Friday, June 08, 2018

Physicist concludes there are no laws of physics.

It was June 4th, 2018, when Robbert Dijkgraaf, director of the world-renown Princeton Institute for Advanded Study, announced his breakthrough insight. After decades of investigating string theory, Dijkgraaf has concluded that there are no laws of physics.

Guess that’s it, then, folks. Don’t forget to turn off the light when you close the lab door for the last time.

Dijkgraaf knows what he is talking about. “Once you have understood the language of which [the cosmos] is written, it is extremely elegant, beautiful, powerful and natural,” he explained already in 2015, “The universe wants to be understood and that’s why we are all driven by this urge to find an all-encompassing theory.”

This urge has driven Dijkgraaf and many of his colleagues to pursue string theory, which they originally hoped would give rise to the unique theory of everything. That didn’t quite pan out though, and not surprisingly so: The idea of a unique theory is a non-starter. Whether a theory is unique or not depends on what you require of it. The more requirements, the better specified the theory. And whether a theory is the unique one that fulfils these or those requirements tells you nothing about whether it actually correctly describes nature.

But in the last two decades it has become popular in the foundations of physics to no longer require a theory to describe our universe. Without that requirement, then, theories contain an infinite number of universes that are collectively referred to as “the multiverse.” Theorists like this idea because having fewer requirements makes their theory simpler, and thus more beautiful. The resulting theory then uniquely explains nothing.

Of course if you have a theory with a multiverse and want to describe our universe, you have to add back the requirements you discarded earlier. That’s why no one who actually works with data ever starts with a multiverse – it’s utterly useless; Occam’s razor safely shaves it off. The multiverse doesn’t gain you anything, except possibly the ability to make headlines and speak of “paradigm changes.”

In string theory in particular, to describe our universe we’d need to specify just what happens with the additional dimensions of space that the theory needs. String theorists don’t like to make this specification because they don’t know how to make it. So instead they say that since string theory offers so many options for how to make a universe, one of them will probably fit the bill. And maybe one day they will find a meta-law that selects our universe.

Maybe they will. But until then the rest of us will work under the assumption that there are laws of physics. So far, it’s worked quite well, thank you very much.

If you want to know more about what bizarre ideas theoretical physicists have lately come up with, read my book “Lost in Math.”


Clara, once known as Nemo said...


this means that Dijkgraf thinks that

- energy is not conserved
- electric charge is not conserved
- the speed of light is not a limit
- gravity can be repulsive

Dijkgraf should get 4 Nobel Prizes for these 4 statements. I am sure that others can provide more statements and more Nobel Prizes for him. He deserves them. Can we organize a collective movement to get this outstanding physicist the prizes he deserves - maybe via twitter? The Nobel comittee could start this october with the first!


hiwa ahmed said...

His claim resembles pseudoscience rather than pure theory .

marten said...

I am an optimist and I have the feeling that Robbert Dijkgraaf is indicating that he is prepared to be sacrified as Landscapegoat.

senanindya said...

Why stop with string theory ?
Let's go one step further and join Maz Tegmark.
Every bit of math describes some universe or another so don't worry, we are sure to find ours if we keep doing math long enough.
What, you don't believe me ?
Prove me wrong, I dare you. 😂

akidbelle said...


I think I understand what you mean; I also think it can be worded with more elegance, beauty, and naturalness:

There is no laws of physics, and this is a natural consequence of quantum theory.

Proof: Obviously, string theory is is a huge superposition state and in order to know which version is correct you have to observe "The" (or "A") universe. So, eventually, as long as you do not observe "The" (or "A") universe there is no laws of physics.


I do not understand why it would be worth a Nobel prize (though there is only one Nobel measurement per year and in the meantime I also have it :)


Michael Sharples said...

So is this the "only game in town" playing its final hand?

It's not a very good one is it?

Rob van Son (Not a physicist, just an amateur) said...


I think you are too harsh. It is not that Robert Dijkgraaf does not believe in empirical science.

Robert Dijkgraaf regularly appears on television trying to teach the audience useful physics. In these shows he stays pretty close to the "canon" of physics. Nothing out of the normal. In this respect, he does not abolish the "Laws of Nature".
Here is one such appearance about black holes (mind you, this is prime time TV on a national channel):

There is an old philosophical discussion about what it means to have a "Law of Nature". My impression is that Robert Dijkgraaf let himself get carried away into a philosophical discussion going from string theory into pure mathematics. Somewhere along that route he lost the concept of a "Law of Nature".

I am convinced that when he is back at work, he will use the same laws as everyone else.

Sabine Hossenfelder said...

