Abstract
Why do we think in the absence of immediate or remote needs if cognitive operations are costly, and what is it with problems that make resolving them satisfying? One possibility is that problem difficulty drives resolution satisfaction. The more uncertainty resolved, the more satisfying is the experience. Another possibility is that the cognitive satisfaction stem from evidence of resolution, and is predicted by resolution confidence. We tested these hypotheses in two online mathematical-logical forced-choice tasks involving 30 and 40 problems, and 96 and 102 participants, respectively. In both experiment, participants were generally more satisfied, the easier the problem was that they correctly solved and we found that resolution satisfaction was highly related to resolution confidence ratings. Moreover, difficulty and confidence interacted to bestow a small but reliable satisfaction boost such that high confidence responses to hard problems were more satisfying than high confidence responses to easy problems. Our results suggest that cognitive satisfaction comes out of a state of comprehension confidence, modulated by surprise. We propose that thinking is the unavoidable consequence of a cognitive system natively set on comprehension and that comprehension is natively rewarded and therefore sought. We also found that uncertainty did not explain curiosity, as neither objective difficulty nor confidence substantially predicted curiosity or costly information seeking. Taken together, our results suggest that the cognitive drive aspires for a reliable state of explanation, rather than a reduction in uncertainty. Our findings imply that congruence expectation incentivize information seeking, and explains the reluctance to seek contradictory evidence.