I won't debate on the wavefunction collapse

[Mentz114]

Vanesch, your posts are well argued and backed by knowledge and understanding. But I have one question I'd like to raise,

But the fact that we don't see superpositions of macroscopic classical states isn't an argument. Quantum theory, by itself, is entirely capable of explaining WHY we don't see them.

Is quantum theory the first physical theory where things exist consistently within the theory ( probability amplitude, superposed states ..) but which, the theory itself tells us, cannot be observed ? Does this very idea not conflict with the notion of objective reality ?

 

[meopemuk]

Consider polarisation states. Is |45 degrees> a superposition of |90 degrees> and |0 degrees > ? Is "spin along X" not a superposition of "spin along z up" and "spin along z down" ? Isn't "short light pulse" not a superposition of "red light" "green light" "blue light", "yellow light" ... ?

Yes. "Spin along x" state in quantum mechanics is a superposition of "spin along z up" and "spin along z down" states. But what this superposition means for experiment? Does it mean that somehow we can see both "spin along z up" and "spin along z down" states simultaneously? No, not at all. That's not what quantum mechanics says. It says that if we perform measurements in the "spin along x" state, then sometimes we will find its "spin along z" up and sometimes the "spin along z" will be down. We will always find a definite value of "spin along z". However, it is impossible to say which of the two possible values (up or down) will be found each time. We can only predict the probabilities by taking squares of superposition coefficients.

This is what we see in experiment. Now you may say: "these experiments indicate that before the spin has been measured it was in a superposition state in which both "spin along z up" and "spin along z down" states coexisted together. Our measurement induced a random collapse of this superposition state to one of its components." This is Copenhagen interpretation. It tells you something about the system before the measurement was performed. So, you have a good reason to be sceptical about this statement, because it is impossible to verify it experimentally, even in principle.

Eugene.

 

[lightarrow]

I wanted to say that individual systems can be only found in a definite state (dead or alive; spin up or spin down) and they are never found in a complex superposition. The idea of quantum superposition is needed only when we try to describe an ensemble of identically prepared states and to "explain" why measurements in such an ensemble are not reproducible (sometimes we find a dead cat other times the cat is alive; sometimes we measure spin up and other times the spin is down).

Quantum mechanics doesn't say that it is possible to see half-dead half-alive cat. This superposition is a necessary and important ingredient of theory, but not something that can be directly observed. I hope I made myself clear now.

Eugene.

I agree with you and I would add: could a "measurement" have any meaning if it wouldn't be described by a well defined state?

 

[Hurkyl]

I agree with you and I would add: could a "measurement" have any meaning if it wouldn't be described by a well defined state?

Why not? I consider a CNOT gate to be the most basic example of a measuring device.

 

[lightarrow]

Why not? I consider a CNOT gate to be the most basic example of a measuring device.

Hello Hurkyl.
Sorry but I would ask you to explain better what exactly it's measured in that case.

 

[vanesch]

Is quantum theory the first physical theory where things exist consistently within the theory ( probability amplitude, superposed states ..) but which, the theory itself tells us, cannot be observed ?

No, the very first theory that did this was Newtonian mechanics, where the concept of force was introduced, but force cannot be observed, only displacement (and derivatives) can. It gave a lot of headaches to Newton himself, and his "action at a distance" which was so different than what one would intuitively think of a force, such as a push or a pull. So you could argue that Newtonian mechanics is just a tool which helps us calculate the observed displacements of objects, and uses a mathematical tool for that which is "force" (or one of its equivalents, potential energy, interaction, ...) but that this is nothing physical. We only observe displacements of things. Things don't really "interact" because we cannot observe directly their "interaction" but only their displacements, and it seems to be a convenient thing in the theory to talk about interactions (forces, or interaction potentials or whatever), but in "reality" the sun doesn't attract the Earth physically, and there is no force of gravity or electrical forces or whatever, all these are constructions of a theory. We only have that objects suffer displacements, period.
Newtonian theory tells us that forces cannot be observed. Only positions, velocities and accelerations can. The reason that a force cannot be observed in Newtonian physics, is that it suffers vector addition. As such, a force plus its opposite force has exactly the same consequence as no force at all. There's no "preferred decomposition" of zero force, and hence its components cannot have the slightest observable effect.

