Serial mechanisms in lexical access: the rank hypothesis

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Computational Linguistics

Background:

• Word frequency significantly impacts lexical access time.
• Existing models often use logarithmic functions to represent this frequency effect.
• The underlying mechanisms driving this relationship require further investigation.

Purpose of the Study:

• To investigate the precise relationship between word frequency and lexical access time.
• To evaluate whether rank order or log frequency better predicts access times.
• To extend lexical access models to account for error rates and age of acquisition.

Main Methods:

• Analysis of group and individual lexical access time data.
• Comparison of model fits using frequency rank order versus log frequency.
• Development of an extended search model incorporating error rates and age of acquisition.

Main Results:

• Word frequency rank order provides a superior fit to lexical access time data compared to log frequency.
• The proposed search model successfully accounts for error rates and latencies.
• Age of acquisition is interpreted as cumulative frequency, further refining the model.

Conclusions:

• Lexical access time is more accurately modeled by word frequency rank order than by log frequency.
• Frequency-ordered serial comparison models offer a robust explanation for the observed effects.
• The extended search model provides a comprehensive account of lexical access, including error patterns and developmental effects.