When Hugh Everett, a PhD student, wrote some bizarre papers about "alternative foundations" of quantum mechanics in the late 1950s, he was allowed to speak about it in Copenhagen etc. Niels Bohr et al. saw that there was no valid physics in the papers whatsoever. The chap only wanted to be a critic whatever it costs, regardless of the absence of any evidence, and Bohr knew it was too little. He discussed these things with the chap's adviser, John Wheeler, and everyone agreed that this guy shouldn't continue as a professional physicist simply because he didn't have enough talent and understanding for that.
In the following 55+ years, there has been no development that would strengthen the case for Mr Everett's musings. On the contrary, the experimental evidence and the modern crystallization of the quantum mechanical insights have made it increasingly clear that these musings were fundamentally wrong and proper quantum mechanics is the only viable theory to describe the relevant phenomena – which are really all processes in Nature.
However, what has changed is that the Academia has been literally flooded by mediocre, incompetent pseudo-intellectuals similar to Hugh Everett – usually much worse and more inadequate than Everett himself. Sean Carroll is a textbook example of this sign of the degradation of the contemporary Academia which is why he wrote another text defending the indefensible,
The illustration of choice used to summarize and promote this "interpretation" of quantum mechanics is this Wikipedia picture of Schrödinger's cat (a four-legged cat; Schrödinger would simultaneously maintain several two-legged ones, too).
What is this picture supposed to convey?
It suggests that there are two worlds (or two similar copies of our world), and not just one, after a cat is observed. In one of them, the cat is alive. In the other world, it is dead. This childish picture is supposed to solve some "problems" that the defenders of this picture believe to exist in quantum mechanics. There are no problems in quantum mechanics, of course.
At the same time, salesmen like Carroll offer you lots of incredible statements such as the statement that this "many worlds interpretation" directly follows from quantum mechanics, is directly justified by quantum mechanics (more justifiable in quantum mechanics than in classical physics), and unlike proper quantum mechanics, it doesn't introduce any new physical laws. All these statements are untrue. They are really the polar opposite of the truth.
First, many worlds surely don't follow from quantum mechanics.
Quantum mechanics is the universal framework of modern physics which is a natural science. As every natural science, physics predicts or explains the observations that are actually being made in one Universe. Deterministic physics (which may only work approximately in Nature) gives unambiguous predictions. Quantum physics (which is valid everywhere in the real world) predicts phenomena by quantifying the probabilities of different outcomes. Outcomes with probabilities close to 100% will probably happen, those close to 0% will probably not. The precise value of the probability may be extracted from the experiment if we repeat it many times (but using the knowledge of probabilities, one can make precise predictions even if we perform many inequivalent experiments!).
The calculation of the probabilities is the quantum mechanical output replacing the unambiguous classical predictions. If you can't say anything about the values of probabilities of some phenomena, then you have no scientific theory whatsoever!
This is the case of the many worlds "interpretation", too.
The many worlds "interpretation" is an artistic picture that simply has nothing to do with science. Note that the picture with the split film of the cat doesn't tell you that the cat will survive with the 73% probability or anything like that. It just tells you that there are two possibilities. It shows you the set of potential histories, not any nontrivial information about the actual history.
And even this information-free statement is really wrong or demagogic; and it can't be derived from the many worlds "interpretation", anyway.
It is wrong because there aren't just two possible outcomes. There are infinitely many outcomes. The cat also has a nonzero probability to morph into a small version of Elvis Presley and be hit by some air molecules that bring it or him to the Moon. The probability is small but it is nonzero. (I could invent more likely outcomes "in between", too.)
So the film should really be split into infinitely many strands at each spacetime point, including strands describing extremely unlikely histories. (In fact, we can't even say that the splitting would occur at each spacetime point because some observations are nonlocal in character. There exist no well-defined rules that would allow us to describe the "cardinality" of the tree at all.) The very fact that only two "reasonably likely" histories were cherry-picked is nothing else than demagogy. Such a qualitatively different treatment of "reasonably likely" and "very unlikely" outcomes is only possible if we first calculate the probabilities and then use different treatments according to their value (and this decision is always about being useful, not being true; fundamentally, Nature doesn't treat high and low probabilities qualitatively differently).
The many worlds "interpretation" doesn't allow us to calculate – or imprint – the generally different probabilities of different outcomes into the "strands" of the film. Even the fans of this religion admit it's the case but some of them say that they are "working on a fix" which is supposed to be enough. (A similarly "modest" fix makes Genesis compatible with all the detailed data about the cosmic microwave background, the DNA, and genetics.)
What they don't realize is that the inability of this theory or "interpretation" to calculate any probabilities isn't just a moderate vice or disadvantage. It is a complete, rigorous proof that this philosophy has nothing to do with the empirical data or science whatever and people who are defending it don't have the slightest clue what they are talking about. Everything that modern science predicts are probabilities or their functions. If your ideas don't predict any probabilities, they don't predict anything at all. They have nothing to do with science.
