3

reading my course I stumbled upon this fact:
Let $f\in\text{Isom}(\mathbb{H}^n)$ be an hyperbolic function (exactly 2 fixed points on $\partial\mathbb{H}^n$ and no fixed point in the interior). Suppose these fixed points are $p$ and $q$ and let $\ell$ be the unique geodesic from $p$ to $q$.
Then the geodesic is preserved under the action of $f$, i.e. $f(\ell)=\ell$.

This seems intuitive but yet I cannot prove it.

I class the only justifications were:

  • Isometries preserve angles.
  • Since $p$ and $q$ on the border are fixed so is the geodesic from one to the other.

But how can we show that if $x\in\ell$ then $f(x)\in\ell$. I cannot find a way to tackle this problem.

Edit:
What I tried since: Let's work in the upper half-space $\mathcal{U}^n$ where the distance is given by $$ \cosh^{-1}(d_{\mathcal{U}^n}(x,y))=1+\frac{\|x-y\|^2}{2\cdot x_n\cdot y_n}\qquad x,y\in\mathcal{U}^n $$ Then since $f$ has to be such that $f(0)=0$ and $f(\infty)=\infty$ and has to preserve the distance, we can (can we ?) say that $f$ is of the form: $$ f(z)=\lambda z $$

If this correct then $x\in\ell\implies x=a\cdot e_n$.
So $f(x)=\lambda a \cdot e_n \in\ell$. cqfd?

Edit n°2:
I found something else. Assuming the same conditions that in previous edit. Then since $f(\infty)=\infty$, $f\in\text{Stab}(\infty)$ and we showed that $\text{Stab}(\infty)\simeq\text{Sim}(\mathbb{R}^{n-1})$.
So we can in a way say that $f$ is in the similarity group of $\mathbb{R}^{n-1}$ but every $\psi\in\text{Sim}(\mathbb{R}^{n-1})$ can be uniquely written as $$ \psi=\lambda\cdot g+ v\qquad\lambda\in\mathbb{R}_{>0},g\in O(n-1),v\in\mathbb{R}^{n-1} $$ Since $f(0)=0$ then here $v=0$ and since it preserves $+\infty$ we can argue that $g=id$. Hence $f(z)=\lambda z$ ?

I'm sure there's an easier way out there, I would be thankfull if someone could point it out to me :)

Leo
  • 492
  • 2
    (1) A geodesic is determined by any two distinct points on it. (2) Hyperbolic isometries map geodesics to geodesics. :) – paul garrett Aug 28 '21 at 18:38
  • @paulgarrett Ohhh yeah, should have thought about this sooner hahaha thank you very much !! :) – Leo Aug 28 '21 at 20:22
  • It is nice that sometimes resolutions are very simple. It doesn't always have to be tortuous. :) To a great degree, "the trick" in math is looking at things from a great many angles until one discovers a viewpoint from which a thing "is trivial". It is disingenuous to say "it's trivial", of course, when the discovery of the potential "triviality" is highly non-trivial. But there is a fashion... :) – paul garrett Aug 29 '21 at 01:23
  • 1
    There is also often a pseudo-zen principle that "doing the obvious thing" fails... (Sure, sometimes things are more straightforward, and the obvious thing succeeds. How to know which it is?) But/and then the sorta insidious thing is that at a professional level "the obvious thing" is revised to be the (not-really-obvious) thing that did succeed. This dynamic can be used by "teachers" to be mean to students, but it also can be quite interesting... – paul garrett Aug 29 '21 at 01:26

0 Answers0