Should we consider string theory to be science? (Philosophy)

For the discussion of the sciences. Physics problems, chemistry equations, biology weirdness, it all goes here.

Moderators: gmalivuk, Moderators General, Prelates

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Should we consider string theory to be science? (Philosophy)

Postby Twistar » Tue Feb 09, 2016 6:40 pm UTC

Disclaimer: This is going to be a long post (4 pages in word).. Sorry about that. I just really want to see the discussion addressed from a particular point of view from which I never see it discussed. Most of the post is to set the stage for a few (hopefully) direct and clear questions at the end.

Apparently there has been much debate in string theory and philosophy of science circles in the past couple of years over whether we should consider string theory to be science or not. A recent nature comment makes the argument that this not some pseudo-scientific ivory tower discussion that we should all brush off, but rather that this is a very important time when, in the midst of public denial of scientific conclusions such as climate change and evolution, scientists are forced to more clearly answer the question of "what do we mean by a scientific theory."

I agree that this is a very important discussion for the same reasons. A few notes on my background and where I'm coming from with my thoughts on these issues.
-I'm working on my Ph.D. in physics but in an unrelated experimental field rather than theoretical/mathematical physics like string theory. I do a bit of quantum optics so I know quantum field theory as far as it concerns the photon field but any other quantum field theory I only know by analogy, and I only know the standard model and beyond the standard model from various colloquium talks and colleague discussions. I basically don't know anymore string theory than you can quickly google or see on t.v...
-I'm even less qualified in terms of philosophy. When I was a freshman working on my Bachelor's I took a class called philosophy and physics. Even though it was just one class the ideas really stuck with me. In particular they stuck with me because it seems like that class taught me more philosophy of science than at least 90% of the scientists I speak with or whose comments I read.. Anyways I’m going to run away with what little philosophy of science I do know, so feel free to correct or criticize me on these points.

I want to give a synopsis of my main takeaways from the course because that is what is guiding my thoughts on this issue and I can’t find anyone approaching the questions of string theory from the same direction. This will be a historical summary of the philosophy of science in terms of a few philosophers.

-Francis Bacon (~1600): Inductivism. Basically we come up with a hypothesis about the world and then we go around checking if things agree with that hypothesis or don’t agree with that hypothesis. If things agree then the theory is right. The biggest logical problem with this theory is that just because we see instances that confirm our theory it NEVER means that we might see an example that refutes it. Say my theory is that “all geese are white”. I look all over and only see white geese, but there may be a black goose on another continent. The point is we can never logically take the theory to be true. Another major problem is that there is no discussion of where the hypothesis comes from.

-Karl Popper (~Early 1900s): Falsificationism. Popper looked at the problems of Inductivism and realized it couldn’t work. However, his argument was that for a theory to be scientific there must be a possibility that one could observe something that would disprove the theory. So for example, the white geese example above could be considered to be a scientific theory because it could be falsified by the observation of a black goose. Popper was trying to demarcate science from things like psychoanalysis at the time. It seems to me that he would argue psychoanalysis is not scientific because you can basically explain away anything by just changing your description. Apparently almost all scientists think the philosophy of science ended here which is pretty frustrating for me when I’m trying to read about string theory…

-Duhem and Quine (~early 1900’s? I don’t really know actually): Theories can’t be falsified. Popper rejected the idea that a theory could be proven by looking at enough example. The Duhem-Quine thesis rejects the idea that a theory can be falsified by a falsifying instance. They argue this by claiming that any hypothesis about the physical world comes with auxiliary hypothesis that can always be rejected in light of a presumably falsifying instance. For example, upon observation of an apparently black goose the scientists might be able to say “that’s not what I meant when I said goose” or “that goose was painted black so it doesn’t count” or such. A more realistic example was the apparent observation of faster than light neutrinos. A strict falsificationist would say the measurements falsify the special theory of relativity. Of course, NO SCIENTIST said that (this hints at the problems with falsificationism) but rather, everyone said “there must be something wrong with the experiment”. So in other words, there was an auxiliary hypothesis going into the special theory of relativity which say “if you’re experiment is working right, then you will never see something faster than the speed of light”. THIS was the unstated hypothesis scientists chose to reject INSTEAD OF rejecting the theory of special relativity. The point is that this can be done with any theory and shows falsificationism to not be a great metric for science either.

Thomas Kuhn (1960’s): Paradigms and revolutions. I think I’m really simplifying all of these philosophies so apologies but this is how I understand them. Kuhn looked at the prior results that theories cannot be confirmed by looking at confirming instances or denied by looking at falsifying instances. That was a weird state of affairs.. Kuhn’s big conclusion to this was that scientists basically do what they feel like. In other words, science is a sociological affair. He introduced the idea of scientific paradigms and investigated how, in a sociological sense, those paradigms could be overturned. Examples include the Copernican revolution when scientists chose to accept the theory that the earth revolves around the sun rather than vice-versa and the shift from Galilean to Special/General relativity. Kuhn said that basically at some point the scientific community comes to a consensus that they are more interested in a particular set of ideas than another and eventually there is a tipping point when the paradigm shift happens. This was a very damning criticism of science and it wasn’t taken lightly. It basically said that science is just what people want and that science doesn’t really get at any deeper truths of nature than any other field of study. I think Kuhn definitely realized this. In my class this was introduced as being a philosophy which was sort of in line with post-modern thinking going on at the time.

