This article is a revised version of a paper read first at the seventeenth Medieval and Renaissance Music Conference at the University of Reading in July 1989, and then at a Conference of Thirteenth-Century Music and Poetry organized by Susan Rankm and Mark Evenst at Emmanuel College, Cambridge, in August 1989. I am grateful to participants at both meetings for their responsive comments, and particularly to John Stevens for his attentive scrutiny of a later draft.Google Scholar

1 It is difficult to do justice in a single definition to the formal features of all the surviving thirteenth-century rondeaux, especially some of the examples in Jean Renart's Guillaume de Dole. For recent descriptions of the rondeau or rondet de carole (as the thirteenth-century examples of the form are often termed), see Bec, Pierre, La lyrique française au moyen âge (XIIe–XIIIe siècles) Contribution à une typologie des poétiques médiévaux, 2 vols. (Paris, 1977–8), 220–8, and Eglal Doss-Quinby, Les refrains chez les trouvères du XIIe siècle au début du XIVe (New York, 1984), ch. 2, esp. pp. 62–70.Google Scholar

2 Renart, Jean, Guillaume de Dole ou Le roman de la rose, ed. Felix Lecoy, Les classiques français du moyen âge, 91 (Paris, 1979), lines 1579–84; also Nico H J. van den Boogaard, Rondeaux et refrains du XIIe siècle au début du XIVe (Paris, 1969), rondeau 9Google Scholar

3 See, for example, Alfred Jeanroy, Les origines de la poésie lyrique en France au moyen âge (3rd edn, Paris, 1925); Friedrich Gennrich, Rondeaux, Virelais und Balladen aus dem Ende des XII, dem XIII und dem ersten Drittel des XIV Jahrhunderts, mit dem überlieferten Melodien, 3 vols (i and ii, Gesellschaft für romanische Literatur, 43 and 47, Dresden, 1921; Göttingen, 1927; iii, Summa musicae medii aevi, 10, Langen bei Frankfurt, 1963), van den Boogaard, Rondeaux et refrains.Google Scholar

4 The belief that refrains originated in rondeaux, first put forward by Alfred Jeanroy (see Les origines, 102–26, 387–401, 406–26), has died hard, and survives even now in many modern descriptions of the refrain, despite van den Boogaard's clear demonstration of its inadequacy as an explanation For detailed discussion of the often complex inconsistencies in the history of modern scholarship on the refrain, see my doctoral thesis, ‘Interpolated Lyric in Medieval Narrative Poetry’ (University of Cambridge, 1988), ch. 1, pp 30–46.Google Scholar

5 This is refrain no. 65 in van den Boogaard's bibliography (hereafter vdB). He lists the six contexts as follows (a) vdB rondeau no. 168; (b) chanson no. 584 in G Raynauds Bibliographie des altfranzosischen Liedes, ed. Hans Spanke (Leiden, 1955; hereafter R), (c) motet 435 in Friedrich Gennrich, Bibliographie der ältesten französischen und lateinischen Motetten, Summa musicae medii aevi, 2 (Darmstadt, 1958; hereafter M), (d) Ml 143a; (e) La court de paradis, ed. Eva Vilamo-Pentti, Annales Academiae Scientiarum Fennicae, B 79, 1 (Helsinki, 1953), lines 359–60, (f) Jaques Bretel, Le tournoi de Chauvency, ed. Maurice Delbouille, Bibliothèque de la Faculté de Philosophie et Lettres de l'Université de Liège, 49 (Liège and Paris, 1932), line 3186.Google Scholar

6 See, for example, Mark Everist, ‘The Refrain Cento: Myth or Motet?’, Journal of the Royal Musical Association, 114 (1989), 164–88, who defines (p. 164, n 1) the refrain in precisely such terms: ‘In this study, the word refrain is italicized when it is used in the sense of a single structure migrating from genre to genre, to distinguish it from the “refrain” in its modern sense of an element which repeats within a single lyric’, see also pp. 172–3. Although there are cases where this distinction applies, there are also many occasions when a single refrain enjoys both characteristics, and I have decided not to force the distinction typographically.Google Scholar

