Title

A Comparative Investigation into the Mechanisms that Support Multidimensional Categorization

Abstract

Numerous studies have reported dissociations between rule-based (RB) tasks, which are assumed to engage declarative decision rules, and information-integration (II) tasks, which are assumed to engage associative category learning mechanisms. However, it remains unclear whether successful RB- and II-learners reliably deploy distinct strategies. The experiments reported here explored novel means of assessing which mechanisms participants use to solve RB and II tasks. We conducted a series of cross-species comparisons between humans, who are believed to possess both declarative and associative mechanisms, and pigeons, who are believed to possess only associative mechanisms. In accordance with a prominent dual-systems framework – the COVIS categorization model (Ashby et al., 1998) – we found that pigeons appeared to solve both RB and II tasks using associative mechanisms, whereas humans appeared to solve RB tasks through unidimensional decision rules and likewise explored unidimensional rules in the early stages of II tasks. However, contrary to the predictions of COVIS, II tasks did not reliably encourage human participants to abandon declarative mechanisms; instead, most II-learners appeared to develop alternative declarative strategies. We also found that pigeons, despite showing no evidence of declarative processing, robustly succeeded at tasks that challenged even human learners, suggesting that associative learning mechanisms merit far greater attention than they often receive in the broader category learning literature.

Papers  related

O'Donoghue, E. M., Broschard, M. B., Freeman, J. H., & Wasserman, E. A. (2022). The Lords of the Rings: People and pigeons take different paths mastering the concentric-rings categorization task. Cognition, 218, 104920. https://doi.org/10.1016/j.cognition.2021.104920

O'Donoghue, E., & Wasserman, E. A. (2021). Pigeons proficiently switch among four tasks without cost. Journal of Experimental Psychology: Animal Learning and Cognition, 47(2), 150–162. https://doi.org/10.1037/xan0000287