TL;DR
This paper proves that "generic" finitely generated groups, in a specific probabilistic model, do not admit any smooth volume-preserving actions on compact manifolds. This is achieved by combining existing results on group actions on Hilbert spaces with a new criterion for groups that lack finite quotients, and it has implications for understanding the rigidity of group actions.
What Problem Does This Paper Address?
The paper tackles a question at the intersection of geometric group theory and topology: can arbitrary groups act on manifolds? Specifically, it addresses a conjecture by Gromov suggesting that "random" groups should not have smooth actions on compact manifolds. The challenge lies in defining what "random" means for groups and then proving that such groups lack the necessary properties to act non-trivially on manifolds. Existing work by Zimmer and others has explored actions of specific classes of groups (like lattices in Lie groups) and their rigidity, but a general statement about random groups was missing. The paper aims to bridge this gap by providing a concrete result for a well-defined model of random groups.
Methods
The core of the paper's argument relies on a clever combination of existing theorems and a new characterization of groups.
- Property (FHM) and (FCP): The authors introduce and utilize "Property (FHM)" (fixed point property on Hilbert manifolds) and "Property (FCP)" (fixed point property on continuum products). These properties essentially state that any action of such a group on certain geometric spaces must have a fixed point. The paper shows that if a group has property (FCP), any volume-preserving action on a compact manifold must preserve an $L^2$ metric.
- Zimmer's Theorem: A key tool is Zimmer's theorem (specifically Theorem 1.7 in [Z2]), which relates actions with discrete spectrum to actions that are measurably conjugate to actions of compact groups.
- No Finite Quotients: A crucial new element is the focus on groups with "no finite quotients." This means there are no non-trivial homomorphisms from the group to any finite group. This property, when combined with others, is shown to imply that any representation of the group on $L^2(M)$ must be trivial.
The proof of Theorem 1.1 proceeds by showing that if a group has property (FHM) and no finite quotients, then any volume-preserving action on a compact manifold must be trivial. This is done by first showing that such an action preserves an $L^2$ metric (Lemma 2.2), then using Zimmer's theorem to deduce discrete spectrum (Proposition 2.3), and finally using the "no finite quotients" property to show the action must be trivial (Proposition 2.4).
The paper also explores methods for constructing groups that satisfy these hypotheses, particularly focusing on producing torsion-free groups with no finite quotients using techniques from Ol'shanski, and exploring torsion tricks for groups with no actions by homeomorphisms.
Datasets
Not discussed in the paper. The paper deals with theoretical constructions of groups and their properties, rather than empirical evaluation on datasets.
Key Insights and Ideas
- The Power of "No Finite Quotients": The paper highlights that the absence of finite quotients is a surprisingly powerful constraint on group actions. It effectively "kills" finite-dimensional representations, which are often the building blocks of non-trivial actions.
- Connecting $L^2$ Metrics to Triviality: The insight that an action preserving an $L^2$ metric on a compact manifold, combined with property (FHM) and no finite quotients, forces the action to be trivial is central. This connects abstract geometric properties to concrete topological actions.
- Constructing "Bad" Groups: The paper provides explicit methods for constructing groups that satisfy the hypotheses, particularly focusing on torsion-free groups. This demonstrates that such groups are not mere theoretical curiosities but can be realized.
- Genericity and Randomness: The paper connects its findings to the notion of "generic" groups in a probabilistic sense, suggesting that Gromov's conjecture about random groups is well-supported by these results.
Key Results
- Theorem 1.1: A finitely generated group with no finite quotients and property (FHM) admits only trivial volume-preserving actions on compact manifolds.
- Proposition 2.3: A group with property (FCP) has any volume-preserving action on a compact manifold with discrete spectrum.
- Proposition 2.4: A group with no finite images has any discrete spectrum action that is trivial.
- The paper demonstrates methods for constructing finitely presented, torsion-free groups with no finite quotients, which then satisfy the conditions of Theorem 1.1.
Limitations and Future Work
- Definition of "Generic": The paper acknowledges that the notion of "generic" used is specific to a particular model of random groups and that there isn't a universally agreed-upon definition of a "typical" group.
- Torsion-Free Groups: While the paper produces torsion-free groups, it notes that proving certain constructions (like those in Section 3.3, Method One) hold for all torsion-free hyperbolic groups is still an open and difficult question.
- Homeomorphism vs. Diffeomorphism Actions: The paper primarily focuses on smooth (diffeomorphism) actions. While Section 4.1 touches on homeomorphism actions, a broader understanding of which groups can act by homeomorphisms on manifolds is still an area of inquiry.
- Gromov's Conjecture: The paper reduces Gromov's conjecture to proving that "generic" random groups preserve a smooth volume form (Conjecture 4.3). This remains an open problem.
- Random Groups with Finite Quotients: A key question raised is whether there's a density threshold for random groups to have no finite quotients with positive probability (Question 4.4). Answering this would yield many hyperbolic groups that are not residually finite.
Summary updated