27

Missionary Traditions in Mesoamerica

otto zwartjes

Although even today missionaries are active in linguistic fieldwork, the

chronological focus of this chapter is on lexicographical production during
the colonial period. Its geographical focus is the culturally defined area that is
traditionally called Mesoamerica. To the east and south, Mesoamerica
extends to the Caribbean Sea and the southern borders of the area where
Mayan languages are spoken, excluding the Caribbean islands and the south-
ern part of central America. The frontier to the north is more problematic.
The current border between the United States and Mexico is not the appro-
priate frontier, because it does not correspond to earlier borders of the
Spanish territories; moreover, the lexicographical traditions from what is
sometimes called the Greater Southwest, in other words northern Mexico,
are quite different from those of the central part of Mesoamerica. Those
traditions will therefore be treated in Chapter 28.
In this chapter, the main focus will be on the most important lexicogra-
phical works of Tarascan; Nahuatl; the Oto-Manguean languages, such as
Zapotec, Mixtec, Otomi, and Matlatzinca; and the Mayan languages. Apart
from these greater works, some experiments in linguistic documentation and
teaching will be discussed as well, such as an introduction to Mixe. An
overview of the dictionaries to be discussed will be followed by accounts of
their sources; their organizational principles; some innovative language-
instruction tools and pedagogical approaches; and some topics related to
the content of the dictionaries: did missionary lexicographers in Mesoamerica
mainly try to find equivalents for lemmata from Western culture, or did they
also attempt to give a record of the world of the other, or did they do both?

Historical Overview
The first pioneering dictionaries in New Spain appeared in the sixteenth
century, most of them being the work of Franciscans in the Central Valley,

579

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

mainly studying Nahuatl, and Dominicans working in the region of Oaxaca,

describing and documenting Oto-Manguean languages such as Mixtec and
Zapotec. An important lexicographical tradition was developed on the
peninsula of Yucatán and Guatemala, where Mayan languages were
documented.
Around 1545 an anonymous author or authors compiled a trilingual
Spanish–Latin–Nahuatl dictionary by copying the 15,260 entries of an early
edition of Elio Antonio de Nebrija’s Spanish–Latin dictionary (for which see
Chapter 14) and then adding about 11,000 Nahuatl glosses to them. This, the
very first substantial dictionary of any language of the Americas, remained in
manuscript throughout the period under discussion.1 In 1547, the Franciscan
Andrés de Olmos completed his Nahuatl grammar, which includes
a vocabulary of 2,062 words, and also remained in manuscript during the
colonial period.2 The first printed dictionary of the New World – the second
printed dictionary of any language other than those of Christian Europe, after
Pedro de Alcalá’s dictionary of Arabic – was completed in 1555 by the
Franciscan Alonso de Molina, containing 13,866 Spanish headwords (drawn
from an edition of Nebrija published in 1545) with 29,742 Nahuatl
equivalents.3 This was a monodirectional Spanish–Nahuatl dictionary. In
1571, Molina produced a bidirectional version, with 17,410 Spanish entries
and 37,433 Nahuatl equivalents in its Spanish–Nahuatl section, and 23,625
Nahuatl entries in its Nahuatl–Spanish section.4 A new bidirectional diction-
ary of Spanish and Nahuatl was published by Pedro de Arenas in 1611; it was
reprinted frequently in the seventeenth and eighteenth centuries.
Lexicographers of Nahuatl in the seventeenth and eighteenth centuries did
not always continue describing and teaching the prestigious variety of the
Central Valley, changing their focus to more peripheral varieties, such as
those spoken in Tierra Caliente or Jalisco.5

1
Téllez Nieto, Acercamiento filológico, includes a partial edition; Hamann, Translations of
Nebrija, 44–7, gives an account of the dictionary, identifying its source as the pirated
edition of Nebrija, dated 1516 and probably printed around 1520, and illustrates a page of
the manuscript.
2
Olmos, Arte de la lengua mexicana (1993), is an edition; entry counts for this wordlist and
those of Molina are from Clayton and Campbell, ‘Alonso de Molina as lexicographer’,
336–8; see also Hamann, Translations of Nebrija, 47–8.
3
Molina, Aqui comiença un vocabulario (2001), includes a facsimile of the 1555 edition; for the
source, see Hamann, Translations of Nebrija, 48–50.
4
Hamann, Translations of Nebrija, 59.
5
See Zwartjes’ entries on Juan Guerra, Arte de la lengua mexicana (1692); Manuel Pérez,
Arte de el idioma mexicano (1713), and Jerónimo Tomás de Aquino Cortés y Zedeño, Arte,
vocabulario, y confessionario en el idioma mexicano in Corpus de textes linguistiques
fondamentaux.

580

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

Printed dictionaries of other Mesoamerican languages followed very soon.

The first was Maturino Gilberti’s work on Tarascan, published in 1559, and
offering 6,254 Tarascan–Spanish and 13,668 Spanish–Tarascan entries, making
it the first bidirectional dictionary printed in the New World.6 Its Spanish
wordlist is a close copy of that of the 1555 edition of Molina’s Nahuatl
dictionary.7 It was followed by Juan Baptista de Lagunas’ Tarascan–Spanish
dictionary of 1574. A Spanish–Zapotec dictionary by Juan de Córdova was
published in 1578, drawing on the 1571 edition of Molina: its 28,352 entries
made it the most extensive of the Mesoamerican dictionaries of its day that
had Spanish as a source language.8 A Spanish–Mixtec dictionary published by
Francisco de Alvarado in 1593 drew both on Córdova’s work and also directly
on the 1571 Molina.9
These printed bilingual dictionaries were sometimes elaborately marked
up with equivalents in a third language. As early as 1557, Andrés de Castro
added 25,000 or more words of Matlatzinca to the margins of a copy of the
1555 edition of Molina’s Nahuatl dictionary.10 Glosses in Otomi were added
to copies of the 1555 and 1571 editions of Molina’s dictionary and to a copy
of Gilberti’s Tarascan dictionary.11 A similar annotated printed book must
have been the parent of the extant manuscript of Alonso Urbano’s large
Spanish–Nahuatl–Otomi dictionary, dated 1605, for the Spanish–Nahuatl
part of each of Urbano’s entries is copied from the 1555 edition of Molina’s
dictionary.12
Other wordlists were never printed during the colonial period and are
conserved in manuscript form. Thomas Smith-Stark lists nine or ten from the
period before 1611.13 The earliest is a Kaqchikel–Spanish wordlist, giving
forms in K’iche’ and Tz’utujil when they differed from the Kaqchikel form,
which was compiled by Domingo de Vico, who died in 1555; this is the oldest

