Because the strings that are said to quiver at the
core of elementary particles are too small to detect —
probably ever — the theory cannot be experimentally
confirmed. Nor can it be disproven: Almost any
observed feature of the universe jibes with the
strings’ endless repertoire of tunes.
In that case, it's not science. At most, mathematics.
I really don't agree with that conclusion, though I recognize that it's not far from the traditional definitions of science.
The point is, let's pretend for the sake of argument that this article's main idea winds up being absolutely true: imagine that someone produces a rigorous mathematical proof that any consistent physical theory that incorporates both quantum mechanics and gravity (general relativity) must be some variety of string theory.
Now, we live in a universe where quantum mechanics is an experimental fact, and we live in a universe where general relativity is an experimental fact (each within their own domains of experimental accessibility). Would it truly be unscientific at that (thus far imaginary) point to conclude that string theory was a correct description of reality? The only alternative that I could see would be to abandon the idea that mathematics is able to describe our universe at all. And that doesn't feel like science, either.
[Note: As a professional string theorist, I'm hardly a disinterested party. But it does mean I spend a lot of time thinking about this stuff.]
If GR and QM can be derived from string theory, but the base theory doesn't provide any extra prediction of its own or simplifies the other two theories, then what good does it serve? Can we at that point say it's a "correct description of reality" or just a clever mathematical trick?
But it does. It predicts extra dimensions and it predicts supersymmetry. However it might be that these are only visible beyond our achievable energy levels such that we cannot verify them. The goal would then be to find what other indirect (as in not actually observing the strings) feature string theory predicts, that is hopefully within our reach, and verify that.
I'm not really sure what your point is. People are making the effort, both experimentally and theoretically. The LHC had some minimal potential to discover supersymmetry, but it has not done so thus far. This does not suffice as a disproof, though, due to the low energy levels. Simultaneously, string theorists have long labored to find more indirect effects that could be detected.
"Make the effort" is like complaining that cancer hasn't been cured because people just haven't tried hard enough. It's one of the hardest intellectual problems the human race has ever faced.
Not exactly. I mean, claiming that the LHC had "minimal potential" after all the "we shall see when the LHC shows the data" is quite disingenious.
The "effor" has nothing to do with cancer and has a lot to do with the REAL WORLD we live in. Physics either talks about REAL experiments or is otherwise void of value.
There are lots of speculative nonsensical theories which can be brought up and have no relation to the real world. As long as they are not falsifiable, they are not Physics.
> I mean, claiming that the LHC had "minimal potential" after all the "we shall see when the LHC shows the data" is quite disingenious.
The potential for discovering supersymmetry through the LHC was indeed minimal. There is nothing disingenuous about that.
The LHC is conducting a vast number of experiments on a vast array of topics. Perhaps the main goal was to uncover the Higgs boson, which unlike supersymmetry could be disproved if it wasn't found by the LHC.
> The "effor" has nothing to do with cancer
It's an analogy...
> As long as they are not falsifiable, they are not Physics.
They are theoretical physics, though.
Again, I don't know what your overall point is. That we should just all give up and end theoretical physics right now for good?
The bottom line is that if you think people aren't putting in the effort, you're wrong. That's my my point here.
In that hypothetical it might be a reasonable conclusion, but it would still be an unscientific one, since it would depend purely on your understanding of a concept rather than empirical evidence. And this distinction would only matter to a philosopher, since the conclusion would have no practical implications either way.
I don't agree that it would be completely independent of empirical evidence. In fact, it would be relying on a vast body of empirical evidence: the century or more of data that has been collected to verify general relativity and quantum mechanics.
It would be an indirect conclusion, but mainstream physics is already filled with those (e.g., the existence of black holes).
Here's the problem: I could easily make up another theory based on quantum mechanics and relativity, with the same amount of empirical support (none). The trivial other theory would not be in the form of math, but pure math is not a sufficient condition for scientific theories like quantum mechanics and relativity, and there would be no other meaningful difference.
