Reflections on diminutions in the polyphonic music of Trecento1 A text presented as part of the critical reflection on the final documentation

The National Norwegian Artistic Research Fellowships Programme Candidate: Jostein Gundersen May 2009

An early version of this text was presented as a final exam for the "Corso di perfezionamento in lArs Nova in Europa", Universit Roma Tor Vergata, 2008.

Index

Introduction 1. Definitions of diminutions 2. Diminutions and counterpoint treatises 3. Diminutions as improvisatory element of performance 4. Some observations on the vocabulary in Codex Faenza 5. Conclusions

2 2 4 8 12 19

"...la divisione non si fa se non per aprire la sententia della cosa divisa..." Dante, Vita Nova2

Introduction:
When I applied for the Research Fellowship, I was convinced that Codex Faenza reflected an improvisatory practice. I made it a precondition for my project that I would improvise diminutions in the style of Codex Faenza to compositions of the Trecento. In the following reflections I intend to question my understanding of the relation between composition and improvisation in Trecento.

1. Definitions of diminutions:
Diminution is a term used frequently to describe ornaments in polyphonic music. Writers usually refer to passages of notes that are faster than the text declamation in vocal music (melismas), or faster than the pace of strict counterpoint in vocal or instrumental music. Attempts have been made at more exact definitions. Grove Music defines diminution as
"A term used in the context of improvised embellishment during the Renaissance and Baroque periods to describe a melodic figure that replaces a long note with notes of shorter value. Diminution is close in meaning to the English division, the Italian passaggio, the Spanish glosa, and the French double. Whereas specific graces such as trills or appoggiaturas were applied to single notes, diminutions served to decorate the transition from one note of a melody to the next with passage-work, giving scope for virtuoso display."
3

Dante Alighieri, Vita Nova, a cura di Luca Carlo Rossi, Milano, Arnoldo Mondadori Editore, 1999, chapter 7/13 (p.67) 3 Greer Garden and Robert Donington. "Diminution." In Grove Music Online. Oxford Music Online, http://www.oxfordmusiconline.com/subscriber/article/grove/music/42071 (accessed November 24, 2008).

Grove Music does not include medieval music in the definition of the term. The Oxford Companion to Musics article on "ornaments an ornamentation"4 discusses music up to 1600 in three short paragraphs, of which one paragraph covers the middle ages, mentioning one medieval theorist (Hieronymus de Moravia) and one medieval composer (Machaut), but no sources. "Diminution" is in this entry again equaled with "passaggi", and is described as "compund ornamentation". In both entries diminutions are considered ornaments, which in turn are considered improvisatory and nonstructural in all available entries on the subject, either implicitly or explicitly.5 Grove Music gives more information on diminution in the middle ages under the entry "Counterpoint"6, mentioning four of the five Ars Nova treatises that describe diminutions in more detail (see chapter 4). Contrapunctus diminutus is equaled to "figured composition", and it is pointed out that counterpoint and diminished counterpoint were largely separate in theory and terminology up to Tinctoris. The chapters on counterpoint up to 1600 are written by Klaus-Jrgen Sachs. In his referential study on counterpoint in the 14th and 15th centuries, Klaus-Jrgen Sachs defines contrapunctus diminutus followingly:
"...die Technik des Diminuirens, d.h. des Einfgens von Zwischentnen in die Gegenstimme eines vorhandenen Note-gegen-Note-Satzes..."7

The definition is deduced from the medieval treatises on counterpoint. Sachs writes a history of counterpoint as moving towards a "vollausgebildete Contrapunctus-Lehre", where contrapunctus means a strict two-part note-against-note counterpoint, with interchanging perfect and imperfect consonances and contrary movement between the voices as hallmarks. This counterpoint was the fundament of contrapunctus diminutus. By the end of the 15th century, the term contrapunctus usually covered
4

McVeigh, Simon and Neal Peres Da Costa. "Ornaments and ornamentation." In The Oxford Companion to Music, edited by Alison Latham. Oxford Music Online, http://www.oxfordmusiconline.com/subscriber/article/opr/t114/e4894 (accessed November 24, 2008). 5 See for example the definition of "embellishment": "That element in music which is decorative rather than structural". Robert Donigton. "Embellishment." In Grove Music Online. Oxford Music Online, http://www.oxfordmusiconline.com/subscriber/article/grove/music/08765 (accessed November 24, 2008): 6 Klaus-Jrgen Sachs and Carl Dahlhaus. "Counterpoint." In Grove Music Online. Oxford Music Online, http://www.oxfordmusiconline.com/subscriber/article/grove/music/06690 (accessed November 24, 2008). 7 Sachs, Klaus-Jrgen. Der Contrapunctus Im 14. Und 15. Jahrhundert. Untersuchungen Zum Terminus, Zur Lehre Und Zu Den Quellen. Edited by Han Heinrich Eggebrecht. Vol. XIII, Beihefte Zum Archiv Fr Musikwissenschaft. Wiesbaden: Franz Steiner Verlag, 1974. p. 142

