The Banach–Tarski Paradox: Cutting a Sphere into Two Spheres

In 1924, the Polish mathematicians Stefan Banach and Alfred Tarski proved one of the most counterintuitive theorems in the history of mathematics: a solid ball in three-dimensional space can be cut into a finite number of pieces, the pieces moved around using only rotations and translations, and reassembled into two solid balls, each the same size as the original.

The original. Out of one ball, you get two. Mass is doubled. Volume is doubled. Nothing else changes.

This is the Banach–Tarski paradox. It is not a logical paradox in the strict sense — there’s no contradiction in the mathematics. It is a theorem, with a rigorous proof, that produces a conclusion so violently at odds with physical intuition that the word “paradox” seems unavoidable. The pieces aren’t anything you could cut with a knife. The proof relies on the Axiom of Choice in a way that produces sets so jagged that they don’t even have a well-defined volume.

This article is about what the paradox says, why it works, and what it has taught mathematicians about the limits of intuition.

What the theorem actually says

Take a solid ball $B \subset \mathbb{R}^3$ of radius 1. The Banach–Tarski theorem states that there exists a partition

$B = A_1 \cup A_2 \cup A_3 \cup A_4 \cup A_5$

into five disjoint subsets, plus rigid motions (combinations of rotations and translations) $g_1, g_2, g_3, g_4, g_5$, such that

$g_1(A_1) \cup g_2(A_2) = B' \text{ (one full ball)}$

and

$g_3(A_3) \cup g_4(A_4) \cup g_5(A_5) = B'' \text{ (another full ball)},$

where $B'$ and $B''$ are non-overlapping balls each of radius 1.

The transformations $g_i$ are rigid — they preserve distances. They don’t stretch, compress, or distort the pieces. They only rotate and translate. Yet the result has twice the volume of the original.

The number of pieces is five, but with cleverer constructions, four are enough — Banach–Tarski themselves used a slightly different formulation. With the Axiom of Choice, the duplication can be arranged in many ways. The exact minimum number of pieces is a separate question; the existence of such a decomposition is the heart of the theorem.

Why this is possible

The mathematical machinery behind Banach–Tarski rests on three pillars:

1. The free group $F_2$

Consider words made from two symbols $a$ and $b$ and their inverses $a^{-1}, b^{-1}$. Concatenate words, with the rule that $a$ and $a^{-1}$ cancel adjacent to each other (and similarly for $b$). The set of all such “reduced” words forms a group called the free group on two generators, written $F_2$.

This group has a special property: it can be partitioned into pieces that can be rearranged to give two copies of itself. Specifically, the words starting with $a$, those starting with $a^{-1}$, those starting with $b$, those starting with $b^{-1}$, plus the empty word, partition $F_2$. By multiplying the right pieces by appropriate elements, you can rearrange them to form two complete copies of the group.

This is the algebraic core of the paradox. A free group on two generators is “self-similar” in a way that allows duplication by rearrangement.

2. The Axiom of Choice

The free group $F_2$ embeds nicely into the rotation group $SO(3)$ — the group of rotations of three-dimensional space. There exist two specific rotations $a$ and $b$ that, when combined freely, never produce the identity rotation (other than the trivial way). Effectively, $F_2$ acts on the unit sphere $S^2$ as a free group action — except for some “bad” points where rotations behave badly.

Using the Axiom of Choice, one can pick one representative from each orbit of this action. The resulting set is the source of the paradox. It’s a non-measurable subset of the sphere — meaning it has no well-defined area or volume in the standard Lebesgue sense.

The Axiom of Choice is a separate axiom in set theory, asserting that one can simultaneously pick one element from each member of an arbitrary collection of nonempty sets. It seems innocuous, but it’s known to produce non-constructive proofs: it asserts the existence of objects without giving any way to write them down.

3. The Hausdorff paradox

Felix Hausdorff in 1914 had already shown a weaker version: the unit sphere can be decomposed into pieces that can be rearranged to form two copies of itself, except possibly for a “small” leftover (a countable set). Banach and Tarski extended this to the full ball without exceptions, using the Schröder-Bernstein theorem on equidecomposability.

Combine the three ingredients — free group structure, Axiom of Choice, Hausdorff’s paradox — and you get the Banach–Tarski theorem.

What “non-measurable” means

The reason the paradox doesn’t break physics is that the pieces $A_1, A_2, \ldots, A_5$ are non-measurable: they don’t have any well-defined volume. They aren’t fractals (those have measurable Hausdorff dimensions). They aren’t smooth shapes. They are pathological subsets of space, defined using the Axiom of Choice in a way that makes them essentially incomprehensible — you cannot describe them, draw them, or compute with them.

In standard measure theory, you assign a “size” (length, area, volume) to subsets of Euclidean space. Reasonable subsets — open sets, closed sets, Borel sets, Lebesgue measurable sets — all have well-defined sizes. The pieces in Banach–Tarski are outside this collection. The paradox doesn’t say “two copies of measure $V$ equal one copy of measure $V$”; it says “the original ball, of measure $V$, can be partitioned into pieces, none of which has any defined measure, that rearrange into two balls of measure $V$.”

Volume isn’t being violated. The pieces are too pathological for “volume” to apply.

What it teaches about set theory

Banach–Tarski is one of the canonical reasons why mathematicians take the foundations of set theory seriously. The theorem is a direct consequence of the Axiom of Choice. Without the axiom, the construction fails — you cannot pick the orbit representatives.

This raises a philosophical question: is the Axiom of Choice “true”? The standard answer in modern mathematics is yes — it’s been adopted as part of the standard axiom system (Zermelo-Fraenkel set theory with Choice, ZFC) — but its consequences are sometimes uncomfortable, as Banach–Tarski demonstrates.

