Cipheron on 23/5/2022 at 12:12
Quote Posted by demagogue
o My biggest motivation is probably the fact that MWI uses the actual existing rules of QM to answer the measurement problem, the rules that were already experimentally confirmed and in place by 1927 (well, plus the relativistic stuff that came a bit later), and have only been reconfirmed for the next ~100 years without any hint of any "new rules" that other interpretations seem to require. It just strikes me as a very clean answer that the existing rules of QM also appear to answer the measurement problem (it's a long story to explain why; I'm just saying it here, but to do it properly, one would need to walk through it) and that answer looks like the vanilla QM computations that physicists do literally every day. That feels cleaner to me than insisting that QM needs to have these other esoteric rules that no one has ever found or needed for their day to day work over the last century.
Sorry, but to me this seems like more evidence that it's the wrong solution.
Firstly, if it doesn't add anything new then it could just be hand-waving away the problem instead of solving the problem. So we have issues and those issues are magically resolved by just saying that thing split into multiple universes constantly, where those different universes just happen to have the different possible solutions to our equations. No result of the theory is ever wrong, even when the predictions don't match, since those outcomes must have happened in another universe.
Plus, not adding anything new is not the correct approach. And that's because *we know quantum mechanics is wrong* from first principles. It isn't compatible with relativity, thus it is inherently wrong, no matter how well we match observations. Any theory that claims to have solved the dilemmas of quantum physics without actually squaring it with relativity must in fact be wrong or at least woefully incomplete, so just cramming inconvenient outcomes of the equations off in separate universes is BS.
So the "no new rules" bit is not in fact a good bit of science, it in fact smacks of unfalsifiable gibberish that merely hand waves a lot of results as occurring in universes we can't see, which is an explanation you could use to prove just about anything.
Pyrian on 23/5/2022 at 15:59
Quote Posted by Cipheron
Firstly, if it doesn't add anything new then it could just be hand-waving away the problem instead of solving the problem. So we have issues and those issues are magically resolved...
That's a strange way of describing "matches the evidence".
Quote Posted by Cipheron
No result of the theory is ever wrong...
No. MWI is observationally indistinguishable from QM (being an interpretation and all), but that just means that anything that breaks QM also breaks MWI, and vice-versa.
Quote Posted by Cipheron
It isn't compatible with relativity, thus it is inherently wrong, no matter how well we match observations.
If QM matched large-scale observations, we wouldn't
have a GUT problem. And there's no guarantee that when/if the two theories are reconciled, general relativity will be the one left standing. (BTW, I specify
general relativity because special relativity doesn't have this issue.)
Quote Posted by Cipheron
Any theory that claims to have solved the dilemmas of quantum physics without actually squaring it with relativity must in fact be wrong or at least woefully incomplete...
This argument applies equally to the standard Copenhagen interpretation of QM.
Quote Posted by Cipheron
...so just cramming inconvenient outcomes of the equations off in separate universes is BS.
That's not what MWI does.
Cipheron on 23/5/2022 at 18:31
Quote Posted by Pyrian
That's a strange way of describing "matches the evidence".
No. MWI is observationally indistinguishable from QM (being an interpretation and all), but that just means that anything that breaks QM also breaks MWI, and vice-versa.
If QM matched large-scale observations, we wouldn't
have a GUT problem. And there's no guarantee that when/if the two theories are reconciled, general relativity will be the one left standing. (BTW, I specify
general relativity because special relativity doesn't have this issue.)
This argument applies equally to the standard Copenhagen interpretation of QM.
That's not what MWI does.
I think it's *exactly* what it's doing:
Quote:
The many-worlds interpretation (MWI) is an interpretation of quantum mechanics that asserts that the universal wavefunction is objectively real, and that there is no wave function collapse.[2] This implies that all possible outcomes of quantum measurements are physically realized in some "world" or universe.
So if we measure an electron to be at one point, and that's as a result of waveform collapse, MWI says that in another reality another version of us measured the electron to have collapsed at a different point. Effectively it does away with the need to explain what happened by saying everything happened, just most of them happened in places we have no access to, the alternate realities.
