Skip to main content Accessibility help

Melodic dialects in Old Hispanic chant



Aside from Don Randel's study of the responsory verse tones, there have been few comparative analyses of Old Hispanic chant melodies. Such comparison requires new methods because of the paucity of surviving manuscripts, the limited sharing of repertoire between them and the nature of the notation. This article examines variants in specific opening and cadential contexts, across the Old Hispanic corpus. In these contexts, cantors chose from a system of interchangeable melodic shapes, which vary by manuscript. Some manuscripts cluster in their choices of these shapes, in ways that confirm Randel's findings, with four melodic dialects in evidence (‘Leon’, ‘Rioja’, ‘Toledo A’ and ‘Toledo B’). Other manuscripts, however, do not fit securely into any of these four dialects, instead showing a certain degree of permeability between the dialects. Although the types of variants we have identified, including differences in notation and melody, may appear ‘insignificant’ in comparisons of individual chants, they emerge as significant markers of melodic dialects in comparisons of large data sets.



Hide All

1 For a summary introduction, see John A. Emerson (with Jane Bellingham and David Hiley; with Bennett Mitchell Zon), ‘Plainchant, §11: Restoration and Reform in the 19th Century’, Grove Music Online, Oxford Music Online. (accessed 4 September 2014); and Hiley, David, Western Plainchant: A Handbook (Oxford, 1993), 622–9. See also Bergeron, Katherine, Decadent Enchantments: The Revival of Gregorian Chant at Solesmes (Berkeley, 1998).

2 See Randel, Don M., The Responsorial Psalm Tones for the Mozarabic Office (Princeton, 1969).

3 Brou analysed the laudes, establishing melodic types and identifying repeat structures. ‘L'alleluia dans la liturgie Mozarabe: Étude liturgico-musicale d'après les manuscrits de chant’, Anuario musical, 6 (1951), 3–90. Brou's ‘Le Joyau des antiphonaires latins’, Archivios leoneses, 8 (1954), 7–114, establishes stylistic traits and observes formulas in several chant genres. Randel's Responsorial Psalm Tones remains a seminal work of melodic analysis of Old Hispanic chant, and more recent articles such as ‘Responsorial Psalmody in the Mozarabic Rite’, Études grégoriennes, 10 (1989), 87–116 and ‘Las formas musicales del canto viejo-hispánico’, in El canto mozárabe y su entorno: Estudios sobre la música de la liturgia viejo hispánica, ed. Ismael Fernández de la Cuesta et al. (Madrid, 2013), 83–94 have broadened our understanding of musical process in Old Hispanic chant. Zapke has studied aspects of melodic structure and form in the fragment Zaragosa M-418 in El antifonario de San Juan de la Peña (siglos X-XI). Estudio litúrgico-musical del rito hispanico (Zaragoza, 1995). Nadeau undertook significant analysis of the chants in Silos 4 in ‘Pro sonorum diversitate vel novitate: The Singing of Scripture in the Hispano-Visigothic Votive Masses’, Ph.D. diss., Cornell University (1998). Hiley analysed the cadences of the Old Hispanic responsories in ‘Office Responsories in the León Antiphoner: Are they all “Original” Melodies?’, in El canto mozárabe y su entorno, 405–12; Hornby and Maloy analysed certain genres of the Lenten Masses in Music and Meaning in Old Hispanic Lenten Chants: Threni, Psalmi, and Easter Vigil Canticles (Woodbridge, 2013).

4 BN10, Silos 3, Silos 4, Silos 5, Sal, Sant, T3, T4 and T7 have such colophons. For details, see Appendix.

5 The most striking example is T6, an eleventh-century manuscript whose text script is characteristic of the Toledo area, but some of whose multiple scribes used vertical notation. For an argument that vertical notation was used in the south until the end of the eleventh century, see Zapke, Susana, ‘Dating Neumes According to their Morphology: The Corpus of Toledo’, The Calligraphy of Medieval Music, ed. Haines, John (Turnhout, 2011), 91–9.

6 The shorter version of the Appendix may be found at the end of the printed article. A longer version, with full bibliography, is available at at ‘Article’.

7 The possibility that L8’s melodic dialect originated in the Islamic south was raised (though not ultimately accepted) by Randel, based on documented immigration from the south and on the use of a southern model for some of L8’s prefatory material. On this southern model, dated to 806 (based on the computus), see Wagner, Peter, ‘Der mozarabische Kirchengesang und seine Überlieferung’, Spanische Forschungen der Görresgesellschaft, 1 (1928), 102–42, at 102; Anglès, Higini, ‘La música medieval en Toledo hasta el siglo XI’, Gesammelte Aufsätze zur Kulturgeschichte Spaniens 7, ed. Honecker, M., Schreiber, Georg and Finke, H. (Münster, 1938), 1–68; Cordoliani, A., ‘Les textes et figures de comput de l'Antiphonaire de León’, Archivos Leoneses, 8 [1954], 260–83. The model's origin has been placed in Beja: de Urbel, J. Pérez, ‘El Antifonario de León: El escritor y la época’, Archivios Leoneses, 8 (1954), 115–44; and idem, ‘El Antifonario de León y su modelo de Beja’, Bracara Augusta, 22 (1968), 213–25.

