Part V — Chapter 18

Cantor & Dedekind

Taming the infinite: how two visionary mathematicians constructed the real numbers, revealed the hierarchy of infinities, and laid the foundations of modern analysis and set theory.

18.1 — Dedekind: Life and Education

Julius Wilhelm Richard Dedekind (1831–1916) was born in Braunschweig, Germany, the youngest of four children of a professor of law. From an early age he displayed a talent for logical reasoning, initially studying chemistry and physics before turning to mathematics at the University of Göttingen in 1850.

Key Dates in Dedekind's Life

  • 1831 — Born in Braunschweig, Duchy of Brunswick
  • 1850 — Enters the University of Göttingen
  • 1852 — Earns his doctorate under Gauss with a thesis on Eulerian integrals
  • 1854 — Habilitation; begins lecturing at Göttingen
  • 1855 — Gauss dies; Dirichlet succeeds him; Dedekind attends Dirichlet's lectures
  • 1858 — Appointed to the Polytechnikum in Zürich; conceives the idea of Dedekind cuts
  • 1862 — Returns to Braunschweig as professor at the Technische Hochschule
  • 1872 — Publishes Stetigkeit und irrationale Zahlen (Continuity and Irrational Numbers)
  • 1888 — Publishes Was sind und was sollen die Zahlen? (What are numbers and what should they be?)
  • 1916 — Dies in Braunschweig at the age of 84

At Göttingen, Dedekind had the extraordinary fortune to study under the last of the great classical masters: Carl Friedrich Gauss, who supervised his doctoral thesis on Eulerian integrals, and then Peter Gustav Lejeune Dirichlet, whose crystal-clear exposition of analysis and number theory profoundly influenced Dedekind's mathematical style. Dedekind later edited and published Dirichlet's lectures on number theory (the famous Vorlesungen über Zahlentheorie), adding his own groundbreaking supplements on ideal theory.

Dedekind was a quiet, methodical thinker who never married and lived modestly with his sister. He spent nearly his entire career at the Technische Hochschule in Braunschweig — a relatively minor institution — yet produced some of the most far-reaching foundational work of the nineteenth century. His emphasis on structural, conceptual definitions over computational formulas anticipated the abstract algebra of the twentieth century.

18.2 — Dedekind Cuts and the Construction of the Reals

In 1858, while teaching calculus in Zürich, Dedekind realized that the standard presentation of the real number line relied on geometric intuition rather than rigorous arithmetic foundations. The result of his reflections, published in 1872, was one of the most elegant constructions in all of mathematics: the Dedekind cut.

Definition: Dedekind Cut

A Dedekind cut is a partition of the rational numbers $\mathbb{Q}$ into two non-empty sets $A$ and $B$ such that:

  1. $A \cup B = \mathbb{Q}$ and $A \cap B = \emptyset$
  2. If $a \in A$ and $b \in B$, then $a < b$
  3. $A$ contains no greatest element

The set of all such cuts is defined to be the set of real numbers $\mathbb{R}$. The cut $(A, B)$ represents the real number that “separates” $A$ from $B$.

The key insight is that every real number, whether rational or irrational, can be uniquely identified with the set of all rationals less than it. For a rational number $q$, the corresponding cut is $A = \{ r \in \mathbb{Q} : r < q \}$. For an irrational number like $\sqrt{2}$, the cut captures it exactly even though no rational number “sits” at the boundary.

Example: The Dedekind Cut for the Square Root of 2

Define the cut $(A, B)$ as follows:

$$A = \{ r \in \mathbb{Q} : r < 0 \text{ or } r^2 < 2 \}$$

$$B = \{ r \in \mathbb{Q} : r > 0 \text{ and } r^2 \geq 2 \}$$

Since $\sqrt{2}$ is irrational, no rational number $r$ satisfies $r^2 = 2$. Thus $A$ has no greatest element: for any $a \in A$ with $a > 0$ and $a^2 < 2$, we can always find a larger rational $a' \in A$ with $a < a'$ and $(a')^2 < 2$.

Proof that A has no maximum: Given $a \in A$ with $a > 0$ and $a^2 < 2$, let $\varepsilon = \frac{2 - a^2}{2a + 1}$. Then $a' = a + \varepsilon$ satisfies:

$$(a')^2 = a^2 + 2a\varepsilon + \varepsilon^2 < a^2 + (2a+1)\varepsilon = a^2 + (2 - a^2) = 2$$

So $a' \in A$ and $a' > a$, confirming $A$ has no greatest element. This cut therefore represents the real number $\sqrt{2}$.

