Parsing With Parallelism: A Spreading-Activation Model of Inference Processing During Text Understanding
The past decade of research in natural language processing has universally recognized that because natural language input is almost always ambiguous with respect to its pragmatic implications, its syntactic parse, and even its lexical
analysis (i.e., choice of correct meaning, or word-sense, for an ambiguous word), processing natural language input requires decisions about word meanings, syntactic structure, and pragmatic inferences. The lexical, syntactic, and pragmatic levels of inferencing are not as disparate as they have often been treated in both psychological and artificial intelligence research. Furthermore, these three levels of analysis interact to form an interpretation of text. For example, the choice of a word-sense affects subsequent pragmatic inference decisions or syntactic structure decisions.