both contrapunctus and contrapunctus diminutus. 8 These two terms serve to divide between no more than two levels: Strict note-against-note counterpoint, and any counterpoint with more than one note against another. If the treatises teach theory of composition, diminished counterpoint includes nearly all surviving polyphonic works. Evidently, Sachs definition is not compatible with the descriptions of ornamentation and diminution in Grove Music and The Oxford Companion of Music, as it would imply that the majority of notes copied into Trecento manuscripts are nonstructural and reflect an improvisatory tradition. These discrepant descriptions result from the emphasis on the aspects of performance on the one hand, and on counterpoint theory on the other.

2. Diminutions and counterpoint treatises

With a few exceptions (see chapter 4), Ars Nova treatises on composition deal only with the rules of note-against-note counterpoint. Strict counterpoint as described above is in the medieval treatises freqently described as the fundament of diminshed counterpoint9, and is often compared to the construction of a fundament.10 Several scholars have accordingly applied reductive analysis to Trecento repertoire,11 uncovering a "Skeleton" (Horsley, Bent) or a "Gerstsatz" (Huck). As Huck points out in his analysis of Quando i oselli from Codex Rossi, the diminutions in cantus (Huck calls them figurations) usually hide parallel perfect consonances in the underlying counterpoint.12 Perforce the counterpoint cannot be considered a "vollausgebildete Contrapunctus-Lehre", in which parallel perfect consonances do not play a part.
8 9

ibid. passim, particularly chapters II.6. (pp. 45-56) and IV.2. (pp. 140-169) See for example Prosdocimus de Beldomandi: "contrapunctus proprie sumptus alterius comuniter sumpti fundamentum", in Herlinger, Jan: Contrapunctus. Critical text, translation and introduction. Lincoln, University of Nebraska Press, 1984 pp. 30-31. 10 Sachs, Der Contrapunctus, pp. 84-85. 11 For example Bent, Margaret, "Ciconia, Prosdocimus, and the workings of musical grammar as exemplified in O felix templum and O Padua" in Johannes Ciconia. Musicien de la transition. Vendrix (ed). Turnhout, Belgium, Brepols Publishers pp. 65-108; Sara Fuller, "On Sonority in FourteenthCentury Polyphony: Some Preliminary Reflections", in Journal of Music Theory, vol. 30, 1986, pp. 3570 (particularly p. 47ff.), Jane Flynn, "The Intabulation of De toutes flours (B31) in the Codex Faenza as Analytical Model" in Leach, Elizabeth E., Machaut's music: new interpretations. Woodbridge, Boydell Press. 2003, pp. 175-191. Also Oliver Huck, "Die Musik des frhen Trecento", Hildesheim, Olms, 2005, p.113ff. 12 Huck, op.cit. p. 114

Yet parallel perfect consonances happen all the time in the Trecento repertoire, as do parallel dissonances, not only in the Gerstsatz, but also in the diminutions.13 It is clear that the rules of strict counterpoint as formulated above were not considered a necessity by all composers. An interesting witness in theory to this practice is the counterpoint treatise of Paolo da Firenze.14 One of his counterpoint examples basically consists of parallel fifths and octaves, "legalised" by a counterpoint of two notes in cantus against one note in tenor. When the tenor has a rising stepwise movement, the cantus goes from the fifth to the octave on each tenor note, thus securing contrary movement. When the tenor descends, the cantus leaps down from the octave to the fifth, and moves through the sixth to the next octave consonance. A glance at the compositions of Paolo reveals such counterpoint in bar 39 of the ballata "Amor da po che tu", and in the two last bars of the piedi of the ballata "Amor, dedimmi". There are also numerous cases where diminutions in cantus hide parallel perfect consonances in the underlying counterpoint. Sachs considers Paolos treatise as belonging between Klanghschrittlehre and contrapunctus. 15 (This classification of contrapuntal carachteristics should not be confused with historical/chronological development. Considering the birthyear of Paolo, according to Fallows ca1355,16 Paolos treatise must be later than for example the Berkeley treatise (1375), which according to Sachs represents the "Kernlehre"17.) If the term "diminution" is derived from contrapunctus diminutus, it is a paradox that musicology applies it to describe figurations (to use a more neutral word) appearing in music that musicology disqualifies as true contrapunctus. I am not suggesting that another term ought to be employed for figurations in all Trecento music that does not follow the strict rules of note-against-note counterpoint (although an investigation into the terminological multitude of the Trecento would be interesting). Yet we should be aware that Trecento music does not always employ the rules of strict counterpoint,
13

