What qualifies as an observer in quantum mechanics?

[DennisN]

It is information that cannot travel faster than light, as Brian Cox points out in his book.

According to the Schrodinger equation, at t=0 a micro (and in principle macro) system is at position 1. Then at t=1, it is potentially everywhere else in the universe (until observation occurs). That's why we can observe the rock to be at position 1 at t=1, and then at position 2 at t=2, because at t=2 it is a potentiality to be there.

Ok, it was a quote regarding the wave function (as I suspected). As long as the velocity of any information (incl. the velocity of a "rock" or any massive particle) does not exceed the speed of light, I'm fine, and so is relativity.

 

[meBigGuy]

According to the Schrodinger equation, at t=0 a micro (and in principle macro) system is at position 1. Then at t=1, it is potentially everywhere else in the universe (until observation occurs). That's why we can observe the rock to be at position 1 at t=1, and then at position 2 at t=2, because at t=2 it is a potentiality to be there.

I have trouble with the words here. If at t=1 you observe it at position 1 it is entangled with the observer. It cannot then just appear at 2 unless it (and the observer) agree was in superposition of 1 and 2 in which case it was not at position 1, but potentially at 1 and 2. Or am I missing something.

 

[StevieTNZ]

I have trouble with the words here. If at t=1 you observe it at position 1 it is entangled with the observer. It cannot then just appear at 2 unless it (and the observer) agree was in superposition of 1 and 2 in which case it was not at position 1, but potentially at 1 and 2. Or am I missing something.

It is still in a potentiality to be at position 1 and 2 at t=1, even if its observed to be a position 1 at that time. That's because no collapse of wave function occurs, in accord with the Schrodinger's equation. Therefore at t=2, it can stay at position 1 where it is observed to be, or go to position 2, if the observer is observing that position.

 

[meBigGuy]

It seems it can show up at 1 or 2, but not 1 then 2 if it was observed at 1. How is the entanglement by observation at 1 disentangled or rendered non-existant?

 

[StevieTNZ]

If that was the case, then nothing would move.

 

[meBigGuy]

That's different. The application of other influences changes everything. I don't think that is what was initially being claimed, or at least implied. It won't move from 1 to 2 in absense of a force in conjunction with the environment.

 

[StevieTNZ]

The author of Quantum Philosophy produces that example as it is possible, in principle, according to QM. No force is needed to move the rock from position 1 to 2. Roland claims it is a tunneling effect, with a very small probability (but not zero) of occurring.

I invite the honourable member to read the book, in particular pages 190-193.

 

[Ken G]

I would say that focusing on either decoherence, or even the coupling of two quantum systems, repeated for Avogadro's number of interactions, is still not the issue that separates the interpretations, and really has nothing to do with the measurement problem. As bhobba and stevendaryl pointed out, it's just basic quantum mechanics that a quantum system will become entangled with its environment, and if we average over all the information in the environment that we, the physicists, are simply not choosing to track, then of course we get a mixed-state treatment of the result. The mixed state can then be used to predict experiments, statistically, and we get lovely agreement, because statistical agreement is all we are shooting for anyway.

The measurement problem is something quite different. It is the question, what determines the outcome that we actually perceive in any single experiment? Is it information that actually exists in the environment, that causes that outcome but we just weren't tracking it? (That sounds like an ensemble interpretation, or perhaps Bohmian as well.) Is it that there is no quantum world in the first place that would need to have a particular outcome to be caused, because the methods of physics end in the macro domain and the quantum world behaves statistically because it is basically a figment of our imagination? (That sounds like Copenhagen, which I like because it is the most overtly skeptical about our own theories.) Is it that the entire issue is moot, because all outcomes actually do occur, and it is just an illusion of our minds that we exist in a sector where only one happened? (That sounds like many worlds, yet note how close it is to Wigner's consciousness being responsible for the collapse.) So we need an interpretation not to understand the equations we are writing down, we need it to understand our experience of the experiment. That's why they are so subjective, they are actually explaining a fundamentally subjective aspect of physics.

 

[meBigGuy]

As I learn more I become less able to relate to the interpretations. They seem contrived. Trying to explain things that can't be explained. I always favored relational interpretations, but they aren't really saying anything more than things are relative and related, which is pretty obvious.

There is a thought experiment/analogy I like called the (quantum) spooky socks that makes a strong point (for me, anyway) about what is real and what exists. I'm not allowed to post the link, but I'm interested in how the statement made by spooky socks relates to interpretations. It is epitomized by the difference between "filling the drawers" and "preparing the chest".

