1.73 (1965)

Solved

Are there only finitely many conjugacy classes of maximal periodic subgroups in a finitely generated linear group over the integers?

Progress

Not always. The extension $H$ of the free group on free generators $a, b$ by the automorphism $\varphi : a \to a^{-1}$, $b \to b$ is a linear group over the integers. For every $n \in \mathbb{Z}$ the element $c_n = \varphi b^{-n} a b^n$ has order 2 and $C_H(c_n) = \langle c_n \rangle$. Let the dash denote an isomorphism of $H$ onto its copy $H'$. As a counterexample one can take the subgroup $G \leqslant H \times H'$ generated by the elements $a, \varphi, a'\varphi'$, $b b'$. Indeed, $G$ contains the subgroups $T_n = \langle c_0, c'_n \rangle, n \in \mathbb{Z}$, which are maximal periodic (even in $H \times H'$). If $T_n$ and $T_m$ are conjugate by an element $x y' \in G$ (where $x \in H$ and $y' \in H'$), then $c_0^x = c_0$ and $c'_n = c_m^{y'}$, whence $x \in \langle c_0 \rangle$, $y' \in b^{m-n} \langle c_m \rangle$. Since $xy' \in G$, the sums of the exponents at the occurrences of $b$ in $x$ and $y$ (in any expression) must coincide; hence $m = n$. (Yu. I. Merzlyakov, 1973.)

Comments

All comments are the responsibility of the user. Comments appearing on this page are not verified for correctness. Please keep posts mathematical and on topic. If you want to submit a proof (or a partial proof), please use the dedicated proof submission form rather than posting it in the comments.
Order by newest first or oldest first.

No comments yet. Be the first to comment.

Proof claims

Proof claims are the responsibility of the submitter. Appearance here does not mean the claim has been checked for mathematical correctness. Moderators only screen for spam, abuse, and obviously low-effort submissions.

No proof claims yet.