▲ 72 points
back
125 comments
I think interpretations fall more within the realm of philosophy. What physicists wants are theories that can yield correct and precise predictions. For instance, if we have, let's say, 10 different interpretations of QM, all of them unfalsifiable, yet all of them provide the same accurate physical predictions, then in terms of theory, they are equally suitable approximations of reality. What is of greater interest are innovative theories that can generate improved and more specific predictions, such as those for higher energy scales where gravity comes into play and other areas. Therefore, as a physicist, I would concentrate on developing theories that can yield more accurate and precise predictions.
I wouldn’t say it’s philosophy so much as a split in what you think the goal of physics is. Is the goal to create the most accurate models or to understand how the universe actually works?
The notion that we can fully understand how the universe works may be a overly optimistic belief. There's no inherent reason to assume the universe is structured in a way that human brains can comprehend. Even if we are fortunate enough to have a universe that is understandable by humans in principle, there is no guarantee that a falsifiable theory that explains everything exist. Some people assume that we could eventually formulate a theory of everything or have a complete understanding of QM, but there is no scientific evidence to support this belief. All we have are equations, like the Schrödinger equation, which provide accurate predictions of physical phenomena.
You are describing philosophy:
because we can’t measure the goals anywhere in the universe; the universe (in the form of humans) has to decide with its consciousness what is more valuable (this is philosophy).
The storyline that “science now accepts their ideas” is questionable editorializing imho. Neither Bohmian mechanics nor many worlds are falsifiable, they are just interpretations. They are cool but they aren’t really something you can reject. And to be honest, an unscientific sample of the physicists I know has the majority basically subscribing to some version of “shut up and calculate”
Sean Carroll claims many-worlds is “super duper falsifiable”
“ Right, so speaking of which, look, many-worlds says there is a wave function or a state vector, it evolves all the time under the Schrödinger equation. So all you need to do to falsify the many worlds-interpretation is to do an experiment where the wave function is not under the Schrödinger equation. These experiments are ongoing. Roger Penrose makes predictions that we should see them. There’s other theories of objective collapses that says we should see them. So that’s just one way. Also, you could find evidence for dynamical variables other than the wave function, ’cause those don’t exist in many-worlds. So there’s plenty of ways in which you could experimentally do these things. They’re hard experiments to do, and they may never converge on anything, but they’re there in principle. If you care about the philosophy of it rather than the practice of it, there’s zero question that many-worlds is completely 100% super-duper falsifiable.”
https://www.preposterousuniverse.com/podcast/2021/04/14/ama-...
> all you need to do to falsify the many worlds-interpretation is to do an experiment where the wave function is not under the Schrödinger equation.
Which, Carroll conveniently fails to note, happens every time we make a measurement on a system that is not in an eigenstate of the observable being measured. For example, every time an H polarization measurement is made on a V+ polarized qubit, which happens all the time in quantum computing, the result is either H+ or H-, with 50% probability of each; but before the measurement the qubit was in state V+. The time evolution from V+ to H+ or H- is not time evolution under the Schrodinger equation.
MWI proponents explain this away by saying there is another branch of the wave function where the result is the other one (for example, H- if we observed the result H+), so that the whole time evolution is still under the Schrodinger equation. (Note that "the whole time evolution" here has to include the entire universe, not just the qubit being measured.) But MWI proponents also have to say that this other branch of the wave function is in principle unobservable. In other words, the MWI itself says it is not falsifiable.
There are objective collapse theories that lead to different physical predictions. Copenhagen is not actually a well defined theory because it does not define measurement.
If we measured objective collapse it would falsify MWI.
> There are objective collapse theories that lead to different physical predictions.
Yes, that's true. So far all of the ones whose predictions can be tested have failed the tests. But if one were to succeed, yes, that would falsify the MWI.
In a general, philosophical sense it probably is non-falsifiable. But... where does it end?
