Set Theory

Set theory is the study of collections—sets—and the question of what collections exist.

12 frameworks, from 1874 to today · Mathematical Logic · Mathematics

still a working view ended · each leads to its entry below

Overview

Set theory is the study of collections—sets—and the question of what collections exist1. That sounds modest, but it turned out to be the framework in which nearly all of modern mathematics can be written down, and the place where mathematics discovers its own limits1. A set is just a collection of things, considered as one object: the set of prime numbers, the set of points on a circle, the set of all functions from one set to another1. Almost every mathematical object can be coded as a set, so a theory of sets became a candidate foundation for the whole subject1. The history of set theory is therefore two histories braided together: a mathematical history of what can be proved about infinite collections, and a foundational history of how mathematics secured its own consistency—and learned that it could not fully do so.

Cantor's transfinite arithmetic

The subject begins with Georg Cantor's work on trigonometric series in the 1870s. Investigating when a series can fail to converge at infinitely many points, Cantor was led to compare infinite collections by size1. His criterion was the one-to-one correspondence: two sets have the same cardinality (the same number of elements, for infinite sets) if their members can be paired off with none left over12. On this criterion, the even numbers and the natural numbers are the same size, since n pairs with 2n2. But in 1874 Cantor proved that the real numbers cannot be paired with the natural numbers: any list of reals omits some real12. So there are at least two sizes of infinity, and the continuum is strictly larger than the countable12.

Cantor then built an arithmetic of infinite sizes1. He introduced ordinal numbers—the counting numbers extended past the finite, with ω as the first infinite ordinal, followed by ω+1, ω+2, and onward—and showed that the ordinals are themselves well-ordered, so every nonempty collection of them has a least member3. He formulated the continuum hypothesis (CH): that there is no size strictly between the countable and the continuum4. He believed it true but could not prove it4. His attempts to well-order the continuum led him to the well-ordering principle, which he could not establish, and around 1883 he began to suspect that the difficulty lay in an assumption about sets themselves. The work was controversial: Leopold Kronecker, who held that only constructively given mathematical objects exist, opposed it sharply, and Henri Poincaré and others initially resisted the actual infinite1. Cantor's set theory was nonetheless taken up by mathematicians who found in it a language for analysis and topology1.

Paradoxes and the axiomatic response

The crisis came from within. In 1897 Cesare Burali-Forti derived a contradiction from the collection of all ordinals, and in 1901 Bertrand Russell found that the set of all sets that are not members of themselves is a member of itself exactly when it is not. Gottlob Frege's Grundgesetze, which had tried to derive arithmetic from a principle of set formation, was undermined by Russell's paradox just as its second volume appeared. The naive assumption—that any describable collection is a set—had to go15.

Ernst Zermelo's response, published in 1908, was to replace naive comprehension with a list of axioms governing which sets exist: pairing, union, power set (the set of all subsets), infinity, separation (subsets of a given set defined by a property), and choice. Abraham Fraenkel and Thoralf Skolem refined the list in the 1920s, adding replacement (the image of a set under a definable function is a set) and making the language first-order—quantifying only over sets, not over properties5. The result, Zermelo–Fraenkel set theory with Choice, or ZFC, became the standard foundation15. Skolem's formulation also produced Skolem's paradox: if ZFC has a model at all, it has a countable one, yet ZFC proves the existence of uncountable sets. The resolution is that "uncountable" is relative to the model—a lesson in how formal theories underdetermine their subject matter.

A parallel line, von Neumann–Bernays–Gödel set theory (NBG), keeps proper classes—collections too large to be sets, such as the class of all sets—as objects, allowing a finite axiomatization and a smoother treatment of class-sized constructions156. It proves the same theorems about sets as ZFC167.

The constructible universe and the first independence results

David Hilbert's 1900 list of problems included the continuum hypothesis as its first entry4. The first breakthrough came from Kurt Gödel in 193882. Gödel defined the constructible universe L: start with the empty set and iteratively add only those sets definable from what is already there, using formulas of set theory. L is a model of ZFC, and inside it every set is definable in a controlled order, which forces the continuum hypothesis and the axiom of choice to hold18. So if ZFC is consistent, it cannot refute CH18. Gödel's method—building an inner model, a transitive class satisfying the axioms—became a fundamental technique1. It also introduced the axiom of constructibility ($V = L$), the hypothesis that every set is constructible, and with it the study of inner models: subuniverses of the set-theoretic universe that satisfy the axioms and can be analyzed.

The other half of independence came from Paul Cohen in 196392. Cohen invented forcing: begin with a model of set theory and adjoin a new "generic" set, one that meets every condition in a suitable partial order but is not already in the model19. By choosing the partial order, one can make CH fail, or make it hold, or control the cardinal arithmetic in many ways. Forcing showed that CH is independent of ZFC—neither provable nor refutable—and it opened a general method for constructing models with prescribed properties942. Cohen received the Fields Medal in 1966. Forcing and inner models together became the twin engines of independence theory: inner models show what must hold, forcing shows what can fail.

Large cardinals

Meanwhile set theorists asked how far the universe of sets extends. ZFC proves the existence of ever larger infinite cardinals, but some cardinals are so large that their existence cannot be proved in ZFC110. A measurable cardinal, for instance, is one carrying a countably complete nonprincipal ultrafilter—roughly, a consistent way of deciding "large" versus "small" for its subsets. Stanisław Ulam introduced measurable cardinals in 1930; Dana Scott proved in 1961 that their existence implies V ≠ L, so they lie beyond the constructible universe. A hierarchy of large cardinal axioms followed—strong, Woodin, supercompact, huge cardinals, and others—each asserting the existence of a cardinal with strong reflection properties. These axioms are not provable in ZFC, but they form a linearly ordered scale of consistency strength: each implies the consistency of those below it. Large cardinal theory became a central industry, and the question of whether there is a natural stopping point, or a largest sensible large cardinal, remains open.

