Thursday, January 17, 2013

Sean Carroll, Copenhagen, and consensus

The recently published poll about the interpretations of quantum mechanics was discussed by Sean Carroll under this dramatic title:
The Most Embarrassing Graph in Modern Physics
I would agree with him that the diverse results of the poll are somewhat strange and a reason for discouragement but I disagree with everything else he writes – what the poll means and doesn't mean, why we should be concerned, and so on. He's clearly a "consensus scientist" and make no doubts about it, all consensus scientists are hacks or at least sleezy jellyfish who spend their time by optimizing the solution to the task how to swim in the society so that they maximally benefit but they are surely not approaching Nature with the scientific integrity.

To show his embarrassment, he picked question 12 whose results look (almost) least embarrassing to me among all the questions.

The question and the composition of answers looks like this:

How did the participants of the poll do? They did pretty well but not very well. First, the most accurate answer, Consistent Histories, was picked by 0% of the participants; an excuse is that this superior interpretation is presented in no popular books and in no undergrad classes so superficial enough "researchers" simply remain unfamiliar with it. But the second most accurate answer, the Copenhagen interpretation, was preferred by 42%, by far the largest group.

There is a lot of room for philosophical flavors which may be idiosyncratic and many of the people are saying the same thing, physically speaking. One must distinguish different degrees of wrongness. If you look at the groups that have demonstrated that they have really misunderstood something and they are sharply wrong – because they believe in the existence of effects that demonstrably don't exist – you will find out that only 9% who believe in the "objective collapse" and perhaps also 18% who believe in the "many worlds" are sharply wrong.

Some of these wrong people use a scaled-down version of the interpretations (such as MWI) from which the demonstrably wrong claims have been removed. If you look at the graph with some degree of tolerance, you will almost certainly conclude that these answers show that most of the participants may actually understand all the physical questions correctly, they just differ in their preferred language, formalism, or in their philosophical accents. (The first commenter on Carroll's blog is suggesting pretty much the same thing.)

I would find questions where the answers would show a more widespread ignorance but from a global viewpoint, the results of the poll were a pleasant surprise for me.

Sean Carroll apparently thinks that the goal of the scientific research is a consensus and he wrote an incredible amount of fundamentally wrong statements that seem to boil down to this fatal delusion of his. Let us look at them:
Inevitably some killjoy will loudly grumble that “scientific questions aren’t decided by voting!”, but that misses the point.
The people who realize that scientific questions aren't decided by voting are merry, pretty people who – first of all – understand an important point about science. On the contrary, it's the people who try to criticize or sling mud on this key, refreshing, and profound finding who are obnoxious sourballs. And who are badly wrong.

The merry, pretty people aren't missing any point. To say the least, their proposition about the irrelevance of the consensus doesn't prove that they are missing a point.
A poll of scientists isn’t meant to decide questions, it’s meant to collect data — mapping out the territory of opinion among people who have spent time and effort thinking carefully about the relevant questions.
The key point that Sean Carroll – and other folks – don't understand is that the poll results are not scientific data at all. At most, they are sociological data. But these are totally different things. In science, polls simply don't represent evidence. Scientific questions can't be decided by voting. Scientific questions must be decided by observations, experiments, and logical reasoning and calculations applied to these observations and experiments as well as the universal laws of logic and probability.

