Characterizing semantic space: neighborhood effects in word recognition.

Area of Science:

• Psycholinguistics
• Cognitive Psychology
• Computational Linguistics

Background:

• Understanding the mental lexicon's structure is key to word recognition research.
• Lexical semantics, particularly its structural characteristics, remains challenging to define and measure empirically.
• Existing models of lexical semantics (e.g., feature sharing, category membership, associations, co-occurrences) lack sufficient empirical validation.

Purpose of the Study:

• To empirically test and compare two distinct models of lexical semantics.
• To investigate the effect of semantic neighborhood size on visual word recognition.
• To determine if semantic neighborhood measures derived from association and co-occurrence models can predict lexical processing performance.

Main Methods:

• Conducted experiments examining visual word recognition.
• Defined semantic neighborhood size using two models: one based on word associations, the other on global co-occurrences.
• Measured performance on lexical decision and word naming tasks.

Main Results:

• Semantic neighborhood size, as defined by both association and co-occurrence models, significantly impacts word recognition.
• Both tested semantic neighborhood measures were found to be predictive of performance in lexical decision and word naming.
• The findings support the application of these semantic measures in understanding lexical activation.

Conclusions:

• Semantic neighborhood size is a crucial factor in visual word recognition.
• Models defining semantics via associations and co-occurrences offer viable approaches for empirical investigation.
• These findings contribute to a more refined understanding of the mental lexicon's structural properties and word recognition processes.