That's what I meant by it sweeps inconvenient results under the carpet - most of the hypothetical observations occur in the other universes we have no access to. So we've just shifted along the need to explain why the waveform collapsed this way in our *specific* universe. It doesn't in fact explain anything.
demagogue on 24/5/2022 at 11:26
Yeah, there's no way I can keep this reasonably short, and by "short" I mean a wall of text not quite big enough to defend the North from the White Walkers. :sweat: I mean I've already written a post out that long, but it's not easy to just summarize.
I can put the chapter headings at least!
I'll do that. Then one can just skip to the ones they care about.
But keep in mind each bullet point is like 1 or 2 pages or more by itself.
o As long as we're all learning and thinking about QM, it's all good! ----------
o Here's the thing with the charge of “by saying everything happened”: That's literally what the path integral form of QM says you have to do. [...] The criticism of "sweeping the problem under a rug", like the Occam's Razor argument, is kind of symmetrical on this topic because it depends on what you think is the "rug" and what you're "sweeping" under it. Another way to look at MWI is it's exactly not sweeping what the math is insisting we do under the rug. I get your argument, but because of the symmetry of it, it opens itself to being attacked by the same logic. I went into a lot more detail in my long version, but this was the punchline.
----------
o As for the Relativity, Pyrian handled that fine. Gravity is a problem (the normalization trick doesn't work), but it's not unique to MWI. Interestingly I think MWI gives one a good picture of why renormalization is necessary (as in you can see pretty concretely why the path integral with all the insane layering of superpositions would sum to infinity; in MWI they really layer like that), but other interpretations have their own versions as well, so I wouldn't even say that's a plus in its favor.
----------
o Incidentally, (https://www.youtube.com/watch?v=A0da8TEeaeE) Sabine's video on the Principle of Least Action just came out.It has some thoughts about its connection with wave function collapse and measurement. I thought it was kind of fitting here, and as an added bonus Maupertuis make a cameo appearance!
-----------
o “Those aren't your only two options!”Yes, that's right. I tried to respond to this in my previous post. But you make a good point reemphasizing it. I'm not favoring it because it's better than other options because other options include "possibly countless other options you haven't even imagined and in fact may not even be able to imagine, or even if you could, that could never be figured out in theory or by experiment". And if I was saying or implying that I was, then I strike it. I just want to say it's just what I'd bet on right now, even with that caveat included.
----------
o Re: Falsifability & a bet!If it were falsified, or if anything could be experimentally said about "quantum fundamentals" in any way at all, expected or not, that would be awesome of course because we'd learn something fundamental about our universe. I don't want you to think I've lost the baby for the bathwater, or however that idiom works in this context.
So I'd bet on MWI right now, but I wouldn't bet much on it, maybe the cost of a Coke. On that note: I am ready to bet you one Coke or a comparably-priced beverage of your choice, sir or ma'am, that if ever an experiment is well designed and conducted that could actually falsify the MWI or its negation, that it will be vindicated to the standard of some respectable sigma. Let's do this the old fashioned way. XD
----------
o Speaking of possible falsifiability (maybe?), an article I saw today in Quanta Magazine was nicely timed: (https://www.quantamagazine.org/puzzling-quantum-scenario-appears-not-to-conserve-energy-20220516/) Puzzling Quantum Scenario Appears Not to Conserve Energy According to that article, there are, mathematically speaking at least, special combinations of sine waves that can make a wave packet move faster, or have greater momentum, than their fastest k constituents, just by the weirdo way they sum. The thought experiment is, you could have a system of light in a blackbody chamber, and then let a half-silvered mirror emit a beam with that special sine wave combination. Then if the measurement collapses the system in that state, you've got a photon with more energy than the photon(s) that went into the chamber, and would that violate conservation of energy? And if it doesn't, where did the extra energy come from? (Assuming it's even physically possible; someone needs to do an experiment to see if it is.) I need to think more about it, but on first impression assuming it'd be possible arguendo, I wonder if the energy is conserved across the multiverse, but the extra energy added in one branch (defined by the WF collapse for the special outgoing beam) is taken from other branches (not outgoing). It sounds kind of similar to the argument that quantum fluctuations can violate energy conservation inside one branch; but in MWI they'd also be taking it from one branch to another so it's still conserved over the multiverse.