6
Gilberti, Vocabulario en lengua de Mechuacán (1997), is an edition; entry counts are from
Monzón, ‘Tarascan lexicographic tradition’, 173.
7
Hamann, Translations of Nebrija, 51–5.
8
Córdova, Vocabulario Castellano–Zapoteco (1942), is a facsimile; for the source, see
Hamann, Translations of Nebrija, 62; for the entry count, see Smith-Stark,
‘Lexicography in New Spain’, 26.
9
Alvarado, Vocabulario en lengua mixteca (1962), is a facsimile; for the sources, see
Hamann, Translations of Nebrija, 73–4.
10
Schuller, ‘Unknown Matlatsinka manuscript vocabulary’; further information in
Hamann, Translations of Nebrija, 51, 109, and 172 n. 25.
11
Hamann, Translations of Nebrija, 50–1, 55, 59, 108–11.
12
Hamann, Translations of Nebrija, 77–8; Urbano, Arte breve dela lengua otomí y vocabulario
trilingüe (1990), is a facsimile edition.
13
Smith-Stark, ‘Lexicography in New Spain’, 19–21; see also Hernández, Lexicografía
hispano-amerindia, 29–35 and passim.

581

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

extant wordlist of a Mayan language.14 It was followed by a Spanish–

Kaqchikel wordlist of the 1570s (it is based on the 1571 edition of Molina’s
Nahuatl dictionary, and a manuscript of c. 1578 is extant, so it must have been
composed between those dates) by Juan Alonso.15 A bidirectional dictionary
of Tzeltal and Spanish, with 7,769 Spanish entries, was compiled by Domingo
de Ara around 1560.16 A Spanish–Maya dictionary of about 9,000 or 9,500
entries and a bidirectional dictionary of Maya and Spanish are both attributed
to Alonso de Solana, and belong to the period around 1580.17 The manuscripts
of these dictionaries are known respectively as ‘Diccionario de Motul II’ and
‘Diccionario de San Francisco’, the latter taking its name from the Franciscan
convent of that name in Mérida. Another Spanish–Maya dictionary was
compiled by Gaspar González de Nájera by 1582.18 The Maya–Spanish dic-
tionary of 15,975 entries compiled by Antonio de Ciudad Real at the end of the
century was the most extensive of the dictionaries with a Mesoamerican
source language; a seventeenth-century source reports that its compilation
took forty years.19 An anonymous compiler produced a Spanish–Tzotzil
dictionary of 8,077 entries at the end of the sixteenth century or the beginning
of the seventeenth, now extant only in a twentieth-century transcript.20
A bidirectional dictionary of Spanish and Tarascan was compiled by an
anonymous Augustinian friar at the end of the sixteenth century, offering
about 22,000 Spanish–Tarascan and 27,000 Tarascan–Spanish entries.21 The
otherwise obscure Dionisio de Zúñiga Marroquín compiled a bidirectional
dictionary of Spanish and Poqomchi’ with a Latin–Poqomchi’ section around
1608.22 To these may be added the Matlatzinca–Spanish and Spanish–
14
Smith-Stark, ‘Lexicography in New Spain’, 24; Hernández, ‘Indigenismos’, 69–71;
Hernández, ‘Vocabularios Hispano–Mayas’, 142–3.
15
Hernández, ‘Vocabularios Hispano–Mayas’, 141–2; Hamann, Translations of Nebrija, 60.
16
Ara, Vocabulario de lengua tzeldal (1986), is an edition (in the introduction of which a date
c. 1560 is assigned); see also Hernández, ‘Vocabularios Hispano–Mayas’, 140–1, and, for
the entry count, Laughlin in Great Tzotzil Dictionary of Santo Domingo Zinacantán
(1988), I.28.
17
Hernández, ‘En torno al vocabulario hispano–maya’ (entry count of 9,500 at 115; cf. the
lower counts reported by Smith-Stark, ‘Lexicography in New Spain’, 26); see also
Hamann, Translations of Nebrija, 60–1.
18
Hernández, ‘Vocabularios Hispano–Mayas’, 134.
19
[Ciudad Real], Calepino maya de Motul (1995), is an edition; for the entry count and the
forty years, see Smith-Stark, ‘Lexicography in New Spain’, 26, 28.
20
Great Tzotzil Dictionary of Santo Domingo Zinacantán (1988) presents a facsimile of the
extant transcript, with a freely modernized edition; Tzotzil–English and English–Tzotzil
dictionaries based on the edition; a grammar; and an introduction: entry count I.28.
21
Diccionario grande de la lengua de Michuacán (1991) is an edition; entry counts from
Monzón, ‘Tarascan lexicographic tradition’, 181.
22
Smith-Stark, ‘Lexicography in New Spain’, 55 (and see 76 for information about the
manuscripts).

