The Two Rules That Can’t Both Be Right
Quantum mechanics is the most successful theory in the history of physics, and it rests on two rules that contradict each other. That is the uncomfortable position quantum mechanics has occupied for a century. Two mathematical rules underpin every calculation in the field, and they contradict each other. The first rule says a quantum system is described by a wave function — a quantity encoding everything about a particle’s position, momentum and spin — which evolves according to Schrödinger’s equation. That equation is a wave equation, and its solutions obey the superposition principle: if two states are possible, their sum is also possible, and superpositions cannot be destroyed. The second rule, the measurement postulate, says that when you measure a system in a superposition, the wave function collapses and only one outcome appears, selected at random. One rule preserves superpositions forever. The other destroys them on contact with a detector.
The contradiction is not a technicality. Any detector is itself made of atoms, so Schrödinger’s equation should apply to it too. Apply the equation to atom plus detector plus experimenter, and the result is a laboratory in a superposition of having measured two different outcomes — two experimenters in the same room, each an exact copy of the original, differing only in what they now know. No collapse appears anywhere in the mathematics. The measurement postulate has to be imposed by hand, as if a different physics governed the observer and the observed.
In 1957, Hugh Everett III, then a physics Ph.D. student at Princeton University, proposed deleting the measurement postulate entirely. [1] Schrödinger’s equation would be universally valid. The wave function would never collapse. Detectors would be ordinary physical objects obeying the same rules as atoms. The price: the universe becomes a superposition of branches, one for each possible outcome — a multiverse. Everett’s proposal is usually called the Many Worlds interpretation, but it is not an interpretation in the ordinary sense. It is a different theory, one that predicts a multiverse as an inevitable consequence of the superposition principle rather than positing one as a convenience.
A Philosopher Draws the Line That Physics Forgot
A parallel effort has arrived at a sharper version of the same question from a different house. Baptiste Le Bihan, a philosopher at the University of Geneva’s Department of Philosophy, published a framework in September 2025 that asks whether multiverse hypotheses can ever receive empirical support at all — and answers by splitting the category in two. [2] Le Bihan distinguishes fragmented multiverses, in which universes are maximally isolated with no unifying structure whatsoever, from holistic multiverses, in which all universes are embedded in a single structure — a universal wavefunction, an inflating field, cyclic dynamics — that can connect them causally, spatiotemporally, or through non-causal relations.
The distinction carries real weight. Fragmented multiverses are epistemically sealed: nothing inside one universe can carry information about the others, so no observation could ever confirm them. Holistic multiverses satisfy what Le Bihan calls an epistemic transparency thesis — at least some properties of the whole must be accessible from within the parts. Everettian quantum mechanics is his prime example of the holistic kind, because the branching structure of the wave function grounds regularities we actually observe, such as the frequencies predicted by the Born rule. The observed frequencies are the grounded consequence; the multiverse is the grounding structure. What is observed is never the structure itself, only its shadow cast into our branch.
Le Bihan’s typology of empirical signatures separates direct causal traces — a signal or collision from another universe imprinting on our past light cone, as some models of eternal inflation allow — from indirect signatures by grounding, where features of our universe hold because of the larger structure. He borrows a standard from Bas van Fraassen: a quantity is empirically grounded if its value can in principle be determined from measurement results, through theoretically posited connections, consistently and non-trivially — meaning the determination could have come out differently and contradicted the theory. That is a lower bar than direct observation, and it is the bar much of physics already clears. The atom was accepted on exactly this basis before anyone saw one.
The Objections That Don’t Survive the Split
Le
Bihan isolates three complaints that critics run together: that multiverses posit unobservable entities, that they lack current empirical support, and that they could never in principle receive such support. Only the third is serious, and it fails against holistic scenarios. His historical counterexample is the heliocentric model: Copernicus implied stellar parallax as a test in principle in 1543, but the effect stayed undetectable for centuries — indirect support came only in 1725 with Bradley’s discovery of stellar aberration, and the first direct parallax measurement was achieved by Bessel in 1838, nearly 300 years later. Heliocentrism was accepted on explanatory and unifying grounds long before any test was technically possible Inflation followed the same arc, proposed to solve the horizon, flatness and monopole puzzles and only later matched against the cosmic microwave background’s nearly scale-free pattern of temperature fluctuations.
Le Bihan also sets aside the fine-tuning argument that often props up multiverse claims. The parameters cited as fine-tuned appear in effective theories, not fundamental ones, so their values could be artifacts of an incomplete description — patching over ignorance of a deeper dynamics that a confirmed theory of quantum gravity might supply. Fine-tuning is at best motivational, he argues, never decisive on its own. And he names the objection he generalizes as the epistemic isolation objection: if universes cannot reach each other, no evidence can cross. That objection is fatal to fragmented multiverses and irrelevant to holistic ones.
The analogy Le Bihan reaches for is the early-twentieth-century debate over island universes. Until the 1920s it was unclear whether spiral nebulae belonged to the Milky Way or were distant systems then construed as other universes. Henrietta Leavitt’s work on Cepheid variables supplied a new empirical method — standard candles for cosmic distance — and Edwin Hubble applied it in 1923-24, showing many nebulae lie outside our galaxy. What had been called other universes were reclassified as parts of a broader one. Le Bihan concedes the analogy is imperfect, since galaxies are causally connected to us while holistic universes may not be. But the lesson holds: what counts as “our universe” is partly a matter of observational limits and conceptual framing, and a holistic multiverse that gained empirical support might one day be redescribed as a more expansive universe. That possible relabeling does not make the current idea idle — before the reconceptualization, the unobservable regions would still instantiate the features we associate with distinct universes.
What remains is a qualified defense, and Le Bihan states it plainly: fragmented multiverses stay empirically inaccessible, while certain holistic multiverses could in principle be supported by the same epistemic standards used everywhere else in physics. The Many Worlds theory that began as a fringe proposal now sits inside a framework that tells us what kind of evidence would count for it — and why the absence of that evidence today is not a verdict.