Andalusian Music and the Cantigas de Santa Maria

7/28/2019 Andalusian Music and the Cantigas de Santa Maria

1/12

ANDALUSIAN MUSIC AND THE CANTIGAS DE SANTA MARIA

Manuel Pedro Ferreira

When the question of the relationship between the Cantigas de Santa Mariaand MedievalHispano-arabic music is raised, the shadow of Julin Ribera's partial musical edition of theCantigas, published in 1922, cannot be avoided1. In Ribera's edition, the Alfonsine songsare presented as derivatives of classical Arab music, and transcribed according to whatRibera thought was typically Arab; as a consequence, their original notation was oftendisregarded. Ribera's approach came under heavy criticism from professional musicologists,amongst them Higinio Angls, who in his monumental work of 1943-58 buried &endash;seemingly for good &endash; the scholarly pretensions to read Arab music into the

Cantigas2. In the past half-century (1943-1993), the pro-Arab stance has therefore beenconfined to the performing domain as a kind of colouristic exoticism, of doubtful historicalseriousness, which is sometimes made vaguely respectable through mention of the Islamicinstruments depicted in one of the manuscript sources of the Cantigas3. This tendency to

value instrumental colour can be explained not only on the basis of Ribera's claim that therepertory has an "orchestral" character4, but also in relation to the history of the modern"early music" movement; in actual practice, instrumental colour has been served as a kindof dressing added to Angls's transcriptions, which have been generally accepted by theperformers.

A lot has none the less changed after Ribera's edition in our knowledge of Medieval Arab

music; and our understanding of the original notation of the Cantigashas progressed in thepast few years. The time has come to re-evaluate the Arab question from a scholarly pointof view.

The first thing to do is to rephrase the question, substituting "Andalusian" for "Arab".Moorish-Andalusian (or Ibero-Arab) music is not just, or even mainly, Arab musicper se. Itis a hybrid Western tradition which evolved independently from oriental trends from theninth-century onwards and reached its highest level of integration of Western and orientalelements in the twelfth-century5. The originality of Andalusian music, when compared withother Western Medieval traditions, is to be sought primarily in the aspects of form andrhythm. Form represents the Peninsular indigenous element; rhythm, the Arab one.

The question of musical form in the Medieval Andalusian song has generally been ignored;recently, Vicente Beltrn and, most importantly, David Wulstan, have faced the problemand attempted to give it a solution6. Both take as their starting point the formal structure ofthe poems, to which they remain anchored as their only secure evidence; naturally, sincethe textual data gives minimal musical information, their conclusions cannot be firmlyfounded from a musical point of view. I have therefore taken the opposite approach: start

with the surviving music from Moorish Andalusia (Al-Andalus).

My work was made possible by the recent publication, by James Monroe and BenjamimLiu, of nine survivingazjaland muwashshahatcomposed in Al-Andalus between c. 1100 and

the mid-fourteenth century7. As complementary data, I used the analysis of a representativesample both of today's North African music of Andalusian origin8 and of the muwashshah


2/12

oral tradition in general9. The results of this undertaking, which I will summarize here, willbe published in detail elsewhere10. The main conclusion is that the muwashshahand thesymmetrical zajalseem originally to have had mainly two kinds of formal scheme,corresponding to the virelai (AA || BBB AA , AB || CCC AB) or to a special kind ofrondeau (AB || BBB AB , [...] AAA BA). The zajalproper, textually assymetrical, must

have had related schemes, for the zajaland the muwashshahare two faces of the sametradition11.

At this juncture, the example of the Cantigas de Santa Mariahas to be taken into account.Poetically speaking, most of them have the zajalform. Musically, they tend to present the

virelai form, which as we have seen is also characteristic of the muwashshah. In itself, thisfact is not conclusive, for the French virelai could be taken as both the poetical and musicalmodel for the Cantigas; the only problem with this hypothesis is that the virelai hardlyexisted in France before c.1300, while the Cantigaswere composed before 1284; this factled Willi Apel to propose a Spanish origin for the virelai12. Moreover, in the Cantigastheinfluence of the French rondeau is slight when compared with the important presence of

the reverse kind of rondeau (AB || BB AB), also characteristic, mutatis mutandis, of themuwashshah; this is found in more than seventyCantigas13. Since this last form is virtuallyunknown elsewhere in Medieval Europe14, it is probably indigenous; and since the Cantigas

were mostly composed in a cultural environment where the Ibero-Arab presence wasstrongly felt, it probably derives from the zajalor its mozarabic counterpart. The Cantigas deSanta Mariaappear therefore, from a formal point of view, to encapsule typical features ofmedieval Andalusian music: the virelai form and what I propose to call the Andalusianrondeau.