A stronger form of "non-observability", which prompted Einstein to his statement, is found in general relativity. The typical trap in doing GR is to take one's coordinate system too seriously, and to think that because one has introduced x,y,z,t, that these are things that can be observed. In GR, you cannot observe directly the spacetime manifold, and certainly not a coordinate system over it. You can only measure such things as eigentimes, and other intervals.

Does this very idea not conflict with the notion of objective reality ?

No, this idea conflicts with the notion of naive realism: that reality is what is observed, and that it is all there is to it.

 

[vanesch]

It tells you something about the system before the measurement was performed. So, you have a good reason to be sceptical about this statement, because it is impossible to verify it experimentally, even in principle.

This is true in a way. But the problem with that attitude is that, when applied to the extreme, you end up in solipsism. Indeed, the whole point of an ontological hypothesis is to have a claim of objective existence of something loose from observation. If you insist on only taking for real what is observation, then you run into the following problem: you cannot even take for real the banana on the table. Indeed, you don't observe "a banana", you only observe some yellow light in your eye. But then, you don't observe "yellow light in your eye", your brain only observes some nerve pulses coming from your eye.
You can say: yes but I can touch the banana. True. But how do you know that you touch the banana ? Because receptors under your finger skin send out nerve pulses to your brain. So you're not really observing the touch of the banana, your brain only observes nerve pulses from your fingers.
You can also eat the banana. You will taste it, but I can set up the same story, that the entire sensation of "eating a banana" is just the result of nerve pulses arriving in your brain. And then, do you really observe your brain ? Or do you assume you have a brain because you have read about it, and you've seen motion pictures or even bodies being dissected etc... ?

So why do you think that there really is a banana on the table ? Because it is a (small piece of) theory which is consistent with all your observations. You have the theory "there's a banana on the table" and that theory gives rise to some ontology (a REAL banana on a REAL table), and from it, you can deduce a lot of observations: if you touch it, you will feel something that feels like a banana ; if you look at it, you will see some yellow light ; if you eat it, it will taste like a banana etc...
So this theory "there's a banana on the table" explains very well, in a consistent way, many observations. As such, it is a good theoretical construct. But of course, it is only that: a construct. You could just as well argue that "banana's" are of course useful theoretical constructs, which are very useful to explain observations, but that they are only that, and don't, of course, correspond to a genuine ontology, because you can never verify experimentally that a banana really exists! You can only verify experimentally that you have subjective experiences which probably result from nerve pulses on something you probably have, which is a brain. But there's no way that the observation of these nerve pulses proves that there is really a banana on the table.

 

[Mentz114]

Vanesch:

No, this idea conflicts with the notion of naive realism: that reality is what is observed, and that it is all there is to it.

Agreed.

No, the very first theory that did this was Newtonian mechanics, where the concept of force was introduced, but force cannot be observed, only displacement (and derivatives) can.

I'd like to say that a quantity that can be calculated using experimental data is 'observable'. For instance by measuring times, distances and masses, I can in principle put a number on F=ma for a given situation.

Is it not the case that in principle there is no experiment that would allow me to calculate what the quantum phase was at the time of measurement ?

So why do you think that there really is a banana on the table ? Because it is a (small piece of) theory which is consistent with all your observations. You have the theory "there's a banana on the table" and that theory gives rise to some ontology (a REAL banana on a REAL table), and from it, you can deduce a lot of observations: if you touch it, you will feel something that feels like a banana ; if you look at it, you will see some yellow light ; if you eat it, it will taste like a banana etc...
So this theory "there's a banana on the table" explains very well, in a consistent way, many observations. As such, it is a good theoretical construct. But of course, it is only that: a construct. You could just as well argue that "banana's" are of course useful theoretical constructs, which are very useful to explain observations, but that they are only that, and don't, of course, correspond to a genuine ontology, because you can never verify experimentally that a banana really exists! You can only verify experimentally that you have subjective experiences which probably result from nerve pulses on something you probably have, which is a brain. But there's no way that the observation of these nerve pulses proves that there is really a banana on the table.

