Adelic descent for continuous localizing invariants

Grigorii Konovalov · 2025

Read Adelic descent for continuous localizing invariants by Grigorii Konovalov (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-2507-20359.txt.

Introduction

Grigorii Konovalov, a solo author, proves a version of adelic descent for continuous localizing invariants in a 2025 math paper. The work builds on a known result about solid quasi-coherent sheaves on schemes of finite type over the integers. Those sheaves satisfy adelic descent, a gluing rule that rebuilds global data from local pieces at each prime. Konovalov extends that descent picture from sheaves to the stated class of invariants.

Catalog record

CREATORGrigorii Konovalov
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE122 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-20359

Browse the full library · About the Tower of Babel

Full text

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

Adelic descent for continuous localizing invariants Grigorii Konovalov Abstract . We prove a version of adelic descent for continuous localizing invariants. 1. Introduction Let X be a scheme of finite type over \mathbb{Z} . Recall from [ 1 ] , Theorem 4.2.4, that solid quasi-coherent sheaves on X satisfy adelic descent. Namely, for any tuple 0\leq k_{0}<\ldots k_{m}\leq d , where d denotes the dimension of X , we have the corresponding adelic ring \begin{split}L_{k_{0}}\ldots L_{k_{m}}\mathcal{O}_{X}\;\simeq\;\prod_{\begin{subarray}{c}\mathfrak{p}_{m}\in X\\ \mathrm{dim}\overline{\{\mathfrak{p}_{m}\}}=k_{m}\end{subarray}}{\mathrel{{\mathop{\text{\phantom{a}}}\limits}}\!\!}\>(\ldots\prod_{\begin{subarray}{c}\mathfrak{p}_{1}\in\overline{\{\mathfrak{p}_{2}\}}\\ \mathrm{dim}\overline{\{\mathfrak{p}_{1}\}}=k_{1}\end{subarray}}{\mathrel{{\mathop{\text{\phantom{a}}}\limits}}\!\!}\>(\prod_{\begin{subarray}{c}\mathfrak{p}_{0}\in\overline{\{\mathfrak{p}_{1}\}}\\ \mathrm{dim}\overline{\{\mathfrak{p}_{0}\}}=k_{0}\end{subarray}}{\mathrel{{\mathop{\text{\phantom{a}}}\limits}}\!\!}\>\mathcal{O}_{X,\mathfrak{p}_{0}})_{\mathfrak{p}_{1}}\ldots)_{\mathfrak{p}_{m}}\;\in\;\mathrm{CAlg}(\text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}}))\;,\end{split} where {\mathrel{{\mathop{\text{\phantom{a}}}\limits}}\!\!}\>(-)_{\mathfrak{p}} denotes the functor given by localizing at the prime \mathfrak{p}\in X first and then competing at it. In fact, the formation of adelic rings defines a functor \textstyle{[0<1]^{d}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\mathrm{CAlg}(\text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}}))} \textstyle{0\leq k_{0}<\ldots k_{m}\leq d\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\mathrel{\rotatebox[origin]{90.0}{$\in$}}} \textstyle{L_{k_{0}}\ldots L_{k_{m}}\mathcal{O}_{X}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\mathrel{\rotatebox[origin]{90.0}{$\in$}}} . In particular, we get a symmetric monoidal functor (1.0.0.1) \text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}})\longrightarrow\mathrel{{\mathop{\lim}\limits}}L_{k_{0}}\ldots L_{k_{m}}\mathcal{O}_{X}-\mathrm{Mod}_{\text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}})}\;, which happens to be an equivalence—see Theorem 4.2.4 of [ 1 ] . All of the categories appearing in ( 1.0.0.1 ) are compactly generated, but they are not suitable for studying localizing invariants because the category \text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}})^{\omega} satisfies a version of the Eilenberg swindle. Namely, for any fixed object P\in\text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}})^{\omega} , there exists a countable product \prod_{\mathbb{N}}P\in\text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}})^{\omega} . Nevertheless, Efimov showed ( [ 2 ] ) that any localizing invariant uniquely extends from the category of compactly generated stable categories to the category of presentable dualizable stable categories, and Clausen and Scholze defined (8th lecture in [ 3 ] , 13th lecture in [ 4 ] ) the category of nuclear modules over an analytic ring (more generally, nuclear objects in a symmetric monoidal category), which is a presentable stable dualizable category and is expected to produce correct localizing invariants in many situations including those coming from formal geometry. For a certain nice class of localizing invariants such as K-theory, that expectation was proven in [ 5 ] (Theorem 7.9) in the case of a ring completion. It therefore should be reasonable to study localizing invariants of the adeles using the nuclear modules of Clausen-Shcolze. In this paper, we prove that the diagram \textstyle{[0<1]^{d}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \textstyle{\mathrm{CAlg}(\mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}})} \textstyle{0\leq s_{0}<s_{1}<\ldots<s_{m}\leq d\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\mathrel{\rotatebox[origin]{90.0}{$\in$}}} \textstyle{\mathrm{Nuc}(L_{s_{0}}\ldots L_{s_{m}}\mathcal{O}_{X},X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}})\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} \scriptstyle{\mathrel{\rotatebox[origin]{90.0}{$\in$}}} , which we basically get from the diagram ( 1.0.0.1 ) by passing to nuclear objects, is mapped to a limit diagram by any stable localizing invariant, see Theorem 3.0.1 . We also prove that this diagram is a limit diagram in both \mathrm{CAlg}(\mathrm{Pr}^{L}_{\mathrm{st}}) and \mathrm{CAlg}(\mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}}) , see Theorem 3.4.1 . The latter in fact follows directly from Theorem 4.2.4 of [ 1 ] by passing to rigidification, but the argument presented in this paper is in a sense independent of that theorem. We note that versions of adelic descent for localizing invariants have already been recorded in the literature, see [ 6 ] and [ 7 ] . In the former paper, the author studies localizing invariants of perfect modules over the adeles and proves the adelic descent in that context. In the latter paper, the author does all the algebra within the category \text{QCoh}(X)-\mathrm{Mod}(\mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}}) and proves a version of adelic descent for localizing invariants of dualizable categories. Namely, for a category \mathcal{C}\in\text{QCoh}(X)-\mathrm{Mod}(\mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}}) , [ 7 ] constructs a diagram consisting of products iterated completed localizations of \mathcal{C} computed inside \text{QCoh}(X)-\mathrm{Mod}(\mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}}) and proves the descent statement for it, see Theorem 1.4 of loc. cit.. On the other hand, our approach is to do all the algebra inside \text{QCoh}(X_{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}{\vbox{\hbox{$\scriptscriptstyle\blacksquare$}}}}}) (which has already been done in [ 1 ] ) and see what kind statement about localizing invariants we can get as an output. At the categorical level, the difference between our version and Kim’s version ( [ 7 ] ) is in a sense similar to the difference between Clausen-Scholze’s category of nuclear modules and Efimov’s category of nuclear modules in the case of a ring completion, which is well explained in the introduction to [ 5 ] . That difference may disappear after passing to a nice localizing invariant—like it does in the case of a ring completion, see Theorem 7.9 in [ 5 ] —but that seems to need a proof. The paper is structured as follows. In the second section, we summarize some results about nuclear objects and derive some relevant corollaries about nuclear modules over the adelic rings. In the thrid section, we state and prove out main theorem. 1.1. Notation and conventions 1.1.1. In this paper, we work in the setting of \infty -categories. We adopt the usual convention that the word category means (\infty,1) -category, and all the standard categorical notions and constructions, such as functors and (co)limits, should be understood in this context. In particular, we shall typically refer to derived completion simply as completion. 1.1.2. We will denote by \Pr^{L}_{\mathrm{st}} the (\infty,1) -category of stable presentable categories and continuous (colimit preserving) functors. We will denote by \mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}} the (\infty,1) -category of stable presentable dualizable categories and strongly continuous functors. 1.1.3. By default, all limits of (stable presentable) categories are computed inside \mathrm{Pr}^{L}_{\mathrm{st}} . Limits computed inside \mathrm{Pr}^{\mathrm{dual}}_{\mathrm{st}} will be denoted by \mathrel{{\mathop{\lim}\limits}} . 1.1.4. Because we need the results of Appendix A, [ 1 ] , we work in finite type over \mathbb{Z} . It is also possible to work in finite type over a countable field using the ultrasolid vector spaces over the field as the base category. 1.1.5. We use notation of [ 1 ] for completions. It is summarized in the next four points. 1.1.6. Let Z\subseteq\spec(A) be a closed subset defined by an ideal I\subset A . We will use A^{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\wedge$}}}{\vbox{\hbox{$\scriptstyle\wedge$}}}{\vbox{\hbox{$\scriptscriptstyle\wedge$}}}{\vbox{\hbox{$\scriptscriptstyle\wedge$}}}}}_{I} to denote the (derived) completion of A at the ideal I . The same goes for the completion of a (solid) module over A . We note that the functor of completion (-)^{\mathrel{\mathchoice{\vbox{\hbox{$\scriptstyle\wedge$}}}{\vbox{\hbox{$\scriptstyle\wedge$}}}{\vbox{\hbox{$\scriptscriptstyle\wedge$}}}{\vbox{\hbox{$\scriptscriptstyle\wedge$}}}}}_{I} at the ideal I\subset A only depends on the closed subset Z\subseteq\spec(A) , but not on the ideal I cutting out Z . To emphasize that, and for other reasons, we will often call the functor of completion at I the functor of completion at Z and denote it (-)^{\overset{\sim}{\!\!\phantom{\ldots}}}_{Z} . 1.1.7. More generally, for a specialization closed subset T\subseteq\spec(A) (which can be thought of as a subset equal to the union of all its subsets which are closed in \spec(A) ), we will define a completion at T and denote it (-)^{\overset{\sim}{\!\!\phantom{\ldots}}}_{T} , see section 1.4 in [ 1 ] . 1.1.8. Similarly, for a generalization closed subset E\subset\spec(A) , we will define a completion at E and denote it (-)^{\overset{\sim}{\!\!\phantom{\ldots}}}_{E} , see section 1.7 in [ 1 ] . This will turn out to be equivalent to the completion with respect to the ultrasolid analytic ring structure on the ring \Gamma(E,\mathcal{O})\;\mathrel{{\mathop{=}\limits}}\;\mathrel{{\mathop{\mathrm{colim}}\limits}}\Gamma(U,\mathcal{O})\;, where the colimit is taken over all Zariski open subsets U\subseteq\spec(A) containing E and ordered by (opposite) inclusion. See Definition 1.1.5.2 in [ 1 ] for the definition of the ultrasolid analytic ring structure. 1.1.9. For a prime ideal \mathfrak{p}\subset A , we will denote by {\mathrel{{\mathop{\text{\phantom{a}}}\limits}}\!\!}A_{\mathfrak{p}} the local ring A_{\mathfrak{p}} completed at the maximal ideal. Similarly, for any (solid) A -module M , {\mathrel{{\mathop{\text{\phantom{a}}}\limits}}\!\!}\>M_{\mathfrak{p}} denotes the localization of M at \mathfrak{p} followed by the (derived) completion at the maximal ideal of A_{\mathfrak{p}} . 1.1.10. We use Lurie’s notation for categories of modules over an algebra as opposed to the traditional derived category notation. For example, \mathrm{Mod}_{A}

Read the complete text