Research efforts to model speech perception in terms of a processing system in which knowledge and processing are distributed over large numbers of highly interactive--but computationally primative--elements are described in this report. After discussing the properties of speech that demand a parallel interactive processing system, the report reviews both psycholinguistic and machine-based attempts to model speech perception. It then presents the results of a computer simulation of one version of an interactive activation model of speech, based loosely on the COHORT model, devised by W. D. Marslen-Wilson and Welsh (1978), which is capable of word recognition and phonemic restoration without depending on preliminary segmentation of the input into phonemes. The report then notes the deficiences of this model, among them its excessive sensitivity to speech rate and its dependence on accurate information about word beginnings. It also describes the TRACE model, which is designed to address these deficiencies, noting that it allows interactive activation processes to take place within a structure that serves as a dynamic working memory. The report points out that this structure permits the model to capture contextual influence in which the perception of a portion of the input stream is influenced by what follows it as well as by what precedes it in the speech signal. (FL)