Exercise 7.2.1

Give a direct proof of α + α = α by expressing ω α as a disjoint union of two sets of cardinality α .

Answers

Lemma 1. If α and β are ordinals and α > β then there is a γ < α such that | β | | γ | .

Proof. Clearly for γ = β we have γ = β < α and | β | = | γ | so that | β | | γ | is true. □

Lemma 2. If a set ( A , ) is isomorphic to ordinal α and α > β for another ordinal β then there is an a A such that | β | | X | for the set

X = { x A x a } .

Proof. First let f be the isomorphism from α to A . Clearly by Lemma 1 there is an ordinal γ < α such that | β | | γ | . Now let a = f ( γ ) so that clearly a A , and let

X = { x A x a } .

Now, we claim that X = f [ γ ] . So consider any x X so that x a . It then follows that f 1 ( x ) < f 1 ( a ) = γ since f 1 is an isomorphism since f is. Hence f 1 ( x ) γ so that clearly x = f ( f 1 ( x ) ) f [ γ ] . Since x was arbitrary it follows that X f [ γ ] . Now consider any x f [ γ ] so that there is a δ γ such that x = f ( δ ) . Then δ < γ so that x = f ( δ ) f ( γ ) = a since f is an isomorphism so that by definition x X . Hence f [ γ ] X . This shows that X = f [ γ ] . Then, since f is bijective, it follows that | β | | γ | = | f [ γ ] | = | X | . □

Lemma 3. For initial ordinals α and β , if | α | | β | , then α β .

Proof. Suppose that | α | | β | but that α > β . Then clearly β is isomorphic (and therefore equipotent) to an initial segment of α so that | β | | α | . Then by the Cantor-Bernstein Theorem we have | α | = | β | . However since α is an initial ordinal and β < α it cannot be that | α | = | β | . Thus we have a contradiction so that it must be that α β as desired. □

Lemma 4. For any ordinal α and any infinite initial ordinal Ω where Ω < ω α , there is a γ < α such that Ω = ω γ .

Proof. We show this by induction on α . For α = 0 we have ω α = ω 0 = ω so that there is no infinite initial ordinal Ω such that Ω < ω α = ω . Hence the hypothesis is vacuously true. Now suppose that, for every infinite initial ordinal Ω < ω α , there is a γ < α such that Ω = ω γ . Consider any infinite initial ordinal Ω < ω α + 1 . Then Ω < ω α + 1 = h ( ω α ) so that Ω is equipotent to some subset of ω α by the definition of the Hartogs number. From this it clearly follows that | Ω | | ω α | and hence Ω ω α by Lemma 3 since both Ω and ω α are initial ordinals. If Ω = ω α then we are finished but if Ω < ω α then by the induction hypothesis there is a γ < α such that Ω = ω γ so that we are also finished.

Now suppose that α is a nonzero limit ordinal and that for every β < α and infinite initial ordinal Ω < ω β there is a γ < β such that Ω = ω γ . Consider then any infinite initial ordinal Ω < ω α . Then since ω α = sup { ω β β < α } it follows that Ω is not an upper bound of { ω β β < α } so that there is a β < α such that Ω < ω β . But then by the induction hypothesis there is a γ < β such that Ω = ω γ . This completes the transfinite induction. □

Lemma 5. For ordinal α > 0 and an ordinal β < ω α there is an ordinal γ < α such that | β | γ .

Proof. First, if β is finite then clearly β < ω so that β ω . Then β ω since ω is transitive (since it is an ordinal number). Hence | β | | ω | = 0 (i.e. γ = 0 so that γ < α ). On the other hand if β is infinite then by Theorem 7.1.3 β is equipotent to some initial ordinal Ω . Clearly Ω is infinite since β is and clearly Ω < ω α since β < ω α and ω α is an initial ordinal. It then follows from Lemma 4 that there is a γ < α such that Ω = ω γ . Then we have | β | = | Ω | = | ω γ | = γ so that | β | γ is true. □

Lemma 6. Every infinite initial ordinal is a limit ordinal.

Proof. Suppose that α is an infinite initial ordinal and that it a successor so that α = β + 1 . It was shown in Lemma ?? that | β | = | β + 1 | = | α | , but since clearly β < α this contradicts the fact that α is an initial ordinal. Hence α must be a limit ordinal. □

Lemma 7. For well ordered sets A and B either | A | | B | or | B | | A | (or both in which case | A | = | B | ).

Proof. By Theorem 6.1.3 we have:

Case: A and B are isomorphic. Let f be the isomorphism from A to B . Then clearly f is a bijection so that | A | = | B | . Also since f is injective | A | | B | . Clearly also f 1 is bijective from B to A so that | B | | A | as well.

Case: A is isomorphic to an initial segment of B . Then if f is the isomorphism clearly f is an injective function from A to B so that | A | | B | .

Case: B is isomorphic to an initial segment of A . Then if f is the isomorphism clearly f is an injective function from B to A so that | B | | A | .

