Encoding interference effects support self-organized sentence processing.

Area of Science:

• Psycholinguistics
• Cognitive Science
• Computational Linguistics

Background:

• Cue-based retrieval theories posit that syntactic dependencies, like subject-verb agreement, are vulnerable to interference from similar noun phrases.
• These models struggle to explain 'encoding interference,' where interference occurs before retrieval, suggesting limitations in current theories.
• Existing models inadequately address interference effects that arise during the initial encoding phase of sentence processing.

Purpose of the Study:

• To propose and evaluate a self-organized sentence processing model as an alternative explanation for encoding interference.
• To provide new behavioral evidence supporting the existence of encoding interference in subject-verb number agreement.
• To demonstrate the unifying potential of self-organization in explaining similarity-based interference in sentence comprehension.

Main Methods:

• Implementation of a novel self-organized sentence processing model.
• Conducting two self-paced reading studies to gather behavioral data on subject-verb number agreement.
• Utilizing a semantic similarity manipulation to elicit and measure encoding interference effects.

Main Results:

• New behavioral evidence for encoding interference was successfully obtained.
• Experimental results demonstrated greater compatibility with the proposed self-organizing account.
• The self-organizing model provided a more parsimonious explanation for encoding interference compared to extensions of cue-based models.

Conclusions:

• Self-organization offers a unifying framework for understanding similarity-based interference in sentence comprehension.
• This approach reduces parsing to feature match optimization, without distinguishing between encoding and retrieval.
• The findings challenge traditional distinctions and advocate for a more integrated view of sentence processing mechanisms.