This paper scrutinises the limits of a posteriori induction in acquiring modal knowledge. I focus on my similarity-based account (Roca-Royes [2017]); an inductive, non-rationalist epistemology of modality about concrete entities. Despite the explanatory merits of the account in relation to a vast range of modal claims, this inductive epistemology has been found incapable of yielding knowledge of a certain, other range of modal claims. Here, two notions of knowability are distinguished which reveal some of these limitations to be not only accidental to the method but also virtuous. Additionally, the scrutiny suggests a recipe for increasingly pushing back, as modal enquirers, some of these limits. Limits will irremediably remain. But, as modal epistemologists, it is to explain what lies beyond these irremediable limits (not within) that we should look somewhere else.
