A note on Martindale lemma. (Q2880073)

Let \(R\) denote a prime ring with extended centroid \(C\), maximal ring of quotients \(U\), and central closure \(S\); and let \(C\{X_1,X_2,\dots,X_m\}\) be the free \(C\)-algebra in noncommuting indeterminates \(X_1,X_2,\dots,X_m\). The Martindale lemma of the title asserts that if \(\{a_1,a_2,\dots,a_n\}\) is a \(C\)-independent subset of \(S\) and \(\{b_1,b_2,\dots,b_n\}\subseteq S\) such that \(\sum a_ixb_i=0\) for all \(x\in S\), then \(b_1=b_2=\cdots=b_n=0\).NEWLINENEWLINE The authors give several nonlinear generalizations of this result, the principal one reading as follows: Let \(n\), \(m\), \(t_1,t_2,\dots,t_n\) be fixed positive integers; and for \(1\leq i\leq n\) and \(1\leq j\leq t_i\), let \(f_{ij}(X_1,\dots,X_m)\in C\{X_1,\dots,X_m\}\) with zero constant term, and let \(b_{ij}\in U\). For each \(i=1,2,\dots,n\) and \(x_1,x_2,\dots,x_m\in R\) define \(F_i(x_1,\dots,x_m)=\sum f_{ij}(x_1,\dots,x_m)b_{ij}\), the sum being over \(j= 1,2,\dots,t_i\). If \(a_1,a_2,\dots,a_n\) are \(C\)-independent elements of \(U\) and \(\sum a_iF_i(x_1,x_2,\dots,x_m)=0\) for all \(x_k\in R\), then for each \(i=1,2,\dots,n\), \(F_i(x_1,\dots,x_m)=0\) for all \(x_k\in C\).