Can a computer be an observer?

[Karl Coryat]

Hello all, I am writing an article on the definition of "observer" in a quantum mechanics context.

From what I know about QM, most consider that an inanimate apparatus and even individual particles can function as the "observer" in quantum measurements. I also know that the 1991 experiment by Mandel, Wang, and Zou pretty much established that the experimental setup is what determines whether quanta behave as particles or waves, not what the experimenter actually determines or chooses to learn.

However, I came across a 1996 talk by the late computer scientist Seymour Cray, who described a then-recent experiment which would seem to contradict this. A computer performed wave/particle duality experiments and stored the results in memory. This data then appeared to remain in macroscopic superposition until a human experimenter actually checked the files, thus suggesting (incredibly) that a machine cannot function as a quantum observer in the capacity that a human observer can.

I have spent literally hours searching Google Scholar (and regular Google) for this experiment, turning up absolutely nothing. I am starting to think that Cray was mistaken, that this experiment never happened.

http://americanhistory.si.edu/collections/comphist/montic/cray.htm" of Cray's talk -- he discusses the experiment in the sections "Wave/Particle Duality and Computers" and "Giving Meaning to Binary Data."

Whether you know anything about this experiment or not, any insight is greatly appreciated!

 

[cesiumfrog]

You see how confused I am now. I'm getting ready for the question and answer session, so if any of you can help me with this, I'd like to hear about it.

If anything, it sounds like your speaker may have misunderstood (and further mangled) something he had only heard indirectly, it's mentioned more as an aside than as something he had any personal expertise on. (Science journalism tends to be oversimplified and excessively sensationalised, and it's difficult for typical reporters themselves to get a subtle understanding of cutting edge work.)

 

[Fredrik]

I'm not familiar with the experiment you're talking about, but it sounds like this Cray guy has misunderstood something. It's certainly plausible that the memory was in a superposition for a while, but it certainly didn't have to interact with a human for the coherence of the superposition to be delocalized into the environment.

As for the title question, I'll just quote myself:

The "observer" doesn't have to be human. Any system that interacts significantly with its environment qualifies as an observer.

https://www.physicsforums.com/showthread.php?p=2589421

 

[hamster143]

It's certainly plausible that the memory was in a superposition for a while, but it certainly didn't have to interact with a human for the coherence of the superposition to be delocalized into the environment.

Depends on the interpretation.

 

[computerphys]

According to Wikipedia ( http://en.wikipedia.org/wiki/Observer_(quantum_physics) ) "In quantum mechanics, the observer and the system being observed became mysteriously linked so that the results of any observation seemed to be determined in part by actual choices made by the observer."

I am not sure if that is right or wrong, but supposing it is true, your question "Can a computer be an observer?" would be translated into "Can a computer make a choice?".

If not (and if Wikipedia is correct... and if my logic is also correct...), then Seymour Cray would be right.

 

[alxm]

That's a pretty poorly written Wikipedia article.

By the "choices of the observer", all they mean is which property the 'observer' chose to measure (if any). Which ultimately means:
- "What kind of interaction [between the "observing" system and "observed" system] occured, if any"?

Now if you're a human conducting an experiment where you, say, change polarization filters, then you could say
"your choice affects the observed system". But that's because you change the nature of the interaction.
Nothing to do with the choosing in itself, nothing to do whether or not you look at the results. Etc

 

[cesiumfrog]

It's certainly plausible that the memory was in a superposition for a while

That's being overly generous. In context, he's suggesting that megabytes of data stored in memory and disk (not to mention all the circuit activity) is in a macroscopic superposition (of completely different classical states) for an arbitrary number of hours, never decohering with the room-temperature environment, until the moment that the computer is directed to execute (using a chosen basis) an analysis routine on this archived data. His sense is not of some extreme minority interpretation (human consciousness mediated collapse), nor in the sense of something possible in principle and just astronomically low probability, but in the usual QM sense of repeatably preparing a system in a superposition. It's flat wrong.