Let us now turn our attention to rhythm. Rhythm is intrinsically linked with the musicalnotation of the manuscripts. The notation has been variously described by different

authors, depending on the interpretative model used to approach it. Hendrik Van der Werf,for instance, confronted the notation with the late-thirteenth-century Franconian system,and inevitably concluded that the alfonsine notation is not Franconian15, which is hardlysurprising since this system was formulated in writing only around 1280, when most of theCantigaswere in the process of being copied16. It does not follow, though, that the alfonsinenotation lacks a mensural character, for there were mensural systems in existence beforeFranco of Cologne. On the contrary, I think that it can be proved that the mensuraldimension is an important one &endash; I have dealt with this problem elsewhere17&endash; regardless of how we choose to interpret it.

Interpretation is about ways to make the data historically intelligible. Angls was right when

he accepted the notation as it stands without trying to force it into preconceived moulds, asRibera has done; he also realized that the rhythms written down by the copyists were oftenequivalent to the contemporary French patterns known as "rhythmic modes", but that this

was not always true. Unable to accept Ribera's hypothesis of an Arab derivation, hechampioned the theory of a folkloric origin for the cases of non-modal rhythm; needless tosay, the "folk music" label could embrace everything, and because of this generality couldbe neither proven nor challenged; it was an easy way out of the problem. In my own work,I have expanded the framework of possible preexisting models: French developments ofmodal rhythm, troubadouresque isosyllabism and the rhapsodic rhythm found in thecantigas d'amigo; since even this large range of possibilities does not exhaust the rhythmic

variety found in the repertory, I had eventually to confront the long-discredited hypothesis

of an Andalusian connection.


3/12

According to one of the leading specialists in Arab music, Baron Rodolphe d'Erlanger, "lerythme est, en musique arabe, l'lment principal et prpondrant de toute composition

vocale ou instrumentale". He also remarks that the rhythmic system used by Arabmusicians today is substantially the same as it was during the first centuries of Islam18. Thissystem is based on the principle of periodicity: the repetition of a rhythmic period defined

by the number and quality of the attacks and the time elapsing between them. This time isstrictly measured, meaning that it is counted in units of time. Among the ancient musictheorists, Al-Farabi ( 950) is the only one who tries to describe the actual musical practice,instead of following Greek music theory19; he eschews the Greek definition of the basictime-unit as the shortest perceptible time value, choosing instead as time-unit a compoundtime, as Arab musicians do today20. According to Al-Farabi, a rhythmic period is typicallycomposed of two identical rhythmic cycles. A cycle is a repeated rhythmic patternsuperimposed on a given meter. From an abstract point of view, each cycle has a basicform in which all the attacks are separated by equal time-intervals, and the last attack isfollowed by a silence of the same length (the disjunction). In actual practice, this basicscheme gives way to more complex rhythmic patterns which have the status of standard

metric fillings. These metric fillings can be varied over a wide range, and two differentvariants can be joined together in a period. Al-Farabi himself lists a large number ofrhythmic periods derived from each of the seven basic meters, and describes theconventional variation procedures which lead to them; his list is not exhaustive; he simply

wants to show how these variation procedures work in practice21.

In the following examples, the spacing between two commas ( ' ' ) illustrates the minimumtime-unit; if an audible attack marks the beginning of a time-unit, it will be represented ( | '

); time signatures will be used for convenience, the minimum time-unit being equivalent toa quaver.

When the chosen meter is the "First Thaqil" (or "First-Heavy"),

(a) (4/2) | ' ' ' | ' ' ' | ' ' ' ' ' ' '

doubling of the attacks will produce the following pattern;

(b) (4/2) | ' | ' | ' | ' | ' | ' ' ' ' '

adding a loud attack to allow a proportional disjunction will change it into:

(c) (4/2) | ' | ' | ' | ' | ' | ' | ' ' '

If we reproduce this cycle twice in a row, we have one of the forms of the rhythmic period"First-Heavy" listed by Al-Farabi.

