There’s a new popular book out this week about the interpretation of quantum mechanics, Adam Becker’s *What is Real?: The Unfinished Quest for the Meaning of Quantum Physics*. Ever since my high school days, the topic of quantum mechanics and what it really means has been a source of deep fascination to me, and usually I’m a sucker for any book such as this one. It’s well-written and contains some stories I had never encountered before in the wealth of other things I’ve read over the years.

Unfortunately though, the author has decided to take a point of view on this topic that I think is quite problematic. To get an idea of the problem, here’s some of the promotional text for the book (yes, I know that this kind of text sometimes is exaggerated for effect):

A mishmash of solipsism and poor reasoning, [the] Copenhagen [interpretation] claims that questions about the fundamental nature of reality are meaningless. Albert Einstein and others were skeptical of Copenhagen when it was first developed. But buoyed by political expediency, personal attacks, and the research priorities of the military industrial complex, the Copenhagen interpretation has enjoyed undue acceptance for nearly a century.

The text then goes to describe Bohm, Everett and Bell as the “quantum rebels” trying to fight the good cause against Copenhagen.

Part of the problem with this good vs. evil story is that, as the book itself explains, it’s not at all clear what the “Copenhagen interpretation” actually is, other than a generic name for the point of view the generation of theorists such as Bohr, Heisenberg, Pauli, Wigner and von Neumann developed as they struggled to reconcile quantum and classical mechanics. They weren’t solipsists with poor reasoning skills, but trying to come to terms with the extremely non-trivial and difficult problem of how the classical physics formalism we use to describe observations emerges out of the more fundamental quantum mechanical formalism. They found a workable set of rules to describe what the theory implied for results of measurements (collapse of the state vector with probabilities given by the Born rule), and these rules are in every textbook. That there is a “measurement problem” is something that most everyone was aware of, with Schrodinger’s cat example making it clear. Typically, for the good reason that it’s complicated and they have other topics they need to cover, textbooks don’t go into this in any depth (other than often telling about the cat).

As usual these days, the alternative to Copenhagen being proposed is a simplistic version of Everett’s “Many Worlds”: the answer to the measurement problem is that the multiverse did it. The idea that one would also like the measurement apparatus to be described by quantum mechanics is taken to be a radical and daring insight. The Copenhagen papering over of the measurement problem by “collapse occurs, but we don’t know how” is replaced by “the wavefunction of the universe splits, but we don’t know how”. Becker pretty much ignores the problems with this “explanation”, other than mentioning that one needs to explain the resulting probability measure.

String theory, inflation and the cosmological multiverse are then brought in as supporting Many Worlds (e.g. that probability measure problem is just like the measure problem of multiverse cosmology). There’s the usual straw man argument that those unhappy with the multiverse explanation are just ignorant Popperazzi, unaware of the subtleties of the falsifiability criterion:

Ultimately, arguments against a multiverse purportedly based on falsifiability are really arguments based on ignorance and taste: some physicists are unaware of the history and philosophy of their own field and find multiverse theories unpalatable. But that does not mean that multiverse theories are unscientific.

For a much better version of the same story and much more serious popular treatment of the measurement problem, I recommend a relatively short book that is now over 20 years old, David Lindley’s *Where does the Weirdness Go?*. Lindley’s explanation of Copenhagen vs. Many Worlds is short and to the point:

The problem with Copenhagen is that it leaves measurement unexplained; how does a measurement select one outcome from many? Everett’s proposal keeps all outcomes alive, but this simply substitutes one problem for another: how does a measurement split apart parallel outcomes that were previously in intimate contact? In neither case is the physical mechanism of measurement accounted for; both employ sleight of hand at the crucial moment.

Lindley ends with a discussion of the importance of the notion of decoherence (pioneered by Dieter Zeh) for understanding how classical behavior emerges from quantum mechanics. For a more recent serious take on the issues involved, I’d recommend reading something by Wojciech Zurek, for instance this article, a version of which was published in Physics Today. Trying to figure out what “interpretation” Zurek subscribes to, I notice that he refers to an “existential interpretation” in some of his papers. I don’t really know what that means. Unlike most discussions of “interpretations”, Zurek seems to be getting at the real physical issues involved, so I think I’ll adopt his (whatever it means) as my chosen “interpretation”.

