arriving Vulgate. Yet the Vulgate differs in only small ways from the minority versions. The local musicians must then have set considerable store in the small differences. In rejecting the Vulgate’s similar were on their own details, they insisting preferences; they were them with care. In effect, they have left accurate transcriptions recording of the local oral deliveries. to neumes

If

Elegerunt’s close multiples tell us this much, they also do one thing more. They answer a question about the nature of the melos during oral transmission. Did a considerable improvisatory license persist throughout the purely oral phase, or was there a turn toward more calculated methods of melodic production? These multiples say

that the oral melos became stable and remembered. There is no better explanation of the musical ties between Elegerunt's independently neumed regional recensions than as parallel outgrowths of a melodic formulation that before neumatic fixing acquired a fixed, memorized

profile.

Speculative this is, and speculative it is bound to remain. Yet an oral Elegerunt that was the source for the written ones seems to have been “memorized,” not “improvised.” In Nettl’s spectrum of improvisatory processes, its singers were no longer “improvising upon something,” they were “performing a version of something.” 74 At face value, this applies only to a single chant, and it is one whose lineage is not Gregorian but Gallican. Alongside it can perhaps be placed the processional Antiphon Deprecamur te Domine, another case of close multiples, whose lineage is properly Gregorian. This might end here, with the two isolated cases, but consideration must also be given to the general correspondences in musical style that link these two chants with others in their categories of Gallican Offertories and Gregorian processional Antiphons. The Vulgate Elegerunt is related to Offertories of its distinctive text-type (non-psalmic, centonate librettos) in the Carolingian-Gregorian recension, 75 and to the general style of the Gregorian Offertories. Similarly, Deprecamur te, in its manifestations as “Carolingian” and Beneventan close-multiples, is related to other processional Antiphons of presumable Italic origin that circulate with the main Gregorian collection. In this way, suppositions about a memorized melos during later oral stages can be extended to considerable portions of both those major repertories. In the end, the question is one of memory: of how sizable a role it played during oral transmission, and how plausible that role would be. It has long been presumed that there was a significant reliance on memory in the stage before neume writing began, and that this was carried over into the relationship between singers’ memories and the techniques of campo aperto neuming that was the essence of the early written transmission. 76

74

75

Nettl, “Thoughts on Improvisation,” 9. Such as Angelas Domini; according to Paléograpbie musicak,

formulas

to Factus at

vol. 15,

167,

it lends

repente.

76 Dom Cardine wrote of a time when “la notation n’était pas encore inventée: la

mélodie des textes chantés était entiérement confiéeá la mémoire”; “Vue d’ensemble sur le chant grégorien,” Études grégoriennes 16 (1977): 173. Similarly Huglo has written of “les mélodies du répertoire liturgique conservées dans toutes les mémoires,” and of an early neumed stage when the tradition, “reste essentiellement basé sur la mémorisation” (p. 34); “Tradition orale et tradition écrite dans la transmission des mélodies gégoriennes,” in Studien zur Tradition in der Musik. Kurt von Fischer zum 60. Geburtstag, ed. Hans Heinz Eggebrecht and MaxLütolf (Munich: Musikverlag Katzbichler, 1973), 3 1-42; cf. 35, 34. Huglo accurately distinguished three stages: an early one of pure oral tradition based on memorization; a mixed one, where memory received the support of campo aperto neumes; and then a stage where the pitchspecific notation released the singers from dependence on memory.

Concerning memory-use after neumes were introduced, the nature of the neumes speaks for a considerable input from memory. These give precise details of duration and delivery with figures like the tractulus, episema, liquescences, oriscus, and quilisma (using their later Germanic names). 77 They show information about pitch less well; there are accurate charts of ups, downs, and repetitions, but there is little or nothing about pitch-levels and interval-widths. Inasmuch as the factors of length and ornament are represented in detail, it can scarcely be that just the pitch choices were left to the 78 vagaries of “re-improvisational” or “reconstructive” performers. It must rather be that the musicians who devised the economical neumatic system of graphing melodic events considered the specifics of pitch to be secure enough in professional memories to sustain this somewhat casual treatment by the early neumes. Essential to the system, therefore, was a dependence on memory for much of the information about pitch. At the point when the repertory was cast in neumes the singers’ memories were stocked with the full complement of melodic information. 79 Turning to the purely oral deliveries of a prior stage, it may be that the lack of

a

handy

written

means

for

consolidating

and

reviewing

the free end of kept plainchant the improvisatory spectrum throughout their oral transmission. Singers may have wanted it that way, as something that gave scope to their operations. Yet it may also be that the memory-usage that was the partner of the early campo aperto neumes was itself the continuation of a usage during a prior, purely oral stage. The close multiples of Elegerunt and Deprecamur have suggested this in their own way. It can also be observed that if the developers of the campo aperto system were engaged in first-time fixings of melodic substances that until then were produced as freely-improvised deliveries, they might have been less

compositional

deliveries

decisions

near

willing to consign as much of the fruit of their labors to memory as they did. The neumes in fact supply about the level of support one would expect for

practiced

77

Michel

(1954):

in

Huglo,

professional musicians whose memories were handling an essentially concretized repertory.

“Les

noms

des

neumes et

leur

already

origine," Études grégoriennes

I

53-7. 78 See note 18. 79 Jacques Handschin put this with characteristic clarity: “je ungenauer die Notierung, um so mehr müssen wir die alten Sänger ästimieren, die mit einer so rudimentären Gedächtnisbeihilfe die Melodien richtig zu singen vermochten.” Musikgeschicbte im Überrblick (Lucerne: Raber, 1948), 128.

Another sign of a turn toward fixity in later oral stages may be seen in the nature of certain melodies. Within the written repertory’s assortment of styles and procedures there are obvious distinctions to be made. For some kinds of melos, the chances of memorized oral antecedents are greater, with others less. For florid soloists’ chants where the fabric is an apparent elaboration of a “psalmodic” framework,

techniques of free, improvisational delivery may have prevailed throughout oral transmission; the versions fixed by the Carolingian neumators may be faithful images of former improvisational the

tional deliveries. On the other hand, for chants where “centonate” processes have a role, the situation may be different, and to some extent this affects the assessment of “psalmodic” chants where there is a considerable centonate component. Centonate chants have fabrics that are more or less densely woven with short formulas and patterns which may reappear in other contexts and combinations throughout an extensive repertory. Where there is a substantial centonate input, the chance of free improvisatory survival is lessened. The ultimate origins of centonate procedures may lie in age-old improvisatory techniques where select melodic elements were “hot-wired” in from volatile, artful memories. Yet in such fine-tooled creations as the fifth-mode Graduals or eighth-mode Tracts, the neumed entities that appear on ninth- and tenth-century parchment are likely to incorporate considerable amounts of compositional and editorial tinkerings that entered during written stages. For such musical texts there are analogies in the realm of ancient literary epic. But they are not with the Iliad or Odyssey. They are with such poetry as the Aeneid, which is

replete

with

archaic,

epic-type gestures,

but whose

deployment

of

those gestures is the result of a cooly calculated written process. Taken together, the deliberate centonate constructions and the florid psalmodic elaborations amount to only a small part of the repertory. For the large bulk of plainchant, the conditions of oral delivery may have been still different. The rationale for most Gregorian chants lies in distinctive melodic profiles, in formulations that

unique as to melodic detail.

80

In

Byzantine chant there is a parallel phenomenon. The chants are called “idiomela,” that is, “distinctive,” “unique,” “original” melodies. Hundreds are

usage, and

even a name

for the

of chants may draw their basic materials from a common fund of conventional modal-melodic gestures, but each of them is turned 80

Thus Hubert Sidler characterized the Gregorian Offertories as unique in their both as to melody and structure; Studien zu den alten Offertorien mit ibren Versen (Fribourg: Verlag des Musikwissenschaften Instituts der Universität, 1934), 7.

shapes,

distinct melodic entity, with singular twists given to the familiar gestures. Each chant is notable for those twists. This says that the idiomelic chants were by nature “memorable,” and since for many of them the written tradition goes back to the earliest noted states, there is the chance of a memorability reaching farther back. A further indication of this may be seen in another common feature of idiomelic repertories: certain pieces emerge as favorites and become musical models to which new texts are accommodated. In the Byzantine terminology again, a uniquely-profiled idiomelon becomes a prototype or generator automelon, the basis for imitated melodies or prosomoia. Automelic chants are generally cut of the same stylized cloth as the idiomelic chants that make up the bulk of their particular category. But the automela were so well fixed in choristers’ and congregational memories that in earlier hymn-books they rarely appear with neumes. It was sufficient to identify the model by its text-incipit and supply the fresh text. Automelic prototypes and their adaptations are also common in the West, with their most familiar examples among the 81 Gregorian Alleluias, as in the type Dies sanctificatus. The Byzantine musicians persisted longer in relying on memory, so that most of the model-melodies and even many of the adaptations were transmitted without neumes until after the fall of the Empire in the fifteenth century. The Latin musicians turned sooner to the regular neuming of both models and adaptations, so that even in the oldest neumed Graduals and Antiphoners the imitated versions are written out. Yet the Eastern and Western practices have in common that where the idiomelic chants with their distinctive profiles were by nature simply “memorable,” the automelic chants were by nature “remembered” and “memorized.” With these respected models, the chance that memorability reaches back of the earliest neumings is good. The role taken by prototypes and adaptations in filling out the Gregorian roster of the later eighth century indicates a similar role for verbatim

out as a

memory

81

during

a

prior

oral stage.

82

Schlager, Tbematiscber Katalog der ältesten Alleluia-Melodien (Munich: Verlag, 1965), 38-39; there are recent discussions of the practice by M. Huglo, “Antiphon,” The New Grove Dictionary of Music and Musicians, vol, 1, 473-76; K.-H.

Walter Ricke

Huglo, “Gradual,”

vol. 7,

anciens,” Études grégoriennes 82

603; Jean Claire, “Les formules centons des alleluia (1981); 3-4 and 12 pages of charts. imitations are well represented among the Latin Offertories, 20

and their in the group that includes Viri Galilei where five texts are accommodated to the same melody; Baroffio and Steiner, “Offertory,” vol. 13, 516. Of particular interest is the Offertory Posuisti for St. Gorgonius of Metz, with music based on the

tories,

Prototypes as

From diverse viewpoints (close multiples, neumatic technique, idiomelic-automelic practice) an image of the later oral transmission has emerged in which much of the melos was memorized rather than freely improvised. There have long been common-sense reasons for supposing this was the case, but it has never been taken so near to proof. Now the final point to be considered is that of plausibility. Were the memories of church singers without the support of musical notations capable of managing the amounts of melodic substance that this implies? The eighth- and ninth-century Gradual contained about 560 chants: 70 Introits, 118 Graduals, 100 Alleluias, 18 Tracts, 107 Offertories, and 150 Communions. 83 The figure can be expressed in hours in performance, and so as a function of tempo. Something can be inferred about Carolingian plainchant tempos from the nuance-rich neumations of the Gregorian Mass Propers that are found in manuscripts like Saint-Gall 359 and Laon 239. If the scribes took pains to indicate such details of length and ornament, it was because the niceties were supposed to be manifested in performance, and the deliveries must have gone at a pace moderate enough to allow those effects to be heard. Estimates of duration can be based on present-day performances, where Introits take about 3 minutes each (much of the time for the psalm-tone and repeated Antiphon); Graduals about 3 ½ minutes; Alleluias with verse about 2 minutes; Tracts, 5; Offertories with verses, 8; Communions, 2 to 3. A tally that may run generously high would put the aggregate for the late-eighth-century Mass Propers at about 35 hours of moderately-paced music, the largest single segment (about 14 hours) going to the Offertories and their verses. Comparisons with other musical repertories are bound to be misleading, with different conditions (polyphonic, instrumental, written, sound-recorded, etc.) always involved. For the thirty-five hours of “memorable” plainchant, a corresponding portion of Beethoven’s output would be the symphonies, sonatas, and quartets; of Wagner’s output, the canon of ten mature dramas minus one, as from Tannbäuser through Parsifal. Wagner is the better comparison, with music that is text based and often moderate in pace. The figures for the Mass Propers need to be augmented by those for the Office, where the first full Antiphoner with neumes (dating from about 1000) contains about six hundred Greater Responsories, a

Offertory Angelus Domini; Posuisti may have been compiled soon after the translation of the saint’s body to Metz around 754; Hesbert, Antipbonak Missarum Sextuplac, CVI and no. 148bis. 83 Huglo, Les livres de cbant liturgique, 102.

number that tends to increase

thereafter

rapidly

notations rendered fresh

pitch-specific practical. 84 The Antiphoner’s

calls

on

as

the

availability

of

compositional essays Carolingian singer’s memory

more

the

ory would be limited, however, by the considerable amounts of stereotyped matter in the refrains and verses of the Greater Responsories, and by the thousands of Office Antiphons that are modeled on between four and five dozen melodic prototypes. At another rough estimate, the Antiphoner would add thirty-five or forty hours of distinctive melos, so that, in a provisional tally of the music of the Gregorian Propers, about 800 might come to seventy-five or eighty hours of memorized matter. This would correspond to the selection of Beethoven’s instrumental works plus the full Wagnerian canon. In a final accounting, there would be further reductions. Although prototype melodies are less common among the Mass chants than those for the Office, they are well represented among the Graduals and Offertories, in addition to the Alleluias mentioned above. It must also be supposed that among the psalmodic chants and the centonate chants there was, for the diverse reasons already given, less dependence on fixed-memorized matter than among the idiomelic chants. That further lightens the burden of Carolingian memory.

In

a

recent

paper I

proposed

as

a

major accomplishment of

the later

Carolingian liturgists during eighth century the shuttingdown of what remained of improvisatory freedom in the “Gregorian” plainchant environment; it was done through the promulgation of an authoritative melodic repertory whose substances were concretized in neumes. 85 In the present paper I have tried to look behind the introduction of neumes toward a stage where transmission was entirely unwritten. The true face of plainchant orality will never be

revealed,

and efforts

to

view

From the

it, by

whatever means,

are

bound to

situation examined here it appears speculative. that substantial amounts of Latin plainchant while still in oral transmission acquired the status of song stored in memory. The argument depends on “close multiples,” where a chant survives in parallel readings that are musically related but notationally independent. For the Offertory Elegerunt apostoli there are three versions remain

84 Paul F. Cutter, vol. 14, 759. 85

“Responsory,”

“Charlemagne’s Archetype

of

rare

The New Grove

Dictionary of Music and Musicians,

Gregorian Chant,”

1-30.

and “Vulgate”) that evidently represent the melodic preferences of different Gallican-Carolingian regions. There is a chance they were taken out of the oral state and put into writing at a relatively recent time, perhaps during the later ninth through eleventh centuries; or even that they were fashioned as fresh compositions in the medium of neumes at that time. Yet the local Gallican traditions that Elegerunt's multiples represent were by then largely obsolete because of the spread of the Gregorian musical usage through Carolingian liturgical dependencies during the later eighth to early ninth centuries. It is likelier therefore that the regional versions existed at earlier times of continuing Gallican vigor, and perhaps at times before neumes. Despite the apparent independence of their written traditions, the three musical substances are very much alike, and that adds the thrust to their testimony. In the narrow range of differences between them can be guaged the precision with which the locally-prized oral deliveries were recorded. More than that, in order for them to emerge so much alike from the independent processes of transmission, the “oral” substance that they shared must itself have had the shape of a whole, remembered melody. This may apply only to the Gallo-Gregorian offertory Elegerunt, for which an unusually rich documentation survives. Yet arguments of liturgical-musical analogy suggest extending its application to large parts of the early repertory. If some of this is granted, it would seem worthwhile to seek out and analyze other cases of close multiples. It is by their means, and perhaps their means alone, that we may gain a clouded window on the oral past.

(Aquitainian, Beneventan,

Princeton

List

of

Works Cited

Manuscripts Albi, Bibliotheque Municipale Rochequde, MS 44. Baltimore, Walters Art Gallery, MS W. 6. Baltimore, Walters Art Gallery, MS W. 11. Benevento, Biblioteca Capitolare, MS 33. Benevento, Biblioteca Capitolare, MS 34. Cambrai, Bibliotheque Municipale, MS 75 (76).

Cambridge, Corpus

Christi

College,

MS 473.

Einsiedeln, Stiftsbibliothek, MS 121 (1151). Leon, Biblioteca Catedral, Cod. 8. London, British Library, Add. 12194. London, British Library, MS Cotton Caligula A.

14.

University

London, British Library, Harleian MS 4951. Madrid, Academia de la Historia, MS Aemil. 30. Montpellier, Faculté de Médecine, MS H. 159. Oxford, Bodleian Library, MS Bodley 775. Oxford, Bodleian Library, MS Douce 222. Paris, Bibliotheque de 1’Arsenal, MS 1169. Paris, Bibliotheque Nationale, fonds latin, MS 776. Paris, Bibliotheque Nationale, fonds latin, MS 780. Paris, Bibliotheque Nationale, fonds latin, MS 903 Paris, Bibliotheque Nationale, fonds latin, MS 9436. Paris, Bibliotheque Nationale, fonds latin, MS 10508. Paris, Bibliotheque Nationale, fonds latin, MS 10511. Paris, Bibliotheque Nationale, fonds latin, MS 17305. Paris, Bibliotheque Nationale, fonds nouv. acq. latines, MS 1177. Paris, Bibliotheque Nationale, nouv. acq. latines, MS 1235. Rome, Biblioteca Angelica, MS 123. Vatican City, Biblioteca Apostolica Vaticana, MS Vat. lat. 4770. Zurich, Biblliothek Cantonale, MS 23.

Modern Editions and

Literature

Secondary

“

Chant .” The New Oxford History of Music vol. 2 127 London : Oxford University Press 1954 Antifonario visigótico mozárabe de la Catedral de León. Facsimiles musicales vol. 1 Monumenta Hispaniae Sacra, Serie litúrgica, vol. 5, part 2. Barcelona and Madrid : Consejo Superior de Investigaciones Cientificas 1959 Antiphonale Missarum Mediolanensis Rome : Desclée et Socii 1935 Apel Willi The Central Problem of Gregorian Chant .” This JOURNAL 9 ( 1956 ): 118 27 Gregorian Chant Bloomington : Indiana University Press 1958 Terence The Ambrosian Alleluias London : The Plainsong and Bailey Mediaeval Music Society 1983 Baroffio Giacomo Bonifacio and Ruth Steiner Offertory .” The New Grove Dictionary of Music and Musicians vol. 13 513 17 London : Macmillan

Angles Higini Gregorian .

,

92

,

,

-

.

.

,

.

,

.

,

.

.

,

“

.

,

-

.

.

.

_____

.

,

.

,

.

,

.

“

,

.

,

-

,

1980

.

,

,

.

“

Cardine Eugène De l’édition critique du Graduel .” In Higini Anglés ed., Atti del Congresso Internationale di Musica Sacra 187-91 Rome 1950 Reprint. Toumai: Desclée, 1952 émiologie grégorienne .” Études grégoriennes n ( 1970): 1 158 Vue d’ensemble sur le chant grégorien .” Études grégoriennes 16 ( 1977): 173 92 Claire Jean L’Evolution modale dans les répertoires liturgiques occidentaux," Revue grégorienne 40 ( 1962 ): 196 210 229 45 ; 41 ( 1963 ) 7 29 49 62 77 102 + tables. Les formules centons des alleluia anciens .” Études gregoriennes 20 ( 1981 ): 3 4 + 12 pp. of charts. Les repertoires liturgiques latins avant l’octoéchos. I: L’Office férial romano-franc .” Études grégoriennes 15 ( 1975): 5 192 .

,

,

.

.

,

.

“

-

_____ .

.

“

. ______

-

.

“

.

,

-

-

,

-

_____

-

,

-

,

. "

-

“

______ .

-

.

Music in Greek and Roman Culture Baltimore : Johns Hopkins University Press 1989 Connolly Thomas H. Introits and Archetypes: Some Archaisms of the Old Roman Chant .” This JOURNAL 25 ( 1972): 157 74 Corbin Solange Die Neumen Palaeographie der Musik nach den Plänen Leo Schrades vol. 1 sect. 3. Cologne : Amo Volk-Verlag 1977 Les neumes .” In Histoire de la musique vol. 1 689 98 Edited by ______." Roland Manuel Encyclopedic de la Pléiade, no. 9 Paris : Gallimard Comotti , Giovanni

.

.

.

,

“

,

-

.

.

,

.

,

.

,

,

-

.

,

,

.

1960

.

,

.

.“

Les notations neumatiques en France a l’époque carolingienne .” Revue d'histoire de I'église en France 38 ( 1952): 226 28 Corbin Solange with Madeleine Bernard Répertoire de manuscrits médiévaux Paris : contenant des notations musicales, vol. 3: Bibliothéques parisiennes Centre Nationale de la Recherche Scientifique 1974 Cutter Paul F. Musical Sources of the Old-Roman Mass Musical Studies and Documents, vol. 39 American Institute of Musicology 1979 .“ The Old-Roman Chant Tradition: Oral or Written? This JOURNAL 20 ( 1967 ): 167 81 .“ Oral Transmission of the Old-Roman Responsories? The Musical Quarterly 62 ( 1976 ): 182 94 .“ Responsory .” The New Grove Dictionary of Music and Musicians vol. 14 759 64 London : Macmillan 1980 An Unknown Complete Ninth-Century Gradual and Emerson John Antiphoner from Southern France: Albi, Bibliothèque Municipale Rochequde, MS 44 .” Unpublished paper read at the Annual Meeting of the American Musicological Society Baltimore 1988 Ferand Ernest Improvisation .” In Die Musik in Geschichte und Gegenwart vol. 6 cols. 1096 1100 Kassel : Bärenreiter 1957 Ferretti Paolo Estetica gregoriana, ossia Trattato delle forme musicali del canto gregoriano vol. 1 Rome : Pontificio Istituto di Musica Sacra 1934 Fischer Rupert Offertoires neumés avec leurs versets d'aprés les manuscrits Laon 239 et Einsiedeln 121 Solesmes : Abbaye Saint-Pierre 1978 Frere Walter Howard ed. Graduate Sarisburiense London : The Plainsong and Mediaeval Music Society 1894 Gauntlett Stathis Rebetika Carmina Graeciae Recentioris: A Contribution to the Definition of the Term and the Genre rebetiko tragoudi through Detailed Analysis of its Verses and of the Evolution of its Performance D. Phil, diss., Oxford University, 1978 Athens, Greece : Denise Harvey 1985 Graduate Triplex Solesmes : Abbaye Saint-Pierre 1979 Le Graduel romain, Édition critique par les moines de Solesmes. Vol. 2: Les sources Solesmes : Abbaye Saint-Pierre 1957 Handschin Jacques Musikgeschichte im Uberrblick Lucerne : Räber 1948 Hansen Finn Egeland The Grammar of Gregorian Tonality 2 vols. Copenhagen: Dan Fog Musikforlag 1979 ed. Hi39: Montpellier: Tonary of St. Benigne of Dijon 2 vols. Copenhagen : Dan Fog Musikforlag 1974 Hesbert René-Jean Antiphonale Missarum Sextuplex Brussels : Vromant ,

______

-

,

,

.

.

.

.

,

.

,

.

.

,

”

______

-

.

”

______

-

.

_____

,

-

.

,

.

,

“

.

,

,

,

.

,

.

“

.

,

,

-

.

,

.

,

.

,

.

,

.

,

.

,

.

,

.

,

.

,

.

,

.

.

.

,

.

.

,

.

.

,

,

.

.

.

,

_____

,

.

,

.

.

,

,

.

,

1935

.

.

.

.

“

Die Anfänge der abendländischen Notenschrift .” In Festschrift Rudolf Elvers edited by Ernst Herttrich and Hans Schneider 271 88 Tutzing : H. Schneider 1985 Die Anfänge der Bearbeitung .” In Bearbeitung in der Musik. Colloquium Kurt von Fischer zum 70. Geburtstag edited by Dorothea Baumann Schweizer Jahrbuch für Musikwissenschaft, Neue Folge, 3 ( 1983 ):

Hucke

Helmut

,

.

,

,

-

.

.

,

“

______

.

,

,

15

-

20

.

“

Fragen .” Die Musikforscbung 41 ( 1988 ): 304 30 im Gregorianischen Gesang .” Kirchenmusikalisches Improvisation Jahrbuch 38 ( 1954): 5 8 Musikalische Formen des Offiziumsantiphonen ,” Kirchenmusikalisches . _ _ _ _ Gregorianische

-

.

“

._ _ _ _

-

.

“

. _______

37 ( 1953): 7 33 Die Texte der Offertorien .” In Speculum Musicae Artis: Festgabe für Heinrich Husmann zum 60 Geburtstag edited by Heinz Becker and Reinhard Gerlach 193 203 Munich : Wilhelm Fink Verlag 1970 . _ _ _ Toward a New Historical View of Gregorian Chant .” This

Jahrbuch

-

.

“

______ .

,

-

.

,

.

,

“

33 ( 1980 ): 437 67 Der Übergang von mündlicher zu schriftlicher Musiküberlieferung ___ im Mittelalter .” In Report of the Twelfth Congress of the International Musicological Society, Berkeley, 7977 edited by Daniel Heartz and Bonnie Wade 180 190 Basel : Bärenreiter 1981 Zur Entwicklung des christlichen Kultgesangs zum Gregorianischen Gesang .” Römische Quartalschrift 48 ( 1953): 172 85 Hughes David G. Communication .” This JOURNAL 41 ( 1988 ): 578 79 Communication .” This JOURNAL 42 ( 1989): 435 37 Evidence for the Traditional View of the Transmission of Gregorian Chant .” This JOURNAL 40 ( 1987 ): 377 404 Huglo Michel Altgallikanische Liturgie .” In Karl Gustav Fellerer ed., Geschichte der katholischen Kirchenmusik vol. 1 219 33 Kassel : Bärenreiter

Journal

-

.

“

.

,

-

.

,

.

,

“

_____

.

-

.

“

-

.

,

“

-

. _____

.

“ . _____

-

.

“

.

,

,

-

,

.

,

,

1972. “

. _ _ _ Antiphon .” The New Grove Dictionary of Music and Musicians vol. 1 471 81 London : Macmillan 1980 _ _ _ Gallican Rite, Music .” The New Grove Dictionary of Music and Musicians vol. 7 113 24 London : Macmillan 1980 ,

,

-

.

.

,

“ .

-

,

.

,

,

.

“

._ _ Gradual .” The New Grove Dictionary of Music and Musicians vol. 7 601 9 London : Macmillan 1980 Les livres de chant liturgique Typologie des sources du moyen âge occidental Directeur L. Genicot. Fasc. 52. Tumhout : Brepols 1988 Les noms des neumes et leur origine .” Etudes grégoriennes 1 ( 1954): ,

,

-

.

.

,

._ _ _

.

.

,

.

,

“

.

___

53 67. Review of Finn Egeland Hansen, The Grammar of Gregorian Tonality ._____ This JOURNAL 37 ( 1984 ): 416 24 _ _ _ Le tonaire de St. Bénigne de Dijon .” Annales musicologiques 4 ( 1956): —

.

-

.

“ .

718-

.

“

Tradition orale et tradition écrite dans la transmission des mélodies In Studien zur Tradition in der Musik: Kurt von Fischer zum 60. Geburtstag edited by Hans Heinz Eggebrecht and Max Lütolf 31 42 Munich : Verlag Katzblicher 1973 ___ ..

grégoriennes .”

-

,

,

,

.

.

.ed., ______ Musicologie médiévale: Notations et séquences. Actes de la Table Ronde du C. N. R. S. àl'Institut de Recherche et d'Histoire des Textes: 6-7 septembre 1982 Paris : Centre Nationale de ia Recherche Scientifique 1987 Husmann Heinrich Tropen- und Sequenzenhandschriften Répertoire international des sources musicales, vol. B V' Munich : G. Henle Verlag 1964 John the Deacon. S. Gregorii Magni Vita In Jacques Paul Migne ed., series Latina vol. 75 col. 59 242 Paris : Patrologiae cursus completus Gamier 1862 Kelly Thomas Forrest The Beneventan Chant Cambridge : Cambridge University Press 1989 Klöckner Stephen Analytische Untersuchungen an 16. Introiten im I. Ton des altrömischen und des fränkisch-gregorianischen Repertoires hinsichtlich einer bewussten melodischen Abhängigkeit .” Beiträge zur .

.

,

.

.

,

.

.

,

.

,

-

.

.

.

,

.

,

.

,

.

,

.

.

,

“

.

,

Gregorianik 5 ( 1988 ): 3 95 Lambillotte Louis Antipbonaire de Saint Grégoire. Facsimile du manuscrit de Saint-Gall Brussels, 1851 ; 2nd ed., Brussels : CH.-J.-A. Greuse 1867 Levy Kenneth The Byzantine Sanctus and its Modal Tradition in East and West .” Annales musicologiques 6 ( 1958 63 ): 7 67 Charlemagne’s Archetype of Gregorian Chant .” This JOURNAL 40 -

.

.

,

.

.

,

“

.

,

-

-

.

“

_____ .

( 1987 ):

1

-

30 Communication .” This JOURNAL 41 ( 1988 ): 575 78 .“ A Gregorian Processional Antiphon .” SchweizerJahrbuch für Musikwissenschaft Neue Folge, 2 ( 1982 ): 91 102 The Italian Neophytes’ Chants .” This JOURNAL 23 ( 1970): 181 227 Latin Chant outside the Roman Tradition .” The New Oxford History _____ . of Music (new edition), vol. 2 edited by Richard L. Crocker and David Hiley 69 110 Oxford : Oxford University Press 1990 Old-Hispanic Chant In Its European Context .” España en la Música de Occidente vol. 1 3 14 Madrid : Ministerio de Cultura 1987 On the Origin of Neumes .” Early Music History 7 ( 1987 ): 59 90 Plainchant .” The New Grove Dictionary of Music and Musicians vol. 800 5 London : Macmillan 1980 14 Rome and the Legacy of Gaul .” Early Music History 4 . _____ Toledo, .

“

-

_____ .

.

_____

,

-

.

“

-

_____ .

.

“

,

,

-

.

,

.

,

“

. _____

-

.

,

,

.

,

“

-

. _____

.

“

. ____

,

-

.

,

.

,

“

( 1984 ):

49 99 The Trisagion in Byzantium and the West .” In Report of the Eleventh Congress, International Musicological Society, Copenhagen, 1972 edited by Henrik Glahn Søren Sørenson and Peter Ryom 761 65 Copenhagen : Bärenreiter 1974 Lipphardt Walter Gregor der Grosse und sein Anteil am römischen Antiphonar .” In Higini Angles ed., Atti del Congresso Internazionale di Musica Sacra, Roma, 1950 248 54 Tournai : Desclée 1952 Lord Albert B. Oral Poetry .” In Princeton Encyclopedia of Poetry and Poetics edited by Alex Preminger et al. 591 92 Princeton : Princeton University Press 1965 The Singer of Tales Cambridge : Harvard University Press 1960 Nettl Bruno Thoughts on Improvisation: A Comparative Approach .” The Musical Quarterly 60 ( 1974): 1 19 -

.

“

.

_____

,

-

,

,

,

.

.

,

“

.

,

,

-

.

,

,

.

“

,

,

-

.

,

.

,

. _____

.

,

“

,

.

-

.

.

“

of Tradition and Transmission .” In Cross-Cultural Perspectives Music edited by Robert Falck and Timothy Rice 3 19 Toronto : University of Toronto Press 1982 Notopoulos James A. Modern Greek Heroic Poetry New York : Ethnic Folkways Library Album FE 4468 1960 Ott Charles ed. Offertoriale sive Versus Offertoriorum Cantus Gregoriani Tournai : Desclée 1935 Paléographic musicale: Les principaux manuscrits de chant grégorien, ambrosien, mozarabe, gallican, publiés en fac-similés phototypiques par les Bénédictins de Solesmes Edited by A. Mocquereau J. Gajard et al. Series 1 20 vols. Solesmes : Imprimerie Saint-Pierre Toumai: Desclée, and Berne: Lang, 1889 Reprint. Vols. 1-15. Berne: Lang, 1968 74 Series 2 (monumentale), 2 vols. Toumai: Desclée, 1900 24 Reprint. Berne: Lang, 1970 Planchart Alejandro Enrique The Repertory of Tropes at Winchester 2 vols. Princeton : Princeton University Press 1977 Randel Don Michael An Index to the Chant of the Mozarabic Rite Princeton : Princeton University Press 1973 Rankin Susan Neumatic Notations in Anglo-Saxon England .” In Michel Huglo ed., Musicologie médiévale: Notations et séquences. Actes de la Table Ronde du C. N. R. S. àl'Institut de Recherche et d'Histoire des Textes: 6-7 septembre 1982 129 44 Paris Centre Nationale de la Recherche Scientifique _____

Types

on

-

.

,

,

.

,

.

,

.

,

.

,

,

.

,

.

,

,

,

,

-

-.

.

-

.

.

.

.

,

.

,

.

.

,

.

,

“

.

,

,

-

1987

,

.

,

,

.

Roederer Charlotte ,

“ .

Can We

Identify

an

Aquitanian

Chant

Style?

”

This

27 ( 1974 ): 75 99 Schlager Karl-Heinz Thematischer Katalog der ältesten Alleluia-Melodien Munich : Walter Ricke Verlag 1965 Sidler Hubert Studien zu den alten Offertorien mit ihren Versen Fribourg : Verlag des Musikwissenschaftlichen Instituts der Universität 1939 Stablein Bruno Gregorius praesul,’ der Prolog zum romischen Antiphonale." In Musik und Verlag: Karl Vötterle zum 65. Geburtstag 537 61 Kassel : Bärenreiter 1968 . _____ Monumenta Monodica Medii Aevi, vol. 2: Die Gesänge des altrömischen Graduale Kassel : Bärenreiter 1970 Zur Frühgeschichte des römischen Chorals .” In Higini Anglès . _____ ed., Atti del Congresso Internazionale di Musica Sacra, Rome, 1950 271 75 Toumai : Desclée 1952 Treitler Leo ‘Centonate’ Chant: übles Flickwerk or E pluribus unus? This JOURNAL 28 ( 1975): 1 23 Communication .” This JOURNAL 41 ( 1988 ): 566 75 The Early History of Music Writing in the West .” This JOURNAL 35 . _____ ( 1982 ): 237 279 Die Entstehung der abendländischen Notenschrift .” Die Musikforschung 37 ( 1984): 259 67 Homer and Gregory: The Transmission of Epic Poetry and Chant .” The Musical Quarterly 60 ( 1974): 333 72 Oral, Written, and Literate Process in the Transmission of Medieval Music .” Speculum 56 ( 1981 ): 471 91 Reading and Singing: On the Genesis of Occidental Music-Writing Early Music History 4 ( 1984 ): 135 208

JOURNAL

-

.

.

,

.

.

,

.

,

.

.

,

‘“

.

,

-

.

,

.

,

.

.

,

“

,

-

,

.

.

,

“

”

.

,

-

.

“

-

. ____

.

“

-

.

“

. _____

-

.

“

. _____

-

.

“

. ____

-

.

“

."

_____

-

.

“

Turco Alberto Melodie-tipo e timbri modali nell’ Studi gregoriani 3 ( 1987 ): 191 241 Les répertoires liturgiques latins .” Études .

,

-

Antipbonale

romanum ."

.

“

( 1979 ): 177 van der Werf Hendrik Communication .” This JOURNAL 42 ( 1989): 432 4 The Emergence of Gregorian Chant: A Comparative Study of Ambrosian, Roman, and Gregorian Chant. Volume 1: A Study of Modes and Melodies Rochester : the Author 1983 Wagner Peter Einführung in die Gregorianischen Melodien. Vol. 1: Ursprung und Entwicklung der liturgischen Gesangformen 3rd ed. Leipzig : Breitkopf and Härtel 1911 Vol. 2: Neumenkunde, 2nd ed. Leipzig: Breitkopf and Härtel, 1912 Vol. 3: Gregorianiscbe Formenlehre. Leipzig: Breitkopf and . ____

gregoriennes

18

- 221.

“

-

.

.

,

. ____

.

,

.

.

,

,

.

.

,

.

Härtel, 1921

.

ABSTRACT The study considers various means of approach to the Gregorian melos during its oral transmission, before the conversion to neumes. Among these are intuitive analysis, based on the Carolingian received text; analogy with Balkan oral epic (“Homer and Gregory”); and “multiples” or parallel readings. An approach by way of a rare case of “close multiples” is explored in depth. The Gallo-Gregorian Offertory Elegerunt apostoli survives in parallel readings that are close in their musical substance but may be independent in their neumation. It suggests that during a later oral stage this particular chant, and perhaps a good deal of its cognate “idiomelic” repertory as well, had become melodically stable and memorized, and was no longer freely improvised. There have been common-sense reasons for supposing this, but nothing else takes it so near to proof.

[13] The Transmission of Western Chant in the 8th and 9th Centuries: Evaluating Kenneth Levy’s

Reading

of the Evidence EMMA HORNBY

the later

In

19th

and earlier 20th century, when

the Work concept was thoroughly ingrained in people’s outlooks, scholars assumed that the original version of Gregorian chant was composed and written down. The search for the Urtext

was real and pressing, alPope Gregory the Great—composer, conduit, figure-head, patron, or legend?—was and remains much discussed. By the third quarter of the 20th century, ideas of an oral origin of the chants became the accepted wisdom, championed most famously by Helmut Hucke and Leo Treitler. And soon after that, in a magisterial series of articles, Kenneth Levy shed doubt upon this new hypothesis, coming

though

the role of

1

2

This review-article grew out of extensive collaborative discussion with Rebecca Maloy, University of Colorado. Rebecca read, commented on, and contributed significantly to the paper at various draft stages. I gratefully acknowledge my debt to her; any mistakes and misinterpretations of the evidence that remain are my

own.