What if we took the measurement 5ns, 500ms, 5 seconds, 5 hours later, because of, you know, things. Billions of universes springing up every second does not seem plausible to me and it does sound like an escape hatch: we don't know why it happened that way, so let's just say there is another unobservable universe where it happened the other way round.
Which is nice, I don't have to think about it anymore. But... it kind of reminds me of the flying arrow or a turtle you can never reach. If you are thinking in the same problem space you will never reach the correct conclusion. So, what we really need is the different mode of thought (I wish I would know to suggest something ;) ), not an escape hatch.
it’s not “billions of universes springing up”, it’s just simple wavefunction evolution according to the laws we already know.
“collapse” seems much more arbitrary, especially as we are able to construct larger and larger objects that are best described by wavefunction mechanics. it seems obvious we just become entangled with the experiment through decoherence. idk, it probably is non-falsifiable
> Billions of universes springing up every second does not seem plausible to me
Why? Do you have any reason for this, or is it just a gut feeling?
If it’s just a gut feeling, I’ll note that this was exactly the reaction most of the intelligencia had to Capurnicus and Galileo. Turns out the universe is a caster place than we previously knew.
I mean, obviously. If you let a system interact with the external world that system on its own will not follow the schrodinger equation.
Carroll is suggesting we try to observe this behavior from a sufficiently large isolated system.
> If you let a system interact with the external world that system on its own will not follow the schrodinger equation.
Not only that--even the combined system + what it is interacting with won't.
> Carroll is suggesting we try to observe this behavior from a sufficiently large isolated system.
But you can't, even with the entire universe, if measurements have single results. In the MWI, they don't; any measurement has all possible results. But we don't observe that, we observe measurements to have single results, so the MWI has to jump through hoops to explain that discrepancy away--and in doing so, makes itself unfalsifiable.
That's not very convincing! He's only talking about falsifying it vs extensions of quantum theory. The interesting question is whether it can be falsified vs other interpretations of quantum theory.
Why is no one ever asking how to falsify Copenhagen vs other interpretations? What did Copenhagen ever do to deserve this privilege?
People do think about that. But it's rich of Carroll to claim that MWI is falsifiable, as if it is quantum theory.
Versus all others it’s more parsimonious (as has been noted elsewhere in this discussion) which is typically more desirable as a description of nature. So if it needs less and the others don’t have any more reason or evidence to be taken more seriously as a description of what is happening then shouldn’t it rank first as the best current interpretation of QM?
IIRC the pilot wave theory from Bohmian is good for single particles, but it's difficult to generalize to multiple particles. In my opinion most physicist don't think it's a good idea.
The many-words interpretation of Everett is quite different. I'm still not sure if it's just claiming that the "hallucinated" Hilbert spaces in a measurement are actualy real copies of the "universe".
Disclaimer: I prefer the "shut up and calculate" interpretation, and I hope that "something-something-decoherence" will fix all the problems one day. Perhaps I like "many-words" without knowing.
> I'm still not sure if it's just claiming that the "hallucinated" Hilbert spaces in a measurement are actualy real copies of the "universe".
I'm not sure what you mean by "'hallucinated' Hilbert spaces", but MWI is just the combination of the claims that
- The wavefunction is a (the) real physical object
- The Schrodinger equation always holds
- The Born rule is correct (though there's some hope this might be derivable from QM)
The "worlds" are then a statmech phenomenon.
The main problem with Bohmian mechanics is trying to extend it to quantum field theory
frankly i don’t see at all how decoherence can rescue copenhagen and if anything it seems highly suggestive of a MWI interpretation in practice
Technically, if you find a system that doesn't satisfy the Schrödinger equation, you have falsified many worlds. On the other hand, is the wave function collapse falsifiable? Note that MWI makes less assumptions than Copenhagen.
If you find a system that doesn't satisfy the Schrödinger equation, you have falsified quantum mechanics. The point is that many worlds isn't falsifiable independently of any other interpretation. It's "just" an interpretation.