Arithmetic on Cuts

Dedekind showed that the arithmetic operations on real numbers can be defined directly in terms of cuts. Given two cuts $\alpha = (A_1, B_1)$ and $\beta = (A_2, B_2)$:

Addition of Dedekind Cuts

The sum $\alpha + \beta$ is the cut $(A, B)$ where:

$$A = \{ a_1 + a_2 : a_1 \in A_1, \, a_2 \in A_2 \}$$

and $B = \mathbb{Q} \setminus A$. One must verify that this indeed defines a valid Dedekind cut: $A$ is non-empty, $B$ is non-empty, every element of $A$ is less than every element of $B$, and $A$ has no greatest element.

Multiplication of Dedekind Cuts

For cuts $\alpha, \beta > 0$ (meaning $0 \in A_1$ and $0 \in A_2$), the product $\alpha \cdot \beta$ is the cut $(A, B)$ where:

$$A = \{ r \in \mathbb{Q} : r < a_1 \cdot a_2 \text{ for some } a_1 \in A_1, a_2 \in A_2 \text{ with } a_1, a_2 > 0 \} \cup \{ r \in \mathbb{Q} : r \leq 0 \}$$

The cases where one or both cuts are negative are handled by reducing to the positive case via sign rules. This makes $(\mathbb{R}, +, \cdot)$ into an ordered field.

The Completeness Property

Theorem: Completeness of the Dedekind Reals

The real number system $\mathbb{R}$, constructed via Dedekind cuts, satisfies the Least Upper Bound Property: every non-empty subset of $\mathbb{R}$ that is bounded above has a least upper bound (supremum) in $\mathbb{R}$.

Proof: Let $S \subseteq \mathbb{R}$ be non-empty and bounded above. Each element of $S$ is a cut $(A_s, B_s)$. Define:

$$A^* = \bigcup_{s \in S} A_s$$

and $B^* = \mathbb{Q} \setminus A^*$. We verify that $(A^*, B^*)$ is a valid Dedekind cut:

  • Non-empty: $A^*$ is non-empty because $S$ is non-empty, so some $A_s$ is non-empty. $B^*$ is non-empty because $S$ is bounded above by some cut $(A_M, B_M)$, so $A^* \subseteq A_M$ and $B^* \supseteq B_M \neq \emptyset$.
  • Ordered: If $a \in A^*$ and $b \in B^*$, then $a \in A_s$ for some $s$, and $b \notin A_s$ (since $b \notin A^*$), so $b \in B_s$, hence $a < b$.
  • No maximum: If $a \in A^*$, then $a \in A_s$ for some $s$. Since $A_s$ has no maximum, there exists $a' \in A_s \subseteq A^*$ with $a' > a$.

Now $(A^*, B^*)$ is an upper bound for $S$: for every $s \in S$,$A_s \subseteq A^*$, so the cut $s \leq (A^*, B^*)$. And it is theleast upper bound: if $(A', B')$ is any upper bound with $A' \subsetneq A^*$, then some $a \in A^* \setminus A'$ exists, but $a \in A_s$ for some $s$, contradicting that $(A', B')$ is an upper bound for $S$. $\blacksquare$

This completeness property is precisely what distinguishes the reals from the rationals. The rationals have “gaps” — for instance, the set of all rationals whose square is less than 2 has no rational least upper bound — and Dedekind's construction fills in every such gap.

Example: Embedding the Rationals

Every rational number $q \in \mathbb{Q}$ corresponds to the cut:

$$q^* = \bigl(\{ r \in \mathbb{Q} : r < q \}, \, \{ r \in \mathbb{Q} : r \geq q \}\bigr)$$

The map $q \mapsto q^*$ is an order-preserving injection that respects addition and multiplication: $(p + q)^* = p^* + q^*$ and $(pq)^* = p^* \cdot q^*$. Thus $\mathbb{Q}$ embeds as a dense subfield of $\mathbb{R}$.

18.3 — Dedekind's Theory of Ideals

Dedekind's second great contribution was to algebraic number theory. The problem of unique factorization in number rings had been a central concern since Kummer's work on Fermat's Last Theorem. In rings like $\mathbb{Z}[\sqrt{-5}]$, unique factorization of elements fails: for example,

$$6 = 2 \cdot 3 = (1 + \sqrt{-5})(1 - \sqrt{-5})$$

where all four factors are irreducible. Kummer had introduced “ideal numbers” to restore uniqueness, but his approach was somewhat ad hoc. Dedekind reformulated the entire theory in terms of ideals — certain subsets of the ring itself.

