The Resonant Dynamics of Speech Perception: Interword Integration and Duration-Dependent Backward Effects

Abstract

How do listeners integrate temporally distributed phonemic information into coherent representations of syllables and words? During fluent speech perception, variations in the durations of speech sounds and silent pauses can produce different pereeived groupings. For exarnple, increasing the silence interval between the words "gray chip" may result in the percept "great chip", whereas increasing the duration of fricative noise in "chip" may alter the percept to "great ship" (Repp et al., 1978). The ARTWORD neural model quantitatively simulates such context-sensitive speech data. In AHTWORD, sequential activation and storage of phonemic items in working memory provides bottom-up input to unitized representations, or list chunks, that group together sequences of items of variable length. The list chunks compete with each other as they dynamically integrate this bottom-up information. The winning groupings feed back to provide top-down supportto their phonemic items. Feedback establishes a resonance which temporarily boosts the activation levels of selected items and chunks, thereby creating an emergent conscious percept. Because the resonance evolves more slowly than wotking memory activation, it can be influenced by information presented after relatively long intervening silence intervals. The same phonemic input can hereby yield different groupings depending on its arrival time. Processes of resonant transfer and competitive teaming help determine which groupings win the competition. Habituating levels of neurotransmitter along the pathways that sustain the resonant feedback lead to a resonant collapsee that permits the formation of subsequent. resonances.