It isn't just the title, the whole article is poorly written. It gives the impression that faster-than-light communication is possible.
From the first sentence:
"[Scientists] conducted an experiment they say proves one of the most fundamental claims of quantum theory — that objects separated by great distance can instantaneously affect each other’s behavior."
Emphasis mine. This is totally wrong.
The second sentence:
"The finding is another blow to one of the bedrock principles of standard physics known as “locality,” which states that an object is directly influenced only by its immediate surroundings."
Again, this is wrong. Quantum physics is entirely consistent with locality. It contradicts "local hidden variable models." Despite the similar name, this is different from locality. (Local hidden variable models are local, classical models.)
To add to that, Bell's theorem says that either locality or hidden variables may be true, but not both. We could construct a formulation of quantum mechanics that includes hidden variables, but must account for the state of the universe at arbitrarily distant locations.
I posted this b/c I was hoping someone would dive into these quotes:
The Delft researchers were able to entangle two electrons separated by a distance of 1.3 kilometers, slightly less than a mile, and then share information between them.
...
Researchers like Dr. Hanson envision a quantum communications network formed from a chain of entangled particles girdling the entire globe. Such a network would make it possible to securely share encryption keys, and know of eavesdropping attempts with absolute certainty.
This implies that entanglement allows faster-than-light communication (transmitting arbitrary information across the entangled link). I thought that was a no-no, and that entanglement still didn't allow FTL communication despite FTL-like properties.
While everything you said is true, I feel its worth pointing out entanglement is stronger than simply being able to generate the same number in two locations, a feat easily accomplished via classical means.
Edit:
Entanglement is akin to the following... Assume Alice and Bob are distantly separated. Alice has an arbitrary float variable called x. Alice and Bob both have a function rand() generating floats between zero and one, with a shared seed so that they can both generate the same random numbers.
Classical correlation is that they both can call rand() with their identical seeds to get the same random numbers.
Entanglement is that Bob's rand() is actually ((x + rand()) % 1), where Bob instantly and magically has access to Alice's variable x. No information can be transmitted, because the distribution of random numbers hasn't changed, and Bob doesn't have direct access to x. But something is different, and if Bob is clever, he might be able to do some fun stuff.
This isn't a perfect analogy, but perhaps it helps.
> While everything you said is true, I feel its worth pointing out entanglement is stronger than simply being able to generate the same number in two locations, a feat easily accomplished via classical means.
Really? That's not random, that's pseudo-random, so the secret is still just the seed distributed before hand. True randomness would be entirely uninvolved from other state, including location or time. This is really just combining the sharing and the generating into one process so you CAN share true randomness without sharing other state.
It's the nyt, so they are at best imprecise and at worst wrong. Entanglement can be thought of as a resource that allows you to do unexpected things (win the CHSH game), but sharing information isn't one of them.
Good old Aaronson, thanks. My recollection of "Quantum Computing since Democritus" informed my frowny-intuition about the NYT article ... should have gone to his blog directly no doubt.
The linked article from Scott Aaronson is quite accessible and very worth reading. My favorite quote is:
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.
What I still struggle to understand is the realism part. The NYT article seems to imply that, when all loopholes are closed, we will have the final prove to live in a non-local world. Realism isn't mentioned.
The finding is another blow to one of the bedrock principles of
standard physics known as “locality,” which states that an object is
directly influenced only by its immediate surroundings.
I always thought that local realism meant one of the following:
- no realism
- no locality
- neither realism nor locality
It doesn't say anything about FTL communications. I know that's a subject that has been associated with entanglement many times, but I'm not sure this is one of those times.
In other words, let's not blame the Times for something they did not actually say.
Instead what I inferred from this is that since entanglement can only be resolved once per pair of particles, a communications system based on entanglement might provide perfect forward secrecy even in an environment in which very large numbers are able to be factored by quantum computers.
(copied from my response to a similar comment above):
But since the 1970s, a series of precise experiments by physicists are increasingly erasing doubt — alternative explanations that are referred to as loopholes — that two previously entangled particles, even if separated by the width of the universe, could instantly communicate.
The problem is with the usage of the word "communicate". Normally communication would mean "transmit information". Quantum entanglement does not allow this. With entangled particles a measurement on one instantly changes the statistical behavior of the other. However there's no way to see this effect until you compare the statistics between the two particles (which probably requires an ensemble of particle pairs to be measured, not just one), so you can't actually use the effect to transmit information.
