Abstract
For an arbitrary quiver \(Q=(I,\Omega )\) and dimension vector \(\mathbf {d} \in \mathbb {N}^I\) we define the dimension of absolutely cuspidal functions on the moduli stacks of representations of dimension \(\mathbf {d}\) of a quiver Q over a finite field \(\mathbb {F}_q\), and prove that it is a polynomial in q, which we conjecture to be positive and integral. We obtain a closed formula for these dimensions of spaces of cuspidals for totally negative quivers.
Similar content being viewed by others
Notes
More precisely, the conjecture of Deligne and Kontsevich is phrased in terms of counting geometrically irreducible \(\overline{\mathbb {Q}_l}\)-representations of \(\pi _1^{arith}(X)\); the two problems are equivalent by the Langlands correspondence, [14].
In terms of plethystic functions, we may also write \(\sum _d C_{Q,d} z^d=\text{ Log }_z\text{ Exp }_{t,z}\left( tz(1-z)^{-1}\right) \).
In an effort to unburden the notation, we drop the subscript \(+\) in \(\mathfrak {n}_+\).
References
Borcherds, R.: Generalized Kac–Moody algebras. J. Algebra 115, 501–512 (1988)
Bozec, T.: Quivers with loops and generalized crystals. Compos. Math. 152(10), 1999–2040 (2016)
Bozec, T., Schiffmann, O., Vasserot, E.: On the number of points of nilpotent quiver varieties over finite fields, (2017). arXiv:1701.01797
Deligne, P.: Comptage de faisceaux l-adiques, in De la géométrie aux formes automorphes (I) (en l’honneur du soixantième anniversaire de Gérard Laumon, Astérisque 369, 285–312 (2015)
Davison, B., Meinhardt, S.: Cohomological Donaldson-Thomas theory of a quiver with potential and quantum enveloping algebras, (2016). preprint arXiv:1601.02479
Deng, B., Xiao, J.: A new approach to Kac’s theorem on representations of valued quivers. Math. Z. 245, 183–199 (2003)
Hausel, T.: Kac’s conjecture from Nakajima quiver varieties. Invent. Math. 181, 21–37 (2010)
Hausel, T., Letellier, E., Rodriguez-Villegas, F.: Positivity for Kac polynomials and DT-invariants of quivers. Ann. Math. (2) 177(3), 1147–1168 (2013)
Hua, J.: Counting representations of quivers over finite fields. J. Algebra 226(2), 1011–1033 (2000)
Hua, J., Xiao, J.: On Ringel-Hall algebras of tame hereditary algebras. Algebr. Represent. Theory 5(5), 527–550 (2002)
Kac, V.: Infinite root systems, representations of graphs and invariant theory. Invent. Math. 56(1), 57–92 (1980)
Kang, S.-J.: Quantum deformations of generalized Kac-Moody algebras and their modules. J. Algebra 175(3), 1041–1066 (1995)
Kontsevich, M.: Notes on motives in finite characteristic, (2007). arXiv:math/0702206
Lafforgue, L.: Chtoucas de Drinfeld et correspondance de Langlands. Invent. Math. 147(1), 1–242 (2002)
Maulik, D., Okounkov, A.: Quantum groups and quantum cohomology, (2012). arXiv:1211.1287
Mozgovoy, S.: Motivic Donaldson-Thomas invariants and McKay correspondence, (2011). arXiv:1107.6044
Okounkov, A.: On some interesting Lie algebras, Conference in honor of Victor Kac, IMPA (2013). https://www.youtube.com/watch?v=H8rCJ7ls1K4. Accessed Nov 2015
Schiffmann, O., Vasserot, E.: Cohomological Hall algebras of quivers: generators, (2017). arXiv:1705.07488
Sevenhant, B., Van den Bergh, M.: A relation between a conjecture of Kac and the structure of the Hall algebra. J. Pure Appl. Algebra 160, 319–332 (2001)
Acknowledgements
We are grateful to B. Davison and A. Okounkov for some stimulating discussion and correspondences, and to B. Deng and J. Xiao for explanations concerning their work [6]. The work of the first author started during his postdoctoral appointment at MIT, before being supported by the LABEX MILYON (ANR-10-LABX-0070) of Université de Lyon, within the program “Investissements d’Avenir” (ANR-11-IDEX-0007) operated by the French National Research Agency (ANR). The second author was partially supported by ANR (Grant 13-BS01- 0001-01).
Author information
Authors and Affiliations
Corresponding author
Appendix A. A trivial variant of Kac’s conjecture
Appendix A. A trivial variant of Kac’s conjecture
Proposition A.1
Let \(Q=(I,\Omega )\) be any quiver, and let \(Q^{re}=(I^{re},\Omega ^{re})\) be the full subquiver of Q whose vertices are real. Then for any dimension vector \(\mathbf {d}\in \mathbb {N}^I\) we have
Proof
It is enough to show that \(A_{Q,\mathbf {d}}(0)=0\) whenever \(\mathbf {d}\) is not supported on \(I^{re}\) (and then we are reduced to the usual Kac conjecture proved by Hausel). Let \(H_{Q,\mathbf {d}}(t) \in \mathbb {N}[t]\) be the polynomial counting all isomorphism classes of representations of dimension \(\mathbf {d}\). The relation to \(A_{Q,\mathbf {n}}(t)\) reads
Evaluating (A.1) at \(t=0\), we see that \(A_{Q,\mathbf {d}}(0)=0\) for all \(\mathbf {d}\not \in \mathbb {N}^{I^{re}}\) if and only if \(H_{Q,\mathbf {d}}(0)=0\) for all \(\mathbf {d}\not \in \mathbb {N}^{I^{re}}\). Writing \(H_{Q,\mathbf {d}}(t)=c_{Q,\mathbf {d}} + t K_{Q,\mathbf {d}}(t)\) with \(c_{Q,\mathbf {d}} \in \mathbb {N}\) and \(K_{Q,\mathbf {d}}(t) \in \mathbb {N}[t]\) we immediately see that \(H_{Q,\mathbf {d}}(0)=0\) of and only if \(q\;|\; H_{Q,\mathbf {d}}(q)\) for all prime powers q. Let us now fix a finite field \(\mathbf {k}\) and \(\mathbf {d}\not \in \mathbb {N}^{I^{re}}\). Let \(i \in I \backslash I^{re}\) be an imaginary vertex for which \(\mathbf {d}_i \ne 0\), and let \(h \in \Omega \) be an edge loop at i. Define an action of the additive group \(\mathbb {G}_a(\mathbf {k})=(\mathbf {k},+)\) on the set \(Ind_{Q,\mathbf {d},\mathbf {k}}\) of isomorphism classes of indecomposable representations of dimension \(\mathbf {d}\) by setting
Considering the eigenvalues of \(x_h\), we see that this action is free. This proves that \(q\;|\; H_{Q,\mathbf {d}}(q)\) as wanted. \(\square \)
Remark
A more interesting variant of Kac’s conjecture in the context of quivers with edge loops is proved in [3]: the constant term of the nilpotent Kac polynomial \(A^{1}_{Q,\mathbf {d}}(t)\) is equal to the multiplicity of the root \(\mathbf {d}\) in the generalized Borcherds algebra \(\mathfrak {g}_Q\) associated to Q in [2].
Rights and permissions
About this article
Cite this article
Bozec, T., Schiffmann, O. Counting absolutely cuspidals for quivers. Math. Z. 292, 133–149 (2019). https://doi.org/10.1007/s00209-018-2155-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00209-018-2155-5