Skip to main content
Notice removed Authoritative reference needed by CommunityBot
Bounty Ended with no winning answer by CommunityBot
deleted 97 characters in body
Source Link

What I am asking may apply to a much more general setting and I am interested in the underlying level of generality of these statements, mostly with canonical references. However I state the question how it occurred in my readings.

Let $G$ be the groups of points of a reductive group on a non-Archimedean local field, $N$ a unipotent subgroup and consider $\chi$ a character of $N$. Take the induced representation $\pi = \mathrm{Ind}_N^G(\chi)$.

1. There is a continuous decomposition of the form

$$\pi = \int_X \pi_x dx$$

where the $\pi_x$ are irreducible representations. I am searching for a more precise statement: what can be said about the space of parameters $X$, the underlying representations spaces of $\pi_x$ and the measure $dx$? What properties are known on $\pi_x$ in this specific setting (are they e.g. smooth? spherical if $\pi$ is spherical? etc.)

2. There is a spectral decomposition formula of the form, for any functions $\phi_1, \phi_2 \in \pi$, $$\langle \phi_1, \phi_2 \rangle = \int_X \langle A_x \phi_1, A_x \phi_2 \rangle dx$$ where the $A_x = V_\pi \to V_{\pi_x}$ are linear maps, more precisely intertwining operators so that $A_x \circ \pi = \pi_x \circ A_x$. Why is this true, and how far can we reconstruct $\phi$ from the knowledge of the $A_x \phi$?

I hope the question is precise enough and I will do my best to clarify it in comments otherwise. Any reference dealing with such theories is welcome, I do not know where to start. Thanks!

What I am asking may apply to a much more general setting and I am interested in the underlying level of generality of these statements, mostly with canonical references. However I state the question how it occurred in my readings.

Let $G$ be the groups of points of a reductive group on a non-Archimedean local field, $N$ a unipotent subgroup and consider $\chi$ a character of $N$. Take the induced representation $\pi = \mathrm{Ind}_N^G(\chi)$.

1. There is a continuous decomposition of the form

$$\pi = \int_X \pi_x dx$$

where the $\pi_x$ are irreducible representations. I am searching for a more precise statement: what can be said about the space of parameters $X$, the underlying representations spaces of $\pi_x$ and the measure $dx$? What properties are known on $\pi_x$ in this specific setting (are they e.g. smooth? spherical if $\pi$ is spherical? etc.)

2. There is a spectral decomposition formula of the form, for any functions $\phi_1, \phi_2 \in \pi$, $$\langle \phi_1, \phi_2 \rangle = \int_X \langle A_x \phi_1, A_x \phi_2 \rangle dx$$ where the $A_x = V_\pi \to V_{\pi_x}$ are linear maps, more precisely intertwining operators so that $A_x \circ \pi = \pi_x \circ A_x$. Why is this true, and how far can we reconstruct $\phi$ from the knowledge of the $A_x \phi$?

I hope the question is precise enough and I will do my best to clarify it in comments otherwise. Any reference dealing with such theories is welcome, I do not know where to start. Thanks!

What I am asking may apply to a much more general setting and I am interested in the underlying level of generality of these statements, mostly with canonical references. However I state the question how it occurred in my readings.

Let $G$ be the groups of points of a reductive group on a non-Archimedean local field, $N$ a unipotent subgroup and consider $\chi$ a character of $N$. Take the induced representation $\pi = \mathrm{Ind}_N^G(\chi)$.

1. There is a continuous decomposition of the form

$$\pi = \int_X \pi_x dx$$

where the $\pi_x$ are irreducible representations. I am searching for a more precise statement: what can be said about the space of parameters $X$, the underlying representations spaces of $\pi_x$ and the measure $dx$? What properties are known on $\pi_x$ in this specific setting (are they e.g. smooth? spherical if $\pi$ is spherical? etc.)

2. There is a spectral decomposition formula of the form, for any functions $\phi_1, \phi_2 \in \pi$, $$\langle \phi_1, \phi_2 \rangle = \int_X \langle A_x \phi_1, A_x \phi_2 \rangle dx$$ where the $A_x = V_\pi \to V_{\pi_x}$ are linear maps, more precisely intertwining operators so that $A_x \circ \pi = \pi_x \circ A_x$. Why is this true, and how far can we reconstruct $\phi$ from the knowledge of the $A_x \phi$?

Any reference dealing with such theories is welcome, I do not know where to start. Thanks!

Notice added Authoritative reference needed by Desiderius Severus
Bounty Started worth 50 reputation by Desiderius Severus
Title
Link

Continuous decomposition of representation Why and functional analysishow is a representation "continuously decomposable"?

Source Link

Continuous decomposition of representation and functional analysis

What I am asking may apply to a much more general setting and I am interested in the underlying level of generality of these statements, mostly with canonical references. However I state the question how it occurred in my readings.

Let $G$ be the groups of points of a reductive group on a non-Archimedean local field, $N$ a unipotent subgroup and consider $\chi$ a character of $N$. Take the induced representation $\pi = \mathrm{Ind}_N^G(\chi)$.

1. There is a continuous decomposition of the form

$$\pi = \int_X \pi_x dx$$

where the $\pi_x$ are irreducible representations. I am searching for a more precise statement: what can be said about the space of parameters $X$, the underlying representations spaces of $\pi_x$ and the measure $dx$? What properties are known on $\pi_x$ in this specific setting (are they e.g. smooth? spherical if $\pi$ is spherical? etc.)

2. There is a spectral decomposition formula of the form, for any functions $\phi_1, \phi_2 \in \pi$, $$\langle \phi_1, \phi_2 \rangle = \int_X \langle A_x \phi_1, A_x \phi_2 \rangle dx$$ where the $A_x = V_\pi \to V_{\pi_x}$ are linear maps, more precisely intertwining operators so that $A_x \circ \pi = \pi_x \circ A_x$. Why is this true, and how far can we reconstruct $\phi$ from the knowledge of the $A_x \phi$?

I hope the question is precise enough and I will do my best to clarify it in comments otherwise. Any reference dealing with such theories is welcome, I do not know where to start. Thanks!