Any predictive unification of quantum field theory and general relativity will be mathematical in nature.
A unification which is not mathematical is guaranteed to never be predictive and to never aid in the construction of a unified theory which is predictive, because the problems to be solved here are indeed entirely mathematical (divergent integrals).
String theory is a unification which is mathematical. Its direct empirical support is indeed currently none. However, as Hawking stated, it remains the only serious viable theory of everything right now. Every other competing theory has been found to have serious intrinsic issues, let alone lack of experimental support.
People take string theory seriously not because it's some vague unification. They take it seriously because it solves extremely difficult mathematical problems. Even if it turns out to be wrong, it is fairly likely that the mathematical insight gained will be invaluable. Indeed, it has even led to at least one awarding of the Fields Medal for advances in pure mathematics, to Edward Witten.
Your point about the need for prediction is relevant, but string theory is currently in the class of ideas which have no prospect of making meaningful predictions anytime soon. You might want to argue that the mathematical formulation makes the possibility that this will change in the future more likely, but I honestly don't see how that follows.
> I could easily make up another theory based on quantum mechanics and relativity [...]. The trivial other theory would not be in the form of math[...]
Can you clarify that at all? I honestly don't know how to reconcile your first sentence with your second one: you clearly mean something different by some of those words than I do, but I can't puzzle out what.
Are you suggesting that you'd make up a theory that said, for example, "The universe looks behaves according to quantum field theory in the following set of circumstances: [long list of conditions], and behaves according to general relativity in the following distinct set of circumstances: [other long list], and consists of a troupe of dancing angels under all other circumstances [including, for example, black holes and the early universe]." Because I'd hesitate to call that (or anything like it) in any sense a scientific "theory". It has zero explanatory power, because it's an archetypical example of "overfitting the data": it's a description of reality, not an explanation, and there is literally zero reason to expect anything that you used to fill in the gaps between the observed data to be true.
If that is the sort of thing you're talking about, then it sounds like your main claim is that there's no meaningful difference between that and (in our imagined case) a unique mathematical framework that gave rise to both quantum and gravitational observations as distinct limiting cases (while also making explicit claims about how the universe must behave in possibly-unobservable situations like black holes and the early universe). I don't really know how I would answer such a claim (if it's what you're trying to suggest), except to suggest an upgrade to a philosophical system that isn't deliberately obtuse. (I know that there is no a priori reason to expect the universe to have a mathematical description. But by this point, given the truly amazing success of those descriptions to date, the burden of proof really does have to rest on anyone who claims that those successes are just some sort of coincidence.)
My main point is that there is only one meaningful difference: the presence of the unique mathematical framework. This alone would not be sufficient to put that theory in the category of science.
Why? It comes down to the basic requirement that the theory should make some new and falsifiable prediction. Without that, the theory would at best be a mathematical unification (admittedly quite an impressive one). If using it isn't substantially better than using the older theories, all you've really done is invented a mathematical object that behaves like relativity in one context and quantum mechanics in another (and doing that without introducing loads of hidden variables seems to be enormously difficult). There is only something noteworthy and new in that mathematical sense. Is it interesting? Absolutely. Is it a scientific theory? Not really.
Yes, that is what Peter Woight at Columbia keeps saying on and on: http://www.math.columbia.edu/~woit/wordpress/. It is a sad world in which scientists dare (in the XXI Century) call "science" something which cannot be either verified of falsified...
Stephen Bond had that great article (well, lousy on the AI side) about "Positivist Programming", and how the philosophical education of scientists is often so poor they're unknowing logical positivists.
The whole achievement of post-positivism and falsifiability seems to be lost on people who seem to think "some math" == "rigor". String theory at least seems like a harmless example, I'm more worried about the public perception of economics.
Yes, economics is the most hyped pseudo-science and just because it "is mathematics, man! so it cannot be wrong...".
Like Varoufakis claiming he uses "Game Theory...".
