From what I can piece together, from the posts here and a variety of other sources, the potential audio advantages/disadvantages of 48 vs. 96 kHz sampling are mostly in the realm of theory or are essentially miniscule differences around the edges [EDIT: assuming a "decent" ADC]. Particularly for concert recording, I am not sure if you'll ever hear a real, audible difference between the two. If that is true, then it comes down to the more practical considerations people have mentioned and, in the end, personal preference...For what it's worth, the AES recommends 48 kHz for most applications.
With respect to an earlier question about hearing above 24 kHz, I have a more solid opinion on that. I reviewed the literature on the range of human hearing and found that a large body of research, stretching back for decades, supports 20 kHz as the upper limit, and that's in young, healthy individuals (the upper limit decreases, often substantially, with age). There are a few papers that suggest that some people can hear pure tones above 20 Khz, up to a maximum of 24 kHz. Those papers described studies that were conducted with very small samples under ideal laboratory conditions (and with high SPL signals), so their conclusions may not be entirely reliable and are decidedly not relevant for "real world" conditions. So a more liberal interpretation of the data would peg the upper bound at 24 kHz, while a more conservative interpretation would favor 20 kHz...