New Scientist today has a feature article headlined
How to think about… The multiverse
The idea of an infinite multitude of universes is forced on us by physics.
It starts off quoting Sean Carroll:
“One of the most common misconceptions is that the multiverse is a hypothesis,” says Sean Carroll at the California Institute of Technology in Pasadena. In fact, it is forced upon us.”It is a prediction of theories we have good reason to think are correct.”
The problem with this claim is that it’s simply not true. There is no model that “we have good reason to think [is] correct” that predicts a multiverse of universes with different physics (i.e. fundamental constants). I’ve written about this many times, see for instance Theorists Without a Theory. In case you were thinking of interpreting Carroll’s claim in some other way, the article goes on to invoke Alexander Vilenkin:
“The so-called constants of nature, like the mass of the electron or Newton’s gravitational constant, will have different values in different bubbles,”
To be fair to New Scientist, I haven’t read beyond the headline and first few paragraphs of this article, since the rest is behind a paywall. Maybe the later part of the article (which most people can’t read) explains what is wrong with the Carroll and Vilenkin claims.
For the latest on the models supposed to give us different physics in different parts of the multiverse, you might want to take a look at this new paper on the arXiv, and Cumrun Vafa’s talk about it this week at Strings 2018. The paper and talk conjecture that the supposed metastable dS solutions of the string landscape don’t really exist (they are in the “swampland” of things that aren’t solutions of string theory). If this is true and you want to save string theory, as Vafa explains, you need to invoke different sorts of supposed solutions to string theory, with the CC replaced by a “quintessence” mechanism.
At the end of the talk (1:01), Eva Silverstein tries to explain what is wrong with Vafa’s arguments. He responds “I’m not saying you’re wrong, you might be right, this might be also be right”. This shows clearly the fundamental problem of the subject: there is no well-defined theory here, just a bunch of conjectures about what one might be, with no way to tell whether Vafa or Silverstein is right, and no way to extract well-defined predictions from the mass of possible conjectures.
Update: Thanks to those who sent me a copy of the full New Scientist article. It’s short, and the part behind the paywall is even worse than the part publicly available, just adding to the confusion by invoking “many-worlds”, with more from Sean Carroll. Our doppelgangers doing exciting stuff in other universes make an appearance, although Carroll expresses a lack of interest in what they’re up to.
Update: The Strings 2018 talks and videos are available here or at this Youtube channel, and 4 gravitons has a blog posting. As usual with Strings 20XX conferences, very little about actual string theory there. For an overview of the state of the field, you might want to watch this video of the 50 years of string theory session, moderated by David Gross. Dan Harlow was the only speaker raising the elephant in the room question: “is string theory still a useful candidate as a theory of HEP physics?” (and also asked whether they should finally rename the conference series). Gross read off submitted questions for the panel, most of which were asking about the elephant in the room. The panelists each found a different way of avoiding dealing with the question. Other questions asked about the hot “is there a dS string vacuum?” issue, responses were “maybe yes, maybe no”, with no indication of any way to resolve this.
I give an alternative view in a post earlier this week to Prospect magazine’s science blog. https://www.prospectmagazine.co.uk/science-and-technology/the-problem-with-multiverse-theories-theyre-just-not-science
What I really don’t understand is what theorists like Sean Carroll think they’re gaining by holding these views.
I think it’s very clear what Carroll is gaining here: he has written a whole book (From Eternity to Here) claiming to explain a problem about time via different physics in different universes. Most people think this is not science, so he needs to convince them that it is science, that the multiverse is “forced on us by science”.
What harder to understand is why he never responds to arguments explaining the problems with this, just ignores them, only responding to straw man arguments. Even harder to understand is why journalists put his claims front and center, ignoring the problems with them. As I said, I haven’t seen the rest of this piece, but typically pieces like this are almost entirely quotes from multiverse proponents, with little if anything from people pointing out the problems.
“One of the most common misconceptions is that the multiverse is a hypothesis,” says Sean Carroll
Well, I can actually agree with Sean Carroll for once! As the start of the Wikipedia page on “Hypothesis” says: “For a hypothesis to be a scientific hypothesis, the scientific method requires that one can test it.”
So, I’m totally in agreement with Sean Carroll – it fails the criteria for a hypothesis. Dunno what you’d call it, really. An “idea”? A “bit of fun”? Don’t care, really.
“What [is] harder to understand is why he never responds to arguments explaining the problems with this, just ignores them, only responding to straw man arguments.”
Isn’t cognitive bias the obvious explanation?
If we already believe something to be true, we tend to evaluate new input, including critical questions, in terms of how it confirms our beliefs, and if we believe it not to be true, we look for ways the new input helps us avoid having to believe it.
I think this can give rise to situations in which people talk past each other.
What is odd here is not that “cognitive bias” is a factor, but that well-known arguments are completely ignored. This is not the way academia is supposed to work (you aren’t supposed to get to ignore arguments you don’t like) and it’s not the way journalism is supposed to work (journalists are supposed to be aware of and present arguments of both sides).