No, I don't like this. It is sophisticated ( in the worse sense) and goes against experience. There's an easy test for the reality/existence of a banana.

 

[Hurkyl]

Is it not the case that in principle there is no experiment that would allow me to calculate what the quantum phase was at the time of measurement ?

You can't get complete information, but you can get partial information.

No, I don't like this. It is sophisticated ( in the worse sense) and goes against experience.

That doesn't make it wrong.

There's an easy test for the reality/existence of a banana.

Please elaborate: propose an experiment that distinguishes between a real banana and the solipsist hypothesis.

 

[Mentz114]

Hurkyl:

You can't get complete information, but you can get partial information.

Ok. I'm doing some reading about this.

No, I don't like this. It is sophisticated ( in the worse sense) and goes against experience.

That doesn't make it wrong.

But it doesn't make it physics.

Please elaborate: propose an experiment that distinguishes between a real banana and the solipsist hypothesis.

You cannot be serious ! Solipsism is a ridiculous idea and not worth discussing.

My point is that QM is different from earlier theories, that's all, but we stray into pure philosophy. We have to accept that we can believe our senses and that there is an objective reality or physics has no meaning or purpose.

Another thing worth mentioning about the original topic 'wave function collapse' is that it is possible to start with the SE and correctly predict the dynamics of quantum particles without mentioning probablity amplitudes, Hilbert space or wave function collapse. So these things do not have the special position they seem to have assumed.

 

[Hurkyl]

But it doesn't make it physics.

If we were discussing physics, the discussion would have ended with "QM predicts superpositions, and is supported by empirical evidence". This has been a debate of metaphysics for quite some time.

You cannot be serious ! Solipsism is a ridiculous idea and not worth discussing.

You're the one who claimed that there was an easy test to tell the difference between the solipsist hypothesis and the real banana hypothesis.

 

[vanesch]

You cannot be serious ! Solipsism is a ridiculous idea and not worth discussing.

Mmmm... doesn't sound like a very convincing argument to me

My point is that QM is different from earlier theories, that's all, but we stray into pure philosophy. We have to accept that we can believe our senses and that there is an objective reality or physics has no meaning or purpose.

But that's the whole point ! This is a statement that one should adhere to naive realism and that it is forbidden to think of the consequences of not doing so. Rather dogmatic, no ?

Personally, I don't find solipsism "unworthy of discussion" ; I only find it not a very useful ontology hypothesis, because we stop immediately. Ontology should be useful (although, and that is what the possibility of solipsism illustrates, up to a point arbitrary and fixed by convention). As such, the question is: what's the most useful ontology hypothesis that we can make, that helps us "understand" (= give us an intuitive feeling) for our subjective observations ?

"there's a banana on the table" is a very useful ontology hypothesis in daily life.
When doing quantum mechanics, what we take as "real" (to help our intuition) may be set differently.

Another thing worth mentioning about the original topic 'wave function collapse' is that it is possible to start with the SE and correctly predict the dynamics of quantum particles without mentioning probablity amplitudes, Hilbert space or wave function collapse. So these things do not have the special position they seem to have assumed.

Ah ? We can use the SE without the entire hilbert space mechanism ?

 

[meopemuk]

If you insist on only taking for real what is observation, then you run into the following problem: you cannot even take for real the banana on the table.

This example makes clear that one needs to carefully draw the line between what is considered observable effects and what is "stuff" of theory. You vividly demonstrated that drawing this line inside our brain leads to solipsism. But it would be just as silly to draw this line in such a way that all theoretical stuff gets promoted to the rank of observable effects. Shall we say that forces are real? what about electromagnetic fields? wave functions? Hilbert spaces? quantum fields? curved space-time..?

I think that you provided an excellent example of the separation observable/theoretical in your previous post:

Newtonian theory tells us that forces cannot be observed. Only positions, velocities and accelerations can.

I basically agree with this Newtonian philosophy: Experiment tells us about certain observables of material particles (position, velocity, spin, momentum,...). This is what objectively exists. The rest of physics (forces, wave functions, Hilbert spaces...) is just mathematical stuff that we invented to describe these observations.

Eugene.

 

[Mentz114]

From: Wiki

Solopsism is the philosophical idea that "My mind is the only thing that I know exists". Solipsism is an epistemological or metaphysical position that knowledge of anything outside the mind is unjustified.