8 Randel raises the possibility that this dialect – though not the manuscripts – could have its ultimate origin in Galicia, due to the influx of Galician immigrants, first into Eastern León, then into the Rioja. Responsory Tones, 95–7.

9 See the online version of Appendix at at ‘Article’.

10 It is not universally accepted that these sources represent such liturgical continuity. Zapke has characterised the notation as having a ‘practically symbolic value’ and asked ‘whether the copies kept in the Museo de los Concilios y de la Cultura Visigoda were ever really intended to offer a melodic representation’. See ‘Notation Systems in the Iberian Peninsula: From Spanish Notations to Aquitanian Notation (9th–12th Centuries)’, in Hispania Vetus: Musical-Liturgical Manuscript from Visigothic Origins to the Franco-Roman Transition 9th–12th Centuries, ed. Susana Zapke (Bilbao, 2007), 189–242, at 209. Elsewhere she gives a more positive assessment: ‘the Toledan corpus gives the impression of being a late creation, prompted by political-cultural forces attempting to legitimate and conserve the authenticity of the Visigothic rite originating from Toledo’ (‘Dating Neumes’, 96), also calling for ‘detailed analysis of the script and notation, as well as a comparison of melodies in northern manuscripts’ (‘Dating Neumes’, 97).

11 BN01, T4, T7 and Toledo 35.3 (a liber sacramentorum, the celebrant's book, containing only two notated chants, with annotations linking it to Santa Eulalia). Furthermore, Mundó believed that Cinc and T3 may have been copied by the same hands as T4. ‘La datación de los codices litúrgicos visigóticos toledanos’, Hispania sacra, 18 (1965) 1–24, at 17–18.

12 See Hornby and Maloy, Music and Meaning, ch. 1.

13 Possible theories about the relationship between the two liturgical traditions are given in Pinell, Jordi, ‘El problema de las dos tradiciones del antiguo rito hispánico. Valoración documental de la Tradición B en vistas a una eventual revisión del ordinario de la Misa Mozarábe’, in Liturgía y musica mozarábes (Toledo, 1978), 3–44; and Janini, José, Liber Misticus de Cuaresma (Cod. Toledo 35.2, hoy en Madrid Bibl. nac. 10.110) (Toledo, 1979), xxix–xxx. See also the summary discussion in Hornby and Maloy, Music and Meaning, 303–14.

14 Zapke, ‘Notation Systems’, 192.

15 The Catalan fragments Tarragona, Archivio Histórico Archidiocesano, Fragmento 22/1, and Montserrat, Biblioteca de la Abadía contain the Franco-Roman rite, with Gregorian melodies, as Zapke acknowledges in ‘Coexistencia de signos y funciones en la cultura visigótica escrita: notas marginales’, Études grégoriennes, 40 (2013), 283–91, at 287. However, elsewhere, even within the same article, she claims that these are ‘the two oldest liturgical-musical specimens of the Hispanic rite’. See ‘Coexistencia de signos’, 284, and ‘Dating Neumes’, 91, n. 3. This may lead to some confusion. Catalonia is relevant to a study of Visigothic notation, but not to a discussion of the Old Hispanic chant.

16 The manuscripts in Randel's ‘Rioja’ melodic dialect that Zapke attributes to Navarre are M-418, S3, S6 and S7; those in her Castile-León region include A30, A56, BL51, S3, S4 and S5. See Zapke, ‘Notation Systems’, 198–9.

17 For example, the orational BL52 contains an incipit (sometimes notated) for the antiphon, alleluiaticum or responsory that precedes each office oration, but no other office chants and no Mass chants.

18 The formulaic threni, for example, survive in only two manuscripts, L8 and T5. Their texts have been compared in Jordi Pinell, ‘El cantos de los treni en las misas cuaresmales de la antigua liturgia hispanica’, Eulogia miscellanea liturgica in onore di p. Burkhard Neunheuser O.S.B., Analecta Liturgica 1, Studia Anselmiana, 68 (Rome, 1979), 317–65; in Brou, Louis, ‘Le psallendum de la messe et les chants connexes’, Ephemerides liturgicae, 61 (1947), 1354; and their melodies in Hornby and Maloy, Music and Meaning, ch. 2.

