How many multiverses?

People reach for the multiverse to explain fine-tuning: with enough universes trying different constants, some are bound to be habitable, and of course we live in one of those. But the word covers at least four very different ideas, which the physicist Max Tegmark sorted into levels.

Why reach for other universes at all?

Two engines drive the idea. The first is fine-tuning: if the constants look chosen, an ensemble unmakes the choosing. The second is physics itself — inflation and string theory both predict variety without being asked to. Behind both sits the old Copernican drift: every previous demotion of Earth's specialness enlarged the cosmos (not the center, not the only sun, not the only galaxy), and the multiverse is simply the last demotion available — not the only universe either.

Level I — beyond the horizon

Space may simply go on far past what we can see. Given enough of it, distant regions repeat every arrangement of matter — including, somewhere, another you. Same physics, different scenery.

Tegmark once priced the repetition: in an infinite space, an identical copy of our entire observable region is expected within roughly 10^(10^115) meters — a number so large that writing it out defeats the observable universe itself. That sounds like fantasy, yet Level I is the mildest claim on this page: it asks only that space continues beyond our horizon and that matter shuffles, which cosmology already suspects for independent reasons.

A caution even here: because the expansion accelerates, those far regions are causally disconnected from us forever — no signal, no visit, no check. Level I's defense is parsimony (it needs no new physics, just more of the same space), not observation. That trade — believable mechanism, untestable extent — is the signature of every level below.

Level II — other bubbles

In eternal inflation, space keeps sprouting new bubble universes, and each can freeze in with different constants. This is the level that actually speaks to fine-tuning: a vast lottery of physics, and we occupy a winning ticket.

The mechanism came from elsewhere and that's its strength. Alan Guth invented inflation to fix problems with the Big Bang's initial conditions; Andrei Linde noticed it never has to stop everywhere at once. Each ended patch becomes a bubble with its own vacuum energy, particle masses, even dimensionality. Weinberg's anthropic bound on Λ — observers only where Λ stays small — is exactly the kind of reasoning Level II licenses: not proof, but a map of what to expect.

The evidence status is honest and thin. Inflation itself has indirect support and unresolved rivals; bubbles beyond our horizon are in principle unobservable; and every prediction runs through the measure problem — how to count observers across infinitely many bubbles. Level II is the scientifically serious multiverse and still, today, a bet.

Levels III and IV

Level III is the many-worlds reading of quantum mechanics — every possibility realized in a branching wavefunction. Level IV is the boldest: every mathematically consistent structure exists as its own universe. The further you go, the harder they are to test — which is the standing objection to all of them.

One distinction matters for this site: Level III multiplies outcomes, not laws. Every quantum branch of our universe shares our constants, so many-worlds says nothing about why the constants permit life. Only Levels II and IV vary the physics itself — which is why fine-tuning arguments live there and not in the quantum branches.

Level IV, Tegmark's Mathematical Universe Hypothesis, dissolves the tuning question in the most radical way: if every consistent mathematics exists, a universe like ours was guaranteed to be among them. Its critics reply that it explains everything and therefore nothing — unless someone finds a measure that explains why simple, orderly physics like ours is what a typical observer should expect.

Which levels are science?

Roughly: Level I is conservative extrapolation, Level II is speculative physics, Level III is interpretation of known equations, Level IV is philosophy with equations attached. None has been observed, and none is forbidden. What keeps the whole tower honest is that each level makes different mistakes — so the arguments are separable, and you can buy one floor without the rest.

The case against

George Ellis has argued the sharpest version: once a theory's claims sit forever beyond observation, it has left science for metaphysics — however elegant the mathematics. The multiverse, in this view, explains the unexplained with the unobservable. Defenders reply that many accepted entities (quarks, the interiors of black holes) are observed only through their effects, and that the boundary of the testable has moved before.

Lee Smolin offers the rival that plays by stricter rules: cosmological natural selection, in which universes reproduce through black holes and constants drift toward fecundity — a theory that, unusually for this debate, makes risky claims about our own universe's parameters. It may be wrong, but it is wrong in the way scientists prefer: by being checkable.

And the deep point, often missed: an ensemble does not dissolve fine-tuning, it relocates it. The question 'why these constants?' becomes 'why this ensemble, with this distribution, counted this way?' — three unknowns stacked where one stood. The multiverse remains the most promising answer we have precisely because it is the only one that could, in principle, turn a mystery into a prediction. Whether it ever will is the live argument.

What would convince you?

A fair way to hold each level is to name, in advance, what evidence would move you. Level II has a candidate: a collision with a neighboring bubble could leave a circular bruise in the microwave sky — searched for, not found so far. Level IV offers nothing of the kind, and that is itself information: it belongs to a different folder than physics. If no imaginable observation could change your mind about a multiverse, your belief is not about universes — it's about what counts as an explanation.

One last calibration, because the word poisons debates so easily: believing in some multiverse does not oblige you to believe in every multiverse. You can hold that space continues beyond our horizon and find quantum branching extravagant; you can accept eternal inflation's bubbles and call Level IV poetry. The levels are a menu, not a package deal — the honest reader picks line by line, and says so.

The evidence also weakens as the levels rise. A larger region beyond our horizon extends a geometry already observed nearby. Inflationary bubbles extrapolate a mechanism with some empirical successes but an uncertain eternal regime. Quantum branches inherit tested quantum dynamics while adding an interpretive claim. The mathematical multiverse begins from a philosophical identification. Sharing one noun should not grant all four the same evidential credit.

Where the toy simplifies

The Tuner shows one universe at a time; every share link mints another, and the graveyard of broken ones is this site's own pocket Level II — a lottery of settings with the winners visible. Real bubbles can't be visited; ours can. The physics of why only some settings produce observers is the honest part.

A multiverse can explain why our universe is habitable. Whether that counts as an explanation or a dodge is still fiercely argued.