Moreover, there can't be any fix. There can't even exist a candidate theory that would "extract" the probabilities from something else. In proper quantum mechanics, probabilities are fundamental which means that they are pretty much directly predicted from the mathematical formalism. There is no useless "intermediate agent" that would extract probabilities from something else. There can't really be any. A theory is either exactly deterministic (plus admitting complete knowledge in principle) or it is not. If it is exactly deterministic, it is possible to produce unequivocal predictions out of the information about the initial state, at least in principle. That's the case of classical physics. In this setup, all predicted probabilities are either 0% or 100%. That's no good for the experiments backing quantum mechanics where almost all the probabilities are strictly between 0% and 100%.
Or the theory may fail to be strictly deterministic, or prohibit the observers from knowing the exact required information, even in principle. In that case, the probabilities may be in between 0% and 100%. But such continuous, a priori arbitrary probabilities may only be calculated from other probability-like continuous quantities that are present in the fundamental formalism of the theory such as the classical probability distributions on the phase space or the density matrix (which may be obtained from a state vector if the state is pure).
The very illustration with the split film of the cat as well as many explicit statements by the many worlds apologists make it absolutely clear that one of the key assumptions of the whole many worlds movement is that probabilities are not fundamental. But it's just mathematically impossible to derive a priori continuous numbers interpretable as probabilities from a formalism that rejects any fundamental quantities with a directly probability-like interpretation! Probabilities may only be calculated from other continuous numbers of a similar kind and all such continuous numbers are probability-like because probabilities are their functions! If there aren't any, your formalism just can't be capable of predicting probabilities.
Note that this simple argument – pure maths – really eliminates all attempts to replace quantum mechanics with some ideas where probabilities are not fundamental. This includes the Bohmian and other hidden-variable theories. The pilot wave theory needs to assume that the classical particle it envisions has a position in the initial state that is statistically distributed according to the wave function (renamed as guiding wave). Only the correct time evolution of the probabilistic distribution is guaranteed by the mechanical equation of motion including the effect of the guiding wave.
They don't have any mechanism explaining the right probabilistic distribution of their classical particle in the initial state. Of course, there can't be any. They prefer not to talk about this point at all even though this point really means that their claim that they have derived the "random generator" of quantum mechanics from something else is completely fraudulent. The same "random generator" must still exist somewhere to produce the classical particle in a position that is correctly distributed. It must be distributed according to the wave function which proves that if this condition is guaranteed, the wave function has a fundamentally probabilistic interpretation, anyway!
But let me return to the many worlds. Its defenders often say something like
Proper quantum mechanics dictates the evolution of operators in the Heisenberg picture (or, equivalently, the direct Feynman path integral formula to calculate the probability amplitudes; or the evolution of the state vector in Schr̦dinger's picture Рthe latter picture is the most popular approach among the superficial and confused people). The complex probability amplitudes may be directly squared (in absolute values, using the right Born's rule etc.) to calculate the probabilities of various observations or outcomes or perceptions etc. Quantum mechanics works totally universally. It doesn't matter what object we want to observe and which aspect of it we measure. Quantum mechanics predicts the probabilities using the same universal formulae and the time evolution is always given by the same equations, too.
What about the many worlds "interpretation"? Just look at the marketing illustration at the top. Aside from Heisenberg's or Schr̦dinger's dynamical equations, there is a new process added on top of it Рthe splitting of the film, the splitting of the worlds to make them "many". When writing this promotion of "many worlds", Carroll must have forgotten about something, about the "many worlds", right?
Believe me that to produce a film that splits into two in the triangular junction is much harder than to produce an ordinary film with a single story of a cat. It's also harder to insert such a film with "junctions" into a film projector. The film projector sometimes jams, and so on. It doesn't work well. Try it if you don't believe me.
Now, to produce another copy of the Universe is even harder than to produce a plastic foil with junctions. It's even a bit harder to produce many copies or \(\exp(10^{120})\) copies you might need to describe the multiplicity of outcomes of every arrangement of processes that may occur within the patch of the visible Universe. But OK, you might rightfully say that it's not your job to produce copies of the Universe. Nature can do it easily.
Yes, Nature can do anything if the laws of physics say that it should happen. The problem is that one can't even define the rules when the hypothetical splitting should take place and how the two or three or \(N\) or \(\infty\) (also unknown) copies attached to the junction should differ from each other. And the problem isn't just that we don't know how often and finely the film should be "split" etc. We may easily see that whatever the rules are, they contradict some known facts about Nature. Why?
The really key feature of the quantum evolution is that it always allows parts of the state vector to interfere in the future and affect various other measurements in non-classical ways.
We know this even from the double slit experiment which, as Feynman observed, contains all the qualitative wisdom about quantum mechanics if you think about the experiment carefully enough. We can't assume that the particle goes either through slit A or slit B (with the strict, classical assumptions about the alternatives) if we don't observe it. The actual initial state implies nonzero probability amplitudes for both intermediate histories and their mutual interference affects the existence and location of the interference minima and maxima on the photographic screen!
