Latin American Choral Music - Table of Contents

References Cited

From Manuscript to Performance:
The Metepec Villancicos of Antonio de Salazar

Sherrill B. L. Blodget  

The goal of my session today is to present a performance of three unpublished  villancicos by Mexican Baroque composer Antonio de Salazar (16501715), and to discuss some of the challenges I encountered while preparing performance editions of them.  1 It is my hope that highlighting these issues, and demonstrating our solutions, will shed some light on performance of the Baroque villancico in general. The finished editions will be published through the University of Arizona and made available to choral ensembles for performance.

Introduction to Antonio Salazar and the Metepec Villancicos

Antonio Salazar served as maestro de capilla  at Puebla Cathedral from 1679–1688, and at the Mexico City Cathedral from 1688 until the end of his life.  He was an extremely prolific composer of motets, masses, hymns, Te Deums, and  villancicos, including a number of settings of poetry by Sor Juana Inéz de la Cruz.  Despite his renown, few of his works have yet been published.  The Metepec villancicos, named for the Chapel in which they were found in San Diego, Metepec near Puebla, Mexico, are manuscript copies of three of Salazar’s  villancicos.   The copy dates on the manuscripts are 1771 and 1773, but the original dates of composition have not been determined.  The manuscript copies were originally transcribed in 1984 by Wayne Moss, director of the Texas Baroque Ensemble.   In preparing performance editions of these pieces I used Dr. Moss’s transcriptions as a reference as well as working directly from the manuscript copies.

Issues in Transcribing and Creating Editions

In preparing editions of the villancicos, a number of challenges arose. Some of these are common to any Baroque transcription project, while others are more particular to the Latin American Baroque and the Salazar Manuscripts.  In dealing with the text, there were three areas that presented problems.  First, I found inconsistencies in underlay and spellings. There were a number of instances in the manuscripts in which the text or underlay differed between parts, offering contrasting poetic interpretations and text settings.   Though it is possible that Salazar intended for there to be concurrent multiple texts, it is more likely that the scribe who copied the works, made mistakes. In these cases, I examined each part and picked the one that was most consistent, and which made the most poetic sense.

Second, I noticed the use of old spellings in general. There were many instances of 17th Century Spanish spellings, which differ from modern spellings, at times making it difficult to understand the words. In the editions I used modern Spanish spellings to make them more accessible to the modern performer. Figure 1, which shows four examples from Al salir el sol, demonstrates these issues with differences in the spelling of "espejo" "tributa" and "union," and the use of "brillar"in the tiple voices and "arbol" in the tenor.

Figure 1. Four examples of text settings from   Al Salir el Sol , mm. 28-30; 49-51

Tiple 1:  Son espejo a su brillar con que en dulce union,
            Su fragransia trivuta, trivuta el pensil


Tiple 2:  Sun espexo a su brillar, con que en dulse union,
               En el sol su fragransia trivota trivota el pensil.

Example 2

Tenor:   Son espexo a su arbol, con que dulse  unio, con que dulce,
              Su fragransia tributa tributa el pensil.

Example 3

Alto:   Sonn espexo a su arbol. Con que en dulse unio.

Example 4

Example 5

Perhaps the biggest challenge with the text is the translation. The combination of Old Spanish, allegorical texts, and the inconsistencies between the parts, made translating the works very difficult.   I worked with a number of scholars including Francisco Grijalva, Susan Tattershall, and one of our students, Ernesto Amaya, to create the best translations possible. Clearly it is important for the performers to understand what they are singing, as well as for the audience to know what they are hearing.  The editions will include notes about the meaning of the more metaphorical sections. 

Notation: Musica Ficta and Continuo

As with the text, the manuscripts contained instances of inconsistencies between parts in the use of accidentals, and in some cases harmonies that appeared to be too uncharacteristic of Salazar to be correct.  In these cases I used the context within each piece and a general knowledge of Baroque harmony to determine an acceptable solution. 

As is common to most Baroque  villancicos, the manuscripts contained un-texted bass parts designated as acompañamiento.  None of these contained figures or ficta indicating harmonies for realization.  As Robert Stevenson noted in his “Christmas Music from Baroque Mexico,” the continuo realization would have followed seventeenth century rules, but differed based on the instrument used to play it, for example, guitar, harp, organ. I provided a suggested continuo realization in the editions that would work for organ or harpsichord, the latter most appropriate in Victoria por el Silencio.