If 58% (or fewer or more) people at the conference (and similar conferences) failed to have found the best answer, it only suggests that the answer is too hard for many people – or that the people are just not competent enough. Or some linear combination of these two statements. But it surely doesn't mean that no one knows the right answer. It certainly doesn't imply that science is ignorant about the right answer.
I’ll go out on a limb to suggest that the results of this poll should be very embarrassing to physicists. Not, I hasten to add, because Copenhagen came in first, although that’s also a perspective I might want to defend (I think Copenhagen is completely ill-defined, and shouldn’t be the favorite anything of any thoughtful person).
The Copenhagen interpretation may only be called "ill-defined" by those people who don't want to learn the truth but who demand the accepted laws of physics to confirm and pay lip service to their misguided, pre-existing, intrinsically classical misconceptions. Because they don't hear any compliments about their misconceptions coming from Copenhagen, they declare the discoveries coming from Copenhagen – arguably the most important discoveries in the 20th century science – ill-defined. Needless to say, Sean Carroll is a textbook example of such an anti-quantum bigot. In reality, the Copenhagen interpretation says everything one needs to know to do physics, everything by which the framework of quantum physics supersedes the now defunct framework of classical physics. If I simplify a bit, it says that the basic sketch of physics is to probabilistically predict the values of observables in the future out of results of observations in the past according to Hilbert-space, linear-algebra-based rules it describes in detail. That's it. All of science may be reduced to this basic set of procedures, at least in principle. Every observer who knows how to deduce values of observables from his perceptions may start to verify the predictions of quantum mechanics – this intellectual skyscraper with the foundations built in Copenhagen – and be sure that everything works. Nothing is missing here. There isn't any "extra physics" outside observers, physics independent of observations. Physics is a set of quantitative rules to probabilistically relate various observations.
The embarrassing thing is that we don’t have agreement.
No, this is a ludicrous usage of collective guilt. The people who choose – pretty much – the right answer have nothing to be embarrassed about. It's only the people who choose the wrong answers who should be embarrassed. In particular, I think that those 42% – and perhaps a few other groups – have no reason to feel guilt. They can't do much about their deluded colleagues' opinions – these questions, however fundamental, are simply too difficult or too conceptually new for too many people.

Even if I were wrong about the foundations of quantum mechanics – which isn't a realistic scenario at all but I still have to mention it because the validity of some claims about quantum mechanics is an independent question from the validity of claims about the consensus – it would still be almost certainly true that at least one group of the participants has nothing to be embarrassed about. And it's the people who are right who matter in science. They're the roots of a ramified tree of science that will exist in the future. People who are wrong today are just dead ends whose impact on proper science will converge to zero. There's no "equivalence" between the truth and untruth in science.

We may discuss whether it's because most people are too intellectually limited or stubborn or because the insights are objectively too hard and new but obviously, we don't have a method to answer which of those answers is right. Only the relative sociological statement, "foundations of quantum mechanics are too hard and new for most people", may be defended by referring to similar polls.
Think about it — quantum mechanics has been around since the 1920′s at least...
And yet — we don’t understand it.
No, there is no evidence that "we" don't understand it. Indeed, this statement is wrong. Instead, what the diversity of the poll results show is that "at least some groups of people don't understand it". But the diversity of the results doesn't mean that no one understands it! In the U.S., about 50% of the people believe that the biological species were created by a Creator or a Creation Operator or whatever was Her name within several days. But this fact doesn't mean that we – scientists et al. – don't understand that the right explanation of the origin and diversity of species is Darwin's evolution. Of course we understand it. People who say wrong things can't "eradicate" the understanding by those who actually do understand an issue. They can't do it, not even if they call themselves scientists, not even if their environment calls them scientists, not even if they really try to be scientists! It just doesn't matter. What matters is that someone has accumulated and processed the evidence correctly and deduced the right conclusions – which may the case whether or not this individual or these individuals enjoy a prestigious status within the society.
I put it this way: here in 2013, we don’t really know whether objective “wave function collapse” is part of reality (as the poll above demonstrates).
Again, this is nonsense. We do understand whether an objective wave function collapse occurs. It doesn't. Heisenberg and others have understood it doesn't since the beginning in the mid 1920s. The poll can't demonstrate anything else. Quite generally, and we have already said it, polls can't demonstrate anything about questions in natural science. Only scientific evidence may have certain implications for the truth values of propositions about science. By repeated proclamations of the kind above, Sean Carroll proves that he has nothing in common with science. He is just a sleazy slick jellyfish swimming among genuine scientists in the scientific institutions (and among other jellyfish) that is emulating statements by others in such a way so that he optimizes his own benefits. He's an immoral jerk. But every scientist knows that the diverse results of a poll don't imply that science doesn't understand the corresponding question. Genuine science just doesn't give a damn about polls – and just to be sure, this is the good news and only sourballs may doubt it.
I’m optimistic that we will, however. And I suspect it will take a lot fewer than another eighty years. The advance of experimental techniques that push the quantum/classical boundary is forcing people to take these issues more seriously.
I don't think so. Similar experiments that demonstrate that quantum mechanics really works, including all of its counterintuitive predictions – and I mean the Copenhagen quantum mechanics and/or any other intrinsically probabilistic scheme using the maths of quantum mechanics to determine probabilities of some propositions about Nature – and that none of the proposed new effects implied by the candidate "replacements of the proper Copenhagen quantum mechanics" exists have been around for more than half a century, too. Even the "tests of EPR phenomena" that were accompanied by pretty much the current formulations and motivations have been done since 1976 or so. And lots of directly relevant experiments have been done since the 1920s – and some of the key "smoking guns" proving that quantum mechanics is right and its (more or less) classical competitors are wrong are even older.

