On the apparent unbindability of overt third-person pronouns in Japanese

Abstract

Contrary to the widely held view that Japanese overt third-person pronouns such as kare and kanozyo cannot function as bound variables, it has been sporadically reported in the literature that there are cases in which they can receive a bound-variable interpretation. The present paper attempts to provide an account of why Japanese third-person pronouns can be construed as bound variables only in a subset of the contexts in which bound pronouns in English can occur. I argue that Japanese overt third-person pronouns should be analyzed as epithets, claiming that they can function as bound variables only when Condition B and the so-called anti-logophoricity constraint are simultaneously satisfied. I also claim that the apparent insensitivity of the referential use of kare/kanozyo to the anti-logophoricity constraint is attributed to the fact that Japanese (but not English) allows a structure in which a null pronoun is juxtaposed with an appositive epithet phrase.