Talk by Alina Ahrens (A2), University of Bremen, on “Action-effect learning: An attempt at replication and generalization”
Time: Thursday, 1-2 pm
Location: Liebermeisterstraße 18 (ZITh), R. 318, and via research colloquium zoom.
Abstract: According to ideomotor theory, actions are represented through their effects. For goal-directed action to occur, one must therefore learn an association between an action and its effect (i.e, the outcome). Through this association, the action becomes activated once the representation of its effect becomes anticipated and thereby activated. This process of acquiring bidirectional action-effect (A-E) associations, and their influence on later actions, is commonly studied under the label of A-E learning. Numerous studies show that such associations can be acquired through repeated experience of action-effect pairings, that they subsequently influence action selection and performance in later tasks, and that they generalize to similar effects. However, many of these studies employ free-choice test phases or forced-choice learning phases, which are design choices that have recently been criticized for potentially reflecting strategic decision-making rather than genuine A-E learning, and for being confounded with stimulus-stimulus learning. We recently failed to find evidence for effect generalization, and even for fundamental action-effect learning, across two conceptual replications: in one, we replaced a forced-choice learning phase with a free-choice phase (Ahrens et al., 2026, Psychological Research); in the other, we used a design combining the learning and test phases into a single phase. In my talk, I will outline the methodological challenges of A-E learning research, present the results of these two failed replication attempts, and discuss possible reasons for the contradictory findings within the literature.
