Image of functoriality for general spin groups (Q404035)

Consider a quasi-split general spin group \(G\), written as \(\mathrm{GSpin}(2n)\) or \(\mathrm{GSpin}(2n+1)\), over a number field \(k\). The principle of functoriality predicts a transfer of irreducible automorphic representations from \(G(\mathbb{A}_F)\) to \(\mathrm{GL}(2n, \mathbb{A}_F)\). As is customary, we denote such a lifting of representations by \(\pi \mapsto \Pi\). The local components \(\Pi_v\) of \(\Pi\) can be pinned down at the unramified places, which leads to the notion of \textit{weak transfer}. The main purpose of this article is to establish the weak transfer when \(\pi\) is cuspidal and generic with respect to some given Whittaker datum. It turns out that \(\Pi\) is an isobaric automorphic representation of \(\mathrm{GL}(2n)\), whose isobaric summands \(\Pi_i\) satisfy the condition that their twisted \(\mathrm{Sym}^2\) (resp. \(\Lambda^2\)) \(L\)-function has a pole at \(s=1\) in the even (resp. odd) case. In view of the upcoming work by Hundley and Sayag on automorphic descent for \(G\), the image of \(\pi \mapsto \Pi\) is completely characterized. The applications include a finer description of Kim's transfer from \(\mathrm{GL}(4)\) to \(\mathrm{GL}(6)\), as well as information on the reducibility of parablically induced representations of non-exceptional \(p\)-adic groups. The proof is based on an application of the converse theorem of \textit{J. W. Cogdell} and \textit{I. I. Piatetski-Shapiro} [Publ. Math., Inst. Hautes Étud. Sci. 79, 157--214 (1994; Zbl 0814.11033); J. Reine Angew. Math. 507, 165--188 (1999; Zbl 0912.11022)]. Note that the split case has been considered by the authors in [Duke Math. J. 132, No. 1, 137--190 (2006; Zbl 1099.11028)]. The new ingredients for the quasi-split case are (i) a result on local \(L\)-functions and normalized intertwining operators (Proposition 3.13), and (ii) the stability of \(\gamma\)-factors (Proposition 3.15). The latter serves to wash out the contributions from ramified places upon a highly-ramified twist, in a comparison between the Langlands-Shahidi \(L\)-functions and the \(L\)-functions of pairs that appear in the converse theorem.