The Idiom Processing Advantage is Explained By Surprisal

Area of Science:

• Psycholinguistics
• Cognitive Science
• Computational Linguistics

Background:

• Idioms are processed faster than literal phrases, sparking debate on chunking vs. compositionality.
• Existing research offers conflicting conclusions regarding the impact of compositionality on idiom processing.
• Surprisal theory, an expectation-based model of sentence processing, offers a potential unifying framework.

Purpose of the Study:

• To investigate idiom processing through the lens of information update and surprisal theory.
• To determine if surprisal can explain the idiom processing advantage.
• To examine the role of compositionality versus predictability in idiom comprehension and production.

Main Methods:

• Comprehension experiments on verb-object idioms.
• Production experiments on verb-object idioms.
• Analysis of idiom processing using surprisal metrics and comparison with matched literal phrases.

Main Results:

• The idiom processing advantage is largely explained by lower surprisal values for idioms compared to literal phrases.
• Idioms exhibit lower surprisal, indicating greater predictability within an expectation-based framework.
• The idiom advantage is most prominent on the noun component of verb-object idioms.

Conclusions:

• Surprisal theory provides a more general framework for understanding idiom processing advantages than compositionality alone.
• The predictability of idioms, quantified by surprisal, is a key factor in their rapid processing.
• Future research should consider information-theoretic approaches to further elucidate language processing mechanisms.