ProofComplete

Sigma-Algebras and Measures - Key Proof

ProofProof of Continuity of Measures

We prove the continuity from below property of measures: if A1A2A_1 \subseteq A_2 \subseteq \cdots are measurable sets, then μ(n=1An)=limnμ(An)\mu\left(\bigcup_{n=1}^{\infty} A_n\right) = \lim_{n \to \infty} \mu(A_n)

Proof: Define B1=A1B_1 = A_1 and for n2n \geq 2, let Bn=AnAn1B_n = A_n \setminus A_{n-1}. Then the sets BnB_n are pairwise disjoint, and n=1Bn=n=1An\bigcup_{n=1}^{\infty} B_n = \bigcup_{n=1}^{\infty} A_n

By countable additivity of μ\mu: μ(n=1An)=μ(n=1Bn)=n=1μ(Bn)\mu\left(\bigcup_{n=1}^{\infty} A_n\right) = \mu\left(\bigcup_{n=1}^{\infty} B_n\right) = \sum_{n=1}^{\infty} \mu(B_n)

Now observe that for each kk: Ak=n=1kBnA_k = \bigcup_{n=1}^{k} B_n

By finite additivity: μ(Ak)=n=1kμ(Bn)\mu(A_k) = \sum_{n=1}^{k} \mu(B_n)

Therefore: μ(n=1An)=n=1μ(Bn)=limkn=1kμ(Bn)=limkμ(Ak)\mu\left(\bigcup_{n=1}^{\infty} A_n\right) = \sum_{n=1}^{\infty} \mu(B_n) = \lim_{k \to \infty} \sum_{n=1}^{k} \mu(B_n) = \lim_{k \to \infty} \mu(A_k)

This completes the proof.

Remark

The continuity from above property requires an additional finiteness condition. If A1A2A_1 \supseteq A_2 \supseteq \cdots and μ(A1)<\mu(A_1) < \infty, then: μ(n=1An)=limnμ(An)\mu\left(\bigcap_{n=1}^{\infty} A_n\right) = \lim_{n \to \infty} \mu(A_n)

The finiteness condition is necessary. Without it, the result can fail. Consider counting measure on N\mathbb{N} and let An={n,n+1,n+2,}A_n = \{n, n+1, n+2, \ldots\}. Then AnA_n \downarrow \emptyset, but μ(An)=\mu(A_n) = \infty for all nn, so the limit is 0=μ()\infty \neq 0 = \mu(\emptyset).

To prove continuity from above, we use the result just proven. Note that A1AnA1n=1AnA_1 \setminus A_n \uparrow A_1 \setminus \bigcap_{n=1}^{\infty} A_n. By continuity from below: μ(A1n=1An)=limnμ(A1An)\mu\left(A_1 \setminus \bigcap_{n=1}^{\infty} A_n\right) = \lim_{n \to \infty} \mu(A_1 \setminus A_n)

Since μ(A1)<\mu(A_1) < \infty, we can subtract to get: μ(A1)μ(n=1An)=μ(A1)limnμ(An)\mu(A_1) - \mu\left(\bigcap_{n=1}^{\infty} A_n\right) = \mu(A_1) - \lim_{n \to \infty} \mu(A_n)

Thus μ(n=1An)=limnμ(An)\mu\left(\bigcap_{n=1}^{\infty} A_n\right) = \lim_{n \to \infty} \mu(A_n) as claimed. These continuity properties are essential for proving convergence theorems in integration theory.