I don't think there's a good English word for the concept of an influence that isn't causal and doesn't transmit information. That's why it's often called "spooky action at a distance" and why attempts to describe it with simple English sentences are likely to lead to confusion.
Yea, communicate is the wrong word. Perhaps "could continue to share state" would be better, but perhaps that would be confusing to the non-quantum-inclined audience.
That is a different quote, which means a different thing.
The quotes you posted above refer to human communications, and don't mention FTL. That's what I responded to.
The quote you posted now, is clearly referring to the fact that the particles both react to one measurement instantly, regardless of difference. It makes no mention of people communicating with one another FTL.
Edit: I agree with tgflynn that the Times just overloaded the word "communicate."
This does NOT imply FTL communication. The actual key is arbitrary/random, so no information is transferred. However, both can read the same particle state. So it's more like they sent the particles to each other (or from the center), "caching" whatever value is there in two separate locations. However, you don't know when the entanglement is broken, so you STILL can't communicate FTL.
Can you give a link that clearly describes "measurement-as-a-cascade-of-entanglements process" in some reliable source? Best, some QM textbook.
Basically exactly the same as this http://www.flownet.com/ron/QM.pdf (5.3), but in a well-known textbook or some review paper in a journal with good reputation. I run from time-to-time into a discussion which just needs such pointer...
And there is some evidence that even Bohr and Heisenberg actually understood this.
Basically, the physics community bifurcates among those who deny this is true (e.g. Lubos Motl) and those among whom it is considered common knowledge. Indeed, when I submitted my paper to Physics Today back in 2000 it was rejected on the grounds that it was nothing new.
The hard part of QM is not understanding it. The hard part is accepting that it is true. Because people really want to believe that the reason that we can make reproducible measurements is that these measurements are a faithful reflection of a underlying metaphysical reality that is actually "out there" in some sense. But that turns out not to be the case. As David Mermin so eloquently puts it (http://arxiv.org/abs/quant-ph/9609013) we are in fact made of "correlations without correlata."
Yeah, it feels like it is just common knowledge (usually given as example during the lecture). But it's difficult to find this one in a form of nice reliable formal publication.
I wish some student somewhere would publish a review paper in some reputable place, clearly describing this cascade of entanglements from start to the end using an example of photomultiplier. Cascade starting from a photon and ending with a 'classical bit' sitting in the memory cell...
Why do you think that these quotes imply FTL? They talk about a network allowing communication over a large distance, but they don't specify a communication speed. Why would you assume that means that it would be instantaneous/faster than light?
A more parsimonious reading of the article would assume that the author knows, and assumes you know, that FTL communication is impossible, and that the interesting aspect of the communication possible through an entangled quantum network is, as stated, its security and resilience to eavesdropping.
The interaction between the particles is instantaneous. There is no mention of a non-FTL communication betwixt the particles. And then there's this sentence in the same article:
But since the 1970s, a series of precise experiments by physicists are increasingly erasing doubt — alternative explanations that are referred to as loopholes — that two previously entangled particles, even if separated by the width of the universe, could instantly communicate.
So I don't think my reading was insufficiently parsimonious.
this may not elucidate a clear answer, but I've heard it described as such and thought it might help:
entanglement works sort of like creating a matching pair of gloves. You can put each glove in separate boxes and ship one to New York, one to San Francisco. You open the box in New York, finding the right-hand glove and know instantly that the left-handed mate is in San Francisco.
So too bits can be entangled, and measuring one, gives you a "cheater's peek" at what the other's value is -- they're integrally linked.
This seems immensely useful regarding encryption, because in my mind, this lets you mimic the public-, private-key model on the small scale.
Can someone convince me that there aren't hidden variables? Because my puny monkey brain hears "a variable can be one of two states, but in the beginning, it's in a super-position of them. When measured, the state is determined." and it's a relative easy thing to get. But....
"There are two particles entangled together such that when one becomes up or down, the other becomes the opposite, and measuring either will determine both, instantaneously, regardless of distance" is a much harder thing to swallow.
Occam's Razor says that there are hidden variables that are pre-selecting state, and we just don't know about them until we measure. ESPECIALLY when one of the requirements for the system is that we can only measure, not influence the outcome of the measurement.
So please, someone who knows more than I do - why can't there be hidden variables being set at entanglement that pre-determine the measurement outcome?
Bell's theorem is the keyword you should look up[1]. If you are more a CS type than a physicist understanding the CHSH game might help.
