A shorter version of this paper was read at the Eighteenth Annual Conference on Medieval and Renaissance Music, Egham, in July 1990 I would like to thank Hollace A Schafer and Andrew Tomasello for comments on the present version. Abbreviations for manuscripts cited in this article are listed in the AppendixGoogle Scholar

1 The only collections from the first half of the fifteenth century that may be linked definitively to a composer whose music they transmit are the two manuscripts devoted to the songs of Oswald von Wolkenstein Vienna, Nationalbibliothek, Cod 2777, and Innsbruck, Universitatsbibliothek, without call number For an assessment of Oswald's musical models see Welker, Lorenz, ‘New Light on Oswald von Wolkenstein Central European Traditions and Burgundian Polyphony’, Early Music History, 7 (1987), 187–226 The best-known music manuscripts from the fourteenth century to demonstrate convincing associations with a composer are the ones supervised by Machaut Lawrence Earp, ‘Machaut's Role in the Production of Manuscripts of his Works’, Journal of the American Musicological Society, 42 (1989), 461–503, explores some of the limits of authorial control over accurate transmission of the composer's own worksGoogle Scholar

2 Aubry, Pierre, ‘Iter hispanicum II Deux chansonniers français à la Bibliothèque de l'Escorial’, Sammelbande der Internationalen Musikgesellschaft, 8 (1906–7), 517–34, dated EscA c 1460 and thought it was the work of a single conservative scribe Heinrich Besseler, ‘Studien zur Musik des Mittelalters, I Neue Quellen des 14 und beginnenden 15. Jahrhunderts’, Archiv fur Musikwissenschaft, 7 (1925), 167–252 (pp 241–2), suggested that the manuscript was in two layers, each the work of a different scribe In ‘Escorial Liederbucher’, Die Musik in Geschichte und Gegenwart (Kassel, 1939–51), ii, cols. 1520–3, Besseler refined his description and asserted provenance at the court of Philip the Good during the 1430s In 1958 EscA was published in facsimile as Codex Escorial Chansonnier, Documenta musicologica, zweite Reihe Handschriften-Faksimiles, 2 (Kassel, 1958), with a brief Nachwort by Wolfgang Rehm describing the structure of the manuscript Rehm followed most of the readings in EscA for his edition of Binchois's songs, Die Chansons von Gilles Binchois (1400–1460), Musikalische Denkmaler, 2 (Mainz, 1957) The most complete description of EscA appeared in Walter H Kemp, ‘The Manuscript Escorial V III 24’, Musica disciplina, 30 (1976), 97–129 Kemp also took EscA as the basis for his exploration of Binchois's style in Burgundian Court Song in the Time of Binchois The Anonymous Chansons of El Escorial, MS V III 24 (Oxford, 1990)Google Scholar

3 The full call number is Munich, Bayerische Staatsbibliothek, Handschriften-Inkunabelabteilung, MS galloram monacensis 902, olim Musiksammlung, MS 3192 Previous sigla include MuM (Besseler, ‘Studien‘), Mu1 (Rehm, Die Chansons) and Mu (Kemp, Burgundian Court Song)Google Scholar

4 The other Italian sources for early songs by Binchois are PC2, PC3, P4917 and BooGoogle Scholar

5 The only source for one of the three ‘late’ works (Comme femme desconfortée) that dates from before 1460 is the fragment M9659, probably from the 1450s See Petzsch, Christoph, ‘Fragment mit acht dreistimmigen Chansons, darunter Lochamer-Liederbuch Nr 18’, Die Musikforschung, 27 (1974), 319–22 Later sources for Comme femme desconfortée, Tout a par moy and Je ne vis onques la pareille are listed in Rehm, Die Chansons, 63–4, with additions in David Fallows, ‘Binchois, Gilles de Bins [Binch, Binche] dit’, The New Grove Dictionary of Music and Musicians (London, 1980), ii, 709–22 Writers have questioned the attribution of these chansons to Binchois on grounds of style and the lateness of the sources Moreover, Tout a par moy and Je ne vis onques la pareille receive conflicting ascriptions in contemporary manuscripts No such conflicts arise for Comme femme desconfortée, but only one of eleven manuscripts provides an attributionGoogle Scholar

