No Faster-Than-Light Travel

[Username17]

Any time where you can know the spin of a particle, the entanglement is destroyed. The only time entanglement exists is when two particles are in a superposition - they're really the same thing, a superposition exists because you don't know the states of the entangled particles. In order to send information via spin-coupling, you need to know what the spin is at your end much in the same way you need to know, at your end, if you've just made a dot or a dash when sending morse code, otherwise you're sending information free noise - and to do that means pulling it out of a superposition, which means the entanglement is destroyed. Removing the entanglement just gives you a probabilistic collapse from a superposition to a known state, meaning you get random noise at both ends which no information can be encoded in. It's effectively identical at both ends, but there's no information transmitted because it's randomised. There is absolutely no way to get around this. Any time you want to manipulate a particle, you remove the entanglement and such manipulation is what is needed for meaningful information to be encoded. Indeed, if you assume the hidden variables interpretation the appearance of entanglement and action-at-a-distance is just an illusion anyway.

So... thoughts?

That's factually wrong. Even if it were true that you could only break it down into non-superpositioning (which is far from clear), then all that would mean is that you'd have to use really a lot of entangled particles.

To a first approximation, it appears that Neutrinos pass very slightly faster than light. Relativity doesn't fail, the world doesn't end, it's just that there are tachyons and you deal with that. Instantaneous transfer of states exists, the Copenhagen Interpretation is crap, and things move faster than light all the time without traveling backwards in time and then crawling over at C.

-Username17

[DSMatticus]

To a first approximation, it appears that Neutrinos pass very slightly faster than light.

The neutrino thing is perhaps hasty. We have plenty of instances where neutrinos have been observed to behave at the velocities we expect, including an exploding star where travelling faster than c would have meant the arrival time being off by years. And we didn't see it being off by years, we saw it not being off from our predictions at all. Under normal conditions, we already know neutrinos obey the speed limit. We may have found some edge cases where neutrinos flip the fuck out, or maybe someone screwed up the equipment somewhere. I'd say either hesitant optimism or jaded cynicism are the appropriate responses at this point in time.

Quantum entanglement is a weird thing, but it can't actually transmit information FTL. Whether or not they remain superpositioned (i.e., entangled) after a measurement is actually irrelevant to the question. To be able to transmit information using quantum entanglement, you have to be able to distinguish between a random result, and a result determined by anti-correlation. You can't do this.

I know this is like a year old, but:

You pass information by reflipping it again before the next scheduled check by the other side.

You can't pass information this way, because the other side has no way of determining that you've flipped, except to flip it themself; and at that point, they can't tell if they're getting the results of a random flip (and sending you their anti-correlation), or getting the results of your anti-correlation (from your random flip).

So you need scheduled flips, and scheduled flips are all you have, because that's the only way you can tell the other side has flipped at all. But even with scheduled flips, the only information you can send is a 1 or 0 randomly. You can't even send a terminating signal. You can't embed a message in this in anyway.

[Username17]

Flipping particles isn't a 100% perfect operation, and there is random noise. But it's not completely random. It has a probability distribution and that probability distribution is mirrored perfectly on the other side. The thing is, the EPR paradox sunk the Copenhagen interpretation a long time ago. Not because information can't go faster than light, but because it can.

The demands of causality to be necessarily in the same reference frame we ca experience is intuitively reasonable, but it's physically unsupported. There is no actual reason to believe that things are never caused by events that occur microseconds in the future except human hubris. Local Realism is false, and Einstein's attempt to save causality was a failure. Things don't have to have causes in the past to have effects in the present.

-Username17

[Username17]

But to get this back on track:

if you assume the hidden variables interpretation the appearance of entanglement and action-at-a-distance is just an illusion anyway.

Let's look up that Local Hidden Variable Theory.

In quantum mechanics, a local hidden variable theory is one in which distant events are assumed to have no instantaneous (or at least faster-than-light) effect on local ones.
According to the quantum entanglement theory of quantum mechanics, on the other hand, distant events may under some circumstances have instantaneous correlations with local ones. As a result of this it is now generally accepted that there can be no interpretations of quantum mechanics which use local hidden variables.