Since these cases are exhaustive by Theorem 6.1.3 clearly the result has been shown.

Note that this did not require the Axiom of Choice. □

Corollary 8. If A and B are well ordered sets then | A | | B | if and only if | B | < | A | .

Proof. ( ) Suppose that | A | | B | . Then it follows from Lemma 7 above that | B | | A | . Suppose that | B | = | A | . Then there is a bijection f from B to A . But then clearly f 1 is also a bijection and therefore injective. Hence by definition | A | | B | , a contradiction. So it cannot be that | B | = | A | . Hence | B | < | A | by definition as desired.

( ) We show this by proving the contrapositive. So suppose that | A | | B | . Also suppose that | B | | A | so that by Lemma 7 above | A | = | B | . Thus we have shown that

| B | | A | | A | = | B | | B | | A | | A | = | B | ¬ ( | B | | A | | A | | B | ) ¬ ( | B | < | A | ) ,

thereby showing the contrapositive. □

Main Problem.

The following proof is similar to the proof of Theorem 7.2.1.

Proof. Suppose that A 1 and A 2 are disjoint sets that are both equipotent to ω α for some ordinal α . Then there are bijections f 1 and f 2 from A 1 and A 2 , respectively, to ω α . We define a well-ordering of A = A 1 A 2 as follows: for a and b in A we let a b if and only if

  • a and b are in A 1 and f 1 ( a ) < f 1 ( b ) , or
  • a and b are in A 2 and f 2 ( a ) < f 2 ( b ) , or
  • a A 1 and b A 2 and f 1 ( a ) f 2 ( b ) , or
  • a A 2 and b A 1 and f 2 ( a ) < f 1 ( b ) .

First we show that is transitive. So consider a , b , and c in A such that a b and b c .

Case: a A 1

Case: b A 1

Case: c A 1 . Then f 1 ( a ) < f 1 ( b ) < f 1 ( c ) so that f 1 ( a ) < f 1 ( c ) and hence a c .

Case: c A 2 . Then f 1 ( a ) < f 1 ( b ) f 2 ( c ) so that f 1 ( a ) f 2 ( c ) is true and hence a c .

Case: b A 2

Case: c A 1 . Then f 1 ( a ) f 2 ( b ) < f 1 ( c ) so that f 1 ( a ) < f 1 ( c ) and hence a c .

Case: c A 2 . Then f 1 ( a ) f 2 ( b ) < f 2 ( c ) so that f 1 ( a ) f 2 ( c ) is true and hence a c .

Case: a A 2

Case: b A 1

Case: c A 1 . Then f 2 ( a ) < f 1 ( b ) < f 1 ( c ) so that f 2 ( a ) < f 1 ( c ) and hence a c .

Case: c A 2 . Then f 2 ( a ) < f 1 ( b ) f 2 ( c ) so that f 2 ( a ) < f 2 ( c ) and hence a c .

Case: b A 2

Case: c A 1 . Then f 2 ( a ) < f 2 ( b ) < f 1 ( c ) so that f 2 ( a ) < f 1 ( c ) and hence a c .

Case: c A 2 . Then f 2 ( a ) < f 2 ( b ) < f 2 ( c ) so that f 2 ( a ) < f 2 ( c ) and hence a c .

Since all cases imply that a c and a , b , and c were arbitrary this shows that is transitive.

Now we show that for any a and b in A , that either a b , a = b , or b a and that only one of these is true. So consider any a and b in A . Then we have

Case: a A 1

Case: b A 1 . Then clearly exactly one of the following is true: a b if f 1 ( a ) < f 1 ( b ) , a = b if f 1 ( a ) = f 1 ( b ) since f 1 is a bijection, and b a if f 1 ( b ) < f 1 ( a ) .

Case: b A 2 . Clearly a = b is not possible since A 1 and A 2 are disjoint. Then if f 1 ( a ) f 2 ( b ) then a b and if f 1 ( a ) > f 2 ( b ) then b a , noting that these are mutually exclusive.

Case: a A 2

Case: b A 1 . Then again clearly a = b is not possible since A 1 and A 2 are disjoint. Then if f 2 ( a ) < f 1 ( b ) then a b and if f 2 ( a ) f 1 ( b ) then b a , noting that these are mutually exclusive.

Case: b A 2 . Then clearly exactly one of the following is true: a b if f 2 ( a ) < f 2 ( b ) , a = b if f 2 ( a ) = f 2 ( b ) since f 2 is a bijection, and b a if f 2 ( b ) < f 2 ( a ) .

Thus we have shown that is a total (strict) order on A .

Now we show that is also a well-ordering. So let X be a nonempty subset of A . Let B = f 1 [ X A 1 ] f 2 [ X A 2 ] , noting that this is a nonempty set of ordinals. Then let B has a least element α .

Case: α f 1 [ X A 1 ] . Then we claim that x = f 1 1 ( α ) is the -least element of X , noting that x A 1 . Note also that clearly then f 1 ( x ) = f 1 ( f 1 1 ( α ) ) = α . So consider any y X .

Case: y A 1 . Then y X A 1 so that f 1 ( y ) f 1 [ X A 1 ] so f 1 ( y ) B . Thus f 1 ( x ) = α f 1 ( y ) since α is the least element of B . Clearly if f 1 ( x ) = f 1 ( y ) then x = y since f 1 is bijective. On the other hand if f 1 ( x ) < f 1 ( y ) then by definition x y . Hence in either case we have x y .

Case: y A 2 . Then y X A 2 so that f 2 ( y ) f 2 [ X A 2 ] so that f 2 ( y ) B . Thus f 1 ( x ) = α f 2 ( y ) so that by definition x y . Hence again x y is true.

Case: α f 1 [ X A 1 ] . Then it has to be that α f 2 [ X A 2 ] . Then we claim that x = f 2 1 ( α ) is the -least element of X , noting that x A 2 . Note also that clearly then f 2 ( x ) = f 2 ( f 2 1 ( α ) ) = α . So consider any y X .

Case: y A 1 . Then y X A 1 so that f 1 ( y ) f 1 [ X A 1 ] so f 1 ( y ) B . Thus f 2 ( x ) = α f 1 ( y ) since α is the least element of B . Now, it cannot be that f 2 ( x ) = α = f 1 ( y ) for then α would be in f 1 [ X A 1 ] . So it must be that f 2 ( x ) = α < f 1 ( y ) so that by definition x y so that x y is true.

Case: y A 2 . Then y X A 2 so that f 2 ( y ) f 2 [ X A 2 ] so that f 2 ( y ) B . Thus f 2 ( x ) = α f 2 ( y ) . Then if f 2 ( x ) = α = f 2 ( y ) then x = y since f 2 is bijective. On the other hand if f 2 ( x ) = α < f 2 ( y ) then x y by definition. In either case we have x y .

Hence in all cases we have shown that X has a -least element so that is a well-ordering of A .

Now we show by transfinite induction that α + α = α for all ordinals α . First it was already shown in a previous chapter that 0 + 0 = 0 . So now consider any α > 0 and suppose γ + γ = γ for all γ < α .

Then consider two disjoint sets A 1 and A 2 that are both equipotent to ω α and the well-ordering on A = A 1 A 2 as defined above, also again letting f 1 and f 2 be the isomorphisms from A 1 and A 2 , respectively, to ω α . Now let a be any element of A and define

X = { x A x a } .

Let X 1 = X A 1 and X 2 = X A 2 so that clearly X 1 and X 2 are disjoint and X = X 1 X 2 . From this it follows from the definition of cardinal addition that | X | = | X 1 | + | X 2 | .

If a A 1 then define β = f 1 ( a ) ω α so that β < ω α . It follows from this and Lemma 5 that there is a γ < α such that | β | γ since α > 0 , noting also that γ < α by the remarks following Definition 7.1.8. Now consider any x 1 X 1 so that also x 1 X . Then by definition x 1 a so that by the definition of we have f 1 ( x 1 ) < f 1 ( a ) = β since x 1 A 1 and a A 1 . Hence f 1 ( x 1 ) β so that f 1 [ X 1 ] β since x 1 was arbitrary. Hence, since f 1 is bijective, we have | X 1 | = | f 1 [ X 1 ] | | β | γ . Next, consider any x 2 X 2 so that x 2 X and hence x 2 a . Then, again by the definition of , we have that f 2 ( x 2 ) < f 1 ( a ) = β since x 2 A 2 and a A 1 . Hence f 2 ( x 2 ) β so that f 2 [ X 2 ] β since x 2 was arbitrary. Thus we have | X 2 | = | f 2 [ X 2 ] | | β | γ since f 2 is bijective.

A similar argument shows that | X 1 | γ and | X 2 | γ for some γ < α in the case when a A 2 . However, in this case we must set β = f 2 ( a ) + 1 , noting that β ω α since f 2 ( a ) ω α and ω α is a limit ordinal by Lemma 6.

Thus in all cases we have

| X | = | X 1 | + | X 2 | γ + γ (by property (d) of in section 5.1) = γ (by the induction hypothesis since  γ < α ) < α .

Thus we have shown that | X | < α = | ω α | for any a A , and hence | ω α | | X | by Corollary 8 since ω α and X are both well-ordered. If δ is the ordinal isomorphic to ( A , ) (which exists by Theorem 6.3.1 since we have shown that is a well-ordering), then it follows from the contrapositive of Lemma 2 that δ ω α and hence | A | = | δ | | ω α | = α . Thus we have

α + α = | A 1 | + | A 2 | = | A | α .

Since obviously 0 α it follows again from property (d) in section 5.1 that

α = α + 0 α + α .

Hence, by the Cantor-Bernstein Theorem we have that α = α + α , which completes the inductive step. □

User profile picture
2024-07-15 11:42
Comments