To mention but one example: "Io me son uno che per le frasche" by Jacopo da Bologna (the version of Florence, Biblioteca Nazionale Centrale, Panciatichiano 26), tempora 1-3, 8, 16, 29-30 (Gerst), 46, 48-49, 51-52, 62, 62-63, 64, 65-66. There are also parallel 7ths in bars 20 and 33. I follow the edition Die mehrfach berlieferten Kompositionen des frhen Trecento, Oliver Huck, Sandra Dieckmann (ed.), Olms, 2007 14 "Ars ad discantandum contrapunctum secundum magistrum Paulum de Florentia", published by Pier Paolo Scattolin: "I trattati di Jacopo da Bologna e Paolo da Firenze", Quadrivium, vol. XV - Fasc. 1, 1974, pp. 63-79. The example referred to in the text is on p. 69 15 Sachs, op.cit., p. 115 16 David Fallows. "Paolo da Firenze." Grove Music Online. Oxford Music Online. 29 May. 2009 <http://www.oxfordmusiconline.com/subscriber/article/grove/music/20842>. 17 Sachs, Der Contrapunctus, p. 145

and that diminutions consequently are not necessarily an extension of or secondary to strict counterpoint. Terminology was not always uniform in the 14th century, as is evident from Sachs study. Contrapunctus diminutus ("De diminutione contrapuncti"/de Muris)18 seems to be synonymous with a considerable number of historical terms, including contrapunctus large sumptus and cantus fractibilis ("Contrapunctus"/Prosdocimo19), discantus mensurabilis floribus adornatus (Petrus frater dictus Palma Ociosa, 133620), voces in partes dividere or verbulare (Berkeley treatise/Goscalcus)21, fractura contrapuncti or biscantus ("Ad sciendum artem cantus")22, discantare,23 etc. To my knowledge, there is no scholar who has suggested semantic differences between the terms. Sachs gives account of five treatises before Tinctoris that deal with diminshed counterpoint: the above mentioned Palma Ociosa (1336), "De diminutione contrapuncti", the Berkeley treatise (1375), "Ad sciendum artem cantus" and "Regulae de contrapunto" by Antonius de Leno (beginning of 15th century).24 The five treatises have different strategies of systematization of diminutions. Petrus displays diminutions according to the four main mensurations. Antonio da Leno shows how to write diminutions of two notes against one, then three notes against one. After that he deals with three notational issues: Prolation, alteration, and the punctus divisionis (or ponte de divisione, as he calls it), then turns to the proportions sesquitertia, sesquialtera, dupla, dupla superpartiens (8:3), tripla, quadrupla, before
18

The author of "Cum notum sit" and its second part "De diminutione contrapuncti" is generally believed to be de Muris. See Sachs, Der Contrapunctus, p. 181ff. 19 Prosdocimus de Beldomandi/Herlinger, pp. 30-31 20 Sachs, Der Contrapunctus pp. 48-49 and 140-141 21 ibid p. 145 and pp. 148-153, and Ellsworth, Oliver The Berkeley Manuscript, Lincoln, University of Nebraska Press, 1984, p. 120ff. (For a short discussion of the attribution of the treatise to Goscalcus, see the introduction, pp. 13-15). p. 120ff. 22 Sachs, Der Contrapunctus, pp.153-154 23 In some medieval treatises, discantare takes on the same meaning as contrapunctus diminutus. Sachs gives the name "eigentlicher jngerer Discantus" to this usage of the term. Ibid.. p. 36ff, particularly p. 41 24 ibid, on Petrus, pp. 140-142, "De diminutione", p. 45f. and pp.143-147, Berkeley/Goscalcus, p. 145, p. 148-153, "Ad sciendum artem cantus" pp.153-154, Antonius de Leno, p. 141f. Sachs does not publish the musical examples of Petrus. A transcription of the treatise with the music in modern notation has been published by Johannes Wolf: "Ein Beitrag zur Diskantlehre des 14. Jahrhhunderts", Sammelbnde der Internationalen Musikgesellschaft 15. Jahrgang, 1914, pp. 504-543. Sachs does not publish much of the problematic examples in Berkeley, and only gives a small portion of Antonio de Lenos examples. The date of Lenos treatise is suggested in Gallo, F. Alberto and Andreas Bcker. "Antonius de Leno." In Grove Music Online. Oxford Music Online, http://www.oxfordmusiconline.com/subscriber/article/grove/music/01062 (accessed November 24, 2008).

