Language learning, language use and the evolution of linguistic variation.

Area of Science:

• Cognitive Science
• Linguistics
• Computational Linguistics

Background:

• Linguistic universals emerge from the interplay of language learning and use.
• Linguistic variation is typically constrained by linguistic or sociolinguistic factors.
• The extent to which unpredictable variation reflects statistical learning biases is debated.

Purpose of the Study:

• Investigate the relationship between statistical learning biases and linguistic variation.
• Explore how language learning, transmission, and use shape language structure.
• Understand the mechanisms behind the scarcity of unpredictable variation in natural language.

Main Methods:

• Reviewed experimental work on linguistic variation and statistical learning.
• Introduced a Bayesian model for learning and transmission of linguistic variation.
• Conducted an artificial language learning experiment with adult participants.

Main Results:

• Learner biases can influence language structure, but the relationship is complex.
• Weak biases can significantly impact language structure through cumulative transmission.
• Strong biases may have limited effects, and language use can also reshape systems.

Conclusions:

• The interaction between learning biases, transmission, and use is crucial for shaping language structure.
• The scarcity of unpredictable variation is not solely explained by statistical learning biases.
• A comprehensive understanding requires integrating insights from learning, transmission, and language use studies.