Making Sure Our Lab Equipment Isn't Tricking Us

An anonymous reader writes "In a newly published paper, MIT researchers propose an experiment that may close the last major loophole of Bell's inequality. The test is to see whether, as far-fetched as it sounds, a particle detector's settings conspire with events in the shared past to determine which properties of a particle to measure — a scenario that implies that a physicist running the experiment does not have complete free will in choosing each detector's setting. MIT’s David Kaiser says, 'It sounds creepy, but people realized that's a logical possibility that hasn't been closed yet. Before we make the leap to say the equations of quantum theory tell us the world is inescapably crazy and bizarre, have we closed every conceivable logical loophole, even if they may not seem plausible in the world we know today?' The test involves quasars, telescopes, and lots of deep, deep space. It was published this week in the journal Physical Review Letters (pre-print available at the arXiv)."

Comments

Please!

[NotQuiteReal]

Don't anthropomorphize the machines. They hate that, and will go back into their past to get you!

Re:Please!

[interkin3tic]

If only I had mod points today... I had some last week, so I'll be right back.

Sincerely,
A confocal microscope

again with the assumptions.

[clovis]

From the article:
The idea, essentially, is that if two quasars on opposite sides of the sky are sufficiently distant from each other, they would have been out of causal contact since the Big Bang some 14 billion years ago, with no possible means of any third party communicating with both of them since the beginning of the universe — an ideal scenario for determining each particle detector’s settings.

Why would you assume that if they're 14 billion years apart that it would be any different than 14 seconds apart in time, at least in regard to entanglement?
" with no possible means of any third party communicating" makes me think "we don't know of a means to communicate"
Could the outcome of the experiment could show either action at a distance, or some faster-than-light communication without excluding either possibility?
If it does happen that entanglement went away, it would be most interesting.

Re:again with the assumptions.

[Goldsmith]

Exactly right.

There are two possibilities:
1) The universe is infinite, and it would be possible to find two quasars which never shared a quantum state.
2) The universe is not infinite and it is not possible to find two of anything which have never shared a quantum state.

They've completely failed to close this loophole.

Re:again with the assumptions.

[MozeeToby]

You cannot have entanglement without interaction, you cannot have interaction between two things that lie outside of each other's light cones.

To be fair, you can't have interaction outside of your light cone without also having faster than light communication. But you can't have faster than light communication without also having the possibility of sending messages back in time. And you can't have the possibility of sending messages back in time without breaking causality. So, on the one side you have the assumption that causality doesn't exist and faster than light communication is possible (both of which are contradicted on scales from pico-meters to billions of lightyears), on the other side you have the assumption that information can't travel faster than light (which again, seems to be supported by every experiment and observation made in human history).

Re:again with the assumptions.

[BlueKitties]

All experimental evidence suggests that the transfer of information faster than the speed of light in a vacuum is impossible. Scientists understand full well that there is always a possibility some new discovery could upend even the most fundamental view of the universe, but until evidence to suggest otherwise comes along, scientists can only work with the experimental data they have -- and so far, all of that suggests FTL transfer of information is impossible.

Therefore, drawing from the incredible amount of experimental data, it is reasonable to conclude that such stars could not have conspired with phenomenon on Earth, since the information would need to have left Earth, reached the star, then returned. The universe simply isn't old enough for such a round trip to have happened. If the max speed you can travel is 1,000 miles in an hour, and the universe is one hour old, then it is impossible to make a round trip to a destination 1,000 miles away. Therefore, if someone arrives from that destination, we know it is impossible for them to have conspired with someone from where you are at, since such a round trip would have taken at least two hours, and the universe is only one hour old.

Re:again with the assumptions.

[clovis]

Turns out that if I had read the pdf to the end, I would have seen that they discussed the alternate expected outcome's implications in detail.

Re:again with the assumptions.

[camperdave]

How do entangled particles "seem" to communicate? Everything I've ever read points to one particle just being the opposite of what the other one is (eg. up spin on one, down spin on the other).

Re:again with the assumptions.

