Is Inflationary Cosmology Science?

[tl;dr: Check out this article in Scientific American by Ijjas, Steinhardt, and Loeb suggesting that inflation isn’t science; this response by Guth, Kaiser, Linde, and Nomura that was co-signed by a bunch of people including me; and this counter-response by the original authors.]

Inflationary cosmology is the clever idea that the early universe underwent a brief period of accelerated expansion at an enormously high energy density, before that energy converted in a flash into ordinary hot matter and radiation. Inflation helps explain the observed large-scale smoothness of the universe, as well as the absence of unwanted relics such as magnetic monopoles. Most excitingly, quantum fluctuations during the inflationary period can be amplified to density perturbations that seed the formation of galaxies and large-scale structure in the universe.

That’s the good news. The bad news — or anyway, an additional piece of news, which you may choose to interpret as good or bad, depending on how you feel about these things — is that inflation doesn’t stop there. In a wide variety of models (not necessarily all), the inflationary energy converts into matter and radiation in some places, but in other places inflation just keeps going, and quantum fluctuations ensure that this process will keep happening forever — “eternal inflation.” (At some point I was a bit skeptical of the conventional story of eternal inflation via quantum fluctuations, but recently Kim Boddy and Jason Pollack and I verified to our satisfaction that you can do the decoherence calculations carefully and it all works out okay.) That’s the kind of thing, as we all know, that can lead to a multiverse.

Here’s where things become very tense and emotional. To some folks, the multiverse is great. It implies that there are very different physical conditions in different parts of the cosmos, which means that the anthropic principle kicks in, which might in turn imply a simple explanation for otherwise puzzling features of our observed universe, such as the value of the cosmological constant. To others, it’s a disaster. The existence of infinitely many regions of spacetime, each with potentially different local conditions, suggests that anything is possible, and therefore that inflation doesn’t make any predictions, and hence that it isn’t really science.

This latter perspective was defended in a recent article in Scientific American by three top-notch scientists, Anna Ijjas, Paul Steinhardt, and Avi Loeb. They argue that (1) the existence of a wide variety of individual inflationary models, and (2) the prediction of a multiverse in many of them, together imply that inflation “cannot be evaluated using the scientific method” and that its proponents are “promoting the idea of some kind of nonempirical science.”

Now, as early-universe cosmologists go, I am probably less inclined to think that inflation is part of the final answer than most are. Many of my colleagues are more or less convinced that it’s correct, and it’s just a matter of nailing down parameters. I am much more concerned about the fine-tuning difficulties that make inflation hard to get started in the first place — in particular, the hilariously low entropy that is required. Nevertheless, inflation has so many attractive features that I still give it a fairly high Bayesian credence for being correct, above 50% at least.

And inflation is indubitably science. It is investigated by scientists, used to make scientific predictions, and plays a potentially important explanatory role in our understanding of the early universe. The multiverse is potentially testable in its own right, but even if it weren’t that wouldn’t affect the status of inflation as a scientific theory. We judge theories by what predictions they make that we can test, not the ones they make that can’t be tested. It’s absolutely true that there are important unanswered questions facing the inflationary paradigm. But the right response in that situation is to either work on trying to answer them, or switch to working on something else (which is a perfectly respectable option). It’s not to claim that the questions are in principle unanswerable, and therefore the field has dropped out of the realm of science.

So I was willing to go along when Alan Guth asked if I would be a co-signer on this response letter to Scientific American. It was originally written by by Guth, David Kaiser, Andrei Linde, and Yasunori Nomura, and was co-signed by an impressive group of physicists who are experts in the field. (A quick glance at the various titles will verify that I’m arguably the least distinguished member of the group, but I was happy to sneak in.) Ijjas, Steinhardt, and Loeb have also replied to the letter.

I won’t repeat here everything that’s in the letter; Alan and company have done a good job of reminding everyone just how scientific inflationary cosmology really is. Personally I don’t object to ISL writing their original article, even if I disagree with some of its substantive claims. Unlike some more delicate souls, I’m quite willing to see real scientific controversies play out in the public eye. (The public pays a goodly amount of the salaries and research budgets of the interested parties, after all.) When people say things you disagree with, the best response is to explain why you disagree. The multiverse is a tricky thing, but there’s no reason to expect that the usual course of scientific discussion and investigation won’t help us sort it all out before too long.

