Format | |
---|---|
BibTeX | |
MARCXML | |
TextMARC | |
MARC | |
DataCite | |
DublinCore | |
EndNote | |
NLM | |
RefWorks | |
RIS |
Files
Abstract
Let G be a (split) reductive group over Fq, T a maximal torus, T′⊂T a subtorus, and M=ZG(T′) -- a reductive subgroup of G. Let P denote a parabolic subgroup of G with Levi factor M; let U=Ru(P) (resp., Uop=Ru(Pop)); let ¯G/U (resp., ¯G/Uop) denote the affinizations of the quasi-affine homogeneous spaces G/U (resp., G/Uop). We propose a construction for a non-abelian Fourier transform S(¯G/U(Fq),C)→S(¯G/Uop(Fq),C), where S denotes a certain restricted subspace of functions on ¯G/U(Fq). We conjecture that this transform is involutive (i.e., satisfies the analogue of Fourier inversion) and provide evidence for this conjecture in concrete low-rank cases. Along the way, we will prove a finite-field Fourier transform for affine quadratic cones, and construct the complete set of finite-field normalized intertwining operators for the standard Borels in SL3.