Abstract
Anaphora is a common, universal, and critical aspect of communication. For quite some time, linguists and psycholinguists have applied introspective and behavioral techniques to the study of anaphora, yielding a large body of evidence and generating a number of theories and models of anaphoric usage and processing. The recent advent of the use of electrophysiological techniques to study language processing creates a unique opportunity to reevaluate these theories and models on the basis of neurophysiological data. This paper examines whether two key characteristics of anaphoric relationships (i.e. the underlying linguistic mechanism and the nature of the anaphor) have observable effects on neurophysiological processing. An exhaustive review of electrophysiological studies testing anaphoric constructions leads to the novel conclusion that the effects of these characteristics are reducible to differences in the time-course of the availability of information and the ability to form expectations about anaphoric relationships. This review also demonstrates that while theories of anaphoric usage are generally supported, finer-grained models of the time-course of anaphoric processing require revision to account for the new electrophysiological data.