582

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

Matlatzinca dictionaries of Diego Basalenque, of which the manuscripts are

dated 1640 and 1642 respectively, and anonymous eighteenth-century diction-
aries of Kaqchikel and K’iche’ with Spanish, one of which has been edited as
Vocabulario en lengua 4iche otlatecas (the character 4 represents the same
ejective consonant as the combination k’ which is now used: see the section
on organizational principles below).23
It is remarkable that the most comprehensive dictionaries – and also the
best, from our modern perspective – were composed in the period 1555–1650
(the same applies to South America, but not to the Philippines: see Chapters 26
and 29 respectively). It is surprising that some of the minor grammars were
often reprinted (such as the five editions of Vázquez Gastelú’s grammar of
Nahuatl), whereas the major lexicographical works were generally not
reprinted during the colonial period: there were, for instance, no reprints of
Molina’s dictionary after the 1571 edition, and Olmos’ work was never printed
at all.24 This is probably a consequence of the changing nature of the pedago-
gical tools in language instruction. By the 1630s, Miguel de Guevara’s ‘Arte
doctrinal y modo general para aprender la lengua Matlaltzinga’ and Diego de
Nágera Yanguas’ Doctrina y enseñança en la lengua maçahua were combining
information about vocabulary and about grammar, and Augustín de
Quintana’s Confessonario en lengua mixe, con . . . un compendio de voces mixes
would likewise mix genres in the following century. When there was a need for
an update to the dictionary record – for instance, when words of the diction-
aries of centuries before were no longer in use, when archaic usage had to be
avoided, or when specific dialectal varieties were missing in the pioneering
works – authors tended not to publish new dictionaries as independent works,
but included them as sections of other publications, such as Cortés y Zedeño’s
Arte, vocabulario, y confessionario (1765).
The variety of the lexicographical works which I have surveyed above is
great, and in particular the works that translate from the indigenous language
to Spanish are far from uniform. Almost every lexicographer followed his
own methodology which was developed for the specific features of the
language under study.
When dictionaries are bilingual and bidirectional, the two sections are
seldom equal in length. So, for instance, the Spanish–Nahuatl first part of

23
Basalenque, Arte y vocabulario de la lengua matlaltzinga vuelto en la castellana (1975), and
Basalenque, Vocabulario de la lengua castellana vuelto en la matlaltzinga (1975), are editions
of the first two; for the latter, see the table in Diccionario k’iche’ de Berlín (2017), 12.
24
For the several reprints of Vázquez Gaselú’s grammar, see O. Zwartjes’ entry for the
author in Corpus de textes linguistiques fondamentaux.

583

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

Molina’s dictionary of 1571 contains 121 folios, whereas the Nahuatl–

Spanish second part contains 162 folios. By contrast, the Matlatzinca–
Spanish part of Diego Basalenque’s work is less comprehensive than the
Spanish–Matlatzinca part. Maturino Gilberti’s Spanish–Tarascan section is
twice as big as his Tarascan–Spanish section, and, unlike Molina, Gilberti puts
the section of his dictionary which starts with Spanish in second place.25
Some lexicographers composed works other than dictionaries. Alonso de
Molina, for instance, also published a Nahuatl grammar and a Doctrina
christiana breve traduzida en lengua mexicana. Juan de Córdova wrote
a grammar as well as a dictionary of Zapotec. Grammar and dictionary
were not always by the same person: for Mixtec, for instance, we have an
author who composed the first printed grammar, namely Alonso de los
Reyes, whereas another missionary, Francisco de Alvarado, composed the
dictionary. But it was not unknown for a single author to compose a whole
suite of grammatical and lexicographical texts, as in the case of Basalenque,
whose dictionaries of Matlatzinca are accompanied by a note on the use of
the Roman alphabet to write Matlatzinca, ‘Cartilla Matlaltzinga’; a grammar,
‘Arte de le lengua Matlaltzinga’; and a treatise on particles, to which we shall
return, ‘Tratado de las particulas de la lengua Matlaltzinga’.26 This means that
sometimes there is a tight relation between a dictionary and a grammar or
other texts, which results in a frequent use of cross-references, whereas in
other cases the works can be quite different from each other.

Sources
Although other sources were used, most Mesoamerican dictionaries were, as
we have seen, inspired by the Spanish–Latin Vocabulario of Antonio de
Nebrija, first published in 1495. Some authors followed Nebrija’s entries
strictly, while others used the Vocabulario more freely, adding new entries,
suppressing others, adapting Nebrija’s work to the languages they studied, or
just following their own creativity. As was the case with the use of Nebrija’s
Latin grammar, there was not just one version which served as a model for
missionary lexicographers. As Byron Ellsworth Hamann observes in The
Translations of Nebrija, after the editio princeps thirty-four further editions of
the Vocabulario were published in nine European cities by the early 1600s;
their ‘lists of entries were in constant flux’, and ‘the constantly changing
25
Monzón, ‘Tarascan lexicographic tradition’, 173.
26
Basalenque, Arte y vocabulario de la lengua matlaltzinga (1975), 7–12 (‘Cartilla’), 13–117
(‘Arte’), 119–42 (‘Tratado’).