Imre Lakatos (~1970): Progressive research programs. This is where my knowledge of the philosophy of science ends so I’m definitely missing out on more modern stuff so I would be interested to learn about that. But here goes. Lakatos took what Kuhn said and tried to recover something that looked more like science that these paradigms. I think Lakatos also tried to take a look at how scientists actually DO science and how they choose to accept or reject different theories. Lakatos said that science has to have progressive research programs in the sense that science should progress or move forward always. There are a couple of questions, what is science moving towards and how do we tell if science is progressing? He gives a few criteria which I think are realy nice and really close to what scientists actually do. Say we have two theories A and B. There will be 3 criteria to compare the theories.
1) Explanatory power: If one theory has all of the explanatory power of another AND THEN SOME we accept that theory. A great example is Einstein’s theory of gravity versus Newton’s Einstein’s theory explains everything that Newton’s does and then some so in that regard Einstein’s theory is better.
2) Experimental Corroboration: Bearing in mind the fact that Inductivism fails, we still think a theory is better the more experimental corroboration it has. Every time I see another white goose it DOES give me more confidence that my theory is useful*. That is, experimental corroboration doesn’t PROVE a theory to be TRUE, but we still like it when it happens. Note the intentional use of the word “corroboration” rather than “confirmation” or “proof”.
3) Occam’s Razor: Scientists prefer simpler theories to more complex ones if they’re given the choice. This is again a very pragmatic criterion. We wouldn’t choose a simpler theory if it had less corroboration or explanatory power, but if A and B are the same on criteria 1 and 2 then we might as well take the simpler one since it will be easier to work with and talk about.
Lakatos says that, in practice, scientists use these criteria to choose between different explanations of reality.
edit: 4) Ah, here's a question. Maybe the answer to all of this is no. But what about all of these questions IN PRINCIPAL? Like, is the problem just that we don't yet know how to do the math to reduce string theory or to pick out a subset of the 10^500 worlds? If that's the case then this again feels like 2a). So does string theory IN PRINCIPLE make these predictions but we just can't figure out how to massage them out of the formalism?


And now let me give my synthesis (as someone hoping to one day be a scientist) of all of this and how I define science.
Twistar (~2010): Given that we can’t accept or reject a theory as being true we need to abandon the notion of finding “true” theories. That was a red herring all along. I think this is pretty clear. We BELIEVE that gravity will pull our feet towards the ground when we get out of bed in the morning but what if it doesn’t? We can just never logically exclude that possibility. So forget about truth. Instead, here is what I think science does in one sentence. Science searches for the model of reality which most closely matches our experience. And then I guess there should be more sentences with references to Lakatos’ criteria.



Anyways, enough philosophy of science. Now with all of that background I want to get to my questions about string theory which I have not been able to answer by looking on the internet. I’m basically trying to use Lakatos’ criteria for assessing string theory. Most people use Popper’s which is frustrating because it is old and naïve. I am probably being naïve by being so into Lakatos but I haven’t done any research beyond that point in the philosophy of science unfortunately.
String Theory Questions
1) Does string theory make ANY predictions about the world? Like for example, does string theory predict that the sun will exert a gravitational force (via the bending of space time) on the earth. Does it make the same predictions as the standard model? If string theory really does make no predictions about the world, then that is kind of strange. I always thought of string theory as an extension of the standard model and in that sense that it would at least make the same predictions as that. I always assumed the motivation of string theory was to explain the same things as the standard model and then some.
2) Does string theory make any NEW predictions about the world? This question is in DIRECT reference to Lakatos’ criterion 1. People often say string theory makes no predictions. I’m trying to discern if they mean to say string theory makes no novel predictions.
2a) Also bear in mind that just because we might need an outrageous amount of energy to test the prediction is not a reason to discount that as a prediction. The Bose-Einstein condensate was predicted about 70 years before it was experimentally realized but no one said that prediction was unscientific. There are COUNTLESS more examples of this in science.
3) People talk about how there are 10^500 versions of string theory or whatever in relation to the multiverse. They say that we can just pick whichever one best fits our data. First of all, this argument seems to be at odds with the claim that string theory makes no predictions.. because apparently we can pick whichever model fits the data which would be a property of a theory that makes predictions. Anyways I’m struggling to form a question out of this. I guess it’s something like this. Are some subset of these worlds consistent with what we do observe? Could we make more observations to further discern which of these worlds we live in?

I don’t know… it’s getting tough here and I’m losing clarity in my thought. But my basic question is this: Does string theory make predictions about the world or not? No qualifications on those predictions like “they are testable” or “they are unique” or “they are falsifiable”. Like, is string theory trying to model the physical world or not? Then, after answering that, the next question would be to start applying Lakatos’ criteria to those predictions. My suspicion is that string theory IS a scientific theory but at this point it isn’t very useful yet, but that it definitely will be useful in the future as we both understand string theory better and become more technologically advanced to perform more advanced experiments. I’m trying to affirm or deny this suspicion.
And I also don’t want to hear about Bayesian comparisons of different theories. People resorting to that usually don’t seem to be aware that they can use Lakatos’ criteria to compare theories. I’m not very sympathetic the Bayesian stuff but that’s for a different thread, in any case, as it is, I am interested in hearing an analysis from this particular Lakatos perspective and not Bayesian.
*Note that early on people thought about science as searching for the “true” theory of nature. I have sneakily slipped in here that we are now looking for a “useful” theory of nature. This is critical and I’ll say more on it a few paragraphs down.

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Tue Feb 09, 2016 6:57 pm UTC

LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Tue Feb 09, 2016 7:42 pm UTC

doogly wrote:I recommend reading here:
http://backreaction.blogspot.com/2015/1 ... ience.html


Doogly, great! Thank you! That's exactly the type of article that I was looking for. I think it definitely addresses most of my questions. And it looks like my suspicions we're basically correct.

Now I guess here is the next question which maybe also motivates my interest in the subject.
I thought the unification of gravity and quantum field theory was still an open question which people had some ideas about. Apparently it sounds like this has already been done though.. So I know there's all sorts of offspring of different quantum field theories and there's also like quantum loop gravity and stuff. Here is my question. Are all of those quantum gravity theories consistent with all physical observation thus far? In other words do those theories really consistently make the same predictions as the standard model and general relativity? If so that's amazing and we're much further along than I thought we were!

And also, if all of that is the case then what is the big deal? Ok, the theories are consistent with everything so far, we obviously can't choose between until the experiments come along and we might have to wait a long time for that but that's just how it works... The theorists should obviously keep working on the theories, and they should of course use non-empirical criteria to try to decide which path they should go down but obviously in the end it's going to be determined by experiment.. This all seems very clear to me so I'm confused about what all the confusion is about [\stuff physicists say]. Like, no, you can't use your non-empirical criteria to decide which theory is "right" because of what I said above, we need to abandon the notion of truth and instead search for the model which most closely resembles our experience. And you obviously CAN'T use a non-empirical criterion to discern that.

