(finally!)Falsifying String Theory by Jacques Distler

marcus

a theory is not science---it is predictively empty---unless there is possible some future observation which it could not accept

a specific measurement which, if it turned out against it, would refute the theory

It looks like Distler may have come through.

http://arxiv.org/abs/hep-ph/0604255
Falsifying String Theory Through WW Scattering
Jacques Distler, Benjamin Grinstein, Ira Z. Rothstein
4 pages, 2 figures

"We show that the coefficients of operators in the electroweak chiral Lagrangian can be bounded if the underlying theory obeys the usual assumptions of Lorentz invariance, analyticity and unitarity for all scales. Violations of these bounds can be explained by either the existence of new physics below the naive cut-off of the the effective theory, or by the breakdown of one of these assumptions in the short distance theory. If no light resonances are found, then a measured violation of the bound would falsify string theory."

On the surface of it, this seems to offer a TEST!

I hope this is the real thing, and would welcome it warmly in that case.

The absence of such a paper is the main thing that Peter Woit has been complaining about. And it is presumably what Feynman was objecting to when he said "String theorists don't make predictions, they make excuses."

Now this unsatisfactory situation may be changing. Let's hope.

francesca

For light Higgs they mean the usual prediction about 115 Gev, isn't it?

so, viceversa, if they find Higgs would we worry about Lorentz invariace?

marcus

Maybe you have found a "catch" in Distler's article. I hope someone else can answer this more definitely. To me it now seems a little fishy.
Instead of falsifying string theory, it looks like it might merely cause the theorists to switch to a DSR version of relativity---which several of the non-string quantum gravitists have already done!
The falsification would then be of string-with-strict-Lorentz
but not of (if it can be formulated) string-with-deformed-Lorentz.

Maybe that is still all right? In a sense, any kind of falsifiability is progress. Any hard prediction which, if denied by Nature, must force a major reorganization of the theory

marcus

Peter Woit said (off blog) something that makes me think he might agree with francesca, interestingly enough. It may be that, at bottom, Distler paper does not really have to do with a falsifiable test of string theory.

In any case maybe one should be cautious about drawing conclusions about what the Distler paper accomplishes or does not. Woit will probably write about the paper in a day or two, and then people will discuss at Not Even Wrong blog.

selfAdjoint

I think this passage that francesca quoted:
Boils down to four boolean propositions.
H: There is no light Higgs boson (how light they mean they don't say).
N: Physics is Analytic, Unitary and Lorentzian (N for "nice").
B: The bounds they calculate are violated.
F: String theory is falsified.

And the paragraph asserts that
H => (N AND B => F)
H and B, and to a lesser degree N are empirical questions that could be verified at LHC. If you find both H and B then the only way to deny F would be to falsify N. The paper's citation on this point is to a discussion of tests of Lorentz invariance, but you could show Analyticity or Unitarity is violated instead. If H is false, that is if there turns out to be a light Higgs, then the whole argument is vitiated.

This is certainly a valid proposal for a test of falsification: whether the falsifaction stands or falls is put up to empirically testible propositions.

marcus

Thanks! that lays the issues out clearly and squarely.

marcus

Peter Woit had some what I think is rather perceptive comment on Distler's paper today:

Sample quote:

"...It turns out that the author’s proposal isn’t actually a proposal to falsify string theory at all, but a proposal to falsify the idea that physics satisfies Lorentz invariance, analyticity and unitarity at high energies. This would falsify our standard ideas about QFT, but I don’t think it would actually falsify current ideas about string theory...."

http://www.math.columbia.edu/~woit/wordpress/?p=385

Any reactions?

Chronos

I had the same impression, marcus. Distler's proposal would not be a stake in the heart of string, but would rule out certain models. I seriously doubt it is even possible to falsify the fundamental premises of ST. It is a slippery fish.

selfAdjoint

Based on my reading in my post #5, this is false. Distler, Grinstein and Rothstein mention the possibility of falsifying Lorentz invariance just in the spirit of examining all the options. But the TITLE of the paper refers to falsifying string theory (specifically the string theory amplitude matrix).

