An elementary account of Selberg's lemma (Q1100285)

The author offers proofs of the following known results. Let G be a subgroup of GL(n,R), where R is a finitely generated integral domain of characteristic \(p\geq 0\). a) G is residually finite. b) If \(p=0\) then G is torsion-free by finite. c) If \(p>0\) then G is a finite extension of a group whose finite elements are all unipotent. a) is due to \textit{A. I. Mal'cev} [Mat. Sb., Nov. Ser. 8, 405-422 (1940; Zbl 0025.00804) = Transl., II. Ser., Am. Math. Soc. 45, 1-18 (1965)], b) appears in \textit{A. Selberg} [Contrib. Function Theory, Tata Inst. Bombay 1960, 147-164 (1960; Zbl 0201.366)], effectively in \textit{Yu. I. Merzlyakov} [Algebra Logika 3, No. 4, 49-59 (1964; Zbl 0244.20048)] and in \textit{M. I. Kargapolov} [Algebra Logika 6, No. 5, 17-20 (1967; Zbl 0252.20042)] and the obvious variant c) of b) in \textit{V. P. Platonov} [Dokl. Akad. Nauk BSSR 12, 492-494 (1968; Zbl 0228.20018)] and the reviewer [Proc. Lond. Math. Soc., III. Ser. 20, 101-122 (1970; Zbl 0188.062)]. a) is an easy consequence of Hilbert's Nullstellensatz (which can be dressed up as field theory of course) and b) and c) of the Krull intersection theorem. Here the author presents `field theoretic' proofs of these results. The basic trick is to reduce to R being a suitable localization of a polynomial ring, a trick that essentially also appears in Mal'cev's original 1940 paper. This reduction looses information about the group G and hence usually needs to be avoided.