<p>A fundamental aspect of intentional action is the ability to generalize learned action-effect relations to novel situations. However, the mechanisms underlying such categorical generalization, particularly in the context of voluntary action selection, remain contested. Across four free-choice experiments, participants first learned that pressing each of two keys produced pictures from one of two categories. The learning experience ranged from repeated presentation of a single dog or car image to one-time presentation of many diverse animal or machine images. At test, participants saw novel pictures from the learned categories and freely chose either key. Results for the total sample consistently demonstrated a significant generalization effect: participants preferentially selected actions congruent with the category of novel test stimuli, even when each learning image appeared only once. At the same time, this group-level effect was partly driven by a subset of participants exhibiting near-deterministic choice patterns, and the statistical evidence for this effect was weakened after excluding them. These findings show that learned action–effect relations can guide voluntary choices for novel category members and that repeated exposure to the specific exemplar is not necessary for such generalization. Individual differences further suggest that explicit action–category knowledge and its strategic use contribute to generalization. From Exemplars to Categories: Generalization of Action-Effect Learning in a Free-choice Task.</p>