Cumulative semantic cost without successful naming

Abstract

Accessing semantic information has negative consequences for successive recovering attempts of similar information. For instance, in the course of picture-naming tasks, the time required to name an object is determined by the total number of items from the same category that have already been named; naming latencies increase proportionally to the total number of semantically related words named previously. This phenomenon is called cumulative semantic cost (or interference). Two picture-naming experiments with children (4-11 years old, 229 participants) investigate whether having successfully named the previous within-category items is a necessary condition for the cumulative semantic cost to appear. We anticipated that younger children would have a larger rate of nonresponses compared with older children, reflecting the fact that younger children have not yet consolidated many lexical representations. Our results confirmed this prediction. Critically, we also observed that cumulative semantic cost was independent of having successfully retrieved previous within-category lexical items. Furthermore, picture trials for which the previous within-category item elicited a nonresponse showed the same amount of cost as those picture trials for which the previous within-category item elicited a correct naming event. Our findings indicate that it is the attempt to retrieve a lexical unit, and not the successful retrieval of a specific lexical unit, that causes semantic cost in picture naming. This cost can be explained by a mechanism of weakening the semantic-to-lexical mappings of semantic coordinate words. The findings are also discussed in the context of retrieval-induced forgetting effects in memory recall research.

Keywords: Interference/inhibition in memory retrieval; Lexical processing; Picture naming in children; Psycholinguistics; Semantic priming; Word production.

Fig. 1

Study 1. a Developmental trend of nonresponses: Percentage of nonresponses in each School-level. b Cumulative semantic costs: Mean naming latencies by ordinal position within-category in each school level. As can be seen, naming latencies increase with ordinal position within category in all the school levels. All error bars are standard error of the mean. PS = preschool; 1ry Sch = primary school

Fig. 2

Study 2. a Developmental trend of nonresponses: Percentage of nonresponses in each school level. b Cumulative semantic costs: Mean naming latencies by ordinal position within category in each school level. As can be seen, naming latencies increase with ordinal position within category in all the school levels. All error bars are standard error of the mean. PS = preschool; 1ry Sch = primary school