And it's a very bad interpretation IMO as it fails to provide a convincing interpretation of the probabilities QM produces, which are the main predictive content of the theory. If you believe every outcome happens then one thing happening with higher probability than another loses meaning. The decision theory argument put forward by some many worlds proponents fails to solve this as it just provides a calculation you can do that produces the Born probabilities. It doesn't provide a convincing interpretation of them.
Edit: Also the claim many worlds is somehow more parsimonious in its assumptions than Copenhagen is highly dubious. It assumes an infinite multiverse that is splitting into infinite variants in every infinitesimal instant, in possibly the grossest violation of conservation of momentum and energy (and therefore the corresponding symmetries) that could be imagined.
Of course it would falsify quantum mechanics, because MWI is nothing but the most austere version of QM. But people always speak of Copenhagen as if it was the standard interpretation of QM and everything else must be measurable different from it. In another world, MWI would have been the first consensus interpretation, and people postulating unfalsifiable wave collapses would be the weirdos.
Also, MWI doesn't assume infinite worlds, it predicts them. It's a consequence, not a postulate.
> If you believe every outcome happens then one thing happening with higher probability than another loses meaning. The decision theory argument put forward by some many worlds proponents fails to solve this as it just provides a calculation you can do that produces the Born probabilities. It doesn't provide a convincing interpretation of them.
What further interpretation is needed? To the extent that we have a number we can measure, we have a calculation that can predict it. More interpretation would be nice, but fundamentally every interpretation of QM has this problem, it's not a unique problem with MWI.
> It assumes an infinite multiverse that is splitting into infinite variants in every infinitesimal instant, in possibly the grossest violation of conservation of momentum and energy (and therefore the corresponding symmetries) that could be imagined.
It assumes continuous unitary evolution that preserves momentum and energy in exactly the way you'd expect, exactly the way we already assume they work. Accepting that you can reasonably make calculations about infinitesimal changes is literally the foundation of physics, it's how Newton was able to make a theory of gravity (and people were just as unhappy about it then).
> If you find a system that doesn't satisfy the Schrödinger equation, you have falsified quantum mechanics.
No, you haven't. Standard "shut up and calculate" QM does not say time evolution always happens according to the Schrodinger equation. It only says that happens when a measurement is not being made.
The evolution a system undergoes when you measure it in the Copenhagen interpretation does not follow the schrodinger equation, that's his point.
> Neither Bohmian mechanics nor many worlds are falsifiable, they are just interpretations.
If we accept that (I’m not sure I do[1]), then that must be true of all interpretations, including the once most popular, Copenhagen/collapse. And as valuable is the "shut up and calculate" refusal to interpret anything, I suspect it is at least in part a way to avoid the possible demise of the collapse interpretation. A compromise of sorts.
[1]: https://www.lesswrong.com/posts/DFxoaWGEh9ndwtZhk/decoherenc...
I would argue that “shut up and calculate” is correct specifically because QM is a model. What hubris it would take to believe that we’ve discovered how the universe actually works (and there’s the issue of inconsistency with GR).
And since it’s just a model, it doesn’t matter what “interpretation” you pick, since all models are wrong anyway.
> What hubris it would take to believe that we’ve discovered how the universe actually works
Isn't that what happened with the interpretations that came before many worlds? Weren't most people believing in an actual wave function collapse back then, with basically no one telling them of their hubris?
Genuine question: how soon did the "shut up and calculate" approach, as a refusal to interpret the maths, started getting traction?
Don’t know the specifics about this particular case
However, this is a typical dichotomy in academic/research settings
The interplay between the practical (like “calculate” or “build”); and searching for meaning (like coming up with a new model for some phenomenon)
My impression is that the majority in a particular field, usually focus on the practical, applying known “tried and true” models to novel things
And then, there are only a few who are successful at convincing others of using their models
There are a lot of people who are technically capable of creating new models. But it is incredibly difficult to make them go mainstream
For a layman in physics, what do you mean by “shut up and calculate”, and does it always align with existing interpretations?
