The “Many Worlds Interpretation” Makes No Sense
People repeatedly insist to me that it’s basically a fact, proven by quantum mechanics, that we live in a multiverse. In fact, even some academics spread this viewpoint on social media platforms such as YouTube, and it has become rather popular there. The people who push this narrative on social media platforms usually claim it is the “simplest” and “most reasonable” viewpoint of the natural world based on our best scientific theory of nature, therefore we should just accept the multiverse as reasonable.
Are these people correct? Well, you have seen the title, so you know what this article is about: No! The Many Worlds Interpretation is in fact more complicated than traditional quantum mechanics and entirely incoherent. Whatever you have been told about the supposed reasonableness of the Many Worlds Interpretation — even, sadly, by people with a real education who should know better — is just patently false.
A Desire for Beauty
One of the biggest selling points of MWI is that, as the advocates claim, it is just quantum mechanics but made “simpler.” This simplicity comes from denying one of the fundamental axioms of quantum mechanics: the Born rule. You see, in quantum mechanics, you compute the motion of particles using the Schrodinger equation whereby they evolve not as point-like particles but like waves up until you go to make a measurement. At the moment of measurement, you then square the wave to get a probability distribution of where the particle will end up once you measure it. This final step is known as the Born rule.
Since this is a probability distribution, naturally, you will also update the probability distribution based on what you actually observe. This is known as the reduction of the state vector, or sometimes misleadingly referred to as the collapse of the wave function. This is just a feature of probability itself and is not an additional “postulate.” There is no “collapse postulate.” The reduction of the state vector follows from the definition of probability, and the Born rule gives you the probabilities.
MWI proponents do not like the Born rule due to the discontinuous nature of it. There is a sudden “jump” of the wave function when you make a measurement. There are two separate problems here that should be separated to avoid confusion.
You see, MWI proponents like to pretend their only competition is the Copenhagen interpretation. They will point out that the jump caused by measurement is arbitrary because, well, what is a measurement? Why should big objects like measuring devices even play a role in a fundamental theory of nature? John Bell had made similar criticisms.
What exactly qualifies some physical systems to play the role of ‘measurer’? Was the wavefunction of the world waiting to jump for thousands of millions of years until a single-celled living creature appeared? Or did it have to wait a little longer, for some better qualified system . . . with a PhD?
— John Bell, “Against ‘Measurement’”
While this is indeed a legitimate criticism of the Copenhagen interpretation, it turns out it is hardly a big deal at all, because you can simply replace “measurement” with “interaction” and you do not run into any inconsistencies. You do not need to posit a whole multiverse to get rid of the reference to measurement in quantum mechanics. Simply replacing “measurement” with “interaction” gets you into the relational interpretation, which many MWI proponents seem to act like simply does not exist.
The founders of the theory expressed this relational character in the “observer-measurement” language. This language seems to require that special systems (the ob- server, the classical world, macroscopic objects…) escape the quantum limitations. But neither of this, and in particular no “subjective states of conscious observers”, is needed in the interpretation of QM. As soon as we relin- quish this exception, and realize that any physical system can play the role of a Copenhagen’s “observer”, we fall into relational QM. Relational QM is Copenhagen quantum mechanics made democratic by bringing all systems onto the same footing.
— Carlo Rovelli, “Space is blue and birds fly through it”
Yet, as I said, there are two problems here, not just one. The other problem is the discontinuous nature of quantum mechanics: the fact that there is linear evolution according to the Schrodinger equation that seems to jump to something nonlinear at the moment of an interaction predicted by the Born rule. By treating measurement as merely an interaction like any others, the relational interpretation expands the discontinuity even further: every interaction has a discontinuous jump.
Is this a problem? Why is this a problem? Why can this not just be how nature works? There is no problem with saying that particle motion is mathematically discontinuous. There is no contradiction in the mathematics nor any philosophical issues making it coherent. The distaste for it is clearly an argument from some arbitrary notion of mathematical beauty rather than a deep philosophical problem in the theory.
Indeed, quantum mechanics had actually been formulated by Werner Heisenberg using matrix transformations prior to Erwin Schrodinger deriving his famous equation. They are mathematically equivalent, but Schrodinger, too, chose to promote his formulation instead of other formulations partially because he found the discontinuity in them to be ugly.
Schrödinger dismissed Born’s probability interpretation…if Born was right, then there was no way to avoid quantum jumps…he wrote to Born: ‘I have, however, the impression that you and others, who essentially share your opinion, are too deeply under the spell of those concepts (like stationary states, quantum jumps, etc.), which have obtained civic rights in our thinking in the last dozen years; hence, you cannot do full justice to an attempt to break away from this scheme of thought…I can’t imagine that an electron hops about like a flea’, he once memorably said.