[BlueKitties]

But the conspiring forces can't know what the quasars are sending off! That's the point. Nothing is capable of containing information pertaining to what the quasars have sent, because it's impossible to have made a round trip to find out ahead of time. Therefore, even if the forces can influence the scientist's choice of quasar, the mysterious force still has no idea which quasar will say what, so it can't rig the experiment.

As an analogy, imagine there are three settings: A, B, and C. The mysterious force wants to rig the experiment by forcing the scientist to select setting B. The scientist, to get around this, allows three quasars to select a setting. One of the quasars will pick A, one B, and one will select C. The scientist will then pick one of the quasars to set the settings. The mysterious force can choose which quasar the scientist will pick, but the mysterious force doesn't know which quasar will send which setting, therefore it can't force the scientist to pick B.

Superdeterminism

[Warbothong]

While interesting, it doesn't solve the most glaring assumption of Bell's inequality which is that the Universe is non-deterministic.

It's perfectly plausible that the Universe is deterministic, and hence the behaviour of the particles *and the experimenters* is pre-determined, ie. there is no choice in which measurement to take. Taking the determinism of the observers into account tends to be called "superdeterminism", and is necessarily a global property: either the whole Universe is superdeterministic, or nothing is deterministic. Bell's inequalities demonstrate this, since they cannot be explained by a *local* deterministic model, ie. a model which only involves properties of the particles (known as 'local hidden variables').

Note that superdeterminism doesn't necessarily rule out 'free will'. Personally I find the most elegant explanation of free will to be irreducibility: an irreducible process has no 'shortcuts'; the only way to predict its result is to run the process from start to finish. If, say, my mind is a deterministic but irreducible process, then a powerful-enough computer could predict my decisions exactly. However, I can still be said to have 'free will' because the computer can't take any shortcuts in its calculations: the only way it can predict my decisions is to run a perfect simulation of me and see what decisions that simulation makes, but in that case it's still (a perfect simulation of) 'me' making the decisions.

Re:Superdeterminism

[jfengel]

Yeah, superdeterminism was my first thought in reading this. It sounds like they're pushing any superdeterminism all the way back to the time of inflation, but since that's exactly what superdeterminism would predict, I don't see that they've contradicted anything.

It's intellectually unsatisfying to think that superdeterminism could relate to something as supremely complicated as a scientific apparatus: the whole state two measuring apparatuses conspires to yield opposite results on particles that were, up to that instant, completely identical, without any communication. But I think it makes more sense than trying to impose some outside "free will" force that also makes itself visible only on the most carefully isolated particle experiments yet also just happens to manifest as something we see numerous orders of magnitude larger as "what we think", despite layers of purely chemical interactions in between.

We're still obligated to explain the larger-scale version of "free will", in that the phenomenon that we believe it exists is real, and I think your way of looking at it is good as any. And superdeterminism doesn't contradict that.

Superdeterminism still doesn't satisfy, but I suspect that "satisfaction" is a purely human property. The equations yield the right answers, and that's all you get. Like classical dynamics, free will is an idea that we're going to keep expecting to see, even though we'll always get out unsatisfying answers when we try to explain corner cases.

News will report it as proof of Free Will

[Remus Shepherd]

A year from now you should expect to hear about this research again, but it will be delivered as a dramatic result: "Scientists have proven Free Will exists!", or "Scientists have disproven Free Will!" The experiment won't actually do this, but that's how the press will report it.

The thought that some hidden variable may affect not only both sides of the universe but our own minds is frightening. It would really shake things up. So I expect that QM and 'free will' will come out triumphant in this test. Whether it's an actual assessment of Free Will or not will be the interesting argument afterward.

Re:News will report it as proof of Free Will

[TheCarp]

I am kind of hoping they will close the loophole and find the equality is still violated. If so, then i actually expect the headline "Do far away stars disprove quantum mechanics?"

Re:News will report it as proof of Free Will

[Immerman]

As it is classical physics had already ruled out the possibility of free will centuries ago, and QM only inserted a bunch of dice-rolling that could give the illusion of it. For free will to actually exist we must assume some sort of non-physical "soul" that can manipulate the outcome of the dice.

