Discrete serial versus cascaded processing in lexical access in speech production: Further evidence from the coactivation of near-synonyms

In 5 cross-modal picture-word interference experiments, the authors investigated the time course of lexicalization in speech production. Participants named pictures of simple objects while hearing distracter words at different stimulus-onset asynchronies (SOAs). Distracters were semantically related or phonologically related to either the picture name or its nontarget near-synonym. Compared with an unrelated control, effects from all distracters were obtained at a late SOA (150 ms) and vanished shortly thereafter (at SOA = 300 ms). These findings conflict with the notion of an early selection of a single element. Rather, multiple lexical representations appear to remain active until late in the production process if a near-synonymous lexical competitor is present. The authors discuss the scope of these observations and their relevance for discrete two-stage and cascaded models of lexical access.