A weak form of epistemological solipsism states that the agent has no proof of anything beyond the senses. This can be raw observation, at the level of "I see red", "I am not aware of a proof". A stronger form states "No proof exists", this is falsifiable in as far as anything is. In order to falsify it, a proof must be provided.

This is what I understand by Solopsism and I admit I only vaguely see what is meant, while finding it offensively ilogical. I don't want to get into metaphysics, not my bag.

You're the one who claimed that there was an easy test to tell the difference between the solipsist hypothesis and the real banana hypothesis.

No, you misunderstand me. I would not try and answer any question that relates to or involves Solopsism any more than count angels on pinheads.

Quote:
Originally Posted by Mentz114

You cannot be serious ! Solipsism is a ridiculous idea and not worth discussing.

Mmmm... doesn't sound like a very convincing argument to me.

You are right, of course. May I amend it to

"Solipsism is a ridiculous idea and not worth discussing further as a useful ontological model"

This is a statement that one should adhere to naive realism and that it is forbidden to think of the consequences of not doing so. Rather dogmatic, no ?

If Maxwell, Planck, Lorentz and Einstein were 'naive' realists, I'm proud to be counted in their company. What's wrong with being dogmatic in keeping physics grounded in experiment ?

Ah ? We can use the SE without the entire hilbert space mechanism ?

Well, yes. It's De Broglie-Bohm theory.

I appreciate the time and trouble you've taken to put me right on the metaphysics, it has been stimulating and edifying. Now it's back to the equations (reality !) for me.

 

[rewebster]

I basically agree with this Newtonian philosophy: Experiment tells us about certain observables of material particles (position, velocity, spin, momentum,...). This is what objectively exists. The rest of physics (forces, wave functions, Hilbert spaces...) is just mathematical stuff that we invented to describe these observations.

Eugene.

let's see---do I agree with Newtonian/Galilean philosophy, or is it that my philosophy is Newtonian/Galilean physics?

 

[Hurkyl]

This example makes clear that one needs to carefully draw the line between what is considered observable effects and what is "stuff" of theory. You vividly demonstrated that drawing this line inside our brain leads to solipsism. But it would be just as silly to draw this line in such a way that all theoretical stuff gets promoted to the rank of observable effects. Shall we say that forces are real? what about electromagnetic fields? wave functions? Hilbert spaces? quantum fields? curved space-time..?

It is because of this arbitrariness that I think all this talk of "existing" and "reality" is usually just an expression of cognative bias.

This is what I understand by Solopsism and I admit I only vaguely see what is meant, while finding it offensively ilogical. I don't want to get into metaphysics, not my bag.

If you don't want to get into metaphysics, then you shouldn't jump into a discussion about what is "real" and what isn't. It is physics to postulate entities that explain experiment, but it is metaphysics to postulate about the "reality" of those entities.

It's amazing how much QM inspires people to turn into metaphysicists. I suppose SR did the same thing, though.

 

[Count Iblis]

Another thing worth mentioning about the original topic 'wave function collapse' is that it is possible to start with the SE and correctly predict the dynamics of quantum particles without mentioning probablity amplitudes, Hilbert space or wave function collapse. So these things do not have the special position they seem to have assumed.

Thought experiment: A classical computer simulates an observer measuring the z-component of the spin of an electron in state $$\frac{1}{\sqrt{3}}\left |\uparrow\right\rangle+\frac{\sqrt{2}}{\sqrt{3}}\left |\downarrow\right\rangle$$

The computer numerically computes the time evolution of the many particle wave function of the observer as it interacts with the electron. Assuming functional artificial intelligence (the idea that an exact simulation of an observer generates a real conscious observer), what is the probability that the spin is found in the state $$\left |\uparrow\right\rangle$$ ?

 

[meopemuk]

Another thing worth mentioning about the original topic 'wave function collapse' is that it is possible to start with the SE and correctly predict the dynamics of quantum particles without mentioning probablity amplitudes, Hilbert space or wave function collapse. So these things do not have the special position they seem to have assumed.

