Explicit computation of the generic degree of the generalized Verschiebung in rank two

Yuki Kondo, Yasuhiro Wakabayashi · 2025

Read Explicit computation of the generic degree of the generalized Verschiebung in rank two by Yuki Kondo, Yasuhiro Wakabayashi (2025) 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-2509-03993.txt.

Introduction

In a 2025 mathematics paper, Yuki Kondo and one co-author compute the generic degree of the generalised Verschiebung map in rank two. In positive characteristic, the Frobenius map raises coordinates to the p-th power and pulls bundles back along it. The Verschiebung is the induced map on moduli spaces of stable bundles, and its generic degree counts preimages over a typical point.

Earlier work on dormant opers supplies the key input. A conjecture of the second author predicted a direct link between counting higher-level dormant opers and these degrees, and the rank two case is resolved affirmatively here. Consequently, for general curves of genus above one in characteristic above two, the degrees can be determined explicitly by counting certain combinatorial objects. One degree is computed, for genera three through ten, as a polynomial in p. Such explicit formulas turn an abstract moduli question into concrete, checkable data.

Catalog record

CREATORYuki Kondo, Yasuhiro Wakabayashi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE103 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-03993

Browse the full library · About the Tower of Babel

Full text

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

Explicit computation of the generic degree of the generalized Verschiebung in rank two Yuki Kondo, Yasuhiro Wakabayashi Abstract . The purpose of this paper is to apply previous work on dormant opers to the study of the moduli space of stable bundles in positive characteristic. We affirmatively resolve the rank 2 2 case of a conjecture proposed by the second author, which predicts a direct relationship between the number of higher-level dormant PGL 2 \mathrm{PGL}_{2} -opers and the generic degree of the generalized Verschiebung map for rank 2 2 stable bundles induced by Frobenius pull-back. As a consequence, we obtain a procedure for explicitly determining these generic degrees in the previously unexplored range of genera by counting certain combinatorial objects. † † footnotetext: 2020 Mathematical Subject Classification : Primary 14H60, Secondary 14G17; † † footnotetext: Key words: positive characteristic, stable bundle, generalized Vershiebung, dormant oper 1. Introduction 1.1. Counting problem for the generalized Verschiebung map Let k be an algebraically closed field of characteristic p>2 and X a smooth projective curve over k of genus g>1 . For a positive integer N , we denote by X^{(N)} the N -th Frobenius twist of X over k , i.e., the base-change of X along the N -th iterate of the Frobenius automorphism of \mathrm{Spec}(k) . Pulling-back stable bundles on X^{(N)} via the N -th relative Frobenius morphism F^{(N)}_{X/k}:X\rightarrow X^{(N)} gives rise to a rational map (1.1) \displaystyle\mathrm{Ver}_{N}^{n}:SU_{X^{(N)}}^{n}\dashrightarrow SU_{X}^{n} between the moduli spaces SU_{X^{(N)}}^{n} and SU_{X}^{n} of rank n>1 stable bundles with trivial determinant on X^{(N)} and X , respectively. We refer to this rational map as the ( N -th) generalized Verschiebung map . It captures the dynamics of stable bundles under Frobenius pull-back and has been studied from various perspectives in, e.g., [ BrKa ] , [ dSa ] , [ Gie ] , [ JoPa ] , [ JRXY ] , [ Las ] , [ LasPa1 ] , [ LasPa2 ] , [ Li1 ] , and [ Oss2 ] . See also [ DuMe ] , [ LanPa ] , [ Li2 ] , [ Li3 ] , [ Li4 ] , [ Oss1 ] , and [ Yan ] for studies on the geometric structure of \mathrm{Ver}^{n}_{1} itself. It is known that \mathrm{Ver}^{n}_{N} is generically finite, which allows us to define its generic degree \mathrm{deg}(\mathrm{Ver}_{N}^{n}) , a quantity that reflects the complexity of this rational map. Although the Frobenius action on the cohomology ring of the moduli stack of vector bundles has been investigated in [ CaNe ] and [ NeSt ] , not much is known about \mathrm{deg}(\mathrm{Ver}_{N}^{n}) . It therefore arises as a natural question to know how much this value is explicitly. Note that \mathrm{Ver}^{n}_{N} factors as a composite of N rational maps obtained as Frobenius twists of \mathrm{Ver}^{n}_{1} : (1.2) \displaystyle SU_{X^{(N)}}^{n}\dashrightarrow SU_{X^{(N-1)}}^{n}\dashrightarrow\cdots\dashrightarrow SU_{X^{(1)}}^{n}\dashrightarrow SU_{X}^{n}. This implies the identity \mathrm{deg}(\mathrm{Ver}_{N}^{n})=\mathrm{deg}(\mathrm{Ver}^{n}_{1})^{N} , which reduces the problem to the case N=1 . As far as the authors are aware, the case (n,g)=(2,2) is the only one for which the value of the generic degree has been explicitly determined so far. For a genus- 2 curve X , the rational map \mathrm{Ver}^{2}_{1} is relatively more accessible for analysis. In fact, the moduli space SU_{X}^{2} (as well as SU_{X^{(1)}}^{2} ) admits a natural compactification by semistable bundles, which is canonically isomorphic to \mathbb{P}^{3} , and its boundary can be identified with the Kummer