What Is the Continuum Hypothesis?

Is there a size of infinity strictly between the number of whole numbers and the number of points on a line? Georg Cantor asked exactly that in 1878. He guessed the answer was no. Nobody has ever proven him right. Nobody has ever proven him wrong either. And here's the part that took almost a century to land: nobody ever will, not because the problem is too hard, but because the question falls completely outside what our usual rules of math can settle.

That guess is called the continuum hypothesis, CH for short. To see why it's such a strange claim, and why it broke mathematics in such an interesting way, we need to build up to it piece by piece.

If you haven't read our piece on the different sizes of infinity, it's worth a quick detour first. We're about to reuse its two main tools: countable infinity and Cantor's proof that the real numbers form a bigger infinity than the counting numbers.

Somewhere Between Counting and the Real Line, Is There a Gap?

We already know two different sizes of infinite set. The natural numbers $1, 2, 3, \ldots$ are the smallest kind of infinite: you can list them one at a time, forever. Mathematicians call that size $\aleph_0$, read "aleph-zero."

The real numbers are bigger. Cantor's diagonal argument shows there's no way to list every real number one at a time, no matter how clever your list is. There are always more reals than your list catches. We write the size of the real numbers as $2^{\aleph_0}$, and it's also called $\mathfrak{c}$, the continuum.

So we have two rungs on a ladder: $\aleph_0$ at the bottom, $2^{\aleph_0}$ higher up. Now here's the natural question. Every infinite set of real numbers we actually know how to build, the rationals, the irrationals, the algebraic numbers, all the ones mathematicians have poked at for centuries, turns out to be exactly one size or the other. Never something in between. Try it yourself below.

Notice that nothing ever lands in the gap. Not the rationals, which feel dense but are secretly countable. Not the irrationals, which feel like almost nothing is missing but turn out to be exactly as big as $\mathbb{R}$ itself. Every candidate you can name snaps to one rung or the other. That's not proof there's nothing in the gap. It's just a hundred and fifty years of circumstantial evidence.

Every Real Number Is Just an Infinite Sequence of Yes-or-No Answers

Before we go further, it helps to actually see why the real numbers have size $2^{\aleph_0}$ and not, say, $3 \times \aleph_0$ or $\aleph_0^2$. The notation $2^{\aleph_0}$ isn't decoration. It's counting something specific: the number of ways to answer "yes" or "no" to infinitely many questions, one question for every natural number.

Here's the connection. Pick any number between 0 and 1. You can find it by repeatedly asking "is it in the left half or the right half of what's left?" First question splits $[0,1]$ into $[0, 0.5]$ and $[0.5, 1]$. Second question splits whichever half you picked into quarters. Keep going forever, and the sequence of left/right answers pins down exactly one real number. Flip that around: every infinite sequence of yes/no answers corresponds to exactly one point on the line.

Try building a number this way yourself.

Sixteen clicks got you down to one part in 65,536. That's already a real interval, not a real number. A full real number needs an infinite sequence of choices, one bit for every natural number, forever. The set of all such infinite sequences is written $2^{\aleph_0}$: two choices, repeated $\aleph_0$ times. That's why the reals have that exact cardinality, and it's why $2^{\aleph_0}$ is the notation mathematicians settled on instead of just calling it $\mathfrak{c}$ and moving on.

Cantor's Guess: There's Nothing In Between

Now we can state the continuum hypothesis properly. It says there is no set $S$ with

In plain words: no infinite set of real numbers is bigger than the naturals and smaller than the reals at the same time. Every infinite subset of $\mathbb{R}$ is either countable (size $\aleph_0$) or already as big as $\mathbb{R}$ itself (size $2^{\aleph_0}$). There's no third option.

There's a second way to say the same thing that will matter in a second. Mathematicians number the infinite cardinal sizes in order: $\aleph_0$, then the very next one up, $\aleph_1$, then $\aleph_2$, and so on, forever. $\aleph_1$ is defined to be the smallest infinite size that's bigger than $\aleph_0$, whatever that size turns out to be. The continuum hypothesis is exactly the claim that

Read that carefully. It's not just saying "nothing sits between $\aleph_0$ and the reals." It's pinning the reals to one very specific rung: the very next one after $\aleph_0$, no skipping.

The Staircase and the Elevator

Here's a way to picture that pin. Think of the aleph numbers as a staircase: $\aleph_0$, $\aleph_1$, $\aleph_2$, $\aleph_3$, climbing one flight at a time, forever. Climbing the staircase is the slow, safe way to build bigger infinities. Every step is well understood and every step lands exactly one flight up.

Taking the power set, going from a set to the set of all its subsets, is a completely different move. It's an elevator, not a staircase. It takes you from $\aleph_0$ up to $2^{\aleph_0}$ in one jump, and it's not obvious in advance which floor that elevator opens on. The continuum hypothesis is the specific claim that the elevator, starting from the ground floor, always opens exactly one flight up: floor $\aleph_1$, never higher.

Play with the staircase and the elevator below.

One thing is actually forbidden here, and it's worth knowing why. A rule called König's theorem guarantees the elevator can never land on a floor like $\aleph_\omega$, the first rung you'd only reach by climbing infinitely many finite flights. Every other rung above $\aleph_0$ is fair game as a resting spot for $2^{\aleph_0}$, at least as far as the axioms are concerned. CH just says it happens to be the very first one.

Here's the Part That Breaks Your Brain

So which floor is it really? Given everything above, you'd expect this to be settled by now. It isn't, and the reason is the genuinely strange part of this whole story.

In 1940, Kurt Gödel built a very specific, minimal universe of sets, called $L$, the constructible universe. Every standard axiom of set theory holds inside $L$, and inside $L$, the continuum hypothesis is true. That proved CH can never be disproven from the standard axioms: if it could, that disproof would also apply inside $L$, where CH holds. Contradiction avoided, CH survives.

Then in 1963, Paul Cohen invented an entirely new technique called forcing, and used it to build a different universe of sets where every standard axiom still holds, but CH is false. That proved CH can never be proven either, for the same reason in reverse. Cohen won a Fields Medal for it, the only one ever awarded for work in mathematical logic.

Put those two results together and you get something unusual. The continuum hypothesis is independent of the standard axioms of mathematics. Both "CH is true" and "CH is false" are perfectly consistent with everything else we normally assume about sets. Neither one breaks anything.

This isn't the same as an unsolved problem like the Riemann hypothesis, where we just haven't found the proof yet. It's closer to what happens with Euclid's parallel postulate: assume it and you get ordinary flat geometry, drop it and you get equally valid curved geometries instead. Neither choice is "more correct." CH works the same way, one level up, in the axioms of set theory itself rather than the axioms of geometry. If you want the sibling result, where a similar kind of built-in blind spot shows up inside arithmetic instead of set theory, that's Gödel's incompleteness theorem.

The Short Version

The continuum hypothesis asks whether any infinite set sits strictly between the size of the natural numbers and the size of the real numbers. Cantor guessed no. That guess is exactly the same as saying $2^{\aleph_0} = \aleph_1$: the reals sit on the very next rung after the naturals, with nothing squeezed in between and no skipped floors either.

Gödel showed you can never disprove that guess from the standard axioms of set theory. Cohen showed you can never prove it either. Both outcomes, CH true and CH false, build into perfectly consistent universes of math. The continuum hypothesis isn't a hard problem waiting for a clever proof. It's a genuine fork in the road that the axioms simply refuse to choose for us.

All visualizations are interactive React components running entirely in your browser, using plain SVG and requestAnimationFrame. No libraries beyond React.