Which Interpretation of Quantum Mechanics Is Correct?
Quantum mechanics predicts experimental outcomes with flawless accuracy, but physicists still cannot agree on what the mathematics says exists, and no experiment has yet distinguished between the leading pictures of reality that all reproduce the same predictions.
The question — what exactly is unknown, and why it resists an answer
Strip away the philosophy for a moment and state the problem in its bluntest form. Quantum theory gives you a wavefunction, evolving smoothly and deterministically under the Schrödinger equation, spreading into superpositions of possibilities. Then, whenever you measure the system, the theory tells you — via the measurement postulate — to throw away that smooth evolution and instead draw a single outcome at random, with a probability given by the squared amplitude of the wavefunction, the Born rule. Two irreconcilable-looking pieces of dynamics, stitched together by fiat, with no consensus on where the stitching happens, why it happens, or whether it happens at all.
An "interpretation" of quantum mechanics is an attempt to tell a coherent story about reality that reproduces this two-part recipe as an approximation, an illusion, or a genuine feature of nature — without simply asserting it as a brute postulate. The three major families of answer — the Copenhagen tradition, Everettian many-worlds, and de Broglie–Bohm pilot-wave theory — each do this in radically different ways. And here is the vertiginous part: all three, when worked out carefully, make exactly the same predictions for every experiment anyone has so far performed or seriously proposed. This is not a minor technical coincidence. It appears to be a structural feature of the theories, at least for the versions that have survived scrutiny.
So the open question is really two nested questions. First: is there any possible experiment — not merely a thought experiment — that could distinguish between these pictures, given how carefully each has been constructed to reproduce ordinary quantum statistics? Second, and more mathematically pointed: can the Born rule, the probabilistic heart of quantum mechanics, be derived from more basic principles within each interpretation, or must it simply be assumed? Everettians in particular have spent decades trying to derive probability from a theory that, on its face, has no randomness in it at all — only branching certainty. Whether that derivation succeeds is itself contested.
What would an answer even look like? It would have to be one of a small number of things: a genuine empirical test, sensitive to some difference in dynamics at a regime we haven't yet probed (extreme macroscopicity, extreme gravity, or something else entirely); a mathematical proof that one interpretation is logically forced by facts we already accept, foreclosing the others; or — the deflationary possibility that many physicists quietly hold — a permanent verdict that the question is metaphysical rather than scientific, and that physics has reached a place where its own formalism underdetermines the reality beneath it. That last possibility is uncomfortable to sit with, but it must be taken seriously as a live outcome, not dismissed as failure.
What we know — the settled machinery, and the seam where it strains
To see why the interpretations diverge exactly where they do, it helps to walk through the standard story as it's actually taught, and notice precisely which move is the contested one.
A quantum system is described by a wavefunction, a mathematical object that assigns a complex amplitude to every possible configuration. Left alone, this wavefunction evolves under the Schrödinger equation: linear, deterministic, reversible, no randomness anywhere in it. If an electron can be "here" or "there," the Schrödinger equation is entirely happy to evolve it into a superposition of here-and-there, with no preference, no collapse, no choosing.
The trouble starts the moment you couple that electron to something that records its position — a Geiger counter, a photographic plate, a physicist's retina. Because the equation is linear, the superposition doesn't resolve; it spreads. The detector, too, ends up in a superposition of having-detected-here and having-detected-there. Carry the logic forward with genuine rigor, as Erwin Schrödinger did in his famous 1935 thought experiment, and the physicist herself, and the cat in the box, end up correlated with the electron in a superposition of two macroscopically distinct worlds. Nothing in the bare Schrödinger equation stops this cascade. Yet nobody has ever observed being in a superposition of having-seen-two-different-things. Every measurement, without exception, yields one definite outcome.
The measurement postulate is the textbook's patch for this. It says: when you measure an observable, the wavefunction "collapses" onto one of the possible outcomes, with probability given by the Born rule, . This is not derived from the Schrödinger equation. It is bolted on, an extra rule that only applies — mysteriously — during "measurement," a word the postulate itself never precisely defines. This is the seam. Every interpretation of quantum mechanics is, at bottom, a proposal for what is really happening at that seam.
