Uniformly bounded representations of \(\operatorname{SL}(2,\mathbb{R})\) (Q1626420)

Uniformly bounded representations have been studied for a very long time; in fact, results on unitarizability of uniformly bounded representations appeared as early as 1947 and extensions of these results in 1950 showed that, on amenable groups, all uniformly bounded representations are unitarizable. The converse is the famous Dixmier's problem, still unsolved. In an important paper, \textit{L. Ehrenpreis} and \textit{F. I. Mautner} [Proc. Natl. Acad. Sci. USA 41, 231--233 (1955; Zbl 0064.02604); Ann. Math. (2) 61, 406--439 (1955; Zbl 0066.35701)] constructed two analytic families \(\pi_{\lambda, \epsilon}\), \(\lambda \in \mathbb C\), \(\epsilon \in \{0,1\}\), of representations of \(\mathrm{SL}(2, \mathbb R)\) as analytic continuations of complementary series representations. These are uniformly bounded for \(| \lambda| <1/2\) and most of these are not similar to unitary representations. This work, in turn, led to the Kunze-Stein phenomenon on noncompact semisimple Lie groups a little later. Subsequently, considerable efforts of many authors have gone into studying uniformly bounded representations. More recently, Pisier has made considerable advances towards the solution of Dixmier's similarity problem and has found connections between uniformly bounded representations and multipliers of the Fourier algebra. The present paper goes back to a detailed study of uniformly bounded representations of \(\mathrm{SL}(2,\mathbb R)\). It is concerned with the norms of the Kunze-Stein version of the Ehrenpreis-Mautner uniformly bounded representations \(\pi_{\lambda, \epsilon}\). (The norm of a uniformly bounded representation \(\pi\) of \(G\) is defined as the infimum of all constants \(C\) such that \(\| \pi(x) \| \leq C\) for all \(x\in G\).) Kunze and Stein gave estimates for these norms and here the exact norms are computed. The main result gives sharp estimates for these norms and shows that the Kunze-Stein representations have optimal norms.