Some Notes on the Inverse Voice in Mayan Languages and Epigraphic Mayan


David F. Mora-Marín
University of North Carolina
Chapel Hill



In a seminal article on the pragmatics of voice constructions in Akatek Mayan, Zavala (1997) identified, for the first time, the inverse voice in a Mayan language. The inverse voice is a type a voice construction defined functionally on the basis of its discourse contexts, in which the agent and patient are both topical, but the patient is more topical than the agent. Moreover, Zavala was able to show that there exist two voice constructions in Akatek that fit the pragmatic profile of an inverse: one that is conventionally described as a passive, and the other consisting of a “semantically transitive but syntactically intransitive” construction. Examples (1a-1b) illustrate the difference between an active transitive clause and an intransitive clause that is nonetheless semantically active, i.e. an inverse construction (Zavala 1997:458). Note that in (1a) the transitive verb ‘to carry’ is inflected with both ergative (Set A) and absolutive (Set B) person markers, as expected of an active transitive verb, and both ‘the boy’ and ‘the bull’ are core arguments of the verb. In (1b), in contrast, the verb is ‘to go’, an intransitive verb, and is inflected only with the absolutive person marker; also, there is only one core argument of the verb, ‘the boy’, and the semantic agent is now an oblique argument, expressed as the possessor of a relational noun -uu, used to express agents. In such inverse clauses the verb is thus intransitive and the agent must be expressed, albeit indirectly.[1]


(1)       Akatek (Zavala 1997:458)

            a.         Active transitive

                        max     Ø-y-ii                   toj                        naj      unin    no’      wakax

                        cmp     b3-a3-carry     dir:thither       ncl      boy      ncl      cow

                        ‘The bull took the boy away’.

            b.         Inverse

                        max     Ø-too              naj       unin     y-uu                no’      wakax

                        cmp     b3-go               ncl      boy      a3-rn:by         ncl      cow

                        ‘The bull took the boy away’.

                        (Lit., ‘the boy went by the bull’.)


Previously, I (Mora-Marín 2007a) discussed examples of this construction in other Mayan languages at the CILLA III conference: Huastec, Ch’ol, K’ichee’, Awakatek. However, in the proceedings paper derived from that presentation (Mora-Marín 2007b), I left out the discussion of the inverse voice in order to focus on the main topic of the paper, inchoatives. This note thus aims to merely provide a few examples of inverse constructions à la Zavala (1997) from a few additional Mayan languages and discuss their historical significance.


The first example comes from Potosino Huastec. Although the translation is not suggestive of a semantically transitive construction, the clauses seen in (2a-b) are morphosyntactically analogous to those described for Akatek by Zavala (1997). Huastec employs a preposition k’al ‘with, to, of, in, by, for’ to convey a variety of indirect roles, such as comitative, benefactive, recipient, demoted agent, and demoted object (Edmonson 1988:523-524, 531). In (2a-b) k’al is used to introduce a semantic cause, as in (2a) with ‘sun rays’, or semantic agent, as in (2b). Edmonson (1988:491-492) suspects it is likely that this preposition is historically related to the noun root k’aal ‘possession’, attested unpossessed as k’aal-aab ‘property, possession’, and possessed as k’aal in 7in k’aal ‘her/his property; hers’. I agree with this suggestion, and further hypothesize that its case-marking role may have in fact originally involved a construction of the type erg-k’aal ± np. As shown below, a similar development is seen in Ch’ol. Also, as a sidenote, it seems likely that Acalán/Yokot’an k’a (Keller and Luciano 1997:) is a borrowing of Huastec k’al (or k’a:l).


(2)       Potosino Huastec (Edmonson 1988:242, 297)

            a.         7in      7ot’-ool           7u        7okoob          xuh(uu)l-in-Ø

                        a3        skin-poss        a1        arm                  spotted-vn-cmp

                        k’al                    7i         k’ak’al

                        prep:by           of        sun.ray

                        ‘The skin of my arm is spotted by the sun’s rays’.

            b.         k’atz-utzuul                 7an      lukuk

                        stirred.up-rep              the       dirt

                        k’al                     7an     bel-al

                        prep:by           the       travellers

                        ‘The dirt (is) very stirred up by the travellers’.


