How to “Fake” a Language

Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 HOW TO “FAKE” A LANGUAGE Lyle Campbell Abstract On the basis of six cases, general observations are made about attempts to fabricate a language to deceive a fieldworker. Key Words: fake languages, fieldwork, Pipil, Aguacateco II, Central America. Resumen Basado en seis casos, se hacen observaciones generales sobre intentos de falsificar lenguas para engañar al investigador. Palabras clave: lenguas falsificadas, trabajo de campo, pipil, aguacateco II, América Central. 1. Introduction In the course of several years of fieldwork in Central America and Mexico seeking potential speakers of endangered languages, I encountered on several occasions individuals who attempted to fabricate a language, to create spontaneously what they hoped I would take to be an indigenous language. The number of cases in the sample considered here is not large; nevertheless, the goal of this paper is to attempt to make some general observations about how these individuals have attempted to fake a language. It may be valuable to be able to spot a fake and to distinguish it from real languages, particularly since new, previously unidentified languages have continued to tune up in this region. For example, Terrence Kaufman discovered Sakapulteko (Sacapultec), Sipakapense (Sipacapa) (Kaufman 1976), and Teko (Teco) (Kaufman 1969), three new Mayan languages; I discovered Jumaytepeque, a previously unknown Xinkan language (cf. Campbell 1979); and Roberto Zavala Maldonado (2014) recently discovered a previously unknown Mixe-Zoquean language of the Zoquean branch in Chiapas, Mexico. Clearly, then, one cannot conclude that anything different or unexpected must be a fake. The fakes I have encountered appear to exhibit defining characteristics ________ Universidad de Hawaii <[email protected]> Recepción: 06/06/2014- Aceptación: 19/06/2014 64 ESTUDIOS DE LINGÜÍSTICA CHIBCHA that distinguish them readily from real languages, and these earmarks alone are sufficient for distinguishing the deceptions from real languages. 2. The sample The sample considered has six cases of individuals who have attempted to fake a language. These include: Fake “Pipil” of Comapa, Guatemala Fake “Pipil” of Panchimalco, El Salvador Unidentified “language” of La Trinitaria region, Chiapas, Mexico Las Cumbres, Chiapas “mystery language” Guatemala “mystery language” Aguacatec II (Stoll 1884) This may seem to be a small number, particularly upon which to base any generalizations, but at the same time also a relatively large number, given that faked languages are not generally expected. However, I have interviewed the hundreds of consultants over many years while seeking potential speakers of languages once reported to have been spoken in these areas and now either extremely critically endangered or already extinct – though that remained to be established at the time of the investigations. Given the large number of potential speakers of various languages that I consulted, the number of actual attempted language fakings seems not very remarkable to me. Also, since all the cases considered here are from Central America and Mexico, it is uncertain whether the dominant campesino culture, to which all these fakers belong, may affect their behavior and therefore constrain the generalizations drawn from the sample. Thus it is not clear, if similar kinds of language fakings occurs in other parts of the world, whether they would have similar or different characteristics. 3. Motives? Why would a person try to fabricate a language, why one would engage in such deception? Whatever the full range of motives might be, two significant ones almost certainly involve money and status. Most of the cases considered here involve persons hoping to make money from their deception. Since it is often difficult to make contact with and establish confidence with potential speakers in situations where the language may no longer have any speakers or has extremely few, one strategy I sometimes used for helping to overcome these difficulties was to let it be known in a town or an area of interest that information about any local indigenous language was extremely Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 CAMPBELL: How to “fake” a language 65 important, important enough to pay more than local standard wages for people who know the language to talk with me. This strategy was often useful for making contacts and being able to work with real speakers, but it also on rare occasion attracted fakers who hoped to cash in on the funny gringo’s strange interest. The other main motive was personal status, prestige. In two cases, particularly elderly individuals had a local reputation for being the repository of local cultural and linguistic knowledge, a reputation which gave them a status in their communities that they clearly enjoyed. For example, in Comapa, Departament