IAI has a brief interview of David Deutsch on his advocacy for the many-worlds interpretation of quantum mechanics. (Warning: possible paywall.) Deutsch has a history of showing little patience with other interpretations, and this interview is no different. A lot of the discussion centers around his advocacy for scientific realism, the idea that science is actually telling us about the world, rather than just providing instrumental prediction frameworks.
Quick reminder. The central mystery of quantum mechanics is that quantum systems seem to evolve as waves, superpositions of many states, with the different states interfering with each other, all tracked by a mathematical model called the wave function. But when measured, these systems behave as localized particles, with the model only able to provide probabilities on the measurement result. Although the measurement results as a population show the interference patterns from the wave function. This is often called the “wave function collapse”.
Various interpretations attempt to make sense of this situation. Many deny the reality of what the wave function models. Others accept it, but posit the wave function collapse as a real objective event. Some posit both a wave and particle existing throughout. The Everett approach rejects wave function collapse and argues that if we just keep following the mathematical model, we get decoherence and eventually the same observations. But that implies that quantum physics apply at all scales, meaning that it’s not just particles in superpositions of many states, but measuring equipment, labs, people, planets, and the entire universe.
Reading Deutsch’s interview, it occurred to me that my own structural realist outlook, a more cautious take on scientific realism, is reflected in the more cautious acceptance I have of Everettian quantum mechanics. People like Deutsch are pretty confident that there is a quantum multiverse. I can see the reasoning steps that get them there, and I follow them, to a point. But my own view is that the other worlds remains a possibility, but far from a certainty.
I think this is because we can break apart the Everettian proposition into three questions.
- Does the mathematical structure of quantum theory provide everything necessary to fit the current data?
- If so, can we be confident that there won’t be new data in the future that drives theorists to make revisions or add additional variables?
- What effect would any additions or changes have on the broader predictions of the current bare theory?
My answer to 1 is yes, with a moderately high credence, maybe around 80%. I know people like Deutsch and Sean Carroll have this much higher. (I think Carroll says his is around 95% somewhere on his podcast.) And I think they have defendable reasons for it. Experimentalists have been stress testing bare quantum theory for decades, with no sign of a physical wave function collapse, or additional (hidden) variables. Quantum computing seems to have taken it to a new level.
But there remain doubts, notably about how to explain probabilities. I personally don’t see this as that big an issue. The probabilities reflect the proportion of outcomes in the wave function. But I acknowledge that lot of physicists do. I’m not a physicist, and very aware of the limitations of my very basic understanding of the math, so it’s entirely possible I’m missing something, which is why I’m only at 80%.
(Often when I make the point about the mathematical structures, it’s noted that there are multiple mathematical formalisms: wave mechanics, matrices, path integrals, etc. But while these are distinct mental frameworks, they reportedly always reconcile. These theories are equivalent, not just empirically, but mathematically. They always provide the same answer. If they didn’t, we’d see experimental physicists trying to test where they diverge. We don’t because there aren’t any divergences.)
If our answer to 1 is yes, it’s tempting to jump from that to the broader implications, the quantum multiverse. (Or one universe with a much larger ontology. Some people find that a less objectionable description.)
But then there are questions 2 and 3. I have to say no to 2. The history of science seems to show that any claims that we’ve found the final theory of anything is a dubious proposition, a point Deutsch acknowledges in the interview. All scientific theories are provisional. And we don’t know what we don’t know. And there are the gaps we do know about, such as how to bring gravity into the quantum paradigm. It seems rational to wonder what kind of revisions they may eventually require.
Of course 3 is difficult to answer until we get there. I do doubt any new discoveries would drive things toward the other interpretations people currently talk about, or overall be less bonkers than the current predictions. Again given the history of science, it seems more likely it would replace the other worlds with something even stranger and more disconcerting.
So as things stand, there’s no current evidence for adding anything to the structure of raw quantum theory. That does imply other worlds, but the worlds remain untestable for the foreseeable future.
To be clear, I don’t buy that they’re forever untestable. We can’t rule out that some clever experimentalist in the future won’t find a way to detect interference between decohered branches, to recohere them (which has been done but only very early in the process), or some other way we haven’t imagined yet.
My take is the untestability of the other worlds means that Everettian quantum mechanics, in the sense of pure wave mechanics, shouldn’t be accepted because we like the worlds, or rejected because we dislike them. For now, the worlds should be irrelevant for a scientific assessment. The only question is whether anything needs to be added to the bare theory, a question, it should be noted, we can ask regardless of whether we’re being realist or antirealist about any of this.
