Ishii’s conjecture and Bridgeland stability conditions for dihedral reflection groups
Shu Nimura · 2026
Read Ishii’s conjecture and Bridgeland stability conditions for dihedral reflection groups by Shu Nimura (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-2605-22474.txt.
Introduction
Nimura gives a fresh proof of Ishii's conjecture for every dihedral reflection group, seen through Bridgeland stability rules. The route runs through the derived McKay link. That link turns the claim into a geometric build of stability rules on the root stack of the top resolution, taken along the discriminant divisor's strict transform. The proof thus ties a group theory conjecture to a concrete stability construction.
Catalog record
| CREATOR | Shu Nimura |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 119974 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.22474
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Ishii’s conjecture and Bridgeland stability conditions for dihedral reflection groups Shu Nimura Abstract . We provide a new proof of Ishii’s conjecture for any dihedral reflection group G ⊂ G L 2 ( ℂ ) G\subset GL_{2}(\mathbb{C}) from the viewpoint of Bridgeland stability conditions. Our strategy is to reduce the problem, via the derived McKay correspondence, to a geometric construction of Bridgeland stability conditions on the root stack of the maximal resolution along the strict transform of the discriminant divisor. 1. Introduction 1.1. Motivation One of the central themes in the McKay correspondence is the relationship between birational geometry and moduli spaces of representations of noncommutative algebras. Craw and Ishii proposed a conjecture [ 11 ] on the crepant resolutions of \mathbb{C}^{3}/G for any finite subgroup G\subset SL_{3}(\mathbb{C}) : every crepant resolution arises as a moduli space of \theta -stable representations by varying King stability \theta\in\Theta(G) [ 22 ] . The conjecture was proved for abelian groups in SL_{3}(\mathbb{C}) by [ 11 ] and for general finite groups in SL_{3}(\mathbb{C}) by [ 39 ] . Furthermore, inspired by the DK hypothesis, Ishii proposed the following conjecture in dimension two, extending the Craw–Ishii conjecture to the non-crepant setting: Conjecture 1.1 ( [ 15 ] ) . Let G\subset GL_{2}(\mathbb{C}) be any finite subgroup. Then, a resolution of \mathbb{C}^{2}/G is dominated by the maximal resolution of the pair (\mathbb{C}^{2}/G,B) if and only if X\cong M_{\theta} for some generic \theta\in\Theta(G) . For the definitions, see §2. Here, we only note that if B=0 (equivalently, if G acts freely in codimension one), the crepant resolution, when it exists, is both minimal and maximal. Ishii proved the “if” part in full generality, and also the “only if” part for arbitrary abelian subgroups and for small finite subgroups G\subset GL_{2}(\mathbb{C}) , i.e., when G acts freely in codimension one. Later, [ 9 ] proved the conjecture for dihedral reflection groups. This paper aims to give a new proof of Ishii’s conjecture in the case of dihedral reflection groups via Bridgeland stability conditions and wall-crossing phenomena. 