of Jutiapa, Guatemala (near the El Salvador border) Pipil (Nahuate, a Uto-Aztecan language) had been spoken formerly, but no longer had speakers there. When I visited Comapa in the 1970’s to determine whether any speakers of the language might yet be found, one elderly woman, Doña X (I do not give her real name for reasons of protecting privacy) enjoyed the position of being a symbol of the lost indigenous heritage of the town – she greatly enjoyed this position and the townspeople respected her as an important representative of their heritage and past. Unfortunately, Pipil had ceased to be spoken there at least two generations before her time. She knew only five words of Pipil, learned from her grandmother, who was also not a speaker of the language but who remembered some words just because the novelty had interested her. With great enthusiasm on the part of townspeople, I was taken to Doña X, and Doña X’s reputation as a venerated representative of the town’s Indian past was at stake. To save face, she attempted to fake answers to my wordlist questions about how do you say such-and-such in the language. That she did not know the language and that she was attempting to invent answers became apparent almost immediately (as took place with all the other fakes I encountered, also – see more details below). Another motive that I have sometimes suspected but for which I have no direct evidence might be “recreational”: I suppose it is possible that someone might gain enjoyment from a practical joke of trying to put one over on the foreigner. I have not detected this motive in the cases that I encountered, though perhaps I would not recognized it, if it were present. I have also heard rumors from other parts of the world of purposeful deception in order to lead the linguist off the track, away from the real language, which the fakers wished to conceal. I have no experience with that kind of deception. 4. Characteristics of faked languages It is to be assumed that some fakers may be more creative, more persistent, or just smarter than others, and so that there may be variation in the kind and number of faked forms different fakers might offer a researcher. Nevertheless, it seems that it can Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 66 ESTUDIOS DE LINGÜÍSTICA CHIBCHA accurately be asserted that the faked language material will depart very little if at all from the phonetics and structural properties of the dominant language(s) spoken by the “faker,” although the faker’s behavior usually does depart from the patterns of behavior of consultants with real knowledge of an endangered language. True consultants tend to have difficulty remembering certain words, and when asked, strain to remember. They often ask for clarification of the question asked. Initially they tend to have more lapses of memory and strain more to recall, but remembering some linguistic material often seems to stimulate the recollection of additional things in the language so that their performance at answering questions sometimes improves as the interview progresses. Very rarely do fakers volunteer forms that have not been asked for, whereas sometimes with consultants of real languages, one elicited form will trigger memory of other associated word or construction which the consultant then volunteers. The fakers I have experienced typically exhibit just the opposite behavior. They usually express confidence initially, they do not ask for clarifications, and initially they show no signs of struggling to remember, with little hesitation (except to imagine fabricated forms). However, as time goes on in the interview, the fakers exhibit greater and greater difficulty in keeping up the act, in coming up with additional created stuff – in short their performance gets worse as the interview progresses, with less confidence and more hesitation. 4.1. Phonology In almost all the cases considered, the faked material corresponded phonologically and phonetically exactly to Spanish, the dominant language of these fakers, and contained no non-Spanish phonetic material and violated none of the syllable patterns or canonical shapes of Spanish morphemes. Only in one case did a faker produce anything with a non-Spanish sound. In this case, the faker lived in an area in southern Chiapas where there were Tzeltal speakers and he produced some forms with glottalized (ejective) ts’ and č’, though not with the other sounds of the glottalized series of Tzeltal (p’, ɓ, t’, k’) (examples below). 