Steinberg's theorem for crystallographic complex reflection groups (Q1712489): Difference between revisions

This nice paper deals with complex reflection groups which topic has been intensively researched in the literature. The elaboration of the background to this topic is due (among others) to the work of authors G. Bagnera (1905), H. F. Blichfeldt (1905), H. H. Mitchell (1914), N. Bourbaki, C. (1968), Chevalley, (1955), A. M. Cohen (1976). Here, I would just like to highlight the paper of \textit{G. C. Shephard} and \textit{J. A. Todd} [Can. J. Math. 6, 274--304 (1954; Zbl 0055.14305)] who published a list of all finite irreducible complex reflection groups (up to conjugacy). In their classification, they separately studied the imprimitive groups and the primitive groups. The authors deal in a narrower sense with the crystallographic complex reflection groups that are classified by \textit{V. L. Popov} by determining lattices they stabilize [Commun. Math. Inst., Rijksuniv. Utrecht 15, 89 pp. (1982; Zbl 0481.20030)]. The main result of the article is a further development of \textit{R. Steinberg}'s theorem [Can. J. Math. 12, 616--618 (1960; Zbl 0099.36802)]: Theorem. Let \(W\) be a Coxeter group (of finite or infinite order) or a finite complex reflection group acting on \(V\). Then a vector \(v\) in \(V\) is fixed by some nonidentity group element of \(W\) if and only if \(v\) lies on a reflecting hyperplane for \(W\). The authors show that Steinberg's theorem holds for most crystallographic complex reflection groups but not all. They classify those groups in a family for which non-regular orbits all lie on reflecting hyperplanes, thus determining when Steinberg's theorem holds and also determined those crystallographic reflection groups \(G\ltimes\Lambda\) for which Steinberg's theorem fails. The central results of the paper are summarized in the following theorems (using the notation of the authors and of Popov [loc. cit.]): Theorem. Suppose \(W =G(r,p,n) \ltimes \Lambda\) is a genuine crystallographic complex reflection group acting on \(C^n\) for some lattice \(\Lambda\) with \(r,p,n \geq 1\). Then \(W\) has the Steinberg property if and only if \(W\) is \([G(r, 1,1)]_1\), \([G(3, 1,n)]_1\), \([G(4, 1,n)]_{1,2}\), \([G(6, 1,n)]_1\), \([G(4, 2,n)]_{1,2}\), \([G(4, 2, 2)]_3\), \([G(6, 2,n)]_1\), \([G(6, 2, 2)]_2\), or \([G(6, 3,n)]_1\). Theorem: Let \(W =G(r,p,n) \ltimes \Lambda\) be a crystallographic reflection group whose linear part \(\mathrm{Lin}(W)\) is the complexification of a finite Coxeter group. Then \(W\) has the Steinberg property if and only if \(W\) is \([W(A_{n-1})]_1^\alpha\), \([G(2, 1,n)]_1^\alpha\), or \([G(2, 2, 3)]_1\).