I find it curious the Agrawal-Obied-Steinhardt-Vafa paper you link to starts with three “naturality” assumptions, there called Criterion 1, Criterion 2 and an unlabelled definition, that certain quantities (Delta, c, and alpha, resp.) are all ~O(1). Also, the paper doesn’t exactly shout it from the rooftops that certain things that have been held to be true about string theory just don’t seem to work. My conception of how sciences work is that if you make a real advance that seems to turn all accepted wisdom/best working current theory on its head, then it’s seen as a big thing, you make a lot of noise, and (sadly) put out a press release trumpeting how everything we knew was wrong, and we have a much better idea now…
I subscribe to New Scientist and the whole mutliverse article is only a few paragraphs long. It is one of 13 articles in the issue that start with “How to think about …” Besides MV there is “Time”, “Genes”, “Consciousness”, “Particles”.
The MV article brings up eternal inflation, string theory as our “best stab at a ToE” and “Then there is the quantum multiverse, predicted by the “many worlds” interpretation of quantum theory”
No contrary opinions are given…
The problem is that the new “swampland” paper doesn’t show that well-advertised claims about string theory don’t work, it just says “maybe they don’t”, without a lot of justification. Not enough for a press release…
Thanks. Some people did send me a copy, I added some comments to the posting.
if they aren’t saying that kind of thing, then maybe they should be. Urs Schreiber is now highlighting that there are serious problems turning up up for the KKLT approach, see eg the quotes in comments on this G+ post, taken from Danielsson-Van Riet, arXiv:1804.01120. Here’s a sample:
Note the citations of work in pure QFT, which is surely the kind of thing that should make anyone sit up and take notice, regardless on one’s position on string theory (I have no skin in the game: ST is just interesting mathematics to me).
The problem with this whole subject is that basically what people are doing is arguing about whether or not solutions exist to a set of unknown equations, based on a very complex web of guesses and information about limits of these equations. You then end up with the kind of thing you quote, arguments invoking “not unthinkable” and “the most natural assumption”. All that’s clear here is that the technical issues are very complex and murky. My best guess from all I’ve seen is that there’s no reason to believe there is a well-defined problem here, all evidence is with Vafa’s “maybe you’re right, maybe I’m right”.
I also don’t think it matters who is right: either infinitely complicated metastable dS string vacua exist, making string theory useless for predictions, or they don’t, and you don’t have solutions to work with, which is equally useless.
I’ve wasted far too much time following these discussion for nearly 15 years, see
and put up with a lot of personal insults from string theorists who take the attitude that my refusal to go farther down this rabbit hole proves that I’m not competent and shouldn’t be listened to. For some recent examples, see for instance
This illustrates well what has always been a major problem with string theory research: the constructions are so complex and ill-understood that they provide lots of intricate and challenging technical problems to work on. People get deeply involved in this and a whole field of research is generated, with practitioners tending to take the attitude that anyone dismissing work on these constructions should not be listened to because of their lack of expertise.
We’re now 30-some years and tens of thousands of papers down the road from the first attempts to construct a “string vacuum” that looked like the real world. You don’t need to be an expert in all the intricacies of every attempted construction to notice that the output of all this activity is no explanation of anything about the real world, just a bunch of excuses for not being able to get such a thing.
After writing that last comment, took a look at tonight’s new hep-th papers on the arXiv. At least two of them are about the “are there dS solutions” question. This is now officially the latest fad of a faddish subject.
Lifted from the comments on your blog, April 15, 2004
serenus zeitblom says:
April 15, 2004 at 4:07 am
Peter, you should look at http://arxiv.org/abs/hep-th/0404075
It’s clear that the claim that string
theory can produce “anything” just isn’t
true. In particular, it is very far from
clear that it can produce a deSitter
background. The KKLT proposal essentially
suggested that you could get dS out of string
theory if you were willing to tolerate extremely
contrived models involving known gadgets such
as fluxes. Now it seems likely that even
this is not true.
Or in the words of Danielsson-VanRiet 18 p. 4:
I remember reading that and noticing that it’s completely untrue. This critic of string theory has never agreed that “the landscape exists and we more or less know its properties.” I have no idea whether the unknown equations of string theory have the claimed metastable dS solutions, suspect it’s not even a well-defined question. What I’ve always argued is that it doesn’t matter: whether you have too many solutions or none, you can’t say anything about the real world.
Yes, nothing ever changes on this issue, as for many others about string theory.
I added some links and comment re: the Strings 2018 conference that just ended.
We’ve started a discussion on this topic at my ‘Cosmology and Astrophysics News’ Facebook group (that’s GROUP, not PAGE :). I’m taking the same case as the author of this page, others are not…
All welcome to join in, oh and I post Cosmo-astro news every day as do some others…never a dull moment 😀
Just wondering that you wonder about Sean Carroll, et al.
We’re living now with alternative facts – so alternative science is the logical result.