Today’s Hype

Hype about string theory and fundamental physics seemed to have been dying down recently, with only three editions here so far in 2019 of This Week’s Hype. Today however brings a bumper crop of the highest quality, with new examples from a few of the world’s most prominent theoretical physicists. Today’s hype neatly exemplifies the two main current genres of hype about string theory and supposed new fundamental physics. The first is the old-school genre of string theory hype we’ve now been seeing for 35 years: “string theory makes a testable prediction” (no, it doesn’t). The second is the new, post-modern variety: no actual theory, just a grandiose claim that space and time have been replaced, although it’s unclear by what.

Trend: Cosmic Predictions from the String Swampland

In the subtitle of this APS Physics piece from Cumrun Vafa, we learn that the string swampland has “led to testable predictions about dark energy.” If you read the article trying to find the testable predictions of string theory, you’ll get to Figure 3, where the caption says “The colored curves are string theory predictions for dark energy for different values of c.”

The problems with this include:

  • This is based not on a theory or calculation, but on a conjecture (see e.g. here) that consistent string vacua have certain properties.
  • Many experts disagree with this conjecture. In particular it would imply that the well-known supposed metastable string vacua of KKLT are inconsistent, and that is a matter of controversy.
  • This conjecture is not the sort of thing that can be proved one way or another, since it is not about something well-defined. The non-perturbative formulation of string/M-theory necessary to get a well-defined answer to such questions about string vacua remains unknown. You can make various conjectures about the behavior of this unknown theory, but then your swampland conjecture is a conjecture about a conjecture.
  • The conjecture involves an unknown constant “c”. Unless you know what “c” is, you don’t actually have a prediction.

In his conclusion, Vafa writes:

In the next 5–10 years, we may know, for example, whether dark energy is constant or not. If it is, that could pose a serious blow to string theory. But if dark energy is found to be changing, could that observation be the first experimental evidence for ideas emanating from string theory?

The most likely possibility over the next 5-10 years is that measurements continue to be compatible with constant dark energy. I don’t believe for a minute that 10 years from now after that result is in you will see Vafa or anyone else giving up on string theory or even admitting it has suffered a “serious blow”. On the other hand, if there is any evidence for a varying dark energy, Vafa or others will surely claim it as “evidence for string theory”, which it will not actually be.

Over the last 15 years I’ve often written here about this “Swampland philosophy”, which never made much sense to me. I didn’t understand back in 2005 and still don’t understand now why conjectures that behavior you don’t observe in the real world might be inconsistent with some other conjectures about an unknown M-theory are supposed to be of interest. The sociological motivation here is rather clear though: the KKLT-based “anthropic landscape” philosophy has not worked out well for the field, and the hope is to disentangle the subject from that morass. A good explanation of what is going on is provided by this (stolen from Will Kinney, who also has a lot to say about the whole swampland business):

Cosmic Triangles Open a Window to the Origin of Time

This is a different sort of hype, with no direct relation to string theory. In this genre of hype, you don’t have any connection between your calculation and either experiment or a fundamental theory, but this doesn’t stop you from making grandiose claims. What’s behind this particular article is this paper, which develops a nice calculational method exploiting conformal symmetry. What’s not made clear in the Quanta article is that this has no connection to anything measurable. As the authors of the article explain:

In this paper, we have worked under the lamppost of weakly broken conformal symmetry. This has allowed us to derive particularly clean insights into the analytic structure of inflationary correlators. However, it also restricts the strength of the couplings between the inflaton and additional massive fields. This makes the observational challenge to detect these effects enormous.

In other words, this is about speculative models in which the observable effects described at length in the Quanta article would be unmeasurably small.

The post-modern hype come into play with the argument that these conformal-symmetry based calculations somehow tell us how to replace space and time.

This suggests that the temporal version of the cosmological origin story may be an illusion. Time can be seen as an “emergent” dimension, a kind of hologram springing from the universe’s spatial correlations, which themselves seem to come from basic symmetries. In short, the approach has the potential to help explain why time began, and why it might end. As Arkani-Hamed put it, “The thing that we’re bootstrapping is time itself.”