For a recent summary of his role, see David Hiley, Western Plainchant: A Handbook (Oxford: Clarendon Press, 1993), 502-14. Exemplified by the session chaired by Leo Treitler on “Transmission and Form in Oral Traditions,” at the IMS Congress in 1977. See Daniel Heartz and Bonnie Wade, eds., Report of the Twelfth Congress of the International Musicological Society, Berkeley, 1977 (Basel: Bärenreiter, 1981), 139-211. 1

2

DOI: 10.4324/9781315090283-16

Oral and Written Transmission in Chant

up with Hucke

new

rationales for the view that had held sway until Treitler and it. 3 While the overturning of received wisdom was in

challenged

1980s, Levy’s cumulative use of different kinds of evidence—documentary, paleographical, and musical—to reinvigorate the older theory that chant had been notated in Francia since the mid 8th century also appealed widely. 4 Twenty years later it is almost inevitable that the next generation of scholars should wish to synthesize and evaluate the positions held by their seniors. Both Levy’s earlynotation paradigm and theories of oral transmission have wide currency among chant specialists, who inevitably have to consider the merits of both. Special impetus was given to the need for scholars to come to terms with Levy’s work by the publication in 1998 of his Gregorian Chant and the Carolingians; and in response to the appearance of that work, James McKinnon, in his own book The Advent Project, urged that Levy’s idea of “a notated Carolingian archetype [should] become the

itself attractive in the

subject of the probing debate that it merits.” 5 My purpose in this review-article is to examine Levy’s the

main

hypotheses,

that underlie them, and his interpretation of and to demonstrate why I am not yet convinced by his arguments

assumptions

the

evidence,

9th

centuries. This does not

about the transmission of Western chant

during

the 8th and

that he is wrong, but only that I can equally plausible ways of reading the available mean

alternative and, I hope, 8th- and 9th-century evidence. Many of the ideas in this article have been expressed before, but the chief elements of Levy’s hypothesis

see

have not

previously been

evaluated

in what follows that readers will

systematically in

already

one

be familiar with

6

place. I assume Levy’s writings. 7

3 All references in this article are to the collected volume of Levy’s essays, Gregorian Chant and the Carolinians (Princeton: Princeton Univ. Press, 1998). Hucke’s exasperation with what he saw as a backwards historical step was patent. See “Gregorianische Fragen,”

Musikforschung 41 (1988):1326-30. 4 The English A capella hypothesis and subsequent pro-instrumental backlash come to mind as a similar overturning and reestablishment of previous assumptions. See Daniel Leech-Wilkinson, The Modem Invention of Medieval Music: Scholarship, Ideology, Performance (Cambridge: Cambridge Univ. Press, 2002). 5 James McKinnon, The Advent Project: The Later-Seventh-Century Creation of the Roman Mass Proper (Berkeley: Univ. of California Press, 2000), 442. 6 In much of his work, Leo Treitler has presented hypotheses opposed to those of Levy. The newly written chapter introductions in Treitler’s recent volume of collected essays summarize some of their differences in approach. See Leo Treitler, With Voice and Pen: Coming to Know Medieval Song and How It Was Made (Oxford: Oxford Univ. Press, 2003). 7 In his more recent work, Levy has overturned another fundamental premise of chant scholarship, namely that the Frankish melodies have their ultimate origins in Rome. He has proposed instead that a substantial portion of the offertory repertory is Gallican in origin. See Kenneth Levy, “A New Look at Old Roman Chant,” Early Music History 19 (2000): 81-104; and idem., “A New Look at Old Roman Chant II,” Early Music History 20 (2001): 173-98. Rebecca Maloy is engaged in evaluating these ideas, whose

Oral and Written Transmission in Chant

1. Historical Introduction Prior to the reign of Pippin (who seized the Frankish throne in 751), chant in Gallo-Roman Gaul and later in Frankish Gaul appears to have been lector chant, consisting usually of whole psalms sung by a soloist on an ad hoc basis. 8 Documentary evidence suggests that it was during or after the visit of Pope Stephen and his entourage to Francia in 754—55 that Pippin and his clerics were encouraged to adopt Roman chant. This entailed the adoption not simply of new texts and melodies, but also of

a new

institutional model—the Schola

of clerics whose chief the proper

The as

performance

adoption is

duty was

of chant and

of Roman schola chant did not

attested, perhaps

more

Cantorum—consisting in community, to

themselves, liturgy. 9

to devote

colorfully

than

proceed without incident, accurately, by Carolingain

and Italian reports of the continuing differences between Roman and Frankish chant. 10 Despite these differences, the Frankish Mass Proper

sufficiently firmly fixed by the time of the earliest surviving (late 9th century) to appear in largely consistent versions in manuscripts from as far apart as Brittany and Saint Gall. Scholars are still trying to establish how the Carolingian reformers melodies

were

notated Graduals

11

achieved the perceived level of melodic stability. Three alternative hypotheses are described by Treitler: that the oral tradition had settled into a fixed state; that the notators of the early manuscripts faithfully copied written exemplars, and that the uniform appearance of the manuscripts does not reflect the oral performance practice; or that the consistency of the sources is a result of the rules, melodic forms, and formulae which helped to maintain the fixity of the Levy’s “Toledo, Rome and the Legacy of Gaul,” Gregorian Chant and the Carolingians, 31-81, but her findings will not be anticipated here. 8 James McKinnon, “Lector Chant versus Schola Chant: A Quesdon of Historical Plausibility,” in Janka Szendrei and David Hiley, eds., Laborare Fratres in Unum: Festschrift László Dobszay (Hildesheim: Weidmann, 1995), 201-11. We lack a clear understanding of this native style of singing, its level of complexity, and the extent to which it prepared Frankish singers for performing the Roman chant repertory. 9 On the Roman Schola, see Joseph Dyer, “The Schola Cantorum and its Roman Milieu in the Early Middle Ages,” in Peter Cahn and Ann-Katrin Heimer, eds., De Musica et de Cantu: Helmut Hucke zum 60. Geburtstag (Hildesheim: Olms, 1993), 19-40. By the later 9th century John the Deacon maintained that the barbaric Franks

origins

lie in

10

had been unable to manage the subtledes of Roman chant, while the Saint Gall “Life of Charlemagne” (probably by Notker Balbulus) instead promoted the view that the Roman cantors had been willfully obstructive. The three surviving graduals from Rome (the earliest of which is dated 1071) directly attest to a melodic tradition (“Old Roman”) different from, yet related to, the “Gregorian” (Frankish) one. 11 On the consistency of the manuscript tradition, see Levy, Gregorian Chant and the Carolingians, 85; David Hughes, “Evidence for the Traditional View of the Transmission of Gregorian Chant,” Journal of the American Musicobgical Society 40 (1987): 377-401. However, see the discussion below for a nuancing of this view.

oral tradition. 12 The second of these hypotheses entails a supposition that the state of the manuscripts does not reflect the 9th-century performing tradition. Since this performing tradition is accessible only

through the notational record, the hypothesis has rarely been pursued (although the flexibility demonstrated in the early manuscripts in the matter of small rhythmic, melodic, and ornamental details will

further

be discussed further

below).

The first and third

hypotheses are closely related, differing only in explanation of how the observed consistency of the manuscripts may have come about. Some believe that highly trained clerical memories would have been capable of maintaining the

that the third entails

some

tradition intact and have turned

seeking metaphors, analogies,

to the

work of various other scholars

and models for their characterization

of the processes involved. The work of Parry and Lord on oral epic a starting point for Treitler’s study of oral traditions; 13 more recently

formed

he and others have drawn

on

the work of

Mary

Carruthers

on

memory training in the Middle Ages 14 and of David Rubin on the cognitive processes in operation in oral traditions. 15 Edward Nowacki’s structuralist approach reveals the power of linguistic models as an explanatory 16

A further as an oral process. the in extensive comparative contribution to field, grounded important is of Theodore chants, analysis many Karp’s Aspects of Orality and Formularity in Gregorian Chant (Evanston: Northwestern Univ. Press, 1998). Here, formulaic procedure forms the foundation for understanding the musical grammar—and oral nature—of the repertory. tool in the consideration of chant

that the Frankish repertoire of Gregorian during the reign of Pippin (in paleofrankish

By contrast, Levy believes chant

was

notation)

fixed in and

Charlemagne. invokes

neumes

subsequently (in “gestural” notation) during While there is

suggestive

the

reign

of

solid evidence that this occurred, Levy evidence to support his case—from documentary no

accounts, and from the notational appearance and melodic variants of

12 Treitler, With Voice and Pen, 231. Defended by Max Haas, Mündlicher

13

P.

Lang, 1997), 30-34. 14 Mary Carruthers,

The Book

Überlieferung

und alt-römische Choral (Bern:

of Memory (Cambridge: Cambridge

Univ. Press,

1990),

esp. 1-45. 15David Rubin, Memory in Oral Traditions (Oxford: Oxford Univ. Press, 1995). Chaps. 5-7, entitled “Combining Constraints,” “The Transmission of Oral Traditions,” and “On Remembering,” are particularly helpful (90-174). See also David Rubin, “Cognitive Processes and Oral Traditions,” in Heartz and Wade, Report of the Twelfth Congress, 173-80. 16 Edward Nowacki “Studies on the Office Antiphons of the Old Roman Manuscripts" (Ph.D. diss., Brandeis Univ., 1980). Nowacki has most recently and most explicitly explored these ideas in “Linguistic Perspectives in the Analysis of Plainchant,” presented ,

at

the Irish World Music Centre, Univ. of Limerick, March 2004.

the chants in

surviving

neumed

manuscripts

from the late

9th

century

onwards. 17

why scholarly emotions have tended to run high in this adopted chant tradition carried by memory for almost too years before being fixed in notation is a very different prospect from a chant tradition whose dissemination and stability was aided by the use of neumes showing the outlines of the melody. For scholars trying to It is clear

debate. An

establish to what extent the chant tradition recorded in Frankish

manuscripts

maintains the Roman

suppressed Gallican elements, or accurately reflects tradition it was supposed to be, the presence or absence of

notation is of the greatest moment.

2. The

Uniformity of the

“The

Gregorian

Tradition

repertory of Introits, Graduals, Tracts, Offertories a remarkable degree of fixity around 850.

and Communions attained

Whether that not

we we

believe that this

formalized in

opinion.”

fixity was represented by an actual textstate archetype, or only by a consensus of tradition any imaginable documentary form, is a matter of

could call the

18

The idea that the

9th-century chant tradition was fixed and stable is of chant scholarship. 19 For Levy, this includes small details of rhythm, ornament, and pitch: the early neumings are “exact enough about duration and ornament so that, despite the approximate profilings of pitch motions, they can be seen to represent

one

of the sacred

cows

down to fine detail, and meant to sound the This premise is linked by Levy to a second: occasion.” every that melodic accuracy of the kind demonstrated by the earliest extant manuscripts would not have been possible without the aid of notation. Ergo, there must have been a late 8th- or early 9th-century written

music that

was

quite fixed,

20

same on

21

17 Those examining the origins of notation tend to concentrate on the core repertory of Mass Proper chants. This is understandable, as the earliest surviving complete Antiphoners are rather later than the earliest neumed Graduals. Levy claims that “the historical situation should be a parallel for the Office,” and this is supported by the survival of from what appear to have “two notated fragments dating from the late ninth century been Office antiphoners, one with Breton neumes and one with German.” David Hiley, “Notation, §III, 1; History of Western Notation: Plainchant,” in New Grove Dictionary of Music and Musicians 2nd ed. vol. 18 (New York: Grove’s Dictionaries, 2000), 84-119; 88 .

.

.

(hereafter, NGII). 18Richard

1990),

Crocker, New Oxford History of Music,

2

(Oxford:

Oxford Univ. Press,

214. 19 The various forms taken by this assumption are beautifully summarized by Treitler, With Voice and Pen, 143-49. 20 Levy, Gregorian Chant and the Carolinians, 195. 21 Pfisterer challenges Levy to defend this supposition. See Andreas Pfisterer, Cantilena Romana: Untersuchungen zur Überlieferung des gregorianischen Chorals (Beiträge zur Geschichte der Kirchenmusik 11. Paderborn: Schöningh, 2002), 97.

archetype in order to achieve the observed melodic uniformity. Both premises need examination before Levy’s conclusions can be accepted. Surviving notated graduals

as

broadly representative of the

It has been said that the earliest

tradition

surviving chant manuscripts

reflect

uniform chant tradition, and this claim will be discussed cussed below. First, however, it is important to establish to what extent the existence of

a

early manuscripts are representative of the 9th-century tradition. surviving notated sources are few in number and date only from the end of the 9th century. This, McKitterick considers, casts doubt on their value as witnesses to pan-imperial musical practice in the earlier years of the 9th century. Levy refers to Carolingian royal capitularies and ecclesiastical legislation insisting on correct Christian texts and uniformity of Christian observance, but these instructions seem not to have been consistently observed. McKitterick lists aspects of liturgy that, as surviving documents attest, were not as uniform as the propagandists liked to suggest: the Bible, the Homiliary, Sacramentaries, collections of canon law, and monastic rules. It seems to her unlikely that musical practice was uniform and stable when other aspects of liturgy the

The first

22

were

not.

Is it

possible

that musical

practice

was

uniform in the

Carolingian

era? After all, Roman chant was not just musically and textually but also practically different from what had gone before. When the Franks took on Roman chant, they took on not only the melodies and their texts, but also the very concept of properization and schola chant. 23 This might mean that music was the one aspect of liturgy to be adopted consistently,

since it was not remotely familiar either in substance or in kind. “Throughout Charlemagne’s legislation and the Carolingian conciliar decrees, it is only with reference to the chant that the Roman practice is specifically mentioned and ordered to be followed.” 24

Rosamund McKitterick, Review of Gregorian Chant and the Carolingians, in Early History 19 (2000): 279-91. 23 “Properization” is McKinnon’s term for the development of having particular chants consistently attached to particular feasts. The term was introduced in James McKinnon, “Properization: The Roman Mass,” in Laszlo Dobszay, ed., Cantus Planus, Papers Read at the Sixth Meeting, Eger, Hungary September 1993 (Budapest: Hungarian Academy of Sciences, 1995), 15-22. The concept is more fully developed in idem, The Advent Project. McKinnon’s association of the beginnings of properization with the later 7th-century Roman Schola Cantorum has been challenged by some as being too late; nevertheless, having specialized singers maintaining an annual cycle of proper chants does not seem to predate Carolingian times in northern continental Europe. 24 Yitzhak Hen, “The Royal Patronage of Liturgy in Frankish Gaul,” Henry Bradshaw Society Subsidia 3 (Woodbridge: Boydell and Brewer, 2001), 85. This monograph otherwise 22

Music

deconstructs “the common text-book narrative of the Romanization and unification of the Frankish liturgy under Charlemagne and his successors” (151).

Carolingian

commentaries of the

9th century, however,

make it

clear that the melodic tradition

represented by the late 9th-century manuscripts is not identical to that found in the early 9th century, and that the chant tradition in the time of Pippin and Charlemagne was not considered to have been uniform melodically (or in the repertoire included) even by those writing a generation or two later. For example, Helisacher wrote a letter ca. 814-22 about revising the ways responsories and their verses were combined so that they would make textual sense. In the 830s Amalar wrote in his Prologus antiphonarii a se composti that he marked Roman chants with “R” in the margin, Frankish ones with “M” (for “Metz”), and his own chants with “IC” (seeking the “indulgentia" and “caritas” of possible critics); the Antiphonaries of his province did not represent a uniform tradition, either textually or (presumably) musically. 25 Ago bard of Lyons, temporarily replaced by Amalar his Italian exile (835-38), attempted to undo Amalar’s changes during to the Antiphonary during his absence, being opposed to the use of non-biblical texts in the liturgy, which he considered to be recent. 26 There are also clear indications in Aurelian’s Musica Disciplina that the tradition was not fixed in the 840s. New composition continued, overlaying melodies that Aurelian considered to be authoritative: “certain people wish to join the verses of these antiphons with their reciprocal portion by means of some new-fangled Tone but they are mistaken."27 Further comments by Aurelian, which show that the tradition he knew in the 840s was melodically different from that found in the earliest practical sources, will be discussed below. All of these examples indicate that the repertory was not fully fixed in one authoritative version until after the 840s, if then. 28 Detailed discussion of the musical unity, or lack thereof, of the chant tradition was not a principal concern of 9th-century commentators. As Anders Ekenberg makes clear, the Carolingian commentators were instead interested in the unity of the choir signaling the unity and community of Christians, and in the connection between earthly song and the eternal

praise

of the

angels. 29

Their

interpretations

Edward Nowacki, “The Gregorian Office Antiphons and Journal of Musicology 4 (1985/6): 243-75; 263-64. 25

the

of the

Comparative Method,”

26

Ibid 253. Aurelian of Réôeome, The Discipline of Music (Musica disciplina), trans. Joseph Ponte (Colorado Springs: Colorado College, 1968), 41. 28 See also Michel Huglo, “Les remaniements de l’antiphonaire Grégorien au IXème siècle: Hélisacher, Agobard, Amalaire,” Culto Cristiano e Politica Imperiale Carolingia. Convegno del centro di studi sulla spiritualitá medievale 18, Todi 1977 (Todi: Edizioni Scientifiche .,

27

Italiane, 1979), 87-120. 29 Anders Ekenberg, Cur Cantatur? Die Funktionen des liturgischen Gesangs nach den Autoren der Karolingerzeit (Stockholm: Almquist and Wiksell International, 1987) [Bibliotheca Theologiae Practicae 41], 171-82.

meaning liturgy

of chant focused

and

While the

on

on

the need for

early 9th-century

and music

sources

treatises), “offentsichtlich hat

Qualität

des

holy mystery

of

between cor, vox, and opus.30 mentioned here do signal Carolingian

for external musical correctness

concern

die innerer

the role of chant in the

continuity

Gesanges bemüht;

(also indicated by

man

tonaries

sich noch mehr

ohne diese inner

um

Qualität

verfehle der Kirchengesang seine Aufgabe." The Saint Gall “Life of Charlemagne” usually attributed to Notker does not mention the situation at the time of writing (88os) but in a familiar passage notes the distress of Charlemagne himself at the differences in the ways the different provinces, cities, and even smaller localities chanted in the late 8th or early 9th century. The teaching of the cantor who returned from Rome to Metz spread widely and quickly, and the cantor who (according to this account) remained at Saint Gall also made sure chant was taught and learned there according to the Roman fashion. The anecdote makes clear that Charlemagne’s main concern was that the singing style should be consistent, which it was not already; once again, there is no claim of melodic unity in the early 9th 31

century. Even in the early 10th century there was no single and unified melodic tradition, as the anonymous Commemoratio Brevis makes clear: “I do not prejudge those who quite permissibly sing the same intonations differently; their way is not worse and may perhaps be better.”32 Carolingian commentators, to the extent that they concern themselves with such matters, do not claim a melodically unified tradition, although it would appear that chant unified in style at least, if not in the melodic details, was disseminated from Metz by the end of Charlemange's reign. 33 There were certainly wholesale local revisions of the repertory until at least the 840s, but there was undoubtedly a level of continuity in the chant tradition through the 9th century. 34 However, the commentators’ evidence of

change through the century means that early manuscripts represent the entirety 9th-century Frankish chant tradition, or even the entirety

it would be foolish to assert that the of the

Ibid ,, 182, 189. Ibid 185. 32 Terence Bailey, Commemoratio Brevis de Tonis et Psalmis Modulandis: Introduction, Critical Edition, Translation (Ottawa: Univ. of Ottawa Press, 1979), 107. 33The stylistic unity of the 9th-century tradition is one of the central observations of James Grier who, however, connects this to his hypothesis that neumes were developed in the early 9th century to facilitate the wide transmission of the style. James Grier, “Adémar de Chabannes, Carolingian Musical Practices and Nota Romana," Journal of the American 30 31

.,

Musicological Society 56 (2003): 43-98. 34 “If the notated chant is a fundamentally new repertoire, one would expect rather more to have been made of the change than we in fact find.” David Hughes, “The Implications of Variants for Chant Transmission,” in De Musica et Cantu, 65-74; 66n5.

of the chant tradition at the time

detail. It is

they

were

written, in close melodic

necessary to turn to these early manuscripts melodic tradition they represent.35

now

how unified

a

Compatibility

does not

Even when the

to ascertain

equal consistency

neumes

used for

a

given melody

do have

compatible

melodic outlines in different manuscripts, the conclusions to be drawn are not self-evident. A brief example will serve to illustrate this point. One of the standard cadences in the eighth-mode tracts, used at the half-verse caesura, is shown in Figure 1 The half-verse melisma in .

the

tracts is

in these five

eighth-mode compatible identically time after time

early manuscripts, manuscripts, with occasional variants. 36 However, the sense of melodic flow implied by the early manuscripts is not identical, as the numbers in the righthand column of Table t indicate. The melisma ends with a trigon followed by two dives in the Saint Gall and Einsiedeln manuscripts as well and it is written

in each of these

in Chartres 47. 37 Laon 239 instead has a punctum, two short dives 38 joined together, and a long clivis consisting of two separate tractuli. as

rhythmic detail of this phrase is not always identical in the early manuscripts. For example, the clivis after the pressus minor is sometimes short and sometimes long. This is marked by an episema in the Saint Gall manuscripts and with t (hold) in Laon 239; 39 it is sometimes marked with c (quick) in Saint Gall 359. The manuscripts do not correspond systematically in their rhythmic variants; there was, in other words, no single performance practiced. The

40

It has only been possible for me to consult the five early Graduals/Cantatoria in Paléographie Musicale in preparing this study. 36 For example, Laon 239 sometimes has three puncta at the beginning of the figure rather than two puncta and a tractulus. [For the identification of chant manuscripts, see the Appendix at the end of the present study.] 37 The mark above the trigon in Einsiedeln 121 is just a smudge. Although this is not clearly visible in the facsimile in Paléographie Musicale 4 (Solesmes: Impr. Saint-Pierre, 1894), 89, reproduced here, it can clearly be seen in the color facsimile in Codex 121 Einsiedeln (Weinheim: Odo Lang, 1991), f. 45r. 38 On the varied neume groupings found in early manuscripts, see Pfisterer, Cantilena Romana, 30-33 and 74. 39 The significative letters used in the examples found in this article are defined by Hiley in “Notation,” 99 as follows: c = cito or celeriter = perform rapidly or quickly l = levare rise [in pitch] r = rectitudinem vel rasuram. crispationis = with straight or forthright vibrato-less 35

published

=

. .

[tone] vel sursum scandere = ascend trahere or tenere = drag out or hold st also appears, identified by Pfisterer

s

=

t

=

susum

(Cantilena Romana, 74)

keep going. 40

See also Pfisterer, Cantilena Romana, 74-75.

as

meaning

statim

=

FIGURE

1. Iubilate domino “omnis terra” in the

early manuscripts

Manuscript examples reproduced by kind permission of the Abbey of Solesmes, Einsiedeln Stiftsbibliothek, Saint Gall Stiftsbibliothek, and Laon Bibliothèque Municipale.

The notation of ornaments in Frankish

manuscripts is the subject (University of Limerick). study by He is finding that the appearance of ornaments in early manuscripts is governed by conventions and procedures which do not relate to a copying tradition. Instead, there was a significant element of choice within of a current doctoral

Oscar Mascareñas

This is not the where pressus consistently Figure shown in the same in all five here, but appears place manuscripts there are many other contexts in which freedom is shown in the choice of whether or not to use an ornament in these early manuscripts. Figure 2 illustrates the two ways of articulating a different standard phrase

well established conventions for both scribes and cantors. 41 the

case

in the melisma shown in

1

,

42

Described in “Procedure vs. Fixity: On the Use of the Quilisma and Oriscus in at the Irish World Music Centre, Univ. of Limerick, March

41

Gregorian Chant,” presented 2004. 42

Pfisterer, Cantilena Romana, 74-77 points

manuscripts within

the Saint Gall tradition.

to

the ornamental

independence of different

TABLE 1 The half-verse melisma in the Manuscript

eighth-mode

tracts

Neumes used

Neume

grouping Saint Gall Two

Torculus

Pressus Clivis

Clivis

Trigon

minor

apostrophes joined

359

+

Long 2+54-2+3+2+2 clivis

long second

onto

note

Saint Gall Two

Torculus

Pressus Clivis

Trigon

apostrophes joined

339

Clivis +

Long 2+5+2+3+2+2

long

clivis

second

onto

note

Einsiedeln Two

Torculus

apostrophes joined

i2i

Pressus

Long Trigon

Clivis

minor

clivis

long

+

Long 2+5+2+3+2+2 clivis

second note

Laon 239

Two puncta Tractulus. Pressus Clivis Punctum. Clivis Clivis minor clivis

joined

joined

onto

onto

joined

Chartres

Two

Punctum, Pressus Clivis Trigon

47

puncta

Clivis

leading

to

the

same

oriscus appears

on

to

Clivis

Clivis 3+2+2+2+3+

(long) 2+2

stereotyped the

Long 3+6+1+4+2 clivis

cadence in Laon 239. Sometimes sometimes there is

penultimate syllable;

an no

ornament. 43

To claim that the obscure the

high

early manuscripts

are

essentially compatible

is to

level of variation shown at this level: The details of

and melodic flow are interpreted in different different cantors, and by the same scribe within a single manuscript on different occasions. As far as these aspects are concerned, the

rhythm, ornamentation, ways

by

tradition

When the

was

not stable. 44

same

neumations

are

expressed differently

in later pitched

notations Van der Werf

tradition

was

objected to Hughes’s claim that the Gregorian melodically consistent: “my publication contains

chant many

The oriscus is used in Laudate dominum verse 2, Vinea facta est verse 3, Sicut ceruus 2 and 3, and Attende caelum verses 2 and 3. There is no oriscus in Cantemus domino verses 1, 2 or 3, in Vinea facta est verse 2, in Qui confidunt verses 2 or 3, or in Attende caelum verse 4, 43

verses

44

1,

Hucke, “Gregorianische Fragen,” 327.

FIGURE

2. Mid-verse cadences in Laon 239

chants with variants among their Gregorian versions that are far more substantial than the ones found in the tables selected by Hughes.” 45

point can also be made by considering the pitches of the cadence given in Figures 1 and 2 These vary in pitched manuscripts: ccabGGF in Aquitanian and Beneventan manuscripts; bcacGGF in the Klosterneuberg

This

.

Gradual, GRAZ 807; cca♭GGF in the Cluniac manuscript Fitzwilliam Museum MS.

Cambridge,

369;

and ccacGGF in various

northern French, Norman, and English manuscripts. While it seems sensible to suppose that the Saint Gall and Einsiedeln manuscripts share

a

melodic tradition, one cannot with certainty specify which of pitch outlines they used. One definitely cannot claim

the four known with

that the traditions represented by Laon 239 and Chartres pitches as each other or as the Saint Gall manuscripts. Even when the neumatic outlines of the early manuscripts are

certainty

47 used the

compatible, This, of

same

the

pitches

sung may not have been uniform at all. problem for modern scholars rather than for

course, is a

or 10th-century musicians. Either the pitch content of the chants sufficiently well preserved in 9th- and 10th-century memories that was unnecessary to notate it, or pitch uniformity was not important

9thwas

it

enough

to be recorded

existed, of

in

exactly (in

daseian

or

letter notation, which already

notational system, necessity being the mother The compatible melodic outlines in the early neumed

or

invention). manuscripts do

a new

not tell us which of these alternatives was in

play;

nor

do the different pitch outlines in later manuscripts, since such variation might have been an integral part of the tradition from the beginning, or

might have developed over the

43

42

course

of several

Hendrik Van der Werf, “Communication,” Journal See also Hucke, ibid ., 326.

(1989): 432.

generations.

of the American Musicological Society

The kinds

of variants

that

are

passed

“basically

over as

the

same”

“insignificant”

or

Whole classes of variants get sidelined in this way, but information about how the tradition

was

they too

understood. 46 As well

carry the

as

neumes instead of normal ones and variants in freedom is shown in other matters of detail. notation, rhythmic The important thing was for the cantors to get from one structurally

use

of ornamental

important pitch to the next in a grammatically appropriate way. Structurally important pitches can be identified by comparing different versions of the same chant in different pitched manuscripts, different genre, and cognate chants in different melodic traditions (such Gregorian and Old Roman). The goal tones that all in have common readings might be considered the structural pitches, even when there are otherwise many variants. These structural pitches are sometimes common to all chants within a melodic family, a genre, or a mode. 47 The detail of exactly how the individual cantor moved between these structurally important points appears to have been considered of less importance, especially in formulaic phrases. The same outline can be distributed pitch among several syllables in varying ways; can be repeated pitches repeated more or fewer times; some notes can be omitted or gaps filled in for melodic smoothness; recitation passages appear in different ways, with different reciting notes and different chants in the

same

as

the

treatment of accents.

This is illustrated in

beginning of verse 10 of the second-mode example chosen at random from the has a pes rather than a virga on “-ver-”; this is an pitches being distributed in different ways. Many Figure

3 the ,

tract Deus deus meus, an

repertoire. Laon 239 example of the same further examples of this sort of small melodic variant can be found in other chants. 48 As Hughes states: Although the “Gregorian versions had in principle been stabilized, the oral tradition (despite its constancy in most large matters) was in continuous motion with respect to details.” 49

46 For the following, I am indebted to the classification by David Hughes in “Evidence for the Traditional View.” 47 Identification of the structural pitches forms part of the analysis of the eighth-mode tracts in Emma Hornby, Gregorian and Old Roman Eighth-Mode Tracts (Aldershot;

Ashgate, 2002), Chap.

3. For several further examples of this kind of variant, see Hornby, Gregorian and Old Roman Eighth-Mode Tracts, 26-40. It should be noted that, while Einsiedeln 121 is scarcely legible in the Paléographic Musicale facsimile, reproduced here, it does in fact have the same neumes as the Saint Gall manuscripts, as consultation of the color facsimile in Codex 121 Einsiedeln reveals. It has extra cautionary significative letters pertaining to pitch; 1 (high note) on “Ip-” and “(considera)ver(runt),” i (low note) on “con-” #1 and “et” and s (high note) on “(con)spe(xerunt).” 49 David Hughes, “Variants in Antiphon Families; Notation and Tradition,” in Marc Honegger and Paul Prévost, eds., La musique et le Rite Sacré et Profane: Actes du XIIIe Congreès 48

FIGURE 3. Deus deus mens

Digression: Levy

“Ipsi

vero.

.

.” in the

early manuscripts

and Aurelian

claim of “continuous motion” is in direct contradiction to Kenneth Levy with regard to the relationship between Aurelian’s descriptions of melodic flow in Chapter XIX of Musica disciplina and the actual state of the melodies in the Hartker Codex and other antiphoners. 50 Levy’s finding that there was continuity, in terms

Hughes’s

that made

by

of both melodic outline and nuance, therefore merits a brief digression here. The passage from Aurelian is as follows (in Levy’s translation): In the Responsory Magi veniumt ab oriente, on the twenty-second syllable, that is ‘-tes’ of ‘dicentes’, the melody does not curve through a twisting inflection (anfractus inflexionum) because another syllable

de la Societé International de Musicologie: Strasbourg 1982 (Strasbourg, Association des Publications près les Universités de Strasbourg, 1985), 41-42. L evy, Gregorian Chant and the Carolingians, 187-90.

follows directly with a melodic winding around (circumvolvitur): it would be absurd for that same melodic motion to be repeated on successive syllables, with no pause or other syllable in between. In the Responsory Iste est qui ante Deurn magnas virtutes operatus est, however, on the eighteenth syllable, that is ‘est’, there is a circumvolution (circumvolutionem —a two-pitch ascent) as well as a circumflexion (circumflexionem: two-pitch descent); that is so because between this syllable and the one that winds around after it there is the syllable ‘et,’ which has an acute accent (acutus accentus: a two-pitch ascent) that separates a

the two melodic As

tums.

51

Levy writes, the “-tes” Codex, which is

of “dicentes” in

the Hartker

not a

twisting

Magi indeed has a clivis in (see Fig. 4 reproduced

inflection

,

with permission from Gregorian Chant and the Carolingians, 189). Also, in Iste, the syllable “et” has an acutus accentus, which Levy interprets as being a two-note ascent. This also fits with the melodic outline found in the Hartker codex. However, in Iste on “est” there is a circumvolutionem as well as a circumflexionem. Levy interprets these as a two-note ascent and a two-note

descent, respectively,

and

sure

enough,

this is what

we

find in the Hartker

codex. But Atkinson makes it clear that “Aurelian uses the verbs ‘circumvolvo' and ‘circumflecto’ and the nouns ‘circumvolutio’ and ‘circumflexio' interchangeably, with all four terms referring to a three-note

ascending/descending melodic This

means

gesture, insofar as

that Aurelian knew this melodic

one can

shape

as a

determine.” 52

double

twisting

gesture, perhaps something like GaGFG or GaGaba. Levy interpreted Aurelian’s terminology to fit the melodic state of the chant rather than accurately translating the words and then comparing the result to the melody in the Hartker codex. The details of this piece in the Hartker codex are close to Aurelians's description, but they are not “exactly the same,” despite Levy’s claim to the contrary. 53 In his dissertation, Joseph Ponte discusses Aure51 Ibid 187-88. The original passage is as follows: “In hoc responsorio: Resp. Magi veniunt ab oriente, vicesima secunda eiusdem responsorii syllaba, videlicet ‘-tes’ ut ‘dicentes,' ideo non incurvatur per anfractus inflexionum quia protinus altera subsequitur syllaba quae circumvolvitur. Ideoque absurdum esset si iteraretur duplatio modulationis in duabus syllabis, nulla interiacente morula vel qualibet syllaba. In hoc autem responsorio: Resp. Iste est qui ante Deum magnas virtutes operatus est, idcirco octavadecima syllaba, scilicet ‘est,’ circumvolutionem ac circumflexionem recipit, quia inest inter hanc et illam ea aquae post circumvolvatur, id est ‘et,’ atque in ipsa fit acutus accentus quae has duas distinguit modulationis” (Aureliani Reomensis Musica Discplina, ed. Lawrence Gushee, CSM 2, .,

128). Charles Atkinson, “De Accentibus Toni Oritur Nota Quae Dicitur Neuma: Prosodic Accents, the Accent Theory, and the Paleofrankish Script,” in Graeme M. Boone,

52

on Medieval Music in Honor of David G. Hughes: Isham Harvard Univ. Press, 1995), 40. Levy, Gregorian chant and the Carolinians, 188.

ed., Essays 53

Library Papers

4

(Cambridge:

FIGURE 4.

Responsories Iste est (Common of Confessors) (Epiphany)

Kenneth repr.

Levy, Gregorian by permission.

Chant and the

Carolingians,

and

Magi veniunt

Princeton Univ. Press,

1998;

examples in some detail. There are numerous occasions on which descriptions of melodic outline, or his association of certain chants with particular differentiae, are not reflected in the practical sources

lian’s

Aurelian’s at all. 54

Compatible and incompatible neumings within

a

single manuscript

Figure 5 shows phrase 3c of the second-mode tracts in the early notated manuscripts. 55 Phrase 3, cadencing FEFGGF, is used after the half-verse cadence on C, and phrase 3c, with the decorated cadence FEGabt♭GGF, is used for particular emphasis in the final verses of the chants. 56Phrase 3c is used for emphasis in the first verse of Domine exaudi for the important “et clamor meus” (“Lord hear my prayer and let my cry come to you”); the further musical parallel between “et clamor” and “exaudi” in the first half of the

between the cry of the this is the most

important

verse

reinforces the

connection

that God will hear it; sentiment of the chant and it is emphasized

psalmist and

the

hope

Joseph Ponte, “Aureliani Reomensis, Musica Discipline A Revised Text, Translation, and Commentary,” 3 vols. (Ph.D. diss., Brandeis Univ., 1961). 55 The nomenclature is that used in my forthcoming study of the second-mode tracts. Phrase 3 appears third in the verse, after the half-verse caesura. Phrases 3a and 3b are the versions used with a penultimate and an ante-penultimate accent, respectively; phrase 3c has a decorated cadence. 56These are the in most found diastematic pitches commonly manuscripts. I provide them here to give readers an idea of the shape of the cadence, rather than to claim any particular authority for this exact version, which is not the only one found. Other pitch oudines given in the discussion should similarly be understood as illustrative rather than definitive. 54

FIGURE 5.

