On (anti-)multiplicative generalized derivations. (Q2901987)

Let \(R\) be a semiprime ring with extended centroid \(C(R)\) and nonzero ideal \(I\). Set \(A=\text{ann}(\text{ann}(I))\). Consider functions \(F,f\colon R\to R\) that satisfy \(F(xy)=F(x)y+xf(y)\) for all \(x,y\in R\). The main result in the paper assumes also that \(F(uv)=mF(u)F(v)+nF(v)F(u)\) for integers \(m\) and \(n\), and for all \(u,v\in I\) and shows that if \(R\) is \((n+m)\)-torsion free then \(F\) and \(f\) are additive on \(A\). Further, if either \(m=0\), \(n=0\), or \(R\) is 2-torsion free and also \(m\)-torsion free or \(n\)-torsion free, then \(f(A)=0\) and for some \(c\in C(A)\), \(c=(n+m)c^2\), \(nc[A,A]=0\), and \(F(x)=cx\) for all \(x\in A\). It follows that the identity \(F(uv)=mF(u)F(v)+nF(v)F(u)\) holds on \(A\). One consequence of the theorem is that when \(I\) is an essential ideal of \(R\), then \(F\) multiplicative on \(I\) forces \(f=0\) and the existence of \(c^2=c\in C(R)\) satisfying \(F(x)=cx\) for all \(x\in R\), so \(F\) is multiplicative on \(R\). Alternatively, when \(F\) is anti-multiplicative on \(I\) then again \(f=0\) and \(F(x)=cx\) for all \(x\in R\), but now \(c[R,R]=0\) and \(F\) is both multiplicative and anti-multiplicative on \(R\).