Multiverse Politics

The political campaign for the multiverse continues today with a piece by Amanda Gefter at Nautilus. It’s a full-throated salvo from the Linde-Guth side of the multiverse propaganda war they are now waging, with Linde dismissing Steinhardt’s criticism as based on “a total ignorance of what is going on”. All of the quotes for the article are on the pro-multiverse side. There is a new argument from them I’d never heard before: Guth comes up with this one:

You can create a universe from nothing—you can create infinite universes from nothing—as long as they all add up to nothing. Not only is that a deep insight, it also creates a testable prediction. “Eternal inflation certainly predicts that the average density of all conserved quantities should be zero,” Guth says. “So if we ever became convinced that the universe has a nonzero density of electric charge or angular momentum, eternal inflation would no longer be an option.”

The article is subtitled “Why the majority of physicists are on one side of a recent exchange of letters”. One way to interpret this claim is just that 33 is more than 3, but the reason for this is clear: while Guth, Kaiser, Linde and Nomura decided to go on a political campaign, drumming up signatures on their letter, Ijjas, Loeb and Steinhardt didn’t do this, but instead put together a website discussing the scientific issues.

Where the majority of physicists stand on the Guth-Linde claims is an interesting question, one that I don’t think is addressed anywhere by hard numbers. My anecdotal data is that the majority of those I’ve ever talked to about this don’t think the Guth-Linde multiverse claims are science, but don’t see any reason to waste their time arguing with pseudo-science. They hope it will just go away by itself, as it becomes ever clearer that the multiverse is, scientifically, an empty idea.

Unfortunately, I don’t see this going away and I think it’s now doing very serious damage to physics and its public image. There’s a political campaign now being waged, and one side is very determined to win and putting a lot of energy into doing so. Those on the other side need to step up and make themselves heard.

Update: Guth and Linde brought their publicity campaign to New York this past Saturday (video here) where they told a large World Science Festival audience that string theory is beautiful, it predicts the multiverse, inflation has made lots of predictions that have all worked out, and they have (more or less…) the full theory that does all this wonderful stuff. Nothing from them about any thing less than utterly glorious and well-defined about the string theory landscape/eternal inflation product they were pushing. Also on the panel were three philosophers (Jim Holt, David Albert and Barry Loewer) who did an admirable job of trying to push back by pointing out obvious inconsistencies. They at least got Guth and Linde to admit that there was this “measure problem” thing still to be fixed. Physics is in a very weird state indeed now that physicists have adopted untestable metaphysical speculation as their program, with philosophers the ones trying to engage in more normal scientific practice.

There was one multiverse skeptic, George Ellis, who unfortunately didn’t engage with this and was diverted onto other topics. One string theorist was there (Veronika Hubeny), who explained about AdS/CFT duality, without anybody bringing up the fact that this has nothing at all to do with what Guth and Linde were promoting in the rest of the discussion.

This entry was posted in Multiverse Mania. Bookmark the permalink.

36 Responses to Multiverse Politics

  1. Anonymous Professor says:

    The cynical view: As certain people saw their chances at a Nobel Prize slowly ebbing away, they decided to start a propaganda campaign to restore their chances.

    I don’t know whether this is what everybody else is thinking and nobody is saying, or whether I’m just too cynical and completely off the mark here.

    But this is why giving Nobel Prizes only to experimentally confirmed results is a great idea.

  2. Another Anon says:

    I must say, I find something rather disappointing about major scientists trying so hard to defend their pet theories. Aren’t we supposed to be involved in the search for the truth, whatever that might be? Isn’t that why we all got involved in science in the first place, not for personal glorification? If you have a pet theory, and it is disproved, you should be satisfied that it is all part of the process of uncovering the truth of how the universe works. Fighting for your own theory tooth-and-nail seems very low rent.

  3. Don Terndrup says:

    Both inflation and multiverse ideas bother me when the defenders seem to treat these as firm scientific conclusions. Wouldn’t it be more proper to call them both hypotheses, with various levels of supporting empirical and theoretical evidence?

