Poincaré–Birkhoff–Witt theorems in higher algebra

Omar Antolin Camarena, Lukas Brantner, Gijs Heuts · 2025

Read Poincaré–Birkhoff–Witt theorems in higher algebra by Omar Antolin Camarena, Lukas Brantner, Gijs Heuts (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-2501-03116.txt.

Introduction

A classic algebra theorem says every element of a certain wrapped-up algebra can be written in one tidy ordered form. Omar Antolín-Camarena, Lukas Brantner, and Gijs Heuts carry that theorem into higher algebra in this 2025 paper. Their setting uses spectral Lie algebras, a modern form where algebra is done with spaces rather than plain sets. The tidy form is the Poincaré-Birkhoff-Witt theorem.

The proof flows from a basic link between two rule systems for combining things, called operads. One rule system turns out to be the other with a Lie action divided out. That link itself comes from a deeper tie among a family of such systems. The team also proves a version for wrapped algebras built over a base. Their route gives a plain way to build the higher wrapped algebras too.

Catalog record

CREATOROmar Antolin Camarena, Lukas Brantner, Gijs Heuts
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-03116

Browse the full library · About the Tower of Babel

Full text

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

Poincaré–Birkhoff–Witt theorems in higher algebra Omar Antolin Camarena, Lukas Brantner, Gijs Heuts Abstract . We extend the classical Poincaré–Birkhoff–Witt theorem to higher algebra by establishing a version that applies to spectral Lie algebras. We deduce this statement from a basic relation between operads in spectra: the commutative operad is the quotient of the associative operad by a right action of the spectral Lie operad. This statement, in turn, is a consequence of a fundamental relation between different 𝔼 n \mathbb{E}_{n} -operads, which we articulate and prove. We deduce a variant of the Poincaré–Birkhoff–Witt theorem for relative enveloping algebras of 𝔼 n \mathbb{E}_{n} -algebras. Our methods also give a simple construction and description of the higher enveloping 𝔼 n \mathbb{E}_{n} -algebras of a spectral Lie algebra. 1. Introduction The classical Poincaré–Birkhoff–Witt theorem states that the universal enveloping algebra of a Lie algebra \mathfrak{g} admits a natural filtration whose associated graded is the symmetric algebra on the underlying vector space of \mathfrak{g} . This statement may be phrased in the language of operads; it is essentially equivalent to the fact that the quotient of the associative operad by the right action of the Lie operad on it is the commutative operad. In this paper, we provide an extension of this statement to the world of higher algebra (see Theorem 1.2 ), replacing the classical associative, Lie, and commutative operads by their lifts to the \infty -category of spectra. Correspondingly, we deduce a Poincaré–Birkhoff–Witt theorem for spectral Lie algebras ( Corollary 1.10 ). In fact, these results are consequences of a more basic relation between different \mathbb{E}_{n} -operads, namely Theorem 1.6 , which has no ‘classical’ analog. In particular we deduce a result in the style of Poincaré–Birkhoff–Witt for \mathbb{E}_{n} -algebras, which we make explicit in Corollary 1.12 . Finally, our methods provide a straightforward construction of the higher enveloping algebra of a spectral Lie algebra \mathfrak{g} . In Theorem 1.8 we show that the \mathbb{E}_{n} -enveloping algebra of \mathfrak{g} may be calculated as the