surface associated to X . Moreover, the base locus of \mathrm{Ver}^{2}_{1} coincides with the moduli space of certain flat bundles known as dormant \mathrm{PGL}_{2} -opers . By this observation, an explicit computation of the total number of such flat bundles, established in [ Moc ] , leads to the following formula: (1.3) \displaystyle\mathrm{deg}(\mathrm{Ver}_{1}^{2})=\frac{1}{3}\cdot(p^{3}+2p) See also [ Oss1 , Theorem 1.3] , [ LanPa , Corollary] for this computation. For arbitrary genus g , the main result of [ HoWa ] provides an upper bound for \mathrm{deg}(\mathrm{Ver}^{2}_{1}) , based on the idea of K. Joshi involving a correspondence with certain Quot schemes (cf. [ Jos ] , [ JoPa ] ). However, this upper bound has shown to be suboptimal (cf. the remark preceding [ HoWa , Corollary 3.2.4] ), and the overall structure of \mathrm{Ver}^{n}_{1} remains mysterious (even in the case n=2 ) despite its fundamental significance. In order to advance the discussion beyond this point, we propose a new approach that connects the problem to the enumeration of certain combinatorial objects, via the theory of higher-level dormant opers developed in [ Wak7 ] . 1.2. Comparison with dormant \mathrm{PGL}_{2} -opers of higher level In the remainder of this introduction, we outline the contents of this paper. A central object in our discussion is a dormant \mathrm{PGL}_{2}^{(N)} -oper (for N\in\mathbb{Z}_{>0} ), which is a dormant \mathrm{PGL}_{2} -oper enhanced by an action of differential operators of level N-1 , introduced by P. Berthelot in the theory of arithmetic \mathcal{D} -modules (cf. [ Ber ] ). Let \overline{\mathcal{M}}_{g} denote the moduli stack of genus- g stable curves over k , and let (1.4) \displaystyle\mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}} (cf. ( 3.11 )) denote the moduli stack parametrizing pairs (X,\mathscr{F}^{\spadesuit}) consisting of a curve X\in\overline{\mathcal{M}}_{g} and a dormant \mathrm{PGL}_{2}^{(N)} -oper \mathscr{F}^{\spadesuit} on X . As a generalization of [ JoXi , Proposition 3.3] (see also [ JRXY ] , [ JoPa ] , [ Wak6 ] ), we show that a higher-level version of Cartier descent yields (upon fixing a theta characteristic) a correspondence between dormant \mathrm{PGL}_{2}^{(N)} -opers and certain rank 2 stable bundles that are called maximally F^{(N)} -destabilized (cf. Definition 4.1 , Proposition 4.3 , and ( 3.10 )). Let a be a k -rational point of SU^{2}_{X^{(N)}} corresponding to such a stable bundles, and let b be the point in \mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}} determined by the associated dormant \mathrm{PGL}_{2}^{(N)} -oper. Then, one can verify the existence of a canonical isomorphism (1.5) \displaystyle\eta_{N}:T_{a}SU^{2}_{X^{(N)}}\xrightarrow{\sim}T_{b}\mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}} between the tangent spaces at a and b , respectively (cf. Theorem 4.5 , (i)). Moreover, for each N^{\prime}\in\mathbb{Z}_{>0} with N^{\prime}\leq N , the following square diagram commutes: \textstyle{T_{a}SU^{2}_{X^{(N)}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\eta_{N}} \scriptstyle{\sim} \scriptstyle{d\mathrm{Ver}^{2}_{N\Rightarrow N^{\prime}}|_{a}} \textstyle{T_{b}\mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{d\Pi_{N\Rightarrow N^{\prime}}|_{b}} \textstyle{T_{a^{\prime}}SU_{X^{(N^{\prime})}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\sim} \scriptstyle{\eta_{N^{\prime}}} \textstyle{T_{b^{\prime}}\mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}}} (cf. Theorem ( 4.5 ), (ii), for the precise definitions of the various notations). In other words, the local behavior of (the first N-1 arrows of) the sequence ( 1.2 ) is fully compatible with that of the sequence (1.10) \displaystyle\mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}}\rightarrow\mathcal{O}p_{N-1,g}^{{}^{\mathrm{Zzz...}}}\rightarrow\cdots\rightarrow\mathcal{O}p_{2,g}^{{}^{\mathrm{Zzz...}}}\rightarrow\mathcal{O}p_{1,g}^{{}^{\mathrm{Zzz...}}} induced by reducing the level of dormant \mathrm{PGL}_{2} -opers. This compatibility, together with the generic étaleness of \mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}}/\overline{\mathcal{M}}_{g} established in [ Wak7 , Theorem C, (i)] , leads to the following statement, which forms the first main result of this paper. Theorem A (cf. Corollary 4.11 ) . Let \Pi_{N} denote the natural projection \mathcal{O}p_{N,g}^{{}^{\mathrm{Zzz...}}}\rightarrow\overline{\mathcal{M}}_{g} , which is finite and faithfully flat (cf. [ Wak7 , Theorems B and C, (i)] ). Then, the following inequality holds: (1.11) \displaystyle\mathrm{deg}(\Pi_{N})\geq\frac{\mathrm{deg}(\mathrm{Ver}_{1}^{2})^{N-1}\cdot p^{g-1}}{2^{2g-1}}\cdot\sum_{\theta=1}^{p-1}\frac{1}{\sin^{2g-2}\big(\frac{\pi\cdot\theta}{p}\big)}. Moreover, if X is general (cf. Section 1.4 ), then this inequality “ \geq ” becomes an equality “ = ”. In particular, the resulting formula gives an affirmative solution to a strong form (for n=2 ) of the conjecture stated in [ Wak7 , Conjecture 10.26] (cf. Remark 4.12 ). 1.3. Explicit computations via edge numberings on a trivalent graph A key point of the

Read the complete text