— Manjit Kumar, “Quantum : Einstein, Bohr and the Great Debate About the Nature of Reality”
Yet, even Schrodinger had come to reconsider this position when he realize that he simply kicked the can down the road so to speak. The discontinuity he removed in Heisenberg’s formulation between interactions was shifted to the moment of measurement. No longer where there “jumps” from interaction to interaction, but there was now a “jump” between the wave formulation described by the Schrodinger wave equation and its discontinuous “collapse” into a particle when measured.
[W]e are so used to thinking that at every moment between the two observations the first particle must have been somewhere, it must have followed a path, whether we know it or not. And similarly the second particle must have come from somewhere, it must have been somewhere at the moment of our first observation…This habit of thought we must dismiss. We must not admit the possibility of continuous observation. Observations are to be regarded as discrete, disconnected events. Between them there are gaps which we cannot fill in. There are cases where we should upset everything if we admitted the possibility of continuous observation. That is why I said it is better to regard a particle not as a permanent entity but as an instantaneous event. Sometimes these events form chains that give the illusion of permanent beings — but only in particular circumstances and only for an extremely short period of time in every single case…The gaps, eliminated from the wave picture, have withdrawn to the connection between the wave picture and the observable facts. The latter are not in one-to-one correspondence with the former.…We must, so it seems, give up the idea of tracing back to the source the history of a particle that manifests itself on the plate…We cannot tell where the particle was before it hit the plate. We cannot tell through which opening it has come. This is one of the typical gaps in the description of observable events, and very characteristic of the lack of individuality in the particle.
— Erwin Schrodinger, “Nature and the Greeks and Science and Humanism”
There seems to be a fundamental discontinuity in quantum mechanics that is impossible to get rid of. If you claim particles do not “hop like fleas” between interactions, then you end up just shifting this discontinuity to the moment of measurement (observation). If you remove it in one place, it just shows up elsewhere.
These difficulties are common to all attempts to formulate quantum mechanics in the language of quantities non-linear with respect to the wave function; throw the latter out of the door and it will come back through the window. This persistence of the wave function is an expression of the fact that the language of the linear theory is that of the essential nature of quantum mechanics; translation into any other tongue will always cause the poetry of this linearity to sound weaker than in the original.
— Dmitry Blokhintsev, “The Philosophy of Quantum Mechanics”
MWI proponents are a continuation of this mindset. They, just like Schrodinger, desperately desire to restore the poetry of continuous evolution. Unlike Schrodinger, however, they do not see the “collapse” of the wave function upon observation as evidence that they should give up on this approach. They are willing to sacrifice anything to achieve this, even if it involves making the theory more complicated and rendering it to be entirely incoherent.
Complexity for Nothing
The Schrodinger equation succeeds in getting rid of discontinuities up until the moment of measurement. At that point, the Born rule has to be applied which involves squaring the wave function in order to compute the probabilities of where a particle may end up. If this is the last discontinuity remaining, then why not, you know, just delete the Born rule? Just get rid of it, allow the Schrodinger equation to continue evolving and just see where that takes you.
When a photon hits a beam splitter, it has a 50% chance of being reflected and 50% chance of passing through. Hence, if you shine a whole beam of photons at it, the beam will appear to split in two directions where the two new beams have half the intensity of the original.
If a single photon is fired at the beam splitter and two detectors are placed on both ends, there would then be a 50% chance of a detection on one of the detectors, and a 50% chance upon the other. After the photon leaves the laser and before it hits the detector, you can imagine it ceases to be a particle but spreads out like a wave. This wave would, in a sense, take both paths, up until it hits the two detectors, then it would “collapse” back down into a single particle one detector or the other.
At least, this is how it is commonly depicted. If you take Heisenberg’s matrix formulation more seriously, then the particle simply jumps from the laser to one of the detectors. There is no continuous wave evolution. It is important to keep in mind that the Schrodinger equation is just one of many mathematically equivalent formulations of quantum mechanics, so how seriously the actual waves should be taken is up for debate. However, we will stick with this wave language for now.
MWI proponents argue that we should keep the waves but delete the “collapse.” It is just not true, they claim, that the wave “collapses” into a particle randomly at one of the detectors. Rather, the wave does indeed interact with both detectors, and both detectors are set off. The claim that the photon has a 50% chance of setting off one detector or the other is therefore false: the wave has a 100% chance of setting off both detectors! There is no longer any probability. There is only the absolute certainty of the completely continuous evolution of the wave. The Born rule is therefore entirely unnecessary.
