Recently, cognitivist accounts about art have come under pressure to provide stronger arguments for the view that artworks can yield genuine insight and understanding. In Gregory Currie’s Imagining and Knowing: Learning from Fiction, for example, a convincing case is laid out to the effect that any knowledge gained from engaging with art must “be judged by the very standards that are used in assessing the claim of science to do the same” (Currie 2020: 8) if indeed it is to count (...) as knowledge. Cognitivists must thus rally to provide sturdier grounds for their view. The revived interest in this philosophical discussion targets not only the concept of knowledge at the heart of cognitivist and anti-cognitivist debate, but also highlights a more specific question about how, exactly, some artworks can (arguably) afford cognitive import and change how we think about the world, ourselves and the many events, persons and situations we encounter. This paper seeks to explore some of the ways in which art is capable of altering our epistemic perspectives in ways that might count as knowledge despite circumventing some standards of evidential requirement. In so doing we will contrast two alternative conceptions of how we stand to learn from art. Whereas the former is modelled on the idea that knowledge is something that can be “extracted” from our experience of particular works of art, the latter relies on a notion of such understanding as primarily borne out of a different kind of engagement with art. We shall call this the subtractive conception and cumulative conception respectively. The cumulative conception, we shall argue, better explains why at least some insights and instances of knowledge gained from art seem to elude the evidential standards called for by sceptics of cognitivism. (shrink)