  4. Peter Woit says:

    Anonymous Professor,

    I hadn’t thought of that, but it makes a lot of sense. It seemed to me that Guth/Linde and their supporters were more and more engaged in a serious publicity campaign, to make a two-pronged case that inflation is settled science and the multiverse is conventional science, but I couldn’t see a particularly good reason for this. They may very well feel that Steinhardt’s activities threaten to deny them the Nobel that is rightly theirs, and this would explain the nature of their highly unusual response to the SciAm article (organizing a letter of support from people who might be thought to have influence with the Nobel selection committee).

  5. Adam says:

    The majority of physicists don’t give a damn about this “controversy”. The majority of physicists is not doing theoretical physics, and most of them are doing experiments on subject very far of inflation/cosmology or what not.
    And even among theorists, quite a few of them (me included), are not doing HEP/QG/… Reducing “physicists” to “string theorists” and affiliated is quite tiring, but I guess this is just usual business in pop science…

  6. RP says:

    I am just a mathematician who is somewhat versed in the philosophy of science, but I do wonder: Does it makes sense to classify a statement X as a ‘testable prediction’ if the verification of X depends on scientists, at some point in time, becoming ‘convinced that the universe has a nonzero density of electric charge or angular momentum’? Clearly we are not talking about things which are ‘directly measurable’, in any reasonable sense of the expression.

    The whole idea behind falsifiability is that scientific theories should be answerable to nature in as direct a way as possible, so certainly without the interference of other theories. This is even the vital fact about scientific theories that ensures that they have real content, and are about something: since it is conceivable that they are false, but in fact all our experiments have shown them to be true, they tell us something real about nature.

    If instead of making our theories answerable to nature, we make them answerable to just the conviction of scientists, we lose this essential quality of science. It ceases to tell us something real about nature, and only tells us something real about the ideas of scientists instead. This is exactly what the notions of ‘testability’ and ‘falsifiability’ were designed to avoid.

    In other words, the whole notion of ‘testable’ is being hollowed out if it is allowed to stand in the way Guth is using it. Not just being hollowed out, I would say, but completely subverted. I think it is a terribly sad state of affairs that this type of sophistry is being peddled by a serious scientist writing for a serious public.

  7. atreat says:

    In the linked article, the Linde-Guth side argue that inflation is not really a theory, but rather a “class of models, a sweeping principle, a paradigm” containing a multitude of testable models. As I understand it, Linde-Guth do not disagree with Steinhardt that inflation can support any/all outcomes. Yet, they want to take credit for only those models which have continued to survive nature’s verdict. They say the, “key is to figure out which model of inflation is right,” but why must one of them be right?

    Reading this article and then going back and reading Steinhardt’s “fact checking” page where he shows all these failed paper’s by Linde-Guth and company with “simple models” that have since been experimentally ruled out just leaves me with wondering what explanatory power can Linde-Guth claim of inflationary theory?

    Is modern physics really just about coming up with creative mathematical paradigms describing a whole zoo of potential ways nature could go and then when a new observation is made or a new experiment conducted just shuttering that part of the zoo newly out of sync with nature’s evidence? That seems… so sad. Modern physicists reduced to nature’s ambulance chasers 🙁

  8. Peter Woit says:

    I don’t think the problem with the Guth claim is the “became convinced” part, which he could easily reword in a different way that would avoid what I think bothers you.

    The problem is the nature of the falsifiability claim itself. The universe as a whole is well-known to be electrically neutral, and it’s very hard to imagine an experimental result that would show overall non-neutrality. So much so that I’m pretty sure no one is looking for such a thing. I’m not sure what if any experimental signal there would be for a global angular momentum of the universe, maybe there is such a thing, but as far as I know there isn’t and no one is looking.

    Saying that “my theory is falsifiable because X would be evidence against it” where X is something highly implausible that no one expects to see or is looking for is a standard bogus argument. To see the problem, note that I have a theory that everything that happens is determined by the Jolly Green Giant. This theory is falsifiable, because in principle by observing the CMB, one might see the written message “I’m the Jolly Red Giant, not the Jolly Green Giant”.

    The other problem with Guth’s claim is that, if someone were to observe a global charge or angular momentum of the universe, I suspect that by the next day Linde and others would have a version of an inflationary model which produced universes with a global charge or angular momentum. Personally, if the above mentioned CMB message were found, I would just say that it really is the Jolly Green Giant, it’s just that he’s a Jolly Green Giant who likes to play tricks.

