Philosophical Issues, 23, Epistemic Agency, 2013 WHY WE CANNOT RELY ON OURSELVES FOR EPISTEMIC IMPROVEMENT Kristoffer Ahlstrom-Vij University of Kent, Canterbury Abstract There is something very appealing about the idea that we are epistemic agents. One reason—if not the main reason—is that, while we are undoubt- edly fallible creatures, us being epistemic agents that do things means that it might just be within our power to improve and thereby do better. One important way in which we would want to improve is in relation to our well- established tendency for cognitive bias. Still, the proper role of epistemic agency in us avoiding or correcting for cognitive bias is highly limited. In fact, what we know from empirical psychology—particularly with respect to our tendencies for overconfidence—suggests that we cannot rely on our- selves for epistemic improvement, and have good reason to impose significant constraints on our ability to exercise such agency in ameliorative contexts. 1. What is Epistemic Agency? As human beings, we are not merely passive recipients of information. We interact critically with our surroundings, mull over our data, reflect on the merits of our beliefs given our evidence, collect more information when we feel that is needed, consult others who we believe to be informed on the relevant matters, and so on. In conducting inquiry thus, we are doing things, and are as a result properly called agents. More specifically, we are epistemic agents, in that we are doing things in pursuit of specifically epistemic goals. This tells us something about what epistemic agency is. It tells us that the domain of epistemic agency encompasses all the things that we do in the pursuit of specifically epistemic goals. At the same time, describing epistemic