6 Rehm, Die Chansons, contains 55 songs securely attributed to Binchois EscA and M902 transmit 25, three of which were not included in the edition Depuis le congié que je pris (copied in EscA by scribe B) also appears in M902, where it lacks a verso folio, every surviving verso in that manuscript contains a Binchois attribution Adieu ma très belle maistresse (EscA, scribe A) was almost certainly headed with Binchois's name in Tr92 before the leaf it appears on was trimmed An additional incomplete virelai was copied on f 19 of M902 For further discussion and a transcription of the virelai see Slavin, Dennis, ‘Binchois’ Songs, the Binchois Fragment and the Two Layers of Escorial A' (Ph D dissertation, Princeton University, 1987) One song included in the edition, Liesse m'a mandé salut (EscA, scribe B), has conflicting attributions to Binchois and Grossin (see below, note 49) I have counted it as Binchois's, but its authorship is questionable The apparent absence from EscA of Binchois unica is misleading only five chansons (one by Binchois) are ascribed in EscA, therefore the only works that can be identified as his are those with concordancesGoogle Scholar

7 The ‘Burgundian’ features of EscA are explored in greater detail in Kemp, ‘The Manuscript Escorial V III 24’, and idem, Burgundian Court Song In a recent review of the latter, Rob C Wegman, Music and Letters, 72 (1991), 264–8, refined Kemp's interpretation of the spellings of the two Dutch pieces in EscA, suggesting that they point towards Ghent or Bruges (and away from Brussels)Google Scholar

8 Kemp, Burgundian Court Song, adds as many as 14 anonymous unica from EscA to the Binchois corpus on grounds of style Although I will argue below for a refinement of Kemp's stylistic criteria for such attributions, they represent a major contribution to the study of fifteenth-century song The anonymous songs in EscA were published in Anonymous Pieces in the Chansonnier El Escorial, Biblioteca del Monasterio Cod V III 24, ed Walter H Kemp, Corpus mensurabilis musicae, 77 (American Institute of Musicology, 1980)Google Scholar

9 For information on the careers of Fontaine and Vide see Wright, Craig, ‘Fontaine, Pierre’, and ‘Vide, Jacobus’, The New Grove Dictionary, vi, 696–7, and xix, 712Google Scholar

10 For an outline of Binchois's career see Fallows, ‘Binchois’, 709–10 The only other extant chanson collection from Binchois's lifetime that might have been copied within the ‘Burgundian’ orbit is M9659 (see above, note 5)Google Scholar

11 Six-lined staves, black notation and the occasional use of C and F clefs in combination indicate that scribe A was somewhat old-fashioned, even Italianate, in his copying technique or that he worked from Italian-style exemplars Binchois himself may have been influenced by southern traditions, after all, his early songs survive only in manuscripts from Italy, albeit in sources that employ five-line staves However, these traits should not be taken to imply a sojourn in Italy by Binchois since few northern manuscripts copied during the 1420s are extant and such a visit remains undocumentedGoogle Scholar

12 All of the three-voice pieces follow this order of voices, as do the three cantus/tenor duets, in the only duet for equal voices, Par tous les alans/Cheluy qui vous remerchira, the voice that ends in the upper octave is at the top of the rotated opening, Dueil angoisseus, rage demeseurée, for which five voices are transmitted in EscA, is the only song in which the cantus does not appear first Large-format manuscripts such as Ox, which often align voices vertically on one leaf (without needing to be rotated), may have inspired the vertical arrangement in EscAGoogle Scholar

13 Kemp, Burgundian Court Song, viii, observed that the strongest indication of the relative chronology of the two repertories is their implied mensurations 11 of 28 songs copied by scribe A display major prolation, but only three of 34 songs copied by scribe BGoogle Scholar

14 J'ayme bien celui qui s'en va is attributed to Fontaine only in Ox, where it appears without a contratenor In Q15 it is anonymous and has a different contratenor from the ‘contratenor trompette’ in EscA Since all of the other attributable songs copied by scribe A are by Binchois, we must consider the possibility that this composer added a voice to the duet by Fontaine (This view was advanced in Vivian Safowitz, ‘Trumpet Music and Trumpet Style in the Early Renaissance’, M Mus dissertation, University of Illinois at Urbana, 1965) On the other hand, there are no stylistic grounds for attributing the contratenor trompette to Binchois Heinrich Besseler argued for attribution to Dufay in ‘Die Entstehung der Posaune’, Acta musicologica, 22 (1950), 8–35Google Scholar

