Human Cognition Colloquium: Human communication as a functional window into human cognition

Colloquium | January 29 | 3 p.m. | 1102 Berkeley Way West

 Kyle Mahowald, Stanford University

 Department of Psychology

Language is an ideal test bed for exploring many core questions about the origins and structure of human cognition, learning, and culture. Whereas many cognitive tools are similar across cultures, there is wide diversity among human languages. To state that observation statistically: humans languages have some fixed parameters (universals) but also a large number of degrees of freedom. Thus, the set of world languages offers an elegant natural experiment. In this talk, I will use ideas from computer science about efficient communication and ideas from psychology about cognitive constraints on human language processing to generate hypotheses about efficient language structure and how it informs our understanding of human cognition. In the first section, I will focus on the two-way relationship between domain-general cognitive structure and functional language pressures. Applying these ideas to syntax, I show that, across 37 world languages, the distances between dependent words are minimized: evidence of cognitive pressure at play in large-scale language structure. Second, I will focus on how communicative pressures constrain language. For instance, consistent with predictions from Shannon’s information theory, languages are optimized such that words that convey less information are shorter and easier to produce/understand. In both a corpus experiment and a behavioral experiment, word shortenings like chimpanzee -> chimp are more likely to occur when the context is predictive. Third, I will discuss my ongoing and future work in meta-science and best practices in psychology methods. I conclude with future research directions: first, how corpora of academic papers can be used to study interpersonal dynamics in language and, second, how we can use the tools of AI and NLP (focused on artificial neural systems) to understand cognitive processes.