Multiple commutator formulas. (Q375865)

Let \(R\) be a commutative ring and let \(A\) be a direct limit of \(R\)-algebras each of which is module finite over \(R\). Let \(\mathrm{GL}_n(A)\) be the general linear group over \(A\) and let \(E_n(A)\) be its elementary subgroup. For each two-sided \(A\)-ideal \(I\) let \(\mathrm{GL}_n(A,I)\) denote the \textit{congruence subgroup of level \(I\)} and let \(E_n(A,I)\) denote the \textit{relative elementary subgroup of level \(I\)}. Now let \(I_i\) (\(0\leq i\leq m\)) be \(A\)-ideals and suppose that \(n\geq 3\). In answer to a question of Stepanov and Vavilov the authors prove the following multiple commutator formula. \[ [E_n(A,I_0),\mathrm{GL}_n(A,I_1),\ldots,\mathrm{GL}_n(A,I_m)]=[E_n(A,I_0),E_n(A,I_1),\ldots,E_n(A,I_m)]. \] This extends all published results of this type. The proof is based on a method of localization and completion of general linear groups introduced by Bak. The original versions of commutator formulae like the above are due to Bass as part of the classical foundations of algebraic \(K\)-theory. They involve restrictions on \(n\) imposed by the so-called \textit{stable range} of \(R\). Later Suslin and others were able to remove these restrictions. However the constraint \(n\geq 3\) cannot be removed. It is well-known that when \(n=2\) and \(R\) has ``few'' units commutator formulae of the above type can break down completely.