 

[Fredrik]

That's being overly generous. In context, he's suggesting that megabytes of data stored in memory and disk (not to mention all the circuit activity) is in a macroscopic superposition (of completely different classical states) for an arbitrary number of hours, never decohering with the room-temperature environment, until the moment that the computer is directed to execute (using a chosen basis) an analysis routine on this archived data. His sense is not of some extreme minority interpretation (human consciousness mediated collapse), nor in the sense of something possible in principle and just astronomically low probability, but in the usual QM sense of repeatably preparing a system in a superposition. It's flat wrong.

OK. I didn't click the link in #1, so I assumed that we were talking about a few bits, and not necessarily a typical computer memory.

Depends on the interpretation.

Are you saying that there is an interpretation in which the "observer" needs to be human?

 

[Dmitry67]

Yes
http://en.wikipedia.org/wiki/Interp...rpretation:_consciousness_causes_the_collapse

In his monumental treatise The Mathematical Foundations of Quantum Mechanics, John von Neumann deeply analyzed the so-called measurement problem. He concluded that the entire physical universe could be made subject to the Schrödinger equation (the universal wave function). Since something "outside the calculation" was needed to collapse the wave function, von Neumann concluded that the collapse was caused by the consciousness of the experimenter.[21] This point of view was later more prominently expanded on by Eugene Wigner (see Quantum mind/body problem).

 

[Fredrik]

I know about that, but it would be hard to take that idea seriously even if we didn't know anything about decoherence, and now that we do I think it's impossible.

 

[GeorgCantor]

I know about that, but it would be hard to take that idea seriously even if we didn't know anything about decoherence, and now that we do I think it's impossible.

How does decoherence explain the transition from a mixed state to single outcomes? There still appears to be a MP, unless you believed in the MWI.

 

[Dmitry67]

In MWI no explanation is needed
In BM, hidden particles tag the outcome which is 'real'. Dont ask me why it becomes real :)

 

[Karl Coryat]

Thank you for all of the commentary. It is very helpful.

But this is something I find confusing about QM. On the one hand there is the mainstream decoherence camp, who insist that the phenomenon that Cray describes could not be possible. Then there is the (smaller) Von Neumann/Wigner camp, who I imagine would say those experimental results are not only possible, but probable, if computers are not "conscious." So, you would think that one of them would have tried this experiment, if in fact it has never been done the way Cray describes.

In other words it would seem that the described experiment would falsify or support Von Neumann/Wigner, and/or suggest that decoherence theory is incomplete. So I would be surprised if no one actually has tried it, to settle this significant dispute and make some headway in QM interpretation.

 

[computerphys]

...the described experiment would falsify or support Von Neumann/Wigner, and/or suggest that decoherence theory is incomplete. So I would be surprised if no one actually has tried it, to settle this significant dispute and make some headway in QM interpretation.

What difference would a human find between a superposition and a definite state of the computer data?

How can you know the state of data before a human measure?

Here, the question is "Can a computer be an observer?". In Schrödinger's cat, "Can a cat be a quantum observer?"

Where can I find a reliable definition of "quantum observer"?

 

[Fredrik]

In MWI no explanation is needed

This is not at all true. (I'll elaborate a little below).

How does decoherence explain the transition from a mixed state to single outcomes? There still appears to be a MP, unless you believed in the MWI.

I don't know what you mean by "MP". Decoherence explains the appearence of collapse in the framework of the MWI. The interaction between the system and its environment selects a preferred basis for the Hilbert space, and that basis also identifies the "worlds" in which stable records of the state of the system can exist. A memory in your brain is such a stable record, and there can be no consciousness without memories, so decoherence identifies the worlds which can contain conscious observers and they are the same as the worlds where the measurement had a well-defined result.

