Weyl groups, the Dirac inequality, and isolated unitary unramified representations (Q2076499)

Let \(G\) be a simple algebraic group split over a non-archimedean local field, and \(\mathbb{H}\) be the graded affine Hecke algebra defined by G. Lusztig. The extended Dirac inequality [\textit{D. Barbasch} et al., Acta Math. 209, No. 2, 197--227 (2012; Zbl 1276.20004)] states that the length \(\lvert \nu \rvert\) of the central character \(\nu\) of a unitary simple \(\mathbb{H}\)-module does not exceed \(\lvert h^\vee_{\mathcal{O}^\vee}/2 \rvert\) for certain nilpotent orbits \(\mathcal{O}^\vee\), here \(h^\vee_{\mathcal{O}^\vee}\) is a semisimple element attached to \(\mathcal{O}^\vee\). Using this inequality, the paper under review finds out all unitary \(\mathbb{H}\)-modules with central character \(\nu\) such that \(\lvert \nu \rvert = \lvert h^\vee_{\mathcal{O}^\vee}/2 \rvert\) in which \(\mathcal{O}^\vee\) takes the regular, subregular and sub-subregular orbits respectively, and asserts that these are all unitary \(\mathbb{H}\)-modules with \(\lvert \nu \rvert\) ``large''. In contrast, this paper also studies the unitarisability problem related to the minimal orbit in type \(F_4\). By a result of D. Barbasch and A. Moy, the results of this paper can also be applied to corresponding representations of the \(p\)-adic group \(G\). There seems to be some typos in the paper, for example, \begin{itemize} \item in page 4, ``the transpose anti-automorphism \(x \to x^t\) of \(C(\mathfrak{t}^\vee)\) is defined to be the identity on \(\mathfrak{t}^\vee\)'' should be modified to be ``the negative of the identity on \(\mathfrak{t}^\vee\)''; \item in page 8, the LHS of the formula above formula (4.1), ``\(w(\nu)/2\)'' should be ``\(w(\nu)\)''; \item in page 9, the RHS of the inequality below formula (4.2), ``\(\min\{\sigma \in \operatorname{Irr} W \mid \Hom_W[\sigma, \overline{V}] \ne 0\}\)'' should be ``\(\min \{\operatorname{d}(\sigma) \mid \sigma \in \operatorname{Irr} W, \Hom_W[\sigma, \overline{V}] \ne 0\}\)''; \item in page 9, the LHS of the formula above formula (4.3), ``\(\langle \sigma, \sigma \rangle_{-1}^{\mathsf{ell}}\)'' should be ``\(\langle \sigma, \sigma^\prime \rangle_{-1}^{\mathsf{ell}}\)''. \end{itemize}