I usually groove on Woit's posts but I found this one ill tempered and unsound. His other criticisms of the paper, that it doesn't cover the whole case for SST falsification, is also off base; this paper has citation links to previous papers by others making the case in other ways; this paper was just a neat way to reach the same conclusion.

marcus

If Woit has over-reached in his initial critique of the paper then this has a potential for lively controversy! I wonder if he has received any reaction on the blog by now. Will check presently.

notevenwrong

I know what the title of the paper says, my claim is that it is inaccurate and way overhyped, and I explain why in the posting. I've had a discussion involving e-mail from Distler and one of his co-authors. Distler just engaged in his usual melange of arrogance, sneering, refusal to address issues, and personal attacks on my competence (which has something to do with my perhaps ill-tempered tone in discussing this, and anything involving him in general), but his collaborators seem to be perfectly reasonable people who behave professionally. Based on these discussions, I see no reason to change my claim about this paper. What they are hoping to do is to get some bounds on electroweak scattering amplitudes, purely based on 3 assumptions: unitarity, analyticity and Lorentz invariance, something which would be interesting.

The assumptions they are making are standard assumptions about QFT, so if their bounds were violated it would be indication of a failure of our usual ideas about QFT. String theory is just completely irrelevant to what they are doing, note that the papers they are quoting to back up their calculations are not string theory papers. I suspect that if their bounds were actually experimentally falsified, what most string theorists would say is not that string theory has been falsified, but that QFT has been falsified, and that this implies one has to do string theory.

If their argument works, and you find violations of the bounds they promise to work out, the consequence would be that in one's high energy theory one has to give up Lorentz invariance, analyticity or unitarity. Unitarity is pretty fundamental to the whole idea of quantum mechanics, so people are unlikely to give that up. While perturbative string theory amplitudes in flat space have the assumed Lorentz invariance and analyticity properties, it is not at all clear that this is true of non-perturbative string theory, or of string theory in certain kinds of backgrounds. Some string theorists have speculated about Lorentz non-invariant backgrounds, and, as I mentioned in my posting, if non-perturbative string theory really does, as often claimed, involve getting rid of standard space-time, you quite possibly would also lose analyticity and/or Lorentz invariance.

josh1

Hmm, very, very interesting.

marcus

this is a clear and probably a fair outline of the reasoning in the paper. It seems to me to be an accurate straightforward paraphrase---although one has to leave room for Distler to suddenly jump up and say No that is not our argument! Maybe he would object to something (but I can't see what.)

In this case one has to point out that prop N can not be empirically verified. It is not something that one can OBSERVE.
Therefore the paper's argument does not provide a way to falsify string theory based on empirical observation.

Prop N is intriguing and appealing. My guess is that it is wrong but the main thing is that it is not something one can show is true by experiment. Physics is a human construct consisting of theories and many of these theories are at present formulated using mathematical models that are Lor. Invar. and Unitary and Analytic (in the sense of analytic functions of a complex variable). But not ALL theories are.

Also as a side comment keep in mind that parts of modern mathematics use functions which are NONanalytic, and operators which are NONunitary, and investigate symmetries which are modified Lorentz and depart from Lorentz invariance in some fashion.

It would be impossible to prove that Physics at the present moment is "Nice"----that is, that all the competing human theories use mathematics which is unitary, analytic, Lor invar. I think in fact that it is demonstrably false. I know of physical theories that have not been disproved which involve departures from exact unitarity (because of using a realistic quantum clock) and theories which use other classes of functions besides the analytic, and of theories which use a modified Lorentz invariance.

To prove that Physics is "Nice" one would have to falsify these theories, which has not been done. Also one would have to prove that no similar theories could ever be constructed in the future.

It seems that Prop N is not something that one can OBSERVE to be true, nor ought one decide arbitrarily to BELIEVE that it is true for subjective reasons such as that one simply feels attracted to it and likes it.

And therefore I conclude that the logical scheme of the paper, which selfAdjoint kindly outlines for us here, has no chance of proving F (i.e. falsifying string)
H => (N AND B => F)
because there is no way to verify N.

There could well be more controversy in store, so I plan to keep watching and maybe the picture will change.