Chevalley–Eilenberg homology of the n -fold loop object of \mathfrak{g} , showing that our construction agrees with the one of Knudsen [ Knu18 ] which relies on factorisation homology. We will now review the results of this paper in more detail. The associative and commutative operads have evident lifts to the world of higher algebra; indeed, they can easily be defined in the \infty -category of spectra (or any symmetric monoidal \infty -category with coproducts). The case of the Lie operad is less straightforward, but Salvatore [ Sal98 ] and Ching [ Chi05 ] have shown that it also admits a lift to the \infty -category of spectra. Indeed, if we define \mathbb{L} to be the Koszul dual K(\mathbf{Com}) of the commutative operad, then the operadic suspension s\mathbb{L} (see Section 2 ) is an operad whose homology is concentrated in degree 0 and reproduces the classical Lie operad in abelian groups. Moreover, this construction recovers the more classical Koszul duality between the commutative and Lie operads in the derived \infty -category \mathrm{Mod}_{R} of any commutative ring R . To explain our first result, recall that operads in the \infty -category \mathrm{Sp} of spectra may be defined as associative algebra objects in the \infty -category of symmetric sequences in spectra (cf. [ Kel05 , Tri ] [ Bra , 4.1.2] ). The monoidal structure is given by the composition product (see Section 2 for a brief review). In particular, given a map of operads \mathcal{O}\to\mathcal{P} , we may interpret \mathcal{P} as an \mathcal{O} -bimodule in symmetric sequences. Notation 1.1 . In what follows, we will denote the associative operad \mathbf{Ass} by \mathbb{E}_{1} and the commutative operad \mathbf{Com} by \mathbb{E}_{\infty} , to make the notation consistent with our later results on \mathbb{E}_{n} -operads. We take all of our operads to be nonunital (without emphasising this in the notation), meaning that their term of 0-ary operations is trivial. We will denote by \mathbf{1} the trivial operad, which has a unit in arity 1 and nothing else. This is also the nonunital \mathbb{E}_{0} -operad. We can now state our first main theorem: Theorem 1.2 . There is a commutative square of operads in the \infty -category \mathrm{Sp} as follows: {\lx@inpgf@ignorespaces s\mathbb{L}} {\lx@inpgf@ignorespaces\mathbf{1}} {\lx@inpgf@ignorespaces\mathbb{E}_{1}} {\lx@inpgf@ignorespaces\mathbb{E}_{\infty}.} Moreover, the induced map \mathbb{E}_{1}\circ_{s\mathbb{L}}\mathbf{1}\to\mathbb{E}_{\infty} is an equivalence of left \mathbb{E}_{1} -modules. Remark 1.3 . In the statement of the theorem, \mathbb{E}_{1}\circ_{s\mathbb{L}}\mathbf{1} denotes the relative composition product of \mathbb{E}_{1} and \mathbf{1} over {s\mathbb{L}} . For an operad \mathcal{O} with a right module \mathcal{M} and a left module \mathcal{N} , such a relative composition product may be computed via the bar construction \mathcal{M}\circ_{\mathcal{O}}\mathcal{N}=\mathrm{colim}_{\mathbf{\Delta}^{\mathrm{op}}}(\mathcal{M}\circ{\mathcal{O}}^{\circ\bullet}\circ\mathcal{N}). Remark 1.4 . The square of Theorem 1.2 is not a pushout square of operads (cf. Proposition 4.3 ). In fact, Theorem 1.2 will follow as a limiting case of a statement about the relation between different \mathbb{E}_{n} -operads (see Theorem 1.6 below). To be precise, if \mathbb{E}_{n} denotes the usual (nonunital) operad of little n -cubes in the \infty -category of spaces, then we will be interested in the operad \Sigma^{\infty}_{+}\mathbb{E}_{n} in spectra. Throughout this paper, we