Quantum realism mounts a charge. Prepare phenomenological defenses.

December 5, 2011 2:11 PM

A mixed (superpositioned?) state of buzz among those working in quantum foundations over a new paper by Matt Pusey asserting that quantum states are real physical objects and not simply statistical probability distributions. Matt Leifer does a balanced contextualization and explication. A giddy article in nature news and David Wallace support and summarize.

This type of reasoning, seeking to develop experimental tests to rule out certain interpretations (and accompanying metaphysical baggage) of quantum theory, follows in the footsteps of the famous Bell's Theorem, which rules out certain kinds of hidden-variable theories. This new paper by Pusey, et. al. asserts that what is commonly referred to as the Copenhagen interpretation (i.e. quantum mechanics is merely a way of describing the things in our head that we know about the system) is untenable.

Here are some more rapidly developing encapsulations and generalizations to the original Pusey, et. al. result. Grain of salt reminder: terms like "physically real", "physical property", and "statistical property" need to be read carefully within how they are defined (or not well defined) within each paper. On-going discussion at shetl-optimized.

posted by wjzeng (42 comments total)

Everyone should read the shetl-optimized link before commenting. It cleared up a lot of my confusion about what, on the face of it, seemed to be a pretty classic philosophy of science debate (this isn't).
posted by phrontist at 2:24 PM on December 5, 2011

...their finding is similar to the notion that an individual coin being flipped in a biased way — for example, so that it comes up 'heads' six out of ten times — has the intrinsic, physical property of being biased, in contrast to the idea that the bias is simply a statistical property of many coin-flip outcomes.

So Einstein was sorta right: God plays the universe with loaded dice.
posted by three blind mice at 2:26 PM on December 5, 2011

Could some clarify how this differs from Hugh Everett's Interpretation (aka Many Worlds)?
posted by justkevin at 2:37 PM on December 5, 2011

justkevin: Everett's proposition wasn't an empirical one. This is, so it wouldn't have any bearing on the truth of Everett's idea.
posted by phrontist at 2:40 PM on December 5, 2011

Here's a more layman-readable writeup about the paper on Ars Technica.
posted by danny the boy at 2:42 PM on December 5, 2011 [1 favorite]

I really need to re-read my P-Chem/QPhysics texts. I was barely hanging on through most of that.
posted by Slackermagee at 2:46 PM on December 5, 2011

PBR me, ASAP.
posted by AElfwine Evenstar at 2:49 PM on December 5, 2011 [1 favorite]

justkevin: Could some clarify how this differs from Hugh Everett's Interpretation (aka Many Worlds)?

From Matt Leifer's blogpost linked in this MeFi post:

Pretty much all of the well-developed interpretations that take a realist stance fall under option 3, so they are in the psi-ontic camp. This includes the Everett/many-worlds interpretation, de Broglie-Bohm theory, and spontaneous collapse models. Advocates of these approaches are likely to rejoice at the PBR result, as it apparently rules out their only realist competition, and they are unlikely to regard anti-realist approaches as viable.

And when asked about Many-Worlds in comments, Leifer replies:

The theorem has no consequences for many-worlds because, firstly, the wavefunction is already ontic in many worlds and, secondly, the theorem is based on a framework in which measurements only have single outcomes, so it does not apply to many-worlds.

To understand the context, it's better to read Leifer's post.
posted by Kattullus at 2:51 PM on December 5, 2011 [2 favorites]

Motl doesn't like it.
posted by empath at 2:53 PM on December 5, 2011

empath: Motl doesn't like it.

You know even though many other around did link to it, I really did want to leave out linking to that rant...
posted by wjzeng at 2:58 PM on December 5, 2011

Motl doesn't like it.

I'm not physicist, but it certainly seems like that post leads with a lot of name-calling.
posted by verb at 3:09 PM on December 5, 2011 [1 favorite]

They're staring at Nagarjuna's void, and mistaking it for a real thing. Both sides of the debate.
posted by stonepharisee at 3:09 PM on December 5, 2011

Keep those contest entries coming! mefi is fun right now
posted by telstar at 3:15 PM on December 5, 2011

Hasn't Copenhagen in serious trouble for years (decades?) now?

They're staring at Nagarjuna's void, and mistaking it for a real thing. Both sides of the debate.

I'm pretty sure they're proposing experiments and developing a theoretical construct that will allow for the fruitful interpretation of the results of these experiments. I don't think this has anything to do with Buddhist philosophy, as nice as Buddhist philosophy might be.
posted by mr_roboto at 3:16 PM on December 5, 2011

So are they claiming that all the possible physical states of a system are all physically real? I'm confused. Isn't that basically a strong MWI of quantum mechanics?
posted by AElfwine Evenstar at 3:18 PM on December 5, 2011

You know, this wasn't something they covered at Copenhagen Interpretation Fantasy Camp.
posted by zomg at 3:41 PM on December 5, 2011 [1 favorite]

Great first post by the way.
posted by AElfwine Evenstar at 3:53 PM on December 5, 2011

if I measure the spin of one of a pair of entangled particles, then that measurement automatically and instantaneously sets the spin of the other... even if it's on the opposite side of the Universe.

[ - ars technica ]

IANAQP (or a P at all) but this is one of my biggest pet peeves. Repeat after me:

QUANTUM ENTANGLEMENT DOESN'T ALLOW FTL COMMUNICATION

QUANTUM ENTANGLEMENT DOESN'T ALLOW FTL COMMUNICATION

QUANTUM ENTANGLEMENT DOESN'T ALLOW FTL COMMUNICATION*

Arrrgle bargle bargle.

* unless FTL communication somehow falls out of this paper, which seems unlikely.
posted by BungaDunga at 3:55 PM on December 5, 2011 [1 favorite]

That sentence didn't say anything about FTL communication, though?
posted by ymgve at 3:59 PM on December 5, 2011

Does the writer of the Ars article imply that entanglement allows FTL communication somewhere? That wasn't my read, but maybe I already know that's not what entanglement is about...
posted by danny the boy at 4:01 PM on December 5, 2011

I read up and down both sides of all of these articles, and I didn't find the lyrics to "Never Gonna Give You Up" anywhere. Is this some sort of a joke?
posted by Xoebe at 4:02 PM on December 5, 2011 [1 favorite]

That sentence didn't say anything about FTL communication, though?

You're right, I don't think it does. But most people will read it that way, I think. Nonlocality sounds like instantaneous classical communication, when it isn't. It's deeply weird anyway, but it's not going to build you an ansible.
posted by BungaDunga at 4:08 PM on December 5, 2011 [1 favorite]

I may have the history wrong, but I don't think that Bell's Theorem was a super-big deal when it was published — it became a big deal when Alain Aspect and collaborators were able to construct the first correlated-photon experiment and demonstrated that photons can't be described by some hidden-variable theories. There's apparently a test proposed here, too, but I'll have to read later to see what it looks like.

It's a seductive mistake to confuse a model of reality for reality itself.
posted by fantabulous timewaster at 4:43 PM on December 5, 2011

Between this and the ongoing debate over the "superluminal" neutrinos, this is turing out the be a great year for physics weirdness/awesomeness.
posted by lekvar at 4:44 PM on December 5, 2011

If I measure the spin of one of a pair of entangled particles, then that measurement automatically and instantaneously sets the spin of the other... even if it's on the opposite side of the Universe.

QUANTUM ENTANGLEMENT DOESN'T ALLOW FTL COMMUNICATION


I think you're slightly misinterpreting it. If you measure it, somehow (via "spooky action at a distance") the spin of the particle light-years away from its entangled twin really does align the same way (as far as we can tell) immediately. However, you don't get to tell the first particle HOW to align. There's no way to send a message when there's no way to force the particle to say what you want it to.
posted by chimaera at 5:02 PM on December 5, 2011

Motl doesn't like it.

He lost me at "an incredibly embarrassing article called: Quantum theorem shakes foundations by Eugenie Samuel Reich (it's a female name even though Samuel doesn't look like one)"

Actually, he lost my interest at "dimwits" (third word!) then he lost my respect with his gender bias.

Carrying on...
posted by m@f at 5:15 PM on December 5, 2011

The worst thing about exciting science news is that reading about feels important and not like procrastination at all. It's been four hours since I've gotten any work done. Damn you, wjzeng!
posted by Kattullus at 6:27 PM on December 5, 2011 [1 favorite]

So like... reality is a thing?
posted by Saxon Kane at 6:56 PM on December 5, 2011

photons can't be described by some hidden-variable theories.