25 Comments

25 thoughts on “Is Inflationary Cosmology Science?”

  1. Seems to me that the arc of science since antiquity can fairly be described as the triumph of anthropicism over anthropocentricism. That Inflation should continue that trend I would view as a feature, not a flaw.

    And your point about it being okay for a theory to make untestable predictions is an excellent one. Simple Hubble-type expansion predicts that there are stars today which will vanish beyond our light cone tomorrow. Is it unscientific to think that those stars will continue existing, even though testing is literally impossible? And what of the past-the-event-horizon dynamics of stellar- or galactic-scale black holes? Do they render Relativity unscientific?

    The proper response, clearly, is to note that the theory is (presumably) useful over all domains where we can test it, and that it makes these other predictions that we can’t test. It’s a good starting point for inquiry in settings we can’t test…but only a fool would take that to mean that we know what happens in untestable situations…and it’s absurd to suggest that that which is untestable somehow contaminates our understanding of that which we’ve already tested.

    Cheers,

    b&

  2. Well said Ben, and surely, just because something may be untestable for the moment, as our knowledge and understanding increases, and as even new paradigm shifting ideas come to light as a result of new data, that may not mean they will always be untestable? Surely an open mind is the more salubrious approach?

    G

  3. The granularity of mass fluctuations resulting (or not) in primordial black holes are certainly hypotheses testable by the prevalence of SMBH seeds.

  4. I recall Dirac’s problem of the ‘void’, or positron, that could not be tested but clearly predicted by his beautiful equation. He was being buried by doubters.

  5. Eric Falkenstein

    “quantum fluctuations during the inflationary period can be amplified to density perturbations that seed the formation of galaxies and large-scale structure in the universe.”

    I don’t understand how an initial density fluctuation could ever create accumulations that lead to stars, etc. The initial trajectories pre-inflation (before 10^-36 seconds) would be amplified during inflation, but even after, as there’s no wind resistance slowing things down, everything would be less attracted to its neighbor over time. Quantum fluctuations may create electrons and protons that lead to atoms–and thus small scalar density fluctuations increasing–but these atoms would all fly further and further away from each other, never creating a star.

    /not a physicist

  6. The issue of anthropic-multiverse explanation of physical constants being essentially random is IMO the most problematic. But this from a sociological perspective, rather than scientific one. Namely, if you teach your students that coupling constants of the Standard Model are determined via the random choice of a particular bubble universe, you are implicitly discouraging them from trying to look for any alternative explanations for the values of those parameters. And if you moreover base this “explanation” on an untestable conjecture of other bubbles existing beyond the observable horizon, then you are getting dangerously close to in fact teaching your students to be indoctrinated into a particular dogma — belief that a question of the values of SM constants has been “answered” by downgrading it onto the level of a random contingency, with zero experimental support. It is the famous “Multiverse did it” answer, equivalent to the religious “God did it” answer. It is a Bad Idea ™ to teach students something like that and call it “science”. Instead, one should at least be honest enough to one’s students and call it an ideology.

    So the issue here is not that anthropic multiverse scenario is unscientific, but that it is unethical.

    Best, 🙂
    Marko

  7. Marko, I think your concerns, though not unreasonable, are overemphasized.

    Consider, for example, gravitational waves. For seemingly forever, they were in a similar position: an inescapable conclusion of the best theory going, but everybody assumed that it was impossible to ever actually detect them. Students were taught as much…and, yet, here we are in 2017 and just over a year ago we actually did detect gravity waves — and everybody’s aflutter over the implications for new “channels” into astronomical observation.

    Why should anthropic multiverses present any more of an impediment to pedagogy?

    Now, consider all the other examples I presented in my first post up above…are any of them better or worse fits for bad pedagogy than anthropic multiverses?

