Questioning the Logical Necessity of MWI for QM

[michael879]

Id like comments on this statement because although I see it as true, the fact that MWI isn't generally accepted while QM is suggests it may be false.

The way I see it, if QM is an accurate description of the universe (i.e. it is not just an approximation of some underlying theory), then MWI is the only logical explanation. This comes mainly from the phenomena of superposition and decoherence. Observation in QM is the act of a macroscopic object interacting with a particle in a superposition. The wave function of the particle collapses into a delta function and it is the delta function that is observed. The way observation takes place is that some particle in the macroscopic object interacts with the observed particle and then interacts with all of the other particles in the object (which then at almost the speed of light interacts with the entire planet).

The problem with this is that there appears to be a discontinuity in the act of observation. A single particle interacting with another particle does not collapse either wave function. What happens, according to QM, is that a two particle system is made that itself is in some superposition (of course the system's wave function can sometimes be two delta functions).

These two results seem to be contradictory, since the macroscopic object interacting with a particle is simply a large number of particles interacting with each other. The only way to resolve this "paradox", that when YOU interact with a particle it appears to decohere but when a particle interacts with another particle neither appears to decohere, is to accept MWI. This provides an answer to this problem by saying that when a particle interacts with another particle, the local universe splits into each possible collapse and from the particle's "view", the other one has decohered (and it is always decohered). This explains why when we interact with a particle it appears to decohere by saying that the universe splits around us and in each universe the particle has decohered. To an observer outside of the light cone surrounding the event however, the entire system would be in a superposition.

 

[dst]

Does it even make sense to say "observer out of the light cone"? I don't know much QM so I'm on shaky ground here. What constitutes an observer?

I honestly can't see how MWI is anything more than a clever rephrasal of "it happens but that also can happen".

The only way to resolve this "paradox", that when YOU interact with a particle it appears to decohere but when a particle interacts with another particle neither appears to decohere, is to accept MWI.

Why is there a paradox? Can some sort of observer effect come from many many particles interacting, an emergent phenomenon if you will? It makes no sense to define a temperature for two particles but as soon as we get into many many particles it is quite clear.

 

[peter0302]

MWI is nothing more than saying that the wavefunction continues to evolve indefinitely without "collapsing" due to observation. The state of a particle, however, is relative to the observer, hence, it was originally called "relative state" / universal wavefunction. In other words there is a universal wavefunction describing EVERYTHING, and we each see only a sliver of it for each particle, and the slivers never interact after an irreversible measurement.

The informal way of saying it is that "everything that can happen does happen" but it's a little more complicated than that.

The only way to resolve this "paradox", that when YOU interact with a particle it appears to decohere but when a particle interacts with another particle neither appears to decohere, is to accept MWI.

It's not really the onlly way to resolve the paradox, but it's one way of resolving it. In _any_ interpretation, decoherence happens when there's been an irreversible measurement, i.e., when the particle has interacted with enough of its environment to count as creating "information." So a collision with a photon detector counts, regardless of whether it's viewed by the human.

That's an important point, one Zelinger emphasizes. You can throw away the information or ignore it, but if the information _exists_ there's been wavefunction collapse. So it's really not a matter of whether YOU interact with the particle so much as it is whether you're CAPABLE of observing a measurement.

That's why delayed choice experiments are so fascinating, because on they seem to be "reversing" measurements that were made already.

 

[michael879]

Does it even make sense to say "observer out of the light cone"? I don't know much QM so I'm on shaky ground here. What constitutes an observer?

I honestly can't see how MWI is anything more than a clever rephrasal of "it happens but that also can happen".

Whats wrong with talking about light cones? Its a spherical region of space around the "event" expanding at c. Anything outside of this "cone" could not have possibly interacted directly or indirectly with any of the particles involved in the "measurement".

Yes, MWI does provide a nice deterministic answer to QM but that's irrelevant.

An observer is not well defined but basically, my definition is that observing a particle is some macroscopic object interacting with that particle. That is what's required to collapse the wave function.

Why is there a paradox? Can some sort of observer effect come from many many particles interacting, an emergent phenomenon if you will? It makes no sense to define a temperature for two particles but as soon as we get into many many particles it is quite clear.

The paradox is that when some object interacts with some particle one of two things can happen from our point of view.

