Relation of orbital integrals on \(SO(5)\) and \(PGL(2)\) (Q1269758): Difference between revisions

Let \(G\) be the special orthogonal group defined over a local field \(F\) by an antidiagonal form with upper triangular minimal parabolic subgroup, and let \(H\) be the group \(PGL(2)\) over \(F\). Let \(f\) and \(f'\) be spherical functions on \(G\) and \(H\), respectively, whose Satake transforms are equal, and let \(\Psi (\alpha, f)\) and \(\Psi' (\alpha, f')\) be the associated Fourier orbital integrals corresponding to an element \(\alpha \in F^\times\). In this paper the author proves that these Fourier orbital integrals satisfy the relation \[ \Psi (\alpha, f) = (\theta, \alpha) \psi (\alpha) |\alpha|\Psi' (\alpha^{-1}, f') , \] where \((\theta, \alpha)\) is a Hilbert symbol with \(\theta \in F^\times\) and \(\psi\) is a character of \(F\). This is a local fundamental lemma needed to compare the geometric sides of the global Fourier summation formulas or relative trace formulas on these groups. This relation leads to conclusions about a well-known lifting of representations from \(PGL(2)\) to \(PGSp(4)\), and this lifting produces counterexamples to the Ramanujan conjecture.