Wilmott has a great example of the uselessness of mathematics in the real world when it comes to things that cannot be properly modeled: "when a magician says he is going to guess the card that the person from the audience chose, what is the probability that he guesses right?"
Wrong model (or too simplistic) => absurd/uninformative conclusions.
But "they use cohomology, so they cannot be wrong".
Also relevant is Woit's essay "Towards a Grand Unified Theory of Mathematics and Physics" [http://www.math.columbia.edu/~woit/mathphys.pdf]. Reading that, I can imagine that there might be statements in physics that are mathematically provable, but not experimentally testable. However, in his opinion, this clearly isn't one of them.
Very true. At most we can say that these guys found a mathematical theory that is consistent with what the universe looks like. I imagine that there could be an infinite number of such theories, if you are creative (and good at math) enough.
What if you considered (mathematically) the set of ALL theories that satisfied certain basic criteria (requires a small number of assumptions compared to the number of outputs it produces, doesn't change "arbitrarily" at some point in time... things like that) and which also matched the currently accepted physics of general relativity (in the limit of large sizes) and the currently accepted physics of quantum mechanics (in the limit of small sizes). And what if, when considering that set of theories, you were able to prove that any theory matching those criteria had to be equivalent (isomorphic) to a certain theory.
(This is like saying that the theory of basic algebra using equations and the theory of curves on graphs using geometry can be proved equivalent (isomorphic) by defining how to graph an equation and how to derive the equation for a graph. Once you've done that, it hardly matters whether you think the algebra is the "real theory" or the graphs... they are both equivalent and thus both equally "real".)
Suppose you found that, and thus proved (mathematically) that string theory was isomorphic (in certain key ways) to any theory that could apply to our universe, assuming that quantum mechanics and general relativity are roughly accurate. Why, then you WOULD actually have demonstrated that string theory isn't a "waste of time", that it is a useful way to make predictions about our universe.
And if I understand the article, this proof is precisely what these researchers are attempting to do. So far, they have constructed such proofs for much simpler universes than ours and a way of slowly working their way up to proving it for our universe.
That seems unlikely, because for that to hold they also need to show that there is nothing better than relativity and QM to describe the known universe. Then, you are back to square one, finding a better unifying theory, or even worse, proving that no such theory exists.
I don't think that you've really understood the sort of limits that the previous commenter (and the original article) were trying to describe. To a very good approximation, QM does describe our universe: any "better" theory must necessarily be equivalent in the appropriate limits (to many decimal places of precision, I might add). In the same way, relativity does describe our universe, and any "better" theory must be equivalent in the (different) appropriate limits. As I understand it, that is the level at which those are being assumed in this argument. The whole point of the theorems that these groups are trying to prove is that any "better" theory that matches both of these limits must be equivalent to string theory.
I think that would be even harder to achieve than a simple unification of QM and Relativity. Who knows what kind of phenomena we don't know about? There could be millions of theories that are consistent with current physics and still possible, given particular observations. In a few words, nobody can anticipate the kind of phenomena that could be observed by future physicists.
Except the history of such proofs suggests we not take them too seriously, because they necessarily depend on assumptions that may or may not apply to our universe. Hawking's proof that our universe contains a singularity in its past depended on a positive pressure assumption that is violated in inflationary universes but that was believed to be compellingly reasonable at the time, for example.
The important thing is that our inability to imagine a universe that violates some assumption tells us nothing about the likelihood that the universe violates it. What we can or cannot imagine is simply irrelevant to the way the universe actually is. No one could imagine a universe in which anything like Bell's Inequalities could be used to test local causality until Bell derived them.
There was nearly 200 years of philosophy arguing that no such test could possibly exist, culminating with the positivists just decades before it was found, and 2000 years of philosophy before that arguing that no such universe as one in which Bell's Inequalities were violated could exist, because no one could imagine such a universe, or such a test.
This is why science and math are different disciplines. Science is the discipline of publicly testing ideas by systematic observation, controlled experiment and Bayesian inference (if you read that definition carefully you'll find it's a lot closer to "Anything goes" than Popperian hypothetico deduction). Math is the discipline of formal deduction from axiomatic premises, or something like that. Math is an incredibly useful tool in science, but so is plumbing.