1) If the object is small enough, they form a multi-particle system and no wave function collapses

2) If the object is large, the particle's wave function collapses to a delta function.

Since large objects are simply made up of smaller objects, this seems to be a paradox. Why don't we form a multi-particle wave function with an electron when we measure its spin?

 

[michael879]

That's an important point, one Zelinger emphasizes. You can throw away the information or ignore it, but if the information _exists_ there's been wavefunction collapse. So it's really not a matter of whether YOU interact with the particle so much as it is whether you're CAPABLE of observing a measurement.

I emphasized "you" just to make the point that a conscious being is never aware of being in a multi-particle superposition. I wasnt suggesting intelligence is required to collapse a wave function.

It's not really the onlly way to resolve the paradox, but it's one way of resolving it. In _any_ interpretation, decoherence happens when there's been an irreversible measurement, i.e., when the particle has interacted with enough of its environment to count as creating "information." So a collision with a photon detector counts, regardless of whether it's viewed by the human.

What are the other ways? This is the only interpretation I've heard that resolves it.

Also, I am claiming that there's no such thing as absolute decoherence. I suppose you could be right and a wave function just magically decides to collapse when it has interacted with enough particle. I find that hard to believe though and relative decoherence in MWI makes a lot more sense. Although I measure an electron as spin up, anyone outside my light cone will "observe" (really bad word here because you can't actually observe this) my entire light cone as being in a superposition between me measuring spin up and me measuring spin down and which state the observer ultimately observes is random.

 

[Ken G]

There is nothing "magical" in wave function collapse, that is an unfortunate byproduct of how a lot of people erroneously describe the Copenhagen Interpretation. We observe wave function collapse, it's just that simple-- and we do it on purpose. We intentionally collapse a wave function when we select a measuring device that yields the appropriate decoherence. We might detect photons, for example, by intercepting them using a photographic plate. Why? Because we know a photographic plate will induce the required decohering of the photon position basis. Once the decoherence has been accomplished, intentionally, the photon enters what we would describe as a "mixed state"-- it could still have hit anywhere on the plate, we just don't know until we look, but the distinction doesn't matter because the coherences are lost. Finally, we look, see only one outcome-- the wave function has collapsed. That's just the observed fact.

Now this is embarassing when one interprets quantum mechanics as a "theory of everything", because it is a nonunitary evolution of the wave function. But that's no problem for the Copenhagen Interpretation-- nothing magical, of course the state did not evolve unitarily, it was not a closed system. The introduction of the photographic plate brought in a lot of noise and all the history of that plate, and we have no interest in tracking it, so we choose to average over it like a poker player analyzing his winning chances. That's where the "collapse" occurs, it's all part of how we do science. I like to say that "the projection of the MWI onto the scientific method yields the Copenhagen Interpretation".

So the issue is not if MWI is "right" in some philosophical sense, the issue is, it is science? I say science is demonstrable, and MWI is not, so MWI is not a part of science. What can be stated as more categorically true is that the distinction between MWI and CI doesn't matter at all to science, it simply does not show up anywhere in the application of science. That's why the MWI is never actually used to analyze quantum mechanics experiments, which I think is what you meant in the OP when you said it wasn't "generally accepted". It's just vestigial overhead for any real quantum investigation, but it's fine to apply if it fills a gap in how you like to imagine reality. Personally I think it is taking our axioms too seriously-- we invented the axioms to create a simulacrum of reality, they were not supposed to be reality. That gets into the history and philosophy of science, and is more a matter of personal taste.

 

[michael879]

There is nothing "magical" in wave function collapse, that is an unfortunate byproduct of how a lot of people erroneously describe the Copenhagen Interpretation. We observe wave function collapse, it's just that simple-- and we do it on purpose.

I think you misunderstood me, I probably shouldn't have called it "magical". What I meant can be illustrated with the following thought experiment:

take a particle in some coherent superposition and one by one add particles to this system so that they all form some type of combined wave function. What your saying is that at some point the original particle's wave function collapses (i.e. when it reaches macroscopic levels) right? That discontinuity seems strange to me and I don't see what would make it occur..

