Quantum Essentials · Module 1 · Chapter 6 of 17
What do the interpretations say, and how does popular culture show them?
Physicists agree on what quantum theory predicts. They don't agree on what it says about reality. This chapter builds up the puzzle step by step, explains the six main answers in depth, compares them, and shows how films get them wrong.
The short answer (tap to open)
Physicists agree on what quantum theory predicts but not on what it says about reality. The puzzle is measurement: the main equation never picks one result, yet we always see one. Copenhagen, Many-Worlds, information-based views, pilot waves, collapse models and others answer differently. They all predict the same results, so for now the choice rests on other grounds.
Predict
Do physicists agree on what's really going on?
After this page you can
- explain the measurement problem, step by step;
- describe six main interpretations, and where each gets its odds;
- compare them on the questions they disagree about;
- spot what films and headlines get wrong about them.
Part A · The puzzle
What is the measurement problem?
A1. What do you already know?
From chapter 2: a qubit's state is like a direction, and measuring it asks a question with two answers, such as "up or down?". The angle between the state and the question sets the odds. That rule is the Born rule.
Our example for this whole page is a qubit tilted so that a measurement gives "up" 70% of the time and "down" 30%. Every view on this page agrees on those numbers. What they argue about is what happens when you make one measurement.
A2. How do quantum states change?
Between measurements a quantum state changes by one rule, the Schrödinger equation. It's smooth and completely predictable. Give it a state now and it tells you exactly what the state will be later. There's no randomness in it at all.
It has one more property that matters here. It never chooses. If you start with a mix of possibilities, it carries every one of them forward side by side, each with its own weight. Physicists call this linearity. Keep that word in mind, because it's the source of the whole puzzle.
A3. What happens when you run a measurement through that rule?
Step 1 of 4
The qubit, before
A qubit at 70% up and 30% down. A detector waits in its "ready" position. So far nothing has interacted.
All four steps, in words
- The qubit, before. A qubit at 70% up and 30% down. A detector waits in its "ready" position. So far nothing has interacted.
- The qubit meets the detector. Apply rule 1 to qubit plus detector. The result is "qubit up and detector reads up" with weight 70%, plus "qubit down and detector reads down" with weight 30%. The detector hasn't picked a result. It has joined the mix. Physicists say the two are now entangled.
- You look at the detector. You're made of atoms too, so apply rule 1 again. Now it's "detector reads up and you see up" plus "detector reads down and you see down". The "and" has spread to you.
- But you only ever see one. Every real measurement gives you one result, never a blend. Rule 1 alone never gets you there. Something is missing between the equation and your experience.
The detector, the camera and you are all made of atoms, and atoms obey the same equation. So each new link in the chain just picks up the same "and". Nothing in the maths ever selects one result. The Hungarian-American mathematician John von Neumann set out this chain in 1932.
A4. What is the textbook's patch, and why is it a problem?
Textbooks fix this by adding a second rule. When a measurement happens, the state jumps to one outcome, picked with the Born-rule odds. Physicists call this jump collapse.
So quantum mechanics runs on two rules. Rule 1: smooth, predictable change when nobody measures. Rule 2: a sudden random jump when someone does.
The trouble is that the rules disagree about the same event, and the theory never says when to switch from one to the other. Is it at the detector? The camera? Your eye? Your brain? Each is just more atoms obeying rule 1. This unanswered question is the measurement problem.
A5. What does Schrödinger's cat show?
In 1935 Erwin Schrödinger made the gap impossible to ignore. Put a cat in a sealed box with a radioactive atom. If the atom decays, a detector breaks a vial of poison.
Apply rule 1 to the whole box and after an hour you get "atom decayed and cat dead" plus "atom intact and cat alive", each with a weight. Schrödinger thought that conclusion was absurd. He wasn't saying cats really are half dead. He was showing that something is missing from the theory. Optional page 3a goes through it in full.
A6. Which three questions must every answer settle?
- Does one outcome really happen, or do they all?
- If one happens, what picks it: real chance, or detail we can't see?
- Is the wavefunction a real thing, a summary of something deeper, or our knowledge?
Keep these three in mind. Each interpretation below answers them differently, and at the end you'll answer them yourself.
