Group schemes and their Lie algebras over a symmetric tensor category

Dave Benson, Julia Pevtsova · 2025

Read Group schemes and their Lie algebras over a symmetric tensor category by Dave Benson, Julia Pevtsova (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-02031.txt.

Introduction

Dave Benson and Julia Pevtsova build group theory over abstract tensor settings in 2025. Group schemes are geometric objects that act like groups in algebra. The authors ask what their tangent space at the identity looks like in a general tensor category. That space turns out to carry a restricted Lie algebra structure.

It can also be read as special distributions on the group or as invariant derivations on its ring. The second half tests the theory on one small category in characteristic two. Abstract settings like this unify group behavior across fields where classic tools fail.

Catalog record

CREATORDave Benson, Julia Pevtsova
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE100 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-02031

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Group schemes and their Lie algebras over a symmetric tensor category Dave Benson, Julia Pevtsova Abstract . We investigate the theory of affine group schemes over a symmetric tensor category, with particular attention to the tangent space at the identity. We show that this carries the structure of a restricted Lie algebra, and can be viewed as the degree one distributions on the group scheme, or as the right invariant derivations on the coordinate ring. In the second half of the paper, we illustrate the theory in the particular case of the symmetric tensor category 𝖵𝖾𝗋 4 + \mathsf{Ver}_{4}^{+} in characteristic two. 2020 Mathematics Subject Classification 18D10 (primary), 14L15, 17B45, 18M05, 18M20 (secondary) 1. Introduction Let \mathcal{C} be a rigid k -linear symmetric tensor category, and write {\vec{\hskip 0.56905pt\mathcal{C}}} for its ind-completion. For example, if \mathcal{C} is the category {\nu\hskip-1.13809pt\mathit{ec}}(k) of finite dimensional vector spaces then {\vec{\hskip 0.56905pt\mathcal{C}}} is the category {\mathcal{V}\hskip-1.13809pt\mathit{ec}}(k) of all vector spaces. The theory of group schemes and the theory of supergroup schemes are examples of a more general theory, namely group schemes defined over {\vec{\hskip 0.56905pt\mathcal{C}}} . In the case of group schemes we have \mathcal{C}={\nu\hskip-1.13809pt\mathit{ec}}(k) , while in the case of supergroup schemes we have \mathcal{C}={\mathit{s}\nu\hskip-1.13809pt\mathit{ec}}(k) , the category of finite dimensional super vector spaces, with {\vec{\hskip 0.56905pt\mathcal{C}}}={\mathcal{SV}\hskip-1.13809pt\mathit{ec}}(k) all super vector spaces. Our goal here is to write a version of the theory for general \mathcal{C} . We begin the paper with some introductory material on commutative algebras and group schemes over finite tensor categories, in the first few sections. The reader can also consult the references [ 10 , 11 , 32 ] . Much of the theory mimics what happens in the case where {\vec{\hskip 0.56905pt\mathcal{C}}}={\mathcal{V}\hskip-1.13809pt\mathit{ec}}(k) , with suitably categorical versions of the definitions. But there are some subtleties. One of them is the definition of tangent space, which we give in Section 6. For this purpose, we introduce a categorical notion \mathcal{E}_{\mathcal{C}} of the ring of dual numbers. We take the right adjoint of the tensor product functor, and apply it to the tensor identity to produce a dualising object E_{\mathcal{C}} , and then \mathcal{E}_{\mathcal{C}} is defined to be \mathds{1}\oplus E_{\mathcal{C}} , with augmentation the projection \eta\colon\mathcal{E}_{\mathcal{C}}\to\mathds{1} , and kernel E_{\mathcal{C}} squaring to zero. This is a commutative algebra over \mathcal{C}\boxtimes\mathcal{C} , Deligne’s tensor product of \mathcal{C} with itself. The tangent space is the first part of the algebra \mathsf{Dist}(G) of distributions at the identity on a group scheme G over {\vec{\hskip 0.56905pt\mathcal{C}}} . We give three different interpretations of the Lie algebra \mathsf{Lie}(G) of G . The first is the tangent space at the identity, the second is the degree one part of the algebra of distributions, and the third is the right invariant derivations of the coordinate ring \mathcal{O}(G) . Effectively, we are translating Theorem 12.2 of Waterhouse [ 33 ] , to the context of affine group schemes over a symmetric tensor category. Theorem 