Proposition. Every polynomial has finite absolute Gaussian moments [fdip-002W]

Let \(Q\) be positive self-adjoint on a finite real inner-product space \((E,h)\), with weight \(w_Q=e^{-h(x,Qx)/2}\) and canonical volume. Every actual polynomial function satisfies \[\int _E |P(x)|w_Q(x)\,d\lambda _E(x)<\infty .\] This holds for every finite product and power of polynomials, including \(P^2\), and therefore also for integration against the normalized law.

The envelope estimate and finite-sum argument are direct deductions from Keith Conrad's [conradGaussianIntegral, Section 2, PDF pp. 1-2] and [tao2011measure, Theorem 1.7.15, pp. 200-201, preliminary version]. The orthogonal reduction agrees with [friedli2017statistical, Lemma B.54, p. 504]. TauCeti's TauCeti.integrable_eval_mul_gaussianEnvelope supports the one-variable polynomial envelope with mean zero and variance \(w=1/a>0\). The finite products, coordinate substitution and algebra closure above are separate arguments; no inverse-covariance identity is used.