Bell inequality violation finally done right | Hacker News Reader

Bell inequality violation finally done right

scottaaronson.com

109 pointsrvern46 comments

Worth reading because it has an interesting, new (to me, at least) and approachable explanation of local realism.

Money quote:

    Perhaps the best way to explain local 
    realism is that it’s the thing you believe in, 
    if you believe all the physicists babbling 
    about “quantum entanglement” just 
    missed something completely obvious.

Forget the realism part. Can we at least get the locality part?

In my mind, all the quantum observables are not real properties, therefor it's obvious/trivial that they don't exist until measured. That is, I have no problem accepting they are not "real".

What gets to me is the locality aspect of it.

Of course I'm not a physicist (read: I am layman).

Not a physicist either, but I think "local realism" is a conjunction of "local" and "realism". Bell inequality experiments show that "local realism" is false, and so we have to reject either "local" or "realism", but we don't necessarily have to reject both.

Best i can tell local realism basically means that if something blows up at the other side of the universe, it should not affect whatever is happening on this side until we can observe said event via the reception of energy/particles from said event.

Well that's the problem with local realism. All sorts of things do affect local conditions at a distance. If entanglement between particle pairs is not affected by any of the physical forces as described then whatever happens between them is just a matter of fact, but no where have I seen this mean we'll ever get ansibles or teleportation. It just means the universe as a substructure we're not fully able to grasp and that it may be just one of those things that connects everything despite the appearance of separation of objects and states.

that quote,and pretty much whole blog post, is among the best illustrations of orthodoxy. Speaking about obvious - the paper contains S>2 for entangled case. For some, probably obvious reason, the paper doesn't contain S for non-entangled case. We know that is should be <=2. Why not to show it? Their setup is exactly one very suitable to run without entanglement. From there i stand it actually would be >2 in their setup even for non-entangled case.

What problem do you have with the orthodoxy?

Einstein et. al.'s original paper on the EPR paradox is very clear and understandable, I recommend it: http://journals.aps.org/pr/pdf/10.1103/PhysRev.47.777

Locality (as opposed to non-locality) is a strange concept if you think about it. Because how would different points in space "know" to run the same set of equations? Hence all points have something in common, giving rise to the possibility (or non-weirdness) of non-locality. But of course, this is completely non-scientific :)

> Locality (as opposed to non-locality) is a strange concept if you think about it. Because how would different points in space "know" to run the same set of equations?

I'm sorry but I don't understand your point, can you explain it further?

Also one weirdness of QM is not only that it is non-local, it is that it is non-local allowing instantaneous remote 'interaction' BUT without allowing to send information instantaneously (not faster than light).

Well, there isn't much to explain really. Physicists have long had an "intuition" that says that non-locality is somehow a strange concept. However, intuition is often misleading. So I'm flipping the "argument" around and I say that it is more intuitive to think of locality as a strange concept, and to think of non-locality as something which is more natural.

Physicist's "intuition" (as you put it) is educated by codifying our observations into mathematical equations, using those equations to make predictions, and performing experiments to verify those predictions.

Simply 'flipping around the "argument"' is playing with words and has no usefulness from a scientific point of view; it is definitely not "more intuitive" (as you put it) since it has nothing to do with science.

It's an interesting thought exercise IMO. Keep in mind that QM is widely regarded as "counterintuitive" even by scientists, and as far as we can tell, it's the intuitions that are wrong. I.e. it's not QM that's crazy. QM is normal. We are crazy.

Actually it's the other way around. Because of translation invariance we have the law of conservation of momentum. Because of temporal invariance we have the law of conservation of energy. Every symmetry gives rise to a physical conservation law, from which forces can be derived.

See Noether's theorem.

* every differentiable symmetry of an action

(Sorry to be nitpicky, but the asterisk is very important haha)

> Because how would different points in space "know" to run the same set of equations?

This is not locality. This is translational invariance.

However, equations have parameters.

The principle of locality is that these parameters come from nearby regions of space. Basically, things are only affected by what's near them.

It's violated by quantum mechanics.

