Multiple commutator formulas for unitary groups (Q2357008): Difference between revisions

The main result is the following multiple commutator identity: \[ [\cdots[[\mathrm{EU}(2n, I_0, \Gamma_{0}),\mathrm{GU}(2n,I_1,\Gamma_1)], \mathrm{GU}(2n, I_2, \Gamma_2)],\dots, \mathrm{GU}(2n, I_m, \Gamma_m)] \] equals \[ [\cdots[[\mathrm{EU}(2n, I_0, \Gamma_0), \mathrm{EU}(2n, I_1, \Gamma_1)], \mathrm{EU}(2n, I_2, \Gamma_2)],\dots, \mathrm{EU}(2n, I_m, \Gamma_m)]. \] Here, \(n\geq3\) and the \((I_j,\Gamma_j)\) are form ideals in the form ring \((A,\Lambda)\), where \(A\) is a module finite algebra over the commutative ring \(R\) (or a direct limit of such algebras). The \(E(2n, I_j, \Gamma_j)\) are the relative elementary subgroups of the principle congruence unitary groups \(\mathrm{GU}(2n, I_j, \Gamma_j)\). `This result contains all previous results on commutator formulas for classical-like groups over commutative and finite-dimensional rings.' Specifically, the result covers the symplectic case, the special orthogonal case and the special unitary case. Other bracketing patterns are allowed, like \([\cdot,[\cdot,[\cdot,\cdots[\cdot,\cdot]]\cdots]\), provided the same pattern is used in both multiple commutators. The proof is long and combines several techniques.