Involution on the Graded Grothendieck Ring of Varieties and 𝔻 \mathbb{D} -Singularities

Andrew Burke · 2025

Read Involution on the Graded Grothendieck Ring of Varieties and 𝔻 \mathbb{D} -Singularities by Andrew Burke (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-2508-17587.txt.

Introduction

Andrew Burke studies an algebraic ring built from geometric shapes in a mathematics paper from two thousand twenty-five, written alone. The ring collects classes of varieties, which are shapes defined by polynomial equations, graded by dimension. Burke shows this graded ring is a quadratic extension of the part spanned by smooth, complete shapes.

That structure gives a natural symmetry operation on the ring. Burke shows it works alongside symmetric power operations, up to elements that multiply to zero. He then studies shapes that are smooth up to cut-and-paste equivalence. Applications cover ways of completing open shapes and the irrationality of certain zeta functions named for Kapranov.

Catalog record

CREATORAndrew Burke
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE102 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-17587

Browse the full library · About the Tower of Babel

Full text

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

Involution on the Graded Grothendieck Ring of Varieties and 𝔻 \mathbb{D} -Singularities Andrew Burke Abstract . We realize a graded variant K 0 ​ ( Var k dim ) K_{0}(\Var_{k}^{\dim}) of the Grothendieck ring of varieties as a quadratic extension of the subring K 0 ​ ( Var k sp ) K_{0}(\Var_{k}^{\spr}) spanned by classes of smooth and proper varieties. As such, there exists a natural involution 𝔻 \mathbb{D} on K 0 ​ ( Var k dim ) K_{0}(\Var_{k}^{\dim}) . We show that 𝔻 \mathbb{D} commutes with the symmetric power operations Sym m \Sym^{m} up to zero divisors. Moreover, we study varieties which are smooth up to cut-and-paste relations, which we call 𝔻 \mathbb{D} -singular varieties, and we give applications to compactifications of varieties and the irrationality of Kapranov zeta functions. 1. Introduction Let k be a field, assumed to be of characteristic zero. The Grothendieck ring of varieties K_{0}(\Var_{k}) is the abelian group generated by classes [X] of varieties X over k , modulo the cut-and-paste relations \displaystyle[X]=[Y]+[X-Y] for each closed subvariety Y of X . It is a ring, with product [X]\cdot[Y]=[X\times Y] . This ring was first mentioned by Grothendieck in a 1964 letter to Serre in relation to motives. It features prominently in the theory of motivic integration, introduced by Kontsevich in 1995 and developed by Denef and Loeser. Here, integrals are valued in a completed localization of K_{0}(\Var_{k}) . Despite its simple definition, the Grothendieck ring of varieties K_{0}(\Var_{k}) remains largely mysterious. It is a complicated ring with many zero divisors; see [ Poo02 ] , [ Kol03 ] , [ Bor18 ] , [ HL18 ] . Nevertheless, these zero divisors seem to witness important geometric information; see [ KS18 ] , [ LS24 ] . The Grothendieck ring was connected to birational geometry in [ LL03 ] via an isomorphism K_{0}(\Var_{k})/([\mathbb{A}^{1}_{k}])\cong\mathbb{Z}[\SB_{k}] , where \SB_{k} is the monoid of stable birational equivalence classes. This isomorphism is exploited in [ NS19 ] to prove that stable rationality specializes in smooth families. A modified argument in [ KT19 ] removes the adjective ‘stable.’ The two perspectives are unified in [ NO21 ] by considering a Grothendieck ring of varieties K_{0}(\Var_{k}^{\dim}) , graded by dimension, which can detect birational type. In this paper, we study K_{0}(\Var_{k}^{\dim}) as an object worthy of independent interest. Explicitly, it is the graded abelian group generated in degree d by classes [X]_{d} of varieties X of dimension at most d , modulo the cut-and-paste relations \displaystyle[X]_{d}=[Y]_{d}+[X-Y]_{d}, for each closed subvariety Y of X . It is a graded ring, with [X]_{d}\cdot[Y]_{e}=[X\times Y]_{d+e} . There is a graded subring K_{0}(\Var_{k}^{\spr}) of K_{0}(\Var_{k}^{\dim}) spanned by the classes [X]_{n} of smooth and proper varieties X of dimension n . Note that K_{0}(\Var_{k}^{\spr}) is a proper subring, as, for instance, it does not contain the classes [\mathop{\mathrm{Spec}}k]_{1} or [\mathbb{A}^{1}]_{1} . Nevertheless, these two classes are solutions of the integral quadratic equation \displaystyle\big(\hskip 2.84544ptx^{2}-[\mathbb{P}^{1}]_{1}\cdot x+\big([\mathbb{P}^{1}\times\mathbb{P}^{1}]_{2}-[\mathbb{P}^{2}]_{2}\big)\hskip 2.84544pt\big)\in K_{0}(\Var_{k}^{\spr})[x]. Our main structural result says that K_{0}(\Var_{k}^{\dim}) is a quadratic extension of K_{0}(\Var_{k}^{\spr}) . Theorem 1.1 . There is an isomorphism of graded rings \displaystyle K_{0}(\Var_{k}^{\dim})\cong\frac{K_{0}(\Var_{k}^{\spr})\big[\tau,\mathbb{L}\big]}{\big(\hskip 2.84544pt\tau+\mathbb{L}-[\mathbb{P}^{1}]_{1},\hskip 5.69046pt\tau\cdot\mathbb{L}-\big([\mathbb{P}^{1}\times\mathbb{P}^{1}]_{2}-[\mathbb{P}^{2}]_{2}\big)\hskip 2.84544pt\big)}. We view \tau and \mathbb{L} as indeterminates of degree one on the right side; they map to [\mathop{\mathrm{Spec}}k]_{1} and [\mathbb{A}^{1}]_{1} , respectively. Put more simply, as an abelian group, we have \displaystyle K_{0}(\Var_{k}^{\dim})\cong K_{0}(\Var_{k}^{\spr})\oplus\tau\cdot K_{0}(\Var_{k}^{\spr}). Moreover, there exists an involution \displaystyle\mathbb{D}:K_{0}(\Var_{k}^{\dim})\rightarrow K_{0}(\Var_{k}^{\dim}), which fixes K_{0}(\Var_{k}^{\spr}) pointwise and interchanges \tau=[\mathop{\mathrm{Spec}}k]_{1} and \mathbb{L}=[\mathbb{A}^{1}]_{1} , as well as projections \pi_{1},\pi_{2}:K_{0}(\Var_{k}^{\dim})\rightarrow K_{0}(\Var_{k}^{\spr}) according to the above direct sum decomposition. Though our formalism only works in the graded setting, one finds reflections of it on K_{0}(\Var_{k}) . For example, Bittner’s involution on K_{0}(\Var_{k})[\mathbb{L}^{-1}] (see [ Bit04 ] ) is an ungraded manifestation of \mathbb{D} . Our setup is more flexible because \mathbb{D} fixes the classes of smooth and proper varieties, and it does not require the inversion of \mathbb{L} , thereby losing birational data. Our formalism also yields simple proofs of the main structural results of [ LL03 ] and [ Zak19 ] on K_{0}(\Var_{k}) (see Subsection 2.4 ). For the remainder of the introduction, we discuss three main ways we use the involution \mathbb{D} and projections \pi_{1},\pi_{2} defined via Theorem 1.1 . 