Not signed in (Sign In)

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

  • Sign in using OpenID

Discussion Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

Welcome to nForum
If you want to take part in these discussions either sign in now (if you have an account), apply for one now (if you don't).
    • CommentRowNumber1.
    • CommentAuthorJon Beardsley
    • CommentTimeSep 27th 2017

    Hi all,

    I suspect this is basic stuff, and could be sussed out from something that you all have written (perhaps Urs’ big book), but if I have a manifold M is there an object, I’ll call it X for now, such that (some version of) maps from M to X classify twists of de Rham cohomology on M? In other words, the trivial map should just give us regular de Rham cohomology, and non-trivial maps should give us twists. I’m thinking in analogy with a map YPic(R), where R is a ring spectrum, giving me twists of R-homology of Y.

    Apologies if this is very basic, or perhaps nonsensical.

    • CommentRowNumber2.
    • CommentAuthorUrs
    • CommentTimeSep 27th 2017
    • (edited Sep 27th 2017)

    Which twists are you thinking of, besides those classified by B(/2)? Do you want to include twists by closed 3-forms, which twist not the -graded but the /2-graded de Rham complex?

    • CommentRowNumber3.
    • CommentAuthorJon Beardsley
    • CommentTimeSep 27th 2017
    • (edited Sep 27th 2017)

    Well I don’t really know this stuff very well, I thought you could generally twist the differential of the de Rham complex by a 1-form. I’m not sure where this fits into what you’re saying. I guess it must be the twists classified by B(Z/2) that you mention?

    • CommentRowNumber4.
    • CommentAuthorDavidRoberts
    • CommentTimeSep 28th 2017
    • (edited Sep 28th 2017)

    The Z/2 twists would correspond to a two-fold covering space. Presumably one could take a flat connection on that, i.e. some closed 1-form to make some sort of twisting.

    • CommentRowNumber5.
    • CommentAuthorUrs
    • CommentTimeSep 28th 2017
    • (edited Sep 28th 2017)

    Okay, I see. Right, so as David says, the /2-twisting is for the unoriented case (“pseudo-forms”). Consider the sheaf on CartSp of crossed complexes which to test space U assigns the crossed complex which is the de Rham complex on U (shifted to some degree n) with the groupoid B/2 acting on any k-forms by sending them to their negative. Write Bn/(/2)SmoothGrpd=Sh(CartSp) for the image of this under the Dold-Kan correspondene for crossed complexes. By construction this comes with a canonical forgetful morphism

    Bn/(/2)B/2

    This morphism, regarded as an object in the slice -topos over B/2, is the coefficients for /2-twisted de Rham cohomology in degree n. (We could instead use sheaves of spectra on CartSpace instead, to unify the varying degrees more elegantly, but this is a minor technical point.)

    For the twists by closed differential 1-forms, consider the homomorphism of semifree dgc-algebras

    (dh1=0)(dh1=0dωn=h1ωn)h1h1

    Under Lie integration this again yields a morphism in Sh(CartSp), and, regarded as an object in the slice over its codomain, this is the coefficient for twisted de Rham cohomology with twists by closed 1-forms, I think.