Is the Wave Function Real? Evidence from the Frankenstein Photon Experiment

[yoda jedi]

This is from another thread, yes? How could I have known what your view is? I would be interested to see why you believe that QM is incomplete and a limitation of our knowledge, rather than a reflection of nature. Spectracat refutes that very nicely I think.

refuted ? very far to be settled...

 

[RUTA]

A name I made up to describe a particle consisting of half of Alice plus half of Bob. You wouldn't really expect to be able to create one of these. But I say you can. You can split Alice into 2 halves (an H> and a V> component) and recombine them to recreate the original Alice. Ditto for Bob. So I say you can swap identical components as well. The big caveat being that you must NOT be able to tell which is which.

Looking at your set up, I don't see how the outcomes for Chris and Dale will differ at all from those of Alice and Bob, given the location of Chris and Dale's measurement devices. If they move the devices so as to intercept the crossing routes and thereby measure pure |V> or |H> components, then of course you'd get something different. But, that would be a different set up and therefore be represented by a different wave function.

I fail to see why this set up in any way entails wave function realism. I read the caption to the picture, but the conclusion doesn't follow necessarily from the premises. You write,

"These statistics should not be possible unless the Wave Function is itself real. That is because Chris and Dale were created by combining Wave Function components."

I think you're tacitly assuming the reality of the wave function when you talk about "splitting it" and "recombining it." But, I can compute the same wave function for Chris and Dale by simply following the algorithm using the particular placement and orientation of the specific types of equipment in the set up without any reference to a wave function "doing" this or that through the devices.

 

[DrChinese]

Looking at your set up, I don't see how the outcomes for Chris and Dale will differ at all from those of Alice and Bob, given the location of Chris and Dale's measurement devices...

I fail to see why this set up in any way entails wave function realism. I read the caption to the picture, but the conclusion doesn't follow necessarily from the premises. You write,

"These statistics should not be possible unless the Wave Function is itself real. That is because Chris and Dale were created by combining Wave Function components."

I think you're tacitly assuming the reality of the wave function when you talk about "splitting it" and "recombining it." But, I can compute the same wave function for Chris and Dale by simply following the algorithm using the particular placement and orientation of the specific types of equipment in the set up without any reference to a wave function "doing" this or that through the devices.

Maybe you are right, and this is sort of what I am asking. I am saying that if you split a wave function into 2 components, those components can be mixed and matched with components from other particles. In the "right" circumstances, you could create an entangled photon from these components.

Now, my argument is: it is not JUST that we are creating the 100% probability of there being a photon at Chris and another at Dale - no question about that. You would expect that by basic logic. The question is: can you have a superposition formed from DIFFERENT particles? You seem to agree that this is feasible, but not surprising. And yet, I don't recall ever seeing a reference saying as much (that it is feasible). (You do see comments from time to time, on the other hand, that particles do not interfere with each other, but I do not think this is rigorously true.)

Assuming it is feasible: how can you combine wave functions - i.e. potentials - to form a single particle - which acts as if it is in a superposition of wave states - unless the underlying wave functions are real, physical entities? I.e. more real than simple "knowledge" of the system.

 

[RUTA]

Maybe you are right, and this is sort of what I am asking. I am saying that if you split a wave function into 2 components, those components can be mixed and matched with components from other particles. In the "right" circumstances, you could create an entangled photon from these components.

Now, my argument is: it is not JUST that we are creating the 100% probability of there being a photon at Chris and another at Dale - no question about that. You would expect that by basic logic. The question is: can you have a superposition formed from DIFFERENT particles? You seem to agree that this is feasible, but not surprising. And yet, I don't recall ever seeing a reference saying as much (that it is feasible). (You do see comments from time to time, on the other hand, that particles do not interfere with each other, but I do not think this is rigorously true.)

Assuming it is feasible: how can you combine wave functions - i.e. potentials - to form a single particle - which acts as if it is in a superposition of wave states - unless the underlying wave functions are real, physical entities? I.e. more real than simple "knowledge" of the system.