1.2. Bridgeland stability Bridgeland stability conditions were introduced by Bridgeland [ 6 ] inspired by Douglas’s work on \Pi -stability [ 13 ] . Bridgeland stability conditions play a central role in the study of triangulated categories arising in birational geometry, representation theory, enumerative geometry, and mirror symmetry. Bridgeland stability can be regarded as a generalization of classical notions of stability, such as Gieseker stability for sheaves on a variety and King stability for representations of a noncommutative algebra. In the context of birational geometry of surfaces, Toda interpreted the surface MMP in terms of wall-crossing phenomena for moduli spaces of Bridgeland stable objects [ 36 ] . We briefly review Toda’s work. Let Y be a smooth projective surface. The statement is formulated in the ‘real part’ of \mathrm{Stab}(Y)_{\mathbb{R}} of the stability manifold \mathrm{Stab}(Y) . This space fits into the following cartesian diagram (see also §2) {\lx@inpgf@ignorespaces\mathrm{Stab}(Y)_{\mathbb{R}}} {\lx@inpgf@ignorespaces\mathrm{Stab}(Y)} {\lx@inpgf@ignorespaces\mathrm{NS}(Y)_{\mathbb{R}}} {\lx@inpgf@ignorespaces K^{\mathrm{num}}(Y)_{\mathbb{C}}^{\vee}.} \scriptstyle{\lx@inpgf@ignorespaces\Pi_{\mathbb{R}}} \scriptstyle{\lx@inpgf@ignorespaces\Pi} Here, we fix the notation: for a Bridgeland stability condition \sigma=(Z,\mathcal{A}) , M_{\sigma}([\mathcal{O}_{y}]) denotes the algebraic space parametrizing Z -stable objects E\in\mathcal{A} with [E]=[\mathcal{O}_{y}] in K^{\mathrm{num}}(Y) for y\in Y . Toda proved the following theorem: Theorem 1.2 ( [ 36 ] ) . Let Y be any smooth projective surface. For every smooth contraction Y\to X , we can associate a connected open subset U(X)\subset\mathrm{Stab}(Y)_{\mathbb{R}} such that (1) For any \sigma\in U(X) , M_{\sigma}([\mathcal{O}_{x}])\cong X . (2) If X,X^{\prime} are related by a single blowup, then \overline{U(X)}\cap\overline{U(X^{\prime})} is nonempty and has real codimension one. To prove this, Toda constructs a local section of \Pi_{\mathbb{R}} \mu^{\prime}_{\mathrm{geom}}:A^{\mathrm{tot}}:=\bigcup_{Y\to X}\overline{A^{\dagger}}(X)\to\mathrm{Stab}_{\mathbb{R}}(Y), where the union runs over all smooth contractions Y\to X , and \overline{A^{\dagger}}(X) is a subset of the Neron-Severi group \mathrm{NS}(Y)_{\mathbb{R}} . U(X) is defined to be the image of the interior A^{\dagger}(X) . On the other hand, in the context of representation theory of algebras, Bayer–Craw–Zhang constructed a continuous map [ 2 ] , \mu^{\prime}_{\mathrm{alg}}:\Theta(G)\times\Lambda_{1}\to\mathrm{Stab}([\mathbb{C}^{2}/G]), where \Lambda_{1} is a certain space of additional parameters. Via \mu^{\prime}_{\mathrm{alg}} , one can identify King stability and Bridgeland stability, allowing one to interpret Ishii’s conjecture in terms of Bridgeland moduli (see §4.1 for the details.) 1.3. Main results Our strategy for Ishii’s conjecture is to construct an analogous map to \mu^{\prime}_{\mathrm{geom}} for a certain DM stack \mathcal{Y} and to compare \mu^{\prime}_{\mathrm{geom}} and \mu^{\prime}_{\mathrm{alg}} via the McKay correspondence. To state our results more precisely, we prepare some notations. Let G be a dihedral reflection group and set H:=G\cap SL_{2}(\mathbb{C}) . Let Y_{H}:=H\text{-Hilb}(\mathbb{C}^{2}) be the H -Hilbert scheme defined by [ 19 ] . Y_{H} admits a natural action of G/H\cong\mathbb{Z}_{2} , and we consider \mathcal{Y}:=[Y_{H}/\mathbb{Z}_{2}] , which can be regarded as the second root stack over the quotient variety Y:=Y_{H}/\mathbb{Z}_{2} along the branch locus. Y is the maximal resolution in Ishii’s conjecture. We write \pi:\mathcal{Y}\to Y for the coarse moduli map, or the quotient map. In this situation, the universal family of H\text{-Hilb}(\mathbb{C}^{2}) with G/H -action induces an exact equivalence \Phi:D_{c}(\mathcal{Y})\cong D_{c}([\mathbb{C}^{2}/G]), where D_{c}(-) denotes the bounded derived category of coherent