Proceedings of the Annual Meeting of the Cognitive Science Society

This paper investigates whether typical experimental practices in cognitive psychology instantiate Hempel's paradox. Using a hypothetical priming experiment, it argues that a failure to observe both concept activation and judgment bias—usually taken as disconfirming—can logically count as confirmation via the contrapositive of the hypothesis. Two studies asked participants to judge whether four possible experimental outcomes confirmed, disconfirmed, or were irrelevant to a hypothesis linking concept activation to biased judgment. Participants consistently failed to treat the contrapositive outcome as confirmatory, both in original and replication contexts. Latent class analyses suggested reliance on satisficing rather than formal logical reasoning. Simulations using a large language model showed partial alignment with human judgments, especially for disconfirmatory outcomes, but clear divergences for confirmatory cases. Overall, the results suggest that standard psychological experiments can give rise to Hempel's paradox and that both humans and LLMs systematically deviate from formal confirmation logic.