The Strange Shape of Infinity

Cantor and the discovery of different infinities

Before the 1870s, mathematicians generally treated infinity as a single idea, a useful shorthand for something without end rather than an object that could be studied in its own right. Georg Cantor changed that by asking a deceptively simple question: what does it mean for two infinite collections to be the same size? His answer was that two sets have the same size, or cardinality, if their elements can be matched up perfectly, one to one, with nothing left over on either side. This definition works perfectly well for finite sets, matching chairs to people in a room, but Cantor pushed it into territory no one had rigorously explored, applying it to sets that never end.

The first surprise was that this definition does not behave the way intuition expects. The set of even numbers can be matched perfectly with the set of all counting numbers, one to one, since every counting number n pairs with the even number 2n, even though the even numbers look like only half of the whole. Cantor showed that the rational numbers, which include every fraction and seem vastly more numerous than the counting numbers, can also be placed in a one to one correspondence with them, by arranging all possible fractions into an infinite grid and counting along its diagonals. Infinite sets, it turned out, could be exactly the same size as a proper subset of themselves, a property so strange that it is now used as part of the very definition of what it means to be infinite.

This gave rise to the concept of countable infinity, meaning any set that can be matched one to one with the counting numbers 1, 2, 3, and so on. Cantor labeled the size of this infinity aleph null, the first in a hierarchy of infinite numbers he would go on to define. The question that mattered most, though, was whether every infinite set was countable in this sense, or whether some infinities were fundamentally too large to be counted even in principle.

The diagonal argument

In 1891 Cantor published a proof that answered this question decisively, using a technique now called the diagonal argument. Consider the set of all real numbers between zero and one, written as infinite decimal expansions. Suppose, for the sake of argument, that this set could be listed completely, matched one to one with the counting numbers, so that every real number in this range appears somewhere on an infinite list. Cantor then constructed a new number by going down the diagonal of that list, taking the first digit of the first number, the second digit of the second number, and so on, and changing each digit to something different.

The number constructed this way cannot be anywhere on the original list, because it differs from the first listed number in at least its first digit, from the second listed number in at least its second digit, and so on for every entry on the list, no matter how far it extends. But this new number is still a real number between zero and one, which means the original list, however it was constructed, failed to include every real number after all. Since this contradiction arises no matter how the list is built, no such complete listing can exist, and the set of real numbers cannot be matched one to one with the counting numbers.

The conclusion is that the real numbers form an uncountably infinite set, strictly larger than the countably infinite set of counting numbers, even though both are infinite. This was a genuinely new kind of mathematical fact: not just that infinity exists, but that infinities come in different sizes, with some infinite sets provably bigger than others. Cantor went on to show that this hierarchy never stops, that for any infinite size there is always a larger one, obtained by considering the set of all subsets of the original set, a result known as Cantor's theorem.

Hilbert's hotel and the texture of infinite arithmetic

The mathematician David Hilbert later popularized a thought experiment that makes the strangeness of infinite sets vivid without requiring any formal proof. Imagine a hotel with infinitely many rooms, numbered one, two, three, and so on, and every room is occupied. A new guest arrives and asks for a room. In a finite hotel this would be impossible, but in Hilbert's infinite hotel the manager simply asks every current guest to move from room n to room n plus one, freeing up room one for the newcomer, with every existing guest still housed.

The thought experiment gets stranger from there. If infinitely many new guests arrive at once, the manager can still accommodate them by asking the guest in room n to move to room 2n, freeing up every odd numbered room for the new arrivals, infinitely many of them, all at once. Even if infinitely many buses arrive, each carrying infinitely many new guests, there is still a way to rearrange everyone and fit them all in, using the same diagonal style enumeration Cantor used for the rational numbers. What the hotel cannot do, however, is accommodate a single new guest if the guests already occupy a genuinely uncountable set of rooms, the kind indexed by real numbers rather than counting numbers, because no systematic shifting scheme of that kind exists for an uncountable collection.

