MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Hi, I am troubled with the following question: Does there exist a finite order automorphism of a free group, $f\in Aut(F)$, such that it fixes no non trivial conjugacy class and no non trivial centralizer, i.e. $f(g)$ is not conjugate to $g$ and $f(g)\neq g^{-1}$ for any $g\in F$ ? Can we find such, so the same holds for any of its non trivial powers $f^i$?

By inspection of the finite order automorphisms of $F_2$, this cannot be an automorphism of $F_2$.

Moreover, if there exists such an automorphism is it possible to find one with as large order as we want (possibly after moving to a larger free group)?

share|cite|improve this question
What about $x\rightarrow y, y\rightarrow z, z\rightarrow w$ in $F_3 = \langle x, y, z\rangle? – Igor Rivin Apr 25 '12 at 20:46
Sorry, that was supposed to be $z\rightarrow x.$ – Igor Rivin Apr 25 '12 at 20:55
Doesn't that give a general proof that such a thing does not exist (take an arbitrary $x,$ then write the word $x f(x) f^2(x)...?$ – Igor Rivin Apr 25 '12 at 22:12
@Igor: the word you write could be trivial. For the order 3 automorphism of $<x,y>$ given by $x \to y$, $y \to y^{-1} x^{-1}$, the word $x f(x) f^2(x)$ is trivial. This is an order 3 rotation of an egg beater. – Lee Mosher Apr 25 '12 at 22:19
But upon looking at your (@Lee's) answer, I see that I was not THAT far off... – Igor Rivin Apr 26 '12 at 0:36
up vote 19 down vote accepted

No such automorphism exists. Every finite order automorphism of a finite rank free group has a nontrivial fixed conjugacy class. To see why, you can represent the free group automorphism as a simplicial automorphism $f : G \to G$ of some finite connected graph having no vertices of valence 1. Take any vertex $p \in G$ of valence $\ge 3$. Let $\gamma$ be an immersed edge path with initial oriented edge $E_0$ having initial vertex $p$, and terminal oriented edge $E_1$ having terminal vertex $f(p)$, such that the initial direction of $f(E_0)$ does not equal the terminal direction of $E_1$. The valence condition on $p$ lets you make this choice; if you made a bad choice of $\gamma$ then, since the valence of $f(p)$ is $\ge 3$, you could concatenate with some immersed path from $f(p)$ to $f(p)$ so as to change your terminal direction. Letting $k$ be the order of the simplicial automorphism $f$, it follows that $f$ fixes the conjugacy class corresponding to the immersed loop $\gamma * f(\gamma) * ... * f^{k-1}(\gamma)$.

share|cite|improve this answer
@Lee: In the second phrase "finite order" is missing? – Mark Sapir Apr 25 '12 at 21:28
Oops, yes, fixed that. – Lee Mosher Apr 25 '12 at 21:29

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.