Computational Comparison of Children and Apes on a Non-Verbal False Belief Task

Margeaux CiraoloUniversity of Hawaii at Manoa, Honolulu
Samantha O'HanlonUniversity of Hawaii at Manoa, Honolulu, HI, USA
Leonidas DoumasUniversity of Edinburgh, Edinburgh, UK

Abstract

The key difference between the cognitive abilities of humans and other animals may be the ability to reason relationally; models of relational reasoning are one way to demonstrate this proposed difference. The present work uses the DORA model to simulate a task designed to assess the theory of mind capabilities of 4- and 5-year-old children and apes. In the original experiment, the apes and children successfully completed a number of control tasks in which they used cues from an experimenter to reason about the hiding location of a reward. However, only the children succeeded on the critical manipulation in which it was necessary to infer what the experimenter knew. The simulations presented herein demonstrate that the apes’ performances across all tasks can be accounted for by simple rule use. Conversely, the 4-year-olds succeeded via relational inference and learning; 5-year-olds alone had the requisite relational structures predicated beforehand.

Files

Computational Comparison of Children and Apes on a Non-Verbal False Belief Task (372 KB)



Back to Table of Contents