English

Non-semistable exceptional objects in hereditary categories: some remarks and conjectures

Category Theory 2018-06-01 v3 Algebraic Geometry Representation Theory

Abstract

In our previous paper we studied non-semistable exceptional objects in hereditary categories and introduced the notion of regularity preserving category, but we obtained quite a few examples of such categories. Certain conditions on the Ext-nontrivial couples (exceptional objects X,YAX,Y\in \mathcal A with Ext1(X,Y)0{\rm Ext}^1(X,Y)\neq 0 and Ext1(Y,X)0{\rm Ext}^1(Y,X)\neq 0) were shown to imply regularity-preserving. This paper is a brief review of the previous paper (with emphasis on regularity preserving property) and we add some remarks and conjectures. It is known that in Dynkin quivers Hom(ρ,ρ)=0{\rm Hom}(\rho,\rho')=0 or Ext1(ρ,ρ)=0{\rm Ext}^1(\rho,\rho')=0 for any two exceptional representations. In the present paper we use this property to show that for any Dynkin quiver QQ there are no Ext-nontrivial couples in Repk(Q)Rep_k(Q), which implies regularity preserving of Repk(Q)Rep_k(Q), where kk is an algebraically closed field. We study this property in other quivers. In particular in any star quiver with three arms QQ for any two exceptional representations ρ,ρ\rho, \rho' we have Hom(ρ,ρ)=0{\rm Hom}(\rho,\rho')=0 or Ext1(ρ,ρ)=0{\rm Ext}^1(\rho,\rho')=0 provided that ρ\rho or ρ\rho' is a thin representation. In the previous version we asserted falsely that this holds for any two exceptional representations (without imposing the restriction that one of them is thin) for extended Dynkin quivers E~6,E~7,E~8\widetilde{\mathbb E}_6, \widetilde{\mathbb E}_7, \widetilde{\mathbb E}_8 .

Cite

@article{arxiv.1405.2943,
  title  = {Non-semistable exceptional objects in hereditary categories: some remarks and conjectures},
  author = {George Dimitrov and Ludmil Katzarkov},
  journal= {arXiv preprint arXiv:1405.2943},
  year   = {2018}
}

Comments

Lemma 5.5 in version 2 is false, Claus M. Ringel pointed a counterexample to us. The claim in the proof "Now the arguments are the same as the arguments after Figure (46) in the proof of Lemma 5.4." is false: one of the arguments is not applicable. Here we have done the necessary corrections. We have already proved conjecture 7.3 in the previous version 2, it is removed in this version

R2 v1 2026-06-22T04:12:23.243Z