Homotopy uniqueness of classifying spaces of compact connected Lie groups at primes dividing the order of the Weyl group (Q1327328)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Homotopy uniqueness of classifying spaces of compact connected Lie groups at primes dividing the order of the Weyl group
scientific article

    Statements

    Homotopy uniqueness of classifying spaces of compact connected Lie groups at primes dividing the order of the Weyl group (English)
    0 references
    0 references
    0 references
    14 September 1995
    0 references
    Let \(G\) be a compact connected Lie group, with maximal torus \(T\) and Weyl group \(W= N/T\) (\(N\)= normalizer of \(T\) in \(G\)). Then \(G\) is said to be \(p\)-convenient if \(H^* (BG; \mathbb{Z})\) is \(p\)-torsion free and the \(\text{mod-}p\) cohomology of \(BG\) is given by the \(\text{mod-}p\) invariants of \(W\) in \(H^* BT\). It is well-known that for the primes not dividing the order of \(W\), \(G\) is always \(p\)-convenient, and that this holds for a strictly larger class of primes (e.g. for all odd primes, if \(G\) is simply connected simple and classical). Main result: If \(G\) is either a product of unitary groups or \(p\)- convenient and at the same time a product of simple simply connected groups different from \(\text{SU} (n)\), or having a connected center, then \(H^* (BG; \mathbb{Z}/p)\) as an unstable algebra over the \(\text{mod- }p\) Steenrod algebra determines the homotopy type of the \(p\)-adic completion of \(BG\) among \(p\)-complete spaces. This is a generalization of a similar result obtained by \textit{Dwyer}, \textit{Miller} and \textit{Wilkerson} [to appear] for primes not dividing the order of \(W\). Basically the problem may be first technically reduced to the case of a product of unitary (and special unitary) groups, where explicit computations may be carried on, by passing to a suitable maximal rank subgroup, like in [\textit{S. Jackowski}, \textit{J. McClure} and \textit{B. Oliver}, Ann. Math., II. Ser. 135, 183-270 (1992; Zbl 0758.55004)]. Secondly and even more important, given an algebraic unstable algebra isomorphism between \(H^* (X; \mathbb{Z}/p)\) and \(H^* (BG; \mathbb{Z}/p)\) (where \(X\) is \(p\)-complete), there is a general strategy for building ``homotopically'' the Lie theory for \(X\) to realize it by a map, deviced by Dwyer-Miller-Wilkerson and Jackowski-McClure-Oliver [loc. cit.]. On one hand there are two ways of computing the \(\text{mod-}p\) homology of function spaces of the form map \((BV,X)\) and map \((BV, BG)\) (where \(V\) belongs to a certain class of subgroups of \(G\), generalizing the elementary abelian \(p\)-groups): by Lannes \(T\)-theory, as a functor of the unstable algebra \(H^* (X; \mathbb{Z}/p)\), and by Lie theory, i.e. in terms of centralizers of group representations of \(V\) into \(G\). On the other hand there is a \(\text{mod-}p\) decomposition of \(BG\), as a homotopy colimit taken over the full subcategory of the category of \(G\)-orbits, having as objects \(G/P\), where \(P\) is a so-called \(p\)-stubborn \(p\)-toral subgroup (a version of the above \(V\)) contained in \(N\). The main steps go as follows. The interplay between \(T\)-theory and representation theory provides a maximal torus for \(X\). Then obstruction theory and vanishing results for the (twisted) cohomology of the symmetric groups in low dimensions are used to extend the above map from \(BT\) to \(BN\), hence to construct maps from \(BP\) into \(X\), for any \(P\)- stubborn subgroup \(P\) of \(G\) which is contained in \(N\). It is remarkable that the extension to \(BN\) may always be chosen to give the standard map in \(\text{mod-}p\) cohomology (under the initial identification of \(H^*X\) with \(H^* BG\)) and that this in turn implies that the above diagram of maps from \(BP\) to \(X\) is coherent, up to homotopy. Finally the extension problem from the 1-skeleton of the homotopy colimit to \(BG\) is obstructed by standard derived functors of certain inverse limit functors. The author shows then that the above functors corresponding to \(X\) are isomorphic to those corresponding to \(BG\), where the derived obstruction groups were known to vanish, in order to get his realization result.
    0 references
    compact Lie group
    0 references
    classifying space
    0 references
    Weyl group
    0 references
    \(p\)-convenient
    0 references
    \(p\)- completion
    0 references
    homotopy
    0 references
    unstable algebra
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references