The objects for which quantum mechanics was shown to hold are already so large, complex, and diverse that it makes really no sense to try to increase this cutting edge further. Nevertheless, you will still find crackpots who will tell you, for example, that the Copenhagen interpretation is ill-defined. They say such cosmically preposterous things because quantum mechanics doesn't really agree with their basic world view, with their ideology that underlies everything they want to know or learn. So they will just never accept it, regardless of the progress in experiments that study the foundations of quantum mechanics, regardless of any amount of indisputable evidence or rigorous proofs.

In other words, Sean Carroll and many others are hopeless bigots who can't ever accept quantum mechanics in its genuine, fundamentally non-classical form. I find it very likely that such people will be around in 80 years from now, too. Their percentage in various institutions may even be higher than it is today. But that doesn't imply that science will cease to exist. Science is done by those who do it right and who are not bigots, however small the set of these people is.

And that's the memo.


  1. WFT, I thougth only non physicists, such as Sklivvz on Physics SE for example, believe that answers to physics issues are decided politically by polls, with every dimwit that has no clue about the topic allowed to vote, instead of experiments and theoretically correct considerations, the answers to scientific questions should be homogenized, etc ... Now Sean Carroll says similar things too ...?

    The usage of "we" in sentences as "we know (or dont know) that ..." alway confused me, but when Sean Carroll says "we dont understand QM" and means the whole scientific community (or even the whole mankind) by this it seems very wrong and inappropriate to me. The only thing he could say (in my opinion) that he does not understand it (if this should really be the case) but it is certainly not true that nobody understands it!

    Why does he not like the Copenhagen interpretation, what does he favor instead?
    Maibe the 0% for the consistent histories are due to people not exactly knowing what it is, never heard about before, etc ...?
    I have never heard about this at university and only learned about it by a certain nice TRF article :-)

  2. Yup, Dilaton, I noticed that Carroll shares this "science is settled by polls" many years ago and this pernicious delusion is very important for him because he's been always among the first ones at the battlefront to ideologically and politically abuse the "stamp of a scientist" to vote about political things, among the top champions of the climate hysteria - despite his not-at-all energy-efficient cars etc. - and similar things.

    It's sad and this attitude of his isn't really shared by most of the research physicists. But it *is* shared by most of those who actually have power these days.

  3. Wow, completely baffled. I completely agree with you. Science is not decided by voting, see Feynman and the Emperors nose (

    Furthermore, most of my friends (ok they are lumberjacks and such so they don't care) can't answer correctly to 21-3*5+6 on facebook. I don't conclude that humanity does not understand math, I conclude some people don't care for it and don't bother learning it.

  4. Exactly.

    The percentage of people who understand a question depends on the question - its "difficulty" but also other attributes. The most esoteric and technical questions in string theory - and other fields, whether or not they are close to experiments - are only understood by a dozen of people in the world, maybe less.

    It's questionable whether the percentage of "informed people" is increasing. It depends how you match the questions between two times. I think that for fixed questions, it's probably still increasing a bit, at least for questions that still matter for things people are working on.

    But if one expected progress - i.e. if you compare current cutting-edge physics questions and the current people's knowledge about them with the 1930 cutting-edge questions and the 1930 people's knowledge of them - I am afraid that the progress has clearly been negative. People understand the cutting edge research of their time less and less clearly. Perhaps they are getting less intelligent or interested; perhaps the research is just getting harder and too hard.

    But whenever the number of people who follow some observations and arguments and maths behind a theory is greater than 0, and 1 is enough, science may work and do progress. It's complete nonsense that the people who follow it must be greater than 3.5 billion or any of the preposterous things Carroll seems to suggest. There are just people who got it - competence, honesty, hard work etc. - and people who have not. Institutions should try to support the former more than the latter, it's meritocracy, but the former don't have to "wait" for the majority of the latter!

  5. I don't know anything about quantum theory, so maybe you can explain something to me. I noticed the votes total 129%. Is that due to some quantum effect? Maybe spooky answers at a distance?