Phrase 3c in the

early manuscripts

melody attached to it. 57 Phrase 3c is also used in verses signal a departure in the tract text from the familiar Vulgate text. Saint Gall 359 has a normal way of writing this phrase. In Domine audiui the phrase is used outside the normal formal context (last verses), and the scribe has added the letter s (high note), reminding the cantor to go higher than normal for phrase 3 (i.e. to use phrase 3c). Domine exaudi verse 1 is also an unexpected context for this phrase, 59 and the cantor is reminded to keep moving with the letter c (quickly). This phrase was normally remembered rather than copied from place

by means

of the

1, 2, and 4 of Domine audiui to 58

57 I am currently preparing for publication a study of the use of melodic form to emphasize and link important words and concepts in the second-mode tracts. 58The significant differences between the Vulgate text of Habakkuk 3 and the textual tradition found in the second-mode tract appears to have been a major factor in the choice of unexpected phrase shapes; they appear to act as a signal to the Frankish cantors that the usual textual flow is to be avoided at certain points. See Emma Hornby, “Rhetoric and Exegesis in the Second-Mode Tracts, and the Way in which these Chants Function as ‘Readings’ of their Texts” in Cantus Planus: Papers Read at the 12th Meeting Lillafüured, 2004 (forthcoming). 59 There is an extra note at the beginning of the melisma in all manuscripts.

place; there was one way of performing it, which was reinforced by cautionary significative letters in the unusual contexts. Laon 239 indicates a normal way of performing and writing the phrase. In Domine audiui, the scribe left out one or more of the significative letters in verses 2 and 5, having had the phrase in the previous verse each time (for example, n—join together), but it seems likely to me that this is scribal rather than indicating a different way of performing the 60 phrase. In Domine exaudi verse 1 the phrase is performed in a different to

the

way, with the second note sung shorter (there is no t, which indicates a long note, on this occasion) and the third note long (marked

t). This special case, with its rhetorical emphasis, performance associated with it by the cantor.

with

Chartres 47 has

had

a

particular

normal way of writing and performing the phrase. has a long first note (the extra one) marked with t and then the letter n to join together the phrase in one continuous flow. As in Laon 239, the performance of phrase 3c in Domine exaudi verse 1 is differentiated from the norm. There are two possible interpretations of this evidence. Either the relevant scribe had in mind one way of performing this standard Domine exaudi

a

verse 1

melisma and wrote it

identically

rather than

of direct

the

one

melisma in the

each time it occurred

copying),

or

the

(an

cantor

act of memory

mechanically

way every time it occurred. The case of Domine exaudi verse 1 stimulated a different neuming in all three manuscripts, copied once more in Domine exaudi verse 6 in Chartres 47, suggesting that use of the same neumes on other occasions was not a mechanical copying act. The occasional use of cautionary significative letters in Laon 239 and Saint Gall 359 in Domine audiui also suggests that the phrase was being considered within its immediate context rather than as a standard shape to be copied without critical input. Saint Gall 339 strikingly corroborates this interpretation of the evidence. Each of the four verses of Domine audiui has a slightly different neumation; another is found in De necessitatibus verse 2, and yet another in Domine exaudi verse 6 (see Table 2 ) 6l These seem to be six different ways of performing the same melisma. It is impossible to tell whether the scribe was writing down a

copied special

same

same

.

set of

he

they occurred example should be

possibilities

thought each

as

to him or

sung. It

writing

exactly how likely to me that

down

seems more

60 Examples of scribes omitting the ends of familiar melismas that they had recently copied in full, as well as the more wholesale omission of both letters and neumes in Montpellier H159 (the Dijon Tonary), may be found in Hornby, Gregorian and Old Roman Eighth-Mode Tracts, 153-68. 61 It is not possible to read the neumes of Domine exaudi verse 1 in the Paléographie

Musicale facsimile with any confidence; I have therefore not included it in my comments here.

TABLE 2

Phrase 3c cadence in Saint Gall 339 Chant and

Neumes used

verse

Long clivis Long clivis

Domine audiui vi

Porrectus

Torculus

Domine audiui v2

Porrectus

Torculus+episema on

Domine audiui V4 Domine audiui V5

Clivis, virga

Torculus (joined to)

Short clivis

Porrectus+

Torculus

Long clivis

(and other final verses)

episema on

De necessitatibus\2

Clivis, virga+episema

Torculus+episema on

Long

Porrectus+

last note Torculus

Short clivis

last note

Domine exaudi v6

last

note

clivis

episema on last note

the

manuscript provides

a

(non-exhaustive) range

the melisma from which the moment of These

performance, phrases are not sufficiently legible

facsimile of Einsiedeln

of ways of

performing

cantor, away from the manuscript and in would select as the spirit moved him.

a

121

in the

for them to be worth

Paléographie Musicale reproducing here.

Consultation of the color facsimile, however, reveals that there many nuances in the notation of this melisma (see Table 3 ).

are

The cadence appears in one version in the first verse of Domine verses. The significative letters indicate the high

audiui and in final

pitches

characteristic of this cadence. The omission of the second significative letter in Deus deus meus and Qui habitat, and of both in Domine

audiui,

probably

does not indicate

but rather that the scribe,

on

a

different

performance practice

these occasions, considered the reminder

unnecessary. As in Saint Gall 339, slightly different neumations are used in verse 2, Domine audiui verse 4, and Be necessitatibus verse

Domine audiui 2,

suggesting flexibility

verse 1

(if 1 have

in

read the

performance practice. The r in Domine exaudi script correctly) indicates a particular performance

practice for this moment—without vibrato. Such differences in nuance do not indicate fauldines in the written transmission of the pieces but demonstrate that even within cadences as standard as this one there was room for interpretative maneuver. This can be observed to a great extent in Saint Gall 339 and Einsiedeln 121 and to a lesser extent in the other three manuscripts considered here. The scribes of these manuscripts captured melodic nuances that for them distinguished individual instances of a formula or were part of a flexible repertoire of possible ways of singing that formula.

TABLE 3

Phrase 3c cadence in Einsiedeln Chant and

121

Neumes used

verse

Domine audiui vi

Porrectus+s

Torculus+s

Domine audiui final verse De necessitatibus

Porrectus

Torculus

Clivis+episema Clivis+episema

Porrectus+s

Torculus+s

Clivis+episema

Porrectus+1

Torculus

Clivis+episema

Porrectus+1

Torculus

Clivis+episema

Domine exaudi final verse

Porrectus+1

Torculus+s

Clivis+episema

Domine audiui V2 Domine audiui V4 De necessitatibus V2

Clivis, virga

Torculus Torculus Torculus

Clivis+episema Clivis+episema Clivis+episema

Domine exaudi vi

Porrectus+1

Torculus+s

Clivis+episema

final

verse

Qui habitat final

verse

Deus deus meus final

verse

Porrectus+c Porrectus+st under 2nd note over

3rd note and r/s under 2nd notea a

If this letter is s (= sursum), it makes litde melodic sense, since it appears the lowest note of the porrectus.

next

to

Brief mention should be made of the

found

far for

implications

of the variants

early notations. One of the is that different manuscripts

semiological approaches assumptions of semiologists carry the same melodies, to be performed in the same way. This is axiomatic for comparing and interpreting neume shapes in different manuscripts and in different notational traditions. The level of variation found in the treatment of stereotyped phrases in the early manuscripts that the semiologists concentrate on, and the inconsistent treatment of standard pitch structures in these manuscripts, 62 suggest that such a methodology should be approached with extreme caution. 63 so

fundamental

62

around the interval of a minor third, which can be left as an open filled in with an ornament, with no apparent pattern. Marie-Noëel Colette, “La sémiologie comme voie d’acces a la connaissance de 1’interprétation au moyen age,” in Michel Huglo, ed., Musicologie Medievale: Notations et Sequences (Paris: H. Champion, 1987), 121-28, gives a nuanced view of the value of semiological studies, cautioning that it is necessary to consider each manuscript on its own

Especially

third, filled in, 63

to

or

Pieces that do not appear in consistent versions in the Frankish tradition Some chants

Alleluias

were

very much less firmly fixed in the Frankish tradition.

example. Their repertoire was not fully developed when the Franks adopted Roman chant; McKinnon hypothesizes that only a handful (about 15) of alleluia texts were stably associated The rest, he thinks, were with particular melodies at this time. in both melodies and ad hoc The of variation in an list. conveyed degree are

the obvious

64

texts used for alleluias may reflect their lack of authoritative Roman aura.

The

chant

eighth-mode

seems

to have

tract Beatus uir is a

been

adopted

similar

at St Denis

case.

during

A version of the the visit of

Pope

Francia in the 750s. This undermined the authority of the rather different standard version of the chant, which did not become established in all places in northeastern Francia; there are at least two local versions, connected to St Denis and Noyon, as well as a different chant, Ecce uir, used at Corbie. 65 The lack of authority of a particular chant can be a reason for its unstable transmission. Some chants are less consistent because they fit less well with modal theory, and the attempt to adapt them post-dated the authoritative establishment of the repertory. People in different places came up with different solutions. Similarly, some chants were altered because they could not be notated in diastematic notation. Different scribes transposed the melody at varying points to make the intervals within a 66 phrase or section correct. Again there are good reasons—modal ambiguity or notational impossibility—for chants to have been altered.

Stephen

to

On the

large scale

most chants are, as

Karp puts it, “relatively fixed."67

It is very rare to find large-scale variants in the formulaic chants I work on, but those I have found are instructive.

terms; some are less susceptible to the semiological approach than others. Nancy Phillips cautions that semiology is based on so few manuscripts that one cannot assume that their evidence has wider relevance. “Notation und Notationslehren von Boethius bis zum 12. Jahrhundert,” in Frieder Zaminer and Thomas Ertelt, eds., Die Lehre vom einstimmigen liturgischen Gesang (= Geschichte der Musiktheorie 4) (Darmstadt: Wissenschaftliche Buchgesellschaft, 2000), 293-624. See also Cornelia Kohlhaas, “Dialog oder Rückzug ins Ghetto? Gregorianische Semiologie und Musikwissenschaft—einige Anmerkungen,”

Beiträage zur Gregorianik 30 (2000): 50-51. 64 James McKinnon, “Alleluia,” NGII, 1: 387. 65 See Emma Hornby, “The Transmission History of the Proper Chant for St. Gregory: the Eighth-mode Tract Beatus uir,” Plainsong and Medieval Music 12 (2003): 97-127. 66 On transposition, see Pfisterer, Cantilena Romana, 14-25. Several examples can be found in Karp, Aspects of Orality and Formularity, esp. 225-69, although Pfisterer takes issue with the details of some of Karp’s reconstructions. See also Rebecca Maloy, “The Roles of Notation in Frutolf of Michelsberg’s Tonary,” Journal of Musicology 19 (2002}: 67

See the

examples and

discussion in

Karp,

ibid .,

36-52.

In Domine exaudi that

begins

verse

and ends like

end of the final

verse

the usual melisma

(see Example

1

).

on

of the

5, most 4f

manuscripts

have

a

hybrid phrase

(a phrase shape associated with the second-mode tract), but instead of having

phrase a

penultimate syllable

it has

a

shorter melisma

This combination of the shorter melisma with the 4f

cadence is also found in Domine audiui verse 4. 68 Domine audiui V4 possibly has the hybrid to signal a departure of the canticle text from the version sung regularly in the Frankish offices. 69 It is not at all clear why the hybrid is used in the penultimate verse of Domine exaudi on “manducare

panem meum”—perhaps the use of the phrase with the expected pattern in the previous verse cued it, or perhaps the penultimate verse of the very similar tract Domine audiui acted as an associative cue. If the stimulus is opaque to me, it was also opaque to the scribe of Laon 23g, who followed the short EFGFG melisma with the short cadence that more usually follows it (this phrase shape is labelled 4a in my forthcoming accent

analysis).

Qui habitat verse 9 Laon 239 has the normal phrase 3 on “ad lapidem” where the other manuscripts have a special C cadence following the DCD(E)FG rise common to both contexts (see Fig. 6 ). 70 The In

distribution in Laon 239 is also different from that of the other early manuscripts. The scribe used the normal shape, stimulated by the memory prompt of the many other phrases of the type, rather than this C cadence otherwise found only in Qui habitat V4. Both of these examples are cases where the normal way of singing the chant is not the most obvious one according to the formulaic structure of the genre: A phrase is formed of a hybrid combination of elements, or a phrase is very unusual. In each case one of the early manuscripts does something within the musical grammar of the genre but inconsistent with the rest of the tradition. Many further examples of this kind of variant, dubbed Parallelstellenangleichungen, are provided by Pfisterer. 71 He considers that such variants usually represent a move towards text

more uniform, stereotyped melodies, and that this process is incompatible with a primarily written transmission. If one were to undertake a wider survey of the large-scale variants in Frankish chant, I suspect that one would usually find good reasons

The shorter melisma is also found twice with a short cadence which is associated /— accent pattern; Domine exaudi verse 4 has the hybrid phrase with the accent pattern /—. Domine audiui V4 and Domine exaudi V5 both end with a penultimate accent, however, so that is not the motivation in these instances. 68

with

a

69 70

three

See n58. Einsiedeln 121 is almost illegible in the Paléographie Musicale facsimile, but consultation of the color facsimile reveals that it has the same melodic outline as the other

manuscripts. 71 Cantilena Romana, 33-72. Pfisterer,

EXAMPLE 1.

Comparison

of Domine exaudi v5, “manducare panem meum”

with similar

verse-ending phrases

for them rather than just laissez

faire or memory failure. Variants would chants without Roman authority, to chants that are difficult reconcile with modal theory, and to chants with unexpected

be connected to

to

melodic outlines where there

are

clear

cues

to follow less counterintuitive

alternatives. 72

The axiom that the The

manuscripts

9th-century

chant tradition

show that while the

was

general

fixed needs revision.

outlines of most

Mass Proper chants were probably rote memorized and usually firmly fixed, the surface details were much more open to variation. Sometimes alternative—but equally musically valid—readings are found in different

places, usually because the normal version lacked authority, modal propriety, or was otherwise counterintuitive. Even within apparently uniform chants, consistent neumes can have multiple pitch realizations in later

72

manuscripts.

In later

manuscripts, copying errors are probably the other major factor.

FIGURE

6. Qui habitat “ad

lapidem”

in the

early manuscripts

"Evidence for the traditional view” that the kinds were compatible with mishearing or misre-membering rather than miscopying. 73 I would nuance that conclusion: the kinds of variants found here are compatible with certain degrees of

Hughes

wrote in

of variants he encountered

scribal and

performative freedom being in operation, in the details of exactly one can move from syllable to syllable while continuing to perform what is recognizable as the same piece, in the details of ornamental and rhythmic performance, and in the possible formulaic responses to particular textual and formal situations. 74 Rather than as-

how

73 In general, copying errors can be identified by being visual misinterpretations of exemplar, or mistakes such as repeating or omitting lines of text, individual words, or neumes. A variant that arose orally (however many generations of copying may have intervened between its emergence and the notated source in question) would be connected to different ways of singing or hearing the chant rather than a mistake in reading it. For example, a quilisma is often replaced by a normal note or vice versa; an oscillating figure such as GFGaG might be replaced by GaGFG or perhaps GEFG. Examples of copying errors and non-visual variants are given by Hughes in “Evidence for the Traditional View”; summaries may be found on 381-82 and 400-401. See also Pfisterer, Cantilena Romana, mana, 73-75, for examples of scribal errors; these are very rare. 74 As is evidenced by the examples given by Pfisterer, Cantilena Rormana, 33-72.

the

suming a fixed melodic tradition from which scribes deviated in largely “insignificant” ways, we should instead consider this as a tradition that was broadly—but not entirely—fixed on the large scale, but within which there

was

a

considerable amount of freedom. The versions of the

manuscripts do not—on the detailed scale performance but are precipitations blueprints into writing of particular ways of applying the stylistic conventions of the Frankish chant idiom. As we uncover those conventions, we gain a richer understanding of the performing tradition of Frankish chant in the 9th century.

chants found in the earliest

—constitute

3. The

for future

Impossibility of Memorization

Levy

invokes Guido’s

claiming

early 11th-century

assertion that

even

with

it took ten years to learn the repertoire of chant, that this suggests that the transmission of the tradition without

the aid of

neumes

notation would have been well-nigh impossible. 75 One may make a passing qualification of the significance of Guido’s statements: a) the assistance of neumes may well have removed the urgency of memorization, in consequence of which the mnemonic techniques atrophied,76 and b) since he was advocating the new technology of staff notation, Guido may well have exaggerated the shortcomings of the old system in his promotion of the efficiency of the new. Was such a feat of memory

possible for medieval singers? The first step must be a consideration of the mental processes by which the chants could have been remembered. In addressing this question, Levy proposes different models of transmission for different types of melodies. Levy considers that the unique, individually crafted (“idiomelic”) chants would have been remembered verbatim. Other chants are elaborations of psalmodic frameworks and these, Levy

thinks, may have been settled into

fixity

improvised

until

they were neumed,

before that time. Some chants

these chants the “referential

memory” required

or

may have

centonate, and for for their transmission

are

would have “meant very random access to great quantities of intricately turned melos.” Levy thinks it unlikely that centonate chants could easily

have been composed orally or memorized. 77 Chants with different kinds of structure would obviously have required different memorization techniques. One would not expect simple melodies sung regularly by the whole monastic community to require the Levy, Gregorian Chant and the Carolingians, 193-94. Levy himself observes that “the written technology that became the partner memory, by its availability and exercise rendered memory more fallible” ( ibid 137). 75

76

.,

77

Ibid ., 208.

of

as highly complex soloist’s or schola chants 78 However, making a generalized distinction between verbatim and referential memory is too simplistic. As David Rubin makes clear, the memory processes involved in oral traditions same

effort of memorization

sung

only occasionally.

rely

combinations of

on

“by

heart,”

meaning and

cues

and constraints. When

people usually (and

learning something unconsciously) use its

often

structure to aid recall.79

approaches to chant analysis have much to contribute understanding of how 9th-century Frankish chant might conceivably have been carried stably on a broad scale without notation (we have already established that it was highly flexible in matters of detail). Karp found a much smaller degree of uniqueness in many “idiomelic" chants than did Levy, instead identifying formulaic structures and outlines in the articulation of many phrases of different genres and modes. 80 This immediately raises the prospect of a musical tradition with more overarching principles of melodic construction than had previously been supposed to be the case. Rebecca Maloy’s work on offertories introduces the idea that particular melodic shapes can have more than one distinctive role, distinguished by the position of the 81 segment in the phrase and what precedes it. This entails a further nuancing of our understanding of formulaic processes beyond clearly formulaic genres. My work on the eighth-mode tracts explicitly connects Various modern to our

”

formulaic construction of

Levy’s

rized).

to mnemonic

processes (in direct contradiction

claim that centonate chants could not

easily have psalm

Once the soloist has recalled the familiar

been

memorized).

text and re-

how to divide it into

phrases, the formal context, accent patterns, facilitate the proper choice of melodic phrase type and its proper realization, although the general principles are sometimes overturned by textual cues, assonance, and rhyme. 82 and number of

The

syllables

publication dates

of all of this research

preclude Levy’s having

taken them into account in Gregorian Chant and the Carolinians. There is no clear reason, however, why he should not have taken previous analytical work into account; three scholars associated with such approaches are

Helmut Hucke, Leo Treitler, and Max Haas. In his semi-

78 Helmut Hucke, “Der Übergang von mündlicher zu schriftlieher Musiküberlieferung im Mittelalter,” in Report of the Twelfth Congress, 180-91; 184. 79 See Rubin, Memory in Oral Traditions. The three main classes of constraint and cue (meaning, imagery, and sound) are dealt with in Chaps. 2-4, respectively (15-89). The results of combining them are discussed in Chap. 5 (90-121). 80 Karp, Aspects of Orality and Formularity, 59-97. 81 Rebecca Maloy, “The Offertory Chant: Aspects of Chronology and Transmission” (Ph.D. diss., Cincinnati Univ., 2001). 82 See the discussion in Hornby, Gregorian and Old Roman Eighth-Mode Tracts, 143-50.

nal work

on

the

antiphons,

Edward Nowacki found that rather than displaying

the rich variation and contrast

suggested by a first glance at the mostly obligatory and conditioned by mode, text numbers of syllables, and accent pattern. The melody type, length, a clear indication of a genre “thrift,” antiphons display near-perfect and of stable oral transmission. 83 Some of composed orally capable these findings are implicit in earlier analyses of plainchant (that of the second-mode tracts by Hans Schmidt comes to mind, with its consideration repertory, the variation is

of the role of textual accent in musical structure; the role of text in the musical oudine of chant

also

was

explored by

Helmut Hucke). 84

analytical evidence in support of a core repertory of chant that could have been carried in a broadly stable fashion without the aid of

The

notation is

mounting hypothesis.

Whether

or

and merits discussion

not the Mass

Proper

by

chants

advocates of the

were

earlynotation

notated in the

9th

century, one must be able to account for the existence of the six unnotated Graduals published by Hesbert. 85 These date from the late 8th- to the late 9th century. If the chants were indeed transmitted with the aid of notation at this time, what was the purpose of these unnotated books? They show conclusively that the putative notation of the tradition at this time did not obviate the need for an accurate memory of it. 86 As Hucke wrote: “Even if one wishes to suppose that there were in certain localities chant books with neumes as early as the 9th century, chant books without neumes were written at least until the 10th century. We must be able to explain the beginning of chant transmission in the Frankish Empire without assuming the use of neumes.” 87 The written chant texts themselves, as Treitler suggests, may have acted as sufficient .

aides mémoires.

.

.

only do the Antiphonale Missarum Sextuplex continuing existence of a remembered tradition, but as Levy himself freely acknowledges, 9th- and 10th-century neumed notations cannot be sung from unless one already knows the melody. sources

88 Not

suggest the

83 “Studies Nowacki,

on

the Office

Antiphons.” See

also idem, “The

Gregorian

Office

Antiphons.” 84 Hans Schmidt, “Untersuchungen zu den Tractus des zweiten Tones aus dem Codex St Gallen 359” (Ph.D. diss., Univ. of Bonn, 1954); Hucke, “Gregorianische Fragen.” 85 René-jean Hesbert, Antiphonale Missarum Sextuplex (Brussels: Vromont, 1935). 88 This evidence functions, of course, as a double-edged sword. The continuation of unneumed sources into the 10th-century shows just as clearly that the presence of unneumed manuscripts does not preclude the contemporary existence of neumed ones. 87 Cited by Leo Treitler, “Communication,” Journal of the American Musicological Society 41 (1988): 567. Hucke also referred to the Metz Tonary, dating from the later 9th century, as showing that notation was not considered essential in Metz at that time. See “Gregoriansiche 88

Fragen,” 308-9. Treitler, “The “Unwritten” and “Written” Transmission,” 160.

The role of memory existed. 89

was

central, whether

or

not a

notated

exemplar

The

disputed question is how powerful cantors’ memories could emphasis placed on the correct learning of chant within monastic education, 90 one cannot simply assume that such a feat was impossible, especially when this cultural consideration is added to have been. Given the

the mnemonic characteristics of the chant repertory that have been by Nowacki and Karp, among others. There are alternative

revealed

explanations of the evidence to Levy’s fixity displayed by the earliest

melodic within son

a

assertion that the level of

Graduals was only possible tradition neumed since the time either of Pippin or of his

Charlemagne.

4. Notational Characteristics

Paleofrankish

notation and

of the Sources

gestural

notation

Levy identifies two distinct styles of early neuming: the paleofrankish type, where each point visited by the pen denotes a pitch (/ represents two pitches, low high); and the gestural type, where the pen strokes denote movement (/ represents one high or higher note). 91 The conceptual differentiation of these two styles of notation is a valuable contribution to our understanding of the pedigree of early notation. Levy believes that Paleofrankish notation was used in the first of fixing the chant tradition in the reign of Pippin, with its graphic emphasis being particularly useful for the task of editing and compiling.

92

A

exemplar, Levy maintains, originated in the Trier (during Charlemagne’s reign) in nuance-poor gestural notation, and it was this archetype that was copied and recopied, gradually developing into the different calligraphic styles that we observe in the earliest surviving manuscripts. new

region

authoritative

around 800

89 Levy’s table makes it clear how important the role of memory continued to be (Gregorian Chant and the Carolingians, 118). 90 See Susan Boynton, “The Liturgical Role of Children in Monastic Customaries from the Central Middle Ages,” Studia Liturgica 28 (1998): 194-209; “Training for the Liturgy as a Form of Monastic Education,” in G. Ferzoco and C. Muessig, eds., Medieval Monastic Education (Leicester: Leicester Univ. Press, 2000), 7-20. 91 Levy, Gregorian Chant and the Carolingians, 115ff. 92 But see Leo Treitler, “The Early History of Music Writing in the West,” Journal of the American Musicological Society 35 (1982): 237-79, for an alternative classification of neumatic notations as predominantly iconic or symbolic. According to this classification, Paleofrankish notation, together with the Messine, Breton, Aquitainian, Beneventan, and Nonantolan types, is distinguished from the other notational families by using the virga to signify an ascent (a pictorial representation of the melody) rather than using the virga to

represent a high note (a conventional association).

The identification of Paleofrankish notation with

Pippin’s

reforms

of the mid 8th century must be examined more closely. It should be noted immediately that the surviving Paleofrankish manuscripts are no older than the

manuscripts containing gestural

identification of “two neumations,

one

notation. 93

earlier,

one

confident

Levy’s

later” is

speculation

rather than fact. 94

The idea that is not borne out

a

by

Paleofrankish

exemplar

existed in

the melodic variants found in the

versions of Mass

Proper chants, which do

not

Pippin’s surviving appear

to

time

Paleofrankish vary

10th-century versions of the chants. 95 Had an there been authoritative mid 8th-century Paleofrankish exemplar, the surviving examples of Paleofrankish notation presumably would have descended from it. If the copying tradition were faithful, then the close melodic relationship between the surviving 9th- and 10th-century Paleofrankish sources and the gestural ones can only be explained by supposing that the chant was so well established in Francia in the time of Pippin that it required little or no revision in Charlemagne’s reign.

much from other

9th-

and

Paleofrankish notation, with the melodic tradition it carried, retained currency throughout. This hypothesis is made untenable by the reports of the Carolingian commentators cited above, who make no claim of a

melodically unified tradition in the 9th century. On the contrary, they clearly indicate revisions to the repertory during the 9th century. 96 Alternatively, one might suppose that after the authoritative Paleofrankish exemplar had been replaced notationally by 9th-century gestural styles of notation, and melodically by the 9th-century revisions to the chant repertory, some north-east Frankish cantors transliterated a 9th-century

93 The Paleofrankish manuscripts are dated “between the mid-9th and the eleventh twelfth centuries” (Levy, Gregorian Chant and the Carolingians, 117 and 119). Hiley gives a table of many of the examples of neumes dated to the 9th century in “Notation,”89. He writes elsewhere that there are 9th-century examples of Paleofrankish, German, French, and perhaps Spanish neumes, and that the earliest examples of Laon, Breton, and Aquitanian neumes date from the mid-10th century. See Western Plainchant, 363. 94 Levy, ibid ., 120. 95 No one has yet published a detailed comparison. The comparative examples given in the following articles suggest that Paleofrankish chants are not significantly different from their equivalents in gestural notation: Ewald Jammers, “Die palaeofrankische Neumenschrift,” Scriptorium 7 (1953): 235-59; Charles Atkinson, “De Accentibus Toni Oritur Nota Quae Dicitur Neuma,” 38-39; Bruno Stablein, “Schriftbild der einstimmigen Musik,” Musikgeschichte in Bildern (Leipzig: Deutcher Verlag fur Musik, 1975), 106-9; Wulf Arlt in Christoph Stiegemann and Matthias Wemhoff, eds., 799; Kunst und Kultur in der Karolingerzeit (Mainz: P. von Zabern, 1999), 838-42; Wulf Arlt, “Anschaulichkeit und Analytischer Charakter: Kriterien der Beschreibung und Analyse früher Neumenschriften," in Musicologie Médievale: Notation et Sequences, 29-55. 96 See the discussion, above. or

gestural exemplar into an old and obsolete chant notation. It is appreciate the motivation behind such a transliteration. 97

difficult

to

More serious doubt is

by

cast on

Levy’s Paleofrankish-archetype hypothesis

Atkinson’s close association of Paleofrankish notation with

the intellectual monastic culture of the

early 9th century. 98 Atkinson reinvigorates theory” origins with reference to Paleofrankish script, and his connection of this script with the renewed interest in the study of grammar in early 9th-century Carolingian monasteries is compelling. 99 It is noteworthy that this part of Levy’s theory was dropped entirely by Grier in his recent article arguing in favor the “Accent

of the

of

neume

early-neumed-archetvpe hypothesis.

100

Paleofrankish notation was used in certain northeast-Frankish (including that of St. Amand) for a variety of casual purposes. Chants thus notated appear in the margins of manuscripts not intended to have notation or are squeezed between the lines of text, and the notation was otherwise used for neumes attached to classical texts, Old Testament lections, music theory, tropes, and sequences. These might be considered to be “occasional or private jottings.” 101 If Atkinson is right, the conceptual foundations of Paleofrankish script emerged in the early 9th rather than the mid 8th century. Rather than descending from a notated exemplar in Paleofrankish notation in the time of Pippin, Paleofrankish notation was used locally and occasionally in northeast Frankish monasteries during the 9th century. There is still work to be done on why a small number of monastic scribes in a geographically limited area of Francia chose, on certain occasions, to use a different kind of notation rather than the normal gestural one. monasteries

102

97 We do occasionally encounter the kinds of copying errors to which such a transliteration would inevitably lend itself, but not in sufficient quantity, I would suggest, to support the scenario. For example, / is sometimes used in paleofrankish notation where a one-note virga was used in the gestural model; this sign would usually indicate a two-note pes in Paleofrankish notation. See Levy, Gregorian Chant and the Carolingians,

132. 98

Atkinson, “De Accentibus Toni Oritur Nota Quae Dicitur Neuma." See also idem, on Music and Grammar and the Advent of Music Writing in the West,” in Sean Gallagher, James Haar, John Nádas, and Timothy Striplin, eds., Western Plainchant in the First Millennium: Studies in the Medieval Liturgy and its Music (Aldershot: Ashgate, 2003),

“Glosses

199-215. 99

the

Atkinson, “De Accentibus Toni Oritur Nota Quae Dicitur Neuma,” 19-20. On of Anglo-Saxon grammarians to the Carolingian Renaissance, and of Carolingian

importance

monastic culture to the development of music theory, see Joseph Dyer, “The Monastic Origins of Western Music Theory,” in László Dobszay, Pétér Halász, János Mezei, ” and Gábor Prószéky, eds., International Musicological Society Study Group “Cantus Planus. read of at the Third Papers Meeting, Tihany, Hungary, 1988 (Budapest: Hungarian Academy

Sciences, 1990), 199-225. 100 Grier, “Adémar de Chabannes.”

Levy, Gregorian Chant and the Carolingians, 117-20; 131. At least some scribes knew both types. “Deutschen Neumen waren in Essen— nach Ausweis weiterer Quelien—bekannt und wurden hier in Einzelfällen durch paläe101

102

precludes an exemplar with gestural notation having Charlemagne. If such an exemplar was compiled during Charlemagne’s reign, then the years around 800 seem the most likely; in the 790s, he began to leave the actual conduct of military campaigns to his sons and others and spent much time on his estate at Aachen, surrounded by clerics and scholars. Arguments ex silentio are notoriously suspect, but silence there certainly is. Not a single trace of a notated Gradual from the early 9th century has survived, either as a fragment or even as a mention in a medieval library catalogue. Survival rates from the 9th century are so slight as to deny the possibility of However, one might still observe, drawing conclusions from this. 103 with McKitterick, that since “9th-century accounts of chant refer to would they not also have referred to singing masters and teachers the wonderful new writing system if it were being used to the extent that Levy suggests?” 104 One might also expect the archetype, or close copies of it, to have been preserved (in Aachen?) with the same sort of reverence enjoyed by the Gregorian Sacramentary (the earliest surviving Ms from ca. 812) and other documents of Cambrai dates 164, copy, the Carolingian reformers. None of this

existed in the time of

.

“Close”

and

“essentially

.

.

identical”

neumings

in

different manuscripts

to the similarities

Levy points among many manuscripts in their of into groupings pitches compound neumes, seeing this as a clear indication that the manuscripts share a common written ancestor. This needs consideration on two levels: how equivalent the groupings of pitches to which Levy refers are; and whether they are indeed indicauve of a common notational origin. Levy asserts that the neumations of the processional antiphon Deprecamur are “essentially identical.” 105 Only four sources are compared, and reproductions of three of them are not given. A closer look at the relevant volumes of Paléographie Musicale reveals a slightly different story. “(Depre)ca(mur)” has a pes and a virga in three of the manuscripts but a single penstroke in Angelica 123, the Bologna Gradual reproduced in PM 18;

“do(mine)”

has

a

single pen-stroke

in Saint Gall 359

ofränkische Neumen ersetzt. Wie denn mehrere paläeofränkische Einträge in den Düsseldorfer Handscriften Indizien dafür bieten, dass sie von Händen geschrieben wurden, die in Neumen des deutschen Rereichs geschult waren.” (Arlt, Kunst und Kultur, 839). The Wolfenbüttel fragment, Cod. Guelf 510 Hemst., is the first known example of a Gradual with Paleofrankish neumes (dated ca. 900). It originates from Korvey, founded from Corbie (where gestural neumes were certainly in use), “. .der Notator mit deutschen Neumen vertraut oder gar zunächst in dieser Schrift geschult war” (Arlt, Kunst und Kultur, 842). 103 Levy, Gregorian Chant and the Carolinians, 83-86. 104 McKitterick, Review of Gregorian Chant and the Carolingians, 289. 105 Levy, Gregorian Chant and the Carolingians, 22. .

pes and two strophici in Angelica 123, a virga and a porrectus Gradual (PM 16), and a tractulus then a porrectus in Chartres 47 (PM 11). 106 These neume shapes are melodically compatible, but they are not “essentially identical.” It is also worth noting that the chant is mostly neumadc in melodic density, with a few notes for each syllable. If a single syllable has two notes, the second lower than the first, one would expect all manuscripts to use their local version of a clivis, regardless of whether or not they had access to a common

(PM 1),

in the

a

Noyon

exemplar; correlation need not imply a relationship between manuscripts. A second example given by Levy is that of the Offertory Factus est repente.

neumed two

Levy’s argument

is that the chant fell out of

use

in the north

around the year 800 (after the writing of the unnotated Mont-Blandin AMS source, which contains it). Its appearance in Beneventan manuscripts

indicates that this region had taken on Frankish chant by ca. 800, and that “the neumations are close enough to indicate a common written source.” 107 Levy uses a selection of Benevantan manuscripts to illustrate the similarity of their melisma with that found in a manuscript from Prum (dating from ca. 1000). He does not mention, however, the 11th-century Canosa Missal, the early Beneventan source that is elsewhere so central to his argument. In this manuscript the central part of the melisma is omitted. 108 Pfisterer writes that this is a scribal error: The scribe leaped from the DEFGG figure to the Gabcc one, a mistake that

arose

through copying

the first of these seems

an

adiastematic

identically notated figures

to have

been written for

professional musician, The Prüm

which may

a

exemplar

and

mistaking

for the second. This manuscript

church

explain

dignitary

rather than

the lack of later

a

109 correction.

therefore, has a quite different melisma to that found in the one Beneventan source with which it is roughly contemporaneous; the other Beneventan books used by Levy date from much later and are not scribally derived from the earlier Beneventan

manuscript,

source.

fundamentally, the appearance of Factus est repente in the manuscript, the very manuscript that Levy uses to make his neumatic comparisons, indicates that while the chant was not widely used north of the Alps, it was certainly still sung in this region. Why then More

Prum

106

This fragmentary leaf in PM 11 does not contain the whole chant and is not always securely legible; the World War II destruction of the manuscript means that we will never

be certain of what

107 108

neumes were

written here.

Levy, Gregorian Chant and the Carolinians, Compare ibid 98 and 103.

102.

.,

109

See Pfisterer, Cantilena Romana, 73, and 77-85, for further arguments the Gregorian tradition was transmitted to Benevento in writing.

Levy’s view that

against

time this chant could have gone south to Benevento without neumes) be around the year 800? It could instead have been transmitted south, for whatever reason, at any time until the 11th century (the time of the first Beneventan manuscript containing it). It could even have gone south in oral form, have been

must the

only possible

(with

notated in

or

Benevento,

and

subsequently have been reintroduced (before manuscript) in certain German centers.

the notation of the Prüm Pfisterer thinks it

unlikely that Factus est repente was ever a part of the repertory; the two Frankish sources and the Beneventan ones represent places in which this chant penetrated the repertory, rather than the sole survivals of a previously universal tradition. 110 The multiplicity being of possible interpretations of the data demonstrates that Levy’s case is not strengthened by the Factus est repente neumations. It should be clear that defining how “close” or “essentially identical” core

neumations

might

be is not

finds agreement that

an

exact

two neume

undertaking.

However, when

groups in different

sources are

one

indeed

exactly compatible, is this invariably indicative of a common notadonal origin? In syllabic or neumatic chants, the answer must be no; a syllable with

will tend to be notated either with a pes or a clivis in all Identical neumations of longer melismas might indicate manuscripts. a common origin, although Treider instead suggests that the pitches are so grouped because they were sung the same way and remembered in fixed chunks. Jeffery questions whether one could prove this: “It He also questions its likelihood, may be so, but how is one to know?” given that “the alternative is the easily believable possibility that uniformity results from simple copying, ultimately from one or a few related two notes

111

112

written

exemplars.”

However,

one can

in certain

of the written tradition which rather than

copying

as

the

cases

identify

characteristics

demonstrate remembering

clearly primary activity. 113

Levy’s

identification

Ibid 79. 111 Treider, “Communication,” Leo 569. Elsewhere, Treider invokes Schlager’s evidence of word boundaries coinciding with neume boundaries in texted melismas (see Karlheinz Schlager, “Die Neumenschriften im Licht der Melismentextierung,” Archiv für Musikwissenschaft 38 [1981]: 296-316). “Because of the scruple regarding neumeboundaries and word-boundaries the aggregative property of neumes with respect to their consdtuent notes was shifted, in a way, to words. As Schlager suggested, the words could have functioned as a kind of notation.” See “The ‘Unwritten’ and ‘Written Transmission'," mission’,” 160. 112 Peter Jeffery, Re-Envisioning Past Musical Cultures: Ethnomusicology in the Study of Gregorian Chant (Chicago: Univ. of Chicago Press, 1992), 33. 113 This is most visible when scribes were prompted by the opening of a phrase to complete it in the standard way, although in fact a different cadence is found in most manuscripts. When a mistake of this kind is crossed out and corrected, it is a clear indication that the scribe was using an exemplar, but glancing at it only intermittently. Uncorrected variants of this kind indicate that the scribe who originally made this change (which may have gone through several generations of copying before its appearance in 110

.,

.