In the ensemble interpretation, there's no need, and no possibility, to describe the measurement process itself. State preparation and measurement are primitives, i.e. they're not defined in other terms, and they aren't (and can't) be described as dynamical processes.

This is the main reason why MWI intepretations are worth taking seriously. They can explain some things that the ensemble interpretation can't.

 

[Dmitry67]

Frederik, measurement problem has 2 components:
1. Why we don't observe macroscopic objects in superposition
2. What outcome becomes real, or is cat dead or alive?

By saying "no explanation is needed" I meant N2, not N1
N1 (decoherence) is in fact an important thing to study

 

[GeorgCantor]

Frederik, measurement problem has 2 components:
1. Why we don't observe macroscopic objects in superposition
2. What outcome becomes real, or is cat dead or alive?

By saying "no explanation is needed" I meant N2, not N1
N1 (decoherence) is in fact an important thing to study

The word "real" has been undefinied for at least a century. What does it even mean anymore? Even in a MWI setting, what does it mean for a system to be real? Had we been able to seal it off from the environement and keep it in superposition and avoid entanglement, would the Moon be still there? No. Is the Moon real when it's not decohered? No. Does the MWI look like a relational interpretation. I think it does as it sets a limit to what can be observed as soon as a entanglement with the environment takes place.

 

[GeorgCantor]

This is not at all true. (I'll elaborate a little below).


I don't know what you mean by "MP".

The measurement problem.

Decoherence explains the appearence of collapse in the framework of the MWI. The interaction between the system and its environment selects a preferred basis for the Hilbert space, and that basis also identifies the "worlds" in which stable records of the state of the system can exist. A memory in your brain is such a stable record, and there can be no consciousness without memories, so decoherence identifies the worlds which can contain conscious observers and they are the same as the worlds where the measurement had a well-defined result.

In the ensemble interpretation, there's no need, and no possibility, to describe the measurement process itself. State preparation and measurement are primitives, i.e. they're not defined in other terms, and they aren't (and can't) be described as dynamical processes.

This is the main reason why MWI intepretations are worth taking seriously. They can explain some things that the ensemble interpretation can't.

Yes, i agree with that. MWI is the interpretation that explains the most. And given the number of realities, it's nearly as powerful as "God did it".

 

[hamster143]

Frederik, measurement problem has 2 components:
1. Why we don't observe macroscopic objects in superposition
2. What outcome becomes real, or is cat dead or alive?

By saying "no explanation is needed" I meant N2, not N1
N1 (decoherence) is in fact an important thing to study

What does it mean to "observe macroscopic objects in superposition"?

For me, "to observe" means to form a memory of the outcome of the interaction. Whether I observe electrons or macroscopic objects, my memories are definite-valued. Even if there is a superposition of states of "me" formed at the moment of observation, I have no way of knowing that and I can devise no experiment to determine that.

 

[alxm]

Then there is the (smaller) Von Neumann/Wigner camp, who I imagine would say those experimental results are not only possible, but probable, if computers are not "conscious."

I don't think that's a correct representation of Wigner & Co's view; they were imagining consciousness emerging from quantum properties, not vice-versa.

But more importantly, Wigner abandoned his ideas when the process of decoherence began to clarify (in the 1970's IIRC). Since, "quantum consciousness" ideas have increasingly become the domain of crackpots and new age charlatans. I don't know of a single big-name physicist (or perhaps even reputable physicist) today who believes in quantum-consciousness ideas. (Roger Penrose is a gifted mathematician, but he is not a physicist) I do know several who've openly criticized it (Penrose's friend Stephen Hawking being one)

It's a dead idea. It was never more than speculation to begin with, and now it's debunked speculation.

 

[Karl Coryat]

I don't know of a single big-name physicist (or perhaps even reputable physicist) today who believes in quantum-consciousness ideas.

Well, there's http://en.wikipedia.org/wiki/Henry_Stapp" , who believes that consciousness is a series of measurements that the brain performs on its own superposed physical states (if I understand him correctly).