Causality can suck it.

-Username17

[mean_liar]

Bell laid out a convincing disproving of Local Hidden Variables a long time ago. He remained a strong proponent of causality, because LHV isn't the only way to maintain causality.

What has not been determined is the final disproving of causality-maintaining Non-local Hidden Variables; ie., a pilot-wave, or decoherent histories. Wikipedia should have a lot of the competing, causality-maintaining theories.

[DSMatticus]

Flipping particles isn't a 100% perfect operation, and there is random noise. But it's not completely random. It has a probability distribution and that probability distribution is mirrored perfectly on the other side.

No, see, that's not the point; whether it's perfect or imperfect, there isn't enough going on to create an encoding scheme. Each side is producing random noise that is perfectly mirrored by the other side, so the only thing you can 'send' using quantum entanglement is random noise. And it turns out it's impossible to create an encoding scheme out of random noise (not at all surprising, I think there's a formal proof somewhere) that can send messages.

Which is why this:

Not because information can't go faster than light, but because it can.

isn't a problem. There are already tons of effects in the universe that propagate FTL, but we've shown that they can't be used to transmit information, which means they don't violate causality. Quantum entanglement is another such effect. It is an effect that propagates FTL and carries only random noise; no information is transmitted.

[Username17]

The. Noise. Is. Not. Random. It happens or not at a specific time. That's a bit. A zero or a 1.

And since you can check to see if noise has been sent at all, that's meta information. And meta-information is still information. Causality can suck it. Causality is dead the moment that something propagates faster than light, whether we can use it or not.

The "no information" theories require actually no information. Not "no useful information", and not "no comprehensible information". But the Bell experiments shot the "spooky action at a distance" whiners in the nuts. It's a done deal. There are quantum entanglement communication systems that already exist. They just suck right now. But they exist.

-Username17

[DSMatticus]

It happens or not at a specific time.
...
And since you can check to see if noise has been sent at all, that's meta information.

Okay, this is the part where things break down. You can't actually check to see if you've been sent something. I think you're under the impression that you can tell the difference between "being the dude who collapses the superposition" and "seeing the results of the other guy's collapse." If we could do that, we could build an FTL communication device.

But I'm under the strong impression that this in fact impossible, and you'd have to explain/link/cite it to me.

The "no information" theories require actually no information.

Random noise actually isn't information. It's weird, but it isn't. Random noise is not a description of the state at the other end, which is the definition of information in the physics sense. (The information that their random noise is perfectly mirrored with your random noise propagates in your head, at whatever speed you happen to be travelling; or at c, even.)

[Stahlseele]

Did somebody say faster than light?
http://www.reuters.com/article/2011/09/ ... FH20110923

[Blicero]

Did somebody say faster than light?
http://www.reuters.com/article/2011/09/ ... FH20110923

The neutrino thing is perhaps hasty. We have plenty of instances where neutrinos have been observed to behave at the velocities we expect, including an exploding star where travelling faster than c would have meant the arrival time being off by years. And we didn't see it being off by years, we saw it not being off from our predictions at all. Under normal conditions, we already know neutrinos obey the speed limit. We may have found some edge cases where neutrinos flip the fuck out, or maybe someone screwed up the equipment somewhere. I'd say either hesitant optimism or jaded cynicism are the appropriate responses at this point in time.

[Username17]

Okay, this is the part where things break down. You can't actually check to see if you've been sent something. I think you're under the impression that you can tell the difference between "being the dude who collapses the superposition" and "seeing the results of the other guy's collapse." If we could do that, we could build an FTL communication device.

You don't check to see if you've sent something.

In 1964, Bell [3] formalized the notion of two-particle nonlocality in terms of correlations among probabilities in a scenario where one of a number of a measurements are performed on each particle. He showed that the results of the measurements that occur quantum physically can be correlated in a way that cannot occur classically unless the type of measurement selected to be performed on one particle affects the result of the measurement performed on the other particle.

