In section 4.1.3 of Kleiner and Leeb's paper on symmetric spaces and euclidean buildings, there's a result about pairs of rays from the same point initially spanning a flat triangle (or being degenerate).

One way to look view this result is that the following function of distances at $t$ between the rays is initially linear $$d(\alpha(t),\beta(t))\text{ where }\alpha,\beta\text{ are rays with }\alpha(0)=\beta(0)$$ Furthermore, that initial linear part of this function must have a slope according to one of the finitely many angles in what they'd call $D(\theta(\alpha),\theta(\beta))$.

My question:

Is $d(\alpha(t),\beta(t))$ piecewise linear for all $t\in[0,\infty)$?

If so, do the slopes involved correspond to the elements of $D(\theta(\alpha),\theta(\beta))$?

Another way of asking this:

Their argument shows that $\alpha,\beta$ initially span a flat triangle, for example $\alpha(0), \alpha(t_0), \beta(t_0)$.

Can you pick such a $t_0$ and then keep showing that there is $t_1 > t_0$ where you get a flat quadrilateral $\alpha(t_0), \alpha(t_1), \beta(t_1), \beta(t_0)$?


Pick any time $t_0$ where $\alpha$ and $\beta$ are, locally around $t_0$, each in a fixed chamber : $\alpha(t_0) \in C$ and $\beta(t_0) \in C'$. Then if you consider an apartment $A$ containing $C$ and $C'$, for $t$ close to $t_0$, $d(\alpha(t),\beta(t))$ is just the Euclidean distance between the two rays inside $A$, so in particular it is an affine function.

Furthermore, these small rays inside $A$ also have $\theta(\alpha)$ and $\theta(\beta)$ as asymptotic directions, hence the slope of the distance function also lies in $D(\theta(\alpha),\theta(\beta))$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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