We can never strictly say that some observable (like the "which slit" information about position) took a completely specific value in the classical sense. The eigenvector may always interfere with another one. Now, if you think about it for a second, this totally elementary and universal feature of quantum mechanics in general and of the double slit experiment in particular is nothing else than the statement that the worlds just never split in the classical sense!
The cat may be in the state \[
\ket\psi = 0.6 \ket {\text{alive}} + 0.8 i \ket{\text{dead}}.
\] I chose the probabilities non-uniform (36% and 64%, thanks to Pythagoras for these nice numbers) and I chose the amplitudes to be complex (with different phases) because I want to emphasize that the phases matter as much as the ratios of the absolute values! If you perform another measurement of an operator that doesn't commute with the binary operator having eigenvalues zero (dead) and one (alive), the relative phase will matter for the predictions as much as the ratio of the absolute values!
So the "dead cat" and the "alive cat" are two intermediate histories whose mutual interference may in principle affect the probabilities of later measurements just like the mutual interference of "slit A" and "slit B". Just like we can't assume that the particle in the double slit experiment picked one of the possible slits (and intermediate histories), we can't assume it about the cat.
The knowledge of the existence of interference of the state vector in quantum mechanics is the same thing as the knowledge that the world never fundamentally splits to two classical histories.
In other words, people who really think that the two worlds separately exist in the classical sense are misunderstanding the lesson 1 of of the undergraduate quantum mechanics lecture that begins with the double slit experiment. They're still not getting the point that the histories (basis vectors of the intermediate-time Hilbert space) can't be assumed to be "mutually exclusive" in the classical sense (what I mean is that you can't really assume that only one of the intermediate histories took place, and to calculate the weighted average of the probabilities from the two histories to get the probabilities of the final outcomes).
Now, classical physics and classical interpretations do follow from quantum mechanics as a good approximation for large objects including cats. The interference patterns you could create by reinterfering the "dead" and "alive" cat are so fine, chaotic, and unpredictable (especially because the microstate of the environment that the cat has interacted with can't be exactly followed and measured and one would need to know it very precisely) that in practice, one may use the classical interpretation for observables for which the classical interpretation is approximately OK and that tend to "classically xerox" the information about themselves by interacting with and by entangling many degrees of freedom of the environment. The interference patterns are too fine or average out in the real measurements. Equivalently, the aforementioned operators "not commuting" with the operator of the "livelihood of the cat" are extremely unnatural and hard to operationally measure. The classical reasoning is therefore OK.
But all these classical interpretations are just approximate. And clearly, if we are using them, we are not "interpreting" anything about quantum mechanics. We are only "interpreting" classical physics and all of our comments only become (approximately) valid if all the special features of quantum mechanics (re-interference...) become unmeasurable!
This conclusion is also equivalent to the reasons why another statement by Carroll, the statement that (using my words) "quantum mechanics directly opens the door for multiple universes while classical physics doesn't".
Again, this claim is untrue. Again, it's the polar opposite of the truth. The truth is exactly the other way around. Multiple universes that co-exist in the classical sense may be chosen as a way to visualize the classical probabilistic distribution but this way to visualize what's going on breaks down exactly when the quantum mechanical effects are still important.
(I didn't say "become important" because these words would indicate that classical physics is the normal and one needs to go very far, in an unnatural domain, for the quantum weirdness to emerge. The truth is exactly the opposite. Quantum laws are valid everywhere and they're completely normal and you have to go to special limits or corners if you want classical physics to become at least approximately usable!)
If you throw a classical die, you may imagine that there are six mutually different universes that were split off their parent universe, just like on the picture of the "film juncture". In one of them, you got "1", and so on. The subsequent evolution in each of these 6 universes is independent of the others. The "tree of films" is the tree of potential histories and the set of potential histories may be imagined to "really exist" somewhere in classical physics (at least discrete classical physics, to avoid subtleties with continuous trees and probability densities).
However, that's exactly what quantum mechanics doesn't allow because the right predictions of (probabilities of) future outcomes in a specific Universe – the only Universe or every Universe – are only obtained if you take the complex amplitudes in front of all the alternatives ("all the different universes") and their interference into account, according to the right formulae.
Again, the truth is the opposite than what these many worlds kibitzers are saying. In classical statistical physics, the probability distributions wouldn't be normally visualized as "really existing" in different copies of the world because the 19th century physicists knew that such an expansion of the idea about the "size of the world" would bring them nothing whatever scientifically. It wouldn't have increased their ability to explain or predict observations. But it was possible to imagine the branching tree of the classical histories and transitions between them.
Quantum mechanics made it impossible to visualize the alternatives in this way because the alternatives [possible intermediate-time features of the history of any experiment] are no longer mutually exclusive in the classical sense. So even the "potential" for many worlds in the classical sense is deleted by the basic features of quantum mechanics.
