ISL(Anna Ijjas,Paul Steinhardt,Abraham Loeb) have put together a Fact-Checking page in response to the letter.


In their letter, the authors claim four successful “predictions” of inflation to have been proven right by observations: (1) critical mass density; (2) nearly “scale-invariant” perturbations; that are (3) “adiabatic,” and (4) “Gaussian.” What the letter omits is that, before the observations were made, some of the same authors who signed the rebuttal wrote papers claiming exactly the opposites of (1) through (4) are also compatible with inflation.

One cannot claim a theory predicts an observation if the observation and the opposite outcome are both compatible with the theory.

I have read the detailed explanations and examples of the arguments in the link, unfortunately, I don’t have enough knowledge to tell if these arguments are legitimate.

  • $\begingroup$ Inflation predicts all of those things really well, but other non-inflationary models may also be able to replicate what we observe. $\endgroup$
    – astronat
    Commented Jul 21, 2018 at 19:20
  • $\begingroup$ I'm voting to close this question as off-topic because physics.SE is not an appropriate location for peer review. $\endgroup$
    – ACuriousMind
    Commented Sep 13, 2018 at 15:46

1 Answer 1


The simplest models of inflation are so-called canonical ("normal" kinetic energy term), single field (a single field, the inflaton, is responsible for both driving the inflationary expansion and generating the initial density perturbations), slow roll (smooth, gradually decreasing potential energy function) models. These generically predict near-spatial flatness and a nearly scale invariant spectrum of Gaussian, adiabatic density perturbations. Primordial gravitational waves may or may not be generated at observable levels by these models. If any of these predictions are disconfirmed, then canonical, single field, slow roll inflation is false.

Inflation is a broad and phenomenologically rich paradigm, and there's much more to it than these simplest of models. If the universe turns out not to be near-critical, then there's a model for this: so-called open inflation which was studied by Linde and others in the 90's before the universe was seen to be accelerating. If the perturbations are non-Gaussian, then depending on the type, either the kinetic term is noncanonical, there is transient non-slow roll behavior, or there are multiple fields contributing to the generation of perturbations (the curvaton, for example). If the spectrum is not nearly-scale invariant, then we either have non-slow roll behavior across some range of scales, or there are multiple fields driving inflation such that there are rather abrupt changes in "direction" in field space. Another possibility is hybrid inflation which can generate a very blue spectrum ($n > 1.5$), but these models require an additional auxiliary field to end inflation, and so are not single field models. If the perturbations are non-adiabatic, then there's likely an isocurvature component, for instance, via the curvaton mechanism.

The point being that there are enough knobs to turn across the inflationary model space to accommodate a wide range of predictions. But, the "generic" predictions: flatness, near scale invariance, Gaussian, adiabatic perturbations are the hallmark of the most basic and best studied family of models. It is doubtless that this is the inflation that Guth, Linde and others are referring to in their statement, if not explicitly stated.

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – ACuriousMind
    Commented Sep 13, 2018 at 15:44
  • $\begingroup$ The simplest models of inflation are canonical, single field, slow roll inflation. Do inflationary models other than these simplest models which were mentioned in your answer also produce the required number of e-folds to solve the horizon and flatness problems? $\endgroup$
    – Forge
    Commented Sep 20, 2018 at 21:35
  • $\begingroup$ Yes. All the alternatives that I mentioned are viable models with respect to the horizon and flatness problems. $\endgroup$
    – bapowell
    Commented Sep 21, 2018 at 2:28

Not the answer you're looking for? Browse other questions tagged or ask your own question.