Hacker News new | comments | ask | show | jobs | submit login
What every physicist should know about string theory (scitation.org)
44 points by paulpauper 7 days ago | hide | past | web | favorite | 47 comments

What I find most interesting about these articles showing up on Hacker News is how discussion of abstruse details irrelevant to the bulk of progress in physics or mathematics gets almost all the air time.

Compare the number of articles on string theory and physics beyond the standard model to those on exotic materials, advances in turbulence, orbital mechanics, modern elasticity, solar system formation, or foundations of quantum mechanics. Is it that there is no tradition of popular writing that makes people think that they have any insight beyond the wall of knowledge necessary to understand these fields?

String theory and things similar to it evolved to be as intellectually interesting as possible. Normally you have a mixture of humdrum "experiment theory" and exciting abstractions, but without the humdrum experiments to motivate all the humdrum calculations only the most attractive parts get studied. As a result, string theory is 99% mind-blowing, while the rest of the items on the list are 1% mind-blowing and 99% "numerical simulation of heat transfer in a honeycomb lattice." Nobody would ever write a boring string theory paper because there is no possible motivation for doing so. This is not to say that string theorists are somehow neglecting a duty, just that there is almost certainly a vast mine of untapped results that are not presently being worked on because there is little motivation for them.

One thousands years from now when humanity is building star-bridges out of beyond-standard-model particles, I guarantee that there will be a Journal of Taychon Engineering whose articles never make it on PBS.

Seems to me one solution would be for you and others who think like you to regularly post links on HN to articles on these topics you believe are being unjustly neglected.

... and see them being completely ignored. (That was the whole point, I think.)

Where are the articles on Boyce–Codd normal form, Java bytecode, SOLID etc..

This is still a 'social' network.

I don't think I've heard of BCNF for a while, thanks for the mention. I doubt anyone uses more than 1NF. I've posted it here: https://news.ycombinator.com/item?id=19140206

I would be concerned if an engineer I worked with put forward a DB design that wasn’t normalized to 3NF without a really good reason. Designs that are 1NF but not 3NF are generally either hyper-optimized (better have a really good reason) or just bad. It’s not hard to get a DB into 3NF.

I remember having to learn how to normalize my data a lot when I was learning SQL, because the only tools available for use in my environment was a good ole' tool called Microsoft Access & VBA.

String theory is the buzziest of science buzzwords.

It's just the wrong audience altogether.

Peter Woit's critique of this article: http://www.math.columbia.edu/~woit/wordpress/?p=8068.

I've never found that Peter Woit's critiques of string theory match up very well to the reality of what string theorists are doing.

Woit's critiques are of string theory as an extension of the Standard Model. But most string theorists are not working on string theory trying to get predictions about current energy frontiers. Rather, they are working on understanding the mathematical structure of string theory and seeing what it can teach us about analogous mathematical structures in quantum field theory. (A now classic example: The first calculable models of 4-dimensional confinement came out of string theory.)

Should be called string hypothesis.

I would accept theory as the right word. A mathematical structure that has been assembled for testing is generally referred by practitioners to as a model or, if there is a collection of models and mathematics linking them, a theory.

Unfortunately, a fully fleshed out theory with centuries of matching experimental data is still called a theory.

I think that is more of a namespace collision than anything else.

Even in mathematics, there are different definitions of "theory". One is that a theory is a body of knowledge, like set theory or automata theory. This is where I would group string theory as well.

Another definition from mathematical logic is that a theory is a set of axioms and deduction rules. Fascinatingly, those have in common with scientific theories that they can (if sufficiently complex) only ever be falsified, not verified.

I believe the point (excuse the pun) is that it may well be a mathematical theory but it's not (yet?) a scientific one.

Did any physics enthusiasts read this paper? https://arxiv.org/abs/1001.0785

"Every" is a bit rich, considering we're likely 100s of years from testing it directly.

It isn’t clear that any amount of time would lead to it being falsifiable.

It has to agree with every prior measurement in Quantum Mechanics and Relativity. That, in itself, is impressive.