 

[bhobba]

As I learn more I become less able to relate to the interpretations. They seem contrived. Trying to explain things that can't be explained. I always favored relational interpretations, but they aren't really saying anything more than things are relative and related, which is pretty obvious.

Now you are starting to understand the truth of the situation - all interpretations suck in their own inimitable and peculiar way. As you think about and compare interpretations you get the feeling all you are doing is mapping bits of the same big elephant and we have a long way to go before we see it whole.

I am an advocate of decoherence as being a big step forward in interpretations, but I think we have a long way to go. I believe string theory will eventually have something very important to say about it - but only time will tell.

Thanks
Bill

 

[Ken G]

Personally, I think there is a silver lining to the unsatisfactory character of the interpretations-- it is trying to tell us that we don't just need a better theory, we need a better idea of what a theory is. Too long we have gotten away with imagining that physics is happening "out there", with no reference to us, and we are just flies on the wall, taking notes. But that's never what physics really was, or is, it's just an idealization that we got away with for a few centuries. Real physics is done by a physicist, we involve ourselves in nature before we try to figure nature out. This is an inescapable aspect of just what physics is, the breakthrough in science where we stopped being flies on the wall (watching the wheels of the cosmos) and started being players. We realized that we can interact with nature, and use that interaction to figure nature out. That's what we call experiments! But there's a price to pay-- it means we have to realize that physics involves a physicist. That's the primary philosophical impact of quantum mechanics, and the source of the dissatisfaction with our interpretations. Let us not miss the opportunity to learn the appropriate lesson-- we must recast physics as something that we are doing to understand nature, we must understand the role of the physicist in what physics has always been. Perhaps the next great theory will have that flavor-- but if so, string theory is not getting that message.

 

[meBigGuy]

Can't we define observation as any action that entangles entities, one of which we might call a measuring device? For example, in the 2 slit experiment, it is the entanglement of a some measurement device (possibly just a particle) with the "test" particle at 1 slit that destroys the interference pattern. If the entanglement (measurement) is weak, the pattern is only partially destroyed. I've always thought that any interaction qualified as observation. The "strength" of the observation depends on the degree of interaction.

 

[bhobba]

Can't we define observation as any action that entangles entities, one of which we might call a measuring device?

Not really - measurement is a kind of entanglement but the converse is not necessarily true.

I think the best way to define it is simply when a system, environment, and measuring apparatus is in a mixed state after decoherence. Interpretations differ on what you can infer from that but the QM formalism is clear - the result must be one of the 'elements' of the mixture and its probability is its proportion in the mixture.

But, as I think Ken was pointing out, pinning it down is a slippery issue.

Thanks
Bill

 

[meBigGuy]

There needs to be a result or at the least a probability of a result? Entanglement in general doesn't mean there is a result (yet). Does that work, or am I just digging myself in deeper.

 

[Ken G]

I think you're pretty close, but we can be more precise. What a "result" means is that some space of mutually exclusive definite states ("eigenstates") are having their mutual correlations completely scrubbed by the interaction with the environment. Not just any environment will do that in regard to any set of definite states, it requires a very purposefully chosen environment to do that.

Consider, say, a momentum measurement. Nature doesn't usually do those on her own, if we want a quantum system, like a particle, to have a definite momentum, we need to do something quite purposeful to that system that just isn't going to happen naturally (though it depends on what kind of precision we will tolerate). That's what I mean by the "role of the physicist", we have designed very special interactions to scrub coherences in a very specific way, such that the state of the quantum system becomes a perfectly mixed state in regard to, say, momentum. That's basically an environmental interaction that must produce states of definite momentum, even though we don't know which one it will produce. So that's a measurement, it's not just any entanglement, it's a very strong entanglement that will produce a perfectly mixed state (no remaining correlations whatsoever between some set of definite states), and even more, it has to be a mixed state with regard to some set of definite states that we actually recognize as something physical (like momentum, or location, etc.).

It has even been said that ultimately, all we ever do are position measurements (the location of some meter, etc.), though it's probably not important to try to be that specific about it. The point is, it has to be an interaction of an extremely special type, a small subclass of the things nature does to its quantum systems. We as physicists can't really create any instrument that nature couldn't make on her own, but still, we go to a lot of trouble selecting those instruments very carefully, until we have the ones that are good enough to be considered measuring devices. By restricting the tools of physics to such a small subclass of natural phenomena, we try to use the processes we understand to try to figure out the ones we don't. But what's hiding in those processes we think we understand? That part we can never get at with physics, because we always have to use something we think we understand, chosen from that special class of interactions, the measurements, to try and understand everything else. And we are surprised there is a "measurement problem"?

