Latin American Choral Music - Table of Contents

References Cited

An Appraisal of Spanish Colonial Era Music with Náhuatl Texts
David Leedom Shaul

This paper looks at Spanish colonial era music with texts in the indigenous Náhuatl language (the language of the Aztecs and numerous other groups in Mexico and El Salvador).  Pieces with Náhuatl texts include:

  1. a Christmas lullaby by Gaspar Fernandes and three other pieces from the Fernandes codex from the Oaxaca cathedral;
  2. two motets in honor of the Virgin Mary attributed to Hernando Franco;
  3. a communion hymn in Santa Eulalia M. MD. 7.

I will discuss three examples as an introduction to a lesser known field of Spanish colonial music output.  The style some of these pieces includes features forbidden in European counterpoint of the time (parallel fourths and fifths), but which were common in the musical practice of two noted high cultures of Meso-America (Aztecan, Mayan).  I will argue that not only do these features refer to native musical practice, but also index nativeness since the composers of the pieces with such features were probably Indians themselves.

Notice is taken of native Aztec music composed and/or written down after the Contact.  These pieces are in Romances de los Señores de la Nueva España and Cantares Mexicanas.  A native solmization is found in both these works, and I propose a solution to how this system works following what we know about native musical practice, and illustrate with part of a hymn to Santa Maria in this style.

The Christmas Lullaby of Fernandes

The well-known Christmas lullaby "Xicochi, xicochi conetzintle" by Gaspar Fernandes (1565-1629) is fairly well known.  The composer is known for composing villancicos that used a dialect of Spanish attributed to persons of African heritage, African rhythms, as well as percussion instruments.  Such villancicos were known as negritos, and were composed by other composers as well, and such ethnically influenced villancicos existed well before Fernandes' time (Stevenson 1994).

It is not unusual that Fernandes wrote pieces with texts in Náhuatl, the language of the Aztecs and a major interlanguage between Native Americans in colonial New Spain.  The four pieces in the Tlaxcalan dialect of Nahuatl (in the Fernandes codex:  folios 99v-100; 133v-134; 219v-220; 202) have the 6/8 rhythm that is typical of the negrito villancicos (eighth note-quarter note, eighth note-quarter note) that accent the third and fifth of the six beats.

Xicochi, as an example of Fernandes' Nahuatl language works, is very simple (only 21 measures of 6/8 time), and is not technically a villancico (no estribillo or refrain).  The pattern of an eighth note followed by a quarter note not only gives the sense of a rocking motion appropriate for a lullaby, but the text is also cleverly arranged to have the main stress (antepenultimate in Náhuatl) fall on quarter notes, preserving the primary word-level feature of Náhuatl prosody.

The brief text of this piece has been mistranslated in program notes over the years.  Here is the text.

Xicochi, xicochi conetzintlé [:]
ca ōmitzhuihuixocoh in angelosmeh.

This translates as the following.

Go to sleep, go to sleep revered baby;
the angels already rocked you.

The verb huihuixo- means 'to shake' or 'to rock an infant', and the latter sense is the preferred meaning here.  It should be noted that this translation supersedes translation guesses given for this piece ("cry no more, for angel are here," and "surely angels will carry you on your path").

The piece has two soprano parts (tiples), an alto part, and a bass.  There are three cases of parallel fifths between the bass and alto, and more fifths between these two parts are prominent, although separated by another interval (often an octave or sixth).  Between the alto and lower soprano (tiple 2) there is only one parallel fourth.  One instance of a parallel fourth occurs between the two soprano parts, and there is voice crossing (for 2.5 measures in measure 6 and 7; 3.5 measures, four measure before the end).  There are also consecutive octaves throughout the work.

One of the major features of most if not all Native American musics is the lack of triple meter (except in marked contexts such as hemiola in Puebloan songpoems of the American Southwest (Shaul 2002, 177, 180).  Even Bierhorst in his survey of pre-Cortesian music misses this major stylistic trait.  In Xicochi, a duple pulse may be created within the 6/8 meter.  Yet the triple meter of Xicochi could be interpreted a duple meter (with primary stress on the first and thirds beats of the six), so one could consider that the piece "fits" into the indigenous desideratum for duple meter.