What do you mean by "split the wave function" and "match components from other particles?" If you wanted to "mix a photon with an electron" what would that mean? You have two different sources and two different detectors. So, you're already thinking in a realist fashion about "entities" (wave functions and particles) moving through the experimental equipment. Imagine yourself constructing the experiment. Do you EVER watch "screened off" entities moving through the apparatus? By definition, no. All you can observe are the pieces of equipment and the experimental outcomes. And, you can construct the wave function for the experiment using only information about what you observe. What are the pieces of equipment? Where are they located relative to one another in space? How are they oriented relative to each other? What is the relative temporal sequence of observed events, e.g., when do you turn on source, rotate a mirror, observe a detection event? You construct the wave function using answers to questions like these, so replace your wording about "splitting the wave function" and "matching components" with wording about the experimental set up and you can say goodbye to your "real" quantum entities :-)

 

[DrChinese]

What do you mean by "split the wave function" and "match components from other particles?" If you wanted to "mix a photon with an electron" what would that mean? You have two different sources and two different detectors. So, you're already thinking in a realist fashion about "entities" (wave functions and particles) moving through the experimental equipment. Imagine yourself constructing the experiment. Do you EVER watch "screened off" entities moving through the apparatus? By definition, no. All you can observe are the pieces of equipment and the experimental outcomes. And, you can construct the wave function for the experiment using only information about what you observe. What are the pieces of equipment? Where are they located relative to one another in space? How are they oriented relative to each other? What is the relative temporal sequence of observed events, e.g., when do you turn on source, rotate a mirror, observe a detection event? You construct the wave function using answers to questions like these, so replace your wording about "splitting the wave function" and "matching components" with wording about the experimental set up and you can say goodbye to your "real" quantum entities :-)

I agree with your perspective about context of experiment. What I am questioning is the issue about components, even in a relational format. I would say that all components of the wave functions contribute to the outcome, even when classical logic would say that only the "selected" component did. In other words, classical logic says that Chris and Dale are only ever composed of Alice OR Bob. But I say they are composed of pieces of both Alice AND Bob. Otherwise, they wouldn't be entangled when the component probabilitie amplitudes are from different particles.

I would think you, of all people, would agree with that assessment. But I am not sure it is a unique or surprising result. As I mention, I haven't seen references on it.

 

[RUTA]

I agree with your perspective about context of experiment. What I am questioning is the issue about components, even in a relational format. I would say that all components of the wave functions contribute to the outcome, even when classical logic would say that only the "selected" component did.

What do you mean by "contribute?" Formally, all the components contribute to the unique specification of the wave function, so you certainly don't mean "contribute" in the mathematical sense. Are you making the ontic case for Hardy's "half-empty waves" as in the context of interaction-free measurement (IFM)?

 

[Demystifier]

DrChinese, if I understood you correctly, you suggest that your (proposed) experiment gives STRONGER evidence (for wave-function reality) than other existing experiments. If I am right, it suggests that you CAN imagine how other experiments could be true without reality of the wave function, but you just cannot imagine how your particular experiment could be that.
If this is so, could you explain how the standard double-slit experiment could be true without the wave function being real? If you could explain that, I believe it would be much easier to answer your questions on this thread.

 

[RUTA]

DrChinese, if I understood you correctly, you suggest that your (proposed) experiment gives STRONGER evidence (for wave-function reality) than other existing experiments. If I am right, it suggests that you CAN imagine how other experiments could be true without reality of the wave function, but you just cannot imagine how your particular experiment could be that.
If this is so, could you explain how the standard double-slit experiment could be true without the wave function being real? If you could explain that, I believe it would be much easier to answer your questions on this thread.

There is a detailed explanation of the twin-slit experiment without invoking any "quantum entities" (waves, particles, or anything else) in “Reconciling Spacetime and the Quantum: Relational Blockworld and the Quantum Liar Paradox,” W.M. Stuckey, Michael Silberstein & Michael Cifone, Foundations of Physics 38, No. 4, 348 – 383 (2008), quant-ph/0510090. We have also done it using path integrals over graphs, see section 3.4 of arXiv 0908.4348.

 

[LostConjugate]

Can the lack of any angular momentum in the hydrogen atom ground state be explained without the wave function?

 

[RUTA]

Can the lack of any angular momentum in the hydrogen atom ground state be explained without the wave function?

