Split ergativity and transitivity in Chol

In this paper I examine two types of splits in the otherwise ergative-absolutive person marking system of Chol Mayan. First, Chol shows what is known as a “split-S” system: subjects of unergatives pattern as transitive subjects (ergative), while subjects of unaccusatives pattern as transitive objects (absolutive) (Gutiérrez Sánchez, 2004, ). This pattern is obscured in the non-perfective aspects, where we find an aspect-based split (see Vázquez Álvarez, 2002, ): both transitive and intransitive subjects receive the same morphological marking. Specifically, in the non-perfective aspects, all intransitive subjects pattern with transitive subjects. I argue below, extending Laka's (2006) proposal for Basque, that the appearance of these splits results from differences in syntactic structure, rather than from different rules of Case assignment or agreement. I show that these different structures fall out from independent facts about the language; neither split represents a departure from Chol's ergative system. While the analysis focuses on Chol, I propose that a biclausal analysis of split ergativity may account for splits in a number of unrelated languages. Finally, I examine other language with splits that lend themselves to this type of analysis, and suggest a grammatical basis for the universal tendency for languages with aspectual splits to always retain the ergative pattern in the perfective aspect.