Hacker News new | past | comments | ask | show | jobs | submit login
Lost Causes of Theoretical Physics (2003) (archive.org)
98 points by scottlocklin on March 2, 2018 | hide | past | favorite | 53 comments



Not a physicist, but isn't "Bohmian Mechanics" (1) what's also known as "Pilot Wave Theory", which is now gaining experimental support (2) ?

Clarke's First Law: "When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong."

1. https://en.wikipedia.org/wiki/De_Broglie–Bohm_theory

2. https://www.quantamagazine.org/pilot-wave-theory-gains-exper...


Pilot waves, multiverse and Copenhagen are fundamentally ineligible for experimental support, at least with respect to each other. They lead to the same numbers, meaning than an experiment that agrees with one will also agree with the others. It's something like trying to tell the difference between 2x=2y and x+2=y+2.

The "support" in the article is really just a physical system that obeys (as a macroscopic approximation) something like pilot waves: "just like the universe" if pilot waves are right, and "just like those silly nonlocalists" if pilot waves aren't.


Right. It doesn't provide evidence for pilot wave theory any more than rolling some balls around on a rubber sheet provides support for Glgeneral relativity.


While the current formulation details of the pilot wave theory doesn't allow for new predictions, a better formulated one will. I strongly believe we'll see the breakthrough in the coming 10-15 years. Quantum mechanics are statistic predictions and it's about time we treat them as such.


> While the current formulation details of the pilot wave theory doesn't allow for new predictions

Actually pilot wave theories do have new predictions, namely quantum non-equilibrium. We just have no idea if we can actually design an experiment for it.


Why do you strongly believe that?


Yes it is. The author presents his dogma as the one and only truth, which you're not allowed to deny if you want to be a "scientist". It's similar to labelling people as climate change "deniers" and "anti-vaxxers" etc etc.


That's ridiculous. It's on you to provide evidence for your position. Absent that, it's free for other scientists to criticise you on that ground.


[flagged]


What a strange comment, given that he's obviously right. You're not a scientist, are you?


> Rigorous Feynman Path Integrals

It depends on how rigorous you want the treatment to be.

The whole machinery of Feynman diagrams is derived from Path Integrals and Wick's Theorem, so they are rigorous, at least as rigorous as standard perturbation theory using operatorial quantum mechanics.

A mathematical physicist won't be amused, though, that's correct. But I wouldn't say that Path Integrals are less rigorous than the average technique used in Theoretical Physics, and particularly in many-body theory.


There's been a lot of progress nonperturbatively defining quantum mechanics and quantum field theory using resurgence lately. See, for example, the very recent https://arxiv.org/abs/1802.10441


I think it's incorrect to say that Feynman diagrams are as rigorous as standard perturbation theory. Perturbation theory in QM has a completely rigorous description. It's controlled approximations to known mathematical objects. By contrast, the derivation of perturbation theory in quantum field theory involves applying reasonable well-understood transformations to mathematical objects which are largely _unknown_. Nothing protects you from garbage-in, garbage out. (Perturbation theory turns out perfectly sensible looking answers for theories like 6 dimensional scalar field theory, which doesn't actually exist.)

That said, Feynman integrals are something mathematical physicists have a reasonable level of comfort with these days. There's a well-developed theory of Euclidean path integrals which works just the way Feynman integrals are supposed to. The rigorously constructed examples are lower dimensional field theories, like scalar field theory in 3d, but they don't really differ in kind from the 4d theories used in particle physics. It's just that the analysis gets significantly harder in 4d.


The many worlds interpretation is just that, and INTERPRETATION, not a predictive theory. The article seems really upset by that, but is it any reason to label it a lost cause? I guess you can’t really study it or preform experiments on it but that shouldn’t be any suprise for any interpretation of quantum mechanics.

I might be missing something here...


His argument is that it's useless as a research topic because you cannot perform experiments about it.

You can still believe it if you like, but its slightly less useful than trying to convince everyone to replace pi with tau..


QM is a mathematical model. Many worlds is an intuition to understand that model. There is a pedagogical question on if many worlds is a useful mental model to have. There is also a philosophical question of if many worlds better captures the essence of the model relative to Copenhagen.

We can also flip it around. Why should we think of QM in terms of the Copenhagen when it makes to testable predictions.


Wouldn't you consider the Born rule to be a testable prediction (though in an admittedly weak sense of simply being postulated)?

If the Born rule were found to be violated, it would invalidate Copenhagen but not MWI, which IMO means Copenhagen is more falsifiable.


> If the Born rule were found to be violated, it would invalidate Copenhagen but not MWI, which IMO means Copenhagen is more falsifiable.

All quantum mechanical interpretations depend upon the Born rule. A violation of the Born rule would rule them all out, except perhaps Bohmian mechanics which allows the possibility of non-equilibrium states, but we have no idea if those can be created.


