Episode 32 - What does it mean? The Interpretations

Tomlin Hawthorne

You remember Schrödinger's cat. A sealed box, a radioactive atom, a flask of poison. Until the box is opened, the theory says the cat is in a blend of alive and dead.

Tomlin Hawthorne

Now let's add a person. Your friend is in a laboratory with the box. The laboratory door is shut. You're outside in the corridor.

Tomlin Hawthorne

Your friend opens the box and looks. She sees one thing or the other. For her, it's settled.

Tomlin Hawthorne

But you haven't opened the laboratory door. From where you stand, the whole lab is one big sealed box. With an atom in it, and a cat, and a friend.

Tomlin Hawthorne

So here's the question. Before you open the door: is your friend in a blend of having seen a live cat and having seen a dead one?

Tomlin Hawthorne

If you say yes, you're saying a person can be in two states of mind at once, and that she'll only become definite when you look. And then who's looking at you?

Tomlin Hawthorne

If you say no, she saw something definite when she opened the box, then her looking counted as a measurement and the atom's decaying didn't. Why? What's she got that an atom hasn't?

Tomlin Hawthorne

There's no agreed answer. That puzzle was posed in 1961 by a physicist named Eugene Wigner. And where you come down on it decides, more or less, which of the rival accounts of reality you'll find yourself defending. Today I'll lay them out. You can choose.

Tomlin Hawthorne

Last time you drew a grid, three by three, and counted. Five sames in nine, whatever instruction card you tried. And nature gives four and a half. So the particles aren't carrying instructions. That's Bell's theorem, and you proved it on a sheet of paper.

Tomlin Hawthorne

So here's where we are. We have a theory, quantum mechanics. It's a century old. It has never made a wrong prediction. Its accuracy, in the best cases, is like measuring the width of North America to within a hair. Every piece of electronics you own depends on it.

Tomlin Hawthorne

And its creators couldn't agree what it says about the world. Nor can anybody now. Today's question: what is quantum theory telling us about reality?

Tomlin Hawthorne

That's the situation. And I think you may be better placed to handle it than most physics students. They're trained to calculate. You're trained to hold several readings of one text in your head and weigh them. This is that.

Tomlin Hawthorne

First, the text. Let me state the problem as cleanly as I can, because it often gets lost in the fog.

Tomlin Hawthorne

Quantum mechanics has two rules for how things change.

Tomlin Hawthorne

Rule one. When a system is left alone, its wavefunction evolves smoothly and predictably, according to Schrödinger's equation. There's no chance in it at all. Superpositions stay superpositions. The electron goes through both slits. The atom is decayed-and-not. This rule is as deterministic as anything in Newton.

Tomlin Hawthorne

Rule two. When a measurement is made, something else happens. The superposition vanishes. You get one definite result. Which one is random, with the odds given by the wavefunction. This is usually called the collapse.

Tomlin Hawthorne

Two rules. And they're utterly different in kind. One is smooth, one is a jump. One is certain, one is chance.

Tomlin Hawthorne

So you'd expect the theory to tell you, very precisely, when to use which. What counts as a measurement?

Tomlin Hawthorne

It doesn't say. Nowhere in the theory is there a definition. A measuring device is made of atoms. Atoms obey rule one. So why doesn't the device just go into a superposition too, of having read this and having read that? And the person reading the device? She's made of atoms as well. That's the corridor outside the lab.

Tomlin Hawthorne

This is called the measurement problem. In practice, physicists have an excellent nose for when to switch rules, and they always get the right numbers. But a nose isn't a theory.

Tomlin Hawthorne

So. Five witnesses.

Tomlin Hawthorne

Bohr's view, as near as anyone can pin it down, was this. The wavefunction isn't a picture of what's going on. It's a tool for predicting the results of experiments. An experiment has to be described in ordinary language: this dial pointed here, that counter clicked. The quantum formalism connects one such description to the next. And to ask what the electron was really doing in between, when nobody was measuring, is to ask a question that has no meaning. There's no deep reality to describe. Or if there is, it's not the business of physics.

Tomlin Hawthorne

It's a position of great discipline. It refuses to say more than can be checked. And the cost is that it needs a line. On one side, the quantum system. On the other, the classical measuring apparatus, the big, solid, ordinary world. And nobody has ever been able to say where the line is, or why there should be one. The apparatus is made of atoms.

