A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal

summary

The gist

A finitely generated extension of Caprace’s example of a non-inner-amenable group is also not properly proximal This note records a finitely generated extension of Caprace’s example of a…

In short

The paper constructs a finitely generated group G that is not of property (T), not inner amenable, and not properly proximal. It uses an explicit ICC subgroup of GL4 to demonstrate the existence of such a group, showing that the failure of proper proximality follows from the co-amenability of its normal subgroup.

Key concepts

Non-Inner Amenability
A group is not inner amenable if every way you try to define an average (a mean) over the group by looking at its conjugates results in a trivial average, meaning the only possible mean is the one that assigns all weight to a single element (the identity). This property shows the group lacks certain types of averaging behaviors.
Proper Proximality
A group is properly proximal if it exhibits strong mixing properties related to its subgroups. The paper proves a group is *not* properly proximal by showing that a specific normal subgroup within it is co-amenable, which forces the larger group G to also lack this strong mixing behavior.
Co-Amenability
A subgroup L of G is co-amenable if its complement in G (the quotient group) has a simple structure, specifically being isomorphic to the infinite cyclic group Z. This structural relationship between the subgroup and the larger group is key to proving that G itself cannot be properly proximal.

Terminology used across episodes

This episode discusses

The paper

A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal · Read on arXiv

Institute for Advanced Study in Mathematics · Harbin Institute of Technology

Transcript

Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.

Ada: We're looking at the paper: "A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal".

Title and authors: Ada: Now, let's look at the actual construction described in "A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal". The authors set up a group G defined as t to the power k times a vector v, plus some matrices A and an element s.

Emmy: They describe it as having this structure: G equals t to the power k times a vector v, plus some matrices A and an element s. They build this group by adjoining a scalar dilation to a subgroup Γ which is F2

t, t minus one: three with SL3 over F2 t minus one as its structure

DTDW20, Section five <ref:2609.30158#pg1>.C: .

Kolya: So they are taking something known and doing some construction on top of it to get this group G that has all these properties. What's the payoff there?

Ada: The payoff is that every element in G has a very specific form, t to the power v times some matrices A and an element s with an exponent k. They show how matrix multiplication works out for these elements, which verifies closure under products and inverses

BIP21, Theorem four point three(iii): <ref:2609.30158#pg1>.

Emmy: Specifically, they establish a split exact sequence where G is built by taking this normal subgroup Γ and adjoining the element s to get G Gamma with an abelian quotient isomorphic to Z.

Kolya: That split exact sequence is key because it lets them use the structure of the quotient to argue about co-amenability, right?

Ada: Exactly. They use that fact that a normal subgroup with an amenable quotient is co-amenable

BIP21, Proposition four point one zero(two): , and since G/Gamma is isomorphic to Z, it means G has a normal subgroup Γ that is co-amenable in G BIP21 <ref:2609.30158#pg2,amenable quotient is co-amenable>.

Emmy: And this directly leads to the conclusion that because of this relationship, the group G cannot be properly proximal

BIP21, Proposition four point one zero(two): <ref:2609.30158#pg2,properly proximal BIP21, Proposition 4.10(2>.

Kolya: So they're showing that the structure they built—this specific extension—is inherently resistant to being properly proximal because of how it relates to its quotient.

Ada: It really highlights that you can build groups with very concrete algebraic descriptions, involving matrices and scalar dilations, and still end up with a group that lacks property (T) and fails proper proximality.

The paper's summary: Emmy: Now for the part where the authors discuss how their work improves on prior ideas. They focus on the connection between proper proximality and co-amenability, which is a major result they build upon BIP21.

Ada: They use that established fact—if H is properly proximal and L is a co-amenable subgroup of H, then L must also be properly proximal

BIP21, Proposition four point one zero(two): —to show that their normal subgroup Gamma in G inherits this property because G/Gamma is abelian <ref:2609.30158#pg2,properly proximal BIP21, Proposition 4.10(2>.

Kolya: So they are essentially showing that the structure they constructed respects the rules governing proper proximality under co-amenability, even when you add a dilation extension to it.