Theoretically at least we could monitor every quantum effect that influences the brain and look for statistical anomalies, but even if no anomalies were found that wouldn't rule out free will - rolling 2,5,3 is statistically equivalent to rolling 3,2,5, but the effect on the game can be very different.

Also, being scary and disruptive is hardly an argument against the accuracy of a theory - if it were then nuclear physics wouldn't exist.

Re:News will report it as proof of Free Will

[Immerman]

I think your teachers didn't really understand what they were talking about. I would say rather that GR simply makes "now" a much more complicated concept than in Newtonian physics. At any given point in space, time is still presumed to behave as it does in Newtonian physics. The fundamental difference is that GR recognizes that you can't change your reference position and velocity in space without also changing your position and velocity in time - the two concepts are interwoven rather than being independently absolutes as postulated by Newton.

It seems you also fail to understand free will - if the forked minds had free will then it must be possible for one to chose to do something different than the other. They may give the *appearance* of free will due to chaotic interactions with the environment, but if their behavior is predetermined for a given sequence of inputs then there is no free will involved, only variation in the inputs. Self-modification changes nothing in that regard - given identical inputs the forked minds will modify themselves in identical ways.

And no, it would not be like a person having two brains - at a minimum forked brains would be subjected to different quantum noise which would cause their states to rapidly diverge. If you introduced some source of random noise into the functioning of the AIs you could cause them to diverge unpredictably as well, but that has to come from a hardware source, not software (there's a reason I singled out digital software AI, both analog and quantum computing is inherently noisy) Note too that nothing says that the noise grants free will, in either the AIs or brains, it only grants non-deterministic behavior (and that presumes the noise source is itself non-deterministic). That is necessary, but not sufficient, for free will. Free will requires that the mind be able to choose it's path, such that another mind subjected to identical inputs (noise included) has the possibility of following a different path. (As an aside it's probably worth pointing out that at present there is no evidence that we have free will ourselves - it is quite possible that free will is simply an evolutionarily advantageous illusion)

Show me a convincing artificial mind and I'll accept your speculation as to the validity of the strong AI theories it's based upon, until then they don't even qualify as theories. Theory requires evidence, you have only hypothesis.

how about detector poisoning?

[swschrad]

a much simpler explaination... the detector material is still groaning from the last collision and doesn't have its calibrated act together for the next one. you detect subatomic particles, after all, by watching what happens when they distort a known material, and extrapolate from the distortion what whacked into it. whacking things causes them to go off kilter. from black bands and reduced light in fluorescent light tubes to bright-bloom in old TV cameras to getting wacky when you leave a dark room and are sun-blinded, this has been a known phenonema as long as we have been around.

Re:What?

[Remus Shepherd]

You are neglecting the possibility of true dualism -- that we have a soul. If a portion of our decisions comes from an extra-universal source that does not follow the deterministic rules of the universe, that would provide us with true free will. Sadly, such a thing may be innately non-provable.

Re:How about not-quite-random numbers?

[Immerman]

AC has it right - *any* source of (psuedo)random numbers from causally connected sources is suspect, and no number of repetitions will rule out forces that retroactively ensure consistency. Frank and Bob could be simultaneously flipping coins on the opposite side of the planet, yet still be causally connected by the fact that their results will determine the settings for an experiment that the universe decrees must be consistent with certain principles. Even this experiment won't be able to close the loophole completely - it could be that the experiment and its settings were already fully defined in the first few instants of the universe while the not-matter that would eventually become the quasars was still causally connected. Basically this is a negative-evidence only experiment - it can't show that hidden variables aren't controlling QM, but it could potentially show that they are. Unless of course the hidden variables ensure that the choice of quasars will be such that the experiments yield consistent results. Though if they do it multiple times with different quasar pairs the odds of that being possible drop dramatically.

Re:How about not-quite-random numbers?

[TechyImmigrant]

>Who comes up with the formula for the PRNG?

For all practical purposes, I do.