Syntactic ambiguity resolution in discourse: modeling the effects of referential context and lexical frequency

Area of Science:

• Psycholinguistics
• Computational Linguistics
• Cognitive Science

Background:

• Sentences with temporarily ambiguous reduced relative clauses pose processing challenges.
• Discourse context is known to influence syntactic ambiguity resolution.
• Previous research yielded conflicting results on referential context effects.

Purpose of the Study:

• To investigate how discourse context affects the processing of temporarily ambiguous reduced relative clauses.
• To examine the interaction between discourse bias and verb properties in sentence comprehension.
• To test a constraint-based competition framework for parsing and interpretation.

Main Methods:

• Presented sentences with temporarily ambiguous reduced relative clauses preceded by biasing discourse contexts (main clause vs. relative clause).
• Monitored eye movements during reading to measure processing difficulty in disambiguating regions.
• Employed regression analyses to examine interactions between context, verb frequency, and reading times.
• Modeled reading times using a constraint-based competition framework.

Main Results:

• In relative clause biasing contexts, ambiguous reduced relatives were processed as easily as unambiguous or full relatives.
• Discourse context effects at ambiguity points interacted with the past participle frequency of the ambiguous verb.
• Reading time data supported a constraint-based competition model of parsing.

Conclusions:

• Discourse context effectively resolves temporary syntactic ambiguity in reduced relative clauses.
• A constraint-based competition framework successfully integrates multiple linguistic constraints during parsing.
• This framework reconciles previously conflicting findings on referential context effects in ambiguity resolution.