Abstract In Egon Brunswik’s theory of probabilistic functionalism, human prediction is conceptualized as an inductive inference process, in which cues are utilized as a lens to predict the probabilistically associated criterion in the environment. Dynamic cognitive adjustment, driven by the uncertainty of the individual and the substitutability of the environment, is based on vicarious functioning, the principle of learning from the frequency of co-occurrences. However, previous models of vicarious functioning, the multiple-regression lens and the fast-and-frugal lens, fail to explain how the individual reduces uncertainty while learning ecological cue validities. We therefore developed a diffusion lens model of vicarious functioning that captures dynamic cognitive adjustment to environments with multiple probabilistic and substitutable cues. A superstatistics approach allowed us to account for uncertainty reduction over time by an increasing sensitivity of the drift rate to the ecological validity of the cues. Additionally, the non-decision time is assumed to increase linearly with the number of presented cues to account for cue substitutability in the environment. The resulting model was validated by successfully fitting it to response time and choice data previously collected across multiple-cue probability learning tasks in diverse environments and scenarios. This suggests that the diffusion lens model can explain cognitive adjustment from an initial absence of knowledge to a near-perfect approximation of the probabilistic environment.
Scholten et al. (Tue,) studied this question.