The problem I want to raise in this paper will not be a problem for anyone who does not share both the following concerns or prejudices. The first is that we should take a materialist, extensional approach to linguistic meaning, specifically, the approach that is suggested by the work of Donald Davidson. The second is that social analysis must be emancipatory, and that this requires the conceptual possibility of postulating social structures that are in some sense hidden by discourse. The problem I want to address lies in a possible conflict between these concerns, both of which I will clarify, though not justify (except in so far as we take their compatibility to be apart of their respective justifications). The problem is this: Does the methodological core of Davidson's semantics—the principle of charity—permit interpretation informed by a concept of ideological concealment?
As with many of our philosophical problems, the existence of this one is quite precarious. We can avoid avoiding it only if we move along a rather specific route, which, I am afraid, will constitute a good chunk of my paper.