TheoremComplete

The SchrΓΆder-Bernstein Theorem

The SchrΓΆder-Bernstein theorem is one of the most elegant and useful results in set theory. It provides a practical way to prove that two sets have the same cardinality by showing injections in both directions, which is often easier than constructing an explicit bijection.

TheoremSchrΓΆder-Bernstein Theorem

If there exist injections f:Aβ†ͺBf: A \hookrightarrow B and g:Bβ†ͺAg: B \hookrightarrow A, then there exists a bijection h:Aβ†’βˆΌBh: A \xrightarrow{\sim} B. Equivalently, if ∣Aβˆ£β‰€βˆ£B∣|A| \leq |B| and ∣Bβˆ£β‰€βˆ£A∣|B| \leq |A|, then ∣A∣=∣B∣|A| = |B|.

This theorem is remarkable because it allows us to establish equipollence without finding an explicit bijection, which can be difficult for complicated sets. The proof, while not trivial, is constructive and provides a method for building the bijection.

Proof

Let f:A→Bf: A \to B and g:B→Ag: B \to A be injections. We will construct a bijection h:A→Bh: A \to B.

Step 1: Partition A into layers

Define a sequence of subsets of AA by tracing ancestry through gg and ff:

  • Let C0=Aβˆ–g[B]C_0 = A \setminus g[B] be elements of AA that are not in the range of gg
  • For nβ‰₯0n \geq 0, let Cn+1=g[f[Cn]]C_{n+1} = g[f[C_n]] be the (n+1)(n+1)-th generation descended from C0C_0

Define C=⋃n=0∞CnC = \bigcup_{n=0}^\infty C_n to be all elements with finite ancestry.

Step 2: Define the bijection

Define h:A→Bh: A \to B by:

h(a)={f(a)ifΒ a∈Cgβˆ’1(a)ifΒ a∈Aβˆ–Ch(a) = \begin{cases} f(a) & \text{if } a \in C \\ g^{-1}(a) & \text{if } a \in A \setminus C \end{cases}

Here gβˆ’1g^{-1} denotes the inverse of gg restricted to its range (which is an injection, hence invertible on its range).

Step 3: Verify h is well-defined

For a∈Aβˆ–Ca \in A \setminus C, we claim a∈g[B]a \in g[B], so gβˆ’1(a)g^{-1}(a) is defined. Suppose not, i.e., a∈C0a \in C_0. Then a∈Ca \in C, contradicting a∈Aβˆ–Ca \in A \setminus C. For subsequent generations, if a∈Cna \in C_n for any nn, then a∈Ca \in C. Thus a∈g[B]a \in g[B].

Step 4: Verify h is injective

Suppose h(a1)=h(a2)h(a_1) = h(a_2). We consider cases:

  • If both a1,a2∈Ca_1, a_2 \in C: Then f(a1)=f(a2)f(a_1) = f(a_2), so a1=a2a_1 = a_2 since ff is injective
  • If both a1,a2∈Aβˆ–Ca_1, a_2 \in A \setminus C: Then gβˆ’1(a1)=gβˆ’1(a2)g^{-1}(a_1) = g^{-1}(a_2), so a1=a2a_1 = a_2 since gg is injective
  • If a1∈Ca_1 \in C and a2∈Aβˆ–Ca_2 \in A \setminus C (or vice versa): Then f(a1)=gβˆ’1(a2)f(a_1) = g^{-1}(a_2), which means a2=g(f(a1))∈g[f[C]]a_2 = g(f(a_1)) \in g[f[C]]. But elements of g[f[C]]g[f[C]] are in CC, contradicting a2∈Aβˆ–Ca_2 \in A \setminus C

Step 5: Verify h is surjective

Let b∈Bb \in B. We must find a∈Aa \in A with h(a)=bh(a) = b. Consider two cases:

  • If g(b)∈Cg(b) \in C: Then g(b)∈Cng(b) \in C_n for some nβ‰₯1n \geq 1, so g(b)∈g[f[Cnβˆ’1]]g(b) \in g[f[C_{n-1}]]. Thus g(b)=g(f(c))g(b) = g(f(c)) for some c∈Cnβˆ’1βŠ†Cc \in C_{n-1} \subseteq C. Since gg is injective, b=f(c)b = f(c). Set a=ca = c. Then a∈Ca \in C and h(a)=f(a)=f(c)=bh(a) = f(a) = f(c) = b

  • If g(b)βˆ‰Cg(b) \notin C: Set a=g(b)∈Aβˆ–Ca = g(b) \in A \setminus C. Then h(a)=gβˆ’1(a)=gβˆ’1(g(b))=bh(a) = g^{-1}(a) = g^{-1}(g(b)) = b

Therefore hh is a bijection.