.

stereotyped melismas in chants such as supports the contention that stereotyped melismas were recognized as such and could be both performed and notated in fixed ways. 115 It does not necessarily imply that this stereotyping was linked to a notated tradition as early as ca. 800.

of similar

neume

the Justus

utpalma graduals 114

The

groupings

purpose of early

notated

in

manuscripts

The earliest Mass have been identified

Proper manuscripts containing gestural notation by Jeffery and appear to have been written for the

of the celebrant. “If it is indeed correct that many or most of the earliest Graduals are actually components of primitive ‘Missals,’ then the great majority of surviving non-fragmentary sources are Missals of some sort—books intended for use by the celebrating priest and not by the choir of the Hardly any surviving sources were practical copies intended to serve the singers who actually performed the chants.” 116 Primitive gestural notations, then, may have developed as aids for members of the clergy who were not necessarily gifted musicians and who may have required written assistance to maintain (or at least attempt) the uniformity demanded by the Carolingian reformers. If they wished to sing or recite within a private Mass the parts of the Mass usually performed by schola or cantor, their lack of specialist training may have made written assistance essential. 117 A few further remarks about the hypothetical neumed archetype are pertinent here. Whether or not one accepts the view of the Cardine school that the notational styles found ca. 900 have a common origin, and the view of Levy that it would have taken at least 50 years for the original to develop into the different regional styles, it does not necessarily follow that the entire repertory had been notated prior to ca. 900 or that there was a neumed exemplar. In fact, we cannot assume that there was a gradual process of paleographic change over at least 50 years; instead there may have been a sudden transformation of styles use

.

.

.

extant manuscript) was not using an exemplar but was remembering how the phrases of these chants should go. See Hornby, Gregorian and Old Roman Eighth-Mode Tracts, 155-68 for examples and discussion. Many further examples are given in Pfisterer, Cantilena

an

Romana, 33-72.

Levy, Gregorian Chant and the Carolingians, 227ff. See Hornby, Gregorian and Old Roman Eighth-Mode Tracts, chap. 5 for further examples. 116 Peter Jeffery, “The Oldest Sources of the Graduale: A Preliminary Checklist of MSS Copied Before About 900," Journal of Musicology 2 (1983): 317. See also Hiley, “Notation," 89, for a table including many of the earliest examples of notation. 117 See also Treitler, “Reading and Singing,” 161. ll4

115

due to

ideas. 118

new

regions had

Further, one may accept the possibility that different exemplars (particularly obvious in the early Saint

local

without the existence of a single authoritative archetype.119 This is consonant with Hughes’s observation that the idea of notation might travel from place to place without commitment to any particular notational style or form. If a neumed exemplar existed, it does not seem to have been used very much. Hughes concludes this from his work on the nature of the variants found in early manuscripts, Treitler has written about the likely interaction between copying and remembering, and Levy hints at it himself: Instead of just copying a model, he writes, the scribes were actively realizing the notational method. Levy also states that dictation have may played a major role, giving scribes the freedom to personalize

Gall

manuscripts)

120

neumadc

121

shapes. hearing rather

or

If such scribal freedom, reliance on remembering copying, and critical acumen must be invoked

than

in order to account for the kinds of melodic variation and

independence found neumed archetype? 5. Hints

of Notated

in

surviving manuscripts,

Graduals and

Antiphonaries

calligraphic

whither the

in the

common

Documentary

Record

Levy’s

final category of evidence is the 8th- and

9th-century

documents

that mention chant in greater or lesser detail. A small example is Charlemagne’s instruction in the Admonitio generalis of 789, exhorting

clergy to teach the boys psalmos, notas, and cantus. 122 Levy views this verification of late 8th-century notation. Haas has cast doubt on Levy’s reading of this passage. Each discipline is identified separately (“psalmos, notas, cantus, compotum [computus], grammaticam”); aspects to do with chant would be subsumed under “cantus.” “Notas” instead probably refers to Tironian shorthand. Hiley points out that

all as

123

a

marginal

note in one of the

manuscripts

of the Admonitio Generalis

relates “notas” to the activities of the notarius, teaching of writing.124

possibly implying

118

the

Idem, “Communication” (1988), 571. Indeed, Hucke writes that Cardine meant to refer to local exemplars with a central notational (rather than melodic) exemplar lying behind them. See “Gregorianische 119

Fragen,” 328. Hughes, “The Implications of Variants for Chant Transmission,” 72. Levy, Gregorian Chant and the Carolingians, 137-38. 120 181

123 124

I bid., 92-93. Haas, Mündliche Überlieferung und altrömischer Choral, 152-53. Hiley, Western Plainchant, 364.

Abbot Helisacher More extended space is Helisacher’s letter

given by Levy

to a

consideration of Abbot

(ca. 814-22). 125 Zijlstra refutes

on

two

grounds

the

contention that this letter implies the widespread use of notation. 126 The first is Helisacher’s central concern about the inconsistencies within the tradition, which are shared by other writers such as Agobard of Lyon, Amalar of Metz, and Aurelian of Reome. 127 The inconsistencies they bemoan seem to speak against the possibility that the tradition was already fixed in writing. Zijlstra examines the vocabulary used in different sections of the letter to provide his second ground for reservation about Levy’s hypothesis. Sentences which Levy thinks refer to the notation of music, such as “And so it came about, with divine grace

helping,

that what

reason

and

authority approved

was

placed

in that

work, and what was missing was supplied on the basis of multiple sources,” 128 use vocabulary associated with the correction of texts, but not music, seen earlier in the letter. The document is primarily concerned with the appropriate arrangement and notation of the texts so same

that

responsories and their verses might be correctly sung, a concern by Aurelian as being central to Helisacher. 129 Bullough and Corrêa also interpret this key passage as meaning that musicians were summoned to make and to teach new musical settings of liturgical texts, to be passed on orally with copies of the texts. 130 I remain unconvinced confirmed

that this passage refers

to

notation; Helisacher’s

concern was

that correct and authoritative chant texts

scribes

or

(not corrupted by the vice of cantors) should be written down, and that chants (including the melodies) were to be taught to Helisacher's

the arrogance of

the correct

cantors

by the masters

of the art of

melody

(melodiae artis

magistros)

whom he had summoned.

Aurelian

of Réôme

In his Musica disciplina Aurelian of Réôme unequivocally expected notation to be used on many more than the handful of isolated occa125

In “Adémar de Chabannes,” Grier

comes

to

the

same

conclusion

as

Levy,

but

without

presenting any additional evidence. 126 Marcel Zijlstra, “Communication,” Journal of the (1997): 238-42.

American

Musicological Society

50

See the discussion above. “Ita res divina amminiculante gratia sucessit, ut quod auctoritas et ratio vindicabat in eodem opere poneretur, et quod deerat plurimorum documento suppleretur.” Text and trans. are from Levy, Gregorian Chant and the Carolingians, 182. 129 “Communication,” 241. Zijlstra, 130 Donald A. Bullough and Alice H. Corrêa, “Texts, Chant, and the Chapel of Louis the Pious,” in Peter Godman and Roger Collins, eds., Charlemagne's Heir (Oxford: 128

Clarendon, 1990), 508.

manuscript Valenciennes 148 gestural). The treatise may confirm the existence of notation by the 840s, 131 but Levy considers that it also suggests the existence of a neumed Antiphoner by that time.

sions where

(sometimes

neumes

are

found in the

Paleofrankish and sometimes

Aurelian makes comparisons between chants from "isolated points repertory,” which Levy claims implies a neumed Antiphoner through which the reader would thumb. 132 In a culture where novice monks were expected to know and to be able to quote any verse of any psalm as required, the mental flexibility required to remember and compare chants sung at different times of the year was perhaps not quite so unbelievable. If this mental agility were indeed beyond the

in the

ability

explain the almost total 9th-century writer mentions not listed in surviving Carolingian

of the average cantor, it might help to impact of Aurelian’s work: “no other

lack of

the text

quotes

or

from

library catalogues.” 133 The notated examples

it, and

it is

in Valenciennes

148

are

of

noeane

formulas

These were not notated in early Antiphoners, and Levy sees this as evidence that other examples were not notated because Aurelian expected his readers to have access to them in neumed Antiphoners. One may observe firsdy that notation was clearly expected by Aurelian for large numbers of examples of introit verses which did not receive it in Valenciennes 148, despite their lack of notation in early Antiphoners. More importantly, the main subject of the treatise is and introit

the

verses.

will

pertain

antiphons and guidelines for getting antiphon right. Naturally the musical examples regardless of whether or not they were regularly

text treatment of

appropriate

the link between

verse

to

this,

and

notated elsewhere.

lengthy verbal descriptions in Chapter XIX recognizably manuscripts.134 But one need not invoke a neumed transmission to explain this, and one wonders why, if the notated tunes were readily available, Aurelian bothered with the verbal descriptions. Would the point not have been made equally well by use of notated examples?135 Aurelian’s

refer to the melodies found in later

Levy’s final

claim—that the list

given

in

Chapter

XIX of what

a

singer needs to have in his memory is remarkably restrained and indicates the use of notation for the rest of the chants—ignores the fact that the list is of the very chants that the discourse is about.136

“Although

131 As Grier acknowledges, the text seems to be an amalgam, and we do not know whether the references to notation are original or not. “Adémar de Chabannes,” 81-82. 132 Levy, Gregorian Chant and the Carolingians, 192. 133 Jane Bellingham, “Aurelian,” in NGII, 2: 185. 134 Although they are not as close in their details as Levy claims; see above. 135 Treitler, “Reading and Singing,” 160. 136 Levy, Gregorian Chant and the Carolingians, 193.

anyone may be called by the

perfect unless he the melody of all

name

has

both of the modes and the of the

responsories.” l37 documentary

The

of singer, nevertheless, he cannot be in the sheath of his heart

implanted by memory the verses through all the verses

modes, and of the differences antiphons, introits, and

of the

evidence does not confirm the existence of

neumed Graduals in the

early years of the 9th century, and indeed, Aurelian's painstaking descriptions of melodic shapes suggest that notation was not yet the norm in the 840s.

6. Conclusions

Levy’s methodology is inspirational. The answer to the question of why notation began to be widely used in Western Europe may lie concealed within the documentary background (coupled, as McKitterick states, with knowledge of the historical situation), within the evidence of parallel chants in related melodic traditions (although these comparisons should be undertaken at the level of whole genres

when and

rather than isolated chants, as Nowacki has cautioned in another context),138 and within the evidence of the notational record itself. It may a notated archetype emanating from Charlemagne’s court ca. 800 indeed formed the model for widespread copying of notated Graduals, but the results of the present study do not support such a conclusion. The nature of the melodic flexibility attested to by the earliest

be that

manuscripts,

the apparent emergence of notation for the

use

of monastic

music theorists and nonmusical celebrants in the mid

9th century, tantalizing possibility supported by analytical findings that medieval memories were capable—given unremitting training and enormous —of stably maintaining the Frankish chant tradition, peer pressure 139 suggest to me that if there was a written archetype at all, it did not emerge until the mid or later 9th century. Frankish chant may have been notated in the time of Charlemagne or even Pippin. Levy’s work has forced chant scholars to reexamine their assumptions and to take these possibilities seriously. We cannot idly claim "oral transmission” without rigorous justification. Whether or not Levy turns out to have been right in his reading of the evidence, his

and the

Aurelian of Réôme, The Discipline of Music (Musica disdplina), trans. Ponte, 45. 138 “The Gregorian Office Antiphons,” 270-75. This contrasts starkly with Nowacki, Levy’s comparisons of selected neumations of individual chants, discussed above, from which he draws conclusions about the transmission of the whole tradition. 139 See Rubin, Memory in Oral Traditions, 122-45, for a discussion of the processes by which oral traditions are handed down from person to person. The concept of “nets” is particularly helpful here; material forgotten by one person is supplied correctly by another 137

(133-35).

scholarship

deserves celebration for the

that it has forced

questions

others to confront.

University of Durham Appendix Facsimile editions from which the

figures are reproduced:

Paléographie Musicale 1: ‘Le Codex 339 de la Bibliothèque de Saint-Gall (Xe Siècle): Antiphonale Missarum Sancd Gregorii’ (Solesemes , 1889 ) Paléographie Musicale (IXe-XIe Siecle): 1894 )

‘

4 : Le Codex 121 de la Bibliotheque d’Einsiedeln ’ Antiphonale Missarum Sancti Gregorii (Solesmes , ‘

Antiphonale Missarum Sancti Gregorii, IXeBibliothèque de Laon ( Solesmes 1909 ) 11 : Musicale Paléographie Antiphonale Missarum Sancti Gregorii, Xe Siècle: Codex 47 de la Bibliothèque de Chartres (Solesmes 1912 ) Paléographie Musicale Series 2, 2 : Cantatorium, IXe Siècle: N° 359 de la Bibliothèque de Saint-Gall ( Solesmes 1924 ) Paléographie

Musicale

10 :

’

Xe Siècle: Codex 239 de la

,

’

,

‘

’

,

ABSTRACT Since the 19th century, scholars have been attempting to discover origins of Gregorian chant and to establish when musical notation began to be widely used in its redaction. For almost 30 years, Kenneth Levy’s scholarship on the subject has been hugely influential. He hypothesizes that Gregorian chant was notated in the time of Charlemagne (742-814), or even Pippin (714-768). There are alternative ways of reading the 8th- and 9th-ccntury evidence, however, and largely oral transmission of the Gregorian melodies until the later 9th century the

cannot be ruled out.

[14] Chant research at the turn of the century and the analytical programme of Helmut Hucke EDWARD NOWACKI

The end of a century, like the end of any era, is a time to take stock and lay plans. This is no less true of plainsong and medieval music than of any field of human endeavour. So it comes as no surprise to encounter in our professional literature exhortations to re-envision our subject and to contemplate its direction in the twenty-first century. Richard Crocker has been particularly cogent in recommending the abandonment of lost causes and the adoption of new or 1 neglected research methods. The attempt to read the oldest melodic texts of plainchant as the trace of the sumptuous oral tradition that preceded them, however tantalizing, 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. Without wanting to deny the good sense of these timely ruminations, it is essential for the balance and thoroughness of our stock-taking to consider complementary methodologies, especially ones at risk of being discarded before they have been exhausted or even recognized. One such methodology is presented in Helmut Hucke's 'Gregorianischer Gesang in fränkischer und altrömischer article Adler

in Archiv

2

für Musikwissenschaft in 1955. The is the kind of seminal work that the educational philosopher Mortimer considers primary teaching, in contrast to the secondary transmission of Überlieferung', published

1

Richard Crocker, 'Gregorian Studies in the Twenty-First Century', Plainsong and Medieval Music,

2

Vol. 12, pp. 74-87, used here by permission.

4

(1995),

33-86.

DOI: 10.4324/9781315090283-17

Oral and Written Transmission in Chant

the

thought

of others. In the first

case

the teacher is the thinker him-

or

herself,

from the original thinker, digesting that learning, and passing it on to his or her fellow learners. Primary thinkers in Adler's view should be read in their own words. The visionary potential of their writing and its attention to fundamentals not yet assumed as common knowledge overcome any impediments that the difficulty of the subject matter in the latter,

just

a

better student,

learning

and the writer's style may present. These superior educational qualities are retained even when primary thinkers have not yet solved all the attendant 3

later to change their mind. Taken as a whole, Hucke's work might appear to be some of the least neglected in the field, yet it is striking that this particular article, a masterpiece

problems

or were

of analytic thought laying out a completely new approach to the comparative study of Gregorian chant, is rarely mentioned. Some of the article's neglect may be due to its lapidary density and liberal use of jargon, now dated, from 4

the milieu of the Deutsches Volkslied-Archiv, where Hucke worked as a research assistant under Walter Wiora from 1948 to 1950. But that is compounded in the English-speaking world by the language barrier and the assumption that the gist of Hucke's thought can be obtained from the one article that he published in the

English language.

5

Yet another reason for the article's neglect is the general lack of interest in the kind of imaginative analysis of melodic texts that Hucke practised. Many of his readers have concentrated on his treatments of the history of musical genres and practices, the external history of Gregorian chant. But of equal interest to him and, I would argue, the area of his most original contribution has been the inner history of chant, the history of its musical style as it can be inferred from the comparative study of its texts. 6 The pioneering principles and methods adumbrated in the article include the rejection of the hypothesis that different oral traditions of singing could have come about in the single cultural milieu of Rome, open-mindedness to the possibility of convergent evolution of melodic forms (contrary to the familiar philological expectation of divergence), the hypothesis that tunes could be generalized and transmitted apart from their particular instances, and the role of verbal similarity in prompting the inadvertent replication of melodies. Of deeper significance are the article's high analytical demands, requiring an ability to recognize the extraordinary and the ordinary on the basis of extensive knowledge 3

Mortimer

4

David

5

6

J. Adler, How to Read a Book (New York, 1940), 48-64. Hiley's summary references to it in Western Plainchant: A Handbook (Oxford, 1993), 79 and 539, are exceptional. Peter Jeffery, in his detailed examination of Hucke's thought in Re-Envisioning Past Musical Cultures: Ethnomusicology in the Study of Gregorian Chant (Chicago, 1992), does not mention it, limiting his digest to works of Hucke’s published since 1963. Helmut Hucke, Toward a New Historical View of Gregorian Chant', Journal of the American Musicological Society, 33 (1980), 437-67. I use 'text' in its strict philological sense as anything written. Words in alphabetic script and melodies in musical notation constitute the complete text of plainchant.

Oral and Written Transmission in Chant

of the

repertory's vocabulary

of musical formulas,

phrase-syntax

and tonal

language.

Perhaps most significant of all is the article's visionary, if tacit, assumption of structuralist perspective, insisting that comparisons between the two traditions of Gregorian chant must take account of the synchronic relationships within each tradition. This view seeks evidence of affinity not only in the material similarities between members of isolated cognate pairs, but also in the relational similarities between the two traditions viewed as autonomous systems. Each tradition is seen as able to function - that is, to be coherent and comprehensible in consequence of the internal network of paradigmatic relationships among its -

current members without regard to their genealogy. When that network, expressed in such terms as same, different, typical, unique, etc., is found to be analogous to the corresponding network in the cognate tradition, the relational similarity is interpreted as a sign of a deeper affinity than that exhibited by

material similarities between individual cognate items. In view of Crocker's recent admonitions, directed not to Hucke, but perhaps to some of Hucke's readers, it is important to remember that Hucke's domain of interest in this article, as in much of his later work, has been the period from the Frankish reception of Roman chant beginning in 754 through to the

copying of the Old Roman manuscripts in the eleventh and twelfth centuries. While he occasionally speculates on the state of the Roman prototype just prior to the Frankish reception, his main focus is on the similarities and differences between the Frankish and Old Roman traditions as they are read in their texts. The condition of the common prototype and its prehistory are outside his domain of enquiry. That is not to say that he forgoes historical speculation, for he persistently reads the texts of each tradition as testimony relevant to the history of its cognate. The means by which he does so entail no probing of the unknowable, but only laying out the conceivable hypotheses and ranking them by probability. These alternatives are sometimes worked out in thought experiments. He imagines the logical, psychological and musical

surviving

consequences of considering certain melodic phenomena as having originated in one of the branches and compares them to the consequences of considering the phenomena as merely received. In the light of such lucid comparisons, the more probable alternative is usually clear and often compelling. Where does the work forecast by Hucke's article but stiff unaccomplished stand with respect to Crocker's vision of a more humanistic, critical musicology of plainchant? Clearly, historical understanding and critical appreciation are compatible; in fact, the former promotes the latter by forming part of its allimportant context. Moreover, with its insistence

diachronic

on

synchronic analysis

as

a

preliminary enquiry, Hucke's approach to historical questions is more attentive to context than non-structuralist approaches are. And what of Crocker's agnosticism towards the oral prehistory of plainchant? Here it must be stressed that Hucke regards the prehistory of the Old Roman tradition as to

entirely oral, for he reads the texts of the Frankish tradition as recording part of that prehistory, at least obliquely. For those disappointed in the sobering

not

realization that we will never know how Gregorian chant before the Carolingians be as it is how it was originally shaped in the halls of the Roman the recognition that analogous information can be had for schola cantorum Roman chant between 754 and 1071 (the date of Bodmer C 74, the earliest Old Roman manuscript) should be no small consolation. The case for reading the came to

-

-

Old Roman texts as transmitting something of the prehistory of the Frankish tradition is harder to make for obvious chronological reasons; Hucke makes it, and the results are more credible than mere speculation based on theoretical assumptions. More readings of this kind need to be done before the method is discarded as obsolete. What follows is a translation of the article with commentary. First, however, three technical terms require special comment. They are bearbeiten, umsingen and zersingen. Hucke's use of bearbeiten is clear. It refers to the work of deliberately changing, modifying or adapting melodies in a regime of literalness. Those who perform such work are arrangers (Bearbeiter) and their work is an arrangement (Bearbeitung). He implies that such work employed the instrumentality of musical notation, but does not rule out that memorization by rote may have sufficed. The term is opposed to umsingen, which refers to the change, modification or adaptation of melodies in oral tradition, and whose effect is to save them from 7 becoming obsolete artefacts by absorbing them into the current melodic style. I have translated this term as 'to transform orally', implying a more or less deliberate intent, but without the technology and assumptions of writing. Zersingen, to change over the course of time through constant inaccurate singing, has no precise English equivalent. It may connote a range of viewpoints from descriptive neutrality to disapprobation, stopping short, however, of 'corrupt', which in any case may be expressed by the ordinary German word verderben. I have translated zersingen as '(cause or allow) to deteriorate, lapse or mutate', depending on the implied judgment.

Gregorian

chant in Frankish and Old Roman transmission

In the year 1891 the Benedictines of Solesmes published in the second volume of their musicale a melody for the gradual 'A summo caelo' that comes from a in the Vatican Library [lat. 5319] and that differs from the current Gregorian 8 version of this tune as well as from the Ambrosian one. A voluminous footnote informs us that the manuscript in question is a complete book of chants for the Mass furnished throughout with melodies of this same remarkable type, that yet another such book exists [San Pietro F 22] along with one for the Office [San Pietro B 79], and that the melodies beyond a doubt were once sung, at no less a place than St Peter's. The author presumes that the melodies are late urban Roman distortions of the Gregorian.

Paléographie manuscript

7

8

Walter Wiora, 'Das

produktive Umsingen

deutscher Kirchenliedweisen in der Vielfalt

Stile', Jahrbuch für Liturgik und Hymnologie, 2 (1956), 47-63. Paléographie musicale, 2 (Solesmes, 1891), 6.

europäaischer

Since then only a few further musical examples from this trove of melodies have been In all probability that is one of the reasons for the peripheral attention that has been paid to them in the research community. And when scholars found themselves in the predicament of having somehow to fit them into the history of Gregorian chant, some adopted the hypothesis of Paléographie musicale, while others followed the hypothesis advanced by Dom Andoyer that they were examples of a pre-Gregorian melodic dialect. 9 Only in 1950 was interest turned anew to the puzzling melodies. In a paper given at the Congresso internazionale di musica sacra in Rome, Bruno Stäblein advanced a new hypothesis: that this trove of melodies (in the meantime one further book of Office chants had been discovered in London [British Library, Add. 29988]) was the real Gregorian chant. Between 650 and 680 these melodies were reformulated, and this reformulation is the chant that we commonly designate as 'Gregorian' today. The real Gregorian chant continued to be cultivated only in St Peter's and in the Lateran until the thirteenth century. For this chant Stäblein introduced the term 'Old Roman'. 10 Two years later the discovery of a further Old Roman book of Mass chants was announced, a book copied in the year 1071 for Santa Cecilia in Trastevere [Bibl. 11 Bodmeriana, C 74]. And most recently Michel Huglo has placed the deliberations the Old Roman chant upon an entirely new foundation. 12 Huglo first put concerning a list of verbal variants and differences in the together liturgical order by which all Old Roman manuscripts can be distinguished from all Gregorian ones; he is then able to refer to a whole series of manuscripts without melodies that bear witness to the Old Roman chant indirectly. His investigation comes to the following conclusion; the Old Roman chant is indirectly verifiable back to around 800. It figured in Roman and middle Italian churches up to the thirteenth century as the official liturgical chant. Outside of Italy it is indirectly attested through fragments of a Fulda manuscript of the ninth century, whose script shows Anglo-Saxon influence, and further through what Amalarius reports shortly after 844 on the particulars of a manuscript that Pope Hadrian I had sent to a monastery at Corbie. 'Gregorian' chant is likewise attested indirectly since the ninth

published.

century, then in staffless notation since the tenth century, and

in notation legible to us since the eleventh and significantly in the Frankish kingdom first of all. We therefore speak of this transmission in what follows as the Frankish. In Rome itself the Frankish version of the Gregorian melodies is not transmitted at all before the thirteenth century. These circumstances compel us to conclude that the Frankish transmission does not come from Rome, but arose in connection with the introduction of Gregorian chant in the kingdom of the Franks. -

Hucke's enthusiastic, if preliminary, embrace of Huglo's just published theory can be explained by its ability to dispose of the claim in Paléographie musicale 2 that the Old Roman chant was a late urban Roman corruption of the Gregorian that is, a manifestation of isolated and relatively recent negligence of little value

-

9

10

Raphael Andoyer,

'Le chant romain

antégrégorien',

Revue du chant

grégorien,

20 (1911-12), 69-75

and 107-14. Bruno Stäblein, 'Zur

Frühgeschichte des römischen Chorals', in Atti del congresso internazionale di musica sacra, Roma, 1950 (Tournai, 1952), 271-5. See also idem, 'Alt- und neurömischer Choral', in Kongressbericht Lüneburg 1950, 53-6; idem, 'Zur Entstehung der gregorianischen Melodien', Kirchenmusikalisches Jahrbuch, 35 (1951), 5-9; and idem, 'Choral', in Die Musik in Geschichte und

Gegenwart (Kassel, 1952), II, cols. 1265-1303. Jacques Hourlier and Michel Huglo, 'Un important témoin du chant "vieux-romain": le Graduel de Sainte-Cécile du Trastévère', Revue grégorienne, 31 (1952), 26-37. 12 Michel Huglo, 'Le chant "vieux-romain": Liste des manuscrits et témoins indirects', Sacris Erudiri, 6 (1954), 96-124. 11

as testimony to the origins of the chant or its authentic international tradition. Huglo argues that the tradition is neither isolated nor recent, but one that reaches back to the ninth century and affects a larger region, central Italy, and some of its outposts. In the interim Huglo's thesis has been discounted by most scholars, who fail to be persuaded that a textual tradition only of words necessarily

entails the exact melodies that appear with those words some two centuries later. This very reservation is already implicit in Hucke's article, for he alleges numerous cases of change in the Old Roman melodic tradition after its separation from the Frankish. If such changes were as common as he claims, we can hardly assume that the melodies notated over Old Roman words in the eleventh century are the same as those to which those words were sung in the ninth. Hucke's remark that 'in Rome itself the Frankish version of the Gregorian melodies is not transmitted at all before the thirteenth century' is literally accurate but oversimplifies the probable history of the sources. Klauser had made the observation, to which Hucke has never objected, that the papal court had begun to order

liturgical books from Reichenau in the mid-tenth century, beginning a long complex history of Roman dependence on ultramontane sources. In all probability such dependence entailed the reintroduction into Rome of a tradition of singing that had temporarily been cultivated away from its place of origin, thus providing an explanation for the co-existence of two rather different melodic dialects within the same cultural environment. While Hucke always regarded as completely untenable Smits van Waesberge's hypothesis, later defended by van Dijk, and a similar one of Stäblein, each alleging the emergence of two melodic dialects within the city of Rome, there is nothing in Hucke's reasoning that would have 13

and

ruled out their co-existence as a result of later reintroduction. 14 Again, Hucke's article implicitly acknowledges this possibility, for it observes that the four tracts of Holy Saturday in Old Roman transmission appropriate more or less exactly the melodies transmitted in Frankish sources. The liturgists following the Old Roman tradition could have obtained the melodies by consulting foreign sources, but is seems more likely that they observed them being sung at Roman foundations the international (i.e., Frankish) melodic tradition. In the interim other using of reintroduction, probably through intra-urban contact, have been symptoms observed in the Old Roman manuscripts. 15 13

14

15

liturgischen Austauschbeziehungen zwischen der römischen und der fräankisch-deutschen 8. bis zum 11. Jahrhundert', Historisches Jahrbuch der Görres-Gesellschaft, 53 (1933), 169-89; reprinted in his Gesammelte Arbeiten, ed. Ernst Dassmann (Münster, 1974), 139-54. Joseph Smits van Waesberghe, 'L'Etat actuel des recherches scientifiques dans le domaine du chant grégorien', in Troisième congrès international de musique saerée, Paris, 1957 (Paris, 1959), 206-25; idem, 'Neues über die Schola Cantorum zu Rom', in Zweiter International Kongress für katholische Kirchenmusik, Wien, 1954 (Vienna, 1955), 111-19; S. J. P. van Dijk, 'The Urban and Papal Rites in Seventh-, and Eighth-Century Rome', Sacris Erudiri, 12 (1961), 411-87; Stäblein, 'Zur Frühgeschichte'; idem, Die Gesänge des altrömischen Graduale Vat. lat. 5319, Monumenta Monodica Medii Aevi 2 (Kassel, 1970); Helmut Hucke, review of Stäblein, Die Gesänge, in Musik und Altar, 24 (1972), 138-41. See, for example, Thomas H. Connolly, 'The Graduale of S. Cecilia in Trastevere and the Old Roman Tradition', Journal of the American Musicological Society, 28 (1975), 413-58; and Edward Nowacki, 'Constantinople-Aachen-Rome: The Transmission of Veterem hominem', in De Musica et Theodor Klauser, 'Die Kirche

vom

One thing that Hucke does not claim is that there is such a thing as Frankish chant or that Gregorian chant is of Frankish origin. Nothing could be further from his intent. He maintains only that Gregorian chant, a presumptive Roman creation, is known to us only in the form in which it was edited by Frankish cantors in the ninth and tenth centuries and by Roman cantors in the eleventh and twelfth. By insisting that the chant transmitted in both manuscript traditions is Gregorian, Hucke maintains philological symmetry, refusing to characterize either tradition as remote from the common ancestor. His argument is based on the important distinction between the tangible raw materials of research, Frankish and late

more

Roman

manuscripts (the

term Old Roman is

property that they

a

misnomer) and the intangible intellectual

may attribute regional provenance confidence to the manuscripts, by no means do such attributions automatically attach themselves to the manuscripts' contents. Applying the more general term Gregorian to the chant and the more particular terms Frankish and Old Roman to the channels of transmission, as Hucke does in his concise title, is the most accurate and responsible way of presenting this complex set of conditions. or

late dates with

transmit. While

we

some

Chant research now sees itself facing the task of subjecting the testimony on the introduction of Gregorian chant in the kingdom of the Franks to a new and scrupulous testing and above all of examining the Old Roman melodies even preliminarily. What does the comparison of the melodies have to say about the bibliographic findings? Are the Frankish melodies the result of a consistent arrangement (Bearbeitung), and if so, an arrangement of the Old Roman melodies that lie before us? Are the Old Roman melodies then still the same as at the time of the separation of the two transmissions; are they perhaps even still the same as at the time of Gregory? Do they represent then the true, authentic Gregorian chant? Or are they corrupted versions of older melodies, or, as Walther Lipphardt thinks, late witnesses to an improvisational practice that is transmitted 16 Is the earliest state perhaps the one to us in an earlier state in the Frankish melodies? represented by the Frankish itself; did the Franks then simply appropriate Roman chant in its then-current state, and does the Old Roman transmission preserve nothing more than later versions of those same melodies? Since the oldest manuscript is in private hands and not available [Bibl. Bodmeriana, MS C 74, now available in facsimile], we will take as a basis for our investigation the second oldest manuscript, Vat. lat. 5319, written around 1100, 17 and for the Frankish transmission, the Vatican Edition, which very accurately follows the oldest witnesses. Furthermore, we shall limit ourselves for the most part to a specific part of the repertory, the graduals, of which a few typical examples will be given. Let us consider first a psalmodic tune (Weise) that from the perspective of formal design assumes a special place among the gradual melodies. (See Ex. 1 .) Almost a quarter of all graduals are sung to this tune, which is assigned to the second mode. First

16

17

Cantu: Studien zur Geschichte der Kirchenmusik und der Oper: Helmut Hucke zum 60. Geburtstag, ed. Peter Cahn and Ann-Katrin Heimer (Hildesheim, 1993), 95-115. Walther Lipphardt, 'Gregor der Grosse und sein Anteil am römischen Antiphonar', in Atti del

congresso internazionale di musica sacra, Roma, 1950 (Tournai, 1952), 248-54. the author was able to satisfy himself that the differences between manuscripts Phillipps 16069 [now Bibl. Bodmeriana MS C 74] are negligible. The belong closer together than Vat. lat. 5319 does with the thirteenth-century

Through spot-checks

Vat. lat. 5319 and two

manuscripts manuscript Archivio

di San Pietro F 22, which exhibits

slight

variants.

Ex. 1 consider the Old Roman transmission, designated with the letter R. The response has five segments. The tune is adapted to various texts by inserting repercussions for extra syllables and by dividing and contracting melismas to fit the accents of the words. The verse exhibits exactly the same form. Only its first two segments have distinct melodies of their own; with the fourth segment, and sometimes with the third, it merges with the melody of the response. This third segment is found, apart from two exceptions about which we will speak presently, only in pieces that have a longer text, altogether nine responses and five verses. The appearance of this segment in responses does not make its appearance obligatory in the respective verses and vice versa. This segment differs from the other segments of our tune, however, mainly by using melodic material of the gradual of the fifth mode according to principles of stylization that are usual in this melodic complex. Phrase 3 is thus an insert by which the tune can be expanded in case of need.

Stilisierung Hucke adopts a familiar term of musical analysis not entirely apposite to the phenomenon that he is trying to describe. For in its usual meaning stylization refers to a kind of transformation by which musical forms lose their original function, retaining only its external tokens. The new forms become accepted as conventional allusions to the old, yet admit a certain hypercultivation that is incompatible with the forms' original functional intent. In the case of the melody-type for the second-mode gradual, Hucke's evidence does that is, the not so much suggest hypercultivation as simple stereotyping when correct a fixed of surface into state, singing would patterns merger single With

-

permitted variety within the general parameters of the type. Scholars uniformity is a sign of a young tradition, not yet varied, or a mature one, where frequent repetition has caused related versions to become not merely similar but identical. The former assumes divergent evolution, the latter convergent. With his use of the term stylization, Hucke clearly votes for have

debate whether such

convergence in this case. Of the third segment Hucke says only that it uses 'melodic material of the gradual of the fifth mode according to principles of stylization that are usual in this [second-mode] melodic complex'. Here he may have intended to imply a disconnection from function, where a passage was made to fit, or was recognized as apt, in a context for which it was not originally composed. He cannot be blamed for failing to anticipate the work of Leo Treitler, whose view on the matter he has come to share, in which such apparent cross-modal (even

cross-generic) mixing is viewed as surface stereotyping that is, the rendition of passages whose underlying tonal structure is native to two modes (or genres) in ways that are not merely similar but identical, thereby giving the impression to observers considering only the musical surface of quotation or importation of material from one mode (or genre) into another. With the benefit of forty years' hindsight Hucke's readers may regret that he mentioned mode at all, for the old Gregorian Proper is a pretheoretical tradition whose tonal structures are not aptly described by the terms of the octonary system. In particular their common recitation on c, their projection of the a-c dyad as the primary interval of reference, and their projection of F-a and c-e as secondary intervals unite the two 'modes' more than they divide them. To say that the middle segment of the second-mode gradual makes use of melodic material from the graduals of the fifth mode is both inaccurate, -

since the material is at home in each, and understated, because the apparent mixing affects more than just the one segment. Yet if we prescind from Hucke's old-fashioned terminology, we may discover a still durable methodological principle: the importance of observing underlying tonal structures especially in cross-modal and cross-generic contexts what Treitler later called formulaic systems and of acknowledging the more ephemeral nature of the forms that they sometimes assume on the musical surface Treitler's stereotyped -

-

-

'formulas'. In its

methodology

if not its conclusions, Hucke's

analysis

was

prescient. The other approximately twenty graduals following this tune exhibit certain peculiarities, of which here let only the following be mentioned: the gradual response for the vigil of Christmas and the verse of the gradual for Wednesday in Holy Week make use in each case of the inserted segment 3, but leave out segment 4. Evidently these texts were fitted to the tune later, when segment 3 was no longer understood [by Roman cantors] as something to be inserted only in cases of longer texts. Now compare the Frankish version, designated with F in Ex. 1 The differences do not affect the melodic structure; they are found instead mainly within the articulating melismas. The Frankish version [of those melismas] exhibits a tendency towards wide, emphatic, space-spanning melodic motions; the Old Roman rather towards melodies .

made up of small members (see vestras, portae). This observation asserts itself over and over again in the comparison of the two transmissions. Both versions appear securely formed. It is out of the question that one is a deteriorated (zersungene) version of the other. Rather what lies before us is a reconstitution of the melody, and this impression gains in strength as we compare the others. For they reveal that the distribution of individual segments of the verbal text under the melodic segments differ in the same piece. For example, the Old Roman and Frankish versions of one gradual co-ordinate the words and melodic segments as follows:

can

R

I

n

Domine Deus

converte

in

IV et

nos

F

Domine Deus virtu turn

ostende

et

salvi erimus

et

salvi erimus

faciem tuam

virtu turn converte

nos

et

ostende faciem tuam

In other words it is the tune in general that has been transferred from one repertory into the other, not particular instances of it. But more: in Old Roman chant, segment 4 is missing twice, even though the insert, segment 3, is present. In Frankish transmission that is the case in six additional gradual responses and in four additional verses. The [Frankish] arranger, in other words, did not recognize that segments 1, 2, 4 and 5 form the skeleton of the tune and that the melismas of the third segment, typical of an entirely different group of melodies, are only an insert in cases of need. We may conclude from this that the version in the Old Roman repertory is the older, and that what we have before us in the Frankish repertory is a later arrangement of that version.