Tomlin Hawthorne

Most working physicists, most of the time, adopt a rough version of this without thinking about it much. The physicist David Mermin, who gave us the red and green lights, summed up the attitude in a phrase that stuck. Shut up and calculate.

Tomlin Hawthorne

The second witness says something so extravagant that for thirty years it was hardly discussed.

Tomlin Hawthorne

Princeton, 1957. A graduate student called Hugh Everett, twenty-six, sharp, impatient, fond of a drink. He asked a simple question. What if we just drop rule two?

Tomlin Hawthorne

No collapse. Ever. Rule one is the whole of physics. The wavefunction always evolves smoothly, for atoms, for instruments, for cats, for people.

Tomlin Hawthorne

Then what happens when you measure? Take the atom in a blend of decayed and not decayed. The detector interacts with it, and by rule one, the detector goes into a blend: clicked, and not clicked. You look at the detector. And you go into a blend: a you that saw a click, and a you that didn't.

Tomlin Hawthorne

Both outcomes happen. And each version of you sees one definite result, and remembers one definite past, and is entirely unaware of the other. From the inside, it looks exactly as though the wavefunction collapsed, at random. But nothing collapsed. The world has, in effect, branched.

Tomlin Hawthorne

And it does that at every quantum event, everywhere, all the time. This is the many-worlds interpretation.

Tomlin Hawthorne

Its attraction is real. It has no measurement problem. It has one rule, not two. There's no mysterious line and no special role for observers. It takes the equation that we know is right, and it believes it, all the way.

Tomlin Hawthorne

Its cost is also obvious. A staggering number of other worlds, none of which can ever be detected. And a puzzle about probability: if everything happens, what does it mean to say something had a thirty per cent chance?

Tomlin Hawthorne

Everett's supervisor took the thesis to Copenhagen to show Bohr's circle. They dismissed it. One of them called Everett indescribably stupid. He'd already left physics by then. He went to work for the Pentagon, calculating the consequences of nuclear war, and made a good deal of money in computing. He never published another word on quantum mechanics. He smoked three packs a day, and died of a heart attack at fifty-one. He'd asked that his ashes be put out with the rubbish.

Tomlin Hawthorne

His idea was rescued in the seventies and is now, by some counts, the most popular view among physicists who work on the foundations. His son, by the way, is the singer of the band Eels.

Tomlin Hawthorne

This goes back to Louis de Broglie in 1927, and it was worked out fully in 1952 by an American, David Bohm. It's called the pilot wave theory.

Tomlin Hawthorne

There's a real wave, and there's a real particle. The wave spreads out and goes through both slits. The particle goes through just one, like a sensible bullet. But the wave guides the particle, pushing it this way and that, steering it toward the bright stripes and away from the dark ones. The randomness is only our ignorance of exactly where the particle started.

Tomlin Hawthorne

No collapse, no branching, no line. It's deterministic. Einstein ought to have loved it.

Tomlin Hawthorne

He didn't. He called it too cheap. And the reason is the cost. For it to work, the wave guiding a particle here has to depend, instantly, on what's happening to its entangled partner over there. The theory is openly, robustly non-local. Spooky action is built into the machinery. After Bell, we know something of the sort is unavoidable. Bohm's theory just doesn't hide it.

Tomlin Hawthorne

Bohm had a hard road. He'd been called before the House Un-American Activities Committee in the McCarthy years and refused to name names. Princeton dropped him. He ended up in exile in Brazil, where the American authorities took away his passport, and then in London. His theory was ignored for a generation, partly on the physics and partly, I think, because of who he was. It was John Bell who took it seriously. It was reading Bohm that set Bell on the road to his theorem.

Tomlin Hawthorne

The fourth witness says: the collapse is real, and it's physics.

Tomlin Hawthorne

This is the proposal of three Italians, Ghirardi, Rimini and Weber, in 1986. Rule two isn't something that happens when you measure. It's something that happens spontaneously, at random, to every particle, very, very rarely. A single electron might go a hundred million years without collapsing. So for one electron, you never notice.

Tomlin Hawthorne

But a cat is made of a vast number of particles, all entangled with each other. If any one of them collapses, they all do. And with that many particles, one collapses every fraction of a millionth of a second. So big things are always definite. Small things are almost always fuzzy. The line that Bohr couldn't locate is drawn by the mathematics.

