A universal non-embedding theorem for 3 3 -manifolds
Giulio Belletti, Renaud Detcherry · 2026
Read A universal non-embedding theorem for 3 3 -manifolds by Giulio Belletti, Renaud Detcherry (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-2604-22387.txt.
Introduction
Belletti and Detcherry prove a universal non-embedding theorem for three-manifolds. Given two compact oriented three-manifolds N and M, with M meeting only a mild hypothesis, they find a hyperbolic three-manifold N-prime closely related to N that does not embed in M. For instance, when M is a rational homology sphere, N-prime can be chosen Y-k-equivalent to N for any k at least one. Their construction uses three-manifolds with involved Frohman-Kania-Bartoszyńska ideals, plus strong approximation for quantum representations.
Catalog record
| CREATOR | Giulio Belletti, Renaud Detcherry |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53728 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.22387
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A universal non-embedding theorem for 3 3 -manifolds Giulio Belletti, Renaud Detcherry Abstract . We prove that given two compact oriented 3 3 -manifolds N N and M , M, with M M satisfying only a mild hypothesis, there is a hyperbolic 3 3 -manifold N ′ N^{\prime} arbitrarily “closely related” to N , N, and such that N ′ N^{\prime} does not embed in M . M. For instance, as a weak version of our main theorem, if M M is a rational homology sphere then for any k ≥ 1 k\geq 1 the 3 3 -manifold N ′ N^{\prime} can be chosen to be Y k Y_{k} -equivalent to N . N. Our techniques rely on the construction of 3 3 -manifolds with complicated Frohman–Kania-Bartoszyńska ideals, using the strong approximation for SO 3 \mathrm{SO}_{3} -Witten-Reshetikhin-Turaev quantum representations of mapping class groups of surfaces. 1. Introduction Broadly speaking, the goal of this article is to study the following question: Question 1 . ( 3 -manifold embedding problem) If N and M are two compact oriented 3 -manifolds (possibly with boundary), how to check whether N can be embedded inside M? Note that in the above, we make no assumption on the boundaries of N and M, which can have arbitrarily many connected components of arbitrary genera. We will restrict ourselves to connected manifolds; while embedding non-connected manifolds into each other presents some subtleties and cannot always be reduced straightforwardly to the connected case, to obstruct embeddings it suffices to do so for a single connected component. While Question 1 has rarely been stated in this general way in the literature, there are many techniques to approach this problem, coming from every corner of 3 -dimensional topology: among others, one can get obstructions from homology, as well as the hyperbolic (or simplicial) volume or prime and JSJ decompositions. On the other hand, algorithmic techniques arising for instance from normal surface theory can give partial answers to this question. We will give a comprehensive overview of classical techniques that have been applied to the 3 -manifold embedding problem in Section 2 . Because 3 -manifold topology as well as embeddings between manifolds of the same dimension are very rigid, it is expected that embeddings between 3 -manifolds are rare. Our main results are a confirmation of this heuristic. We show that for any N and M, (with some mild hypothesis on M ) there exist 3 -manifolds N^{\prime} that are arbitrarily close to N but do not embed in M. To make sense of the notion of being arbitrarily close, we use the so-called Y_{k} -equivalence relations on 3 -manifolds, as well as the profinite topology on profinite completions of fundamental groups. For F a compact oriented surface, we will denote by I(F) its Torelli group, that is, the kernel of the natural map \mathrm{Mod}(F)\longrightarrow\mathrm{Aut}(H_{1}(F,\mathbb{Z})), where \mathrm{Mod}(F) is the mapping class group of F. For G a group, we denote by \Gamma_{k}G the k -th element of its lower central series, defined by \Gamma_{1}G=G and \Gamma_{k+1}G=[\Gamma_{k}G,G]. Definition 1.1 . For k\geq 1, we say that two 3 -manifolds N and N^{\prime} are Y_{k} -equivalent, and we write N\underset{Y_{k}}{\sim}N^{\prime} if one