So the issue is not if MWI is "right" in some philosophical sense, the issue is, it is science? I say science is demonstrable, and MWI is not, so MWI is not a part of science. What can be stated as more categorically true is that the distinction between MWI and CI doesn't matter at all to science, it simply does not show up anywhere in the application of science. That's why the MWI is never actually used to analyze quantum mechanics experiments, which I think is what you meant in the OP when you said it wasn't "generally accepted". It's just vestigial overhead for any real quantum investigation, but it's fine to apply if it fills a gap in how you like to imagine reality. Personally I think it is taking our axioms too seriously-- we invented the axioms to create a simulacrum of reality, they were not supposed to be reality. That gets into the history and philosophy of science, and is more a matter of personal taste.

Yea I agree with you, its clearly just an interpretation and not a theory. It makes no testable predictions and we could never know if it were true (although if QM is false so is it). I guess my original thought was a little too extreme (personally, I use both interpretations all the time). However explaining quantum effects at macroscopic levels (i.e. the act of observation) seems, in my opinion, to only make sense in the MWI. The difference between CI and MWI here (as I understand this) is that in CI, the wave function collapses absolutely while in MWI, the wave function only collapses locally as you entangle with it (If I have this wrong then I apologize for this post).

 

[Ken G]

I think you misunderstood me, I probably shouldn't have called it "magical". What I meant can be illustrated with the following thought experiment:

take a particle in some coherent superposition and one by one add particles to this system so that they all form some type of combined wave function. What your saying is that at some point the original particle's wave function collapses (i.e. when it reaches macroscopic levels) right? That discontinuity seems strange to me and I don't see what would make it occur..

The "magical" business is pretty common, I wasn't hanging that on you. As for the thought experiment, the discontinuity in the collapse is no different from all the other types of discontinuities that we invoke in our analysis of physics. How does a particle go from "moving" to "stationary", for example? Isn't that a discontinuity too? We aren't bothered by these things, we are used to idealizations. Yet suddenly when it's quantum mechanics, it's some kind of crisis.

Yea I agree with you, its clearly just an interpretation and not a theory.

Yes, it is designed to solve a "problem", but is it really a problem that needs solving?

However explaining quantum effects at macroscopic levels (i.e. the act of observation) seems, in my opinion, to only make sense in the MWI.

It all depends on if you imagine yourself as outside the observation, or inside it. MWI only resolves anything in the latter case-- but the latter case isn't science, it's philosophy. So the question is, are we seeking a scientific explanation, or some more general thing we can call an explanation? Aren't there other ways of explaining things if we go outside of science?

The difference between CI and MWI here (as I understand this) is that in CI, the wave function collapses absolutely while in MWI, the wave function only collapses locally as you entangle with it (If I have this wrong then I apologize for this post).

I don't think there's any problem with your understanding or mine, I think it's just that there's a lot going on. There are actually two kinds of "collapse" that are often confused. The first is from decoherence, and happens in CI and MWI, because it is physically real. It happens when the coupling to the measuring apparatus destroys coherences between the basis states of the measurement, resulting in a mixture of state-apparatus consistencies. That mixture has many of the properties of a "collapse", but it's there in the MWI too. Then there's the projection onto the scientific method-- the apparatus is interpreted classically, which means it has to yield only one result. That is the final "collapse", the one that distinguishes MWI and CI, but you can't do science without it. To achieve it, the act of conceptualizing an "outcome" forces us to be outside the system (which forces us to ignore our physical connections to the device that allow us to read its outcome), breaking the unitariness, and explaining the source of the collapse. The CI didn't do it, the CI merely reacts to what we did when we chose the particular mode of analysis that we always use. Our brain has no choice but to interpret itself separately from the measurement, it's something about how consciousness works.

 

[dst]

Whats wrong with talking about light cones? Its a spherical region of space around the "event" expanding at c. Anything outside of this "cone" could not have possibly interacted directly or indirectly with any of the particles involved in the "measurement".

Yes, MWI does provide a nice deterministic answer to QM but that's irrelevant.

An observer is not well defined but basically, my definition is that observing a particle is some macroscopic object interacting with that particle. That is what's required to collapse the wave function.

The issue is that since no information would even be reaching our observer outside the light cone, then as far as they are concerned, nothing actually exists therein. Is that the same thing as saying a "superposition of states"? A tangential question, can we consider a vacuum to be in a "superposition of states"? A "measurement" causing it to collapse and produce, perhaps, a particle?

