Animacy effects in picture naming and bilingual translation: Perceptual and semantic contributions to concept-mediation

These experiments investigate perceptual and semantic factors underlying animacy effects in semantically-driven language production tasks, and explore how differences in semantic organization affect the degree to which concepts are shared across languages. McRae, de Sa, and Seidenberg (1993) hypothesized that animate concepts exhibit greater intercorrelations among their semantic features than do inanimate concepts. McRae et al.'s model predicts that animate items should exhibit slower response latencies than inanimate items when presented in categorized lists as a result of response competition. Experiments 1 (bilingual picture naming) and 2 (bilingual translation) compared response latencies for animate and inanimate items presented in categorized and randomized sequences. Translation and picture naming performance were contrasted because they engage similar semantic processes but differ in terms of perceptual processing requirements.; In Experiment 1, pictures of animate items were named slower than pictures of inanimate items, and semantic categorization facilitated response latencies for both types of items. In Experiment 2, animate items were translated faster than inanimate items, and semantic categorization did not affect response latencies. Findings imply that animacy effects in picture naming and translation reflect both perceptual and conceptual factors. Theoretical conclusions drawn from these results rest on the assumption that both tasks are semantically-driven. Results of Experiments 3-5 provide empirical support for this assumption: no animacy effects were observed for target items in a monolingual word naming task, and the animacy effects observed in Experiments 1 and 2 were replicated under conditions in which semantic activation was required before response output. In Experiment 6, visual similarity ratings obtained for the target pictures suggested that animacy effects obtained in Experiments 1 and 4 were influenced by perceptual factors.; The absence of a categorization by animacy interaction in Experiments 1 and 2 does not directly support McRae et al.'s (1993) model. However, these findings do suggest that animate concepts share larger proportions of semantic features than do inanimate concepts, as reflected by facilitated translation latencies for animate items. Higher visual and semantic similarity for animate items results in slower response latencies when the same items are presented in picture naming tasks.