It's not so impressive if all you're doing is generalizing frameworks or cherry-picking the subsets that happen to agree with experiment. You have to show something new.

For example, (going from memory here, so don't quote me), your basic (Von Neumann's?) formulation of quantum mechanics was basically a generalization of Newtonian physics that specified an unreasonable number of non-real possibilities, and you can only recover meaningful results by taking Newtonian physics as an axiom a la the 'classical limit'. This would have just been mathematical over-complication had it not been for the fact that this model predicted things that differed from Newton physics in measurable ways.

So to reiterate, if your theory is unfalsifiable, it is not impressive that it agrees with every measurement. All you'd have achieved is restating what is already known in a language nobody wants to use.

It also has developed some interesting mathematical tools and perspectives, but as a TOE I think the growing skepticism is warranted. The worst outcome of course being that M-Theory is accurate, but untestable. It would be a bitter irony for humanity if our best theory required a galactic-scale collider to test it!

It would be a bitter irony for humanity if our best theory required a galactic-scale collider to test it!

Doesn't that also mean that (in such a hypothetical situation) our best theory says that our previously best theories will give accurate predictions up until the point at which we observe something equivalent to a galactic-scale collider? If so, that doesn't sound so bad to me.

The trouble is that strings is more of a framework for physics to fit in to (like quantum mechanics) than it is a physical theory. There isn't a specific list of string particles, just like there isn't a list of Schrodinger equation particles. You have to add in whatever particles you discover, and then the theory tells you how they move.

On the bright side, if there was a single particle that didn't move like a string, string theory would be falsified. On the dim side, it's difficult for anything to not move like a string, because string theory is so powerful as a framework that you can get it to make just about anything consistent. What you really want in physics is a brittle framework that only allows you to slot in a few different behaviors.

It's kind of funny because programmers want the opposite thing that physicists do. When your client asks for a behavior that you have painted yourself into a corner against ever implementing your software has been "falsified," an expensive disaster that you work very hard in advance to prevent! If you made universes for a living, strings would be an ideal framework.

We just need to make contact with a hyper-dimensional sentience who can give us the low down on how everything works.

I don’t think most PhD physicists know what a diffeomorphism is.

I very much doubt that. I certainly learned it in a physics undergraduate degree.

Ok, let me guess without looking it up... A diffeomorphism is a smooth map that lets you switch coordinate systems.

(I sure hope that's not wrong.)

Pretty close. It's a smooth map with smooth inverse. If you've chosen chosen coordinates, it will transform them for you.

I think most physicists probably have a reasonable intuitive idea what a diffeomorphism is, since most have probably learned something about relativity.


I found HN somewhat amusing when it came to physics, now I find it increasingly irritating. I don't think the same arrogant know-it-all attitude coming from laypersons would take place if the topic were neurosurgery or the theory of integro-differential equations. For some reason physics gets the "they don't know what they're doing" treatment by people that haven't bothered to get any formal training on it.

Well I have, and even though I'm not too keen on superstring theory I just couldn't dismiss it like this. It makes sense, it's important and it's a possibility worthy enough to be explored. Will people here just try to understand that they're talking about another profession with a really high skill access barrier that takes many years of hard work to become just acquainted with? Can we please just admit that if we don't know what we're talking about, we really don't know what we're talking about?

> Can we please just admit that if we don't know what we're talking about

Poster you're replying to admits they're a layman. As am I.

Perhaps the people writing articles for the laymen should do a better job explaining the redeeming aspects of String Theory. Because like the OP, I too have been left with the impression (from reading articles targeting laymen) that String Theory is clever math in search of a reality it can describe.

Physics Today isn't for laymen. It's a trade journal, for physicists.

Laypeople can go ahead and express opinions about physics they don't understand, in exactly the same way a layperson can read HN and then go on to express opinions about the merits of Java vs Ruby without ever trying either language. It's a little bit silly, though.

Wait, is Java point particles and Ruby superstrings, with LISP being quantum fields?

Sure! Any two things I don't understand are equivalent to me!

Which it indeed is.

I partially feel the same way. At the same time, there are a lot more people who studied physics commenting on Hacker News than neurosurgeons. So you have people who have passed the skills barrier commenting negatively. Does that encourage people who have not passed the skills barrier to feel safe to comment?

I think one of the frustrations is probably that the tone, and level of nuance and confidence expressed seems to be inversely correlated with level of expertise in the field.

So for example from the current thread:

(someone with training in the field): "It makes sense, it's important and it's a possibility worthy enough to be explored."


(admittedly uninformed person): "As a layman from a distance, [...] it's total garbage."

I like to read about physics and as far as i understand the energies needed to confirm string theory are so far out that we will potentially never achieve them. I also read that a lot of physicists believe that string theorists have lost touch with reality.

Is there any nearer term benefit that may come out of string theory ? I am honestly curious. Somehow I find it hard to believe that so many very intelligent people are chasing something misguided.

I don't think there's any direct near-term benefit that will come out of it, especially of the engineering variety. Our civilization doesn't even make use of the subnuclear physics described by quantum chromodynamics, which is now nearly 40+ years old. It's likely to be a long time before we 'need' correct physical models on shorter distance scales.

That said, we're not devoting the kind of resources to string theory that we devote to engineering ready problems. Probably somewhere in the ballpark of $10M/year worldwide funding provided by various governments, with a similar amount provided by private donors. This is rounding error. In the US, we spend about 10x this on the National Endowment for the Arts.

You can make an argument that all money spent on extensions of the Standard Model (and there are a lot of them besides string theory) has been wasted. But string theory at least has taught us far more about the structure of the mathematical models we use to describe physics than any of its competitors. This is why people continue to study it, even in the absence of experimental confirmation.


I'm just tired of nonsense. I don't understand what's the point in writing op-eds pontificating about things you have no idea, and now they're coming in every day because for some reason many in the HN community think that you can get an education in theoretical physics by reading popsci and blogs.

I know everybody has an opinion about everything under the Sun, me having an opinion about theirs is only natural. I'm just pointing out the fact that laypersons feel increasingly entitled to dismiss not only the professional work of experts, but whole bodies of knowledge available to them but that they can't be bothered to study. As I don't know whether the professional fields of these people allow for winging it, I'm assuming that maybe they think that you can do that in physics. Well you can't, so if you have a problem with some physical theory the proper way is to engage like everybody else, not trying to kill whatever you don't like because how it sounds from the proudest and laziest outside.

This new sense of entitlement from the opinionated laypersons is actually an antiscience stand, this is what gets you anti-vaxxers and shuts down funding of honest research.


People say it’s not falsifiable but I think it’s underappreciated how difficult it is to create a theory where the standard model plus gravity just falls out of it.

The theories are all falsifiable in that they make lots of testable predictions.

The problem is that there is a family of related theories which all produce the same testable predictions at energies we can probe and there’s no way to use an experiment to rule any of them out.

I'm not sure you can say that the standard model and gravity fall out of a theory if an almost infinite amount of other stuff falls out of it too.

Currently the theory "predicts" around 10^500 vacuum models, all of which have different properties.

That's quite a large number, and makes specific predictions rather difficult.

There was a vague hope that maybe something stringy would fall out of the LHC. That didn't happen. There's now an even more vague hope that it might fall out of Son Of LHC instead.

Outsiders might wonder if this is really just throwing string at the wall and hoping some of it sticks.

It hasn't yet, but it's always possible it might.

Physics has always been throwing math at reality and seeing what sticks. So why is it invalid to come up with a more universal mathematical object and seeing how it converges. Sure 10^500 but how how quickly are we throwing away 10^450 after experiments x, y and z? Perhaps they've invented too flexible an object and so it will never settle down, or perhaps they've invented the perfect mathematical object that will slowly but surely settle down on our reality as long as we keep feeding it experiments. A lot of science is discovering what can't be. Whatever is left is the discovery of something new.

> Outsiders might wonder if this is really just throwing string at the wall and hoping some of it sticks.

Isn't that part of it? Smash particles together at increasingly higher energies, see what happens. I don't think we can pretend to predict everything that will happen.

Applications are open for YC Summer 2019

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