The overall style of the piece fits other examples of "native" polyphony: 

  1. parallel fourths, especially in the upper two voices;
  2. voice crossing;
  3. parallel fifths, especially in the lower most two voices;
  4. consecutive octaves;
  5. (possibly) basic duple meter.

These traits are also found prominently in other examples of Nahuatl language polyphony.  Critical examination of Fernandes' other Nahuatl texted pieces will show how closely he conformed to these traits that seem to index nativeness in the context of Spanish colonial music in New Spain.  I now turn to other examples where the same traits not only index the indigenous text, but also probably the ethnicity of the composers of the pieces.

The Náhuatl Motets of "Hernando Franco"

Stevenson (1968, 204-6) suggests that the two celebrated pieces attributed to Hernando Franco (1532-1585), a chapel master at the Mexico City cathedral, are in fact the work of a Native American composer.  He bases this idea on the fact that the composer is referred to as Don Hernando Franco.  The title Don was adopted by the indigenous nobility as they scrambled to convert to Catholicism, save their land and elite status, and get coats-of-arms from the King of Spain.  By contrast, commoners did not have this title in colonial times.  Stevenson notes that the chapelmaster Hernando Franco never earned the title.

The two pieces In ilhuicac cihuapillé  and  Dios itlazonantiné  are from a manuscript called the Valdés Codex (folios 121v-123), first published in 1934 (Saldivar, Historia de la Música en México),  date from the mid 1500s.  The date on the manuscript of 1599 is probably when the pieces were copied; the paper the pieces are on have a watermark that was made in Madrid ca. 1561-1571 (Stevenson 1968, 204, note 101).

Stevenson quipped that "our composer liked ternary meter" (Stevenson 1968: 206), although only one piece is in triple meter (In ilhuicac cihuapillé).  Although Don Hernando had mastered "the prickly system" of triple mensuration still current in Spain and Mexico, Stevenson notes that he had "a certain clumsiness in his part writing ... his management of suspension, of consecutive fifths and octaves --  not to mention other niceties  --  is frequently distressing" (Stevenson 1968: 207).

A feature of the texture of  In ilhuicac cihuapillé   that is prominent is the use of parallel fifths, especially in the tenor and bass.  This interval is common in many indigenous North American musics, and was a feature of pre-Cortesian music in Mexico.  Of the two pitched drums used to accompany Aztec song cycles the huehuetl, which was retuned higher with each succeeding piece in the song cycle, was always pitched at a fifth (between the rim and center).  The other drum, a slat or log drum called the teponaztli,  had two keys that could, among other intervals, be pitched as a perfect fifth.  It follows, then, that parallel fifths as a feature common in the native music style, are to be expected from a native composer, and that other infelicities "within a Spanish Renaissance context" (Stevenson 1968, 107) are also indigenous traits used to naturalize the imported polyphonic style.  The two pieces of Don Hernando show the same traits as the Fernandes lullaby.

The Communion Hymn in Santa Eulalia M. MD. 7

A body of Renaissance music from late 16th century Guatemala includes a communion hymn in Náhuatl.  I will refer to this music as the Guatemalan repertory.  The Guatemalan repertory dates between 1570-1600 (Stevenson 1964 and 1970; Baird 1981; Borg 1985). It was copied and composed for use in three towns in Huehuetenango Department, Guatemala (Santa Eulalia; San Juan Ixcoy; San Mateo Ixtatán).  Most of the manuscripts containing this repertory ended up in the Lilly Library at Indiana University.  Stevenson (1970: 51) stated that there were songs in Náhuatl and Mayan languages (Chuj, Kanjobal, Jaltec), and Borg (1985:66) noted that the fifteen manuscripts included "vernacular motets."  Of the text with Native American language texts, the following obtain:

  1. Jesu Christo to esi caol (8/7);
  2. Itech nepa sacramento (7/9);
  3. Magalhi vinac Dios (7/44);
  4. Vachon loh (3/24);
  5. Bay magalhi (7/37);
  6. Vac jat santa (6/7);
  7. Santa Maria uch (6/14).

