Section 1.4

The Borel σ\sigma-Field

[6/67%]
Show
LaTeX
Problem 1.4.1

Show directly that σ(A3)=σ(A3∗),σ(A4)=σ(A7)\sigma \left(\mathscr {A}_{3}\right)=\sigma \left(\mathscr {A}_{3}^{*}\right), \sigma \left(\mathscr {A}_{4}\right)=\sigma \left(\mathscr {A}_{7}\right), and σ(A4∗)=σ(A10)\sigma \left(\mathcal{A}_{4}^{*}\right)=\sigma \left(\mathcal{A}_{10}\right).

?
Remark.
?

Notation: A3=\mathscr {A}_{3}= intervals of the form [x,∞)[x, \infty ), A4=\mathscr {A}_{4}= intervals of the form (x,∞)(x, \infty ), A7=\mathscr {A}_{7}= intervals of the form [a,b)[a, b), and A10=\mathscr {A}_{10}= closed subsets of R\mathbb {R}. An asterisk on a collection (e.g. A3∗\mathscr {A}_{3}^{*}, A4∗\mathcal{A}_{4}^{*}) denotes the same form of interval restricted to rational endpoints.

Problem 1.4.2

All amc subsets of R\mathbb {R} are Borel sets. All subsets of R\mathbb {R} that differ from a Borel set by at most countably many points are Borel sets. That is, if the symmetric difference CΔBC \Delta B is amc and B∈BB \in \mathfrak {B}, then C∈BC \in \mathfrak {B}.

?
Problem 1.4.3

The Borel σ\sigma-field on (0,1](0,1] is denoted by B(0,1]\mathscr {B}_{(0,1]} and is defined as the σ\sigma-field on (0,1](0,1] generated by the rsc subintervals of (0,1].B(0,1](0,1] . \mathscr {B}_{(0,1]} may be equivalently defined by {B∩(0,1]∣B∈B}\left\{ B \cap (0,1] \mid B \in \mathscr {B}\right\}.

?
Problem 1.4.4

B\mathscr {B} is generated by the compact subsets of R\mathbb {R}.

?
Problem 1.4.5

B\mathscr {B} is not generated by the following:

?
(a)

Any finite collection of subsets of R\mathbb {R}.

(b)

The collection of real singletons.

(c)

The collection of all finite subsets of R\mathbb {R}.

(d)

The collection of all amc subsets of R\mathbb {R}.

Problem 1.4.6

The representations (1)-(5) give equivalent formulations of Bˉ\bar{\mathscr {B}}.

?
Remark.
?

Definition: the extended Borel σ\sigma-field on R‾\overline{\mathbb {R}} is denoted by Bˉ\bar{\mathscr {B}}, and is defined by Bˉ≡σ({(a,b]:a,b∈R‾,a<b})\bar{\mathscr {B}} \equiv \sigma \left(\left\{ (a, b]: a, b \in \overline{\mathbb {R}}, a<b\right\} \right). Sets in Bˉ\bar{\mathscr {B}} are called extended Borel sets.

Just as different authors use different definitions for B\mathscr {B}, so too Bˉ\bar{\mathscr {B}} has various definitions. Some other definitions are (1) the class of sets that are B\mathscr {B}-sets or are B\mathscr {B}-sets enlarged by one or both of ±∞\pm \infty, (2) the σ\sigma-field on R‾\overline{\mathbb {R}} generated by the class of intervals of the form (a,b)(a, b), [−∞,a)[-\infty , a), and (b,+∞](b,+\infty ], where a,b∈Ra, b \in \mathbb {R}, (3) the σ\sigma-field on R‾\overline{\mathbb {R}} generated by intervals of the form [−∞,x)[-\infty , x), x∈Rx \in \mathbb {R}, (4) σ(B∪{{−∞},{+∞}})\sigma \left(\mathscr {B} \cup \left\{ \left\{ -\infty \right\} ,\left\{ +\infty \right\} \right\} \right), and (5) the σ\sigma-field on R‾\overline{\mathbb {R}} generated by the open subsets of R‾\overline{\mathbb {R}}, sets of the form AA, A∪(x,∞]A \cup (x, \infty ], A∪[−∞,x)A \cup [-\infty , x), and A∪[−∞,x)∪(y,+∞]A \cup [-\infty , x) \cup (y,+\infty ], where AA is an open subset of R\mathbb {R} and x,y∈Rx, y \in \mathbb {R}.

Claim 2 (Section 1.4): B=σ(A1)=⋯=σ(A10)=σ(A1∗)=⋯=σ(A8∗)\mathscr {B}=\sigma \left(\mathcal{A}_{1}\right)=\cdots =\sigma \left(\mathcal{A}_{10}\right)=\sigma \left(\mathcal{A}_{1}^{*}\right)=\cdots =\sigma \left(\mathcal{A}_{8}^{*}\right), where A1,…,A10\mathcal{A}_{1}, \ldots , \mathcal{A}_{10} and A1∗,…,A8∗\mathcal{A}_{1}^{*}, \ldots , \mathcal{A}_{8}^{*} are the classes of intervals defined in the remark to problem 1.4.1 (with An∗\mathcal{A}_{n}^{*} the same form as An\mathcal{A}_{n} but with rational endpoints, for n=1,…,8n=1, \ldots , 8).

The natural extension of Claim 2 to R‾\overline{\mathbb {R}} holds in the extended Borel setting: Bˉ\bar{\mathscr {B}} is generated by any of the analogously-extended versions of A1,…,A10,A1∗,…,A8∗\mathcal{A}_{1}, \ldots , \mathcal{A}_{10}, \mathcal{A}_{1}^{*}, \ldots , \mathcal{A}_{8}^{*} as well.