Rob van Son,

I'm sure he is a great guy and so on, but if he publicly goes and proclaims nonsense I'll public go and proclaim it's nonsense.

Mark Martens said...

Like a child whose favorite toy has just been taken away, you expression is perfect.

Honza K. said...

Maybe the professor wanted to say something like "There no laws in the nature, that fits into our models". Must be really hard to realize that 20 years of research have no ending in "theory of everything". Moreover if great "minds", best scientists are involved. As one guy in ESO videos used to say "nature surprises us beyond our imaginations". And that's the best part of it.

Neil Rickert said...

For several years, I have been saying that there are no laws of nature. However, there could be no doubt that there are laws of physics.

Perhaps I should add my reason. From attempting to understand human cognition, it becomes clear that there is no method by which we could discover laws of nature. The best we could hope to do is make up our own laws such as would enable us to better understand the world that we experience.

Sagittarius A said...

" is extremely elegant, beautiful, powerful and natural,” yeah my dope too ;-D I'm only a layperson and taxpayer and I wonder why do we pay all these eggheads :-/

Udi Fuchs said...

Did any of you read the article? The title says that "There Are No Laws of Physics." I guess that is suppose to trigger your curiosity and make you read the article.

If you read the article, you will see that all that he is saying is that there are no fundamental laws of nature that can determine the 20 or so free parameters in the standard model of particle physics.

Initially there was hope that String Theory would fix these parameters. This doesn't seem to be the case. Instead there appear to be a landscape of models allowed by string theory. Quantum Field Theory also allows for a landscape of models. One of these models is the standard model that describes our world. Having to fit these parameters to experiment does not mean that the theory is not predictive.

I didn't find anything controversial in Dijkgraaf's article.

Y'all should be more careful before you make fun of someone. Otherwise, the joke is on you.

Matthew Rapaport said...

This a very good point! Often the editors choose the title for its shock value and the author's actual claims are reasonable in their proper context.

I do not know this physicist and have not read the article, but thanks for pointing this out.

Sabine Hossenfelder said...


Yes, I read it, thanks for asking, see my comments about string theory being unique and such. Nothing controversial about the multiverse, no, nothing controversial whatsoever.

Georg said...

“The universe wants to be understood and that’s why we are all driven by this urge to find an all-encompassing theory.”

we have to decide whether this field should named wanting-field or
urgity. The quanta of that field would then be named wantons
or urgons.


Udi Fuchs said...

The word "multiverse" doesn't appear in Dijkgraaf's article.

Physicists spent about 5 decades studying the landscape of Quantum Field Theories before converging into the standard model.

I don't know if String Theory describes nature. But if one wants to research this topic, then studying this landscape seems a reasonable thing to do.

There might be other people that made controversial statement about this subject. But beyond the title, I don't see here nothing controversial whatsoever.

Uncle Al said...

Robbert Dijkgraaf's fears are well-founded.

1) Organic molecules passing through a grating diffract (DOI 10.1002/prop.201600025)
2) Any attempt to localize the process collapses it.
3) Hund's paradox, DOI:10.1103/PhysRevLett.103.023202
4) A resolved chiral molecular beam must racemize when diffracted, or not diffract.
5) Nanogram sample enantiomeric excess is quantifiable to large differential sensitivity,

6) Some of quantum mechanics’ beauty is demonstrably WRONG.

The only defense is not looking - the stock and trade of contemporary theory.

Sabine Hossenfelder said...


The string theory landscape is a type of multiverse. Yeah, right, QFT also has a landscape and hence a multiverse and look what we did there. We just picked the one that fits the data and no paradigms had to be changed or shifted.

Peloton said...

My limited understanding of this maelstrom is that there is general confusion between the concepts of reality, laws, and models of nature. It is ok to have multiple models, even models that overlap in part or whole and agree with testable reality. These models may lead to more understanding and models with increasing coverage of the natural world. However a model that is so flexible that it could describe many different realities is NOT getting us closer to the goal of understanding THE one and true physical reality of the cosmos. If the models are increasing in coverage and leading to laws of nature that correspond to a theory of THE reality then that is progress and it is perfectly OK to have many non contradictory models to cover different portions of reality. Does this make any sense?

sean s. said...

I'm just going to lurk in the back ground here and watch the conversation. ... But on second thought; I have a question: what exactly would a "law of physics" be? (Assuming one exists.) I always thought the "laws of nature/physics/chemistry/etc. were simply descriptions of what we observe or what we expect to observe.

Is that not correct?

sean s.

Nate said...


