Hochschild (co)homology and cyclic homology via a graded Euler characteristic with applications to higher preprojective algebras

Jon Wallem Anundsen and Mads Hustad Sandøy · 2026

Read Hochschild (co)homology and cyclic homology via a graded Euler characteristic with applications to higher preprojective algebras by Jon Wallem Anundsen and Mads Hustad Sandøy (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-2606-26255.txt.

Introduction

Jon Wallem Anundsen and Mads Hustad Sandøy extend a shortcut for computing Hochschild and cyclic homology, invariants that describe an algebra's structure but are notoriously hard to calculate. Etingof and Eu had shown the job becomes easy for preprojective algebras of ADE type using a graded Euler characteristic.

The generalization covers higher preprojective algebras, including those built as tensor products of type-A hereditary algebras. For these, knowing just the algebra's center and its zeroth homology determines the graded structure of all the rest, in every degree. The proof again runs through the graded Cartan matrix, turning a hard computation into bookkeeping.

Catalog record

CREATORJon Wallem Anundsen and Mads Hustad Sandøy
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE142 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.26255

Browse the full library · About the Tower of Babel

Full text

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

Hochschild (co)homology and cyclic homology via a graded Euler characteristic with applications to higher preprojective algebras Jon Wallem Anundsen and Mads Hustad Sandøy Abstract . Computing the structure of the Hochschild (co)homology and the cyclic homology of an algebra can be hard work, but Etingof and Eu [ 20 ] showed that it can be done surprisingly easily for preprojective algebras of ADE Dynkin type, at least if one only wants to know the graded vector space structure of each Hochschild cohomology group. Their method is based on exploiting strong structural features of such a preprojective algebra via a graded Euler characteristic that can computed using the algebra’s graded Cartan matrix. In this paper, we present a generalization of the method used by Etingof and Eu to higher preprojective algebras. We also apply our generalization to the higher preprojective algebras of the 2 2 -representation finite algebras that arise as tensor products of representation finite hereditary algebras of type 𝔸 \mathbb{A} . For this, it turns out to be enough to know the graded vector space structure of the center and the zeroth Hochschild homology to be able to deduce the graded vector space structure of the Hochschild (co)homology and the cyclic homology in all other degrees. 2020 Mathematics Subject Classification 16E40, 16S37, 16W50, 16G20, 16G60, 18G80 1. Introduction Hochschild (co)homology and cyclic homology are all notorious for how difficult they can be to compute for non-trivial examples. It is thus quite striking when a method allows all three of these forms of (co)homology to be computed surprisingly easily for otherwise seemingly complicated classes of algebras, as is the case in [ 20 ] and in the works [ 24 , 48 ] : In particular, under the assumption that the base field has characteristic zero, the Hochschild (co)homology and cyclic homology was computed for preprojective algebras of ADE Dynkin algebras in [ 20 ] . Using essentially the same approach, [ 24 ] achieved similar results for a “higher” version of preprojective algebras of ADE Dynkin algebras in all types except type \mathbb{A} , which was instead later done by [ 48 ] up to assuming the veracity of two conjectures. The simplifications their method yields can be quite staggering. Indeed, in [ 20 ] the method allowed one to infer the Hochschild (co)homology and cyclic homology in all degrees from knowing the zeroth Hochschild homology. In other words, it completely removed the need to do explicit computations using complexes! Of course, for each of these classes of algebras, the various forms of (co)homology are periodic, and so even with a naive approach it would suffice to compute them in finitely many degrees. However, the simplifications afforded by the method cuts down the number of degrees one needs to compute dramatically: e.g. for [ 20 ] it goes from five to one. The goal of the present paper is to generalize the method or approach used in the aforementioned papers. However, to explain how and why we do this, we first give a fairly “handwavy” review of the method itself: Let \Lambda=\oplus_{i\geq 0}\Lambda_{i} be a positively graded finite dimensional K -algebra for a field K of characteristic zero, and let \overline{\HH}_{i}(\Lambda) and \overline{\HC}_{i}(\Lambda) be its i th