  6. I suppose they were allowed multiple options and the options aren't quite mutually exclusive - or at least the participants decided they were not.

  7. I remember schoolmates with zero interest to math, and clearly not bright on any level... When our math teacher would ask them to compute 13+17, they said that they don't know. She would then change tactic and ask: You have 7 dollars and 53 cents (actually rubles and kopecks were involved) and you buy a lunch for 5 dollars 75 cents, and before she even formulate the question they would immediately say that 1.78 is left. It had become a running joke...

    Mathematician Gelfand used to claim that one can teach ordinary folks rather advanced math, one should just formulate it in appropriate terms. For instance, Gelfand claimed that heavy drinkers understand fractions. It goes like this: "If you ask them, 'Which is larger, 2/3 or 3/5?' it is likely
    they will not know. But if you ask, 'Which is better, two bottles of vodka
    for three people, or three bottles of vodka for five people?' they will
    answer you immediately. They will say two for three, of course."

  8. Lubos, I object to you calling these cretins "sourballs". Sourballs are a wonderful candy I enjoyed as a kid. You should not taint this pleasant memory of mine. ;-)

    Perhaps you could start referring to the likes of Carroll and Shmoit as "shitballs" or "dingleberries"?

  9. Does Feynman's "sum over histories" explanation of quantum electrodynamics that he presents in his semi-popular lectures on QED (The Strange Theory of Light and Matter) have anything to do with consistent histories?

  10. Well, they're histories but they're far from the same thing. Feynman's histories are summed over to get physically interesting predictions, consistent histories aren't summed over. Feynman's histories are continuous and all of them while a key point of consistent histories is that they are coarse-grained, clumped into "groups".

  11. Aah, dingleberries is funny, I had to quite google a while to find out what it is ;-).

    I never thought that sourballs could be pleasant as you say :-D.

    I derieved the term from the German word "Miesepeter" (note that the second part Peter occuring in the word, it is very suitable ;-) ... Peter F I like of course), which means something like "sourpuss" in English according to Leo, a grumpy person who is against everything and complains about everything without reason, etc ...

  12. Lubos - I've never found a powerful rebuttal to the claim that the layman of little knowledge must base his judgements about scientific questions on "scientific consensus" since it is "necessary" for him to make some judgements, e.g., about anthro. global warming. I could outline my attempts, but I don't want to make this tediously long. Do you know of a powerful rebuttal?

  13. Dear Smoking Frog, nope, I don't know a powerful rebuttal, especially because I think that the claim is half-right, half-wrong.

    A laymen who can't verify things reliably enough, with the full-fledged scientific method, must obviously rely on some approximate and not quite reliable or scientific ways to judge scientific questions.

    Looking for majorities among "official scientists" while considering oneself a "complete idiot" - and regardless of all other considerations - is one such a strategy, not the only one and not necessarily the cleverest one. But once again, there's no "quite reliable" way to reach scientific conslusions if one doesn't do it by the proper scientific method himself.

  14. Dear Lubos,

    I completely agree that the Copenhagen interpretation is the correct one. ( I'm not familiar with consistent histories. ) Decoherence theory allowed us to understand the measurement process more deeply, but clearly nothing is invalidated from the earlier insights.

    I think that a lot of misunderstanding is coming from the fact that most people are talking about wavefunctions, and they think that these are physical waves. The reason behind this is that in one particle nonrelativistic QM, the wavefunction is defined on R^3 and one can think that this is the physical space. But this is wrong, because if I go to many particle nonrelativistic QM, then the wavefunction is defined on R^3N . It is clear that this is not the physical space. This is the configuration space of the system. The wavefunction is the generalisation of a classical probability distribution.

    This becomes even more evident in second quantised language. The "wavefunction" is no longer a function at all, it is an element of an abstract Hilbert space.

  15. OK, thanks.

    not the only one and not necessarily the cleverest one.

    I think that's true, but I've found it hard to defend. I've found that most proponents of consensus-based judgement have little or no idea of what the skeptic scientists say. They even have little idea of what the "consensus" says. Now, I believe that, in some cases, esp. global warming, it is possible to "smell" which side is more likely right, but that's hard to defend.

  16. Dear Rezso, I agree almost completely. Well, in QFT, the Hilbert space is not too abstract. It may be represented in many ways - Fock space, space of wave functionals of the field configurations, and so on.