I mean, the pi/tau thing is _purely_ aesthetic. That's a big waste of time for sure.

Having a more coherent metaphor for the evolution of quantum systems might help with physical insight, though. If e.g. the universe isn't differentiable across spacetime, we need a better understanding of GR. Many worlds insists it is differentiable (and local), so thinking that way could help integrate the two.

It isn't an important thing, but pi/tau is an unrealistically silly thing to compare it to.


If a theory isn't predictive, it isn't falsifiable. If it isn't falsifiable, then the theory isn't truly scientific.

This is a major issue/concern in advanced theoretical physics. Is it really physics if you can't experiment or is it just mathematics?

The foundation of science is falsiability within the scientific method. It's the core tenet.


But you need to do a lot of speculative work before falsifiability becomes possible. And that speculative work is going to involve groping around in the dark being guided by mathematical logic (and a certain amount of politics, fashion, and wishful thinking.)

Considering physics as if MWI could be true has the potential to lead to mathematical elaboration and eventual falsifiability.

Considering physics as if there's no need to interpret QM because there's nothing to interpret (because Copenhagen is the correct interpretation) is a statement of faith, not rigour - because as a position, it's no more falsifiable than MWI, or Bohm, or any of the other interpretations.

The point of mathematical physics is to generate new falsifiable predictions. It's still science, it's just embryonic proto-knowledge - which is why it's called research, not teaching.


But what if, after spending more time working on and thinking about it, someone does find a way to make predictions based on it or on an extension/modification of it?

If that's a possibility (I can't say, I'm not a physicist) then I'd be a mistake to discard it on the basis that it's just not predictive now.


> If a theory isn't predictive, it isn't falsifiable. If it isn't falsifiable, then the theory isn't truly scientific.

That's an overly anemic view of science. Scientific theories have both predictive and explanatory power. Interpretations of QM don't change the former, but they do change the latter.

Copenhagen, favoured by this article's author for some unfathomable reason, has little explanatory power. Many worlds and Bohmian mechanics have excellent explanatory power, and their predictive power equals Copenhagen.


For sure, I’m by trying to say it’s a real theory, it’s just a way to interpret the quantum mechanics that we can test in a way that provide us with some sort of intuition when working with the physics.


After skimming the whole thing, I was relieved to see that the argument against Many Worlds was the only one I considered invalid. It took the form of "It's not even a testable theory", rather than "it has mathematical deficiencies".

Well, of course, no one's saying it's testable necessarily, but for various aesthetic and intuitive reasons it still seems valid. Just don't expect predictions from it and you'll be fine.


> After skimming the whole thing, I was relieved to see that the argument against Many Worlds was the only one I considered invalid.

You should consider the Bohmian criticisms invalid as well then. Many have argued that de Broglie-Bohm is "many worlds in disguise", so if the criticisms the author levels against Bohmian mechanics are true, then they apply to MWI, but if they don't apply to MWI, then they don't apply to Bohmian mechanics either.

Not because they are the same theory, because they are not ontologically the same, but the formalisms are interchangeable.


Oh yeah. Probably, yeah. I glossed over that one because its criticisms were strangely technical for a theory that, iirc, just rephrases regular quantum theory.


Valid and true are two very different things - and we've seen from history that the human brain can be lead into believing that essentially anything is intuitive or aesthetic.


Sure, hence the reason physicists tend to say MWI is valid and good, rather than 'true'.


It does make different predictions though. WMI essentially says there is only unitary evolution and nothing like a physical collapse. That means everything unitary evolution predicts will happen, unlike other "interpretations".

This matters in at least two cases:

(1) In a finite system (perhaps our universe, but doubtful), those other worlds will eventually come back into contact. They never go away and with only a finite number of states available, the splitting of worlds can't continue forever.

(2) From a subjective perspective, there is always a future version of yourself in some "branch". Since you will become all future versions, you are subjectively immortal.


> The many worlds interpretation is just that, and INTERPRETATION, not a predictive theory.

How does the many worlds interpretation INTERPRET the Born rule? Shouldn't that be a key feature of any QM interpretation?


From other sections it seems clear that the author is a devout adherent of The Copenhagen Interpretation, aka objective wavefunction collapse; this is a close relative of the baseline non-Interpretation “Shut Up And Calculate.” People like that are generally utterly unmoved by anything short of a rigorous mathematical formulation, which works for them in QM where “rigorous mathematical formulation” is pretty much all of what matters.

While I respect the adherence to formalism, the notion that our imperfect view of QM (it doesn’t mesh with GTR after all) represents an ontology to hang your hat on is less respectable. Wavefunction collapse seems like a massive cop-out, and it won’t satisfy anyone except the most rigid logical positivist.

