Morphosyntactic vs. morphosemantic functions of Indonesian –kan

Levin and Rappaport Hovav (1998) distinguish two basic types of morphological operations on verbs: those that modify the Lexical Conceptual Structure (LCS, the semantic representation) vs. those that alter the syntactic expression of arguments without changing the semantic representation. The Locative alternation is an example of the first type; passive and applicative are examples of the second type. Levin and Rappaport Hovav state that the specific affixes used for operations of the first type are frequently the same as those used to mark aspectual distinctions, but are always different from the affixes used for operations of the second type. They know of no language in which the same affix is used to mark alterations to LCS in some forms and changes in the syntactic expression of arguments in other forms. In this paper I will present evidence showing that the Indonesian suffix -kan functions in both of these ways. This suffix has been the topic of much discussion over the past 40 years, and a source of considerable confusion. I will argue that in order to understand the various functions of the suffix we must recognize that it has two very different modes of operation in the grammar. These two modes can be identified by the syntactic (monotransitive vs. ditransitive) and semantic (displaced theme vs. benefactive) patterns that they create. Adopting the terminology of Ackerman (1992) and Sadler and Spencer (1998), I refer to operations that modify LCS as MORPHOSEMANTIC, and to operations that modify only the syntactic expression of arguments as MORPHOSYNTACTIC.1