I don't see how you can make the distinction between a macroscopic object and a single particle, since as we say, one is just a more complex arrangement of the other.

The paradox is that when some object interacts with some particle one of two things can happen from our point of view.

1) If the object is small enough, they form a multi-particle system and no wave function collapses

2) If the object is large, the particle's wave function collapses to a delta function.

Since large objects are simply made up of smaller objects, this seems to be a paradox. Why don't we form a multi-particle wave function with an electron when we measure its spin?

Now I see what you mean. Can't you just brute force the problem and apply QM to the whole thing?

As I said, why can it not simply be an emergent phenomenon that simply happens? Does god get overwhelmed trying to decide the outcome of too many particles? :yuck:

 

[peter0302]

What are the other ways? This is the only interpretation I've heard that resolves it.

Cramer's TI and Bohmian Mechanics also "resolve" it. TI says that the wavefunction collapse is the result of two waves cancelling each other out. In Bohmian mechanics there is no collapse - the particle always had a fixed momentum and position.

 

[Hurkyl]

The "magical" business is pretty common, I wasn't hanging that on you. As for the thought experiment, the discontinuity in the collapse is no different from all the other types of discontinuities that we invoke in our analysis of physics. How does a particle go from "moving" to "stationary", for example? Isn't that a discontinuity too? We aren't bothered by these things, we are used to idealizations. Yet suddenly when it's quantum mechanics, it's some kind of crisis.

This discontinuity is clearly of a different kind, since there are degrees of "moving", and a continuous transition from them into "stationary".

And this is different than other 'discontinuities' that we see in physical analysis. For example, in classical mechanics, we might treat a gas as if it were a continuum. But when we do this, we do not suspend the laws of classical mechanics for the sake of analysis; we know that "under the hood", the gas is still made up of zillions of tiny particles that obey the laws of (classical) particle mechanics, and we are using a methodology known to approximate the bulk properties of the gas to sufficient accuracy. In fact, I imagine the people who really understand this stuff have discarded their "continuum fluid" intuition, and replaced it with a "continuum approximation to a particulate fluid" intuition.

The Copenhagen interpretation requires one to momentarily suspend the normal time evolution of state of the universe (or of the system, or whatever) so that you can replace it with an eigenstate of an observable. And that is why it's a 'crisis'.

Then there's the projection onto the scientific method-- the apparatus is interpreted classically, which means it has to yield only one result. That is the final "collapse", the one that distinguishes MWI and CI, but you can't do science without it. To achieve it, the act of conceptualizing an "outcome" forces us to be outside the system

If "yield only one result" is the only requirement you are putting onto an apparatus, then you do not need to invoke a collapse, or suspend quantum mechanics -- unless you are making additional philosophical assumptions that would imply that.

I would like point out, though, that the only time collapse is relevant, even in CI, is when an "observation" occurs and analysis continues -- i.e. when the observation is part of the analysis.

 

[masudr]

There is nothing "magical" in wave function collapse...

Well it is, kind of. Reducing a superposition to a single (or set of) state (this is usually called collapse) is a nonunitary process. Schrodinger time evolution is a unitary process. There is no way a unitary process can lead to nonunitary changes in the state vector.

So claiming that the macroscopic and microscopic systems interacting somehow end us up with an eigenstate of the measurement operator is sort of magic, because that can't happen by just time evolution of the two interacting systems.

 

[Ken G]

This discontinuity is clearly of a different kind, since there are degrees of "moving", and a continuous transition from them into "stationary".

But there is a similarly continuous transition from classical treatments into quantum mechanical ones. A classic example is a moving hydrogen atom emitting a photon-- the electron is necessarily treated quantum mechanically, but the proton is almost always treated classically. Such hybrid approaches are perfectly common in physics-- the idea that the practice of physics is formal and axiomatic is a fantasy, so we should never be bothered when faced with "transitional treatments".

And this is different than other 'discontinuities' that we see in physical analysis. For example, in classical mechanics, we might treat a gas as if it were a continuum. But when we do this, we do not suspend the laws of classical mechanics for the sake of analysis; we know that "under the hood", the gas is still made up of zillions of tiny particles that obey the laws of (classical) particle mechanics, and we are using a methodology known to approximate the bulk properties of the gas to sufficient accuracy.