I'm not too confident about any of this though, so I'm just referencing the article & thought experiment here as an interesting case to think about in this context.
----------
o What's the value of an interpretation at all? It's not going to really change how people "do" QM. The method is probably still going to be what we've always been doing. But I think the two values are:
(1) They're good mental models for making sense of what the math is doing, to build up good intuitions, even aside from their truth value.
(2) The philosophical thing. Science is a human enterprise to understand more about reality with models we can use. Interpretations don't really add much to that, so textbooks don't need to deal with it. Philosophy, what's the Truth of our world, is another kind of human enterprise to understand on a human level, as best we can, what reality is actually like, or what's less-wrong in our understanding of it. Interpretations are good for that. It's not what scientists typically say is their day job, but that's fine. It doesn't have to be.
----------
o Alright, the "measurement problem". I didn't come up with that term. It's just the common label that's used for the discussion. A more useful framing might be: "one of the rules of QM isn't like the others", as in "Everything up until measurement is smooth, no discontinuities, linear / unitary evolution, and then there's this discontinuous, irreversible thing." Okay, fine: nature is as nature does. There's no sense in lecturing her to change her act.
But it does raise a question, what's the relationship between these two types of rules, if any? It's not a problem per se; just a question you can look at and think "hmm". I don't really think much of the semantics of it, whether one wants to call it a “problem” or a “feature” or “Bob”. That's not really the best answer here though. I think it's better to frame it in similar terms as you're framing it.
----------
o So it's interesting how you phrased it, "wavefunction collapse is a central feature of QM".I think MWI-types would almost agree with this except I think they'd want to word it as "wavefunction collapse is the central feature of the transition from QM to classical physics." And they'd even want to put it more strongly to underline this point. Definitely there's the Quantum Zeno Paradox to explain, but also the fact that only one eigenvector is ever measured, the fact that probability follows the Born rule. All of those are real outcomes of measurements. I don't think they'd say they're illusions as such.
Something is physically happening that looks like a wavefunction collapse.
----------
o I have an idea to try to explain this succinctly. I don't know how well I'll do it. But, as a thought experiment, consider if there was no wavefunction collapse. The Schroedinger Cat image they always show is a living a dead cat in superposition (although it's not 2 cats; it's like 2^10^50 cats in a superposition, along with that many Earths, Milky Ways, etc.), as if, if that really existed, a person could actually see all those cats in the same space. But if that mixed state were real (and MWI says it is real) it doesn't look like that image at all. It will look like a vacuum state at every point because at every point the value of the all the systems in superposition will cancel each other out.
----------
o Bottom line: No structure (structure = maintained positive amplitude in the WF) would survive over time without "wavefunction collapse". QM as we know it wouldn't work. Wave function collapse is
the central feature or requirement, like you literally just said. And MWI would say, yes, that's exactly right. That's the answer. You don't need some extra mechanism to "kill" mixed superpositions and "protect" classical physics from being "destroyed" by them because mixed superposition states already "destroy" their connection to classical physics (in their basis) all by themselves; they obliterate all positive amplitudes by themselves. And then MWI types will want to conclude: in fact all of those things
are in that superposition. But there aren't "observers" "in" "there" to see it as a superposition exactly because of that superposition destroys all positive amplitudes.
----------
o So what about those observations of wave function collapse then?So the part where there are no observers in a mixed superposition state I think is the easy part to explain. No one sees it because mixed states wipe out the positive amplitudes. Maybe the thing that takes more explanation is: but then why are there observers in simple states? Where do the "point-like particles" with a definite value that we measure that sticks around (Quantum Zeno Effect, etc.) come from?
This also took a lot of space to explain. I don't know if I can do a bullet point version well.
