The Existence of Nothing

The sold-out “debate” held Wednesday night here in New York is now available for viewing online, see here. I just watched most of it, and one of many things I couldn’t figure out is what if any propositions were being debated. Lots and lots of the usual multiverse mania, and endless flights of speculative fancy and empty, meaningless argumentation. I’d guess this left much of the audience thinking there’s not much difference between what well-known scientists do and what stoned college students do when they’re talking late at night. Amongst all this, a few topics stood out as completely missing:

  • Any significant discussion of what our best theories really say about the vacuum. There are all sorts of interesting things you could say about the vacuum of the Standard Model QFT, but no one seemed interested in this topic.
  • Any legitimate connection to experimental test. The Planck results to be released the next day were referred to by Eva Silverstein, who claimed that CMB observations could test the sort of thing she was talking about. In actuality, there seems to be zero prospect that Planck or any other such observations will test the speculative ideas about string cosmology she was referring to.
  • Any indication that the multiverse and string theory are not settled science that all physicists now agree on. The problems with this picture of the world went completely unmentioned as far as I could tell.
  • Any mention of the disinvited David Albert.
This entry was posted in Multiverse Mania. Bookmark the permalink.

34 Responses to The Existence of Nothing

  1. P says:

    “In actuality, there seems to be zero prospect that Planck or any other such observations will test the speculative ideas about string cosmology she was referring to.”

    This isn’t true. Like anything, it is model dependent, but there are string cosmology models which predict a large scalar to tensor ratio, e.g. axion monodromy. I believe the predictions for r in these models are large enough that they could be ruled out by Planck.

  2. Peter Woit says:


    I don’t doubt that you can come up with “string cosmology” models that “predict” just about anything imaginable. People are free to do this and go on stage and tell the public about how this is an inspirational example of science in action, but I think this is quite misleading.

  3. P says:


    This is precisely how model building – in QFT or in string theory – works; there are many choices, and indeed a wide variety of things can be predicted (though not “anything imaginable”, as you say). This is true in both particle physics and cosmology.

    Inflation needs a scalar field, and certain types of scalar fields appear in string theory. In considering one as a candidate inflaton, predictions are made and these models can be ruled out as experiments become more precise. This could be true of axion monodromy at Planck. Such ideas are perfectly honest and are how science / model building works, in both QFT and string theory.


  4. pred says:

    I suppose the key criterion with the notion of ‘predictions’ is falsifiable tests. What falsifiable statements do the predictions (e.g. of string cosmology) make?

  5. Peter Woit says:


    The question of whether string theory satisfies the criterion to be science has been debated endlessly here (and elsewhere). If you look into this, you’ll see that falsifiability by itself cannot be such a criterion. For a trivial example analogous to P’s axion monodromy, I can tell you that I have a (rather complicated…) model of the early universe, which predicts that Steven Hawking’s initials will be visible in the polarization data inside the large cold spot that Planck has found. This is a testable prediction (which will be falsified next year…), but if I were to spend my time coming up with a detailed model designed to do this and studying it, you might sensibly consider what I’m doing to not be science.

  6. pred says:

    I saw the SH initials (was it the WMAP data?). Most ‘predictions’ are actually retrofits to data, i.e. you tell me the answer and I’ll demonstrate why my theory fits the facts perfectly. It gets embarrassing if the facts change, because someone discovered a calibration error etc., but most retrofit theories are able to backtrack on things like that. There has to be more than one falsifiable prediction, and they all have to come from the same model (no fiddling with endless adjustable parameters).

  7. Peter Woit says:


    The SH was found in WMAP I think.

    Yes, typically people try and justify their models by fitting them to facts. The problem I’m describing is a bit different though. The question of whether something is science has to do with whether it is giving non-trivial, reliable information about the physical world. Being able to make accurate falsifiable predictions is one characteristic of successful science. But my point is that just because you make a falsifiable prediction, that doesn’t by itself mean you are doing science. Astrologers make falsifiable predictions all the time…

  8. Paul Wraight says:

    I thought Luke Barnes comments in his blog conveyed a lot of sense.

  9. Z says:

    I wish they had discussed the question of if Mathematics exists and the notion of logical consistency (do axioms exist? What about incompleteness theorems?), since presumably, any deep fundamental theory must ultimately arise from some mathematical structure Some of these issues go back to Plato’s time, but I think they still have relevance to these musings.

  10. cormac says:

    There was a series of seminars on the results in Cambridge today. Very impressive; it’s now SWH, by the way!

  11. cormac says:

    Sorry, that was a bit obscure. What I mean is this: in relation to the comments above on the initials SH being imprinted on the cosmic microwave background, as measured by WMAP, Paul Shellard joked at a seminar yesterday that one can now read the initials SWH in the PLANCK spectrum. It was a fun way to finish his lecture, but I couldn’t see really it!

  12. Maciej says:


    I think you are saying that falsifiability is not a sufficient criterium to say that a theory is scientific. Clearly. However it is essential and ST does not even fullfill this.


  13. Peter Woit says:


    Yes. But the question of string theory itself is complicated by not even knowing what the theory is. The multiverse, as being promoted now at events like this one, is the ultimate in unfalsifiabilty.

  14. Maciej says:

    I see.

    Well in that case (i.e. that we do not even know what ST is) it is even easier to claim that ST is not scientific since we don’t know what we are talking about.

  15. dark says:

    Is the multiverse a way of avoiding the conclusion that string/m-theory predicts the wrong dimensionality (4 vs 10/11)?