There are alternative axiom systems that reject Choice or replace it with a weaker version. The most famous is Solovay’s model (1970), in which all sets of real numbers are Lebesgue measurable. In this model, the Banach–Tarski paradox doesn’t go through — you can’t construct the non-measurable pieces. But Solovay’s model also fails to support some standard results in functional analysis (like the Hahn-Banach theorem) and is not the working framework of most mathematicians.

The pragmatic position is: the Axiom of Choice is too useful to give up. It produces some mind-bending consequences (like Banach–Tarski), but those consequences live in a corner of mathematics that doesn’t intersect with practical applications. Real-world physics, statistics, and engineering all work just fine inside ZFC, paradox or not.

How the paradox is used

Although Banach–Tarski itself doesn’t have direct applications, it has motivated several lines of mathematical research:

Equidecomposability theory: When can two sets be partitioned into pieces that can be rigidly rearranged into each other? Banach–Tarski shows that in 3D this notion is too flexible. In 2D, the Banach–Tarski paradox doesn’t hold (because the rotation group of the plane is amenable — a technical algebraic property), but other equidecomposability questions remain interesting.

Amenable groups: A group is amenable if it admits an invariant probability measure. The fact that $SO(3)$ is non-amenable is what makes Banach–Tarski possible. The rotation group of the plane $SO(2)$ is amenable — that’s why the paradox doesn’t transfer to 2D. Amenability has become a central concept in modern group theory and operator algebras.

Geometric measure theory: Modern measure theory studies subsets of Euclidean space with various notions of dimension and measure (Hausdorff, Minkowski, packing). Banach–Tarski shows the limits of measure-extension procedures and motivates careful analysis of which sets can be assigned reasonable measures.

Higher-dimensional analogues: The paradox holds in any Euclidean space of dimension 3 or higher. In dimension 2 and below, it fails. The reason — amenability of low-dimensional rotation groups — gives a precise geometric explanation.

Why this matters philosophically

The deepest reason Banach–Tarski has captured mathematical and philosophical imagination is that it reveals a clean tension between geometric intuition and set-theoretic logic.

Geometric intuition says: rearranging pieces of a ball can never increase volume. This is conservation of matter, conservation of measure, common sense.

Set-theoretic logic says: if you allow the Axiom of Choice and accept the existence of non-measurable sets, then geometric intuition fails. Volume doesn’t conserve under arbitrary subsets.

The resolution is to recognize that “intuitive geometric volume” only applies to “intuitive geometric subsets” — that is, measurable ones. Once you allow yourself to construct non-measurable sets via abstract logical principles, the rules change.

This is one of the recurring themes in foundational mathematics: when mathematics generalizes beyond what physical intuition was built on, intuition becomes unreliable. Cantor’s infinities had taught the same lesson in a different domain. Gödel’s incompleteness taught it in logic. Banach–Tarski teaches it in geometry.

The right response, mathematicians have largely concluded, is not to reject the strange consequences but to understand them — to learn what their existence tells us about which intuitions can be trusted in which contexts. Banach–Tarski doesn’t say that a ball can really be doubled. It says that, in the strange world of arbitrary set-theoretic constructions, geometric measure can fail to behave the way physical volume behaves.

That distinction is, in many ways, the central lesson of twentieth-century foundations of mathematics: rigorous mathematics goes places where intuition cannot follow, and the mathematics is responsible for telling you when intuition has stopped applying. Banach–Tarski is the cleanest possible example of intuition stopping. The pieces of the ball cannot be drawn, cannot be measured, cannot be physically constructed. But mathematically — within the framework of ZFC — they exist.

Strange as that sounds, it’s the truth that mathematics has been telling us, in increasingly sharp ways, since the late nineteenth century. The map is not the territory. The math is not the world. Both are sometimes deeply, productively, irretrievably weird.

Frequently asked

Is Banach–Tarski actually a paradox?

Not in the logical sense — it's a genuine theorem with a rigorous proof, no contradictions. It's called a paradox because the conclusion violates physical and intuitive expectations so dramatically. The mathematics doesn't allow you to actually duplicate a real-world ball. The 'pieces' are mathematically defined but physically impossible — they can't be cut with a knife or constructed in any physical sense.

Does this break physics?

No. The pieces required by Banach–Tarski are not measurable in the Lebesgue sense — they don't have a well-defined volume. Real physical objects always have measurable volume. Mathematical paradoxes about non-measurable sets don't transfer to physical reality, where conservation laws and the structure of matter prevent such constructions.

Can you avoid the paradox?

Yes — by rejecting the Axiom of Choice or by restricting attention to measurable sets. Without the Axiom of Choice, you cannot construct the non-measurable pieces that Banach–Tarski requires, and the paradox doesn't go through. Most mathematicians accept the Axiom of Choice anyway, because rejecting it costs more than the paradox is worth: you'd lose much of modern functional analysis and topology.

Continue reading

Blog

Cantor's Infinities: Some Are Larger Than Others

At the end of the nineteenth century, Georg Cantor proved that infinity comes in different sizes. The result outraged his contemporaries and became a foundation of modern mathematics.

Read →Blog

The Monty Hall Problem and Why Probability Tricks Your Brain

A game-show puzzle from the 1990s that fooled thousands of mathematicians — and what it reveals about how badly our intuition handles conditional probability.

Read →Blog

Gödel's Incompleteness: Mathematics That Can't Prove Itself

Kurt Gödel's 1931 theorems proved that any mathematical system powerful enough to describe arithmetic contains true statements it can never prove — and can never certify its own consistency.

Read →