Are we certain that this question is closed? Or is it that some physicists just want to make it go away, because it's so messy?

 

[cesiumfrog]

Are we certain that this question is closed? Or is it that some physicists just want to make it go away, because it's so messy?

If you want to go ahead and write articles on this topic, we're not going to stop you. But it isn't mainstream physics. (Scientists love being wrong, but you don't learn science by rejecting the current expert consensus a priori.) You might as well write about psychic abilities, climate denial, homeopathy and perpetual motion machines. Decoherence isn't just some idly conjectured postmodern philosophical interpretation of the QM formalism, rather decoherence is grounded in the experimental science of actually checking what can and what cannot be maintained in coherent quantum superpositions. Big molecules tangled in a lump of wet meat at room temperature cannot. (That, and you ignore progress in neuroscience.)

 

[Fredrik]

David Mermin has also said that he believes consciousness to be relevant, and not in the passive way I described earlier. His "Ithaca" interpretation starts out sounding a lot like a many-worlds interpretation, but then he says that we can avoid the conclusion that there are many worlds because of...something that we still don't understand about consciousness! I think that's an absurd claim. It seems like he's just desperately trying to avoid the MWI while at the same time insisting that the correlations between subsystems are "real".

 

[Karl Coryat]

Cesiumfrog, I understand your point about perpetual motion machines, etc. However, I keep coming across statements where physicists talk of the "unease" that is commonly felt regarding QM, or admit their reservations about certain explanations at the deepest level.

Lee Smolin writes, "The question that comes up in these interpretations revolves around what actually causes the collapse of the quantum wavefunction...The principle of decoherence is, to many, the explanation -- interaction with the environment causes the quantum collapse. Even more significantly, physicists are able to solve the equations, perform experiments, and practice physics without resolving the questions of what exactly is happening at a fundamental level, and so most physicists don't want to get near these bizarre questions with a 20 foot pole."

Forgive me, but I just don't see this kind of fundamental unease happening in discussions about perpetual motion.

What I would like to understand is, how serious are these ongoing foundational questions within the legitimate physics community? Is it really just the crackpots and charlatans who bring them up, or does Smolin have a point? I hope I am not getting on people's nerves by raising this topic.

 

[cesiumfrog]

What I would like to understand is, how serious are these ongoing foundational questions within the legitimate physics community?

It might interest you to note that if you were to study the philosophy of science in any depth, you would discover genuine and profound "unease" regarding whether we can fundamentally define what is and is not science. And yet, most scientists are not troubled by (or even aware of) this.

The common unease about QM is merely that the true laws of the universe turn out to be quite foreign from the intuition that monkeys are born with. So what? I'm sure some people find the notion of a spherical Earth to be disturbing on some deep level.

The problem of consciousness is far more philosophical than scientific: you wouldn't take seriously anybody proposing that new emergent quantum physics is necessary to explain how a mosquito reacts to the stimuli it can sense.

 

[Fredrik]

The measurement problem.

Ah, in that case, I'll just add that the ensemble interpretation doesn't have a measurement problem...and this:

The "measurement problem" isn't even a consequence of QM. It's a consequence of assuming that both of the following statements are true (in addition to all the axioms of the standard version of QM):

a) A state vector is a mathematical representation of all the properties of a system.
b) A measurement has only one classical result.

So if the "measurement problem" bothers you, just stop insisting that both of the above must be true. Drop a) and you have the "ensemble interpretation". Drop b) and you have some sort of "many-worlds interpretation". Another option is to try to replace the standard version of QM with a theory that's equivalent to QM in the sense that it makes the same predictions, but is still a different theory because it's defined by different axioms. Such a theory may or may not have a "measurement problem".

Yes, i agree with that. MWI is the interpretation that explains the most. And given the number of realities, it's nearly as powerful as "God did it".

