Knowledge of Statistics or Statistical Learning? Readers Prioritize the Statistics of their Native Language Over the...

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Computational Linguistics

Background:

• Implicit statistical learning is crucial for visual processing and reading.
• Previous research primarily used artificial stimuli like pseudo-fonts or atypical words.
• The role of statistical learning with language-like materials remains underexplored.

Purpose of the Study:

• To investigate if individuals extract local statistical regularities from materials resembling their native language.
• To determine if average bigram frequency influences lexical decisions in a language-specific context.
• To compare the influence of average versus minimal bigram frequency in visual word processing.

Main Methods:

• Two experiments were conducted with Italian speakers.
• Participants performed lexical decisions on letter strings modeled on the Italian lexicon.
• Stimuli were designed such that real words could be distinguished from foils by average bigram frequency.

Main Results:

• No evidence was found for participants utilizing average bigram frequency to distinguish words from foils.
• Lexical decisions were significantly guided by minimal bigram frequency.
• This minimal bigram frequency cue is rooted in the participants' native Italian language.

Conclusions:

• Implicit statistical learning from language-like materials may not rely on average bigram frequency as previously assumed.
• Native language statistics, specifically minimal bigram frequency, play a more significant role in visual word processing.
• Findings challenge existing models of statistical learning and visual word recognition.