Oddly enough, a lot of MWI proponents stop here as if they just made a great argument, but it obviously makes no sense. We do not see both detectors going off 100% of the time. We only see a single detector go off, and if we repeat the experiment many times, we find the probability distribution exactly matches the 50–50 split predicted by the Born rule. Clearly, we need the Born rule to make correct predictions! You can’t just delete it, that makes no sense! It was not added there because it’s funny, it is an essential component of the theory.
If you point this out, the common response from MWI proponents is to say that you branch with the multiverse as well whenever there is a measurement. When the wave hits the detector, the universe splits in two where in one universe you are seeing the outcome on one of the detectors, and in the other universe you are seeing the outcome on the other detector. So that is why you only see a single outcome.
On the surface, if you don’t think about it too hard, you might convince yourself that makes sense as an explanation. Yet, it actually evades the question entirely. The question was not “why do I only see one outcome,” but the question was specifically why is there a 50–50 split? Why do we get this very specific probability distribution if not for the Born rule?
A common answer is the notion of self-location. Imagine if you awake and find yourself in an unknown building. You recognize the building enough to know its rough location, and in that location, there are two buildings, so you must be in either one of them. What is the probabilities you would assign that you are in one building and not the other?
The question is actually meaningless because probability always implicitly references an ensemble and so it is meaningless with a single event. David Albert had a lengthy discussion with the MWI proponent Sean Carroll on this specific issue pointing out how probability in such a situation is not even applicable.
However, putting this aside, an intuitive answer may be to just assign equal probabilities to all possible outcomes. Since there are two possible outcomes, then you assign each an equal probability of 50%. In a similar way, when the universe splits, you do not know which branch of the universe you are in, so if you assign equal probabilities in this case of the beam splitter, you can “self-locate” yourself by assigning equal probabilities to each branch, reproducing the 50–50 split. This is also known as the epistemic separability principle (ESP).
If you don’t think about it very hard, that seems to make a lot of intuitive sense on the surface. Yet, the moment you apply any scrutiny, this viewpoint falls apart. Consider, for example, an imperfect beam splitter that has a ~33% chance of reflecting the photon and a ~67% chance of allowing it to pass through. There are two possible outcomes. If you assign each “branch” of this multiverse equivalent probabilities, then you get the wrong answer. Clearly, this simple self-location cannot work.
Indeed, there is no simple way to reproduce the predictions of the Born rule. A lot of MWI proponents pretend like the Born rule can be derived just by some sort of intuitive and preexisting axioms so that, if you delete it, you end up with a simpler interpretation with less axioms. Yet, this is just patently untrue. If you discard the Born rule as an assumption, it is simply impossible to then derive the Born rule without some sort of different assumption.
The problem how to derive the experimental content of quantum mechnics from the abstract framework of the MWI is addressed by Lev Vaidman. He reviews attempts to derive Born’s rule in other approaches to quantum mechanics as well. Vaidman’s conclusion is clear: Born’s rule cannot be derived from the other postulates of quantum theory without additional assumptions.
— Per Östborn, “Born’s rule from epistemic assumptions”
One could try to modify the ESP to develop one specifically for quantum mechanics (ESP-QM), yet doing so always requires making assumptions just as arbitrary as the Born rule itself entirely for the purpose of deriving the Born rule and no other justifiable reason. Deleting the Born rule thus has no benefit to making quantum mechanics “simpler” as you always end up reintroducing some new assumption just as arbitrary as the Born rule itself for no other reason than to derive the Born rule from it.
According to Sebens and Carroll, their result establishes that ‘the Born rule is the uniquely rational way of apportioning credence in Everettian quantum mechanics’. Here we dispute this claim by arguing that Sebens and Carroll misrepresent what is accomplished by their derivation of the Born rule…There is no cogent step that leads from ESP to ESP-QM in a quantum physical context…the plausibility of ESP-QM depends entirely on the empirical success of quantum mechanics (QM) — and therefore, indirectly, on assuming the Born rule. There is no basis for viewing ESP-QM as an independently attractive principle of rational reasoning. Establishing that the Born rule can be derived from it does not solve the probability problem of Everettian quantum theory.
— Carroll Richard Dawid & Simon Friederich, “Epistemic Separability and Everettian Branches: A Critique of Sebens and Carroll”
There is an even more problematic angle to this: a derivation is more mathematically complex. Not only is the number of assumptions in MWI equal to any other interpretation of quantum mechanics (the claim it has less assumptions has been repeatedly demonstrated to be patently false as already shown), but MWI ends up adding greater mathematical complexity to quantum mechanics. No longer can we just assume the Born rule, we have to derive it!
You have to add other assumptions to Many Worlds, about what a detector is and how the universes split and so on, which for all practical purposes amounts to the same as updating the wave-function. In most cases these prescriptions are actually more complicated than the measurement update. So multiverse theories are either simple but don’t make predictions, or they make predictions but are more complicated than the generally accepted theories.
