TheoremBase

A countable set is a countable union of one-point sets, so countable subadditivity reduces every claim to the measure of a single point; for Lebesgue measure that is bounded by the measure of arbitrarily small closed balls, and absolute continuity then transfers the vanishing to the measure itself.

Proof

Throughout, each result cited is universally quantified over the data appearing in its own statement and is applied to the data named here. Let ι\iota be the canonical map from N\mathbb{N} to R\mathbb{R}, positive with positive inverse by claim 3 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field; this is the only use of that symbol here. For x∈Rqx\in\mathbb{R}^{q} the one-point set {x}\{x\} belongs to B(Rq)\mathcal{B}(\mathbb{R}^{q}) by Pairs of Euclidean Points: Coordinate Projections, Pairings, the Product Measure on a Euclidean Space, Borel Norm Functions and Finite Sets (v).

Step 1 (A countable set is Borel and its measure is bounded by the sum of the point masses). Let α\alpha be a measure on (Rq,B(Rq))(\mathbb{R}^{q},\mathcal{B}(\mathbb{R}^{q})) with α({x})=0\alpha(\{x\})=0 for every x∈Rqx\in\mathbb{R}^{q}, and let A⊆RqA\subseteq\mathbb{R}^{q} be countable. If A=∅A=\varnothing, then A∈B(Rq)A\in\mathcal{B}(\mathbb{R}^{q}) and α(A)=0\alpha(A)=0 by Measure, Measure Space, and Probability Measure. Otherwise, by Countable Set there is a sequence (am)m∈N(a_{m})_{m\in\mathbb{N}} whose set of terms is AA, so that

A=⋃m∈N{am},A=\bigcup_{m\in\mathbb{N}}\{a_{m}\},

which belongs to B(Rq)\mathcal{B}(\mathbb{R}^{q}), a σ\sigma-algebra being closed under countable unions. By claim 4 of Basic Properties of a Measure,

α(A)≤∑m∈Nα({am})=0,\alpha(A)\le\sum_{m\in\mathbb{N}}\alpha(\{a_{m}\})=0,

the series having every term 00; and 0≤α(A)0\le\alpha(A) by Measure, Measure Space, and Probability Measure, so α(A)=0\alpha(A)=0.

Step 2 (Claim 1). If μ\mu is atomless then μ({x})=0\mu(\{x\})=0 for every xx, so Step 1, applied with α=μ\alpha=\mu, gives A∈B(Rq)A\in\mathcal{B}(\mathbb{R}^{q}) and μ(A)=0\mu(A)=0.

Step 3 (Claim 2). Let x∈Rqx\in\mathbb{R}^{q} and j∈Nj\in\mathbb{N}. By claims 2 and 3 of Properties of the Euclidean Norm on Rn\mathbb{R}^n: Square, Distance, Vanishing, Coordinate Bounds, Homogeneity, Triangle Inequalities and Expansion of a Square one has dE(x,x)=∥x−x∥=0d_{E}(x,x)=\lVert x-x\rVert=0, so x∈Bˉ(x,ι(j)−1)x\in\bar{B}(x,\iota(j)^{-1}) and {x}⊆Bˉ(x,ι(j)−1)\{x\}\subseteq\bar{B}(x,\iota(j)^{-1}). By claim 3 of The Lebesgue Measure of a Closed Ball in Rn\mathbb{R}^n and claim 2 of Basic Properties of a Measure,

0≤λq({x})≤κq(ι(j)−1)q,0\le\lambda_{q}(\{x\})\le\kappa_{q}\bigl(\iota(j)^{-1}\bigr)^{q},

where κq\kappa_{q} is the real number of claim 2 of that lemma; in particular λq({x})\lambda_{q}(\{x\}) is a real number. The sequence (ι(j)−1)j∈N(\iota(j)^{-1})_{j\in\mathbb{N}} converges to 00: given a positive real ε\varepsilon, claim 3 of The Archimedean Property of the Real Numbers provides j0∈Nj_{0}\in\mathbb{N} with ι(j0)−1<ε\iota(j_{0})^{-1}<\varepsilon, and every j≥j0j\ge j_{0} satisfies ι(j0)≤ι(j)\iota(j_{0})\le\iota(j) by claim 6 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field, hence ι(j)−1≤ι(j0)−1<ε\iota(j)^{-1}\le\iota(j_{0})^{-1}<\varepsilon on multiplying by ι(j0)−1ι(j)−1\iota(j_{0})^{-1}\iota(j)^{-1}, which is nonnegative by claims 7 and 5 of Elementary Order Arithmetic in an Ordered Field, using claim 5 of Elementary Arithmetic in an Ordered Field and claim 2 of Elementary Order Arithmetic in an Ordered Field. Consequently (κq(ι(j)−1)q)j\bigl(\kappa_{q}(\iota(j)^{-1})^{q}\bigr)_{j} converges to 00 by claims 2 and 3 of Arithmetic of Limits of Real Sequences, the power being a finite product. Hence λq({x})≤η\lambda_{q}(\{x\})\le\eta for every positive real η\eta, and Comparison of Real Numbers with Arbitrary Positive Slack (iii) gives λq({x})=0\lambda_{q}(\{x\})=0.

The second assertion of claim 2 now follows from Step 1, applied with α=λq\alpha=\lambda_{q}, which satisfies λq({x})=0\lambda_{q}(\{x\})=0 for every xx by what has just been proved.

Step 4 (Claim 3). Suppose μ\mu is absolutely continuous and let x∈Rqx\in\mathbb{R}^{q}. The set {x}\{x\} belongs to B(Rq)\mathcal{B}(\mathbb{R}^{q}) and satisfies λq({x})=0\lambda_{q}(\{x\})=0 by claim 2, so Absolutely Continuous Probability Measure on Euclidean Space (i) gives μ({x})=0\mu(\{x\})=0. As xx was arbitrary, μ\mu is atomless.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…