Definition: Ideal

Let $R$ be a commutative ring. A subset $I \subseteq R$ is an ideal if:

  1. $I$ is non-empty (in particular, $0 \in I$)
  2. If $a, b \in I$, then $a - b \in I$ (closure under subtraction)
  3. If $a \in I$ and $r \in R$, then $ra \in I$ (absorption)

Example: Failure of Unique Factorization Restored by Ideals

In $\mathbb{Z}[\sqrt{-5}]$, consider the ideals:

$$\mathfrak{p}_1 = (2, 1+\sqrt{-5}), \quad \mathfrak{p}_2 = (2, 1-\sqrt{-5})$$

$$\mathfrak{p}_3 = (3, 1+\sqrt{-5}), \quad \mathfrak{p}_4 = (3, 1-\sqrt{-5})$$

Then the ideal $(6)$ factors uniquely as:

$$(6) = \mathfrak{p}_1 \mathfrak{p}_2 \mathfrak{p}_3 \mathfrak{p}_4$$

Even though elements do not factor uniquely, ideals do. This is Dedekind's fundamental insight.

Theorem: Unique Factorization of Ideals (Dedekind, 1871)

In the ring of integers $\mathcal{O}_K$ of an algebraic number field $K$, every non-zero ideal factors uniquely (up to order) as a product of prime ideals:

$$I = \mathfrak{p}_1^{e_1} \mathfrak{p}_2^{e_2} \cdots \mathfrak{p}_r^{e_r}$$

A ring with this property is now called a Dedekind domain. This theorem generalized the Fundamental Theorem of Arithmetic from $\mathbb{Z}$ to all rings of algebraic integers.

Dedekind's approach was revolutionary because it shifted the focus from individual elements to structural properties of sets of elements. This conceptual leap — thinking in terms of abstract structures rather than concrete calculations — would become the hallmark of twentieth-century algebra, influencing Emmy Noether, Emil Artin, and the Bourbaki school.

Definition: Prime Ideal

An ideal $\mathfrak{p}$ of a ring $R$ is prime if $\mathfrak{p} \neq R$ and whenever $ab \in \mathfrak{p}$, either $a \in \mathfrak{p}$ or $b \in \mathfrak{p}$. This generalizes the notion of a prime number: in $\mathbb{Z}$, the ideal $(p)$ is prime if and only if $p$ is a prime number.

18.4 — Cantor: Life and Early Work

Georg Ferdinand Ludwig Philipp Cantor (1845–1918) was born in Saint Petersburg, Russia, to a Danish father and a Russian mother from a musical family. When he was eleven his family moved to Germany, and he studied at the universities of Zürich, Göttingen, and Berlin, receiving his doctorate from Berlin in 1867 under Ernst Kummer and Karl Weierstrass.

Key Dates in Cantor's Life

  • 1845 — Born in Saint Petersburg, Russia
  • 1856 — Family moves to Germany
  • 1867 — Doctorate from Berlin on number theory (ternary quadratic forms)
  • 1869 — Appointed to the University of Halle
  • 1872 — Uniqueness theorem for trigonometric series; meets Dedekind
  • 1874 — First proof that the reals are uncountable
  • 1878 — Shows bijection between line and plane; correspondence with Dedekind
  • 1884 — First serious episode of depression
  • 1891 — Publishes the diagonal argument
  • 1895–97 — Publishes Beiträge zur Begründung der transfiniten Mengenlehre
  • 1918 — Dies in a sanatorium in Halle

Cantor's early mathematical work was on number theory, but he soon turned to analysis, specifically the theory of trigonometric series. The question that launched his revolution was deceptively simple: under what conditions does a trigonometric series representation of a function have to be unique?

Uniqueness of Trigonometric Series

In 1870, Heine had asked: if two trigonometric series converge to the same function, must their coefficients be equal? Cantor proved the answer is yes: if

$$\frac{a_0}{2} + \sum_{n=1}^{\infty} (a_n \cos nx + b_n \sin nx) = 0$$

for all $x$, then all coefficients $a_n = b_n = 0$. Cantor then extended this result to allow a finite number of exceptional points, then a countably infinite set of exceptional points, provided the set had a particular structure. This investigation led him to study the “derived sets” of point sets — the beginnings of point-set topology — and ultimately to the theory of infinite sets.

