Characterization of \(k\)-commuting additive maps on rings. (Q2512798)

Let \(R\) be a 2-torsion free ring with identity element 1, center \(Z(R)\), an idempotent \(e\neq 0,1\) and \(1-e=e'\). For any \(x,y\in R\) let \([x,y]_1=xy-yx\) and for an integer \(k>1\) let \([x,y]_k=[[x,y]_{k-1},y]_1\). If \(f\colon R\to R\) is additive and if \([f(y),y]_k=0\) for all \(y\in R\) call \(f\) a \(k\)-commuting map, and when \(f\) is 1-commuting call \(f\) commuting. Let \(Z(R)_k=\{r\in R\mid [r,a]_k=0\) for all \(a\in R\}\), so \(Z(R)=Z(R)_1\). The main result of the authors is that if each of \(Re\) and \(Re'\) has no nonzero left annihilator, and for some integer \(k>0\), \(Z(eRe)_k=Z(eRe)\), and \(Z(e'Re')_k=Z(e'Re')\), then the following are equivalent: i) \(f\) is \(k\)-commuting; ii) \(f\) is commuting; iii) there are \(z\in Z(R)\) and an additive \(h\colon R\to Z(R)\) so that \(f(x)=zx+h(x)\) for all \(x\in R\). Corollaries of this result show that the conclusion holds when \(R\) is prime with 1 and has a nontrivial idempotent, or when \(R\) is a von Neumann algebra.