Event Details

The Cognitive Science Distinguished Lecture Series

Presents its First Speaker!

Wednesday, March 22nd 5:30-7:00

Theater/Dance Room 200

Macalester College

Pizza & Salad Dinner

Sashank Varma

School of Interactive Computing & School of Psychology

Georgia Institute of Technology

  “Understanding Graded Words: Behavioral Experiments and Computational Modeling”

Abstract

 Language understanding and mathematics understanding are two fundamental forms of human thinking. The current study investigated whether the magnitude representations that anchor understanding of numbers are also recruited to understand graded words. These are words that come in scales (e.g., Anger) whose members can be ordered by the degree to which they possess the defining property (e.g., calm, annoyed, angry, furious). Experiment 1 used the comparison paradigm to find evidence that the distance, size, and boundary effects that are present for numbers, and considered evidence of the recruitment of magnitude representations, are also present for words. Experiment 2 collected similarity ratings and used multi-dimensional scaling to find converging evidence for these effects in human understanding of graded words. Experiment 3 looked for the same effects in the vector word representations of Large Language Models such as BERT. The failure to find them suggests that human understanding of graded words transcends the mere statistical structure of the linguistic environment. This research informs the role of machine learning models in cognitive psychology theorizing.

Biography

Dr. Varma earned his B.S. in Mathematics and Cognitive Science from Carnegie Mellon University, his Ph.D. in Cognitive Psychology from Vanderbilt University, and completed a post-doctoral fellowship at Stanford University. He was the Huebner Chair in Education and Technology at the University of Minnesota before moving to Georgia Tech.