I think you may be a bit confused: when the string theorist talk about the landscape they're not talking about the space of possible string theories, they are talking about an assertion that every possible string theory literally exists. This would be like saying that not only does the standard model of QFT exist, but the QFT's associated to every possible combination of constants/gauge groups/matter fields also exists in a very strong and literal sense: as in we could find points in the "universe" where these other theories govern the laws of physics, not the standard model.

I suspect that you think the string theorist are making a completely reasonable philosophical shift, really just making a move about what they consider in their ontology. But you have to understand they are literally saying that all possible versions of physics tangibly exist. That is what they are referencing when they talk about "the landscape"

Ken Mitchell said...

Many "laws of physics" are observations of how the universe behaves rather than fundamental principles. I suspect that some of our "laws of physics" may eventually be proven to be true only for local conditions. For example, can we be absolutely certain that properties in deep space, far away from any star's gravity, will behave in exactly the same way?

Keep an open mind, and be prepared for a few surprises.

Ian Miller said...

There are two possible explanations for the claim. The first is the claimant cannot see any laws of physics, so he denies them. The second, and more reasonable, is that he is criticising the word "law". The problem arises in that a number of "laws" are found to have exceptions, e.g. Einstein showed that Newton's laws have to be amended. Apart from that possibility, the statement is rather silly, and I am not going to turn off the light.

Mitchell said...

Udi Fuchs said

"all that he is saying is that there are no fundamental laws of nature that can determine the 20 or so free parameters in the standard model of particle physics"

Then why does he talk about duality so much? No, he's saying that the landscape itself is more fundamental than physical laws, "because" the laws are always part of a specific string theoretic framework and each such framework only covers part of the landscape.

You could instead look for a new kind of description that is uniformly valid across the entire landscape. See his final paragraph: "Perhaps there is a radical new framework..." But it seems like less than one reader in ten understood his message.

Q said...

Ideologies that have no basis in reality find ways to dismiss reality itself. When Marxists were thoroughly defeated on the battlefield of ideas they found a way to dismiss logic altogether by inventing polylogism. Seems like this is the direction string theory is headed.

tytung said...

A multiverse supporter should not believe that our universe have beautiful laws, as the number of universes with ugly laws most probably far outnumber universes with beautiful, simple laws. Right?

nemo said...

Lost in string...

Udi Fuchs said...

Sabine wrote:

"We just picked the one that fits the data and no paradigms had to be changed or shifted."

The paradigm shift that Dijkgraaf is talking about is that you don't need to pick one model, because all models are connected. Today we assume that quarks fundamentally come in 3 colors. We could decide to use 7 colors instead, and if we understood the landscape well enough we would be able to explain all existing experiments with this new model. (Mitchell, This is the duality that he is talking about.)

The downside of using 7 colors would be that the calculations would be way more complicated. The upside is that it could give us new insight that might help us explain other experiments or device new experiments that we didn't consider before.

Udi Fuchs said...


I am not confused at all. I am fully aware that the landscape and the multiverse got a lot of bad press. Some string theories make public statements that surely deserve this bad press. This paper by Dijkgraaf is not one of them. Read it and judge it on its own merit, you might learn something.

jim_h said...

"Everybody loves a winner." "Nothing succeeds like success." So far, string theory hasn't won, and everyone is piling on, ridiculing the whole concept, hey it's a party. But if the LHC finds something next year, a lot of people are going to wake up with a sort of intellectual hangover. Just my uneducated layman's view.

Sabine Hossenfelder said...


Great, then let's wait until a string theorist calculates the proton mass by building it up from seven quarks.

Lawrence Crowell said...

I think the criticism that the string landscape can't predict the exact outcome of our quantum field world is similar to saying biological evolution is wrong because it does not predict the occurrence of every species. Evolution tells us there are related characteristics between species, clades of genomes as we now know, but it does not predict the evolution of a squid or a nematode as something determined. The landscape is similar, and far more vast in its set of possible configurations. Quantum gravity in general, not just superstring theory, will have to be tested by observations of gravitational waves and to look for possible quantum effects there. Without something along those lines all of quantum gravity, not just string theory, is untestable and represent the end of science.

S.E. Grimm said...

I read the article of Robert Dijkgraaf and my interpretation is:
The “landscape” is the underlying field-continuum that creates phenomenological reality. This in contrast to phenomenological physics that is keeping on to relate phenomena like particles and forces to each other (standard model). If I wipe out all Robert’s statements about string-theory and related concepts the resulting text is state of the art quantum physics. His statement about the laws of nature (and I suppose he included the constants too) is not ridicule. In mathematical physics “physic laws and constants” are thought to emerge from the mathematical model we use. In other words, the article of Robert Dijkgraaf describes a (supposed) promising future of mathematical physics because of recent insights.

