Homepage › Solution manuals › Terence Tao › An Introduction to Measure Theory › Exercise 1.4.4 (Finite Boolean algebra is atomic)

Exercise 1.4.4 (Finite Boolean algebra is atomic)

Let X be a set, and let ℬ be a Boolean algebra over X. Show that if ℬ is finite, then it is atomic and has cardinality of 2n for some n ∈ ℕ.

Answers

Since B is finite with N := #B, we can enumerate its elements as (An)n=1N. Define

∀ ⁡1 ≤ n ≤ N : An′ := A n ∖ ⋃ i∈{1,…,N}∖{n}Ai.

Sorting out the empty sets, consider

B := {An′ : 1 ≤ n ≤ N,(A n′≠∅)}.

We argue that the elements of this set, enumerated as (Ak′)k=1N′ , are atoms that generate B. We verify the atom criteria from Example 1.4.7. First of all, it is obvious that the elements A1′,…,AN′′ are disjoint.

∀ ⁡1 ≤ n < m ≤ N′ : A n′∩A m′ = (A n ∖ ⋃ i∈{1,…,N}∕{k}Ai)∩ (Am ∖ ⋃ i∈{1,…,N}∕{m}Ai) ⊆ (An∕Am)∩ (Am∕An) = ∅.

Also, it is easy to verify that their union gives us the ambient space

⋃ i=1N′ An′ = ⋃ i=1NA n = ⋃ B.

Finally, any set of B can be represented as a union of some of these atom sets. To see why, pick an arbitrary An ∈B,An≠∅. Then

An = ⋃ i∈KnAi′where K n = {i ∈ ℕ : 1 ≤ i ≤ N′,A i ∩ An≠∅}.

We shortly verify this. If x ∈ An then either x ∈ An′ or x ∈⋃ ⁡ i≠nAi. In the former case, x is contained in the right-hand side due to An ∩ An′≠∅. In the former case, there exists i≠n such taht x ∈ Ai and thus Ai ∩ An≠∅. Now pick an arbitrary x from the right-hand side. Then x ∈ Ak such that Ak ∩ An≠∅. But Ak′ := Ak ∖ ⋃ ⁡ i∈{1,…,N}∕{k}Ai, which is only possible when n = k, and we have x ∈ An.

User profile picture
2020-07-16 00:00
Comments

Let B be a finite Boolean algebra, so we can model B as a sequence of its elements ( E i ) i = 1 n .

Define the set F ′ = { F : F ∈ B ∣ F ∩ E = ∅  or  F ∩ E = F  for all  E ∈ B } . Since F ′ ⊆ B , it is also finite. We claim F ′ is a partition of X .

The elements being disjoint is enforced by the conditions set.

Let x be an arbitrary element of X . Since B is a Boolean algebra, it covers X , and there must be an element E 0 ∈ B such that x ∈ E 0 . Define the function f : B × B → B as follows:

f ( A , B ) = { A ∩ B , if  x ∈ A ∩ B , A ∖ B , if  x ∉ A ∩ B .

Then the sequence G 0 = E 0 and G i = f ( G i − 1 , E i ) is well defined and finite, ending with n .

Since we started with an element containing x , G n must also contain x . Now assume to the contrary there is an E different from G n such that E ∩ G n ≠ ∅ . Then E ∩ G n ∈ B , and so is G n ∖ E . x is in either one of them, and then at some point the algorithm would have chosen one of these sets (or one contained in them) over G n , a contradiction.

Thus, G n ∈ F ′ , and F ′ covers X and thus is a partition. Since it is finite, we can enumerate it and call it ( F α ) α ∈ I with I = { 1 , … , m } for some m ∈ ℕ .

Now we have to prove that every element of B is in A ( ( F α ) α ∈ I ) .

Consider an arbitrary E ∈ B . Since F ′ covers X , it also covers E in a disjoint manner. Let F E = ⋃ ⁡ α ∈ I F α  where  F α ∩ E ≠ ∅ , and assume to the contrary F E ≠ E . As F ′ covers X , we must have E ⊂ F E . But then there must exist at least one F i such that F i ∩ E ⊂ F i and F i ∩ E ≠ ∅ , a contradiction of F i being a member of F ′ . So E = F E , and thus E ∈ A ( ( F α ) α ∈ I ) .

Since every F i ∈ B , which is closed under union, every ⋃ ⁡ α ∈ J F α , with J ⊆ I , is in B , and both algebras are equal.

User profile picture
2024-10-17 12:42
Comments