584

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

nature of Nebrija’s “Castilian–Latin dictionary” has been underappreciated, but

it is of fundamental importance’. Nebrija’s dictionary, Hamann continues,
‘gave birth to multilingual offspring’, which had offspring of their own: ‘The
later we move through the sixteenth century, the more complicated these
genealogies become . . . in some cases, dictionaries are descended from two or
even three progenitors.’27 As we saw above, Alvarado’s Mixtec dictionary of
1593 draws on Córdova’s Zapotec dictionary of 1578, and also directly on
a source of Córdova’s, namely Molina’s Nahuatl dictionary in its edition of
1571; Hamann points out that it also draws directly on an edition of the ultimate
source of Córdova’s and Molina’s work, the Vocabulario of Nebrija.
References in dictionary prefaces to ‘Antonio’ as a forerunner often refer
generically to any later dictionary in the Nebrija tradition; likewise, ‘Calepino’
is often used as a generic term for dictionaries in the long and complex Calepino
tradition, and these too were sources for missionary lexicographers in
Mesoamerica (for Calepino, see Chapter 14, and for the contrast between the
Nebrija and Calepino traditions, see Chapter 26). Lagunas explicitly mentions the
name of Calepino in his dictionary of Tarascan. As he informs his reader in
the prologue, the text ‘could be called a small dictionary [dictionarito], because the
author (for the benefit of the students) imitates the order and manner of
Ambrogio Calepino. Though in the order of letters it is impossible to do this in
this language, given the different meanings.’28 As Cristina Monzón notes, it is
remarkable that Lagunas does not mention Gilberti, who was apparently one of
his sources, but Calepino, although the difference between Lagunas’ work and
that of Calepino is enormous. Monzón suggests that Lagunas decided to mention
Calepino as an authority, and concludes that ‘it seems plausible to postulate that
Lagunas was also acquainted with the Regio 1502 [the first edition of Calepino] that
was on the shelves of the library of the Colegio de Santa Cruz de Tlatelolco’.29
Spanish-speaking lexicographers working in the New World (and in the
Philippines) did not always take one of Nebrija’s dictionaries as their model:
as we have seen, Gilberti’s Spanish–Tarascan dictionary of 1559 is a translation
of Molina which replaces the Nahuatl equivalences with Tarascan: ‘Any debt
which Gilberti owes to Nebrija was acquired via Molina.’30 The shorter

27
Hamann, Translations of Nebrija, 3.
28
Lagunas, Arte y dictionario, 1 (second sequence of pagination), ‘podra se llamar
Dictionarito, porq[ue] el auctor (p[ar]a mas p[ro]uecho d[e] los estudia[n]tes) sigue
e[n e]l el orde[n] y modo del Ambrosio Calepino. Au[n]q[ue] e[n e]l orde[n] d[e] las
letras es imposible e[n e]sta le[n]gua, por los distinctos significados’; translated in
Monzón, ‘Tarascan lexicographic tradition’, 189.
29
Monzón, ‘Tarascan lexicographic tradition’, 191.
30
Smith-Stark, ‘Lexicography in New Spain’, 8.

585

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

wordlists characteristic of the eighteenth century are generally not derived

from Nebrija or from Molina’s dictionary.31
The section of a bidirectional dictionary that starts with the indigenous
language was not necessarily the inverted version of the section which starts
with Spanish, so even if Nebrija was the source for one section of a bidirectional
dictionary, the other might be an original compilation. So, for example, in
Molina’s bidirectional dictionary of 1571, ‘the Nahuatl/Spanish side is a new
compilation rather than a reversal of the Spanish/Nahuatl section’.32 The same
is true of Gilberti’s Vocabulario of 1559: the Spanish–Tarascan section follows the
first edition of Molina’s dictionary closely, whereas the Tarascan–Spanish
section is ‘a creation by Gilberti and his team since he did not [invert] the
Spanish–Tarascan entries’.33 In the lexicographical tradition of Mayan lan-
guages, Ciudad Real’s dictionary of Maya and de Ara’s dictionary of Tzeltal
are monodirectional, starting with the indigenous language, and as far as we
know, they never had counterparts starting with Spanish. In these dictionaries,
no trace of Nebrija can be found.34 In the historical introduction of Michael
Dürr and Frauke Sachse to their edition of the Vocabulario en lengua 4iche
otlatecas, Nebrija is not identified as a source.
In Hamann’s Translations of Nebrija, only brief references are included to
bilingual dictionaries that translate to Castilian from another language.
Without any doubt, the study of Hamann is important, but seen from
the perspective of a historian of lexicography, the sections which translate
from the indigenous languages to Spanish are often much more interesting,
since the author had to make important decisions about how to organize the
lemmata, in particular for polysynthetic languages with a great number of affixes,
such as Nahuatl, or languages which have a great number of prefixes, such as
Matlatzinca. How to separate the headwords or roots from other elements, such
as bound morphemes and pronominal prefixes, was a great challenge for
missionary lexicographers; we now turn to their solutions to these problems.

Organizational Principles
The alphabetical sequences by which headwords were ordered in the
Mesoamerican missionary dictionaries cannot all be treated in detail

31
Yáñez Rosales, ‘Presencia y ausencia de Antonio de Nebrija’.
32
Karttunen, ‘Nahuatl lexicography’, 2658.
33
Monzón, ‘Tarascan lexicographic tradition’, 173.
34
Hernández, Lexicografía hispano-amerindia, 139–55.

586

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

here.35 Missionaries followed Nebrija’s model, but also developed a special

alphabet for the languages under study. They gave new values to existing
letters; created new combinations of letters, resulting in digraphs and
trigraphs; and added diacritics whenever necessary and whenever the
printing facilities were able to reproduce them. In some cases they devel-
oped new letters, such as the frequently quoted alphabet devised in the mid
sixteenth century by Francisco de la Parra to represent uvular and glotta-
lized phonemes in languages such as K’iche’ and Kaqchikel: A B C CH
E H Y K L M N O P Q R T V X Ɛ 4 4H TZ.36 Sometimes the digraphs are
ordered alphabetically, which means for instance that <th> is found under
the letter <t> or immediately after it, and sometimes they appear at the
end, after the last letter of the Spanish alphabet, as in Basalenque’s dic-
tionary of Matlatzinca.
In pre-modern times, suprasegmentals were not always recognized or
described. The tones in Oto-Manguean languages were described relatively
late, and before the grammars of the Jesuits Antonio del Rincón (1595) and
Horacio Carochi (1645) vowel length and the presence of the glottal stop as
a phoneme were not described systematically.37 This underdifferentiation
often led to ambiguity, or gave a false image of polysemy when words which
were not in fact homophones were written as homographs. One example is
Molina’s entry auatl, which is translated as ‘enzina, roble, gusano lanudo.
o espina’ (‘oak, woolly caterpillar, or thorn’). This entry represents three
different Nahuatl nouns: āhuatl ‘oak’, āhuātl ‘woolly caterpillar’, and ahhuatl
‘thorn’ (where the first of the h’s represents a glottal stop).38
Apart from these questions of how to reproduce the phonology of
Mesoamerican languages with the Roman alphabet, the organizational pro-
blems confronted by the missionary lexicographers were morphological: the
concept of the root; polysyntheticity and affixation; the particle; the treat-
ment of derivational processes. After these have been discussed, this section
will conclude with some remarks on citation forms, markedness, and the
structure of the dictionary entry.