All of which means that while my credence in austere quantum mechanics is 80%, the credence for the other worlds vacillates somewhere around 50%. In other words I’m agnostic. This resonates with the views I’ve seen from a number of physicists, such as Stephen Hawking, Sidney Coleman, John Preskill, and most recently, Brian Cox, which accept the Everett view but downplay the other worlds. Even Sean Carroll notes in one of his AMAs that he doesn’t really care so much about the other worlds, but the physics at the core of the theory.
But maybe I’m missing something. Are the questions I raised above as easy to separate as I’m thinking? Or are there problems with pure wave mechanics I’m overlooking?
SelfAwarePatterns 
Personally, I prefer the objective wave-function collapse interpretation. I think it’s elegant and exciting, and probably the most “naive” realist view, which is a virtue in my opinion. The wave function is there in the maths until it collapses, and it’s there in reality until it collapses.
It also fits my wider idea of reality, which is probably part of why I like it. It fits pretty nicely with my belief that things don’t exist separate from their interactions, and that reality is making it up as it goes along. It also seems to reflect the general trend in nature that things tentatively explore multiple paths before “choosing” one that is optimal, like if you watch lightning in slow motion, or how evolution tentatively explores multiple paths before choosing.
It seems odd to me to answer “yes” to 1. Doesn’t the wave function collapse show that we don’t have everything necessary to account for all the data? We have no explanation for why it collapses the way it does each time, we just have probabilities for it doing so. It may be that there’s no way to account for how it collapses, which seems the answer from taking WF collapse as real and from MW, but I think it’s fair to say there’s a gap in our theory.
But I don’t know very much about quantum mechanics. I did one module on it at uni, but it wasn’t very deep and didn’t get into this stuff. I’ve been meaning to read up on it properly at some point, and even have a book waiting to be read!
LikeLiked by 1 person
I have a few blockers with the physical collapse. One is if it’s reality, it’s a very mysterious thing unlike anything else in physics with no proposed mechanism. The second is that it’s one of the scenarios that requires non-local dynamics. And third, experimentalists have been pushing the boundary on ever larger systems held in quantum superposition, raising the question of when exactly the collapse is supposed to happen. They’ve gotten to tiny macroscopic objects. Of course, that doesn’t mean it isn’t reality, but if so we’re reaching the point where macroscopic objects would be disappearing in it.
Have you heard of Carlos Rovelli’s relational quantum mechanics interpretation? It takes a non-real stance toward the wave function, but otherwise fits a lot of the ideas you discussed. One aspect of it I like is that like Everett, it doesn’t confine quantum mechanics to the micro-realm. Ultimately it doesn’t work for me because it posits that reality is relational (which would be fine) but only exists during the interactions, not before, after, or even in between. I find too many structural gaps in that approach. But a lot of people are enthusiastic about it.
Answering yes to 1 seemed implausible in the early decades of quantum theory. Erwin Schrodinger developed his famous cat thought experiment to demonstrate what the math does if there isn’t a collapse. But David Bohm worked out the beginnings of decoherence theory when rediscovering pilot-wave theory, which Hugh Everett adapted to a wave only view. Decoherence was further developed in the 1970s and 80s by H. Dieter Zeh and Wojciech H. Zurek. In short, if we keep following the math, then as the system interacts with the environment and becomes entangled with it, the interference effects become broken up, fragmented, decohered, with their effects no longer detectable. It explains the appearance of the collapse, but not the collapse itself.
Everett just accepts this as reality. Most contemporary collapse interpretations incorporate decoherence into their framework. But the question is, with decoherence, why do we still need to posit a collapse? The only work it seems to do is get rid of all the other worlds. Again, doesn’t mean it isn’t reality, but the data doesn’t seems to be forcing it anymore.
I’ve reviewed several books on quantum mechanics over the years. At a layman’s level, one of the best is Chad Orzel’s How to Teach Quantum Physics to Your Dog. It’s campy, but has one of the best non-mathy explanations of decoherence I’ve come across.
It’s hard to find an even handed description of the interpretations though. Jim Baggott does a pretty good job with Quantum Reality, although he’s biased against realist interpretations, and hates many-worlds. Sean Carroll provides a many-worlds pitch in Something Deeply Hidden, and Rovelli pitches his RQM in Helgoland. Carroll gets into the math in his second Biggest Ideas of the Universe book, which I own but haven’t read yet. There are also some excellent intro college courses you can watch on Youtube. It’s a fascinating subject that you can definitely go down a rabbit hole on.
