Classical theories assume that unconscious automatic processes are autonomous and independent of higher-level cognitive influences. In contrast, we propose that automatic processing depends on a specific configuration of the cognitive system by top-down control. In 2 experiments, we tested the influence of available attentional resources and previously activated task sets on masked semantic priming in a lexical decision task. In Experiment 1, before masked prime presentation, participants were engaged in an easy or hard primary task that differentially afforded attentional resources. Semantic priming was attenuated when the primary task was hard, that is, when only little attentional resources were available. In Experiment 2, a semantic or perceptual induction task differentially modulated subsequent masked semantic priming. Hence, unconscious automatic processing depends on the availability of attentional resources and is susceptible to top-down control.
2
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
Dallas and Merikle (1976a, 1976 b) demonstrated that when participants were presented with a pair of words for over 1 s and subsequently cued to pronounce one of the words aloud (postcue task) semantic priming effects occurred. Humphreys, Lloyd-Jones, and Fias (1995) failed to replicate this postcue semantic priming effect using word pairs that were semantic category co-ordinates. The aim of Experiment 1 was to determine if the disparate postcue task results reported by these researchers could be accounted for by the prime-target contexts or cue types engaging different attentional processes or a combination of these factors. A postcue pronunciation task was used and word pairs presented were taken from an associate-semantic context and a semantic category context. In the Dallas and Merikle condition the line cue flanked the location in which the target word was previously shown. In the Humphreys et al. condition the cue wordUPPERorlowerwas centrally presented and indicated the location in which the target word previously appeared. Results demonstrated that the occurrence of semantic and associate-semantic priming effects under postcue task conditions varied for the two cue types. Experiment 2 investigated if these results were attributable to a between subject manipulation of cue type. Using a fully repeated measures design priming effects were evident for top located targets in both the associate-semantic and semantic prime-target contexts. Experiment 3 used a between subjects design to rule out the possibility that carry over effects between cue and context conditions contributed to the postcue task priming effects. Priming was evident for top located targets in an associate-semantic and semantic context for the line cue. For the word cue there was priming for top located targets from an associate-semantic context and a reverse priming effect for top located targets from the semantic context. Possible explanations for the occurrence of priming effects under postcue task conditions are discussed.