**Update**: For another take on much the same subject, out in the UK now is Philip Ball’s *Beyond Weird*. The US version will be out in the fall, and I think I’ll wait until then to take a look. In the meantime, Natalie Wolchover has a review at Nature.

**Update**: There’s a new review of What is Real? at Nature.

**Update**: Jim Holt points out that David Albert has a review of the Becker book in the latest New York Review of Books. I just read a print copy last night, presumably it should appear online soon here [Review now available here].

**Update**: Some comments from Adam Becker, the author of the book.

I won’t try to rebut everything Peter has said about my book—there are some things we simply disagree about—but I would like to clear up two statements he makes about the book that are possibly misleading:

Peter says that I claim the answer to the measurement problem is that “the multiverse did it.” But I don’t advocate for the many-worlds interpretation in my book. I merely lay it out as one of the reasonable available options for interpreting quantum mechanics (and I discuss some of its flaws as well). I do spend a fair bit of time talking about it, but that’s largely because my book takes a historical approach to the subject, and many-worlds has played a particularly important role in the history of quantum foundations. But there are other interpretations that have played similarly important roles, such as pilot-wave theory, and I spend a lot of time talking about those interpretations too. I am not a partisan of any particular interpretation of quantum mechanics.

Second, I don’t think it’s quite fair to say that I paint Bohr as a villain. I mention several times in my book that Bohr was rather unclear in his writing, and that sussing out his true views is dicey. But what matters more that Bohr’s actual views is what later generations of physicists generally took his views to be, and the way Bohr’s work was uncritically invoked as a response to reasonable questions about the foundations of quantum mechanics. It’s true that this subtlety is lost in the jacket flap copy, but that’s publishing for you.

Also, for what it’s worth, I do like talking about reality as it relates to quantum mechanics. But I suppose that’s hardly surprising, given that I just wrote a book on quantum foundations titled “What Is Real?”. I’d be happy to discuss all of this further over email if anyone is interested (though I’m pretty busy at the moment and it might take me some time to respond).

Peter,

Try this. Imagine a modern-day equivalent of Ptolemy’s model for the orbital motions of the planets based on very large numbers of epicycles. We develop a computer programme to calculate planetary orbits with a very high accuracy. We all agree that the planets are ‘real’ (they continue to exist when nobody’s looking at them through a telescope) and that the physical situation we are describing (planets moving with certain speeds in certain directions across the night sky) is ‘real’. But I don’t think anyone would consider the epicycles themselves to be ‘real’, in the sense that they represent real physical forces or processes that somehow govern or determine these motions.

We would say that the epicycles are a convenient way of *representing* the planetary motions, but they shouldn’t be taken literally as aspects of the real physics. Likewise, some theorists have questioned the status of the wavefunction in QM. They don’t necessarily deny that there is ‘real physics’ going on, but they argue that the wavefunctions are like epicycles – they are a convenient way of summarising what we know and calculating probabilities, but they shouldn’t be taken literally as representing the real physical states of quantum systems.

Let’s wait for the Quantum Computers to fail in interesting fashion.

That is a testable, well, not prediction. But outlook.

I was actually astonished to see experiments on evidence of superposition on large molecules come back positive:

“We observe high-contrast quantum fringe patterns with molecules exceeding a mass of 10 000 amu and 810 atoms in a single particle.”. Does the universe know what a human cognitive apparatus labels a “single body”? Why doesn’t it superpose the single body into a mismash different “single bodies”?Jim,

I understand there’s a complex issue of the relationship of reality and our models of it, but it’s exactly this deep philosophical issue that I don’t want to enter into discussions of, because I don’t see that it sheds any light on the non-trivial measurement problem (the relation of quantum and classical). It may very well shed light on some of the debates between interpretations that have gone on over the years, but ones that I’ve long ago lost interest in.

Hope this isn’t OT and/or too basic to merit inclusion in the discussion. Apologies if so.

