Section 4.3

Translation Invariance of Lebesgue Measure

[11/64%]
Show
LaTeX
Problem 4.3.1

Let μ:Bk→R‾\mu : \mathscr {B}^{k} \rightarrow \overline{\mathbb {R}} denote a measure that assigns finite measure to bounded Bk\mathscr {B}^{k}-sets. Furthermore, suppose that μ\mu is translation invariant: μ(B⊕x)=μ(B)\mu (B \oplus \boldsymbol {x})=\mu (B) for all x∈Rk\boldsymbol {x} \in \mathbb {R}^{k}. Then there is c>0c>0 such that μ(B)=cλk(B)\mu (B)=c \lambda_{k}(B) for all B∈BkB \in \mathscr {B}^{k}.

?
Problem 4.3.2

Let A∈LA \in \mathscr {L} be such that λˉ(A)>0\bar{\lambda }(A)>0, and let c∈[0,1)c \in [0,1). There exists an open interval UU such that λˉ(A∩U)⩾cλ(U).1\bar{\lambda }(A \cap U) \geqslant c \lambda (U) .^{1}

?
Problem 4.3.3

Let A∈LA \in \mathscr {L} contain an open interval. Then there exists a>0a>0 such that (−a,a)(-a, a) is contained in D(A)={x−y:x,y∈A}D(A)=\left\{ x-y: x, y \in A\right\}.

?
Problem 4.3.4

Let A∈LA \in \mathscr {L} be such that λˉ(A)>0\bar{\lambda }(A)>0. Then there exists a>0a>0 such that (−a,a)(-a, a) is contained in D(A)={x−y:x,y∈A}D(A)=\left\{ x-y: x, y \in A\right\}.

?
Problem 4.3.5

Let AA be a dense subset of R\mathbb {R}. Then cA={ca:a∈A}c A=\left\{ c a: a \in A\right\} is dense for any c≠0c \neq 0.

?
Problem 4.3.6

Let ξ\xi be an irrational number.

?
(a)

Let A={n+mξ:n,m∈Z}A=\left\{ n+m \xi : n, m \in \mathbb {Z}\right\}. Then AA is a dense subset of R\mathbb {R}.

(b)

Let B=\left\{ n+m \xi : n, m \in \mathbb {Z}, n even \right\}. Then BB is a sense subset of R\mathbb {R}.

(c)

Let C=\left\{ n+m \xi : n, m \in \mathbb {Z}, n odd \right\}. Then CC is a sense subset of R\mathbb {R}.

Problem 4.3.7

For x,y∈Rx, y \in \mathbb {R} write x∼yx \sim y iff x−y∈A={n+mξ:n,m∈Z}x-y \in A=\left\{ n+m \xi : n, m \in \mathbb {Z}\right\}, where ξ\xi is a fixed irrational number as in the previous exercise. Then ∼\sim is an equivalence relation, and hence R\mathbb {R} may be partitioned into disjoint equivalence classes.

?
Problem 4.3.8

We now invoke ( ACA C ) to form a set E0E_{0} consisting of exactly one element from each of the equivalence classes in the previous exercise. We will now show that E0∉LE_{0} \notin \mathscr {L}.

?
(a)

There exist Borel subsets of E0E_{0}.

(b)

Let F⊆E0F \subseteq E_{0} be a Borel set. Then D(F)D(F) cannot contain any nonzero elements of AA, where AA is the set in Exercise 4.3.6.

(c)

By (b), there cannot exist an open interval containing the origin that is contained in D(F)D(F), hence λ(F)=0\lambda (F)=0.

(d)

From (c), we have λ∗(E0)=0\lambda_{*}\left(E_{0}\right)=0.

(e)

If a1a_{1} and a2a_{2} are distinct elements of A={n+mξ:m,n∈Z}A=\left\{ n+m \xi : m, n \in \mathbb {Z}\right\}, then E0⊕a1E_{0} \oplus a_{1} and E0⊕a2E_{0} \oplus a_{2} are disjoint.