he finally shows how to notate the pause. "Ad sciendum artem cantus" is systematic in the intervallic relations, and uses the same tenor line for several examples. Both tenor and discantus in "De diminutione contrapuncti" are held within the space of one hexachord. Goscalcus is a special case: As Petrus, he gives examples in the four mensurations, but he only writes the counterpart, and not the tenorline. In addition, there are many long ascending ligatures interchanging with notation of shorter note values. At first the ligatures seem to function as division lines between the various examples. The ligatures are all different in ambitus, but with seemingly no connection to the verbulas. Except for Sachs, who found some consequence in the ambitus of all the examples, I do not know of attempts at explaining the examples from the point of view of counterpoint. Berger tells us that medieval musicians had a "memorial archive" that covered the three areas chant, music theory treatises (on intervals, solmization and the hexachord), and, from the 13th century, counterpoint.25 She believes the music student, having memorised all note-against-note progressions, would be taught diminutions not by way of rules to be memorised, but through numerous examples. The examples would facilitate the students invention of his own diminutions, rather than repeating formulae.26 This could help to explain the lack of a tenor part in Goscalcus examples. He probably expects the student to have memorised the preceding diagram with possible contrapuntal sonorities,27 from which one can imagine an underlying tenor to the diminutions. Goscalcus demands of a master that he respects the counterpoint when making a discant. He is aware of a common notion that diminutions (which he calls "verbula") work as intermediary notes, which then allow two or more consecutive perfect consonances to appear in the counterpoint. But, as he writes, in so doing, one "does not proceed masterfully."28 Evidently, to Paolo diminutions were not merely the tools of a cantor, but belonged to the training of a magister. The examples of diminution by Petrus allows parallel fifths and octaves in his diminutions, perhaps not so since he is writing 40 years earlier than Goscalcus.

25

Berger, Anna Maria B., Medieval music and the art of memory, Berkeley, University of California Press, 2005, p. 45. 26 ibid, pp. 152-154 27 Ellsworth, op. cit., pp.116-117 28 "...non magistraliter procedit", ibid, pp. 130-131

Antonio da Leno, on the other hand, is a later writer than Goscalcus, but still allows the kind of progressions the latter deems "non magistraliter".29 What the diminution treatises have in common is the method, not the style. They all follow more or less the scheme outlined by Berger, in that they teach intervals and proportions first, then counterpoint, and finally extend the counterpoint with diminutions.30 I do not know any study comparing the styles of the diminution treatises with the Trecento repertoire. It is also outside the scope of these reflections. Such a study would be helpful to deteremine to which extent the treatises are connected to the composition techniques of the Trecento, and vice versa.

3. Diminutions as improvisatory element of performance

Grove Music, in its description of diminutions as performative element, relates to a later repertoire of treatises on improvisation. There are reasons to consider diminutions in Trecento as signs of improvisation in performance as well. Perhaps the most important is that the transmission of Trecento polyphony often displays great variance in the diminutions, where there is more than one source for a composition. A comparison of early Trecento compositions that are transmitted in more than one source shows that there is much more variance in the transmission of cantus diminutions than in the tenor lines.31 The cases of a stable transmission of cantus diminutions with a variance in tenor are not few (I counted for example a little more than 20 in the works of Jacopo da Bologna), but not nearly as many as the variant readings of the diminutions where the tenor is the same in all sources. Furthermore, tenor-changes are very often minor rhythmic differences, whereas the top voices have a higher proportion of melodic to rhythmic divergences. The variance in Trecento

29

Sachs, Der Contrapunctus, p. 142f. According to Sachs, Antonio base himself "eindeutig auf dem Contrapunctus." Yet in the first example, there are numerous cases of parallel octaves hidden by intermediary sixths in the diminution. 30 Antonius de Lenos treatise starts with counterpoint, but he refers to a preceding, lost chapther in the treatise that must have treated solmisation: " Dinanzi se dito dele mutazione a presso dirassi dele voxe che se trovan nel contraponto." Saggi musicali italiani, Andreas Giger (ed.): http://www.chmtl.indiana.edu/smi/quattrocento/LENREG_TEXT.html. Accessed 29.05.2009 31 I follow the edition of Oliver Huck and Sandra Dieckmann (ed.), Die mehrfach berlieferten Kompositionen des frhen Trecento. Olms, 2007

diminutions would be a subject for a thorough study,32 based on the ductus of scribes rather than style of composers, as composers are only accessible to us through scribes.33 Another argument for not attempting a definition of style based on composer, is the number of works transmitted without ascription, or with conflicting ascriptions between the sources. Other arguments for the improvisational character of diminutions have been set forth by scholars: Brooks Toliver regards the melismas of Codex Rossi as improvisatory on grounds of their formlessness: "They seldom end on the pitches that began them...they possess little coherence in the way of imitation, voice-exchange, augmentation, and/or diminution of a melody or theme, or even any casual musical repetition."34 She concludes that the singers probably did not remember the melismas, but performed them differently each time. I find Tolivers observations difficult to accept. I do not see how the melismas in Codex Rossi are "formless". In light of Bergers writing on the medieval art of memory, the conclusion that the singers did not remember their passages seems a little weak. Assuming that the performers indeed did give the diminutions a new shape in each performance, it is just as likely that they did so because it was an open parameter, and not beacuase of memory problems. Wilson suggests that the early Florentine repertoire is in an improvisatory style.35 Contrary to Toliver, he stresses the "more theoretically schooled artistry," and less improvisational elements in the works of Piero and Jacopo.36 (Piero is represented in Rossi Codex with at least two madrigals, and Pirotta speculates about the attribution of three more songs in the manuscript to Piero.37) Wilson puts forward the possibility of a mutual influence between the northern madrigal and the florid lauda repertoire of Florence, and shows how the lauda repertoire shares melodic figures with the madrigals. Since these figures also appear in Codex Faenza, he suggests that
32

