What is a Cognitive Process Model? A Disambiguation

Abstract

Our article proposes a general framework for cognitive process models. A large number of models, including accumulator, exemplar, diffusion, heuristic, and network models, are called process models, and working definitions of the term are numerous. Yet, scientific concepts of process models differ. Researchers lack consensus about which models rightfully merit the label. The literature lacks consensus about which aspects related to form (simple form, probabilistic form), content (reaction time, discrete time), or data (process data, behavioral data) matter for process models. The theoretical framework proposed in this paper defines process models as characterized by four dimensions: (a) they have a tri-modal conceptual scope which individuates the information entering the cognitive system, the phenomena leading to the behavior of interest, and the behavior to be modeled; (b) they allow precise, testable predictions to be derived for both, the behavior and the process; (c) the process predictions can be derived separately from the behavioral predictions, and specifically without reverse inference from the behavioral data, and (d) the information transformation in the model is plausible in the sense of being compatible with the contemporary body of knowledge about human cognitive capacities. The framework can be applied to cognitive models before or after they are empirically tested. It provides dimensions on which process models can be compared. It constitutes a basis for a taxonomy of cognitive models. It offers a checklist to help scholars build new cognitive process models. And it can advance currently unresolved debates about model types.

Keywords: cognitive process model, computational model, Marr’s levels, definitions