Greater Early Disambiguating Information for Less-Probable Words: The Lexicon Is Shaped by Incremental Processing

Area of Science:

• Linguistics
• Cognitive Science
• Information Theory

Background:

• The lexicon evolves balancing production ease and communicative accuracy.
• Zipf's law of abbreviation suggests word length correlates with probability, indicating communicative efficiency pressures.
• Not all sounds contribute equally to disambiguation; information content depends on lexical context and position.

Purpose of the Study:

• To investigate if less-probable words contain sounds with higher disambiguating information.
• To determine if this information distribution is influenced by the incremental perception of speech.
• To examine cross-linguistic patterns in word structure related to communicative efficiency.

Main Methods:

• Analysis of a diverse language dataset.
• Quantification of sound information content within words.
• Correlation analysis between word probability, length, and information contribution of sounds.

Main Results:

• Less-probable words, beyond being longer, possess sounds conveying greater disambiguating information.
• This effect is more pronounced at the beginning of words, where sounds have maximal impact.
• Evidence supports the lexicon's optimization for efficient communication through information-rich sounds in less frequent words.

Conclusions:

• Word structure is shaped by communicative efficiency, favoring informative sounds in less frequent words.
• The incremental nature of speech perception influences this optimization, particularly at word onsets.
• Findings provide cross-linguistic support for information-theoretic principles governing lexicon evolution.