Some examples of DG-Lie formality transfer

Marco Manetti, Gabriele Rossetti · 2026

Read Some examples of DG-Lie formality transfer by Marco Manetti, Gabriele Rossetti (2026) 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-2601-19698.txt.

Introduction

Marco Manetti and Gabriele Rossetti work in homological algebra on a transfer principle. Their objects are DG-Lie algebras, algebraic structures that mix a grading, a differential, and a bracket. Such algebras control deformation problems across mathematics. A key fact is formality: some of these algebras are equivalent to much simpler ones.

A formality transfer theorem says formality passes along maps between algebras. The authors give a convenient reformulation of that theorem. They also slightly generalise it and show applications. The examples demonstrate where the cleaner statement does work the old one could not. Tools like this quietly support large parts of modern algebra and geometry.

Catalog record

CREATORMarco Manetti, Gabriele Rossetti
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-19698

Browse the full library · About the Tower of Babel

Full text

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

Some examples of DG-Lie formality transfer Marco Manetti, Gabriele Rossetti Abstract . We give a convenient reformulation, a slight generalization and some applications of the formality transfer theorem for DG-Lie algebras. 1991 Mathematics Subject Classification 17B56,17B70,16T15 1. Introduction Let f\colon L\to M be a morphism of DG-Lie algebras over a field of characteristic 0 and let \mathcal{P} a property of DG-Lie algebras that is invariant under quasi-isomorphisms. By a forward \mathcal{P} -transfer theorem we mean a statement asserting that if the DG-Lie algebra L has the property \mathcal{P} and the morphism of graded Lie algebras f\colon H^{*}(L)\to H^{*}(M) satisfies certain conditions, then also M has the property \mathcal{P} . Similarly, by a backward \mathcal{P} -transfer theorem we mean a statement asserting that if M has the property \mathcal{P} and the morphism of graded Lie algebras f\colon H^{*}(L)\to H^{*}(M) satisfies certain conditions, then also L has the property \mathcal{P} . For instance, when \mathcal{P} = homotopy abelianity , that is the property of being quasi-isomorphic to a DG-Lie algebra with trivial bracket, it is well known, and largely used in deformation theory, the following \mathcal{P} -transfer theorem. Theorem 1.1 . Let f\colon L\to M be a morphism of DG-Lie algebras: (1) if M is homotopy abelian and f\colon H^{*}(L)\to H^{*}(M) is injective, then also L is homotopy abelian; (2) if L is homotopy abelian and f\colon H^{*}(L)\to H^{*}(M) is surjective, then also M is homotopy abelian. The reader can see [ 21 , Cor. 6.1.3] and [ 1 , 5 , 10 , 13 ] for various proofs of Theorem 1.1 and applications to deformation theory. Other properties usually considered in literature are: • the formality, cf. [ 2 , 20 ] ; • the quadraticity, cf. [ 3 , 6 ] . • a fixed bunch of singularity types of the associated formal Kuranishi family, cf. [ 17 , 22 ] . It is essential to observe that Theorem 1.1 is specific for homotopy abelianity and does not immediately generalize to other properties. For instance, when \mathcal{P} = formality , it is easy to give examples of morphisms f as above with M formal, f\colon H^{*}(L)\to H^{*}(M) injective and L not formal; one of them is described in the Subsection 3.4 below. As already well understood, see for instance [ 7 , 8 , 12 , 16 , 23 , 26 ] , the right context for obstruction to formality of DG-Lie algebras is Chevalley–Eilenberg cohomology theory; the classical (non-graded) definition [ 4 , 9 , 11 ] generalizes in a obvious way to the graded case and give cohomology groups H_{CE}^{p}(L,M) , where L is a graded Lie algebra and M is a graded L -module (Definition 2.2 ). The main goal of this paper is to prove and produce some applications of the following formality transfer theorem. Theorem 1.2 (=Theorem 2.5 ) . Let f\colon L\to M be a morphism of DG-Lie algebras and consider the structure of H^{*}(L) -module on H^{*}(M) induced by the morphism of graded Lie algebras f\colon H^{*}(L)\to H^{*}(M) . (1) If M is formal and f_{*}\colon H^{2}_{CE}(H^{*}(L),H^{*}(L))\to H^{2}_{CE}(H^{*}(L),H^{*}(M)) is injective, then also L is formal. (2) If L is formal and f^{*}\colon H^{2}_{CE}(H^{*}(M),H^{*}(M))\to H^{2}_{CE}(H^{*}(L),H^{*}(M)) is injective, then also M is formal. Notice that the first item of Theorem 1.2 was already proved in [ 19 ] and has been already applied in the proof of Kaledin–Lehn formality conjecture for polystable sheaves on Enriques and bielliptic surfaces [ 2 ] . In this paper we add item 2 and we provide a different and slightly simpler proof of item 1. For reference purposes, we write two obvious corollaries of the