Well-known epistemologies of science have implications for how best to understand knowledge transfer (KT). Yet, to date, no serious attempt has been made explicate these particular implications. This paper infers views about KT from two popular epistemologies; what we characterize as incommensurabilitist views (after Devitt 2001; Bird 2002, 2008; Sankey and Hoyningen-Huene 2013) and voluntarist views (after van Fraassen 1984; Dupré 2001; Chakravartty 2015). We argue views of the former sort define the methodological, ontological, and social conditions under which research operates within ‘different worlds’ (to use Kuhn’s expression), and entail that genuine KTs under those conditions should be difficult or even impossible. By contrast, more liberal voluntarist views recognize epistemological processes that allow for transfers across different sciences even under such conditions. After outlining these antithetical positions, we identify two kinds of KTs of two kinds present in well-known episodes in the history of ecology—specifically, successful model transfers from chemical kinetics and thermodynamics into areas of ecological research—which reveal significant limitations of incommensurabilitist views. We conclude by discussing how the selected examples support a pluralistic voluntarism regarding KT.