Phillip Ball has a long aggrieved essay about the Many-Worlds Interpretation, which is, as Sean Carroll notes, pretty bad. Ball declares that Many-Worlds is “incoherent, both philosophically and logically,” but in fact, he’s got this exactly backwards: Many-Worlds is, in fact, a marvel of logical and philosophical coherence, while Ball’s objections are incoherent and illogical.
The fundamental problem with Many-Worlds is that every argument about it devolves very quickly into stoner dorm-room bull session nonsense about parallel worlds and identity and morality. But none of that is physics. It’s all science-fiction window dressing touted to sell books. Human identity and consciousness are not subjects that physics can say anything sensible about– cognitive scientists can barely say anything sensible about identity and consciousness, and that’s what they do for a living.
Physics is about fundamental rules and interactions, and those are the only things you can treat rigorously. And Many-Worlds is, in fact, a rigorous and logical treatment of these issues. It’s most clearly explicable in thinking about a single particle with two possible states. Could be an electron that’s either spin-up or spin-down, or a photon that goes through either the left slit or right slit of a Young’s double slit experiment. Doesn’t really matter.
The sticking point is that quantum theory predicts that single particles can exist in multiple states at the same time– the electron can be both spin-up and spin-down, or the photon can go both left and right. This is weird, but experimentally confirmed, because we can do interference experiments that unambiguously show the contribution of two distinct states– a photon going through a single slit makes a single stripe, but open a second slit, and you see an interference pattern characteristic of waves passing through both slits.
Of course, while quantum mechanics excels at predicting the probability distribution for the final states, in everyday life, we only see a single outcome of a given measurement. This is where the various interpretations of quantum physics come in. Collapse-type theories, of the sort evidently favored by Ball, hold that something happens in the detection process that physically forces the particle to choose one final state, while Many-Worlds and its variants hold that the particle continues to occupy multiple states at the same time, and the superposition simply expands to incorporate the detector– so now you have an electron that’s both spin-up and spin-down and a detector that’s showing both spin-up and spin-down, with those states entangled together in the appropriate manner.
If everything in the universe exists in multiple states at once, though, why don’t we see that? This is where the solid philosophical grounding of Many-Worlds becomes apparent, because an essential part of the whole business is that you can’t say what the state of a thing “really” is without specifying how you measure that. This is an idea that’s central to a lot of the best physics we have– I’ve been teaching from Peter Galison’s Einstein’s Clocks, Poincaré’s Maps again this term, and a point I hammer on to the students is that relativity comes out of the realization that you can’t talk meaningfully about simultaneity without explaining how you establish the timing of events. Poincaré goes on at length about this in The Measure of Time, and Einstein’s 1905 paper on Special Relativity makes the same point. Once you realize that simultaneity needs to be defined via a concrete mechanism, relativity is inevitable.
And this is what (my preferred understanding of) Many-Worlds does with the measurement problem. The question “Why don’t we see superposition states?” is meaningless without specifying how you would detect that a particular object is in more than one state. But we already said that: You measure some sort of interference effect that arises from two separate components of the wavefunction.
How do you measure an interference effect? Well, you look for some oscillation in the probability distribution. But that’s not a task you can accomplish with a single measurement of a single system– you can only measure probability from repeated measurements of identically prepared systems.
If you’re talking about a simple system, like a single electron or a single photon in a carefully controlled apparatus, this is easy. Everything will behave the same way from one experiment to the next, and with a bit of care, you can pick out the interference pattern. As your system gets bigger, though, “repeated measurements of identically prepared systems” become much harder to achieve. If you’re talking about a big molecule, there are lots more states it could start in, and lots more ways for it to interact with the rest of the universe. And those extra states and interactions mess up the interference effects you need to see to detect the presence of a superposition state. At some point, you can no longer confidently say that the particle of interest is in both states at once; instead, it looks like it was in a single state the whole time.