LikeLiked by 1 person
I’m ok with it having no mechanism, because I think it may be just that fundamental to reality that there’s nothing deeper. I think it’s the frontier between future and present, possibility and actuality. That’s basically just a hunch though.
Re when collapse occurs and larger objects being held in superposition, I’d have to know the details of the experiments, but I’d guess we could we could see it as a relative collapse i.e. we might have two particles that interact, measure, and collapse one another, but those two together remain uncollapsed relative to a third party. So they may fit example each know they other’s position relative to themselves, but the relative position of both to a third party may remain undetermined.
I’ve heard of Rovelli’s interpretation and I’m meaning to look into it eventually, but haven’t properly yet. What I’ve heard I’m very keen on though.
Part of my difficulty is that I’m pretty set that I need to learn QM via the maths. I figure that’s what they’re all interpreting, so I’d better understand that before getting into the interpretations. But I’ll get round to it eventually, I hope…
LikeLiked by 1 person
Given the history of science, I’m leery of declaring anything fundamental. I think we can talk in terms of something being more fundamental or less. But it doesn’t seem like we can ever be confident that we’ve hit bedrock. And I guess my hunch is that actual bedrock would be composed of very simple elements.
A relative collapse is actually an improved way to look at it. If we also take the collapse, or at least the appearance of the collapse, to be something reducible, then the math gives us a mechanism: entanglement. For any observer, phenomenological collapse happens when they become entangled with the quantum system in question. It seems like a clean answer for the Wigner’s friend thought experiment. (Cleaner than Wigner’s initial solution: https://en.wikipedia.org/wiki/Wigner%27s_friend )
A key fork is whether you think the relative states should be reconcilable (as they are in special and general relativity) or if you’re okay with the idea that they’re not. Reconciliation may drive you toward Everett. (His theory’s first published name was “The Relative State Formulation”.) If you reject the idea of an objective reality that always reconciles, then Rovelli’s Relational Quantum Mechanics may seem more attractive. Although part of RQM’s solution involves taking an antireal stance toward the wave function.
Learning the math is definitely a good idea. As I noted in the post, my own understanding is pretty basic. I know what the variables mean (or at least I have at various points) but don’t ask me to solve any of the equations. But even just that basic understanding changed my views substantially. That said, don’t expect a panacea. Arguments about QM have been going on for a century despite the math.
LikeLiked by 1 person
Clearly I am not as well-read on the subject as yourself, but I have a pretty idiosyncratic take on the whole thing, which you may be able to cure me of. I’m familiar w/ Deutsch’s and Carroll’s and Rovelli’s views, and I find Rovelli’s most useful, and I don’t worry about a definition of “exist”.
Here’s the short version: I have not seen any reason to conclude that there are no hidden variables. Bell’s (thoroughly proven) theorem simply states that you can’t have hidden variables and locality at the same time. I take “locality” to mean particles w/ point-like locations. I’m fine with non-locality. Electrons are not particles, ever. They don’t start as waves and collapse into particles. It’s just that when they interact, they only interact with one other thing. When we measure “where” they interact, we’re just measuring the center of the thing they interacted with. It’s like if we throw a yard stick into a crowd and see who catches it. We don’t learn if they caught it on their left side or right side, or which end of the stick they grabbed. We simply learn who caught the stick and put the stick’s location at that person’s location. The stick does not collapse into a ball, or a particle.
So is there reason to think this is wrong?
*
LikeLiked by 1 person
Hey, we all have our views and the reasons for them. Our best bet is to share those reasons and maybe learn from each other.
I did use the word “exist” above in my description of RQM, but my real beef is RQM’s structural gaps. Why do measurement results have the values they do? In my view RQM simply doesn’t have an answer, except to say that’s just the way things are. (This isn’t unique to RQM. It’s my chief issue with all antireal interpretations.)
“I have not seen any reason to conclude that there are no hidden variables.”
There may well be undiscovered variables. The question is whether there’s anything in the current data motivating them, or if we’re only adding them to force things back into our preconceived expectations.
“I take “locality” to mean particles w/ point-like locations.”
“Locality” has a number of different meanings. In the sense you’re identifying, rivers, mountains, and planets are non-local entities. I used that sense in the post when noting the particle-like nature of quantum entities when measured. But like a wave, the state of rivers, etc. at any point in their extent is only directly influenced by adjacent points.