I disagree completely. Schroedinger equation allows you to calculate the wave function. However, the wave function does not "correctly predict the dynamics of quantum particles". It tells us only probabilities of this or that outcome. However, in real measurement only one outcome gets realized out of the whole range of possibilities. Nobody can tell (SE certainly doesn't tell that) which outcome will be realized. This choice is completely random and unpredictable. So, you cannot avoid talking about probability amplitudes and wave function collapse in quantum mechanics.

Eugene.

 

[Mentz114]

I disagree completely. Schroedinger equation allows you to calculate the wave function. However, the wave function does not "correctly predict the dynamics of quantum particles". It tells us only probabilities of this or that outcome. However, in real measurement only one outcome gets realized out of the whole range of possibilities. Nobody can tell (SE certainly doesn't tell that) which outcome will be realized. This choice is completely random and unpredictable. So, you cannot avoid talking about probability amplitudes and wave function collapse in quantum mechanics.

Eugene.

Hi Eugene,
my understanding of the dB-B theory is that given the initial conditions, the trajectory of the particle is determined. But the initial conditions ( like quantum phase) cannot be known completely. The theory is thus interpreted ensemble-wise. Given a distribution of initial conditions, the outcome is found to be the same as with Copenhagen. No probablitiy amplitudes, no imaginary numbers.

I myself don't have any preference for one or another interpretation of the SE as long as experiments are not contradicted.

 

[meopemuk]

Hi Eugene,
my understanding of the dB-B theory is that given the initial conditions, the trajectory of the particle is determined. But the initial conditions ( like quantum phase) cannot be known completely. The theory is thus interpreted ensemble-wise. Given a distribution of initial conditions, the outcome is found to be the same as with Copenhagen. No probablitiy amplitudes, no imaginary numbers.

I myself don't have any preference for one or another interpretation of the SE as long as experiments are not contradicted.

The fact is that experimental outcomes in quantum physics are random. No theory can predict them. Various "interpretations" try to "explain" this unpleasant fact and make it easier to swallow.

One interpretation says that the wave function collapses upon interaction with the measuring apparatus. Another interpretation says that this interaction creates a whole new world. Yet another interpretation says that the randomness occurs because of uncontrolled initial conditions. There are dozens of ways to invent excuses for our ignorance about nature's behavior, but none of them can go around the simple fact that measurements are random and unpredictable. I think that the easiest and the most honest "interpretation" is to say that we simply don't know the reason of this randomness, then shut up and calculate the probabilities.

Eugene.

 

[Mentz114]

Eugene, I agree with everything in your post #55 above. My point is that one can get a statistical interpretation without the mathematics of probability amplitudes, at least for simpler problems.

M

 

[meopemuk]

Eugene, I agree with everything in your post #55 above. My point is that one can get a statistical interpretation without the mathematics of probability amplitudes, at least for simpler problems.

I am sorry, I probably misinterpreted what you said. I am not intimately familiar with dB-B approach (is it de Broglie - Bohm?). So, you are saying that it can describe two-slit interference without adding at some point two complex numbers (the amplitudes for passing through the left and right slit) and taking their square? Interesting.

Eugene.

 

[Count Iblis]

I think that all that can be done is dropping a few postulates. Wavefunction collapse is, of course, a nonsensical postulate or a theory claiming to be a fundamental theory rather than just a phenomenological description of Nature.

The Born rule can be derived from a much weaker postulate. All you need is a postulate that says that if the wavefuntion is in an eigenstate of an observable, then measuring it will yield the corresponding eigenvalue with probability 1.

 

[vanesch]

The fact is that experimental outcomes in quantum physics are random. No theory can predict them. Various "interpretations" try to "explain" this unpleasant fact and make it easier to swallow.

