Let $G$ be a connected reductive group defined over a field $F$. Let $T$ be a maximal torus of $G$ which is defined over $F$, $A_0$ the maximal $F$-split subtorus of $T$, and $X_0$ the cotorsion free subgroup of $X = X(T)$ corresponding to $A_0$. If $L$ is a finite Galois extension of $F$ over which $T$ splits, then $X_0$ can be described as

$$X_0 = \{ \chi \in X : \sum\limits_{\sigma \in \textrm{Gal}(L/F)} \chi^{\sigma} = 0 \}$$

Now, let $\Phi$ be the set of roots of $T$ in $G$, and $\Delta$ a base for the root system $(\langle \Phi \rangle_{\mathbb{Q}}, \Phi)$, where $\langle \Phi \rangle_{\mathbb{Q}}$ denotes the vector subspace of $\mathbb{Q} \otimes_{\mathbb{Z}} X$ spanned by $\Phi$ .

The base $\Delta$ comes from a total ordering $<$ of $\langle \Phi \rangle_{\mathbb{Q}}$, or even of $X$. Suppose that $<$ is compatible with $X_0$ in the sense that if $\chi, \chi' \in X$ with $\chi + X_0 = \chi' + X_0$ with $\chi \not\in X_0$ and positive, then $\chi'$ is also positive. Then $<$ induces a total ordering on $X/X_0$.

Suppose $\Delta$ is coming from such a total ordering. Let $\overline{\Delta}$ be the image of $\Delta$ in $X/X_0$, if necessary removing the zero element.

It may not be the case that the reduction of $\Delta - \Delta \cap X_0$ modulo $X_0$ is injective. Nevertheless, let $\overline{\Delta} = \{\gamma_1, ... , \gamma_t \}$ for distinct $\gamma_i \in X/X_0$. These are positive elements in the total ordering on $X/X_0$. Is it the case that the $\gamma_i$ are linearly independent?

I have tried to prove this without success. I was thinking that $\overline{\Delta}$ should be a base for a possibly nonreduced root system. The elements of the image of $\Phi$ in $X/X_0$, removing zero, are integer-linear combinations of $\overline{\Delta}$ which are either all positive or all negative, but the linear independence does not seem obvious.



As explained in my answer to your previous question (https://mathoverflow.net/q/255159), the map $X\to X/X_0$ is the restriction $X(T)\to X(A_0)$. In your present question, you gave those character groups compatible orders as in "Borel Tits, Groupes réductifs". Denote $_{F}\Delta$ (respectively $\Delta$) the base of the root system $\Phi (G,A_0)$ (respectively $\Phi$) corresponding to this order. Then $\overline{\Delta} =~ _{F}\Delta $ (by "Borel Tits, Groupes réductifs, 6.8"). Being a base of a root system, they are in particular linearly independent.

In summary, as you guessed, the image of $\Phi$ in $X/X_0$ (removing zero) is itself a root system (namely the root system $\Phi (G,A_0)$, also called the relative root system), and $\overline{\Delta} =~ _{F}\Delta $ is a base of it.


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.