Another example is the "Heavy-Ramal" meter,

(a) (3/2) | ' ' ' | ' ' ' ' ' ' '

with another attack added for continuity, this changes into

(b) (3/2) | ' ' ' | ' ' ' | ' ' '

and with doubling of the second attack, it becomes


4/12

(c) (3/2) | ' ' ' | ' | ' | ' ' '

which corresponds to another form of the cycle mentioned by Al-Farabi.

Variation can also produce syncopation: if we take the continuous pattern, double all

attacks and then drop out the fifth, the result is

(d) (3/2) | ' | ' | ' | ' ' ' | '

This is one of several syncopated cycles in Al-Farabi's list. If repeated once, we have ahomogeneous rhythmic period; if combined, for instance, with the continuous double-attack cycle, we have an heterogeneous variant also listed in Al-Farabi; if combined withthe basic "Ramal" cycle, we have the rhythm al-hafif, used in a thirteenth-century

Andalusian composition which survives in today's oral tradition22.

It is important not to forget that the musical tradition that Al-Farabi describes travelled

West from Badgdad to the Andalus, where it found fertile ground. Furthermore, a numberof features distinguish Arab rhythmic periodicity from medieval Western Europeanrhythmic tradition: the larger scale of some cycles and rhythmic periods, the use ofsyncopation and the importance of quadruple meter may be mentioned. This means that

when we find a medieval Spanish repertory, written for the most part probably in Toledoor Seville (next door to a Moorish-Andalusian environment), and using large-scale cycles orperiods, with syncopated patterns or in quadruple metre, we may reasonably conclude thatit reflects the influence of Arab music.

The Cantigas de Santa Mariaare such a repertory. In the "Heavy-Ramal" meter, thecombination of variants (c) and (d) listed above produces the rhythmic period found in

CSM92 (ex. 1). If we take the above-mentioned form (c) of the cycle "First-Heavy" anddouble the second attack, we encounter a variant found in CSM42423 . If, in the secondpresentation of this rhythmic variant, we add a final attack for support, as recommended by

Al-Farabi, we will have a long rhythmic period identical to that found in CSM25 (ex. 2)24.The long rhythmical period which begins CSM100 has two versions which differ in thesecond half (ex. 3); the initial version survives in the rhythm al-Btayhiof the Andalusiantradition (ex. 4)25; both versions can be described as heterogeneous periods made up of twoof the "First-Heavy" cycles listed in Al-Farabi26. In this same song, there is anotherheterogeneous rhythmic period which shares its second half with the second version of thefirst period (ex. 5). The first half presents a cycle that is another variation on the "First-Heavy" meter27, and is found in the rhythm al-qa'im wa-nisfof the Andalusian tradition (ex.

6)28. CSM353 uses exclusively this same cycle. CSM116 uses a related rhythmic period,made up of this same cycle followed by the basic form of the "First-Heavy" meter (ex. 7).This period is strikingly similar to that found in two sister-compositions by Juan delEncina, "Seora de hermosura" and "Una saosa porfa", which share the same melodicopenings29; and is reproduced almost exactly in the first version of "Norabuena vengas" inthe Cancionero de Palacio30. CSM109 exhibits a more complex period based on the samecycle, produced by repeating part of it in the middle of the period &endash; a variationprocedure also mentioned by Al-Farabi (ex. 8).

The medieval French rhythmical theory and the alternative models mentioned above areunable to explain these seemingly anomalous facts, whereas they make complete sense in

the light of Arabian rhythmic theory and its influence on Andalusian song. Given the


5/12

historical context, one cannot but reach the conclusion that, at least, the above-mentionedCantigaswere influenced by Ibero-Arab music.

