Main Clause Phenomena in Korean (vs. V2 in German)

- We examine universals and parameters in Main Clause Phenomena (MCP).
- A structural ambiguity implies semi-Roots are optional operation in Korean.
- We discuss the typology and semantic properties of MCP in Korean.
- Korean MCP and German V2 can be captured by a single principle.
- We propose a unified syntactic analysis based on timing-of-Merge account.

The goal of this paper is to examine what kind of universals and parameters exist in Main Clause Phenomena (MCP), conducting a comparative study between Korean and German. First, we show that complements allowing MCP reveal a structural ambiguity between root or subordinating clauses, which implies that these cases of semi-Roots are only an optional operation in Korean. Second, we discuss the typology of MCP with respect to true Roots and semi-Roots in Korean. Third, we discuss three semantic properties of MCP in Korean: (i) non-presupposition is a prerequisite in triggering MCP, supporting earlier accounts; (ii) the (non-)presuppositional property reflected in complementizer selection precisely patterns with the (non-)factivity of matrix predicates; and (iii) various semantic-pragmatic sources for MCP such as illocutionary forces, expressive/emotional attitude, mirative evidentials, (anti)honorification, and mood choice in Korean are identified. Fourth, we show that Korean complements with MCP and German V2 constructions exhibit striking parallels in terms of distribution, which implies that MCP in these typologically unrelated languages can be captured by a single principle. Finally, we propose a unified account for the syntax of true Roots, semi-Roots, and non-Roots, building on Wurmbrand's (2012, 2014) timing-of-Merge account. This dynamic syntax allows us to capture the crosslinguistic variation such as structural ambiguity of MCP constructions in Korean and the mismatches between syntactic and LF/PF properties in V2 in German.The current analysis has important consequences: First, this study has empirical confirmation that the types of MCP-trigger must be extended to comprise the newly identified cases in Korean. Second, the structural ambiguity for MCP constructions in Korean gives us the generalization that MCP operate on syntax, even though they are semantically or pragmatically motivated. Finally, the comparative analysis of genetically distant languages opens the possibility to provide a principled account of the MCP-activating force heads across languages, incorporating the empirical findings of prior research on individual factors such as emotional attitude, evidentiality, or mood.