Thanks for this article, the previous description of the experiment I read omitted the 'discard the result unless a measurement show that the entanglement was successful' step, leaving me very confused.. I didn't know that this was possible, that said I still don't have any clue about HOW this is possible.

Money quote: "if people had just understood and believed Bohr and Heisenberg back in 1925, there would’ve been no need for this whole tiresome discussion"

Prof. Aaronson is teaching a grad seminar this Fall at MIT on Computation and Physics. Here's hoping he makes the lecture notes public ;)

This is a really good explanation... but I can personally only square most of this entanglement non-locality with a many-worlds interpretation. Which is itself hugely problematic.

What is problematic about the many-worlds interpretation. Even without non-locality, it solves the non-determinism problem, which seems far more problematic to me.

Well, the only one of the many-worlds that actually matters is this particular one, right here. What can we say/predict about this exact one. The others don't matter to me in the slightest...

EDIT: well that's not a great explanation of why I dislike the many-worlds hypothesis, let me try again. Like some other ideas, including the existence of God, it's not (yet?) falsifiable. Which doesn't make it false, of course. But it's a really extraordinary idea. It's not just many worlds, it's a fuck-ton. And what can we do with all of them? Are they for nothing? Does nothing that happens matter because it happened differently in a separate but equally real "world"?

I guess, I can believe that "god plays dice with the universe", but I can not believe that "god makes infinite copies of the whole universe and tries all possible quantum outcomes".

The interpretation we use does not have any effect on our ability to make predictions; the equations are the same.

EDIT: response to parent edit

Almost by definition interpretations are not falsifiable. The many world interpenetration is literally equivalent to the other interpretations, because they all describe how we understand the meaning of the same mathematical model; and it is that model that is making predictions.

In my opinion (as a mathematician, not a physicist), the many worlds interpretation is the more natural way to read quantum theory. That is to say, if I were tasked to design a universe given quantum physics as the specification, my naive implementation would correspond to the many worlds interpretation.

An equally valid interpretation is that we are a computer simulation, and quantum physics describes the rounding behavior of said simulation. As far as I am concerned the only point of having an interpretation is to help build an intuition for the models.

Why not?

But it's a really extraordinary idea. It's not just many worlds, it's a fuck-ton.

Another way of looking at it is that there's only one "world": the amplitude distribution over quantum configuration space. https://www.quora.com/How-does-observation-collapse-probabil... has a decent explanation/argument.

> it's a fuck-ton

Or zero. (But it's definitely not one.) Personally I find the zero-worlds view to be the most intuitively satisfying:

http://www.flownet.com/ron/QM.pdf (Or the movie version: https://www.youtube.com/watch?v=dEaecUuEqfc)

David Mermin's Ithaca interpretation is essentially the same thing:

http://arxiv.org/abs/quant-ph/9609013

The slogan is that "we are correlations without correlata".

Do you believe god exists and has a psychology? I'm trying to understand if you have trouble believing because you don't think god would create such a universe, or if you have trouble believing because such a universe is innately unbelievable.

If the latter, then I would ask, "Do you believe the ocean is deep, even though you can only see the surface?"

The latter.

I don't really understand the ocean metaphor, since humans have seen below the surface, quite a distance with remote-controlled submersibles.

Perhaps a better analogy would be, do you believe that matter exists beyond the observable universe?

Because of the expansion of the universe, it is theoretically possible for me to get in a space ship and travel away from you at such a high speed that we would never be able to meet again, even if at some point we both decided to head back toward each other as fast as possible, and even if we were both immortal. Our light cones would no longer intersect.

Probably you would assume that I continue to exist, despite that for all intents and purposes I cannot affect your life in anyway and therefore "don't" exist.

Similarly if I crossed the event horizon of a black hole.

I believe that a qubit is in superposition (i.e. exists in multiple miniature "universes"), because that theory produces results that are consistent with experiment. When the qubit's universes de-cohere with mine, such that for all intents and purposes I can now only ever observe the qubit in one particular state, why should I assume that the other universes have ceased to exist? Decoherence is a continuous process, just like traveling outside of a light cone, so why treat them any differently. It's rather arbitrary to say that increased distance in space preserves existence, while increased distance in phase space does not.