That being so, perhaps there are other traces of this influence. Again, the first thing to do isto look more closely at what seems to be a rhythmic anomaly from a French-centered

perspective: the dotted rhythm, which is impossible to write within the normal usage ofthirteenth-century French notational systems. It can be observed in seventeen of theCantigas31; in two of them it is used to the exclusion of any other rhythmic pattern32. The

way the Toledo and Escorial MSS solve the notational problem is similar: they add a brevisto the long, and then write an isolated brevis; or they use vertical lines after the long tosignal its ultra mensuramquality, and then write an isolated brevis (Ex. 9)33. The rhythmicalmeaning of these procedures is clear from the different ways the scribes chose to writedown the same musical idea, whether in the same manuscript, when a phrase is rewrittenseveral times, or in different manuscripts which have the same song; comparative workshows that a long with a brevis attached to it is rhythmically equivalent to a long followedby a double bar; it also shows that this augmented long is equivalent to a long followed by a

ligature cum opposita proprietate, or a binary oblique ligature followed by a brevis or a doublebar34. Sometimes the Escorial MSS substitute what seems to be a semibrevis for the brevis35,but this can easily be explained as a case of notational inertia &endash; forms of the Toledonotation which are reproduced without translation in the Escorial notation (ex. 10).

The important presence of the dotted rhythm in this repertory, given that it is ignored inthe surviving Galician-Portuguese troubadour songs and in all the remaining writtenEuropean music, can be explained through the influence of the Andalusian tradition. Wehave seen that one of the Andalusian rhythmic cycles uses dotted rhythm; in the MiddleEast, it is also found in the rhythm Sufiyan36; both derive from classical Arab rhythmicpractice. In some of the surviving Medieval Andalusian songs37, dotted rhythm is pervasive:

it tends to be associated with the sucessive occurrence of a long and a short syllable (ex.11). This probably means that dotted rhythm was a standard declamation procedure inIbero-Arab song, and that it may have influenced the composers of the Cantigas.

Another feature of the Andalusian tradition is the use of a five-beat metric pattern alreadylisted in Al-Farabi. Among the seven basic musical meters acknowledged by this theorist,three have five beats per cycle; each of them has a variant which is similar to the French3rd rhythmic mode, except that the first long has only two units of time instead of three(ex. 12). This rhythmic pattern surfaces in a Hispano-Arab song which has been identifiedas a muwashshahand was partly transcribed, in the sixteenth-century, by Francisco Salinas(ex. 13a)38. The pattern's influence on folk music is attested to by several traditional songs

which reached us in polyphonic settings by Encina, Anchieta and others; its survival maybe illustrated here by the song "Tan buen ganadico" as transcribed by Juan del Encina (Ex.13b)39. It can also be found in CSM 223 &endash; alternative interpretations of the notationleading, in my view, to unsatisfactory results (ex. 13c).

This last case may not be the only one. It happens, on the one hand, that some melodies(Prologue, ns 10 and 105) or isolated phrases (cf. ns 38, 41) in the Cantigas de Santa Mariaare notated in such a way that both the five-beat and the six-beat transcription is possible.On the other hand, CSM339 has a phrase which is clearly reminiscent, from both a melicand a rhythmic point of view, of the Ibero-Arab song quoted by Salinas (ex. 14); itsnotation indicates the third rhythmic mode, which implies a six-beat meter instead of a

five-beat one; this suggests that the use of the third rhythmic mode could, in some cases,


6/12

be seen as a rhythmic variant based on the "Light-Ramal" meter, or indicate a notationaladaptation of an original five-beat pattern.

Although the presence of the five-beat meter in the Cantigascannot be proven withabsolute certainty due to notational ambiguity, the important presence in this repertory of

Andalusian forms and Arabic rhythmic features makes it historically plausible, and helps toexplain the relatively generous use of the third rhythmic mode by Alfonso's collaborators.From this point of view, the preponderance of the second rhythmic mode over the first inthe Cantigas, especially in the Toledo MS, could also derive from the coincidence between,on the one hand, the French second mode and, on the other, the fundamental form of the

Arab "Light-Ramal" meter.

In short: although Angls rightly identified a strong French flavor in the Marian Cantigas,Ribera was also justified in pointing out its debt towards the Andalus. To these importantinfluences one could add those of liturgical music, the troubadours and the Galician-Portuguese love song. We have to conclude that this extraordinary Marian collection

juxtaposes and combines a number of musical styles which we are just beginning toidentify.