Part B · The six main answers, in depth
What are the six main answers?
Each section follows the same pattern: where it came from, the core idea, our 70/30 qubit measured step by step, where the 70% comes from, what supporters like, the main objections, and where it stands today. Five more views get the same treatment on optional page 6b.
B1 · 1 of 6
What does Copenhagen say?
Copenhagen
Niels Bohr, Werner Heisenberg, Max Born · 1920s
"Use the maths to predict results. Don't ask what lies underneath."
Where it came from
Worked out in the 1920s around Niels Bohr's institute in Copenhagen, with Werner Heisenberg, Max Born and others. It was never one fixed doctrine. Today "Copenhagen" usually means the approach most textbooks teach.
The idea
Quantum theory is a tool for predicting what our instruments will record. Bohr insisted that experiments must be described in everyday terms: a detector clicked, a needle moved. So Bohr treats the measuring device as an ordinary classical object.
The wavefunction encodes the odds of what we'll find. Asking what the particle is "really doing" between measurements has no definite answer in this view.
Measuring our 70/30 qubit, step by step
- A wavefunction describes the qubit, giving 70% up and 30% down.
- The detector, an ordinary classical object here, registers a result.
- You update the wavefunction to "up". This collapse is a rule for using the theory, not a physical process the theory describes.
Where the 70% comes from
The Born rule is simply taken as a basic rule of the theory. Max Born proposed it in 1926 and shared the 1954 Nobel Prize for it.
What supporters like
- No experiment has ever contradicted it.
- It asks for nothing beyond the maths, and it matches how most physicists actually work day to day.
Main objections
- Where is the line between the quantum system and the classical detector? Detectors consist of atoms too.
- It seems to give measurement, and perhaps observers, a special role that nothing else in physics has.
Where it stands today
The most chosen view in the 2025 Nature survey, at 36%. Many who pick it treat it as the practical default rather than a firm belief.
B2 · 2 of 6
What does Many-Worlds say?
Many-Worlds
Hugh Everett · 1957
"Every outcome happens, each in its own branch."
Where it came from
Proposed by Hugh Everett in his 1957 Princeton PhD thesis, supervised by John Wheeler. Physicists largely ignored it until Bryce DeWitt popularised it in the 1970s and gave it the name "many worlds".
The idea
Take rule 1 seriously and drop rule 2. The whole universe has one wavefunction, and it always changes smoothly. Nothing ever collapses.
So the "and" in step A3 is real. Both outcomes happen. A process physicists call decoherence (chapter 3) makes the branches stop interfering with each other almost instantly, so each one behaves like a separate world.
Measuring our 70/30 qubit, step by step
- The qubit starts at 70% up and 30% down.
- The detector interacts with it and the universe splits into two branches: one where the detector reads up, and one where it reads down.
- Each branch contains a version of you who sees one result. Each version experiences a single outcome, which is why measurement feels like collapse.
Where the 70% comes from
This is the hardest part. Branches aren't counted one by one; each carries a weight. Supporters such as David Deutsch and David Wallace argue that a rational person should bet according to those weights, which recovers the Born rule. Whether that argument works is still debated.
What supporters like
- No collapse and no special role for observers. One equation does everything.
- It sits naturally with cosmology, where there's no outside observer to collapse the universe.
Main objections
- If every outcome happens, what does "70% likely" mean? Physicists call this the probability problem.
- It implies an enormous number of worlds, and what counts as a branch is only approximately defined.
Where it stands today
15% in the 2025 Nature survey, which offered it as one option with consistent histories. It's popular among some cosmologists and quantum computing theorists.
B3 · 3 of 6
What do information-based views say?
Information-based
Anton Zeilinger, Časlav Brukner and others · from 1999
"The wavefunction is information about what you could find. Nature limits how much there is."
Where it came from
It grew out of Copenhagen and John Wheeler's slogan "it from bit" (1989). In 1999 Anton Zeilinger proposed that the most basic quantum system carries exactly one bit of information, and he and Časlav Brukner built on that idea. Other physicists, such as Lucien Hardy in 2001, later rebuilt quantum theory's maths from a few simple rules about information.