15 Another difference between the scribes is the care and competency with which they recorded the poetic texts While the spelling of the French and Dutch texts copied by the earlier scribe is generally good, there are many inconsistencies and errors in the work of scribe B (who copied only French texts)Google Scholar

16 See the concordance tables in Kemp, Burgundian Court Song, xGoogle Scholar

17 If the contratenor of Fontaine's J'ayme bien celui qui s'en va is by Binchois (see note 14) there would be no overlap among the concordances of non-Binchois songs copied by the two scribesGoogle Scholar

18 The Binchois songs in Lucca and Reina are late additions to manuscripts that were copied before the 1430s. The prominent position in Lucca of Dueil angoisseus, rage demeseurée on ff 1^v–2, where it is a palimpsest, was probably due to the fact that the piece that was scraped away began with the letter D – the only initial D in this section of the manuscript I am grateful to John Nádas for this suggestionGoogle Scholar

19 As a tenor partbook, PC3 stands outside this analysisGoogle Scholar

20 For this chanson and those that follow I shall refer to readings in a variety of manuscripts Access to the transcriptions and critical report in Rehm, Die Chansons, will be helpful, but not all of the variants are noted therein Full presentation and analysis of the variant readings may be found in Slavin, ‘Binchois’ Songs' The errors shared by the contratenors of Adieu, adieu mon joteulx souvenir in Ox and R1411 are (1) repeated e semibreves in bars 34–5 (the other readings have g semibreves, the es sound against repeated d's in the cantus), (2) the d' in bar 28 (the other readings have c[#]'), (3) the sharp preceding and aligned with the d' semibreve in bar 27 (In EscA a c#' in bar 28 creates a double leading-note’ cadence to G The sharp in Ox and R1411 probably is a remnant of a similar cadence – and the correct pitch in bar 28 – in a parent source A parallel example is bar 14, where a sharp in Ox and R1411 was aligned with the b semibreve, it should apply to the following c' – as is clear in EscA - resulting in another double leading-note cadence to G)Google Scholar

21 The weak-beat cadences in Ox are created by rests in all of the voices at the beginning of the rondeau By following the midpoint without pause, the second section shifts the cadences to ‘strong’ beats In EscA all of the cadences fall on strong beats because there are no rests at the opening and semibreve rests appear in all of the voices at the start of the second section Ignoring the rests in Ox would produce a weak-beat ending, unless we insert semibreve rests at the midpoint The reading in EscA efficiently orientates the phrase structure to the underlying metrical organizationGoogle Scholar

22 The other songs with cadential displacement are Les très doulx jeux du viaire ma dame and Nous vous verens bien Malebouche, both of which end on ‘weak’ beats Neither, however, begins with weak-beat cadences or with rests Rehm's transcriptions (Die Chansons, 21 and 28) ‘correct’ the displacements Alternative transcriptions are discussed in Slavin, ‘Binchois’ Songs', 18–22Google Scholar

23 Aside form Adieu mes très belles amours, the chansons transmitted with only two voices in EscA are unicaGoogle Scholar

24 Its place in the alphabetical arrangement of scribe A does not imply that Jamais tant was copied relatively early Indeed, since this rondeau spans two gatherings it may have been one of the last pieces copied by that scribeGoogle Scholar

25 Presumably the unruled pages held introductory material such as a coat of arms or a table of contents, most likely, the first leaf of gathering 4 (originally gathering 1) was removed by scribe B when he enlarged the manuscript by adding gatherings 1–3 This explanation seems more likely than Rehm's suggestion that the scribe removed the leaf to mend a copying error (Codex Escorial, Nachwort) I am grateful to Stanley Boorman for thought-provoking suggestions regarding the missing leafGoogle Scholar

26 Dueil angoisseus, rage demeseurée. Jamais tant que je vous revoye and Adieu, adieu mon joieulx souvenir transmit varying contratenors, but the divergent readings are stemmatically related The contratenors of Adieu mes très belles amours resemble each other no more than voices crafted independently to a pre-existent cantus/tenor duetGoogle Scholar

