Equivariant irrationality of very general symmetric Verra fourfolds
Aideen Fay · 2026
Read Equivariant irrationality of very general symmetric Verra fourfolds by Aideen Fay (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-2605-30439.txt.
Introduction
Aideen Fay proves no Z/2-birational map links a very general complex symmetric Verra fourfold to projective 4-space. She uses the theory of atoms, from Katzarkov, Kontsevich, Pantev, and Yu, in the equivariant setting of Cavenaghi, Katzarkov, and Kontsevich. The result shows that a generic member of this symmetric family cannot be mapped to projective 4-space by a birational map respecting the twofold symmetry. The argument carries the involution through the theory of atoms. This places a sharp limit on symmetric birational models of Verra fourfolds.
Catalog record
| CREATOR | Aideen Fay |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50870 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.30439
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Equivariant irrationality of very general symmetric Verra fourfolds Aideen Fay Abstract . We prove that a very general complex symmetric Verra fourfold is not ℤ / 2 \mathbb{Z}/2 -birational to ℙ 4 \mathbb{P}^{4} , using the theory of atoms introduced by Katzarkov–Kontsevich–Pantev–Yu ( 14 ) in the equivariant setting as in Cavenaghi–Katzarkov–Kontsevich ( 6 ) . 1. Introduction In their seminal work, Katzarkov–Kontsevich–Pantev–Yu ( 14 ) introduce the theory of atoms , which combines Hodge theory and Gromov–Witten invariants to define birationally invariant properties of smooth complex projective varieties. This is underpinned by Iritani’s blowup formula ( 13 , Theorem 1.1) , which dictates the additive decomposition of quantum D -modules under blowups along smooth centres. This approach has already settled long-standing conjectures on the irrationality of certain Fano varieties—most recently for cubic 14 ; 10 and Gushel–Mukai 4 fourfolds. Progress to date leaves the Verra fourfold as the last of the classical Fano fourfolds of K3 type with index >1 5 , whose rationality remains unknown. Realized as a double cover \pi\colon X\to\mathbf{P}^{2}\times\mathbf{P}^{2} branched along a smooth divisor B of bidegree (2,2) , a Verra fourfold X naturally admits two quadric fibrations over \mathbf{P}^{2} . The discriminant locus of each fibration is a plane sextic curve whose double cover, when smooth, is a degree-two K3 surface ( 15 ) . This explicit K3 geometry presents a unique challenge to the birational invariants introduced above, which remain inconclusive here 3 . If we shift perspective, we notice that the branch locus B of a symmetric Verra remains invariant under swapping of the \mathbf{P}^{2}\times\mathbf{P}^{2} factors. In this case the factor-swap lifts to an automorphism \epsilon:X\longrightarrow X , and hence defines an action of the group G=\mathbb{Z}/2=\langle\epsilon\rangle on the cohomology of X . This observation motivates us to adopt the G -equivariant obstructions introduced by Cavenaghi–Katzarkov–Kontsevich ( 6 ) to prove \mathbb{Z}/2 -irrationality of the very general symmetric Verra fourfold. Definition 1.1 ( \mathbb{Z}/2 -rationality) . Let Y be a smooth projective variety of dimension n with an involution \iota:Y\to Y . The pair (Y,\iota) is \mathbb{Z}/2 -rational if there is a birational map \varphi:Y\dashrightarrow\mathbb{P}^{n} and a regular involution \sigma:\mathbb{P}^{n}\to\mathbb{P}^{n} such that \sigma=\varphi\circ\iota\circ\varphi^{-1} on the Zariski dense open subset where \varphi^{-1} is defined. Otherwise (Y,\iota) is \mathbb{Z}/2 -irrational . Theorem 6.4 . Let X be a very general