Hucke describes a tendency, already nascent in Roman performances, to forget the structural principles that originally guided the composition of this melodytype and to use segment 3 incorrectly as an alternative to segment 4. He leaves his readers to speculate whether the Franks erred spontaneously or mistook errors in the singing of their Roman teachers as legitimate examples. To appreciate his argument, one has to recognize that he has limited his consideration in this case (he does not do so elsewhere in the article) to divergent change, inexplicably neglecting the possibility that the greater regularity of the Old Roman versions could be the result of later correction. In the conclusion of the article he corrects this fault by allowing for the possibility in Old Roman transmission of 'standardizing intervention'. That should warn readers who wish to follow Hucke's thought programmatically that he did not intend to exclude the possibility of convergence in Old Roman transmission under a regime of self-correction that occasionally suppressed variation and reasserted a condition of relatively strict uniformity. 2 shows three graduals from a group of about nine melodies. Let us first compare the two Old Roman melodies in staves 1 and 3 with each other. The first segments [Juravit Dominus/Exaltabo te Domine] with the exception of their initia are alike. Alike also are the beginnings of their respective second segments [et non paeni-/quoniam suscepi-]; these segments differ then in their respective continuations. The respective third segments, with the exception of their entrance figures and closing melismas, are almost note-for-note the same [in aeter-/inimi-]. The closing segment is shorter in the second melody [because it has fewer syllables], but shows a certain similarity to the latter half

Example

Ex. 2 of the corresponding phrase of the first melody [cf. super me with -nem Melchisedech]. The closing melismas are alike. About half the melodies offer a similar picture; in the rest the differences go further, and furthest of all, I would say, is the gradual for Quinquagesima [Tu es Deus, not reproduced], where only the initium and the passage from the end of the closing melisma of the third segment on are typical, and everything else in that melody is unique. Exsurge Domine, for Monday of Holy Week, the third melody reproduced here, presents a special case. The opening section follows exactly the gradual for the Saturday after the Fourth Sunday of Lent [cf. Tibi Domine derelictus est in Old Roman transmission, not reproduced in the article] but stands here a fifth too high. That is evidently a sign of melodic mutation (Zersingungserscheinung). We have here nine [Old Roman] melodies of which not one is fully like another, but between which correspondences arise this way and that over shorter and longer spans.

What is this group of melodies? Have diverse melodies developed convergently into one; have they been sung together? Or has a single uniform tune been sung apart, into

divergent

versions? Or is it

a

case

of

an

improvisatory,

free series of

typical

melodic

segments strung together in various orders? This last-named formal principle has affected these melodies to a small degree, but does not lie at their foundation. That is made clear by the fact that various typical melismas have not been used ad libitum, but rather individual typical melismas repeatedly have been replaced by singular ones. Some melodies of this group evidently at one time formed a more or less uniform tune, more uniform in any case than the form in which it is transmitted to us, and this tune has been allowed to mutate quite egregiously (ist recht stark zersungen worden).

The evidence is that about half the group of nine is

moderately stereotyped

in

its use of formulas for structurally well-defined positions in the melody, whereas the remaining half substitute unique melismas in those positions. Formulas do not 'wander'; when they recur, they do so in the same position. The implicit argument is that such purposeful positioning would occur only if the formulas were understood as having a function in a once well-defined melodic type. The fact that not all melodies are stereotyped in all their parts is taken as evidence that they have fallen into decline. In my opinion, such a judgement is premature until one is able to generalize about the underlying formulaic system that regulates the melodies in the Old Roman branch. Perhaps the allegedly mutated versions are simply less stereotyped but nevertheless plausible versions of that underlying structure. Perhaps they are en route not from a stereotyped tune but towards one. On the other hand, in some cases, say, the gradual for Quinquagesima mentioned above [Tu es Deus], the possibility is not ruled out that unique melodies have appropriated parts of this tune, in this case a characteristic melisma and the closing segment. Manifestations of oral transformation (Umsingung) and melodic mutation (Zersingung) can be pointed out in the graduals rather often, and the more one pursues these manifestations, the clearer becomes the multiplicity of their forms and layers. Let us add just one more remarkable example of the mutation of the tune here under discussion. A rather extensive part in the gradual verse for Saturday after the Fourth Sunday of Lent that does not stand out of its melodic context as unusual in any way recurs note-for-note in the gradual of Septuagesima. Both times this fragment of melody it is hard to call it anything else is sung to the same text, in opportunitatibus in tribulatione. A scribe or singer working one day evidently fitted the same melodic fragment to the same text, and the Frankish tradition follows the Old Roman exactly. -

-

In this brief remark Hucke does not challenge the assumption then current in German folksong research that inadvertent melodic changes prompted by the words are evidence of decline rather than a normal dynamic of healthy creativity.

is important in drawing attention to one of the prompting composer-singers in oral tradition to repeat themselves, resulting in convergent melodic change: the singing of identical words to identical melodies. The occurrence in Frankish transmission is judged by Hucke, plausibly, to be the result of copying the Roman melodies by rote. Hucke's remark has had little follow-up in subsequent chant scholarship. Connolly notes the phenomenon in an Old Roman introit and gradual but, neglecting Hucke's proposal, attributes it to the survival of ancient forms from a time when liturgical singing was allegedly uniform and undifferentiated into distinct melodic genres. And Jeffery in his general review of Hucke's work, does not mention it, stating instead that no one has looked at chant systematically from this perspective. 19 Clearly the view is of seminal importance, for it adds substance and detail to the general hypothesis that composers in oral tradition, unable to deliberate at leisure during performance, were susceptible to a variety of promptings from the melodic and verbal context that sometimes caused them to take the path of least resistance, performing what was foremost in their memory, provided it met basic liturgical and artistic standards. This is only to say that in the quest to understand how chant melodies came to be as they are, some account has to be taken of the oral condition of the early repertory and of the concessions that singers made to that condition, producing something not cheaper or compromised, but exhibiting a different mixture of the psychological and cultural promptings that influence all art.

Nevertheless, the observation forces

18

In the Frankish tradition eight of the nine melodies of our group are closely dependent on their Old Roman counterparts. (In one, the first gradual for Wednesday after the Fourth Sunday of Lent, there is a different verse text in Frankish transmission, and the melody of the response bears not the slightest resemblance to the Old Roman melody; it is anomalous also on stylistic grounds.) Altogether five graduals have a different melody in one tradition than in the other. For the relationship of the remaining Frankish melodies to the corresponding Old Roman ones, the first two melodies in Example 2 are representative. The initia of the two melodies in Frankish transmission [staves 2 and 4] are the same. The first syllable of the second segment bears an ornament in the second melody. The following melismas in the two Frankish melodies are the same, just as they are in the Old Roman ones; the two closing melismas [of the second segment] differ from each other in the one tradition as they do in they other, and the closing melisma is unique [differing from the norm in its own synchronic system] in the one tradition as in the other. The third segment in the two Old Roman melodies up to the closing melisma is almost the same; in the Frankish melodies it is exactly the same. The closing melisma of the third segment in one tradition as in the other is unique. The melodic movement of the closing segment in the second melody [super me] is similar to that of the second half of the closing segment of the first melody [-nem Melchisedech] in both traditions. The final melismas in the one tradition as in the other are the same. And yet the two traditions differ from each other in particulars throughout.

Of all of Hucke's 18

19

methodological

contributions to the

comparative study Gregor-

Thomas H. Connolly, 'Introits and Archetypes: Some Archaisms of the Old Roman Chant', of the American Musicological Society, 25 (1972), 157-74.

Jeffery, Re-envisioning

Past Musical Cultures, 91.

Journal

ian

chant, this

scholarship

is the one of most

general programmatic significance, yet chant hardly begun to realize the benefit that a persistent application yield. In a word, the method is structuralist. It assumes that the

has

of it would between the two dialects is revealed best not in the material resemblance of isolated items to their respective cognate versions, but in the resemblance of each dialect's network of relationships (Crocker's 'context') to the corresponding

affinity

network in the other dialect. Making the case for certain historical hypotheses can be difficult when based on the tenuous material similarities exhibited by isolated cognate pairs. Hucke's claims are based on the much stronger foundation of massive relational similarities observations that where chants are the same in one dialect, they are the same in the other; that where they differ in one -

dialect, they differ in the other; that where they are anomalous in one dialect, are anomalous in the other, and so forth. Though the analysis of such

they

networks demands an extensive synoptic grasp of each tradition of the kind recommended by Crocker as a basis even for a critical appreciation of particular no easier comparative method has yet been proposed that will reveal chants the full extent of the relations underlying the common history of the two traditions. -

-

As a further distinguishing characteristic of the Frankish version of the melodies vis-à-vis the Old Roman, one can observe the sharper prominence of a tonal structure manifested in the F-A-C triad; moreover, the Frankish melodies have a rhythmic stress [immediately] before their conclusion, while the Old Roman ones tend rather to taper off conjunctly. But now the third example [of Ex. 2], The Frankish arranger noticed that the melody begins too high and has therefore transposed the initium to the normal range. His arrangement then follows not the tune in general, but this particular melody in its mutated (zersungenen) form, finding itself on that account a fifth too low when the Old Roman version regains the correct tessitura. It rescues itself from this dilemma through an irregular closing melisma that permits it to end in the right range after all. Once again we see in a particular case beyond a doubt that the Old Roman version is the older and the Frankish the more recent. For the Frankish version cannot be explained except as the arrangement (Bearbeitung) of a melody mutated in just this way. In a few additional cases we find in Frankish graduals similar examples of transposition as here, but none so revealing as this one. In general terms, nevertheless, it is the case with the exception of the tune just discussed and the fifth-mode verse melody to be discussed presently that melisma Av in the Frankish version corresponds to melisma A in the Old Roman, that melisma Bv in the Frankish corresponds to melisma B in the Old Roman, and so forth, that unique melismas in the Frankish version stand in the place of unique melismas in the Old Roman, and that in addition we encounter smaller differences. -

-

This is an example of the thought experiment, one of Hucke's most enduring contributions to chant research. The two versions are obviously cognate in some way. Moreover, the affinity between them is not of the kind in which each realizes its dialect's version of the same general melody-type; rather, it seems to reflect an attempt to transmit a unique melody, a fact demonstrated in the way the two chants track each other closely at the interval of the fifth. The more likely explanation of their eventual divergence is that a breach of tradition

has occurred in one of the branches, while the other accurately transmits the form of an already distorted prototype. That both branches should have independently broken with tradition, leaving no trace of the original shape, is of far lower probability. Assume now that the Frankish version is the accurate one, transmitting an already deteriorated prototype that began in the normal tessitura for melodies of its type, descended to an unusually low range in the interior, then lurched back to the correct range at the end. That hypothesis would leave us without an explanation for the anomalously high beginning of the Roman version. On

simply

the other hand, if the Old Roman version is taken as the reflects the condition of the prototype, one that began too version

deliberate and consistent attempt to rectify the perceived anomaly by transposing the melody to the normal starting range and remaining in a constant transposition until an impasse is reached near the end, where the imperatives of a conventional cadence require a return to a more familiar tessitura. To be sure, this argument entails the assumption of serial revision, lacking in advance planning. But the hypothesis that the discrepancies are the result of deliberation in the Old Roman branch, and that the intent of can

be understood

that accurately high, the Frankish

one

as a

its

high beginning is to compensate in advance for its low conclusion, is inconsistent with the style of revision otherwise encountered there. Had the Old Roman cantors possessed such subtlety or interest, their work surely would have reached more deeply into the substance of the melody. In the case of smaller discrepancies between the two traditions there arises each time the question whether they can be attributed to standardization [convergent change] by the [Frankish] arrangers or to melodic mutation [divergent change] in the Old Roman tradition after it separated from the Frankish. In this connection we may want to consider the second segments of the beginnings of the gradual-verses displayed in Example 3 In the first two Old Roman melodies this segment [indicium meum and et baculus tuus] does not make a very well-shaped, sharply etched impression, in contrast to the two Frankish melodies beneath them, in which the segment both times is note-for-note the same. In the next lower Frankish melody, Vias tuas, the beginning of the second segment is too high in comparison to the two preceding melodies, but does return to their range, even imitating the melodic movement of those two melodies for a moment on the word mihi (cf. baculus tuus in staff 4). Compare now the Old Roman transmission of the last melody: the concerned segment [notas fac mihi] exhibits a very well-formed and secure shape, to which the Frankish transmission of the same verse stands in close relation, but to which the Frankish transmission of the other two verses does not. And this same melodic shape returns twice more in Old Roman chant in two further gradual verses. Had standardization lain at the foundation of the Frankish version, then this standardization presumably would not have stopped half way to its goal, but would have included also the entirely irregular beginning of the second verse, Virga tua. But even in the unusual shape of this segment, the Frankish version follows the Old Roman tradition with customary fidelity. The similarity of Vias tuas in Old Roman transmission with two further melodic segments [in that tradition] has no counterpart in Frankish chant. That permits us to conclude that we have here an example of later oral transformation (Umsingen) in the Old .

Ex. 3

Roman tradition. In the

course

of this process of

chants, among them Vias tuas, received

orally transforming

a common new

the

melody,

three

version.

thought experiment. Hucke observes that the melodic judicium meum and et baculus tuus seem to have acquired a vivid profile and to have merged into a single melody, possibly over-generalizing a detail in the prototype that is now lost or vestigial

Example 3 presents segments

another

in staves 3 and 4 over the words

This prompts him to consider whether overgeneralization, and merger were not more generally at work in the formation of the Frankish dialect. Seeking additional evidence for that

in Old Roman transmission.

standardization,

he considers the third pair of examples in staves 5 and 6 over the words notas fac mihi and observes that they have a rather vivid profile in both traditions. If the processes invoked to explain the similarity of judicium meum and et baculus tuus in Frankish transmission were of general validity, we should expect them to have affected melodies like the one over notas fac mihi, causing it to be applied by over-generalization to yet other Frankish examples. Yet this has not occurred. To the extent that over-generalization and merger have

hypothesis,

occurred, they have done so only in the Old Roman branch, to examples not illustrated here. The Franks, on the other hand, give every evidence of adhering scrupulously to the prototype. Even the peculiar beginning of Virga tua is not ironed out, but preserved in all its oddity. If the Franks had set out to standardize that verse by assimilating et baculus tuus to judicium meum, surely they would have so

applied

the process also to the verse's odd

beginning.

suggests that the initial hypothesis is false, and that the

Their failure to do of judicium

similarity

and et baculus tuus is evidence not of merger, but of a similarity already prototype. That supports Hucke's general view that the Franks assimilated Roman chant as literally as they could, even preserving its irregularities, and that uniformity in the Old Roman branch, when not corroborated by comparable uniformity in the Frankish, is probable evidence of later convergent evolution

meum

in the

in Rome.

Example 4 shows the two graduate for the and St Sylvester [Ecce sacerdos]; above and

feasts of St John the Evangelist [Exiit sermo] below are the Old Roman melodies, in the middle the Frankish. In Frankish transmission both graduate are note-for-note the same, with the exception of inserted repercussions and a double distropha in place of a tristropha. The similarity goes so far that the melismas are split or combined, and that a tristropha is even divided into a distropha at the end of one melisma and a single note at the beginning of the next (and vice versa). The Old Roman transmission on the contrary exhibits small variations at the end of the first segment and in the first melisma of the second. 20 Then the melisma on (discipu-) lus il(-le) is not literally contracted for su (-is), but varied. The third segments [in Old Roman transmission] differ from each other utterly. I should mention one other gradual that follows the melody of Ecce sacerdos in Old Roman transmission; it is the one for Holy Thursday [Christus foetus est]. In Frankish transmission it follows the two reproduced here [on die middle staff of Ex. 4]. The example of the two graduals is characteristic of the relationship of the two traditions. Among the many possibilities for fitting a text to a given melody, the one exercised in the Frankish transmission makes the most use of art. The astonishing fixity of the Frankish tradition reveals itself in the faithful transmission of this artful technique of adaptation. 21 The Old Roman tradition, on the contrary regardless whether the two melodies originally 22 corresponded exactly in it or not is [now] less firm, less fixed, less exact. -

-

20

21

22

evidently misread the melisma at the beginning of the second segment over (di-) sci (-pulus). In fact it is identical to the one over (di-) e (-bus) in staff 3. The error has been corrected in the example. Compare, for example, the versions notated in staffless notation in St Gall, MS 339, in Paléographie musicale (Solesmes, 1889). In Rome, as a comparison of the manuscripts Vat. lat. 5319 and Archivio di San Pietro F 22 shows, the formation of variants (albeit unessential ones) began to occur as late as the twelfth

Hucke

and thirteenth centuries.

Ex. 4

23 Example 5 reproduces two gradual verses of the fifth mode in both traditions. The author of this contribution has already dealt with the gradual verses of the fifth mode in another place. 24 The relationship of the two transmissions of these melodies is especially interesting. We find in some of them differences which, at least in the domain of gradual responses and their verses, are unusual. The melodies reproduced in the example show such differences. The question now arises, where did they originate, and why did they come about? The attempt to answer these questions should be based on what we have already determined about the way in which the Frankish tradition was edited and about the fixity of the two transmissions after their separation: that the redaction of the Frankish tradition in general followed the form and structure of the Old Roman very faithfully; that the Frankish transmission is very fixed and immutable; and that the Old Roman transmission is less fixed and on a more limited scale permits us to detect the growth of variations and signs of oral transformation (Umsingungserscheinungen). In Old Roman transmission all gradual verses of the fifth mode employ the same closing melisma. In Frankish transmission several typical melismas appear side by side. The closing melismas must have been standardized subsequently in Old Roman transmission, for it is improbable that a uniform closing melisma would have been differentiated into several in the Frankish redaction, or in the course of Frankish transmission. In the case of verbal variants, inserted passages were redted on a reciting tone in Frankish transmission, unlike Old Roman. The arranger, in other words, supplied what was needed not in the style of the melodies, but by making use of the simplest prindple available to him. Consequently, inserted passages in Frankish transmission have the appearance of stylistic foreign bodies. In view of this observation, a stylistically correct further development of the gradual verse of the fifth mode in Frankish transmission is hardly conceivable. Accordingly, one is inclined to see the remaining differences between the two traditions as the further development of the Roman tradition after the splitting off of the Frankish rather than as the further develop-

23'Three' (erroneously)

24

Helmut Hucke, 5-8.

in the

original.

'Improvisation im gregorianischen Gesang',

Kirchenmusikalisches Jahrbuch,

38

(1954),

Ex. 5 ment of the Frankish or as the result of its deliberate redaction.

So, for example, when a group of initial melismas in Old Roman transmission matches its Frankish counterpart less exactly than others do, or especially when some melodies are differently divided, as the examples in Example 5 show, such differences can be attributed to the effectiveness of stylization principles still at work [in Rome] after the splitting apart of the two traditions.

By 'a group of initial melismas in Old Roman transmission [that] matches its Frankish counterpart less exactly than others do', Hucke evidently refers to certain melodies whose affinities with each other within their own tradition are stronger than the one-to-one affinities that they exhibit with their cognate versions in the Frankish tradition. With this observation he anticipated a whole line of enquiry, referred to as synchronic, which recognizes relationships within traditions as constituting systems whose imperatives often overshadow genealogy as an explanation of melodic form. Like Saussure, who first proposed the distinction between the diachronic and synchronic approaches to the study of language, Hucke recognized that diachronic studies of forms, tracing their genealogies backwards to assumed points of origin, provide a chronicle of change without an explanation. 25 For cantors were not musicologists and could not have been motivated by historical considerations. Their norms of style and form were 25

Fredric Jameson, The Prison-House Formalism (Princeton, 1972), 5-6.

of Language:

A Critical Account

of

Structuralism and Russian

derived from the collective attributes of their current repertory. Accordingly, scholars who wish to understand the artistic and psychological determinants of or who merely want to explain anomalies in the genealogical musical form record have to consider the lateral relationships of chants to others in their current repertory, and not only their affinities with single cognate versions in other dialects. -

-

In regard to the observations made so far, one ought to consider five graduals that have a different melody in one tradition than in the other. The gradual Ego dixi Domine belongs in Frankish transmission to the melodies of the first mode but closes in the fifth mode. The verse begins in the fifth and closes in the second. It can hardly be assumed that such an irregular melody as this would have been constructed in Frankish transmission. The only possible explanation is that it is the copy of a lapsed (zersungenen) Roman melody that in Old Roman transmission was subsequently replaced by the [standard] tune of the second mode. Also in the case of the graduals Liberasti nos Domine and Benedicam Dominum, which belong to the seventh mode in Frankish transmission, one that the use of the tune of the second mode in Old Roman transmission is of more recent date. In the opposite way, the gradual Domine refugium has a melody of the seventh mode in Old Roman transmission and the [standard] tune of the second mode in Frankish transmission. Here the situation is different to the extent that the Frankish redaction or transmission did not have to draw on its own creative resources to make use of the tune of the second mode in place of another one. This psalmodic can assume

Ex. 6

tune was available for use whenever needed and could be fitted to any text without a

second thought. (In fact one Frankish source does transmit the gradual Domine refugium 26 with a melody of the seventh mode. ) Here evidently one melody has been replaced another in Frankish tradition. A special case finally is represented by the already by dted gradual Venite filii, which in Old Roman transmission follows a tune of the third mode and whose melody in Frankish transmission stands all by itself stylistically among graduals. Here the [Frankish] arranger appears to have preferred a melody of his own tradition to the Roman alternative. As a last melodic example the first verse of the tract Vinea facta est in both traditions is reproduced in Example 6 The relationship of these two melodic versions to each other differs from everything that we have observed so far. All four tracts for the prophesies of the liturgy of Holy Saturday behave in this way. The one tradition follows the other note for note with few variants. What we have here is an example of the later appropriation of melodies out of the Frankish tradition into the Old Roman. With that, let our series of examples come to a close. In summary, the comparison mainly of graduals allows us to say the following concerning the relationship of the Frankish and Old Roman traditions. The Frankish melodies are in general consistent arrangements of the Old Roman, whereby the structure of the melodies is preserved, but particular melodic shapes are extensively transformed. The melodic style of the Frankish transmission has more momentum, is more goal-directed, more space-spanning, more emphatic, the Old Roman more small-membered, finer, more playful. In Frankish transmission the major triad F-A-C is more prominent as a tonal structure; and the rather unstressed tapering off of the Old Roman melodies and their segments stands in contrast to the corresponding stressed, deliberate cadences of the Frankish. The Old Roman melodies appear decorative, the Frankish sculpted. Few pieces have a different melody in the one tradition than in the other. In the majority of cases [where such differences occur] it is probably a different melody substituted later in Old Roman transmission, but that has happened in Frankish transmission too. As far as we can say after considering the graduals, no new melodies were created in either tradition after their separation. 27 In one gradual, finally, the Frankish tradition appears to have preserved a melody of its own. While the melodies in Frankish transmission from the beginning on must have been preserved with a truly wonderful fidelity, the Old Roman tradition exhibits the traces .

26 27

Paris, Bibliothèque nationale, MS lat. 10511. It goes without saying that much new creation ocurred outside the central the Mass, but that is not the subject of this article.

canon

of chants for

intervention and the growth of variations after the separation of the It also adopted individual Frankish melodies. Nevertheless, the Old Roman tradition was already in an advanced stage of fixity when the redaction of the 28 Frankish tradition took place. Both the manner in which this redaction was undertaken and the astonishing fidelity with which both traditions matched each other after they had lived apart for a long time makes written transmission at the time when they of

standardizing

two traditions.

separated

a

necessity.

sweeping claim about the necessity of musical notation must be regarded conjecture unsupported by argument or evidence, a mere opening for future exploration. When Hucke did revisit the issue, he formed the opinion that musical notation in the Frankish tradition did indeed begin to emerge in the ninth century, but only for peripheral uses, not for notating the repertory of cantor and schola. In his view, notation was applied to that purpose only in the late ninth and early tenth century, that is, roughly at the time of the oldest surviving Frankish cantatoria and gradualia. In reconsidering the possibility of musical notation in the Old Roman tradition before the late eleventh century, Hucke seems to have repudiated the view hastily expressed in this article. Instead he came to believe that the use of musical notation in the Old Roman tradition was fairly recent, since the earliest of the tradition's known manuscripts are copied in central Italian notation of the eleventh century, apparently in imitation of the current books of the international This

as a

29

(Frankish) tradition.

He argues that the scribes of the Old Roman books would

style of script if they had had one of their own back to the ninth or tenth century. 30 Given these revisions of his views, it is appropriate to reread the last sentence of Hucke's text above and enquire into its status as a reflection of his mature thought. The manner in which the Frankish redaction was undertaken does indeed seem to have required a regime of literalness to a degree. Whether that could have been enforced without musical notation remains a matter of vigorous scholarly debate. Hucke's more recent view was that musical notation was not an absolute necessity to satisfy Frankish objectives in the ninth century objectives that he finds obliquely revealed in Aurelian of Réôme's discussions of the choral repertory, written several decades before the earliest fully notated chant books. And we can safely conclude that the 'fidelity with which both traditions matched each other after they had lived apart for a long time' is not especially 'astonishing', nor did Hucke eventually consider it so. It seems rather

not have borrowed this mature

dating

-

28

29

30

On tradition in the Roman schola cantorum, see the Helmut Hucke, 'Die Tradition des gregorianischen Gesanges in der römischen Schola Cantorum', in Zweiter Intemationaler Kongress für katholische

Kirchenmusik, Wien, 1954 (Vienna, 1955), 120-3. Helmut Hucke, 'Die Anfänge der abendländischen Notenschrift', in Festschrift Rudolf Elvers

(Tutzing, 1985),

271-88.

von mündlicher zu schriftlicher Musiküberlieferung im Mittelalter', Proceedings of the Twelfth Congress of the International Musicological Society, Berkeley, 1977 (Kassel, 1981), 180-91. For empirical confirmation see John Boe, 'Chant Notation in Eleventh-Century Roman Manuscripts', in Essays on Medieval Music in Honor of David G. Hughes, ed. Graeme M. Boone (Cambridge, Mass., 1995), 43-57.

Helmut Hucke. 'Der Übergang in

in the mixed oral and written media of

to reflect the normal

fidelity possible

transmission used

the two traditions.

by

Following upon the comparison of the gradual melodies, one may be permitted to make the following claims concerning the time and place of the redaction of the Frankish transmission: it appears out of the question that it arose inside the Roman orbit. The manner and character of the redaction requires distance from the Old Roman tradition and a different stylistic environment. It awakens less the impression of an intentional revision arising out of aesthetic intent otherwise it would have intervened more deeply into the structure of the melodies and not adopted obviously corrupted ones with the same loving thoroughness as others than that of a literal translation out of a foreign -

-

tonal idiom into one's

own.

This conclusion has proved a stumbling block for some readers, leading them to dismiss the whole article as interesting but speculative and now thoroughly dated. Even those receptive to Hucke's thought have found difficulty with the description of the Frankish reception as a work of translation. It would be a mistake, however, to read this term literally, for it is only a figure of speech presenting a moderate alternative to Stäblein's claim that Gregorian chant is a deliberate recomposition of the Old Roman repertory intended to give it a more cosmopolitan style. Moreover, rejecting the article because of a quibble about its conclusions would ignore the relatively little emphasis that Hucke placed upon them. The article is not about conclusions, but about evidence and methods. It proposes to read the texts of Gregorian chant in Frankish and Old Roman transmission as an intertextual dialogue. Theirs is a kind of quiescent testimony that Hucke wants to prod into eloquence. What might sympathetic readers of this testimony conclude in the 1990s? In my opinion, the grounds for scepticism that I have been able to articulate do not touch the substance of Hucke's underlying meaning: that Gregorian chant in Frankish transmission is chant of Roman origin filtered through the ears, mouths and pens of Frankish singers and scribes, and that most discrepancies with the prototype were unintentional and inadvertent. However, not all Roman chant at the time of the Frankish reception was of well-established tradition. As recent research of James McKinnon has shown, much of it may be no older than the late seventh century and some as recent as the mid-eighth. Some of what the Franks received could have been in a nascent state, differing in the performances of different Roman informants over the multi-year period of the contact. Frankish singers may have had to choose versions, perhaps to standardize

them, perhaps even to change versions that they judged to be erroneous. And in some cases at least, they seem to have collaborated in their composition. 31 How much of Gregorian chant in Frankish transmission is the product of such Frankish initiative is the great area of uncertainty to which Hucke's method can be profitably applied with satisfying, albeit probabilistic results. 31 McKinnon, 'The Eighth-Century Frankish-Roman James Musicological Society, 45 (1992), 179-227.

Communion

Cycle', Journal of the

American

Since the melodies of Offices introduced in the first half of the eighth century have the same relationship to each other in both traditions as older ones do, one can in any case not place the redaction before the introduction of those Offices. Accordingly, the results of our melodic investigation point in the same direction as the results of the bibliographical testimony on the investigations of Michel Huglo and the inspection of the literary 32 introduction of Gregorian chant in the kingdom of the Franks.

With this translation I hope to have rehabilitated one of the twentieth century's most powerful yet under-utilized approaches to the study of Gregorian chant, and with the commentary to have elucidated some of its difficult language and demonstrated its perennial relevance. In particular the attention that Hucke pays to chants as members of systems reminds us that the field of Gregorian studies possesses one of the most enviable historiographical tools in all the humanities: to compare two completely parallel yet autonomous derivative traditions. The fruits of applying that tool to Gregorian chant in Frankish and Old Roman transmission will not only be a better grasp of their respective histories but a deeper appreciation of every chant as art in context.

the

ability

University of 32

Cincinnati

See Helmut Hucke, 'Die Einführung des gregorianischen Gesanges im Frankertreich', Römische Quartalschrift für christliche Altertumskunde und Kirchengeschichte, 49 (1954), 172-87.

[15] Ways of Telling Stories

SUSAN RANKIN

The best stories begin with

something familiar, and mine is no exception:

more perfect art of chant, which now almost the whole of France loves, introduced by Pope Stephen, when he had come to Pippin, the father of Emperor Charlemagne, in France in order to take back from the Lombards that which belonged to St. Peter. [The chant was introduced] on request by the same Pippin, through his clerics, and since then its use has everywhere and always grown strong. 1

The

was

1 “Cantilenae vero perfectiorem scientiam, quam iam pene tota Francia diligit, Stephanus papa, cum ad Pippinum, patrem Caroli magni imperatoris, in Franciam pro iustitia Sancti Petri a Langobardis expetenda venisset per suos dericos petente eodem Pippino invexit, indeque usus eius longe lateque convaluit.” Walafridi Strabonis, Liber de exordiis et incrementis quarundam in observationibus ecclesiaticis rerum, ed. Aloisius Knoepfler (Munich: Stahl, 1890), ch. 26, 84. For the translation of this passage and another below, I am indebted to Ritva Jacobsson. The Pope in question is Stephen II, who had come to France in 753 to ask Pippin III for aid and for protection of the church’s interests in Italy; on their meeting see Rosamond

McKitterick, The Frankish Kingdoms under the Carolingians, 751-987 (London: Longman, 1983). Fundamental studies of the liturgical exchange between Rome and the northern countries indude 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, and Rosamond McKitterick, The Frankish Church and the Carolingian Reforms, 789-895 (London: Royal Historical Society, 1977); see also the brief survey of the relation of the political to the musical situation in the eighth century by Kenneth

DOI: 10.4324/9781315090283-18

Oral and Written Transmission in Chant

by Walafrid Strabo, monk of Reichenau and Fulda, and intimate of the Carolingian court. His version of the famous story, written before 850, is certainly the least dressed up, and probably the closest to historical truth. But it is not accuracy which initially concerns me, rather the “story” of this story. Its next telling is in the words of a Roman, and contains many insults against the Franks:

—this

as

told

2

Alpine bodies do not properly make the sweetness of the melody they have adopted resound, since with their voices they make high-sounding noises like thunder. Because the barbarian fierceness, belonging to a drinker’s throat, emits rough hard voices when it attempts to produce a soft tone with inflections and repetitions, and does so with a kind of natural roar of the voice, sounding confused, as if you were to throw carts down steps. And so, through bewildering and terrible bawling, it [the adopted melody, i.e., Roman 3 chant] rather disturbs the listeners’ minds, which it ought to please .

.

.

Gregorii Magni written between 872 and 875—John the Deacon gives the major credit for the establishment of the Roman chant north of the Alps not to Pippin III but to his son Charlemagne. His account has a new twist: following the model of Bede, John explains how the business of introducing Roman chant was accomplished. In this

version—incorporated

into his Vita

4

First, two Frankish

cantors were

trained in Rome and returned

Levy, “Toledo, Rome and the Legacy of Gaul,” Early Music History 4 (1984): 49-99, esp. 49-50. Finally, the history of the “Roman chant story” is studied in some depth in Stephen J. P. Van Dijk, “Papal Schola versus Charlemagne,” in Organicae voces: Festschrift Joseph Smits van Waesberghe, angeboten anlässlich seines 60. Geburtstages, 18. April 1961, ed. Pieter Fischer (Amsterdam: Instituut voor Middeleeuwse Muziekwetenschap, 1963), 21-30. 2 Pupil of Hrabanus Maurus at Fulda, appointed teacher in 829 of the future Charles the Bald, and in 838 Abbot of Reichenau, Walafrid was closely in touch with the Carolingian court during much of his life; his direct personal knowledge of past and present elements of Carolingian theological politics is beyond question. 3 “Alpina siquidem corpora, vocum suarum tonitruis altisone perstrepentia, susceptae modulationis dulcedinem proprie non resultant, quia bibuli gutturis barbara feritas, dum inflexionibus et repercussionibus mitem nititur edere cantilenam, naturali quodam fragore, quasi plaustra per gradus confuse sonantia rigidas voces jactat, sicque audientium animos, quos mulcere debuerat, exasperando magis ac obstrependo conturbat.” Johannes Diaconus, Vita Sancti Gregorii Magni, ed. Jacques-Paul Migne, Patrologia Latina 75 (Paris: n.p., 1862), col. 91, 4

preface explains how, on the eve of the feast of St. Gregory in 873, the Pope, knowing Longobard Vitae of Gregory, had expressed surprise at the lack of one composed in Rome, and straightway commissioned John to produce one. To this end, John had access to the papal archives. The intended four books (corresponding to the four of Gregory’s Liber pastoralis) were completed soon after Easter 875. On John the Deacon’s literary work see Max The

the Saxon and

Manitius, Geschichte der lateinischen Literatur des Mittelalters 1: Von Justinian bis zur Mitte des zehnten Jahrhunderts, Handbuch der Klassischen Altertumswissenschaft, Abt. 9, t. 2, 1 (Munich: Beck, 1911), 689-95.

Oral and Written Transmission in Chant

north and

to

teach. But their influence did

Charlemagne

had

to start

Roman cantors, who ever after.

two

The

all

not

last

beyond their

own

lifetimes

again. Then Pope Hadrian sent effective, and everyone lived happily

over

were more

version of the chant story was written very shortly after John’s, in the 880s, and by someone who reacted strongly against John’s The story forms part of Notker Balbulus’s Gesta Karoli, a life of next

probably scorn.

5 Charlemagne composed at the behest of Charles the Fat (Emperor, 881-87). With concrete physical evidence, it is possible to prove that Notker knew John’s Life of Gregory, because it is his own hand which begins every gathering of the St. Gallen house copy. Plate 1 shows p. 54 of St. Gallen,

Stiftsbibliothek 578, with John’s version of the story of the introduction of Roman chant into Gaul; Notker’s hand can be seen in the left-hand column, lines 1-14. 6 A crucial difference between Notker’s and John’s accounts is the reversal of the roles of Frankish and Roman cantors. Notker blamed the Romans

teaching, but his story had to end correctly; so he introduced the two Frankish cantors after the Romans. John had placed the two Roman cantors at Metz; Notker sent one to Metz, and one to accompany Charlemagne. In a fourth version of the story, part of which appears in the margins of the St. Gallen copy of John the Deacon’s Vita, the story takes on a more specific local content. In the upper margin of p. 54 of Cod. 578 ( plate 1 ), beside a neume composed of a for bad

7

torculus and names

the

two

two

puncta, the hand of the monastic historian Ekkehard

successful

Casus sancti Galli,

cantors as

completed

IV

Petrus and Romanus! 8 Here and in his

in the middle years of the eleventh century,

5 It was apparently during a visit to the Abbey of St. Gallen between 6 and 8 December 883 that the Emperor made the request of Notker. Charles the Fat’s predecessor as Emperor, Charles the Bald, had visited Rome late in 875, to be crowned Emperor. John the Deacon’s prominence in the Roman situation is such that his Vita is likely to have been brought to Charles’s attention. Thus, it is likely that (if not already by other means) a copy of the Vita was carried north of the Alps by the early part of 876. 6 On Notker’s hand see Susan Rankin, “Ego itaque Notker scripsi,” Revue bénédictine 101

(1991): 268-98. 7 A translation of this portion of Notker’s account can be found in Van Dijk, “Papal Schola versus Charlemagne,” 27. It was probably Notker himself who enlarged the number of Roman cantors mentioned in the story from two to twelve. Far from trying to snipe at the Romans, however, he must have had in mind that the twelve apostles went out from Jerusalem to preach the gospel, but since Rome was the center of the Christian church, and the seat of St. Peter, he transferred the idea. The biblical passage which tells how the ministry of the apostles began reads “And they were all filled with the Holy Ghost, and began to speak with other tongues, as the Spirit gave them utterance” (Acts 2.4). The idea that the Roman cantors should teach different ways of singing thus had a sound biblical model! 8 The neume appears in the main text beside the phrase “duos in Galliam cantores emisit.”

