Argument Structure and Antipassivization in Inuit

This paper examines some seemingly disparate uses of the morpheme -si in two very closely related Inuit (Eskimo) languages: Inuktitut (spoken in Northeastern Canada) and West Greendlandic. I will use the term "Inuit" to refer to both West Greenlandic and Inuktitut as they pattern together. The morpheme -si is traditionally associated with antipassive morphology, although it also appears in several other environments that do not, at first, seem related to antipassivization. I will show that there are actually systematic patterns underlying the distribution of -si, indicating that the uses of of the morpheme in question are not as disparate as initial observations have assumed. I will discuss the implications of these patterns for a theory of the morphosyntax of antipassivization. Antipassivization is canonically a valency-changing operation that intransitivizes a transitive verb by "demoting" the direct object to an oblique Case or omitting it altogether. In Inuit, the antipassive is marked by a suffix which has several allomorphs. Many researchers claim that the specific allomorph which appears is lexically governed to some extent, and is at least partially abitrary or idiosyncratic, and not necessarily conditioned by phonological or other factors. Johnson (1980) and Bittner (1987) argue that the different antipassive morphemes are not allomorphs, but actually are used in different discourse contexts. The most common overt antipassive marker is -si, the one that we will focus on in this paper