7 I say ‘seems’ since van den Boogaard does not express this principle explicitly, although he is harsh on Gennrich's listing of many refrains ‘qui de toute évidence ne le sont pas’ (p 18) and alludes (with tantalizing concision) to such key problems as the importance of the distinctions between ‘préexistants’ and ‘nouveaux’ refrains (p. 16) and between ‘variation’ and ‘variante’ (p 20), and to the especial difficulty of identifying refrains in motets (p 23, n. 21). A similar observation to mine about the number of refrain unica (specifically in motets) is made by Everist, ‘The Refrain Cento’, 170. In fact, modern disagreements over the identity of refrains are numerous. For a range of examples (including differences of opinion between Gennrich, van den Boogaard and Hans Tischler (The Montpellier Codex, Recent Researches in the Music of the Middle Ages and Early Renaissance, 2–7, Madison, 1978), see Doss-Quinby, Les refrains, 126–33 and 266–74 Her own conclusions about motet centons are in turn queried by Everist, ‘The Refrain Cento’, 175 (n. 40) and 179.Google Scholar

8 There is still no full treatment of the refrain on this basis, the study by Doss-Quinby being primarily of the texts For an attempt to initiate such an enquiry, see Butterfield, ‘Interpolated Lyric’, ch 1, esp. pp. 18–30.Google Scholar

9 Thus although Doss-Quinby begins by considering the ‘caractérisation des refrains indépendamment de leurs contextes’ (Les refrains, ch. 1), she concludes the chapter by commenting on the inadequacy of such a procedure in defining refrains and goes on to concentrate in the rest of her book on refrains within their contexts.Google Scholar

10 Les refrains, 55.Google Scholar

11 ‘The Refrain Cento‘Google Scholar

12 However, refrains are often distinguished visually from their contexts by being supplied with music, blank staves, or blank spaces for staves (on unfinished manuscripts), as well as by the use of enlarged initials and other forms of emphasis. In collaboration with Mark Everist, I am preparing a study of the manuscript layout of all thirteenth- and early fourteenth-century narrative contexts for refrainsGoogle Scholar

13 Les refrains, 55Google Scholar

14 In this respect refrains are unlike proverbs, since their words are often not so much gnomic as clichéd.Google Scholar

15 For extended critical surveys of the use of linguistic models in musical analysis, see Powers, Harold S., ‘Language Models and Musical Thought’, Ethnomusicology, 24 (1980), 1–60, and Benjamin Boretz, ‘Meta-Variations: Studies in the Foundations of Musical Thought, I’, Perspectives of New Music. 8 (1969), 1–74 The application of such a concept as syntax to medieval music is considered by Leo Treitler, ‘Musical Syntax in the Middle Ages. Background to an Aesthetic Problem’, Perspectives of New Music, 4 (1965), 75–85, and Ritva Jonsson and Leo Treitler, ‘Medieval Music and Language. A Reconsideration of the Relationship’, Studies in the History of Music, 1 Music and Language (New York, 1983), 1–23Google Scholar

16 Wilkins, Nigel, The Lyric Works of Adam de la Hale Chansons, Jeux partis, Rondeaux, Motets, Corpus mensurabilis musicae, 44 (Rome, 1967); Hans Tischler, The Montpellier Codex; Jacquemart Giélée, Renart le nouvel, ed Henri Roussel, Société des anciens textes français (Paris, 1961), Le roman de la poire par Tibaut, ed. Christiane Marchello-Nizia, Société des anciens textes français (Paris, 1984) The references to music editions and to Marchello-Nizia respectively supplement and replace those supplied by van den BoogaardGoogle Scholar

17 See van den Boogaard, Rondeaux et refrains, 21Google Scholar

18 Ibid., 20, n 17Google Scholar

19 See lines 6698 and 6840 in Roussel's editionGoogle Scholar

20 See vdB no. 78, where a further motet occurrence is noted (M352)Google Scholar

21 Gennrich's Rondeaux, Virelais und Balladen is the major catalogue of refrain melodies, but he has frequently been castigated for his inaccuracies and inconsistencies (see note 7 above and, especially, Willi Apel, ‘Rondeaux, Virelais and Ballades in French 13th-Century Song’, Journal of the American Musicological Society, 7 (1954), 121–30). In particular (unlike van den Boogaard with the texts) his work cannot be used to consider fine degrees of melodic variation.Google Scholar

