N
Is it possible to get qualitative data about the restriction of a free group automorphism to a characteristic subgroup, given assumptions on the initial automorphism? The setting is as follows: $F$ is a finitely generated free group, and $\varphi\in \operatorname{Aut}(F)$ is a free group automorphism with some properties. Given a finite index, characteristic subgroup $N \lhd F$, can we say that the restriction of $\varphi$ on $N$ has any of the nice properties $\varphi$ might have? I know that…
N
Let $G_\Gamma$ be a graph product of cyclic groups, and let $A_\Gamma$ be the right-angled Artin group associated to the same graph $\Gamma$. Denote by $\operatorname{PSA}(A_\Gamma)$ and $\operatorname{PSA}(G_\Gamma)$ the pure symmetric automorphism groups, i.e. the subgroups generated by partial conjugations. Recall that a partial conjugation is an automorphism $\pi_K^v$, where $v\in V(\Gamma)$ and $K$ is a connected component of $$ \Gamma\setminus\operatorname{st}(v), $$ which conjugates…
N
Let $\newcommand\G{(\text{Sym}(\mathbb{N}),\circ)}\G$ be the group of all bijections $\varphi:\mathbb{N}\to\mathbb{N}$ with composition as group operation. Is there a group homomorphism $f:(\mathbb{R},+) \to \G$ such that the kernel $\text{ker}(f)$ is at most countable?
N
I will teach a course entitled Foundations of Algebra for undergraduate students in the coming term for Computer Science students and I have a little bit looked for some concrete examples of applications of algebra in CS. I found the following very interesting weblog https://blog.computationalcomplexity.org/2008/11/barringtons-theorem.html in which I saw for the first time a Theorem named Barrington which is one in the area of computational complexity theory. I do not want to state the Theorem…
N
I'm working with infinite, connected, vertex-transitive graphs, and I'm interested in polynomial growth as a property. Rather than assuming polynomial growth directly (a strong global condition), I'd like to understand what weaker or more local conditions are known to imply it. Question. For an infinite, connected, vertex-transitive graph, are there known local or geometric conditions (e.g. curvature-type conditions, isoperimetric/Følner conditions, or conditions on the local structure around a…
N
Let G=(V,E) be an infinite, connected, vertex-transitive graph with polynomial growth, i.e. |B(r)| ≤ C r^D for some constants C, D < ∞, where B(r) is the ball of radius r in the graph metric. Question. Must such a graph be quasi-isometric to a Cayley graph of a finitely generated group? Equivalently, are there infinite, connected, vertex-transitive graphs of polynomial growth that are not quasi-isometric to any Cayley graph? Context / what I already know. I'm aware that vertex-transitivity…
N
Let $G = A[\widetilde{A}_2]$ be the affine Artin group of type $\widetilde{A}_2$ generated by $\{a, b, c\}$. That is, the group has relations $aba = bab$, $aca = cac$, and $bcb = cbc$. Consider $\varphi$ to be the automorphism of $G$ defined by $\varphi(a) = b$, $\varphi(b) = c$, and $\varphi(c) = a$, and let $\alpha \in G$ be such that the relation $$\alpha \varphi(\alpha) \alpha = \varphi(\alpha) \alpha \varphi(\alpha)$$ holds. Let $H_\alpha$ denote the subgroup generated by $\{\alpha,…
N
In Cohomology and extensions of von Neumann algebras I (Publ. RIMS 16 (1980)), Sutherland considers a Borel map $α: G → Aut(M)$ and a unitary $α$-2-cocycle $w$, and shows that after stabilizing, (after passing from $M$ to $M ⊗ B(L²(G))$), the twist $w ⊗ 1$ becomes a coboundary. The shape of this is familiar: $B(L²(G))$ contains $L^∞(G)$, which as a $G$-module under translation is the coinduced module $Coind_e^G(ℂ)$, and Shapiro's lemma gives $H^n(G, Coind_e^G(A)) ≅ H^n(\{e\}, A) = 0$ for $n ≥…
N
Is $H^n_M(\mathbb{R}, \mathbb{Z}) = 0$ for $n \geq 1$, with $\mathbb{Z}$ discrete and trivially acted on? $H^n_M(G, A)$ denotes Moore cohomology, the cohomology of the complex of Borel cochains $G^n \to A$ for a locally compact second countable $G$ acting on a Polish $G$-module $A$. I expect the answer is yes, from an identification of Moore cohomology with discrete coefficients with the singular cohomology of the classifying space, together with $B\mathbb{R}$ being contractible, but I have not…
N
Is every finite subgroup of an hyperbolic Kac-Moody group over $\mathbb{C}$ contained in a finite (but nonzero)-dimensional subgroup?
N
Let $q=p^f$, and let $n\ge2$. I am studying the exponent of $\mathrm{SL}_n(\mathbb F_q)$ as part of a more general problem concerning linear groups over finite local rings. For $\mathrm{GL}_n(\mathbb F_q)$, I use the standard formula $$\exp\bigl(\mathrm{GL}_n(\mathbb F_q)\bigr) = p^{\lceil\log_p n\rceil} \operatorname{lcm}_{1\le d\le n}(q^d-1).$$ Equivalently, $$\exp\bigl(\mathrm{GL}_n(\mathbb F_q)\bigr) = p^{\lceil\log_p n\rceil} \prod_{d=1}^n\Phi_d(q),$$ where $\Phi_d$ denotes the $d$-th…
N
My question is the following: Consider a one relator group $G_w=\left\langle F_r|w\right\rangle$ where $w$ is a primitive word in $F_r^2=\left\langle x^2\mid x\in F_r\right\rangle$ (a primitive word is one which is contained within some basis). Is $G_w^2=\left\langle x^2\mid x\in G_w\right\rangle$ is necessarily free? A colleague of mine recently presented a proof using the above as a lemma. However the proof of that lemma erroneously assumed that $G_w^2\cong F_r^2/\langle\langle…