Hilbert's hotel is not a physical possibility, no actual hotel could have infinitely many rooms, and that is precisely the point. It is a tool for building intuition about how countably infinite sets behave under the kind of operations that would be impossible for any finite collection, while also marking, through its limits, the boundary where countable infinity gives way to something strictly larger.

From mathematics to the physical universe

Cantor's infinities are mathematical objects, defined and proved within formal set theory, and nothing in his work claims to describe physical space. Whether the universe we actually live in is spatially infinite is an entirely separate question, one that belongs to cosmology and depends on measurement rather than proof. General relativity allows for universes of several different overall shapes, and which shape matches our universe depends on its average density and geometry, properties that can in principle be measured through the behavior of light traveling across cosmic distances.

The key observable is the curvature of space on the largest scales. A universe with positive curvature would be finite, curving back on itself the way the surface of a sphere does, so that traveling far enough in any one direction would eventually return you to your starting point, though this would require a scale completely impractical for anyone to traverse. A universe with negative curvature would be open and infinite, saddle shaped in its geometry. A universe with exactly zero curvature would be flat, and a flat universe can still be infinite, extending without edge or boundary in every direction, although a flat universe could also in principle have a finite extent if it has an unusual global topology, a possibility that is harder to test directly.

Cosmologists measure this curvature primarily through the cosmic microwave background, the faint afterglow of the early universe, by analyzing the characteristic sizes of temperature fluctuations imprinted across the sky. The angular size of these fluctuations depends sensitively on the overall geometry light has traveled through on its way to us, which makes the cosmic microwave background one of the most direct available tests of the universe's large scale shape.

What Planck actually measured

The European Space Agency's Planck satellite, which operated from 2009 to 2013, produced the most precise full sky maps of the cosmic microwave background to date, improving substantially on earlier results from the WMAP mission. Analysis of the Planck data found the curvature of the universe to be consistent with exactly flat within the mission's measurement uncertainty, a result reported in Planck's cosmological parameter papers published by the Planck Collaboration. This measurement places the universe's density very close to what cosmologists call the critical density, the value that exactly balances the expansion of space against its own gravitational pull.

A flat measurement of this kind calls for a careful reading. A flat measurement within current precision does not prove the universe is infinite, because a universe that only appears flat within our observable limits could in principle still be curved on a larger scale than we can currently probe, or it could have a finite volume despite being locally flat if its overall topology wraps around in a way we have not detected. What the Planck results do establish is that if the universe is curved, that curvature is extremely slight, smaller than current instruments can detect, and that a genuinely infinite flat universe remains fully consistent with every observation made so far.

This leaves cosmologists in a position common throughout the history of the field: a sharp measurement narrows the range of viable models dramatically without settling the deepest version of the question. The observable universe, the portion from which light has had time to reach us since the Big Bang, is necessarily finite, roughly 93 billion light years across including the effects of cosmic expansion. Whether the universe as a whole continues indefinitely beyond that observable horizon, in a genuinely infinite expanse, is a question current technology cannot directly answer, and may not be answerable even in principle from inside the universe we are trying to measure.

Why the mathematics and the cosmology stay separate

It is tempting to treat Cantor's proof that infinities come in different sizes as somehow bearing on whether physical space is infinite, but the two questions operate in different domains and should not be collapsed into one another. Cantor's result is a theorem, true within the axioms of set theory regardless of anything about the physical world, established by pure logical argument rather than by observation. The cosmological question of spatial infinity is empirical, answerable only by measurement, and subject to revision as instruments improve, in exactly the way mathematical theorems are not.

What the two domains share is only a word, infinite, applied to genuinely different kinds of objects: an abstract set defined by logical rules in one case, and the physical extent of space in the other. This is worth stating plainly because popular accounts of cosmology sometimes borrow the drama of Cantor's results, the idea that some infinities are bigger than others, to suggest something mysterious or paradoxical about an infinite universe, when in fact an infinite universe, if it exists, would simply be a single unbounded three dimensional space, mathematically no stranger than the infinite set of counting numbers, with none of the exotic structure that distinguishes countable from uncountable sets.