A novel modular connectionist architecture is presented in which the networks composing the architecture compete to learn the training patterns. As a result of the competition, different networks learn different training patterns and, thus, learn to compute different functions. The architecture performs task decomposition in the sense that it learns to partition a task into two or more functionally independent vii tasks and allocates distinct networks to learn each task. In addition, the architecture tends to allocate to each task the network whose topology is most appropriate to that task, and tends to allocate the same network to similar tasks and distinct networks to dissimilar tasks. Furthermore, it can be easily modified so as to...

In recent work on sentence processing, lexical frequencies have been proposed as a primary mechanism for syntactic and lexical disambiguation. In this paper, we instead focus on the consequences that the structural configuration of lexical knowledge has for the time-course of parsing. We concentrate on reduced relative clauses and propose a new lexical-structural analysis for those verbs that are difficult in this construction, manner of motion verbs. The interaction between the proposed lexical structure and the competitive attachment parser (Stevenson, 1994b) explains the persistent difficulty of this construction with a manner of motion verb, even in disambiguating contexts (for example, a reduced relative in object position) or with non-ambiguous past participle verb forms. Weighted influences on the activation competition are possible with other verbs, and the model can therefore also explain data that demonstrate contextual effects on reduced relatives with simple transitives. Co...

Linguistic ambiguity is the greatest obstacle to achieving practical computational systems for natural language understanding. By contrast, people experience surprisingly little difficulty in interpreting ambiguous linguistic input. This dissertation explores distributed computational techniques for mimicking the human ability to resolve syntactic ambiguities efficiently and effectively. The competitive attachment theory of parsing formulates the processing of an ambiguity as a competition for activation within a hybrid connectionist network. Determining the grammaticality of an input relies on a new approach to distributed communication that integrates numeric and symbolic constraints on passing features through the parsing network. The method establishes syntactic relations both incrementally and efficiently, and underlies the ability of the model to establish long-distance syntactic relations using only local communication within a network. The competitive distribution of numeric ev...

This paper presents a massively parallel parser that pre- dicts critical attachment behaviors of the human sentence processor, without the use of explicit preference heuristics or revision strategies. The processing of a syntactic ambiguity is modeled as an active, distributed competition among the potential attachments for a phrase. Computationally motivated constraints on the competitive mechanism provide a principled and uniform account of a range of human attachment preferences and garden path phe- nolnena.