first item of Theorem 1.2 . Corollary 1.3 . Let f\colon L\to M be a morphism of DG-Lie algebras. Suppose that M is formal and there exists a morphism of H^{*}(L) -modules s\colon H^{*}(M)\to H^{*}(L) such that sf=Id_{L} . Then also L is formal. Corollary 1.4 . Let f\colon L\to M be a morphism of DG-Lie algebras. Assume that M is formal, that f is injective and that f(L) is a direct summand of M as L -modules. Then also L is formal. The definition of modules over a DG-Lie algebra is completely unsurprising and will be given in the next section for completeness. To the best of our knowledge, almost all known interesting applications of the above results rely on Corollary 1.4 . In the last section we illustrate some further applications of formality transfer: in particular, we prove Theorem 3.1 about the equivalence of Lie and associative formality for universal enveloping algebras, and Theorem 3.3 , about the persistence of Kodaira–Spencer formality under free action of finite groups. 2. Proof of the main theorem Throughout the whole paper, every differential graded (DG) vector space is intended to be defined over a field \mathbb{K}\, of characteristic 0, \mathbb{Z} -graded and with the differential of degree +1 . Given a homogeneous vector x in a DG-vector space, we shall indicate its degree by \overline{x} . We assume that the reader is also familiar with the notion of (differential) graded Lie algebra. For the benefit of the non-expert reader it is worth recalling that a DG-Lie algebra is formal if it is quasi-isomorphic to its cohomology algebra. Proposition 2.1 . For a DG-Lie algebra L over a field of characteristic 0, the following conditions are equivalent: (1) L is formal; (2) there exists a span L\leftarrow M\rightarrow H^{*}(L) of surjective quasi-isomorphisms of DG-Lie algebras; (3) there exists an L_{\infty} -morphism H^{*}(L)\rightsquigarrow L lifting the identity in cohomology. Proof. This is a well known result; a proof can be found, for instance, in [ 21 ] . ∎ Let (L,d,[-,-]) be a DG-Lie algebra. By definition, an L -module is a DG-vector space M equipped with a morphism of DG-vector spaces M\otimes L\xrightarrow{\;*\;}M,\qquad m\otimes x\mapsto m*x, such that m*[x,y]=(m*x)*y-(-1)^{\overline{x}\,\overline{y}}(m*y)*x. For instance, if f\colon L\rightarrow M is a morphism of DG-Lie algebras, then M is an L -module via the adjoint representation m*x=[m,f(x)] . Given a DG-Lie algebra L and an L -module M , we recall from [ 19 ] the construction of the associated Chevalley–Eilenberg double complex (CE(L,M)^{*,*},\delta,\overline{\delta}) . We have: (1) CE(L,M)^{p,q}=\operatorname{Hom}^{q}_{\mathbb{K}}\,(L^{\wedge p},M) ; in particular CE(L,M)^{p,q}=0 for p<0 and CE(L,M)^{0,q}=M^{q} . (2) \overline{\delta}\colon CE(L,M)^{p,q}\rightarrow CE(L,M)^{p,q+1} is the natural differential of the complex \operatorname{Hom}^{*}_{\mathbb{K}}\,(L^{\wedge p},M) , that is (\overline{\delta}\phi)(x_{1},\dots,x_{p})=d(\phi(x_{1},\dots,x_{p}))-\sum_{i=1}^{p}(-1)^{\overline{\phi}+\overline{x_{1}}+\dots+\overline{x_{i-1}}}\phi(x_{1},\dots,dx_{i},\dots,x_{p}) (3) \delta is the natural extension of the classical Chevalley–Eilenberg differential (with an inessential different choice of signs motivated by the décalage isomorphisms, cf. [ 19 , Section 7] ). More precisely: (a) for m\in CE(L,M)^{0,q}=M^{q} we have (\delta m)(x)=(-1)^{q}m*x ; (b) for \phi\in CE(L,M)^{1,q}=\operatorname{Hom}_{\mathbb{K}\,}^{q}(L,M) we have (\delta\phi)(x,y)=(-1)^{q+1}\left(\phi(x)*y-(-1)^{\overline{x}\;\overline{y}}\phi(y)*x-\phi([x,y])\right); (c) for p\geq 2 and \phi\in CE(L,M)^{p-1,q}=\operatorname{Hom}_{\mathbb{K}\,}^{q}(L^{\wedge p-1},M) we have: \begin{split}(\delta\phi)(x_{1},\ldots,x_{p})&=(-1)^{q+p-1}\sum_{i}\chi_{i}\,\phi(x_{1},\ldots,\widehat{x_{i}},\ldots,x_{p})*x_{i}\\ &\quad+(-1)^{q+p}\sum_{i<j}\chi_{i,j}\,\phi(x_{1},\ldots,\widehat{x_{i}},\ldots,\widehat{x_{j}},\ldots,x_{p},[x_{i},x_{j}]),\end{split} where the \chi_{i},\chi_{i,j}\in\{\pm 1\} are the antisymmetric Koszul signs, determined by the equalities in L^{\wedge p} : x_{1}\wedge\cdots\wedge x_{p}=\chi_{i}\;x_{1}\wedge\cdots\widehat{x_{i}}\cdots\wedge x_{p}\wedge x_{i}=\chi_{i,j}\;x_{1}\wedge\cdots\widehat{x_{i}}\cdots\widehat{x_{j}}\cdots\wedge x_{p}\wedge x_{i}\wedge x_{j}\;. Definition 2.2 . Given L,M as above, the Chevalley–Eilenberg cohomology of L with coefficients in M is the cohomology of the total complex \operatorname{Tot}^{\Pi}(CE(L,M)) (see [ 28 , 1.2.6] ): H^{*}_{CE}(L,M)=H^{*}(\operatorname{Tot}^{\Pi}(CE(L,M))),\quad\operatorname{Tot}^{\Pi}(CE(L,M))^{n}=\prod_{p+q=n}CE(L,M)^{p,q}. It is worth to point out that, when both L and M have trivial differential, then \operatorname{Tot}^{\Pi}(CE(L,M)) is the direct product of the subcomplexes (CE(L,M)^{*,q},\delta) . We denote by (E(L,M)_{r}^{p,q},d_{r})

Read the complete text