I don't think Umbran is denying the possibility of distinctions, only that "old school" vs. "new school" may not be the right distinction.
Mechanically, there is a certain unifying element to pre-3e games, just as there are mechanical unifying elements to post 2e games. One can take a random page of mechanics from any TSR or WotC book, and I would hazard that over 90% of readers would know which era the rule comes from.
If every game truly is strongly flavoured by its underlying mechanics, then this is a meaningful distinction because the mechanics are meaningfully distinct.
I admit, I consider it an oversimplification myself. Trying to lump white-box D&D, 1st Edition, 2nd Edition, and BECMI into one category and 3E, 4E, and all their offspring into another seems to break down when you look closely--and any attempt to maintain an 'old school'/'new school' dichotomy collapses completely if one dares to look beyond the *D&D family.
Any system of categorization, no matter how nominally useful, breaks down when you look closely. Witness the recent (2008-2009) articles published re: species and classification of living beings via kingdoms. Indeed, even the classification of "living" vs. "non-living" breaks down when examined closely.
That doesn't mean that categorization is useless.....merely that it is not absolute.
IOW, the perils of categorization are no different, and no greater, than those of using language itself. The refusal to categorize is the refusal to define, and leaves language meaningless.
Now, Umbran could be imagining that the statement "Some games are old school, and some games are new school" presupposes no possibility that some games are both, or occupy a liminal space between these categories, or exist outside of these categories. Rather as though the statement "Some animals are birds, and some animals are fish" precluded some animals from being mammals, reptiles, or amphibians.....or precluded the existence of plants, fungi, etc.
If one imagines that all games
must be OS or NS, that these are sharply divided categories, and that there is no grey space between them, then one has, indeed, created a false dichotomy. But the false dichotomy doesn't exist because one recognizes the usefulness of the OS/NS labels. It exists because one imagines those categories to be (or to be meant to be taken as) absolutes.
They are not.
(Note also that the poll heading this thread demonstrates that they are understood to not be absolutes by at least a third of the EN Worlders who responded, as well as by the OP.)
RC