Universals of reflexives

The locality conditions on reflexives vary widely from language to language. In English, a reflexive (e.g. himself) must take a local antecedent. In contrast, in Asian languages like Chinese, a reflexive (e.g. ziji) can take an antecedent indefinitely far from its antecedent. Furthermore, in European languages, a reflexive is bound in a less restricted domain than in English, but not in an indefinite domain as in Chinese. The purpose of this dissertation is to develop a unitary analysis of locality conditions on the binding of reflexives among languages within the general framework of Government and Binding theory. The core idea explored here is that reflexives in all languages obey the same locality conditions and that apparent long distance reflexives involve local successive cyclic head movement.Chapter 1 provides a general idea of the ways in which languages seem to be different with respect to long distance reflexives in terms of locality conditions. I suggest that long distance binding between reflexives and their antecedents is in fact local binding in much the same way that a local relation holds between wh-elements and the original and intermediate movement sites.In Chapter 2, taking Chinese as a starting point, I propose that the differences between ziji 'self' and ta ziji 'himself' in Chinese hinge on whether the reflexive form is X$\sp{\rm o}$ or X$\sp{\rm max}$. X$\sp{\rm o}$ reflexives like ziji will involve head movement from INFL to INFL, thus licensing long distance binding. In contrast, X$\sp{\rm max}$ reflexives like ta ziji or himself adjoin to X$\sp{\rm max}$ projections. Therefore, X$\sp{\rm max}$ reflexives can have only local antecedents.Chapter 3 explores the idea of proposing a simple set of principles governing feature percolation to allow a unified treatment of a number of related phenomena. I argue that four sets of data are being accounted for by the proposed Feature Percolation Principles. The sets of data are (1) the differences in "blocking effects" of long distance reflexives in East Asian and European languages; (2) the so-called "sub-command" facts in East Asian languages; (3) the that trace effects in English and other languages; and (4) Pro-drop parameter among languages.In Chapter 4, I first discuss a number of problems with the X$\sp{\rm o}$/X$\sp{\rm max}$ hypothesis. I then compare the head movement analysis proposed in this thesis with a number of previous approaches to long distance reflexives. I argue that these alternative proposals are inadequate conceptually and empirically and, as a result, the head movement analysis is to be preferred.U of I OnlyETDs are only available to UIUC Users without author permissio