The Copenhagen framework, associated loosely with Niels Bohr and Werner Heisenberg though never codified into one canonical document, essentially declines to answer the question of what's "really" happening. It treats the wavefunction as a bookkeeping device for predicting the statistics of measurement outcomes, not as a description of an objective physical reality that exists independent of observation. The classical world of the laboratory — pointers, screens, human eyes — is taken as an unanalyzed given, a stage on which quantum probabilities play out, and the collapse is simply the update of our knowledge, or of the formalism, when a definite outcome is registered on that classical stage. Bohr resisted the idea that there was a deeper story to tell; asking what the electron was doing "in itself" between preparation and measurement was, to him, a question without physical meaning, since only the results of interactions with classical apparatus are physically accessible at all.
This was long the working orthodoxy, useful precisely because it lets you stop asking uncomfortable questions and get on with calculating. But it leaves an obvious sore spot: what counts as "classical apparatus," and why should the universe care where you draw that line? This is the notorious measurement problem, and it connects directly to the deeper puzzle of the quantum measurement problem as a frontier in its own right.
Here is where one of the genuinely settled pieces of modern physics enters, and it is worth being precise about what it does and does not accomplish: decoherence. In the 1970s and 1980s, physicists including H. Dieter Zeh and later Wojciech Zurek showed, using ordinary quantum mechanics with no extra postulates, that a quantum system almost never stays isolated. It interacts with its environment — stray photons, air molecules, the vibrations of the apparatus itself — and this entanglement spreads the coherent superposition out into correlations with an enormous number of environmental degrees of freedom, so fast and so thoroughly that the interference between the "here" and "there" branches becomes, for all practical purposes, permanently unrecoverable. The reduced state of the system, once you average over the environment you can't track, looks for all the world like a classical mixture of definite outcomes with definite probabilities.
Decoherence is real, calculable, and experimentally verified with exquisite precision — Serge Haroche's group watched it happen in real time, photon by photon, inside a microwave cavity, earning a share of the 2012 Nobel Prize. It explains why we never see macroscopic superpositions in practice, why the world looks classical, and why interference terms vanish for all systems we can actually track. What it does not do, and this is the crucial, frequently blurred point, is explain why any single outcome occurs at all. Decoherence turns a superposition into an improper mixture — a sum over branches that remains, mathematically, a single enormous entangled state of system-plus-environment-plus-everything-it's-touched. It does not select one branch as "the one that happened." It dissolves interference; it does not collapse the wavefunction. Whether something else does select an outcome, or whether all the branches persist and merely become unable to interfere with each other, is exactly the question decoherence leaves open — and it's exactly here that the great interpretations part ways.
A physicist can calculate, to twelve decimal places, the magnetic moment of the electron, and the universe agrees with her every time. She can predict the interference pattern of a single molecule fired one at a time through a double slit, the decay half-life of a muon, the spectrum of light from a hydrogen atom born in the first minutes after the Big Bang. Quantum mechanics is the most successful predictive framework humans have ever built. And yet, if you ask that same physicist what actually happens when a measurement occurs — what the wavefunction is, whether the electron had a definite position before she looked, whether somewhere an equally real copy of her just measured the opposite result — she will very likely shrug, or laugh, or tell you that the question is not one physics is equipped to answer. Some will tell you it isn't even a real question.
This is strange. In no other mature science does the working machinery run so far ahead of any agreed story about what the machinery describes. We do not argue about whether rivers are made of water molecules. We argue, sometimes bitterly, about whether the equations of quantum theory describe an objective world of branching universes, a world guided by a hidden wave, or no world at all beneath the level of what we can observe. Three quarters of a century of extraordinary experimental ingenuity — closing loophole after loophole, testing Bell inequalities across ever greater distances, watching decoherence unfold in real time — has narrowed the field of viable stories without eliminating any of the major contenders. That fact alone is the puzzle.
svg[The seam in the standard recipe: linear evolution up to measurement, then an unexplained jump]
Where it breaks down — the same data, three incompatible worlds
Here is the uncomfortable fact at the center of this frontier: the breakdown isn't in the data. The data behave beautifully. The breakdown is that the data radically underdetermine the story about what's producing them.