Such a study has partly been done by Huck, see footnote 30. He does not analyze the stylistic differences between the diminutions in the sources, but has made the material available for further investigation. 33 See for example Jacopos "Di novo giunto", where the scribe of Reina is the only one to notate repeated pitches across the brevis-unit (7-8, 37-38, 45-46), the scribe of Panciatichi is the only to use triplets (40, 44), and cancels 3 and shortens 2 of the 11 breaks in Reina. Huck, op.cit., pp. 63-65 34 Toliver, Brooks: "Improvisation in the Madrigals of the Rossi Codex" in Acta Musicologica Vol. 64 Fasc 2, 1992, pp.165-176, p. 167 35 Wilson, Blake McD., "Madrigal, Lauda, and Local Style in Trecento Florence" in The Journal of Musicology Vol. 15, No. 2: pp. 137-177. p. 139 36 ibid. p. 139 37 Il codice Rossi 215 della Biblioteca Apostolica Vaticana, Introductory study and facsimile edition by Nino Pirrotta, LIM, Lucca, 1992 p. 103f

an "instrumental style, perhaps the improvisatory style of the Florentine organists, may have been a stylistic tributary for both repertoires."38 Jane Flynn follows Wilsons opinion in her (reductive) analysis of Machauts "De toutes flours", and relativises her work to the probability that vocal performances included improvisation. In the process of improvising ornaments, she suspects that the "original" vocal model of the diminution at some points were simplified to its underlying counterpoint, and then re-ornamented.39 Timothy McGee considers the diminutions in Faenza Codex "rare examples of ornamentation and improvization".40 Unlike most other scholars, he does not always stress the importance of counterpoint, but also considers other origins of diminutions. In his book on medieval ornaments,41 he connects Codex Faenza to a general Italian practice throughout the middle ages that remained as "close as possible to the vocal and ornamental style disseminated from Rome as early as the eighth and ninth centuries and probably in practice there from much earlier."42 In my opinion, McGees attempts at drawing lines across enormous spans of time makes it difficult to draw any clear conclusions. Still, the possibility that diminutions, or some figures of diminutions, are born of a performance practice independent of counterpoint cannot be rejected. The fact that we do not have a second source for any of the diminutions in Codex Faenza limits our ability to see the consistency in the transmission of the diminutions. The one song that was copied twice in the codex ("Jour mour lanie" on 43r-v and "Iorleuie" on 50r-v) in fact shows a more stable transmission of the diminution than of the tenor. There is one rhythmic variant reading in tempus 32 of the cantus, reported by Plamenac,43 versus more substantial discrepancies in the tenor lines in tempora 18-20, 29-32 and 34, in both rhythm and pitch.44 Apart from that, the notation is different, and according to Memelsdorff by two different scribes.45
38 39

Wilson, op.cit., pp. 161-162 Flynn, op.cit. p. 176 40 McGee, Timothy J. "Instruments and the Faenza Codex." Early Music 14, no. 4 (1986): 480-90, p. 480. He confirms his view in "Ornamentation, national styles, and the Faenza Codex" in Early Music New Zealand, vol. 3 Pt. 2, 1987, pp.3-14. In his writings known to me, he does not deliver any direct arguments for the opinion that Codex Faenza reflects an improvisatory style. 41 McGee, Timothy J., and Randall A. Rosenfeld. The Sound of Medieval Song: Ornamentation and Vocal Style According to Treatises. Oxford: Clarendon Press, 1998. 42 ibid., p. 152 43 Plamenac, Keyboard Music of the late Middle Ages in Codex Faenza, AiM, 1972, p. XVIII 44 In tempus 18 the first note of "Jour mour lanie" has a quarter note and two eight notes, whereas "Iorleuie" has a punctuated quarter note and an eight note. The third note of "Jour mour lanie" is left