Next is Ch’ol, with a few examples provided in (3a-c). In general, the Ch’ol examples appear to be consistently translated into Spanish or English as semantically transitive clauses, whether in the work of non-native speakers (e.g. Aulie and Aulie 2009[1978]) or native speakers (e.g. Vázquez Álvarez 2011). The semantic agent of these expressions is the possessor of the relational noun cha7an (cha7añ), which is used to express possession when inflected with ergative/possessive pronominals (i.e. Set A), as in (3a-d), but used to express other semantic roles (benefactive) or verbal complements (purpose clauses) when used as a preposition. Note that the translations by Aulie and Aulie (2009[1978]) vary between intransitive (passive), as in (3a), and semantically transitive, as in (3b-d), whereas they are consistently semantically transitive in the work of Vázquez Álvarez (2011), as in (4a-c).


(3)       Ch’ol Mayan (Aulie and Aulie 1978:62, 14, 18, 29)

            a.         ñaj-ä-y-em-Ø                                      k-cha7añ

                        forget-inch-epn-partc-b3            a1-rn:possession

                        jini                   tsa7      bä        i-su[b]-be-y-on

                        dems               cmp     rel      a3-say-indir-epn-b1

                        ‘Ya está olvidado lo que me dijo’.

                        ‘It is forgotten [by me] what s/he told me’.

            b.         käch-äl-Ø                    k-cha7añ                     jini        chitam

                        tie-stat-b3                   a1-rn:property        dems   pig

                        kome               ñajt      mi7      cha7len            xämbal

                                                                   mi        i-cha7len

                        because           far        inc       a3-make          walk(ing)

                        ‘Tengo el cerdo amarrado porque va muy lejos’.

                        ‘I have the pig tied up because it goes/walks far’.

                        (Lit. ‘It is tied up by me the pig because it walking-makes far’,

                        or ‘It is tied up the pig of mine because it walking-makes far’)

            c.         Max=to           (7)añ-ik           k’ajty-i-bil-Ø                k-cha7añ

                        neg=still         exist-subj        ask-tvzr-partc-b3    a1-rn:property

                        ba-ki=ora          mi         kaj         i-tilel.

                        what=time       impf     begin   a3-come-inc

                        ‘No le he preguntado para cuándo va a venir’.

                        ‘I have not asked him what time he will come’.

                        (Lit. ‘S/he is not yet asked by me what time s/he will come’)

            d.         chuk-bil-Ø                   i-cha7an                         i-machit

                        seize-partc-b3           a3-rn:property           a3-machete

                        ‘Tiene el machete agarrado en la mano’.

                        ‘S/he has the machete seized/grasped in his hand’.

                        (Lit. ‘It is seized by him/her his/her machete’ or ‘It is seized the machete of his/



(4)       Ch’ol Mayan (Vázquez Álvarez 2011:149, 216, 334)

            a.         k’ajal-ety=äch=me                   i-cha7añ

                        remember-b2=affr=pre        a3-rn:property

                        ‘Yes, he remembers you’.

                        (Lit. ‘You are indeed remembered by him’.)

            b.         tz’ej-chok-o-bi(l)-Ø                            k-cha7añ

                        sideways-dep-tvzr-partc-b3         a1-rn:property

                        ‘I put it in sideways’.

                        (Lit. ‘It is/was placed sideways by me’.)

            c.         majch             kuch-u(l)-Ø                 k-cha7añ

                        who                 carry-stat-b3             a1-rn:property

                        jiñ                    ix7ä                 Mikolas

                        foc                  that                  Nicolás

                        jiñ                    jiñ=i

                        foc                  that=fin

                        ‘Whom I was carrying, was Nicolás. It was him’.

                        (Lit. ‘Who was carried by me, it was Nicolás. It was him’.)


Acalán/Yokot’an exhibits similar constructions, which are sometimes translated in a semantically transitive manner, but other times not. Two grammatical morphemes are involved. One is k’a in contemporary Yokot’an, <kal> in Acalán, almost certainly related, probably through diffusion, to Huastec k’al ~ k’aal. The other is the demonstrative pronoun base jin. Osorio May (2005) includes examples, most of which were not translated as semantically transitive clauses. In Acalán <kal> is attested only as a possessed relational noun <ukal> or <ukalob> (Smailus 1975:25, 40), for example, used to express demoted or underlying agents. In the variety of Yokot’an from Tapotzingo, it is still used as a possessed relational noun with various case functions, including the introduction of cause and purpose complement verbs and nouns, as well as as a ‘dative pronoun’, as in k’än-ä-Ø u-k’ä (want-stat-b3 a3-rn:property) ‘lo necesita (s/he needs it) (Lit. ‘it is needed/wanted by him’) or u-na<j>y-an u-c’a (a3-forget[pas]-inc a3-rn:property) ‘se le olvida (s/he forgets it) (Lit. ‘it is forgotten by him’), in which rather than dative, the relational noun u-k’a serves to express the semantic agent or experiencer (Keller and Luciano 1997:268-269). In the variety from Tecoluta documented by Osorio May (2005), it has clearly undergone further grammaticalization to a preposition in the context of passive clauses where it introduces the demoted agent, as in (5a). However, in the context of in psychological verb constructions like those reported by Keller and Luciano (1997), it still functions as a relational noun, as in (5b-c), one that expresses the semantic agent or experiencer. I should note that Neither Keller and Luciano (1997) nor Osorio May (2005) analyzes the /j/ of najy ‘to forget’ as a passivizer infix; however, the verb najy is inflected in the manner typical of passives in the completive, with -i, and incompletive, with -an.