And that’s it. You appear to have picked out a single possibility at the point where your system becomes too big for you to reliably detect the fact that it’s really in a superposition. Ball’s complaint that Many-Worlds needs an explicit criterion for “when the universe splits” is the incoherent and illogical demand, because it’s not properly defined. There’s no magical difference between combined and split universes, there are just superpositions that are small enough to measure and superpositions too big to be measurable. How big is that? Well, there’s not a sharp line because that’s not a meaningful question in the abstract– there’s only the question of what you can measure, and how you can measure it. Without a clear and rigorous definition of those things, you’re in the same situation as electrodynamics before relativity, floundering around because the idea of absolute universal simultaneity is fundamentally incoherent.
What about the parallel worlds and the differences in identity and morality? Who cares? All that stuff is just a collection of foggily defined emergent phenomena that arising from vast numbers of simple quantum systems. Absent a concrete definition, and most importantly a solid idea of how you would measure any of these things, any argument about theories of mind and selfhood and all that stuff is inescapably incoherent.
——
(I should note that this goes for Many-Worlds popularizers like Max Tegmark, as well– Ball is correct that most of the pop versions of Many-Worlds don’t make a great deal of sense. But challenging pop versions of Many-Worlds on the basis of pop philosophy of mind just throws incoherence against incoherence like a battle scene in a comic-book movie. It generates a lot of noise, but doesn’t really amount to much. If you want to claim that Many-Worlds is philosophically incoherent as physics, then you need to talk about physics, and that demands a level of rigor that’s missing from Ball’s piece.)
(I’m also shorting the other interpretations a bit here, mostly for reasons of space. As Sean points out in his post, there are, in fact, versions of collapse theories that make sensible and rigorous statements about how things work, and hint at ways we might be able to test them. There are also consistent theories in the Bohmian vein that approach the whole business from a very different angle, but again ground their work properly in discussions of what you can measure and how you can measure it.)
(Finally, the above probably sounds more strongly in favor of Many-Worlds than my actual position, which shades toward agnosticism. But nothing makes me incline more toward believing in Many-Worlds than the gibberish that people write when they try to oppose it.)
Agree that MWI is philosophically and logically coherent, even simple and elegant. It’s just not science.
You can certainly make the claim that in the absence of testable predictions, Many-Worlds isn’t proper science. But it’s really hard to thread the needle of declaring Many-Worlds unscientific on those grounds but keeping any of the other interpretations as proper science. You can argue that objective-collapse theories can in principle make testable predictions, but we’re still a long way from actually testing any of those. And to the extent that those theories do in principle predict something different than Many-Worlds that in itself constitutes a sort of prediction for Many-Worlds (namely “Many-Worlds predicts: not that”).
I’m not sure I understand your preferred version of the many worlds interpretation, Chad. (I’ll admit I’m not sure I understand *any* version of the MWI, to be honest.)
How does it cope with the doyenne of wave function collapse, Schrodinger’s cat? Because the little I understand of the MWI seems to imply that both states (decay/not decayed) should be carried forward in 50/50 representation, one leading to “live” and one leading to “dead”. I’m not sure how talk about interferece effects relates to detecting if a cat is breathing – or am I making an assumption which isn’t justified?
If cats are too “squishy”, how about a slightly more “physicsy” scenario: You have a particle which you’ve prepared in a 50/50 superposition of up/down spin. You pass the particle through an apparatus with a magnetic field such that a spin up particle will curve right, and a spin down particle will curve left. At the appropriate location on the other side of the apparatus you have two identical detectors, each with sensitivity such that a single non-superimposed particle is sufficient to trigger it. Each detector is linked to a separate LED, which turns on when the detector is triggered. You pass a single superimposed particle through the apparatus, and the right detector’s LED turns on. How is that event interpreted in you preferred version of the many worlds hypothesis? Or is there something about this scenario that’s subtly nonsensical?
Each LED is in a superposition of on and off states, entangled with the state of the electron, which is in a superposition of “curved left” and “curved right,” entangled with the spin. Nothing happens to the original spin superposition, it just becomes embedded in a much larger entangled state, which includes whatever apparatus you’re using to measure the state.
