On the central limit theorem in product spaces (Q1908597)

In a separable Banach space \(B\) a mean zero random vector \(X\) is said to satisfy central limit theorem (CLT) in \(B\) if \((X,X_n,n\geq 1)\) is a sequence of i.i.d. random vectors and the normalized sums \(n^{1/2} (X_1+ \cdots +X_n)\) converge weakly to a \(B\)-valued Gaussian random vector. Suppose that \(E\) and \(F\) are separable Banach spaces, \(X\) and \(Y\) are independent symmetric \(E\)- and \(F\)-valued random vectors, respectively. The author studies the CLT for \(X \otimes Y\) in the injective and projective tensor product spaces \(E\check\otimes F\) and \(E\widehat\otimes F\). Special attention is paid to \(l_2 \widehat \otimes l_2\). In addition, he gives two counterexamples. The method used here is the isometric isomorphism of the continuous bilinear functionals to its original tensor product space.