Lie algebras and cohomology of congruence subgroups for \(\mathrm{SL}_n(R)\). (Q392407): Difference between revisions

Let \(R\) be a commutative ring that is free as an Abelian group, but not necessarily of finite rank. Let \(p\) be a prime. The author studies the Lie algebra associated to the filtration of \(\mathrm{SL}_n(R)\) by \(p\)-congruence subgroups. He also obtains some homological results. Unfortunately the paper makes incorrect claims. The first claim is about vanishing of congruence kernels. The author tells it is proved in a paper of \textit{H. Bass} et al. [Publ. Math., Inst. Hautes Étud. Sci. 33, 59-137 (1967; Zbl 0174.05203)], but not where in that paper. The second claim is that maps of the type \(\mathrm{SL}_n(R)\to\mathrm{SL}_n(R\otimes_{\mathbb Z}\mathbb Z/p)\) are surjective. For this he refers to a book of \textit{G. Shimura} [Introduction to the arithmetic theory of automorphic functions. Princeton, Iwanami Shoten Publishers and Princeton University Press (1971; Zbl 0221.10029)], again without any further specification. To get a counterexample one may use a paper of \textit{T. Vorst} [Math. Ann. 244, 193-204 (1979; Zbl 0426.13005)]. Take \(A=\mathbb Z[X,Y]/(Y^2-X^3-6)\). It is regular, hence \(NK_1\)-regular. But \(A/2A\) is not seminormal, so its \(NK_1\) does not vanish by the main result of the paper of Vorst. It follows that \(\mathrm{SL}_n(A[T])\to\mathrm{SL}_n(A[T]\otimes_{\mathbb Z}\mathbb Z/2)\) is not surjective for \(n\) large.