(5)       Yokot’an (Osorio May 2005:20, 252, 258)

            a.         jäts’-k-i-Ø                   ch’ok               k’a                  7ix-pet

                        hit-pas-cmp-b3       child                prep              ncl-Petrona

                        ‘El niño fue golpeado por Petrona’.

                        ‘The boy was hit by Petrona’.

            b.         na<j>y-i-Ø                              kä-k’a-la                          ni         buk

                        forget<pas>-cmp-b3         a1-rn:property-pl      det      shirt

                        ‘Se nos olvidó la camisa’.

                        ‘We forgot the shirt’.

                        (Lit. ‘The shirt was forgotten by me’)

            c.         yaj-i-Ø                                    kä-ka                               7ix-mita7

                        esteem-cmp-b3                 a1-rn:property           ncl-Carmita

                        ‘Estimé a Carmita’.

                        ‘I appreciated Carmita’.

                        (Lit. ‘Carmita was appreciated by me’)


Similar constructions also appear in K’ichee’, as illustrated in (6). K’ichee’ employs the relational noun -umaal to express the agent.


(6)       Kichee(Mondloch 1978:31):

            x-Ø-b’ee                        lee       wuuj                w-umaal

            cmp-b3-go                   det      book                a1-rn:cause

            The document went by me.


Finally, (7) provides an example from Awakatek present in a passage published in McArthur (1979). I have preserved McArthur’s transcription and glossing, for the most part.


(7)       Awakatek (McArthur 1979:223)

            Cmi’x              nin        wex       cu’n

            shirt                 and      pants    all

            el                          tzaj-tz                                 cy-ak’un                  chij

            went.out          in.this.direction           a6-rn                

            ‘They took only shirts and pants’.

            (Lit. ‘the shirts and pants went out by them’.)


Some comments are in order. First, given the wide distribution of these constructions (Huastecan, Greater Tzeltalan, Greater Mamean, Greater K’ichee’an), it is likely that proto-Mayan exhibited at least one type of inverse voice construction in which an intransitive or intransitivized verb plus relational noun construction functioned as a semantically transitive clause. Second, there is no shared relational noun across the subgroups of Mayan represented by the languages just cited, with the exception of Huastec k’al ~ k’aal and Yokot’an k’a, though this is almost certainly a case of diffusion, perhaps in the context of maritime trade along the Gulf Coast, in which ‘possession/property’ would have been a topic of typical linguistic interaction. That said, it is likely that proto-Mayan would have employed an abstract noun with a basic meaning of ‘cause’ or ‘doing’ or ‘property’ as a relational noun in constructions of this type.


One last topic is worth pursuing at this point, at least in a cursory manner. As already noted, previously Mora-Marín (2001:102, 222; 2007a) suggested the possibility that some Epigraphic Mayan constructions were similar to those discussed by Zavala (1997) for Akatek. The construction in question involves, in general, the following structure: [Intransitive Verb + Intransitive Subject]clause1 + [7u-T526(-ji-ya) + Proper Name of Person]clause2. The intransitive verb is most of the time a passivized transitive root or an inchoative in -V1y based on a variety of roots (nominal, transitive, intransitive, positional, adjectival) (Mora-Marín 2007a, 2007b, 2009). Only rarely does an active transitive verb appear in this position, though this may take place (Mora-Marín 2004). The intransitive subject of the first clause is most often inanimate, and is sometimes possessed by a phrase referring to a human referent; however, it may also be animate, and if so, it refers to a human referent (e.g. in clauses referring to taking of office). In the second clause, one finds the 7u-T526(-ji-ya) glyphic expression which appears to indicate either cause or agency. The final proper name phrase typically refers to a human actor in position of greater power relative to the human actor or possessor mentioned in the first clause. To test whether these constructions function like an inverse voice, pragmatically, it would be necessary to study their contexts for topicality and determine whether the subject (or perhaps its possessor) in the first clause is of greater topicality than the subject of the second clause.



