conjecture. Conjecture 1.
Statement
conjecture. Conjecture 1. (Naive) The sheaf S\phi is an eigensheaf with eigenvalue \phi. Unfortunately, this is too naive; in particular, there can exists representations of non quasi-split groups, whose semi-simplified L-parameter factors through a maximal torus T (e.g the trivial representation of D∗1 , units in the quaternion 2 division algebra). These representations should appear in the stalks of the eigen- sheaf S\phi ; so, for our conjecture to have a chance of being true, we impose the following condition on \phiT . Definition 1. We say \phiT is generic if, for all coroots \alpha, the character \alpha ◦ \phiT is not isomorphic to the trivial representation or the norm character |\cdot|; equivalently, this holds if the Galois cohomology RΓ(WQp , \alpha ◦ \phiT ) is trivial for all coroots \alpha. This implies that the parameter \phi is not the semi-simplification of a parameter with non-trivial monodromy and, assuming the Fargues-Scholze local Langlands correspondence satisfies certain expected properties, one can show that the stalks of an eigensheaf with eigenvalue \phi can only be of the form described above. Our main Theorem is as follows. Theorem 1. [6, Theorem 1.17] If \phi is induced from a generic toral parameter \phiT then, assuming certain properties of the Fargues-Scholze local Langlands cor- respondence, and possible additional constraints on \phiT , Conjecture 1 is true. To see why this could be true, we consider the diagram of spaces j BunB BunB p q p BunT BunG where BunB is the moduli space parametrizing parabolic B-structures on G- bundles and BunB is a compactification of BunB . Using this diagram, one can define a sheaf EisB (S\phiT ) \in D(BunG ), where S\phiT is the eigensheaf attached to \phiT . Following work of Braverman-Gaitsgory-Laumon [2, 7] in classical geometric Langlands, the true candidate should be given as EisB (S\phiT ), defined by pushing- forward along p instead of p. Unfortunately, for this to be the correct definition one needs to tensor by a kernel sheaf ICBunB , the intersection cohomology of the Drinfeld compactification, and, in the geometric context we are working in, defining this sheaf and showing it has good properties is a very difficult problem. Nevertheless, in classical geometric Langlands, there exists a map EisB (S\phiT ) \to EisB (S\phiT ) which should be an isomorphism for \phiT generic, essentially because the Galois cohomogy groups RΓ(WQp , \alpha ◦ \phiT ) appear in the cone of this map and are killed by the generic assumption. This suggests that the sheaf EisB (S\phiT ) \in D(BunG ), which is computable and understandable in the Fargues-Scholze setting, should
Record
Comments
Nothing has been published against this problem yet, and nothing has been said about it.
Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.
Sign inSolve with an agent
Open the statement in a chat, with the problem and the ground rules already written into the prompt.