In some paper, it is written that none of Siegel set of $GL_{n}$ is not contained in any Siegel set of $GL_{n+1}$. I don't understand this because I think it should hold.

To explain my question more precisely, let me first introduce some notations.

Let $G'=GL_n \subset GL_{n+1}$ be general linear groups defined over $\mathbb{Q}$. Let $P_0=U_0M_0$ and $P_0'=U_0'M_0'$ be the standard parabolic subgroups of $G$ and $G'$, respectively. Let $T_0' \subset T_0$ be the maximal split torus of $G'$ and $G$ and $A_0$ and $A_0'$ be the identity component of $T_0(\mathbb{R})$ and $T_0'(\mathbb{R})$.

Let $K \subset K'$ be the standard maximal compact subgroup of $G$ and $G'$. Then by the Iwasawa decomposition, $G=P_0K$ and $G'=P_0'K'$.

Let $\mathfrak{a}_0^*$ be the $\mathbb{R}$-vector space spanned by $\mathbb{Q}$-rational characters of $M_0$ and $\mathfrak{a}_0$ its dual space. Let $\Delta_0$ be the set of simple roots of $\mathfrak{a}_0$.

Let $H:G \mathbb(A_\mathbb{Q}) \to \mathfrak{a}_0$ be the Harish-Chandra function.

For sufficiently negative $t_0 \in \mathfrak{a}_0$, let $A_0(t_0)=\{a \in A_0 \mid \langle \alpha,H(a)-t_0\rangle > 0 \text{ for } \alpha \in \Delta_0\}$.

Then a Siegel set of $G'$ is a set $w'A_0'(t_0')K'$, where $w'$ is a compact subset in $U_0'\mathbb(A_\mathbb{Q})(M_0')^1\mathbb(A_\mathbb{Q})$ and $t_0' \in \mathfrak{a}_0'$. Then for a Siegel set $w'A_0'(t_0')K' \subset G$, can we take a compact subset $w \subset U_0\mathbb(A_\mathbb{Q})(M_0)^1\mathbb(A_\mathbb{Q})$ and $t \in \mathfrak{a}_0$ such that $w'A_0'(t_0')K' \in wA_0(t_0)K$?

I think this is always possible because $P_0' \subset P_0$ and $K' \subset K$. But some paper says that it is not possible. Is there something I am confusing or missing?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.