1.1. Symmetric Powers The graded Grothendieck ring of varieties K_{0}(\Var_{k}^{\dim}) is equipped with symmetric power operations \displaystyle\Sym^{m}:K_{0}(\Var_{k}^{\dim})\rightarrow K_{0}(\Var_{k}^{\dim}),\hskip 14.22636pt[X]_{d}\mapsto[\Sym^{m}X]_{md}, for each m\geq 0 , where \Sym^{m}X=X^{m}/S_{m} . By Proposition 4.5 , the classes \tau and \mathbb{L} behave well with respect to \Sym^{m} : \Sym^{m}(\tau\alpha)=\tau^{m}\Sym^{m}(\alpha) and \Sym^{m}(\mathbb{L}\alpha)=\mathbb{L}^{m}\Sym^{m}(\alpha) for \alpha\in K_{0}(\Var_{k}^{\dim}) . Motivated by the fact that \mathbb{D} interchanges \tau and \mathbb{L} , we conjecture that symmetric powers commute with involution. Conjecture 1.2 . \Sym^{m}\circ\mathbb{D}=\mathbb{D}\circ\Sym^{m} for each m\geq 0 One reduces to checking whether \mathbb{D}[\Sym^{m}X]=[\Sym^{m}X] for smooth and proper varieties X . This is really a question about the singularities of symmetric powers \Sym^{m}X , of which we ask a more refined version in Conjecture 5.24 . It is clearly true for curves and holds for surfaces by the formula of [ Göt01 ] . We have the following partial answer to Conjecture 1.2 . Theorem 1.3 . The operations \Sym^{m} and \mathbb{D} commute, up to zero divisors. More precisely, the image of \Sym^{m}\circ\mathbb{D}-\mathbb{D}\circ\Sym^{m} lies in \displaystyle\ker\bigg(K_{0}(\Var_{k}^{\dim})\rightarrow K_{0}(\Var_{k}^{\dim})\big[(\tau\mathbb{L})^{-1},(\mathbb{L}^{n}-\tau^{n})^{-1}_{n\geq 1}\big]\bigg). Our proof identifies the localization K_{0}(\Var_{k}^{\dim})\big[\mathbb{L}^{-1},(\mathbb{L}^{n}-\tau^{n})^{-1}_{n\geq 1}\big] with a suitable graded Grothendieck ring of stacks \Kstk . Following arguments of Ekedahl [ Eke09a ] , [ Eke09b ] , we show the class of \Sym^{m}X is equal to the class of the stack quotient [X^{m}/S_{m}] in \Kstk . We conclude by showing the class of any smooth and proper Deligne-Mumford stack is fixed by an extension of the involution \mathbb{D} to \Kstk[\tau^{-1}] . The generating function for the classes of symmetric powers of a variety X of dimension n is called the Kapranov zeta function of X : \displaystyle Z_{X}^{\dim}(t)=\sum_{m\geq 0}[\Sym^{m}X]_{mn}\cdot t^{m}\in K_{0}(\Var_{k}^{\dim})[[t]]. If k is finite, then Z_{X}^{\dim}(t) specializes under the point counting measure to the zeta function of X ; this is a rational function by a theorem of Dwork. It is natural to ask whether Z_{X}^{\dim}(t) is also rational. Dwork’s argument shows Z_{X}^{\dim}(t) is rational when X is a curve, but [ LL04 ] shows when X is a surface that Z_{X}^{\dim}(t) is rational if and only if \kappa(X)=-\infty . We extend one implication to arbitrary dimension. Theorem 1.4 . Let X be a smooth and projective variety of dimension >1 and non-negative Kodaira dimension \kappa(X)\geq 0 . Then the graded Kapranov zeta function Z_{X}^{\dim}(t)\in K_{0}(\Var_{k}^{\dim})[[t]] of X is (pointwise) irrational. In fact, the ungraded Kapranov zeta function Z_{X}(t)\in K_{0}(\Var_{k})[[t]] of X is also (pointwise) irrational, if we assume a positive answer to Conjecture 1.2 . In fact, a recently posted preprint [ She25 ] independently proves irrationality unconditionally in the ungraded setting. Our proof of Theorem 1.3 suggests that symmetric power operations may be better behaved in \Kstk . It would be interesting to know whether the stacky Kapranov zeta function Z_{X}^{\Stk}(t)\in\Kstk[\tau^{-1}][[t]] is irrational. Our argument fails in this setting. 1.2. Gluing Morphisms One can define a relative graded Grothendieck group of varieties K_{0}(\Var_{X}^{\dim}) over a base scheme X . Here, the classes are of varieties equipped with a morphism to X , which the cut-and-paste relations are required to respect. We

Read the complete text