Of course, we may be most interested in actual-world problems, e.g. interpreting modern physics, addressing salient ethical and political issues, etc. But there doesn't seem any reason why they couldn't in principle be addressed just as well from an empirically neutral position which entertained our actual situation as a merely hypothetical scenario. Indeed, given sufficient imaginative and rational powers, the armchair philosopher (or even the disembodied, floating-in-the-void philosopher) should be capable of achieving a kind of "limited omniscience", knowing everything there is to know about the various possibilities (except for which one happens to be actual -- but never mind that one little fact).
It might be objected that science brings to light new possibilities that would otherwise seem inconceivable -- e.g. space-time relativity. But this is merely to note that experience is a useful imaginative aid; it plays no necessary role in the actual justification of our philosophical beliefs. Einstein's theory is enough; it need not be borne out by the empirical data. His conceptual scheme alone is enough to show how space and time could turn out not to be absolute. (Unless there's really a hidden contradiction in there, in which case ideal rational reflection should suffice to expose the impossibility.)
Am I wrong? (And do you have to conduct an experiment before you can tell?)
P.S. Thanks to Jack for getting me thinking about this topic.