Plate 1: Stiftsbibliothek St. Gallen, Codex sang. 578, p. 54.

Reproduced by

permission.

Ekkehard

explains

and remained two

at St.

that Romanus fell ill of

travelling north, journey. Keeping one of

from Rome, he

the monks the Roman

antiphoners brought

chant. The other book was taken In

identifying

fever while

Gallen rather than continue his

taught

by Romanus

the “successful

Ekkehard has returned

a

to Metz.

travelling

9

cantors”

two

as

Romans,

the substance of John the Deacon’s story, rather than Notker’s. Yet, in respect of the destiny of the cantors, Ekkehard’s story has taken a new direction, rendering it especially significant in the history to

of St. Gallen. While the outline remains

substantially

the

same

from

one

version of the story to another, each new variation casts light on the situation in which it was told—and, following the circle around, the context of each may help to explain the variations. By the mid-eleventh century, there had grown up at St. Gallen a perception that the abbey had received and maintained

an

In both

of Roman chant, allowing it

especially good transmission

rival Metz; and Ekkehard’s story came about.

implicit

and

explicit

can

be read

senses

this

as

intended

to

to

reveal how this

perception has persisted

in the

modern world. The famous Cantatorium, published by Père Lambillotte in 10 1851 as the Antiphonaire de Saint Grégoire, is now thoroughly the worse for

much has it suffered the attentions of reverend

pilgrims. But, as did the Roman-Frankish Jacques Handschin asked fifty years ago, how chant come to St. Gallen? 11 Was it direct from Rome, or via Metz, or how? wear, so

What lies behind the nuance-rich St. Gallen neumes? To what

patterns of manuscript sources, and the

musically

contain, typical? Against

of

a

background

notated

extreme

extent are

the

repertories they

interest in and contact

9 s Petrum et Romanum. Sed St. Gallen 578, p. 54, upper margin: “Sub audis. Romanum febre infirmum. nos Sancti Gallenses quidem retinuimus. Qui nos cantilenas

Karolo iubente edocuit. et antiphonarium e suo exemplatum. In cantario. sicut Rome est. Iuxta apostolorum aram locavit.” These words read as notes made in relation to (and possibly the preparation of) the longer account in Ekkehard’s Casus: Ekkehardi IV, Casus Sancti Galli, ed. Hans F. Haefele, in Ausgewählte Quellen zur deutschen Geschichte des Mittelalters 10 (Darmstadt: Wissenschaftliche Buchgesellschaft, 1980), ch. 47, 106-08. 10 P. Louis Lambillotte, Antiphonaire de Saint Grégoire, facsimilé du manuscrit de Saint-Gall (VIIIe siècle) (Paris: Poussielgue-Rusand, 1851); a superior facsimile was subsequently published as Cantatorium No 359 de la Bibliothèque de Saint-Gall (IXe siècle), Paléographie musicale, 2nd ser. 2 (Tournai: Société de Saint-Jean L’Évangéliste, Desclée, 1924). 11 Jacques Handschin, “St. Gallen in der mittelalterlichen Musikgeschichte,” Schweizerische Musikzeitung 85 (1945): 244-48; see also his “Geschichte der Musik in der Schweiz bis zur Wende des Mittelalters,” in Schweizer Musikbuch 1, ed. Willi Schuh (Zurich: Atlantis, 1939), 11-53. The most substantial modern critique of St. Gallen’s importance is by Rombaut van Doren, Étude sur l'influence musicale de l'abbaye de Saint-Gall, VIII. au XI. siècle (Louvain: Librairie universitaire, 1925).

with St. Gallen

Abbey’s place

manuscripts,

as a

musical

one

center

might imagine

that the story of the thoroughly told; and

had been well and

yet, nowhere in modern literature can one disentangle the specific situation of St. Gallen from the general picture of early medieval music, so often has

secondary literature tended to take St. Gallen as its point of departure. One need only cite the semiological neume studies. It is not that I want to move the spotlight somewhere else, but that I want to understand what is special about the musical traditions of St. Gallen, and to be able to distinguish these from other institutional patterns in the Carolingian and Ottonian periods.

the

"Then

we, too,

who until

now

drank from the troubled

waters

of the

go back to the clarity of the source ...":thus Charlemagne. As point of a new evaluation of the Abbey of St. Gallen as a musical

stream, must

the

starting

we need to go back to the sources. It is entirely characteristic of the late twentieth-century situation in St. Gallen that one should go looking

center,

for

something, and immediately find hard rock—as in the case of Notker’s copying of John the Deacon’s text. While its stock of classical texts was much diminished by the raids of delighted humanists visiting the Council of Constance in the early fifteenth century, the Abbey’s copies of patristic texts and liturgical books are uniquely well-preserved. Through study of the Abbey as a center in itself, and the minutiae of its musical sources—set within the larger context of contemporary St. Gallen books—we have a chance to build up a new framework of knowledge about how the monks of St. Gallen perceived and treated the Roman chant, and what significance they attributed to written sources. In the remaining part of this study I will tell something of my own changing picture of the musical history of the Abbey of St. Gallen, based on the manuscripts in the Stiftsbibliothek. There are simply more early St. Gallen sources with musical notation than many people realize, a total of at least six dating from the late ninth and first half of the tenth century.

St. Gallen,

Stiftsbibliothek Codex 342

The most substantial

early

source

copied

at

St. Gallen

notation—besides the famous Cantatorium—is such, it

chants,

provides as

well

as

containing

musical

Gradual, Cod. 342. As melodies for the Introit, Offertory, and Communion a

the Graduals and Alleluias found in the Cantatorium, and

is thus of considerable interest.

The Gradual

belongs within a composite manuscript, whose other main

parts are a Kalendar and Sacramentary. The Kalendar (pp. 2-20) dates from the first part of the eleventh century (before 1034), 12 and is immediately followed by gatherings containing prayers, chants, and other liturgical material added by various late tenth- and eleventh-century hands (pp. 21-108). The Gradual (pp. 109-272) and Sacramentary (pp. 277-843) were dated by both Scherrer and the art-historian Adolf Merton to the tenth century; 13 Merton specifically associated the Gradual and Sacramentary with each

other, and dated the latter led 15

14

early as 920. Misreading of Merton’s comments the indexer of his study to place the Gradual in the eleventh and several subsequent publications, notably Bruckner’s as

century; Schreibschulen der Diözese Konstanz (1938) and Le Graduel Romain (1957), have affirmed the later date. 16 12

P. Emmanuel Munding, Die Kalendarien von St. Gallen (Beuron, Hohenzollern: Beuroner Kunstverlag, 1948); Munding dates this Kalendar between 1031 and 1034 on the grounds that it includes an entry by the first hand for Aribo of Mainz (d. 6.IV.1031), but has the Remaclus feast—observed in St. Gallen after 1034—added by a later hand. In fact, the Aribo entry is not part of the first layer of copying of the Kalendar. Using these kind of criteria (rather than palaeographical guesses), it is hard to bring the date limits for copying this Kalendar any closer than 993 (the Kalendar includes an original entry for Udalrici, canonized 993) and 1034. 13 Gustav Scherrer, Verzeichniss der Handschriften der Stiftsbibliothek von St. Gallen (Halle, 1875), 120; Adolf Merton, Die Buchmalerei in St. Gallen vom neunten bis zum elften Jahrhundert (Leipzig: Hiersemann, 1923), 53. 14 Merton’s commentary is mainly concerned with the decoration of initials in the Sacramentary, but those parts of his text which refer to the Gradual are without ambiguity as to his assessment of the date; “Eine ganz verspätete Arbeit karolingischen Charakters ist der Codex 342 der St. Galler Stiftsbibliothek. Nach einem jüngeren Teil (Seite 1 bis 108), der unter anderem einen Kalendar des beginnenden 11. Jahrhunderts enthält beginnt der Da die Schrift relativ späten ursprünglich erste Hauptteil, ein Graduale (Seite 109 bis 272). Charakter hat, gehört der Codex sicher schon tief ins 10. Jahrhundert. Die enge .

.

.

.

.

.

.

.

.

Verwandtschaft mit den Werken von der Jahrhundertwende ist wohl um das Jahr 920 noch denkbar” (Die Buchmalerei in St. Gallen, 53-54). 15 Merton, Die Buchmalerei in St. Gallen, 105: “Codex 342, Graduale, saec. XI, Mitte, und Sacramentar, saec. X in St. Gallen.” 16 Albert Bruckner, Schreibschulen der Diözese Konstanz: St. Gallen, Scriptoria Medii Aevi Helvetica: Denkmäler Schweizerischer Schreibkunst des Mittelalters 2-3 (Geneva, 1936 and 1938), 2:98; Le Graduel Romain: Édition critique, 2: Les sources (Solesmes: Abbaye Saint-Pierre, 1957), 131 (GAL 7). The comments made in the latter appear to draw on the work of Munding, but he, albeit without exact dates for the Gradual and Sacramentary, had remarked on the fact that letter forms used in the Sacramentary were conspicuously older than those of the Kalendar, continuing “Somit ist das Kalendar etwas jünger als das Graduale und Sakramentar” (Die Kalendarien von St. Gallen, 13). Two other studies are concerned with more than details of drawings in the Codex: Derek H. Turner, “Sacramentaries of Saint Gall in the Tenth and Eleventh Centuries,” Revue bénédictine 51 (1971): 186-215 (where the Gradual is linked with the Kalendar and dated in the eleventh century), and Die Abtei St. Gallen: Ausgewählte Aufsätze in überarbeiteter Fassung von Johannes Duft, eds. Johannes Duft, Peter Ochsenbein, and Ernst Ziegler (Sigmaringen: Jan Thorbecke, 1990), 254 (where the whole book is dated to the late tenth and early eleventh century).

It

probably

was

for other

Paléographie musicale chose variously,

but

mainly

to

to the

however, that the editors of publish the later Gradual, Cod. 339 (dated reasons,

early

eleventh

century)

as an

example

of

an

missarum notated in the

Antiphonale splendidly detailed St. Gallen neumes, rather than this. Like the Cantatorium, the first layer of Cod. 342 has no liturgy (Gall, Otmar, Wiborada), nor indeed for many others eleventh-century liturgical books of St. Gallen. Further, Cod. 342 contains extensive marginalia dating from the eleventh and late twelfth centuries, rendering it much less clear to read and work with than the later manuscripts. And third, the scholars of Solesmes doubted its St. Gallen provenance, stating in Le Graduel Romain, “La série des versets alléluiatiques diffère de celle des autres manuscrits de Saint-Gall.” will return to these points in the detailed discussion below. for local saints

included in the

17

I

Date of copying of the Gradual

My assessment of the date of copying of the Gradual is based on two categories of palaeographic evidence—characteristics of the text hand and of the decorative initials—as well the

Sacramentary.

as on

The Gradual

the association between the Gradual and

was

copied by one main text hand, with

responsibility

for the music notation shared out among several different scribes. This main hand is not as gracious and well-formed as that of the Cantatorium (which, dated on palaeographical criteria alone, belongs to the first quarter of the tenth century). Nor was the first scribe of the 342 Gradual are

by

as

well-educated

as most

liturgical scribes

of St. Gallen; here there

spelling mistakes and words omitted, frequently occasioning corrections musical scribes and later, eleventh-century emenders. The hand is typical of St. Gallen work of the Ottonian

between the letters and

period, showing less differentiation

vertical aspect than those of the late ninth different sense, those of the eleventh century), mixing a more

century (or, in a round and straight d’s, and finishing descending minims (on n, m, t) with a short, but pronounced horizontal stroke. It has none of the wedged and

squashed qualities of the eleventh-century hands (especially those of the liturgical codices, as, for example, Hanker's). Characteristics which place 18

17

Le Graduel Romain, 131. 18Fine examples of such hands illustrate just what had changed in St. Gallen calligraphy over 100 years: see, for example, the reproduction of a page from Wolfenbüttel, Helmst. 1008 in Hartmut Hoffmann, Buchkunst und Königtum im ottonischen und frühsalischen Reich (Stuttgart: Hiersemann, 1986), vol 2, pl. 215; this was one of the several liturgical books copied for Bishop Sigebert of Minden between 1022 and 1036. Its script, which is more regular and well-formed than that of any of the liturgical books remaining in the Stiftsbibliothek, and may thus stand as a prime example for the period, has a much more square and “horizontal” appearance than the tenth-century scripts with which it could so easily be confused.

this

script

earlier rather than later include the

use

of three forms of

a:

the

the open a, U (which appears in the Cantatorium also, but which harks back to St. Gallen scripts of the early ninth century), and a frequently misformed a, f (almost like an insular standard “St. Gallen” form A

,

19

majuscule). The hand of the Sacramentary, while easily confused with the heavier high grade St. Gallen script of the second half of the tenth century, still retains the variation of stroke directions and rounder St. Gallen work,

through

a

showing

general absence

quality of earlier

only through elegance, but high grade of ligatures, in particular the ubiquitous rt ligature. its

not

20

All that stands between this hand and that of the Folchart Psalter

(copied in the last quarter of the ninth century), for example, is the effort (not always successful) to stop the descending r and s strokes from going below the horizontal base line, and the squareness of the little finishing strokes on descending minims. 21

As with the discussion of text hands, any attempt to date decorative immediately encounters an extreme stability of practice. Some general

work

trends

are

nevertheless discernible. The most

in the Gradual

are

third Christmas

those for the

significant

beginning (Ad

decorated initials

levavi, p. 109), for the plate 2 ) and for Easter

te

(Puer natus est, p. 120, see None are more than simple, but large and decorative (Resurrexi, p. 192). inked capitals. Besides the formal similarity between these initials and securely mass

dated earlier work, they clearly do not share with the late tenth- and eleventh-century decorative work a tendency to fill up every available inner space with weaving interlace, as in the Kalendar of Cod. 342, which must, on the grounds of internal evidence, be dated in the first third of the eleventh century. 22 This same interest in knotting and weaving patterns, which often fill the entire available space, is apparent in the large decorated

19 For reproductions of this see Merton, Die Buchmalerei in St. Gallen, pls. XLVII-XLVIII; Bruckner, Schreibschulen der Diözese Konstanz vol 2, pl. XLVI; and, most recently, Duft, Die Abtei St. Gallen, pls. 18 and 19. 20 For examples of St. Gallen script “on either side” of that of the Sacramentary, pl. 209 and 195 in Hoffmann, Buchkunst und Königtum 2, suffice. Pl. 209 (Stuttgart, Landesbibliothek HB VII 57, fol. 1) shows a very late example of the most characteristic late carolingian “Grimald” hand: since this is the hand of Notker Balbulus, it can be dated before 909, but the evident shakiness of the writer’s hand, as well as the link with younger types of script in the same manuscript, suggest that this was written close to 909. Pl. 195 (London, British Library Add. 21170, fol. 2), dating from the second half of the tenth century, shows the same types of strokes as those of the Sacramentary, but now making a more solid impression. 21 St. Gallen, Stiftsbibliothek 23; for reproductions and commentary see Peter Ochsenbein and Beat von Scarpatetti, Der Folchart-Psalter aus der Stiftsbibliothek St. Gallen (Freiburg-im-Breisgau: Herder, 1987). 22 See n. 12 above.

Plate 2: Stiftsbibliothek St. Gallen, Codex sang. 342, p. 120. Reproduced by permission.

initials in Hartker’s

Antiphoner and the Gradual, Cod. 339. The cleaner, foliage-dominated lines of the initials in Cod. 342 correspond to those of the Cantatorium and many other late ninth- and early tenth-century sources. Lastly, the decorative initials in the Sacramentary, although painted in silver and gold, show the same degree of mediation between interlace and foliage, and in many cases use the same basic outlines as those in the more

Gradual.

Although palaeographical study can never provide exact information about dates, it is certainly beyond dispute that the two main parts of Cod. 342, the Gradual and Sacramentary, were copied in the tenth century. Moreover, they are more closely tied to examples of late ninth- and early tenth-century work, than to anything of ca. 1000. 23

The

liturgy of the Gradual

As noted above, the writers of Le Graduel Romain voiced doubts about the origin of Cod. 342, on the grounds that it did not contain, except

St. Gallen as

marginal additions,

not

feasts for St. Gallen, and that the Alleluia series did sources. The first objection can be

concord with that in other St. Gallen

dismissed,

once

the date of

copying

is rectified: had Cod. 342 been

in the eleventh century, it would indeed have been odd that

saints

were

included in the first

and includes less

“pure

exactly the

same

layer.

But it is

festal series

as

a

no

much earlier

copied

St. Gallen

manuscript,

the Cantatorium. This

more or

of the local, Irish, Saxon, the later liturgical books of St.

Roman” series of feasts included

and other saints whose feasts

none

amplified marginal material—most of it added by one hand in the late tenth or early eleventh century—is collated with the later Graduals, Cod. 339 and 376, it becomes clear that the marginal incipit lists (usually including Introit, Epistle, Gradual, Gospel, Offertory, and Communion) bring Cod. 342 directly into line with the later manuscripts by supplying all that was missing from the earlier liturgical books of St. Gallen. Gallen. And, when the

The story of the Alleluia verse series leads to a similar conclusion, in that it illustrates a process of liturgical change, the development from a received Roman-Frankish

century. First,

liturgy

some

to a

St. Gallen “house version”

indication of

stability

comes

during the

from the Alleluia

tenth verses

for the week following Easter; the Cantatorium, Cod. 342, and the eleventh-

23 While hands and writing styles of the ninth century may be linked with dated extant charters of the monastery (now preserved in the Stiftsarchiv), so few examples of tenth-century charters of St. Gallen survive that it is difficult to use them as more than a general guide

to

changing script practices, let alone achieve

any degree of

certainty concerning dates.

century Graduals (of which Cod. 339 and 376 may be taken all have the same list: Pascha

as

representative)

Alleluia V. Pascha nostrum

Feria II

V'. Surrexit dominus

Feria III

V. Obtulerunt

Feria IV

V. Surrexit altissimus

Feria V

V. Cantate domino

Feria VI

V. Eduxit dominus

Sabbato

V. Laudate

Dom. I post Pascha

V. Pascha nostrum

vere

discipuli

pueri

In the series of post-Pentecost Alleluias, however, there are some differences not only between Cod. 342 and the other St. Gallen sources, but also between the Cantatorium and the of Alleluia verses are Codex 359

Diligam

te

3 In te domine

speravi

etc.

The three series

Codices 339 and 376

1 Verba mea

1 Verba

2 Deus iudex

3 Deus iudex

3

4

Diligam

te

mea

Diligam te

4 Domine deus meus

2 Domine deus

meus

5 In te domine

5 In te domine

speravi

speravi

6 Omnes gentes gentes this in (from point agreement)

4 Omnes

24

follows: Codex 342

1 Deus iudex 2

as

eleventh-century Graduals.

6 Omnes gentes

All three series

although

are of the type which follows the numerical order of psalms; the series in Cod. 359 matches a standard well-known model,

25 appear to. Crucially, the 359 series is not retained in the later books. Since the set of Alleluias copied in Cod. 342 matches that

neither of the other

two

in Cod. 339 and 376, with a small adjustment to the order of the first four pieces (Domine deus meus being moved forwards in the later series), it seems a relatively small step to see the 342 series as an expansion of that in 359, and the later series as a rearranged version of the second list. In fact, a series of Hebrew numerals in the margin of Cod. 342, p. 257—beside the

24

On the

post-Pentecost Alleluia lists see Michel Huglo, "Les listes alléluiatiques dans les Grégorien,” in Speculum Musicae Artis: Festgabe für Heinrich Husmann zum 60. Geburtstag, eds. Heinz Becker and Reinhard Gerlach (Munich: Fink, 1970), 219-27; the difference between various St. Gallen lists is here mentioned in passing (225), 25 Several possibilities are tabulated in Huglo, Les livres de chant liturgique, Typologie des sources du moyen âge occidental 52 (Turnhout: Brepols, 1988), 105.

témoins du Graduel

first four Alleluias—reads 1 (Verba mea), 3 (Deus iudex), 4 (Diligam te), 2 (Dominus deus meus), suggesting exactly the rearrangement found in 339 and 376.

Unfortunately, the annotating hand cannot easily be dated!

many interesting examples of alterations to an older liturgical practice evident in the emendations and marginalia of Cod. 342; it is a source which would richly repay detailed study. I will mention just two There

cases

are

here. The

Trinity

Gradual Benedictus

es

domine has the

verse

of 342 (as Benedicite deum caeli in both the Cantatorium and the main well as in west Frankish books). The later St. Gallen Graduals follow an east text

using the verse Benedictus es in throno; and, in the ap ropriate place margin of the tenth-century Gradual (342, p. 241), the main eleventh-century emending hand has added the verse Benedictus es in throno—with the rubric Aliter (plate 3 ). On p. 133 this same marginal hand has appended a list of proper items for the feast In cathedra sancti Petri (18 January). By the time of the transfer of Roman liturgy north in the eighth century, this old Roman feast had disappeared Frankish tradition in

in the

from Roman use—hence its absence from Cod. 359 and from the main corpus of Cod. 342. But it had remained in the Gallican liturgy, and its inclusion by the marginal hand indicates a process of re-integration of Gallican material into The musical notation With this

new

a

liturgy entirely based

on a

Roman model.

of Cod. 342 of Cod. 342, the

seeming

evidence of origin and the

liturgical

understanding of the complexities

contradiction between

palaeographical

content disappears. A context for the production of the book has thus been provided. This is especially significant since the Gradual reveals a practice of music copying which has not, to my knowledge, been observed in any other music book of the early Middle Ages. The musical notation of the first layer was accomplished by five different scribes, each working through whole gatherings, as follows: Notator

Gathering

A

pp. 109-26 pp. 127-46

I II

B

pp. 147-66 pp. 167-82

III

C

IV

B

pp. 183-202

V

D

pp. 203-22, 223-42

VI + VII

B

pp. 243-62 pp. 263-72

VIII IX

E All by later scribes

Plate 3: Stiftsbibliothek St. Gallen, Codex sang. 342, p. 241. Reproduced by permission.

The handover from break from Even more

one

music scribe

to

another takes

place exactly at

the

26 another, often right in the middle of a chant.

gathering striking than the way to

one

in which

gatherings were simply given out

to different scribes, however, is the realization that, at the time when this Gradual was copied, the monastery at St. Gallen could boast a minimum of

five music scribes, all competent to notate the complex chants of the mass. The similarities and differences between the notating hands—which are

really quite scribes

distinct—raise

questions

about the

process. All five found in the Cantatorium;

copying

able to use the whole range of neumes while differentiation between the hands depends

are

on

observation of

signs principally through their ways of writing the oriscus, quilisma, and pes quassus, as well as the axes of script that each can be distinguished. Each has idiosyncrasies: B, for example, consistently writes the used, it is

all the

trigon upside

down

∵, rather than the

more

usual

marked differences in the work of these five scribes of

and litterae

are to

∴.

But the most

do with the

Where scribe C

use or

letters

consistently episemas significativae. and abundandy (even some of the rarest, such as f), scribes A, B, and D use none at all (and E a few). Only scribes C and D use episemas regularly; they appear with some frequency in E’s work, but only very rarely in portions copied by A and B. What this variance might indicate about the exemplar(s) available to these scribes, and about scribal freedom and competence, belongs to future speculation and intensive study of the notation. At the very least, it is clear that the more musically trained the scribe, the more likely he was to notate a specific reading of each chant. not

uses

The later history of Codex 342 The Gradual, which appears to have been associated with the attached Sacramentary from the beginning, continued to be actively used over the hundred years; the process of “updating” continued into the twelfth and thirteenth centuries. But the most important phase of expansion of both Gradual and Sacramentary belongs to the end of the tenth and next two

early

26

eleventh century. An intensive

project

to

bring

the basic Roman-

Such a division of copying tasks is, of course, entirely familiar in text copying. Similar procedures have been recognized in a fourteenth-century music book compiled at the monastery of Engelberg: see Wulf Arlt, “Repertoirefragen ‘peripherer’ Mehrstimmigkeit: Das Beispiel des Codex Engelberg 314,” in Trasmissione e recezione delle forme di cultura musicale: Atti del XIV Congresso della Società Internazionale di Musicologia, Bologna 1987, 1: Round Tables, eds. Angelo Pompilio, D. Restani, Lorenzo Bianconi, and F. Alberto Gallo, 2: Costituzione e conservazione dei repertorii polifonici nei secoli XIV e XV (Torino: EDT, 1990), 97-123, and Engelberg Stiftsbibliothek Codex 314, facs., ed., and commentary by Wulf Arlt and Matthias Stauffacher (Winterthur: Amadeus, 1986), 25-27 and passim.

liturgical framework of the Gradual into line with a more developed local liturgical practice involved the addition of a host of saints’ feasts, as well as the supplementing, clarification, and alteration of the main corpus. Eventually, with a final phase of expansion of the Sacramentary, the Kalendar was added. In this as in other things, Cod. 342 stands at a point of changeover in the process of book-copying, from a standardized Roman-Frankish liturgy to a more specific version, tailored to the needs and traditions of the local community. Nevertheless, the larger part of the mass chant

Frankish

repertory was notated earlier than the more well-known Cod. 339; Cod. 342 is the earliest extant St. Gallen Gradual. And, because of its easily-observable

layers, it offers a special chance for the study of changes in liturgical practice during the Ottoman period.

St. Gallen,

and

developments

Stiftsbibliothek Codex 18, pp. 21-40

A third substantial

early source notated at St. Gallen appears not to have by musicologists until now: it is a collection of processional and antiphons hymns in one gathering—a typical libellus—now pp. 21-40 of a composite manuscript, catalogued as Codex 18 in the Stiftsbibliothek. The Stiftsbibliothek catalogue of 1875 describes the contents of pp. 21-40 as “Antiphonen mit Neumen; und einige Hymnen von Theodulph und Ven. Fortunatus.” Of course, the hymns referred to are Gloria laus (sung on Palm Sunday) and Crux fidelis for the Adoration of the Cross (sung on Good Friday). There has never been any doubt expressed about the date of copying of these pages: both Scherrer and Bruckner placed them in the tenth century without demur, Bruckner describing the hand as “very fine and beautiful.” One hand copied the text and neumes throughout the whole (excepting nine lines added on pp. 39-40 by a twelfth-century hand). Although an entirely different hand from that of Cod. 342, this shares many been noticed

27

28

29

characteristics with the main text hand of the Gradual, and with the notation of the Gradual’s first music scribe (A). Thus, the processional libellus may also be

placed in the early to middle years of the tenth century.

27 I intend shortly to publish of processional pieces. 28 28

10.

a

report

on

this

source,

with detailed analysis of the repertory

Scherrer, Verzeichniss, 6. A. Bruckner, Schreibschulen der Diözese Konstanz 2:57: “von sehr feiner, schöner Hand,

Jh.”

codicologically and in terms of content, the collection is complete one gathering are copied antiphons and hymns for the major processions of the Temporale, Sanctorale, and Common, under the following rubrics: Both

in itself: within

[For the Major Litany (25.IV), rogations, Sanctorale, and Common] Rubric Page 21-24 [Antiphonae in laetania maiore] 24- 25 Antiphonae ad pluviam postulandum 25

Ad serenitatem

25

Ad mortalitatem

25- 26

De

26

In natalitiis sanctorum

26- 27

Ad

poenitentia

27-28

reliquias ducendas Antiphonae de quacumque tribulationem

28

De

uno

confessore

[Communion antiphons] Rubric Page 28-29

Ad communicandum in natale domini

29

Item ad communicandum

[in pascha]

[Sanctorale] Rubric Page 29- 30 Antiphona quando volueris 30

De sancta Maria

30- 31

De sancto Petro

31

De sanctis

31

De sancto Gallo

apostolis

[Easter procession] Rubric Page paschali ad processionem

31- 32

In dominica sancta

32- 33

De dominica resurrectione

[Palm Sunday procession]

Page

Rubric

33

Versos Thiotolfi

presbiteri in palmis

[Maundy Thursday] Rubric Page 33-35 Antiphonae ad mandatum in cena domini [Palm Sunday procession] Rubric Page 35-37 Antiphonae in palmis

[Good Friday, Adoration of the Cross] Rubric Page 37

Ad salutandum

37-39

Versos Furtunati

[sic] presbiteri [Improperia, added in s. xii]

39-40

Although

crucem

the

phenomemon

of

a

separate book of processional chants

did not emerge until the end of the eleventh/early twelfth century, 30 the bringing together of this type of material in a discrete part of the Gradual

(usually at the end) may already be observed in ninth-century books: the Compiègne Mass and Office Antiphoner (Paris, Bibliothèque Nationale lat. 17436) and another book of the same type from Albi (Albi, Bibliothèque municipale 44), both copied in the late ninth century, are the most obvious early examples. But the actual content of such collections is variable, in 31

terms

of both feasts entered and the actual items included. In this

St. Gallen libellus is the earliest

example

of

a

specific

sense,

the

east-Frankish pattern

and repertory, its selection of pieces and occasions conforming closely to those brought together at the end of two Graduals from the Benedictine monasteries of Einsiedeln

(Einsiedeln, Stiftsbibliothek 121, copied in the century) and St. Emmeram of Regensburg Staatsbibliothek lit. 6: copied ca. 1000). And the ten-folio (Bamberg, libellus has the standard medium format of graduals copied at St. Gallen over a period of two hundred years; it could easily have been intended as a supplement to such a book. Notwithstanding this possibility, one aspect of the St. Gallen libellus suggests a lack of fixed procedure in the copying process. For the jumbled order of the pieces is somewhat surprising; all of the material relating to Easter feasts, for example, is mixed up:

second half of the tenth

32

Easter

Sunday procession Sunday procession (Gloria laus) Holy Thursday Mandatum antiphons Palm Sunday procession (antiphons) Good Friday (Ad salutandem crucem) Palm

30

Pierre-Marie Gy, “Collectaire, Rituel, Processionnal,” Revue des sciences philosophiques et théologiques 44 (1960): 441-69 (esp. 467), and Huglo, Les livres de chant liturgique, 10. The Compiègne Antiphoner is edited in Antiphonale missarum sextuplex, ed. René-Jean Hesbert (Rome: Herder, 1935) as C; for details of the Albi manuscript I am indebted to John Emerson, who will shortly publish an edition. 32 Le Codex 121 de la Bibliothèque d’Einsiedeln, facs. ed., Paléographie musicale 4 (Solesmes: Abbaye Saint-Pierre, 1894; repr. Bern: Herbert Lang, 1974); facs. ed., Die Handschrift BAMBERG Staatsbibliothek Lit.6, Monumenta Palaeographica Gregoriana 2 (Münsterschwarzach: Studien des Gregorianischen Chorals, 1988). 31

The libellus may thus represent some sort of experimental copy, the attempt to bring together, in a clearly laid-out form, a full collection of processional

pieces. Unlike the later Einsiedeln and Regensburg books, the practice of copying this material together was not followed in later St. Gallen books, which instead follow the older models, collecting the major litany and votive antiphons after an Alleluia fascicle, but placing the Palm Sunday, Good Friday, and Easter Sunday processional chants with the rest of the proper for those feasts. One other detail of the content of this libellus intensifies one’s

sense

that

represent internal practice at St. Gallen: this is its inclusion of the fraction antiphon Emitte spiritum (for Christmas), commonly found in it

might not

Frankish and Italian books, but, outside of Cod. 18, entirely absent from St. Gallen books. Where the Easter equivalent, Venite populi, is regularly west

copied after the Easter propers, Emitte spiritum never accompanies the Christmas propers. Against this foreign element, however, must be balanced the inclusion of an antiphon De sancto Gallo, “Suffragante domino beato Gallo.” What the

processional libellus actually represents as liturgical practice a puzzle: and yet, this in itself is extremely interesting, since once again it reveals a situation of some fluidity concerning the liturgy of the monastery, and the way it was written in books. Here, it is important to distinguish between what the monks actually performed in their liturgy,

thus remains

and the content of their books. It is far from clear that these

two

matched

before the eleventh century.

Codification of the Repertories of Tropes and Sequences

Moving

over

from the Roman-Frankish chant

repertories of prominent place in any

to

the

new

trope and sequence, St. Gallen again occupies a catalogue of early sources. For the sequence, it can boast two of the earliest extant: Paris, Bibliothèque Nationale lat. 10587, a fragment of Notker’s Liber Ymnorum, and St. Gallen, Kantonsbibliothek Vadianische Sammlung 317, fols. 13v-15v, where a group of six Notker sequences are copied, with some

musical notation. 33 The Parisian

source was

only recently recognized

33 A detailed study of both sources is made in Susan Rankin, “The Earliest Sources of Notker’s Sequences: St. Gallen, Vadiana 317, and Paris, Bibliothèque Nationale lat. 10587,”

Early Music History 10 (1991): 201-33.

34

by a German paleographer as wrongly dated in the standard literature. His arguments imply that this may be the earliest extant example of St. Gallen notation. The fragment consists of eight folios, containing the prefatory material and all

parts of four sequences (Natus ante saecula, Hanc concordi, Laus tibi christe qui humilis, Festa christi) from Notker’s Liber Ymnorum. These pages belonged to a high quality book, written on fine, or

smooth

parchment by an excellent hand. source are of significance in the present discussion. a First, although small fragment, it appears to represent a discrete of the Liber Ymnorum, thus showing the format of the model from

Three facts about this

copy which later east-Frankish sequence sources took their cue. Second, it has the special two-column ruling—wide for text, and thinner for music—used in east-Frankish sequence collections, implying that this type of presentation allowing for music notation goes right back to the beginnings of the writing down of sequences at St. Gallen. Third, its melodic readings, down to the precise details of neumation, are identical with those of the eleventh-century St. Gallen

implying that the written transmission of sequences was closed, that is, it did not allow of variation introduced from practice (apart from the addition of episemas and litterae significativae). The dedication copy of Notker’s Liber Ymnorum was prepared in 884; but Notker speaks in his preface of discussions with his teachers Iso and Marcellus (Moengal), both of whom died in 871. The use of musical notation at St. Gallen may hypothetically be traced back to this point. The other early sequence source, now in the Vadiana collection, is perhaps less interesting in itself, except as an illustration of another type of transmission in St. Gallen. On some pages of a composite codex, a hand of the early tenth century copied six of Notker’s sequences. Musical notation was added in the margins for part of one only—and then by a hand unused to copying neumes. This tiny collection includes sequences for Epiphany, Ascension, SS. Laurence, Gall, and Otmar, and for a martyr. Although musical notation can hardly be described as a major feature of this source, its fitful appearance, copied by an unpracticed hand, underlines the implications of the Paris fragment, that is, that at St. Gallen, music was recognized as an integral part of the written transmission of sequences from the beginning sources,

on.

Given what

we now know of Notker’s involvement with the copying of surprise to discover such a stable form of written transmission—later adopted in its format, contents, and details by the whole east-Frankish realm. It was one of Notker’s many interests/duties to plan and

books, it is

34

no

Hoffmann, Buchkunst und Königtum 1:390 and 2: pl. 202.

the work of other scribes, 35 and he himself may have been responsible for conceiving the small-format, double-column layout of sequence oversee

collections. Not so the tropes, however. It

these achieved

an

equal degree

was to be another hundred years before of stability of transmission. There are three

separate trope collections of the tenth century associated with St. Gallen, two actually copied at (and presumably for) the monastery itself, and a third—the earliest—with very strong connections to the monastery, on grounds of repertory, context, and provenance. The trope collection in Vienna, Nationalbibliothek 1609 constitutes, in fact, the earliest east-Frankish

example extant. It was probably copied at Freising in the very early years of the tenth century, but may stem directly from a St. Gallen exemplar. 36 It contains a concise repertory of 100 or so trope elements, and has all the appearance of a specially-compiled repertory, sufficient for the main temporale feasts liturgical year. It could be argued that its total lack of tropes for saints’

of the

feasts is the result of the tropes

were

seems more

often

gathering’s loss of at least two folios; but, since saints’ integrated into the temporale cycle in St. Gallen books, it

likely that no tropes for saints’ feasts were ever included.

