In recent blogging (here and elsewhere), Mohan Matthen has been reflecting on the 'decline' of general philosophy of science (GPOS, as opposed to the flourishing of the philosophy of particular sciences--designated as 'PoX'). On the sociological front: there is no doubt that in Anglophone philosophy the resurgence of post-Lewisian analytic metaphysics (PLAM, which as I have argued elsewhere and with follow-ups I can't locate, but see also here and here) often has only tenuous link to science and philosophy of science) has displaced GPOS [maybe not numbers-wise, but as central to discussion in the discipline]. (In Europe the story is largely different; so-called analytic philosophy and philosophy of science remain more mutually co-identified.) Now, in order to revive GPOS it matters quite a bit to figure out i) where GPOS comes from and what it was about; ii) what its zenith was; iii) what went wrong? iv) what the aim of GPOS might be. (Now as was revealed in comments over Mohan's earlier post, he and I disagree about where it comes from (he thinks Vienna and a concern with unobservables, I think Philadelphia and Paris with an attempt to distinguish physics and metaphysics); in his view the zenith was Vienna and its consolidation Stateside; whereas I think the many-sided debates among Kuhn, Feyerabend, Popper, Toulmin, Polyani, Hanson, were the zenith.) But here I respond with some (lightly connected observations) to Mohan's diagnosis of the second question.
1. Now in his post over at itisonlyatheory, Mohan writes, "A good bit of General Philosophy of Science is influenced by the desire to provide a response to scepticism that is better than mere "dogmatism". And this is a problem." In so far Mohan is correct, then his subsequent and very interesting argument (epistemology deals with content-general arguments, while much GPOS fails to realize it can only deal with content-specific arguments, dealing with measurement, experiment, etc.) is on target and devastating: much GPOS would be based on a flawed self-understanding (however tacit)! One thing I like about Mohan's observation is that realism/anti-realism debate is not really a debate within GPOS (and this fits my lazy prejudices nicely).
3A. I take the LSE/Leeds/Bristol commitment to Structural Realism as one response to this post-Kuhnian debacle. In many ways it is the only fully general GPOS *program* left standing, although Van Fraassen gives it a great run for its money (and, of course, there are some terrific, idiosyncratic GPOS figures -- Mark Wilson, Jody Azzouni, George Smith, Bill Wimsatt, etc). 3B. There has, of course, been an explicit revival of neo-Kantianism following the work of Friedman, but it remains physics-focused; 3C. Pragmatism has been constantly renewed (including by my colleagues in Ghent). 3D. I take the folk who think they can discern a topic neutral account of, say, causation (explanation, mechanism, etc) as other responses, although in some respects these lower the ambition(s) of GPOS.
4. It's only when from his arm-chair Quine makes science an extension of common sense that epistemology (naturalized and eventually Gettierized) regains a foothold within GPOS. But from the vantage-point of GPOS, analytic epistemology is vulgar because it has always been too respectful of ordinary practice (and locutions).
5. One final thought here. In my view GPOS never properly theorized its own relationship with science. The quasi-transcendental assumption blocks inquiry into the multi-level relationship between science and philosophy. This suited the proponents of so-called scientific philosophy just fine, because they used the authority of science to settle or close debates within philosophy and to model philosophy on science. I dub this move (and its related arguments), "Newton's Challenge to Philosophy." But once the shared program died (pace points 3A-D above), GPOS was left with few resources to prevent its displacement by analytic metaphysics.
No doubt this is all too terse, but I have gone on long enough.