4.2. Vocabulary As mentioned above, fakers run out of steam after a short while, not able to continue to come up with new “words.” They often resort to repetitions of already offered phonological sequences but with different supposed meanings from when first offered. They often crib from Spanish, sometimes offering archaic or rare Spanish words they suppose the foreign interviewer may not know, or they revert to words that are essentially Spanish, but changed somewhat in form to make them not identical to Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 CAMPBELL: How to “fake” a language 67 Spanish. To cite just one example (see below for others), when asked for the equivalent of lechuza ('barn owl'), one faker gave [áβis], essentially equivalent to Spanish avis from aves 'birds'. Nevertheless, vocabulary fabrication seems to be the main thing fakers attempt to do, even if they are unable to keep it up for very long. Very characteristic of the behavior of these fakers is that when something already asked for and provided earlier in the interview is re-elicited, the consultant very rarely can repeat exactly what was given earlier, often giving something with no connection whatsoever to the form offered when elicited the first time. Some fakers do sometimes include the occasional “word” that appears to have its origin in some indigenous language. This is not surprising, however, since those who grew up in a region where some indigenous language was spoken could pick up the odd real word from the language. What is perhaps more surprising than the presence of the occasional indigenous looking word is the fact that there are so few such words in these cases of faked languages. (See examples below.) 4.3. Morphosyntax The faked languages I have encountered exhibit none of the recurrent parts of “words” that might be associated with inflectional or derivational morphology in true languages. Fakers seem incapable of fabricating morphology. Some may rarely be able to invent something odd when asked for plurals, but if asked for what might be paradigmatically related forms, say verbs with different pronominal subjects or with different tenses or aspects, virtually nothing consistent or coherent comes out. When asked for longer phrases or sentences, most fail completely after only a couple of feeble attempts, and whole sentences or longer phrases are not volunteered. 5. An example: case study To illustrate language faking, I present here the entire corpus collected in an interview from one faker, before the interview was ended. Most of the other cases were remarkably similar to this one, in length and in content. This one is from Las Cumbres, Chiapas, from Don X (identity withheld for privacy reasons), who was 48 years old in 1985 when this was collected. He reported that his father had also spoken this language, though what he produced was clearly faked. He could not name the supposed language, and I have just called it “Chiapas [faked] mystery language.” This is an interesting case, because Don X was the only faker encountered able to produce any sound not limited to what is found in Spanish. He had lived his whole life in a region where Southeastern Tzeltal had been spoken (Campbell 1987); he was not a speaker of Tzeltal, but nevertheless produced forms with glottalized Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 68 ESTUDIOS DE LINGÜÍSTICA CHIBCHA (ejective) ts’, č’, and k’ (examples below), though the other sounds of the glottalized series of Tzeltal (p’, ɓ, t’) do not appear in his “words”. He also produced forms with glottal stop ([ʔ]), a phoneme of Tzeltal but only marginally present in local Spanish, as in semi-emphatic pronunciations of [siʔ] sí 'yes' and [noʔ] no 'no'. Note that where Standard Spanish lacks /š/, the Spanish of Guatemala and Chiapas has /š/, spelled <x>. The entire corpus, with commentary on the forms follows. Note that multiple forms for the same gloss indicate instances where the equivalent to the gloss was given more than once during the interview. The items are given first with the form or forms given for the fake mystery language, spelled phonetically, then spelled in Spanish orthography in parentheses, if an equivalent in Spanish orthography is available. This is followed by the English gloss, and then by the Spanish gloss which was used to elicit the form or forms in square brackets ([...]), and this is followed finally by commentary if any is relevant. Note that here the Americanist phonetic symbols most commonly used in linguistic descriptions of Mesoamerican languages are employed. The equivalents in the IPA (International Phonetic Alphabet) where symbols differ are: Americanist š č y r r̃ IPA ʃ tʃ j ɾ r Stress is marked as an acute accent ( ́ ) over the vowel, rather than as a small straight line ( ̍ ) before the stressed syllable. These forms are presented in the order in which they came up in the interview with the sole exceptions of putting seemingly related forms adjacent to one other for easier comparison. miš (mix) 'cat' [gato] (See also mišmul 'cat' below.) Spanish mix is used to call cats; mix is also behind a loanword from Spanish for 'cat' in a number of indigenous languages of Mesoamerica; cf. Tzeltal mis 'cat'. kaw (cau) 'horse' [caballo] No indigenous language of the region has /kaw/ for 'horse' but some of the languages have longer words that involve loanwords from Spanish caballo 'horse', which start out with kaw, e.g. Tzeltal kawayu 'horse'. mayóx (mayoj) 'cattle' [ganado] pikté, bikté (picté, bicté) 'chicken' [gallina] Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 CAMPBELL: How to “fake” a language 69 gwáro (guaro) 'cigarette' [cigarro] In local Spanish guaro means 'hard liquor'. This is an example of substituting some other Spanish word somehow associated with the meaning of the item asked for from the wordlist, probably in hopes the foreign interviewer might not notice it. gwáro (guaro) 'I smoke (it)' [lo fumo] The form given first for 'cigarette' was repeated later in the interview for the meaning 'I smoke'. It might be noted that none of the indigenous languages of this region has a native /r/ phoneme. wišts’iy 'chayote' [chayote] This is one of the few forms with a glottalized consonant. It appears to be a deformed form suggested by Spanish huisquil, an alternate name for chayote common in Guatemala and parts of Chiapas. makaniár (macanear) 'to plant' [sembrar] Macanear is a Spanish verb which generally means 'to exaggerate', 'to tell tall tales', 'to lie', but in some places it has a connection with 'to clear weeds'. The change of unaccented e to i is very common in non-standard varieties of Spanish. r̃éyna (reina) 'maize, corn' [maíz] In Spanish, reina is 'queen'; possibly 'maize', extremely importance to subsistence in the region, suggested something like 'queen (of cultivated plants)' to Don X. aβéxa (abeja) 'wasp' [avispa] Abeja is just 'bee' in Spanish. čapulín (chapulín), briŋkasakáte (brincazacate) 'grasshopper' [saltamonte] Chapulín is a common word for 'grasshopper' in Mexico. Brincazacate (also brinca zacate) is extremely rare but also means 'grasshopper' in small regions of Chiapas, apparently based on analogy in Spanish with saltamonte, another common word for 'grasshopper'. Saltamonte is composed of salta 'jump' + monte 'weeds, wild plants'; brincazacate contains brinca 'leap, jump' + zacate 'grass'. ts’iy̥, ts’iç (ts’ij) 'dog' [perro, chucho] This “word” has a voiceless “y” at the end, more or less equivalent to the palatal fricative [ç], though more approximant without much friction. It is the phonetic equivalent of Spanish <j> or <g> before [i] as spoken in this region (as in Spanish jinete 'rider, horseman'), though this does not occur word-finally in Spanish. However, this is one of only four “words” in the corpus with a glottalized consonant, [ts’]. It is actually quite similar to the word for 'dog' in several Mayan languages, including Tzeltal ts’iʔ 'dog', though it ends in glottal stop in Mayan languages, not in /y/ or voiceless /y/. Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 70 ESTUDIOS DE LINGÜÍSTICA CHIBCHA kič’éx 'deer' [venado] This is another example with a glottalized consonant, [č’]. It is vaguely suggestive of the word for 'deer' in several Mayan languages in this region, čih in Tzeltal, čex, čeh, or other cognates that reflect Proto-Mayan *kehx 'deer' in others, but no Mayan language has a glottalized consonant in the word for 'deer'.1 palo (palo) 'tree' [árbol] In Spanish everywhere palo is a colloquial word for 'tree', though in Standard Spanish the meaning of palo is usually restricted more to 'stick' or 'wood'. áβis (avis) 'barn owl' [lechuza] This example was mentioned above. It is a slightly modified version of Spanish aves 'birds', which apparently the question about 'barn owl' suggested to the mind of the consultant trying to invent words for the language on the fly. animális málus, animáles málos (animales malos) 'weasel' [comadreja] Here, as with the previous two examples and seen again in the next three examples, it seems the consultant was starting to run out of imagination for inventing new vocabulary items, reverting more to Spanish. This case is clearly Spanish animales malos 'bad animals', repeated first with very slightly modified final-syllable vowels and then later with no vowel modification at all (animal-es malo-s [animal-plural bad-plural]). mananéra, bananéra (mananera, bananera) 'banana' [guineo, plátano] This case clearly involves the Spanish word bananera, the feminine adjectival form of 'banana', as in república bananera 'banana republic'; the masculine bananero means 'banana plant', 'banana tree', 'banana plantation'. The form here was offered first with a deflected sound, with initial b modified to m, and then later repeated but with no modification at all, with just initial b. sústo (susto) 'I am afraid' [tengo miedo] This is just straightforward Spanish susto 'fright'. kwéntas, duðósas (cuentas, dudosas) 'to pay' [pagar] The two forms offered on different repetitions of the question are straightforward Spanish, cuentas 'accounts' and dudosas, the feminine plural version of the adjective dudoso 'doubtful, dubious', possibly suggested