Theory of Binding Summation
The theory of binding investigated here consists of three principles which determine the distribution of bound elements—anaphors, pronouns, and names. Section 1 demonstrates that the relation between an antecedent and a bound element relevant for the operation of these principles is that of c-command. Section 2 addresses the issues involved in specifying opaque domains for anaphor binding, concluding that the notion of accessible θ-SUBJECT gives the closest fit to the facts. In section 3 it is noted that while the S paradigms for pronoun and anaphor binding are analogous, the NP-paradigms are not. To account for these facts of pronoun binding, it is necessary to formulate the notion of opaque domain differently from that of Principle A. This is achieved by utilizing the notions of θ-domain and government, with the result that Principles A and B overlap with respect to opaque domains for the S-paradigm, but diverge in the appropriate way for the NP-paradigms. The extension of binding theory to empty categories is considered in section 4. While there is some motivation for concluding that WH-trace is subject to Principle C, it is questionable whether PRO is subject to the binding theory in any way.
Furthermore, there is some evidence that NP-trace should be treated as an empty category analogue of a lexical anaphor and hence subject to Principle A. In this way, the binding theory contributes to a typology of empty categories, under which empty categories may be identified functionally or derivationally. The question of which levels of representation are relevant for the three principles of the binding theory is raised in section 5. Discussion establishes that the c-command relation of section 1 may be insufficient for cases involving Ā-bounding. This suggests that S-structure by itself is not the appropriate level for any of the principles. Equative constructions provide evidence that Principle C should apply to a different level than Principles A and B.
This article has attempted to provide an overview of some of the fundamental issues confronting a theory of binding. At various points, alternative proposals have been considered. At the level of grammatical analysis it is often difficult to distinguish between alternatives which cover the same range of data. Yet these alternatives may not be equivalent in terms of what they suggest for processing or acquisition of language. Consider for example the issue of defining opaque domains for Principles A and B. If the definition is identical for both principles, then we might expect that a child who has acquired the S-paradigm for anaphors will also have the S-paradigm for pronouns given that Principle B is operative in his/her grammar. If, however, the definitions are different as proposed above, we might expect to find differences in the acquisition of the S-paradigms for pronouns and anaphors. For one study that argues in favor of the latter case, see Solan to appear.
