A meta-analysis of 26 studies indicated that automatic semantic priming can occur without association. Priming did not vary substantially with differences in variables that affect automatic versus strategic processing, such as time spent processing the prime and target, relationship proportion, and task (except that average effects were smaller in the naming task). Although category coordinates were investigated in the majority of studies, synonyms, antonyms, and script relations also demonstrated priming; functional relations showed greater priming, and essential and perceptual relations showed less. The average effect size for semantic priming was smaller than that for associative priming, suggesting that there is an "associative boost" from adding an associative relationship to a semantic one. The implications of these findings for the modularity thesis and for models of semantic priming are discussed.

No abstract

The level of representation accessed when inferences are made during sentence comprehension was examined. The inferences investigated included antecedent assignment for both definite noun phrase anaphors and pronouns and also instrument inferences. In making these inferences, a listener must access the inferred element, whether an antecedent or an instrument, in either a linguistic form representation or a discourse model. The level of representation involved in these inferences was determined by exploiting differences in the lexical decision and naming tasks, which were argued to exhibit differential sensitivity to representationallevels. In three experiments, the priming of antecedent and instrument targets in the lexical decision task was compared with priming of the same targets in the naming task. Differences in the patterns of activation across the two tasks indicated that all three types of inferences required accessing elements in a discourse model. Three control experiments ruled out simple context or congruity checking as an explanation for our results. The following conclusions were also supported by these studies:(1) Antecedent assignment occurs immediately after processing an anaphor; (2) antecedent assignment involves inhibition for the inappropriate antecedent rather than facilitation for the appropriate antecedent; (3) although subjects do not make instrument inferences when they hear isolated sentences containing verbs that strongly imply certain instruments, the inferences are made when sentences are preceded by a context that mentions the instrument. 611Understanding sentences in discourse involves integrating information in the sentence currently processed with information in a memory representation of the previous discourse. Sometimes the need for integration is explicitly marked, as when a sentence contains an anaphoric expression. Often it is less explicit (and necessary for maintaining coherence), as when actions mentioned in the sentence imply agents or instruments that were mentioned earlier in the discourse. A great deal of psycholinguistic research has been devoted to investigating how often, at what point in the sentence, and under what circumstances integrative processes occur. A basic question that has received relatively little attention, though, is that of the form of the representation involved in integrative processes. Although a common assumption is that integrating an element in the currently processed sentence requires changing the activation of an element in a memory representaThe research presented here was supporte

No abstract

The modularity hypothesis was evaluated in this review of 25 studies that investigated context effects in early lexical processing. Selection criteria restricted the review to priming studies and to studies that presented the target word before or at the end of the prime word. Meta-analysis was used to provide summary information about the 17 studies in the review for which effect sizes could be calculated. Overall, the analysis revealed a small effect of context on lexical access. Results provide a disconfirmation of the modularity hypothesis; although multiple interpretations of a word were sometimes accessed, the contextually appropriate interpretation was consistently more strongly activated. The lack of significant heterogeneity of variance suggests that the variability observed across studies was due to sampling error; however, task, timing of target presentation, meaning frequency, type of target, and type of context were identified as potential moderator variables.