Ada: That's right. And they also use a standard elementary consequence of property (T) groups acting on sets, which is Lemma two point one BIP21 <ref:2609.30158#pg2,also use a standard elementary consequence of property (T>. This lemma helps them show that any conjugation-invariant mean on G must be the point mass at the identity BIP21.

Emmy: That's how they rigorously prove non-inner amenability. They show that if you have an invariant mean, it has to concentrate all its measure on a finite set of elements, and then they show that this contradicts the idea that it could be atomless.

Kolya: So the improvement isn't just finding a new group; it’s using these existing structural theorems in a new way to nail down the non-inner amenability proof.

Ada: It is. And they also use their knowledge of property (T) groups to rule out the ICC property by contradiction, which was handled by showing that assuming an element has a finite conjugacy class leads to a mean that isn't the point mass at the identity

BHV08, Proposition four point two: <ref:2609.30158#pg1>.

Emmy: So they are using those deep tools—the property (T) group machinery—to rule out these properties one by one, which is a very thorough approach.

The paper's improvements: Ada: So to wrap things up on "A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal", the main implication is that they’ve constructed a concrete example of this type of group using an explicit ICC subgroup of GL4 (F2 (t)) that has no property (T).

Emmy: This confirms the existence of these groups, and it shows that they are not just theoretical curiosities; they are actual algebraic objects we can describe precisely.

Kolya: For someone who only listens to this show, it means there's a class of groups where you can build them up algebraically and still have them exhibiting non-inner amenability and failing proper proximality.

Ada: Right. The paper demonstrates that the group G is finitely generated, not of property (T), ICC, but it fails both inner amenability and proper proximality.

Emmy: It’s a nice piece of structural knowledge about how these properties interrelate in linear groups; it shows that you can find instances where these conditions don't all go along together.

Ada: So we're done with this paper, "A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal". It’s a fascinating demonstration of the complexity that can arise even in relatively simple settings.

Emmy: Indeed. We'll be back next time to look at some other interesting papers from arXiv.

Conclusion: Ada: So we’ve just talked about "A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal". Basically, they built a very specific group G using some matrix multiplications and scalar dilations, and they proved it has all these properties at the same time.

Emmy: It’s a nice piece of structural work because it shows you can construct these groups in a concrete way, not just as abstract symbols. The key was that they used a dilation extension on this subgroup Gamma, which is co-amenable in G because the quotient G divided by Gamma is just the integers.

Kolya: From my side, what’s interesting is how they handle the non-inner amenability part. They show that every invariant mean on G must be concentrated at a single point, which really rules out any kind of "spread out" averaging mechanism in these groups.

Ada: Exactly. And then they use that to rule out the proper proximality part. Because Gamma is co-amenable in G, and there’s a theorem linking co-amenability to proper proximality that applies here, G can't be properly proximal either.

Emmy: It’s a neat way to show how different structural properties—like being non-inner amenable and not properly proximal—are linked by the underlying algebraic structure of the group itself.

Kolya: The caveat is that they rely on some established results about property (T) groups for that contradiction, which is fine, but it does tie the result to that specific area of mathematics.

Ada: Yeah, so we’ve seen how these specific constructions can yield groups without property (T) while still having these very strong algebraic features like non-inner amenability. It proves there are many ways to build complex structures in linear groups that don't behave the way we expect them to.

Emmy: It makes you think about where the boundaries are on what we can construct when we start modifying known groups with extensions and dilations.

Kolya: Definitely a good example of how specific choices in construction—like this one involving the F2 group and the SL3 structure—can dictate all these different behaviors.

Ada: Right, so that’s "A finitely generated Linear group without property (T) that is neither inner amenable nor properly proximal". It’s a solid example of how concrete algebra can lead to non-trivial results about group dynamics.

Emmy: Next up, we’ll be looking at the Rank-One HCIZ Integrals and Multiplicative Convolutions paper. We study how spectral averaging interacts with the logarithm in large dimensional rank one...

More episodes

← Home