We need wave functions (or path integrals) to generate the probability amplitude, so I assume that "without the wave function" is an ontic statement. The answer would probably be "no" because what is meant by "a hydrogen atom in its ground state" entails a wave function (or equivalent) perspective (model). If you rather ask about the nature of "sources" and the measurement of quantum angular momentum per experimental observations, then, yes, one can avoid wave function realism. Essentially, you have to define what you mean by "a hydrogen atom in its ground state" via experimental observations (which is true of all physics, but most such identification is tacit).

 

[LostConjugate]

We need wave functions (or path integrals) to generate the probability amplitude, so I assume that "without the wave function" is an ontic statement. The answer would probably be "no" because what is meant by "a hydrogen atom in its ground state" entails a wave function (or equivalent) perspective (model). If you rather ask about the nature of "sources" and the measurement of quantum angular momentum per experimental observations, then, yes, one can avoid wave function realism. Essentially, you have to define what you mean by "a hydrogen atom in its ground state" via experimental observations (which is true of all physics, but most such identification is tacit).

Instead of saying in it's ground state what if I just said how can a particle be some distance r from the nucleus with no angular momentum if it were not a wave or membrane of sorts.

 

[DrChinese]

DrChinese, if I understood you correctly, you suggest that your (proposed) experiment gives STRONGER evidence (for wave-function reality) than other existing experiments. If I am right, it suggests that you CAN imagine how other experiments could be true without reality of the wave function, but you just cannot imagine how your particular experiment could be that.
If this is so, could you explain how the standard double-slit experiment could be true without the wave function being real? If you could explain that, I believe it would be much easier to answer your questions on this thread.

Yes, good point. I think the double slit explanation tends to invoke the idea of interference of the various paths the particle takes to the screen. I.e. there is self interference. And you could say that proves the reality of the wave function, with the logic that the photon took a discrete path. For the Bohmian, there is a pilot wave portion and that is real as well.

In my Frankenstein version, we must have an interaction between the probability from Alice and the probability from Bob to get entangled Chris. Similar for Dale. I don't see how that would occur if the probabilities represented knowledge of the system and nothing else.

I guess you could say that this is just words, and doesn't mean anything. Sorta like the "entangled photons that have not interacted in the past". On the other hand, that is a pretty interesting experiment and certainly worthly of consideration when thinking about entanglement.

 

[Fredrik]

What does any of this have to do with the wavefunction being "real" or not? What does that even mean?

 

[DrChinese]

What does any of this have to do with the wavefunction being "real" or not? What does that even mean?

My question has to do with the idea that state has to do with knowledge of the system. I think the state is an accurate and complete description of the system.

 

[RUTA]

My question has to do with the idea that state has to do with knowledge of the system. I think the state is an accurate and complete description of the system.

By "system" do you mean the experimental devices or "a quantum entity" or some combination thereof or ...?

 

[RUTA]

Yes, good point. I think the double slit explanation tends to invoke the idea of interference of the various paths the particle takes to the screen. I.e. there is self interference. And you could say that proves the reality of the wave function, with the logic that the photon took a discrete path. For the Bohmian, there is a pilot wave portion and that is real as well.

The wave function need not have any ontic status, e.g., the pilot wave. You don't even need a "wave" function, i.e., you can compute the twin-slit outcome using the transition amplitude from the path integral. So, the twin-slit experiment doesn't imply the wave function is "real" in the ontic or formal sense.

 

[DrChinese]

By "system" do you mean the experimental devices or "a quantum entity" or some combination thereof or ...?

Ha! I am not sure... I guess what I am saying is that the probability function (WF) can be manipulated in all kinds of strange ways.

1. The WF can be split, recombined, combined with the WFs from other particles, entangled with particles it has never been in contact with, including those that have never even existed at the same time. I guess that makes it pretty real. I realize that is a subjective statement.

2. If it can be recombined, I guess I would say that collapse is NOT an irreversible process. At least not at the point that the WF is split, so it would need to be later - when the complete final context has been determined.

