Goodwillie calculus via adjunction and LS cocategory (Q505342): Difference between revisions

From MaRDI portal
Created claim: MaRDI profile type (P1460): MaRDI publication profile (Q5976449), #quickstatements; #temporary_batch_1710362833254
Importer (talk | contribs)
Changed an Item
Property / arXiv ID
 
Property / arXiv ID: 1209.2384 / rank
 
Normal rank

Revision as of 14:34, 18 April 2024

scientific article
Language Label Description Also known as
English
Goodwillie calculus via adjunction and LS cocategory
scientific article

    Statements

    Goodwillie calculus via adjunction and LS cocategory (English)
    0 references
    0 references
    20 January 2017
    0 references
    This paper brings together two concepts, homotopy calculus of functors and Lusternik-Schnirelmann (LS) (co)category. Homotopy calculus approximates a homotopy functor \(F\) by a tower of functors \[ F\longmapsto(\ldots\to P_nF\to P_{n-1}\to\ldots \to P_1F). \] The LS category of a space \(X\) is roughly the least amount of contractible open sets required to cover \(X\). According to the author, this paper arose from the observation that an alternative definition of LS cocategory (an appropriately dualized notion of LS category) can be stated in terms of homotopy calculus of functors. In more detail, in the construction of the functors \(P_nF\), there is an intermediate construction of functors \(T_nF\). Iterating \(T_nF\) and taking the homotopy colimit of the natural transformations between these iterates gives \(P_nF\). The main result of the paper is that there exists a pair of adjoint functors \((R_n, L_n)\) such that \[ T_nF\simeq R_nFL_n. \] The equivalence is an equality for functors that are \textit{strongly reduced}, namely \(F(*)=*\). In particular, for the identity functor \(\mathbb I\), \(T_n\mathbb I=R_nL_n\) is a monad. The functors \(L_n\) and \(R_n\) go between the category of spaces and the category of cubical diagrams missing the last space that are coreduced (all spaces indexed on non-singletons are contractible). They are given essentially by making a punctured cubical diagram into a reduced one and by taking the homotopy limit, respectively. A reformulation of the main result is that \(X\) has symmetric LS cocategory \(\leq n\) iff \(X\) is a homotopy retract of \(T_n\mathbb I(X)\). This leads to some nice consequences. For example, the author deduces that \(T_nF\) and \(P_nF\) take values in spaces of LS cocategory \(\leq n\) and that Whitehead products of lengths \(n+1\) vanish in spaces \(T_nF(X)\) and \(P_nF(X)\) for all \(X\) and for \(n\geq 1\). There are also the corresponding dual functors \(R^n\) and \(L^n\) and the author uses these to develop a dual homotopy calculus, thereby extending work done in [\textit{R. McCarthy}, Contemp. Math. 271, 183--215 (2001; Zbl 0996.19005)]. She defines \(T^n\) as \(L^nFR^n\), and this definition allows for the iteration that produces \(P^nF\), exactly dualizing the original homotopy calculus setup. This dual iteration is the piece that has been missing from the theory. The author also establishes the natural dual of her main result which says that \(X\) has symmetric LS category \(\leq n\) iff the map \(T_n\mathbb I(X)\to X\) has a section up to homotopy. The author elucidates the relationship between functor calculus and LS cocategory very nicely and provides ample background on cubical diagrams, functor calculus, and LS category to make for a largely self-contained exposition. She also states some interesting conjectures, specifically regarding the nilpotence of \(T_nF\), and further potential avenues of investigation.
    0 references
    0 references
    LS category
    0 references
    LS cocategory
    0 references
    Goodwillie calculus
    0 references
    homotopy limit
    0 references
    nilpotence
    0 references

    Identifiers

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