Conjecture on transfers of test functions

Let EFFE \geq F' \geq F be a tower of number fields, let SS be a finite set of places of FF containing the infinite places, and let SS' and S0S_0 be the places of FF' and EE above SS. Let Π\Pi be a cuspidal automorphic representation of GLn(AE)\mathrm{GL}_n(\mathbb{A}_E), and suppose ΠσΠ\Pi^{\sigma}\cong\Pi for every σGal(E/F)\sigma\in\mathrm{Gal}(E/F). A transfer is a triple of compactly supported smooth functions fS0f_{S_0}, hSh_{S'}, and ΦS\Phi_S satisfying the trace identity of Definition of the paper for every irreducible generic unitary representation πS\pi_S of GLn(FS)\mathrm{GL}_n(F_S); a function is of positive type in the sense used there. Transfer conjecture. There exist fS0Cc(GLn(ES0))f_{S_0}\in C_c^{\infty}(\mathrm{GL}_n(E_{S_0})) and hSCc(GLn(FS))h_{S'}\in C_c^{\infty}(\mathrm{GL}_n(F'_{S'})) admitting a transfer ΦSCc(GLn(FS))\Phi_S\in C_c^{\infty}(\mathrm{GL}_n(F_S)) of positive type such that the transfer identity holds for every irreducible generic unitary representation πS\pi_S of GLn(FS)\mathrm{GL}_n(F_S) and

tr(ΠS0)(fS0)0.\mathrm{tr}(\Pi_{S_0})(f_{S_0})\neq 0.

The conjecture asserts that sufficiently many transfers can detect every Galois-invariant cuspidal representation; the paper notes that it is known in some special settings, including n=2n=2 and cyclic extensions of prime degree, but remains open in general.

Sources & referencesView supporting material

Primary source

Jayce R. Getz, “An approach to nonsolvable base change and descent”, arXiv:1701.01766 (2017).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.