— Sabine Hossenfelder, “The Multiverse: Science, Religion, or Pseudoscience?”
What’s even worse is that it is not just more complex, but entirely metaphysical. Any assumption you introduce to derive the Born rule from, to give an “underlying story,” is not something that could ever be confirmed, so there can be an infinite number of possible stories and no way to distinguish between them. It is like trying to explain general relativity by introducing angels that push on spacetime to curve it, and then arguing over the mathematical properties of the angels. Je n’avais pas besoin de cette hypothèse-là.
Defining a probability– typicality measure using Born’s rule is allowed, but this breaks the symmetry or democracy between the branches; the theory does not contain enough ingredients to unequivocally fix the problem. There are infinite ways for defining probabilities, and the theory cannot decide because there is no way to decide.
— Aurelien Drezet, “An Elementary Proof That Everett’s Quantum Multiverse Is Nonlocal: Bell-Locality and Branch-Symmetry in the Many-Worlds Interpretation”
You can see, now, how the claim that MWI is “simpler” is not just false, but completely false: it is more complicated! And for what purpose? The motivation for making quantum mechanics more complicated is, again, purely an argument based on mathematical beauty. They simply dislike the discontinuous jumps. We already know that we can just replace “measurement” with “interaction” if we wish to get rid of that language without running into contradictions and without positing a grand multiverse. It just requires us to stop trying to get rid of the discontinuous jumps simply because we do not think they’re mathematically elegant!
The Observability Problem
There is yet another major issue with MWI. Even if we accept it is more complicated than other interpretations of quantum mechanics, it at least is coherent, yes? So it is not that unreasonable to believe in it. Sadly for MWI proponents, however, the issues go much deeper than that. MWI is not even coherent.
Consider how humanity discovered magnetic fields. You can drop some iron filings near a magnet and the iron filings will spontaneously take the shape of the magnetic field. You cannot see the field itself, but you can see the effect it has on the iron filings. From that, you can work out the mathematical description of the field itself.
However, imagine if someone came along after you had discovered magnetic fields and told you: “I think the iron filings aren’t real. Only the field exist.” You might be a bit puzzled at this proposal. You derived the field from the iron filings, you had to have seen the filings in order to know anything about the field. In fact, the field is just a mathematical construct, it has no observable properties on its own. Without the iron filings, not only would you not be able to derive the field, there would be nothing visible to derive at all!
Advocates of the MWI are even more bizarre than this. In quantum mechanics, the evolution of the Schrodinger equation is used to predict where particles will be found. The wave equation itself was discovered by analyzing the behavior of particles and is applicable to all particles. MWI proponents then come along and say: get rid of the particles. We only need the evolving waves.
Yet, again, the evolving wave is not visible. It is a purely mathematical construct used to predict together how visible particles move. If you just have the wave, then how do you derive the wave in the first place? Even worse, the Schrodinger equation is applicable to all particles, so what MWI proponents argue is that everything is merely the invisible waves we use to predict the behavior of visible particles. Nothing else exists besides the invisible waves.
If we presume we know of the Schrodinger equation really does represent a real wave, then we only know of this real wave because of the effects it has on visible particles. The wave itself cannot be seen. If you delete all the visible particles and just have the wave, then you would be claiming that the entire universe is composed solely out of an invisible substance with no observable properties. How can such a viewpoint possibly hope to explain the world in which we observe?
The gigantic, universal ψ wave that contains all the possible worlds is like Hegel’s dark night in which all cows are black: it does not account, per se, for the phenomenological reality that we actually observe. In order to describe the phenomena that we observe, other mathematical elements are needed besides ψ: the individual variables, like X and P, that we use to describe the world. The Many Worlds interpretation does not explain them clearly. It is not enough to know the ψ wave and Schrödinger’s equation in order to define and use quantum theory: we need to specify an algebra of observables, otherwise we cannot calculate anything and there is no relation with the phenomena of our experience. The role of this algebra of observables, which is extremely clear in other interpretations, is not at all clear in the Many Worlds interpretation.
— Carlo Rovelli, “Helgoland: Making Sense of the Quantum Revolution”
Quantum fields are both quantum and fields. The “quantum” aspect implies that, when observed, they can only take values in particular discrete states known as eigenstates. These states have observables associated with them and thus are the observable particles we observe. In MWI, quantum fields evolve continuously without ever “jumping” to an eigenstate, and thus it does not even seem very meaningful to talk about where to place the observables.
I recommend watching the lecture below where this problem is explained in much greater detail, but the gist of it is as I have already stated: MWI simply has no observables and therefore makes no sense to say that it can interpret the reality that we observe. It is not simply that it is more mathematically complex that other interpretations, it is hardly even an interpretation at all. It is not even coherent.