Vocal Forces and Range

As mentioned, the Metepec  villancicos contained un-texted bass parts.  However, although an SSAT voicing is common in this repertoire, there are exceptions.  I believe the case must be examined in the context of each piece, which I did in making these decisions for the Metepec villancicos (see my solutions, below).


Another question, which must be addressed either in creating an edition or in performing from one, is whether to transpose the parts. The argument for transposing includes the following two points. First, there is evidence that many of the Baroque Latin American villancicos would have been performed a fourth below where they are notated. The ranges and tessituras of the written parts are often very high, particularly for modern tenor.  Second, if one is doubling the vocal lines with instruments, modern instruments have different ranges and capabilities than the original ones, and may be far easier to play in different keys.

Solutions for the Metepec villancicos

It should be noted that while every effort was made to maintain the integrity of Salazar’s intentions based on the evidence in each score, editorial decisions do reflect the typical make up of the modern SATB choir, in an effort to make performance of these pieces more accessible.  For example:

Al Salir el Sol: Manuscript parts: SSAT and  acompañamiento.
Edition/performance:  SSAT, Acompto; at notated pitch.
Al Coronarse Reyna: Manuscript parts: SAT and  acompañamiento.
Edition/performance:  SATB, Acompto; G Maj—P4 lower than notated.

Victoria Por El Silencio:Manuscript parts: T solo, SATB,  acompañamiento.
Edition/performance: T, SATB, Acompto; G Maj--P4 lower than notated.

Al Salir el Sol
The manuscript parts were for SSAT, with the  continuo part essentially doubling the tenor an octave below.  This doubling was a good indication that the tenor was intended to be the lowest vocal part, and thus I ultimately scored the edition for SSAT.  However, in our performance I  included two basses to give the choral sound more depth. The ranges in  Al Salir were not particularly demanding, and the key was fine for our instrumental accompaniment of recorders and sackbuts. I therefore left the key as the original, G major.

A Coronarse Reyna
The manuscripts were scored for SAT and continuo. In contrast to Al Salir, the continuo part did not double the tenor line, but was fully independent.  Due to this difference, and the fact that SAT pieces were not particularly common in the repertoire, I scored the edition for SATB, placing the text underlay in the bass part as it fit logically. I also made the decision to score  A Coronarse down a perfect fourth in the edition, placing it in G major, largely due to the high tessitura of the original soprano and tenor parts, and to make it possible to double on recorders.  This did cause an issue for the bass part, which then lies very low.

Victoria Por el Silencio
The manuscripts included a specifically designated  Bajo part in addition to the acompañamiento part.  Though it was un-texted, the  Bajo part was not doubled by the  Continuo.  By all indications, it represented a separate bass vocal part. Therefore, I included a bass part in the score, placing the text underlay to correspond with the other parts.  As with  A Coronarse, I transposed Victoria down a perfect fourth due to the high tessituras in all parts.

Performance Issues

My discussion thus far has been directed at decisions that must be made when creating a transcription or an edition from a manuscript. In the majority of cases however, choral ensembles will be working from published editions of one sort or another. I would like now to touch on some performance decisions that will still need to be addressed to bring those editions to performance. 


Other than the possibility of transposing down a perfect fourth, which is in keeping with the noted Spanish tradition, the final decision of whether to transpose will rest with the abilities of the choir and the instrumentation used.  Because pitch was a relative thing in the Baroque conductors may assume some leeway without drastically altering the composer’s original intent.
Vocal Forces

In considering vocal forces, we need to ask if the singers were men, women, boys, or a combination thereof. Many of the New World choirs likely consisted of men and boys, and Stevenson, Russell and others have demonstrated that some villancicos would have been performed in convents. Indeed, some of Salazar’s manuscripts contain the names of singing nuns written on the parts, and he is known to have taught in the Trinity Convent in Puebla (Stevenson 1974). Therefore, the use of women’s voices for the soprano and alto parts is very acceptable.

