Abstract
According to a recent argument, we should de-idealise analytic epistemology: when theorising our epistemic practices and obligations, we should stop abstracting away epistemically relevant features of knowers, such as their social identities. This paper puts in dialogue two apparently unrelated arguments for non-ideal epistemology - McKenna (2023) and Basu (2023) - focusing on the obligation to engage with challenges to our beliefs. I argue that while these are welcome first steps towards de-idealisation, we need to de-idealise further by acknowledging the role that groups (of a very peculiar kind) play in arguments for de-idealising. Many of the things we want to say about knowers in a non-ideal epistemology, I show, turn on these knowers' group membership and the epistemic authority ascribed to their group. Acknowledging this incurs the burden of theorising such groups. Accepting this burden, I aim to show, is both necessary and worthwhile for a fully non-ideal epistemology.