35
Fuller accounts include Smith-Stark, ‘Phonological description’, 12–21, and Monzón,
‘Tarascan lexicographic tradition’, 183–5.
36
Smith-Stark, ‘Phonological description’, 17; M. Dürr and F. Sachse in Diccionario k’iche’
de Berlín (2017), 21.
37
Karttunen, ‘Nahuatl lexicography’, 2659; see also Smith-Stark, ‘Phonological descrip-
tion’, 15 (Rincón and the glottal stop), 23 (Rincón, Carochi, and vowel length), 24–6
(tone).
38
Molina, Vocabulario en lengua castellana y mexicana (1571), fo. 9r, cited in Karttunen,
‘Nahuatl lexicography’, 2658.

587

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

As for the concept of the root, the Tarascan lexicographical tradition can
be distinguished from the others, since missionary lexicographers working in
this tradition started from the beginning to distinguish roots from words.
Although Molina describes the distinction between the roots and the ‘servile
letters’ in his Nahuatl grammar, there is no systematic treatment of roots in
his dictionaries. Gilberti’s Vocabulario, by contrast, has a macrostructure
designed to engage with the presentation of roots. Its first section contains
the Tarascan words with Spanish translations, followed by an intermediate
section which is a list of ninety-two roots, followed by the Spanish–Tarascan
section. An addendum contains 123 additional Spanish–Tarascan items.39 The
intermediate section containing roots laid the foundations for later lexico-
graphers. The section is preceded by a short introduction, in which Gilberti
informs his readers that ‘the following contains certain verbs in alphabetical
order, which some people wish to call roots, because it seems that, separated
from the limbs or, better, from those that do service, the root will be left
without meaning, like a trunk that has no branches, but is ready to produce
them’.40 The second dictionary of Tarascan, by Lagunas, also describes roots:
‘they are like a base or a foundation to build upon, or like roots ready to
produce, or first and true etymological positions, on which are constructed
and built or produced the true edifice or productive branches for the compo-
sition of verbs, verbal nouns, and adverbs with the elements of the
interpositions’.41
Although the concept of the root had already been mentioned by Molina
and had been explored further by Gilberti, Lagunas has a very creative and
original lexicographical approach. He explains in his prologue that he marks
the root with the symbol of a cross, after which the lemma starts with the
root in Tarascan, followed by its translation. Next, more complicated words
are given, and Lagunas gives in the margin in italics the several affixes in
alphabetical order, which are explained separately (he often called them

39
Monzón, ‘Tarascan lexicographic tradition’, 173.
40
Gilberti, Vocabulario en lengua de Mechuacan (1559), fo. 80r, ‘S I G V E N S E C I E R T O S V E R B O S
P O R E L A L - | phabeto, a los quales algunos quieren llamar rayzes: porque parece que
apartados los miembros, o para mejor dezir las seruiles quedara la rayz sin significar
nada, como el tronco sin ramos: solamente dispuesto a producirlos’, translated in
Monzón, ‘Tarascan lexicographic tradition’, 174.
41
Lagunas, Arte y dictionario, 38 (second sequence of pagination), ‘son como fundamento,
o vasas para edificar, o como rayzes aptas a produzir, o primeras posiciones ethymo-
logicas. i. verdaderas, sobre quien se arman y edifican, o produzen el verdadero edificio,
o ramos productiuos de la composicion en los verbos y nombres verbales y aduerbios,
mediante los materiales de las Interposiciones’, translated in Monzón, ‘Tarascan lexico-
graphic tradition’, 176.

588

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

‘interposiciones’, but we also find terms such as ‘particulas’ and

‘preposiciones’).42 So, for instance, the root andà appears in its place in the
alphabetical sequence of roots, marked by a cross, and translated into
Spanish.43 The sublemma which follows (not marked by a cross) is the
word andambezcanitomines which is analysed and translated, with the ‘inter-
posicion’ bez beside it in the margin. The next sublemmata, andacazcani and
andahcazingani, are similarly treated, with the ‘interposiciones’ cazca and
cazin beside them in the margin, and so on, so that the user can find
a given root in one alphabetical sequence, and can then see the ‘interposi-
ciones’ with which it can be combined in a secondary alphabetical sequence.
The forms given in the margin are also listed in a section of the Arte which is
in effect a separate glossary of ‘interposiciones’.44 Lagunas was aware that he
was innovating: as he observes in his prologue, his dictionary is ‘briefer and
more profitable, and [written] in a curious manner and in alphabetical order,
and not so much according to the letters, but rather in the profitable phrases,
following the Arte and “interposiciones”’.45
The problems presented by polysyntheticity and affixation are summed up
in an example given in Richard Andrews’ Introduction to Classical Nahuatl:
Molina’s translation of Spanish monacordio (‘clavichord’) into Nahuatl is
petlacalmecaueuetl.46 This ‘word’ is in fact a complete phrase, meaning literally
‘it is an upright drum with strings that has the form of a wickerwork coffer’.
In Nahuatl, the absolutive ending -tl always indicates nominal predication,
whose translation is a phrase in English. For a lexicographer, it will not make
any sense to include separate lemmata for the forms in polysynthetic lan-
guages which correspond to English phrases like ‘I see the house’, ‘I see the
mountain’, ‘I see the forest’, and so on. The word ‘dictionary’ is in fact
a misnomer when polysynthetic languages are concerned – or at least,
a dictionary of a polysynthetic language must be, as has been said in the
context of Inuktitut, ‘a dictionary without words’.47 It has to be observed that
the concept of dictionarium is in fact a collection of dictiones, and if we follow

