It is fairly uncontroversial that when you encounter disagreement with some view of yours, you are often epistemically required to become at least somewhat less confident in that view. This includes political disagreements, where your level of confidence might in various ways affect your voting and other political behavior. But suppose that your opponents don’t comply with the epistemic norms governing disagreement – that is, they never reduce their confidence in their views in response to disagreement. If you always reduce (...) your confidence, but your opponents never reduce theirs – and everyone participates in the political process accordingly – then it seems like the deliberative process will be unfairly skewed in favor of your opponents. In this paper, I do two things. First, I try to explain how this can be so, even though the process by stipulation represents everyone’s beliefs equally. Second, I defend the view that in such cases, you should remedy the unfairness by voting out of accord with your beliefs. By introducing a distinct state that I call a “personal take,” which you can vote on the basis of in such a case, I explain how doing this need not be problematically insincere, nor incoherent from the inside. The discussion has a number of independent upshots for both democratic theory and the epistemology of disagreement. (shrink)