Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $(J,\pi)$ be a cuspidal type in $SL(2,F)$, $F$ is a non-Arch. local field and let $I$ be the Iwahori subgroup of $SL(2,F)$.

Is there any possibility that $J\subset I$ or even a subgroup?

share|improve this question

4 Answers 4

up vote 5 down vote accepted

Here is a novel, explicit answer to your question :

Benedict Gross and Mark Reeder have recently discovered a family of supercuspidal representations, called "simple supercuspidal representations", of simply connected, split, almost simple, reductive groups. These representations are examples of what you are looking for.

Here's an example of a simple supercuspidal representation, and the general case is similar. Consider $G(\mathbb{Q}_2) = SL(2,\mathbb{Q}_2)$. Let $I$ be an Iwahori subgroup of $SL(2,\mathbb{Q}_2)$, and let $I^+$ be the pro-unipotent radical of $I$. For example, we can take

$$I = \begin{bmatrix} \mathfrak{o} & \mathfrak{o}\\\ \mathfrak{p} & \mathfrak{o} \end{bmatrix}$$

$$I^+ = \begin{bmatrix} 1 + \mathfrak{p} & \mathfrak{o}\\\ \mathfrak{p} & 1 + \mathfrak{p} \end{bmatrix}$$

Let $\chi$ be an "affine generic character" of $I^+$. Since we are dealing with $SL(2,\mathbb{Q}_2)$, this just means the following : Let $\eta$ be the character of $\mathbb{Z}_2$ given by $$\eta : \mathbb{Z}_2 \rightarrow \mathbb{C}^*$$ $$2 \mathbb{Z}_2 \mapsto 1$$ $$1 + 2 \mathbb{Z}_2 \mapsto -1$$ Then define the character

$$\chi : I^+ \rightarrow \mathbb{C}^*$$ $$\begin{bmatrix} a & b\\\ 2c & d \end{bmatrix} \mapsto \eta(b) \eta(c)$$

Then $Ind_{I^+}^{SL(2,\mathbb{Q}_2)} \chi$ is a supercuspidal representation (which Gross and Reeder call a "simple supercuspidal representation", since it is so "simple" to define). The general situation is similar : If $G$ is simply connected, split, almost simple, $F$ is a $p$-adic field, $I \subset G(F)$ is an Iwahori subgroup, $I^+ \subset I$ is the pro-unipotent radical, and $\chi : Z I^+ \rightarrow \mathbb{C}^*$ is an "affine generic character" (where $Z$ is the center of $G(F)$), then $Ind_{Z I^+}^{G(F)} \chi$ is a supercuspidal representation (called a "simple supercuspidal representation"). For more details, check out Section 9 of https://www2.bc.edu/~reederma/AdjointGamma.pdf

share|improve this answer
    
In your example with ${\mathbb Q}_2$, I think you need either to extend your character to the center $\pm I_2$ first, or replace ${\mathbb Q}_2$ by ${\mathbb F}_2 ((X))$. –  Paul Broussous Jul 6 '11 at 19:08
    
In this case (unless I'm mistaken) I don't think I need to, since $-1 \in 1 + \mathfrak{p}$, since $-1 = 1 + 2 + 2^2 + 2^3 + 2^4 + ...$ Thus, the center of $SL(2, \mathbb{Q}_2)$ is already in $I^+$, so $Z I^+ = I^+$. –  Moshe Adrian Jul 6 '11 at 19:46
    
Sorry, you're totally right ! –  Paul Broussous Jul 7 '11 at 8:05

Yes this is indeed possible. You may read the proof of theorem (2.1.) of

Bushnell and Kutzko, The admissible dual of ${\rm SL}(N)$ - I Ann. Sci. École Norm. Sup. (4) 26 (1993), no. 2,

where the supercuspidal representation are described as induced representations from a subgroup wich can always be taken of the form $U({\mathfrak A })\cap {\rm SL}(N)$, where $\mathfrak A$ is a principal order in ${\rm M}(N,F)$.

But of course, you may replace the group $J$ by a maximal compact subgroup by inducing the type to a maximal compact subgroup $K$ containing $J$.

share|improve this answer

While possible it only can happen if the quadratic extension associated to the type is ramified.

share|improve this answer

The Iwahori is a type for the trivial representation on the Levi for $SL_2(F)$. Moreover all the types for $SL_2(F)$ of non supercuspidal bernstein components are subgroups of the Iwahori. There is a paper by Kutzko where he describes all the types for $SL_2(F)$.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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