This is in fact not the difficulty at all. The difficulty resides in the fact that quantum theory as it is usually formulated, needs an arbitrary "transition point" (also called, the Heisenberg cut) where the fundamental dynamical rules *completely change*. The Schroedinger equation stops working, amplitudes give rise to probabilities when the a priori vectorial quantity (the state vector) gets components "with a meaning", the so-called preferred-basis problem. There is no way to give, in this picture, a *physical explanation* of the functioning of a measurement device, and hence the measurement basis in which one has to expand the state vector in order to transform it in a list of probabilities. There is no fundamental way to explain why a "position measurement" apparatus actually measures positions! At a certain point, you have to decide that quantum theory is no longer working "the usual way", that the system (apparatus ...) no longer has a quantum description (state vector), and that you are in "the classical domain".
So you have to decide then that quantum theory is NOT applicable to certain systems, although they are build up of atoms and particles and all that which ARE described by quantum theory. This is very well possible, but this is not an "interpretation" of a theory, it is a modification of its applicability domain. And no-one succeeded in writing down a sensible way in which this transition might occur, without inducing a lot of other problems.

THIS is the difficulty ; not so much that nature is or isn't random at a certain level.

 

[meopemuk]

This is in fact not the difficulty at all. The difficulty resides in the fact that quantum theory as it is usually formulated, needs an arbitrary "transition point" (also called, the Heisenberg cut) where the fundamental dynamical rules *completely change*. The Schroedinger equation stops working, amplitudes give rise to probabilities when the a priori vectorial quantity (the state vector) gets components "with a meaning", the so-called preferred-basis problem.

I don't see it as a difficulty at all. It appears as a difficulty only for those who want to see quantum mechanics (or whatever theory of nature they have in mind) as a complete and comprehensive description of the world, which encompasses everything including physical systems, measuring devices, minds of observers, whole universe, etc. In my opinion, it is too grandiose vision of the role of physics, and all paradoxes of quantum mechanics are clear evidence (for me) that this is not what theoretical physics is about.

I believe that the role of theoretical physics is much more modest. Its goal is to describe and predict observations made on physical systems by measuring apparatuses. The idea is to produce numbers which can be compared with results of experiments. In each well-defined experiment there is a clear separation between the measuring apparatus and the observed physical system. Only the physical system needs to be described by the wave function. There is no need for a description of the measuring apparatus, neither quantum nor classical description. Actually, the wave function of the system already takes into account the kind of measuring device that is used in the experiment. For example, when we write the wave fuinction of an electron in the position representation we already use the fact that the electron is observed by a device measuring position.

Of course, if we like, we can decide to include the measuring apparatus as a part of the physical system and shift the "Heisenberg cut". But then we will be speaking about a different experimental setup, whose description requires a completely different wave function.

In my opinion, this understanding of the limited role of physics is the most important lesson of quantum mechanics. We could have dreamed about precise and comprehensive description of the whole world in the days of classical Laplacian determinism. After the discovery of quantum mechanics, we should forget such dreams.

Eugene.

 

[vanesch]

The Born rule can be derived from a much weaker postulate. All you need is a postulate that says that if the wavefuntion is in an eigenstate of an observable, then measuring it will yield the corresponding eigenvalue with probability 1.

But the difficulty resides in the following: suppose I give you the physical description of an apparatus. What's the physical description of an apparatus ? I would say, a beginning state vector and a corresponding hamiltonian, no ? That's the quantum mechanical description of an apparatus. Now, given that, and only that, how do you deduce WHICH KIND OF MEASUREMENT BASIS goes with that apparatus ?

One can in fact do that, by fully assuming quantum theory all the way. One will then see that the interaction hamiltonian is such, that certain subspaces of states of the "system-under-study" couple with subspaces of states of the "measurement apparatus" in a kind of coarse-grained Schmidt decomposition of the overall state. Furthermore, if one introduces a quantum description of the environment (thermodynamical heat reservoir), usually, these couples ("subspace of states of measurement apparatus" and "subspace of states of system-under-study") remain stable against interaction with the environment ; again, by assuming that all this has a quantum description and that we don't leave the quantum-domain or the schroedinger equation. If one calls these subspaces of states of the measurement apparatus "the pointer states", then one can see that they are close to "classical states with different outcomes".
The whole above story is called "decoherence" and singles out specific subspaces of states of "macroscopic systems" which remain stable against interaction with the environment. From it, one can derive as such the "stable pointer states" and from this, and the interaction hamiltonian, one can then derive the "measurement basis" that the apparatus applies to the system.
But in order to do all this, we cannot collapse the wavefunction and we have to assume that quantum interactions and state descriptions are valid all the way up.