Something like MWI is on the other end of the philosophical spectrum, so the hostility is not surprising. At the end of the day though, it comes down to this:

If you want to build something based on QM, you need someone like this guy.

If you want to satisfy your human need for an ontology, he won’t help you.


"Objective wavefunction collapse" is not a good "aka" for the CI. There is a great deal of objectivity in it but there's an essential element of subjectivity (shared by neo-CI descendants of the CI such as QBism and Rovelli's relational interpretation) too. Streater's discussion of EPR explains why: https://web.archive.org/web/20151117174141/http://www.mth.kc...

Broadly speaking, the CI and its "radical" psi-epistemic* relatives interpret QM as (generalised) probability theory applied to mechanics: https://telescoper.wordpress.com/2017/06/23/the-quantum-mech... It's hardly a cop-out. See e.g. Rovelli's take on what the ontological ramifications are: https://arxiv.org/abs/1712.02894

* https://www.sciencedirect.com/science/article/pii/S135521981...


I would argue that expecting QM to satisfy a human need for ontology is about as useful as expecting general relativity to satisfy a human need for sexual gratification.

It's easy to find a way to be dissatisfied: just want more out of things. But it doesn't make sense to always expect to get what you want. Some things simply won't provide. As long as QM is providing a means to achieve accurate predictions and precise language, it is doing its job perfectly. It is not responsible for satisfying a (potentially irrational) need for ontological interpretation.


> As long as QM is providing a means to achieve accurate predictions and precise language, it is doing its job perfectly. It is not responsible for satisfying a (potentially irrational) need for ontological interpretation.

What a strange position. Clearly some things exist, since we're here, we're discussing the subject itself, and so on. Defining what "to exist" actually means would seem to be the central rational pursuit of physics. This is an ontology.

The focus on prediction is merely a proxy to the ultimate goal.


And if QM and GTR agreed in their predictions I’d be right there with you, telling people to suck it up. I’m a bit there with you anyway, because nature has no obligation to us, and the predictive power of QM is just incredible.

But... in the absence of a complete theory including gravity, there are huge gaps. It’s just human nature to try and fill those gaps, for better or worse. For worse when people stuff those gaps with god, but for better when they’re uncomfortable with overly pat explanations like objective collapse.


> nature has no obligation to us

This is an excellent point indeed, but what if it also applies to the disagreement between QM and GR? That is, does Nature have an obligation that QM and GR, both being mere reflections, in the human mind, of its particular aspects, should ever become in agreement with each other? What if mathematics (the human logic) can in principle only provide models that are local approximations of reality?


It’s possible, but maybe not likely. I’ll tell you though, if reality is that unpredictable, thst unfriendly to rules, expect a rash of suicides in the physics community the day it’s confirmed.


These discussions are always composed of people who don't know very much laying down philosophical assertions (that they very rarely realize are philosophical), and innocent bystanders posing reasonable what-if questions that really can't be answered with (present) empirical knowledge. Like politics, the content of what is said is dominated by the dynamics of the discussion, with English skills "winning" arguments more often than anything else.

Physics is already approximate; experiments don't have access to infinity significant figures. If you'd prefer a more complicated example then ask anyone who has ever calculated a QED problem to less than infinity terms.


These discussions are always composed of people who don't know very much laying down philosophical assertions (that they very rarely realize are philosophical)

That seems needlessly judgmental.

...and innocent bystanders posing reasonable what-if questions...

They’re rarely reasonable. You get a lot of questions which are philosophical in nature, but which people believe are scientific. “What is time like for a photon?” “What was there before the Big Bang?” “What is spacetime expanding into?” “What’s inside of the event horizon of a black hole?” “Why are the physical constants what they are?”

When you give the scientific answer, “Bad question, like asking what’s North of The North Pole?” People are unhappy. Plus you’re a dick, because you can figure out what they’re asking and try to give a complete and accessible answer, complete with warnings as to the philosophical/metaphysical nature of both question and answer.

...that really can't be answered with (present) empirical knowledge.

Often the questions themselves are essentially unanswerable in principle. It’s not easy to ask good, relevant, scientific questions about QM or STR/GTR. The lack of knowledge is more of a factor when the questions are insightful.

Like politics, the content of what is said is dominated by the dynamics of the discussion, with English skills "winning" arguments more often than anything else.

Depends on the audience... try that in a rigorous environment and be crucified, but on HN the rigor is absent. It’s a good thing too, because it’s Hacker News, not PhysicsForums.

Physics is already approximate; experiments don't have access to infinity significant figures. If you'd prefer a more complicated example then ask anyone who has ever calculated a QED problem to less than infinity terms.

And? You seem to be complaining that on a site full of non-physicists, people don’t act like physicists. Besides, ontologies of QM are philosophical, not scientific, and hardly limited to laypeople.