Consider many-worlds branching. Hugh Everett III proposed in 1957 that the Schrödinger equation is simply always true — there is no collapse, ever, anywhere, for any system, including the entire universe. When the electron interacts with the detector, and the detector with the physicist, and the physicist with the air in the room, the universal wavefunction does not choose a branch. It simply becomes a superposition of a physicist-who-saw-here and a physicist-who-saw-there, each version of her entirely unaware of the other, each experiencing what feels, from the inside, like a single definite outcome. Decoherence explains why these branches, once split, can never practically interfere with each other again: they have become causally sealed off, entangled with vast, divergent swaths of environment. What decoherence supplies for Everett is the mechanism by which the "worlds" become distinct and stable. What it does not supply — and this is the celebrated difficulty of the approach — is any explanation of why a physicist inside one branch should assign a probability to being in that branch at all, since on this view every branch, without exception, definitely occurs. Probability seems to require ignorance about which outcome will happen; Everett's picture, taken literally, has no such ignorance. Everyone finds out everything.
The attempts to solve this — from Everett's own frequency-based argument, to Bryce DeWitt's early proposals, to David Deutsch's 1999 decision-theoretic derivation, refined substantially by David Wallace, arguing that a rational agent who knows she is about to branch should, by the very structure of the decision problem facing her, weight her preferences exactly as the Born rule prescribes — are mathematically serious and have occupied some of the best philosophers of physics for a generation. Adrian Kent, Simon Saunders, and others have pushed back hard, arguing that the decision-theoretic argument smuggles in exactly the probabilistic assumptions it's meant to derive, or that the very notion of "you, the observer, care about your future branches" requires an unjustified identity criterion across a splitting multiverse that has no single fact of the matter about which future self is "you." This is not a settled debate. It is one of the liveliest disputes in the foundations of physics, and reasonable, technically expert people land on opposite sides of it.
Now turn to the de Broglie–Bohm pilot wave picture, developed by Louis de Broglie in 1927 and revived and completed by David Bohm in 1952. Here the story is almost defiantly old-fashioned: particles always have definite positions, at every instant, whether or not anyone is looking. There is no collapse and no branching. Instead, there are two mathematical objects — the particle's actual position, and a wavefunction that never collapses and evolves exactly as ordinary quantum mechanics says, but which acts as a genuine physical field guiding the particle's motion, via a "guidance equation" that determines velocity from the wavefunction's phase. The double-slit interference pattern arises because the wave passes through both slits and shapes the probability of where particles land, even though each individual particle goes through only one slit, following a perfectly definite, if wildly complicated and highly non-classical, trajectory.
Pilot-wave theory is fully deterministic — all the apparent randomness of quantum mechanics comes not from anything genuinely stochastic but from our ignorance of the particle's exact initial position, an ignorance quantified by exactly the Born-rule distribution, a special "quantum equilibrium" that the theory can argue (though not, by most accounts, strictly prove without additional assumptions) is a dynamical attractor that any reasonable initial ensemble will relax toward, much as classical statistical mechanics relaxes toward the Maxwell-Boltzmann distribution. This is an elegant and much-underrated argument, developed particularly by Antony Valentini, who has also explored the more radical possibility that the early universe might not have fully relaxed to quantum equilibrium — which would make pilot-wave theory experimentally distinguishable from ordinary quantum mechanics in principle, though no such deviation has ever been detected, and most cosmological probes so far strongly favor equilibrium having long since been reached, if it wasn't there from the start.
The price pilot-wave theory pays is famous and severe: to reproduce quantum correlations for entangled particles, the guidance equation is explicitly nonlocal. Moving one particle instantaneously affects the trajectory-determining wave everywhere else it has support, in flagrant tension with the intuitive spirit, if not the letter, of relativity. Bohmians reply, correctly, that this nonlocality is not a defect unique to their theory — it's required by any theory, interpretation-independent, that reproduces the violations of Bell inequalities that nature demonstrably exhibits. Which brings us to the piece of settled science that every interpretation must, on pain of being wrong, accommodate.
