The Pit of the Banach–Tarski Paradox: How to Double a Ball

There are problems that feel like they shouldn’t exist. They don’t just resist intuition — they spit in its face. The Banach–Tarski paradox is one of them.

The statement is so wild you’d expect it to be a joke:

You can take a solid ball in 3D space, cut it into finitely many pieces, and reassemble those pieces — using only rotations and translations — to form two balls identical to the original.

Not squashed, not stretched. Just rotated and moved.

One ball becomes two.

When I first heard this, I thought: impossible. Surely someone is cheating. Surely this is wordplay. Surely this is some optical illusion.

Then I stepped into the pit.

Step 1: My First Reactions

“Conservation of volume!” My physics instincts screamed at me. Volume can’t just double.
“Maybe the pieces overlap?” But the theorem insists: no overlap, no stretching.
“Maybe the pieces are infinitely many?” No. Just five pieces.

Everything I tried to hold onto slipped away.

I thought: okay, maybe this is some kind of measure-theory trick. Like the Vitali set.

And yes — but it was worse.

Step 2: Why 3D?

The first surprise: Banach–Tarski only works in dimension 3 or higher.

In 1D, you can’t duplicate an interval.
In 2D, you can’t duplicate a disk. The Lebesgue measure rules too strongly.
But in 3D, rotations and the geometry of the sphere allow something bizarre.

So the paradox lives in the gap between 2D and 3D. Already strange.

Step 3: The Vitali Detour

To understand Banach–Tarski, I tried to understand non-measurable sets.

Take the unit interval $[0,1]$ . Using the axiom of choice, you can pick one representative from each equivalence class of real numbers modulo rationals. This creates a Vitali set $V$ .

$V$ cannot be Lebesgue measurable.
Its “volume” cannot be defined.
If it could, translation invariance would lead to contradictions.

That was already unsettling: sets that exist but have no length.

I thought: maybe Banach–Tarski just uses these weird sets.

Correct — but the way it uses them is terrifying.

Step 4: Free Groups Sneak In

Here’s the key insight: in 3D, rotations are complicated enough to contain a free group on two generators.

That means you can pick two rotations $R$ and $S$ such that any product of them (like $RSR^{-1}S^2$ ) is unique. No cancellations beyond the trivial.

This free group structure allows you to partition the sphere into orbits under the group action.

And then the axiom of choice lets you pick one point from each orbit.

I realized: this isn’t geometry anymore. This is algebra wearing a geometric mask.

Step 5: The Construction (Sketchy and Frightening)

The Banach–Tarski construction roughly goes like this:

Take a ball.
Split off its center. (It can be handled separately.)
On the sphere, use group theory and the axiom of choice to split into paradoxical subsets.
Lift these subsets into the ball (extend radially inward).
Move the pieces around by rotations.
Reassemble into two balls.

Each piece is highly non-measurable — so “volume” never makes sense for them.

That’s how conservation of volume escapes.

The pieces aren’t “real” in any geometric sense — they exist only because of choice.

And yet, the construction is rigorous in ZFC (Zermelo–Fraenkel set theory with choice).

Step 6: Wrong Intuitions I Clung To

“Maybe the pieces are infinite dusts?” Yes, but not in the way I thought. They’re not fractals, not Cantor sets. They’re stranger — they have no measurable structure at all.
“Maybe we could actually build this physically?” No. Physical matter is atomic; Banach–Tarski only works with mathematical continua.
“Maybe this violates physics?” It doesn’t — because physics doesn’t assume the axiom of choice. Banach–Tarski is pure math.

Every intuition I tried was wrong.

Step 7: Why It Doesn’t Break Everything

If Banach–Tarski were literally true in the physical world, you could duplicate gold from a nugget, or bread from a loaf. That’s absurd.

So why doesn’t it break mathematics itself?

Because volume only makes sense for measurable sets. Banach–Tarski’s pieces are not measurable. They slip between the cracks of Lebesgue measure.

Lebesgue measure guarantees translation invariance and countable additivity — but not for non-measurable sets.

That’s the loophole.

Step 8: The Role of the Axiom of Choice

The entire paradox hinges on the axiom of choice.

Without it, you cannot guarantee the existence of these bizarre subsets.

In ZF (without choice), Banach–Tarski fails.
In ZFC (with choice), Banach–Tarski holds.

So the paradox isn’t just about geometry. It’s about the foundations of mathematics: do we accept choice, or not?

I realized: the problem isn’t about cutting balls. It’s about how far we’re willing to stretch existence.

Step 9: Related Monsters

The Banach–Tarski pit connects to other horrors:

Hausdorff paradox: The sphere alone can be partitioned into finitely many pieces, reassembled into two spheres of the same size.
Vitali sets: Non-measurable subsets of the real line.
Paradoxical decompositions: Generalizations to groups beyond rotations.

The theme: whenever a group contains a free subgroup, paradox lurks.

Step 10: Open Questions

Even now, Banach–Tarski raises deep issues:

What if we weaken the axiom of choice? Some models of set theory kill Banach–Tarski.
What about constructive mathematics? There, Banach–Tarski cannot be constructed.
In physics, what happens if spacetime is continuous? Could some “Banach–Tarski-like” effect exist in quantum gravity?

These are not solved. The paradox continues to echo.

Step 11: Climbing Out (Barely)

By the end, here’s what I carried out:

Banach–Tarski is real (in ZFC).
It doesn’t break conservation of volume, because “volume” doesn’t apply to non-measurable sets.
It depends crucially on the axiom of choice.
It reveals that infinity and geometry interact in ways intuition cannot handle.

I went in thinking it was a hoax. I came out realizing it was a theorem, a cornerstone of set-theoretic analysis.

That’s The Problem Pit: a problem that looks like a joke until it takes your foundations away.

Arya’s Algorithms

Search This Blog