A common epistemic principle is that certainty is never justified: there is always some nonzero probability of error. But I’m struggling to see how this principle can itself be justified without ...