    I also agree on the Heisenberg equations' being the right quantum counterpart of the classical equations of motion.

  17. Dear Lubos, thanks, you are right, I forgot that we can use wave functionals in QFT. ( I never used them personally. )

    One more comment on the Heisenberg-picture.
    I suspect that many people believe that QM is some kind of a sloppy, unnatural, ad hoc mathematical construction, which should be replaced by something more fundamental.

    The Heisenberg-picture clearly shows that this is not true. It tells us that QM is nothing more than the noncommutative generalisation of CM. Noncommutative algebras can be represented on Hilbert-spaces, and to preserve the linearity of the structure, we define the expectation value to be Tr(rho*A). The mathematics is natural and unique.

    This also means that one isn't really allowed to come up with sloppy interpretations of QM, because you are not allowed to mess with this mathematical structure.

  18. Uh, it seems that the participants violated unitarity. :(

  19. Off Topic

    Lubos, I have been trying to understand the Little Higgs theories for a near layperson. How do they solve the hierachy problem? What is a top prime? What is a vector like fermion and how is it different from a regular fermion. They seem to come from a symmetry of SU(5)/SO(5). What does the division sign mean?

    Thanks Physics Junkie

  20. Dear Physics Junkie,

    I wrote about this theme e.g. here:

    SU(5)/SO(5) is the coset, quotient, i.e. the set of all equivalence classes {g} of elements in SU(5) ("orbits") such that g1 is equivalent to g2 if g1=g2*h for some h in SO(5).

    Because SU(5) is a 24-dimensional group and SO(5) is 10-dimensional, the coset is 14-dimensional, the difference.

    This is kind of rudimentary maths here and most of the maths in little higgs models etc. is somewhat more difficult so I don't want to discourage you but it may make little sense to study those things without the appropriate math background.

  21. And one more comment. :)

    I think that the Heisenberg picture instantly allows us to falsify the objective collapse theories.

    My argument is the following. Let's start with the linear Schrödinger equation. We can go to the Heisenberg-picture and calculate the equations of motion. We will simply obtain the classical equations with hats.

    In the hbar=0 limit these equations reproduce the classical results.

    Now, let's modify the Schrödinger equation with a nonlinear term, to allow objective collapse. But now, we can no longer go to the Heisenberg picture, because unitarity is lost. This means that we have lost any correspondence with the classical theory.

    This clearly shows that the mathematical structure of QM is rigid.

  22. I at least morally agree with that. ;-)

  23. And to rub it in, he occupies Feynman's office...Feynman who would have spit about the idea of deciding science by poll.
    The problem with Sean is that his overactive PC leads him to support this consensus idea about truth. I don't really like the Copenhagen interpretation, but so what. Nature doesn't care what I like. I heard Jim Hartle talk about consistent histories, and it makes sense to me.

  24. The question about your favorite interpretation of QM is like asking, “Who is your favorite actress?”. It implies a mapping of possible answers into a framework that you are already carrying around in your head. Of course different people will give different answers; they have different frameworks.

    That is why the question is not only unscientific but profoundly unhelpful. Really grasping the essence of QM requires dismantling a large part of that framework rather than distorting QM to try an get a fit. Even talking about an interpretation of QM is usually the beginning of a journey into fantasyland.

  25. well, you can also have other senses too: when the consensus says :"it is warming" and you are freezing, it is time to use the saying : whom shall i believe, you or my lying eyes?

  26. anna v - Yes, but I'm far from persuaded that we're freezing.

  27. Something I forgot to mention, but which I think is important: The "consensus" argument is deceptive in that it exploits a confusion between a common meaning of "consensus" and an uncommon one. Consensus is usually about what to do about something, not what is true of something. The former usage is more or less unobjectionable, but this is not true of the latter. For example, if 9 out of some 10 people think my wife is at the supermarket, while the other thinks she's at her sister's house, the mere fact of consensus (about what is true) does very little to falsify that 10th person's belief, but if the 10 must find her, it does make sense to look first in the supermarket, if only because a consensus about what to do may be necessary.

  28. Very perceptive, never seen this explanation before but it strikes me as accurate.

  29. Eugene S - THANK YOU! Some people have rejected my argument, and some have said it was of no importance. Very few have agreed with it. I think that's because most people's understanding of language is insufficiently subtle. They may understand something in usage, but they don't notice that they are understanding it.