John Bell in 1964 proved a theorem showing that no theory obeying local realism — the assumption that distant systems have pre-existing definite properties unaffected by faraway measurement choices, and that no influence travels faster than light — can reproduce all the statistical predictions of quantum mechanics for entangled particles. Experiments by Alain Aspect in the 1980s, and definitive "loophole-free" versions by groups including Ronald Hanson's in Delft and others in Vienna and Boulder in 2015, closed essentially every conceivable experimental escape hatch and confirmed: nature violates Bell's inequalities, exactly as quantum mechanics predicts. This means something has to give up either locality or realism (or both, in various ways depending on how you slice it) — and interpretations differ dramatically in which they sacrifice. Copenhagen tends to give up realism (there's no definite value before measurement). Bohm gives up locality explicitly and unapologetically. Everett gives up the naive picture of "one outcome" altogether, and with it the framing of the dilemma, since in a branching multiverse there's no unique distant outcome for a local one to be correlated with in the classical sense.
This is where it breaks down: not one experimental anomaly, but a genuine, structural, mathematically demonstrable degeneracy. Three, and really many more than three, complete pictures of reality, each internally consistent, each compatible with every confirmed experiment including the most stringent tests of Bell nonlocality ever performed, and no known experiment that tells them apart.
The contenders — weighed, honestly, for their appeal and their trouble
It's worth laying these out not as a menu of curiosities but as serious scientific research programs, each with working physicists devoted to developing its mathematics, and each with real costs.
Copenhagen and its descendants (including modern variants like QBism, developed by Christopher Fuchs, Rüdiger Schack, and others, which treats the wavefunction as an agent's personal degree of belief rather than an objective feature of the world) have the virtue of asking for the least metaphysical commitment. If you only care about predicting outcomes, you don't need to resolve what's "really happening" between preparations and measurements — and arguably that restraint has been enormously productive, letting generations of physicists build lasers, transistors, and MRI machines without waiting for a foundations seminar to conclude. The cost is that it treats "measurement" as a primitive, undefined term in a fundamental physical theory, an intellectual debt most physicists agree is uncomfortable even if they're willing to defer paying it. QBism sharpens this into a virtue, arguing that quantum mechanics was never meant to be a theory of an observer-independent world at all — but this move strikes many physicists as changing the subject rather than answering the original question, trading a puzzle about physical reality for a theory of personal credence.
Everettian many-worlds has surged in respectability over the past thirty years, championed by David Deutsch, David Wallace, Simon Saunders, and, historically, argued for by figures including Bryce DeWitt who coined the "many-worlds" name itself. Its appeal is austere elegance: it takes the Schrödinger equation utterly literally and adds nothing. No collapse, no hidden variables, no extra postulate beyond the bare mathematics everyone already agrees describes microscopic systems, now applied without exception to the whole universe. Decoherence does real, honest work here, explaining the branching structure. Its cost is metaphysical extravagance on a staggering scale — an unfathomable, continuously multiplying profusion of parallel realities, including for every quantum event in every star and every synapse throughout cosmic history — and the unresolved question of where probability comes from in a theory where, strictly, nothing is left to chance. Critics also press on the "preferred basis problem": why does decoherence pick out branches corresponding to definite positions of pointers and cats, rather than some exotic superposition-of-superpositions basis? The honest answer is that decoherence, driven by the specific form of physical interactions (position-dependent forces, local interactions in space), does pick out an approximately stable "pointer basis" tied to quantities like position — a real result, not hand-waving — but whether this fully and rigorously grounds the appearance of a single classical world, rather than presupposing part of what needs explaining, remains argued over.