    Also consider from the other direction. Another way of summarizing Evolution is to state that all of biology is the selective amplification of random initial parameters. It’s no chance that all apex aquatic predators are torpedo-shaped, but it is chance that sharks, dolphins, squid, and tuna all happen to be apex aquatic predators. Does that present any sort of pedagogical stumbling blocks?

    Just as it’s no surprise that the best hunters in the water have body shapes that optimize propulsion, it should be no surprise that organisms capable of modeling the universe are found in an universe whose physics is amenable to the eventual evolution of such organisms. The real scientific pedagogical stumbling block would be if the universe we found ourselves in was one in which life was impossible….

    Cheers,

    b&

  8. I’m not physicist, but loved the articles. I don’t think a group of physicists signing a letter is a good idea, even done with good intentions, to the outsiders it seems like argument from authority.

  9. Mitchell Porter

    “there’s no reason to expect that the usual course of scientific discussion and investigation won’t help us sort it all out before too long.”

    I disagree. They have been saying that inflation is Not Even Wrong for a few years now, first in papers on arxiv, now in popular science media. They didn’t stop the first time Guth et al rebutted the claim that inflation isn’t science, why would they stop now?

    This isn’t an ordinary case of “my model is better than your model”. Their claims are so bizarre – inflation isn’t science, *and* it is disfavored by observation (I mean, how do those two propositions even coexist?) – that I have had to seek special explanations for this behavior. Ijjas has a background in theology, Steinhardt has played a bigger role in ekpyrotic theory than in inflationary theory, and Loeb is a contrarian who likes theoretical pluralism. My “theory” is that something like that coalition of personal incentives, is at work here.

  10. I get concerned when anyone seeks to resolve a scientific question by seeking a larger number of signatories. I am reminded of Einstein’s response to One Hundred Authors Against Einstein–“If I am wrong, one will be enough.”

  11. I agree that “I have more signatures” doesn’t look good, unless, perhaps, these signatures are on the author list of a rebuttal article.

    Yes, it is true that Steinhardt has been grinding this axe for a long time, e.g. Steinhardt had an article on the same theme in Scientific American back in 2011 or whatever.
    Interestingly, Andreas Albrecht, together with Steinhardt one of the early non-Guth inflationists, is also quite sceptical about inflation now, but is more gentlemanly about it.

  12. The issue of anthropic-multiverse explanation is similar to the issue of dinosauric-multiverse explanation, so I don’t see a problem there.

  13. Ben,

    “Consider, for example, gravitational waves. For seemingly forever, they were in a similar position: an inescapable conclusion of the best theory going, but everybody assumed that it was impossible to ever actually detect them. Students were taught as much… Why should anthropic multiverses present any more of an impediment to pedagogy?”

    Oh no, there is a big difference. In the case of gravitational waves, students were taught that they are in principle observable, but that this is extremely challenging technologically. In contrast, the other bubble universes are postulated to reside outside our observable horizon, and are therefore unobservable even in principle, by definition, and irrespective of any technology one may have at ones disposal. This difference is the very crux of the whole issue here.

    “Now, consider all the other examples I presented in my first post up above…are any of them better or worse fits for bad pedagogy than anthropic multiverses?”

    You mean stars and galaxies beyond the observable horizon? While one also cannot test their existence and properties (anymore), it is not an extraordinary claim that they exist, as long as they are assumed to have properties similar to stars and galaxies we still see inside the horizon. It is a leap of faith, but not a big one. Regarding other bubble universes, I could accept their postulated existence with no fuss, as long as they are assumed to have properties similar to our own bubble — same physical constants, similar matter content, etc. — they would just represent more of the same stuff, which is also not an extraordinary claim. But note that this picture is *not* what is being proposed in the anthropic multiverse scenario — other bubbles are in fact assumed to have quite different properties (which is an extraordinary claim!!), and based on this unverifiable conjecture one tries to “explain away” that the properties of our universe are random and that one should not look for an explanation.

    “It’s no chance that all apex aquatic predators are torpedo-shaped, but it is chance that sharks, dolphins, squid, and tuna all happen to be apex aquatic predators. Does that present any sort of pedagogical stumbling blocks?”