reduced Hochschild and cyclic homology, respectively. Note that \overline{\HH}_{i}(\Lambda)=\HH_{i}(\Lambda) holds for i>0 . In brief, the method begins by using that for any n there is a short exact sequence 0\to\overline{\HC}_{n-1}(\Lambda)\to\overline{\HH}_{n}(\Lambda)\to\overline{\HC}_{n}(\Lambda)\to 0 by [ 44 , Theorem 4.1.13] , splicing these together to form the Connes long exact sequence 0\to\overline{\HH}_{0}(\Lambda)\to\overline{\HH}_{1}(\Lambda)\to\overline{\HH}_{2}(\Lambda)\to\overline{\HH}_{3}(\Lambda)\to\cdots, and then attempting to leverage the graded Euler characteristic \sum_{i=1}^{\infty}a_{i}x^{i}=\sum_{i=1}^{\infty}\sum_{j=0}^{\infty}(-1)^{i}\dim\overline{\HC}_{i}(\Lambda)_{j}x^{j}. In the papers cited above, this is done by using that (A) \Lambda is positively graded such that \Lambda=\oplus_{i\geq 0}\Lambda_{i} with \Lambda_{0} semisimple. Hence, if C_{\Lambda}(x) is the graded Cartan matrix of \Lambda , one has by [ 21 ] that \prod_{i=1}^{\infty}(1-x^{i})^{-a_{i}}=\prod_{s=1}^{\infty}\det C_{\Lambda}(x^{s}). However, this by itself is not sufficient, and so further strong properties of \Lambda must be invoked. In the case of [ 20 , 24 , 48 ] , these properties include that the algebras in question are almost Koszul in the sense of [ 7 ] : Roughly speaking, a finite dimensional algebra is almost Koszul if it is as close as possible to actually being Koszul (as in [ 51 ] , but see also [ 6 ] ) while also satisfying that its simple modules are periodic. Indeed, let \Lambda=\oplus_{i=0}^{p}\Lambda_{i} be almost Koszul and recall that it then has a “dual” algebra \Lambda^{!}:=\oplus_{i=0}^{q}\Ext^{i}_{\gr\Lambda}(\Lambda_{0},\Lambda_{0}\langle i\rangle). Moreover, recall that if p,q\geq 2 and \Lambda is connected and basic, then both \Lambda and \Lambda^{!} are selfinjective and even Frobenius. In this case, (A) is actually part of the assumption that \Lambda is almost Koszul, but we also obtain the following: (B) \Lambda has a minimal graded projective resolution P^{\bullet}_{\Lambda^{e}}\twoheadrightarrow\Lambda of \Lambda considered as a bimodule that is “piece-wise linear” by [ 58 ] (and “twisted periodic” by [ 29 ] ); and (C) the graded Cartan matrices of \Lambda and \Lambda^{!} satisfy C_{\Lambda}(x)C_{\Lambda^{!}}(-x)^{T}=I+(-1)^{q}x^{p+q}P^{p+q} for a permutation matrix P corresponding to the product of the Nakayama permutation of \Lambda and the inverse of the Nakayama permutation of \Lambda^{!} ; see Proposition 3.14 of [ 7 ] . Note that (C) allows one to reduce to whichever one of C_{\Lambda}(x) and \C_{\Lambda^{!}}(x) is easier to work with, and it often results in significant simplifications. Let \mu be the Nakayama automorphism of \Lambda and recall that {}_{1}\Lambda_{\mu^{-1}} is the bimodule obtained by twisting the right \Lambda -action of \Lambda by \mu^{-1} , i.e. if \lambda,\lambda^{\prime\prime}\in\Lambda and \lambda^{\prime}\in{}_{1}\Lambda_{\mu^{-1}} , then \lambda\cdot\lambda^{\prime}\cdot\lambda^{\prime\prime}=\lambda\lambda^{\prime}\mu(\lambda^{\prime\prime}) . With this in mind, we note that in all of [ 20 , 24 ] and [ 48 ] , the algebras satisfy that (D1) \Omega^{m+1}_{\Lambda^{e}}\Lambda and {}_{1}\Lambda_{\mu^{-1}} are isomorphic as bimodules for some integer m\geq 2 ; and (D2) \mu has finite order. In the following example, we try to show how all of this comes together by sketching what is done in [ 20 ] . Example 1.1 . Let \Lambda be an ADE Dynkin preprojective algebra; see [ 26 ] , [ 13 ] , but also note that e.g. [ 54 ] is a readable source discussing various equivalent definitions of preprojective algebras. Then, by [ 55 ] , we have that \Omega^{3}_{\Lambda^{e}}\Lambda\cong\Lambda_{\mu^{-1}} so that (D1) holds; and (D2) holds since it is well known that \mu^{2}=1 . This forces a certain self-duality of a portion of the Connes long exact sequence for \Lambda via the formula \HH_{i}(\Lambda)\cong D\HH_{5-i}(\Lambda) that follows by also using that \Lambda^{e}:=\Lambda^{\op}\otimes\Lambda is Frobenius since \Lambda is Frobenius. In fact, more can be said: Note that it is well known that \Lambda^{!} is in this case isomorphic to the trivial extension \Delta(A^{!}) for A=KQ and Q an orientation of an ADE Dynkin quiver of the same type as \Lambda ; see e.g. [ 7 ] . As a consequence, we have that \Lambda^{!}=\oplus_{i=0}^{2}\Ext^{i}_{\gr\Lambda}(\Lambda_{0},\Lambda_{0}\langle i\rangle), i.e. \Lambda^{!} is of highest degree q=2 . We now know that C_{\Lambda^{!}}(x)=I+x(M+M^{T})+x^{2}I for M the adjacency matrix of A^{!} . We can thus in principle use (A) and (C) in combination to compute the graded Euler characteristic of \overline{\HC}_{*}(\Lambda) . For instance, one can find the values of

Read the complete text