Pilot-wave theory offers the most conservative ontology of the three in one sense — particles simply have positions, full stop, no observer needed to bring them into being, no branching multiverse — while paying for that conservatism with an explicit, built-in nonlocality that most physicists find viscerally uncomfortable, even knowing Bell's theorem says some nonlocality is unavoidable for anyone. Its champions, including Sheldon Goldstein, Detlef Dürr, Nino Zanghì, and historically John Bell himself — who wrote admiringly that Bohm's theory showed determinism was "not only possible but ... gives a much more valuable and instructive form of quantum mechanics" than the version he'd been taught — argue it single-handedly dissolves the measurement problem, since there's never any ambiguity about what a "measurement" is: it's just one more physical interaction between particles, guided like any other. Its cost, beyond nonlocality, is that extending the guidance equation cleanly to relativistic quantum field theory, with particle creation and annihilation, has proven technically thorny, though there has been real progress, including work by Dürr, Goldstein, Roderich Tumulka, and Zanghì on Bohmian formulations of quantum field theory using stochastic particle creation events.
There is also a fourth family worth naming honestly, because it is the one genuine attempt to make an interpretation empirically distinguishable rather than merely philosophically distinct: objective collapse models, such as the Ghirardi–Rimini–Weber (GRW) theory and its continuous variant developed by Philip Pearle and refined by Angelo Bassi, GianCarlo Ghirardi, and others. These modify the Schrödinger equation itself, adding a tiny, genuinely random, spontaneous collapse process to every particle's wavefunction, occurring on average roughly once every hundred million years for an isolated particle, but which becomes overwhelmingly frequent and effectively instantaneous for macroscopic collections of many particles entangled together — precisely producing definite outcomes for cats and pointers while leaving single electrons alone for eons. This is not merely an interpretation; it is a different physical theory, making predictions that in principle diverge, ever so slightly, from standard quantum mechanics for sufficiently large or sufficiently isolated systems. That difference is small, but it is not zero, and that is precisely what makes this program experimentally exciting in a way the other three, strictly, are not.
What would settle it — the experiments and proofs actually on the table
This is the part of the story that keeps the question alive rather than merely philosophical, and it deserves to be taken seriously rather than treated as a coda.
The most concrete route runs through objective collapse models, precisely because they are not mere interpretations but modified dynamics. GRW-type theories predict that superpositions of increasingly massive objects should decohere — collapse — at a rate that grows with mass, in a way subtly different from ordinary environmental decoherence. This means that if you could isolate a sufficiently massive object exceptionally well from its environment, and still saw its quantum superposition survive far longer, or shorter, than either standard quantum mechanics or GRW predicts, you would have real, discriminating data. Experimentalists including Markus Arndt's group in Vienna have pushed matter-wave interference to molecules of tens of thousands of atomic mass units. Space-based proposals such as MAQRO have been designed explicitly to test collapse models with nanoparticles in free fall, shielded from decoherence by the vacuum and isolation of orbit, where standard quantum mechanics and GRW-type dynamics would diverge measurably. Optomechanics experiments — mirrors and membranes cooled near their quantum ground state, some pioneered by groups including Markus Aspelmeyer's in Vienna — are creeping toward masses and coherence times where such bounds become meaningful. As of the mid-2020s, experiments have ruled out large swaths of GRW parameter space but have not excluded it entirely, nor confirmed standard quantum mechanics to arbitrary precision at macroscopic scales. This is a genuinely open experimental frontier, and it is realistically decades, not centuries, from either confirming a deviation or squeezing collapse models into such a narrow, fine-tuned corner that most physicists would abandon them as unmotivated.
For pilot-wave theory, the analogous opening is Valentini's "quantum non-equilibrium": if the early universe had not fully relaxed to the Born-rule distribution, relic signatures might survive in the cosmic microwave background or in primordial gravitational waves, subtly deviating from standard inflationary predictions. This is a long shot, constrained already by how thoroughly standard quantum statistics fit existing cosmological data, but it is a real, falsifiable proposal rather than a philosophical stance, and future CMB polarization missions could tighten or loosen the noose further.
For Everettian quantum mechanics, there is, as far as anyone has shown, no analogous experimental handle, precisely because it is constructed to reproduce standard unitary quantum mechanics without modification — its distinctiveness is entirely in what it says exists, not in any new equation. The frontier here is mathematical and philosophical: whether the Deutsch–Wallace decision-theoretic derivation of the Born rule is actually valid, and whether "branch counting" or any other proposal can give a non-circular account of probability in a fundamentally deterministic, all-outcomes-occur theory. This is not an empirical dead end so much as a different kind of open problem — a question for careful proof and counter-proof rather than for a lab bench, and it continues to generate real technical papers and real disagreement among specialists who all accept the same mathematics of quantum theory.