19 See, for example, Le Graduel Romain, vol. 4 (Solesmes, 1962), and the massive variant tables preserved in the Abbey's atelier.

20 See, in particular, Karp, Theodore, Aspects of Orality and Formularity in Gregorian Chant (Evanston, IL, 1998).

21 Pfisterer, Andreas, Cantilena Romana: Untersuchungen zur Überlieferung des gregorianischen Chorals (Paderborn, 2002). Hughes, David G., ‘Evidence for the Traditional View of the Transmission of Gregorian Chant’, Journal of the American Musicological Society, 40 (1987), 377404.

22 These include: Hughes, David G., ‘From the Advent Project to the Late Middle Ages: Some Issues of Transmission’, in Western Plainchant in the First Millennium: Studies in the Medieval Liturgy and its Music, ed. Gallagher, Sean et al. (Aldershot, 2003), 181–98; idem,‘The Implications of Variants for Chant Transmission’, in De musica et cantu. Studien zur Geschichte der Kirchenmusik und der Oper. Helmut Hucke zum 60. Geburtstag, ed. Peter Cahn and Ann-Katrin Heimer (Hildesheim, 1993), 65–73; Manuel Pedro Ferreira‚ ‘Music at Cluny: The Tradition of Gregorian Chant for the Proper of the Mass. Melodic Variants and Microtonal Nuances’, Ph.D. diss., Princeton University (1997); Hiley, David, ‘The Norman Chant Traditions – Normandy, Britain, Sicily’, Proceedings of the Royal Musical Association, 107 (1980–1), 133. For a methodological perspective, see Dobszay, László, ‘Antiphon Variants and Chant Transmission’, Studia Musicologica Academiae Scientiarum Hungaricae, 45 (2004), 6793.

23 The first note of each neume is of indeterminate pitch relative to the previous note (N=neutral); notes within neumes can be identified as higher than, the same as, or lower than the previous note (H, S and L, respectively).

25 The custom of not notating certain syllabic and unison passages in some Old Hispanic sources is discussed in Brou, ‘Notes de paléographie musicale mozarabe’, Anuario musical, 10 (1955), 23–9. One of the exceptions to the tendency to use the single note, T5’s version of Sollemnem habeatis, is written in a different hand than the rest of the sacrificia. In comparison to the other sacrificia, it is much closer melodically to the only surviving earlier version of the melody (L8).

26 This version also occurs, in other genres, in the early fragment PB99, at the end of the soni Custodi me and Custodi animam, each time on ‘alleluia’.

28 The different layers of scribal activity in Silos 6 are discussed in Ismael Fernández de la Cuesta, ‘El “Breviarium Gothicum” de Silos’, Hispania sacra, 17 (1964), 393–494; and the melodic preferences of the scribes are discussed in Randel, Responsory Tones, 75–6.

29 See the full presentation of the data in online Appendix 2b.

30 See online Appendix 2b. In L8, NH comes on a proparoxytone only two times, and NH never comes on a proparoxytone in a ‘Rioja’ source.

31 In the ongoing work for her Ph.D. thesis, Raquel Rojo-Carrillo has identified a less common paroxytonic version of this cadence that occurs in other genres, though not in the sacrificia. Here the last three syllables have N/NHH/N(H), and they are often preceded by a descent.

32 S6 has it only once, in Formavit.

33 Of the forty-two occurrences of NH+NL+N(H) in L8’s sacrificia, seven have corresponding repertory in the ‘Rioja’ tradition (Ecce ostendit in A30, Haec dicit . . . formans in A30, Haec . . . dabo in BL45 and M418, Omnis qui me in BL45, and Elegit in BL45). Six of these have N/NHL/N(H), and the other has a different cadence type.

34 The other cases are Audi Israhel quia magna, ‘dominum’; Omnis populus, ‘aureis’; Stetit, ‘aureum’; Audi israhel preceptum, ‘temporis’; Aedificavit Moyses tabernaculum, ‘foederis’; Elegit, ‘scientiae’; Vos qui, ‘tetigit’; Locutus est ecce, ‘sapientiae’.

35 Data est lex, ‘infantulus’.

37 As online Appendix 3a shows, this contour rule is followed, with one exception (Alleluia miserere in Sal) among the responsories.

38 The clamores are based on a single melody. See online Appendix 3a.

40 In the other case, the responsory Miserere mei, BL51 has a textual variant (‘deus’ instead of ‘dominus’), with NHL on the penultimate syllable.