To summarize what I feel string/other beyond SM theorists should be doing:
-Exploring the existing theories to understand them better
-Refining the theories, it sounds like some of the theories are still underdetermined. Like if there are 10^500 worlds the open question is raised "why do we observe this particular one". But I consider that process to kind of be part of "explore existing theories"..
-Applying those theories to current open questions, dark matter? black hole information?
-Thinking up ways experimental tests that could discern between one theory or another


Yeah I guess I Just don't see where all the confusion is coming from unless it really just is people having the wrong idea about what science is and what scientists do, which is, I guess again, my main suspicion. If my suspicion is correct then it just even highlights again the importance for scientists and science communicators to become more unified as to what we mean by "science".

The way people talk about it is like there is something fundamentally different about the type of theory that string theory is versus every other theoretical model that has come along.. No one criticized the people who predicted the Higgs boson years ago for being unscientific...

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Tue Feb 09, 2016 8:17 pm UTC

The issue is often that they can but relying on some parameter choices that are essentially begging the question. Consistent with an answer you already know is easy, predicting is hard.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

aph
Posts: 296
Joined: Tue Nov 12, 2013 7:48 am UTC

Re: Should we consider string theory to be science? (Philoso

Postby aph » Wed Feb 10, 2016 4:23 am UTC

Have you read anything from Lee Smolin?
https://en.wikipedia.org/wiki/The_Trouble_with_Physics (it is down in the references of the nature article you linked, along with a few others)

Don't know what to think about string theory, I'm waiting for a technological application. A new type of a powerplant of something :?

---

I'm not sure I agree with criticism of falsifiability as a criterion for a theory. From my understanding, falsifiability means that it needs to be possible, in principle, to design a TEST of the theory. The test will tell how well the prediction holds. Those geese examples seem to confuse the idea.
I like how Feynman put it: https://www.youtube.com/watch?v=EYPapE-3FRw

And Poincare is very interesting: http://strangebeautiful.com/other-texts ... thesis.pdf
He was discussing a century ago how science can't reach absolute certainty and truth.

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Wed Feb 10, 2016 12:30 pm UTC

Smolin does some interesting research occasionally but when he talks about string theory it is clear he is just being cranky.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Wed Feb 10, 2016 5:04 pm UTC

doogly wrote:The issue is often that they can but relying on some parameter choices that are essentially begging the question. Consistent with an answer you already know is easy, predicting is hard.


Hmm, so are people just upset with string theory because it hasn't yet made any "novel" predictions? That's certainly not a reason to discount it as science. Well anyways.. going back to Lakatos' criteria it sounds like all of these beyond SM theories have maybe similar explanatory power as SM + GR, similar experimental corroboration in the sense that all experiments serve as corroboration to all of these above approaches so it comes down to occams razor. But maybe we can replace occams razor with "non-empirical theory assesssment". But anyways, it comes down to that. And here is how I interpret that.
1) You CANNOT say your theory is true based ONLY on non-empirical theory assessment. This is because it is conflict with Lakatos' second criteria.
2) You can and should probably use non-empirical theory assessment to compare two theories with similar explanatory power and experimental corroboration. Theorists must do this to figure out where to spend their efforts. But that doesn't mean the theory is more right.. obviously in the end only experiment will tell.

As for falsifiability.
Yes, Aph, you have correctly characterized falsifiability as the possibility of in principle designing a test of the theory but I think the point is when you actually set up the test you can always reject auxiliary hypotheses. The faster than light neutrinos is a better example than the geese. No one took that experiment as a falsifying example, rather, they instead rejected the auxiliary hypothesis that the experiment was running properly.
I think we can debate and quibble about what falsification means. But it's going to get very epsitemologic if we go down that road and I'm going to start saying things like "in the end all knowledge is based on faith in something" and based on that you can ALWAYS technically/logically/whatervly speaking reject auxiliary assumptions. But I don't think we should do that. I think the key take away is that humans don't have access to the truth. We can't prove something true, and we can't prove something false. So just forget about it. Don't think about trying to prove or disprove theories. That is a category error.

Rather than searching for theories that are "closest to truth" instead look for theories that "best match our experience", where "best" is defined by the criteria of Lakatos. We can then examine falsification from that point of view. I think then the reason theories which are traditionally put in the camp of not-falsifiable are considered to be not scientific is because they have very little to no explanatory power. If I can always modify the theory so that it matches my experience then it isn't really explaining my experience, and it certainly can't be used to make very reliable predictions. So then instead of rejecting the theory because of the logically tenuous criterion of falsifiability we are instead rejecting it based on it's lack of explanatory power.

And then what is even more the point than any of this, is that when scientists (or mechanics or anyone in any technical or even some non-technical fields) is doing their day-to-day work they are never really thinking about the falsifiability of what they are doing. Instead they are messing around with things and trying one thing and then another and just seeing how well it is work. How closely are their ideas coming to describing what they see happening?

We can look at String theory from the falsification-turned-into-a-lack-of-explanatory-power perspective too. You might say if string theory predicts that 10^500 different worlds all happen and thus we can just pick whichever one fits our reality then yes, again we find that string theory doesn't have very much explanatory power. In particular, it doesn't allow us to make predictions. If we could narrow down which world we are in then we could make predictions. I'm sure there are some people trying to figure out how to do that.

It just seems to me like all of the difficulty encountered in these debates just crumbles and vanishes as soon as you reject the notion that science is searching for truth and things start to make a lot more sense. I'll have to have a look at what Poincare has said about it.

Tchebu
Posts: 564
Joined: Thu Feb 22, 2007 12:42 am UTC
Location: Montreal

Re: Should we consider string theory to be science? (Philoso

Postby Tchebu » Wed Feb 10, 2016 8:21 pm UTC

Just thought I'd comment on what string theory takes as input and what it does or does not predict.

