A Note on the Automorphic Langlands Group (Q4796558)

Given a number field \(F\), Langlands conjectured the existence of a universal group, which would play a fundamental rôle in the classification of automorphic forms. In \textit{R. E. Kottwitz}'s formulation [Duke Math. J. 51, 611-650 (1984; Zbl 0576.22020)] it would be an extension \(L_F\) of the absolute Weil group \(W_F\) by a connected reductive group. It is to be expected that it would have many properties in common with \(W_F\) and the absolute Galois group, in particular, local analogues \(L_{F_v}\), for each valuation \(v\) of \(F\); indeed these are already known [loc. cit.]. Here, the author presents a candidate for the global Langlands group \(L_F\). He says it ``appears to be the most optimistic possible guess on the ultimate form of \(L_F\). It is highly speculative, depending on hypotheses for which there is little evidence, in addition to well known conjectures. On the other hand, it seems to conform with what is generally believed about automorphic representations''. The group \(L_F\) is defined (in \S\S 3-4) as a fibre product over \(W_F\) of extensions \[ 1\to K_c\to L_c\to W_F\to 1. \] The indexing set for the product is the set of isomorphism classes of pairs \((G,c)\), where \(G\) is a quasi-split, simple, simply connected algebraic group over \(F\) and \(c\) is a primitive equivalence class of automorphic families of conjugacy classes \(c_v\in{}^LG_v\); \(K_c\) is a compact real form of the simply connected cover of the complex dual group \(\widehat G\); and the extension is given by describing a two-cocycle associated to \(c\). In \S 5 the author discusses various questions concerning the hypotheses in this construction and the implications of it to functoriality and the local Langlands conjecture. In \S 6 he goes on to describe a possible candidate for the complexification of Grothendieck's motivic Galois group.