  9. Anonymous Professor says:

    This year’s Nobel will probably go to LIGO. Judging by past behavior, the Nobel won’t go to an astronomy/cosmology result two years in a row, so the next reasonable chance inflation has is 2019. If this is a campaign for the Nobel, it will have to be a prolonged one. So by 2019, we will probably have a good idea whether it is or not.

  10. Jeff M says:


    Jolly Green Giant? I thought it was turtles all the way down.

    Anonymous Prof, while I think you might well be right, I couldn’t see the Nobel committee considering inflation. If that’s what Linde and Guth are after, it’s a false hope. I’m no big fan of the Nobel, but they have been pretty strict about experimental confirmation.

  11. Antimetrics anonymous says:

    Since the Anonymous Professor has raised the issue lobbying for Nobel, it is worth mentioning the other contribution of Prof Paul S.
    Paul Steindhardt wrote an excellent book on ” Quasicrystals” . Paul S contributed a lot to the theoretical basis of quasicrystals. In fact in one interview Dan Schectmann to a TV station told the interviewer whether anybody shared the Nobel prize with him, when he recd the phone call for the prize announcement. It was obvious he was referring to Paul S. Here is paper in Dec 1984 in PRL pf Levine and Paul S where they have cited the paper of Dans S and others.
    Here is the paper of Dan S and others. in November 1984

  12. Louis Wilbur says:

    There are many more critics to the inflationary paradigm than Amanda Gefter’s article claims. But some of the critics of the inflationary paradigm go too far in claiming that it should be completely abandoned and some of its advocates go too far in claiming it should be thought of as a settled science. The truth is between these two extremes and numerous critics and advocates of the inflationary paradigm fail to realize that non-eternal inflation is much easier to realize than had been previously thought. For instance, the Wikipedia article on eternal inflation points out that “A 2014 paper by Kohli and Haslam called into question the viability of the eternal inflation theory, by analyzing Linde’s chaotic inflation theory in which the quantum fluctuations are modeled as Gaussian white noise [arxiv 1408.2249]. They showed that in this popular scenario, eternal inflation in fact cannot be eternal, and the random noise leads to space-time being filled with singularities. This was demonstrated by showing that solutions to the Einstein field equations diverge in a finite time. Their paper therefore concluded that the theory of eternal inflation based on random quantum fluctuations would not be a viable theory, and the resulting existence of a multiverse is still very much an open question that will require much deeper investigation”. The conservation laws are certainly not evidence of eternal inflation and de-Sitter space is also unstable in quantum gravity (e.g. arxiv 1608.07237).

    All of this shows that non-eternal inflation is vastly superior to eternal inflation. Eternal inflation inherits various pathologies of infinity. In an infinite cosmos there is no way to have a coherent measure (e.g. arxiv 1202.3376 and 1211.1347). This fact follows from basic Cantorian set theory. Because there are many ways to establish bijections between an infinite number of events, probabilities can be anything one wants. In these kinds of infinite situations probability questions are meaningless. For example, what fraction of the positive integers are odd? One might think it is 50 percent but it could be 75 percent if the bijection is {2} pairs with {1,3,5}, {4} pairs with {7,9,11}, and so on.

    Non-eternal inflation is predictive, explanatory, has no multi-mess and no measure problems, unlike eternal inflation which has all of these problems. There are many ways to inflate the cosmos without self-reproduction. These include using some type of fluid to do the inflating (e.g. arxiv 1609.04953, 1601.04773 and 1601.05337) and multi-field inflationary models can easily be non-eternal. The weak gravity conjecture prohibits eternal inflation (e.g. arxiv 0707.3471 and 0805.4520) and the string landscape is also inconsistent with eternal inflation (e.g. arxiv 1404.5543, 1609.00385 and 1504.00056). There are also more traditional ways of doing this (e.g. arxiv 1409.2335). Although inflation theory is popular with many early universe cosmologists, it is not a standard in the same way that General Relativity, the standard model of particle physics, and the LCDM model are. The idea of a strong but short lived vacuum energy is plausible and compelling but its details are almost entirely unknown. Cosmic inflation can be thought of as a faster and stronger version of dark energy but the mechanisms of dark energy are almost completely unknown. If dark energy inflation is a sibling of cosmic inflation, perhaps related by some types of seesaw mechanisms, then it is grossly premature to claim that their details are understood in the same way that the details of General Relativity and particle physics are understood. That means, for example, there is no solid basis to claim that inflation leads to a multiverse and it is grossly premature to claim that the mechanisms postulated in traditional inflationary theory are correct explanations of nature.