That's certainly unfair, and a bit silly. The claim that "God did it" makes no predictions whatsoever (or predicts that every statement that can be made is true), so it doesn't meet the requirements of a theory. The version of the MWI that I have in mind makes all the predictions of QM (which of course is the most successful theory in the history of science), and only a few more, which also agree with experiments. (This isn't enough to allow us to distinguish between the MWI and the ensemble interpretation, because those results are not in conflict with the ensemble intepretation). This version of the MWI is essentially just QM plus the assumption that the theory actually describes a physical system.

However, I keep coming across statements where physicists talk of the "unease" that is commonly felt regarding QM, or admit their reservations about certain explanations at the deepest level.

Statements like that always come from people who have made assumptions that they think have to be true, but that contradict each other, such as the two in my quote above.

Lee Smolin writes, "The question that comes up in these interpretations revolves around what actually causes the collapse of the quantum wavefunction...The principle of decoherence is, to many, the explanation -- interaction with the environment causes the quantum collapse. Even more significantly, physicists are able to solve the equations, perform experiments, and practice physics without resolving the questions of what exactly is happening at a fundamental level, and so most physicists don't want to get near these bizarre questions with a 20 foot pole."

Forgive me, but I just don't see this kind of fundamental unease happening in discussions about perpetual motion.

What I would like to understand is, how serious are these ongoing foundational questions within the legitimate physics community? Is it really just the crackpots and charlatans who bring them up, or does Smolin have a point? I hope I am not getting on people's nerves by raising this topic.

Smolin is making a good point when he says that QM works even if we don't know "what exactly is happening at a fundamental level". But isn't he the one who wrote about how QM means that fossils in the Earth doesn't imply that there were dinosaurs 100 million years ago? (A reference to a talk that I think was based on this paper). If that was him, he's confusing QM with theories that include additional assumptions that people make to be able to interpret the augmented theory as a description of what actually happens. Results like the one about the dinosaurs make people think they feel "fundamental unease" about QM, when in fact it isn't QM that's causing the unease. It's the additional assumptions.

 

[GeorgCantor]

Cesiumfrog, I understand your point about perpetual motion machines, etc. However, I keep coming across statements where physicists talk of the "unease" that is commonly felt regarding QM, or admit their reservations about certain explanations at the deepest level.

That's true and it all comes down to the old question of Einstein:

"Is the Moon there when nobody is observing?"


There is actually one interpretation that claims that the Moon is there all the time, no matter what, and it's based on hidden variables. Just not local hidden variables. Some claim the BM gives a picture of the 'universe' that's easier to imagine, but i sill fail to see how such a universe is imaginable. It was formulated so that people accommodate a worldview that would make them feel more comfortable, namely that things are there and are what they are all the time, whether anybody is measuring or not. But Bohm had a very different philosophy of how the universe operates(everything is one undivided wholeness), one that causes even more raised eyebrows than relational interpretations, MWI, etc. I don't think the BI is supported by more than 10% of the physics community however.

 

[Dmitry67]

Yes. in BM 'universe wavefunction' is exactly the same as in MWI. (In fact, wavefunction is the same in ANY non-collapse int. one can imagine). However, BM declares that only one (what we call in MWI 'branch') forms 'reality'.

BM claims that wavefunction is real, and yet it does not form reality without the particles - which are, in principle, undetectable. 'empty' waves behave exactly the same way the waves with particles behave - and yet they are not 'real'.

in BM the conception of reality is fuzzy - empty waves are 'half real' - they are real, because wavefunction is real in BM, but they are not real, because we just want to leave only one branch. This is even more evident if we imagine different BM theories with the different sorts of BM-particles. MWI is BM(0) with a single type of reality. BM(1) – standard BM with 1 sort of particles - has 2 flavors of reality: real and half-real (empty waves). BM(2) has 4 flavors of reality, BM(3) – 8 etc. For me, the cure BM tries to get rid of extra MWI worlds in a very ugly and contre-intuitive way.