 

[bhobba]

And we are surprised there is a "measurement problem"?

Well once you have a fundamental theory concerning what objects (called measurement devices etc etc) 'register', and those things are built from what that theory is supposed to explain, you are bound to have issues.

Thanks
Bill

 

[Ken G]

Exactly. It's amazing it took this long for us to encounter the conundrum, I guess we had to get to a point where we were bridging suitably large gaps between what we already understood and what we wanted to understand. I believe that concept was at the heart of what Bohr was saying.

 

[stevendaryl]

Exactly. It's amazing it took this long for us to encounter the conundrum, I guess we had to get to a point where we were bridging suitably large gaps between what we already understood and what we wanted to understand. I believe that concept was at the heart of what Bohr was saying.

I think that there is some kind of "measurement problem" that is inevitable, regardless of the laws of physics: Any attempt to measure a quantity requires a physical object to do the measurement, and requires an interaction between the measuring object and the measured object, so it can be difficult (or maybe impossible, in some circumstances) to complete disentangle the two. The original understanding of Heisenberg's uncertainty principle was in terms of a measurement disturbing that which is being measured. For example, to accurately determine the position of an object, you have to "see" the object using very short-wavelength light, and the light imparts momentum to the object. So attempting to measure position accurately makes the momentum uncertain.

But I don't think that quantum mechanics, and its "measurement problems" are really explained by such a disturbance model. The EPR experiment shows that. If you create two particles with identical (or complementary) properties such as spin or momentum, then you can find out about one without disturbing it, by measuring the corresponding property of the OTHER particle.

So even though the general discussion of the problem of measurement might seem to make the quantum mechanical situation more understandable, I think that looking at the details shows that quantum mechanics has its own problems that are not the same as the generic measurement problems.

 

[Ken G]

Yes, it's more than just the disturbance issue, complementarity comes from the wave/particle duality. That duality is a classic example of the problem of using one thing that you think you understand to try to figure out the other. When waves are aggregates of particle behavior, like a sound wave is an aggregate of air molecules, you can write the equations for the particle motions and show how the wave equation emerges. But what if the particles already obey a wave equation, how can we understand waves using particles then? The wave behavior cannot be said to emerge from the particle behaviors. Or, if we abandon particles, and just try to understand wave mechanics in its own right, so we understand a two-slit diffraction pattern in the language of interfering waves, then we have trouble saying why we only get one tiny "blip" at a time-- the particle behavior doesn't arise completely from the mechanics of waves.

 

[meBigGuy]

I just stumbled across "Experimental motivation and empirical consistency in minimal no-collapse quantum mechanics" .(schlosshauer) http://arxiv.org/pdf/quant-ph/0506199v3.pdf

Puts it all in perspective for me. Doesn't leave much room for hocus-pocus interpretations.

 

[Ken G]

But here's the problem. When they say "the perception of single “outcomes” is likely to be explainable through decoherence effects in the neuronal apparatus", how does it help answer the following question:
What determines which outcome results from an experiment that must yield one of several outcomes?
I see no resolution of that question, anywhere in that paper.

 

[StevieTNZ]

I don't think we can ever know what determines whether +1 or -1 (for example, spin) will result (note that is the outcome +1 over -1, not that one of them becomes real). I don't think we can rid any future theory of randomness, unless we discover hidden variables that allow us to predict with certainty either +1 or -1.

 

[Ken G]

And note we had a similar problem in classical mechanics, predicting weather and so on. But at least we had a theory whereby we could (erroneously) imagine that the weather was actually determined, we just couldn't ever get precise enough information to "know what nature knows about itself." Of course it was always a fantasy that nature really had that information stored in some memory bank somewhere, but we could imagine it. Now, we don't have that any more, we can't even imagine it unless we go beyond our theories and choose an interpretation that allows that.

 

[marksesl]

I don't really mean to imply that a conscious observer is absolutely necessry to collapse the probability wave, but I do have a couple of problems with your answer.