First, the trend these days is to note that spacetime may be emergent from entanglement or Qbit operations. So any talk about "position", "momentum", "angular momentum" is already problematic, and whenever we have a better theory, we'd want to translate that kind of talk to the better theory. But I'm going to do it to keep things simple.
For an observer in the basis of a non-mixed state, MWI looks a lot like Pilot Wave theory, except instead of the particle being pushed around by the wave, the point-like particle is "part of" the wave, and it's taking a truly non-deterministic random walk in a way which respects the Born Rule (from the perspective of an observer in there). When a wave-like object defining a system smashes into another wave-like object in a way that actually shatters their coherence into little wavelets, like spilling their guts overlapping into each other and outwards (maximal entanglement / scattering), positive norm measurements are reduced to isolated points detached from each other.
Then from there the Shroedinger Equation ("SE", or its relativistic cousins) does the only thing it knows how to do, which is in the position basis is to take a positive norm measurement and spread it out as a wave packet based on its curvature. Since we have isolated positive norm points, basically near infinite curvature, they're going to be spread out very fast, at relativistic speeds, so already we need the relativistic equation. But the point is, the SE-like relativistic equation is going to preserve the positive norm measurement based on that single point, oblivious to other positive points out there, that are being spread out right on top of it. (Even as isolated points, it can still maintain bound states though, and nearby points in Hilbert space travel together. So you still get macroscopic objects evolving similarly. Isolated means isolated from the decohered parts of its own system; but you still get entanglement with other points in a consistent history.)
The thing is, in a hot and wet environment, as soon as the SE starts spreading out an isolated point of positive norm, it will almost immediately smash into another system that shatters it into isolated points of positive norm, and for each isolated positive norm point from that shattering, the SE will expand those out, each of them almost immediately being shattered into isolated points again.
----------
o Emergence of Classical Physics, what MWI stands or falls onThe miracle that's supposed to make MWI work, what it stands or falls on, is that the succession of these emergent "points" are the particles of classical physics that instantiate chemistry, biology, natural selection, etc.
Each isolated positive norm point would be a product of or receiving a contribution from the positive norm of the two contributing systems, the actual "entanglement" part, because that's what the SE does, it maintains positive norms over time... You can even start to see why you have to renormalize after the "wave function collapse", what creation and destruction operators would be, or what virtual particles would be & what happens to those wavelengths that don't fit in the bound state or that don't go to zero at infinity, and why we don't see those (Like I was saying, even if you don't believe the theory is true, it can be a useful mental model.)
What would it look like for a “system” made out of those interactions? I was already noting above that in a mixed system, i.e., following maximal entanglement events, there should be no structure maintained in the mixed basis. But the MWI would say there is structure maintained in another basis, and that's the basis of each individual positive norm point followed by its successor positive norm points in entangled states, in a consistent line of history. And what is that going to look like to someone measuring that history from "inside"? It's going to look like a succession of point-like particles following a path that looks like a truly random walk that happens to fit inside a wave-like object's structure. Since each point is “part of the parent expanding wave”, it's going to properly follow a path that manages to respect the Born Rule (well it should over time, over very many entanglement events; I think this is still being hashed out though and some dispute it), while still being perfectly random and non-deterministic from the perspective of a person in there.
I probably need to talk about consistent histories & encryption, why the isolated particles of classical physics really stay isolated, except for a consistent-history succession of connected points in the basis of one of the points. It's already written into the SE though, and the decoherence literature will go into detail about it.
----------
o Concluding remarks XDAnyway, this is not very careful writing on my part. So I would probably want to disown or change a lot of it with some more time, thought, and if I could make it about 20 times longer, but that would probably be worse for our purposes here. Also I'm tired now, so I'll leave what I have here. I also mentioned that I wrote a previous version of this post that was like 20 times as long. It went into a lot more detail responding to the comments I got, and I'm sure I'll get unsatisfied responses again even though I covered them in the longer version. But such is the challenge of this topic however you slice it! But I may put that up somewhere in case anyone is curious and I'll put a link here if I do.
