Infinity, Mathematical Truth, and the Set-Theoretic Multiverse

Hamkins reviews Aristotle’s potential infinity, Archimedes’ method of exhaustion, and Galileo’s examples of one-to-one correspondences between infinite collections. He explains how Cantor clarified that some infinite sets are larger than others.

The hotel’s room-shifting examples show how a countably infinite set can make room for additional guests, an infinite bus, or passengers on an infinite train. Hamkins connects the examples to countability and the union of countably many countable sets.

Hamkins uses a zigzag path through a grid to enumerate pairs of natural numbers, then applies similar reasoning to fractions. Although the rationals are densely ordered, they are still countable.

The discussion distinguishes algebraic from transcendental real numbers and mentions examples such as pi and e. Hamkins also gives a proof that every natural number is interesting, based on the supposed smallest uninteresting number.

Hamkins constructs a real number whose nth decimal digit differs from that of the nth number on a proposed list. This shows the reals cannot be enumerated by the natural numbers, and he discusses the argument’s decimal-representation detail and influence on later diagonal proofs.

Hamkins explains set theory’s dual role as a subject and a foundation for mathematics. He describes the origin of Zermelo’s axioms in debate over the well-ordering principle and introduces the axiom of choice, including the distinction between explicit and nonconstructive choice.

Russell’s examples of choosing one shoe or sock from each pair illustrate why an explicit rule works for shoes but may not be available for indistinguishable socks. Hamkins connects the axiom of choice to different views about mathematical objects and constructive proof.

The speakers review the main ZFC axioms and Hamkins explains examples including extensionality, the empty set, pairing, union, power set, and infinity. He discusses the historical controversy over choice and well-ordering, consistency, and why constructivists may still care whether choice is used.

Hamkins proves that a set’s power set is larger than the set, using a diagonal subset. He recasts the argument through committees and fruit salads, then relates the same structure to Russell’s paradox and the impossibility of a set of all sets.

Hamkins recounts how Russell’s paradox challenged Frege’s project of reducing mathematics to logic. He discusses Frege’s response and argues that set-theoretic foundations can be viewed as fulfilling logicism’s aims, while noting that this interpretation is disputed.

To introduce Gödel’s theorems, Hamkins describes Hilbert’s effort to build a strong mathematical theory and prove its safety using finitary reasoning. He explains the formalist view of proofs as finite sequences of symbols governed by rules.

Hamkins discusses what “finitary” can mean and gives arithmetic theories as examples. He describes the reach of Peano arithmetic for classical number theory while noting that claims about its scope need to be understood in light of incompleteness.

Hamkins explains that sufficiently strong computably axiomatized theories are incomplete and cannot prove their own consistency. The conversation distinguishes computable enumeration from decision and connects undecidable provability questions with the halting problem.

Hamkins distinguishes truth in a mathematical structure from proof in a formal system. He explains Tarski’s disquotation approach, gives modus ponens as a proof rule, and describes soundness, completeness, and the ability to check whether something is a proof.

The discussion returns to incompleteness and the limits on a complete theory of arithmetic. Hamkins defines the halting problem and explains why observing that a program has not halted does not establish that it never will.

Hamkins constructs a program that behaves opposite to a proposed halting decider when run on itself, producing a contradiction. He then uses the halting problem to show that a computable, complete theory of elementary mathematics cannot exist.

Hamkins describes his approach to teaching proofs through interesting statements and elementary arguments. A problem about everyone being pointed at more than they point is solved by tracking money; he uses it to illustrate anthropomorphizing mathematical ideas and notes how the result changes for infinitely many people.

Hamkins considers whether infinity and numbers exist, contrasting abstract objects with physical objects. He says physical existence remains mysterious while mathematical existence can be described through logical properties, and identifies himself with mathematical realism.

Hamkins explains structuralism as treating mathematical objects by their roles in structures, up to isomorphism, rather than by their substance. The discussion uses the number four and Frege’s Julius Caesar problem to illustrate the view.