 

[Demystifier]

... measurement problem has 2 components:
1. Why we don't observe macroscopic objects in superposition
2. What outcome becomes real, or is cat dead or alive?

Actually, 3 components:
3. Why the statistics of measurement outcomes obeys the Born rule?

MWI easily solves 1. and 2., but has serious problems with 3. On the other hand, BI (which some consider to be an "unnecessary" complication of MWI) easily solves 3. as well. In fact, for me the problem 3. is the main reason to introduce this "unnecessary"-complication-of-MWI called BI. If there was no problem 3. in MWI, I would be the first to admit that BI is indeed an unnecessary complication. But since the problem 3. is there, it seems that some complication of MWI is necessary, and that BI is the simplest such complication.

 

[Dmitry67]

Demistifier, I am just curious, how would you answer the question:
Are empty waves real?

 

[cesiumfrog]

Yes. in BM 'universe wavefunction' is exactly the same as in MWI. (In fact, wavefunction is the same in ANY non-collapse int. one can imagine). However, BM declares that only one (what we call in MWI 'branch') forms 'reality'. [...] empty waves are 'half real'

Thanks, that's a really forthright description. Would all proponents of BM find that uncontroversial?

MWI is BM(0) [...] BM(2) has 4 flavors of reality,

Could you point me to a reference on this?

 

[Dmitry67]

Could you point me to a reference on this?

It is just my argument against BM I used in coversation with Demystifier.

Imagine the generalization of BM with N sorts of particles. Say, N=2 and there are RED and BLUE particles. So we have:

1. Empty waves
2. Waves with BLUE particles only
3. Waves with RED particles only
4. Waves with both types of particles inside

What of above is real? The answer was obvious in BM(1), bot not in BM(2)

Finally, one can imagine an inverse BM(1) : only EMPTY waves are real. The question is, how can you tell the difference between BM(1) and iBM(1) ? My point is that BM uses physical axiom, which can't be reduced to formalas and can't be expressed mathematically. That axiom is 'only waves with particles are 'REAL''.

 

[Demystifier]

Demistifier, I am just curious, how would you answer the question:
Are empty waves real?

According to BI - yes, they are real.
But why then don't we see them? Because "we" are (made of) particles, so we are not there to see the empty waves.

Satisfied?

 

[Demystifier]

It is just my argument against BM I used in coversation with Demystifier.

Imagine the generalization of BM with N sorts of particles. Say, N=2 and there are RED and BLUE particles. So we have:

1. Empty waves
2. Waves with BLUE particles only
3. Waves with RED particles only
4. Waves with both types of particles inside

What of above is real? The answer was obvious in BM(1), bot not in BM(2)

Finally, one can imagine an inverse BM(1) : only EMPTY waves are real. The question is, how can you tell the difference between BM(1) and iBM(1) ? My point is that BM uses physical axiom, which can't be reduced to formalas and can't be expressed mathematically. That axiom is 'only waves with particles are 'REAL''.

In BI, there is no axiom that "only waves with particles are 'REAL'". Instead, BI explains why it APPEARS that empty waves are not real, even though they are real. In your variant 4., both blue and red particles would be real, but they would not mutually interact, so observers made of red particles would think that blue particles are not real and vice versa.

 

[Dmitry67]

According to BI - yes, they are real.
But why then don't we see them? Because "we" are (made of) particles, so we are not there to see the empty waves.

Satisfied?

partly :)
So if fact you admit that when I open a box and see a dead cat, then alive cat is real as well but just not detectable Note that it is different from a canonic way to explain BM :)

You say: Assuming that I am tagged (have particels inside) I can interact (observe) only tagged branches. So yes, according to BM, for tagged observer only tagged reality is real.

However, why observers are tagged in the first place? non-tagged observer can observe non-tagged reality as well as tagged observers can observe tagged reality.