The influence of global discourse on the resolution of lexical ambiguity was examined in a series of naming experiments. Two-sentence passages were constructed to bias either the dominant or the subordinate meaning of a homonym that was embedded in a locally ambiguous sentence. The results provided evidence for the immediate (O-msec interstimulus interval) resolution of lexical ambiguity and were subsequently replicated in Experiment 2, in which an 80-msec stimulus onset asynchrony exposure duration was employed for the homonyms. Strong dominant and subordinate biased discourse contexts activated only the contextually appropriate sense of a homonym. In Experiment 3, each sentence of the discourse was presented in isolation. The pattern of activation obtained in Experiments 1 and 2 was found to be contingent on the integration of the two sentences to construct an overall global discourse representation of the text. The results support a context-sensitive model of lexical ambiguity resolution.Over the past two decades, lexical ambiguity research has sustained a theoretical dichotomy in which the processing of an ambiguous word (e.g., a homonym) is either influenced by context or impervious to it. The bulk of the research on whether context can immediately influence lexical ambiguity resolution has produced mixed results. Early studies supported the principles of modularity (see Fodor, 1983;Forster, 1979), in which the process ofword meaning activation is encapsulated and independent ofhigher level influences ofsyntax and semantics (e.g., Onifer & Swinney, 1981; Seidenberg, Tanenhaus, Leiman, & Bienkowski, 1982;Swinney, 1979). For example, Swinney presented passages aurally, each of which contained an ambiguous word, and found that lexical decision times for both contextually appropriate and inappropriate target words were faster than those for unrelated words. Seidenberg et al. (1982) also employed the crossmodal priming paradigm, but used a naming task as an index ofpriming, and found that naming latencies for target words that were related to the dominant and subordinate meanings of words with noun/verb ambiguities were faster than those for unrelated target words.More recent research has presented a challenge for the modular view of ambiguity resolution. Several studies have supported an interactive activation framework (see Marslen-Wilson & Tyler, 1987;McClelland, 1987), in which context can immediately influence meaning activation in such a way that only the contextually appropriate meaning of an ambiguous word is activated (e.g

In two experiments, we examined the influence of situation-evoking stimuli on the resolution of lexical ambiguity. In Experiment 1, we examined situation-evoking stimuli at an early NP position. Readers were asked to establish whether specific entities were likely to participate as agents in contextually defined situations. Naming latencies demonstrated that defined situations headed by likely agents evoked a domain of reference that included only the situation-appropriate meaning of a targeted lexical ambiguity. In contrast, defined situations headed by unlikely agents evoked a domain of reference that did not include either meaning of the intended ambiguous word. In Experiment 2, we examined situation-evoking stimuli at a later direct object position. The specificity of the theme/patient role filler was manipulated, where the linguistic expressions were either specific or general with respect to a given contextual situation. The results showed that contexts with specific situation-evoking stimuli were rated as strongly biased and provided a domain of reference for the immediate resolution of lexical ambiguity, whereas contexts with nonspecific role fillers were rated as ambiguous and provided a domain of reference that was supportive of both meanings of an ambiguous word. The results were discussed within a contextual-feature-sensitive model of language processing.