Tomlin Hawthorne

The cost: you have to change Schrödinger's equation, adding a new term nobody has seen. The benefit: this is the one interpretation that makes different predictions from the others. In principle, it can be tested. There are experiments under way now, trying to put larger and larger objects into superposition, to see if nature gives up at some size. So far, it hasn't.

Tomlin Hawthorne

And the fifth witness questions the questioner.

Tomlin Hawthorne

This is a family of views that have grown up since the 1990s. One version is associated with Carlo Rovelli, whose books I've mentioned. It's called the relational interpretation. Another goes by the name QBism.

Tomlin Hawthorne

The common idea is that a wavefunction isn't a description of a thing in itself. It describes a relationship. It's the state of this system relative to that one. Or, in the other version, it's a statement of what a particular observer is entitled to expect.

Tomlin Hawthorne

Go back to the corridor. Relative to your friend, who's opened the box, the cat is definitely alive, let's say. Relative to you, outside the door, the cat and the friend are still in a blend. And both of those are correct. There's no contradiction, because there's no view from nowhere. No single account of the facts that holds for everybody at once. Facts are relative to an observer, a bit as motion is relative to an observer in Galileo.

Tomlin Hawthorne

The cost is that last sentence. You give up the idea that there's one world with one set of facts. Some people find that an elegant extension of relativity. Others find it a polite name for giving up.

Tomlin Hawthorne

There they are. Five accounts.

Tomlin Hawthorne

There's nothing really there between measurements. Everything happens, and the world branches. There are particles guided by an instant-acting wave. Collapse is a real physical event. Facts are relative to the observer.

Tomlin Hawthorne

And here's the extraordinary thing. With the one exception I mentioned, all of these make exactly the same predictions, for every experiment anyone has ever done or knows how to do. You can't choose between them in a laboratory.

Tomlin Hawthorne

There have been polls. At conferences on the foundations of physics, somebody hands round a questionnaire. Which interpretation do you favour? The results are all over the place. No view gets a majority. A sizeable group ticks Copenhagen, a sizeable group ticks many worlds, there are clusters for each of the others, and quite a few tick: none of the above.

Tomlin Hawthorne

I don't know of another case like it. The most successful theory in the history of science, and a century on, the experts can't agree what it's about.

Tomlin Hawthorne

If you ask me which I hold, I'll be honest: I change my mind. On a day when I think the mathematics should be trusted, I lean toward Everett. On a day when I think physics should only talk about what can be observed, I lean toward Bohr, or Rovelli. Plenty of physicists are certain. They're certain of different things.

Tomlin Hawthorne

The myth of the week is that it's been settled. You'll get this impression from popular books, because authors tend to write up the view they favour as though it were the result. In the last twenty years it's usually been many worlds, which makes for good television. It hasn't been established. Nor has any of the others.

Tomlin Hawthorne

And a sub-myth: that the many worlds are places you could visit, with a version of you who took the other job. The branches can't interact. That's what makes them branches. As far as any experiment is concerned, they might as well not be there, which is rather the objection.

Tomlin Hawthorne

And the question to take away. It's a philosopher's question, and it has a name: underdetermination.

Tomlin Hawthorne

If five theories agree on every possible observation, is choosing between them still science? Or is it philosophy? Or just taste?

Tomlin Hawthorne

Physicists who make a choice appeal to things like simplicity, elegance, economy of assumptions. Many-worlders say: we have the fewest rules. Their critics say: you have the most universes. Each calls that Occam's razor.

Tomlin Hawthorne

You'll have seen this before. Two readings of a poem, each consistent with every word on the page. Two accounts of the causes of a war, each fitting all the documents. In your field, what settles such things? Does anything? And should we expect physics to be different?

Tomlin Hawthorne

Your challenge is a debate, and you need one other person.

Tomlin Hawthorne

Pick one of the five. Whichever you found most appealing, or most outrageous. Take ten minutes to prepare. Then defend it, for five minutes, against your friend, whose job is to attack. What are the facts? What happens to the cat? What's the price?

Tomlin Hawthorne

Then swap. You attack, and they defend, the same one.

Tomlin Hawthorne

And then, if you have the stamina, do it again with the interpretation you liked least. That's often the one people end up respecting.

Tomlin Hawthorne

Next time. After three episodes of mystery, something solid. Literally. The atom is nearly all empty space, a fly in a cathedral. You're made of atoms. So is the floor. So why, when you stand on it, don't you fall straight through?