Sometimes, people or companies choose to cherry-pick some advantages of the products or ideas they are promoting. But what I find remarkable about the defense by individuals such as Sean Carroll is that their claims are downright lies from the beginning to the end. The truth isn't just "independent" from what they're saying. The truth about pretty much every isolated claim is demonstrably the opposite of what they say.
They constantly say that the gold is Å¡it and Å¡it is gold. They constantly say that their philosophy isn't adding anything to QM even though it's the whole point of this philosophy that it is trying to add some extra process – an undefined, undefinable, awkward, unphysical, and ultimately inevitably inconsistent process of splitting the "film". They constantly say that the proper quantum mechanics needs some extra laws to make the predictions even though it's demonstrably the case, according to the basic rules of quantum mechanics, that quantum mechanics extracts the predictions directly from the objects that depend on time through the known universal equations.
They repeatedly claim that their philosophy isn't introducing or depending on any qualitative boundary between the "microscopic quantum phenomena" and "macroscopic classical phenomena" while proper quantum mechanics is postulating such a boundary. The truth is exactly the opposite. Their splitting of the film is only viable if it is only applied to processes that are compatible with the classical interpretation, while the splitting is prohibited for intrinsically quantum observables (like the "which slit" information in the double slit experiment) which means that they need to introduce a qualitatively different treatment for the "two domains", otherwise they instantly get a wrong prediction for the double slit experiment and every other intrinsically quantum experiment. Proper quantum mechanics doesn't introduce any qualitative difference between the "domains". All phenomena are described by the same theory developed in Copenhagen etc. It's only classical physics that is "only valid" (approximately) in a realm behind the boundary. It is valid along with quantum mechanics over there (QM holds everywhere, classical physics only holds approximately in the limit) and the validity of classical physics is also needed for observers to make classical-like statements about their observations but it's just a limitation of classical-like statements about observations, not a limitation of the laws of quantum mechanics which always hold. An observer large enough to take decoherence in his brain for granted may choose the "Heisenberg cut" (the boundary, if interpreted using the words of the founders of quantum mechanics) anywhere in between his brain and the small objects whose interference may be measured, so the location of any would-be boundary is physically inconsequential. The theory really doesn't depend on any features like that and the only "invariant enough" refinement of this boundary, the place where decoherence starts to legitimize the classical interpretation of the probabilities (by erasing the potential for re-interference), is again fully derivable from the same universal laws of quantum mechanics.
("The Heisenberg cut" and the basic foundations of quantum mechanics also clearly, crisply, and cleverly answer the question "how many junctions in the MWI film there should be?" or "how finely the histories should be divided?". The answer is that it is only justified – but still not necessary – to imagine that the history of the world splits into several alternatives if you can and do actually measure a quantity. No measurement means no splitting. Whether one made a measurement or perceived an outcome is really a subjective question – think about Wigner's friend – which is why all questions about "how fine the MWI tree objectively is" are no good because they are based on the fundamentally wrong assumption that such questions may have objective answers. Observation is a subjective experience or act by an observer.)
They constantly say that their philosophy is a "friend" of the fundamental equations of quantum mechanics even though they have demonstrably no positive relation at all. All the equations of quantum mechanics are constraining the values of intrinsically probability-like basic entities (the state vector etc.) while the MWI advocates deny the existence of any concepts that must have a fundamentally probabilistic interpretation. So the claim that the philosophy is a "friend" with the equations of quantum mechanics is a much greater lie than the claim by a creationist who bought a T-shirt with the Standard Model Lagrangian and who says that "particle physics is on their side". In the creationist's case, the two "friends" are at least independent of each other. In Carroll's case, the "friends" are two directly contradicting ideas.
They constantly say that quantum mechanics made it more natural than classical physics to interpret probabilities through many worlds. The truth is that quantum mechanics eliminated the freedom to interpret probabilities in this way that existed in classical physics (classical physics allowed us to imagine that the "set of potential histories" is the set of "real histories" in an appropriately extended world). And so on, and so on. This whole MWI movement is completely irrational, dishonest, and just plain idiotic, and it has contaminated the Academia so thoroughly that there exists no Niels Bohr in 2014 who could just eliminate all these hacks.
I am disgusted by this Å¡it. If you combine it with other equally misguided stuff, it is probably fair to say that this trash has taken over the majority of most of the official institutions that should study conceptual issues in physics or science. The situation outside the Academia is no better. Just look at the dozens of imbeciles at Carroll's blog – only a tiny fraction manages to mention that Carroll's writing is just wrong.
BTW even if the many worlds "interpretation" didn't suffer from the basic lethal conceptual diseases above, and it demonstrably does, there would still be pyramids of technical problems that the picture couldn't overcome. For example, Moshe Rozali wrote at Sean's blog:
In the following 55+ years, there has been no development that would strengthen the case for Mr Everett's musings. On the contrary, the experimental evidence and the modern crystallization of the quantum mechanical insights have made it increasingly clear that these musings were fundamentally wrong and proper quantum mechanics is the only viable theory to describe the relevant phenomena – which are really all processes in Nature.