Despite the Baroque tradition of having one voice per part, there are a number of reasons to assume there may have been more singers performing the   villancicos. For one thing, the alternation of copla and estribillo sections textually suggest a contrast in number of voice parts, which would be heightened by having more than one singer on a part in the estribillo. Secondly, descriptions in the Cathedral records indicate there were fairly large choirs in Puebla and Mexico City, and describe villancico performances as large events and spectacles, both of which suggest more singers were involved. Based on the evidence, I concluded that a smaller ensemble is ideal, with one to a few voices on the  copla lines, and more on the  estribillo sections. 


Were there instruments involved?   Until later in the Baroque, most villancico manuscripts did not indicate specific instrumentation, or obbligato parts. However, literary descriptions of the villancicos, Cathedral pay rosters, Capitular Acts, and musicological evidence clearly suggest the Baroque villancicos were accompanied colla parte or improvisationally by instrumentalists. In a discussion of recordings in "Christmas Music in Baroque Mexico," Stevenson argues that unaccompanied performances are unauthentic and states that "the villancicos of  Salazar...presuppose an orchestral accompaniment, and the vocal parts therefore represent only a skeleton"  (Stevenson 1974: 143)

An extant list of the instruments available in the Mexico City Cathedral in 1691 includes: violin, treble viol, tenor viol, rebeck, bandore, cittern, marine trumpet, harp, clarion, trumpet, trombone (sackbut), double-reed chirimía (shawm), organ, and bajón (bassoon).  Recorders, zither, dulcian, and cornetts have been listed in other documents. Further confirmation of the use of these instruments can be found one of Sor Juana’s villancico texts set by Salazar in 1691. Instrumentalists typically played multiple instruments, and a variety of combinations would likely have been used for different sections of the villancicos.  Stevenson specifically suggested that timbres may have been varied on the coplas.

With modern performances for programs that do not have shawms, dulcians, and viols available, adjustments must be made, keeping the above principals in mind.  In this performance, we are using recorders, sackbuts, rackett, and bass viol as colla parte instruments.   The choice of which to use for which piece was based on the affect of the particular piece. 

A final instrumental decision to be made will be what to use for the continuo group. The most typical Baroque continuo instruments in Mexico and Spain were harp, vihuela, organ, viol, and bassoon. In an article on the villancico in High Baroque Spain, Thomas Taylor suggests that one of the reasons for such portable continuo instruments was that the villancicos were often performed as processionals through the city streets on the way to the Cathedral during feast days (Taylor 1984).

In our performance we use cross-strung harp, organ, harpsichord, and viola da gamba.  While the use of harpsichord is not described as typical, I felt that Victoria por el Silencio, as a battle villancico, needed a more percussive sound in the continuo. As we did not have guitars available for this performance, the harpsichord provides a better substitute than the organ for this particular piece.

Length and Order

Two final performance considerations that must be made are whether to repeat the estribillo between each copla and whether to perform all verses. Originally the villancicos were part of festivals or services, and lyrics were made available as pliegos suletos and the audience or congregation may even have had the text in hand. (Laird 1997:52) 2 Therefore, participants in the past were listening to the whole story as well as the music, perhaps while processing to the Chapel.  For today’s audiences, however, even with changes in instrumentation, the musical repetition can make the performance of an entire villancico a bit long to listen to. In addition, it is not always clear from the manuscripts when and how often the estribillo would have been repeated throughout the villancico, though fairly clear that they would be heard at the beginning and end.   With this in mind, I made the decision to sing the estribillos at the beginning and the end, and to perform only two of the coplas for each villancico, but to print the entire text in the program.


1. This article represents a consolidation of information presented on January 19, 2007 at the International Symposium on Latin American Choral Music at the University of Arizona. This presentation was followed by a live performance of three villancicos of Antonio Salazar titled Victoria por el Silencio, A Coronarse Reyna, and Al Salir el Sol, performed by member of the UA Collegium and the Viola da Gamba Society of Southern Arizona. For a program with the complete list of performers and selections performed see figure 2.

2. Stevenson writes, quoting Lothar G. Siemans Hernandez (1968): Before the performances began, the choirboys [...] passed out to the assembled crowd printed texts of the carols to be sung at the hours of nocturns, so that the public could follow the music and better appreciate the details of the ofen involved poetic texts. At times, these printed libretti for the villancico-cycles were even dropped from the dome in some churches, along with flower petals and colored paper birds." (Stevenson 1974:25)