Negative concord vs. negative polarity in Japanese

This paper aims to explore some of the behaviors observed in Negative Sensitive Items (NSI) in Japanese. In particular, my main focus is placed on sika, which is one of Japanese NSIs. Whether sika can be regarded as an Negative Concord Item (NCI) or not has hitherto been a controversial issue. I argue that sika should be categorized into an NCI. Sika has interesting syntactic asymmetries, i.e. the asymmetries between sika with a hidden Case-marker (i.e. argument) and sika accompanied by a postposition (i.e. adjunct) in multiple NCI constructions. One of my empirical findings with respect to this asymmetry is that sika can co-occur with other NSIs such as wh+mo or 1+Classifier+mo only when the sika phrase is marked with a postposition but not with a hidden Case-marker. To resolve this problem, this paper proposes two different theoretical frameworks, i.e. similarity-based interference and prosody-syntax interaction in sentence comprehension.