">That seems needlessly judgmental.

There's a practical takeaway from my statement - that physics explanations should always be taken with lots of salt (even ones written by physicists, who are prone to throwing out the star baby with the mathwater.) Judgmental, maybe, but not needlessly.

>They’re rarely reasonable.

They're reasonable because they have an answer: it's the one you suggested. People would accept your "north of the north pole" response in a universe where there was no false but well-written alternative. The problem is, outside of very limited audiences and forums mixtures of misunderstood philosophy and science are a lot easier to come by than the truth. (This extends to science journalism, and even to the rare bad press release.)

By the way, if you want to tell people that they're asking the wrong question in a way that keeps them happy, try responding with the knowledge that would make them realize their question was misguided. For example, whenever anything involving moving at the speed of light comes up, my go-to responses are:

- Discussing hyperbolic geometry until it's clear that what looked like a question about pi/4 was actually a question about infinity.

- Mention that most of the equations have a singularity there, and then give some more down-to-earth examples of equations that have singularities.

- Say that we've never seen it happen, and as a result the physics we write down doesn't include it. (This is sometimes the best answer, even though it's the least satisfying to give or hear. If science fiction effect X happened, we'd have to account for it, no matter how badly it would break everything.)

>And?

Physics is all about approximations with known error bounds. (That is, X plus or minus Y, and the equivalent of that but with equations instead of numbers.) That's what's going on at the boundary between physics and metaphysics, and it's the ultimate answer to most "unanswerable" questions.


I am not a physicist, but this seems backwards to me. If anything, it is the Copenhagen interpretation that comes with the largest amount of extra-mathematical baggage while many-worlds interpretation came out of effort to add the least amount of additional assumptions on top of the math.


The price that MWI pays for this minimalism is that it loses some contact with reality, i.e. physical experiments. Basic physical observations like "some measurements are non-deterministic" are quite difficult to explain in the MWI. Reality is inconvenient.


I don't think that's true - it takes that into account, it just looks like it collapses because you're in the world with that state - not seeing the entire picture. MWI accounts for it appearing that way while in actuality the wave function never collapses.

Max Tegmark gives a pretty accessible description of this in his book - Our Mathematical Universe, and Eliezer Yudkowsky does a decent job with a general explanation here (though he avoids talking about MWI): http://lesswrong.com/lw/pd/configurations_and_amplitude/

I think the Copenhagen interpretation is now considered wrong by most physicists. It's strange to me that it was ever really thought of as true - though it makes sense that it would look like what's happening at first.

It's less that MWI is a predictive theory and more that MWI is something that comes out as a result of the theory's prediction (Schrodinger's wave equation that doesn't collapse).


> I think the Copenhagen interpretation is now considered wrong by most physicists.

Recent surveys suggest that most practicing physicists still pick Copenhagen as their default, mainly because that's what they were taught and QM foundations have no implication on their actual work.


There has been an enormous resurrection of the analytic bootstrap. Grep, for example, https://www.icts.res.in/program/NUMSTRINGS2018/talks for bootstrap. The difference is the realization that the bootstrap equations can be solved numerically and can be thought of as a way to test whether certain conformal theories exist at all, rather than as a way to take you from hardly-any-input-at-all to a complete description of the theory.


> Aspect, Dalibrand and Roger (1982)

Aspect, Dalibard and Roger (1982)


Grad school physics was/is your background I assume. I had the same thought when reading.


Yep, I was a graduate student in physics, but not at all in that field. The reason why I recognize Dalibard's name even misspelled is because I was a student in the same institution and I used to know several of his grad students.

Little bit of trivia: he was 23 or 24 when he wrote that paper…


Can anyone explain Feynman path intergrals to a non-physicist? I'm guessing there's a bit more to it than defining a measure on C[0, T] and lebesgue integrating.


I'm not a physicist but as far as I know the key concept of the Everett interpretation is that you can think of universe states as of vectors and perform mathematical operations on them. IMHO giving up the ability to think this way would be a tragic loss. Many other parts of the list would make a huge pity to forget too.


Does it say Mach Principle is incorrect? That is good, as a non physicist I could never wrap my head around the distant stars influencing anything.


The distant stars are important because they make space flat. Remember that general relativity says mass bends space, or more accurately is the same thing as space.

So if all the mass was nearby, space would bend inward around the current point and it would fold in on itself - the universe would be like the inside of a sphere.

Because the mass is distributed evenly at the largest scales, it instead continues flatly in all directions.


> the universe would be like the inside of a sphere.

The surface of which is "flat"? The interior surface, if that's the right word. Like living on the Event Horizon of a black hole. So we don't and hardly can know what's outside or inside or on the outside or inside.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: