Salience-weighted agreement feature hierarchy modulates language comprehension.

Area of Science:

• Neuroscience
• Psycholinguistics
• Computational Linguistics

Background:

• The brain incrementally builds sentence meaning using linguistic cues.
• Subject-verb agreement, where verb marking matches subject features (person, number, gender), is a common linguistic mechanism.
• Cross-linguistic variation in these features suggests differential modulation of language comprehension.

Purpose of the Study:

• Investigate the online processing of subject-verb agreement in Arabic.
• Examine how person, number, and gender features are processed during comprehension.
• Determine if agreement violations elicit differential brain responses based on the violated feature and subject type.

Main Methods:

• Utilized a visual event-related brain potential (ERP) study.
• Analyzed simple intransitive Arabic sentences with full or partial subject-verb agreement violations.
• Measured ERP responses at the post-nominal verb.

Main Results:

• Agreement violations elicited a biphasic negativity-late-positivity ERP effect.
• This effect was systematically graded based on the specific feature violated (person, number, or gender).
• The graded effect differed qualitatively for singular versus plural subjects, reflecting feature salience.

Conclusions:

• Language comprehension is differentially modulated by agreement features based on their language-specific salience.
• Proposed a Salience-weighted Feature Hierarchy to explain cross-linguistic neurophysiological findings in verb agreement processing.
• Findings highlight the dynamic and feature-dependent nature of syntactic processing in the brain.