3
$\begingroup$

Let $F$ be a hyperbolic surface and $p\in F$ be a point. Consider $\pi_1(F,p)$, the fundamental group of $F$ with base point $p$. Let $x,y\in \pi_1(F,p)$ and $z$ be a simple closed curve in $F$ such that $i(z,x*y)\neq 0$ where $i(~,~)$ denotes the geometric intersection number (i.e. minimum number of possible intersection points between the free homotopy classes of each curves).

Q) Is $i(z,[(x*y)^n*(y*x)])\neq 0$ for all $n\in \mathbb{N}$?

My intuition is, as $i(z,x*y)\neq 0$, powers of the same curve intersects $z$ non-trivially and as (pictorially) there is no back tracking, the answer should be true. But I am not sure whether this could give a proof.

(P.S. As Ian Agol pointed out in the comments, this is not true for non-orientable case. Therefore assume $F$ to be orientable.)

A more general question is the following.

Let $z$ be a simple closed curve and $x,y\in \pi_1(F,p)$ such that $i(z,x*y)= 0$ and $y$ is not conjugate to $x^{-1}$. Then $i(z,x)=i(z,y)=0$.

But this general statement is false. A counter example is given in the following picture.

enter image description here

$\endgroup$
4
  • $\begingroup$ Does $*$ denote concatenation? $\endgroup$
    – HJRW
    Commented Dec 28, 2016 at 14:17
  • $\begingroup$ @HJRW Yes. It is the multiplication of the fundamental group. $\endgroup$
    – Cusp
    Commented Dec 28, 2016 at 14:21
  • 1
    $\begingroup$ This is false for a non-orientable surface, but might be true for orientable surfaces. The point is that the fundamental group of a Klein bottle is given by $\langle x, y | x^2y^2=1\rangle$. This relator is conjugate to $(x*y)*y*x$. So if one takes a punctured Klein bottle, and take $x,y$ to be the appropriate closed curves, then $x*y*y*x$ will be homotopic to the boundary. Then any other closed curve (representing $z$) intersecting $x*y$ will not intersect $x*y*y*x$, the peripheral curve. $\endgroup$
    – Ian Agol
    Commented Dec 30, 2016 at 22:31
  • $\begingroup$ @IanAgol Thanks for the remark. But I am mostly interested in orientable surfaces. $\endgroup$
    – Cusp
    Commented Dec 31, 2016 at 1:33

0

You must log in to answer this question.