Chronos

Agreed, marcus, and I agree with Peter. I have seen these smoke and mirror tactics too many times to be swayed by them. Conveniently ignoring observational evidence is . . . convenient.

marcus

Chronos thanks so much for your agreement and support! However saying smoke and mirrors goes well beyond what I can do at the moment. I suspect Distler is dead serious and that he intends (after LHC has found no light Higgs) to hold the gun to the "nice" baby. He may be deluding himself but I suspect that he thinks he has a serious "proof" here of string validity. He can say "BELIEVE IN STRING
OR ELSE ACCEPT THAT PHYSICKS IZ NOT NICE!!!!
And if we do not all bow our heads and accept string then he will shoot the baby and declare that Physicks must be not Analytick, or not Yoonitaree, or not Lorentziak.

I think that a statement like "Physics" is analytic Lorentz invariant unitary is a METAPHYSICAL type statement, not an empirical or observational one. The reality is that some of the theories which have so far not been experimentally falsified ARE and some are NOT. But about "Physics" we cannot observe if she is these things. Nevertheless I suspect that Distler imagines that he has a real argument involving empirical falsifiability, as his title indicates, and that he is SINCERE.

But so far this is just my hunch. I want to wait and see. He could, of course, NOT be sincere (as you suggest) and be doing smoke and mirrors. Or he could turn out to be right! Anything could happen so for now I think just watch and see how it plays.

selfAdjoint

Well Marcus, the last few comments on Woit's discuss N. Of the three components of N. Lorentz Invariance seems to be the shakiest, and also the one easiest to test experimentally, I can't imagine why you think it is "unobservable".

Much the same goes for Unitarity any violation of this would require the abandonment of the Born rule, which considering the trouble it has caused (hi Patrick!) would be a benefit in my view.

Analyticity doesn't mean the functions of physics are eveywhere holomorphic; the only everywhere holomorphic functions are constants! Singularities, such as the poles of meromorphic functions, their locations and their residues play a BIG role in physics. I believe a failure of analyticity would be a proof that the wave function was nowhere differentiable, or at least so throughout some open set in spacetime.

marcus

I'm hip to the distinction between holomorphic and meromorphic---I tend to use the term "analytic" loosely to cover both cases---and also real analytic meaning that you can expand it in a powerseries most places.

I like the general tenor of your comment. And I am glad that people are discussing the extent to which one ought to expect physical theories to use series-expandable functions, and have normpreserving ("unitary") hibertspace operators do their time-evolution and so forth.

More power to them. Let them all discuss these issues. I am making a different point.

Physical theories can not be ultimately verified, only falsified and replaced by better. "Physics" I take to be the collection of theories which have not yet been refuted and discarded.
The statement "Physics is nice (unitar. anal. Lor)" is IMPOSSIBLE TO VERIFY.

IN ORDER TO USE IT IN THE distler SYLLOGISM you provided us, so as to logically falsify string (your statement F) you have to VERIFY niceness. It is not sufficient to be able to falsify statement N, "niceness".

I am very happy if people want to prove to us that physics is not necessarily Lorentz invariant. In fact that seems obvious to me because there are excellent physics theories like Freidel's spinfoam unification which are clearly NOT---they use a modified Lorentz invariance. And so on. Gambini and Pullin have an interesting way of proving physics can not have a unitary time-evolution. Gambini Pullin has not yet been falsified and neither has Freidel. So they are part of physics and they are not nice.

But simply proving that physics is not nice does not make the syllogism work. To get to your F you would have to VERIFY your N and your N can not be verified observationally. Even if it weren't obviously wrong (to a lot of us) it couldn't even in principle be verified observationally. It is a normative statement about the physical models that humans can construct and therefore metaphysics.

I do believe it is highly QUESTIONABLE metaphysics. But if someone stubbornly wants to believe that all human models of physical nature OUGHT to be unitar. anal. and Lor----then he will just go ahead and believe that. I can try persuasion and show examples of theories that are not and which achieve interesting results. But I don't see how I can design an experiment to demonstrate to him that physical theories do not OUGHT to be those things.

Am I missing something? It seems to me that to demonstrate to a steadfast believer I would have to do something like show that Freidel spinfoam is CORRECT. And that is something you cannot do with physical theories, you can only falsify them. If they pass this year then the next year's experiment can still show them wrong.