In the end again I'm not so interested in actually persuading anyone of anything. Everybody should obviously think for themselves and come to the conclusion that seems best to them. The main thing I want to do is just fairly explain how MWI works and why I marginally like it right now. I hope I can at least do that much. And I just like talking about QM in any context, so thank you all for joining along, and feel free to add your own thoughts.
Briareos H on 24/5/2022 at 15:51
I haven't thought about it as much as you have because I haven't yet given myself the time to properly learn the mathematical tools and frameworks. I only know enough to understand the basics and follow along most papers if I hand-wave the hard math. Here I'll just put a rather scientifically worthless opinion, if only to understand yours better. I've always felt that the problem with quantum measurement was an artificial one that shouldn't be solved with ad-hoc theories for non-unitary evolution but rather with changes to the fundamental assumptions on Bell's theorem. On this we agree. I also strongly dislike the notion of measurement needing dedicated ontology within the theoretical framework. In general, hidden variables have always been more attractive to me because they (as a general concept) have been fundamental in refining our understanding of the world. We always end up with hidden variables.
You mention that "[e]very other interpretation you have to bend over backwards with hacks whose only job, it looks like, is to prevent parallel worlds, because they don't actually fix anything else, and they have to break a lot of other stuff to do it, like insisting on absurd new rules for which there is evidently no other basis than people's discomfort with what vanilla QM is telling us", but you also quote Sabine Hossenfelder who is an active proponent of an alternative interpretation which happens not to be a hack, not to break existing frameworks and at the same time happens to provide novel insight into common pain points with QM such as entanglement and nonlocality: superdeterminism. If you lose the assumption of statistical independence and implicitly allow retrocausality, you can use correlations between the hidden variables to recover the funky stuff you get out of collapsing superposed states, and nothing stops you from describing measurement as taking a slice of a continuous process constrained only by the initial conditions, in a way where that measurement is only an approximation which never observes a true eigenstate with its hidden variables.
Any reason you didn't mention it? It doesn't have many advantages over the many-worlds interpretation, but I like the idea that you don't need to consider the sum of all parallel worlds to reconstruct history and get your unitary evolution back. It's currently my favourite interpretation because the more I think about it, the less I can imagine how everything would not be correlated.
On a philosophical level, I acknowledge it can look profoundly uncanny, because 1. the universe would be in a state of pseudo-causal ultimate fine-tuning where all of the past is (and can only be) the exact one that gives us the observable present (BUT: that is exactly what we observe) and 2. it removes free will (BUT: I am struggling to find an issue with that idea, although supposedly philosophers and physicists get all up in arms at the thought).
Regarding the previous discussion about non-determinism vs. determinism. I have no money against the non-deterministic horse, but I honestly believe that removing statistical independence is less jarring than the measurement problem.
demagogue on 24/5/2022 at 20:58
Really quickly, I referenced one Sabine Hossenfelder video just as a kind of ex fortiori argument, because she described a common criticism of MWI in a relatively easy way to understand, but I think MWI people would say it was a strawman argument and she was conceding the central part. Showing one of your most notable critics conceding your central argument is supposed to be rhetorically good. "Even a big critic agrees on this part."
I didn't mean to get into her favored interpretation. I mean my posts are already long, so I know I'm skipping over a lot. Comparing superdeterminism and MWI would be interesting. The factor that's contributing to the perception of a "hidden variable" in MWI would be the location or indexical uncertainty. You can't realistically specify every part of the physical state you're in, so you typically don't know which branch you're in.
It's not really variable. Most branches off of you at time 0 will have someone in there, and all of them can credibly wonder why or what it is that makes it them in there in that state. But there is no deeper why then that they are because the state made it so. From their perspective, I think it would appear superdetermined, fine tuned to exactly put them into that state.
So I think the phenomenology would be the same with superdeterminism, and so valuing one over another would have more to do with parsimony, as in the math is much more straightforward to see superdeterminism emerge in one branch, it's just the Schrodinger Equation & friends, than showing a superdetermined set of oscillators that know how to orchestrate just one state for all, which would be a lot more complicated.
