Reflexivization in Mandarin: the role of zi-ji and its components

This dissertation analyzes reflexivization in Mandarin Chinese. It has two main aims: one is to critically examine the standard assumption in the literature that the Mandarin anaphor ziji is simplex, providing an analysis based on the role of the verbal prefix zi- in the anaphoric system of Mandarin; the second aim is to provide a critical assessment of a claim in the current literature on binding, namely that anaphor binding is in principle always local and that nonlocal binding makes crucial use of an additional mechanism involving discourse factors. I show that zi-ji is in fact morphosyntactically complex based on diachronic and synchronic evidence involving the systematic application of various diagnostics. I conclude that zi-ji consists of a reflexivizing element zi- (also as verbal affix reflexivizing the predicate) and a defective pronominal stem -ji. I also show that the original arguments in favor of a simplex analysis lack force. This complexity is crucial for understanding of reflexivity in Mandarin. It entails that zi-ji ‘REFL-self’ licenses reflexivity by protection (Reuland 2011a). Another major issue not discussed in the theoretical literature is the role of the verbal prefix zi- in the anaphoric system of Mandarin (but see Chief 1997, 1998 for earlier discussion). Zi- functions as an operator on argument structure generating reflexive verbs by identifying two argument variables. Its effect is similar to the bundling operation of Reinhart and Siloni (2005), which is restricted to a subset of Agent-Theme verbs. Zi- also operates on a class of verbs taking clausal complements, identifying the external argument variable of the verb with a free variable in the complement. This explains a highly curious and hitherto mysterious binding pattern: the pronominal ta and the reflexive zi-ji in the complement of such a zi-verb must be bound by the subject of that verb; in the latter case unless a blocking effect interferes. I review a number of influential earlier approaches to binding in Mandarin. For each of them I identify empirical and theoretical problems that warrant pursuing an alternative. Specifically, none of them addresses the role of zi-prefixation. Moreover, they all turn out not to be able to account for the binding pattern in the complements of zi-verbs. The conclusion is that the complexity hypothesis paired with Giblin’s (2016) Agree-based approach is best equipped to provide an explicit and principled account of both nonlocal binding of ziji and the blocking effect, it can be straightforwardly integrated with my account of that peculiar binding pattern. In a nutshell, I show that the complexity of zi-ji provides the basis for a unified analysis in which the contrast between local and nonlocal binding of zi-ji reduces to a timing difference involving the interaction of zi-ji’s constituents with the syntactic environment. There is no qualitative difference in the mechanisms employed for local and nonlocal binding, since both are part of narrow syntax. Overall, then, Mandarin has a verbal reflexivizing affix zi-, a (complex) semireflexive zi-ji (licensing reflexivity without enforcing), and a supercomplex reflexive ta zi-ji (licensing and enforcing reflexivity).