There is a broader, more general possibility worth naming: some no-go theorem, in the spirit of Bell's, might one day show that certain classes of interpretation are logically incompatible with facts we already accept — as Bell's theorem itself did to naive local hidden-variable theories. The 2012 Pusey–Barrett–Rudolph theorem, for instance, showed that under certain assumptions, the wavefunction cannot merely represent an observer's knowledge about a hidden underlying reality — it must correspond to something more directly "real" — which sharply constrains, without wholly eliminating, some flavors of the Copenhagen and QBist stance. This kind of theorem is precisely the sort of progress that can be made even without new experiments: tightening the logical space of viable interpretations by proving that certain combinations of assumptions are simply inconsistent with known quantum phenomena. More such theorems, refining exactly what any viable interpretation must and must not say, are an active and genuinely productive research direction, distinct from experimental tests but no less rigorous.
Why it matters — beyond bragging rights
It is tempting to file this under harmless philosophical hairsplitting, since — after all — every interpretation gives the same predictions, so what does it matter for building better lasers or medical scanners? But that dismissal undersells the stakes in at least three concrete ways.
First, for science itself: quantum computing, quantum cryptography, and quantum metrology are all, at bottom, engineering projects built on entanglement and superposition, and progress on those frontiers has repeatedly been driven by physicists thinking hard about foundational questions — Bell's own theorem arose from foundational discomfort, not engineering need, yet it underlies every device that exploits quantum nonlocality today, from device-independent cryptography to Bell tests used as randomness certifiers. A deeper understanding of decoherence, born from foundational worries about measurement, is now indispensable engineering knowledge for anyone building a quantum computer, where fighting decoherence is the central practical battle. Foundational and applied physics are not as separate as textbooks suggest; today's philosophical irritant is tomorrow's engineering toolkit.
Second, for cosmology and gravity: if you want to write down a quantum theory of the early universe, or of black hole evaporation, or eventually a quantum theory of gravity itself, you need to know what happens when there is no external observer to "measure" anything — since the universe, taken as a whole, has nothing outside it to collapse its wavefunction. This makes the measurement problem inescapable for quantum cosmology in a way it can seem optional for laboratory physics, and it connects directly to unresolved puzzles about the arrow of time and entropy: whether the universe's low-entropy beginning, the direction time seems to flow, and the irreversibility built into collapse (in interpretations that have real collapse) are linked, coincidental, or actually the same problem viewed from different angles. Everettian branching, notably, gives one natural story for why time seems to have a direction — branches only ever proliferate forward, never merge back — that some physicists find suggestive and others find merely decorative.
Third, and perhaps most provocatively, this question is entangled with a genuinely open frontier about whether information itself is a physical, ontologically basic ingredient of reality or a derivative bookkeeping notion — the puzzle of is information physical. QBism leans hard into treating quantum states as information about an agent's beliefs; objective collapse and pilot-wave theories treat the wavefunction (or something like it) as a genuinely physical field regardless of any observer; Everettians treat the universal wavefunction as the single most physically real object there is, full stop, with everything else, including individual "worlds," a derived, coarse-grained description of it. Depending which of these is right, the deepest layer of reality is either irreducibly informational, irreducibly branching, or irreducibly a field guiding literal particles — three utterly different pictures of what the universe fundamentally is made of, hiding underneath data everyone agrees on.
