English

A probabilistic proof of Schoenberg's theorem

Probability 2019-06-14 v1

Abstract

Assume that g(ξ2)g(|\xi|^2), ξRk\xi\in\mathbb{R}^k, is for every dimension kNk\in\mathbb{N} the characteristic function of an infinitely divisible random variable XkX^k. By a classical result of Schoenberg f:=loggf:=-\log g is a Bernstein function. We give a simple probabilistic proof of this result starting from the observation that Xk=X1kX^k = X_1^k can be embedded into a L\'evy process (Xtk)t0(X_t^k)_{t\geq 0} and that Schoenberg's theorem says that (Xtk)t0(X_t^k)_{t\geq 0} is subordinate to a Brownian motion. A key ingredient of our proof are concrete formulae which connect the transition densities, resp., L\'evy measures of subordinated Brownian motions across different dimensions. As a by-product of our proof we obtain a gradient estimate for the transition semigroup of a subordinated Brownian motion.

Keywords

Cite

@article{arxiv.1808.00190,
  title  = {A probabilistic proof of Schoenberg's theorem},
  author = {Franziska Kühn and René L. Schilling},
  journal= {arXiv preprint arXiv:1808.00190},
  year   = {2019}
}
R2 v1 2026-06-23T03:21:14.266Z