"Lecture 20: Bell inequalities and nonlocality" from John Watrous of the University of Calgary [2] is a short and accessible introduction to the CHSH game. fizx mentioned Scott Aaronson's "Bell inequality violation finally done right" [3] which is about the same recent experiment as the NYT article but much better. It contains a concise description of the CHSH game.
So the real problem isn't that "when one is up, then the other will magically be up too." That could be accomplished with local hidden variables (e.g. shared seeds on a PRNG).
The real problem is that when you measure A in the "up" direction, and then B in the "10 degrees east of up" direction, then B seems to know that you measured A in the "up" direction.
That is to say: B's probability distribution as a function of the direction its being measured is dependent on the direction that A is measured. There's no way to construct an "A-independent" probability distribution of B's results for arbitrary directions. The probabilities won't sum to 1 and still match experimental results.
It's unfortunate that "A up" therefore "B up" is a degenerate case of this reality where classicality actually works, because it leads to confusion.
Also, follow the links others have provided for a more formal explanation :)
>The probabilities won't sum to 1 and still match experimental results.
exactly. The probabilities are 1 only in theory. For example theoretical Malus law over all angles would give 1. On practice - i couldn't find single photon Malus law confirmation. A slight deviation from Malus law like cutting the tail at high angles off and thus decreasing the total to below 1 (ie. when/if photons at high angles being lost disproportionally much more intensively than Malus law states) would bring the experimentally measured S to values higher than 2 while still being in the realm of local realism.
But then what about a partially shared seed (sharing only a few digits) ? Couldn't you imagine a hidden variable only influencing the probablity distribution, making it only more likely to see an up on both, although still keeping it partially random ?
But doesn't this then violate "no information faster than speed of light"? I.e. You measure A at angle a, I measure B (entangled with A) at angle b, and the resulting distribution gives me some information about your angle a. If I repeat this with enough particles, I can be reasonably certain about what a is.
Agree with weinzierl, you want to look into Bell's theorem. He essentially set a limit (Bell's inequality) to what local variables could do to influence the measurements on both entangled particles. Bell's inequality has repeatedly been violated in experiments on quantum mechanics, which tells us that local, hidden variable cannot be the cause for the correlations in measurements of entangled particles. Bell's theorem then simply states that local hidden variables cannot ever reproduce all the experiments of quantum mechanics.
It's not strictly correct to say that Occam's Razor says we should have hidden variables. Occam's Razor is the idea that in deciding between two or more similar theories (which both have to be correct), we should prefer the simpler of the theories. Bell's theorem tells us that we should reject the notion of local hidden variables.
Importantly, this does allow the notion of nonlocal hidden variables. One such theory is the de Broglie-Bohm theory, which claims that the wavefunction is an actual physical entity, referred to as the pilot wave or guiding wave. It would be nonlocal because the wavefunction collapse of one particle to measure its properties would necessarily influence the other particle at superluminal speeds. Another nonlocal hidden variable theory is superdeterminism, which claims that everything that ever happened and ever will happen has already been determined, so the correlations are not a result of some spooky interaction, they were just "predestined" to be correlated. This would again be nonlocal because the universe as a whole would have knowledge of the particle properties a priori.
These experiments provide evidence against local hidden variables. The experiments are silent about theories that involve non-local hidden variables. The act of measuring one particle of an entangled pair instantly affects the other particle, instead of this effect needing to propagate at or below the speed of light; this instant interaction ("spooky action at a distance") is incompatible with local hidden variables. As other folks have mentioned, look up Bell's inequalities for more information.
Also, it should be noted that quantum entanglement does not allow information to travel faster than light.
50 comments
[ 324 ms ] story [ 2213 ms ] threadFrom the first sentence:
"[Scientists] conducted an experiment they say proves one of the most fundamental claims of quantum theory — that objects separated by great distance can instantaneously affect each other’s behavior."
Emphasis mine. This is totally wrong.
The second sentence:
"The finding is another blow to one of the bedrock principles of standard physics known as “locality,” which states that an object is directly influenced only by its immediate surroundings."
Again, this is wrong. Quantum physics is entirely consistent with locality. It contradicts "local hidden variable models." Despite the similar name, this is different from locality. (Local hidden variable models are local, classical models.)
The Delft researchers were able to entangle two electrons separated by a distance of 1.3 kilometers, slightly less than a mile, and then share information between them.
...