The attempt to prove there is one consistent theory of the universe is ambitious and interesting, but there are good reasons many scientists are skeptical about it. It has resulted in a great deal of interesting math and no strongly testable predictions over the past century (string theory encompasses so many diverse models of the low energy world that its predictions only ever knock off models, not the family of theories proper).
Testable predictions come in many forms, and "unique consistency" is one of them. It's just not one that many people outside the string community think they can achieve, and my personal bet is that the next century of string theory will involve a widely-believed claim that the uniquely consistent theory has been found, followed by its collapse some time later as a key assumption is found not to be true of the universe it is supposed to describe (that is if loop quantum gravity doesn't take the field on an experimental basis first.)
In principle, nothing that is unproven is impossible. But your belief in this possibility in particular shows that you are the one with a huge imagination.
(I am not a string theorist.) String theory makes plenty of testable predictions, they just aren't new so it has to compete with Quantum Mechanics. It gets into questions of how to choose between multiple theories that make the same predictions.
Scientists need to admit the possibility that their theories are wrong, but that cuts both ways. Just because a theory seems untestable today, that does not mean it is guaranteed to be untestable forever.
If a theory is internally consistent and its predictions match known empirical evidence, then the theory is supported by that evidence just as much as any other theory is. That does not provide a reason to prefer it over other theories, but by itself that is not a reason to dismiss it.
The way I understand it (and I am probably completely wrong, seeing as I am not a physicist, nor a mathematician), string theory is to scientific theories as a class is to objects in OOP.
You can indeed never falsify string theory itself, but you absolutely can falsify a theory derived from string theory by picking parameters for it.
I define science as Popper did. And based on my understanding of Gödel's and Kuhn's work, I distrust mathematical physics and cosmology. Developing clever hypotheses is a great thing, but data rules.
Strictly speaking, nothing can ever be confirmed or disproven, for random events in the universe can always spoil any experiment. QM, by its statistical nature, actually destroyed the definition of science. We can only speak of how likely we believe that something about the physical reality is true.
In mathematics and logic the "rules" are not only known but defined a priori by a human being. Not so with physics, and as a result the "proofs" of physics, what are known as scientific theories, have never been absolute. Quantum Mechanics merely elaborated on that theme.
As though the results of randomly flipping a coin "destroyed the definition of science"?
Dealing with randomness has always been a part of science. Understanding how far that randomness permeated the nature of the universe was a revelation, but it's hyperbolic and sensationalistic to say it as you did.
Not the same. You're conflating two very different types of randomness. The coin is understood to be deterministic, just hard to predict. Quantum mechanics is not like that. It's viewed as fundamentally non-deterministic.
Imagine this. Imagine that tomorrow, a string theory physicist made the discovery that the strings vibrations can actually be predicted. If you know their configuration and certain starting parameters, you can predict the spin of an electron, say.
Would "the definition of science" suddenly not be destroyed... or would it be re-destroyed?
Second of all, you missed the point of this sub-thread. Even one of the quoted scientists claims, "Bell's theorem is the most profound discovery of science."
He didn't say it "destroyed science".
How does the theory of the randomness described in QMT destroy science?
It's true that no hypothesis can ever be confirmed or disproved. At best, we can estimate the probability that some hypothesis is wrong aka inconsistent with our observations.
I used to work with Mukund (Rangamani) and Veronika when I was doing my PhD at Durham Uni (2004-2008) -- it's great to see the work they and others are continuing to do on this. It formed the basis for my thesis (http://etheses.dur.ac.uk/2906/), and Anti de-Sitter spacetimes allow for some great visualizations (I spent much time working in Mathematica to produce the one shown in the slide here: http://bethnalgreenventures.com/2013/08/05/guest-blog-from-m...).
You also learn something new everyday - I'd never come across the term "Fisheye Universe" for AdS space before! :)