If you’re trying to understand the origin of this particular dollop of hype, it’s a good idea to keep in mind something Arkani-Hamed said at a talk about Lance Dixon’s work back in 2013:

… I AM an ideologue. In my defense at least I can say that I’m a serial ideologue, in the sense that I’ll take totally different ideologies and drop the last one without thinking about it, but it’s very important for me personally to be an ideologue when I’m working on something…
So, usually I’ll get up when I talk about scattering amplitudes and give a long introduction about how spacetime is doomed, we have to find some way of thinking about quantum field theory without local evolution in space time and maybe even without a Hilbert space and blah-blah-blah. This is all very high-falutin stuff, this is stuff that Lance wouldn’t be get caught dead saying. I think none of these guys would ever say something that sounds so pretentious, but I have to say it, you know I have to say it, because this is the only way I can get up in the morning, and like “I suck again, OK, here we go, I’m doing it because spacetime is doomed, I swear to God, right”.

I’m actually rather sympathetic to the “bootstrap” philosophy in some general sense, which I’d interpret as “all is unitary representation theory of the conformal group”, i.e. that constraints of conformal symmetry, analyticity and unitarity are almost enough to determine a fundamental theory. The 1960s version of this, trying to get strong interaction physics purely from such general principles, didn’t work out, but I’ve become more and more convinced that representation theory and fundamental quantum field theory are very deeply intertwined. I do think though that to get anywhere you’re going to need to either work top-down (i.e have an actual fundamental theory and derive its implications) or bottom up (i.e. use observations to find the route to better theory).

Will head out soon for a short vacation in Northern California (if it hasn’t burned to the ground). I had thought about stopping by the 2020 Breakthrough Prize Symposium, but decided listening to talks about why supersymmetry/supergravity got a $3 million award would not be good for my blood pressure.

Update: Those interested in KKLT vs. Swampland debates might be interested in the latest from Tom Banks. He makes a detailed case against KKLT/eternal inflation/Landscape models, which string theory Swampland enthusiasts may find appealing. They should however note what he has to say about string theory itself:

A more sensible attitude, which I share, is to accept that string theory defines some models of quantum gravity, but obviously not the one that corresponds to the real world.

This entry was posted in Swampland, This Week's Hype. Bookmark the permalink.

13 Responses to Today’s Hype

  1. S says:

    Curious — when you say you’re sympathetic to bootstrapping, does that include the idea that “the thing that we’re bootstrapping is time itself”? Or are there versions of this that are less exotic? (I haven’t known you to be a big friend to emergentist pictures, so I’m guessing so, but am out of my depth.)

  2. Thomas Van Riet says:

    “The non-perturbative formulation of string/M-theory necessary to get a well-defined answer to such questions about string vacua remains unknown. ” Not sure. If there are regions of string moduli space where we can do computations (weak coupling regimes and their duals) then we can try to uncover patterns in those regions. These patterns can then have deeper meanings, even relate to thermodynamics, black hole thought experiments, etc. If so one can expect these patterns to be there in the full string theory. Then we use holography to check for further non-perturbative evidence. That is the Swampland. Being against that would be against understanding patterns. Sure, some conjectures are quite radical and imply that “established” things (KKLT) are not that established. But that we knew. Check for instance McAllisters’ talk at Strings 2019. He admits at the very end that the separate ingredients of KKLT might seem under control (not my opinion) but it remains to figure out whether the collections of ingredients can coexist. So no. Nothing was established. Neither a landscape and neither a very constraining Swampland. But a Swampland is clearly there. And it is exciting to see what can be proven more rigorously in the already-complicated “lamppost region”. So let us carry on and enjoy it. It is the most healthy hype sofar. Kudos.