22 Lord's highly influential The Singer of Tales (Cambridge, MA, 1960) was itself indebted to two pioneering theses by Parry, L'épithète traditionnelle dans Homère Essai sur un problème de style homérique (Paris, 1928) and Les formules et la métrique d'Homère (Paris, 1928); and to his ‘Studies in the Epic Technique of Oral Verse-Making, I Homer and Homeric Style’, Harvard Studies in Classical Philology, 41 (1930), 73–147 The standard bibliography on the subsequent development of the field is Oral-Formulaic Theory and Research An Introduction and Annotated Bibliography, ed. John Miles Foley (New York and London, 1985) I have consulted, in particular, New Literary History, 3rd quarterly issue (1977), and the studies by Leo Treitler cited elsewhere in these notesGoogle Scholar

23 Wittig, Susan, Stylistic and Narrative Structures in the Middle English Romances (Austin and London, 1978).Google Scholar

24 Treitler's two most important articles are ‘Horner and Gregory: The Transmission of Epic Poetry and Plainchant’, The Musical Quarterly, 60 (1974), 333–72, and ‘Centonate Chant Übles Flickwerk or E pluribus unus?’, Journal of the American Musicological Society, 28 (1975), 1–23 His subsequent work on transmission, and on the relations between music and language, concerns topics which it is not possible to develop here in relation to the refrain, but which would greatly repay further research See note 15 above and Leo Treitler et al., ‘Transmission and Form in Oral Traditions’, International Mustcological Society Report of the Twelfth Congress (Berkeley, 1977) (Kassel, 1981), 139–211Google Scholar

25 ‘Studies in the Epic Technique’, 80Google Scholar

26 Ibid., 118Google Scholar

27 Nagler, Michael, ‘Towards a Generative View of the Oral Formula’, Transactions and Proceedings of the American Philological Association, 98 (1967), 269–311 (p. 270)Google Scholar

28 ‘Homer and Gregory’, 352–68, esp. p 356.Google Scholar

29 Wittig refers to Kenneth Pike, Language in Relation to a Unified Theory of the Structure of Human Behavior (The Hague, 1967), Claude Lévi-Strauss, ‘The Story of Asdiwal’, The Structural Study of Myth and Totemism, ed. Edmund Leach (London, 1967), 1–47, and Lévi-Strauss, The Raw and the Cooked An Introduction to a Science of Mythology, trans. John and Doreen Weightman (New York, 1969). See Wittig, Stylistic and Narrative Structures, 6, n. 6 and n. 7 She sets out her ‘tagmemic approach’ to formulaic analysis in detail in her introduction and ch. 1 (pp. 3–46). My own analysis draws on her work and on Treitler's, but makes no claim to be similarly systematic On structuralist thinking in musicology, see Turnbull, Patricia, ‘Structuralism and Musicology An Overview’, Current Musicology, 27 (1979), 51–64, and Pandora Hopkins, ‘The Homology of Music and Myth Views of Lévi-Strauss on Musical Structure’, Ethnomusicology, 21 (1977), 247–61Google Scholar

30 Wittig, Stylistic and Narrative Structures, 6–9 The search for paradigmatic rules of ‘deep structure’ (a term which originates with Chomsky) forms the third part of Wittig's methodology that is, to ‘go beyond the simple linearity of the surface order to a deeper structural examination in which the infrastructure of formal oppositions is revealed in synchronic organization’ (p. 8)Google Scholar

