Why String Theory Is Not Science (forbes.com)

StartsWithABang writes: Earlier this month, a conference was held devoted to the question of whether untestable scientific ideas like string theory and the multiverse are actually science or not. While many opinions were stated and no one changed their mind, the answer is apparent: unless you're willing to change the definition of science to include "this thing that isn't science," then no, string theory is not science. It's a theory in the sense of a mathematical theory — like set theory, group theory or number theory — but it isn't yet a scientific theory. Of course, it could become science, but that would require that it actually do the things a scientific theory does: make testable predictions that can be validated or falsified.

Comments

Only if you Exclude Technological Limits

[Roger W Moore]

The condition for science is that it has to be testable in principle, NOT that it has to be testable within the limits of current technology. When Higgs came up with his theory there was no accelerator capable of testing it (although we did not know that at the time). So would that make the Higgs mechanism non-science until the 21st century when we built the LHC? Clearly not. So, unless String theory is completely untestable in principle, regardless of potential future technological advances, it is science albeit science which is currently impossible to test with current technology.

Re:Only if you Exclude Technological Limits

[hey!]

The classic example used by logical positivists in the early 20th century was any conjecture about the far side of the Moon. Until we developed spacecraft any statement about the half of the Moon we can't see would have been untestable in practical terms, but it would have been testable in principle.

Re:Only if you Exclude Technological Limits

[sjames]

But that's not what string theory does. Instead, it predicts everything. No, not everything we have observed, EVERYTHING. No negative or positive finding tests string theory, it just suggests more knobs to twiddle. The one shining hope is that if we don't find supersymmetry, it is dead as a theory since it cannot accommodate a universe without.

It could be forgiven all of that if it made things more tractable, but it doesn't.

Re:Only if you Exclude Technological Limits

[mbone]

The condition for science is that it has to be testable in principle, NOT that it has to be testable within the limits of current technology. When Higgs came up with his theory there was no accelerator capable of testing it (although we did not know that at the time). So would that make the Higgs mechanism non-science until the 21st century when we built the LHC? Clearly not. So, unless String theory is completely untestable in principle, regardless of potential future technological advances, it is science albeit science which is currently impossible to test with current technology.

String theory is arguably not science not because it makes predictions we cannot test, but because it basically makes no predictions at all. Originally, when people realized the importance of 10-dimensional manifolds (i.e., of theories with 6 compact dimensions), there was a lot of excitement as people thought (and confidently said) that there would be one and only one suitable such manifold, which would have led to concrete (if maybe hard to test) predictions. But, now, there is a huge number (order 10^500) of such manifolds known, each basically allowing for a separate theory, and we have no idea which could be the right one.

Also, there is the pesky fact that predictions have been made about the foundations of string theory (that, for example, the LHC would detect the supersymmetric partners of existing particles), and they have not been born out by experiment,

Having said that, my personal feeling is that string theory is science, but science that is unlikely to be fruitful. Eventually, unless this changes, something else will come along, and it will cease to be the center of attention for theoretical physics.

Re:Only if you Exclude Technological Limits

[Ramze]

The problem is that its equations are quite elegant and simple -- because its goal was to unify the fundamental forces... and by doing so, created so many variables, functions, and possibilities, you can describe just about any universe that might exist. Each particle exists in 11 dimensions with varying degrees of freedom -- and we have few hints at the shapes of those hidden dimensions. Pinning it down to our universe is hard -- really hard. Like 10 ^ 99 possibilities hard. And, there's still the possibility that it's wrong, but there are a very long list of possibilities to go through before they can figure that out. Even if they find one that matches perfectly, just because the math agrees doesn't mean it says anything about how the universe really works -- just that the math works.

But, say you want to describe fundamental forces -- easy... the equations for light, gravity, strong force, and weak force look identical in string theory except for a function tacked on the end -- same for particles and their properties. (Also, as the energy level reaches a certain point, the forces converge towards a point where they are all equal.... as if they are all aspects of the same force that split into different dimensions.)

The same sort of thing was done to predict the Higgs. Equations were written as if all particles were massless plus some function based on interaction with the Higgs. Without the Higgs, the equations were ugly and none of the equations for the particles looked alike, but factor in the Higgs, and they all fall beautifully into place as identical plus some Higgs function. So much was based on this math, that it was understood that it HAD to be right -- years before the particle from the field was discovered.

The irony is -- string theory is a bit like creating epicycles to make the Earth the center of the solar system in the sense that string theory was created to unify the forces -- and in doing so, necessitated the creation of multiple curled up dimensions. It made some things very easy by postulating something strange. Einstein described gravity as the curvature of spacetime, and this gave string theorists the initiative to do the same with the other forces. It may be right -- it's probably right... it looks right in how elegant the math is... but... there's no way to test it -- yet. The theory says more about what can be rather than what must be, so it'll be a while before it matures enough to be called true "science."