There is of course a way out, and that is by saying that there IS a preferred measurement basis, which is "position measurement". All measurements are then position measurements. When you do that, you can arrive at Bohmian mechanics, but the interpretational issues of Bohmian mechanics are not as simple as one might think at first, because the Bohmian ontology consists of two interacting worlds: the particle/position world (which we are used to from Newton), and on the other hand the quantum-mechancal wavefunction world which continues to evolve with superpositions and all that, just as in non-collapsing MWI quantum theory. This last world influences the former (the particle world), but not vice versa: the particle positions have no influence on the wave world. The problem as many people see it with Bohmian mechanics (except, of course, proponents of this view), is that the interaction between the wave world and the particle world is not relativistically invariant.

 

[Hurkyl]

In my opinion, it is too grandiose vision of the role of physics,

But I'm sure you understand that some people don't give up so easily.

and all paradoxes of quantum mechanics are clear evidence (for me) that this is not what theoretical physics is about.

By the way, to the best of my knowledge there are no paradoxes in quantum mechanics -- only pseudoparadoxes. Much like the twin pseudoparadox of special relativity, you only run into problems if you make unwarranted assumptions.

I believe that the role of theoretical physics is much more modest. Its goal is to describe and predict observations made on physical systems by measuring apparatuses.
...
There is no need for a description of the measuring apparatus, neither quantum nor classical description.
...
Of course, if we like, we can decide to include the measuring apparatus as a part of the physical system

You contradict yourself.

Actually, the wave function of the system already takes into account the kind of measuring device that is used in the experiment. For example, when we write the wave fuinction of an electron in the position representation we already use the fact that the electron is observed by a device measuring position.

It's not the wavefunction that takes that into account. The wavefunction has no idea what we're measuring. It's the choice of representation that takes into account what we're measuring.

In my opinion, this understanding of the limited role of physics is the most important lesson of quantum mechanics. We could have dreamed about precise and comprehensive description of the whole world in the days of classical Laplacian determinism. After the discovery of quantum mechanics, we should forget such dreams.

Eugene.

I take a wholly different lesson -- quantum mechanics emphasizes, more so than any other theory, the need to stick to experimentally meaningful questions when studying physics.

 

[vanesch]

I don't see it as a difficulty at all. It appears as a difficulty only for those who want to see quantum mechanics (or whatever theory of nature they have in mind) as a complete and comprehensive description of the world, which encompasses everything including physical systems, measuring devices, minds of observers, whole universe, etc. In my opinion, it is too grandiose vision of the role of physics, and all paradoxes of quantum mechanics are clear evidence (for me) that this is not what theoretical physics is about.

In a way I agree with you: I think one should remain modest and most probably our understanding of nature is still very remote from what is needed to get an overall universal picture. So all this is "running with the legs we have". However, I extend "theoretical physics" to "there's a banana on the table" too: it is a primitive form of theoretical physics!

So this is my point: once we have learned from quantum mechanics (and we already got a warning shot from relativity) that "theoretical physics" (in other words, trying to make sense of our observations) is not going to be a smooth ride in understanding the meaning of life, the universe and everything, and given that all intellectual activity related to observations is, in one way or another "theoretical physics", then we can only conclude that we don't know ANYTHING about the world and never will. We then realize that we don't even know whether that part of "theoretical physics from kindergarten" which tells us that there is a banana on the table has any sense.

So, or we try to make sense, in as much as we can, about all of it, or we simply say that nothing we ever deduced from observations has any sense, and we can just as well dwell in mysticism. Most people seem to prefer to place some "sane cut" somewhere, between "what's obviously true" on one hand, and what's "just theoretical constructs" on the other hand, but any such cut runs sooner or later in difficulties of logic, because of the grey zone.

I believe that the role of theoretical physics is much more modest. Its goal is to describe and predict observations made on physical systems by measuring apparatuses. The idea is to produce numbers which can be compared with results of experiments. In each well-defined experiment there is a clear separation between the measuring apparatus and the observed physical system.