However, what has changed is that the Academia has been literally flooded by mediocre, incompetent pseudo-intellectuals similar to Hugh Everett – usually much worse and more inadequate than Everett himself. Sean Carroll is a textbook example of this sign of the degradation of the contemporary Academia which is why he wrote another text defending the indefensible,
Why the Many-Worlds Formulation of Quantum Mechanics Is Probably CorrectHoly crap.
The illustration of choice used to summarize and promote this "interpretation" of quantum mechanics is this Wikipedia picture of Schrödinger's cat (a four-legged cat; Schrödinger would simultaneously maintain several two-legged ones, too).
What is this picture supposed to convey?
It suggests that there are two worlds (or two similar copies of our world), and not just one, after a cat is observed. In one of them, the cat is alive. In the other world, it is dead. This childish picture is supposed to solve some "problems" that the defenders of this picture believe to exist in quantum mechanics. There are no problems in quantum mechanics, of course.
At the same time, salesmen like Carroll offer you lots of incredible statements such as the statement that this "many worlds interpretation" directly follows from quantum mechanics, is directly justified by quantum mechanics (more justifiable in quantum mechanics than in classical physics), and unlike proper quantum mechanics, it doesn't introduce any new physical laws. All these statements are untrue. They are really the polar opposite of the truth.
First, many worlds surely don't follow from quantum mechanics.
Quantum mechanics is the universal framework of modern physics which is a natural science. As every natural science, physics predicts or explains the observations that are actually being made in one Universe. Deterministic physics (which may only work approximately in Nature) gives unambiguous predictions. Quantum physics (which is valid everywhere in the real world) predicts phenomena by quantifying the probabilities of different outcomes. Outcomes with probabilities close to 100% will probably happen, those close to 0% will probably not. The precise value of the probability may be extracted from the experiment if we repeat it many times (but using the knowledge of probabilities, one can make precise predictions even if we perform many inequivalent experiments!).
The calculation of the probabilities is the quantum mechanical output replacing the unambiguous classical predictions. If you can't say anything about the values of probabilities of some phenomena, then you have no scientific theory whatsoever!
This is the case of the many worlds "interpretation", too.
The many worlds "interpretation" is an artistic picture that simply has nothing to do with science. Note that the picture with the split film of the cat doesn't tell you that the cat will survive with the 73% probability or anything like that. It just tells you that there are two possibilities. It shows you the set of potential histories, not any nontrivial information about the actual history.
And even this information-free statement is really wrong or demagogic; and it can't be derived from the many worlds "interpretation", anyway.
It is wrong because there aren't just two possible outcomes. There are infinitely many outcomes. The cat also has a nonzero probability to morph into a small version of Elvis Presley and be hit by some air molecules that bring it or him to the Moon. The probability is small but it is nonzero. (I could invent more likely outcomes "in between", too.)
So the film should really be split into infinitely many strands at each spacetime point, including strands describing extremely unlikely histories. (In fact, we can't even say that the splitting would occur at each spacetime point because some observations are nonlocal in character. There exist no well-defined rules that would allow us to describe the "cardinality" of the tree at all.) The very fact that only two "reasonably likely" histories were cherry-picked is nothing else than demagogy. Such a qualitatively different treatment of "reasonably likely" and "very unlikely" outcomes is only possible if we first calculate the probabilities and then use different treatments according to their value (and this decision is always about being useful, not being true; fundamentally, Nature doesn't treat high and low probabilities qualitatively differently).
The many worlds "interpretation" doesn't allow us to calculate – or imprint – the generally different probabilities of different outcomes into the "strands" of the film. Even the fans of this religion admit it's the case but some of them say that they are "working on a fix" which is supposed to be enough. (A similarly "modest" fix makes Genesis compatible with all the detailed data about the cosmic microwave background, the DNA, and genetics.)
What they don't realize is that the inability of this theory or "interpretation" to calculate any probabilities isn't just a moderate vice or disadvantage. It is a complete, rigorous proof that this philosophy has nothing to do with the empirical data or science whatever and people who are defending it don't have the slightest clue what they are talking about. Everything that modern science predicts are probabilities or their functions. If your ideas don't predict any probabilities, they don't predict anything at all. They have nothing to do with science.
Moreover, there can't be any fix. There can't even exist a candidate theory that would "extract" the probabilities from something else. In proper quantum mechanics, probabilities are fundamental which means that they are pretty much directly predicted from the mathematical formalism. There is no useless "intermediate agent" that would extract probabilities from something else. There can't really be any. A theory is either exactly deterministic (plus admitting complete knowledge in principle) or it is not. If it is exactly deterministic, it is possible to produce unequivocal predictions out of the information about the initial state, at least in principle. That's the case of classical physics. In this setup, all predicted probabilities are either 0% or 100%. That's no good for the experiments backing quantum mechanics where almost all the probabilities are strictly between 0% and 100%.