  16. Peter Woit says:

    String/m-theory doesn’t predict any particular observable number of dimensions. It doesn’t really predict anything, and the multiverse pseudo-science is just being used as a way to avoid doing honest science and drawing the standard conclusion that a theory that predicts nothing has to be abandoned.

  17. dark says:

    doesn’t gauge anomaly cancellations require 26 in bosonic, 10 or 11 in string/M-theory?
    hence, a “prediction” if wrong.

  18. Peter Woit says:

    The standard story is that you can use compactification or “branes” to get solutions to the 10/11 d theory that have other numbers of large dimensions at observable energies (e.g. 4).

    Perhaps if string/M-theory were more precisely defined, you could simply show it made an incorrect prediction. The current state of the art is that with what is known about the possible definition of such a theory, you can get pretty much anything you want. String theorists used to argue that once string theory was better understood, one would find that it had certain specific solutions and you would get predictions. The newer ideology is that once string theory is better understood, it will have a landscape of solutions and not predict anything.

    This is getting off topic, and it’s by now a very old story, enough about this.

  19. Nick M. says:

    Peter Woit says:

    “But the question of string theory itself is complicated by not even knowing what the theory is.”

    Thank You so much, Peter, for this comment. When I was an undergraduate at UC Berkeley in the early to mid 2000′s, I remember asking a GSI friend of mine, who was working on String/M-Theory at the time for his doctoral thesis, what he knew of M-Theory, and could he tell me more of what it was all about. He put his index finger to his lips, and whispered in a quite and a somewhat whimsical voice, “Shhhhhh…, the big secret is, is that there really is no M-Theory”. Although he tried to explain what he meant by this at the time, it’s only over the intervening years that I have come to appreciate what exactly it was that he was trying to get across.

    Now per your admonishment, I will say no more on this topic.

  20. Tammie de Cortez Haynes says:

    Dear Dr. Woit

    Your account of this episode was most depressing.
    Pardon me while I make it more so.

    You thought the audience was left thinking there’s not much difference between what well-known scientists do and what stoned college students do.
    I doubt that.

    I suspect the audience was left admiring their own superiority, as among the few smart enough to share in the enlightenment of our greatest scientists.

  21. socrates says:

    Do people actually feel superior for being in the audience of anything? That’s stoned enough for me.

  22. P says:


    Sorry for the delay. Busy weekend.

    The falsifiable test for axion monodromy is a higher than usual scalar to tensor ratio, usually written as r in the literature. I forget the precise numbers, but they’re in the literature and I can look them up if you’d like.

    Make no mistake about it: there are experiments to measure r, axion monodromy predicts that it is larger than many models (including slow roll, I believe), and this is absolutely falsifiable as experiments bound the value.


  23. Bernhard says:


    Is this prediction unique to string cosmology models?

  24. mat noir says:

    Was it Gellman who said: Given 3 variables, I can make an elephant. With four, I can make it walk.

  25. aleph/eleph says:

    John von Neumann apparently, quoted from Enrico Fermi, and it was 4 and 5, not 3 and 4:
    “I remember my friend Johnny von Neumann used to say, ‘with four parameters I can fit an elephant and with five I can make him wiggle his trunk.”

    But Dorothy Parker said it better
    “I like to have a martini,
    Two at the very most.
    After three I’m under the table,
    after four I’m under my host.”

  26. P says:


    No. I believe, in fact, that before the axion monodromy mechanism was understood people thought one couldn’t get large scalar to tensor ratios out of string cosmology.

  27. piscator says:


    I think many people still think that.

  28. P says:


    You may be right. I’m not enough of a string cosmology expert to know if or why there is a debate about axion monodromy as a model.

  29. Randioactive says:

    What of AMS-02?

  30. Jim Akerlund says:

    This is off topic, but FQXi has just put up a new contest for 2013 called “It from Bit or Bit from It”. It runs until 6/28/13.

  31. emile says:


    Please find details and poster related to the SPECIAL CERN-EP Seminar on Wednesday 3rd of April 2013 in Main Auditorium (500/1-001) at 17h00:

    “Recent results from the AMS experiment”
    by Prof. Samuel TING (Massachusetts Inst. Of Technology (US))

  32. Beelzebud says:

    I tried to watch that debate. Watching educated people sit around debating on the definition of nothing is as fun as watching paint dry. As a layman I just don’t quite see what the entire topic has to do with science. I can see why philosophers would discuss it, but what has it got to do with science?

    The only insightful part of it was when the gentleman with the fabulously colored suit explained his model about branching universes, but even then, how does science test something like that?

  33. Raizonator says:

    I watched the debate and there is one nagging question that came to my mind:
    When we talk about our Universe, we mean the ONE (classical) Universe we observe. O.K. But if we extrapolate back in time until the Universe is as big as an elementary particle and QFT “kicks in” then one is forced to speak of a universe being in a “one particle state” (for example). That is the number of universes we are talking about gets uncertain (Heisenberg).
    So with a certain probability it could also be in a zero-particle state, i.e. in a state of “nothingness” if one likes. But what does that mean anyway if nobody (i.e. no classical observer) can be around to do the measurement and find out how many quantum universes there are? That’s where things really get weird, I think, and where I can’t trust extrapolations via (conventional) QFT any more.
    Are these thoughts too naive ?

  34. MS says:

    @Beelzebud: It is a business (not science) thing.