So imagine entangled Alice: We run Alice through a PBS oriented at 0 degrees, getting 2 outputs. We run each of those outputs through PBSs oriented at 45 degrees. We then recombine the 4 outputs into 2 using 2 reverse splitters (a la the Eberly article quoted previously). And finally, recombine these 2 outputs with another suitably oriented reverse splitter. What do we have now? We have our original entangled Alice (at least we hope so).

3. And similarly, we should be able to mix and match any similar permutations of Alice and Bob so we can get Frankenstein photons Chris' and Dale', as long as everything sums to 100% and we have no idea which paths were traced.

So I am asking, if you are a believer of MWI for example (which of course you are not): how does all this splitting and mixing and matching go on and the entanglement can be restored? How does the universe know not to split because we are going to put everything back together again when we are finished with the manipulations? You don't think that might be a tall order for that interpretation?

 

[SpectraCat]

The wave function need not have any ontic status, e.g., the pilot wave. You don't even need a "wave" function, i.e., you can compute the twin-slit outcome using the transition amplitude from the path integral. So, the twin-slit experiment doesn't imply the wave function is "real" in the ontic or formal sense.

That last part seems like semantics ... AFAIK the path integral approach is formally equivalent to solving the Schrödinger equation in the normal way. Also, there are certainly wave-like entities involved in the path integral approach, since each individual path is represented by an amplitude times a complex exponential of the classical action. So, while it may be strictly true that the path integral doesn't involve the wave function as we usually think about it, it does involve wave-like entities.

So all this just moves the discussion to being about the reality of the path-integrals. Perhaps this is worthwhile ... I guess no one would dispute the reality of the paths themselves, since they are physical trajectories in space and time. Furthermore, the classical action along a given path also seems like a real thing, since it is just a time integral over the kinetic and potential energies along that path. Still, I guess that doesn't answer the question of how "real" the path integrals themselves are. They certainly seem real to me though, for all the same reasons that the wave-functions do. I guess that is what I meant when I said it seems like you are drawing a semantic distinction.

 

[mooglue]

I think everyone can at least agree that a wavefunction the square of the probability density for a given system. Would you say "probability" is real? It's not something tangible, but it does carry tangible information. I think that if you can argue that a probability distribution is a real thing, then you can possibly consider the wave function real.

 

[Dmitry67]

No, because in deterministic interpretations there is no 'probability' - at all.

 

[DrChinese]

So all this just moves the discussion to being about the reality of the path-integrals. Perhaps this is worthwhile ... I guess no one would dispute the reality of the paths themselves, since they are physical trajectories in space and time.

This is pretty much the path I am on and hoping to consider. And sincerely, I realize the line between semantics and meaning is razor thin on this one.

If the WF/path is real, then a lot of additional things must be true. For example, a free photon in space that eventually hits my eye must have its WF still propagating in space long after it hits my eye.

 

[Count Iblis]

So I am asking, if you are a believer of MWI for example (which of course you are not): how does all this splitting and mixing and matching go on and the entanglement can be restored? How does the universe know not to split because we are going to put everything back together again when we are finished with the manipulations? You don't think that might be a tall order for that interpretation?

But MWI just assumes that the formalism of quantum mechanics works as usual without invoking a real collapse of the wavefunction after a measurement. It is only that the effective collapse can be pictured as if the universe splits.

 

[RUTA]

That last part seems like semantics ... AFAIK the path integral approach is formally equivalent to solving the Schrödinger equation in the normal way. Also, there are certainly wave-like entities involved in the path integral approach, since each individual path is represented by an amplitude times a complex exponential of the classical action. So, while it may be strictly true that the path integral doesn't involve the wave function as we usually think about it, it does involve wave-like entities.

So all this just moves the discussion to being about the reality of the path-integrals. Perhaps this is worthwhile ... I guess no one would dispute the reality of the paths themselves, since they are physical trajectories in space and time. Furthermore, the classical action along a given path also seems like a real thing, since it is just a time integral over the kinetic and potential energies along that path. Still, I guess that doesn't answer the question of how "real" the path integrals themselves are. They certainly seem real to me though, for all the same reasons that the wave-functions do. I guess that is what I meant when I said it seems like you are drawing a semantic distinction.

