Abstract representation theory of Dynkin quivers of type \(A\) (Q261208)

The paper under review is part of the authors' program to study representation theory of quivers from the point of view of stable homotopy theory, which includes the prequels [\textit{M. Groth} and \textit{J. Šťovíček}, ``Tilting theory via stable homotopy theory'', Preprint, \url{arXiv:1401.6451}, to appear in J. Reine Angew. Math.] and [\textit{M. Groth} and \textit{J. Šťovíček}, J. Pure Appl. Algebra 220, No. 6, 2324--2363 (2016; Zbl 1337.55024)]. Let us give a brief overview of the program before focusing on the paper. Representations of a (finite) quiver \(Q\) in \(k\)-vector spaces correspond to modules over the path algebra \(kQ\). Important results of Gabriel had been revisited by Bernšteĭn, Gel'fand, and Ponomarev using reflection functors [\textit{I. N. Bernstein}, \textit{I.M. Gel'fand}, and \textit{V. A. Ponomarev}, Usp. Mat. Nauk 28, No. 2(170), 19--33 (1973; Zbl 0269.08001)]. Happel showed that for quivers without oriented cycles, reflections between \(Q\) and \(Q'\) induce a triangulated equivalence \(D(kQ) \cong D(kQ')\) between derived categories [\textit{D. Happel}, Comment. Math. Helv. 62, 339--389 (1987; Zbl 0626.16008)]. One would like a richer invariant than the derived category \(D(kG)\). The equivalence \(\mathrm{Mod}(kQ) \cong \mathrm{Mod}(k)^Q\) induces an equivalence between categories of chain complexes \(\mathrm{Ch}(kQ) \cong \mathrm{Ch}(k)^{Q}\). Instead of working with the homotopy category \[ D(kQ) = \mathrm{Ho} \left( \mathrm{Ch}(kQ) \right) \cong \mathrm{Ho} \left( \mathrm{Ch}(k)^{Q} \right), \] one could work with the homotopy theory \(\mathrm{Ch}(k)^{Q}\), thought of as representations of \(Q\) in the homotopy theory \(\mathrm{Ch}(k)\). The authors' program addresses the following question: Which classical constructions and results are \textit{formal consequences of the stability} of the homotopy theory \(\mathrm{Ch}(k)\)? Such constructions are still available if we replace \(\mathrm{Ch}(k)\) with any stable homotopy theory. For example, instead of working over a field \(k\), one could work over an arbitrary ring, a scheme, a dg-algebra, or a ring spectrum. As their main tool, the authors use \textit{derivators}, which encode enough structure of a homotopy theory to compute homotopy (co)limits. For instance, any model category has an underlying derivator; one is stable if and only if the other is. Here are the main results of the paper. For \(Q\) a Dynkin quiver of type \(A_n\) and \(\mathcal{D}\) a stable derivator, Theorem 4.15 yields an equivalence of derivators \(\mathcal{D}^Q \simeq \mathcal{D}^{M_n, \mathrm{ex}}\), where \(M_n\) is a variant of the classical \(k\)-linear mesh category. The superscript ``ex'' refers to exactness conditions, which play the role of the classical \(k\)-linear mesh relations. Using this result, the authors describe abstract reflection functors, Coxeter functors, Serre functors, and Auslander-Reiten translations as arising from symmetries of the mesh category \(M_n\). This extends some of their work from the prequels. A large part of the paper is devoted to spectral bimodules. In tilting theory, Rickard's theorem says that if two \(k\)-algebras \(R\) and \(S\) have triangulated-equivalent derived categories, then there is a two-sided tilting complex, i.e., a chain complex of \(R\)-\(S\)-bimodules whose derived tensor-hom adjunction \(D(S) \rightleftarrows D(R)\) is an equivalence, called standard. The authors refine the notion of tilting complex to the setup of stable derivators. Using the fact that every stable derivator is canonically a module over the derivator of spectra, they describe how a \textit{spectral bimodule} induces an adjunction of stable derivators \(\mathcal{D}^B \rightleftarrows \mathcal{D}^A\), where \(A\) and \(B\) are small categories. The main technical result (Theorem 8.5) says that certain morphisms of stable derivators, called \textit{admissible}, are represented by spectral bimodules via a tensor or hom construction; the proof yields an explicit formula for the representing bimodules. They describe the spectral bimodules representing the reflection functors, Coxeter functors, Serre functors, and the equivalence \(\mathcal{D}^Q \simeq \mathcal{D}^{M_n, \mathrm{ex}}\). They obtain a spectral refinement of Serre duality (Theorem 11.13). They show that for a quiver \(Q\) of type \(A_n\), the map from the spectral Picard group of \(Q\) to the derived Picard group of \(Q\) is a split epimorphism (Theorem 12.6). This relies on the fractionally Calabi-Yau property (Corollary 5.20), which says that the Serre functor \(S : \mathcal{D}^Q \to \mathcal{D}^Q\) and the suspension functor \(\Sigma\) are related by a natural isomorphism \(\Sigma^{n-1} \cong S^{n+1}\). Using their work on \(A_n\)-quivers, the authors prove a result of independent interest for homotopy theory. Stable derivators are known to be an enhancement of triangulated categories in the following sense: If \(\mathcal{D}\) is a strong stable derivator, then the value categories \(\mathcal{D}(A)\) can be canonically endowed with triangulated structures. Here, \textit{strong} refers to a rectification condition, satisfied notably by the underlying derivator of any stable model category. The authors improve the result in Theorem 13.6: If \(\mathcal{D}\) is a strong stable derivator, then the value categories \(\mathcal{D}(A)\) can be canonically endowed with \textit{higher triangulations} in the sense of Maltsiniotis.