(f)

R=⋃{E0⊕a:a∈A}\mathbb {R}=\bigcup \left\{ E_{0} \oplus a: a \in A\right\}, the countable union being disjoint.

(g)

If E0∈LE_{0} \in \mathscr {L}, then λˉ(E0⊕a)=0\bar{\lambda }\left(E_{0} \oplus a\right)=0 for each a∈Aa \in A, hence λˉ(R)=0\bar{\lambda }(\mathbb {R})=0. Therefore, since the assumption that E0∈LE_{0} \in \mathscr {L} leads to an absurdity, it must be the case that E0∉LE_{0} \notin \mathscr {L}, and hence there exists a subset of R\mathbb {R} that fails to be a Lebesgue set.

Problem 4.3.9

In conjunction with the previous exercises, it is not possible to have Lebesgue measure defined for every subset of R\mathbb {R}. That is, a measure extension μ\mu of ΔF\Delta_{F} [where F∈SF \in \mathfrak {S} is the identity function] from the semiring A\mathcal{A} consisting of ∅\varnothing and the bounded rsc intervals (a,b](a, b] to 2R2^{\mathbb {R}} fails to exist.

?
Remark.
?

Argue by contradiction for a quick answer: if such a measure extension μ\mu to 2R2^{\mathbb {R}} exists then the previous work would force μ(R)=0\mu (\mathbb {R})=0, contradicting what we expect of length.

Problem 4.3.10

Let B∈LB \in \mathscr {L} be such that λˉ(B)>0\bar{\lambda }(B)>0. We will show that there is E1⊆BE_{1} \subseteq B with E1∉LE_{1} \notin \mathscr {L} by following the outline below. The set AA appearing below is the set AA from Exercise 4.3.7, consisting of the numbers n+mξn+m \xi, where n,m∈Zn, m \in \mathbb {Z} and ξ\xi is a fixed irrational number.

?
(a)

λ∗\lambda_{*} is monotone and translation invariant.

(b)

B=⋃{(E0⊕a)∩B:a∈A}B=\bigcup \left\{ (E_{0} \oplus a) \cap B: a \in A\right\}, the union being disjoint.

(c)

There exists a∗∈Aa^{*} \in A such that λ∗((E0⊕a∗)∩B)>0\lambda^{*}\left((E_{0} \oplus a^{*}) \cap B\right)>0.

(d)

Let E1=(E0⊕a∗)∩BE_{1}=(E_{0} \oplus a^{*}) \cap B. Then E1⊆BE_{1} \subseteq B and E1∉LE_{1} \notin \mathscr {L}.

(e)

We may actually generalize further and state that any subset B⊆RB \subseteq \mathbb {R} with λ∗(B)>0\lambda^{*}(B)>0 [BB not necessarily in L\mathscr {L}] contains a subset outside of L\mathscr {L}.

Remark.
?

By (b), we have 0<λˉ(B)=λ∗(B)⩽∑{λ∗((E0⊕a)∩B):a∈A}0<\bar{\lambda }(B)=\lambda^{*}(B) \leqslant \sum \left\{ \lambda^{*}((E_{0} \oplus a) \cap B): a \in A\right\}; and by (a) and part (d) of Exercise 4.3.8, λ∗(E1)=λ∗((E0⊕a∗)∩B)⩽λ∗(E0⊕a∗)=λ∗(E0)=0\lambda_{*}(E_{1})=\lambda_{*}((E_{0} \oplus a^{*}) \cap B) \leqslant \lambda_{*}(E_{0} \oplus a^{*})=\lambda_{*}(E_{0})=0, while at the same time λ∗(E1)>0\lambda^{*}(E_{1})>0; invoke the relevant part of the Approximation Theorem.

Problem 4.3.11

Let A∈LA \in \mathscr {L}, let BB be a dense subset of R\mathbb {R}, and suppose that λˉ(A△(A⊕x))=0\bar{\lambda }(A \triangle (A \oplus x))=0 for every x∈Bx \in B. Then either λˉ(A)=0\bar{\lambda }(A)=0 or λˉ(Ac)=0\bar{\lambda }(A^{c})=0. This exercise is independent of the previous exercises.

?