    I don’t see your point. Everything about sharks, dolphins, squid and tuna is observable and verifiable. But if you base your argument on an assumption that in another universe dolphins instead have transparent skin or three eyes, then I’d have a reason to doubt your argument.

    “Just as it’s no surprise that the best hunters in the water have body shapes that optimize propulsion, it should be no surprise that organisms capable of modeling the universe are found in an universe whose physics is amenable to the eventual evolution of such organisms.”

    This is not an issue, I agree with that. It’s just that I find hardly believable that other universes, ones with *different* physics, actually exist. There is no evidence, and it is impossible to obtain any evidence, by the definition of the theory itself. That is the problem.

    Best, 🙂
    Marko

  14. When I was a little kid I said I thought there were lots of universes and everyone told me I was stupid.

    So come on multiverse!

  15. Good to see a new posting on the blog!

    You may recall the debates Ben and I had last year on philosophy of mind. After mulling over ‘The Big Picture’ (which I read on Kindle) and thinking hard, I greatly clarified my thoughts.

    Ben you’ll be pleased to hear that I now have no remaining doubts about physicalism – in fact, I’m now convinced that Sean and yourself were correct – all objective properties are indeed physical properties (i.e materialism is true). The mistake I was making was to confuse my trinitarian ontology (‘information-fields-cognition’) with the materialism-idealism debate.

    I still believe that my triple-aspect ontology (information-fields-cognition) is correct -I still think that there are 3 distinct fundamental (universal) properties-but where I went wrong was to think that there was anything non-physical.

    As regards the multiverse , my views on that have also changed since last year. Whereas I’ve entirely come-around to materialism , I’m now sceptical about multiverses/MWI. So I’ve moved to Sean’s position on one major point (materialism), only to move away from him on other one (multiverses) 😀

    The problem with the multiverse is that it implicitly buys into the ‘block universe’ picture. What I realized is that ‘time’ in quantum mechanics conflicts with the ‘block universe’ picture from relativity.

    Once the ‘block universe picture’ is put into doubt, all the multiverse models (including MWI and inflation) are suspect in my view. I think we need to seriously consider the possibility that there is more than one type of time. Could time in fact have TWO dimensions rather than just one?

    I have been considering again the interesting ideas of philosopher Alfred-North-Whitehead. His ‘process philosophy’ treated time as truly fundamental, and his radical position was that there are no ‘things’ , only ‘processes’,

    The discovery of ‘time crystals’ is extremely interesting – here’s a system that oscillates forever in time (a repeating set of processes in time) with little to no energy input into the system! This really does make me wonder if Alfred-North-Whitehead was right.

    Could the ‘time is a river’ model of reality be correct after all?

  16. Marko:

    It’s just that I find hardly believable that other universes, ones with *different* physics, actually exist.

    Why should that be so hard to believe? What’s so special about 299,792 km/s that you can’t consider it reasonable that another universe could have light that propagates at 300,208 km/s instead?

    There is no evidence, and it is impossible to obtain any evidence, by the definition of the theory itself.

    Eh, you seem to have missed that link in Sean’s original post — the one that goes to a guest post on Preposterous Universe by Matt Johnson proposing that, perhaps, evidence consistent with interactions with other universes could be discerned in the WMAP CMBR data.

    Cheers,

    b&

  17. “…,perhaps, evidence consistent with interactions with other universes could be discerned in the WMAPCMBR data.”

    I think I have to quote Unspiek, Baron Bodissey here who said “We must not confuse statistical probability with some transcendental and utterly compelling force.”

  18. David D Prentiss

    Dear Sean,
    Your discussion that Inflation is necessary is very convincing. Where the argument falls short in my opinion is the insertion of the Inflaton Field to drive the process. A paper on Loop Quantum Cosmology: arXiv:1103.2475v1 [gr-qc] 12 Mar 2011 by Abhay Ashtekar and David Sloan offers a more natural origin for the energy powering inflation from that contained within the Big Bounce Itself. The process also seems to be self limiting doing away with the bizarre continuous inflation producing a large number of other universes. Anyway this paper’s idea seems to produce a simpler route to Inflation and with knotty problems, simpler is usually the better choice.
    Dave Prentiss

  19. Ben,

    “Why should that [universes with different parameters] be so hard to believe? What’s so special about 299,792 km/s that you can’t consider it reasonable that another universe could have light that propagates at 300,208 km/s instead?”