Here you defined observation as interaction/measurement, which can sometimes leave people with the wrong impression, that interaction alone is enough to collapse the probability wave, which it absolutely isn't. Only measurement collapses the probability wave, and then only for the property for which the state of the particle is thus known. Thus in the double slit experiment you could interact with, and measure the particle, until the cows come home, you could knock the heck out of it, but if none of those measurements gives you which path information, then the interference pattern isn't going anywhere. The probability wave ain't going to collapse for just any old measurement, it's got to be specific. The particle somehow seems to know what you're measuring, but not only that, it also seems to know what you may indirectly learn from that measurement. So while the observer may not need to be a conscious one there appears to be more going on here than merely, I measure it, it collapses.

While it's true that we can easily build a photodetector to "see" the particle in question, in what way is this evidence that such a detector could collapse the probability wave? The last time that I checked, every biological photodetector is connected to a biological brain, and who's to say that it's not the latter that actually collapses the probability wave? Is there indisputable evidence that detection alone collapses the probability wave?

It does seem logical that there is no need for a conscious observer, but is there evidence? After all, this is science, right?

Yes, there is evidence because the position of the particle need not even be known. One need only configure a mechanism that distinguishes one slit from the other to lose the interference pattern, but no knowledge of where the particle is, need exist. So, no one can say a human observation is necessary.

 

[Ken G]

However, note the importance of the potentially different meanings of "collapse." Some might take that word to mean "put the system in a mixed state", i.e., "destroy correlations between different eigenstates of some operator." That is certainly what decoherence does, and it does it independently of the presence of any information in a conscious brain.

But that is not the meaning of "collapse" that is required by the measurement problem, because that is no problem at all. The second meaning, the important one, is "take the system from a state where it could have multiple values for a given measurement, and realize only one value." Note how that meaning of "collapse" is much trickier, and is very hard, or impossible, to achieve without a conscious brain.

Indeed, I would say there is an issue of where lies the burden of proof here. You hold that we should regard all occurrences as objectively independent of human perception until such a time that it can be demonstrated that the occurrence requires human perception. That is an impossible standard, because no experiment ever could look like "comes out A if no human perception is ever involved." On the other hand, since all physics experiments ever done did involve human perception, it is easy to argue that if all experiments have property A, then the burden of proof is on those who would claim that A is not an essential property of a physical outcome.

 

[marksesl]

I read you to mean that decoherence can happen independently of an observer, but just what state out of all possible states a wave "collapses" to requires an observer. That seems philosophically valid, and the rest certainly presents a pause for thought. If all experiments are done by a human, how can one ever know if human perception is not a catalyst for the outcomes. There can be no such thing as a specific event unless experienced by a mind.

 

[Nugatory]

If all experiments are done by a human, how can one ever know if human perception is not a catalyst for the outcomes.

We cannot, at least through the methods of empirical science.

However, the assumption that human perception is a necessary catalyst leads to a number of conclusions that most people find wildly unsatisfying. "Is the moon still there when nobody is looking?" tends to draw an automatic "yes of course" from most people most of the time.

 

[marksesl]

This is getting at exactly the crux of the issue that I'm exploring, and many people see a connection between wave-function collapse in the double slit experiment and the nature of reality around us. I take the position while we do not cause decoherence in the double slit experiment, the results are ambiguous (not settled) until known to a mind, which I believe is what Ken G was getting at. Reality is consummated in the mind. Would the moon or anything else really exist if there were no minds to perceive them at all, to distinguish those happenings from an infinite domain of other possible happenings? Even in the double slit, just what position a probability wave collapses to, out of infinite possibilities, may indeed require a mind to perceive it, even if the more general idea of decoherence does not. This gets really weird, I know, so I'm going to bed now. Thanks for your input.

 

[DennisN]

Observers in formal quantum mechanics need not to be human, biological nor what we call "alive" or "conscious". An instrument will do.

Something to think about:
If I did a quantum experiment and recorded it with camera (without watching it), and put the clip on youtube (something like this), I would surely not believe that the first human who saw the clip collapsed the wave function in my experiment. Youtube is cool, but not that cool .

 

[Ken G]

I read you to mean that decoherence can happen independently of an observer, but just what state out of all possible states a wave "collapses" to requires an observer. That seems philosophically valid, and the rest certainly presents a pause for thought. If all experiments are done by a human, how can one ever know if human perception is not a catalyst for the outcomes. There can be no such thing as a specific event unless experienced by a mind.

