Subject Extraction and Labeling Algorithm

This paper demonstrates that subject extraction in Korean, which apparently violates the Subject Criterion (SC), actually fits with the labeling theory (Chomsky 2013) and Criterial Freezing (CF) (Rizzi 1997, 2006, 2015, 2016, 2017). It is shown that subjects are legitimately extracted in Korean by the labeling algorithm, without violating CF. It is suggested that the head of Finiteness Phrase (FinP) satisfying SC makes subject extraction possible through Spec-FinP. The extracted subject NP is properly labeled in the left periphery of CP by sharing features such as a focus feature or a topic feature. Various constructions such as Korean seem-clauses, ECM-clauses, regular complex clauses, and wh-clauses are analyzed with respect to subject extraction. This analysis is based on Rizzi and Shlonsky (2007) and Citko, Germain, and Witkos (2018), in which the notion of criterial satisfaction is extended from Spec-head configurations to head-head configurations.