So. If your measurement is one that cannot occur unless a measurement was taken off schedule, than that is a 1. If it is a measurement that can occur without there having been a measurement on the other end, that's a zero. You have now sent binary information. In 1964. Now repeat an arbitrary number of times to send as many bits as you feel like sending.

I genuinely don't know if it is possible to send information encoded into the specific spins that particles develop. But since I'm not going to be the person actually making these things, I don't even care. Simply the fact that you can send information at all is proof of concept that a sufficiently large particle array could send sufficiently large information strings. Because no matter how particles add up or don't, once you've defined a zero or a one, you can write that down and do it again.

-Username17

[fectin]

But you need poke as well as peek, otherwise you're just listening to static.

[Username17]

But you need poke as well as peek, otherwise you're just listening to static.

It does not matter. Because an experiment provided a positive result, that result can be interpreted as a 1. Not doing it can be interpreted as a 0. Because the Bell Experiment was able to provide a result at all, it is unambiguously true that you can send information.

No matter how it works, it does work, and that means that you can create a setup that transmits information. Since the Bell Experiment was able to distinguish between making a measurement at the far end and not doing that, then at the very least you can have scheduled measurements and then skip or double the schedule in order to send information. The very moment that you can distinguish between binary states - that of doing something and not doing it - you can send information. That's how information works.

We think of information as reading bits, but for data transmission you don't have to have unambiguous bits - you have to be able to tell the difference between any set of states and any other set of states. You can send information through frequency modulation, amplitude modulation, or in this case: state probability modulation.

-Username17

[Maxus]

So the only thing keeping us from faster-than-light communication is making an apparatus which can make an atom freak out on demand, at high-speed, and and read it at the other end.

Wouldn't be that much use on Earth...well, apart from stealthy communication. but there's applications for space, especially if we ever get to Mars.

[Username17]

So the only thing keeping us from faster-than-light communication is making an apparatus which can make an atom freak out on demand, at high-speed, and and read it at the other end.

Wouldn't be that much use on Earth...well, apart from stealthy communication. but there's applications for space, especially if we ever get to Mars.

The big difficulty is getting enough atoms to stay freaked out in the proper way long enough to pull them far enough apart that you would actually care about sending messages from one end to the other.

But yeah, there is also some really cool math about how you can make a quantum communicator that can tell you with 100% certainty if anyone in the universe could have listened in. It can't tell you if anyone did, since symmetry can be broken in other ways - but if the symmetry is still there you know for an immutable fact that noone else got any portion of your message. And that's pretty neat for cryptographers.

-Username17

[Quantumboost]

I recently found a paper by David Deutsch & Patrick Hayden about information flow in quantum entanglement, which indicates that Bell's Theorem doesn't actually hold - which means that you *could* have both causality and locality, which would prevent FTL entanglement communicators. Unless there's some craziness where performing the experiment splits the universe and decides which split you end up in or something.

I can't quite follow all the mathematical bits, but it's fascinating otherwise.

[DSMatticus]

He showed that the results of the measurements that occur quantum physically can be correlated in a way that cannot occur classically unless the type of measurement selected to be performed on one particle affects the result of the measurement performed on the other particle.

What I want to say is that that sentence is written like shit and doesn't quite mean what it's being read to (the actual paper in question goes on to talk about the impossibility of FTL in the very next paragraph).

Since the Bell Experiment was able to distinguish between making a measurement at the far end and not doing that

I don't believe that's something Bell Experiments allow you to do until you compare results; and that means transmitting information classically. You can see the correlation "after the fact" once you have the other person's results, but not before. The no-communication theorem says as much, but I can't even begin to defend the math. I do think I can explain it, though. Here's an example using only one axis of measurement:

Code: Select all

Alice sends 1 (measures):
1) Alice measures the X-axis and the result is +X by random chance (50%). Bob measures the X-axis afterwards, and the result is -X due to the entanglement.
2) Alice measures the X-axis and the result is -X by random chance (50%). Bob measures the X-axis afterwards, and the result is +X due to the entanglement.