Right. For example, one take a many-worlds view, and say that a closed system begins and ends in a pure state, but if the closed system can generate consciousness, it may generate incoherent consciousnesses that each perceive "one universe", but they really only perceive the sector of the pure state that is mutually coherent with that consciousness. In a situation like that, which requires an understanding of consciousness to either accept or refute, we could certainly hold that the closed system evolves unitarily as per quantum mechanics, but that the perception of experimental outcomes is indeed a product of consciousness. That would be essentially Wigner's position that consciousness is responsible for wavefunction collapse, yet with zero "mystical" elements involved-- unless you count many-worlds quantum mechanics as already mystical.

 

[Ken G]

"Is the moon still there when nobody is looking?" tends to draw an automatic "yes of course" from most people most of the time.

Yet it is easy to argue that a universe with no minds in it has not the least reason to attribute any meaning to the phrase "the Moon." In a very real sense, "the Moon" does not exist in a universe like that, a universe like that is just what it is, with no need to describe itself or partition itself in any way.

 

[meBigGuy]

Yet it is easy to argue that a universe with no minds in it has not the least reason to attribute any meaning to the phrase "the Moon." In a very real sense, "the Moon" does not exist in a universe like that, a universe like that is just what it is, with no need to describe itself or partition itself in any way.

You need to read some basic philosophy. Try "The problems of philosophy" by Bertrand Russell. It is available on the web. It explains objectivism. What is "real" what can we "know", what "exists" independent of us?

How you can read the Schlosshauer paper and not relate to classical existence independent of consciousness is beyond me. It is well explained. Start at chapter VI. One can make up all these weird interpretaions, but they don't explain anything that needs explaining beyond people wanting to know why and making up answers (that can't be tested) to feel good about questions they can't answer (yet). What data do we have that isn't fully explaind by QM, a minimal interpretation, and dechoherence, but is explained by some exotic interpretation. None, obviously, or it wouldn't just be an interpretation.

As soon as you make consciousness the basis of existence or reality you are in mumbo-jumbo land. You can play that game but it will never get you anywhere meaningful.

 

[Ken G]

You need to read some basic philosophy. Try "The problems of philosophy" by Bertrand Russell. It is available on the web. It explains objectivism. What is "real" what can we "know", what "exists" independent of us?

I don't think you understand what I'm saying. There are (at least!) two very different meanings to the phrase "exists outside of us," yet people imagine that statement has a clear interpretation. It does not. The two very different meanings I have in mind are:
1) Exists outside of us in the sense that our presence would make no difference if we were a "fly on the wall." This is the usual meaning people give when they talk about trees falling in woods and making noise, but it is a meaning that has little to do with quantum mechanics. Quantum mechanics is about the difficulty of being a "fly on the wall", difficulties that philosophers who don't know quantum mechanics have most likely never even dreamed of.
2) Exists outside of us in the sense that the words could mean something without any intelligences to convey them meaning. When one recognizes this different meaning of "exists outside of us", the one relevant to quantum mechanics, one can see that even stating "a tree falls in the woods" has already begged the question of existence.

How you can read the Schlosshauer paper and not relate to classical existence independent of consciousness is beyond me.

I notice you did not answer the question I posed. Let's try it again: how does that paper account for the fact that I will perceive a single outcome if I do an experiment? Until you can answer that, you have not even scratched the surface of the "measurement problem."

As soon as you make consciousness the basis of existence or reality you are in mumbo-jumbo land. You can play that game but it will never get you anywhere meaningful.

To me, this just sounds like the "mumbo jumbo" objection, closely related to the "navel gazing" objection. They are fallback positions when the ability to answer the question that I just posed is found lacking. But don't shoot the messenger, the fact that we have a perfectly reasonable question on the table, and cannot answer it, is not a fact that can be ducked with the "mumbo jumbo" objection.

 

[Nugatory]

the fact that we have a perfectly reasonable question on the table, and cannot answer it, is not a fact that can be ducked with the "mumbo jumbo" objection.

It is a perfectly reasonable question, but like many perfectly reasonable philosophical questions, it cannot be answered with the methods of empirical science... So is maybe a bit out of scope for this forum.

 

[bhobba]

IIf all experiments are done by a human, how can one ever know if human perception is not a catalyst for the outcomes.

I have zero idea where you get the idea that QM is only about the outcome of experiments done by humans.

It's about any quantum process that leaves a mark here in the classical world - whether done in an experiment devised by humans or not.

I have posted before about the absurd world view you are led to if you don't do that - but even beyond such considerations I know of no textbook on QM that presents such a view - Ballentine for example certainly doesn't.

Thanks
Bill