The people chasing it
The roll call spans a century and includes some of the most consequential and most argumentative minds in physics.
physicist[Niels Bohr|1885–1962|Architect of the Copenhagen tradition, insisted quantum mechanics describes what can be said about nature using classical concepts, not nature "in itself"]
Origins · 1 minHugh Everett III and how this was discovered
physicist[David Bohm|1917–1992|Revived and completed de Broglie's pilot-wave theory in 1952 while politically exiled from the United States during the McCarthy era, showing determinism and hidden variables were logically compatible with quantum statistics]
:::physicist[John Stewart Bell|1928–1990|Proved in 1964 that no local hidden-variable theory can reproduce all quantum predictions, reshaping the entire debate from philosophy into experimental physics, while remaining a sympathetic critic of orthodoxy who admired Bohm's theory}
Alongside them stand the experimentalists who turned foundational disputes into laboratory results: Alain Aspect, John Clauser, and Anton Zeilinger, who shared the 2022 Nobel Prize in Physics explicitly for experiments on entangled photons establishing violation of Bell inequalities and pioneering quantum information science — a rare instance of the Nobel committee honoring work whose original motivation was philosophical to its core. Wojciech Zurek and Dieter Zeh built the modern theory of decoherence. David Deutsch, David Wallace, and Simon Saunders have spent careers making Everett's proposal mathematically and philosophically rigorous. Sheldon Goldstein, Detlef Dürr, and Nino Zanghì have done the same for Bohmian mechanics. GianCarlo Ghirardi, Alberto Rimini, Tullio Weber, Philip Pearle, and Angelo Bassi built and refined objective collapse models into genuinely falsifiable physics. Christopher Fuchs and Rüdiger Schack pressed the case for QBism. None of these people agree with each other, and several of them have said so, in print, with considerable feeling.
:::history 1926 — Schrödinger publishes the wave equation; Born proposes the probabilistic interpretation of the wavefunction. 1935 — Einstein, Podolsky, and Rosen challenge completeness of quantum mechanics; Schrödinger names entanglement and poses the cat paradox. 1952 — Bohm publishes a complete deterministic hidden-variable reformulation of quantum mechanics. 1957 — Everett proposes the relative-state (many-worlds) formulation in his Princeton dissertation. 1964 — Bell proves his inequality, making the foundations debate experimentally decidable in principle. 1982 — Aspect's experiments confirm violation of Bell inequalities under improved conditions. 2015 — Loophole-free Bell tests in Delft, Vienna, and Boulder close remaining experimental escape routes. 2022 — Aspect, Clauser, and Zeilinger receive the Nobel Prize for entangled-photon Bell experiments.
The open frontier
Stand back from the technical detail and the situation is almost paradoxical: quantum mechanics may be the most rigorously tested theory in the history of science, and simultaneously one of the least understood at the level of what it is telling us exists. That combination should unsettle anyone who thinks scientific progress means steadily converging pictures of reality. In this one, singularly important corner of physics, seventy years of extraordinary theoretical and experimental effort have sharpened the alternatives, ruled out entire classes of naive hidden-variable theories, and forced every remaining serious contender to swallow something deeply counterintuitive — nonlocality, or branching multiplicity, or the dissolution of an observer-independent world — without ever forcing a choice among them.
Maybe that choice is coming, on a timescale of decades, from optomechanics experiments finally massive and isolated enough to catch a collapse model in the act, or from a cosmological signature of quantum non-equilibrium, or from some as-yet-unimagined theorem that closes off an entire family of interpretations the way Bell's theorem closed off local realism. Or maybe the honest long-run answer is that empirical physics has reached the edge of what data alone can settle, and the remaining work is philosophical: rigorous, mathematically constrained, but not decidable by any experiment even in principle. Living with that possibility, without either pretending the question is already answered or abandoning the effort to answer it, may be the most intellectually mature stance available right now.
- Can the Born rule be derived, non-circularly, from Everettian branching alone — or does every attempt secretly assume probability to get probability out?
- Is there any physically realizable experiment, even a highly exotic one, that could distinguish Everett, Bohm, and Copenhagen given how carefully each reproduces standard statistics?
- Will optomechanical or space-based tests of objective collapse models find a deviation, or squeeze GRW-type theories out of existence within a generation?
- Does a future theory of quantum gravity demand a specific interpretation, because "measurement" and "observer" cannot be defined without a fixed external background the whole universe doesn't have?
- Is quantum non-equilibrium a real possibility that has left traces in the early universe, or a mathematical curiosity with no cosmological echo at all?
Ask the tutor: If decoherence explains why we never see a superposition, but not why we see any one outcome rather than all of them, what — if anything — is decoherence still missing?