Steinberg groups as amalgams (Q325498): Difference between revisions

Let \(R\) be a commutative ring and \(A\) a generalized Cartan matrix. Denote by \(\mathfrak{St}_A(R)\) the Steinberg group of type \(A\) as defined by \textit{J. Morita} and \textit{U. Rehmann} [Tohoku Math. J. (2) 42, No. 4, 537--560 (1990; Zbl 0701.19001)]. Recall that this group is generated by elementary root unipotents \(x_\alpha(t)\), parametrized by all real roots of the root system \(\Phi=\Phi(A)\), subject to Chevalley relations which are imposed for all prenilpotent pairs of real roots. In the paper under review a new group \(\mathfrak{PSt}_A(R)\), called \textit{pre-Steinberg group}, is introduced. This group has the same generators as \(\mathfrak{St}_A(R)\), but the Chevalley relations are imposed only for classically prenilpotent pairs of roots. It turns out that this group coincides with \(\mathfrak{St}_A(R)\) under mild assumptions on the form of the Dynkin diagram of \(A\) (e.\,g. this is true if every subdiagram with \(\leq 3\) vertices is a Dynkin diagram of a finite root system, see Theorem~1.1). The key virtue of the group \(\mathfrak{PSt}_A(R)\) is that it admits a very compact presentation formulated solely in terms of the Dynkin diagram of \(A\), in which, moreover, no signs are left implicit, see Theorem~1.2. Another advantage of this presentation is that its very form implies the fact that the group \(\mathfrak{PSt}_A(R)\) is a colimit of groups \(\mathfrak{PSt}_B(R)\), where \(B\) runs over all \(1\times 1\) and \(2\times 2\) Cartan submatrices of \(A\) (this fact is a generalization of the classical Curtis-Tits presentation known in the theory of finite groups of Lie type). Based on the Curtis-Tits property of pre-Steinberg groups and \textit{S. Splitthoff}'s earlier result on finite presentation of Steinberg groups (see~[Contemp. Math. 55, 635--687 (1986; Zbl 0596.20034)]) the author deduces finite presentability theorems for a large class of pre-Steinberg groups and Tits' Kac-Moody groups over rings satisfying some finite presentability assumptions, see Theorem~1.4 and Corollary~1.5, respectively.