All of these different interpretations use equivalent mathematics under the hood, and make identical, accurate predictions about what we will see in experiments. Where people get excited is in the stories they tell about what the math “means”. QM is definitely wierd, which makes the stories fun and hard to articulate. But “shut up and calculate” says that there is no truly satisfying story, all of our analogies are flawed, we will just go in circles for ever. So just calculate. And yes, the math works perfectly. A bit sad but quite pragmatic!
> All of these different interpretations use equivalent mathematics under the hood, and make identical, accurate predictions about what we will see in experiments.
This isn’t true at all, especially the former part. The mathematics of these different interpretations is very different, there are some things you can do with some interpretations but not others, etc. they’re not all mutually interchangeable. For example, in Copenhagen you can’t analyse the interaction between the measurement device and the measured system while MWI can do so, no one has managed to make quantum field theory for Bohmian mechanics, etc.
That’s of course we are a storyteller species. We need narratives to understand the world.
Like with “Particles” or “waves”. These are all words borrowed from our daily experience to describe an abstract reality which isn’t easy to comprehend for our story minds.
Agreed that “science accepts their ideas” is a mischaracterization (indeed Everett annd Bohm are probably mutually exclusive so couldn’t both be “accepted”). Better to say that the community now takes both seriously as contenders for a completion of quantum mechanics.
I deliberately say “completion”, not “interpretation” since, without further elaboration, Copenhagen quantum mechanics is only a piece of a model.
You just rejected them!
There are some biographical glimpses of Hugh Everett's family life in his son's autobiography, "Things the Grandchildren Should Know" https://en.wikipedia.org/wiki/Things_the_Grandchildren_Shoul...
The son, Mark Oliver Everett, is better known as E, the leader of the Eels.
I highly recommend the documentary on Hugh Everett, hosted by his son, Mark:
It's an odd thing to have a bio piece on what amounts to a "failed" phycist — at least while he was alive. Not all of life's stories have happy endings.
"Claims that the standard procedure for testing scientific theories is inapplicable to Everettian quantum theory, and hence that the theory is untestable, are due to misconceptions about probability and about the logic of experimental testing. Refuting those claims by correcting those misconceptions leads to an improved theory of scientific methodology (based on Popper's) and testing, which allows various simplifications, notably the elimination of everything probabilistic from the methodology (‘Bayesian’ credences) and from fundamental physics (stochastic processes)." – David Deutsch
pdf, The Logic of Experimental Tests, Particularly of Everettian Quantum Theory https://www.sciencedirect.com/science/article/pii/S135521981...
I love these sort of paradoxes in physics: multiple models, multiple interpretations, multiple meanings and realities, out of seemingly the same experiments and measurements, sometimes even the exact same mathematical formulas
Sometimes it can definitely be seen as a bit ridiculous, like if maybe a formula is taken to mean something slightly different, it could mean the whole Universe is upside down!
However, sometimes creating alternative models, even if weird when taken at face value, can actually make a difference in making better predictions and even finding new practical applications
In the end, all of our models are made up by us
For anyone interested in other theories that were not mentioned and not satisfied with the unfalsifiability of Many Worlds I can recommend reading Carlo Rovelli's book Helgoland and his Relational Quantum Mechanics[0].
Copenhagen isn't any more satisfiable than Many Worlds.
They both say similar things: that the wavefunction evolves according to the Schrodinger equation. One leaves the equation intact (Many Worlds), while the other involves randomly selecting certain parts of it to be "real" or "unreal" without any explanation of how or why that choice is made (Copenhagen).
https://arxiv.org/abs/2101.11052 claims to have derived an experimentally verifiable prediction from many worlds (namely that energy is only conserved in the linear progression of the wave function, and not necessarily in the "branched" or "collapsed" bit that we observe; and so in certain circumstances it might be possible to observe an interaction that fails to conserve energy.
For anyone interested in MWI in a novel, Stephenson’s Anathem is quite nice. I think I may start a reread this evening.