42
Lagunas, Arte y dictionario, 3 (second sequence of pagination).
43
Lagunas, Arte y dictionario, 10 (second sequence of pagination).
44
Lagunas, Arte y dictionario, 144 (first sequence of pagination), ‘De las interposiciones que
ya comiençan por su orden Alphabetico’.
45
Lagunas, Arte y dictionario, 2 (second sequence of pagination), ‘Mas breue y prouechoso
que ningun otro dictionario, y en modo curioso y Alphabetico, y no tanto en las letras
como e[n] las prouechosas sentencias, siguiendo al Arte y interposiciones.’
46
Molina, Vocabulario en lengua castellana y mexicana (1571), s.v. monacordio (petlacalme-
caueuetl is, conversely, a headword in the Nahuatl–Spanish section); Andrews,
Introduction to Classical Nahuatl, 20.
47
Cornillac, ‘De la necessité de concevoir’, 206, ‘un dictionnaire sans mots’.

589

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

Priscian’s definition of the term dictio – pars minima orationis constructae (‘the
minimal part of a constructed sentence’) – there would be no place in
a dictionary for smaller entities, such as ‘clitics’, ‘particles’, or ‘affixes’ (see
below), nor for larger entities, such as the constituent or the phrase.
Agglutinative languages with (mainly) suffixation did not cause major
problems for missionary lexicographers, since suffixation did not affect the
alphabetical sequence. By contrast, prefixation does affect the alphabetical
sequence, and this problem had to be solved, in particular in the dictionaries
which translated from the indigenous language to Spanish. So, for instance,
compiling the Spanish–Matlatzinca section of his dictionary was an easy task,
according to Basalenque. With ‘our Spanish dictionary’ he went to ‘the
bilingual natives’, and started asking them to translate.48 It was much more
complicated to make the inverted Matlatzinca–Spanish section. Matlatzinca,
according to Basalenque, is different from other languages of the region (he
probably compared Matlatzinca with Tarascan, which has mainly suffixes)
since it uses iniciales, or prefixes. There is no infinitive form of the verb which
could be used as the citation form of the lemma. Basalenque found a solution,
and the structure of the entries of this section is unique: the entry for each
root is preceded by three columns, in which he fills in the prefixes which can
be combined with the root. Nouns often, according to Basalenque, take
single prefixes, such as ca, hue, huebe, huebu, and huebete, explained in the
prologue of the Matlatzinca–Spanish dictionary.49 Verbs can take one, two,
or three prefixes. So, an entry for a root which is used in the composition of
verbal and nominal forms might look like this:

qui tu tzitzí yo como cosa de fruta (‘I eat something, as a fruit’)

qui tu tu tzitzí yo doy de comer a otro (‘I give someone else to eat’)
in tzitzí la comida (‘the food’).50
We now turn to the question of the particle, an element which, as Gerda
Haßler observes, falls between grammar and dictionary.51 In most
Mesoamerican missionary grammars of this period, we find a main section
devoted to the eight parts of speech. The final section of the grammar is often
devoted to ‘particles’ or addictiones, an addictio in this sense meaning an
element which is used in combination with a dictio (compare adverb and
48
Basalenque, Arte y vocabulario de la lengua matlaltzinga (1975), 145, ‘nuestro vocabulario
castellano fui preguntando a los naturales ladinos’.
49
Basalenque, Arte y vocabulario de la lengua matlaltzinga (1975), 147.
50
Basalenque, Arte y vocabulario de la lengua matlaltzinga (1975), 318.
51
Haßler, ‘Las partículas entre la gramática y la lexicografía’, 87.

590

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

adnoun). Juan de Córdova’s Arte closes with a list of seventy of these addic-
tiones, arranged alphabetically, some of them labelled as particles and others
as adverbs or interjections (there is no systematic classification here, but most
do not occur in the preceding sections devoted to these indeclinable parts of
speech). Basalenque’s treatment of particles, in a ‘Tratado de las partículas’,
with its own half-title page and prologue, between the ‘Arte’ and
‘Vocabulario’ in his guide to Matlatzinca, is, as far as I have been able to
trace, unique in this period in the tradition of New Spain. After this treatise,
the dictionary does not start immediately: several chapters follow, including
a supplement to the ‘Arte’, so that it is difficult to ascertain whether
Basalenque viewed ‘Tratado’ and ‘Arte’ as completely distinct, but it is
more independent from its grammatical context than any other contempor-
ary treatment of particles.
For an example of the missionary lexicographers’ engagement with the
multiplicity of derivational forms in some Mesoamerican languages, we turn
to the Zapotec dictionary of Juan de Córdova. As we saw above, this is the most
extensive of the early Mesoamerican dictionaries with Spanish as a source
language. In the first aviso (‘monition’) in his prologue, Córdova explains that
his dictionary is in fact ‘more copious’ than others. The main reason is that the
speakers of Zapotec often have an enormous amount of terms in a given
semantic field, which are not known in Western languages. He gives the
example of the words for different sounds, such as the sounds for snakes
when they move along, of birds, the beating of the heart, the cooking of
water, and all possible ‘interjections’ humans can produce. Although the number
of entries in Nebrija’s dictionary related to sounds is impressive, the Zapotec
dictionary adds even more entries that are not drawn from Nebrija. The second
aviso of the prologue explains how the dictionary deals with derivational
morphology. According to Córdova, it would not make sense to give all the
‘derivations and compositions’ for each verb, which would make his dictionary
endless: almost forty ‘vocables between nouns and verbs’ can be derived from
one single ‘principal verb’. He includes a paradigm of the verb whose first-person
singular is tol lŏbaya ‘I sweep’, with a list of all these forty forms.52 This example
demonstrates that Córdova was developing his own strategies and lemmatiza-
tion procedures, and supported his choices by theoretical explanations, different
from Antonio de Nebrija, who also includes derivation patterns quite system-
atically, but without any theoretical explanation in his prologue.

