On the relation of special linear algebraic cobordism to Witt groups (Q303922)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On the relation of special linear algebraic cobordism to Witt groups
scientific article

    Statements

    On the relation of special linear algebraic cobordism to Witt groups (English)
    0 references
    0 references
    23 August 2016
    0 references
    In the article under review the author establishes a new motivic variant of the theorem of Conner-Floyd reconstructing \(K\)-theory from cobordism. More specifically, the article shows that higher Witt theory may be reconstructed from special linear bordism with the Hopf map inverted. Recall that classically, for a space \(X\) we write \(MU^*(X)\) for the complex cobordism groups, \(KU^*(X)\) for the \(K\)-groups, \(MSp^*(X)\) for the symplectic cobordism groups, and \(KO^*(X)\) for the orthogonal \(K\)-groups. Complex \(K\)-theory is complex oriented, i.e. satisfies a Thom isomorphism formula. Since complex cobordism is the universal complex oriented cohomology theory, there is a natural transformation \(MU^*(X) \to KU^*(X)\). The theorem of Conner-Floyd for \(KU\) can now be stated as \[ KU^*(X) \approx MGL^*(X) \otimes_{MGL^*} KU^*. \] Here \(MGL^* := MGL_*(\mathrm{pt})\) denotes the coefficient ring, and similarly for \(KU^*\). (Of course \(MGL^*\) is isomorphic to the Lazard ring by a theorem of Quillen, and \(KU^* = \mathbb{Z}[\beta, \beta^{-1}]\) for some \(\beta\) in degree 2.) There is also a variant for orthogonal \(K\)-theory: even though \(KO\) is not complex orientable, it is symplectically orientable (i.e. satisfies a Thom isomorphism theorem for \textit{symplectic} vector bundles) and hence there is a natural transformation \(MSp^*(X) \to KO^*(X)\) (since symplectic cobordism is the universal symplectically orientable cohomology theory). The Conner-Floyd isomorphism in this setting is \[ KO^*(X) \approx MSp^*(X) \otimes_{MSp^*} KO^*. \] There are various motivic analogs of these results. Essentially one may make sense of a notion of an extraordinary cohomology theory for algebraic varieties. These cohomology theories are bigraded, and satisfy many analogs of the classical Eilenberg-Steenrod axioms. For a smooth algebraic variety \(X\) (over a field \(k\)), typical examples of cohomology theories are algebraic cobordism \(MGL^{*,*}(X)\) (as constructed by Voevodsky), higher algebraic \(K\)-theory \(KGL^{*,*}(X)\) (as constructed first by Quillen; here Bott periodicity makes one of the indices redundant), Symplectic cobordism \(MSp^{*,*}(X)\) and special linear cobordism \(MSL^{*,*}(X)\) (as constructed by Panin-Walter), (higher) hermitian \(K\)-theory \(KO^{*,*}(X)\), and higher Witt theory \(KT^{*,*}\) (as constructed by Balmer). There are then the following motivic Conner-Floyd isomorphisms: \[ KGL^{*,*}(X) \approx MGL^{*,*}(X) \otimes_{MGL^{2*,*}} KGL^{2*,*} \] (Panin-Pimenov-Röndigs) \[ KO^{*,*}(X) \approx MSp^{*,*}(X) \otimes_{MSp^{4*,2*}} KO^{4*,2*} \] (Panin-Walter). In motivic homotopy theory there is an algebraic analog \(\eta\) of the Hopf map. This provides a (graded) endomorphism of any cohomology theory. One may show that Balmer's higher witt theory \(KT\) is given by \(KO[1/\eta]\). In the article under review, the author establishes an isomorphism \[ KT^{*,*}(X) \approx MSL[1/\eta]^{*,*}(X) \otimes_{MSL[1/\eta]^{4*,2*}} KT^{4*,2*}. \] The proof of the result combines the Conner-Floyd isomorphism for Hermitian \(K\)-theory as established by Panin-Walter with the careful study of \(\eta\)-inverted cohomology theories of the author.
    0 references
    0 references
    Witt groups
    0 references
    algebraic cobordism
    0 references
    SL-oriented cohomology
    0 references
    Hopf map
    0 references

    Identifiers

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