Alice sends 0 (doesn't measure):
1) Bob measures the X-axis and the result is +X by random chance (50%).
2) Bob measures the X-axis and the result is -X by random chance (50%).

Alice's choice to measure has not affected Bob's probability distribution. Note that in this example, if Bob had Alice's results, he could look at them and say, "everytime Alice got +X, I got -X, and everytime Alice got -X, I got +X, and everytime Alice chose not to measure I got random results." But because Alice's results are random, from Bob's perspective the act of Alice measuring is exactly as random as Alice having not measured at all until he knows what Alice's results actually were.

Bob cannot differentiate between Alice having measured and Alice having not measured, meaning that that bit of information has not been sent.

The actual Bell experiments are way more complex than this, because they use every angle of measurement, not just along the axes. And it turns out at that point, the correlation is only probabilistic, not deterministic. So it looks something like this:

Code: Select all

Alice sends 1 (measures):
1) Alice measures along the X-axis, and the result is +X by random chance (50%). Bob measures at some non-right-angle to the X-axis, and the closer his chosen angle is to the X-axis the more likely he is to get the complement of Alice's measurement, a negative.
2) Alice measures along the X-axis, and the result is -X by random chance (50%). Bob measures at some non-right-angle to the X-axis, and the closer his chosen angle is to the X-axis the more likely he is to get the complement of Alice's measurement, a positive.

Alice sends 0 (doesn't measure):
1) Bob measures at some angle to the X-axis, and gets a positive by random chance (50%).
2) Bob measures at some angle to the Y-axis, and gets a negative by random chance (50%).

The difference here is that instead of deterministically affecting Bob's result (Alice gets +X, so Bob will get -X), Alice is only probabilistically affecting Bob's result (Alice got +X, Bob's axis of measurement is close to Alice's, so Bob is very unlikely to get a positive and very likely to get a negative).

When Alice doesn't measure, Bob is looking at a 3d probability distribution that is completely symmetrical. No matter what axis or angle he chooses, he will always get positive 50% of the time and negative 50% of the time.

When Alice measures, and get +X, the probability density around +X thins for Bob, and the probability density around -X gets thicker (because his results have the tendency to be the opposite of her's due to the entanglement). The reverse is true when Alice gets -X; the probability density around -X thins for Bob, and the probability density around +X gets thicker. So Alice's act of measurement produces two 3d probability distributions, which are not symmetrical. Unfortunately, Bob has no idea which 3d probability distribution Alice has created for him, because Alice's selection is random. And when you average the two 3d probability distributions Alice can create, they form the completely symmetrical one that Bob looks at when Alice doesn't measure.

Once again, Bob is incapable of differentiating between whether or not Alice has measured, meaning that this bit of information has not been sent.

So... what am I missing?

[Username17]

So... what am I missing?

Last year the Chinese sent a series of quantum entanglement teleports 16 kilometers at a fidelity of 89%. While there are profound limits to how much the system can tell you about itself, as human beings we can have information about it that does not come from the system: we call them "schedules". By contrasting observed reality with the scheduled reality changes, we can have more information than the system itself is capable of giving us.

The assumption is that we have to read qubits in order to pass information. We don't. We can compare a probability distribution of a series of particles to an expected probability distribution based on the scheduled actions at another site. We can get information from things that don't involve any specific correspondence to specific states.

As soon as fidelity got over 50%, instantaneous information transmission became possible. Last year it crept up to nearly 90% over 16 kilometers. Not good enough to contact Mars, but good enough to prove the concept.

-Username17

[DSMatticus]

The assumption is that we have to read qubits in order to pass information. We don't. We can compare a probability distribution of a series of particles to an expected probability distribution based on the scheduled actions at another site. We can get information from things that don't involve any specific correspondence to specific states.