What’s the consensus on probability in MWI? As others have mentioned above, it doesn’t appear that everyone agrees it’s a matter that’s been satisfactorily resolved. To the extent that I understand the pro arguments, I’m not especially satisfied either. The whole notion of somehow equating the “illusion” of probability with frequency seems like slight-of-hand. World mangling doesn’t seem to help much. It’s not clear to me how these supposed solutions do anything but leave us with exactly the same questions about quantum dice we might have with other interpretations. The notion that all these branches are somehow extant also seems to necessitate something like an anthropic principle to explain the fact that I have never experienced bizarre phenomena that are highly unlikely but not excluded. I don’t need a multiverse to explain why I don’t get “heads” ten thousand times in a row with a fair coin, do I? To me that’s not just an aesthetic problem, it’s a problem with legitimately excess baggage.

LMMI and others,

I confess that I have a more basic confusion here. I don’t see why the explanation for probability entering the picture isn’t just the usual one: we only have probabilistic information about the state of the measurement apparatus. This seems too obvious to be any sort of explanation, I’d love to hear why from someone who has thought about this more than I have (which is not much…).

Low Math, Meekly Interacting,

It’s hard to say what the general consensus is, I haven’t run any polls. But I can tell you my opinion.

One thing that Many-Worlds explain better about the quantum dice is why the outcome is in principle unpredictable: there is nothing to predict, as all outcomes happen. This ties in well with the Deutsch-Wallace theorem: since all outcomes happen, you should be indifferent about them, which leads to uniform probabilities to states with equal coefficients. Throw in some unitary invariance, and you have the Born rule. So I think the Many-Worlds explanation for subjective probabilities is pretty satisfactory.

What is not satisfactory, and is probably the source of your unease, is the explanation for objective probabilities, as it is simply lacking. This is not a point against Many-Worlds, as there is no explanation for objective probabilities in single-world quantum mechanics either. Why don’t you get “heads” ten thousand times in a row with a fair coin? Because the probability is small? Because the mod-squared amplitude is small?

I think, however, that it can be explained, most easily in a toy many-worlds theory where worlds can be counted. Consider that after each throw of the coin exactly one world is created with “heads”, and one world is created with “tails”. Then after 10,000 throws there will be 2^10,000 worlds, and in only one of them you see 10,000 heads in a row. In general, the relative frequencies measured in this toy theory will be close to the objective probability of 1/2 in most worlds.

Mateus,

“there is nothing to predict, as all outcomes happen”

But an observer can try to predict whether she will see this or that outcome, right? Because to any observer, she only sees one outcome. Only to a God’s eye point of view do all the outcomes occur. Observers don’t have this viewpoint.

Moreover, what forbids the classical world to adopt a kind of MWI when dealing with uncertainties? If so, we do not need to explain why anything happen anymore, just accept that everything happened and we happened to be in a world where THESE things happened.

Since in the MWI we dragged the observer into the interaction, we can go further and ask why the observer made this measurement choice and not others. We don’t have to stopped at the Outcome MWI. We can have a much bigger world – the Choice MWI, by saying that in fact all the incompatible measurement choices are made, but in different worlds. The problem of probability measure will become more intractable, and the entire situation becomes more absurd.

Peter,

The term “real” is not deep and obscure, it is empty and reduplicative. Similarly with “actual” and “true” and “exists”. Thus: “I have a dog”, “I have a real dog”, “I have a real, actual dog”, “I have a real actual dog that exists” and “It is true that I have a real, actual dog that actually exists in reality. Truly I do.” all say exactly the same thing. And the question of physical existence or physical ontology is a matter of physics, not philosophy.

Here is a simple question of physical ontology. When we first learn electromagnetism, we are taught to use two mathematical vector fields the E field and the B field, to represent the physical electro-magnetic situation. And we are also taught to use the scalar and vector potentials to represent the physical situation. But we are warned that the scalar and vector potentials are not “real”, in the sense that mathematical representations that differ in the values of A and phi can represent the very same physical situation if a gauge transformation carries {A, phi} into {A’, phi’}. In short, there are mathematical degrees of freedom in the mathematical representation that do not correspond to physical degrees of freedom in the physical situation. And when the Aharonov-Bohm effect was discovered, this classical understanding was no longer tenable. How exactly to understand what the physical degrees of freedom might be is a question of physics, and not philosophy. One can ask that question using questions like “Are the scalar and vector potentials physically real?” meaning “Are the physical degrees of freedom isomorphic to the mathematical degrees of freedom in the mathematical representation that uses A and phi?”.