(The first number in parentheses is the number of the Guatemalan repertory manuscripts according to the numeration used by the Lilly Library.)  Of these, only Itech nepa sacramento is clearly in Nahuatl. Baird (1981: 91-92) identifies the texts of Bay magalhi and Magalhi vinac Dios as Nahuatl on the testimony of Dr. Dow Robinson (Summer Institute of Linguistics. Mexico Branch), a Nahuatl expert.  I suggest that these two texts (and the four others listed above) are texts in a Mayan language. Borg suggests that the non-Nahuatl texts in the Guatemalan repertory are in Jacaltec-Chuj Maya (Borg 1985 117), while Baird (1981: 10) notes that all three of the towns in which the Guatemalan repertory were found have Kanjobal Mayan as the native language. 

The piece Itech nepa sacramento is from the manuscript designated Santa Eulalia M. MD. 7 (item 7 of the fifteen manuscripts as per the Lilly Library designations).  It is dated 20 Jan. 1600 in an inscription in Nahuatl by its compiler, one Tomás Pascual, an indigenous official.  (Ironically, the inscription gives the town San Juan Ixcoy as the place of origin, although this manuscript was found in Santa Eulalia.)

The Nahuatl text of Tomás Pascual's hymn and inscription are impugned by Baird (1981, 91) as a corruption "of the elegant, classical Náhuatl of the early Aztecs."  Actually, given the 16th century spelling conventions used for writing Nahuatl (which routinely do not write vowel length), there is nothing in Pascual's inscription or hymn, other than Spanish loanwords, that cannot be easily transcribed into Classical Nahuatl. 

The texture of  Itech nepa sacramento  lacks the rhythmic variation of Don Hernanado's two pieces.  However, it has prominent parallel fourths and fifths, and is in duple meter.  Both of these are characteristic of Meso-American indigenous musical style.  The fact that this hymn is in a tradition far away from the metropolitan center at the same time (mid 1500s) as Don Hernando suggests a style unique to native composers who used features common to their traditional music to create their own naturalized part music.  The style that indexes Fernandes' lullaby not only indexes but underscores the nativeness of the native composers (Don Tomás and Don Hernando).  It is even more unusual that this constellation of traits within the context of western polyphony was shared at the same time (last half of the 1500s) not only in the Viceroyalty of New Spain, but also was to be found in a fairly remote part of Guatemala.

A Reconstruction from the Cantares Mexicanas

The solfège in the Cantares Mexicanas  and the  Romances de los señores de la Nueva España  has long been the object of interpretation (Nowotny 1956, Mendoza 1956: 20-26, Ziehm 1976:267-276,  Bierhorst 1985: 72-79).  Perez de Ribas states "what is sung (and the singing never stops) always corresponds to the color the teponaztli plays" (1944).  The teponaztli is the slit drum, to which we will return.  The point is that the two pitches produced by the teponaztli were the basis of the sung melody.

Scholars interpreting the Nahuatl solfège have taken the two vowels common to the four syllables of the solfège as pitch indicators, and the rest (the consonant shapes) as rhythmic indicators.  The approach taken here is to propose a total pitch indicator solution, based on the acoustic properties of the syllables themselves.  Arranged by relative pitch, from high to low, we have:

  1. ti
  2. to
  3. qui
  4. co

Stevenson (1968:102-103) infers that Aztec music was too narrow in pitch range for European ears, and so a four-note pitch row would fit into this aspect of Aztec musical style.  Moreover, what we do know about Aztec performance practices of the cantares fits in with a four-note pitch row concept.

The huehuetl  and  teponaztli  drums were always used together in accompaniment to  cantares.  Motolinia (Bierhorst 1985: 81), stated that the teponaztli was lower in pitch and was the counterbass (contrabajo) and that the huehuetl was the bass.  This means that the teponaztli would have produced two low notes (the  qui  and  co), with the hueheutl making the two higher pitches (the  ti  and  to).  (The huehuetl was a pitched drum played by the hands striking the rim or center, the center being higher in pitch than the rim.) 