    There is equivalence between the two major sides in this debate because the cosmological community is not mostly nor entirely on one side in the following sense. Cosmic inflation has been quite popular in the sense that very early universe cosmologists find the idea of a short lived but strong non-zero vacuum energy compelling enough to work on. But the idea is very much under development; its practitioners certainly are not claiming that it is a finished product. So the so-called multiverse is certainly not any kind of standard. In her blog on 12/19/15 Sabine Hossenfelder wrote “But the longer the chain of inference, and the less trust you have in the theories used for inference, the less real objects become. In this layered reality the multiverse is currently at the outer fringes. It’s as unreal as something can be without being plain fantasy”. She also wrote in a different blog that the multiverse is something that researchers within the scientific community spend almost no time on. And the problem of inflation’s initial conditions are only partly solved (e.g. arxiv 1601.01918 and 1506.07306).

    It should also be pointed out that string gas cosmology is a very plausible and compelling scenario that is completely consistent with all cosmological observations (e.g. arxiv 1505.02381). String gas cosmology requires that the primordial power spectra be almost scale invariant with a slight scalar mode red tilt whereas in inflation theory the potential must be chosen such that these properties emerge (e.g. arxiv 1608.05079). And string gas cosmology can explain why there are only three large spatial dimensions: the annihilation process cannot take place in more than three dimensions (e.g. arxiv 1105.3247). But, like inflation, string gas cosmology still has lots of work to be done. And it is also possible that the correct explanation for the very early cosmos is something entirely different from all of these approaches. An experimentalist will criticize all of these approaches in that the hypothesized additional scalar fields do not have their existence empirically established. The quantas of the inflaton fields have not been observed in any particle accelerators nor have they been observed in any cosmic rays and the extra dimensions of string theory have not been empirically verified. So all of these approaches still have a long way to go.

  13. Bee says:

    The problem that nobody seems to want to talk about is that rather than trying to find a minimal model that explains the data and leave it at this, there are many hundreds of models for inflation all of which are almost certainly wrong because they contain too many details that aren’t supported by data. As the philosophers have it, these models are severely underdetermined. (Good paper about this here.)

    Theoretical physicists produce these models literally because they can make money with it. They make money with it by getting them published and then using the publications to claim it’s relevant research so it’ll get funded and they can hire more postdocs to crunch out more papers. It’s the same reason why theorists invent dark matter particles and extensions of the standard model. It’s a way to make a living.

    Steinhardt & co have an issue with this because this overproduction crisis tends to crowd out alternative explanations. I agree that that’s a problem, but that doesn’t mean of course that Steinhardt’s alternative is a better explanation for the data…

    What’s really missing here is a scientific criterion to draw the line and say, look, at this point in time it entirely pointless to produce further variants of speculations because the data isn’t there and won’t be there for decades to come. But nobody in the community has an incentive to come up with such a criterion. That’s because along the line everyone makes money with this overproduction. It’s for this reason I’m putting my hope on philosophers to help us out because I find this situation pretty embarrassing for my discipline.

  14. cosmology grad student says:

    Guth et al are not going to get a Nobel until/unless primordial B-modes are detected. The BICEP fiasco means that it’ll be a long time before another group has a claim even if they’re seeing hints.

    The argument in the article about conserved quantities averaging to zero is dumb. The argument about inflation being able to start for generic initial conditions way overstates the argument of the cited paper. But the discussion of probabilities in a multiverse is pretty much an accurate reproduction of what almost everyone in the community–including Paul until a few years ago–believes. The measure problem is that the obvious choices of measure you might think give results that are far from the predictions of the classical theory. The fact that you need to pick a measure in the first place is just of doing quantum mechanics in a potentially infinite universe, and it doesn’t really bother most people in the field.