One doesn't have to view the path integral Z as involving paths (in configuration space or spacetime). Typically, one sees it explained using paths (thus the name), but that's just a convenient way of ordering all possible values of the integration variable (typically evaluated from -infinity to infinity). One can rather view Z as providing a measure of the symmetry of that part of the integrand stripped of the integration variable (we call it the actional). That's why we call Z the symmetry amplitude in RBW instead of the transition amplitude. And, again, there are no quantum entities represented in the actional -- just the experimental equipment.

 

[Fredrik]

My question has to do with the idea that state has to do with knowledge of the system. I think the state is an accurate and complete description of the system.

In that case I guess I just don't understand the argument. Everyone already agrees that the wavefunction describes the system immediately after a measurement, so what you're getting at must be that it does so at all times. I don't see anything in your argument that we can use to reach that conclusion.

Maybe you're arguing for something else entirely. If you're arguing against the state vector being a representation of our "knowledge" of the system, I think the argument would have to start with an explanation of what that means. Personally, I don't think it makes sense to argue against that particular flavor of the CI, because it's so ill-defined and so poorly explained by its proponents that there's nothing to refute.

 

[Demystifier]

2. If it can be recombined, I guess I would say that collapse is NOT an irreversible process. At least not at the point that the WF is split, so it would need to be later - when the complete final context has been determined.

In my opinion, the collapse is a vague and obsolete concept. A modern much better defined concept that to a large extent replaces the concept of collapse is - decoherence. Decoherence is reversible in principle, but irreversible in practice because decoherence occurs when the system interacts with a LARGE number of environment degrees of freedom. When the wave can be recombined (in practice), it simply means that decoherence has not happened before the recombination.

So, does it say anything about reality of the wave function? Well, certainly not directly. Nevertheless, most physicists working with decoherence like to think in terms of some variant of the "many-world" interpretation, in which the wave function is thought of as something "real". Bohmians also can be counted as belonging to this camp, as Bohmian interpretation can also be thought of as a variant of the "many-world" interpretation.

 

[DrChinese]

In that case I guess I just don't understand the argument. Everyone already agrees that the wavefunction describes the system immediately after a measurement, so what you're getting at must be that it does so at all times. I don't see anything in your argument that we can use to reach that conclusion.

Maybe you're arguing for something else entirely. If you're arguing against the state vector being a representation of our "knowledge" of the system, I think the argument would have to start with an explanation of what that means. Personally, I don't think it makes sense to argue against that particular flavor of the CI, because it's so ill-defined and so poorly explained by its proponents that there's nothing to refute.

Fredrik, thanks for your comments.-DrC

 

[DrChinese]

In my opinion, the collapse is a vague and obsolete concept. A modern much better defined concept that to a large extent replaces the concept of collapse is - decoherence. Decoherence is reversible in principle, but irreversible in practice because decoherence occurs when the system interacts with a LARGE number of environment degrees of freedom. When the wave can be recombined (in practice), it simply means that decoherence has not happened before the recombination.

So, does it say anything about reality of the wave function? Well, certainly not directly. Nevertheless, most physicists working with decoherence like to think in terms of some variant of the "many-world" interpretation, in which the wave function is thought of as something "real". Bohmians also can be counted as belonging to this camp, as Bohmian interpretation can also be thought of as a variant of the "many-world" interpretation.

Demystifier, thanks for your comments as well. I know the question was a bit ambiguous, I was curious to see what others thoughts in this area might be.

I still can't get around the idea that: A particle's WF expands in free space even after a particle is detected. (It would do this because there are still path alternatives in existence.) It seems like any point in open space would have a large number of these alternative paths coming through. And that would lead to physically detectible effects of some sort. Or maybe not. I know I'm rambling... maybe I had too much coffee this morning.

 

[my_wan]

Interesting thread. I operate under the assumption of an ontic wavefunction. I also agree with RUTA that these frankenstein particles don't actually provide an empirical distinction that circumvents prior arguments. However, I am impressed with its potential to articulate the distinctions between these opposing viewpoints.

No, because in deterministic interpretations there is no 'probability' - at all.