β– 
Remark

The proof partitions AA into two parts: elements with "finite ancestry" (traced back through the images of ff and gg) and elements with "infinite ancestry." On the first part, we use ff; on the second part, we use the inverse of gg. This clever partition ensures both injectivity and surjectivity.

ExampleApplications of SchrΓΆder-Bernstein
  1. Countability of rationals: There are obvious injections Zβ†ͺQβ†ͺZΓ—N\mathbb{Z} \hookrightarrow \mathbb{Q} \hookrightarrow \mathbb{Z} \times \mathbb{N}, and ∣ZΓ—N∣=∣N∣|\mathbb{Z} \times \mathbb{N}| = |\mathbb{N}|, so ∣Q∣=∣N∣|\mathbb{Q}| = |\mathbb{N}|

  2. Open intervals: The interval (0,1)(0, 1) injects into R\mathbb{R} (obviously), and R\mathbb{R} injects into (0,1)(0, 1) via x↦1Ο€arctan⁑(x)+12x \mapsto \frac{1}{\pi}\arctan(x) + \frac{1}{2}. Thus ∣(0,1)∣=∣R∣|(0, 1)| = |\mathbb{R}|

  3. Power sets: For any infinite set AA, there is an injection Aβ†ͺP(A)A \hookrightarrow \mathcal{P}(A) via a↦{a}a \mapsto \{a\} and an injection P(A)β†ͺAΓ—P(A)\mathcal{P}(A) \hookrightarrow A \times \mathcal{P}(A) via X↦(a0,X)X \mapsto (a_0, X) for a fixed a0∈Aa_0 \in A. However, Cantor's theorem shows these cardinalities are actually different, so the hypothesis of SchrΓΆder-Bernstein fails

TheoremTrichotomy of Cardinals

For any sets AA and BB, exactly one of the following holds:

  1. ∣A∣<∣B∣|A| < |B|
  2. ∣A∣=∣B∣|A| = |B|
  3. ∣A∣>∣B∣|A| > |B|

This is the trichotomy law for cardinals. Its proof requires the axiom of choice (or equivalent principles).

The SchrΓΆder-Bernstein theorem, combined with trichotomy, gives a complete picture of cardinality comparisons. In practice, we often use SchrΓΆder-Bernstein to avoid constructing explicit bijections.

ExampleHilbert's Hotel Paradox

David Hilbert's famous "Grand Hotel" with infinitely many rooms (numbered 1,2,3,…1, 2, 3, \ldots) illustrates SchrΓΆder-Bernstein:

  • The hotel is full (one guest per room)
  • Infinitely many new guests arrive
  • Solution: Move guest in room nn to room 2n2n, freeing up all odd-numbered rooms

This works because Nβ†ͺEvenβŠ†N\mathbb{N} \hookrightarrow \text{Even} \subseteq \mathbb{N} via n↦2nn \mapsto 2n, and Evenβ†ͺN\text{Even} \hookrightarrow \mathbb{N} naturally. By SchrΓΆder-Bernstein, ∣Even∣=∣N∣|\text{Even}| = |\mathbb{N}|.

Remark

The SchrΓΆder-Bernstein theorem holds in ZF (without choice) and is strictly weaker than trichotomy, which requires AC. The theorem is also known as the Cantor-SchrΓΆder-Bernstein theorem or the Cantor-Bernstein theorem, reflecting its discoverers.

The SchrΓΆder-Bernstein theorem is indispensable in set theory and analysis. It allows us to reason about cardinalities in a flexible way, often simplifying proofs significantly. Understanding this theorem and its applications is essential for working with infinite sets.