The starting point for string theory is to study the motion of a string (i.e. a one-dimensional object with some tension/energy-density) in a spacetime which locally obeys special relativity. You then hit this problem with the quantum mechanics hammer to obtain a quantum mechanical version of the theory. So this is the input of the theory: local lorentz invariance and quantum mechanics. Any other thing you've heard string theory says is a consequence of these two basic inputs!

The main early "predictions" or "successes" of string theory come from exploring the consistency conditions of this quantized theory. Whenever you're dealing with a quantum theory that you obtained by quantizing a classical theory you need to worry about quantum effects ruining symmetries of the classical theory. The requirement that this doesn't happen is a very constraining condition on what quantum theories are possible!!

In the case of string theory these consistency conditions lead to the following restrictions as you consider more and more quantum effects that might break the theory and demand that they don't:

1) Spacetime must be of a certain dimension! 26 for "regular" (bosonic) strings and 10 for supersymmetric strings.
2) The spacetime must solve the equations of General Relativity! This is a genuine prediction of string theory, GR wasn't anywhere in our input, but it turns out to be mandatory.
3) Higher order quantum effects give additional corrections to General Relativity, that only come into play at short length scales.

Furthermore, the various oscillation modes of the strings behave like particles when you zoom out and the energy contained in the oscillations is interpreted as the mass of the particles. There's a set of massless states, which covers the full set of the particles we know and love: spin 1/2 fermions, spin 1 bosons, which are the basic ingredients you need to construct something like the standard model, but also a massless spin 2 particle. Upon closer examination this spin 2 particle turns out to behave exactly the way you'd expect a graviton to behave! So we already had the requirement that our spacetime obeys GR, but now we also have quanta of gravity waves. So we have a bona fide theory of quantum gravity, which reduces to classical gravity in the correct limit! The higher oscillation modes all correspond to particles with enormous mass, well above anything we should expect to see in any particle accelerator. (There's also a negative mass state, which a "Bad Thing"(TM) and only goes away if you consider supersymmetric strings, which is why this is the thing people do).

Now this might all seem excellent, except for restriction number 1 above... 10 dimensions? wtf? If you want to have a theory with only 4 large dimensions, you need to roll up (compactify) the remaining 6 into some small curled up shape and this is where things become complicated. There's an enormous number of ways to compactify 6 dimensions of space and you can vary the shape and size of the various parts of it. Each different compactification leads to slightly different physics in the remaining 4 dimensions, different kinds of particles, different strenghts of interactions between them, all of them perfectly internally consistent. So string theory does not predict our universe as the unique possible way that things could be and has no guiding principle that would allow us to pick out the standard model out of all the gazillion possibilities. This is what people are usually referring to when they say string theory makes no predictions.

On a personal level, I disagree with this objection. It's exactly like objecting to quantum field theory, because the standard model isn't the only possible one. You can change the coupling constants, you can change the field content and write down all the theories you'd get from string compactifications! The number of possible consistent quantum field theories is at least as big as the number of possible string theory compactifications, because string compactifications always result in consistent quantum field theories!

So tl;dr
Input of string theory: strings as fundamental object + special relativity and quantum mechanics
Output: Requires equations of General Relativity to hold (+ additional quantum corrections, beyond experimental reach), provides necessary ingredients to construct a theories of particle physics but also includes a graviton.
Claimed Problems: A vast number of consistent 4-dimensional effective theories, with no guiding principle to pick out one possibility over the others, but no more vast than the number of possible quantum field theories to begin with...
Our universe is most certainly unique... it's the only one that string theory doesn't describe.

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Wed Feb 10, 2016 9:18 pm UTC

Thank you Tchebu, awesome explanation!

Question. Are our current observations and measurements of the standard model consistent with
a) exactly one choice of compactification
b) all the choices of compactification
c) some choices of compactification? If so do we know roughly how many? Like is it 6 or 7 or 10^350?

Also another question. When you refer to "an enormous number of ways to compactify 6 dimensions of space" is that number of ways finite or infinite? Is this 10^500 number that is often quoted?

Whether I agree with your objection to the objection that we don't know the proper way to compactify string theory depends on the answer to the 1st question.

Tchebu
Posts: 564
Joined: Thu Feb 22, 2007 12:42 am UTC
Location: Montreal

Re: Should we consider string theory to be science? (Philoso

Postby Tchebu » Thu Feb 11, 2016 5:59 am UTC

Yeah, the 10^500 or whatever is the rough estimate of the number of different topologies the compactified space can have and the kinds of fluxes you can put along various cycles in that space. Each choice leads to a different effective 4D theory.

As for your first question, it's definitely not option b). Most models people actually deal with in research are not the standard model, but some simpler theories that are easier to deal with mathematically, but which have many of the same features as the part of the standard model they want to study (usually QCD, since that's the part we don't have a good handle on). It's also not option a) just because the standard model is a (relatively) low-energy theory and can result from a whole range of higher energy theories. This is not something that's specific to string theory, but is a general fact that a whole class of different high energy theories can look the same at lower energies.

So the answer is c) and although I can't give you the number you're asking for, it's really more of a field theory question than a string theory question. Any compactification which results in a 4D theory, which then flows to the standard model at low energies is fair game. Somewhere along the way you also have to break the supersymmetry that string theory has. So it really becomes a question of figuring out which supersymmetric field theories contain the standard model in their low-energy limit. From there it's not too hard to construct these theories as a string compactification.

I also feel that besides it being an extreme version of the needle in a haystack problem, part of the reason why we still don't know the quantitative answer to your question is that I don't think people are all that interested in pinpointing the exact compactifications to get the SM, since that's not really going to give us extra tools to learn new things. All the stringiness goes away at much higher energies and everything really just reduces back down to vanilla quantum field theory.
Our universe is most certainly unique... it's the only one that string theory doesn't describe.

User avatar
nash1429
Posts: 190
Joined: Tue Nov 17, 2009 3:06 am UTC
Location: Flatland
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby nash1429 » Wed Feb 17, 2016 6:29 pm UTC

You might be interested in String Theory and the Scientific Method and this related article. The book is written by a physicist-cum-philosopher and aims to answer the question: why do physicists believe they don't have empirical evidence for? It's less about the scientific method or the bounds of science per se and more about the mental heuristics of truth in science.

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Thu Feb 18, 2016 4:32 pm UTC

Tchebu wrote:I also feel that besides it being an extreme version of the needle in a haystack problem, part of the reason why we still don't know the quantitative answer to your question is that I don't think people are all that interested in pinpointing the exact compactifications to get the SM, since that's not really going to give us extra tools to learn new things. All the stringiness goes away at much higher energies and everything really just reduces back down to vanilla quantum field theory.


This confuses me. Here's another question.
1)Does string theory describe the physics of phenomena which are not explained by GR or the SM? Quantum mechanics at black holes? Inflation? Dark Energy? Etc.

If it did then it seems to me that if we knew which subset of compactifications led to OUR GR and SM we could then use those compactifications to study physics which is not yet described by GR/SM. That would to me be exactly "extra tools to learn new things."

@nash1429
That article was interesting. I haven't read Dawid's book, but I'll admit that it was conversations and articles about that book that rekindled my interest in this question. As for the article it was pretty good. Here are my thoughts.

1) I agree with the conclusion of the article which seems to be: The article claims that many string theorists accepted the idea that non-empirical justification justifies WORKING on a theory, but it doesn't provide confirmation or corroboration of the theory.
1a) The annoying thing to me about this is that it seems obvious... This is how theory has always worked. It is only recently for some reason people (Richard Dawid?) have started saying we should take these non-empirical criteria to be corroboration of the theory.. The only thing that seems different about string theory compared to others is that the string theorists are just more impatient about getting their theories corroborated than theorists before them.. Or maybe they're not more impatient but because of the nature of the theories it's going to take in their eyes longer to corroborate the theories than it has in the past?

2) It was interesting to me that the philosophers of science said the cutting edge model was this Bayesian confirmation. That seems stupid to me. They're trying to put a quantitative number on something that is not quantitative. And you can use anything for your priors.. I don't know. I don't want to get into an argument about Bayesian inference, that's another thread. I would be curious to see what Bayesian inference provides over Lakatos' criteria. But I think that would maybe be better suited for a different thread.

3)
“You mean confidence that it’s true?” asked Peter Achinstein, an 80-year-old philosopher and historian of science at Johns Hopkins University. “Confidence that it’s useful? confidence that …”

This quote hits on another key point in the article. It has to do with the idea of corroboration and what it means for a theory to be more corroborated. It also hits upon the purpose of science. I think the first important point that I take is that we are being naive if we ever gain confidence that a theory is "true" no matter how many times it's experimentally confirmed, see my above posts about why truth is a red herring. I think rather the utility of science is in ultimately providing new technologies that improve the lot of humanity. Science is able to do that because it produces models which identify predictable, reproducible, reliable phenomena that occur in our world. Once we understand what is going on we can leverage this understanding and start shaping matter and energy towards our benefit. So when we have confidence in a scientific theory it means exactly what the respondent says, it means that we will go ahead and use that knowledge in the next link of science. For example, it is clear that scientists have confidence in GR because we use GR in our satellite GPS systems.


But anyways, I'm struggling again with my train of thought here and have to get going. I think this 3rd point is also a seperate point from what I'm trying to get at in this thread.
I think my main question for this thread is: Does string theory make predictions or not, rather than: what do we do since string theory doesn't make predictions?

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Thu Feb 18, 2016 5:00 pm UTC

"String theory" as a whole makes no unique predictions in particular.

As it becomes more and more all encompassing, it becomes less and less likely that it will be a "one true path" in the sense that if you were to make a multiple choice test, "who is the right quantum gravity? a) string theory b) lqg c) dynamic triangles d) group field theory e) lisi's shit f) .........." that you would answer "string theory, totally, fuck those other guys"

But it does mean that whatever winds up being right, will wind up having elements that look like string theory.

It seems like the working model has really progressed beyond this debate, in a birds vs ornithologists way. It seems like Dawid is just interested in making a more sophisticated and eloquent book out of a discussion that was already tired and overwrought ten years ago with Smolin and Woit.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

quarkcosh1
Posts: 85
Joined: Wed Aug 28, 2013 11:53 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby quarkcosh1 » Thu Feb 18, 2016 5:00 pm UTC

One thing I never see mentioned in these debates is engineering (string theory). One of the motivations for doing science is to eventually find an engineering application for it but how would you do this is you can't prove the theory correct. If any engineering application were found it should also have the effect of proving the theory.

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Thu Feb 18, 2016 5:22 pm UTC

quarkcosh1 wrote:One of the motivations for doing science is to eventually find an engineering application for it

oh no. oh no no no.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

Tyndmyr
Posts: 11443
Joined: Wed Jul 25, 2012 8:38 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Tyndmyr » Thu Feb 18, 2016 5:53 pm UTC

doogly wrote:
quarkcosh1 wrote:One of the motivations for doing science is to eventually find an engineering application for it

oh no. oh no no no.


It's a pretty common motivation. Hardly the only one, of course, but I don't see a problem with quark's statement.

Granted, I also can't imagine a good engineering application of string theory, but string theory ain't really my game, so maybe someone else can?

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Thu Feb 18, 2016 6:05 pm UTC

It is the thinking of Tools. Tools! But this is the haven of the Sweet. We ought not abide Tools.

But there are no foreseeable applications.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

aph
Posts: 296
Joined: Tue Nov 12, 2013 7:48 am UTC

Re: Should we consider string theory to be science? (Philoso

Postby aph » Thu Feb 18, 2016 8:28 pm UTC

Science often follows technology. People tinker and build stuff, then others study the inventions and describe phenomena mathematically. Maxwell developed electromagnetic equations after Faraday's discoveries... James Watt claimed he didn't know much about thermodynamic theories when he invented (or improved) his steam engine. Aerodynamics was influenced by the Wright brothers' invention, and further developed after the successful first flights. Or say, in today's medicine, there are a lot of chemicals we know for certain to cure diseases because the experiments prove it, but we don't necessarily know the exact mechanism of how they work. We're still waiting for science to catch up to those.

