Pointwise multiple averages for systems with two commuting transformations

We show that for every ergodic measure-preserving system (X, X, mu, S, T) with commuting transformations S and T, the average i/N-3 Sigma(N-1)(i,j,k=0) f(0)(S(j)T(k)x)f(1)(S(i+j)T(k)x)f(2)(S(j)T(i+k)x) converges for mu-almost every x is an element of X as N -> infinity for all f(0), f(1), f(2) is an element of L-infinity(mu). We also show that if (X, X, mu, S, T) is an ergodic measurable distal system, then the average 1/N Sigma(N-1)(i=0) f(1)(S(i)x) f(2)(T(i)x) converges for mu-almost every x is an element of X as N -> infinity for all f(1), f(2) is an element of L-infinity (mu).