on a vague analogy to somewhat phonetically similar deuda 'debt', i.e. suggestive of what one pays. Finally, when the equivalent of Spanish tambor 'drum' was asked for, the consultant just gave a paraphrase, 'the thing that is beat for music'. Here it was clear the consultant was struggling severely to come up with anything new at all, and the interview was ended at this point. Not part of the interview based on a formal questionnaire (wordlist), Don X volunteered four forms – this is unusual in that none of the other fakers volunteered Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 CAMPBELL: How to “fake” a language 71 anything that was not asked for. These were volunteered initially when Don X was trying to convince me to accept him for this sort of work. The forms he offered in that context were: tuhmé (tujmé) 'to disfigure someone, make someone ugly' [afear una persona] mišmul 'cat' [gato] (See also miš 'cat' above.) As mentioned, this appears to involve Spanish mix, used to call cats, and mix or mis 'cat' is a Spanish loanword in several of the local indigenous language. k’ahnéʔ 'meat' [carne] This is apparently a partial imitation of Spanish carne [kárne] 'meat', with the sound deflected for disguise – one of only four forms with a glottalized consonant, k’ here. hwiʔmé 'to eat a chicken' [comer una gallina] None of the known indigenous languages of the region has a cluster /hw/ or segment /hw/, though this is very much like local Spanish words with sequences of <jui>, as in juicio [jwísio] 'judgement'. 6. Other faked languages and the case of Aguacatec II There have been a number of “phantom” languages – hoaxes, falsely identified languages, and faked languages – in the history of American Indian linguistics (for examples and discussion, see Campbell 1997: 13-15, 2012: 131-4). A well-known case is that of Aguacatec II, one quite similar to the fakes reported here. Aguacatec II, of central Guatemala, was made up in Antigua by Otto Stoll’s (1884: 166-8) ladina (nonIndian) maid Soledad Barrueto, who was from Aguacatán and claimed to have spoken this language in her youth. Today only Awakateko (Aguacatec), a Mayan language of the Mamean branch, is spoken in Aguacatán, and no non-Mayan indigenous language has ever been encountered there or in the region. Stoll mentioned 300 words that she had given him, but he presented only 68 of them, saying: The other expressions given to me by Soledad Barrueto are too suspicious to reproduce here. Probably they are based on self-deception of this girl who had lived through long years away from her home and not spoken the local Indian language.”2 However, most of the 68 he did present turn out to be very dubious, also. Ultimately, no one other than Stoll ever found anything remotely resembling Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 72 ESTUDIOS DE LINGÜÍSTICA CHIBCHA Aguacatec II in Aguacatán or anywhere else. There is no non-Mayan language (other than Spanish) anywhere near this region of Guatemala, near the proposed homeland of Proto-Mayan. The invented Aguacatec II exhibits forms that fit the characteristics of the faked languages reported here. Some examples of forms that Soledad Barrueto gave to Stoll which are suggestive of Spanish with somewhat modified pronunciation include: báiba 'water' (Wasser); cf. Spanish bebe 'drinks', beba 'drink!' topúy 'mouse' (Maus); cf. Spanish topo 'mole' cojóy 'coyote' (Coyote); cf. Spanish coyote 'coyote' maneré 'morning' (Morgen); cf. Spanish mañana 'morning' However, a very few of the forms given are similar to forms in indigenous languages of Guatemala, not surprising since many non-Indians of Guatemala know a few words from the local Mayan languages, and growing up in Aguacatán with its predominantly Awakateko-speaking population, Soledad Barrueto would no doubt have been exposed to some Awakateko: boch 'pig' (Schwein); cf. Awakateko ɓo:ch jolop 'egg' (Ei); cf. Awakateko k’oloɓ jus 'vulture' (Zopilote); cf. Awakateko quʔs matiox 'thanks' (Danke); cf. Kaqchikel matiox 'thanks' (contains the Spanish loan -tiox from Dios 'God'; Stoll called this a “Mexican [Nahuatl] expression”) cham 'sour' (Sauer); cf. Eastern Mayan *č’am 'sour'. Some forms contain sounds not found in any indigenous language of Guatemala, e.g. f, as in: chorífe 'meat' (Fleisch); reminiscent of Spanish chorizo 'sausage'. lorof 'dust' (Staub). papelif 'little' (klein). bayenef 'narrow' (eng). Many of the forms also have r, though this is not known in Mayan languages apart from those of the K’ichean subgroup of Mayan (several examples with r given above). Most of the forms are polysyllabic; Mayan roots are nearly all monosyllabic. This suggests that at least this is not a Mayan language, made up though it is. Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 CAMPBELL: How to “fake” a language 73 7. Conclusions Here I have reported on several attempts to fake a language to deceive a fieldworker. Although the number of cases is small, I have attempted to point out some commonalities they share. The faked language material will rarely depart from the phonetics and structural properties of the dominant language spoken by the “faker,” though it may contain some things from other languages the faker may be familiar with. Consultants with real knowledge of an endangered language, when interviewed, tend to have difficulty remembering words, strain to remember, ask for clarifications of the questions asked, and often begin to recall more as the interview progresses. Fakers rarely volunteer forms; with consultants of real languages, one elicited form may often trigger memory of another associated form which the consultant may volunteer. The fakers exhibit quite the opposite behavior from the consultants with knowledge of real languages. Initially, fakers typically express confidence, do not ask for clarifications, and initially do not struggle to remember, but as the interview continues, fakers have difficulty coming up with additional invented forms, and perform more poorly as the interview progresses. Fakers may repeat the same thing though the meaning asked for is different. They often rely on Spanish in one way or another, sometimes giving archaic or rare Spanish words, sometimes “words” that are essentially Spanish but changed in pronunciation to disguise them. Not uncommonly, the proffered form appears to be stimulated by something in Spanish that has some association for the faker with some other Spanish form offered in disguise (as in the case of avis, essentially from Spanish aves 'birds', offered when 'barn owl' was elicited). Very often the faker cannot remember what he/she volunteered for something elicited earlier in the interview; re-elicitation of the same thing typically results in something unlike what was given earlier. The faked material almost never involves anything identifiable as bound morphology and rarely anything of more that one “word” or occasionally a very short “phrase.” The interview wordlist elicitation sessions in no case lasted longer than about a half an hour – the fakers I have encountered are not capable of sustaining their inventive activity very long. In short, it turns out to be very easy to detect attempts to fake a language in these situations – they exhibit the characteristics pointed out here. Notas 1 The invention perhaps could have had some stimulation from the name of the language Kiche’, though that seems unlikely, given the huge meaning difference, and since Kiche’ is not spoken anywhere near this region. Interestingly but almost certainly accidentally, the word for ‘deer’ in Kiche’ is k’i:če:ʔ kye:x, Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014 74 ESTUDIOS DE LINGÜÍSTICA CHIBCHA from k’i:- ‘many’ + če:ʔ ‘tree’ (the compound word k’i:-če:ʔ ‘forest’ is the source of the name of the language) and kye:x ‘horse’ – literally ‘deer’ here is ‘horse of the forest’, where originally kye:x meant only ‘deer’; then when ‘horses’ were introduced, both deer and horses were called kye:x; then later, in order to distinguish deer from horses, deer came to be called is k’i:če:ʔ kye:x, literally ‘forest horse. However, in no Mayan language does ‘deer’ sound anything like [kič’éx]. 2 Die übrigen, mir von der Soledad Barrueto gegebenen Ausdrücke sind mir zu verdächtig, um sie hier zu reproducieren. Wahrscheinlich beruhen sie auf einer Selbsttäuschung dieses Mädchens, das durch lange Jahre von seiner Heimat entfernt gelebt und die dortige Indianersprache nicht mehr geredet hatte. References Campbell, Lyle. 1979. “Middle American languages”. In: Lyle Campbell and Marianne Mithun (eds.): 902-1000. _______. 1987. “Tzeltal dialects: new and old”. Anthropological Linguistics 29:549-70. _______. 1997. American Indian languages: the historical linguistics of Native America. Oxford: Oxford University Press. _______. 2012. “The classification of South American indigenous languages”. In: Lyle Campbell and Verónica Grondona (eds.): 59-166. Campbell, Lyle and Verónica Grondona (eds.) 2012. The Indigenous Languages of South America: A Comprehensive Guide. Berlin: Mouton de Gruyter. Campbell, Lyle and Marianne Mithun (eds.) 1979. The Languages of Native America: an Historical and Comparative Assessment. Austin: University of Texas Press. Kaufman, Terrence. 1969. “Teco – a new Mayan language”. International Journal of American Linguistics 35:154-74. _______. 1976. “New Mayan Languages in Guatemala: Sacapultec, Sipacapa, and others”. Mayan Linguistics 1:67-89. Stoll, Otto. 1884. Zur Etahnographie der Republik Guatemala. Zurick: Orell Füssli & Co. Zavala Maldonado, Roberto. 2014. “Polisíntesis en oluteco: una revisión de sus rasgos”. Ponencia presentada en Coloquio sobre Lenguas Otomangues y Vecinas, Oaxaca, Mexico, April 24-26, 2014. Estudios de Lingüística Chibcha (ISSN 1409-245X) 33: 63-74, 2014