will use the short-hand \mathbb{E}_{n} for this nonunital, stable version of the \mathbb{E}_{n} -operad. Given integers m,n\geq 0 , we will generically use the letter \iota to denote the usual morphism \mathbb{E}_{m}\to\mathbb{E}_{m+n} arising from the standard inclusion \mathbb{R}^{m}\to\mathbb{R}^{m+n} of Euclidean spaces. Theorem 1.2 motivates the following: Definition 1.5 (Composition squares) . A commutative square of operads {\lx@inpgf@ignorespaces\mathcal{O}} {\lx@inpgf@ignorespaces\mathcal{P}} {\lx@inpgf@ignorespaces\mathcal{Q}} {\lx@inpgf@ignorespaces\mathcal{R}} is called a composition square if the induced map \mathcal{Q}\circ_{\mathcal{O}}\mathcal{P}\to\mathcal{R} is an equivalence (of \mathcal{Q} - \mathcal{P} -bimodules). We then establish the following fundamental relation between different \mathbb{E}_{n} -operads, with s denoting operadic suspension (see Section 2 ): Theorem 1.6 . Given integers k,m,n\geq 0 there is a composition square of operads in spectra {\lx@inpgf@ignorespaces\mathbb{E}_{k+m}} {\lx@inpgf@ignorespaces\mathbb{E}_{k+m+n}} {\lx@inpgf@ignorespaces s^{k}\mathbb{E}_{m}} {\lx@inpgf@ignorespaces s^{k}\mathbb{E}_{m+n}.} \scriptstyle{\lx@inpgf@ignorespaces\beta} \scriptstyle{\lx@inpgf@ignorespaces\iota} \scriptstyle{\lx@inpgf@ignorespaces\beta} \scriptstyle{\lx@inpgf@ignorespaces\iota} We will formally describe the morphism \beta\colon\mathbb{E}_{m+k}\to s^{k}\mathbb{E}_{m} in detail in Section 3 . For now, let us outline two ways of thinking about it. Without loss of generality we take k=1 – the general case is obtained by a k -fold composition of such maps. First, \beta can be obtained as the Koszul dual of the morphism \iota\colon\mathbb{E}_{m}\to\mathbb{E}_{m+1} , relying on the fact that the Koszul dual of the operad \mathbb{E}_{m} is the operad s^{-m}\mathbb{E}_{m} , see Remark 3.2 . Secondly, under the ‘degree shifting’ equivalence \mathrm{Alg}_{s\mathbb{E}_{m}}(\mathrm{Sp})\simeq\mathrm{Alg}_{\mathbb{E}_{m}}(\mathrm{Sp}), the morphism \beta induces a left adjoint functor \beta_{!}\colon\mathrm{Alg}_{\mathbb{E}_{m+1}}(\mathrm{Sp})\to\mathrm{Alg}_{\mathbb{E}_{m}}(\mathrm{Sp}). which may be identified with the usual bar construction. This second perspective is the one used to construct and characterise \beta in [ HL24 ] . In addition to Theorem 1.2 , we will deduce further limiting cases of Theorem 1.6 : Theorem 1.7 . The following are composition squares of operads, where \sigma denotes the suspension map (see Section 2 ): {\lx@inpgf@ignorespaces\mathbb{L}} {\lx@inpgf@ignorespaces s^{n}\mathbb{L}} {\lx@inpgf@ignorespaces\mathbb{E}_{n}} {\lx@inpgf@ignorespaces\mathbb{E}_{\infty}} {\lx@inpgf@ignorespaces\mathbf{1}} {\lx@inpgf@ignorespaces\mathbb{E}_{n}} {\lx@inpgf@ignorespaces\mathbf{1}} {\lx@inpgf@ignorespaces s^{n}\mathbb{E}_{\infty}.} \scriptstyle{\lx@inpgf@ignorespaces\sigma^{n}} \scriptstyle{\lx@inpgf@ignorespaces\beta} \scriptstyle{\lx@inpgf@ignorespaces\iota} \scriptstyle{\lx@inpgf@ignorespaces\sigma^{n}} The morphism \beta\colon s^{n}\mathbb{L}\to\mathbb{E}_{n} featuring in Theorem 1.7 can be obtained as an inverse limit over k of the morphisms \beta\colon s^{-k}\mathbb{E}_{n+k}\to\mathbb{E}_{n} . Alternatively, it can be described as (a shift of) the Koszul dual of the inclusion \iota\colon\mathbb{E}_{n}\to\mathbb{E}_{\infty} , see Remark 3.2 . Identifying the \infty -categories \mathrm{Alg}_{s^{n}\mathbb{L}}(\mathrm{Sp}) and \mathrm{Alg}_{\mathbb{L}}(\mathrm{Sp}) via an n -fold degree shift, this morphism induces a

Read the complete text