Constraint satisfaction as a theory of sentence processing

Area of Science:

• Psycholinguistics
• Computational Linguistics
• Cognitive Science

Background:

• The constraint satisfaction model proposes parallel processing of syntactic structures.
• Empirical support for this model is questioned regarding its divergence from simpler models.
• Methodological issues also challenge the model's validity.

Purpose of the Study:

• To critically evaluate the constraint satisfaction model of sentence processing.
• To identify its theoretical and empirical shortcomings.
• To assess its explanatory power in light of psycholinguistic evidence.

Main Methods:

• Review and critique of existing empirical evidence supporting the constraint satisfaction model.
• Analysis of the model's predictions against established findings in sentence processing.
• Examination of the model's capacity to explain various syntactic phenomena and psycholinguistic data.

Main Results:

• The model fails to account for diverse syntactic structures, including local, global, discontinuous, and long-distance dependencies.
• It makes inaccurate predictions regarding the timing of syntactic analysis.
• The model cannot explain the independence of lexical and syntactic structure observed in psycholinguistic studies.

Conclusions:

• The constraint satisfaction model is inadequate as a theory of sentence processing.
• Its reliance on lexical entries for syntactic structure is problematic.
• The model lacks restrictiveness and explanatory power, often relying on post hoc correlations.