A Neural Network Model of Lexical-Semantic Competition During Spoken Word Recognition

Visual world studies show that upon hearing a word in target-absent visual context containing related and unrelated items, toddlers adults briefly direct their gaze toward phonologically before shifting semantically visually ones. We present neural network model processes dynamic unfolding phonological representations of words maps them to static internal lexical, semantic, representations. The model, trained on derived from real corpora, simulates this early over semantic/visual preference. Our results support the hypothesis incremental spoken is itself sufficient account for transient preference competitors both Phonological mapped dynamically bottom-up fashion semantic-visual capture effects reported tasks. semantic typically observed later such task does not require top-down feedback or system.