Let X be a Hausdorff topological space, μ a positive measure of norm 1 on X and E a hilbertian subspace of the real hilbertian space LR2(X,μ). We say that E is a gaussian space if the following equivalent conditions are satisfied :
a) for all f∈E, we have ∫Xeifdμ=exp(−∥f∥2/2);
b) for all f∈E with norm 1, the image of the measure μ under f is the measure
(2π)−1/2 e−x2/2dx. on
R.
Suppose E is a gaussian space. Let f1,...,fn be functions whose classes fi belong to E. We define a function : f1...fn : on X (called « Wick’s product » of f1,...,fn) by the formula
(23):f1...fn:=0≤2p≤n∑(−1)pσ∈Ip∑i=1∏p⟨fσ(2i−1)∣fσ(2i)⟩j=2p+1∏nfσ(j), where
Ip is the set of permutations
σ of
{1,2,...,n} such that we have
σ(1)<σ(2),...,σ(2p−1)<σ(2p)σ(1)<σ(3)<...<σ(2p−1)σ(2p+1)<σ(2p+2)<...<σ(n).
Then there exists an isomorphism ϕ from S^(E) onto a hilbertian subspace of LR2(X,μ) which transforms the product f1...fn of f1,...,fn, calculated in S^(E), into (:f1...fn:). Suppose that X is a Souslin space and that there exists a countable family (fn) of functions whose classes belong to E and which separate the points of X. Then ϕ is an isomorphism from S^(E) onto LR2(X,μ). \*
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.
Nothing.