I think we have actually made some progress in making sense of eliminativism, in the sense that we have a clear definition of eliminativism as the view that propositional attitudes are embedded in folk psychology, a false theory whose posits will be eliminated in a matured neuroscience.
We have also agreed that eliminativism faces some problems in construing its philsoophy as genuinely realist. Traditionally, a scientific realist is committed to the truth, as opposed to merely the empirical adequacy, of scientific theories. But if having a true theory means having true beliefs, we have a problem. The Churchlands, like Steven Stich and William Lycan, in the late 80s and early 90s, were moving toward a pragmatism that dethrones truth. Of course, Churchland proclaimed that new theories would provide us some thing better than truth, but what would that be, something more true that truth? Dennett had argued on pragmatic grounds for folk psychology while denying that it was ultimately true. But if truth is replaced by something pragmatic, then these justifications for FP are as good as there can be. As I wrote in 1991:
"The dethronement of truth opens up the possibility of a much looser form of pragmatism: a non-propositional cognitive science may be the best way to go, propsotional attitude attributions are prefectly justified in other contexts, and there just isn't any question of limning the true and ultimate structure of reality. This may not be acceptable to eliminativists like Churchland, but one would like ot know why not."
Belief in the unity of science, for example, which is one of Churchland's fundamental commitments, is undermined by going pragmatist.
However, it is now suggested that perhaps a non-propostional concept of accuracy can save eliminativism. I'm not sure it makes sense; i think the concepts used are parasitic upon the older "folk' concepts. But that is what is at issue.