Existence of Klyachko models for \(\mathrm{GL}(n,\mathbb R)\) and \(\mathrm{GL}(n,\mathbb C)\) (Q413205): Difference between revisions

In 1984, A. A. Klyachko proved that \(\mathrm{GL}(n,\mathbb{F}_q)\) contains a family of subgroups \(G_d\) (\(1 \leq d \leq n/2\)) and one-dimensional complex characters \(\chi_d\) of \(G_d\) such that \[ \sum_{d=0}^{[n/2]} \mathrm{Ind}(\chi_d) = \sum_{\chi \in \mathrm{Irr}(\mathrm{GL}(n,\mathbb{F}_q))} \chi. \] That is, the characters induced by \(\chi_d\)'s are multiplicity-free, pairwise disjoint and contain as constituents all the irreducible complex characters of \(\mathrm{GL}(n,\mathbb{F}_q)\) (they give a model). More general results were obtained later by N. F. J. Inglis and J. Saxl, R. B. Howlett \& C. Zworestine. Later, such Klyachko models were constructed by M. J. Heumos and S. Rallis. They were studied in further detail by others like O. Offen and E. Sayag and C. Nien. \noindent In this paper, the archimedean cases \(\mathrm{GL}(n,\mathbb{R})\) and \(\mathrm{GL}(n, \mathbb{C})\) are dealt with - the authors prove the existence of a Klyachko model. Although the \(p\)-adic case points to a scheme of proof which could be carried over to the archimedean cases too, there are hurdles particular to these fields. For instance, although a theory of derivatives for smooth Fréchet representations has begun to be developed, there is no suitable Leibniz rule formulated as yet. The third author -- in his proof of the Kirillov conjecture for the complex case and many real cases -- developed a theory of adduced representations and this is adapted in this paper.