Or the theory may fail to be strictly deterministic, or prohibit the observers from knowing the exact required information, even in principle. In that case, the probabilities may be in between 0% and 100%. But such continuous, a priori arbitrary probabilities may only be calculated from other probability-like continuous quantities that are present in the fundamental formalism of the theory such as the classical probability distributions on the phase space or the density matrix (which may be obtained from a state vector if the state is pure).
The very illustration with the split film of the cat as well as many explicit statements by the many worlds apologists make it absolutely clear that one of the key assumptions of the whole many worlds movement is that probabilities are not fundamental. But it's just mathematically impossible to derive a priori continuous numbers interpretable as probabilities from a formalism that rejects any fundamental quantities with a directly probability-like interpretation! Probabilities may only be calculated from other continuous numbers of a similar kind and all such continuous numbers are probability-like because probabilities are their functions! If there aren't any, your formalism just can't be capable of predicting probabilities.
Note that this simple argument – pure maths – really eliminates all attempts to replace quantum mechanics with some ideas where probabilities are not fundamental. This includes the Bohmian and other hidden-variable theories. The pilot wave theory needs to assume that the classical particle it envisions has a position in the initial state that is statistically distributed according to the wave function (renamed as guiding wave). Only the correct time evolution of the probabilistic distribution is guaranteed by the mechanical equation of motion including the effect of the guiding wave.
They don't have any mechanism explaining the right probabilistic distribution of their classical particle in the initial state. Of course, there can't be any. They prefer not to talk about this point at all even though this point really means that their claim that they have derived the "random generator" of quantum mechanics from something else is completely fraudulent. The same "random generator" must still exist somewhere to produce the classical particle in a position that is correctly distributed. It must be distributed according to the wave function which proves that if this condition is guaranteed, the wave function has a fundamentally probabilistic interpretation, anyway!
But let me return to the many worlds. Its defenders often say something like
the many worlds "interpretation" is just pure mechanics; it is not adding anything to it.Carroll makes this statement about 5 times and it is a complete lie. It is really the opposite of the truth.
Proper quantum mechanics dictates the evolution of operators in the Heisenberg picture (or, equivalently, the direct Feynman path integral formula to calculate the probability amplitudes; or the evolution of the state vector in Schr̦dinger's picture Рthe latter picture is the most popular approach among the superficial and confused people). The complex probability amplitudes may be directly squared (in absolute values, using the right Born's rule etc.) to calculate the probabilities of various observations or outcomes or perceptions etc. Quantum mechanics works totally universally. It doesn't matter what object we want to observe and which aspect of it we measure. Quantum mechanics predicts the probabilities using the same universal formulae and the time evolution is always given by the same equations, too.
What about the many worlds "interpretation"? Just look at the marketing illustration at the top. Aside from Heisenberg's or Schr̦dinger's dynamical equations, there is a new process added on top of it Рthe splitting of the film, the splitting of the worlds to make them "many". When writing this promotion of "many worlds", Carroll must have forgotten about something, about the "many worlds", right?
Believe me that to produce a film that splits into two in the triangular junction is much harder than to produce an ordinary film with a single story of a cat. It's also harder to insert such a film with "junctions" into a film projector. The film projector sometimes jams, and so on. It doesn't work well. Try it if you don't believe me.
Now, to produce another copy of the Universe is even harder than to produce a plastic foil with junctions. It's even a bit harder to produce many copies or \(\exp(10^{120})\) copies you might need to describe the multiplicity of outcomes of every arrangement of processes that may occur within the patch of the visible Universe. But OK, you might rightfully say that it's not your job to produce copies of the Universe. Nature can do it easily.
Yes, Nature can do anything if the laws of physics say that it should happen. The problem is that one can't even define the rules when the hypothetical splitting should take place and how the two or three or \(N\) or \(\infty\) (also unknown) copies attached to the junction should differ from each other. And the problem isn't just that we don't know how often and finely the film should be "split" etc. We may easily see that whatever the rules are, they contradict some known facts about Nature. Why?
The really key feature of the quantum evolution is that it always allows parts of the state vector to interfere in the future and affect various other measurements in non-classical ways.
We know this even from the double slit experiment which, as Feynman observed, contains all the qualitative wisdom about quantum mechanics if you think about the experiment carefully enough. We can't assume that the particle goes either through slit A or slit B (with the strict, classical assumptions about the alternatives) if we don't observe it. The actual initial state implies nonzero probability amplitudes for both intermediate histories and their mutual interference affects the existence and location of the interference minima and maxima on the photographic screen!
We can never strictly say that some observable (like the "which slit" information about position) took a completely specific value in the classical sense. The eigenvector may always interfere with another one. Now, if you think about it for a second, this totally elementary and universal feature of quantum mechanics in general and of the double slit experiment in particular is nothing else than the statement that the worlds just never split in the classical sense!