7/12

\


8/12


9/12


10/12

1. Julin Ribera, La msica de las Cantigas. Estudio sobre su origen y naturaleza, conreproducciones fotogrficas del texto y transcripcin moderna, Madrid: La Real

Academia Espaola, 1922 (Vol. III of Cantigas de Santa Mara de Don Alfonso el

Sabio , published by the same Academy; volumes I and II appeared in 1889).2. Higinio Angls, La msica de las Cantigas de Santa Mara del Rey Alfonso el Sabio,3 vols., Barcelona, Biblioteca Central, 1943, 1958, 1964 (the last-published volumeis a facsimile edition of MS E).

3. MS E (Library of El Escorial, ms. b.I.2).4. Ribera, op. cit., p. 117: "siendo todas las melodas de las Cantigas destinadas a

ejecucin por varias voces y por orquestra numerosa...".5. For a historical summary, see the Appendix in Harvey L. Sharrer and Manuel Pedro

Ferreira, Cantus Coronatus (forthcoming).6. Vicente Beltrn, "De zjeles y dansas: orgenes y formacin de la estrofa con

vuelta", in Revista de Filologa Espaola, 64 (1984), pp. 239-66; David Wulstan,

"The Muwashshah and Zagal revisited", in Journal of the American OrientalSociety, 102 (1982), pp. 247-64.7. Benjamin M. Liu e James T. Monroe, Ten Hispano-Arabic Strophic Songs in the

Modern Oral Tradition &emdash; Music and Texts, Berkeley: University ofCalifornia Press, 1989.

8. Leo J. Plenckers, "Les rapports entre le muwashshah algrien et le virelai du moyenge", in I. A. El-Sheikh, C. A. Van de Koppel e R. Peters (eds.), The Challenge ofthe Middle East: Middle Eastern Studies at the University of Amsterdam,

Amsterdam: University of Amsterdam, 1982, pp. 91-111; Jozef M. Pacholczyk,"The Relationship Between the Nawba of Morocco and the Music of the

Troubadours and Trouvres", in The World of Music, 25 (1983), pp. 5-16;

id.,"Rapporti fra le forme musicali della nawba andalusa dell'Africa settentrionale ele forme codificate della musica medievale europea", in Culture musicali: quadernidi etnomusicologia, 3:5-6 (1984), 19-42. To the data presented in these articlessome more analytical information was added, based on Moroccan sources.

9. Lois Ibsen al Faruqi, "Muwashshah: A Vocal Form in Islamic Culture", inEthnomusicology, 19 (1975), pp. 1-29.

10.Sharrer and Ferreira, op. cit.11.According to the traditional view, the former derives from the latter, but the

reverse seems now to be more likely: Wulstan, op. cit.; Samuel G. Armistead andJames T. Monroe, "Beached Whales and Roaring Mice: Additional Remarks onHispano-Arabic Strophic Poetry", in La Cornica, 13 (1985), pp. 206-42.

12.Willi Apel, "Rondeaux, Virelais, and Ballades in French 13th-Century Song", inJournal of the American Musicological Society, 7 (1954), pp. 121-30.13.This calculation is based on the tables published by Angls, op. cit., Vol. III/1

Parte, pp. 397-400.14.Friedrich Gennrich, Grundriss einer Formenlehre des mittelalterlichen Liedes als

Grundlage einer musikalischen Formenlehre des Liedes, Halle: Max Niemeyer,1932, pp. 67-68.

15.Hendrik Van der Werf, "Accentuation and Duration in the Music of the Cantigasde Santa Maria", in Israel J. Katz and John E. Keller (eds.), Studies on the Cantigasde Santa Maria: Art, Music, and Poetry, Madison: The Hispanic Seminary ofMedieval Studies, 1987, pp. 223-34.

16.The originals which underlay the final compilation of the Cantigas were mostlywritten (i. e. between two hundred and fifty and three hundred pieces) before 1280.


11/12

The collection is presumed to have been completed or nearly so by the timeAlfonso died (1284). On the dating of the manuscripts, see Manuel Pedro Ferreira,"The Stemma of the Marian Cantigas: Philological and Musical Evidence", inCantigueiros, VI (1994), pp. 58-98.