Definition: Derived Set

Given a point set $P \subseteq \mathbb{R}$, its derived set $P'$ is the set of all limit points (accumulation points) of $P$. One can iterate this operation: $P'' = (P')'$, $P''' = (P'')'$, and so on. Cantor defined the transfinite iterates $P^{(\alpha)}$ for ordinal numbers $\alpha$, pioneering the study of what we now call the Cantor–Bendixson derivative.

It was during this work that Cantor began his fateful correspondence with Dedekind in 1872. Their exchange of letters — warm, intellectually vibrant, and deeply collaborative — would produce some of the most consequential ideas in the history of mathematics.

18.5 — Cantor's Theory of Infinite Sets

The central idea of Cantor's set theory is that the “size” of an infinite set can be measured by placing it in one-to-one correspondence (a bijection) with other sets. Two sets have the same cardinality if there exists a bijection between them.

Definition: Countable Set

A set $S$ is countably infinite if there exists a bijection $f: \mathbb{N} \to S$. A set is countable if it is finite or countably infinite. A set that is not countable is uncountable.

Example: The Integers and Rationals are Countable

The integers $\mathbb{Z}$ are countable via the bijection:

$$0 \mapsto 0, \quad 1 \mapsto 1, \quad -1 \mapsto 2, \quad 2 \mapsto 3, \quad -2 \mapsto 4, \quad \ldots$$

More surprisingly, the rationals $\mathbb{Q}$ are countable. Cantor showed this by arranging the positive rationals in a grid (with $p/q$ at position $(p,q)$) and traversing the grid diagonally, skipping fractions already counted. This “zigzag” enumeration establishes $|\mathbb{Q}| = |\mathbb{N}| = \aleph_0$.

Example: Countable Unions of Countable Sets

A countable union of countable sets is countable. If $S_1, S_2, S_3, \ldots$ are each countable sets, then $\bigcup_{n=1}^{\infty} S_n$ is countable. This follows from a two-dimensional enumeration argument similar to the one for $\mathbb{Q}$. This result implies, for example, that the set of all algebraic numbers is countable (since algebraic numbers of degree $n$ form a countable set for each $n$).

The Uncountability of the Reals

Cantor's most famous and consequential result is that the real numbers cannot be enumerated in this way. His original 1874 proof used a topological argument involving nested intervals. In 1891, he published the far more elegant and influential diagonal argument.

Theorem: The Real Numbers are Uncountable (Cantor, 1891)

There is no bijection between $\mathbb{N}$ and $\mathbb{R}$. Equivalently, the set of all infinite binary sequences is uncountable.

Proof (Diagonal Argument):

Suppose for contradiction that the real numbers in the interval $(0, 1)$ can be listed in a sequence $r_1, r_2, r_3, \ldots$. Write each in its decimal expansion:

$$r_1 = 0.d_{11}d_{12}d_{13}d_{14}\ldots$$

$$r_2 = 0.d_{21}d_{22}d_{23}d_{24}\ldots$$

$$r_3 = 0.d_{31}d_{32}d_{33}d_{34}\ldots$$

$$\vdots$$

where $d_{ij}$ is the $j$-th digit of $r_i$. Now construct a new number $x = 0.x_1 x_2 x_3 \ldots$ by setting:

$$x_n = \begin{cases} 3 & \text{if } d_{nn} \neq 3 \\ 7 & \text{if } d_{nn} = 3 \end{cases}$$

We choose digits 3 and 7 (avoiding 0 and 9) to prevent issues with dual decimal representations such as $0.999\ldots = 1.000\ldots$.

Then $x \in (0, 1)$ but $x \neq r_n$ for every $n$, because $x$ differs from $r_n$ in the $n$-th decimal digit:$x_n \neq d_{nn}$ by construction. This contradicts the assumption that $r_1, r_2, r_3, \ldots$ was a complete listing of $(0, 1)$. Therefore $(0, 1)$ — and hence $\mathbb{R}$ — is uncountable. $\blacksquare$

The diagonal argument is one of the most powerful and versatile proof techniques in mathematics. It has been adapted to prove Gödel's incompleteness theorems, Turing's undecidability of the halting problem, and numerous results in descriptive set theory and complexity theory.

Example: Cantor's 1874 Proof (Nested Intervals)

Cantor's original 1874 proof used a different technique. Suppose $r_1, r_2, r_3, \ldots$ is a sequence of real numbers. Construct nested closed intervals $[a_n, b_n]$ as follows:

  • Start with any interval $[a_1, b_1]$
  • Given $[a_n, b_n]$, find the first two elements $r_j, r_k$ from the sequence that lie in the open interval $(a_n, b_n)$, and set $[a_{n+1}, b_{n+1}]$ to be the interval between them (with the smaller as left endpoint)

By the nested interval property (completeness), there exists a point in $\bigcap_{n=1}^{\infty} [a_n, b_n]$ that differs from every $r_n$, since by construction each $r_n$ is eventually excluded from the nested intervals. Thus the original sequence cannot exhaust $\mathbb{R}$.

18.6 — Transfinite Numbers

Having established that infinite sets come in different sizes, Cantor developed a systematic theory of transfinite numbers to measure and compare these infinities. He introduced two parallel hierarchies: the cardinal numbers, which measure the “size” of sets, and the ordinal numbers, which measure the “order type” of well-ordered sets.

Cardinal Numbers

Definition: Aleph Numbers

The aleph numbers form an ascending sequence of infinite cardinal numbers:

$$\aleph_0 < \aleph_1 < \aleph_2 < \cdots < \aleph_\omega < \cdots$$

  • $\aleph_0$ (aleph-null) is the cardinality of $\mathbb{N}$, the smallest infinite cardinal
  • $\aleph_1$ is the next infinite cardinal — the smallest uncountable cardinal
  • The continuum $\mathfrak{c} = |\mathbb{R}| = 2^{\aleph_0}$ is the cardinality of the real numbers

Cardinal Arithmetic

Cantor developed an arithmetic of infinite cardinals that behaves quite differently from finite arithmetic. For infinite cardinals $\kappa$ and $\lambda$:

Rules of Cardinal Arithmetic

$$\aleph_0 + \aleph_0 = \aleph_0$$

$$\aleph_0 \cdot \aleph_0 = \aleph_0$$

$$\mathfrak{c} + \aleph_0 = \mathfrak{c}$$

$$\mathfrak{c} \cdot \mathfrak{c} = \mathfrak{c}$$

More generally, for infinite cardinals: $\kappa + \lambda = \kappa \cdot \lambda = \max(\kappa, \lambda)$.

However, exponentiation produces genuinely larger cardinals: $2^{\aleph_0} > \aleph_0$.

Example: Why the Line and the Plane Have the Same Cardinality

In 1877, Cantor proved the startling result that $|\mathbb{R}| = |\mathbb{R}^2|$. He wrote to Dedekind: “I see it, but I don't believe it!”

The bijection works by interleaving decimals. Given a point $(x, y) \in \mathbb{R}^2$with decimal expansions $x = 0.x_1 x_2 x_3 \ldots$ and $y = 0.y_1 y_2 y_3 \ldots$, map it to:

$$z = 0.x_1 y_1 x_2 y_2 x_3 y_3 \ldots$$

This shows $\mathfrak{c}^2 = \mathfrak{c}$, confirming that cardinal multiplication by $\mathfrak{c}$ does not increase $\mathfrak{c}$. (A careful treatment must handle the non-uniqueness of decimal expansions, but the result holds.)

Ordinal Numbers

While cardinal numbers measure the raw size of a set, ordinal numbers encode the structure of a well-ordering. The finite ordinals are simply$0, 1, 2, 3, \ldots$, and the first infinite ordinal is $\omega$, which represents the order type of $\mathbb{N}$ under its usual ordering.

Definition: Ordinal Numbers

An ordinal number is the order type of a well-ordered set. The first few transfinite ordinals are:

$$\omega, \quad \omega + 1, \quad \omega + 2, \quad \ldots, \quad \omega \cdot 2, \quad \ldots, \quad \omega^2, \quad \ldots, \quad \omega^\omega, \quad \ldots, \quad \varepsilon_0, \quad \ldots$$

Here $\varepsilon_0 = \omega^{\omega^{\omega^{\cdot^{\cdot^{\cdot}}}}}$ is the first ordinal satisfying $\varepsilon_0 = \omega^{\varepsilon_0}$.

Unlike cardinal arithmetic, ordinal arithmetic is not commutative: $1 + \omega = \omega \neq \omega + 1$.

Definition: Well-Ordering

A well-ordering on a set $S$ is a total order $\leq$ such that every non-empty subset of $S$ has a least element. The natural numbers $\mathbb{N}$ are well-ordered, but $\mathbb{Z}$ and $\mathbb{R}$ with their usual orderings are not. The Well-Ordering Theorem (proved by Zermelo in 1904 using the Axiom of Choice) states that every set can be well-ordered.

Example: Non-Commutativity of Ordinal Addition

Consider $1 + \omega$: place a single element before the natural numbers. The resulting well-ordered set $\{*, 0, 1, 2, 3, \ldots\}$ is order-isomorphic to $\{0, 1, 2, 3, \ldots\}$, so $1 + \omega = \omega$.

But $\omega + 1$: place a single element after all the natural numbers. The set $\{0, 1, 2, 3, \ldots, *\}$ has a last element, so it is not isomorphic to $\omega$. Thus $\omega + 1 > \omega$, and $1 + \omega \neq \omega + 1$.

Example: Ordinal Multiplication

Ordinal multiplication is also non-commutative. We have $2 \cdot \omega = \omega$(infinitely many copies of a 2-element set, which looks just like $\omega$), but$\omega \cdot 2 = \omega + \omega$ (two copies of $\omega$ placed end-to-end), which is strictly larger than $\omega$.

18.7 — Cantor's Theorem and the Continuum Hypothesis

Having shown that the naturals and reals have different cardinalities, Cantor asked the most far-reaching question of all: is there an endless hierarchy of infinities? His answer was resoundingly yes.

Cantor's Theorem (1891)

For any set $S$, the power set $\mathcal{P}(S)$ (the set of all subsets of $S$) has strictly greater cardinality than $S$:

$$|S| < |\mathcal{P}(S)| = 2^{|S|}$$

Proof:

Step 1 (Injection exists): The map $f: S \to \mathcal{P}(S)$ defined by $f(s) = \{s\}$ is an injection, so $|S| \leq |\mathcal{P}(S)|$.

Step 2 (No surjection exists): Suppose for contradiction that $g: S \to \mathcal{P}(S)$ is a surjection. Define the “diagonal” set:

$$D = \{ s \in S : s \notin g(s) \}$$

Since $D \subseteq S$, we have $D \in \mathcal{P}(S)$. Since $g$ is surjective, there exists some $d \in S$ with $g(d) = D$. Now we derive a contradiction by asking whether $d \in D$:

  • Case 1: If $d \in D$, then by the definition of $D$, we need $d \notin g(d)$. But $g(d) = D$, so $d \notin D$ — contradicting our assumption.
  • Case 2: If $d \notin D$, then by the definition of $D$, we need $d \in g(d)$. But $g(d) = D$, so $d \in D$ — again a contradiction.

Both cases lead to contradiction, so no surjection $g: S \to \mathcal{P}(S)$ can exist. Combined with the injection from Step 1, we conclude $|S| < |\mathcal{P}(S)|$. $\blacksquare$

Cantor's theorem immediately yields an infinite hierarchy of infinities with no upper bound:

$$|\mathbb{N}| < |\mathcal{P}(\mathbb{N})| < |\mathcal{P}(\mathcal{P}(\mathbb{N}))| < \cdots$$

In terms of alephs and beth numbers: $\aleph_0 < 2^{\aleph_0} < 2^{2^{\aleph_0}} < \cdots$. There is no largest cardinal number. The beth numbers are defined by $\beth_0 = \aleph_0$,$\beth_{n+1} = 2^{\beth_n}$, so $\beth_1 = \mathfrak{c}$.

The Continuum Hypothesis

Since $|\mathbb{R}| = 2^{\aleph_0}$ and $\aleph_1$ is the smallest uncountable cardinal, a natural question arises: is $2^{\aleph_0} = \aleph_1$? In other words, is there an infinite set whose cardinality is strictly between that of the naturals and the reals?

The Continuum Hypothesis (CH)

Continuum Hypothesis: There is no set $S$ with $\aleph_0 < |S| < 2^{\aleph_0}$. Equivalently, $2^{\aleph_0} = \aleph_1$.

Generalized Continuum Hypothesis (GCH): For every infinite cardinal $\aleph_\alpha$, we have $2^{\aleph_\alpha} = \aleph_{\alpha+1}$.

Cantor spent years trying to prove CH, and his inability to do so was a source of great frustration and anguish. The reason became clear only decades later:

Independence of the Continuum Hypothesis

  • Gödel (1940): Proved that CH is consistent with the axioms of ZFC (Zermelo–Fraenkel set theory with the Axiom of Choice). He constructed the “constructible universe” $L$, a model of ZFC in which CH holds. Thus CH cannot be disproved from ZFC.
  • Cohen (1963): Proved that the negation of CH is also consistent with ZFC, using his revolutionary method of forcing. He constructed models of ZFC in which CH fails. Thus CH cannot be proved from ZFC either.

Together, these results show that the Continuum Hypothesis is independent of the standard axioms of set theory — it can be neither proved nor disproved. Hilbert had listed it as the first of his famous 23 problems in 1900; Cohen received the Fields Medal in 1966 for his resolution.

18.8 — Opposition, Struggle, and Vindication

Cantor's work provoked one of the most bitter intellectual controversies in the history of mathematics. The opposition was led by Leopold Kronecker (1823–1891), a powerful Berlin mathematician who held deep convictions about the nature of mathematical existence.

Kronecker's Opposition

Kronecker famously declared: “God made the integers; all else is the work of man.” He rejected the existence of actual infinite sets and regarded Cantor's transfinite numbers as mathematical nonsense. More concretely, Kronecker insisted that legitimate mathematics should deal only with objects that could be constructed in finitely many steps from the natural numbers — a philosophical position that anticipated twentieth-century constructivism and intuitionism.

Kronecker's opposition was not merely philosophical. As an editor of Crelle's Journal and a professor at Berlin, he wielded significant institutional power. He blocked Cantor's attempts to secure a position at Berlin — the most prestigious mathematics department in Germany — and reportedly tried to delay or prevent the publication of Cantor's papers. He called Cantor a “scientific charlatan” and a “corrupter of youth.”

Other mathematicians also expressed reservations. Henri Poincaré called set theory a “disease” from which mathematics would eventually recover. Charles Hermite wrote of his “horror” at the idea of infinite sets being treated as completed totalities. Even Ludwig Wittgenstein, decades later, would describe Cantor's diagonal proof as “hocus pocus.”

Cantor's Mental Health

Beginning in 1884, Cantor suffered recurrent episodes of severe depression that required hospitalization. The exact causes are debated by historians: some attribute them partly to the relentless attacks from Kronecker and the frustration of being unable to prove the Continuum Hypothesis; others note that there may have been a genetic predisposition to mood disorders in his family. What is clear is that the periods of illness became more frequent and severe in his later years.

During his lucid periods, Cantor remained mathematically active and engaged with the broader mathematical community. He helped found the Deutsche Mathematiker-Vereinigung(German Mathematical Society) in 1890 and was instrumental in organizing the first International Congress of Mathematicians in Zürich in 1897. He also developed an interest in theology, finding resonance between his theory of the infinite and theological conceptions of the Absolute.

His final years were marked by poverty, declining health, and institutional confinement. Georg Cantor died on January 6, 1918, in a sanatorium in Halle, not having lived to see the full vindication of his ideas.

Vindication

Even during Cantor's lifetime, the tide had begun to turn. David Hilbert became the most prominent champion of set theory, declaring in 1926:

“No one shall expel us from the paradise that Cantor has created.”

Bertrand Russell, despite discovering the paradox that bears his name (which necessitated a more careful axiomatization of set theory), recognized Cantor's work as epoch-making. Ernst Zermelo provided the first axiomatization of set theory in 1908, and by the 1920s, Fraenkel, Skolem, and von Neumann had refined it into the ZFC axioms that remain the standard foundation for mathematics today.

Timeline: The Battle Over Set Theory

  • 1874 — Cantor publishes the uncountability of the reals; Kronecker expresses disapproval
  • 1884 — Cantor's first major depressive episode
  • 1891 — Diagonal argument published; opposition intensifies
  • 1897 — Burali-Forti paradox discovered (first set-theoretic paradox)
  • 1899 — Cantor himself discovers paradoxes of “the set of all sets”
  • 1901 — Russell's paradox shakes the foundations
  • 1904 — Zermelo proves the Well-Ordering Theorem using the Axiom of Choice
  • 1908 — Zermelo publishes the first axiomatization of set theory
  • 1918 — Cantor dies; his ideas are increasingly accepted
  • 1922–25 — Fraenkel and Skolem refine ZFC
  • 1940 — Gödel proves consistency of CH with ZFC
  • 1963 — Cohen proves independence of CH from ZFC

18.9 — Legacy: Foundations of Modern Mathematics

The combined work of Cantor and Dedekind transformed mathematics in ways that are difficult to overstate. Their legacy permeates virtually every branch of modern mathematics.

Foundations of Analysis

Dedekind's construction of the reals via cuts, together with Cantor's parallel construction using Cauchy sequences, finally placed calculus on a rigorous footing. The completeness of $\mathbb{R}$ — whether expressed as the least upper bound property (Dedekind) or the convergence of Cauchy sequences (Cantor) — became the cornerstone of real analysis. Every theorem in modern analysis that depends on completeness — the Intermediate Value Theorem, the Bolzano–Weierstrass Theorem, the Heine–Borel Theorem — traces its foundations back to their work.