Now, you’re free to choose to consider only a subset of that ever-expanding entangled superposition state, for the sake of convenience when trying to calculate stuff. So, you can calculate the probability of ending up in the particular piece where the LED on the right is lit, and continue the calculation using just that piece. But that’s no different than, say, calculating the energy levels of an isolated hydrogen atom without worrying about the fact that, (if you’ll forgive a detour through Brian Cox territory) the electron in any real hydrogen atom technically needs to be in an antisymmetric state with every other electron in the universe. While technically true, that doesn’t have any measurable consequences, so you can safely ignore it for the sake of being able to get things done.
The key question is whether there are observable consequences of that superposition. In order for there to be observable consequences, you have to be able to do some measurement– most likely involving an interference effect– that would clearly show the influence of other pieces of the superposition. Doing that means repeating the experiment many times with identically prepared particles, but as the system gets bigger, fluctuating interactions with the environment smear out any effect you might be looking for. So it becomes practically impossible to observe any consequences of the rest of the superposition. Which is what allows you to pick out a particular branch in the first place.
Now, if you want to blow the minds of people who only think about this stuff casually, you can choose to call that isolated piece of the wavefunction a “world.” I really wish you wouldn’t, though, because that creates more problems than it solves, leading to endless arguments about how Many-Worlds must be bunk because how can you create a whole new universe worth of matter every time you measure an electron spin, and so on.
This is kind of similar to the previous question but rather then calling into question Many World,
Why do any of the arguments that you are putting forward at all apply to specifically why I should look to many worlds as being a better interpretation of what is going on with quantum mechanics.
Don’t all the things you’ve said apply to the Copenhagen interpretation just the same.
For example:
“there are just superpositions that are small enough to measure and superpositions too big to be measurable”
and so sometiems we see the wavefunction collapse and sometimes we see the most probable option behave as if its the only option. Or something like that
(Note: I’m not saying I prefer copenhagen or anything nor do i dislike many worlds, just trying to make sense of your argument)
Agreed, speculating about duplicated selves and suchlike is so much narcissism that only obscures the issue.
1) The (admittedly somewhat-ignorant-layperson) question I’ve had all along about MWI is:
Where do you get the energy needed to produce the objects that exist in the unobserved branch of the “split”? (Even assuming that all that splits is an electron at the moment of measurement.)
From your article it appears that the entire concept of splitting even a particle + apparatus sized chunk of the universe is “not even wrong.” What I think you’re saying is: The correct interpretation of MWI is that that the observable world exists in a constant superposition of states, and measurements only select for one outcome out of the superposed state. The other outcomes don’t actually exist in any material sense.
2) Is that (the preceding paragraph) approximately right?
—
3) Next, some wild speculation. I really do want to know if you think the following is reasonable or crap (no need to mince words, I’m not attached to this stuff;-):
Assume a second axis of measurement of time, described as “distance or difference from observed timeline.” In the timeline we observe, a photon goes through slit (a). In another potential timeline that diverges from ours at the moment of measurement, a very slight distance away from the timeline that we observed, the photon goes through slit (b).
The scale of measurement of the 2nd time axis is the quantity of energy that would have caused it to become the actual timeline. For example an infinitesimal (too small to measure at present?) quantity of energy added to (or subtracted from) the system, would have caused the photon to go through slit (b).
These alternate timelines are similar to, but not the same thing as, probabilities. So we might say, “there is an X percent chance of a random thermodynamic fluctuation (or whatever hypothetical source of energy added or subtracted) occurring at the moment of measurement,” that in turn causes the measurement to occur as observed.
I’m inclined to think this could be “useful” if we can quantify the energy differences (or other differences) needed to produce different outcomes. That might give us our scale of measurement for the hypothetical second axis of time.
The way I represent this as a thought-experiment with classical elements is: the independent variable is the difference in electrical power in milliwatts, passing through a solenoid that kicks a billiard ball; and the dependent variable is the distance the billiard ball rolls on a flat surface. The 2nd time dimension (“T2”) for this demonstration, is the plot of energy input in milliwatts needed to produce measurable differences in the final position of the billiard ball.
OK, so whose “physics 101” experiment did I just unknowingly replicate, and what piece of mundane theory did I just twist into a pretzel to make a point that otherwise doesn’t hold water?;-) Or is there anything new and potentially useful about a T2 dimension?