    There might be some of the scrambling for a Nobel that you perceive. But I think a lot of this is genuine frustration with Paul (+ his collaborators, but especially Paul) not grasping this fact about the measure problem, and similarly refusing to accept that it’s okay for different inflationary models to make different predictions. (I think Paul has a reasonable point when he says that many simple monomial models, e.g. lamda phi^4 and m^2 phi^2, are disfavored by the lack of observed B-modes and that models like hilltop inflation which give unobservable levels of B-modes seem more contrived. But there’s a big difference between that and arguing that inflation is useless because different models predict different things!)

    A related note–my sense is that many people, probably even a majority, in the HEP/cosmology theory community are sympathetic to the arguments you and Paul, among others, have made against the string landscape. There’s a reason almost no one works on e.g. KKLT or trying to get the Standard Model out of particular compactifications any more. So I think that it’s a major tactical mistake, at the very least, to attack string theory and inflationary cosmology with the same criticisms. The community has concluded, I think mostly correctly, that they don’t hold up against inflation. I suspect that the result of this is that people are going to be much more inclined to discount what you, Paul, and others have to say about string theory.

  15. Haelfix says:

    One of the problems with this discussion is the equivocation of the word ‘multiverse’. It really does mean many different things.
    One of the initial model independent predictions of inflation was that there would be a specific pattern of correlations of the CMB on Superhorizon scales. This is now an experimental fact.. So you have a choice. Either you give up the Friedman-Robertson-Walker solution of GR as a statement about nature, you give up locality, you accept enormous cosmologically large finetunings of the initial conditions, or you do what everyone in the field does… Namely you buy yourself some time, by some sort of exponential expansion right after the big bang. Thus causally disconnected regions would have had enough time to equilibriate in the past. The upshot is that with exponential expansion, comes an exponential amount of space. This really means our horizon is really a small speck in a larger island universe.
    The next version is a little more speculative. It says that b/c you have such large volumes, you have to start looking carefully at quantum processes that might be exponentially suppressed, but that would become non negligable when you analyze the system as a whole. So things like vacuum decay would become important (for instance the Higgs is currently measured to be metastable, so in such a large universe, almost assuredly you would have some sort of Higgs vacuum decay somewhere if that turns out to be true). Thus you now have an island universe with some horizons having different fundamental physics.
    This then leads to the version of the multiverse that is the most speculative, namely that we are in our particular bubble with xyz finetuned constants, b/c of anthropic selection, due to the aforementioned processes taking place and eliminating most places for life.
    As you can see, each version above makes really quite different statements about the laws of nature.

  16. atreat says:

    re “cosmology grad student”:

    Linde-Guth want it both ways though; to claim the inflationary paradigm is somehow explanatory, but also a multitude of theories capable of explaining anything. That is what critics of Linde-Guth are pointing out. Admitting that the inflationary paradigm contains multitudes of theories is to admit that it contains no real explanator power. It is just a mathematical idea allowing theorists to be employed as nature’s ambulance chasers. That seems a pretty damning point for Linde-Guth…

    I want to hear more about the “measurement problem” as it seems there is still huge disagreement about how to do statistics properly in an infinite multiverse. Is this the same problem of trying to do anthropic arguments in a multiverse? Ie the same as disagreements over the Boltzmann brain gedanken experiment?

  17. Peter Woit says:

    cosmology grad student,
    I think the BICEP2 fiasco has a lot to do with this. Guth and Linde claiming “smoking gun” evidence and collecting their Kavli prize based on a wrong experimental result, then going on to claim that the evidence just doesn’t matter, inflation can give any value for r, likely encouraged Steinhardt to say something. From their point of view, I suspect that after BICEP2 they felt they were so close to a Nobel Prize they could taste it, and aren’t about to give that up.

    I just don’t at all believe that the measure problem “doesn’t really bother most people in the field”, my impression is quite the opposite. What most strikes me about this whole subject though is that you don’t have a clue what space you want your measure on (unless you believe that KKLT describes the real world, which no one does).