All of the tropes copied here reappear in two mid-tenth-century St. Gallen tropers, Stiftsbibliothek 484 and 381. But the contrast between the earlier and later trope collections could not be more extreme. Where the Vienna collection is self-contained, stylistically limited, and copied in a

non-liturgical context, Cod. 484 and 381 may have as many as thirty sets of tropes for a specific Introit, mixing many diverse types of trope; and both Cod. 484 and 381 are, as complete books, specific to their purpose of recording repertories of new liturgical compositions. Almost the tiniest liturgical book I have

exemplars.

ever

seen,

Cod. 484 shows

many tropes as he He simply did not know

work, collecting 37

as

can, at

one

fastidious monk at

evidently from a variety of beginning what the total size

but

the

and subdivisions of his eventual collection would be, and ended up with the oddest, most irregular gatherings anyone could imagine. But he seems in Cod. 484

to

have

35

On this see

36

An

accomplished

his

object

of bringing order to

a

large

and

Rankin, “Ego itaque Notker scripsi,” passim. inventory of the trope repertory appears in Susan Rankin, “Notker und Tuotilo: Schöpferische Gestalter in einer neuen Zeit,” Schweizer Jahrbuch für Musikwissenschaft 11 (1991): 17-42; on the origin of the manuscript see also Hermann Menhardt, “Die Überlieferung des althochdeutschen 138. Psalms,” Zeitschrift für deutsches Altertum und deutsche Literatur 77 (1940): 76-84, and Natalia Daniel, Handschriften des zehnten Jahrhunderts aus der Freisinger Dombibliothek, Münchener Beiträge zur Mediävistik und Renaissance-Forschung 11 (Munich: Arbeo, 1973), 70-72. 37 Such small books are common in the later Middle Ages; this was, for example, a popular format for the book of hours. But the size is certainly unusual for the Carolingian and Ottonian periods, even for a troper.

varied trope collection, for he then proceeded to copy the whole lot out again, in a larger higher quality format, with regular gatherings—this survives as Cod. 381. And here he brought together with the trope repertory a collection of versus, a whole year’s repertory of Introit and Communion verses, a collection of sequences, and various other bits and pieces. In more

than

one

sorting

sense,

Cod. 484 and 381 make visible

a process of gathering and the results of their compilation evident

of

liturgical compositions, streamlined eleventh-century trope collections of St. Gallen, which contain very little not already present in the tenth-century books. 38 Certainly, other tenth-century tropers exhibit in their repertories (and, sometimes, in their codicological presentation also) “collecting” aspects: London, British Library 19768 from Mainz and Paris, Bibliothèque Nationale lat. 1240 from St. Martial-de-Limoges are two obvious examples. in the

In this

more

sense

the selected, concise nature of the collection in Vienna 1609 Notker, and its lack of any

sets it apart. Given its undeniable connection to

trope attributed to Tuotilo, I am inclined to see the Vienna repertory as representing a special Notker “project.” The implication of this hypothesis is

that, in the early years of the tenth century, the whole repertory of St. Gallen tropes

was

was

not

assembled in any

one

book.

Only in

the middle of the century

this exercise undertaken.

The process of exploration and integration of material into written sources is thus evident not only in the main chant sources, but in the

tropers also. The

two

encountered to

books Cod. 484 and 381 tell

eloquently

of the

by enthusiastic trope hoarder, who tried to his material; and the adventures of that most famous

arriviste, Quem

an

quaeritis,

in the St. Gallen

sources

problems bring order liturgical

illustrate the

consequences of the monks’ determination to fashion their liturgy in their own way. Of course, Quem quaeritis is not in the processional libellus in Cod. 18. It appears to have come to St. Gallen in a wave of trope transmission, and was copied as such in Cod. 484 and 381. By the time Hartker copied his Antiphoner (ca. 1000), Quem quaeritis had been put into the Easter morning procession, and there it stayed in the later books. Setting the seven ninth- and tenth-century liturgical books of St. Gallen (Paris 10587, Vadiana 317, Stiftsbibliothek 359, 342, 18, 484, and 381) against those copied at the end of the tenth and the beginning of the eleventh centuries (beginning with Hartker’s Antiphoner, through Stiftsbibliothek 339, 38 On the physical structure and repertory of these two manuscripts see Susan Rankin, “From Tuotilo to the first manuscripts: the shaping of a trope repertory at St. Gall,” in Recherches nouvelles sur les tropes liturgiques, eds. Wulf Arlt and Gunilla Björkvall (Stockholm: Almqvist & Wiksell, 1993), and St. Gallen, Stiftsbibliothek Codices 484 and 381, facs. ed. with commentary by Wulf Arlt and Susan Rankin (Winterthur: Amadeus, forthcoming).

376, and several other mass and office books), we can perceive the major liturgical task of the tenth century as the cultivation and secure recording in

script of a “home” liturgy, with all its Roman and Frankish, old and new, foreign and native, dimensions. The end result of that process was the ability and inclination to produce liturgical books in large numbers, occasionally of an astonishingly high grade. The summit of St. Gallen production of liturgical books was reached in those prepared for Bishop Sigebert of Minden between 1022 and 1036.

Musical Notation in

Non-Liturgical Sources

Mention of two other

completes the story of musical notation at St. Carolingian period. The more significant of these sources is a miscellany, Naples, Biblioteca nazionale IV.G.68, compiled in St. Gallen towards the end of the ninth century. Containing Boethius’s Consolatio, Prudentius’s Psychomachia, and part of Martianus Capella’s De nuptiis philologiae, it also has the texts of fourteen hymns and versus, nine with musical notation. These were added on empty pages by various scribes, sources

Gallen in the late

39

from the late ninth century until well into the tenth. 40 The newest Carolingian verse (Gratuletur omnis caro / nato christo domino) is here

(Prudentius’s O triplex honor). There same pieces, without notation, in a St. Gallen manuscript of ca. 800 (Leiden, Bibliotheek der Rijksuniversiteit, Vos. lat. 69). An outstanding characteristic of the Naples notations is a concern with heighting techniques. Not only are significative letters used in abundance, but the open space above the written text is well exploited as mixed with is

an

hymns by classical

authors

earlier transmission of some of the

representative of a spectrum of pitches. descriptions and discussions of this source dealing with the hymns and versus Capella,” Aevum 34 (1960): 412-14; Dieter Schaller, “Frühmittelalterliche lateinische Dichtung in einer ehemals St. Galler Handschrift,” Zeitschrift für deutsches Altertum 93 (1964/65): 272-91, and Ewald Jammers, “Rhythmen und Hymnen in einer St. Galler Handschrift des 9. Jahrhunderts,” in Festschrift Bruno Stäblein zum 70. Geburtstag, ed. Martin Ruhnke (Kassel: Bärenreiter, 1967), 134-42; this last includes reproductions 39 The main

are

Claudio Leonardi, “I codici di Marziano

of all 40

versus with

music notation.

Königtum (1:402), Hoffmann raises doubts about the origin and subsequent fate of the manuscript; undoubtedly, the whole source merits a careful palaeographical study. On the evidence currently available to me, I have to disagree with his proposal that it is not from St. Gallen. As far as can be judged from extant sources, several of the Latin versus stem from pure St. Gallen traditions (see Schaller, “Frühmittelalterliche lateinische Dichtung”); further, two late medieval entries of the note “Liber Sancti Galli” imply that the Codex had remained at the monastery throughout the medieval period. In Buchkunst und

Despite the sketchiness of these remarks, the Naples source deserved

inclusion

in this outline discussion because of the way in which it broadens picture of musical literacy at St. Gallen in the late ninth and early

the whole

tenth centuries. From

an

earlier

period, there is only one example

of musical

notation in the Stiftsbibliothek: this appears in Cod. 397, the vademecum of Grimald, Archchaplain to Louis the German and Abbot of St.

Gallen, 841-72. 41 The book contains shorter and longer entries made by as many as forty different hands, the last datable entry from 867, three years

departure from court. Musical notation for the beginning Vere quia dignum on p. 30 of this book is unlikely to have had anything

before Grimald’s of

to do with St. Gallen. But the book came to St. Gallen with

Grimald, who

died there in 872. This is sources

precisely

the

point

to which my

of St. Gallen had led back—on the

story of the earliest notated

supposition

that Notker’s

early

attempts writing down sequences resembled those of ca. 900. From this we witness an explosion of interest in musical notation, moreover, on, point at

uniquely precise nature. Right across the spectrum of texts associated with music, sung in the liturgy and in school, notation appears. On

notation of a

the basis of sources which

can

be

seen

and touched, it

can

be deduced that

the consistent and competent integration of musical notation into writing techniques at St. Gallen dates from the last quarter of the ninth century.

My story finishes with a moral, its subject the dedicatee of this book, Hughes. I first encountered him in the Stiftsbibliothek at St. Gallen—surrounded by every available Gradual—painstakingly copying version after version of a melody, checking a word here and a neume there; and this reminded me of nothing so much as one of Ekkehard IV’s anecdotes, which tells how Notker, Tuotilo, and Ratpert had obtained the permission of the abbot to spend the hour between the night offices in the library, collating texts. I believe that that kind of critical work, comparing and evaluating written texts, going back to read the sources, has never lost David

its value

or

interest.

41 For a detailed study of the manuscript see Bernhard Bischoff, “Bücher am Hofe Ludwigs des Deutschen und die Privatbibliothek des Kanzlers Grimalt,” in his Mittelalterliche Studien 3 (Stuttgart: Hiersemann, 1981), 187-212, esp. 201ff. The page with musical notation is reproduced and discussed by Handschin, “St. Gallen in der mittelalterlichen Musikgeschichte.”

[16] Interrelationships Chants Gregorian

among

An Alternative View of Creativity in

Early Chant

Theodore Karp

DURING THE PAST decades scholars have come to realize increasingly that we cannot fully appreciate the nature of surviving Western chant repertoires without reflecting on how these came into being. Conversely, we cannot effectively recreate the nature of chant performance during the centuries preceding the advent of notation without an intimate knowledge of the similarities and disparities among surviving readings for individual chants over the first half-millenium of notation. In this essay, I shall seek to pay tribute to Jan LaRue’s contributions to style analysis by showing how observations concerning interrelationships among chants of different genres and modes may be employed to shed light on musical processes affecting chant in the period both prior and posterior to the development of European musical notation. It is well known that many of the most basic stylistic distinctions in chant go hand in hand with distinctions of liturgical function. Other equally basic distinctions involve questions of modal structure. By placing the categories of mode and liturgical function on the contrasting axes of a graph, one creates a natural system of classification that will account for a large proportion of the stylistic idiosyncracies of Gregorian chant. Such a paradigm has been employed in a seminal article by Leo Treitler, “Homer and 1 Gregory: The Transmission of Epic Poetry and Plainchant,” and I believe that comparable models may reasonably be inferred from remarks by other scholars. As an

1 Leo Treitler, “Homer and Gregory: The Transmission of Epic Poetry and Plainchant,” The Musical Quarterly, LX (1974) , 347.

DOI: 10.4324/9781315090283-19

Oral and Written Transmission in Chant

addendum, Treitler carefully notes that not all liturgical categories exist in all modes; furthermore, a single liturgical category may employ several basic melodic types within a single mode. 2 A number of chant scholars seem to feel that these categories reflect the innate channels of creative activity that were formed during the gestational 3 period of Gregorian chant. In order for this view to remain viable, it is necessary to explain how it is that one may find the same or similar melodic materials in different categories of chant, whether viewed from the standpoint of mode, liturgical function, or both. One of Treitler’s major contributions is that he considers this matter in broad terms. He seeks first to link the phenomenon to the functioning of the memory, as in the work of Frederic C. Bartlett. One characteristic cited is that “a salient detail may be common to interest, and it may serve as a crossing point between them. In that way the theme originally presented may be left and another entered."4

two or more themes or streams of

An equally important underpinning for Treitler’s views is provided by the studies of Milman Parry and Albert Lord dealing with the Yugoslavian verse epic, both for its own sake and as a means towards the understanding of Homer. In his classic book The Singer of Tales, Lord observes that while “themes lead naturally from one to another to form a song which exists as a whole in the singer’s mind. [they] In a traditional poem, therefore, there have a semi-independent life of their own. is a pull in two directions: one is toward the song being sung and the other is toward the 5 previous uses of the same theme.” Lord discusses the inconsistencies that may arise from this twofold pull, some of which may remain constant for a period of many years. In his summation, he points out that the process of oral composition may “actually lead the singer to mix songs, passing from one song pattern to another at a point where the two patterns coincide. Singers recognize that this kind of thing happens, because 6 they criticize other singers for ‘mixing’ songs.” We can understand that similar inconsistencies in arise that become a matter of habit and thus eventually chant, might they might .

.

be

.

.

,

.

represented in later written documents.

The principle of “crossing” was unquestionably operative in chant, not only during the time preceding the development of notation, but for centuries thereafter. It is worth

our while to examine representative samples of this process, both for their own sakes and in order that we may become better aware of the nature of documentation establishing the existence of the process. Proceeding further, we shall realize that there are numerous instances in which this documentation is lacking. These examples

will suggest the need for

a complementary theory, and this, in turn, will reflect on our concerning the nature of early creativity in chant. We may thus arrive at a more flexible and comprehensive view of chant categorization. Among the vast array of tenth- and eleventh-century manuscripts forming the basis

ideas

2

Ibid For example, in his book Gregorian Chant (Bloomington, Ind., 1958), Willi Apel cites instances of interrelationships among chants of different genres (e.g., p. 310). This notwithstanding, he states broadly, “In Gregorian chant each single composition belongs to a liturgical class from which it receives the general characteristics of musical form and style that set it wholly apart from an item belonging to a different class” (p. 305). 4 Treitler, “Homer and Gregory,” p. 345; see also p. 362. 5 Albert Lord, The Singer of Tales (Cambridge, Mass., 1960), p. 94. 6 Ibid ., p. 120. .

3

.

.

.

Oral and Written Transmission in Chant

7

for Karl-Heinz

Schlager’s magisterial catalogue of Alleluia melodies, the Alleluia Respexit only in the early eleventh-century codex Einsiedeln, Klosterbibliothek, MS 121. (It would appear that this chant is also rare in later sources, although known in the Chartrain tradition.) It is cited by Dom Louis Brou as a member of the Dies sanctificatus family and by Schlager as a member of the Redemptionem 8 family. As indicated in Example 1 the two families have much in common. They share the same melody for the opening alleluia and jubilus and are both in second mode. Both have verses that are divisible into unequal halves, and both cadence Dominus is to be found

,

d at the end of the first half. Both verses open with gestures that rise from d and are related to the initial gesture of the alleluia melody. In the Alleluia Dies sanctificatus, the ascent is almost immediate, and there is a balancing descent that parallels the similar motion of the alleluia melody. In the Alleluia Redemptionem, the ascent is delayed by an opening recitation on d, and the return to d is postponed until the conclusion of the first half of the melody. Whereas the second half of the verse to the Alleluia Dies sanctificatus is basically a variation of the first, the second half of the Alleluia Redemptionem verse begins with a slight variation of the alleluia melody and continues with a faithful replication of the jubilus. Thus the second half begins in a context that is comparable to that of the Alleluia Dies sanctificatus and evidences a similar pattern of motion. The first half of the Alleluia Respexit Dominus follows the first half of the Alleluia Redemptionem, but the chant crosses to the material of the Alleluia Dies sanctificatus for the second. Even though the notation of the Einsiedeln source is adiastematic, the neume patterns are so clear that their basic import is unmistakable. Because other early German sources that contain the Alleluia Redemptionem are also adiastematic, I give the readings for the three chants according to the normative versions presented in the Solesmes editions. on

(I

have studied the Alleluia Dies

sanctificatus

in

a

few dozen

sources

and have

find any major variants.) The presence of crossing in the Alleluia Respexit Dominus is clear, since we are able to establish both the chronological priority of the two distinct families and the parallel features that prompted the crossing. yet

to

Phrases occurring in the Alleluia Dies sanctificatus involve crossing in other ways, as well. The formulas closing the two halves of the melody each occur in a sizable number of chants of other genres and modes. When one investigates the manuscript traditions

for these chants, other. The

one

finds several instances in which

one

1 is the

one

labeled “B” in

phrase is exchanged for the most frequently present in

phrase Example readings. However, there is a variant version of the Alleluia Paratum cor 9 meum in which this phrase is replaced by the one labeled “C” in Example 1 and the same is true for individual readings of the graduals Domine praevenisti, Timete Dominum, Posuisti Domine, and Ecce quam bonum (second verse). These variant readings are given together with pertinent normal ones in Example 2 normal

,

.

7

Karl-Heinz Schlager, Thematischer Katalog der ältesten Alleluia-Melodien aus Handschriften des 10. und 11. Jahrhunderts, Erlanger Arbeiten zur Musikwissenschaft, 2 (Munich, 1965). 8 Dom Louis Brou, “L’Alléluia gréco-latin ‘Dies sanctificatus’ de la messe du Jour de Noël,” Revue grégorienne, XXIII (1938), 175, n. 1; Schlager, Katalog, pp. 53, 81. 9 I have not had the opportunity to examine large numbers of readings for each of the fourteen members of the Paratum cor meum-Adducentur family cited in Schlager, Katalog, p. 161; additional examples of this exchange may perhaps be found among them.

Example 1.

the Alleluia the Alleluia Dies sanctificatus.

Comparison of (a) the Alleluia Redemptionem, (b)

Respexit Dominus, and (c)

Conversely, while the verse of the gradual

Universi qui te exspectant normally terminates there are at least two C, phrase readings that end with phrase B instead, and a similar exchange occurs among Beneventan and Aquitanian sources for another first-mode gradual, Os justi. These excerpts are given together with pertinent normal with

ones

in

Example 3

10

.

If we examine phrases B and C closely, we shall note that—as demonstrated in Example 4—they employ the same basic materials, though expressed in greatly differing

degrees of elaboration. The two openings are very similar. The mid-portions both descend from a through f-e-c, and both close with an ascent from d to f and a return. Here again one can establish both the chronological priority for the main tradition of each chant and the circumstances that made crossing natural. 10

The normative versions follow the readings given in the Solesmes editions. The bracketed excerpts from Timete Dominum (f-g) represent variants present in certain manuscript readings. The bracketed notes in the excerpt from Universi (i) represent a conjectural reconstruction of material apparently in the manuscript but trimmed from the facsimile reproduction in Paléographie musicale, XV. The excerpt from Concupivit rex (k) is taken from the Aquitanian gradual and not the one in modern chant books. The former has the same text for the respond as the latter, as well as a related melody. The text for the verse of the Aquitanian chant is longer than that for its counterpart in modern books, and the melodies show no obvious resemblance. I am indebted to Dr. Jeffrey Wasson for bringing this melody to my attention. Transcriptions from several manuscripts are available in his doctoral dissertation, “Gregorian Graduals of the First Mode: An Analytical Study and Critical Edition” (Northwestern University, 1987), II/1, 361—81.

notes in the two

Example 2.

Comparison of the conclusion of the Alleluia dies sanctificatus (a) and related materials: (b) Os justi (Gr. I); (c) Universi (Gr. I); (d) Alleluia Paratum (All. III; London, British Library, Harley 4951); (e) Domine praevenisti (Gr. IV; Nonantola, Abbazia); (f) Timete Dominum (Gr. I; Paris, Bibliothèque nationale, lat. 1132); (g) Timete Dominum (Gr. I; Avignon, Bibliothèque municipale, 181); (h) Posuisti Domine (Gr. I; Paris, ibid., lat. 903); (i) Ecce quam bonum (Gr. I, London, ibid., Harley 4951); (j) Alleluia Surrexit (All. I); (k) Haec dies (Gr. II); (l) Ego dixi (Gr. V).

A further example of crossing occurs in the fifth-mode gradual Ego dixi. In a small group of fifth-mode graduals, including Ego dixi, 11 the penultimate melisma of the verse is the same as the one filling a similar role in the verses of second-mode graduals. In three of the former—Adjuvabit eum, Respice Domine, and Ego dixi—the final phrase begins as does the equivalent passage in second-mode graduals. As illustrated in

Example 5

,

there is

an

ascent from f to c' followed

by a descent from c'

to a. While

11 This group includes Respice Domine, Adjuvabit eam, Timebunt gentes, Vindica Domine, and Bonum est confiteri. The latter chant is nearly identical to Ego dixi throughout the penultimate phrase.

Example 3.

First halfof the Alleluia Dies sanctificatus (a) and related materials: (b) Alleluia Paratum (All. III); (c) Alleluia Adducentur (All. III; (d) Domine praevenisti (Gr. IV); (e) Posuisti Domine (Gr. I); (f) Timete Dominum, respond (Gr. I); (g) Timete Dominum, verse (Gr. I); (h) Ecce quam bonum (Gr. I); (i) Universi (Gr. I;

Benevento, IV.34); (j) Os justi (Gr. I; Benevento, VI.34); (k)

Concupivit rex (Gr. I; Madrid, Biblioteca de la Real academia de la historia, MS Aemil. 51); (i) Tu, puer, propheta (Comm. and Resp. II); (m) Tecum principium (Gr. I); (n) Domine audivi (Tract II), V 5.

Example 4.

The

closing cadences of (a)

Ecce quam bonum and

(b) Universi.

Example 5.

The closing portions of (a) Adjuvabit eum, (b) Respice Domine, (c) Ego dixi, (d) Haec dies, (e) Ego dixi (Avignon, Bibliothèque municipale, MS 181).

Adjuvabit eum and Respice Domine conclude with a normal fifth-mode formula, Ego dixi instead follows accurately the pattern of the second-mode gradual Haec dies. The impetus for the path followed in Ego dixi is clear. Yet the choice of ending is unparalleled and the chant lacks the tonal centricity present in other fifth-mode graduals currently known. In this instance, the presence of crossing is documentable by a clear stylistic anomaly rather than by ascertainable chronology. Understandably, certain medieval musicians viewed the unusual ending of Ego dixi with disfavor and deemed it to be in need of correction. In manuscripts such as Avignon, Bibliothèque municipale, 181, the final portion of the terminal melisma is altered so as to end on f (see the bottom line of Ex. 5 ). There is, so to speak, a “recrossing." In the absence of exhaustive critical editions of individual chants, it is impossible to state at present how frequently the phenomenon of crossing can be documented. It is likely that detailed investigations of formulaic chants will add considerably to our

present fund of knowledge. For example, crossing may

found

be

on rare

occasions

Aquitanian versions of second-mode tracts. Knowledge of later polyphonic settings may also provide valuable clues. A few decades ago, in an edition of a cycle of Mass Propers from one of the Trent manuscripts, Laurence Feininger noted that the verse of the gradual Benedictus es closed with an ornamented version of a formula associated with the Christus factus es subgroup of fifth-mode graduals instead of the 12 ending normally employed. A chant source for this variant version has not yet been discovered, but one may reasonably hope that such documentation will be found in the foreseeable future. It is quite unlikely that the change described would have been introduced by the composer of the polyphony. Often we must deal with interrelated materials without either firm chronological or stylistic guides. Returning to the chants cited in Example 3 one will note that the B phrase of the Alleluia Dies sanctificatus appears in the mode 3 family of the Alleluia Paratum cor meum. Excerpts are provided for the two main members of the fourteen-member family. Phrase B appears also in the mode 4 gradual Domine praevenisti and in two later contrafacta, Benedicta et venerabilis and Benedicta es Virgo Maria. It is further to be found in four graduals of mode 1, Timete dominum, Ecce quam bonum, Posuisti Domine, and Concupivit rex (Aquitanian version). In addition, it occurs in certain versions of the mode 2 chant Tu, puer, propheta, which serves as both a communion and a responsory. Finally, a variant form is to be found among mode 2 graduals of the Justus ut palma family. In brief, the phrase occurs in no less than seven different categories of chant. What was the original context of this material? Almost all the chants cited in Example 3 belong to the layer present in the Antiphonale Missarum Sextuplex, the composite edition of the earliest surviving textual sources for the chants of the Mass Proper. 13 Few are of obviously late date. Because of the liturgical prominence of the Alleluia Dies sanctificatus, one might be tempted to posit that this phrase was first employed in that chant. However, closer examination reveals that in certain important respects this chant is not typical of its family. No other early member of this group begins the melisma on an unaccented syllable, and no other has an extensive internal recitation on f. 14 On the contrary, other chants substitute for this recitation one involving d that serves to introduce the B phrase rather than expand its central portion. Had the Alleluia Dies sanctificatus served as the founding member of its family, surely some other member would have followed its lead regarding the interrelationship between text and music. And had the alleluia family provided the original matrix for this phrase, it is likely that a wider variety of chants would have begun the melisma at the same point in the melody. Instead, most begin the melisma earlier. Even though the graduals of mode 2 are also liturgically prominent, it is unlikely that they provided the original matrix for this phrase; a number of idiosyncratic melodic traits are present in these readings. These may have arisen as an aftermath of the need to notate the chant at a level a fifth higher than normal, but there is no way in which one can be certain of this. The first-mode graduate cited do not belong to feasts

among

,

12 G. Dufay: Missa de Sanctissima Trinitate, ed. Laurence Feininger, Documenta polyphoniae liturgicae S. Ecclesiae Romanae, Ser. 1, No. 4 (Rome, 1949), prefatory page. 13 See the edition by Dom René-Jean Hesbert (Rome, 1935). 14I have examined at least one manuscript reading for each of the chants of this family listed in Schlager, Katalog, p. 78. However, I have not had the opportunity to study the later members of the family.

that

are as

important liturgically

as

those served

by the

chant families

just cited.

The

communion and responsory Tu, puer, propheta is not only isolated, but displays an unstable transmission. 15 One may note that the text itself is used also for one of the members of the large Alleluia Dies sanctificatus family. Questions of tonal framework make it difficult to posit that the phrase originated either with the Alleluia Paratum cor meum family or with the gradual Domine praevenisti. In short, reasons may be found to contradict the possible chronological priority of each of the melodies cited for this

complex. It seems to strain the evidence to consider that this phrase originated at

a

time when

categories of chant were conceived as being stylistically individual. It seems particularly difficult to offer the principle of crossing as an explanation for the communion and responsory Tu, puer, propheta. If chant had originally been created within a set of multiple boundaries delimited in part by liturgical function, how could this chant have served within two different rites? There could hardly have been any confusion in the mind of the singer as to whether he was functioning within the Mass or within the Office. Yet it is estimated that approximately one-fifth of the surviving communions served also as responsories in either the Old Roman or Gregorian repertory. the several

16

I should therefore like to propose a contrasting hypothesis, namely that there was a period, beginning before 600 and continuing somewhat beyond, during which stylistic boundaries were hazy. Variety was undoubtedly present then, both within the various liturgical genres and from genre to genre. However, formulas were not regarded as being proper to any one liturgical category or tonal order. Rather, they formed part of 17 I suspect that there was not a common vocabulary evolved for the singing of psalms. a great awareness of the individuality of the various liturgical categories at first. I suggest that this situation is reflected in the early variability of rubrics. It is thought that the earliest mention of the offertory is to be found in a passage by Augustine. 18 But Augustine has no special name for the chant, merely mentioning the singing of a psalm during the offering of gifts. The rubrics for graduals shift somewhat even during the time of our early textual and musical sources. An introit may be labeled “Antiphona” or “Officium” in early sources. Chants that are now generally called tracts were identified at times as responsories or cantica, while others bearing the rubric “Tractus” are no longer so called. The flexibility of the rubrics seems to have been regularized only at a comparatively late date. 15 There are at least two melodies for the responsory Tu, puer, propheta. The version in the Liber responsorialis, p. 358, is to be found incomplete in the Codex Hartker (Sankt Gall, Stiftsbibliothek, MSS 390-91). Another melody, not listed in the Bryden and Hughes Index of Gregorian Chant, is to be found in the Codex Albensis (Graz, Universitätsbibliothek, MS 201) and the Antiphonale Sarisburiense (Cambridge, University Library, MS MM.ii.9). 16 Helmut Hucke and Michel Huglo, “Communion,” New Grove, IV, 592. 17 When a preliminary form of this essay was presented at the Annual Meeting of the American Musicological Society, Ann Arbor, Mich., 1982, Professor Thomas Kelley kindly pointed out that certain of my results paralleled his findings with regard to Old Beneventan chant. In that repertoire certain melodic formulas recur within the context of different tonal organizations. And Professor Terence Bailey graciously pointed out a useful parallel with the Ambrosian repertoire. It would thus appear that we are dealing with a basic principle operative in several chant repertoires at an early date. 18 Giacomo Bonifacio Baroffio and Ruth Steiner, “Offertory,” New Grove, XIII, 514; Apel, Gregorian Chant, p. 41.

Similarly, it would appear that awareness of tonal order and tonal centricity came being only gradually. While many early chants were tonally centric, others were not. Again, concepts of tonal propriety seem to be a late development. They reached ultimate importance only in the eleventh and twelfth centuries, during which time many of the non-centric chants were reworked in order that they might conform with the newer aesthetic. The Cistercian chant manuscripts of course provide the best-known and most far-reaching examples of melodic reworking, but such revisions are by no means limited to sources associated with this order. In sum, it is possible to view the relative distinctiveness of the various liturgical categories and modes as the result of an ongoing process of revision and individualization, rather than as part of the earliest heritage of the creators of Western chant. I would suspect that the first peak in this into

process

was

reached with the Frankish revision of Roman chant. as this cannot rest on a single example, however extensive.

Obviously, a theory such

interrelationships that seem to be hypothesis. Example 6 is devoted to a formula employed in various

I shall therefore discuss several other sets of consonant with this

guises among different melismatic chants. As is customary with such formulas, the are freest at the beginning and strictest at the cadence. Within this context, strictness refers to the constancy of material among the multiple appearances within a given manuscript. We may note, for instance, that the characteristic cadence of Example 6 c'-g-g-f, appears as b(b ?)-g-g-f in several of the Aquitanian sources. However, our present state of knowledge does not permit a complete account of the variants existing throughout the manuscript tradition for the different formulas. (Unless otherwise stated, the quotations in this and the following examples have been taken from the normative versions provided by the Solesmes editions and have been checked against a handful of sources.) Despite the relative flexibility of the openings, the essential similarity of the different statements seems clear. Most of the citations are noted in Willi Apel’s Gregorian Chant, and I shall supply his symbols in lieu of further references. This material appears in five fifth-mode graduals under Apel’s symbols F17 and F18. It appears in twenty-four graduals of mode 2, being used in both the respond and verse. The near-identity of the first two excerpts is noteworthy, while the following three demonstrate the kind of flexibility present within the two modal families. The formula F10 that occurs in seven graduals of mode 1 is simpler than the plainest of the statements in mode 5, given immediately above it. On the other hand, the formula F1, which occurs in five graduals of mode 7, shares different features of the more elaborate statements in modes 2 and 5. Isolated occurrences are to be found in the mode 3 gradual Speciosus forma, in the mode 1 offertories Jubilate Deo universa terra and Super flumina, and in the mode 3 offertory Deus tu converters, the latter two involving only the cadential area itself. Lastly, the formula occurs in seven tracts of mode 8 and in thirteen tracts of mode 2, not all of these belonging to the layer of chant known before 900. It would appear that the phrase serves the role of an intermediate cadence, regardless of the relationship between the concluding tone and the final. The excerpts selected are from chants cited in the Antiphonale Missarum Sextuplex. In all, chants of six different modes and three different genres are represented. The materials in Example 7 comprising phrases from graduals of five different patterns

,

,

from tracts, alleluias, and offertories of mode 8, are treated with somewhat greater freedom than those in the preceding example. They share a common

modes

as

well

as

approach to

the cadence and

generally introduce this area with material that descends prefer to view some of the looser resemblances

at least twice from c'. One may

Example 6.

of related phrases closing on f: (a) Ego dixi, (b) A summo caelo, (c) Tecum principium, (d) Timebunt gentes (R ), (e) Timebunt gentes ( V ), (f) Ecce quam bonum, (g) Benedictus Dominus, (h) Jubilate Deo universa terra, (i) Benedictus Dominus, (j) Vinea facta, (k) Speciosus forma, (l) Deus tu convertens, (m) Cantemus Domino, (n) Domine audivi, (o) Super flumina Babylonis.

A group

*Transposed up a fourth.

Example 7.

of related phrases closing on g: (a) Tenuisti manum, (b) Ego autem, (c) Benedicam Dominum, (d) Miserere mei, (e) Benedicite et intende, (h) Qui Dominum, (f) Exaltabote, (g) Exsurge sedes, (i) Deus, vitam meam, (j) Sicut cervus, (k) Dilexisti, (l) Alleluia Dominus regnavit et exsultet, (m) Alleluia Confitemini, (n) Emitte Spiritum, (o) Benedictus qui venit.

A group

.

.

.

*Transposed up

a

fifth.

fortuitous, but there still remain stricter ones that cannot easily be shrugged off. For example, the relationship between the gradual Dilexisti and the Alleluia Dominus regnavit (k-l) is strikingly close. While the resemblance between the gradual Deus, vitam meam (i) and the tract Sicut cervus (j) is not particularly strict, it is of interest inasmuch as Dom Louis Brou has previously pointed to more literal melodic relationships linking tracts with other melismatic chants. He cites in particular materials shared by second-mode tracts and as

both the seventh-mode Alleluia Pascha nostrum and the second mode Alleluia Dies 19 The latter relationship is illustrated in the bottom staff of Example 3 , above. I shall explore larger relationships involving second-mode tracts shortly. In Examples 6 and 7 , the passages cited demonstrate the reuse of similar materials in different modal contexts. Other evidence may emphasize the reuse in different liturgical

sanctificatus.

contexts of materials belonging to the same mode. Example 8 presents the openings of two antiphons, a responsory, two introits, an offertory, two communions, an alleluia, and two graduals, all of mode 1. The reader will recognize immediately that the initial gesture constitutes a formulaic pattern common in mode 1 antiphons and introits and not infrequent among responsories. If one were to focus primarily on the characteristic leap of a fifth, associated with either the first or second accented syllable, a very large number of chants come into view, including the mode 6 offertory Confitebor Domino. The selection offered here has been made on the basis of similar long-range goals. Although the internal points of repose may occur on either a, g, or f, all the excerpts come to their principal conclusion on d. And the overall phrase shapes are basically equivalent, even though they obviously differ considerably in their degree of elaboration. This is dependent largely upon the extent of the text to be set. Notwithstanding this variety, many of the individual details occur in more than one genre. For example, the ascent to c'—following the leap of a fifth and its embellishment—occurs in both the introit Inclina domine (e) and the offertory Confitebor tibi Domine (f). The latter shares its cadence structure with the antiphon Canite tuba (b). The longer correspondence between the latter example and the responsory having the same text incipit (c) is undoubtedly a result of their shared text. Shared text, however brief in nature, is apparently at the root of other musical resemblances, such as those which occur at the respective openings of the first-mode gradual Custodi me and the first-mode offertory that begins with the same two words. A briefer comparison of similar nature, involving chants of mode 7 that belong to six different categories, is provided in Example 9 Here the similarities are looser still, but sufficient nevertheless to show a basic shared language. It seems particularly difficult to interpret interrelationships occurring at the openings of chants, such as appear in the last two examples, as the result of crossing. A musician recreating a chant without reference to notation may easily take an unexpected turning in the course of performing and cross to a related melodic theme. But if the early performers of chant did view the liturgical genres as having their individual melodic traditions, it seems strange that they would disregard this distinctiveness at the beginning of a piece. At that point there would be no lack of time for planning ahead and no possible confusion about either genre or tonal class. .

19 Brou, “‘Dies Sanctificatus,’” pp. 204-5. There is an excellent possibility that the Alleluia Dies sanctificatus (or the oldest member of its family) furnishes a classic example of the process of centonization, uniting melodic materials drawn from different matrices.

Example 8.

Related openings of first-mode chants.

9. Example Related

openings of seventh-mode chants.

on

One may approach this general subject in still other ways, focusing, for example, the variety of interrelationships that may cluster about a given piece. Example 10

shows that most of the first-mode introit Exaudi Domine is closely related to the communion Data est mihi, also in mode 1. The beginning, however, constitutes a common third-mode formula appearing in introits, communions, graduals, offertories, and alleluias, while the ending appears as a cadence for responsories, communions, and offertories. The matter of a common language is even reflected in the origin of the entire genre of second-mode tracts. The three earliest representatives of this category—De necessitatibus, Domine exaudi, and Domine audivi—are labeled “Responsorium Graduale” or “Graduale,” not only in the Sextuplex manuscripts, but also in a wide variety of

through the thirteenth century. Thus scholars such as Peter Wagner, Ferretti, and René Hesbert regarded these chants as having originally been graduals with several verses. However, in a study based on the St. Gall tradition for second-mode tracts, Hans Schmidt has declared that “these fundamentally different types never had anything more in common than momentary names and performance."

later sources Paolo

20

This incautious statement has already been rebutted by Helmut Hucke on the basis of the Old Roman tradition. 21 Hucke has shown that in this repertoire the three tracts that are labeled Gradual Responsories share their closing formulas with five first-mode 22 graduals. Furthermore, they share mediant formulas with four of these graduals and four others. 23 The formulas in question are given in Examples 11a and b ,

respectively. Hucke does not pursue this comparison with the Gregorian tradition, mentioning in Frankish sources these melismas have numerous different forms. But if one pursues the leads he has provided on the basis of melodic identities, one finds traces of comparable relationships within the northern-influenced repertoire. Here,

only that

however, the points of correspondence are masked by the greater degree of individuality exhibited. In Example 12 the closing melisma of De necessitatibus is presented in conjunction with melismas occurring at the ends of the responds of Si ambulem, Universi qui te exspectant, and Gloriosus Deus. The order of presentation has been chosen to point up the correspondence between these works and does not imply any chronological progression from close relationship to more distant relationship. Example 13 presents two forms for mediant melismas in De necessitatibus and one for Domine audivi. Sandwiched in between is a comparable passage from the respond of Timete Dominum. The excerpts from the graduals cited in Examples 12 13 each attain a higher peak than do the melismas in the tracts, as is befitting in a contrast ,

-

between first- and second-mode chants. 20 Hans Schmidt, “Untersuchungen zu den Tractus des zweiten Tones aus dem Codex St. Gallen 359” (Ph.D. diss., Bonn, 1954), p. 210; cited after Helmut Hucke, “Tractusstudien,” in Festschrift Bruno Stäblein, ed. Martin Ruhnke (Kassel, 1967), p. 116. A somewhat comparable remark made in a later summary having the same title as the dissertation would indicate that Schmidt is restricting his comparison of second-mode tracts to second-mode graduals; see his article in Kirchenmusikalisches Jahrbuch, XLII (1958), 24. 21 Hucke, “Tractusstudien,” pp. 116-20. 22The tracts are De necessitatibus, Domine audivi, and Domine exaudi, the relevant graduals Si ambulem, Gloriosus Deus, Timete Dominum, Os justi, and Universi qui te exspectant. The four additional graduals include Sciant gentes, Custodi me Domine, Adjutor meus, and 23 Miserere mei Deus.