sheaves whose cohomology sheaves are compactly supported. This provides the natural setting for Bridgeland stability for our purposes since the corresponding Bridgeland moduli spaces are well-behaved (unlike the derived category D_{Z}(-) supported on a fixed compact subscheme Z ). To formulate the analogue of Toda’s work, we introduce \mathrm{NS}_{{\mathrm{orb}}}(\mathcal{Y})_{\mathbb{R}} , the orbifold Néron-Severi group (see §3.3 for the definition). For our purposes, we consider \mathrm{NS}_{{\mathrm{orb}}}(\mathcal{Y})_{\mathbb{C}}:=\mathrm{NS}_{{\mathrm{orb}}}(\mathcal{Y})_{\mathbb{R}}\otimes\mathbb{C}. We define \mathrm{Stab}_{\mathrm{n}}(\mathcal{Y}) to be the space of normalized stability conditions, which fits into the following cartesian diagram (see §3.3) {\lx@inpgf@ignorespaces\mathrm{Stab}_{\mathrm{n}}(\mathcal{Y})} {\lx@inpgf@ignorespaces\mathrm{Stab}(\mathcal{Y})} {\lx@inpgf@ignorespaces\mathrm{NS}_{{\mathrm{orb}}}(\mathcal{Y})_{\mathbb{C}}} {\lx@inpgf@ignorespaces K^{\mathrm{num}}_{c}(\mathcal{Y})_{\mathbb{C}}^{\vee}.} \scriptstyle{\lx@inpgf@ignorespaces\Pi_{\mathrm{n}}} \scriptstyle{\lx@inpgf@ignorespaces\Pi} As in the ordinary surface case, we consider the numerical class of skyscraper sheaves on \mathcal{Y} , which is explicitly described as [\pi^{*}\mathcal{O}_{y}]\in K_{c}^{\mathrm{num}}(\mathcal{Y}) for any y\in Y . Theorem A . For any smooth contraction Y\to X of the maximal resolution Y , we can associate a connected open subset \mathcal{U}(X)\subset\mathrm{Stab}_{\mathrm{n}}(\mathcal{Y}) such that (1) For any \sigma\in\mathcal{U}(X) , M_{\sigma}([\pi^{*}\mathcal{O}_{y}])\cong X . (2) If X and X^{\prime} are related by a single blowup, then \overline{\mathcal{U}(X)}\cap\overline{\mathcal{U}(X^{\prime})} is nonempty and has real codimension one. As a result, the birational geometry of the maximal resolution Y is encoded by wall-crossing phenomena of Bridgeland moduli spaces. More precisely, we will construct a continuous section of \Pi_{\mathrm{n}} \mu_{\mathrm{geom}}:A^{\mathrm{tot}}\times\mathcal{B}:=\left(\bigcup_{Y\to X}\overline{A^{\dagger}}(X)\right)\times\mathcal{B}\to\mathrm{Stab}_{\mathrm{n}}(\mathcal{Y}) where \overline{A^{\dagger}(X)}\subset\mathrm{NS}_{{\mathrm{orb}}}(\mathcal{Y})_{\mathbb{R}} is the counterpart of \overline{A^{\dagger}(X)} appearing in Toda’s construction, and \mathcal{B} is a small open neighborhood of 0 in \mathrm{NS}_{{\mathrm{orb}}}(\mathcal{Y})_{\mathbb{R}} that is independent of the contractions. We can think of \mu_{\mathrm{geom}} as a small deformation of \mu^{\prime}_{\mathrm{geom}} inside \mathrm{Stab}_{\mathrm{n}}(\mathcal{Y}) . A subtle but crucial point is that the real slice \mathrm{Stab}_{\mathbb{R}}(\mathcal{Y}) is too small to compare \mu^{\prime}_{\mathrm{geom}} with \mu^{\prime}_{\mathrm{alg}} . The two local sections do not admit a nonempty common domain inside \mathrm{Stab}_{\mathbb{R}}(\mathcal{Y}) . The deformation by the additional parameter space \mathcal{B} is introduced to produce such a common domain. The main point is that, after the deformation, the two local sections glue along the resulting overlap: Theorem B . After identifying the two stability manifolds \mathrm{Stab}(\mathcal{Y}) and \mathrm{Stab}([\mathbb{C}^{2}/G]) via the derived equivalence \Phi , an explicit change of parameters, and a rotation action, the local sections of \Pi_{\mathrm{n}} \mu_{\mathrm{geom}}\quad\text{and}\quad\mu^{\prime}_{\mathrm{alg}} coincide on their common domain. See Theorem 4.13 for the precise statement. Combining Theorem