41 Randel, Responsory Tones, 75–6.

42 On the interpretation of the NS element of these neumes as involving unisons, by analogy with other early medieval notational systems, see Herminio Gonzalez Barrionuevo, ‘Relación entre la notación “mozárabe” de tipo vertical’, Studio gregorani, 11 (1995), 68–71. Although each unison opening type is gesturally unified, we do not assume that these five neume combinations referred to five melodies, each with a consistent pitch content. Randel makes a cogent argument for not interpreting identical neumes as indicating identical pitches, by comparison with the Gregorian repertory (‘Las formas musicales del canto viejo-hispánico’). Even within the Old Hispanic repertory, the pitch-readable chants in A56 furnish several examples of repeated formulas that occur at different pitch levels within the same chant.

43 At at ‘Article’, we have made a comprehensive survey of all manuscripts included, as well as of the Coimbra and Lamego fragments, Silos fragment 26, and León fragment F-5. Although some manuscripts have insufficient chants beginning with these shapes to permit claims about their melodic tendencies, we have included them in our data tables for the sake of completeness.

44 See, for example, Barrionuevo, González, ‘Algunos rasgos fundamentals de la notación “mozárabe” del norte’, Revista de Musicología, 20 (1997), 38–49, at 43.

45 Barrionuevo, González, ‘La notación del antifonario de León’, in El canto mozárabe y su entorno: Estudios sobre la música de la liturgia Viejo hispánica, ed. de la Cuesta, Ismael Fernandez et al. (Madrid, 2013), 95120, at 113.

46 Zapke, Susana, El Antifonario de San Juan de la Peña (Siglos X-XI): Estudio litúrgico-musical del rito hispano (Zaragoza, 1995), 31. It is contradicted by Barrionuevo, Herminio Gonzalez, ‘Relación entre la notación “mozárabe” de tipo vertical y otras escrituras neumáticas’, Studi gregoriani, 11 (1995), 5112, at 70–1. There is a sign – of unknown meaning – that sometimes appears underneath a unison in L8, as on f. 37v: pointing down to the right, it is usually made with a finer penstroke than the neumes, and is sometimes in a different ink. It is almost always found below and to the right of a steeply angled virga preceding a unison and seems unlikely to indicate a note. It is graphically differentiated from the horizontal tractulus interpreted by Zapke as indicating a rhythmic lengthening rather than a note, but which we have interpreted here as a separate note. Our thanks to Elsa De Luca for fruitful conversation about this graphic symbol.

47 Plus Cinc, whose provenance is less certain.

48 Plus the tradition A manuscript T3.

49 González Barrionuevo, ‘Relación entre la notación “mozárabe” de tipo vertical’, 70–1.

50 Two of three chants; the other has N-NH < texmath/ >. This shape appears rarely, as an alternative to a unison opening, across five manuscripts (L8, BL45, BL51, Silos 3 and Silos 4). It is found in total eleven times across the 1,294 instances of unison openings in the sample (with each chant in each manuscript counting as an instance) – that is, less than 1 per cent of the time. The notation is ambiguous here; it might signal NSH or NHH (see González Barrionuevo, ‘Relación entre la notación “mozárabe” de tipo vertical’, 53).

51 One of three chants; the others have NH and NS.

52 Six of eight chants.

53 The presence of NH-NSH in one manuscript when another has Table 4, no. 17 further supports our interpretation of the horizontal line as signalling a note; it seems more likely to be a variant between N-NSH and NH-NSH (one note different) rather than between a lengthened NSH and NH-NSH.

54 Asencio, Ruiz, ‘Códices pirenaicos y riojanos en la biblioteca de Silos en el siglo XI’, in Silos. Un milenio. Actas del Congreso Internacional sobre la Abadía de Santos Domingo de Silos, Vol. 2: Historia, ed. Fernández Flórez, José A. (Silos, 2003), 177210.

55 Zapke, Hispania Vetus, 300.

56 Mundó, ‘La datación’, 19; José Janini, Ramón Gonzálvez and Anscario M. Mundó, Catálogo de los manuscritos litúrgicos de la cathedral de Toledo (Toledo, 1977), 102.

57 Old Hispanic material is not found in Zapke's other two regions, Aragon and Catalonia. See above, note 15.

58 The sigla are derived from Don Randel, An Index to the Chant of the Mozarabic Rite (Princeton, 1973), except for: HSA (which was still lost when his index was made); L8 rather than his ‘AL’ (L8 combines the manuscript's current location with its shelfmark); and M-418 (thought, when Randel made his index, to have been copied at the monastery of San Juan de la Peña, hence his siglum SJP). All British Library manuscripts have been labelled BL rather than Randel's BM (since they have been moved from the British Museum to the British Library). Only manuscripts mentioned in the main text of the article are included here. For detailed bibliographic apparatus, see the online version of this Appendix at

This article was researched and written collaboratively. Both authors had an equal role in the generation of ideas and in the drafting of the article. We gratefully acknowledge the financial support of the European Research Council.

Related content

Powered by UNSILO

Melodic dialects in Old Hispanic chant



Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed.