Monday, November 07, 2016

Steven Weinberg doesn’t like Quantum Mechanics. So what?

A few days ago, Nobel laureate Steven Weinberg gave a one-hour lecture titled “What’s the matter with quantum mechanics?” at a workshop for science writers organized by the Council for the Advancement of Science Writing (CASW).

In his lecture, Weinberg expressed a newfound sympathy for the critics of quantum mechanics.
“I’m not as happy about quantum mechanics as I used to be, and not as dismissive of the critics. And it’s a bad sign in particular that those physicists who are happy about quantum mechanics, who see nothing wrong with it, don’t agree with each other about what it means.”
You can watch the full lecture here. (The above quote is at 17:40.)

It’s become a cliché that physicists in their late years develop an obsession with quantum mechanics. On this account, you can file Weinberg together with Mermin and Penrose and Smolin. I’m not sure why that is. Maybe it’s something which has bothered them all along, they just never saw it as important enough. Maybe it’s because they start paying more attention to their intuition, and quantum mechanics – widely regarded as non-intuitive – begins itching. Or maybe it’s because they conclude it’s the likely reason we haven’t seen any progress in the foundations of physics for 30 years.

Whatever Weinberg’s motivation, he doesn’t like neither Copenhagen, nor Many Worlds, nor decoherent or consistent histories, and he seems to be allergic to pilot waves (1:02:15). As for qbism, which Mermin finds so convincing, that doesn’t even seem noteworthy to Weinberg.

I learned quantum mechanics in the mid-1990s from Walter Greiner, the one with the textbook series. (He passed away a few weeks ago at age 80.) Walter taught the Copenhagen Interpretation. The attitude he conveyed in his lectures was what Mermin dubbed “shut up and calculate.”

Of course I as most other students spent some time looking into the different interpretations of quantum mechanics – nothing’s more interesting than the topics your prof refuses to talk about. But I’m an instrumentalist by heart and also I quite like the mathematics of quantum mechanics, so I never had a problem with the Copenhagen Interpretation. I’m also, however, a phenomenologist. And so I’ve always thought of quantum mechanics as an incomplete, not fundamental, theory which needs to be superseded by a better, underlying explanation.

My misgivings of quantum mechanics are pretty much identical to the ones which Weinberg expresses in his lecture. The axioms of quantum mechanics, whatever interpretation you chose, are unsatisfactory for a reductionist. They should not mention the process of measurement, because the fundamental theory should tell you what a measurement is.

If you believe the wave-function is a real thing (psi-ontic), decoherence doesn’t solve the issue because you’re left with a probabilistic state that needs to be suddenly updated. If you believe the wave-function only encodes information (psi-epistemic) and the update merely means we’ve learned something new, then you have to explain who learns and how they learn. None of the currently existing interpretations address these issues satisfactorily.

It isn’t so surprising I’m with Weinberg on this because despite attending Greiner’s lectures, I never liked Greiner’s textbooks. That we students were more or less forced to buy them didn’t make them any more likable. So I scraped together my Deutsche Marks and bought Weinberg’s textbooks, which I loved for the concise mathematical approach.

I learned both general relativity and quantum field theory from Weinberg’s textbooks. I also later bought Weinberg’s lectures on Quantum Mechanics which appeared in 2013, but haven’t actually read them, except for section 3.7, where he concludes that:
“[T]oday there is no interpretation of quantum mechanics that does not have serious flaws, and [we] ought to take seriously the possibility of finding some more satisfactory other theory, to which quantum mechanics is merely a good approximation.”
It’s not much of a secret that I’m a fan of non-local hidden variables (aka superdeterminism), which I believe to be experimentally testable. To my huge frustration, however, I haven’t been able to find an experimental group willing and able to do that. I am therefore happy that Weinberg emphasizes the need to find a better theory, and to also look for experimental evidence. I don’t know what he thinks of superdeterminism. But superdeterminism or something else, I think probing quantum mechanics in new regimes is best shot we presently have at making progress on the foundations of physics.

I therefore don’t understand the ridicule aimed at those who think that quantum mechanics needs an overhaul. Being unintuitive and feeling weird doesn’t make a theory wrong – we can all agree on this. We don’t even have to agree it’s unintuitive – I actually don’t think so. Intuition comes with use. Even if you can’t stomach the math, you can build your quantum intuition for example by playing “Quantum Moves,” a video game that crowd-sources players’ solutions for quantum mechanical optimization problems. Interestingly, humans do better than algorithms (at least for now).

[Weinberg (left), getting some
kind of prize or title. Don't
know for what. Image: CASW]
So, yeah, maybe quantum physics isn’t weird. And even if it is, being weird doesn’t make it wrong, and therefore you don’t think it’s a promising research avenue to pursue. Fine, then don’t. But before you make jokes about physicists who rely on their intuition, let us be clear that being ugly doesn’t make a theory wrong either. And yet it’s presently entirely acceptable to develop new theories with the only aim of prettifying the existing ones.

I don’t think for example that numerological coincidences are problems worth thinking about – they’re questions of aesthetic appeal. The mass of the Higgs is much smaller than the Planck mass. So what? The spatial curvature of the universe is almost zero, the cosmological constant tiny, and the electric dipole moment of the neutron is for all we know absent. Why should that bother me? If you think that’s a mathematical inconsistency, think again – it’s not. There’s no logical reason for why that shouldn’t be so. It’s just that to our human sense it doesn’t quite feel right.

A huge amount of work has gone into curing these “problems” because finetuned constants aren’t thought of as beautiful. But in my eyes the cures are all worse than the disease: Solutions usually require the introduction of additional fields and potentials for these fields and personally I think it’s much preferable to just have a constant – is there any axiom simpler than that?

The difference between the two research areas is that there are tens of thousands of theorists trying to make the fundamental laws of nature less ugly, but only a few hundred working on making them less weird. That in and by itself is reason to shift focus to quantum foundations, just because it’s the path less trodden and more left to explore.

But maybe I’m just old beyond my years. So I’ll shut up now and go back to my calculations.


Phillip Helbig said...

"It’s become a cliché that physicists in their late years develop an obsession with quantum mechanics."

On a somewhat similar note, Chandra once said "Cosmology is the graveyard of astronomers." When he became interested in cosmology in his old age, a colleague reminded him of the words spoken in his youth. :-)

Phillip Helbig said...

"It’s become a cliché that physicists in their late years develop an obsession with quantum mechanics. On this account, you can file Weinberg together with Mermin and Penrose and Smolin."

To be fair, Penrose has been like this for at least 40 years, so started when he was at most about our age. I don't know if that means that we are old. :-)

Kevin Van Horn said...

"If you believe the wave-function is a real thing (psi-ontic), decoherence doesn’t solve the issue because you’re left with a probabilistic state that needs to be suddenly updated."

I'm not following you here. In the Everettian view you just have a continuously evolving wave-function, don't you? Or are you talking about the measurement process? In that case, what about Sebens and Carroll's paper on the notion of self-locating uncertainty as the origin of QM probabilities?

J. e. L. said...

It is always too easy to ask a question; answering them can take a bit of work. But if you are making a list of ideas for future posts, I would be very much interested in a short primer on superdeterminism and possible experimental tests of it.

Uncle Al said...

Quantum mechanics glories in clean diagonals. The fun parts arise from off-diagonal terms. Enforcing the former by parameterizing the latter is elegantly wrong.

"To my huge frustration, however, I haven’t been able to find an experimental group willing and able to do that" and so intense chiral Equivalence Principle violations. Euclid had a 2000-year run, Newton got 200 years, both have weak postulates. Neither GR nor QM are whole. Anomalies are diagnostics.

A 250 ml glass holds 125 ml of water. Philosophers rant, [|full> + |empty>]/√2, etc. Demand a 125 ml glass? Temperature and differential linear coefficients of thermal expansion....

Per Arve said...

What about Many-Worlds Interpretation, you write that Steven Weinberg doesn't like that. But, he actually gives the reason that there is a problem with quantum probabilities. Hartle wrote a derivation in 68 that Weinberg dismisses on the ground that it uses an infinitely repeated state. Weinberg using such infinite-dimensional spaces is a 'stretch' that he refuses to take seriously. The derivation holds. I am sure the derivation holds and I have contacted Hartle about it and he says it is correct.

Nevertheless, I was provoked by Weinberg's QM book and started to work on it. I have a version in the arxiv that I am currently revising. I will contact you when it is ready.

Sabine Hossenfelder said...

Kevin, Per,

I have to admit that I'm not sure why it matters to you what I think of the Many World's interpretation. But since you ask, let me put it this way:

You can always remove the problems you have with a troublesome axiom by just dropping the axiom. Indeed, you can remove all troubles with all theories by just removing sufficiently many axioms. Because a theory without axioms doesn't have any problems. (That's Tegmark's approach.)

But the consequence of this is that you end up in a multiverse of one or the other type. In the multiverse the question then is no longer "why this damned axiom" but "why are we where we are" in the multiverse. Then you're told that this isn't a reasonable question to ask. This I think is equally bad as saying it's not a reasonable question to ask what the wave-function does before it was measured. It doesn't really solve any problem. Many Worlds, decoherent histories, and so on, for all I can tell in the end always leave you with a probability distribution for what you measure. Best,


Haus said...

Are you familiar with information theoretic reconstruction of QM? See D'Ariano and Hardy.

Per Arve said...


In response to your response, I want to convince you and the rest of the world that many-worlds is a description of the world as we observe it, nothing more, nothing less. I cannot see that there is a problem with a probability distribution for what will be seen in the future if we can explain why that is. Any other description will deviate from quantum mechanics. If the there is no problem that it solves and there is no sign of its existence it seems me strange to believe in such a process. I hope you are open for further discussion when I am ready to show you that there is no problem, MWI describes the world.


travis said...

Yes, all interpretations of quantum mechanics ultimately leave you with a probability distribution for what you measure, but as far as I know only Many Worlds and Bohemian mechanics (which some would argue is many worlds in disguise) give a consistent theory from which you could hope to eventually explain how the measurement process happens instead of just saying "I know a measurement when I see it." And it has already been proven by David Wallace and others that you can derive the Born rule from many worlds with just the barest assumptions about decision theory. So in that sense many worlds is strictly better than Copenhagen because Copenhagen (which is a vaguely defined interpretation) can already be derived as an approximation of many worlds (which is a well-defined interpretation), that is, Copenhagen is an approximation where you're always ignoring the part of the universal wave function that no longer interacts with you.

Sabine Hossenfelder said...


We're talking past each other. I'm not interested in debating the merit of different interpretations that lead to the same results. I couldn't care less. As I said, I'm a phenomenologist, for me the question is whether there is a different, more fundamental, theory that gives rise to quantum mechanics. This theory *will* deviate from quantum mechanics, that's the reason I'm interested in it. Yes, MWI describes the world. Is all fine by me. Best,


Sabine Hossenfelder said...


I almost certainly heard one-two-three talks by Hardy about this. I have no objection to this. But like Chris Fuchs' QBism I find it kind of pointless. If a reinterpretation doesn't lead to testable new predictions, it's not for me. On that account I have to admit to having a liking for collapse models, just that the ones I've seen are so obviously made-up just to show it's possible that I find them utterly unconvincing. Best,


daedalus2u said...

I am also a fan of non-local hidden variables and superdeterminism too.

I think the "problem" in proposed tests is in specifying an "isolated" system. Systems are not "isolated" and cannot be "isolated", they are bathed in dark energy and a background field of gravitational radiation (both of which are poorly specified) which cannot be ignored (or as yet measured).

The properties of dark energy are pretty much unknown, other than it interacts with the space time metric causing expansion. If it interacts with the space-time metric, then it interacts with *everything*.

Gravitational radiation also interacts with the space-time metric.

An issue with standard treatments of gravitational radiation is that there is no consideration of the “back reaction”, the gravitational radiation produced when gravitational radiation interacts with mass/energy. Ignoring this is necessary to preserve the equivalence principle (equating inertial and gravitational mass), but this is non-physical (in my opinion). Moving any real mass in a gravitational field necessarily generates gravitational radiation. Ignoring this for “test masses” is non-physical and cannot be correct.

akidbelle said...

Hi Sabine,
I like this post but please, don't shut up :-)

What about Cramer's interpretation? I am surprised you do not mention it. As far as I know all theories are time-symmetric.


Sabine Hossenfelder said...


Argh, look, you guys can now go on for a couple of weeks asking me well, what about the transactional interpretation and the modal thing and the Montevideo interpretation and Rovelli's what's it called, or maybe that was the same. But I'm not sure what you think you gain by asking me about my beliefs.

Let me put it this way: I looked at all of them at some point, and didn't find they solve the problem I pointed out. So let me instead ask you in return: How do you think Cramer solves the problem? My problem with the transactional interpretation is that all I could find about it was infinitely vague. I want a derivation that says, here's the assumptions, and there's the limit in which we get QM back and here's the prediction that allows you to test that. And then I want to see the results of that test. Best,


PS: I think time-symmetry is a key-point, but the transactional interpretation doesn't include it correctly.

akidbelle said...


OK, I understand the "Argh". Best,


PS: what do you mean "not correctly"?

Sabine Hossenfelder said...


I don't know what I mean, sorry. That's the thing with intuitions I guess. I tried to make sense of it and I just couldn't. That doesn't mean it's wrong, just that I couldn't see how it would work.

S.E. Grimm said...

All the interpretations of quantum field theory have to do with phenomenological physics (empiricism). However, I am afraid that it isn’t so difficult to prove mathematically that physicists can never understand the underlying reality of the quantum fields with the help of phenomenological physics. Unfortunately, without a general concept of the structure of the quantum fields it is impossible to find the proof. In other words: physicists have to overhaul the foundations of physics, but phenomenological physicists have no idea how this must be done. So nothing will change.

Louis Tagliaferro said...

Science is often mucked up by scientists and there is no way to avoid it. You’ve discussed the reason in many of your blogs and it is human behavior. Support will always be very thin for re-writing the currently accepted standards. I think Max Planck’s correspondence to Einstein regarding the latter’s intention to re-write gravity sums up why, he wrote, “As an older friend, I must advise against it, In the first place, you won’t succeed, and even if you do, no one will believe you.”

There are many good scientific reasons to consider redoing QM and GR but most of the people established in those fields would likely advise against it as Planck did with Einstein.

Quantum Moxie said...

I'm not even trying to make them less weird. After all, weirdness is just as much an aesthetic concept as ugliness. I'd be happy if they were logically consistent, empirically valid, and generally obeyed Occam's razor (and thus explained the most with the least). It's that last point that people too often miss. I think we sometimes overcomplicate things. That's not to say that the universe isn't complicated, just that we tend to forget that we should be starting with a spherical cow (for lack of a better analogy).

Pascal said...

Sabine, can you tell us why you think quantum mechanics should be replaced by a better underlying theory? Is it "just" that it's incompatible with general relatvity or do you have other issues with quantum mechanics?

Arun said...

If you believe the wave-function only encodes information (psi-epistemic) and the update merely means we’ve learned something new, then you have to explain who learns and how they learn.

IMO, this is in some sense simpler than the psi-ontic position, because it is of a similar nature as asking "how does entropy increase in a system governed by time-reversible laws"?

JSV said...

My highlights here: "quantum mechanics … needs to be superseded by a better, underlying explanation", and "…finding some more satisfactory other theory, to which quantum mechanics is merely a good approximation."

Theorists shouldn't be hungry to destructively eliminate ideas that work, like QM/QFT. Instead, they should be looking at general ideas that allow the other theories to work in their respective effective domains, explaining why those domains apply, and telling you what should be outside them, or why it's meaningless.

They should be identifying symmetries and dualities of more general representations, giving context and meaning to our existing abstractions, discovering implicitly hidden images in the algebra, identifying where transformations lose information because their algebra isn't quite right, finding other ways to accurately explain the unintuitive stuff without ruining our chances of really understanding, and so on.

naivetheorist said...


this seems relevant

when Dirac was asked, “What’s the answer to the measurement problem?” his response was, “Quantum mechanics is a provisional theory. Why should I look for an answer in quantum mechanics?”


Tom Andersen said...


I was going to ask 'what about the transactional interpretation' - but I see you have looked at that. :) My view on the transactional interpretation is that it might help explain the mystery of Bell's theorem - something along the lines of advanced solutions can make some underlying quantum substrate able to be causally local but still have the instant connection that QM experiments show exists.

gomesha said...

Two comments: 1) Per, can you point me to this derivation by Hartle that you referred to, please?
2) Sabine, have you ever read the book The Beginning of Infinity, by David Deutsch? I highly recommend it. In it, he constructs a scathing criticism of "instrumentalism". He, in my view rightly, points out that the point of science is explaining the world, not merely predicting. He has many arguments and examples to back up the failings of instrumentalism (and even reductionism) as explanations, but one which I found amusing was the following: suppose there is a magic trick in which the magician saws the woman in half. Even if you predict correctly 100% of the times that the assistant will emerge unscathed, you have not succeeded in explaining the trick.

Per Arve said...


The referense is gomesha J. B. Hartle, American Journal of Physics 36, 704 (1968).


Q said...

Looks like 't Hooft has done some work on this:
(The Cellular Automaton Interpretation of Quantum Mechanics)

piein skee said...

Gomesha - David Deutsch believes there is no difference between science and philosophy and that predictions and real world measurements can be put aside or relegated in favour of 'criticism'. He says rational people will acknowledge critical refutation and put aside their ideas.

But have you ever asked Deutsch how much criticism from independent viewpoints has any of his ideas benefited from? Have you ever asked him to name what core position of his own has he ever conceded due to a criticism on the format he proposes other people operate? His ideas are unworkable even by himself.

Deutsch in my view, because his ideas are never at the scientific standard, never predictive, and for that reason largely ignored by scientists, seeks instead of raising the standard of his work, lobbying for the collapsing of standards generally and presumably getting his ideas in that way.

Recommendation: Don't bother with 'The Beginning of Infinity;'

Stuart said...

Finding a basic postulate for quantum gravity via phenomenology rather than theory based research(which IMO is becoming biased ) will reveal the origins of Quantum Theory .

Galaxy rotation curves & dark energy are phenomena that can point towards QG which in turn will give a general picture of QT.

Hakan Tomaşoğlu said...

i think, for a while, and of course the opinion in my mind about why people so close minded toward other directions rather than their way.

weinberg is an enoughly clever man rather than you and your commentors (except uncle al, he is the one i try to figured out, a different drop in the ocean).

so we need more tolerance, more thinking and open society, indeed the arab spring in the europe must be the answer of this kind spirit. you may see this behaviur brutally oppposite than foundational fathers... where are the repulsive intensity of your society?

please banned this, dont give a fuck...

ppnl said...

Would quantum computers disprove superdeterminism? I think Gerard ’t Hooft's version would kill quantum computers but would any version kill quantum computers?

емиванов фф said...

Remember that Monty Hall problem, or Bertrand's pardox? Or the principle of indifference?

Nobody really understands probabilities, that's the problem with QM. As usual we try to look elsewhere, hoping that the problem somehow will go away. Probabilities are deeply counter-intuitive and, accordingly, breed paradoxes or even outrageous nonsense e.g. multiverses and the like. There are axiomaticized versions but they obviously don't tell anything about the world. "Random" means just that we have decided to look no further. A constrained form of ignorance projected back on Nature itself, perhaps?

Hum Bug said...

Sabine, I like superdeterminism too and would love to hear a clear explanation about how you would test it. Very interesting!

Superdeterminism seems the only clear option to me without real issues within quantum theory.
That said, I like collapse models too since they're more phenomelogical and testable, and the general framework doesn't really need a interpretation. Though it has it's own problems. It does make it easy to keep track on which part of the microscopic-macroscopic we need to investigate and is a useful tool for experimentalist to point to for getting funding to test these realms.

piein skee said...

Quantum 'weirdness' is important because it's how we get to where we are now. It was the weirdness that drove the major debates and differences and schisms even, in the later years of the founding generation (of QM).

It was the weirdness plus the product of those debates that caused the...intellectual panic that in turn drove the emergence of the 'interpretations'. Note: The Copenhagen interpretation is unique in that it is retrospective: it is the way QM was dealt with prior to the interpretations. For that reason it seems self-contradictory and so on, but that's because it wasn't synthetically produced to be a self-consistent explanation, but instead simply as a reflection of the reality people were facing.