Researchers like Dr. Hanson envision a quantum communications network formed from a chain of entangled particles girdling the entire globe. Such a network would make it possible to securely share encryption keys, and know of eavesdropping attempts with absolute certainty.
This implies that entanglement allows faster-than-light communication (transmitting arbitrary information across the entangled link). I thought that was a no-no, and that entanglement still didn't allow FTL communication despite FTL-like properties.
Can somebody help out here?
Edit:
Entanglement is akin to the following... Assume Alice and Bob are distantly separated. Alice has an arbitrary float variable called x. Alice and Bob both have a function rand() generating floats between zero and one, with a shared seed so that they can both generate the same random numbers.
Classical correlation is that they both can call rand() with their identical seeds to get the same random numbers.
Entanglement is that Bob's rand() is actually ((x + rand()) % 1), where Bob instantly and magically has access to Alice's variable x. No information can be transmitted, because the distribution of random numbers hasn't changed, and Bob doesn't have direct access to x. But something is different, and if Bob is clever, he might be able to do some fun stuff.
This isn't a perfect analogy, but perhaps it helps.
Really? That's not random, that's pseudo-random, so the secret is still just the seed distributed before hand. True randomness would be entirely uninvolved from other state, including location or time. This is really just combining the sharing and the generating into one process so you CAN share true randomness without sharing other state.
I believe quantum key sharing requires a classical channel in addition to the quantum channel, hence no information is transmitted faster than light.
A better discussion can be had at http://www.scottaaronson.com/blog/?cat=33
http://plato.stanford.edu/entries/qm-relational/
I'm not super knowledgeable about this (and whether it even meets the criteria), but perhaps you will find it interesting.
There is also a third option that almost no-one mentions, and that's superdeterminism; i.e. that free will doesn't exist.
In other words, let's not blame the Times for something they did not actually say.
Instead what I inferred from this is that since entanglement can only be resolved once per pair of particles, a communications system based on entanglement might provide perfect forward secrecy even in an environment in which very large numbers are able to be factored by quantum computers.
(copied from my response to a similar comment above):
But since the 1970s, a series of precise experiments by physicists are increasingly erasing doubt — alternative explanations that are referred to as loopholes — that two previously entangled particles, even if separated by the width of the universe, could instantly communicate.
I don't think there's a good English word for the concept of an influence that isn't causal and doesn't transmit information. That's why it's often called "spooky action at a distance" and why attempts to describe it with simple English sentences are likely to lead to confusion.
The quotes you posted above refer to human communications, and don't mention FTL. That's what I responded to.
The quote you posted now, is clearly referring to the fact that the particles both react to one measurement instantly, regardless of difference. It makes no mention of people communicating with one another FTL.
Edit: I agree with tgflynn that the Times just overloaded the word "communicate."
http://www.flownet.com/ron/QM.pdf
https://www.youtube.com/watch?v=dEaecUuEqfc
Basically exactly the same as this http://www.flownet.com/ron/QM.pdf (5.3), but in a well-known textbook or some review paper in a journal with good reputation. I run from time-to-time into a discussion which just needs such pointer...
http://arxiv.org/pdf/quant-ph/9605002v2.pdf
I don't know if that exact paper was ever published, but an earlier version was formally published here:
C. H. Adami and N. J. Cerf, “Information Theory of Quantum Entanglement and Measurement,” Physica D 120 (1998) 62-81.
But this paper turns out to be a rehash of an older result (Zurek, 1991)
http://arxiv.org/abs/quant-ph/0306072
which in turn is more or less a rehash of an even older result (1955) by von Neuman:
http://arxiv.org/pdf/1311.7649v1.pdf (See the references for the original source.)
And there is some evidence that even Bohr and Heisenberg actually understood this.
Basically, the physics community bifurcates among those who deny this is true (e.g. Lubos Motl) and those among whom it is considered common knowledge. Indeed, when I submitted my paper to Physics Today back in 2000 it was rejected on the grounds that it was nothing new.
The hard part of QM is not understanding it. The hard part is accepting that it is true. Because people really want to believe that the reason that we can make reproducible measurements is that these measurements are a faithful reflection of a underlying metaphysical reality that is actually "out there" in some sense. But that turns out not to be the case. As David Mermin so eloquently puts it (http://arxiv.org/abs/quant-ph/9609013) we are in fact made of "correlations without correlata."