Aulie, Wilbur H., and Evelyn W. de Aulie. 1999. Diccionario Ch’ol-Español, Español-Ch’ol.  Mexico City: Instituto Lingüístico de Verano. Third edition.

Edmonson, Barbara Wedemeyer. 1988. A descriptive grammar of Huastec (Potosino dialect).  Unpublished Ph.D. dissertation, Tulane University.

Hull, Kerry. 2016. A Dictionary of Ch’orti’ Mayan-Spanish-English. Salt Lake City: The University of Utah Press.

Kaufman, Terrence, and William Norman.  1984.  An outline of Proto-Cholan phonology, morphology, and vocabulary. In Phoneticism in Maya Hieroglyphic Writing, edited by John S. Justeson and Lyle Campbell, pp. 77-166. Institute for Mesoamerican Studies Publication No. 9. Albany: State University of New York.

Kaufman, Terrence, with John Justeson.  2003.  Preliminary Mayan Etymological Dictionary.

Keller, Kathryn C., and Plácido Luciano G. 1997. Diccionario Chontal de Tabasco. Tucson, Arizona: Summer Institute of Linguistics.

Knowles, Susan M. 1984. A Descriptive Grammar of Chontal Maya (San Carlos Dialect). Ph.D. Thesis, Department of Anthropology, Tulane University.

MacArthur, Lucille. 1979. Highlighting in Aguacatek Discourse. In Discourse Studies in Mesoamerican Languages, Volume I: Discussion, edited by Linda K. Jones and Robert E. Longacre, pp. 219–244. Dallas: The Summer Institute of Linguistics and The University of Texas at Arlington.

Mondloch, James L. 1978. Basic Quiché Grammar. Albany, N.Y. : Institute for Mesoamerican Studies, State University of New York at Albany.

Mora-Marín, David F. 2001. The Grammar, Orthography, and Social Context of Late Preclassic Mayan Texts. Unpublished Ph.D. Thesis.  University at Albany, Albany, New York.

—–. 2004. The Preferred Argument Structure of Classic Mayan Writing. In The Linguistics of the Maya Script, edited by Søren Wichmann, pp. 339–364. Salt Lake City: University of Utah Press.

—–. 2007a. The Identification of an Ingressive Suffix in Classic Lowland Mayan Texts. Paper presented at the The CILLA III Conference, October 2007, Austin, Texas.

—–. 2007b. The Identification of an Ingressive Suffix in Classic Lowland Mayan Texts. In Nora England (ed.),Proceedings of the CILLA III Conference, October 2007, Austin, Texas, 1-14. Austin: Archive of the Indigenous Languages of Latin America, Linguistics Department, University of Texas. ( (Accessed January of 2017.)

—–. 2009. A Test and Falsification of the Classic Choltian” Hypothesis: A Study of Three Proto-Cholan Markers. International Journal of American Linguistics 75:115–157.

Osorio May, José del Carmen. 2005. Análisis de la morfología verbal del yokotan chontalen el poblado de Tecoluta, Nacajuca, Tabasco. M.A. Thesis. CIESAS, México.

Smailus, Ortwin. 1975. El Maya-Chontal de Acalán: Análisis lingüíistico de un documento de los años 1610-12. Centro de Estudios Mayas, Cuadernos 9. Mexico: UNAM.

zquez Álvarez, Juan Jesús. 2011. A Grammar of Chol, A Mayan Language. PhD dissertation, The University of Texas at Austin.

Zavala, Roberto. 1997. Functional Analysis of Akatek Voice Constructions.  International Journal of American Linguistics 63:439-474.


[1] Abbreviations: 3 = third person; a = Set A pronominal (ergative), affr = affirmative, b = Set B pronominal (absolutive), cmp = completive, dems = demonstrative, det = determiner, dir = directional, fin = phrase-final enclitic marker, foc = focus marker, impf = imperfective, inc = incompletive, ncl = nominal classifier, neg = negative, parts = (passive) participle, pas = passivizer, poss = possessive suffix, pre = (pre)cautionary, prep = preposition, rep = repetitive, rn = relational noun, stat = stative participle, subj = subjunctive, tvzr = transitivizer, vn = vowel-initial thematic suffix. <7> stands for /ʔ/.