A study of eleven authentic teponaztlis in museums (Castañeda and Mendoza 1933) found the following results in pitches:

Number of Instuments



minor 3rd




Major 2nd




Major 3rd

This would give a certain amount of flexibility in terms of pitch rows, and again, what we know about Aztec performance practices of  cantares  fits into this interpretation.

Motolinia states that the first cantare  performed was the lowest in pitch.  The singers would continue singing while the huehuetl was tuned higher.  Each succeeding song had higher pitches in the huehuetl at least; there is not report of teponaztlis being switched in performance.  Since the teponaztli and huehuetl played together, we can assume heterophony at different pitch levels.

At this point, before presenting a reconstruction, we need to consider form and rhythm.  Bierhorst (1985: 80) assumes that there was no regular meter (despite interpreting the solfège as mainly rhythmic figures with only some pitch indication) and as being heterorhythmic, meaning that varying numbers of syllables of text were gathered by basic pulses.  I further assume a basic duple pulse, which is one of the near universals of Native American music.  Bierhorst (1985: 81), citing Motolinia, states that the beginning of a song melody is low in pitch and slow, and then the song rises in pitch and tempo, and then returns to the lower pitch and tempo (Bierhorst calls these three sections:  verse-refrain-litany;  ethnomusicologists of Native American musics may call the middle section "the rise").

This return form fits with the reconstruction given here (section B of Song 59 of the  Canteares).  The following is the first solfège for section B:

Tocoto tocoto tocoto tocoto.

Then the mansucript states "ynepantla onahci in cuicatl niman ye ontlami" [comes the middle of the song, it finishes as follows]:

tiquiti ticoti tiquiti tocoto.

The pitch, as proposed here, shifts higher, and then returns as per Motolinia to the original pattern at the beginning.

In the reconstruction presented here, the fourth syllables of the solfège are assumed to be a pitch row are C, D, E, and F natural, simply for purposes of illustrating the total-pitch interpretation of the Aztec solfège.  A descant of a fifth has been added to simulate the pitch contribution of the huehuetl, again simply for purposes of illustration, as well as to simulate the heterophonic texture of most Native American musics. 

The way I have grouped text is based on main stress in Nahuatl, which is antepenultimate.  I have assigned a quarter note value and lower pitch to these stressed syllables.  I have also assigned quarter note values to syllables beginning syntactic phrases.  The syllables of prefixes and suffixes (Nahuatl has plenty of both) have shorter rhythmic values, and groups of three syllables before a main stress are gathered rhythmically as triplets.  The effect of glottal stop (a morpheme marking plural in Nahuatl) has been reconstructed as an eighth rest.


I have suggested that there was a style created and used by Native American composers (Aztec and Mayan) that incorporated major features of their pre-Cortesian music.  This included a tendency for duple meter, parallel fourths, and parallel fifths, as well as occasional tritones and voice crossing.  Further, a simple texture of note-for-note seems preferred to rhythmically active polyphony.

Stevenson notes a hymn to Hippolytus (d. 236), the patron saint of Mexico City which has a Nahuatl text (1968:203 and note 98) .  The piece is reproduced in Pedro de Morales' Carta de Padre Pedro de Morales  (Mexico, 1579).  He also notes (1968, 50-51 and note 62) that Sor Juana de la Cruz used the native tocotin dance with a Nahuatl text in her Assumption villancicos of 1687.  The Morales piece (performed 1578 accompanied by flutes and teponaztli) and the de la Cruz piece (1687) can be compared to the work of Don Hernando and Tomas Pascual, and they more closely conform to the strictures of western polyphony of the period.

Opportunities for future research are posed by the instrumental pieces in the Guatemalan repertory, as well as those identified by Stevenson (1968, 165, note 28) for the Valley of Mexico.  In the Panorama de la Música Tradicional  (Mendoza 1956), there are six  tocotines  dated 1620, 1625, 1640, and 1651.  Guitar tablatures dating ca. 1650 in a private collection (Stevenson 1968, 234-235) contain  tocotines  as well as an additional dance genre, the  mitote

There is also the body of music with Mayan texts in the Guatemalan repertory to consider stylistically, as well as linguistically.  Further, there are pieces with texts in Spanish in the Guatemalan repertory as well.