I wish some student somewhere would publish a review paper in some reputable place, clearly describing this cascade of entanglements from start to the end using an example of photomultiplier. Cascade starting from a photon and ending with a 'classical bit' sitting in the memory cell...
https://en.wikipedia.org/wiki/Measurement_in_quantum_mechani...
and
https://en.wikipedia.org/wiki/Quantum_decoherence
?
If you don't like Wikipedia, the references section are chock-full of "real" papers, including this one:
http://arxiv.org/ftp/quant-ph/papers/0306/0306072.pdf
A more parsimonious reading of the article would assume that the author knows, and assumes you know, that FTL communication is impossible, and that the interesting aspect of the communication possible through an entangled quantum network is, as stated, its security and resilience to eavesdropping.
But since the 1970s, a series of precise experiments by physicists are increasingly erasing doubt — alternative explanations that are referred to as loopholes — that two previously entangled particles, even if separated by the width of the universe, could instantly communicate.
So I don't think my reading was insufficiently parsimonious.
entanglement works sort of like creating a matching pair of gloves. You can put each glove in separate boxes and ship one to New York, one to San Francisco. You open the box in New York, finding the right-hand glove and know instantly that the left-handed mate is in San Francisco.
So too bits can be entangled, and measuring one, gives you a "cheater's peek" at what the other's value is -- they're integrally linked.
This seems immensely useful regarding encryption, because in my mind, this lets you mimic the public-, private-key model on the small scale.
"There are two particles entangled together such that when one becomes up or down, the other becomes the opposite, and measuring either will determine both, instantaneously, regardless of distance" is a much harder thing to swallow.
Occam's Razor says that there are hidden variables that are pre-selecting state, and we just don't know about them until we measure. ESPECIALLY when one of the requirements for the system is that we can only measure, not influence the outcome of the measurement.
So please, someone who knows more than I do - why can't there be hidden variables being set at entanglement that pre-determine the measurement outcome?
"Lecture 20: Bell inequalities and nonlocality" from John Watrous of the University of Calgary [2] is a short and accessible introduction to the CHSH game. fizx mentioned Scott Aaronson's "Bell inequality violation finally done right" [3] which is about the same recent experiment as the NYT article but much better. It contains a concise description of the CHSH game.
[1] https://en.wikipedia.org/wiki/Bell's_theorem
[2] https://cs.uwaterloo.ca/~watrous/CPSC519/LectureNotes/20.pdf
[3] http://www.scottaaronson.com/blog/?cat=33
https://simple.wikipedia.org/wiki/Bell's_theorem
The real problem is that when you measure A in the "up" direction, and then B in the "10 degrees east of up" direction, then B seems to know that you measured A in the "up" direction.
That is to say: B's probability distribution as a function of the direction its being measured is dependent on the direction that A is measured. There's no way to construct an "A-independent" probability distribution of B's results for arbitrary directions. The probabilities won't sum to 1 and still match experimental results.
It's unfortunate that "A up" therefore "B up" is a degenerate case of this reality where classicality actually works, because it leads to confusion.
Also, follow the links others have provided for a more formal explanation :)
exactly. The probabilities are 1 only in theory. For example theoretical Malus law over all angles would give 1. On practice - i couldn't find single photon Malus law confirmation. A slight deviation from Malus law like cutting the tail at high angles off and thus decreasing the total to below 1 (ie. when/if photons at high angles being lost disproportionally much more intensively than Malus law states) would bring the experimentally measured S to values higher than 2 while still being in the realm of local realism.
It's not strictly correct to say that Occam's Razor says we should have hidden variables. Occam's Razor is the idea that in deciding between two or more similar theories (which both have to be correct), we should prefer the simpler of the theories. Bell's theorem tells us that we should reject the notion of local hidden variables.
Importantly, this does allow the notion of nonlocal hidden variables. One such theory is the de Broglie-Bohm theory, which claims that the wavefunction is an actual physical entity, referred to as the pilot wave or guiding wave. It would be nonlocal because the wavefunction collapse of one particle to measure its properties would necessarily influence the other particle at superluminal speeds. Another nonlocal hidden variable theory is superdeterminism, which claims that everything that ever happened and ever will happen has already been determined, so the correlations are not a result of some spooky interaction, they were just "predestined" to be correlated. This would again be nonlocal because the universe as a whole would have knowledge of the particle properties a priori.
http://hansonlab.tudelft.nl/loophole-free-bell-test/
Also, it should be noted that quantum entanglement does not allow information to travel faster than light.
This sounds like lazy evaluation.