—-
BTW, it’s axiomatic to my way of thinking, to say that one can’t know the state of a thing without also specifying how it’s measured. This also applies in macro, such as in engineering.
James @ #5: Why do any of the arguments that you are putting forward at all apply to specifically why I should look to many worlds as being a better interpretation of what is going on with quantum mechanics.
To the extent that Many-Worlds is a “better” interpretation, it’s an Occam’s Razor sort of thing– you can see this clearly in Sean Carroll’s post linked above. The main selling point is that the interpretation doesn’t add anything beyond the normal axioms of quantum mechanics: that quantum systems are described by wavefunctions with a certain class of mathematical properties, and that those wavefunctions evolve according to the Schrodinger equation.
“Collapse” interpretations in the Copenhagen mode have a real, discontinuous change take place at the instant of measurement. Thus, they need an additional postulate, something explaining that (and also ideally how) the collapse occurs. If you’re concerned with philosophical and mathematical elegance, that’s a big strike against these interpretations. Of course, it also provide an in-principle possibility of measuring the behavior of whatever mechanism accounts for the collapse, which is one of the reasons why some people prefer these interpretations.
G @ #6: From your article it appears that the entire concept of splitting even a particle + apparatus sized chunk of the universe is “not even wrong.” What I think you’re saying is: The correct interpretation of MWI is that that the observable world exists in a constant superposition of states, and measurements only select for one outcome out of the superposed state. The other outcomes don’t actually exist in any material sense.
I would agree with that, with the caveat that “in any material sense” is a little vague. There’s certainly no creation of “copies” at any point in the Many-Worlds picture as I prefer to think about it. The total amount of stuff in the universe is finite and fixed, and the “other outcomes” are just bits of an ever-expanding entangled state.
In another sense, though, it would be a mistake to say that those other outcomes aren’t “real.” They have exactly as much reality as any other piece of the giant wavefunction of the universe, including the one we “actually see.”
I have to admit, I don’t really follow what you’re getting at in your #3 example, but it’s very early in the morning here…
Chad, it seems to me that you’re just kicking the can down the road with that explanation. I’m total okay with having the entire apparatus be in a single wave function, with the two LEDs being in an entangled state, superimposed in an anticorrelated on/off superposition. Raises no eyebrows from me.
However, all of the accumulated experience from these sorts of experiments indicates that 1) it’s possible to talk about the result of a single run with a single particle and 2) for that single particle run, when the experimenter observes the LEDs, they will see only a single LED lit, rather than observing a superposition of states (whatever that would mean). That’s what I’m really asking: what happens with the MWI that makes the equally probable superpositon of spin states resolve itself into that recognition of the single lit LED in the experimenter head, followed by them shouting down the hall “Hey, Chad, the right one lit up!”
As I understand the MWI, we keep extending the range of the superimposed/entangled states. So the superimposed light from the LEDs result in superimposed states on the experimenter’s retina, which result in superimposed neuronal signals, which result in superimposed vocal cord and sound waves and ear drums and neuronal signals in Chad. So now, theoretically, you have two versions of Chad, one in which he heard “left”, and one in which he heard “right”, and the wavefunction for each keeps evolving independantly. But experience and experiment tells us we only ever observe one. Why? One you get into mental states of physicist, you go past “computational convenience”. Obsevation is the cornerstone of experimental science, and if we throw that into the woodchipper, we might as well take up banking.
Hi Chad- Thanks for taking the time on this, it’s really helpful.
If I understand you correctly: In MWI, the wavefunction continues to exist and entail all possible values or outcomes of a measurement, while the measurement that is actually made selects which value/outcome is observed. Contrast to the Copenhagen interpretation where the wavefunction collapses to the observed value at the moment of measurement. Is that any closer to correct?
Stated that way, the “millions of MEs” problem never arises in the first place. I never thought “millions of MEs” were possible to begin with, which is the popular understanding of MWI that I thought pretty well disqualified it, and as it turns out, the popular version was completely wrong.
The reason I’m trying to “get” MWI now, is that if Hawking et.al. think MWI is correct, there’s a decent probability he’s right, so I need to catch up. Until now I’ve strongly sided with the Copenhagen interpretation, but if I’m understanding MWI correctly, then I end up switching to agnosticism about which is correct.