27 Fallows, David, ‘Binchois’, 718, reached this conclusion, commenting that ‘only the discantus and the tenor are by Binchois since the contratenor is different in each source and entirely omitted in the most authoritative source, [Escorial] V III 24’ The contratenor in Tr92 confirms transmission (at least to that manuscript) of only two voices for Adieu mes très belles amours because it is incomplete and obviously corrupt It reads as a failed ‘compositional’ sketch, the counterpoint breaks down in the eleventh tempus, after which the voice simply gives outGoogle Scholar

28 Seven pieces beginning ‘adieu’ by one composer are surprisingly numerous, despite the ubiquity of the genre Dufay, for example, whose extant songs outnumber those of Binchois 84 to 58 (including the three songs cited in note 6), wrote only three ‘adieu’ chansons (There are 84 songs attributed to Dufay in Gutllelmi Dufay opera omnia, ed Heinrich Besseler, Corpus mensurabilis musicae, 1, vi (Rome, 1964), none of the opera dubia therein begins with ‘adieu‘)Google Scholar

29 P4917 opens with Adieu m'amour and the first layer of EscB (ff 7–27) begins with Adieu, adieu mon joieulx souvenir The latter rondeau also begins a series of 17 three-voice songs in R1411, where it is preceded by two anonymous duos (One of those duos is found elsewhere in three voices, see Fallows, David, ‘Busnoys and the Early Fifteenth Century A Note on “L'ardant desir” and “Faites de moy”’, Music and Letters, 71 (1990), 20–4)Google Scholar

30 Duets with voices a fifth apart or with slower-moving tenors were not rare, but – as witnessed with Adieu mes très belles amours – they often appear with ‘added’ contratenors For a fascinating analysis of differences between versions with two and three voices of a slightly earlier song see Walter H Kemp, ‘A Chanson for Two Voices by Cesaris, “Mon seul voloir”’, Musica disciplina, 20 (1966), 47–56 David Fallows, ‘Two Equal Voices A French Song Repertory with Music for Two More Works of Oswald von Wolkenstein’, Early Music History, 7 (1987), 227–41, surveys duets from the first half of the Fifteenth centuryGoogle Scholar

31 If the ‘thin’ version in EscA is authoritative, an additional pun on ‘adieu’ may have been intended by a textural allusion to an actual departure Such an event would also help to account for the wealth of pieces in EscA that begin ‘adieu‘Google Scholar

32 The vexing issue of ‘authority’ remains unresolved Ultimately we cannot eliminate the possibility that the scribe discarded a contratenor or that the chanson circulated as a duet before it was copied into EscA Nevertheless, although the filiation of Adieu mes très belles amours cannot establish definitive compositional authority for the version in EscA, analysis of the other variants does suggest several reasons for preferring it In the first two bars the cantus in Tr92, MuEm and Stras moves directly from a bar that expresses a triple division of the breve to a duple division One other song by Binchois (Mon seul et souverain desir) displays the same ‘syncopation’ within its opening gesture, but the composer probably did not intend this rhythm for the beginning of Adieu mes très belles amours because the underlay possibilities it presents are very poor (The Latin text in MuEm is better suited to the rhythm of that version, but the contrafact cannot have given rise to the rhythmic change because the same rhythm appears with the French text in other sources) The unusual rhythm and the awkward texting suggest that the version in EscA is closer to the composer's own For discussion of all variants of Adieu mes très belles amours see Slavin, ‘Binchois’ Songs', 36–42Google Scholar

33 The missing contratenor is not ‘significant’ because it may have been present in the exemplar Actually, the facing recto (missing) in EscB may have contained a contratenor This seems unlikely however in view of the fact that the page on which the cantus and tenor were copied is the only verso in EscB on which the scribe copied a tenor part (beneath the cantus), more likely, the recto was missing at the time the piece was copied Therefore the absence of a contratenor in EscB could be due to lack of room on the half opening, alternatively, the piece was reserved for this page because it was the only duet the scribe intended to copyGoogle Scholar

34 I am compelled to point out, however, that the sharp in Adieu mon amoreuse joye (Example 4) appears only in EscA (not in MuEm or Tr87) Fallows, ‘Binchois’, 714, cites another characteristic ‘“cross relation” effect’ in Adieu jusques je vous revoyeGoogle Scholar

35 Esc A reads d, f, g - all coloured breves Tr87 and MuEm read f, g – semibreves, a, d - coloured brevesGoogle Scholar

36 Under both black notes in Tr87 (contratenor, b minim, bar 7, tenor, breve, bar 26) the scribe remarked his error by writing a small ‘v’ (vacua) in the same inkGoogle Scholar