symmetric Verra fourfold, and let \epsilon be its factor-swap involution. Then (X,\epsilon) is \mathbb{Z}/2 -irrational. Equivalently, for every regular involution \iota\in\Aut(\mathbb{P}^{4}) , the pair (X,\epsilon) is not \mathbb{Z}/2 -birational to (\mathbb{P}^{4},\iota) . Remark 1.2 . The same argument applies if \mathbb{P}^{4} is replaced by any smooth projective fourfold P with a regular \mathbb{Z}/2 -action and no cohomology classes of Hochschild degree 2. For instance, one may take P=\mathbb{P}^{2}\times\mathbb{P}^{2} . However, the argument does not apply to an arbitrary smooth rational fourfold. Thus the involution \iota in the statement must be regular and not just rational. In pursuit of this, we verify the G -equivariance of Iritani’s blowup decomposition ( Theorem 5.2 ), and introduce an evaluated G -equivariant version of these obstructions, following Guéré’s work in the non-equivariant setting ( 10 ) . Specifically, we define in this paper what very general means for symmetric Verra fourfolds. Indeed, the natural condition of being Hodge-general is actually empty (see Corollary 3.5 ), but a more refined condition that we call \mathbb{Z}/2 -Hodge generality is not (see Proposition 4.4 ). We refer to Section 4 for the precise definitions. 2. Evaluated Hodge invariants We recall the notation used throughout the paper. We always work with smooth projective varieties over \mathbb{C} . The notation and definitions follow Guéré’s evaluated formulation ( 10 ) of the Hodge-theoretic invariants introduced in ( 14 ) . Definition 2.1 . Let Y be a smooth projective complex variety. A class \gamma\in H^{2p}(Y,\mathbb{Q}) is a rational Hodge class if its complexification lies in H^{p,p}(Y) . We write H^{*}(Y,\mathbb{Q})_{\mathrm{Hdg}}:=\bigoplus_{p}(H^{2p}(Y,\mathbb{Q})\cap H^{p,p}(Y)). For k\in\mathbb{Z} , set H^{(k)}(Y):=\bigoplus_{p-q=k}H^{p,q}(Y)\subset H^{*}(Y,\mathbb{C}). We call H^{(k)}(Y) the Hochschild-degree- k part of the Hodge decomposition. Let \operatorname{NE}(Y)\subset H_{2}(Y,\mathbb{Z}) be the monoid of effective curve classes in Y . For d\in\operatorname{NE}_{\mathbb{N}}(Y) , we associate the formal Novikov variable Q^{d}\in\mathbb{Q}[\operatorname{NE}_{\mathbb{N}}(Y)] which records the curve class of stable maps contributing to a Gromov–Witten invariant. It is assigned the grading \deg Q^{d}=2\,c_{1}(Y)\cdot d . Fix a homogeneous basis \{\alpha_{i}\}_{i=0}^{m} of the rational Hodge classes H^{*}(Y,\mathbb{Q})_{\mathrm{Hdg}} . Let \{t^{i}\}_{i=0}^{m} be formal variables, graded by \deg t^{i}=2-\deg\alpha_{i} . A Hodge parameter is then \tau=\sum_{i}t^{i}\alpha_{i} . We use the completed Novikov–Hodge ring \widehat{R}(Y,K):=K[q^{\prime\pm 1}]\llbracket\mathrm{NE}(Y)\rrbracket\llbracket t^{0},\dots,t^{m}\rrbracket, where K is a number field, q^{\prime} is a formal Laurent variable coming from Iritani’s blowup formula, and the completion is graded; see ( 13 , Section 2.2) . The subring R(Y,K)\subset\widehat{R}(Y,K) consists of series with no non-zero powers of q^{\prime} . The big quantum product is denoted by \star_{\tau} . Let \operatorname{Eu}_{\tau} be the Euler vector field defined as \operatorname{Eu}_{\tau}:=c_{1}(Y)+\sum_{i=0}^{m}\left(1-\frac{\deg\alpha_{i}}{2}\right)t^{i}\alpha_{i}. We then define the operator \kappa_{\tau} by \kappa_{\tau}:=\operatorname{Eu}_{\tau}\star_{\tau}(-). At \tau=0 , this specializes to the small quantum multiplication operator \kappa_{0}=c_{1}(Y)\star_{0}(-). The operator \kappa_{\tau} preserves rational Hodge classes and Hochschild degree ( 10 , Proposition 14) . An evaluation, in the sense of ( 10 , Definition 20) , is a continuous degree-preserving homomorphism \operatorname{ev}:\widehat{R}(Y,K)\longrightarrow S_{K} where S_{K}=F_{K}[b^{\pm 1}] , and F_{K} is the Levi–Civita field over K . Applying \operatorname{ev} to the coefficients of \kappa_{\tau} gives an S_{K} –linear endomorphism of H^{*}(Y,K)\otimes_{K}S_{K} . Passing, if necessary, to a graded field extension L\supseteq S_{K} splitting the characteristic polynomial of \operatorname{ev}(\kappa_{\tau}) , we regard the evaluated operator \operatorname{ev}(\kappa_{\tau}) as an endomorphism of H^{*}(Y,K)\otimes_{K}L . Definition 2.2 (Evaluated generalized eigenspaces and Hochschild rank) . Let \lambda\in L . The evaluated generalized eigenspace is E^{Y}_{\operatorname{ev},\lambda}:=\ker\bigl(\operatorname{ev}(\kappa_{\tau})-\lambda\bigr)^{N},\qquad N\gg 0, with Hodge-class part \mathcal{H}^{Y}_{\operatorname{ev},\lambda}:=E^{Y}_{\operatorname{ev},\lambda}\cap\bigl(H^{*}(Y,\mathbb{Q})_{\mathrm{Hdg}}\otimes_{\mathbb{Q}}L\bigr). We set \rho^{Y}_{\operatorname{ev},\lambda}:=\operatorname{rk}_{L}\mathcal{H}^{Y}_{\operatorname{ev},\lambda}, and define the Hochschild-degree-two rank \nu^{Y}_{\operatorname{ev},\lambda}:=\operatorname{rk}_{L\otimes\mathbb{C}}\Bigl((E^{Y}_{\operatorname{ev},\lambda}\otimes\mathbb{C})\cap\bigl(H^{(2)}(Y)\otimes_{\mathbb{C}}(L\otimes\mathbb{C})\bigr)\Bigr). Now suppose a finite group G acts algebraically on Y . An evaluation is G -invariant if \operatorname{ev}(Q^{g_{*}d})=\operatorname{ev}(Q^{d})\qquad\text{and}\qquad\operatorname{ev}(g^{*}\tau)=\operatorname{ev}(\tau) for every d\in\operatorname{NE}(Y) and every g\in G . For small quantum cohomology all Hodge variables are evaluated at zero, so the second condition is automatic. Definition 2.3 (Invariant rational Hodge rank) . When E^{Y}_{\operatorname{ev},\lambda} is G -stable, define \rho^{G,Y}_{\operatorname{ev},\lambda}:=\operatorname{rk}_{L}\bigl(\mathcal{H}^{Y}_{\operatorname{ev},\lambda}\bigr)^{G}. Thus \rho^{G} is the rank of the G -fixed rational Hodge classes in the evaluated generalized eigenspace. Remark 2.4 . For any G -invariant evaluation map \operatorname{ev} , the endomorphism ev(\kappa_{\tau}) is automatically G -equivariant. This is because the virtual fundamental cycle used in the definition of Gromov–Witten invariants is G -invariant. Indeed, this follows from the G -action on the moduli spaces of stable maps to Y and the G -equivariance of the perfect obstruction theory defining the virtual fundamental cycle. 3. Symmetric Verra Fourfolds Following 8 , we realize the Verra fourfold X as a hypersurface in the smooth toric fivefold \mathcal{F}:=\operatorname{Tot}(\mathcal{O}_{\mathbf{P}^{2}\times\mathbf{P}^{2}}(1,1)) . The total space \mathcal{F} is given by the Cox-coordinate quotient \mathcal{F}=\bigl((\mathbb{C}^{3}\setminus\{0\})\times(\mathbb{C}^{3}\setminus\{0\})\times\mathbb{C}\bigr)/(\mathbb{C}^{*})^{2} with coordinates (x_{i},y_{i},w) and weight matrix (1) \begin{array}[]{ccccccc|c}x_{0}&x_{1}&x_{2}&y_{0}&y_{1}&y_{2}&w&\\ \hline\cr 1&1&1&0&0&0&1&L\\ 0&0&0&1&1&1&1&M\end{array} For f_{2,2}\in H^{0}(\mathbf{P}^{2}\times\mathbf{P}^{2},\mathcal{O}(2,2)) , the equation (2) w^{2}=f_{2,2}(x,y) defines a hypersurface X\in|\mathcal{O}_{\mathcal{F}}(2,2)| , which is smooth for general f_{2,2} . Let \varpi:\mathcal{F}\to\mathbb{P}^{2}\times\mathbb{P}^{2} be the bundle projection, and set \pi:=\varpi|_{X}:X\to\mathbb{P}^{2}\times\mathbb{P}^{2}. Then \pi is the double cover branched along B=\{f_{2,2}=0\}\in|\mathcal{O}(2,2)|. Composing \pi with the two projections from \mathbb{P}^{2}\times\mathbb{P}^{2} gives two maps p_{1}:X\to\mathbb{P}^{2},\qquad p_{2}:X\to\mathbb{P}^{2}, which are the two quadric surface fibrations on X . Let