  3. Luca says:

    It’s a bit sad to see this kind of hype coming from Quanta magazine. At least it is based on a very interesting theoretical work though

  4. Peter Woit says:

    People have different motivations and meanings for “bootstrap”, but in general it involves the idea of not looking for a fundamental Lagrangian, but instead trying to characterize as much as possible a theory by general principles including typically symmetry, analyticity, unitarity. It you look at what these principles are, to a large degree they’re equivalent to saying you want some sort of highest weight (which gives positive energy, analyticity), unitary (unitarity), representation of a space-time symmetry group (conformal group or at least Poincare group).

    Where I part company with bootstrappers is that I don’t think one should assume that the fundamental theory is some complicated, inaccessible thing (e.g. string theory/landscape), with simple representation theory behavior only emerging in a low energy limit. If you find a good representation theory picture that realizes the hopes of the bootstrap program and describes the real world, I think it will actually be a fundamental theory (and you probably can find a Lagrangian version of it if you want).

    The “we’re bootstrapping time itself” business is just meaningless promotional blah-blah.

  5. Peter Woit says:

    Thomas Van Riet,

    If you want, you can carefully examine an area directly under a lamppost and then guess that the rest of the world behaves in the same way. The problem is that if you then start publicly claiming “predictions” based on this guess, you need to make clear to people that all it is is a guess. And also, that when this guess is shown to be wrong, you have no intention of abandoning your theory, all you will do is say “OK, looks like things aren’t the same elsewhere as under the lamppost”.

    Both the swampland and landscape programs have exactly this problem. The arguments about the consistency of metastable vacua like KKLT also take place close to an unphysical lamppost, and are useless for making any legitimate scientific predictions about what the conjectural M-theory predicts about the real world.

  6. Peter Woit says:


    Most of the Quanta article is fine. It should though have made clear how far this is from connecting to any observations. As a general principle, any quotes from Arkani-Hamed about revolutionary changes in our understanding of space and time should come with a warning label about the large renormalization factor that needs to be applied to such claims.

  7. Will Kinney says:

    Glad you liked the meme Peter. 🙂

  8. Peter Woit says:


    I thought it was very enlightening as an explanation of the current swampland/KKLT argument. Thought for a minute about asking for your permission to use it, but decided “why do that, this is the internet, I’ll just steal it.”

  9. DRLunsford says:

    How does the conformal group come into the referenced paper ( That is, what was the motivation for this?

    The problem with a paper like this is that it is too vague.

    BTW there are interesting cosmic effects just from conformal kinematics, e.g. the idea of a cosmic rest frame, and by implication a way to distinguish it (e.g. the CMB). Amazingly, the Hubble relation itself comes right out this idea.


  10. “all is unitary representation theory of the conformal group”

    Well, not quite all. Non-unitary representations are pretty important too, if (rather frustratingly) still very poorly understood. Critical 2D percolation, the plateau transition of the quantum Hall effect, self-avoiding walks are all supposedly logarithmic conformal field theories still waiting for a consistent synthesis as far as I am aware.

    I might wager that something like quantum gravity should also suffer from signatures of “non-unitarity,” i.e. multifractality, turbulence, etc. Why should “spacetime foam” be anything less complex than classical 2D percolation is my question. It would seem to me that the same issues with nonlocality could naturally arise…

  11. Bernhard says:

    “A more sensible attitude, which I share, is to accept that string theory defines some models of quantum gravity, but obviously not the one that corresponds to the real world.”

    But is even that really true? I remember reading an article on the arXiv ( several years ago which argued that “string theory addresses some problems of quantum gravity” but cannot yet be considered a theory of a quantum gravity, even an inconsistent with Nature one.

  12. Thomas Larsson says:

    Matthew Foster,
    AFAIU unitarity is necessary for locality. Geometric phenomena like percolation and SAWs (I know nothing about the QHE) are defined globally, even if they are built from many small building blocks. Percolation is a path between two distant points, and a SAW defines the mass of a linear polymer inside a large sphere, and both definitions involve large-distance properties. Both phenomena can be embedded into families of graph models, a discrete set of which corresponds to unitary reps of the Virasoro algebra with local correlation functions, but percolation or SAWs themselves are inherently non-local.

    So if you believe that quantum gravity is a local theory, you probably need unitarity.

Comments are closed.