Maybe some new phenomena we discover by accident will help find a new version of string theory or a theory of everything or something in that area.

ijuin
Posts: 1110
Joined: Fri Jan 09, 2009 6:02 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby ijuin » Fri Feb 19, 2016 6:16 pm UTC

Truth in science is like a mathematical limit, to be approached closer and closer, but never quite reached. We can build new theories that are more accurate (e.g. GR/SR replacing Newtonian mechanics), but we will never have The Last Word on the nature of the universe.

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Fri Feb 19, 2016 7:16 pm UTC

But you can determine whether such a thing as a limit exists, whereas with "Truth," it is not clear that such a concept even makes any sense.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Fri Feb 19, 2016 7:36 pm UTC

ijuin wrote:Truth in science is like a mathematical limit, to be approached closer and closer, but never quite reached. We can build new theories that are more accurate (e.g. GR/SR replacing Newtonian mechanics), but we will never have The Last Word on the nature of the universe.


I agree with doogly. This is just a silly notion and an attempt to apply mathematics outside of the realm where mathematics makes sense. To get around this confusion just drop the notion of truth entirely. Once you do that a lot of these discussions about string theory and science start to make a lot more sense.

Regarding science and technology. I wasn't trying to say that to pursue science we should only be motivated by possible technological achievements. However, I was saying, that in the end, sciences real worth comes from allowing humans to manipulate our universe in ways that benefit us. When we say "science works" or "we are confident in this theory" I think it is something like saying we are taking a step towards being able to manipulate things better. And abstract mathematical physics theory whatever, is all taking steps towards that. They're just very early steps along the road.

ijuin
Posts: 1110
Joined: Fri Jan 09, 2009 6:02 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby ijuin » Fri Feb 19, 2016 7:43 pm UTC

Well, I meant that "truth" is an idealization that we can only approximate, never achieve. We can come up with "better" theories that make more accurate predictions, but we will never reach the point where there is no room for further improvement.

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Fri Feb 19, 2016 8:46 pm UTC

Right, but you are implicitly assuming that there is some "Truth" that exists in some sense which is meaningful, but not accessible. And the meaningfulness of such a notion is what we are saying is not so.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

aph
Posts: 296
Joined: Tue Nov 12, 2013 7:48 am UTC

Re: Should we consider string theory to be science? (Philoso

Postby aph » Fri Feb 19, 2016 10:37 pm UTC

Twistar wrote:However, I was saying, that in the end, sciences real worth comes from allowing humans to manipulate our universe in ways that benefit us. When we say "science works" or "we are confident in this theory" I think it is something like saying we are taking a step towards being able to manipulate things better. And abstract mathematical physics theory whatever, is all taking steps towards that. They're just very early steps along the road.

Yeah, the goals of science are expressed as 'prediction and control' of phenomena.
The only way we can trust a science (that the theories will help us in achieving a goal) is if its predictions are checked in experiment against reality. In that sense, string theory is not yet 'trustworthy'.

Economics and social sciences have a similar problem with experiment. Usually, the most you can do are correlation studies and often multiple theories can explain the same findings in wildly differing ways. Can't trust a science that doesn't have experiment.

Tchebu
Posts: 564
Joined: Thu Feb 22, 2007 12:42 am UTC
Location: Montreal

Re: Should we consider string theory to be science? (Philoso

Postby Tchebu » Sat Feb 20, 2016 5:05 am UTC

Twistar wrote:
Tchebu wrote:I also feel that besides it being an extreme version of the needle in a haystack problem, part of the reason why we still don't know the quantitative answer to your question is that I don't think people are all that interested in pinpointing the exact compactifications to get the SM, since that's not really going to give us extra tools to learn new things. All the stringiness goes away at much higher energies and everything really just reduces back down to vanilla quantum field theory.


This confuses me. Here's another question.
1)Does string theory describe the physics of phenomena which are not explained by GR or the SM? Quantum mechanics at black holes? Inflation? Dark Energy? Etc.

If it did then it seems to me that if we knew which subset of compactifications led to OUR GR and SM we could then use those compactifications to study physics which is not yet described by GR/SM. That would to me be exactly "extra tools to learn new things."



The thing is, quantum gravity works the same way regardless of whether the rest of physics is governed by the standard model or some other field theory. The current theoretical problems in physics aren't "how does the black hole at the center of our galaxy work?", it's "how do black holes work in general?" Deviations from the standard model won't affect things like the mechanism by which the singularity at the center of a black hole gets resolved or how information about infalling books gets encoded in the subsequent hawking radiation. The key theoretical problem is that quantum gravity is not a quantum field theory in the usual sense. We have to make sense of that conundrum before tackling any specific problems about real-life objects.

String theory does have something to say about quantum gravity on a conceptual level, which is where the current puzzles are, beyond just GR. It doesn't have anything to say about particle physics beyond quantum field theory, although it does provide useful ways of thinking about various featres quantum field theory that are hard to notice or work with otherwise.

For black holes, there's a bunch of explicit calculations you can do that allow you to count the microstates of a black hole and get exact agreement with the thermodynamic black hole entropy derived by Hawking in the '70s. Beyond that, there are some excellent efforts in constructing these microstates explicitly (e.g. the fuzzball proposal). Most of the more recent progress in understanding black holes doesn't use string theory explicitly, but it uses general notions (holographic principle) for which the only explicit examples are known in a stringy context.

Inflation and Dark Energy are roughly in the same category to my mind. In both cases the main puzzle for string theorists is how to obtain a compactification that has an accelerated expansion in the non-compact directions (i.e. deSitter space). This turns out to be hard, with lots of no-go theorems limiting the range of ingredients that can do the job. However there are constructions which are believed to get the job done (most notably the KKLT scenario), or certainly get you most of the way there with some slightly less studied effects invoked at the last step. This story again runs into the embarrassment of riches in terms of the number of possible compactifications.

