Prediction using verb-specific syntactic information in sentence processing

概要

論文の詳細を見る
Recent research has shown that language comprehenders make predictions about upcoming linguistic information using various sources of information (Altmann & Kamide, Cognition 1999; Arai et al., CogPsy 2007; Staub & Clifton, JEP 2006; Van Berkum et al., JEP 2005). Those studies demonstrate that the processor does not only analyze the input that it already received but also projects upcoming unseen elements that are predicted. They also suggest that syntactic structures are built, in an incremental fashion. An important issue concerning any processing theories is granularity of this process. In particular, the current study investigates whether 'fine-grained' lexically-specific syntactic information is used in parsing. The results from two experiments showed that listeners indeed use detailed lexical and morphological information to make predictions about upcoming syntactic structures. The finding is consistent with the sentence processing models that incorporate fine-grained lexical information into their parsing strategy such as constraint-based (MacDonald et al., PsyRev 1994) and probabilistic models (Jurafsky, Cogsci 1996).
社団法人電子情報通信学会の論文
2008-08-01