Differential Function Marking, Case, and Information Structure: Evidence from Korean

The central goal of this article is to propose a systematic description of differental function marking (DM) in Korean, a language in which both subject and object markers may fail to be spelled out. Taking Aissen's theory of DM (Aissen 2003) as a starting point, we show that although its predictions seem mostly consistent with the statistical results of corpus-based research on Korean (and Japanese), this model does not accurately account for the Korean data. We argue that subject and object bareness (the lack of a functional particle) regularly correlates with interpretive effects that should be captured in terms of information structure (focus structure). Adapting Erteschik-Shir's (1997, 2007) framework to represent f(ocus)-structure, we argue that bare subjects and objects in Korean fail to be visible at this level. Consequently, they may be construed neither as active topics nor as foci, and thus must either be left out of f-structure or incorporated within larger f-structure constituents in order to be interpreted. We show that bare objects are never construed as topics or foci and always exhibit a form of semantic incorporation, while leul-marked objects always stand as f-structure constituents construed as focused at some level. Bare subjects, unlike neun-marked topical subjects and GA-marked subjects, can be construed neither as active topics nor as foci, and always occur in tense-deficient clauses construed as thetic and anchored to speech time. We argue that our assumptions correctly predict the results of corpus studies, and we suggest that as regards nominal arguments, f-structure visibility might ultimately stand as the crucial interpretive correlate of functional positions in syntax.