10

The main problem of regarding diminutions as a remnant of a improvisational performance practice, be it the diminutions of Codex Faenza or of any other manuscript, is that our concept of improvisation is not analogous to the that of the middle ages. Margaret Bent problematises the projection of our understanding of improvisation, which according to her includes "the notion of spontaneous, unpremeditated music-making", onto medieval music.46 Although not all musicians and scholars would agree that "unpremeditated" is an appropriate adjective to describe contemporary understanding of the word,47 I agree with Bent that we are likely to assume differences between written and unwritten music that do not apply to a culture where nearly everything would be memorised. As Bent points out, and as anyone can confirm by searching the database "Thesaurvs Mvsicarvm Latinarvm",48 the adjective improvisus ("unforeseen") is very rarely used before 1500,49 and not once in the above mentioned five treatises that describe in more detail methods for diminishing. According to Berger, the musician of the 14th and 15th century could compose polyphonic music in his mind and perform it by heart, without ever having to write it down.50 When a change occurred, was it a lapse of memory, improvisation, or revised composition? The question is impossible to answer precisely. Our division between system (composition) on the one side and arbtrariness (improvisation) on the other does not apply to the middle ages. The medieval way of making music might have possessed an openness on all levels of the "composition", from contrapuntal progressions and musica ficta to diminutions. In this context, diminutions are not
out. In tempus 19 the rhythm of the two versions are opposite, and in tempus 20, the syncoptade rhythm of "Jour mour lanie" is replaced by a quarter note and two eight notes on the same pitches. The variant readings of 29-32 was reported by Plamenac, ibid., to which should be added that "Iourleuie" ends on a G. 45 All references to scribes in Codex Faenza in this article are from Pedro Memelsdorff, in "Motti a Motti: Reflections on a Motet Intabulation of the Early Quattrocento." Recercare: Rivista per lo studio e la prattica della musica antica 10 (1998): 39-68, "New Music in the Codex Faenza 117." Plainsong and Medieval Music 13, no. 2 (2004): 141-61, and information presented during medieval seminars at Escola Superior de Musica de Catalunya, Barcelona, 2004-2008, and Venice, Fondazione Giorgio Cini, Seminari di Musica Antica, May 2007 46 Bent, Margaret. ""Resfacta" and "Cantare Super Librum"." Journal of the American Musicological Society 36, no. 3 (1983): 371-91, p.374 47 Blackburn comments on Bents statement in Blackburn, Bonnie J., "On Compositional Process in the Fifteenth Century" in Journal of the American Musicological Society 40 No. 2, pp. 210-284: "It is hard to believe that any musical result could be obtained if one insists that improvisation be `spontaneous, unpremeditated", p. 258. Bent refers to Apel, though, who goes even further in his definition of improvisation, as an art of performing without aid of manuscripts, sketches, or even memory. Bent, op. cit., p. 374. 48 http://www.chmtl.indiana.edu/tml/start.html 49 Bent, op.cit., p. 375 50 Berger, op. cit., p. 159, or chapter 6, passim

11

more or less arbitrary than the rest of a composition. In the last chapter, I would like to demonstrate that in the paraphrases in Codex Faenza, that I and others have considered improvisations, certain figures are very important for the character of not only single pieces, but groups of pieces.

4. Some observations on the vocabulary in Codex Faenza

There have been made attempts in the past at codifying the diminutions of Codex Faenza. The contributions of McGee51 are unfortunately full of generalizations: While being correct in observing that Codex Faenza does not reproduce shifted rhythms across the unit of measure in the vocal models, his allegation that Codex Faenza does not know this subtlety52 is contradicted by the Kyrie on 88r-90r (measures 121, 124), and the Benedicamus on 97r-v (measure 17). He observes that the intabulations of French pieces have one less level of diminutions than the Italian, and that the Italian diminutions have more sudden changes between quick and fast notes. This might be a general tendency, but ignores the fact that Le ior and Viver ne puis have semiminims (Le ior also has semiminim-triplets) and very abrupt changes of speed. Vice versa, some of the diminutions on Italian songs are as slow as the original, and would then fall into one of McGees descriptions of French style (Rosetta on 50v-52r, Un fior gentil). His account of the use of the rhythm SB-MSB-M in "simple time" (senaria perfecta in the original notation), "compund duple" (senaria imperfecta) or "compound triple" (novenaria) as a "substitute for even quavers in the vocal model" is strange indeed. He counts four instances in the seven French pieces (where the vocal model is known) and no cases at all in the Italian examples.53 To say that this rhythm does not occur in the Italian songs is of course wrong (see list below): The vocal model for Io me son was recognised by Plamenac, and obviously overlooked by McGee. That he did not know Deduto sey (attributed to Zacara only in 1997 by Caraci Vela54) or suspect the connection of Plamenac 42 with

51 52

His principle article in the style of Codex Faenza is "Ornamentation and national styles" ibid., p.8 53 ibid. p.10 54 Vela, M. C. (1997). "Una nuova attribuzione a Zacara da un trattato musicale del primo Quattrocento." Acta Musicologica vol.69 no.2, 1997, pp.182-185.