A usual reply is that physicists don't or shouldn't care about metrics of the size of the universe but rather the simplicity and symmetry of physics. After all, the universe might well be infinite without many worlds, and why do even we care about amounts of stuff? The library of Babel doesn't contain more information than any single book in it. Similarly, pointing to specific areas in a large universe may require adding details.

> "god makes infinite copies of the whole universe and tries all possible quantum outcomes".

MWI says that probability is conserved across the multiverse (unitarity of the global wavefunction)! The only reason we can "create universes" is because we're in the low-entropy time near the Big Bang. As we start getting closer to the heat death of the multiverse, then less "universes" are being created, because there's no room for them. In fact, given enough time, eventually some of the "universes" will actually join back together! If the multiverse is finite and closed (we don't know this), then sooner or later (like e^2^120 years later, according to a back-of-envelope calculation I saw) all of the universes will be lucky enough to join back up (Poincare recurrence), and we'll end up where we started.

I like looking at MWI not as the creation of universes, but as the creation of perspectives on one universe. In any case, it obeys conservation laws.

Well, for starters it requires an infinite number of universes, each with zero 'weight.' Secondly, it causes problems for materialist views of the universe. (I wrote a long term paper about this as an undergrad, but that was, like, 15 years ago and I don't have the cuneiform tablets anymore.)

Briefly, there's no way to explain the statistical correlations between spin polarization states: if we measure the polarization of a photon in a basis that's only, say, 5 degrees offset from its known polarization, the statistical breakdown between the two resulting polarization states is something like 9:1; what if it's only a few arc-seconds off? Then the universe must divide more and more finely.

You quickly realize the universe must be infinitely-divisible (because quantum collapse occurs so often, and each collapse anywhere in the universe must divide the universe, it can't even be discretized at, say, some Planck-like scale).

That said, it does neatly solve local realism. But you have to believe in an immortal soul or something non-material -- I believe that has to be the 'weighting factor'. (Again, that was the uncomfortable conclusion of my paper, wish I could conjure more of it.)

I think that materialism is a dead concept anyways. It just doesn't seem to support itself under the weight of evidence in physics which seems to contradict it. So, either materialism goes or physicists are going to have a fun time with new experiments overturning old ones.

Wow, that's a strong statement. (Are we using the same definition of materialism?)

I use it to mean that there are no outside forces describing the state of the universe beyond what's materially measurable/describable; in other words, probably we have no transmaterial soul.

That's a strong assumption for anyone who's religious, but it's sort of gospel for atheists. That said, it's the many-worlds hypothesis doesn't square with a totally-naive interpretation of Bell's Inequality and materialism/atheism.

So... are you saying that we have to admit there's a God/transmaterial soul / some other magical thing because we don't understand the implication of Bell's inequality? If so, that's a big jump, but I can't say it's wrong. Just that I don't take that leap along with you.

You know, there are more things in heaven and earth, Horatio, than etc etc. For both of us.

>Wow, that's a strong statement. (Are we using the same definition of materialism?)

I'm probably not using the same definition. I'm just of the view that what we think of as material often gets reduced to atoms or particles representing the four forces for which no other 'outside' component can be entertained to exist. To me, this is a over simplification that erases all the other seemingly non-material components of our material world (thoughts, feelings, experiences, math itself, and etc). It's easy to assume it to be true, but every time I bump into this argument I always point out that the number 1 isn't of any singular substance but the number itself exists everywhere there is any form of divisibility or unitary attribute in use.

I have no idea if this makes sense, as it is mostly random typing from a monkey (who has been drinking beer, no less)... but couldn't it be perfectly reasonable if non-locality was the result of something that is non-observable? I mean, if I flip a coin and heads comes up, it's not surprising that tails is down. The head is connected to the tail, so I can't have both coming up at the same time. Imagine that there are inhabitants of a torus world with the coin at one point. It's spinning there. Alice stays to watch the spinning coin and Bob wanders along the torus in the opposite direction. Eventually he sees a spinning coin too. It can't be the same coin, because he's walked a long way in the opposite direction. So he watches the coin. Eventually it stops spinning. Alice sees heads and Bob sees tails. No matter how often they do this, it works out opposite. It's not that one side of the coin is communicating to the other side of the coin at the other "end" of the torus. It's just that there is no other way for it to work out.