Example 10.

The introit Exaudi Domine alleluia (a) and related chants: (b) Miserere mihi; (c) Exsurge Domine non praevaleat, Beatus servus; .

.

.

(d) Alleluia Adducentur; (e) Alleluia Qui sanat; (f) Lavabo inter innocentes; (g) Data est mihi; (h) Manducaverunt; (i) Confundantur, Principes persecuti, Psallite Domino; (j) Dum staret; (k) Sicut fui.

*Transposed down

a

fifth.

11.Exampl (a)e Terminal melisma shared by certain first-mode graduals and second-mode tracts in the Old Roman tradition. (b) Mediant melisma shared by a group of first-mode graduals and second-mode tracts in the Old Roman tradition.

Example 12.

Comparison of the terminal melisma of De necessitatibus (a) and the terminal melismas for the responds of three graduals within the Gregorian tradition: (b) Si embulem, (c) Universi qui te exspectant, and (d) Gloriosus Deus.

between mediant cadences for De necessitatibus (a-b) 13. Example Comparison and Domine audivi, verse 2 (d), and corresponding passage in the

first-mode gradual Timete Dominum (c).

A number of other loose resemblances between second-mode tracts and first-mode

graduals might be examined, but I shall cite only one further. Example 14 compares the openings of De necessitatibus and Universi qui te exspectant, the latter chant also being among those cited by Hucke. Although these comparisons are illuminating when viewed in conjunction with Hucke’s data, they do not point to the origin of the genre of second-mode tracts. Directions for the responsorial performance of De necessitatibus, Domine exaudi, and audivi do survive, particularly among Aquitanian sources. When one investigates Domine these, one finds that (1) the directions do not agree in detail and (2) the manner of performance specified is one that would ordinarily be associated with Office responsories rather than with graduals. These points are made clear in Table 1 which provides the full texts for the opening verses, together with incipits and cues given for the remainder in four important sources. The returns to mid-verse specified in these sources prompt us to inquire whether meaningful musical relationships might not exist between second-mode tracts and second-mode responsories. Before pursuing this lead, it would be well to reflect on the transformation that would be necessary in order to convert music designed for the text ,

structure of a responsory into music suitable for the text structure of a tract.

is

represented visually

in the

simplified diagram given

below

This transformation

as

Figure

1

p. 31 ), which restricts itself to those structural elements that are essential for present purposes. (Figure 1 does not take into account either the number of formulas that may be present within a given respond or the varied restatements that may occur

(see

within

a

sections

owing to textual exigencies.) Obviously, the respective opening equivalent to one another. But the close to the half verse of the tract is to

tract verse are

sought in the close to the responsory verse; these are the two main interior cadences. (Should there be correspondence in the responsory between the end of the verse and the midpoint of the respond, this section will be present within the respond itself; however, this feature is optional.) The third phrase of the tract corresponds to the opening of the second half of the psalm verse and thus has for its counterpart the opening of the responsory verse. Naturally, the close of the tract is equivalent to the close of the respond. Once the nature of the formal equivalences has been established, little comment is needed to elucidate the correspondence between an important group of second-mode responsories and second-mode tracts, as demonstrated in the two upper groups of staves in Example 15 (The responsory family includes, in addition to the chants given in full below, Velum templi scissum est, Auribuspercipe, Beatus Laurentius, Benedictus Dominus Deus Israel, Christi mater, Deletis cunctis substantiis, Impetum inimicorum, Non vos relinquam, and Propter intolerabiles rerum; additional members may perhaps be located among uncatalogued Antiphoners.) In general, the closeness of the correspondence depends in part on the manuscript readings selected and in part on the specific chants chosen. In this example, I am employing the versions of the Lucca Antiphonal, Biblioteca Capitolare, MS 601, drawing on three of the more appropriate chants. Although no two chants are identical throughout one or more phrases, the differences between the responsories and tracts are not significantly greater than the differences to be found within either genre. Certainly the basic similarities between the groups would not have escaped the notice of monastic singers, given the fact that Velum templi scissum est, one of the relevant responsories, and Domine audivi (f) be

.

The 14. Example openings of (a) the tract De necessitatibus first-mode gradual Universi qui te exspectant.

and

(b) the

both belong to the liturgy for Good Friday. The similarities of melodic outline and the numerous correspondences of detail are too great to be the result of mere chance. The fact that De necessitatibus (d) provides a double statement of the first half while Domine audivi contains a twofold statement of the second element (requiring an additional staff in the example) does not weaken the basic resemblances. What is

surprising is to find that a series of loose parallels exists between both soloistic genres and certain second-mode introits. In some instances these are limited to isolated phrases, while in others they are maintained throughout the entire melody, as shown in the lowest staves of Example 15 All three

of these

more

elaborate,

.

representatives of the latter group belong to an early layer and appear in the Sextuplex manuscripts. Vultum tuum (j) served for a variety of feasts, including three Masses for the Virgin: the Birth, the Annunciation, and the Assumption. However, it seems likely that its earliest use was in connection with the feast of St. Agnes. Multae tribulationes (k) served for the Mass for Sts. John and Paul, while Terribilis est (l) functioned as part of the Mass for the Dedication of a Church, originally the Church of Sancta Maria. The descent to A toward the end of Terribilis est and the lesser descent in Multae tribulationes may appear at first glance to represent departures from the practice of the tracts and responsories. Nevertheless, such descents are to be found in these two genres, as demonstrated by the extracts from Qui habitat and Auribus percipe added at the bottom of the final group of staves. The

correspondences observable in the Gregorian tradition for the chants given in Example 15 are not an obvious part of the Old Roman repertoire. In particular, the

Table 1

Responsorial Forms of Performance 1.

for Three Second-Mode Tracts as Indicated in Four French

Manuscripts

De necessitatibus

Paris, Bibliothique nationale, lat. 1132. De necessitatibus meis eripe me, Domine: vide humilitatem meam, et laborem meam: et dimitte omnia peccata mea.

y Ad te Domine y Et enim universi

[Cue] Et dimitte [Cue] De necessitatibus 2.

Domine exaudi

Paris, Bibliotheque nationale, lat. Domine exaudi

1132.

London,

orationem meam, et

Library, Harley 4951

British

Domine exaudi

.

.

[Cue] Domine exaudi

Brussels, Bibliotheque royale, 11.3824. Domine exaudi

.

.

.

.

.

.

.

clamor meus ad te veniat.

y Ne avertas

y

Ne avertas

y

y

In quacumque die

y In quacumque die

[Cue] Et clamor y In quacumque die

[Cue] Et clamor

[Cue] Domine y Quia defecerunt

[Cue]

sum

Domine

y Tuexsurgens

[Cue] Et clamor [Cue] Domine exaudi

[Cue]

[Cue] Et clamor y Quia defecerunt

[Cue] Et clamor y Percussus

Ne avertas

[Cue] Et clamor

[Cue] Domine y

Percussus sum

y Percussus

[Cue] Domine exaudi

sum

[Cue] Et clamor

[Cue] Et clamor y Tu exsurgens

Et clamor

y Quia defecerunt

y

Tu exsurgens

[No cue]

3.

Domine audivi

Paris, Bibliotheque nationale, lat. 1132. Domine audivi auditum tuum, et timui: consideravi opera tua, et expavi. y In medio duorum animalium [Cue] Domine audivi y In eo, dum conturbata fuerit [Cue] Consideravi y Deus a Libano veniet [Cue] Domine audivi y Operuit caelos majestas ejus [Cue] Consideravi [Cue] Domine audivi

Figure 1.

Paris, Bibliotheque nationale, lat. Domine audivi

.

.

.

y In medio duorum animalium

[Cue] Consideravi y In

eo, dum conturbata fuerit

[Cue] Domine audivi y Deus a Libano veniet

[Cue] Consideravi y Operuit

caelos

[No cue]

majestas ejus

776.

Brussels, Bibliothique royale, 11.3824. Domine audivi

.

.

.

y In medio duorum animalium

[Cue] Consideravi y In eo, dum conturbata fuerit

[Cue] Consideravi y Deus

a

Libano veniet

[Cue] Consideravi y Operuit caelos majestas ejus

[Cue] Consideravi [Cue] Domine audivi

characteristic opening is quite different in the Old Roman manuscripts. In considering the interrelationships among the three categories of chants, it is striking to note the lack of any phrase cadencing on f in the introits and the absence of such a phrase from the first strophe of De necessitatibus (d) and Deus, Deus meus (i). Only an irregular form of such a phrase is present in the first strophe of Qui habitat (h).

One could continue to explore the subject of interrelationships among chants of different modes and genres almost indefinitely. The subject is vast and cannot possibly 24 be fully encompassed within the framework of a single essay. In the end, the thesis I have proposed will necessarily remain unprovable. Nevertheless, I believe sufficient

presented to indicate that it is worthwhile to consider that chant a series of melodic procedures common to various liturgical have evolved from may functions and tonal orders. At the very least, we should be cautioned against restricting our vision to a narrow framework and studying within a set of pigeonholes, regardless

evidence has been

of their

seeming comfort or practicality. Northwestern

24

University

The present essay forms one of a series of seven dealing with various problems in the history of Gregorian chant that I hope to issue in book form in the foreseeable future. The penultimate essay, dealing with formulaic usage among Gregorian introits, will discuss a widely distributed formula that occurs among introits, offertories, communions, and responsories of modes 5-8, while the first of the essays will take up problems in the oral transmission of chant that were briefly alluded to at the outset of this article.

15. Example tracts, and introits Responsories,

of the second mode.

Index

(References to music examples and are in bold) Accademia

illustrations

Monteverdiana, recordings

86

Gregorian Chant 87 Aquitania 25 Arbogast, Paul M. 93 recordings

definition 54

-

26 7 , 59 -

-

Musica

accents

70 2

Donatus

on

53 -4, 55

Martianus

prosodic

70 2 , 215 16 , 325 , -

,

-

-

on

on

notation 355 7 -

55 7 -

Capella on 57

-

8

Bach, J.S., Goldberg Variations Bailey, Terence 265 283

xii

168

,

types 53 4 Adler, Mortimer 359 60

Balducci, Sanzio 93 Balkan epics, and Gregorian chant 248 , 280 see also Homeric poems

-

-

Agobard of Lyons

325 , 355

Albarosa, Nino xiii

Bannister, H.M. 16 77 Bannwart, Fr Roman 83 Bardolet, Sebastià M. 92 Bartlett, Frederic C., on memory 126 7 410 ,

Alcuin of York

Disputatio de

disciplina 33

332 3 , 355

-

Isidore of Seville

,

Aurelian of Réôme 10 , 34 , 60 , 231 , 355 on accents 70 2

-

on

-

,

15 20 ,

criticism of 33 , 34 5 Aurelian

84 5

Argentan, Benedictine nuns, recordings 82 Atkinson, Charles xii 214 238 Augustine, St, Vita 284

and

,

Coradini,

Arezzo Schola Cantorum Francesco

accent

melody 36 accent-theory of origins, neumes xii

422

,

vera

philosophia

50

educational reforms 49 50 -

-

,

Alleluias

Frankish tradition 340

Beiträge zur Gregorianik xiii Bellecocq, Marie-Claire 93 Benedetti, Luigi 84 Beneventan scripts see under scripts

Gregorian chant 158 9 Schlager’s catalogue 411

Berg, Alban,

'crossing' 411 , 413 , 420 music examples 412 16 -

Benevento 214 , 225 , 226

-

Wozzeck 168

Amalar of Metz 355

Berno of Reichenau 103

Prologus antiphonarii a se composti 325 recordings 84 Ampleforth Schola, recordings 79 Amsterdam, Schola Cantorum, recordings 86 Andoyer, Dom Raphaël 363 Angelus Domini, relationship 228 229 30 231

Berry, Mary xiii 99 110 Beuron, Benedictine Monks Choir, recordings 82

bistropha 104 107 199 fn30 Blackley, R. John 86 A Guide to Gregorian Chant ,

-

,

,

,

,

Ambrosian chant,

,

,

106

Boethius Consolatio 407

232

Antiphonale Missarum Sextuplex see

under

Bohn, P. 37 59

Hesbert

,

Antiphonale Romanum 76 antiphoners, neumes 309 10 antiphons, Deprecamur, neumes Apel, Willi 16 162 277

Bologna Gradual 350 Brompton Oratory Choir, recordings

-

,

De musica 55

,

350 1 -

Brou, Dom Louis 411 427 Bruckner, Albert 400 ,

80

Oral and Written Transmission in Chant

Le Graduel Romain 391

Romanus letters 103

Schreibschulen der Diözese Konstanz 391 Bruning, Dr Eliseus 80

tonal

Brunner, Lance xiii

see

rubrics, variability 421

performance xiii authenticity 107 8

Cabrol, Fernand 167 caesuras

chant

95 103 132 ,

422

propriety

also Gallican chant; Gregorian chant; Mozarabic chant; Old Roman chant 99 110 -

,

-

,

Canosa Missal 221 222 223 225 , 226 230 , 351

Cardine’s theories 102 103 107

Capella Antiqua of Munich, recordings 85 Capella Musicale del Duomo di Milano, recordings 84

equalist interpretation

,

,

,

,

6

-

,

,

,

,

and memory 126 and notation xii xiii Schola Antiqua 101 107

,

,

Solesmes

276 , 280 chant

performance, theories teaching activity 91

,

102

-

Cardine, Dom Eugène 81 82 89 99 213 247 ,

,

99 , 101 102 107

style

,

,

sound 107 10

102 103 , 107

-

,

Charlemagne administrative reforms 48 50

works

-

Graduel Neumé 83 84 85

admonitio

Is

death 214

,

,

Gregorian Chant Measured Music ? 86 Sémiologie Grégorienne xiii 83 85 87 ,

,

,

,

Gregorian

91

216 17 231 354

generalis chant

-

,

,

208 33 246 9,

archetype

-

23746

Carotta, Alberto 93

Treitler’s

critique

Carruthers,

liturgical

reforms 120 208 216

Mary

322

‘cento’

Notker

music

168

examples

,

on

,

122

Charles the Fat, Emperor 387 and notation 102

term 167

cheironomy,

‘centonate' chants xiv 161 77 , 228 , 307 analysis 168 9

Christmas Introit, puer natus 26 Claire, Dom Jean xiii 82 283

-

,

-

music

-

,

253

examples

,

170 171 172 3 , 229

,

conducting 87

-

,

,

chant

Clervaux, Benedictine Monks, recordings 85

Byzantine 307 8 choir, medieval 100

climacus

-

comparisons 282

-

Clivis

3

419 20

definition 47

,

Gregorian chant neumation 207 critique 208 210 Corpus troporum 26 239 on

hypothesis

421 2 427 432 , 436 -

,

,

,

437 40 -

,

,

music examples 423 6 428 31 , 433 5 , 437 , 441 8 -

-

,

-

Frankish, variants 340 4 342 343 -

,

history,

,

,

distinctions 409 formula

neumes

neumes

Confirma hoc Deus see under Pentecost Offertory Connolly, Thomas H. 277 Corbin, Solange 15 16 35 278

-

,

under

under

Commemoratio Brevis 326

composition 128 9 -

‘crossing’ 410

see

see

,

narratives 385 408

-

Corti, Fosco 84 de Coussemaker, Edmond 16

Crocker, Richard 104 359 361 ,

,

'crossing'

-

memorization 344 7 410

Alleluia melodies 411 413 420

Ratisbon editions 76

music

repertory

chant 410 419 20

-

,

,

-

-

Graduals 415 419 , 420 1 -

,

346

music

rhythm

example 419

Cutter, Paul F. 277

controversies 101 2 , 102 3 106 -

-

,

mensural

,

412 16

,

classification 129 core

examples

interpretations

103 -4

Dabrowski, Jerzy 94

Oral and Written Transmission in Chant

Dallafior, Camillo 93

Graduale Romanum

Deller Consort, Gregorian chant recordings 85 Deprecamur te Domine see under multiples

publications Graduale triplex 104

Deus deus meus, in early manuscripts 332 Dijon Cathedral Choir, recordings 79

Graduals 321

Dobszay,

László

xv

Donatus on

Abbey

106 , 200 228 , 250

,

,

419 , 420 1 -

,

music example 419 grammar learning 50

accents 53 4 , 55 -

Ars maior 51 5

Gregorian

-

on

Bologna 350 'crossing' 415

under Solesmes

see

tones 53

chant

abandonment, temporary 82 99 aesthetics 158 60 ,

Duchêne, Dom Guy 84

-

Alleluia melodies 158 9 -

eight-mode system 129 , 327 , 328 melismas 329 Ekenberg, Anders

and Balkan

epics 248 280 Charlemagne’s archetype 208

325

,

-

glossary

Elegerunt apostoli

and Homeric poems 280 1 Hucke’s research 359 , 360 83

,

under

multiples

,

,

89 90 -

-

episema 92 93 95 103 104 Estrada, Fr Gregori 83 ,

-

-

Ekkehard IV, Casus sancti Galli 387 389 see

33 , 246 9 , 253

of 237 46

critique

-

,

instability 327 9 meaning 263 melodies, origins -

Factus est repente see under Pentecost Laurence 420

Offertory

Feininger,

279 80 -

neumation 207

Ferand, Ernest 277 278

origin 157 409 10 recordings xii xiii 75 -

,

,

Fernandez de la Cuesta, Dom Ismael 83 Ferretti, Bernardino 95

-

Hungarian

-

,

,

,

,

Gregorian aesthetics 159 fixity, oral transmission 260 Fleet, Edgar 87

-

-

87 -

Ferretti, Dom Paolo 33 161 2 167 275 432 Estetica gregoriana xiv 158 160 247 ,

-

,

Archiv-Produktion 82 5 ,

,

60

sources

86 7 -

123

stages 271 2 traditional material, -

1 , 307

reworking of 160

-

1

transmission xiv, 116 , 180 205 -

Flossmann-Kraus, Roswitha Maria 95 96 Frere, Walter Howard 161

model 202 3 -

,

music

184 186 187 , 189 , 192

examples

,

,

194 5 , 198 -

Gajard,

Dom

Joseph

xiii , 79 , 80 , 102

Gallen, Thomas 92

Old Roman,

139 54

comparison

-

variant types 185 201 uniformity 282 323 7 -

Gallican chant 216 , 217 18 , 226 -

Louis 49

Ganshof, Francois Gastoué, Amédée 80 Gautier, Léon 223 Gelb, I.J., A Study of Writing 8 Gevaert, François-Auguste 160 La melopée antique dans le chant de l'église latine xiv

-

,

Gregorian Congress (1904) 77 78 84 99 Gregorian semiology xiii 83 85 87 91 Gregory the Great ,

,

-

John

Hymmonides on 119 20 music legend of 120 123 124 -

Guido of Arezzo 124 , 344

chant 89 90

Goar, Jacques 32 Goethe, Johann Wolfgang

-

von, on Homeric

poems 155 6 Gómez de Quevedo, Francisco 8

Micrologus 103 Gushee, Lawrence 215

-

Göschl, Johannes Berchmans 94 95 ,

,

,

Homer, comparison xiv , 115 , 280 , 281 , 409 images 117 18 , 121 , 125

,

glossary, Gregorian

,

,

,

Haas, Max 345 354 Haberl, Franz X 76 ,

.

Halbig,

Prof H. 80

,

,

126 , 157

,

Handschin, Jacques 15 16 28 9 33 34 60 64 -

,

,

,

,

,

,

,

65 389

Jordan, William 94 Journal

,

Harrison, Frank 107

of the

American

Musicological Society

xv

Hartker Codex 119 121 , 332 333 395 ,

,

,

Kanzler, Baron Rudolph 77 78 79 87 Karp, Theodore xv xvi 340 345 Aspects of Orality and Formularity

Heiman, Lawrence F. 92 3 Helisacher, Abbot 325 -

,

,

letter 355

Hesbert, Dom René-Jean 208 214 218 219 ,

,

,

Kelly, sextuplex 239

missarum

346

,

,

in

Chant xvi 322

Gregorian

,

221 , 227 , 228 , 432

Antiphonale

,

,

-

,

Columba 93

Kurris, A.M.W.J. 93

,

420 , 422

Hilario, César 93

Hiley,

Lam, Basil 87 Lambillotte, Père Louis,

David 354

Holtz, Louis 51 55 Homer,

the Great, 280 281, 409

Gregory

comparison xiv

,

115 ,

Homeric poems Goethe on 155 6

language writing,

-

-

,

,

,

,

,

,

354 , 357

also Balkan

epics

on

Gregorian

chant

archetype 208

-

33 , 246 9, -

319 20 , 323 4 , 349

Hornby, Emma xv Hucke, Helmut xi

-

-

xv

,

Hornby’s critique 320 Hughes’ critique 249 50 Treitler’s critique 237 46

179 , 203 232 238 251 ,

,

,

,

,

252 253 277 278 279 , 345 , 346 , 432 , ,

2

-

,

chant 280 1

Wolf on 155 see

writing 31

and music

,

-

Gregorian

Saint

LaRue, Jan 409 Lavery, Sean 95 96 Leo XIII, Pope 76 77 Levy, Kenneth xiv xv 169 251 277 332 351

,

and

Antiphonaire de

Grégoire 389

,

,

-

,

,

436

-

Gregorian music

response 246 9 Chant and the

chant research 359 , 360 83

-

-

examples 366 369 370 1 375

Gregorian

-

,

,

,

,

377 80

,

,

Hughes, Dom Anselm 80 Hughes, David xiv 245 249 50 251 253 261

on

-

,

,

,

,

,

,

263 278 282 299 , 329 , 330 332 , 408 Huglo, Michel 37 85 238 , 242 363 364 Husmann, Heinrich 223 ,

320

Carolingians

333 345

-

,

,

,

,

,

,

,

memorization 344

Lipphardt, Walter 276 280 liturgy, changes, Second Vatican Council ,

xii 82 ,

Lord, Albert xiii 135 139 280 322 ,

The

,

,

,

Singer of Tales

410

Lowe, Elias Avery 37 8 -

Iliad 136

improvisation,

and oral transmission 128

McElligott, Dom

indexical mode 7

,

Isidore of Seville 51 115

The Advent Project 320 McKitterick, Rosamund 324 350 357

,

on

accents 55 7 -

,

55 7

Etymologiarum on

Maloy, Rebecca 345 manuscripts Bologna Gradual

-

sound xi 108 9 -

,

Jammers, Ewald 15 99 ,

Peter 251 , 265 352

checklists 204 5 311 12 -

,

Vita

-

,

John the Deacon 120 122 157 , 275

consistency 322

,

Gregorii Magni 386 387 388 Hymmonides, on Gregory the Great John of Salisbury 8 Joppich, Fr Godechard 84

John

350 ,

Janssens, Dom 77 79

,

,

Canosa Missal 221 , 222 223 , 225 , 226 230 351

,

Jefferey,

Bernard 79

McKinnon, James 340 382

,

Gradual Laon 103

,

119 20 -

St Gall monastery 103 St Gall Canatorium 93

,

,

Gradual Codex 18: 400 3 Gradual Codex 342: 390 400 , 394 Sacramentary 391 -

Winchester

Troper 27

Maria Einsiedeln, Swiss Schola, recordings 83 Marrocco, W.T. & N. Sandon, Medieval Music 87 Martianus Capella 51 on

accents 57 8 -

De nuptiis Philologiae 57 9 , 407 Mascareñas, Oscar 328 Mass -

early books antiphoners 211 212 craftsmanship 212

-

,

-

13

modes, of representation 5 also eight-mode system Monastères Gregorian chant recordings 85 Montserrat, Schola, recordings 83 Moreschi, Alessandro 76 Morrow, Michael 108 Moscatelli, Claudio Salvatore 93

see

Mozarabic chant 80 , 180 , 284 see

also chant

recordings 83 multiples 282 4

4, 87

-

,

Elegerunt apostoli 284 Anglo-Saxon 294 Aquitaine 287 289

-

208 , 210

305

,

texts 208 , 209

Auxerre 289

propers 208 , 209 Frankish 321

Beneventan 289 90 292 , 293 4 -

-

,

melismas 285

melismas

melodies 301 2 -

eighth

mode tracts 329

music

manuscript examples 328

examples 286

‘Northern’ 295 297

Old-Hispanic

and accent 36

oral transmission 305 9 10, 11

on

origin

-

formula/formulaic system 134 5 function 24

285

Paleofrankish 296

-

versions 297 310 11 -

,

and memory 259 60 representation 5 6

reasons for

-

298 301 -

253 , 321 2

Münsterschwarzach, Schola, recordings 84

-

types 129 130 1 132 4

Murethach of Auxerre 54

variants 253 , 257 8 , 259 , 265 , 271 , 274

music writing

-

-

,

,

-

music

examples

254 6 , 258 , 262 , -

266 70 , 272 3 -

-

Old Roman 263 , 264 memorization chant 344 7 , 410 -

on

Rubin

344

on

memory Bartlett

accent-theory xii

early history 3 -45 and language writing 31 semiotic approach xii see also notation; scripts musical notation musical

see

paleography

notation 4

126 7, 410 -

and chant

performance melody 259 60

126

-

and oral transmission 115 , 309

Merton, Adolf 391

Migliavacca,

2

-

Musica enchiriadis 54

345

on

-

Vulgate 287 , 289 , 299 300 , 305

-

stability

-

-

melody early writers

288 , 290 3 , 295

,

296 , 302 4

Wagner on 161

and

,

Deprecamur te Domine 284 305

survival 212

Levy

,

,

102 , 282

-

,

neumes

Dom André xiii 76 7 78 79 94

Mocquereau,

-

Gradual codex 339: 392 , 395

Luciano 84

missa graeca 213

,

214

Nashdom

Abbey, recordings

Nettl, Bruno 277 305 ,

neumatic groups, and tones 92 separation 91 , 93

writing xii

80

,

,

tractulus 103 104 306 , 329 351

neumes

,

,

accent 15 34

trigon 37 tristropha

theory of origin xii 15 20 26 7

accent

-

-

,

,

criticism of 33 34 5

,

,

,

,

,

at St Gall

,

monastery 389 90 404 407 8 -

,

in

Byzantine 28 32 cheironomy 102 common origin 350 3

single

,

and

manuscript 334 335 6 336 9 -

-

,

,

Deprecamur antiphon English 27

350 1

-

-

function 3 4 9 -

,

Factus est repente 351 2 Gregorian, evidence 213 19 322 3

gestural 350 353 Gregorian chant 207

illustrations 121

Helisacher’s letter 355

-

-

,

-

,

iconicity 6 examples

information encoded 11 Mass books,

early 208 210 memory input 306 music examples 102 104 105

music

,

,

,

origins 47 8 248 ecphonetic notation

,

,

-

,

-

xii xiii

performance

-

St Gall Gradual Codex 342: 397 , 398 , 399

38 9 -

oriscus 93 185 241 247

Notker Balbulus 10 260

,

,

function 92

on

types 96

Gesta Karoli 387

ornamental 106

122 326

Charlemagne

,

Liber Ymnorum 404

pes 12 16 18 25 34 52 , 65 92 93 95 96 point 15

Nowacki, Edward

porrectus 13 17 19 , 20 21 24 25 26 27 , 28 29 242 , 338 , 339 351

Odo of St Maur 32 124

-

,

,

,

,

,

,

,

-

,

,

,

,

,

,

pressus 96 106 328 , 329 pronunciation 95 ,

,

,

,

,

103

,

unattached 93

quilisma 37 92 94 95 96 104 185 228 ,

,

,

,

,

,

,

230 241 247 -

salicus 92 95 106 191

Old Roman chant 128 , 362

adoption

321 386 7

Frankish

reception 361

origins

-

,

363 4 -

,

Olgiati, Antonietta 95 oral composition formulaic hypothesis, Parry xiii mixing of songs 410 in performance 128 137 164 5

7

,

199

,

scandicus 13, 16 17 93 94 95 ,

,

,

similar, in different

single

,

,

-

etymology 106 interpretations 106 repercussive 107 ,

136

transmission, Gregorian, comparison 139 54

,

,

manuscripts

,

350 3

,

strofici 93 4 table 12 13

Gregorian chant

-

,

-

,

-

14

fixation 260 1 -

torculus 13 17 18 64 93 190 329 338 339 387 -

,

,

,

Elegerunt apostoli 305 fixity 260 1 307

,

-

oral transmission xi 3 116

-

-

,

135 9 322

-

,

notes 25 6 , 26

term 10

,

,

,

themes 137

-

,

265 322 346 357

-

punctum 12 15 21 27 33 34 37 40 104 241 ,

,

formulas 137 8

,

,

xv

,

Odyssey

,

,

,

-

,

Paleofrankish 347 9 and

,

,

39 -45 61 3 66 9

oldest 124 106 108 230

,

-

,

-

,

,

-

compatible/incompatible,

,

,

Aurelian of Réôme 355 6

329 338 339

,

,

,

also scripts

see

Clivis 13 14 17 19 20 21 , 24 , 28 , 37 59 ,

,

notation

,

,

,

strata 92 , 96

,

climacus 13 17 21 103 ,

,

-

,

,

104 107

,

campo aperto 189 289 290 298 306 character, distinction 14

,

,

virga 12 15 21 27 34 37 103 104 242

-

,

,

,

309 10

,

,

uncinus 93 95

-

,

antiphoners

,

93 106 185 199

,

,

,

,

,

,

,

,

Old Roman,

comparison

and improvisation 128

139 54 -

Ponte, Joseph 333 4 Pontificio Istituto di Musica Sacra 102

139

learning

-

and memory 115 309 music examples 140 143 145 147 ,

,

also

see

oriscus

see

,

academic

,

under

87

-

musicale 16 75 6 , 83 213 , 350 ,

Pressus

composition,

under

see

neumes ,

neumes

94

prolongation signs 80

recordings

pronunciation, neumes 95 Prudentius, Psychomachia 407

Milman 280 , 410

oral

under

,

,

,

parallel readings see multiples

Parry,

see

Pothier, Dom Joseph 76 77 78

-

,

362 363 392 Paris Schola Cantorum Choir, Parkes, M.B. 36

-

,

Porrectus

,

91

purpose 91 students’ research 92 6 97 8

neumes

Oxford Anthology of Music Paléographie

degrees

foundation 91

multiples

Prüm

formulaic

hypothesis

xiii

,

135 9, 322 -

Benedictine

Abbey 223 Gradual-Troper 223 224 ,

Pater, Giles H. 93

Psalle modulamina 213

Paul the Deacon 123

Pugliese, Benvenuto 93 punctuation

Paul Warnefried 119 Paulinus of Aquileia 49

medieval 35 6

Peirce, Charles xii

signs

-

indexical mode 7

see

under

neumes

quilisma see

under

neumes

punctum

representation, modes Pelizzoni, M. Luigina 95

54

5, 6 , 7

Pentecost Greek Mass 213 , 214 15 -

Pentecost

Offertory

Rajecky, Benjamin 252

Confirma hoc Deus 218 19 223 224 225 6 Factus est repente -

-

,

,

,

Angelus Domini, relationship 228

,

229 30 , 231 , 232

recordings, Gregorian

-

in

Canosa Missal 221 222

Abbey

225

Respighi, Msgr Carlo -

-

,

227 , 229 231 239 , 242 3 -

performance,

oral

,

composition

-

-

,

,

transmission 218 , 219 25 , 220 226 8 , ,

chant xii xiii 75 87

Rella, Don Antonio 77 79

,

in Prüm

Ramm, Andrea von 108 Rankin, Susan xv Ratisbon editions, chant 76

in 128 137 , ,

79

Rubin, David 322 on

memorization 345

Ruhland, Konrad 86

164 5 -

see

also chant

performance

Perosi, Msgr Lorenzo 79 pes

see

under

neumes

Picone, Carmelo 92 Pippin, King 208 321 325 348 ,

,

pitch/es groupings

,

350 -

spectrum 4 5 6 ,

structural 331 , 332 variations 330 Pius X, Pope 91 Church music reform 77 , 99 , 101

plainchant see

alleluia

verse

Congress (1950)

series 395 7 -

Cantatorium 93 , 210 389 390 392 Gradual Codex 18, processional libellus ,

,

,

400 3 -

Gradual Codex 339: 392 , 395

notation 6 7 ,

Sachs, Curt 109 Sacred Music, International St Gall

chant

Ponchelet, René 92

Gradual Codex 342: 390 400 394 -

,

date of copying 392 3 395 later history 399 400 -

,

-

liturgy

395 6 -

notation 397 , 398 399 ,

Kalendar 391 notation 389 90 404 -

,

276

non-liturgical

sources

Isidore of Seville

407 8 -

Sacramentary 391

tropes/sequences repertories 403 Saint-Cyr, André 92 salicus

under

see

-

7

Sarum Gradual 24 , 28

Suñol, Gregory 16

,

-

under

,

,

Stephen II, Pope,

see

-

,

Die Music in Geschichte und

Samson, Joseph 79 Sandon, N. see Marrocco, W.T. de Santi, Angelo 76 77 79 Santo Domingo de Silos Choir, recordings 83 4 Scandicus

xi , 108 9

Stäblein, Bruno 83 276 283 363

neumes

,

on

Spoleto, Duchy of 214 ‘st’ sign, function 94

385

Stevens, Denis 86 strofici

see

under

neumes

5

symbol

neumes

Szendrei, Janka 261

Scherrer, Gustav 391 400 Schlager, Karl-Heinz, Alleluia melodies, ,

catalogue 411

Theodulf of Orléans 49

Schmidt, Hans 346 432

Thibaut, Jean Baptiste 37 38

Schola Antiqua

Thompson,

,

,

chant performance 101 107 recordings 86 101 Tenth-Century Liturgical Chant 99

Claude 93

tones

,

Donatus

,

,

on

53

and neumatic groups 92

109

Schola Cantorum 119 , 281

Torculus see under neumes

tractulus

foundation 321

under

see

neumes

xii , xiii, xiv, 59 , 179 , 203 , 251 , 252 , 253 , 278 , 279 , 280 , 345 , 409 10

Scholes, Percy, Columbia History of Music 79

scripts 14 20 24 Aquitanian 25

Gegenwart 71

visit to France 208 , 321 , 340 ,

Treitler, Leo

-

xi ,

-

,

,

Carolingian archetype, critique 237

27 , 30

-

Beneventan 20 2 20 3 24 26 222

trigon see under neumes

Bari type 221 classification 30 1

tristropha see

-

-

,

,

,

,

under

46

neumes

Turco, Alberto 283

-

diastematic 29 30 , 31 -

iconic 20 , 21 , 22 3 , 25 , 27 , 28 , 30 sources 23 4

uncinus

-

Paleofrankish xii , 29 , 34 , 60 , 61 3 , 64 5 , 66 9 , 71 2 -

symbolic 20

21 , 22 , 24 , 26 , 27 8 , 30 , 31

Second Vatican Council, 82 , 99

signs,

as

signifiers

liturgical changes xii

,

Solesmes Abbey Choir chant performance 99 , 101 , 102 , 107 recordings 79 , 80 2 -

Vespers & Compline

81

-

Emergence of Gregorian Chant 246 van Gerven, Wim 86 van

6

Waesberge,

virga see

under

Wagner,

Peter 16 , 28 , 48 , 79 , 128 , 159 , 275 , 281

432 works

melismas 161

Einführung

in die

gregorianischen

Melodien 179

sound chant

neumes

Medieval Ecclesiastical Chant 86

on

Responsorialis 77

80

Vollaerts, Fr, Rhythmic Proportions in Early

Liber Antiphonarius 78 Liber

photo 78 recordings

Verdun, Treaty (843) 214

Liber Gradualis 76 , 77 , 78 Liber Usualis 77

Smits 364

Vatican Commission, Venray, Friars Minor,

Solesmes Abbey publications Directorium Chori 76 Graduale Romanum 71 , 76 , 82 , 243 , 392 Les Mélodies Grétoriennes 76

neumes

van

-

,

under

de Van, Guillaume 109 der Werf, Hendrik 251 , 278 , 283 , 329 30

—

-

-

see

Gregorianische Formenlehre performance

107 10 -

160 , 162 , 247

xiv, 158 ,

,

Walafrid Strabo 386

Washington, Henry

80

Werner, Eric 38 Westminster Cathedral Choir, recordings 79 Wiesli, Walter 92 Wilamowitz-Moellendorff, U. von 156

Winchester

Troper 27

Wolf, Friedrich August, on Homeric poems 155 Wolf, Johannes 16 words, music, relationship 10 Worth, Henry G. 79

writing, function

8