52
This is found in the dictionary and in the grammar: Córdova, Arte en lengua zapoteca,
fo. 36r.

591

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

The citation forms of verbs presented a more simple problem.53 In Nebrija,

the normal citation form is the infinitive. Since the infinitive is often lacking
in Mesoamerican languages, the first-person singular is usually given, as in
Latin: this was done by Molina for Nahuatl, by Córdova for Zapotec, and by
Alvarado for Mixtec. In Tarascan, by contrast, the lexicographers could find
an equivalent for the infinitive, and, since this form was available, they
selected it as the citation form. Molina also ‘provides information about
transitivity and reflexivity by listing after the verb a sample set of the prefixes
it may take’, so that the verb ihtoā ‘to say something’ is listed in three
successive entries with different prefixes and the appropriate different
forms of the preterite.54
Missionary lexicographers might provide information about features such
as frequency, style, origin, dialect, or sociolect, although their marking was
unsystematic. From Molina onwards, they showed an innovative interest in
metaphorical usage.55 Diaevaluative information is sometimes included (for
instance, when insults are translated, the lemmata can allude to appreciative,
derogatory, or offensive meaning, or taboo words). The use of the expres-
sions ‘lo mesmo’ or ‘idem’ indicates that in the indigenous language no
equivalent is given, and that instead the Spanish loanword is used. In
Molina, this occurs in about 200 lemmata.56 In an anonymous dictionary of
Kaqchikel and K’iche’, a cross is used when a form is not only Kaqchikel but
also K’iche’. Marking of dialectal or linguistic varieties is not, however, the
rule, although diatopical variation is often described in missionary grammars.
So, although the grammar of Mixtec of Alonso de los Reyes describes a great
number of diatopical varieties in the prologue, it is surprising that in
Alvarado’s dictionary of Mixtec, which appeared in the same year, almost
nothing is found related to dialectal variation, beyond a preliminary reference
to Alvarado’s desire to serve people of the high and low parts of the Mixteca
region.57 Grammatical categories were also marked unsystematically, for
instance to distinguish a as a pronoun from a as an interjection.58

53
Overview in Smith-Stark, ‘Lexicography in New Spain’, 60–1.
54
Karttunen, ‘Nahuatl lexicography’, 2658, citing Molina, Vocabulario en lengua castellana
y mexicana (1571), fo. 43r (second sequence of foliation).
55
Hernández, ‘La marca de uso metafórico’; see also Clayton and Campbell, ‘Alonso de
Molina as lexicographer’, 347–8.
56
Karttunen, ‘Nahuatl lexicography’, 2658.
57
Alvarado, Vocabulario en lengua mixteca, fo. 3r, ‘aprouechar a los naturales de los pueblos
de la Misteca alta y baja’; cf. Reyes, Arte en lengua mixteca, sig. ¶7r, ‘las grandes
differencias y modos distinctos de hablar esta lengua’.
58
M. Dürr and F. Sachse in Diccionario k’iche’ de Berlín (2017), 27.

592

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

Since Nebrija was undoubtedly the most important source for most
lexicographers, his entry style was also the most common one in dictionaries
translating from Spanish to a Mesoamerican language. In the tradition of
Mayan lexicography, however, we find totally different lemmatization pat-
terns. One example which illustrates this is Dürr and Sachse’s analysis of the
internal structure in the Vocabulario en lengua 4iche otlatecas.59 In the lemma,
the grammatical function of the word is often indicated first. When verbs are
analysed, the author often adds information related to the conjugation class,
combined with information on the first-person prefixes for transitive and
intransitive verbs. Then, the equivalent meaning of the lemma is given in
Spanish. In most entries, whole phrases are given as examples where the
headword is used in a specific context, and when verbs are discussed several
derivations of the verb are added as well. In other lemmata we also see
related meanings, sometimes antonyms, with sketches of semantic fields
related to the headword. We often also find observations related to ethno-
graphy. It is remarkable that many K’iche’ examples are not translated into
Spanish.

New Tools and Pedagogical Approaches

In New Spain, during the first two centuries, the documentation, teaching,
and learning of the indigenous languages concentrated on the trilogy of
dictionary, grammar, and catechism. These three genres were still in use
until the end of the colonial period, but we see also that authors composed
new learning methods. Most grammars and dictionaries in New Spain were
composed by and for missionaries, but there is an exception: Pedro de Arenas
published his successful Vocabulario manual of 1611 for use in a wide variety of
communicative situations, as the subtitle shows: it contains ‘the most com-
mon and ordinary words, questions and answers, which are used in commu-
nication between Spaniards and Indians’.60
Only a few missionaries followed Arenas’ method. Miguel de Guevara’s
‘Arte doctrinal y modo general para aprender la lengua Matlaltzinga’ of 1638 is
one example. His method of teaching the Matlatzinca language brings para-
digms and wordlists together with idiomatic expressions (‘everyone is laugh-
ing at you!’), proverbs (‘the one who lives well is always as lovely as a rose’),

59
In Diccionario k’iche’ de Berlín (2017), 21–2.
60
Arenas, Vocabulario, title page, ‘palabras, preguntas, y respuestas mas com[m]unes,
y ordinarias que se suelen ofrecer en el trato y communicacion entre Españoles é
Indios’.

