??modified version of "tangent space at a point" for a stratified space?? or something?? something about normal bundles... ??as seems to be going on with flag varieties??... ??any relationship to all that other stuff people do with this sort of thing... perverse sheaves and "intersection homology" and so forth ... ??...

??map from positive roots to weyl group?? seems to work in a_n case... can understand tangent spaces of schubert varieties that way ... ???....

hmm, so the obvious map (a quandle homomorphism or something?) from the positive roots to the weyl group (in fact slightly different from the map that i was thinking of at first, in the a_n case) is (or at least seems to be from casual inspection of some examples) "order-preserving" but not "fully" so... which is perhaps "good" in some sense, in that the order on the roots contains more delicate (dynkin as opposed to coxeter...) information than the order on the weyl group does...