C5

In order to study the group of (outer) automorphisms of

any group G by geometric methods one needs a well-behaved "outer

space" with an interesting action of Out(G). If G is free abelian, the

classic symmetric space SL(n,R)/SO(n) serves this role, and if G is

free non-abelian an appropriate outer space was introduced in the

1980's. I will recall these constructions and then introduce joint

work with Ruth Charney on constructing an outer space for any

right-angled Artin group.

In 1878, Jordan showed that if $G$ is a finite group of complex $n \times n$ matrices, then $G$ has a normal subgroup whose index in $G$ is bounded by a function of $n$ alone. He showed only existence, and early actual bounds on this index were far from optimal. In 1985, Weisfeiler used the classification of finite simple groups to obtain far better bounds, but his work remained incomplete when he disappeared. About eight years ago, I obtained the optimal bounds, and this work has now been extended to subgroups of all (complex) classical groups. I will discuss this topic at a “colloquium” level – i.e., only a rudimentary knowledge of finite group theory will be assumed.

Let $A$ and $B$ be boundaries of CAT(0) spaces. A function $f:A \to B$ is called a {\em boundary isomorphism} if $f$ is a homeomorphism in the visual topology and

$f$ is an isometry in the Tits metric. A compact metrizable space $Y$ is said to be {\em Tits rigid}, if for any two CAT(0) group boundaries $Z_1$ and $Z_2$ homeomorphic to $Y$, $Z_1$ is boundary isomorphic to $Z_2$.

We prove that the join of two Cantor sets and its suspension are Tits rigid.

A virtual endomorphism of a group $G$ is a homomorphism $f : H \rightarrow G$ where $H$

is a subgroup of $G$ of fi…nite index $m$: A recursive construction using $f$ produces a

so called state-closed (or, self-similar in dynamical terms) representation of $G$ on

a 1-rooted regular $m$-ary tree. The kernel of this representation is the $f$-core $(H)$;

i.e., the maximal subgroup $K$ of $H$ which is both normal in G and is f-invariant.

Examples of state-closed groups are the Grigorchuk 2-group and the Gupta-

Sidki $p$-groups in their natural representations on rooted trees. The affine group

$Z^n \rtimes GL(n;Z)$ as well as the free group $F_3$ in three generators admit state-closed

representations. Yet another example is the free nilpotent group $G = F (c; d)$ of

class c, freely generated by $x_i (1\leq i \leq d)$: let $H = \langle x_i^n | \

(1 \leq i \leq d) \rangle$ where $n$ is a

fi…xed integer greater than 1 and $f$ the extension of the map $x^n_i

\rightarrow x_i$ $(1 \leq i \leq d)$.

We will discuss state-closed representations of general abelian groups and of

…nitely generated torsion-free nilpotent groups.

Let G be a finite group, p a prime and S a Sylow p-subgroup. The group G

is called p-nilpotent if S has a normal complement N in G, that is, G is

the semidirect product between S and N. The notion of p-nilpotency plays

an important role in finite group theory. For instance, Thompson's

criterion for p-nilpotency leads to the important structural result that

finite groups with fixed-point-free automorphisms are nilpotent.

By a classical result of Tate one can detect p-nilpotency using mod p

cohomology in dimension 1: the group G is p-nilpotent if and only if the

restriction map in cohomology from G to S is an isomorphism in dimension

1. In this talk we will discuss cohomological criteria for p-nilpotency by

Tate, and Atiyah/Quillen (using high-dimensional cohomology) from the

1960s and 1970s. Finally, we will discuss how one can extend Tate's

result to study p-solvable and more general finite groups.

A finitely generated group has the Haagerup property if it admits a proper isometric action on a Hilbert space. It was a long open question whether Haagerup property is a quasi-isometry invariant. The negative answer was recently given by Mathieu Carette, who constructed two quasi-isometric generalized Baumslag-Solitar groups, one with the Haagerup property, the other not. Elaborating on these examples, we proved (jointly with S. Arnt and T. Pillon) that the equivariant Hilbert compression is not a quasi-isometry invariant. The talk will be devoted to describing Carette's examples.