    Because extraordinary claims require extraordinary evidence. Skepticism is the way good science should always be done. If someone tells me that unicorns exist, I’ll be skeptical until they produce evidence. If someone claims that universes with different fundamental constants exist, I’ll be skeptical about that too, until presented with evidence. And in the latter case, the issue is that not only that evidence isn’t available “so far”, but that one should not even expect that any evidence exists to begin with. That’s an extraordinary claim that is taken completely via a large leap of faith, i.e. not something a responsible scientist would easily accept.

    And btw, the speed of light is almost the only parameter which should remain the same for all bubble universes, since it is assumed that general relativity operates in the same way both across various bubble universes and the intermediate inflating space. Newton’s gravitational constant is another.

    “Eh, you seem to have missed that link in Sean’s original post — the one that goes to a guest post on Preposterous Universe by Matt Johnson proposing that, perhaps, evidence consistent with interactions with other universes could be discerned in the WMAP CMBR data.”

    I didn’t miss it. The colliding bubbles scenarios can indeed be tested by looking at signatures in the CMB, but — nothing has been found! Both the old WMAP data and the new Planck data have been searched for these signatures, and essentially nothing has been detected, despite optimistic hopeful and wishful thinking by Matt Johnson. Let me quote you the key sentence from his guest post: “We did not find any circular temperature discontinuities in the WMAP data.”. That’s it — that’s how observational data is judging the colliding bubbles scenarios. This means that these scenarios are being excluded by the data, ever more precisely as time passes, and therefore such models are getting increasingly falsified. The only remaining viable models are the ones which say that other bubbles have always remained, and will always remain, beyond the observable horizon. Which means that any evidence for their existence is absent by definition of the theory. And I am supposed to take such a big leap of faith, only to conclude that the parameters of the SM have no explanation at all? And moreover teach all that to my students? No way, that’s both irresponsible and unethical.

    Best, 🙂
    Marko

  20. That certain limits to observable phenomena and testable hypotheses may be being reached, doesn´t preclude that certain subtle inferences aren´t there and won´t be made with better more precise measurements by more highly refined instruments in the future such as the Webb telescope, when, we hope fingers crossed, it can finally be launched and correctly positioned next year. A clearer more detailed picture of the very early universe will no doubt both answer, and raise, many questions about the early inflationary period and most significantly nascent galaxy formation.

  21. In the original authors’ reply letter, aren’t they setting up a strawman when they use Guth’s words to argue:

    “And if inflation produces a multiverse in which, to quote a previous statement from one of the responding authors (Guth), “anything that can happen will happen” — it makes no sense whatsoever to talk about predictions.”

    In the first place, whether or not inflationary models involve a multiverse, these models can certainly “talk about predictions” that take place in THIS universe. In fact, they do.

    In the second place, it may be possible for multiverse models to make testable predictions, an idea, as Sean has already pointed out, that is being explored in published papers.

    Some commenters here are remarking that the signed letter from 33 physicists is a bad idea, smacking of an “argument from authority”. I completely disagree. Legitimate scientific ideas are supposed to be submitted to a peer review process that depends upon authorities in the field. It is entirely appropriate for additional authorities in the field to make their assessments known as well – clearly the editors of Scientific American agree. Further, the letter does not depend on authority, but rather provides itemized evidence that some of the concluding statements of the original article are simply false.

  22. vmarko,
    https://arxiv.org/abs/1704.03814 Evidence against a supervoid causing the CMB Cold Spot

    “This suggests a primordial origin for the Cold Spot,
    either from an unlikely fluctuation in the standard cosmology
    or as a feature produced by non-Gaussian conditions in
    the early Universe.”
    i.e. code for it was likely produced by a colliding bubble universe . I guess they didn’t mention “colliding bubble universe” to get it quickly peer review published in MNRAS (already accepted)

Comments are closed.

Scroll to Top