Now you may not be interested in this question. You can take the instrumentalist attitude that no matter how that question is answered, if you know how to use the formalism to make predictions that is all you care about. No one can stop you from adopting such an instrumentalist attitude. But those of us who care deeply about this question do not do so as “philosophers”, whatever that might mean. We do so as physicists, interested in the nature and structure of the physical world. The question “Are the scalar and vector potentials physically real or merely mathematical conveniences?” is slightly poorly phrased, but it is posing these very physical questions.

But Tim, how exactly are you deciding whether it is E&B or A&phi that are, or represent, the “true physical degrees of freedom”? A priori, any of these is just a mathematical object, namely a scalar or vector field in 3-dimensional space. And you have to derive empirical consequences for a priori real objects, such as ultimately dogs, in order to decide which of these mathematical objects is useful or necessary for the description of reality, and before assigning them reality themselves. Without the Lorentz force and Newton’s laws, the reality of the electro-magnetic field is an empty statement. Similarly, without the Schrödinger equation and the laws of quantum mechanics, the reality of the (integrated) vector potential just has no content whatever.

I agree that the question of physical ontology is not a philosophical one. But what physics teaches is that physical reality is hierarchical, interdependent, and not all the same. The quantum mechanical wave function is not real in the same sense as a dog, for example, though the distinction is not peculiar to quantum mechanics as your E&M example illustrates. What Bohr reminded everyone is that the beginning and end of the discussion of physical reality are experimental setups and experimental outcomes described in plain language. By that token, many worlds or Bohmian position variables have physical reality only to the extent that they have empirical consequences (but do they?).

It also needs to be said that in the discussion of physical ontology, “exists” frequently appears in the sense of the mathematical quantifier, which has to be carefully distinguished from the notion of physical reality (which is also different from the mathematical notion of reality). I hope you agree that at least this distinction applies.

Tim Maudlin/Joking,

Sorry, but I really will start ruthlessly suppressing discussion of ontology now. Debates about “is the vector potential real?” have nothing to do with the QM measurement problem (and, if you ask me, trivialize and obscure actual important issues, but that’s a topic for another day…)

Philip Ball’s book has a great discussion of the tricky issues that plague the many worlds hypothesis (he is not a fan)

Thanks Davide,

I may (or may not…) wait until it comes out in the US to take a look at the Ball book. In the meantime, for some idea of his criticisms of MWI, there are other things he has written, for example

https://aeon.co/essays/is-the-many-worlds-hypothesis-just-a-fantasy

The Columbia physicist/philosopher David Z Albert will have a long and pungent review of Becker’s “What Is Real?” in the next issue of the New York Review of Books. The review will be online at nybooks.com tomorrow at noon. (Spoiler alert: Oppenheimer does not come off well.)

Hi Peter,

is the book by N.P. Landsman – Foundations of Quantum Theory – From Classical Concepts to Operator Algebras (2017) on your list of read books about Quantum Physics?

http://www.springer.com/gp/book/9783319517766

I’d be happy to hear your thoughts about it.

Stephane,

Thanks for pointing that out, I hadn’t seen it before and haven’t had a chance to look at it closely. Landsman has thought very carefully and done a lot of work on the question of how classical physics emerges from quantum physics. One thing I can recommend is this article

https://arxiv.org/abs/quant-ph/0506082

From what I remember, Landsman clearly lays out the issues, I think comes down on the side of “there is still something going on here that we don’t understand” (I may be misrepresenting him).

Albert’s NYRB article is more his own re-telling of the history than a book review, imho. (If only Bohm, Everett, and Bohr could have taken some walks together!) But it gives me an opportunity: this novice was quite taken with Albert’s observation (echoed by Coleman in “Quantum Mechanics in Your Face”) that, after a measurement fork in the quantum mechanical garden paths, all observers see a definite result (but not necessarily the same result), with nary a superposition in sight. Is this result significant, trivial, or somewhere in between?

Art,

This is the heart of the measurement problem of quantum mechanics, and to my mind very much significant, with the difficult question that of providing a compelling explanation of how this happens.

By the way, I had somehow missed the fact that Gerard ‘t Hooft has published with Springer a book-length exposition of his heterodox approach to the foundations of quantum mechanics.

Daniel Tung wrote:

Yes, Daniel, that is indeed how it is supposed to work. You describe this as the situation becoming “more absurd.” Well, perhaps. But lots of people found special relativity absurd, and it seems to be true.

To be sure, this multiple infinity of universes does raise some questions about Occam’s razor, but the MWI proponents say the ideas are simple even if the MWI multiverse itself is not so simple.

The real problems are the ones that Peter Woit and I have mentioned, the “preferred-basis” problem as to why the universe “splits” along certain directions in Hilbert space and not in other directions, and the “probability-measure” problem as to how the multiple universes lead to Born’s rule.

From the beginning, various MWI proponents have claimed that MWI automatically produces Born’s rule, then other MWI proponents admit there are problems, then still others claim they have some new way of getting Born’s rule, and so it goes. I have been watching this process for nearly fifty years now, and I fear it reminds me a bit more of theology than physics (not that theology can’t be amusing!).

By the way, superficially there is a simple solution: you just announce that the probability measure on the space of possible universes just

isthe Born rule. Of course, then you are notderivingthe Born rule but just assuming it, just as we do in textbook QM.Furthermore, as physicists, we want the Born rule to be true in some frequentist interpretation, not just as an abstract measure. How do you do that? Well, often MWI proponents have spoken as if there are a large but finite number of universes in the MWI multiverse, and the probability measure is just a relative count of the number of universes in which one or another outcome occurs.

But that just does not work: You actually need an infinite number of universes, not a finite number, so naïve counting does not work. Also, the universes are not actually separate and discrete (that, after all, is the whole point of MWI: to be able to deal with quantum interference).

So, as far as I can see, the end result of decades of work and rather heated debate is that you just have to throw in Born’s rule with a frequentist meaning as an extra assumption in MWI, which largely defeats the point of MWI in the first place.

If you want to dig into this further, David Deutsch is an intense proponent of MWI who has been aware of the two basic technical problems and who, at least on occasion, thinks he has found solutions. Deutsch has done some interesting work connecting MWI to quantum computing and even to time travel and travel to parallel universes (he somehow got Phys. Rev. to publish the latter work!). My own feeling is that if Deutsch, who is indeed very bright, can’t get this bird to fly, and I don’t think he can, then no one can.

But, I am sure, many MWI proponents will have a different take: there is no final agreement in matters of theology!

Dave

From the review in the New York Review of Books by David Albert:

“It wasn’t until sometime in the 1980s that a small and embattled community of physicists, mathematicians, and philosophers, who had learned of the the theory [Bohmian mechanics] from Bell, began to take an active interest in what Bohm had done. His theory is now regarded as one of the two or three most important achievements in the history of our understanding of quantum mechanics.”

Regarded by whom? As an antidote to this, I highly recommend the following posting by Reinhard Werner, a first-rate mathematical physicist, who has made numerous contributions to quantum information, for those of you who do not know him: https://tjoresearchnotes.wordpress.com/2013/05/13/guest-post-on-bohmian-mechanics-by-reinhard-f-werner/

Mark Hillery,

Thanks for the comment and link. The fondness of philosophers (and a few physicists) for Bohmian mechanics has always mystified me.

Note that I’ve added some comments by the author, Adam Becker, to the posting above.

Mark Hillery: Regarded by who? Regarded by the people who actually work on foundations of physics and foundations of quantum theory in particular. A group to which Werner most certainly does not belong. Nor, obviously, does Peter Woit. You can start with Bell’s “On the Impossible Pilot Wave”, since Bell was the leading figure in foundations in the latter half of the 20th century. (If you disagree with the assessment, I would be extremely curious who else you think deserves that honor.)

Lindley’s book is excellent. It’s been clear for some time that Zurek and others’ work on decoherence is the way forward on this question.

The invocation of strings and multiverses for the many-worlds idea is hilarious — many-worlds dates from the 1950s, about 30 years before the rise of string theory as we know it now and 40+ years before the “landscape.” This is like string theory “predicting” Kaluza-Klein, inflation, gauge theory, general covariance, and what-not, long after those made their appearance. What claptrap.

Cophenhagen and its interpretation have roots long before the military-industrial complex. They didn’t know how to fully resolve the problem, although von Neumann isolated it in the measurement operation. Most physicists then didn’t care, and don’t care now, about the ultimate nature of reality implied by QM. They just know they can calculate with it and get amazing results. “Shut up and calculate.”