Disjunctive Metaphysics
I'm inclined to think that we don't know very well what the fundamental structure of the world is. (I am, I sometimes think, like a flea on a dog's back, watching a hair grow and saying "ah, that's how the world works!") Supposing evolutionary theory is true, there would seem to be little pressure to get fundamental metaphysics right. If mainstream cosmology is correct, we've only seen a small portion of time and space in one of possibly many universes. And thoughtful people have disagreed radically about the nature of the world, even given the similar evidence base of our one small vantage point. Some say that the world is wholly material, some that the world has material and immaterial aspects, others that only minds and their contents exist and there are no material things at all. Plato thought there was a world of perfect forms behind the imperfect world we see. Panpsychists think that ordinary objects have mental lives. David Lewis believed in the real existence of alternative possible worlds. Theists typically believe in an afterlife, which must either transpire somewhere or somehow be realized without transpiring anywhere.
Historically, philosophers have exhibited high confidence in various metaphysical views of this sort -- without, as far as I see, much epistemic warrant. Probably, people who lack unwarranted metaphysical self-confidence generally drop out of philosophy -- or at least avoid metaphysics or occupy the small niche of the skeptic. So metaphysics, at least metaphysics of the grand sweeping sort, tends to be written by people who lack a fully realistic assessment of their (and all of our) epistemic shortcomings in this domain.
Furthermore, it's almost inevitable that a number of highly unintuitive things are true about the fundamental structure of the world. Materialism, for example, is fairly unintuitive (even if contemporary philosophical materialists find themselves able eventually to jettison their anti-materialist intuitions). But so is idealism. And so are the various forms of dualism, when their consequences are explicitly displayed (David Chalmers' 1996 book brings this fact out nicely, I think). And intuitions about the infinite, about space and time, about possibility and the openness of the future, etc., often conflict, both within and between people. It seems unlikely, then, that we have a fully coherent set of metaphysical intuitions. If so, to the extent philosophical methodology relies on metaphysical intuitions, it builds from an inconsistent groundwork, and strange results follow.
BUT: Even if my pessimism is justified, I don't counsel abandoning metaphysical speculation. Rather, I advise that we give it a different role in our philosophical cognition. Rather than trying to settle on the one correct metaphysical picture, we should develop arms of a disjunction. We can improve our sense of what the metaphysical possibilities are. We can sophisticate our understanding of those possibilities and discover new possibilities we hadn't considered. We can reflect on the consequences of the different possibilities. Maybe -- although we should resist pushing too hard on this -- we can cultivate a sense of which possibilities seem relatively more likely and which seem relatively more far-fetched. Disjunctive metaphysics would mostly be an opening of the mind to the numerous (and perhaps at first wacky-seeming) possibilities, rather than a narrowing of the mind upon one.
[Update, June 3: In that last paragraph I mean epistemic possibilities, of course, about what metaphysics is correct, not different ways the world might possibly be given the one right metaphysics.]