This article takes a renewed look at the concept of “affordance.” It points out that the concept is being used within the CSCL community in ways which signify an underlying disagreement concerning the exact ontological nature and epistemological status of an “affordance.” Such disagreement, it is argued, is a problem for both design and empirical research. Because HCI discussions of the concept have informed CSCL, views presented within this discourse are discussed. A Merleau-Pontian account of affordances is developed, building on his view of the human being as always already being-in-the world in a non-thematized, pre-reflective correspondence of body and world in the concrete activity. A dynamic, agent-centred, cultural-, experience- and skill-relative, but perception-independent, ontology is proposed for affordances. Toward the end of the article, examples are given of how the Merleau-Pontian account of affordances may shift the focus of empirical research and of design processes within CSCL.