    As for the identification of string theory and inflation, I agree that there are huge terminology problems here, with no precise meaning to either term. But it’s not me who is identifying the two, it’s Linde. See
    where for instance he writes

    “the most interesting recent developments of the theory of eternal inflation are related to the theory of inflationary multiverse and string theory landscape [57, 62, 63, 64, 65, 66, 67]. These developments can be traced back to the very first paper on eternal inflation in the chaotic inflation scenario [57]. It contained the following statements, which later became the manifesto of the string landscape scenario: “As a result, our universe at present should contain an exponentially large number of mini-universes with all possible types of compactification and in all possible (metastable) vacuum states consistent with the existence of the earlier stage of inflation. If our universe would consist of one domain only (as it was believed several years ago), it would be necessary to understand why Nature has chosen just this one type of compactification, just this type of symmetry breaking, etc. At present it seems absolutely improbable that all domains contained in our exponentially large universe are of the same type. On the contrary, all types of mini-universes in which inflation is possible should be produced during the expansion of the universe, and it is unreasonable to expect that our domain is the only possible one or the best one.”

    The problem is that GKLN and ISL aren’t just arguing about the conventional sort of scientific issue that arises if you have a well-defined class of models (say single-field inflaton with a simple potential) and are trying to compare them to experimental data. GKLN are explicitly saying that the potential can be arbitrarily complicated, there can be arbitrary many fields, ISL describe this as “postmodern inflation” and I think very rightly object to it as unfalsifiable pseudo-science. The GKLN letter that they got 29 others to sign onto does not specifically address this issue, but it does contain a defense of the multiverse, by implication claiming that postmodern inflation and its string theory landscape backing are part of what is being defended.

  18. Peter Woit says:


    Yes, but Guth/Linde are very explicit about what they are defending: the multiverse in which you have different physics in each different universe, and use that as an excuse for why your models don’t predict things. This is what the controversy is about.

  19. Blake Stacey says:

    “So if we ever became convinced that the universe has a nonzero density of electric charge or angular momentum, eternal inflation would no longer be an option.”

    I dunno. It seems to me that if the astronomers found that the net angular momentum of the observable universe is nonzero, then the eternal inflationologists would claim this as evidence for their multiverse, because it meant there must be another universe out there with opposite angular momentum to cancel everything out.

  20. Haelfix says:

    It is perfectly logically consistent to accept the inflationary paradigm, and even the subset of inflationary models that give rise to eternal inflation, while simultaneously questioning whether our particular bubble was anthropically selected. That’s why I reject the either/or dichotomy that’s being presented to us.

    The anwer is that we simply don’t know whether our bubble was anthropically selected or not, and the details matter. I mean we could have asked the same question about the orbits of planets in our solar system and why the convenient configuration for the earth in particular . One could have imagined (before the calculations were done) that there was a mechanism during the collapse of primordial interstellar hydrogen that might have led to planets being preferentially chosen in the goldilocks zone. Of course we now know the answer for us was essentially anthropic.

    We’re in the exact same position with the theory of inflation, namely we need more data (gravitational wave observations in particular) to really zero in on the messy details to make statements that are more constrained.

  21. S. says:

    cosmology grad student,

    does there exist any known case of where using the kinds of probabilistic methods that “almost everyone in the community” believe in has led to a correct prediction?

    It seems to me that the practice is pure nonsense, mathematically speaking. We have as much mathematical reason to believe in conclusions reached through such arguments as we have in conclusions reached by inspecting goat entrails.

    Of course, things can work empirically even without being mathematically well-defined (e.g. infinitesimals in the 17th century, parts of QFT still today). But if we don’t have any empirical evidence that they actual work, the fact that lots of people believe in them is more a condemnation of the state of the field than a vindication of the method.

    So is there any reason to believe that “picking a measure in the first place” actually works, rather than that it is just popular?

  22. EJM says:

    If the multiverse idea is correct, and there are an infinite number of universes where anything and everything is possible, then there must be a universe in which most physicists in it believe and sign a letter that the multiverse idea is incorrect.

  23. Peter Woit says:

    The question isn’t really anthropic selection, it’s whether your theoretical framework is rigid enough to be useful to say something about reality, or whether it’s so flexible it can never be wrong and is an empty, useless idea. The problem IJS are pointing out is that the framework GKLN are pushing appears to be compatible with any conceivable future observational results. For any value of r, Linde has a model that will give that value of r, same for other observables. He and Guth seem happy with this situation, argue that it’s normal science. It’s not.