17.Manuel Pedro Ferreira, O Som de Martin Codax: Sobre a Dimenso Musical daLrica Galego-Portuguesa (Sculos XII-XIV)/The Sound of Martin Codax: On theMusical Dimension of the Galician-Portuguese Lyric (XII-XIV Centuries), Lisboa:Imprensa Nacional - Casa da Moeda, 1986; id., "Bases for Transcription: GregorianChant and the Notation of the Cantigas de Santa Maria", in Jos Lopez-Calo(coord.), Los instrumentos del Prtico de la Gloria. Su reconstruccin y la msicade su tiempo, La Corua: Fundacin Pedro Barri de la Maza, 1993, Vol. 2, pp.573-621.

18.Baron Rodolphe d'Erlanger, La Musique Arabe, Tome VI, Paris: Paul Geuthner,1959, pp. 1, 4.

19.George Dimitri Sawa, Music Performance Practice in the Early Abbasid Era 132-320 AH / 750-932 AD, Toronto: Pontifical Institute of Mediaeval Studies, 1989, p.

16.20.D'Erlanger, op. cit., p. 7.21.Sawa, op. cit., pp. 46, 54.22.Liu and Monroe, op. cit., p. 82.23.In Angls's edition, this song is the second in the second Appendix [FJC 2]; its

form belongs to the "Andalusian rondeau" type.24.This is true of the version recorded in the Escorial codices E and T, not of the

version in To.25.D'Erlanger, op. cit., p. 148.26.Sawa, op. cit.,"First-Heavy" cycles ns 11+3 and 11+9.27.This variant is arrived at by adding an attack for continuity, doubling this attack,

and dropping out the first articulation.28.Liu and Monroe, op. cit., p. 82.29.Juan del Encina, Poesa Lrica y Cancionero Musical, ed. R. O. Jones and Carolyn

R. Lee, Madrid: Castalia, 1972.30.The Cancionero de Palacio shows a few striking continuities with the CSM: for

instance, the rhythmic pattern minim-crotchet, minim-crotchet, crotchet-minim,minim-crotchet (or dotted minim), which often reccurs in this repertory, canalready be found in at least ten CSM (34, 46, 104, 199, 232, 295=388, 300, 328, 345and 398).

31.CSM 1, 26, 37, 47, 51, 61, 88, 89, 101, 109, 112, 116, 118, 158, 193, 353 and 393.See also CSM 100, 315 and 352.

32.CSM 118 and 393.33.The double vertical line may also be used at the end of a musical phrase or piece,

with no apparent rythmical consequences (see CSM 123, 159, 160, 341, 386 and394). The Cantigas 88 and 116 use a long with a double vertical bar to mean eitherlong plus brevis, when followed by a brevis, or double long, when followed by along (in the CSM 88, the Toledo MS makes it clear that in the latter case theaugmentation applies to a three-tempora long).

34.See CSM 1, 47, 51, 89, 116 and 393.35.CSM 37, 47, 193 and 353.36.D'Erlanger, op. cit., p. 53.37.Liu and Monroe, op. cit., songs I, III, V (ocasionally in other compositions).38.

Francisco Salinas, De Musica Libri Septem, Salamanca, 1577. It is the song Calvi vicalvi / Calvi aravi ("My heart is in [another] heart / [because] my heart is arabic"),


12/12

quoted by Gil Vicente in both the Comdia de Rubena and the Tragicomedia deDon Duardos. See Emlio Garca Gmez, "La cancin famosa Calvi vi calvi / Calviaravi", in Al-Andalus, XXI (1956), pp. 1-18, 215-16, and Juan Jos Rey, DanzasCantadas en el Renacimiento Espaol, Madrid: Sociedad Espaola de Musicologia,1978, pp. 25-26. Salinas's musical quotation was wrongly transcribed (in 6/8) by

Higinio Angls, op. cit., Vol. III/2 Parte, p. 440.39.Juan del Encina, op. cit., pp. 45, 294, 354 (see also the commentary by ManuelPedro Ferreira, in Cancioneiro da Biblioteca Pblia Hortnsia de Elvas, Lisboa:Instituto Portugus do Patrimnio Cultural, 1989, pp. ix-x); Spanish songs inquintuple time are discussed by Marius Schneider, "Studien zur Rhythmik imCancionero de Palacio", Miscelnea en homenaje a Monseor Higinio Angls,Barcelona: C. S. I. C., 1958-1961, vol. II, pp. 833-41, and Juan Jos Rey, op. cit., pp.30-33.

Andalusian Music and the Cantigas de Santa Maria

Documents