ℓ(I)=⎩⎨⎧b−a0∞I=(a,b),I=∅,I=(−∞,a),(a,∞), or (−∞,∞).
If we fully cover a set A with open intervals in the most conservative way possible, the length of all intervals in the cover would seem to be a good notion of “size” for A. And so it is.
Suppose A={a1,a2,…} is a countable subset of R. Fix some ε>0. For each k∈Z+, define Ik=(ak−ε/2k,ak+ε/2k). Then note
μ∗(A)≤k=1∑∞ℓ(Ik)=2εk=1∑∞2k1=2ε.
Since ε is arbitrary, we have that μ∗(A)=0. ❦
[0.3.4]Theorem(Outer Measure Preserves Set Inclusion Order)#
Suppose A,B are subsets of R such that A⊆B. Then μ∗(A)≤μ∗(B).
[0.3.5]Theorem(Outer Measure is Translation Invariant)#
Let A be any subset of R and t any arbitrary number in R. Then μ∗(A)=μ∗(t+A).
Let (1,4) and (3,5) be open intervals, whose union is (1,5). Then note that
μ∗((1,4)∪(3,5))<μ∗((1,4))+μ∗((3,5)),
because the LHS is 4 and the RHS is 5. The intuition behind this is that the RHS counts (3,4) twice, whereas LHS counts it only once. We will generalize this to show the outer measure on R is countably subadditive.
[0.3.6]Theorem(Outer Measure is Countably Subadditive)#
Suppose A1,A2,… is a sequence of subsets of R. Then
μ∗(k=1⋃∞Ak)≤k=1∑∞μ∗(Ak).
Proof.
If μ∗(Ak)=∞ for any k∈Z+, the result would be immediate, so suppose μ∗(Ak)<∞ for every k∈Z+. Fix ε>0. For each k∈Z+, define a sequence of open intervals A1,k,A2,k,… that cover Ak in such a way that
j=1∑∞ℓ(Aj,k)≤2kε+μ∗(Ak).
Simply consider the collection {Aj,k}j,k∈Z+. Note the sum over these indices yields
k=1∑∞j=1∑∞ℓ(Aj,k)≤ε+k=1∑∞μ∗(Ak).
Note that this collection also covers ⋃k=1∞Ak by construction, and so
Countable subadditivity implies finite subadditivity, since we can let Ak=∅ for every k>n, then we have
μ∗(k=1∑nAk)≤k=1∑nμ∗(Ak).
[0.3.7]Theorem(Outer Measure of Closed Intervals)#
μ∗([a,b])=b−a.
Proof.
Note that, for any ε>0, the sequence of open sets (a−ε,b+ε),∅,∅,… forms an open cover for [a,b], and so
μ∗([a,b])≤(b−a)+2ε≤b−a.
We simply show that μ∗([a,b])≥b−a. Let {In}n∈Z+ be any countable open cover for [a,b]. By compactness of [a,b], there exists a finite subcover {Ik}k=1n. We prove by induction that
k=1∑nμ∗(Ik)≥b−a,
and so ∑k=1∞μ∗(Ik)≥∑k=1nμ∗(Ik)≥b−a, meaning μ∗([a,b])=b−a.
If [a,b]⊆I1, then μ∗(I1)≥b−a, of course. Now suppose we know that, if [a,b]=I1∪I2∪⋯∪In, then
k=1∑nμ∗(Ik)≥b−a.
Suppose [a,b]⊆I1∪⋯∪In+1. Write In+1=(c,d), and without loss of generality suppose b∈In+1. Note that if c≤a, then μ∗(In+1)≥b+a, so we are done. Thus suppose a<c<b<d, and so [a,c]⊆I1∪⋯∪In. By inductive hypothesis, note that
Every interval in R that contains at least two distinct elements is uncountable.
Proof.
Suppose I is an interval containing a,b∈R with a<b. Then
μ∗(I)≥μ∗([a,b])=b−a=0.
Countable sets are measure zero, and so I is not countable. ❦
The outer measure does have some unpleasant properties, however. We would expect that, for any disjoint sets A,B, the outer measure of A∪B is equal to the sum of the outer measure of A and B respectively. Unfortunately, this is not true in general.
There exist disjoint subsets A,B⊆R such that μ∗(A∪B)=μ∗(A)+μ∗(B).
Proof.
Consider the set [−1,1]. For any a,b∈[−1,1], define the equivalence relation
a∼b⟺a−b∈Q.
Then the equivalence relation ∼ partitions [−1,1] in such a way that
[−1,1]=a∈[−1,1]⋃[a]∼.
For each coset, choose exactly one point, then let V be the set of all these points. That is, V∩[a]∼ has a single, unique element for each a∈[−1,1].
Let r1,r2,… be a sequence of points such that [−2,2]∩Q={r1,r2,…}. Then note
[−1,1]⊆k=1⋃∞(rk+V).
Indeed, for any a∈[−1,1], note there is some v∈V∩[a]∼. Then a−v∈Q, meaning there is some k∈Z+ such that a=rk+v, meaning a∈rk+V as desired. By order properties and countable subadditivity of the outer measure, note
since V⊆[−1,1] and rk∈[−2,2] for each k. Again, by order properties of the outer measure, note
μ∗(k=1⋃n(rk+V))≤k=1∑nμ∗(V)≤6.
But also note that
k=1∑nμ∗(V)=nμ∗(V).
Simply choose n∈Z+ such that nμ∗(V)>6, and so
μ∗(k=1⋃n(rk+V))≤k=1∑nμ∗(rk+V).
Note each rk+V is disjoint. Since μ∗(A1∪⋯∪An)=μ∗(A1)+⋯+μ∗(An) in general for disjoint A1,…,An, by induction we have μ∗(A∪B)=μ∗(A)+μ∗(B) in general. ❦
[0.3.10]Remark(Intuition behind Disproof of Additivity of Outer Measure)#
We start by taking [−1,1] and forming equivalence classes that group numbers that differ by a rational together. From each equivalence class, we choose one point, and we take the set of these points to be V. Crucially, note that no elements of V differ by a rational, otherwise two elements of V would be in the same equivalence class, contrary to construction. We then make copies of the set, shifting V by all the rationals in [−2,2], written as a sequence r1,r2,…. The translates r1+V,r2+V,… are all disjoint from eachother, since if they overlapped, there would be two numbers in V that differed by a rational. We will refer to rk+V as a “copy” of V, since it is still isomorphic to V.
If we take infinitely many of these copies and union them together, we note it forms a cover for [−1,1]. Thus the outer measure of this union is greater than 2, establishing the outer measure of V is nonzero.
Now, we look at finitely many copies of V, say n of them. The union of n many copies of V lie in [−3,3], and so the outer measure of this union is bounded above by 6. But the sum of n many outer measures of V is given by n multiplied by the outer measure of V, which is nonzero. This sum grows without bound, so there exists some n such that nμ∗(V)>6, completing the disproof.