Cantor's Construction via Cauchy Sequences

Independently of Dedekind, Cantor also constructed the reals (1872). His approach defines a real number as an equivalence class of Cauchy sequences of rationals. A sequence $(a_n)$ of rationals is Cauchy if for every rational $\varepsilon > 0$, there exists $N$ such that $|a_m - a_n| < \varepsilon$ for all $m, n > N$.

Two Cauchy sequences $(a_n)$ and $(b_n)$ are declared equivalent if:

$$\lim_{n \to \infty} |a_n - b_n| = 0$$

Arithmetic operations are defined term-by-term: $[(a_n)] + [(b_n)] = [(a_n + b_n)]$. This construction is now more commonly used in textbooks, especially in metric space theory, where it generalizes to the completion of any metric space.

Set Theory as a Foundation

Cantor's set theory, once formalized in ZFC, became the standard foundation for all of mathematics. In principle, every mathematical object — numbers, functions, spaces, algebraic structures — can be constructed within set theory. The language of sets and mappings pervades every mathematical textbook written since the mid-twentieth century. When a modern mathematician says “let X be a topological space” or “consider the group G,” they are working within the framework that Cantor created.

Abstract Algebra

Dedekind's ideal theory was the direct ancestor of modern ring theory and module theory. His conceptual, structure-first approach influenced Emmy Noether, who built on Dedekind's ideas to create the abstract algebra of the twentieth century. The very notion of defining mathematical objects by their structural properties rather than their concrete constructions — the hallmark of modern algebra — owes much to Dedekind. His notion of ideals, quotient structures, and homomorphisms became the basic vocabulary of algebra.

Point-Set Topology and Measure Theory

Cantor's study of exceptional sets for trigonometric series led to foundational concepts in topology: derived sets, closed sets, perfect sets, and the Cantor set. The famous Cantor set, constructed by repeatedly removing middle thirds from $[0,1]$, remains a fundamental example in topology, measure theory, and fractal geometry.

The Cantor Set

Define $C_0 = [0, 1]$. At each step, remove the open middle third of every remaining interval:

$$C_1 = \left[0, \frac{1}{3}\right] \cup \left[\frac{2}{3}, 1\right]$$

$$C_2 = \left[0, \frac{1}{9}\right] \cup \left[\frac{2}{9}, \frac{1}{3}\right] \cup \left[\frac{2}{3}, \frac{7}{9}\right] \cup \left[\frac{8}{9}, 1\right]$$

The Cantor set is $C = \bigcap_{n=0}^{\infty} C_n$. It consists of all numbers in $[0,1]$ whose ternary (base-3) expansion uses only the digits 0 and 2. Remarkably:

  • $C$ is uncountable (cardinality $\mathfrak{c}$)
  • $C$ has Lebesgue measure zero (total length removed = $\sum_{n=0}^{\infty} \frac{2^n}{3^{n+1}} = 1$)
  • $C$ is perfect (closed, and every point is a limit point)
  • $C$ is nowhere dense (its interior is empty)
  • $C$ is self-similar (a fractal with Hausdorff dimension $\log 2 / \log 3 \approx 0.631$)

Logic and Computer Science

The diagonal argument became one of the most productive proof techniques in logic and theoretical computer science. Gödel used a variant to prove his incompleteness theorems (1931), showing that any sufficiently powerful formal system contains true statements that cannot be proved within the system. Turing adapted the technique to prove the undecidability of the halting problem (1936), establishing fundamental limits on what computers can compute. The entire theory of computability and computational complexity rests on techniques that Cantor pioneered.

The Cantor–Dedekind Correspondence

The letters exchanged between Cantor and Dedekind from 1872 to 1899 constitute one of the great scientific correspondences. In these letters, major theorems were first articulated, debated, and refined. Cantor's 1873 letter to Dedekind contains the germ of the first uncountability proof; Dedekind's replies helped sharpen the arguments. The 1877 correspondence includes Cantor's astonished discovery that the line and the plane have the same cardinality. The exchange reveals the warmth and mutual respect between two very different mathematical temperaments: Cantor the passionate visionary and Dedekind the patient architect.

Together, Cantor and Dedekind inaugurated a new era in mathematics — one in which the infinite was not merely a convenient fiction or a source of paradox, but a genuine mathematical object to be studied, classified, and computed with. Their work laid the foundation upon which twentieth-century mathematics — from functional analysis to algebraic geometry to mathematical logic — was built. As Hilbert recognized, the paradise they created is one from which mathematics has never looked back.

←Chapter 17: Riemann & FourierChapter 19: Hilbert & Poincaré→
Rate this chapter: