I shall now conclude by summarising the major findings of this book and discussing the theoretical implications of these findings for linguistic theory. The main contributions of this book, I think, are the following: (i) properties of anaphora in Chinese are characterised; (ii) inadequacies of Chomsky's GB theory to account for these properties are demonstrated; and (iii) a neo-Gricean pragmatic theory of anaphora is constructed.

Properties of anaphora in Chinese

Although our analysis has been developed in response to current syntactic and pragmatic theories, implicit in it is a general characterisation of the nature of anaphora in Chinese. Looked at from a relatively theory-neutral, descriptive point of view, anaphora in Chinese has the following properties:

1. (i) Zero anaphors occur extremely freely and often take preference over pronouns.

2. (ii) The use of pronouns is very much restricted.

3. (iii) Reflexives occur very freely and often take long-distance antecedents.

4. (iv) Lexical NPs are frequently used and are often repeated for an intended coreferential interpretation.

5. (v) The assignment of antecedent for anaphora is not so much restricted by grammatical rules as by pragmatic inferences.

6. (vi) There are precise parallels in the assignment of antecedent for reflexives, pronouns and zero anaphors.

7. (vii) There are precise parallels in the assignment of antecedent for intrasentential and discourse anaphora.