# zbMATH — the first resource for mathematics

0-1 laws of a probability measure on a locally convex space. (English) Zbl 0605.60040
The authors introduce the following notions of 0-1 laws for a measure $$\mu$$ defined on the $$\sigma$$-algebra generated by the topological dual E’ of a locally convex Hausdorff space E resp. on the $$\sigma$$-algebra generated by the open subsets of E.
(0) For every x’$$\in E'$$, $$\mu (x;x'(x)=0)=0$$ or 1.
(1) p5 For every $$\mu$$-measurable linear subspace $$F\subset E$$, $$\mu (F)=0$$ or 1.
(2) For every sequence $$x_ n'\in E'$$ and every convex Lusin subspace F of $$R^{\infty}$$, $$\mu (x;(x_ n'(x))\in F)=0$$ or 1.
(3) For every sequence $$x_ n'\in E'$$, $$\mu (x;(x_ n'(x))\in l_{\infty})=0$$ or 1.
(4) For every sequence $$x_ n'\in E'$$, $$\mu (x;(x_ n'(x))\in c_ 1)=0$$ or 1.
(5) For every sequence $$x_ n'\in E'$$, $$\mu (x;(x_ n'(x))\in c_ 0)=0$$ or 1.
(6) There exist no sequence $$x_ n'\in E'$$ such that $$\mu (x;(x_ n'(x))\in c_ 0)>0$$ and that $$\mu (x;(x_ n'(x)\not\in l_{\infty})>0.$$
(7) For every sequence $$x_ n'\in E'$$, if there exists $$(a_ n)\in c_ 0$$ such that $$\mu (x;(a_ n^{-1}x_ n'(x))\in l_{\infty})>0$$, then $$\mu (x;(x_ n'(x))\in c_ 0)=1.$$
(8) For every sequence $$x_ n'\in E'$$, if there exists $$(a_ n)\in c_ 0$$ such that $$\mu (x;(a_ n^{-1}x_ n'(x))\in l_{\infty})>0$$, then $$\mu (x;(x_ n'(x)\in l_{\infty})=1.$$
(9) For every sequence $$x_ n'\in E'$$, if there exists $$(a_ n)\in c_ 0$$ such that $$\mu (x;(a_ n^{-1}x_ n'(x))\in c_ 1)>0$$, then $$\mu (x;(x_ n'(x))\in c_ 1)=1.$$
(10) For every sequence $$x_ n'\in E'$$, if there exists $$(a_ n)\in c_ 0$$ such that $$\mu (x;(a_ n^{-1}x_ n'(x))\in c_ 0)>0$$, then $$\mu (x;(x_ n'(x))\in c_ 0)=1.$$
(11) For every sequence $$x_ n'\in E'$$, if there exists $$(a_ n)\in c_ 0$$ such that $$\mu (x;(a_ n^{-1}x_ n'(x))\in c_ 0)>0$$, then $$\mu (x;(x_ n'(x))\in l_{\infty})=1.$$
(12) For every sequence $$x_ n'\in E'$$, $$\mu (x;(x_ n'(x))\in l_ p)=0$$ or 1 $$(1\leq p<\infty).$$
(13) For every closed convex balanced subset B, $$\mu$$ ($$\cup^{\infty}_{n=1}nB)=0$$ or 1.
(14) For every lower semi-continuous semi-norm N(x) on E (admitting the value $$\infty)$$, $$\mu (x;N(x)<\infty)=0$$ or 1.
(15) For every compact convex balanced subset K, $$\mu$$ ($$\cup^{\infty}_{n=1}nK)=0$$ or 1.
Main results: Theorem 1: The 0-1 laws (2)-(11) are equivalent. Theorem 2: Suppose that $$\mu$$ is a Radon probability measure. Then the 0-1 laws (2)- (11), (13) and (14) are all equivalent. Theorem 3: If $$\mu$$ is a convex Radon measure, then the 0-1 laws (2)-(11), (13), (14) and (15) are all equivalent.
Reviewer: D.Plachky

##### MSC:
 60F20 Zero-one laws 60B05 Probability measures on topological spaces
Full Text:
##### References:
 [1] Borell, C., Convex measures on locally convex spaces, Ark. for Math., 12 (1974), 239-252. · Zbl 0297.60004 [2] Dudley, R. M. and Kanter, M., Zero-one laws for stable measures, Proc. A. M.S., 45 (1974) , 245-252. · Zbl 0297.60007 [3] Hoffmann-J0rgensen, J., Integrability of seminorms, the 0-1 law and affine kernel for product measures, Studia Math., 61 (1977), 137-159. · Zbl 0373.60014 [4] Krakowiak, W., Zero-one laws for stable and semi-stable measures on Banach spaces, Bull. Acad. Sci. Polonaise, 27 (1978) , 1045-1049. · Zbl 0407.60027 [5] Louie, D., Rajput, B. and Tortrat, A., A zero-one dichotomy theorem for r-semi- stable laws on infinite-dimensional linear spaces, Sankya Ser. A, 42 (1980) , 9-18. · Zbl 0487.60020 [6] Sato, H., Banach support of a probability measure in a locally convex space, Lecture Notes in Math., 526, Springer. · Zbl 0344.60005 [7] Takahashi, Y., Hilbertian support of probability measures on locally convex spaces and their applications, Hokkaido Math. /., 10 (1981), 57-74. · Zbl 0469.46032 [8] Zinn, J., Zero-one laws for non-Gaussian measures, Proc. A. M. S., 44 (1974) , 179-185. · Zbl 0309.60022
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.