The cat may be in the state \[
\ket\psi = 0.6 \ket {\text{alive}} + 0.8 i \ket{\text{dead}}.
\] I chose the probabilities non-uniform (36% and 64%, thanks to Pythagoras for these nice numbers) and I chose the amplitudes to be complex (with different phases) because I want to emphasize that the phases matter as much as the ratios of the absolute values! If you perform another measurement of an operator that doesn't commute with the binary operator having eigenvalues zero (dead) and one (alive), the relative phase will matter for the predictions as much as the ratio of the absolute values!
So the "dead cat" and the "alive cat" are two intermediate histories whose mutual interference may in principle affect the probabilities of later measurements just like the mutual interference of "slit A" and "slit B". Just like we can't assume that the particle in the double slit experiment picked one of the possible slits (and intermediate histories), we can't assume it about the cat.
The knowledge of the existence of interference of the state vector in quantum mechanics is the same thing as the knowledge that the world never fundamentally splits to two classical histories.
In other words, people who really think that the two worlds separately exist in the classical sense are misunderstanding the lesson 1 of of the undergraduate quantum mechanics lecture that begins with the double slit experiment. They're still not getting the point that the histories (basis vectors of the intermediate-time Hilbert space) can't be assumed to be "mutually exclusive" in the classical sense (what I mean is that you can't really assume that only one of the intermediate histories took place, and to calculate the weighted average of the probabilities from the two histories to get the probabilities of the final outcomes).
Now, classical physics and classical interpretations do follow from quantum mechanics as a good approximation for large objects including cats. The interference patterns you could create by reinterfering the "dead" and "alive" cat are so fine, chaotic, and unpredictable (especially because the microstate of the environment that the cat has interacted with can't be exactly followed and measured and one would need to know it very precisely) that in practice, one may use the classical interpretation for observables for which the classical interpretation is approximately OK and that tend to "classically xerox" the information about themselves by interacting with and by entangling many degrees of freedom of the environment. The interference patterns are too fine or average out in the real measurements. Equivalently, the aforementioned operators "not commuting" with the operator of the "livelihood of the cat" are extremely unnatural and hard to operationally measure. The classical reasoning is therefore OK.
But all these classical interpretations are just approximate. And clearly, if we are using them, we are not "interpreting" anything about quantum mechanics. We are only "interpreting" classical physics and all of our comments only become (approximately) valid if all the special features of quantum mechanics (re-interference...) become unmeasurable!
This conclusion is also equivalent to the reasons why another statement by Carroll, the statement that (using my words) "quantum mechanics directly opens the door for multiple universes while classical physics doesn't".
Again, this claim is untrue. Again, it's the polar opposite of the truth. The truth is exactly the other way around. Multiple universes that co-exist in the classical sense may be chosen as a way to visualize the classical probabilistic distribution but this way to visualize what's going on breaks down exactly when the quantum mechanical effects are still important.
(I didn't say "become important" because these words would indicate that classical physics is the normal and one needs to go very far, in an unnatural domain, for the quantum weirdness to emerge. The truth is exactly the opposite. Quantum laws are valid everywhere and they're completely normal and you have to go to special limits or corners if you want classical physics to become at least approximately usable!)
If you throw a classical die, you may imagine that there are six mutually different universes that were split off their parent universe, just like on the picture of the "film juncture". In one of them, you got "1", and so on. The subsequent evolution in each of these 6 universes is independent of the others. The "tree of films" is the tree of potential histories and the set of potential histories may be imagined to "really exist" somewhere in classical physics (at least discrete classical physics, to avoid subtleties with continuous trees and probability densities).
However, that's exactly what quantum mechanics doesn't allow because the right predictions of (probabilities of) future outcomes in a specific Universe – the only Universe or every Universe – are only obtained if you take the complex amplitudes in front of all the alternatives ("all the different universes") and their interference into account, according to the right formulae.
Again, the truth is the opposite than what these many worlds kibitzers are saying. In classical statistical physics, the probability distributions wouldn't be normally visualized as "really existing" in different copies of the world because the 19th century physicists knew that such an expansion of the idea about the "size of the world" would bring them nothing whatever scientifically. It wouldn't have increased their ability to explain or predict observations. But it was possible to imagine the branching tree of the classical histories and transitions between them.
Quantum mechanics made it impossible to visualize the alternatives in this way because the alternatives [possible intermediate-time features of the history of any experiment] are no longer mutually exclusive in the classical sense. So even the "potential" for many worlds in the classical sense is deleted by the basic features of quantum mechanics.
Sometimes, people or companies choose to cherry-pick some advantages of the products or ideas they are promoting. But what I find remarkable about the defense by individuals such as Sean Carroll is that their claims are downright lies from the beginning to the end. The truth isn't just "independent" from what they're saying. The truth about pretty much every isolated claim is demonstrably the opposite of what they say.