37 The errors and variants in MuEm are too numerous to explore here They are listed in Rehm, Die Chansons, 66Google Scholar

38 The extra note is a g’ minim after the g’ semiminim in bar 22 The error is not acknowledged in Rehm, Die Chansons, 10, where a rhythmic adjustment in bar 20 accommodates the extra minim while seriously undermining the counterpoint in bar 21 For consideration of a better reading see Slavin, ‘Binchois’ Songs', 96–7Google Scholar

39 Transcriptions of each version would take up too much room for presentation here Three of the contratenors are presented in parallel transcription in Rehm, Die Chansons, 73 All four are transcribed in Slavin, ‘Binchois’ Songs', 46–50 In M902 the cantus, tenor and CT1 of the first section of the ballade appear on the last surviving verso of the manuscript The recto of that opening probably contained at least two of the other contratenors, since the other verso folios in the manuscript contain only cantus voices and the cantus of Dueil angoisseus is followed in M902 by a page-turn indication For a more detailed discussion of the version in M902 see my ‘Genre, Final and Range Unique Sorting Procedures in a Fifteenth-Century Chansonnier’, Musica disciplina, 43 (1989), 115–39Google Scholar

40 Designation of the contratenors by number follows the labels in TrS8Google Scholar

41 CT1 lies above the tenor for 44% of the chanson, SCT is the middle voice only 18% of the time Rehm's references to contratenor ‘types’ in Das Chansonwerk von Gilles Binchois' (Ph D dissertation. University of Freiburg, 1953), chapter 9, summarized in Die Chansons, 9∗–11∗. were based on the seminal investigation of this repertory by Heinrich Besseler, Bourdon und Fauxbourdon (Leipzig, 1950) CT1 corresponds to the type that Besseler labelled Mittelstimmen-contratenor, an earlier style than the Kombinations-contratenor characteristics of the SCTGoogle Scholar

42 As mentioned above (note 18), Duetl angoisseus is an undatable palimpsest in Lucca CT1 and SCT are never found in the same manuscript However, M902 might have contained SCT on two of its missing pages (ff [21] and [22])Google Scholar

43 Although SCT is a substitute for CT2 and CT3, it is not a conflation of those voices If CT1 had already been composed, the sources for SCT were CT1 and CT2, if CT1 belonged to a later stage, SCT contained music from CT2 along with original material Unlike parts labelled ‘solus tenor’, the solus contratenor is not contrapuntally essential This is a crucial distinction in light of the compositional role most likely played by voices called solus tenor See Bent, Margaret, ‘Some Factors in the Control of Consonance and Sonority Successive Composition and the Solus Tenor’, Report of the Twelfth Congress [of the International Musicological Society] Berkeley 1977, ed Daniel Heartz and Bonnie Wade (Kassel, 1981), 625–33Google Scholar

44 In R1411, EscB and Tr88 a sharp dissonance between the tenor and the cantus occurs on the third beat of bar 39 The reading in EscA, MuEm and Lucca is preferable All of the manuscripts that transmit the corrupt cantus were compiled in Italy, Lucca is also Italian, but the correct cantus reading in bar 39 confirms the distinct avenue of transmission suggested by the presence in that manuscript of SCTGoogle Scholar

45 Evidence for an exemplar that employed a six-line stave is compelling an extended passage that follows an octave leap is transposed by a third Moreover, the C4 clef in the lower voices (the same clef used in Esc A) forces the scribe to use leger lines – an unusual step that is avoided by the use in the other five-line manuscripts of C3 Support for a black-notation ancestor for the reading in MuEm is limited to a single black semibreve (cantus, bar 3) that should be voidGoogle Scholar

46 The most consistent example of a song by Binchois in which lower voices are ‘expanded’ (via removal of ligatures and addition of repeated notes) is the reading of Amours merchi de très tout mon pooir in PC2 Some implications of these procedures are discussed in my ‘In Support of “Heresy” Manuscript Evidence for the A cappella Performance of Early 15th-Century Songs’, Early Music, 19 (1991), 178–90Google Scholar

47 Stras survives only in Charles Edmond de Coussemaker's partial inventory (now Bibliothèque du Conservatoire Royal de Bruxelles, MS 56,286) published in facsimile and edited by Albert Vander Linden as Le manuscrit musical M 222 C 22 de la Bibliothèque de Strasbourg, X V^e siècle. Thesaurus musicus, 2 (Brussels, n d) Some of the variants of the reading in Stras therefore may be the residue of careless copying in the nineteenth century, but the systematic changes in the tenor were the work of a fifteenth-century scribeGoogle Scholar

48 To argue that the readings in M902 are earlier in the stemma than those in the portion of EscA copied by scribe B is not necessarily to assert chronological precedence for M902 A late source may transmit a reading that is stemmatically early Nevertheless, contrary to the rough chronology of Binchois's songs advanced above, this section of EscA may have been compiled later than M902 Although the style of the songs in EscA suggests an earlier date than the more ‘advanced’ songs in M902, the stemmatic profile of scribe B as a copyist somewhat ‘removed’ from the repertory he copied underscores the possibility that his work was removed chronologically as wellGoogle Scholar

49 The attribution of Liesse m'a mandé salut posed a problem for two of the four scribes In Ox the Binchois attribution appears over an erased ascription to Dufay, and Grossin's name was added to Tr87 in a different ink and hand from the music (EscA and BU do not ascribe the chanson) The change in Ox might indicate scribal confidence (justified or not) in the new attribution, but it is also possible that the scribe learned of his error and preferred not to leave the song anonymous, the majority of pieces in Ox are attributed, and Liesse m'a mandé salut falls within a succession of 18 attributed songs Discussion of stylistic questions posed by the conflicting attributions is outside the purview of the present study I argued in favour of Grossin in ‘Binchois’ Songs', 164–71Google Scholar

50 Another link between the two readings is an unusual texting pattern they share In bars 21–3 all three voices participate in unison imitation EscA and Ox do not supply a text incipit for the contratenor The use of incipits as cues for lower voices is not stemmatically significant since scribes could have made independent choices, but the choice is usually circumscribed when scribes supply such cues they are for each imitative voice Plains de plours et gemissemens is unique among Binchois's songs in so far as the sources provide cues only for the tenor The lack of a cue for the contratenor was probably an oversight that was transmitted from one manuscript to the other Both scribes needed to lengthen their rastrum-drawn staves by hand in order to make room for the last few notes of the cantus This is almost certainly a striking coincidence, but perhaps both readings were copied from exemplars in which the last notes appeared in an unexpected place, such as the opposite leaf The white long that concludes the cantus in EscA may be a remnant of a void-notation exemplar (such as Ox), or it could be due to scribal carelessnessGoogle Scholar

51 Rehm, Die Chansons, 69, misread the notation of bar 19 and therefore did not notice the error.Google Scholar

52 This rondeau is without clefs both manuscripts indicate pitch relationships with flats at the beginning of the staves The flats indicate the placement of fa and the hexachord pair used in each voice. In Example 9 (based on hypothetical clefs C4♭ F4♭ F4♭) the cantus employs the G and C hexachords while the lower voices use the ones on F and C Binchois's Comme femme desconfortée and the anonymous Tous desplaisirs m'en sont prochains are the only other fifteenth-century chansons (of 20 without clefs in one or more voices) that indicate relative pitches with the same placement of flats as Mon seul et souverain desir Rehm's transcription (Die Chansons, 29) assumes that the flats were placed incorrectly For discussion of issues raised by clefless pieces see Dahlhaus, Carl, ‘Zu einer Chanson von Binchois’, Die Musikforschung, 17 (1964), 398–9, Margaret Bent, ‘Musica Recta and Musica Ficta’, Musica disciplina, 26 (1972), 73–100, and idem, ‘Diatonic Ficta’, Early Music History, 4 (1984), 1–48 The clefless repertory was surveyed also in Timothy C Aarset, ‘The Clefless Signature in Fifteenth-Century Chansons Interpreting Notational Traces of Musical Practice’, a paper read at a meeting of the New England Chapter of the American Musicological Society in February 1981 I am grateful to Prof Aarset for a copy of that paper, which will appear in a forthcoming issue of Musica disciplinaGoogle Scholar

53 This explanation seems more plausible than that proposed by Rehm, Die Chansons, 68 Rehm adopted the readings in EscA for bars 10 and 22–3Google Scholar

54 The rest in BU imposes a graceless interruption of the rhythmic and melodic arch to the cadence in bar 12 In EscA, a different rhythm for one note (the b' minim becomes a dotted semibreve) creates a compelling parallel between bars 6 and 7 (a cadential gesture to G that is quickly undermined by motion in the cantus and contratenor) and bars 10 and 11 (a cadence to A that leads to a phrase overlap) We cannot know whether this parallelism was present in a hypothetical original, but the result is preferable to the awkwardness of BU and to the unnecessary dissonance in Tr87 (on the third beat of bar 9)Google Scholar

55 An objection to the version in Tr87 comes in bar 24, in which the dotted semibreve a’ in the cantus is unacceptably dissonant The pause in EscA is extremely abrupt, in BU somewhat less so The textual interruption in BU (within the word ‘curieuse’) is exceptionally awkward Text declamation is best in EscAGoogle Scholar

56 Rehm, Die Chansons, 35, attempted to make up for the missing beat by adjusting the rhythm of the contratenor in bar 19, resulting in an awkward suspension on the downbeat of bar 21Google Scholar

57 Preliminary investigation of the Dufay songs copied by scribe B, all of which appear in Ox, also points to less than authoritative recension in EscA Both Estrinez moy, je vous estrineray (contratenor bars 5–6) and Craindre vous veul, douce dame de pris (contratenor bars 12–13 and 27) contain errors in EscA that do not appear in Ox Additionally, the cantus of Porray je avoir vostre merchi was copied in the five-line section of EscA beginning with a C2 clef, after five notes scribe B switched to C1 in order to avoid leger lines This switch suggests that his exemplar had six lines Reina and Stras, both six-line sources, employ C2Google Scholar

58 The ruling of gathering 8 (and the openings in gatherings 4 and 6 with music copied by scribe B) demonstrate that scribe A ruled entire gatherings at a time The rastra listed in Figure 1 measure as follows rastrum I (six lines, 19 5 mm), rastrum II (six lines, 18 mm), rastrum III (five lines, 16 mm)Google Scholar

59 The third song is Adieu, adieu mon joieulx souvenir, the distinct versions of which are discussed aboveGoogle Scholar

60 Even if Adieu mes très belles amours was not composed for EscA, its features may bear witness to a tradition of alphabetically ordered manuscripts and the corresponding need for short pieces that begin with AGoogle Scholar

61 Wolf, Johannes, Geschichte der Mensural-Notation (Leipzig, 1904), i, 193, first labelled M902 a ‘Binchois Sammlung’ Hugo Riemann, Handbuch der Musikgeschichte, n/1 (reprint of 2nd edn, New York and London, 1972), 98, also indicated that M902 was limited to works by that composerGoogle Scholar

62 The ordering system employed in M902 - including its implications for reconstructing the manuscript – is explored in Slavin, ‘Genre, Final and Range’ Tinctoris describes the properties of the modes in his Liber de natura et proprietate tonorum (Naples, 1476) Although he mostly refers to mode with regard to chant, chapter XXV contains his famous analysis of how to judge the mode of a polyphonic song The example is Le serviteur, presumably the rondeau attributed to Dufay However, as David Fallows has pointed out (private communication), although the range of the contratenor of the extant setting is authentic, Tinctoris labelled as piagai the range of the song he describedGoogle Scholar

63 Call numbers for the Machaut manuscripts are listed in Earp, Machaut's Role', 497–8 Music by Adam de la Hale appears as a fascicle within Paris, Bibliothèque Nationale, f fr 25566Google Scholar

64 This point is flexible in so far as the extant gatherings might have been followed by fascicles of works by other composers Appropriate analogies might be drawn to poetry anthologies from the thirteenth and fourteenth centuries, such as those of the trouvères, in which sorting by genre operates within sections limited to works by one author – in manuscripts that transmit works by severalGoogle Scholar

65 This point is obvious in theory only Those who transcribe more-or-less intact the ‘best’ version extant in contemporary sources justify the practice by dual appeals to ‘authority’ and to the ‘integrity’ of a version that happens to have survived A E Housman pointedly characterized the essential laziness of that position in his preface to Manilius, Astronomicon (London, 1903), i, p xxxi ‘[the editor] calls one of them “the best MS”, and to this he resigns the editorial functions which he is himself unable to discharge‘Google Scholar