    S./cosmology grad student,
    I really don’t want to host a discussion of “the measure problem” consisting of completely ill-defined quantities and meaningless statements. If there are specific, well-defined calculations that are relevant to the issues in this controversy, fine, but spare us the nonsense if they’re not.

  24. Jonatan says:

    I think one key point raised by IJS is that EVEN once you’ve picked a particular potential and fixed its parameters (selected one model within the inflationary “paradigm”), if eternal inflation apply (which seems to be so for the vast majority of cases), then you’ll get an infinity of CMB observable values in the infinity of pocket/bubble universes.

    If they’re right, then it means that the whole inflation framework is really incapable of predicting anything and has therefore zero explanatory power.

    I’ve never read or heard anything from GKLN directly addressing this issue, which is very suspicious to me. If IJS are wrong about that, they should very clearly point it out.

    Perhaps a cosmologist could enlighten me here…

  25. Jonatan says:

    In other words, is it possible for a given eternaly inflating model (a particular potential with defined parameter values) to predict exactly the same CMB statistical distributions and spacetime geometry in ALL of the infinite pocket/bubble universes ?

    If the answer is yes, than inflation might be said to be predictive.
    If the answer is no (which is I think one of IJS’s claims), than it can’t.

  26. Haelfix says:

    It’s important to distinguish two different things. There is on one hand the underdetermination of the models due to a lack of data. Basically you are given 5 observables in the CMB and you have hundreds of different inflationary potentials with parameters, some of which are allowed to vary. Therefore for any possible measurement you are possibly (but not necessarily) looking at more than one model that fits. This is the inverse problem, but note that this is really no different than any usual pheno problem in physics, and is how the standard model of particle physics came about. There was a time in the past where we didn’t have enough data to really uniquely select the standard model out of the space of all field theories with undetermined parameters. Fortunately, just like the standard model we have strong constraints (renormalizability, naturalness, simplicity, slow roll etc) and so with further data collection you can really start to zero in on the right answer. Indeed, you could even falsify the whole thing, recall that the initial combination of Bicep with other measurements actually ruled out the whole inflationary paradigm.
    On the other hand, there is the concern that given the large volumes of space, you might have bubbles that undergo vacuum decay (or other rare events), and so we might be living in a bubble that just happened to randomly output our observed universe with ‘whatever you want’. To this second point, I would repeat my criticism above. Namely that it depends upon the details, you don’t necessarily get ‘anything you want’, but yes the measure problem appears in this case…

  27. Peter Woit says:

    Sorry, I’m having trouble taking this seriously enough to spend time thinking about. I don’t remember hearing any inflationistas claiming “the initial combination of Bicep with other measurements actually ruled out the whole inflationary paradigm”, I remember everyone congratulating Guth/Linde on their soon to be awarded Nobel prizes for the confirmation of inflation.

    As for the “it’s just like the Standard Model”, my tolerance level has dropped to zero for that rhetorical move in these arguments over theories of questionable testability.

  28. Haelfix says:

    So in one of the original combination plots (that no one took seriously b/c you can’t just combine data sets like that), BAO + Planck + Bicep produced a so called ‘negative running’ of the primordial power spectrum. This sort of thing would be almost impossible to find in eternal inflation.
    See :

  29. vmarko says:


    “There was a time in the past where we didn’t have enough data to really uniquely select the standard model out of the space of all field theories with undetermined parameters.”

    This is still the case — we still don’t have enough data about the neutrino sector to fix a unique Lagrangian for the SM. There is the Dirac/Majorana issue, the direct/inverse mass ordering issue, etc.

    However, no sane physicist is shouting that SM is untestable because of that, because this is just a handful of unknown numbers, which moreover just barely participate in most of the experiments we can perform. In contrast, the inflation scenarios typically have an arbitrary function (the inflaton potential) in the Lagrangian, which amounts to infinitely many parameters to tune. This is untestable almost by definition, since you need an infinite number of experiments to fix the Lagrangian before you even begin testing it (and this is not the only problem with inflation). This situation is nothing like the SM nonuniqueness.

