Proceedings of the Annual Meeting of the Cognitive Science Society

Human adults share the ability to approximate large quantitieswithout counting with newborn infants and non-human species.This ability is supported by the Approximate Number System(ANS) - a primitive and domain-specific cognitive system thatsupports noisy numerical decisions. How does the ANS supportactive exploratory decisions? Using a numerical comparisontask, we found that the amount of active information seekingdoes not simply increase as the decision becomes more difficult.Instead, there seems to be an inverted U-shaped relationshipbetween trial difficulty and how much one chooses to seekinformation. Additionally, this effect is not modulated byparticipants’ performance, suggesting that participants’exploratory decisions based on ANS representations are drivenby the utility of information seeking actions.