31 VdB nos. for the refrains are given, where they exist, in the far left column, followed by ‘(T)’ or ‘(—)’ to indicate whether or not they are also noted as refrains by Tischler (in The Montpellier Codex or The Earliest Motets (to circa 1270) A Complete Comparative Edition, 3 vols., New Haven and London, 1982). The far right column gives the contexts, specifying line numbers in the romans, and voice-parts (mot. = motetus; tr = triplum) in the motets If the context is one of the five ((a)–(e)) for vdB 784, this is noted. The texts given are taken from the first context in that column, from the respective editions of each context; additional occurrences (with further variations not noted) are given after ‘+’ Ovide = L'art d'amours Traduction et commentaire de l'Ars amatoria d'Ovide, ed. Bruno Roy (Leiden, 1974), with folio references to Paris, Bibliothèque Nationale, fonds français 881Google Scholar

32 All of these disputed ‘refrains’ are given in roman rather than italics No 20 is included not because I am claiming it to be a refrain, but precisely to show how the formulaic nature of refrains makes them difficult to distinguish from other lines in a text.Google Scholar

33 The transcriptions in the following examples are based on those of Wilkins (for the Adam de la Halle rondeau); Tischler (for the motet occurrences); and Maria Vedder Fowler, ‘Musical Interpolations in Thirteenth- and Fourteenth-Century French Narratives’, 2 vols (Ph D dissertation, Yale University, 1979) (for the Renart le nouvel refrains). The sigla used by Fowler and Roussel for the Renart le nouvel manuscripts differ, I have used Roussel's (F = fonds français 1593; C = fonds français 372; L = fonds français 1581), except in the case of fonds français 25566, which I have referred to as W. References for the motets are given to the numbering in Tischler, Montpellier Codex (= Mo), as well as to GennrichGoogle Scholar

34 The notation in the Renart le nouvel manuscripts is sometimes mensurally differentiated, and sometimes not. the situation is especially complicated in F, where Fowler has identified four musical scribes Neither the rhythm nor the placing of semitones in the examples can be entirely secure, my comments on the music try to take this into account.Google Scholar

35 The shared refrains are vdB nos 156, 289, 430, 496, 746, 784 and 1074.Google Scholar

36 See van den Boogaard, , ‘Jaquemart Giélée et la lyrique de son temps’, Alain de Lille, Gautier de Châtillon, Jakemart Giélée et leur temps, ed. Henri Roussel and François Suard (Lille, 1980), 333–53Google Scholar

37 See Butterfield, , ‘Interpolated Lyric’, 80–2Google Scholar

38 See Everist, , ‘The Refrain Cento’, 181–2.Google Scholar

39 Wittgenstein, Ludwig, Philosophical Investigations, trans. G. E. M. Anscombe (3rd edn, Oxford, 1978), §66, p 32Google Scholar

40 It should be said that in the absence of a fully cross-referenced study of the melodies comparable to van den Boogaard on the texts, it is not easy to check how common this figure is among refrain melodies in general Systematic work needs to be done, for example, on cadential formulas in refrain melodies.Google Scholar

41 Ideally, as well as extending this kind of work on the connections between verbal and melodic formulas in refrains to many other examples from van den Boogaard's inventory, it would also be illuminating to take account of the full tonal context of refrains cited in this (or any other) work and, in particular, to consider how the overall musical structure of a work influences the shape of a refrain melodyGoogle Scholar

42 VdB rondeau 109, Tischler, Montpellier Codex, ii, 159–60.Google Scholar

43 See the conclusions of Everist in 'The Rondeau Motet Paris and Artois in the Thirteenth Century“, Music and Letters, 69 (1988), 1–22 (pp 6–7). The same view had been assumed (though not fully argued) by Bec, La lyrique française, 226–7Google Scholar

44 ‘Centonate Chant’, 12–15.Google Scholar

45 Reproductions of the major illuminations in this manuscript are included in Le roman de la poire, ed. Marchello-Nizia, planches I–XVIII, 159–78.Google Scholar

46 For a stimulating discussion of this manuscript's ‘performative’ characteristics, see Huot, Sylvia, From Song to Book The Poetics of Writing in Old French Lyric and Lyrical Narrative Poetry (Ithaca and London, 1987), ch 6, esp. pp 174–93Google Scholar

47 I have argued this in ‘Interpolated Lyric’, ch. 3, pp. 87–133.Google Scholar