The genetic matrix of Mayan applicative acquisition*

This article uses data on Mayan applicative constructions to demonstrate the use of a comparative method for language acquisition research. Mayan languages express indirect objects through an applicative su‰x on verbs, a prepositional phrase, or the possessor of the direct object. Mayan children must also acquire language specific lexical constraints on the applicative su‰x. Learners cannot resolve the setting for these parameters through positive evidence. Two-year old children learning the Mayan languages K’iche’ and Tzeltal demonstrate language specific acquisition patterns. Children learning K’iche’ omit the preposition at the head of the indirect object phrase, but retain the ergative cross-reference markers. Children learning Tzeltal sometimes omit the applicative su‰x on the verb, but retain the absolutive crossreference markers. Tzeltal children begin producing the applicative su‰x a year earlier than children learning K’iche’. The Mayan acquisition data refute Crain and Pietroski’s (2002) Continuity proposal. There is no evidence that K’iche’ children extend the applicative along Tzeltalan lines or that Tzeltalan children extend prepositions in the K’iche’ manner. The comparative Mayan data also refute Pinker’s (1989) theory of narrow semantic verb classes in that the applicative su‰x is not constrained by narrow semantic classes of verbs, but rather by the patterns of usage within a given society. The comparative method o¤ers a systematic framework for assessing claims about the nature of children’s language.