They constantly say that the gold is Å¡it and Å¡it is gold. They constantly say that their philosophy isn't adding anything to QM even though it's the whole point of this philosophy that it is trying to add some extra process – an undefined, undefinable, awkward, unphysical, and ultimately inevitably inconsistent process of splitting the "film". They constantly say that the proper quantum mechanics needs some extra laws to make the predictions even though it's demonstrably the case, according to the basic rules of quantum mechanics, that quantum mechanics extracts the predictions directly from the objects that depend on time through the known universal equations.
They repeatedly claim that their philosophy isn't introducing or depending on any qualitative boundary between the "microscopic quantum phenomena" and "macroscopic classical phenomena" while proper quantum mechanics is postulating such a boundary. The truth is exactly the opposite. Their splitting of the film is only viable if it is only applied to processes that are compatible with the classical interpretation, while the splitting is prohibited for intrinsically quantum observables (like the "which slit" information in the double slit experiment) which means that they need to introduce a qualitatively different treatment for the "two domains", otherwise they instantly get a wrong prediction for the double slit experiment and every other intrinsically quantum experiment. Proper quantum mechanics doesn't introduce any qualitative difference between the "domains". All phenomena are described by the same theory developed in Copenhagen etc. It's only classical physics that is "only valid" (approximately) in a realm behind the boundary. It is valid along with quantum mechanics over there (QM holds everywhere, classical physics only holds approximately in the limit) and the validity of classical physics is also needed for observers to make classical-like statements about their observations but it's just a limitation of classical-like statements about observations, not a limitation of the laws of quantum mechanics which always hold. An observer large enough to take decoherence in his brain for granted may choose the "Heisenberg cut" (the boundary, if interpreted using the words of the founders of quantum mechanics) anywhere in between his brain and the small objects whose interference may be measured, so the location of any would-be boundary is physically inconsequential. The theory really doesn't depend on any features like that and the only "invariant enough" refinement of this boundary, the place where decoherence starts to legitimize the classical interpretation of the probabilities (by erasing the potential for re-interference), is again fully derivable from the same universal laws of quantum mechanics.
("The Heisenberg cut" and the basic foundations of quantum mechanics also clearly, crisply, and cleverly answer the question "how many junctions in the MWI film there should be?" or "how finely the histories should be divided?". The answer is that it is only justified – but still not necessary – to imagine that the history of the world splits into several alternatives if you can and do actually measure a quantity. No measurement means no splitting. Whether one made a measurement or perceived an outcome is really a subjective question – think about Wigner's friend – which is why all questions about "how fine the MWI tree objectively is" are no good because they are based on the fundamentally wrong assumption that such questions may have objective answers. Observation is a subjective experience or act by an observer.)
They constantly say that their philosophy is a "friend" of the fundamental equations of quantum mechanics even though they have demonstrably no positive relation at all. All the equations of quantum mechanics are constraining the values of intrinsically probability-like basic entities (the state vector etc.) while the MWI advocates deny the existence of any concepts that must have a fundamentally probabilistic interpretation. So the claim that the philosophy is a "friend" with the equations of quantum mechanics is a much greater lie than the claim by a creationist who bought a T-shirt with the Standard Model Lagrangian and who says that "particle physics is on their side". In the creationist's case, the two "friends" are at least independent of each other. In Carroll's case, the "friends" are two directly contradicting ideas.
They constantly say that quantum mechanics made it more natural than classical physics to interpret probabilities through many worlds. The truth is that quantum mechanics eliminated the freedom to interpret probabilities in this way that existed in classical physics (classical physics allowed us to imagine that the "set of potential histories" is the set of "real histories" in an appropriately extended world). And so on, and so on. This whole MWI movement is completely irrational, dishonest, and just plain idiotic, and it has contaminated the Academia so thoroughly that there exists no Niels Bohr in 2014 who could just eliminate all these hacks.
I am disgusted by this Å¡it. If you combine it with other equally misguided stuff, it is probably fair to say that this trash has taken over the majority of most of the official institutions that should study conceptual issues in physics or science. The situation outside the Academia is no better. Just look at the dozens of imbeciles at Carroll's blog – only a tiny fraction manages to mention that Carroll's writing is just wrong.
BTW even if the many worlds "interpretation" didn't suffer from the basic lethal conceptual diseases above, and it demonstrably does, there would still be pyramids of technical problems that the picture couldn't overcome. For example, Moshe Rozali wrote at Sean's blog:
Sean, I am happy to believe many-worlds if all that is at stake is the definition of the word “exist” , which is likely ambiguous in this context anyways. But I am uncomfortable with what seems to me a description too closely tied to single-particle QM. For example, most descriptions I see talk about a split between finite number of options following a measurement arbitrarily localized in space and time. Is there a way to tell the story of universes “splitting” for measurements of operators with a continuous spectrum, like most interesting operators in quantum field theory? Do we get continuous infinity of universes, and if so is there some regularized version and some sense of cut-off independence?
Many worlds pseudoscience, again
![Many worlds pseudoscience, again]()
Reviewed by MCH
on
July 01, 2014
Rating:
No comments: