Long-lasting semantic interference effects in object naming are not necessarily conceptually mediated

Abstract

Long-lasting interference effects in picture naming are induced when objects are presented in categorically related contexts in both continuous and blocked cyclic paradigms. Less consistent context effects have been reported when the task is changed to semantic classification. Experiment 1 confirmed the recent finding of cumulative facilitation in the continuous paradigm with living/non-living superordinate categorization. To avoid a potential confound involving participants responding with the identical superordinate category in related contexts in the blocked cyclic paradigm, we devised a novel set of categorically related objects that also varied in terms of relative age - a core semantic type associated with the adjective word class across languages. Experiment 2 demonstrated the typical interference effect with these stimuli in basic level naming. In Experiment 3, using the identical blocked cyclic paradigm, we failed to observe semantic context effects when the same pictures were classified as younger-older. Overall, the results indicate the semantic context effects in the two paradigms do not share a common origin, with the effect in the continuous paradigm arising at the level of conceptual representations or in conceptual-to-lexical connections while the effect in the blocked cyclic paradigm most likely originates at a lexical level of representation. The implications of these findings for current accounts of long-lasting interference effects in spoken word production are discussed.

Keywords: language production; lexical retrieval; semantic interference.

FIGURE 1

An example sequence of consecutive trials in the continuous paradigm employed in Experiment 1. Saw and screwdriver are members of the same category of “tools” in ordinal positions 1 and 2, respectively, and separated by a lag of eight items, whereas hand and ear are members of the same category “body parts” separated by a lag of two items. The remaining items are fillers.

FIGURE 2

Mean correct superordinate categorization latencies as a function of ordinal position and lag in the continuous paradigm. Error bars represent 95% confidence intervals (CIs) calculated per Cousineau’s (2005) method.

FIGURE 3

Matrix depicting the novel object stimuli employed in Experiments 2 and 3 with the blocked cyclic paradigm, consisting of sets of four categorically related (arranged in rows) and heterogeneous (columns) contexts that vary according to the relative age dimension.

FIGURE 4

An example sequence of trials over two consecutive cycles in the blocked cyclic paradigm employed in Experiments 2 and 3. Segway, train, jetski, airship are from the homogeneous context (vehicles category), while waistcoat, lamb, airship, house are from a heterogeneous context.

FIGURE 5

Mean correct basic-level naming latencies as a function of context and cycle in the blocked cyclic paradigm. Error bars represent 95% CIs calculated per Cousineau’s (2005) method.

FIGURE 6

Mean correct superordinate categorization latencies as a function of context and cycle in the blocked cyclic paradigm in Experiment 3. Error bars represent 95% CIs calculated per Cousineau’s (2005) method.