Some being the operative word.

Doesn't matter how you slice it; Spooky action at a distance is sufficiently weird that it allows for a lot of play on the other side of the fence.
posted by effugas at 7:22 PM on December 5, 2011

Motl doesn't like it.

Lubos Motl is probably the single largest reason why I don't think String Theory is a valid branch of physics. Even when he's right, he's wrong. His complete failure to explain anything of what he's doing, his inability to fend off criticism of his own field with rational explanation rather than sneering dismissal, and worst of all, his inability to defend physics from obvious cranks despite enormous effort indicates that he's fantastic at inventing new math, less fantastic at pretty much everything else involving physics.

He's an entertaining writer, but so was Groucho Marx.
posted by Slap*Happy at 7:49 PM on December 5, 2011 [1 favorite]

I think you're slightly misinterpreting it. If you measure it, somehow (via "spooky action at a distance") the spin of the particle light-years away from its entangled twin really does align the same way (as far as we can tell) immediately. However, you don't get to tell the first particle HOW to align. There's no way to send a message when there's no way to force the particle to say what you want it to.

Stupid question: what if the signal isn't the direction of the alignment, but rather the timing of when the alignment happens? Basically: Morse code, but instead of pressing down on the transmitter button, we do (or do not) collapse the waveform at set intervals.

Example:
Alice and Bob each have 1000 sets of 4 entangled photon-pairs, and they each possess one of the 4000 total pairs.

Alice tells Bob to begin checking whether any of the photons in his first set spontaneously align themselves in exactly one hundred minutes, and to check each successive set in 250ms intervals.

Over the next 100 minutes, Alice moves to a position 10 light-minutes from Bob.

Upon arrival, Alice begins measuring the spin of her photons (or not) one set at a time, eventually transmitting a traditional SOS (for every two sets, observe only the first set to transmit a dot, observe both to transmit a dash). Bob receives the signal, rushes outside to his telescope, and ten minutes later watches Alice transmit the signal he received.

There's probably several things wrong with my mental model, here. Help me, Obi Wan.
posted by Ryvar at 8:18 PM on December 5, 2011

Alice tells Bob to begin checking whether any of the photons in his first set spontaneously align themselves

Huh? If I understand you properly, Bob is observing the photons here. He would collapse the wave function, even if Alice didn't. I'm pretty sure we don't have any method, theoretical or otherwise, to see if the other entangled particle has been observed solely based on the entangled particle we have.
posted by sbutler at 8:28 PM on December 5, 2011

Disclaimer: it's been a while since I've studied quantum mechanics, and I was never an expert anyway, so...

> spontaneously align themselves

Do you mean something like "wave functions collapse"? The photons are aligned (or anti-aligned or whatever) already, at the start of the experiment.

Bob receives the signal, rushes outside to his telescope, and ten minutes later watches Alice transmit the signal he received.

Ok, so there are two signals here, one is a light signal (?). What is the other one? There's no special signal telling Bob "Your photon's wave function has collapsed". When Bob looks at his photons, the wave functions have collapsed. Either they collapsed already because Alice measured her photons, or they collapsed because he looked at them. I don't think there's any way to know whether they were collapsed before he looked, because to know such a thing would involve a measurement, which itself would collapse the wave function instantly.
posted by -jf- at 8:57 PM on December 5, 2011

So like... reality is a thing?

It's a possibility.
posted by Slithy_Tove at 8:59 PM on December 5, 2011 [2 favorites]

A distinct one, at that.
posted by joe lisboa at 11:18 PM on December 5, 2011

There's probably several things wrong with my mental model, here. Help me, Obi Wan.

Unfortunately you can't tell if the wavefunction has collapsed, all either end can do is observe their particles. You can't even tell if the other party observed their particles first.
posted by atrazine at 1:28 AM on December 6, 2011

Someone please tell me what is flawed about this analogy. I have a bunch of toothpicks in a box, each with one end painted red and one end painted blue. In total darkness, I pull out a toothpick, snap it in half, and seal each half in an envelope. I give one envelope to Alice and one to Bob. Without opening her envelope, Alice can't tell whether her half is red or blue - but as soon as she does, the color of Bob's half is instantaneously known to be the opposite.

What's so spooky and quantum about entangled particles that makes them different from toothpicks in envelopes, other than the fact that they are more mathematically perfect in this behavior (i.e., you have to open the envelope to look inside, unlike a real envelope which is subject to trickery, and the quantum envelopes are provably randomly assigned to Alice or Bob)?
posted by NMcCoy at 2:25 AM on December 6, 2011

NMcCoy: That's exactly what Bell's Theorem is about.
posted by edd at 2:29 AM on December 6, 2011

So basically what I'm getting from the Bell's Theorem thing is that the toothpicks are not broken in half, but put into the envelope at some mystery angle; and the way you measure the toothpick is you tilt the envelope to some angle, then chop it in half, and if the red part of the toothpick is on your left then it's a 1 and if it's on your right then it's a 0; and so if Alice and Bob both keep their envelopes level and chop, they agree 100%, and if Bob tilts his envelope 90 degrees and then chops it they match 50% (totally uncorrelated), and the reason that this is not toothpicks is that if Bob tilts his envelope 45 degrees, the toothpick sides do not match 75% because they are not actually angled toothpicks.

Right?
posted by NMcCoy at 3:07 AM on December 6, 2011

NMcCoy: using your setup, the way I understand it, we need to define some more properties. So we have that the ends are either red or blue. Let's also say that one end is dipped in peppermint and the other plain. And that one end is make from a dense wood and the other a light wood (these are very weird, inconsistent toothpicks!). Each of these combinations exists:

- red, mint, dense
- red, mint, light
- red, plain, dense
- red, plain, light
- blue, mint, dense
- blue, mint, light
- blue, plain, dense
- blue, plain, light

So there are three experiments we can perform on a toothpick end. We can look at it and see the color (but not feel or smell it), smell it in a pitch black room (but not look or feel it) and see if there's mint, or we can weight it still in the envelope (without seeing or smelling it) and see if it's dense or light.

Now you put a toothpick in an envelope and chop it in half. Send one half to Bob and one half to Alice.

- If Alice measures color and Bob measures smell then Alice gets red X_1 times and Bob gets plain X_2 times. X = X_1 + X_2.

- If Alice measures smell and Bob measures weight then Alice gets mint Y_1 times and Bob gets light Y_2 times. Y = Y_1 + Y_2.

- If Alice measures color and Bob measures weight then Alice gets red Z_1 times and Bob gets light Z_2 times. Z = Z_1 + Z_2.

The inequality is that X + Y >= Z. And if this toothpick behaves in a classical way then the inequality holds. But we can run this experiment, and for quantum toothpicks we know that X + Y is not >= Z! So we know that quantum toothpicks do not behave in a classical way, and the conclusion everyone draws from this is that there is spooky action at a distance.

Hopefully, I correctly restated the argument made at this page.
posted by sbutler at 8:30 AM on December 6, 2011

NMcCoy, that's pretty much right. The meaty part of Bell's theorem isn't what happens when the measurements are along the same axis (perfect correlation of remote measurements, apart from experimental noise) or along orthogonal axes (absolutely no correlation between remote measurements), but what happens when the measurement axes are somewhere in between, and partial correlation is expected.

In quantum mechanics, where the usual interpretation is that the polarization isn't defined before it gets measured (subject to the "spooky" constraint that all the measurements must be consistent afterwards), there's a certain amount of correlation predicted as a function of the angle between the axes.

If the particles are born with a certain intrinsic polarization, which they carry with them "locally" to the detector, the correlation as a function of measurement angle is different. For the "perfect correlation" and "absolutely no correlation" cases, it's the same, but for the mixed-axis measurements, the correlation is smaller than the quantum-mechanical case.

What Bell proved was that in any "local intrinsic polarization" model, not just his simple example, the correlations between measurements in the mixed case are the same as or weaker than the correlations in quantum mechanics. In the experiments, one finds correlations which are consistent with the QM prediction and inconsistent with whatever local-variable model is under test.

sbutler: your analogy is missing the "mixed" measurements: some method that gives you partial information about color and partial information about flavor, but not total information about either.
posted by fantabulous timewaster at 9:13 AM on December 6, 2011

From Shtetl-Optimized: [...] one further assumption: that “rational beliefs behave well under tensor products.”

This assumption produces a nice and Champagne-like feeling in my stomach.
posted by springload at 2:25 AM on December 7, 2011