But that's exactly what we don't know "under the hood"-- because it isn't even true. One theory is subsumed by another all the time, all we require is that the contradictions be smaller than our accuracy target. So it is with the "collapse of the wavefunction", it occurs only within some level of precision, which is completely up to us. Our fingerprints are all over the results of science-- always have been, no reason to expect it will ever be different.

In fact, I imagine the people who really understand this stuff have discarded their "continuum fluid" intuition, and replaced it with a "continuum approximation to a particulate fluid" intuition.

That's an important insight, because I suspect the opposite is true-- there must be a rich and powerful new system of "continuum fluid intuition" that fluid experts develop, and the fact that no such thing exists in reality is completely irrelevant to the construction of that powerful intuition.

The Copenhagen interpretation requires one to momentarily suspend the normal time evolution of state of the universe (or of the system, or whatever) so that you can replace it with an eigenstate of an observable. And that is why it's a 'crisis'.

But we know exactly why it does that, it's no different from suspending the need to understand how a card deck was shuffled when you play poker. It isn't a "crisis", it is the best you can do. If you can track the shuffling process, you'll be the best poker player ever, but for everyone else, they do fine suspending the time evolution of the shuffling process and just average over the possibilities.

The incredibly powerful concept of "ergodicity" in thermodynamics works on the same principle-- why was that never a "crisis" for deterministic Newtonian physics? I've never understood why people get so bothered about nonunitariness in quantum measurement, but seemed just fine with irreversibility in Newtonian thermodynamics. It's the same issue all over again-- it's in how you treat the information you cannot track. That's also why the patent office will not look at perpetual motion machines, even though the fundamental equations of physics are reversible (including unitary time evolution).

If "yield only one result" is the only requirement you are putting onto an apparatus, then you do not need to invoke a collapse, or suspend quantum mechanics -- unless you are making additional philosophical assumptions that would imply that.

I think you are wrong about that. A single definite result will not be unitary unless it was the prepared state where you first started your investigation. Yet it is what we observe. It's just that simple-- we observe nonunitary behavior, so we have two choices. Either we embed ourselves in some uber-unitary universe that we cannot even test or demonstrate, just to assuage our sense of pride in our unitarity axioms, or we accept that the unitarity axioms apply to only a part of the process of science (the time evolution of the quasi-closed system prior to measurement)-- and we know how they perform when we average over information we cannot track scientifically (i.e., they perform by removing unitariness). The former choice is MWI, the latter CI, and only one is science.

I would like point out, though, that the only time collapse is relevant, even in CI, is when an "observation" occurs and analysis continues -- i.e. when the observation is part of the analysis.

That depends on which type of "collapse" you mean. The decohering of pure substates into mixed substates has most of the properties of collapse, and that happens all the time, it is a physical effect. Then there is the "final collapse" that you are talking about, where a single result is selected, recovering a pure state for the subsystem, and that indeed only happens as part of an analysis process. That is the science part, where the human brain gets into the picture.

 

[Ken G]

Well it is, kind of. Reducing a superposition to a single (or set of) state (this is usually called collapse) is a nonunitary process. Schrodinger time evolution is a unitary process. There is no way a unitary process can lead to nonunitary changes in the state vector.

That is true on the surface, but the implication is not. The problem is what you mean by "the state vector". Your statement is only true for the state vector of a closed system, so must include not only the measuring device but also the scientist. But when you talk about the nonunitary collapse of that "state vector", you are only talking about the quantum system. This inconsistency in language causes most of the misconceptions about problems with the CI. The CI has no problems-- it simply chooses not to track certain information in the combined quantum+observation system. Nothing it says in any way violates quantum mechanics once you include that feature (yes, feature, not "bug"), which is why it is what gets used. Did Bohr spend much of his life arguing that quantum mechanics is the best science we can produce, and then equip it with an interpretation that contradicts it? No, he was on the "feature not bug" side.

So claiming that the macroscopic and microscopic systems interacting somehow end us up with an eigenstate of the measurement operator is sort of magic, because that can't happen by just time evolution of the two interacting systems.

We know exactly how it happens-- it happens in our treatment of reality. Note we have no other options-- we cannot scientifically treat reality in any other way, there really is information we cannot track.