593

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

and similes (‘the priests are walking from one village to another, like birds,
today here and tomorrow there’).61 Another remarkable text which does not
follow the strict genres of the grammar, the dictionary, and the religious text
is Diego de Nágera Yanguas’ Doctrina y enseñanza en la lengua maçahua of 1637.
The grammatical section is presented as a series of advertencias (‘remarks’),
rather than as a traditional grammar. As the title Doctrina y enseñanza indi-
cates, the work is a mixture of grammar, texts for confession, liturgy, and
a method for learning to speak in daily situations. As Dora Pellicer has put it,
introducing a neologism for a new genre, Nágera Yanguas did not compile
a ‘diccionario’ but a ‘conversacionario’.62 The method is important seen from
a pedagogical standpoint: learners could read the examples aloud, or they
could learn them by heart. I am not aware that there circulated a dictionary
parallel to Nágera Yanguas’ method, since no dictionary of Mazahua has been
preserved. What we know is that Nágera Yanguas included lexicographical
sections, arranged thematically, in his work, for instance numbers, kinship
terms, names related to the house, names of places in the city, and words to
do with sewing, colours, and the parts of the body. The inclusion of a list of
toponyms is not a very common practice, although there are earlier exam-
ples, such as the Mixtec grammar of Alonso de los Reyes, which has several
lists enumerating regions. However, Nágera Yanguas did not copy from
Reyes, but made his own list.
Augustín de Quintana’s description of Mixe in his Confessonario en lengua
mixe, con una construccion de las oraciones de la doctrina christiana, y un compendio
de voces mixes is another experiment from the eighteenth century. His work is
not a traditional ‘Arte’ which has to be combined with a ‘Vocabulario’ and
religious texts in the teaching of Mixe, but a mixture of bilingual texts, detailed
annotations, wordlists, and grammatical rules. The verbal index to Quintana’s
bilingual Instruccion christiana is, in effect, a Spanish–Mixe theological wordlist.

Content
The final topic to be analysed in this chapter is the content of the dictionaries.
Again, they are far from uniform. In the more encyclopedia-like dictionaries
we not only see many Western concepts which are translated into other
languages, but also the other way around.

61
Guevara, ‘Arte doctrinal’ (1862–3), 218–19, ‘Todos se ríen de tí . . . el que vive bien
siempre está hermoso como la rosa . . . los sacerdotes andan de pueblo en pueblo como
los pájaros hoy aquí y mañana allí.’
62
Pellicer, ‘Confesión y conversación’, 31.

594

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 Missionary Traditions in Mesoamerica

Concepts from the Old World might be carried over into dictionaries of
Mesoamerican languages. So, the entry mezquita (‘mosque’) in Nebrija is
also included in the 1571 edition of Molina’s Arte, where we even find
Nahuatl translations: Mahomacalli (‘it is the house of Muhammad’) and
˙
Mahomatlatlatlauhtilizcalli, a polysynthetic construct combining the root
Mahoma with tlatlatlauhtiliztli (‘prayer’) and calli (‘house’). Other lexicogra-
phers might explain mezquita with forms which included Mahoma, as
Urbano did; or, in a spirit of hostility to Islam, as ‘house of the devil’,
using the Spanish word diablo ‘devil’ as a loanword, as Gilberti did; or using
forms which included words for ‘devil’ in a Mesoamerican language, as
Córdova did.63 Loanwords, calques, and literal translations were likewise
used to express Christian vocabulary by lexicographers of K’iche’.64
Mesoamerican gods and goddesses were compared with Greek and
Roman analogies – for instance, Huitzilpochitli was ‘another Mars’ – and
Nebrija’s entries for kinds of divination in ancient Rome became the basis
for entries in Córdova’s dictionary which coined Zapotec words for what
were in fact divinatory techniques of the ancient Romans.65
On the other hand, missionary lexicographers often expanded an entry from
Nebrija, adapting it to the the New World context. Nebrija’s entry altar donde
sacrifican (‘altar where they sacrifice’) became two entries in Alvarado’s Mixtec
dictionary, a Christian altar being chiyo (originally ‘foundation’) and a pagan
one being tayu quacu or tayu dzana. In Córdova’s Zapotec dictionary,
a Christian altar became pecógoláya nitaca missa (‘throne where Mass is
held’).66 Using these translations, pre-Hispanic models were preserved.
Likewise, Nebrija’s relumbrar, o reluzir (‘shine’) is given four different Nahuatl
equivalents in Molina’s dictionary, expressing Nahua categories of reflected
light (tlanextia, peptlaca, pepetzca, tzotzotlaca), so that ‘a whole world of visual
distinctions is opened up by these definitions: the luster of silk and feathers is
distinguished from the sun glittering on the water and the sparkle of precious
stones’.67

63
Urbano, Arte breve dela lengua otomí y vocabulario trilingüe (1990), s.v. mezquita, has
naxæcãmbemcangũ´ . magũ´ nquexæcambenimahoma, which is a calque of the Nahuatl trans-
lation by Molina (mahoma is the final element). Gilberti, Vocabulario en lengua de
Mechuacan, s.v. mezquita, has diabloeueri quahtaqueri (‘house of the devil’; elsewhere,
quahta glosses casa ‘house’). Córdova, Vocabulario Castellano–Zapoteco (1942), s.v. mez-
quita, has lìchi pezè láo (‘house of the devil’; elsewhere, lichi glosses casa ‘house’ and
pezèlàotào glosses diablo ‘devil’).
64
Diccionario k’iche’ de Berlín (2017), 37–40. 65 Hamann, Translations of Nebrija, 86–91.
66
Hamann, Translations of Nebrija, 91–6. 67 Hamann, Translations of Nebrija, 98–9.

595

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press

 otto zwartjes

Missionary lexicographers generally included more specific information

related to the indigenous cultures in their dictionaries which translated from
the indigenous language to Spanish. A clear example is the Vocabulario
otlatecas of K’iche’, which contains lemmata related to food, local religion,
political and social organization, daily life, agriculture, how to participate in
a social event, such as buying meat, or how to have a conversation or
discussion about a certain topic.68
As regarding their content, we can conclude that missionary dictionaries
oscillate between two extremes: on the one hand they tried as best as they
could to translate Western concepts, imposing their own culture on the
indigenous languages, but on the other hand, most frequently in the diction-
aries which translate from the indigenous language to Spanish, we find
meticulous and surprisingly detailed information related to the culture of
the other.

68
Diccionario k’iche’ de Berlín (2017), 41–8.

596

https://doi.org/10.1017/9781316827437.028 Published online by Cambridge University Press