    Best, 🙂

  30. Jonatan says:

    If IJS are right that even once a potential has been chosen and parameters fixed, eternal inflation inevitably produces a whole range of CMB observable outcomes across pocket/bubble universes, than it seems quite clear that eternal inflation is hopelessly untestable and is an empty idea.

    If they’re wrong, proponents of inflation should explain clearly where and why IJS are wrong about that and clearly state that at least some eternally inflating models will produce the exact SAME CMB observable outcomes in ALL pocket universes.

    Is anyone aware whether such explanation has ever been provided ?

  31. Cosmonut says:

    The multiple versions of inflation seem to be a trivial corollary of the fact that GR permits multiple space-time geometries. So in that sense I agree with Linde & Guth that this doesn’t automatically disqualify them.
    However, people didn’t believe GR because we could somehow dream up a metric that explained the perihelion of Mercury or whatever (I daresay with enough people slogging away, we could). What happened was that the appropriate metric could be *independently deduced* from other considerations and *then* the predictions matched.
    GR also permits metrics corresponding to time machines, rotating universes and so on. But nobody is loudly proclaiming that “This is a feature of the theory and we must accept the existence of time machines because other predictions of GR were valid”.
    The question simply focuses on what kind of matter and energy distribution leads to such metrics and whether we can find *independent evidence* for existence of the latter.
    In that respect, I am with ISL. It doesn’t suffice at all to dream up some inflation potential that makes “generic predictions”, declare victory and then insist we need to accept extrapolations like the multiverse. Until we find independent evidence of at least one scalar field and *then* deduce what it implies for cosmology, inflation must remain a provisional theory.

  32. kashyap vasavada says:

    I thought the main argument in favor of inflationary model was that it has three desirable features. It can explain (1) flatness (2) horizon problem and (3) lack of monopoles. Does anyone on this blog know if Steinhardt’s or anybody else’s model does this?

  33. Peter Woit says:

    kashyap vasavada,
    I’m no expert on 1 or 2, but an obvious answer to 3 is that the Standard Model has no monopoles. 3 is only a problem if you believe in GUTs with monopoles (as Guth did when he was looking for a solution to this “problem”). There is no evidence at all for such a GUT (and by now quite a lot of evidence against), so the simplest solution to 3 is that our fundamental theory is not a GUT with monopoles.

  34. Cosmonut says:

    Yes, Steinhardt’s model does indeed handle the flatness and horizon problem. The model has a mechanism whereby the universe is flattened and homogenized during a contracting phase of the universe, before it “bounces” (the bounce is “our” big bang)
    The main problem is the bounce itself – it requires violations of GR and is just “assumed to be possible” – whereas inflation requires unobserved scalar fields, but then it’s just GR.
    More details here:

  35. tulpoeid says:

    I was waiting until the next renewal date to cancel my Nautilus subscription. No need to wait anymore. The amount of money lost will be the same but the point will not be totally lost.

  36. vmarko says:


    “The main problem is the bounce itself – it requires violations of GR and is just “assumed to be possible” […]”

    It’s not such a big stretch to imagine the violation of GR near the Big Bang moment. First, at such energies the difference between GR and Einstein-Cartan gravity becomes nontrivial, and the latter was shown by Poplawski to produce a Bounce instead of a Bang. So a rather miniscule violation of GR is quite enough. In addition, quantum corrections to GR should play a nontrivial role in such a regime, and in some models (such as loop quantum cosmology) one naturally obtains a bounce via the Friedmann equation with a small quantum correction term. So I’m not at all surprised that in string theory something similar might appear, as conjectured by Steinhardt and Turok.

    Big Bounce is a quite ubiquitous concept in cosmology, you can read some history overview here:

    And some of the models which feature Big Bounce also naturally resolve the flatness and horizon problems, so they are very solid alternatives to inflation.

    Best, 🙂

Leave a Reply

Informed comments relevant to the posting are very welcome and strongly encouraged. Comments that just add noise and/or hostility are not. Off-topic comments better be interesting... In addition, remember that this is not a general physics discussion board, or a place for people to promote their favorite ideas about fundamental physics. Your email address will not be published. Required fields are marked *