Characterizations of infinite dimensional Gaussian shift experiments (Q1825501)

The author generalizes the following characterization of the normal distribution to random variables with values in a separable Banach space: Sufficiency of \(\sum^{n}_{i=1}a_ iX_ i(\sum^{n}_{i=1}a_ i\neq 0\), \(a_ i\neq 0\), \(i=1,...,n)\) for a translation parameter resp. independence of \(E_{\theta}(X_ 1X_ 2| \sum^{n}_{i=1}X_ i| n)\) of the translation paramter \(\theta\), where \(X_ 1,...,X_ n\) are i.i.d. (and square integrable in the second case) implies that the \(X_ j\), \(j=1,...,n\), are normally distributed.