That's where the difficulty resides: the clear separation between the measurement apparatus and the observed system. Even von Neumann realized this already in 1932. It is especially important for an instrumentalist, who is to study the physical interaction between the measurement apparatus and the system under study! For the instrumentalist, the "system under study" is the measurement apparatus. It's maybe because I'm an instrumentalist, that I take on this stance, btw

Only the physical system needs to be described by the wave function. There is no need for a description of the measuring apparatus, neither quantum nor classical description. Actually, the wave function of the system already takes into account the kind of measuring device that is used in the experiment. For example, when we write the wave fuinction of an electron in the position representation we already use the fact that the electron is observed by a device measuring position.

Again, except if you want to study the physics of a measurement apparatus!

Of course, if we like, we can decide to include the measuring apparatus as a part of the physical system and shift the "Heisenberg cut". But then we will be speaking about a different experimental setup, whose description requires a completely different wave function.

Right, and where do we stop ? With the "ultimate observation", which is nothing else but subjective experience. Even that, von Neumann realized.

In my opinion, this understanding of the limited role of physics is the most important lesson of quantum mechanics. We could have dreamed about precise and comprehensive description of the whole world in the days of classical Laplacian determinism. After the discovery of quantum mechanics, we should forget such dreams.

I agree with you. But I take this even further: the limited role of ANY knowledge and not only of physics. Quantum mechanics, IMO, showed us up to what point things that we thought we knew, are in fact up to a certain level, imaginary constructions of the mind. It depends or not whether you want to keep a "sanity cut" and keep insisting on the reality of the banana on the table (taking granted the logical difficulties that this will induce), or whether you're willing to just throw up the arms in the air and say that in the end, we don't know ANYTHING for sure (partly my position), or whether you say that whatever strange things your most complete formal knowledge gives you, it must somehow more be related to any "reality" than whatever intuitive idea you had of it (also partly my position).

So I'm in a kind of quantum superposition of the three views on reality and I won't collapse in either.

 

[jostpuur]

By the way, to the best of my knowledge there are no paradoxes in quantum mechanics -- only pseudoparadoxes. Much like the twin pseudoparadox of special relativity, you only run into problems if you make unwarranted assumptions.

Seems redundant terminology. Aren't all paradoxes pseudoparadoxes?

 

[Count Iblis]

Vanesch, that's also how I think about this issue. I.m.o. MWI + decoherence is the most natural thing to assume. If one proposes a real collapse of the wave function then there should be experimental evidence for that to motivate this. One has to demomonstrate that the time evolution of an isolated system is not exactly unitary. Some time ago I read about a proposal to look for such effects in observatons of neutrinos from astrophysical sources. Neutrino oscillations lead to neutrinos of one flavor evolving into a superposition of the three flavors. But if a pure neutrino state evolves into a mixed state then that can be detected as it affects the relative probabilities for detecting the three flavors.

 

[meopemuk]

It's not the wavefunction that takes that into account. The wavefunction has no idea what we're measuring. It's the choice of representation that takes into account what we're measuring.

Yes, I agree.

I take a wholly different lesson -- quantum mechanics emphasizes, more so than any other theory, the need to stick to experimentally meaningful questions when studying physics.

Yes, that's what I wanted to say.

Eugene.

 

[meopemuk]

then we can only conclude that we don't know ANYTHING about the world and never will.
[...] or we simply say that nothing we ever deduced from observations has any sense, and we can just as well dwell in mysticism.

Why such pessimism? We do know something don't we?

Right, and where do we stop ? With the "ultimate observation", which is nothing else but subjective experience. Even that, von Neumann realized.

The place to stop is determined by the concrete experimental situation. Ask the experimentalist conducting the experiment where is the boundary physical system/measuring device and you'll get a pretty accurate answer.

Eugene.

 

[Hurkyl]

Seems redundant terminology. Aren't all paradoxes pseudoparadoxes?

No. In a pseudoparadox, the contradiction is the fault of the arguer; typically he makes a subtle mistake or unwarranted assumptions. In a paradox, the contradiction is the fault of the system.

Examples of actual paradoxes are the Liar's paradox of 'naive' formal logic and Russell's paradox of 'naive' set theory.

 

[Count Iblis]

Hmmm, is there always something "naive" about a system allowing for a paradox? What about the Banach-Tarski Paradox

 

[Count Iblis]

Actually, the Banach-Tarski could be called a pseudoparadox...