12

an Italian motet (suggested convicingly by Memelsdorff in 199855) is understandable, but serves only to prove the danger of generalising on fragmentary knowledge. The following (incomplete) list gives a better image of the employment in Codex Faenza of the rhythm referred to by McGee:
Kyrie 2r-3r (bar 64), Gloria 3v-5r (16), Hont paur (23-24), De tout flors (33), Aspire refus (12), De ce fol penser (4), Jay grant espoir (20), Constantia (58), Viver ne puis (45), Elas mon cuer 39r-40r (7), Deduto sey (22), Plamenac 17 (25), Sangilio (3), Plamenac 23 56 (3), Io me son uno (26-27), Gloria 90r-92v (159-160), Plamenac 42 (64), Plamenac 43, (18), Plamenac 45 (3), Plamenac 46 (20), Ave maris stella (6).

The existence of this rhythm in (at least) three compositions of Italian origin, makes it highly problematic to categorise its usage as part of a French "national performance practice",57 with the imminent danger of creating circular arguments. Michael Kugler distinguishes several figures and their usage in his extensive work on Codex Faenza,58 among them "Triolenwendung" (p. 57ff.), "Initialformel" (p.60ff.), "Tonumschreibungswendung" (89ff.), (p.62ff.), "Quintoktavfloskel" (97ff.), and (75ff.), "Semiminimatriole" "Tonrepetitionen" "auskolorierte

Schluklnge" (108ff.). He also lists various "formulae" and describes their usage.59 One of the problems with Kuglers work is that he, as McGee and most others, stresses the differences between a French and an Italian section of the manuscript, perhaps because he believed Codex Faenza was written by one person, as was the common opinion until Memelsdorffs article in 1998.60 Thus the main parameter of investigation into style has frequently been compositional model and/or origin, which in many cases is not completely clear. It is possible to investigate musical style without primary knowledge of scribes, provided one does not base research on hybrid editions, such as many of the
55 56

Memelsdorff: "Motti a motti" See footnote 65 57 McGee, "Ornamentation and national styles", p.11 58 Kugler, Michael, Die Tastenmusik im Codex Faenza, Tutzing, 1972 59 see f.ex. p.66 in the section on Italian music, or p.151ff. on liturgical music 60 Memelsdorff, 1998, "Motti a motti". Memelsdorffs verdict, that his "hypothesis [of several scribes rather than one for the whole codex] invalidates many, if not all, of the analytical generalizations attempted in the past", with particular reference to McGee and Kugler, p. 42, is quite harsh. Whereas McGees lack of detail in his observations disqualifies many of his results, no matter how many scribes were involved, Kuglers primary interest in distinguishing between the functions of the various figures is in itself not a blind alley.

13

volumes of Polyphonic Music of the Fourteenth Century. 61 Oliver Huck suggests various strategies that I find useful. By separating pitch, rhythm and counterpoint, he arrives at results that are informative for all these parameters, and eventually also leading to conclusions on style.62 In the following, I will draw attention towards a few figurations in Codex Faenza that might look common, but deserve special attention. I will first look at a figure and its variants that appears many times in the manuscript, the majority of the them in gathering 10 (folios 88r-97v)63. The elements of this figure was partially discussed by Kugler, who gave the name "Formel a" to the part corresponding to the second half of my figures A3/a3. He also pointed out that it was frequently used in combination with the first part of the figure corresponding to my A1/a1. Kuglers "Formel a" appears very frequently in Codex Faenza, as do his other micro-formulae,64 too frequently to justify their categorisation as diminutions characteristic to the liturgical music:
The following list of "Formel a" in the manuscript gives an indication of the frequent appearance of Kuglers formulae. (Only the occurences in senaria perfecta are counted here, but it can be found in other divisions as well.) "Formel a" is very often on the penultima (marked with a p). For sake of simplicity, the numbers follow the bar numbering in Plamenacs edition: Hont paur: 3, 21, 70, De tout flors: 3p, 6p, 7p, 18, 20, 29p, 35p, 44p, 62p, Aspire refus: 24, 25, 36, Elas mon cuer 39r :15p, 18p, (28p), (38p), 42p, 44p, De ce fol penser: 5p, Jay grant espoir: 14p, 37p, not in Constantia, (only first part of Figure a in bars 14, 22, 57. This song seems to have a slightly different vocabulary than its neighbours in the ms), Viver ne puis: 5p, 10p, 14p, 43p 65, 47p, Plamenac 18: 10, 16, 18, 20, Plamenac 23: 66 8p, Biance flour: 41p (many occurences in Aquila altera, but it is not in senaria imperfecta), Un fior gentil: 10p

61 62

Loiseau-Lyre, Monaco Huck, 2005, Die Musik des frhen Trecento, chapter I.4.4 (particularly pp.111-113), chapter II.5.4, chapter III.2.4 63 On the gatherings and their order, see Plamenac, "A Note on the Rearrangement of Faenza Codex 117", in Journal of the American Musicological Society, Vol.17, No.1, 1964, pp. 78-81, and Memelsdorff, "Motti a Motti", particularly p. 44 64 Kugler, op.cit. p.151 65 if one considers the 4th note an error, which I do not doubt: The neck of the note corresponds with a b, and the figure would be extraordinary with three repeated cs 66 The Sangilio gap is well known. McGee has suggested that the piece continues from bar 26 into the next piece (Plamenacs no. 23) after the double line in the last system of 54v (the end of Plamenacs Sangilio). See his edition of the piece in McGee, Medieval instrumental dances, Indiana, University Press, 1989, pp.150-159. See also the list of errors in McGees cleffing in this piece, in the review by Brainard, Ingrid, in Dance Chronicle, Vol.15, No.2, 1992, pp. 237-243.