Udi Fuchs said...


I certainly agree with you that if string theorists calculate the proton mass it would be great. This would require a better understanding of string theory, which this article is exactly about.

My problem is with your unwarranted attack on this article. You claim that you don't like group-think, but this is how it feels to me -- it is popular to bash string theory so lets bash this article about string theory, who cares what it actually says.

tytung said...


Evolution is not a good analogy for string landscape, as there are fundamental differences.

1) Evolution can indeed be used to predict which species will survive (even what kind of biological properties will occur) under a given circumstances, via its central principle "survival of the fittest". But string landscape does not do that at all.

2) For biological evolution, you can manipulate the organisms and their genes to see what will happen. You can't do that with universes in a string landscape.
The former endeavour is active while the latter is entirely passive.

3) String theory giving rise to landscape is entirely ad hoc and you can have any predictions, which means it is not falsifiable. Evolutionary theory is not, it is actually falsifiable (despite what you might have read about its falsifiability).

4) Evolution explains the amazing species diversity within one simple framework. String landscape has to make up a diversity of universes to explain one universe.

Sabine Hossenfelder said...


Of course I too am subject to social reinforcement, but I don't know how this is relevant here as I am not making an argument ad populum. I am also not "bashing string theory", as you claim. I am saying that the expectation that you can derive a theory that is unique without drawing on sufficient amounts of observational constraints is nonsense. If you use an insufficient amount of constraints, you get an ambiguous theory. There's nothing new about this, and it's not specific to string theory. New is only that now that's supposedly a deep insight. That's my issue with the article. It's logical nonsense.

Dualities are not specific to string theory. As to your comparison with evolution, well, show me evidence for the evolution tree of our universe then.

Anatol Sevashko said...

S.E. Grimm. We see the need for a new mathematical physics. Albert Einstein saw a solid foundation in a thermodynamic arrow, but he did not use this foundation, he did not have the mathematics for this.

We need to create new mathematics based on the demands of physicists. We can localize mathematics for localize material reality as a stable part of the information field.

Lawrence Crowell said...

@ Tytung,

In the end using one scientific theory as an analogy for another is going to have problems. No two theories have the same structure. I was thinking along the lines of Stephen J, Gould when he said that one could not predict on the basis of what ever early conditions existed on Earth that in so many billions of years there would be the panoply of life we observe. Evolution is not an entirely predictive theory in that sense. I was also thinking of evolution in a classical sense prior to the techniques of genetic manipulation. I might have compared the situation with planetary dynamics, in that we know the configuration of a planetary system is based heavily on initial conditions that are not themselves predicted by Newton's second law with his elementary force of gravitation. This was a problem that Kepler had; he thought the configurations of the planets was grounded in some hard principle, where orbits were configured according to the circumscription of Platonic solids.

The basic idea of the landscape comes from the number of possible topologies for Calabi-Yau manifolds and the number of configurations that D-branes may be wrapped. The estimate of the number of such topologies is 10^{500} to 10^{1000}. We can work out the details of compactifications such as the torus compactification, but this results in a low energy physical vacuum far larger than what exists in this observable world. We are then left with a bit of a serious conundrum on how to work through this immense number of possible topologies.

It is possible (here I am making a “what if” sort of statement that might cause trouble) there is some extremal principle that selects the minimal vacuum configurations for such a landscape. Maybe larger energy physical vacuum configurations for CY/brane wrappings are not stable, so these are then by some means selected out in some form of superselection rule on the grand path integral or partition function over the landscape. Maybe this will reduce the number of physically realizable cosmologies in the landscape or multiverse, the two bearing some relationship, to a much smaller number or maybe for that matter just one. Maybe then all these other cosmologies are sort of off shell configurations. As yet such a principle does not exist.

We will not likely ever have Planck scale physics in the laboratory. This is one problem with quantum gravity, whether it is string theory based or LQG or dynamic triangulation or other hypotheses. I suspect the only way to test quantum gravity is to look for signatures of graviton physics in gravitational waves. In this way we let nature do the heavy lifting. In the near horizon configuration of colliding black holes the vacuum between two horizons or holographic screens just prior and during coalescence may emit gravitons in response to the change in “quantum hair” or if you prefer Mathur's “fuzz ball physics” on these horizons.

Sabine Hossenfelder said...

A lot of people have tried to post comments to this thread with information about their self-made theories of everything. I therefore want to recommend that all newcomers read the comment rules.

I have comment moderation turned on which means that a submitted comment will not appear until I manually approve it, and I will not approve off-topic comments or self-advertisements. I also strongly discourage you post links to sites other than major news outlets or journal pages. The reason is not that I generally block other links, but that I will only approve them after I have inspected the sites they link to, which I might just never come around to do.