Hamkins says mathematical reality is more intelligible to him than physical existence, while stopping short of claiming it is more real. The speakers consider how future advances in physics might change that discussion.

Hamkins contrasts continuing growth in mathematical understanding with philosophy’s enduring questions. He expects mathematics far in the future to be difficult for him to recognize, while suggesting that some ideas may still be explained across historical periods.

Hamkins reflects on his long involvement with Math Overflow and how answering questions in logic and adjacent subjects has helped him learn and contribute. The conversation turns to the continuum hypothesis and Cantor’s interest in it.

Hamkins defines the continuum hypothesis as the claim that no infinity lies between the natural numbers and the reals. He reviews results for open, closed, and Borel sets, and describes extensions of this line of work to more complex definable sets.

The conversation clarifies the development of ZFC and discusses why set theory offered a common foundation across mathematical subjects. Hamkins also explains Hilbert’s tenth problem and its negative solution: there is no algorithm deciding whether every such polynomial equation has an integer solution.

Hamkins recounts Gödel’s construction showing that the continuum hypothesis and axiom of choice can hold in a suitable set-theoretic world, and Cohen’s forcing result showing that the continuum hypothesis can also fail. He introduces forcing as a way to construct alternative set-theoretic universes.

Hamkins contrasts the view of one true set-theoretic universe with his pluralist multiverse view, where statements can vary between universes. He discusses independence, large-cardinal axioms, and why those axioms do not settle the continuum hypothesis.

Hamkins describes how the multiverse view treats the continuum hypothesis as having different answers in related universes. He says philosophical positions need not change the shared mathematics, but can shape research questions, contrasting the pursuit of one universe with study of relations among universes.

Hamkins discusses set-theoretic potentialism and research on forcing extensions and their grounds. He describes set-theoretic geology as investigating how a universe might have arisen through forcing, and compares reasoning through an extension and back to earlier work with complex numbers.

Hamkins describes surreal numbers, introduced by John Conway, as a system containing the natural numbers, integers, rationals, reals, ordinals, and infinitesimals. He explains the staged rule that creates a number in each gap and outlines how the system supports arithmetic.

Hamkins explains why surreal numbers are discontinuous in the sense that they lack the least-upper-bound property and convergent sequences, while still allowing infinitesimal-based calculus. He discusses Conway’s hopes for the system and explains that predicting whether a cell in the Game of Life will ever become alive is equivalent to the halting problem.

Hamkins introduces a result about a “black hole” for the halting problem: a procedure decides almost all instances for a specified family of Turing machines, though the general problem remains undecidable. He connects the idea to approximation algorithms for NP-complete problems.

Hamkins stresses that P versus NP is an asymptotic question and discusses existing algorithms and approximations. The conversation considers mathematicians and simultaneous discoveries before Hamkins describes his preference for clear, simple arguments and playful exploration.

Hamkins describes using playful experimentation and metaphor to explore problems, and contrasts his collaborative practice with solitary work on difficult problems. The speakers discuss Grigori Perelman’s decision to decline prizes and Math Overflow as a source of mathematical collaboration.

Hamkins introduces chess on an unbounded board and explains positions where a player can force a win but cannot guarantee a finite upper bound on the number of moves. He discusses ordinal game values and how collaboration helped construct positions with increasingly large values.

Hamkins says current AI systems have not helped him with mathematical arguments and worries that fluent responses can appear valid without being proofs. The host describes using AI for inspiration and connections, while Hamkins distinguishes chat-based systems from formal proof verification.

Hamkins names transfinite ordinals as his most beautiful mathematical idea, explaining how counting continues beyond omega through successor and limit ordinals. He connects transfinite recursion to constructions including the constructible universe.

Hamkins discusses his work across mathematics and philosophy and identifies the distinction between truth and proof as a central philosophical idea. Truth concerns what is the case; proof concerns how people come to know mathematical facts. The conversation then closes.