Re:Only if you Exclude Technological Limits

[DrJimbo]

What if their goal isn't to make a testable prediction that diverges from the current best theory, but merely to explain more elegantly what's already explained? Shouldn't that count as scientific progress too?

An excellent question! Yes, more elegant explanations of existing phenomena are definitely a big part of science. The unification of electricity and magnetism is an example . But that unification led to new predictions that the non-unified models did not. Yet, even if string theory was able to make the same predictions as the standard model and no new predictions then, hell yeah, it is would be science. The problem is that it makes no predictions. Well, to be more accurate, it makes way too many predictions which is pretty much the same thing.

You see, explaining what has already been explained involves making testable predictions. String theory does not do this which is why it is not science. That doesn't mean it is worthless to pursue.

Re:Only if you Exclude Technological Limits

[Ramze]

Problem is, supersymmetry (SUSY) is the only theory that even attempts to explain why the masses of particles are as small as they are -- including the Higgs particle. Without SUSY, the Higgs, W, and Z bosons become nearly infinitely massive due to loops in their feynmen diagrams. It's the SUSY that cancels out infinities in a lot of equations to make the string theory results make sense.

Many expect SUSY particles (sparticles) to start showing up at a multiple of the Higgs mass -- say... close to the Higgs mass or an order of magnitude higher, but not much higher. Sparticles are a good candidate for dark matter, but they're unlikely to be detected by the LHC.

Also, we know that string theory can give the same answers as other quantum theories for known values... so it's not "wrong" so much as it's a different way to do the same math... but ST requires SUSY because it's a necessary result of the math. If you compute a universe that allows both bosons and fermions, then for each boson, there is complementary fermion and vice versa -- only SUSY predicts that they have the same mass, but clearly they don't -- so the symmetry is broken.

If no sparticles are found at higher energy levels, then someone will have to explain what's wrong with particle physics in general -- because the math works.... so, what is it about the math that is correct that we're incorrectly interpreting as reality? Even if string theory is discounted in favor of another theory... you can wipe out the theory, but not the math. The math is just a different formulation of a problem to get the same answer. If I say 3 + 2 = 5, and the 2 is the sparticle in my theory... then when we find sparticles don't exist, what the heck is it that I'm adding to 3 to make it 5 ?!?!? There must be some unknown physics that string theory is describing as super particles that may actually be something else we don't understand. We don't even know why the symmetry is "broken" in the current theory to begin with.

Re: Climatology

[Anonymous Coward]

Didn't take long for the peanut gallery to weigh in.

Re:Wouldn't it be more properly referred to as

[nashv]

Exactly. It is a model. It might even be a useful model with some explanatory power. But the same can be said of many belief systems. The only difference is that the other belief systems have been shown to be inaccurate by showing their contradictions with reality. With string theory, we are not aware of any specific such contradiction yet.

I don't think we should have any problems with models, as long we understand very clearly that they are only models. Like Newton's laws - they are strictly inaccurate but as approximate models of reality that are valid under some limited set of conditions, they remain useful.

Re:Climatology

[phantomfive]

No. Climatology, even if you are the biggest skeptic, makes testable predictions all the time. For example, predictions that the earth will be two degrees warmer in 100 years. That is completely testable: it will take 100 years to test it, but that's irrelevant.

An untestable theory is one that can never be tested, even with infinite time and resources. For example, "the universe was created as-is five minutes ago." Maybe that's true, but there's no way to test it. Even if you had a time machine, it still couldn't be tested. There is no experiment that can be imagined to test this.

In the case of string theory, the author claims that string theory makes no predictions that distinguish it from the standard model. That is, if you perform an experiment, you will not know if it is supporting string theory, or if it's just a natural result of the standard model.

Re:This has been known for a long time

[vux984]

I think cargo cults are pretty science-ish.

The observed a correlation between airports and cargo planes arriving.

The formed a hypothesis that constructing something that looks like an airport and control tower would bring the cargo airplanes.

They tested the hypothesis, by building the airport etc. It didn't work.

They (correctly) knew that something made the cargo planes come; so they tried to improve their emulation of the airport operations etc.

Sure if was fundamentally wrong. But it WAS the scientific method in action. Observation, hypothesis, experiment...repeat.

Its no different than heliocentric astronomy. We kept trying more complicated and elaborate constructions to predict the planetary motions, but it just kept failing because it was wrong.

Re: Climatology

[Anonymous Coward]

Aye, there's the rub. First, only one planet. Second, difficulties with the "awhile" parameter. Make it too short and you're just testing long-range *weather* forecasts which is not where the controversy is. Make it too long and the theory changes so that the argument becomes "that's an old model, we know better now". Do they really know better, or are they just moving the goal posts?

