Well no, I don't expect us to do without meta-discussions or the like. However, what I'm trying to get at isn't a distinction between substance and non-substance here (though that is where I think lots of philosophy is substantively wrong) but rather a bias towards theories with a certain flavor and complexity.
For instance I think Carnap does a truly excellent job of dealing with many meta-philosophical questions and while he certainly isn't entirely correct about everything (logical probability was a big mistake and he had some problems accepting some of the points Quine got right) I think many of his approaches are a much better choice than the more modern approach of throwing in lots of ambiguous concepts and building complex structures to talk about meta-philosophy.
Now I personally tend to think Carnap's approach of always taking philosophy to be postulating a certain model and then dividing up issues into internal quesitons and external ones is the right approach to take. I'm certainly not saying that others have to agree. However, I'm pointing out that it's odd that the answers a Carnapian approach would give to these problems have such an extremely small mindshare in philosophy today.
Now if I thought that this approach had died out because it had been out argued that would be fine. However, my impression is that it has lost mindshare for exactly the opposite problem: it answers questions to well. The Carnapian answer to the problem doesn't leave you any more interesting philosophy to do. You simply say, "well that's an external question, just pick whatever best suits your purpose," and that's the end of the conversation apart from debating the fundamental validity of the approach.
I probably made an explanatory error in bringing up another example to make my argument. I don't want the focus to be on any particular position because I'm not suggesting one couldn't reasonably disagree with these conclusions. What I am suggesting is that some philosophical analysis provides excellent fodder for others to cite and build upon while other types of analysis leave the problem sterile and dry, and even when the better arguments lie with the latter type of explanation sheer weight of publications will drive consensus opinion toward the earlier one.
Of course you can disagree but the problem is that continuing to have things to say (in general or about the disagreement) isn't evidence of correctness. In fact I think in general it is (weak) evidence against a view (more complex, more places for potential confusion to enter). However, a view which doesn't have anything more to say about an issue will be pushed out of the academic community because you can't keep publishing the same paper over and over again.
My hope with the examples is not that you would say, "hey yah that side is certainly correct" (though I wouldn't mind) but that you would say, "yah that position is not represented in the philosophical world proportionally to it's plausibility."
... Note that once you admit that the [surprise examination] paradox can't be stated formally you can only generate the paradox if I accept some other notion as meaingful/well-defined (rationality etc.). Thus it will always be a perfectly good solution to the paradox to simply be really really stingy and refuse to accept these non-formal notions. But can you see the stingy approach continuing to hold the interest of the philosophical community?
Even if the stingy approach has the better of the argument the approach which offers the opportunity for other philosophers to help build up a logic of announcements for true but unbelievable propositions will be the subject of greater attention by the community. It's this bias towards things which give rise to philosophy papers that troubles me.
Three quick thoughts: (1) Neo-Carnapians are not so rare, in my experience. (2) I doubt conceptual stinginess is ever a good response to anything. And (3) it's not entirely clear why the argument is targeted specifically at philosophy rather than academic publishing in general.
But, on the main point, I just don't think it's true that academic structures are biased against knock-down work that conclusively settles (or dissolves) an issue, leaving no further work to do in its wake. If Jones can successfully close an old and troubling issue X in one paragraph, he could get this published in Analysis and the next time anyone opened their mouth about X their colleagues would all say, "Go read Jones, then find a real issue to work on." Or so I imagine.
The reason this doesn't tend to happen is not because of any structural bias, but because it's not that easy to conclusively settle philosophical problems. Anyway, my core response is as follows: we have plenty of unsolved issues to provide us with "fodder", so there's no incentive to keep plugging away at solved ones.