Once again, I must stress that at the present moment the problems with quantum gravity are of a conceptual nature, not of a "working out the details as applied to the real world" nature, so expectations and requirements must be tailored accordingly. If this was electromagnetism we were talking about, we would be at the stage where we came up with the notion of fields and are trying to derive Maxwell's equations, except we don't even have experiments to guide us. And we're certainly not at the stage where we can hope to understand, say, geomagnetism or radio antennas.
Our universe is most certainly unique... it's the only one that string theory doesn't describe.

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Sat Feb 20, 2016 7:13 pm UTC

aph wrote:The only way we can trust a science (that the theories will help us in achieving a goal) is if its predictions are checked in experiment against reality. In that sense, string theory is not yet 'trustworthy'.


Right, that's not different than what I've been saying or the point of the article above. We certainly don't trust string theory the same way we trust electromagnetism, and I think in this thread we're in agreement that we would all disagree with someone who purported otherwise. However, the fact that we don't 'trust' string theory in the same way we 'trust' electromagnetism is NOT sufficient grounds for saying string theory isn't science.

@Tchebu

Clearly I'm not understanding things correctly because I can't get simple answer to what I am trying to make to be simple questions.

My point is thinking very very basically about string theor and the development of physics in general.

There are some unexplained phenomena that humans have observed. Think things like dark energy or something. We've observed something and our models don't explain it, or we've observed something in contradiction to our models. There are also scenarios that humans HAVEN'T observed but that we are trying to think about. Think black hole information paradoxes. If we observed something about them than it wouldn't really be in contradiction to our "best" model because we don't really yet have a best model.

Say we had string theory worked out in the way that we are thinking it is going to work out... Would string theory make predictions about any of these as yet unexplained phenomena? And not just the examples I've listed but also the unlisted examples on the edges of theoretical physics that I'm (me personally, not the researchers) not smart enough to know or think about right now..

Like if the answer to the question in the above is "it's complicated" or "no" then I am starting to see a reason why we should doubt whether string theory is science. Maybe it is more like math? Maybe what you're trying to say is that string theory is to some nice theory of nature as vector calculus is to electromagnetism.

edit: and if that is the case then I would call string theory math, or mathematical physics. And then there's the question of is mathematical physics a science? Well it's clearly kind of part of science but you're obviously going to run into trouble drawing a demarcation line at that point and you're also getting to the point where it doesn't matter (unless you're trying to convince funding agencies of something or something..)

edit2:
Tchebu wrote:Once again, I must stress that at the present moment the problems with quantum gravity are of a conceptual nature, not of a "working out the details as applied to the real world" nature, so expectations and requirements must be tailored accordingly. If this was electromagnetism we were talking about, we would be at the stage where we came up with the notion of fields and are trying to derive Maxwell's equations, except we don't even have experiments to guide us. And we're certainly not at the stage where we can hope to understand, say, geomagnetism or radio antennas.


I mean yeah, I understand this distinction. String theory is in early stages. But my question is: When string theory gets to later stages is it eventually going to lead to telling us details about the real world.
You say there's no experiments to guide us, but clearly there is something guiding us(/you?) otherwise string theory wouldn't have been come up with in the first place.. presumably if ALL observed phenomena were well described by some model we had we would stop doing theoretical physics. Maybe we would start pushing technologies and limits to see if we could find something that our theory wouldn't explain but I am open to the idea of some sort of an endpoint of reductionist theoretical physics.

aph
Posts: 296
Joined: Tue Nov 12, 2013 7:48 am UTC

Re: Should we consider string theory to be science? (Philoso

Postby aph » Sat Feb 20, 2016 11:42 pm UTC

You want to find if there is any way of knowing if a theory if worth pursuing further, given current track record?

Tchebu
Posts: 564
Joined: Thu Feb 22, 2007 12:42 am UTC
Location: Montreal

Re: Should we consider string theory to be science? (Philoso

Postby Tchebu » Sun Feb 21, 2016 3:16 am UTC

"String theory" has basically the same status as "quantum field theory" in that it's a mathematical framework/toolkit. Within that framework you can have a multitude of specific models. In QFT these models are specified by the field content and the coupling constants between them at some particular energy scale, in string theory they are specified by the choice of compactification.

In principle, once we find a compactification (or family of compactifications) of string theory that reduces to the standard model it automatically brings with it a spectrum of superpartners, kaluza-klein states (particles that have momentum along the compact directions) and higher mode string oscillations, which you can then in principle look for using your nearest solar-system-sized particle accelerator. Measuring the properties of these states will then give you details about the compactification. A stringy construction should also tell you the specific mechanism by which dark energy arises, rather than the current "derp, just some vacuum energy density, because it fits in the equations". Specific compactifications predict specific values of the cosmological constant, so measuring it also tells you more about which compactification we're in. This is very much analogous to how you can write down the general form of the standard model and then use measurements to fill in the free parameters.

From there it depends on how you want to split things. Are the triumphs of particle physics a success of quantum field theory in general or of the standard model specifically with QFT being "math"? The answer in the case of string theory should be analogous.

Personally, I think the fact that the standard model works speaks to both the merits of the SM as a particular model but also of the entire framework of QFT because it turns out to be adequate to describe these phenomena and qualifies both of those as "science". I wouldn't say that QFT is "math" because the real "math" that people in math departments do is even more general and abstract even if it's motivated by problems in physics, though I certainly sympathize with people who feel that studying general properties of QFT is more detached from reality than studying the SM and extentions thereof specifically.
Our universe is most certainly unique... it's the only one that string theory doesn't describe.

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Sun Feb 21, 2016 10:50 pm UTC

Tchebu wrote:"String theory" has basically the same status as "quantum field theory" in that it's a mathematical framework/toolkit. Within that framework you can have a multitude of specific models. In QFT these models are specified by the field content and the coupling constants between them at some particular energy scale, in string theory they are specified by the choice of compactification.

In principle, once we find a compactification (or family of compactifications) of string theory that reduces to the standard model it automatically brings with it a spectrum of superpartners, kaluza-klein states (particles that have momentum along the compact directions) and higher mode string oscillations, which you can then in principle look for using your nearest solar-system-sized particle accelerator. Measuring the properties of these states will then give you details about the compactification. A stringy construction should also tell you the specific mechanism by which dark energy arises, rather than the current "derp, just some vacuum energy density, because it fits in the equations". Specific compactifications predict specific values of the cosmological constant, so measuring it also tells you more about which compactification we're in. This is very much analogous to how you can write down the general form of the standard model and then use measurements to fill in the free parameters.


Ok, so based on these two paragraphs, and in particular the first paragraph, the answer to my questions is a definitive YES. String theory CAN be about the physical universe. And it can also make predictions. Work still needs to be done before those predictions can be made specific enough for an actual experiment to test but that is nothing new.. Maybe the timescale for that theory-experiment linkage is long but in principle this is how it ALWAYS is and has been with theory..

So ok. I think I'm going to walk away with this as my final answer for now and just decide that all of the debate people are having is just because they're confused about what string theory/science does?

User avatar
drachefly
Posts: 197
Joined: Thu Apr 23, 2009 3:25 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby drachefly » Tue Feb 23, 2016 10:02 pm UTC

Twistar wrote:) It was interesting to me that the philosophers of science said the cutting edge model was this Bayesian confirmation. That seems stupid to me. They're trying to put a quantitative number on something that is not quantitative. And you can use anything for your priors.. I don't know.


I find it very funny that you're so down on Bayes when Lakatos is simplified Bayes.

As for priors, the point of gathering evidence is to make them as nearly irrelevant as possible. Having a prior off by six extra orders of magnitude against a theory takes you from wanting, say, a three sigma rejection of the null to wanting six sigma.

User avatar
doogly
Dr. The Juggernaut of Touching Himself
Posts: 5532
Joined: Mon Oct 23, 2006 2:31 am UTC
Location: Lexington, MA
Contact:

Re: Should we consider string theory to be science? (Philoso

Postby doogly » Wed Feb 24, 2016 3:24 am UTC

I say go Feyerabend. It's the sassiest.
LE4dGOLEM: What's a Doug?
Noc: A larval Doogly. They grow the tail and stinger upon reaching adulthood.

Keep waggling your butt brows Brothers.
Or; Is that your eye butthairs?

Twistar
Posts: 445
Joined: Sat May 09, 2009 11:39 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby Twistar » Wed Feb 24, 2016 4:28 pm UTC

drachefly wrote:
Twistar wrote:) It was interesting to me that the philosophers of science said the cutting edge model was this Bayesian confirmation. That seems stupid to me. They're trying to put a quantitative number on something that is not quantitative. And you can use anything for your priors.. I don't know.


I find it very funny that you're so down on Bayes when Lakatos is simplified Bayes.

As for priors, the point of gathering evidence is to make them as nearly irrelevant as possible. Having a prior off by six extra orders of magnitude against a theory takes you from wanting, say, a three sigma rejection of the null to wanting six sigma.


I should probably learn more about Bayesianism before I criticize it, but from what I know I would say, rather, that Bayes (when applied to the philosophy of science) is obfuscated Lakatos.

It seems like with Bayes you're trying to put numerical quantification on something that can't be numerically quantified. Maybe you can jump through all these hoops you're talking about of doing the analysis to mitigate sensitivity to the priors, but at that point what you've done probably amounts to the same thing you would have come up with with Lakatos. Beyond that, I guess there's 2 reasons I dislike the Bayesian approach to philosophy of science.

1) Is there an example you can think of or that has actually happened in reality when a meta-Bayesian analysis of different scientific proposals has allowed scientists to go with one model rather than another when thinking about it from Lakatos' perspective wouldn't have?

2) And this is what really bothers me about it: It allows people like Dawid to make arguments that string theory is correct and then go around telling other people they're correct and they have this bolstered sense of confidence that they are correct because they have "proven it" with numbers. The point is, a Bayesian analysis is not a rigorous proof. It FEELS like one, but it is not one. We have already shown that there is no way to logically prove anything in science. Bayesianism seems like a step backward in that sense because Bayesianists are trying to prove one theory is right over another. Maybe they're not, maybe they're just trying to prove that we should follow one model over another, but if that's the case they shouldn't get antsy when people decide to follow a different model.

So my argument against it boils down to: It is trying to quantify something that shouldn't be quantified, and the process of doing so obfuscates the lines of thinking so that misguided ideas can be bolstered.

This confusion with string theory is EXACTLY the type of thing that makes the Bayesian approach seem obnoxious. After Doogly and Tchebu's responses in teaching me a little bit about the details of string theory it seems very obvious to me (from a Lakatos perspective) that string theory is most definitely science in the same way any physics theories have been science in the past, yet, many pages and hours of confused discussion seem to have arrisen as a result of people trying to think in the Bayesian perspective.



This isn't to say the Bayesian perspective doesn't have any merit. If theorists want to do Bayesian analysis to determine which theories to pursue they should by all means go for it. That's the thing about theory, there is no rigorous way to determine what direction you should go so if a theorist wants to follow Bayes that is his or her prerogative. A good theorist should have a good sense of which direction to go even in absence of any other guiding light and if they think Bayes will lead them to a good place then go for it!

User avatar
drachefly
Posts: 197
Joined: Thu Apr 23, 2009 3:25 pm UTC

Re: Should we consider string theory to be science? (Philoso

Postby drachefly » Tue Mar 01, 2016 12:08 am UTC

As for point 1 - when you zoom out, you can't do all the math, so you just end up with Lakatos. It's like the difference between QM or GR and Newton. Fundamentally, it's all the complicated thing. At the high level, it's the simple thing. It is nice to know how the complicated thing works.

2 - Bayes tells you how to shift your probabilities in light of evidence, and reaching a probability of 1 on a model is a thing it specifically doesn't allow you to do. So anyone who uses Bayes to prove that some model is correct is doing it wrong, and this is blatantly obvious from the math.

That won't stop people from saying that THEY'RE RIGHT it, though. Nothing can do that.


Return to “Science”

Who is online

Users browsing this forum: No registered users and 9 guests