I'm not suggesting that it actually works this way, I'm just giveng my drunken monkey interpretation of how particles could exhibit this behaviour without many worlds ;-)

I'm not sure what you're trying to say, but it appears that you're conjecturing that "everything is local". That doesn't square too well with relativity, though.

Either that or you're talking of a hidden variable theory, which is exactly what Bell's theorem disallows.

Beer is a good tool to try to understand quantum mechanics, but there's stronger stuff that might be more helpful.

That said, you're articulating a version of what's called a 'hidden variable' view of QM, which is actually explicitly what Bell's inequality is meant to disprove. I'm going to write a long post at root to try to explain this more, because I think it's really important to talk in particulars when discussing QM... hopefully it helps.

I'd really appreciate that. I've been looking at Wikipedia articles and I really can't quite understand what "local hidden variable" means in relation to "non-local hidden variable".

I wasn't really suggesting that the system was deterministic. The coin is neither heads nor tails until someone looks at it. I was suggesting that it is the same coin in 2 different places (from the vantage point of the people observing it). However, you can't know a priori that it is the same coin. You have to observe it in both places and communicate your discovery. Since you are trapped in your torus, it will take time communicate it.

But it's not like the coin is sending messages to itself to say, "if someone looks at the other side, make sure it is tails". It just is. Similarly, it is not in a determined state at the point of entaglement. It is only determined when one party looks at it. But then that decides what the other party will see.

I honestly don't see how you would create a test for this because you would not be able to determine if 2 particles were entanlged (the same coin in my silly example) without observing them.

Anyway, I appreciate your patience. I am keenly aware that "crazy theories from people who have no idea what they are talking about" are not particularly helpful. But if you write a blog post on the topic, that's the kind of area I'd like to see explained better ;-)

So, in a nutshell, the phrase 'It is only determined when one party looks at it' is the definition of non-locality.

If the two parties are separated by a significant distance, that implies that one viewer (Alice?) determines what the other viewer (Bob) sees the moment she looks at the coin. But of course, our notion of who 'looks' first is a function of relativity; there are reference frames where Bob looks first.

The fact that they always correspond suggest that either A) the result was predetermined (hidden variable, i.e. local reality) or B) the two particles 'communicate' faster than light. (Note that, because the 'communication' can't actually convey information, it doesn't technically violate relativity, just, you know, that last shred of common sense you hoped applied to physics.

I'd write more -- didn't get to that meta-post, because I recently turned on the 'no procrastinate' flag on HN... and though it's made me way more productive I'm not sure if it's made me happier. C'est la whatever.

I don't have a problem with locality being violated. According to relativity, distance is relative and depends on the reference frame. There will always been some reference frame where their is no distance between points.

Noooooooooooooooo. No.

"distance is relative and depends on the reference frame." does not imply " There will always been some reference frame where their is no distance between points.". It just means that distances may be different.

Special relativity is a linear transformation, so a nonzero distance cannot become zero.

Relativity doesn't help locality violation, far from it. Relativity has issues (EPR) with locality violation.

What's the distance between two points when measured from a third reference frame moving at the speed C?

I don't think you can have a reference point moving at speed C, the universe throws a 'divide by zero' exception.

While true, I thought in QM you had to evaluate all probabilities, even those where the speed is > C.

Nope. In QM you have probabilities where energy conservation is violated for a little bit of time, or where a particle appeared to jump in a distance longer than allowed. This doesn't mean you can have reference frames moving at the speed of light.

You can't use a reference frame moving at c -- it'd be like trying to use a coordinate system where the X and Y axes are parallel to each other.

Also, distance isn't really what locality is about; it's really about the causal sequence of events. If two events are close enough in time and far enough in space that light couldn't get from one to the other (in jargon: they're separated by a spacelike interval), then there'll be some reference frames where one happened before the other, and some where they happened in the opposite order. So if one of the events caused the other, there are reference frames where the cause came after the event.

So nonlocality violations imply that effects can come before causes in at least some reference frames. That's why people don't like it.

Reply on news.ycombinator.com