No, I understand what you're suggesting now. But it all depends on the act of measurement leading to a different probability distribution than the act of non-measurement. And while my familiarity with the math is very, very basic, I was under the impression those probability distributions were identical, and that's what I was trying to explain.

Here's a crappy 2d simplification I drew in a few minutes and/or 9000 mspaint hours that covers the three cases.

In the first diagram (C1), Alice doesn't measure, and Bob's results follow a uniform distribution. In the second diagram (C2), Alice measures and gets +X, and because they're entangled Bob's results want to be the opposite of Alice's, so the results surrounding -X become more likely, and the results around +X become less likely. (I drew this as crappy little white and maroon circles; it isn't actually discrete like that, I should have put a shading gradient over the circle or whatever so it went from 'white' to 'maroon;' pretend). When you look at the third diagram (C3), you notice that it is the exact mirror (well, it's supposed to be; pretend some more) of C2. This is because Alice's two possible results affect the probability distribution in equal, but opposite, ways.

So what's the point of this? C2 and C3 are equally likely to occur, and it turns out that (.5)C2(X,Y) + (.5)C3(X,Y) = C1(X,Y). The probability distributions are identical whether Alice measures or not. There is no experiment Bob can perform at his end to tell if Alice has measured her particle.

[RadiantPhoenix]

So, when you measure the (previously unmeasured) quantum state of the whatever, you get a(n essentially) random result, and someone else who measures the (formerly) entangled particle will probably get a similar result to the one you got?

[DSMatticus]

someone else who measures the (formerly) entangled particle will probably get a similar result to the one you got?

Opposite, not similar, but otherwise yes. When you and the other person are measuring the same thing, it's actually definite instead of just probable.

If you measure the x-axis, and get a +X, you create a spot of zero probability at +X for the other person. You create a spot of twice the original probability at -X. The reverse is true when you get -X. Which means if you and your friend measure the same axis, you will always get directly opposite result-pairs (+X,-X), 100% of the time.

But in this case, the proper way to think of an atom's state isn't just a simple "yes/no?" for each of the three axes. It's actually better modelled like a sphere. If you measure the x-axis and get +X, every result on that sphere 'around' +X gets less likely as well, and every result on that sphere 'around' -X gets more likely.

So imagine that you measure the x-axis, and your friend is measuring along an axis that is really, really close to the x-axis (call this axis M, and let's say it's rotated 5 degrees away from the x-axis). If you get +X, your friend is very, very unlikely to get +M, because +X and +M are 'close' to eachother on the sphere. He's much more likely to get -M, because it's very far away from +X. If you got -X instead of +X, the reverse is true.

So it seems like when you measure your particle, your friend's probability is altered in a way that is dependent on your result. But...
1) Which result you get is random.
2) The two probability distributions each result leads to are perfectly balanced with respect to eachother.
The net result is that you randomly send your friend to 1 of 2 probability distributions that, in tandem, average out to be the same as having not sent your friend to a different probability distribution at all. Ergo, he has no way of detecting whether or not you've measured using probability alone.

Edit: Though there are other promising experiments for FTL communication. I believe we've found that the double-slit experiment holds for entangled particles, which means you can use the act of measurement/non-measurement to create different behavior at the far end. Not just alters the probability, leads to deterministically different behavior. I'm not certain about this, I might have something important wrong here. The experiments are still being made to see if we can use it to transmit information.

[RadiantPhoenix]

So, it's equivalent to having a bunch of hidden random bits that you then randomize?

[DSMatticus]

Hm. Kind of. Depends what you mean, exactly. I'd put it more like equivalent to sending someone random bits to overwrite their random bits.

[Cynic]

Really if we can't go FTL, is there a speed (FTL-X) that can be achieved that's relatively close to FTL.

[Doom]

In theory, sure; there's even a 'solar sail' type engine that might do it...granted, it'll take a long time to get to that speed.

Trouble is, 'near c' is still basically useless for any particular human being wanting to travel the galaxy, even if it's not a problem for machine that don't, like need to eat or have a relatively brief lifespan.