1.1 . Let G be an affine group scheme over {\vec{\hskip 0.56905pt\mathcal{C}}} . Then there are canonical isomorphisms between the following. (i) The tangent space \mathsf{Lie}(G) to G at the identity, (ii) The degree one distributions on G . (iii) The right invariant derivations from \mathcal{O}(G) to \mathds{1} . The definition of a Lie algebra over {\vec{\hskip 0.56905pt\mathcal{C}}} needs some care. There are three distinct notions which are, from weakest to strongest, operadic Lie algebra , Lie algebra , and restricted Lie algebra . The notion of operadic Lie algebra is easy to define using the Lie operad \mathcal{L}\mathit{ie} . We use Etingof’s definition [ 16 ] of a Lie algebra, where non-linear identities are added to the operadic definition. These identities are not satisfied by the free operadic Lie algebra on an object in \mathsf{Ver}_{p} , for example, so Etingof’s notion is strictly stronger than the operadic one. We use Fresse’s definition [ 18 ] of a restricted Lie algebra as a \Gamma\mathcal{L}\mathit{ie} -algebra, namely a divided power algebra over \mathcal{L}\mathit{ie} . Fresse’s definition of restricted Lie algebra implies but is not implied by Etingof’s definition of Lie algebra, so all these notions are distinct. We show that \mathsf{Lie}(G) satisfies the strongest of these—it is a restricted Lie algebra in {\vec{\hskip 0.56905pt\mathcal{C}}} . The equivalence of parts (i) and (ii) of Theorem 1.1 is proved in Proposition 8.6 , and the equivalence with (iii) is proved in Section 11 . The fact that \mathsf{Lie}(G) is a restricted Lie algebra is proved in Section 12 . In the second part of the paper, we illustrate the theory with the particular case of a category {\mathsf{Ver}_{4}^{+}} in characteristic two. This category first appeared in work on topological K-theory in the nineteen sixties [ 1 , 2 , 3 , 4 , 5 ] , later in other K-theories including Morava K-theory [ 23 , 26 , 27 , 28 , 29 , 30 , 34 , 35 ] and was further investigated more recently by Kaufer [ 25 ] , Venkatesh [ 31 ] and Hu [ 20 , 21 ] . This is the second in an infinite sequence of incompressible symmetric tensor categories in characteristic two investigated in [ 6 , 7 ] . The corresponding categories in odd characteristic appear in [ 8 , 9 ] . The second part of the paper reworks the first part more concretely in the special case of {\mathsf{Ver}_{4}^{+}} , so that the reader can see exactly what the theory is saying in this particular example. This means that some details are reproved there with elements rather than functors. Doing these calculations helped the authors see how to formulate Part 1 of this paper. Acknowledgements. The first named author thanks the University of Washington, and both authors thank City, University of London, for their hospitality during the preparation of this paper. This research was partly supported by the NSF Awards DMS-1901854 and DMS-2200832. The authors would also like to thank Pavel Etingof for useful feedback. Part I General theory 2. Symmetric tensor categories For preliminaries on tensor categories, we refer to Deligne [ 13 ] and the book of Etingof et al. [ 17 ] . Let k be an algebraically closed field, and let {\vec{\hskip 0.56905pt\mathcal{C}}} be an abelian category with set indexed direct sums. Let \mathcal{C} be the full subcategory of compact objects in {\vec{\hskip 0.56905pt\mathcal{C}}} , namely the objects X such that the natural map \bigoplus_{i\in I}\mathsf{Hom}(X,Y_{i})\to\mathsf{Hom}(X,\bigoplus_{i\in I}Y_{i}) is an isomorphism, whenever the Y_{i} are objects in {\vec{\hskip 0.56905pt\mathcal{C}}} indexed by a set I . We suppose that \mathcal{C} generates {\vec{\hskip 0.56905pt\mathcal{C}}} . We suppose that \mathcal{C} is a finite symmetric tensor category. Thus \mathcal{C} is an abelian category that comes with a tensor product, which is a functor of two variables, exact in each variable, sending V and V^{\prime} to V\otimes V^{\prime} , together with a symmetric braiding s\colon V\otimes V^{\prime}\to V^{\prime}\otimes V and a simple unit object \mathds{1} , together with maps l_{V}\colon\mathds{1}\otimes V\to V and r_{V}\colon V\otimes\mathds{1}\to V such that the usual commutativity, associativity and unit diagrams hold up to coherent natural isomorphism. The tensor product and symmetric braiding extend naturally to {\vec{\hskip 0.56905pt\mathcal{C}}} . The following diagram commutes for all objects V in {\vec{\hskip 0.56905pt\mathcal{C}}} , see Proposition XIII.1.2

Read the complete text