Studies In Syntactic Ambiguity Resolution PDF Download

Abstract: The use of two relevant sources of information--frequency and contextual plausibility--during on-line lexical and syntactic ambiguity resolution was examined as a function of verbal working memory capacity in four self-paced reading experiments. Experiment 1 demonstrated that lexical ambiguity resolution was more difficult when disambiguating sentences toward the less frequent homograph meaning. Experiment 2 showed that lexical ambiguity resolution was facilitated when homographs were preceded by strong contextual plausibility information. Experiment 3 found that syntactic ambiguity resolution was facilitated when verb bias information supported the syntactic parse, and experiment 4 demonstrated that contextual plausibility cues which disfavored the incorrect syntactic parse eased sentence comprehension. These results replicated previous research. Additionally, verbal working memory capacity did not mediate ambiguity effects in any of the four experiments. These data best support the working memory theory proposed by Caplan and Waters (1999), who hypothesize that on-line sentence comprehension is not mediated by verbal working memory resources, and these data provide evidence against other working memory theories that have been proposed by Just and Carpenter (1992) and Pearlmutter and MacDonald (1995). Ambiguity effects from each experiment were correlated with the other experiments. An individual's ability to use frequency information correlated with the ability to use contextual plausibility information within lexical ambiguities and within syntactic ambiguities. These results support a model of language in which frequency and contextual plausibility cues interact within each linguistic domain. An individual's ability to use homograph bias was not correlated with the ability to use verb bias, and an individual's ability to use contextual plausibility information was not correlated across lexical and syntactic ambiguities. These results support the autonomy theory, which posits that lexical and syntactic ambiguities are processed independently. These results are also discussed with respect to implications for the lexically-based constraint satisfaction theory, which hypothesizes that lexical and syntactic ambiguities are processed through a single mechanism by the same resources.