Determinantal hypersurfaces and representations of Coxeter groups (Q2234343)

For linear operators \( A_0, A_1, \dots, A_n \) on a complex \( N \) dimensional vector space \(V\), consider a homogeneous polynomial defined by \( P(x_0, \dots, x_n) = \det (\sum_{i = 0}^n x_i A_i) \). It defines a hypersurface in the projective space \(\mathbb{P}^n(\mathbb{C}) \), called determinantal hypersurface. This hypersurface (with multiplicities of its components) is interesting even if the space \( V\) is replaced by an infinite dimensional Hilbert space (in that case, the determinant should replaced by singular spectrum). In this paper, the authors study the determinantal hypersurface arising from representations of finite Coxeter groups. So let \( G \) be a general finite group, and take a representation \( (\rho, V) \) of \( G \). Choose a (generating) subset \( T = \{ g_1, \dots, g_n \} \subset G \) and put \( A_i = \rho(g_i) \; (1 \leq i \leq n) \), while \( A_0 = 1_V \) (the identity operator). Then the determinantal hypersurface above is denoted by \( D(T, \rho) \) (with multiplicity). If \( T = G \) (in this case, we don't need \( A_0 \), since \( \rho(e) = 1_V \)) and \( \rho \) being the regular representation, it reduces to the famous group determinant \( D_G := D(G,\mathbb{C}[G]) \) of Frobenius. His old result tells that the irreducible components of \( D_G \) correspond bijectively with irreducible representations, and that each component appears with the multiplicity of the dimension of the corresponding irreducible representation. However, the size of determinant becomes enormously large, therefore, one would like to choose smaller \( T \). If \( G \) is a finite Coxeter group and \( T \) its Coxeter generators of simple reflections, the authors prove that \( D(T, \rho) \) determines \( \rho \) for any finite dimensional representation \( (\rho, V) \) of \( G \) (Theorem 1.1). This is remarkable for Coxeter groups, because for general finite groups and arbitrary generating set, the claim does not hold. It would be interesting to investigate the components of \( D(T, \rho) \) (and its geometry), but it is not included in the present paper. They also study infinite Coxeter groups as well, and give an example of an affine Coxeter group \( \widetilde{C}_2 \).