Descriptive set theory and determinacy

A separate tradition studied the definable subsets of the real line. Émile Borel, René Baire, and Henri Lebesgue around 1900 introduced the Borel sets (built from open sets by countable unions and complements) and the projective sets (obtained from Borel sets by projection and complementation). Nikolai Luzin and his school in Moscow, and later Polish mathematicians including Wacław Sierpiński and Kazimierz Kuratowski, developed the theory. The central questions—are all projective sets Lebesgue measurable? do they have the perfect set property?—turned out to be independent of ZFC.

The key tool was determinacy. A two-player infinite game is determined if one player has a winning strategy111. In 1964 Jan Mycielski and Hugo Steinhaus proposed the axiom of determinacy (AD): every such game on the natural numbers is determined. AD contradicts the axiom of choice, but it implies that all sets of reals are well-behaved—measurable, with the perfect set property, and so on1. Donald Martin proved in 1975 that Borel games are determined in ZFC alone, and later work connected determinacy for projective sets to large cardinals: determinacy hypotheses are equivalent to the existence of large cardinals, giving a striking bridge between two apparently distant parts of the field. The modern subject, descriptive set theory, studies definable sets and their regularity properties, often assuming large cardinals or determinacy axioms to settle questions independent of ZFC112.

Combinatorial set theory and forcing axioms

Combinatorial set theory studies infinitary combinatorics: partition relations, almost disjoint families, trees, and the structure of the continuum. Paul Erdős and Richard Rado developed partition calculus from the 1950s, asking when an infinite set can be partitioned so that no piece has a given property. The continuum itself became an object of combinatorial study: its cardinality, its cofinality, and the many cardinal characteristics (invariants measuring how the continuum can be covered or split) form a rich subject, much of it organized by the Cichoń diagram relating these characteristics.

Forcing axioms generalize the Baire category theorem to the forcing context. Martin's Axiom (MA), introduced by Martin and Robert Solovay in 1970, asserts that a broad class of forcing constructions have generic filters; it is consistent with ZFC and implies many consequences of CH without deciding CH. Stronger axioms—the Proper Forcing Axiom (PFA), and later Martin's Maximum (MM)—assert the same for larger classes of forcings and have strong consequences, including the failure of CH and the regularity of the continuum's cardinal arithmetic. These axioms are the subject of intense current work, and they connect forcing, large cardinals, and descriptive set theory.

The multiverse and the present

The independence results raised a philosophical question: if CH is neither true nor false in ZFC, what is the status of set-theoretic statements? Joel David Hamkins and others have proposed the set-theoretic multiverse: the view that there are many equally legitimate universes of sets, related by forcing and inner-model constructions, and that set-theoretic truth is truth across the multiverse rather than in a single intended universe. This is a live philosophical position, not a settled doctrine, and it coexists with the older view that there is one intended universe and that new axioms—large cardinals, forcing axioms, or principles of definability—should decide the open questions.

Today set theory is a large, active field. Inner model theory seeks to build canonical inner models for large cardinals, extending Gödel's L upward; the ultimate inner model, if it exists, would settle many questions. Forcing remains the main tool for independence, with sophisticated iterations and new forcing axioms. Large cardinal theory continues to map the hierarchy of consistency strength1013. Descriptive set theory and determinacy connect definability to large cardinals1. Combinatorial set theory studies the continuum's structure14. And the foundational questions—which axioms are true, whether the multiverse or the single universe is the right picture, and where the hierarchy of large cardinals ends—remain open, which is why the field that began with Cantor's pairing of infinite collections is still, more than a century later, the place where mathematics examines its own foundations.

References

  1. Set theory (introduction, Applications, Early history, Naive set theory, Controversy, Formalized set theory, Inner model theory, Forcing, Large cardinals, Determinacy, Descriptive set theory), Wikipedia
  2. Cardinal number (introduction, History), Wikipedia
  3. Ordinal number, Wikipedia
  4. Continuum hypothesis (introduction, History), Wikipedia
  5. Zermelo–Fraenkel set theory (introduction, History), Wikipedia
  6. Von Neumann–Bernays–Gödel set theory, Wikipedia
  7. Morse–Kelley set theory, Wikipedia
  8. Constructible universe, Wikipedia
  9. Forcing (mathematics), Wikipedia
  10. Large cardinal, Wikipedia
  11. Determinacy, Wikipedia
  12. Descriptive set theory, Wikipedia
  13. List of large cardinal properties, Wikipedia
  14. Infinitary combinatorics, Wikipedia

The frameworks of Set Theory

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1874-1930
  2. 1900-Present
  3. 1908-Present
  4. 1925-Present
  5. 1930-Present
  6. 1930-Present
  7. 1938-Present
  8. 1950-Present
  9. 1963-Present
  10. 1964-Present
  11. 1970-Present
  12. 2012-Present

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Cantorian transfinite set theoryDescriptive set theoryZermelo–Fraenkel set theoryVon Neumann–Bernays–Gödel set theoryCombinatorial set theoryLarge cardinalInner modelInfinitary combinatoricsForcingDeterminacyForcing axiomsMultiverse

Beyond the field

Where Set Theory's ideas came from in other fields, and what they led to there: the strongest 12 of 54 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Set Theory is a planet in Nousfarer

Its city has a tower for each of these 12 frameworks. Land, walk in, and climb one concept at a time.

Land on Set Theory

Related fields

Fields its ideas are linked to

More in Mathematical Logic

See where Set Theory sits on the Map of Human Knowledge