can obtain N^{\prime} from N by cutting N along an embedded compact oriented surface F with at most one boundary component, and regluing using a gluing map in \Gamma_{k}I(F). It is easy to see that two Y_{1} -equivalent manifolds have the same homology. The Y_{k} -equivalences (and the related J_{k} -equivalences) have been studied in depth in the literature; we refer to [ 26 ] for an overview of the subject. It is conjectured that if two 3 -manifolds are Y_{k} -equivalent for every k, then they are homeomorphic (see [ 26 , Goussarov-Habiro Conjecture] and subsequent discussion); hence Y_{k} -equivalence gives us a good notion of 3 -manifolds being close. For G a group, we denote by \widehat{G} its profinite completion, that is the profinite group defined as the inverse limit \widehat{G}=\underset{N\triangleleft G,[G:N]<\infty}{\varprojlim}G/N. Two finitely generated groups have isomorphic profinite completions if and only if they have the same set of finite quotients; in light of this, we will say that a sequence of groups G_{k} profinitely converges to G if there is a sequence of natural numbers d_{k} , with d_{k}\underset{k\rightarrow+\infty}{\longrightarrow}+\infty, such that G and G_{k} have the same finite quotients up to order d_{k}. It is generally believed that hyperbolic 3 -manifolds are profinitely rigid , meaning that if \widehat{\pi_{1}(M^{\prime})} is isomorphic to \widehat{\pi_{1}(M)} where M is hyperbolic, then M\simeq M^{\prime}. It is known that closed hyperbolic 3 -manifolds [ 21 ] , hyperbolic 3 -manifolds with toral boundary [ 36 ] and closed Seifert fibered or graph manifolds [ 33 , 34 , 35 ] are profinitely almost rigid , meaning that there are only a finite number of homeomorphism classes of M^{\prime} such that \widehat{\pi_{1}(M^{\prime})}\simeq\widehat{\pi_{1}(M)}. In this sense, having fundamental groups arbitrarily profinitely close to \pi_{1}(N) gives another good notion of being close to N. We are now able to state our main theorem: Theorem 1.2 . Let N be a compact oriented 3 -manifold, and M be a very good compact oriented 3 -manifold. Then, there exists a sequence (N_{k})_{k\geq 1} of hyperbolic compact oriented 3 -manifolds, with N_{k}\underset{Y_{k}}{\sim}N and whose fundamental groups \pi_{1}(N_{k}) profinitely converge to \pi_{1}(N), and such that none of the N_{k} embeds in M. The definition of very good 3 -manifold is postponed to Definition 3.12 ; for the moment it suffices to say that any 3 -manifold that embeds in a rational homology sphere is very good, and that we conjecture that all compact oriented 3 -manifolds are very good (see also Proposition 3.13 ). Theorem 1.2 encapsulates both notions of a 3 -manifold N being approximable by 3 -manifolds that are not embeddable in M; in fact we will package the two notions into a single, stronger equivalence relation, the (Y_{k},T_{n}) -equivalence relation, which we will introduce and study in Section 3 . In fact, our techniques allow us to give a probabilistic version of Theorem 1.2 .To state this, we use a variation of the Dunfield-Thurston [ 6 ] model of random 3 -manifolds, which we describe below. For G a finite or countable group, S a generating set of G, and \mu a probability measure on S with support S, we can take x_{1},\cdots,x_{d} independent identically distributed random variables with law \mu and define the associated random walk (g_{d})_{d\geq 0} by g_{d}=x_{1}\ldots x_{d} . Now pick a Heegaard splitting N=C\cup_{\phi}H where H is a handlebody of genus g\geq 2 , C is a compression body, \partial_{-}C\simeq\partial H\simeq F and \phi\in\mathrm{Mod}(F). Then we let N^{\prime} be the random 3 -manifold C\cup_{\phi\circ f}H, where f\in\Gamma_{k}I(F) is given by a random walk in the group \Gamma_{k}I(F) for some k\geq 1. Theorem 1.3 . Let k\geq 1 and let M be a very good 3 -manifold. Let also N=C\cup_{\phi}H be a Heegaard splitting of a compact oriented 3 -manifold of genus g\geq 2 . Given the random 3 -manifold N^{\prime}_{d}=C\cup_{\phi\circ f_{d}}H where f_{d}\in\mathrm{Mod}(F) is a random walk of length d in \Gamma_{k}I(F), we have: \underset{d\rightarrow\infty}{\liminf}\ \mathbb{P}(N^{\prime}_{d}\ \textrm{does not embed in}\ M)>0 In fact, in Theorem 7.1 , we give a version