The question appears to boil down to whether the wavefunction collapses at measurement (Copenhagen) or whether it continues to exist (MWI). If that’s the case, then I’m clearly headed for agnosticism unless there are empirical findings (I’m not aware of any) about this, or unless the math works better one way than the other (which I’d take on trust since I couldn’t understand the math myself).
Re. the T2 dimension thing: that was intended as another way of characterizing the difference between observed outcomes and other potential outcomes, in terms of diverging timelines and the quantity of energy needed to alter the performance of a system such that is comes to resemble the hypothetical result from a different timeline.
The idea is that a small difference in some input to a system causes selection of the observed timeline from one of many possible timelines (for example electrical energy into a solenoid that converts it to kinetic energy that kicks a billiard ball along a ruler). As events in a set of timelines evolve toward observable outcomes (e.g. final position of billiard ball), the timelines increasingly diverge. The quantity of energy added to (or subtracted from) the behavior of a system to produce different observed outcomes might be a useful scale of measurement of the T2 dimension.
What I’m looking for here is a way to characterize the difference between observed outcomes and other outcomes that would have been possible at various points prior to the observation or measurement. The multitude of timelines would roughly represent the evolution of the wavefunction, and the observed timeline would represent either wavefunction collapse under Copenhagen, or a selection process under MWI.
Anyway, if that’s just mush, by all means say so. Bottom line is, I’m looking for thought-experiments that are useful for understanding MWI and also explaining it to friends, so if you have any that are accessible, I’m all ears. (And it’s also too early in the morning for me to be writing this, so I’ll be back this evening to see where I screwed up;-)
I’m keeping Many Worlds on the same shelf as String Theory, right by the Bibles.
While I certainly think the criticism of that author is justified, I am disheartened by the fact that none of you MWI proponents/agnostics don’t try to answer real criticism like the fact that the Born Rule seems to rule out MWI. And there is no answer to the preferred basis problem, everyone just screams “DECOHERENCE”, but decoherence is not sufficient like many scholars have pointed out in numerous papers. Not to mention like David J Baker showed: decoherence is itself dependent on the Born Rule, so it all becomes circular.
Usually when I hear about many worlds I hear talk about the universe splitting on each measurement. This brings to mind a tree like structure of worlds. I think this confuses things.
Instead think of a vast space or continuum of worlds. What we know about our world locates us in that space and probabilistically determines our path through that space. But we can’t ever fully exist at a point in that space and so there are interference effects from near by spaces.
But does that mean all those other spaces exist? Well what do you mean “exist”? In the sense that those other worlds have interference effects they exist. It can be useful to think of them existing. Beyond that you will have to define exactly what you mean by “exist” to get a meaningful answer.
Does the past exist? Does the future exist? Relativity encourages us to thing of the past and future as an existing geometric structure. But is it or is this just a useful stance? Again you will have to pick a definition of exist that allows a meaningful answer.
It is useful to talk about many worlds existing. It is useful to talk about the past existing. If you are an IP lawyer it is useful to talk about superman existing. The word “exist” is a mere tool, a pointer to a vast constellation of meanings. Don’t treat it as a Platonic essence.
I doubt that you could ever visit any of the other worlds any more than you can visit the past. That places strict limits on the utility of claiming that they exist.
David,
“…none of you MWI proponents/agnostics don’t try to answer real criticism like the fact that the Born Rule seems to rule out MWI. And there is no answer to the preferred basis problem, everyone just screams “DECOHERENCE”…”
The Born Rule cannot rule out MWI since it is just a recognition that we only see one small branch of the universal wave function. That is true of all interpretations of QM. The Born rule recognizes the fact but makes no attempt to explain it. The MWI tries to explain it by claiming that all branches of the wave function exist and see themselves as the result of the Born rule.
Decoherence is simply an aspect of the fuzziness of our position in the continuum of worlds. On a large scale interactions are so vast that probabilities look classical rather than the quantum wave like look of the small scale. Essentially all branches of the universal wave function will appear to have classical looking probabilities at the large scale.
I let these comments slide for long enough that I’m moving the response to a new top-level post. Please go check that out.