14

None of Kuglers formulae are particular for the liturgical music. But, as the following example shows, some of the combinations are close to exclusive. (Where more than one note is given in the lower staff, it indicates that the figure in the higher staff occurs (untransposed) on all those notes.)

15

16

Figure A1-3 appear 32 times, figure a1-3 appear 26 times. Of the total of 58 appearances, only six are outside repertoire of diminutions on liturgical tenors. Three are in Elas mon cuer (2xA and 1xa), one in Viver ne puis (A), one in Rosetta 82v (a), and one in De tout flors (a). Of the remaining 52, an overwhelming 47 occur on the folios 88r-97v, the collection of predominantly liturgical music at the end of the manuscript.67 The only variants that have relatively stable functions are A3/a3. They appear a total of 14 times, of which 12 are penultima-sounds. The other motives occur in various contrapuntal situations. Contrary to the impression that diminutions are arbitrary and replaceable, we see diminutions that are used consistently, almost independent of the tenor line. (A1 appears over all notes of the scale, c,d,e,f,g,#g,a,b). Indeed, they create an element of identity for the diminutions between 87r-97v, reinforcing the impression that the section should be seen as a whole.68 made present not only throughout but also at the very beginning (Kyrie 88r bar 1) and close to the end (Benedicamus 97r-v, bar 24-25). With the exception of the variants A3/a3, the figures A and a cannot be considered extensions of a particular contrapuntal progression. The following "figure B" appears extremely seldom:

It is quite unexpected that there is only one case in the whole ms of more than three successive falling third/rising second (Kyrie 79r, tempus 8), considering the rather high rate of the contrary movement rising second/falling third (example below). According to Memelsdorff, the same scribe is responsible for "Soto limperio" and for

67 68

As Memelsdorff has shown, 97r was not originally the last folio of the codex. "Motti a motti", p. 44 Cf. Caldwells suggestion that the three diminutions on 95r-96v are of liturgical origin, in Caldwell, John. "The Organ in the Medieval Latin Liturgy, 800-1500." Proceedings of the Royal Musical Association 93 (1966): 11-24, and Memelsdorffs similar (in part artistic) proposals regarding the two diminutions on 93r-95v, in "Motti a motti", passim, and the recording Faventina. The liturgical music of Codex Faenza 117 (1380-1420), nave, 2007

17

the Kyrie on 79r.69 As with Figure A/a, its employment is quite local, and perhaps one of the strongest traits of Kyrie 79r.

Not regarding counterpoint, this kind of long chain of rising second/falling third appears, in addition to the examples,70 in La dole sere (46-47), In perial sedendo (78-79), Kyrie 79r (16-17), Deduto sey (88-89), and with less than four consecutive rising second/falling third in another ca 25 places in the ms, particularly concentrated in Che pena questa (5 times), and Constantia (3 times). Finally, Figure D, a rhythmic figure that appears in only two songs, and Figure E, a quite resemblant rhythm, that appears much more frequently:

In the intabulation of Deduto sey, figure D is presented 21 times. In Le ior it is used 3 times. Rhythms corresponding to figure E, on the other hand, is used in as many as 17 songs.

69 70

Personal communication The first example, the erased Gloria 26v, has been restored and published by Memelsdorff, "New Music in the Codex Faenza 117"

18

I could have presented many more examples, but hope these modest "word counts" will suffice to demonstrate that Faenza does not have one vocabulary corresponding to two national styles, or one style for liturgical tenors. Some figures are of almost "private" character (Figures B and D), used in only one or two songs. Other figures, (Figures C and E) looking suspiciousy like their more exclusive relatives, are to be seen in numerous diminutions, and cannot be said to belong to only one scribe or only one formal category. Figure A/a appears very often, but largely within one section, and strengthens the hypothesis that the diminutions in that section should be considered unity.

5. Conclusions
Discussion on diminutions often tend to emhasise their secondary character, either as result of counterpoint, or as more or less arbitrary formulae, expressions of improvisational practices. Although good reasons exist for continuing research in both these directions, it has here been my intention to show that diminutions are not necessarily dependent of counterpoint, and that they through consequent employment within a (group of) composition(s) contribute to the identity of that unit as much as any other parameter. In light of Bergers research, we should want to reconsider the relation between composition and improvisation in the late middle ages.

19