The main type of locality in question for QM is whether anything can be directly influenced by something outside of its immediate surroundings. In that sense, a wave can be completely local. In the evolution of the wave function, the state of every point is only influenced by adjacent points, which means it has this type of locality.
At least unless or until there’s a collapse. The interactive non-locality of QM becomes an issue if correlations between entangled particles aren’t set until the measurement event. But if all the possible correlations are set, then the interactions can remain local.
“Electrons are not particles, ever.”
I think it’s important to remember that “particle” is just a label. In the pure wave picture, particles are just momentarily pinched off bits of wave. In the particle picture, the wave can be seen as composed of all the different versions of the particle. Six of one, half dozen of the other. Particle = wave fragment.
Overall, as I noted a few posts back, my motivation is to get the most structurally complete understanding I can, at least to the extent motivated by the data. Bare quantum theory seems to be the best bet for that, for now. The antireal interpretations leave too much unaccounted for, and the other realist ones require additional assumptions that seem more motivated by metaphysical preferences than data. At least, that’s the way it looks to me. But maybe I’m missing something?
BTW, commenting on WordPress sites seem to be messed up right now. I’m hoping replying by email works.
LikeLiked by 1 person
I agree with you about 1, 2, or 3. On 2, the question of hidden surprises, it would be very arrogant to rule them out, given the numerous major revisions in science up to this point.
I’m glad you put that in – the “count” of “worlds” is fuzzy at best, and it is possible to view the math as describing a single universe. It’s just that for all practical purposes, the different observations (of spin up and spin down, for example) might as well be happening in different universes. Given some arbitrary cutoff for the weakness of interaction of the two very-nearly-orthogonal vectors representing the spin-up world and the spin-down world, we can say that the two worlds have split. Sort of like, given the arbitrary cutoff of 18 years old, we can say that a human being is an adult. Except that the 18-year-old cutoff carries much more practical importance, and deserves more debate — while there are nanoseconds at most between the low end and the high end of reasonable cutoffs for decoherence.
LikeLiked by 1 person
It would be arrogant to rule out new data changing the picture. And yet I’m struck by how many people in history do it. Niels Bohr, for instance, largely dug in, insisting that quantum theory was complete in the late 1920s and 30s. I suspect some of it was political, trying to ensure his acolytes and allies wouldn’t be denied Nobel prizes or other recognition. But the effect was to suppress exploration for decades, a suppression even Einstein couldn’t overcome. (To be fair, I’m sure WWII had something to do with it too.)
Thanks. Yeah, the worlds / universes is an ongoing difficulty with trying to describe the Everett approach. People think the multiverse is a postulate to solve the problem, rather than a side effect. Or don’t understand that “world” is more a reference to how an observer experiences their subset of the universal wave function rather than a rigorous mathematical concept.
It’s probably why many physicists focus on emphasizing the bare theory rather than the worlds.
LikeLiked by 1 person
Re “Quick reminder. The central mystery of quantum mechanics is that quantum systems seem to evolve as waves, superpositions of many states, with the different states interfering with each other, all tracked by a mathematical model called the wave function.”
I wonder about this. Schrodinger chose (intuitively, speculatively) to use a waveform for his equation, which at the time was behind matrix mechanics in its development. He later showed that his “wave mechanics” was equivalent to the kludgier matrix mechanics so can we also not say “that quantum systems seem to evolve as matrices”? Are we allowing our familiarity with waves misguide our interpretations?
LikeLiked by 1 person
We can say that. Or bring in Feynman’s path integrals. The main thing is we’ll always get the same answers. Six of one, half dozen of the other. Although depending on what we’re trying to accomplish, the process might be much harder or easier.
Deutsch himself uses the Heisenberg picture in a paper to argue that Everettian quantum mechanics is local in terms of separability. I can’t judge whether he succeeds, but I do know trying to demonstrate it with just wave mechanics is much harder and more controversial.
LikeLike
Tim Anderson, who writes on Substack and Medium, argues the many worlds may be a 2nd time dimension. Instead of the world splitting, the time line splits off. The other time lines could influence us but only at the quantum level. The extra dimension could account for apparent faster than light information exchange and perhaps other quantum weirdness.
LikeLiked by 1 person
Sounds interesting. Looks like his post is paywalled, although I was able to read the beginning. I completely agree with this point.
I also notice he has a personal blog on his website that mostly seems about religion. I’m not religious, but I sometimes wonder what a theology based on an Everettian cosmology might look like.
LikeLiked by 1 person
I’m not a fan of his religious side and I can’t see how any kind of religious beliefs can come out of physics.
LikeLiked by 1 person