Anyway, I like the idea of oscillators being fine tuned to do some jobs, and I think MWI has some of that flavor too. Also, aside from the issue of wave function collapse, there is also the business about distant entangled pairs sharing complementary properties. I think there is some fine tuning involved with that too. But I should get to that later.
Anarchic Fox on 24/5/2022 at 21:11
I might come back later and respond to more of your big post; this is only a partial reply.
Quote Posted by demagogue
o Here’s the thing with the charge of “by saying everything happened”: That’s literally what the path integral form of QM says you have to do.
It really doesn't! The path integral formalism says that the wave function is a sum of contributions from every possible path, with each contribution multiplied by an action-dependent phase factor. Which is neat, and demonstrates the fundamental nonlocality of QM. But then if you make a measurement you're back to wavefunction collapse, using the regular rule of taking the square of the component along an eigenvector. The path integral formulation also has zero to say about measurements where there are only a finite number of outcomes (like spin measurements), unlike the MWI. If you disregard the nuance*, the path integral formulation says that every possible particle path occurs. This is
not the same as saying that every possible
measurement occurs.
*And the nuance is very important. The phase factors attached to the paths tend to cause rapid cancellation everywhere except the stationary path, which is also the classical one. So the path integral formulation has the additional notion of non-classical paths cancelling each other out.
Quote:
Then if the measurement collapses the system in that state, you've got a photon with more energy than the photon(s) that went into the chamber, and would that violate conservation of energy? And if it doesn't, where did the extra energy come from?
Nope. This is a common phenomenon in QM, where the wave function will have a nonzero amplitude in regions that are classically forbidden, i.e. where the potential energy exceeds the particle's energy. It's why quantum tunneling occurs. And in all such cases, if the particle is measured in the classically forbidden region, then it's because the measurement added energy to it.
Quote:
o What's the value of an interpretation at all? It's not going to really change how people "do" QM. The method is probably still going to be what we've always been doing. But I think the two values are:
(1) They’re good mental models for making sense of what the math is doing, to build up good intuitions, even aside from their truth value.
(2) The philosophical thing. Science is a human enterprise to understand more about reality with models we can use. Interpretations don't really add much to that, so textbooks don't need to deal with it.
My attitude is that the underlying physics is weirder than the human imagination can encompass, and most interpretations are an attempt to force the physics into something it
can encompass.
Quote:
o Alright, the "measurement problem". It's just the common label that's used for the discussion.
It's really not, though. It's only called a problem by people who have a solution to offer. The term "measurement problem" will not be found in most textbooks. "Wavefunction collapse" is more common, although I just checked my copies of Shankar and Eisberg & Resnick (everything else is packed), and they don't even have that! (They do have discussions of wave-particle duality, the double slit experiment, all the usual). What you call "the measurement problem," and I'm calling "wavefunction collapse," Shankar simply calls "measurement"! If this discussion is still going by the time I unpack everything (in a month and a half), I can poll about a half-dozen textbooks.
I come back to this because the term itself is prejudicial. It's like calling Newton's 3rd Law "the reaction problem." You've taken one of the basic
postulates of QM and called it a problem.
Quote:
A more useful framing might be: "one of the rules of QM isn't like the others", as in "Everything up until measurement is smooth, no discontinuities, linear / unitary evolution, and then there's this discontinuous, irreversible thing." Okay, fine: nature is as nature does. There's no sense in lecturing her to change her act.
But it does raise a question, what's the relationship between these two types of rules, if any?
The relationship between them is that the Hamiltonian operator occurring in the (time-dependent) Schrodinger equation is the same function of the measurement operators as the classical Hamiltonian is of the phase-space variables.
At this point I became unable to follow your reasoning. I think it's just due to fatigue. I might come back at a later point and try again.
Anarchic Fox on 17/7/2022 at 09:49
I'm not going to offer any more arguments, because I've already made the best ones I can, stated as eloquently as I can, given my current state of knowledge. All I could add now is unimportant detail.
However, I apparently killed the thread dead, and I feel bad about that. So, what do y'all think of parallel worlds, as a trope? Everything Everywhere All at Once is this year's best movie so far, and it is entirely about that concept.
