On Finding Semantic Facilitation in Blocked Picture Categorization: Convergent Response Mapping is Essential

Semantic context effects in picture naming and categorization are central to word production theories. However, unlike naming studies, categorization studies have shown inconsistent results. Recently, Wöhner, Luckow et al. (2024) replicated the inconsistent pattern in blocked categorization in a within-participant and within-item design. Pictures were presented in a semantically homogeneous or heterogeneous context. In the homogeneous context, there was interference for naming, facilitation for naturalness categorization, but no context effect for size categorization. The authors concluded that the inconsistent categorization findings in their own and previous studies could be due either to the use of tasks based on different kinds of features (stored in semantic memory [natural vs. man-made] vs. ad hoc [smaller vs. larger than a standard]) or to a difference in the response mapping for the exemplars from the semantic categories creating the context (convergent vs. divergent). The present study again contrasted the Wöhner, Luckow et al. tasks, but used materials which resulted in convergent response mapping for both categorization tasks. There was semantic interference in naming and semantic facilitation in both naturalness and size categorization. This pattern suggests that convergent response mapping, not the use of a task based on a stored semantic feature, is critical for obtaining facilitation in blocked semantic categorization. Our result provides further support for the notion that semantic interference in blocked word production has its locus at the lexical level and its origin at the semantic level. This conclusion does not depend any longer on data from only a single categorization task.