The interpretations were not scientific because they attempted unique new theory without telling us something unique and new that could be independently checked out. The impact of that was to the open up new less intellectually demanding pathways into the physics mainstream. The result of that was the rise of the 'infinity dependent theory'.

The culminative effect of all that gave rise to the inception and survivability of String Theory. Which morphed into a theory of quantum gravity. Which drove the dissident work into alternative approaches to quantum gravity. The impact of which was to re-wire the conceptualization of physics effectively into a hunt for quantum gravity. I'm obviously talking about frontier fundamental physics.

Everything stems from the apparently shocking weirdness of QM. Even if as people are increasingly saying now, there is no weirdness. Even then, we're stuck with the history which defines the dominant self-concept of physicists.

Unknown said...

Quantum mechanics is exactly what you would expect if you make observations, find that they are numbers, and simple symmetires exist (which of course they do):

k-froe said...

Bee, with some trepidation I believe that "Whatever Weinberg’s motivation, he doesn’t like neither Copenhagen, nor Many Worlds, nor decoherent or consistent histories, ..."

is better read as "... he likes neither Copenhagen, nor Many Worlds, nor decoherent or consistent histories, ..."

"Doesn't" linked with "neither" can induce head scratching.


Kris Krogh said...

"It’s become a cliché that physicists in their late years develop an obsession with quantum mechanics. On this account, you can file Weinberg together with Mermin and Penrose and Smolin. I’m not sure why that is. Maybe it’s something which has bothered them all along, they just never saw it as important enough."

I saw a 1997 KITP public lecture by Murray Gell-Mann where he described how he came to be dissatisfied with mainstream quantum mechanics. A friend had bent his ear about its shortcomings many years earlier. When he finally decided to devote some precious time to the issues his friend had raised, he discovered he couldn't give a good answer. I'm not a fan of the "coarse-graining" approach he subsequently developed with your former colleague, Dr. Hartle, but have to give him credit for working on an alternative when it was still politically incorrect to do that.

I think Dirac was one of those who always had reservations about mainstream quantum mechanics. After staying quiet for a long time, he confessed in his later years to favoring Einstein's position in the Bohr-Einstein debate. (You could say he was a closet realist.) His later semi-classical approach to quantum mechanics isn't well known, but I think it has some nice things about it.

Pepe Luisito said...

Superdeterminism is so elegant, I love it too...huge philosophical problem: it rules out Free Will, and thus,, seriously.

So we need something more...let me go wild...maybe Many Superdeterministic Worlds?

So consciousness can at least leap between universes in each true counscious decision, with its emergent superpowers of plane-jaunting that make decisions have some meaning, instead of being a joke...because deep down nobody really can belive that their moves while playing Chess with a friend are scripted and causally inevitable since the Big Bang...

Peter Morgan said...

I've not followed Backreaction for quite a few years, but your Nature Physics article that came to me online today,, persuades me to return. I hear no cheers. I think I left when the girls were about two, so it came as something of a shock to see them so big.

Can I put words in your mouth about this post? Of course I can. You seem at least a little in accord with my view of QM, that there are so many possibilities that there's not much point right now in committing oneself to one particular possibility. Superdeterminism? Sure. Faster-than-light? Sure, if one's willing to take Lorentzian symmetries as an effective symmetry, only valid at above 10^{-20} meters, say. GRW type collapse or de Broglie-Bohm type theories? Sure, though the issues are not as well understood as they might be. How do I or Buridan's ass choose between these? The other possibility, that there are (the same) statistical regularities without a classical physics-y type of explanation, doesn't seem to kill me any quicker, so I'm OK with that too.

But you seem to have a particular liking for superdeterminism, despite seeing Penrose take a lot of criticism over the years for liking it. The argument I don't see in the literature that somewhat naturalizes superdeterminism is that only stochastic superdeterminism is required, because we only model statistics of experimental results, and a QFT model that includes Wigner's friends Alice and Bob is already stochastically superdeterministic. So, in that sense, not a lot new to see here.