Work

A learning account of cognitive control in word production: A path to new insights into the domain generality of control

Year: 2026
Type: preprint
Abstract: Various computational models have been proposed to explain interference and its resolution in word production. Of these, incremental learning models provide an elegant explanation of the interfering effects of the similarity between a target word and its context. However, such models have not been extended to explain how the production system deals with other kinds of interference, such as the presence of a prepotent distractor. It thus remains unclear whether incremental learning can furnish a general mechanism for explaining interference and its resolution in word production. This paper proposes the Cognitive Control, Production, and Learning Model (CoCo-PaLM), a comprehensive learning-based model of interference and control in word production. In addition to capturing the previously modeled effects, CoCo-PaLM predicts sequential effects of control, as well as effects involving the transfer of control between the language system and non-linguistic systems. More broadly, the model establishes learning as a computational principle for explaining interference and control and offers a framework for dissociating the neural domain-generality of control from its implementation in specific domains.
Language: English
Open Access status: green
OpenAlex ID: W7236376552
Cites:
Cited by:
Related to: 10