Sentential Polarity under Ellipsis

This paper sheds light on a relatively under-studied phenomenon observed in an elliptical construction – polarity reversal under Sluicing. Specifically, we look into a Sluicing construction in which the unstated proposition in the elliptical site bears inverse polarity to the presumable antecedent. After reviewing Kroll’s (2019) latest work among the relevant literature and presenting an array of intriguing empirical facts, we make the following three generalizations and proposals: [1] In exclusive disjunction contexts, in which verum focus does not play a grammatical role, polarity reversal under Sluicing is readily observed. In line with Rudin’s (2019) syntactic condition, we propose that functional categories such as sentential polarity as well as aspect, modal, tense, and complementizer enter syntax without being specified concerning their lexico-syntactic features. Crucially, we propose that identity in ellipsis be computed before such features are determined at LF. Then, it is only the vP that counts in the calculation of identity for the license of ellipsis. [2] In other non-exclusive disjunction contexts, in which verum focus plays a crucial role in yielding relevant interpretations, polarity reversal under Sluicing is not allowed. This suggests that what is involved in so-called Neg-‘raising’ and implicative verb contexts is, in fact, not polarity reversal but polarity concord under Sluicing. [3] Unlike English, Korean allows polarity reversal under (pseudo-)Sluicing relatively freely in non-exclusive disjunction contexts, which is attributed to the fact that it does not require phonological realization of verum focus. This contrastive property is conjectured to be due to the fact that verum focus is given to an auxiliary verb in English while it is attracted to a Wh-XP in Korean.