Semantic interference and facilitation: the role of feature cues and category in naming

Semantic interference effects have been observed in a variety of naming paradigms where categorically related items, (i.e., cow, horse, sheep), elicit longer naming latencies than categorically unrelated items (i.e., book, knife, mirror). However, under certain conditions, semantic facilitation effects (i.e., shorter naming latencies) may be observed from categorically related items depending on the context and order of presentation within a paradigm. Semantic interference and facilitation effects observed in naming are also proposed to be differentially influenced by the correlational nature of the features that comprise these concepts. Using a lexically cued naming paradigm with word pairs designated as either “distinctive” or “shared” features to elicit a target concept which was either related to other concepts within a category or not, evidence for semantic facilitation effects were found for concepts from categorically related items (e.g., farm animals, zoo animals, pets, etc.) when cued by distinctive features. Interestingly, semantic interference effects were not observed in a lexically cued naming paradigm. Likewise, event-related potentials (ERPs) were evaluated, and a significant effect of category (related vs unrelated) was found in the left frontotemporal and right centroparietal regions between 600-1100ms. These ERPs are proposed to represent in the initiation of feature integration beginning approximately 600ms following stimulus presentation and approximately 1200ms prior to naming, indicating an amplitude divergence between categorically related and unrelated concepts. Given the behavioral and EEG data, the following account of semantic and lexical processing is proposed: Categorically related concepts facilitate semantic processing at the superordinate level (i.e., categorical or domain) and features of concepts less likely to co-occur with other concepts (i.e., “distinctive” features) facilitate activation of concepts at the basic level (i.e., specific concept) as measured by naming. Frequent activation of features common among related concepts facilitate subsequent activation of related concepts which facilitates superordinate level semantic processing and distinctive feature cues facilitate access to the basic-level identification required for naming. Categorical level effects are shown to influence naming and neural correlates and when related concepts (or concepts with increased activation of features) are cued by distinctive feature cues, naming latencies and errors are decreased compared to other conditions. Our results suggest when facilitation occurs at both superordinate and basic levels of conceptual processing, naming performance improves.