The idea
The wavefunction doesn't describe a physical wave. It sums up the information available about what measurements will show.
Nature puts a hard limit on that information. A qubit can hold one bit's worth: enough to answer one question with certainty, but not every question. So if it's set up to answer "vertical or horizontal?" for sure, the answer to "diagonal or anti-diagonal?" has to be random. Randomness is the price of the limit, not a gap in our knowledge.
Measuring our 70/30 qubit, step by step
- The available information gives 70% up and 30% down; the qubit's one bit is spent on a different question.
- The detector interacts and records "up".
- The information updates to "up". Nothing physical jumped; what changed is what can now be known.
Where the 70% comes from
Here the Born rule follows from the information limits. Reconstructions such as Hardy's show that once you accept a few reasonable rules about how information behaves, the squared rule is essentially the only one that works.
What supporters like
- It explains why quantum randomness exists, instead of just accepting it.
- It fits naturally with quantum information science, the field behind quantum computing and cryptography.
Main objections
- Information about what? Critics say it describes our knowledge but stays silent on the world that knowledge is about.
- Some see it as Copenhagen in new words, with the same unanswered question about where measurement begins.
Where it stands today
The second most chosen family in the 2025 Nature survey: 17% picked epistemic or information-based views (a group that also includes QBism, on page 6b). Strong among quantum information researchers, especially in Europe.
B4 · 4 of 6
What does pilot wave theory say?
Pilot wave
Louis de Broglie 1927 · David Bohm 1952
"Particles have definite positions, guided by a real wave."
Where it came from
Louis de Broglie presented an early version in 1927 but dropped it after criticism. David Bohm rediscovered and completed it in 1952. John Bell later championed it, and it helped inspire his famous theorem.
The idea
Particles always have definite positions, even when nobody looks. The wavefunction is a real wave that guides them, rather as a surfer rides a wave.
Everything is deterministic. If you knew every particle's exact starting position, the future would be fixed.
Measuring our 70/30 qubit, step by step
- The qubit's particle has a definite position that we don't know. Its guiding wave splits into an "up" part and a "down" part.
- The particle rides one of the two parts. Which one depends only on its exact starting position.
- The detector registers "up". The empty "down" part of the wave carries on but no longer affects anything.
Where the 70% comes from
If starting positions are spread out according to the Born rule, the guiding equation keeps them spread that way forever. Physicists call this quantum equilibrium. Since we can't control exact starting positions, we see exactly the Born-rule odds.
What supporters like
- A clear, picturable world that is fully deterministic.
- Measurement needs no special rule, and you can draw the paths particles take, for example through the double slit.
Main objections
- It's openly non-local. The guiding wave links distant particles instantly, though it can't send messages.
- Physicists have found it hard to extend to relativity, and the "empty" parts of the wave never go away.
Where it stands today
7% in the 2025 Nature survey, with a small but active research community.
B5 · 5 of 6
What does objective collapse say?
Objective collapse
Ghirardi, Rimini and Weber 1986 · Pearle · Diósi and Penrose
"The wave collapses by itself, at random, and faster for bigger things."
Where it came from
GianCarlo Ghirardi, Alberto Rimini and Tullio Weber proposed the first full model in 1986, which physicists call GRW. Philip Pearle and others developed smoother versions, and Lajos Diósi and Roger Penrose proposed collapse linked to gravity.
The idea
Change rule 1 slightly. Every particle has a tiny chance each second of its wavefunction suddenly narrowing to one place. For a single particle this happens about once in 300 million years.
But the particles in an object share one joint state, so one narrowing collapses the whole object. An object with a hundred thousand billion billion particles collapses within a fraction of a microsecond. Small things stay quantum and big things don't.
Measuring our 70/30 qubit, step by step
- Nothing collapses the single qubit by itself, in practice.
- Once it's linked with the detector, which has trillions of particles, a spontaneous collapse hits the detector almost at once.
- The pointer settles on "up". The collapse picks the result at random, with the Born-rule odds built into the rule.
Try it · the collapse clock
How long until a spontaneous collapse? Slide from one particle up to a cat.
Typical time to a collapse: 317 million years
GRW's suggested rate: about 10⁻¹⁶ collapses per particle per second. Time ≈ 1 ÷ (particles × rate).
The collapse clock, in a table
| Object | Particles | Typical time to a collapse |
|---|---|---|
| a single electron | 1 | 317 million years |
| a small molecule | 10^3 | 316,857 years |
| a large virus | 10^10 | 12 days |
| a speck of dust | 10^15 | 10 seconds |
| a grain of sand | 10^20 | 100 microseconds |
| a raindrop | 10^22 | 1 microsecond |
| a cat | 10^27 | 10 picoseconds |
GRW's suggested rate: about 10⁻¹⁶ collapses per particle per second. Time ≈ 1 ÷ (particles × rate).
Where the 70% comes from
The collapse rule builds in exactly the Born-rule odds.
What supporters like
- One world, a real physical collapse, and no special role for observers.
- It makes predictions that experiments can check.
Main objections
- It adds new constants of nature chosen to fit, with no independent reason for their values.
- Collapses would cause a very slight random heating and a faint X-ray glow. Experiments haven't seen these yet, which rules out some versions.
Where it stands today
A minority view, but taken seriously because it's testable. Searches underground and in precision labs continue.
B6 · 6 of 6
What does CSD say?
CSD: Quantum Without Dice
Constraint-Surface Dynamics · Zayn Blore · from 2025
"One world. The odds come from detail we can't see."
CSD is my own work, so treat this section as the author's view. It's here as one of the options, not as the answer.
Where it came from
Constraint-Surface Dynamics (CSD) is a research programme that began in 2025. Its mathematics lives in Lean, a language in which a computer checks every step of every proof. Its papers are open preprints on Zenodo and Qeios and haven't yet been through journal peer review. So far it covers systems with a limited number of basic states, such as qubits, and leaves relativity aside.
The idea
The world is always in one definite state and moves by fixed rules, with no dice anywhere. The space of possible states holds more detail than the wavefunction records.
The wavefunction is a real but incomplete summary, like the temperature of a gas, which the molecules fix but which leaves out where each one is. No preparation can control the missing "fine detail".
Measuring our 70/30 qubit, step by step
- The qubit's full state includes fine detail that nobody knows.
- Measuring links it with the detector. The joint motion carries the fine detail into one of two regions, set by the detector before the qubit arrived.
- The region it lands in is the result, recorded in the detector. "Collapse" is just updating your summary once you know the result.
Where the 70% comes from
The share of possible fine details that lead to "up" is exactly 70%. The only ruler the symmetry of the space allows measures that share, and repeated runs settle on it. Lean checks these steps, starting from CSD's assumed space of fine detail.
What supporters like
- One world, no dice, and no special role for observers.
- The odds come from geometry rather than assumption, and a computer checks the mathematics.
Main objections
- It's new and not yet peer reviewed, and it assumes the space of fine detail exists.
- Like every hidden-detail view it must accept effects that depend on the whole setup and correlations beyond what locality allows. It accepts both openly. How the detector's interaction carves out the two regions is modelled but not yet fully derived.
Where it stands today
One researcher's programme. The CSD glossary goes deeper, with interactive models. CSD glossary
Part C · All compared
How do all the answers compare?