In principle yes, in practice 'probability' is unavoidable in the formalism. Even classical thermodynamics unavoidably relies on 'probability' in the formalism. Even a basic dice role can only be predicted by probability. My personal sense, given my assumptions, is that the wavefunction is an amalgamation of both ontic and epistemic elements. In the dice case an ensemble can be trivially decomposed into real and epistemic parts. In the wavefunction case even our notions of what constitutes a fundamental physical property fails to maintain a distinct identity like dice, as if the realness we empirically percieve in the things around us are epistemic rather than ontic. The only class of theories I know to get around this considers measured properties as emergent, rather than inate.

Local theories that maintain local realism and escape both the Bell and Kochen-Specker Theorems exploit contextuality. Consider how we operationally define Einstein Realism in the context of BI. In essense we take some measurable value, often labeled Alice and Bob, and use them as a proxy for ontic values. Thus any localy realistic theory that doesn't violate these theorems must contextualize these measured values as emergent global properties of an underlying physics. This seems to imply that the wavefunction is, at least in part, a real wave of some sort, with emergent properties we measure and mistakenly assume are inate to the ontic parts. Relational QM is predicated on a similar viewpoint. A lot of abstract models have been formulated in an attempt to demonstrate feasibility, but to date no such model uniquely or fully recovers the formalism of QM.

Although this approach to circumventing the no-go theorems works in principle, it hasn't been demonstrated to be feasible to empirically replicate QFT. Thus my presumptions are just that, presumptions. Yet any argument that attempts to articulate a case for the wavefunction being real is of interest.

Here's the objection I would pose to RUTA's argument. It's true we don't have to view the path integral Z as involving paths, but this presumes the variable we associate with the path is itself not an ontic entity. Fair enough, as this so far would be true for both CI and the contextual relational interpretation mentioned above. Yet what is implied by presuming no ontic entities are involved in defining these properties? It implies that none of the everyday things we interact with are fundamentally ontic in nature. Most of us I presume reject this right or wrong, but once we reject epistemic properties without ontic elements of some sort to define them how do we then reconcile certain physical variables not requiring them. It seems silly to presume that certain physical properties lack an ontic basis at some level yet still cling to the notion that our everyday world contains ontic entities.

Now certainly there exist many variables that merely encode our state of knowledge, thus require no ontic basis in and of themselves. Yet, if we presume the Universe has an ontic basis, even these variables have ontic underpinnings as some state of knowledge about the ontic elements. Now if we try to maintain an ontic Universe while avoiding any ontic realness contained in the wavefunction, this is only possible by professing ignorance of where the ontic elements are contained, which we are if they exist. Furthermore, any objection to the assumption that the wavefunction has real elements depends on our ignorance to merely claim that we don't have to consider variable X real, which is true, but X can be moved to whatever argument is being made, such as the path integral. Thus it becomes an argument from ignorance. CI requires this very real ignorance to justify moving the goal post at will to maintain cogency without explicitly rejecting the Universe has ontic underpinnings.

It is my opinion that in order to avoid ontic realness associated with the wavefunction at any level requires assuming the Universe or any part of it needs no ontic basis. As a personal preference I'm betting the wavefunction contains real components of some sort, even if no particular variable we use to describe it is explicitly ontic in itself.

 

[Dmitry67]

The very word 'real' is ill-defined.
There is a very good example: BM (Bohmian Interpretation). In BM,there are 2 'real' components: wavefunction (exactly the same as in MWI, with all 'parralel' worlds) and hidden particles guided by this wave.

However, in BM only waves with particles inside (non-empty waves) form what we call reality. If we see a dead cat, then there is definitely a wave of alive one. It is real, but not tagged with particles (which can not be detected!) and hence don't form the reality.

What do you call 'real'?

 

[zonde]

I still can't get around the idea that: A particle's WF expands in free space even after a particle is detected. (It would do this because there are still path alternatives in existence.) It seems like any point in open space would have a large number of these alternative paths coming through. And that would lead to physically detectible effects of some sort. Or maybe not. I know I'm rambling... maybe I had too much coffee this morning.

I am trying to get hold of this question of yours but it does not make much sense for me.
However this comment gives me an idea. Do you consider that requirement for "reality" of wavefunction is "reality" of superposition? Meaning that particles can exist in more than one place before measurement.
If that is so then I can try to defend position that wavefunction is not "real" or even pilot-wave is not "real".