Pronominal reference serves as one of the basic sources of structure in normal discourse. It is this structure that the beginning reader must discover and use to develop into a competent adult reader. Many difficulties in comprehension can be directly traced to a failure of such higher level processing. Determination of a proper anaphoric referent presents a complex task for the human language processing system. Under various conditions, reference assignment draws on lexical, syntactic, semantic, and pragmatic knowledge. The critical psycholinguistic questions about pronoun comprehension concern the time course of the related processes. Our research develops a search and selection model of pronoun comprehension and relies on a well-documented distinction between immediate working memory, which has very limited capacity but fast access, and long term memory, which has a large capacity but slower access. Search for the referent is believed to be sensitive to differing access characteristics of the two memory systems. Selection among possible referents is also influenced by different types of information (lexical, pragmatic, etc.) according to the representation system involved. This research will both use and develop eye tracking as a psycholinguistic technique. Eye movement records give a precise, continuous, and multifactored view of the reading process. By conducting a set of closely related experiments on a crucial aspect of language processing, a large amount of convergent data will be collected that will allow the relationship of eye movements to language processing to be modeled and evaluated. The purpose of this project is to develop a model of normal, successful language comprehension. This can later be used as a standard against which to evaluate the development of comprehension, as well as reading dysfunctions.