| Interpretation | One world? | What picks the result? | The wavefunction is… | Collapse? | Deterministic? | Non-local story? | Extra ingredient | Testably different? | Main objection |
|---|---|---|---|---|---|---|---|---|---|
| Copenhagen | Yes | Chance (a basic rule) | A tool for predictions | Yes, as a rule of use | No | Not addressed | A quantum/classical split | No | Where is the split? |
| Many-Worlds | No, many branches | Nothing: all happen | Everything that's real | No | Yes | Arguably no | None | No | What do the odds mean? |
| Information-based | Yes | Chance, forced by the information limit | Information about outcomes | Information updates | No | Not addressed | A limit on information | No | Information about what? |
| Pilot wave | Yes | Exact starting positions | A real guiding wave | No | Yes | Yes | Particle positions | No | Instant links at a distance |
| Objective collapse | Yes | Random spontaneous collapse | Real | Yes, physically | No | Yes | New collapse constants | Yes | No evidence yet |
| CSD: Quantum Without Dice | Yes | Fine detail we can't see | A real but incomplete summary | No; the summary updates | Yes | Yes | A space of fine detail | Not yet | New and unreviewed |
| QBism · Read more in 6b | Yes | Nothing physical; your experience | Your beliefs | Only beliefs update | Not applicable | No | None in the world | No | What is the world itself like? |
| Relational · Read more in 6b | Yes, but facts are relative | Chance | Relations between systems | Relative to each system | No | No | None | No | Relative to what, ultimately? |
| Consistent histories · Read more in 6b | Yes, within a chosen framework | Chance | A tool for histories | No | No | No | Rules for framings | No | Which framework is "true"? |
| Ensemble · Read more in 6b | Silent | Silent | Describes many runs, not one | No | Silent | Silent | None | No | Says nothing about single runs |
| Superdeterminism · Read more in 6b | Yes | Hidden detail, linked to the settings | A summary of something deeper | No | Yes | No | Correlated settings | Possibly | Undermines free choice of settings |
"Non-local" here means the theory's story involves influences between distant places. None of them lets you send a message faster than light (chapter 5).
Part D · Which one are you?
Which one are you?
1. After the measurement, is there one world, or one for each outcome?
2. Was "up" decided by real chance, or by detail we can't see?
3. The wavefunction is…
How each view answers the three questions
- Copenhagen: One world; Real chance; Our knowledge or information.
- Many-Worlds: One for each outcome; Detail we can't see; A real thing in the world.
- Information-based: One world; Real chance; Our knowledge or information.
- Pilot wave: One world; Detail we can't see; A real thing in the world.
- Objective collapse: One world; Real chance; A real thing in the world.
- CSD: Quantum Without Dice: One world; Detail we can't see; A summary of something deeper.
- QBism: One world; Real chance; Our knowledge or information.
- Relational: One world; Real chance; Our knowledge or information.
- Consistent histories: One world; Real chance; Our knowledge or information.
- Ensemble: One world; Real chance; A summary of something deeper.
- Superdeterminism: One world; Detail we can't see; A summary of something deeper.
Part E · What it means
Why do they all agree?
Each interpretation reproduces the Born rule by design, so each predicts the same odds for every experiment done so far. Physicists call this empirical equivalence. Your phone, an MRI scanner and a quantum computer work identically whichever one is true.
That's why the debate has lasted a century. You can't settle it by building a better detector, because all of them agree on what the detector will show. The argument is about what's real underneath, and that has to be judged on other grounds: simplicity, consistency with relativity, and how well each explains the odds.
Could an experiment ever tell them apart?
Collapse models, yes. Because they change rule 1 slightly, they predict tiny extra effects, such as a faint spontaneous glow of X-rays from ordinary matter and a very slight random heating. Experiments deep underground and in precision labs have looked for these and found nothing so far, which rules out some versions and squeezes the rest.
The others, probably not directly. Some physicists have proposed tests in which a quantum computer simulates a small "observer". These "Wigner's friend" experiments probe what any interpretation can consistently say about observers, and teams have run small photon versions since 2019. They sharpen the debate rather than end it.
Does any of this matter for technology?
For building things, no. Engineers designing a quantum computer or an atomic clock use the same maths whichever view they hold. But interpretations shape how people think and explain. David Deutsch, who described the first universal quantum computer in 1985, has argued that Many-Worlds is the most natural way to understand why quantum computers are powerful. Others explain the same machines without it. The rest of this course uses the shared maths and points out where a view adds insight.
Myths and media
Media check
Right, half right or wrong? Call it first.
Card 1 of 5
Doctor Strange in the Multiverse of Madness, 2022: "You can travel between universes."
Check
Three quick ones
Collect
Your aha card
Same numbers, different stories.
Every interpretation predicts the same results. They differ on what's real.
Go further
- 6a · Real-world caseMultiverses on screen: what Everything Everywhere All at Once and Marvel get right and wrong about Many-Worlds
- 6b · A step deeperFive more answers: QBism, relational, consistent histories, ensemble and superdeterminism
- 6c · People storyHugh Everett: he invented Many-Worlds, left physics, and his son became E of the band Eels