I think climatology is in a grey area in this regard. In theory, it's testable and thus science. In practice, it's political, not well tested, and thus not living up to its potential as science.

Does explanation have a role in science?

[UpnAtom]

David Deutsch argues that it is core:
https://www.ted.com/talks/davi...

Also, string theory is surely as testable as quantum mechanics. It's just currently impossible to say which is more valid.

The original meaning of begging the question

[radarskiy]

They start with the premise that string theory is untestable, and come to the conclusion that it is untestable.

Re:Wouldn't it be more properly referred to as

[losfromla]

There are too many floaty numbers inserted and whenever the numbers don't work more floaty numbers and dimensions are added in to make the "theory" work. Books written about this "Not even Wrong", "The Trouble with Physics". There may be others, those are just two I am familiar with.

http://www.amazon.com/Not-Even...
http://www.amazon.com/Trouble-...

No difference = equivalent

[Roger W Moore]

The author claims that there is no test that can be done that would prove String Theory true as opposed to other theories.

Unfortunately the author has proven many times that he does not understand particle physics in previous posts. The problem with String Theory is that there are far too many possible theories to consider (last count I heard it was around 10^500) to make detailed, concrete predictions. The second that we get an experimental signature for something like String Theory that number would collapse and theorists would be able to start studying the detailed predictions of a vastly smaller number of models. This would undoubtedly lead to some clever theorist coming up with signatures unique to String Theory which other, competing models would not have.

If you can't come up with ANY difference it would mean that the theories must be mathematically equivalent for all situations which are possible. We have had this happen in physics before. Matrix mechanics and wave mechanics are both different ways of doing the same Quantum Mechanics. Nobody worries about which is the "right" way because both make mathematically equivalent predictions.

Re:Climatology

[phantomfive]

Yes. Astrology makes valid, testable scientific predictions. Therefore it is a scientific theory (or hypothesis, if you like it that way).
It is unfortunate that it has been proven false many, many times, but not every scientific theory has to be true.

Re: Climatology

[davester666]

Somewhere on Earth, it's cold. Therefore global warming can't be happening. If it was happening right now, it would be warm everywhere. QED

The concept is often unscientific

[Kjella]

For example, the theory that all matter is made up of small, indivisible bits (atomos) is unscientific. Whenever you find a new smallest building block (atom) there's a chance you'll find they're built by even smaller blocks (a core of protons and neutrons surrounded by electrons) and that protons again are made from even smaller particles (quarks). And maybe the quarks are built from superstrings. And maybe the superstrings are build from something we don't even have a name for yet. That doesn't make them bad ideas to guide scientific research and design experiments. Just like causality is a rabbit hole with no end, even if we could explain the whole formation of the universe back to the Big Bang we'd always be looking for what caused the Big Bang. And what caused that which caused the Big Bang. Scientific exploration is an educated guesswork, you take some observations and try to find a system or pattern or formula and if the results don't contradict reality, great. It's obviously even better if you can predict something new, but if I find that E = mc^2 and show a few reproducible examples it's up to the rest of the scientific community to find a contradiction where E != mc^2. I feel it's a bit like that with superstring theory, if we got multiple theories that both come to the same results then either they're different formulations of the same model or there will be distinct differences that are at least hypothetically testable.

Re: Climatology

[lgw]

You're joking, but that is actually what a "Warm Earth" (i.e., not an ice age) looks like: no year-round ice anywhere. It doesn't get much warmer at the equator, but it gets a lot warmer at the poles.

Best way to convert a "non-science" stickler

[doug141]

If you find yourself talking to a person who dismisses talk of multiverses or string theory as no better than talk of the supernatural, just ask them what a person would see while falling into a black hole. They will proceed to tell you their version. Then ask if someone outside the hole can ever verify anything they just said. They will say no, communication won't work from inside an event horizon to the outside. Then ask, if everything they said is all based on conjecture and extrapolating known laws, and can't be experimentally verified, why do they feel it merits discussion?

Re:Evolution

[Oligonicella]

You do not understand what evolution is. "Create life" is not in the description.

Re:This is of course true

[Oligonicella]

Straw argument. You can study evolution with no references to geology or paleontology at all by just studying soft tissue in the lab.

You're correct though, that's historic evidence of evolution. Those phenotypes were modified by changes in the gene sequences like those we've seen happen in labs. And some of those gene sequences we've extracted, mapped out and compared to living organisms. So evolution is a proven.

Re:The sad state of climatology

[Fragnet]

I have studied what the scientists say about this for ten years. But I've also made a point of studying what the sceptics say, looking at the model results compared to actual reality and squaring the science with the hyperbolic press and political statements. Let me tell you, there's a huge discrepancy here. It's almost as if the science is (on the whole) being manufactured to order. This isn't new. It happens in social "science" all the time.