Thanks for listening!

Isaac Rabinovitch said...

When you talk about the Multiverse being useless as a starting point, you make me think of the Mandela Effect, which basically posits that nobody can ever be proven wrong about anything. Cuz somewhere in the Multiverse what they think happened did happen!

mls said...

So, how does one become lost in math? In Chapter 51 of "The Principles of Mathematics" Bertrand Russell explains 'being' and 'existence'. He writes, "Existence, on the contrary, is the prerogative of some only among beings". Russell had actually been a critic of the received view of formalism -- the view leading to an uncountable number of possibilities for natural numbers and a general rejection of the mathematics used by physcists as "not logical". That would refer to the "second-order" nature of infinitesimals. Of course, mathematical physics is also inclined to accept Leibniz' identity of indiscernibles in the sense by which one of the axioms for a metric space is comparable. This ,too, is "not logical".

It should surprise no one that people who "believe" in mathematics are finding themselves in this situation. Replacing an anthropomorphized delusion with a sterile, inanimate delusion does not change the delusional nature of the claims.

In Russell's defense, his position had been motivated by consideration of forms called "negative existentials". The formalists, however, pusued the ideal of a logical calculus, with respect to which words have no meaning and the primitive elements of language cannot be defined. The physcist who envies the "freedom" of the mathematician has little understanding of the nonsenical views promoted in the mathematics community because of logicians and philosophers. Metaphysical beliefs about geometry in the nineteenth century created a false "crisi" leading to arithmetizatio of mathematics and logicism

Seriously? Shall we accept a logic having existential import forcing us to believe in the natural numbers as being substantial outside of space and time? In case you are unaware, those who pursue logics of fictions are not given any respect in the mathematical logic community.

Those who promote this "received view" would hold that I am misrepresenting the situation. I am not. If one is using words like "truth" and "existence" as technical jargon without clearly presenting this fact at the outset,the result can only be a mass confusion. And, it has been going on for generations.

I look forward to reading your book when I can afford a copy. Bit, do not think that this has not been going on in mathematics for a very long time.

senanindya said...

What confuses me about string theorists is that, instead of blathering about multiverses and paradigm shifts, why don't they just say "String theory is a framework which allows many possible theories (depending on details of dimension compactification and what not). We are trying to find one such theory which describes the universe we observe ?"
It seems like this accurately describes a lot of situations in physics from QFT to Lagrangian formulations - you pick a particular theory or Lagrangian, work out the equations and see if they fit the data. Nobody tries to argue that every possible Lagrangian describes some universe somewhere.
Is it the case that string theorists oversold their mission of finding "one unique theory to describe everything" and are doubling down despite the evidence of non-uniqueness.

Udi Fuchs said...

Sabrine said:

"I am saying that the expectation that you can derive a theory that is unique without drawing on sufficient amounts of observational constraints is nonsense."

I tries to find what are referring to in Dijkgraaf's article and I assume that it this:

"All numbers in nature should be determined by physics itself. They are no “constants of nature,” only variables that are fixed by equations (perhaps intractably complicated ones)."

If I understand it correctly, he seems to have a view point closer to the 80's when people thought that string theory would yield a unique solution. This is exactly the opposite extreme of the multiverse.

I do not agree with him, but I cannot prove him wrong.

I do not know how you can determine that it is nonsense. We already know that in Quantum Field Theories, there are very stringent constraints for a theory to be fundamentally consistent. I do expect that these constraints would be stronger in String Theory.

Udi Fuchs said...

senanindya said...

"What confuses me about string theorists is that, instead of blathering about multiverses..."

Most string theorists don't care about the multiverse at all. A quick search for "multiverse" in the arxiv shows a few papers a month. Most of of them are not in hep-th and therefore most probably not written by string theorists.

The latest paper in hep-th mentioning the word "multiverse" in the title or abstract is from July 2017 and it is co-authored by Stephen W. Hawking, who is (was) not a string theories.

There is a paper from June 2017 by some prominent string theories including Brian Greene, which is again a physicist more known for his popular books than for his scientific work (coinsidence? I don't think so.) I don't think that I can describe my opinion of this paper in a polite way, so I will use the words of the owner of this blog and just call it nonsense.

sean s. said...


I do not agree with him, but I cannot prove him wrong. I do not know how you can determine that it is nonsense.

You don’t have to prove him wrong; you don’t have to show it’s nonsense. What ever Dijkgraaf claims Dijkgraaf must prove correct.

He really can't.

sean s.