Symplectic period for a representation of \(\mathrm{GL}_n (\mathrm{D})\) (Q6982718)

The branching problem is one of the most important and fascinating problems in representation theory. The motivation for the paper under review is the symplectic period problem, that is, the classification of irreducible admissible representations of \(\mathrm{GL}_{2n}(F)\) which support a non-zero functional invariant under \(\mathrm{Sp}_{2n}(F)\). Here, \(F\) is a non-Archimedean local field of characteristic \(0\). This problem starts with \textit{A. A. Klyachko} [Math. USSR, Sb. 48, 365--379 (1984; Zbl 0543.20026)] in the finite field case. \textit{M. J. Heumos} and \textit{S. Rallis} [Pac. J. Math. 146, No. 2, 247--279 (1990; Zbl 0752.22012)] first considered the \(p\)-adic field case. One of their results shows that supercuspidal representations cannot have a symplectic period. This result can also be explained using the relative Langlands duality. In fact, as \(\mathrm{GL}_{2n}/\mathrm{Sp}_{2n}\) is a non-tempered spherical variety, it is expected that only residual representations show up in its spectrum.\N\NIn a series of papers, \textit{O. Offen} and \textit{E. Sayag} [Int. Math. Res. Not. 2008, Article ID rnm136, 25 p. (2008; Zbl 1158.22021); J. Number Theory 125, No. 2, 344--355 (2007; Zbl 1147.11028); J. Funct. Anal. 254, No. 11, 2846--2865 (2008; Zbl 1143.22013)] give a classification of unitary representations of \(\mathrm{GL}_{2n}(F)\) which have symplectic periods. Along the way, they also corrected several inaccuracies and close several gaps in previous papers. We refer the reader to [Offen and Sayag, Zbl 1143.22013] for some history of this problem.\N\NThe paper under review considers the inner form version of the symplectic period problem. Indeed, the authors consider irreducible admissible representations of \(\mathrm{GL}_{n}(D)\) where \(D\) is a quaternion algebra over \(F\). There are several new phenomena for this problem. For example, the parabolic induction of two representations without symplectic period can have a symplectic period. Another interesting fact is that when \(n\) is odd, there are supercuspidal representations that have symplectic periods. The main results of this paper give a classification of irreducible admissible representations with a symplectic period when \(n=3,4\). For the reason mentioned above, when \(n=3\), the authors only consider non-supercuspidal representations.