66 For an overview of trends in cadenrial voice-leading see Randel, Don Michael, ‘Emerging Triadic Tonality in the Fifteenth Century’, Musical Quarterly, 57 (1971), 72–86 During the period represented by EscA and M902 (1430s-40s), Binchois most often employed stepwise motion (sometimes decorated to resemble ‘under-third’ motion in the cantus) in the contratenor for cadences in the final position Only Se je souspire plains et pleure ends with the type that gained prominence around the middle of the century, the double-octave cadence By the 1460s, most final cadences of three-voice chansons are evenly divided between two types the octave leap and the double octave For example, of 56 three-voice chansons in Niv, a representative collection from the early 1460s, 22 contratenors conclude with octave-leap voice-leading, 21 end with the double-octave type, 9 contratenors move by step, and 4 employ other voice-leadingGoogle Scholar

67 Statistics for cadence types in EscA include the octave-leap ending of the SCT in Dueil angoisseus, but not Adieu mes très belles amours or Jamais tant que je vous revoye, both of which lack contratenors there For M902 I have not counted Dueil angoisseus because the end of the contratenor is missing and because I believe that additional contratenors would have appeared on the missing leavesGoogle Scholar

68 For insightful analysis of problems posed by conflicting ‘authoritative’ recensions see Taylor, Gary, ‘The Rhetoric of Textual Criticism’, Text Transactions of the Society of Textual Scholarship, 4 (1988), 39–57 Among the works discussed are the three versions of Piers Plowman, each of which can be traced to the author, William LangfordGoogle Scholar

69 Considerations of scribal idiosyncrasy are discussed in the two most important general essays on stemmatics in the musicological literature Margaret Bent, Some Criteria for Establishing Relationships between Sources of Late-Medieval Polyphony', and Stanley Boorman, ‘Limitations and Extensions of Filiation Technique’, Music in Medieval and Early Modern Europe, ed lain Fenlon (Cambridge. 1981), 295–317 and 319–46 respectivelyGoogle Scholar

70 Kemp, Burgundian Court SongGoogle Scholar

71 Although Kemp recognized many of the distinguishing features of the two layers of EscA, he did not differentiate between them regarding the authority of the readings they transmit for Binchois's songsGoogle Scholar

72 Rehm, In, Die Chansons, 45–6, the rhythm of the cantus in bars 18 and 50 of the first version of Dueil angoisseus and bar 18 of the second version should be changed to follow bar 50 of the second versionGoogle Scholar

73 Kemp uses the word ‘fingerprint’ in the sense of a gesture highly characteristic of the work of one composer and rarely found in the music of others He does not imply, however, that the presence in an anonymous work of one such feature assures its attribution, rather, that a combination of several strongly reinforces the possibility of an ascriptionGoogle Scholar

74 The manuscripts that do not contain the fingerprint – EscB and Lucca - are among those compiled furthest from Burgundy, in addition, EscB is the latest sourceGoogle Scholar

75 Kemp, Burgundian Court Song, 20Google Scholar

76 One of the reasons for which it is tempting to accept the lower-auxiliary version of this rhythm as a Binchois ‘fingerprint’, despite its stemmatic instability, is the strongly marked rhythmic impetus it lends the musical phrase, a feature that one associates with Binchois because, as Kemp has pointed out, it appears in several of his works and rarely in music attributed to other composers This association, however, may be due to historical accident – the result of Binchois's long-standing relationship with the court of Burgundy and the survival of many of his works in copies by Burgundian scribes, who may have favoured this figureGoogle Scholar

77 Fallows, David, ‘Two More Dufay Songs Reconstructed’, Early Music, 3 (1975), 358–60, and 4 (1976), 99 In support of En triumphant de Cruel Dueil as an hommage to Binchois, Fallows cites several features of musical style as well as textual quotations of the titles of two poems set to music by BinchoisGoogle Scholar

78 Kemp, Burgundian Court Song, 18, cites only two songs by Dufay in which the ‘fingerprint’ appears, two others have the passing-note version Those statistics are dubious since they are based, apparently, on the transcriptions in Besseler, Guillelmi Dufay opera omnia, without reference to vanancs listed in the accompanying critical notes As mentioned above, at least one reading of an additional song by Dufay, Las/ que feraye? ne que je devenray?, contains the ‘fingerprint‘Google Scholar