Skip to main content

Physicist Slams Cosmic Theory He Helped Conceive

I love apostates, believers in or, better yet, conceivers of a theory who turn against it. They restore my faith in science, because they show that scientists can overcome attachment to their own brainchildren, a feat that is essential for progress and cannot be taken for granted.

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


I love apostates, believers in or, better yet, conceivers of a theory who turn against it. They restore my faith in science, because they show that scientists can overcome attachment to their own brainchildren, a feat that is essential for progress and cannot be taken for granted. Paul Steinhardt, Albert Einstein Professor in Science and Director of the Center for Theoretical Science at Princeton University (is there a cooler title?), is an apostate. I first interviewed him in the late 1980s while researching quasicrystals, a weird, quasi-periodic form of matter once thought to be entirely hypothetical. Steinhardt coined the term "quasicrystals" and was a major investigator of them. In subsequent years I talked to Steinhardt about cosmology and in particular inflation, an idea that he helped refine in the early 1980s. Inflation holds that immediately after the big bang, our universe underwent an almost unimaginably explosive, faster-than-light growth spurt. Lately, Steinhardt has been criticizing inflation and related ideas, notably multiverses, in unusually blunt terms. So I was delighted when he agreed to answer some questions. For related material, see my Q&As with physicists George Ellis, Carlo Rovelli, Edward Witten and Garrett Lisi, as well as articles listed in Further Reading.– John Horgan


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Horgan: What inspired you to become a theoretical physicist? What were your hopes for you personally, and for the field as a whole?

Steinhardt: I was drawn to science as a very young child and have wanted to be a scientist ever since. My dream was always to make discoveries: to reveal some of the secrets that Nature has been hiding from us. I did not care so much about what the field of science is or what approaches to use. I enjoyed everything that I delved into. I just wanted the thrill of discovering things that no one knew before. I was not immediately aiming to be a theorist or a physicist. I loved mathematics and produced several science fair projects based on purely mathematical ideas. But I also had labs of all sorts at home as a child. I did considerable work in university labs (mostly biological) as a teenager and young adult. I am even directly involved in some experimental work today. I always enjoy visiting labs and working closely with experiments whenever I can. Despite enjoying all of these experiences, I realized at some point during my undergraduate experience at Caltech that I had the mentality of a theorist and not an experimentalist. I enjoyed creating and developing ideas more than testing them. I was also incredibly fortunate to have been mentored by Richard Feynman, who, in addition to be generally inspirational, showed me that a theorist has the advantage of being able to work in many diverse areas of science at the same time.

Horgan: Have your hopes been fulfilled?

Steinhardt: It has been a privilege to be part of the scientific enterprise and to have witnessed the remarkable scientific breakthroughs of dedicated, heroic scientists. One has to marvel at what humans can accomplish. On a more personal level I have been fortunate to work with a number of talented students and scientists on some of the most exciting issues of our time. I have experienced firsthand the thrill of discovery in a variety of fields. No matter the field and no matter how small the advance, the feeling of those moments when you know that you are the only person (or one of the few people) to have figured something out or to have uncovered a truth that no one knew before is incredible.

Horgan: You were one of the originators of inflation theory. When and why did you start having doubts about it?

Steinhardt: From the very beginning, even as I was writing my first paper on inflation in 1982, I was concerned that the inflationary picture only works if you finely tune the constants that control the inflationary period. Andy Albrecht and I (and, independently, Andrei Linde) had just discovered the way of having an extended period of inflation end in a graceful exit to a universe filled with hot matter and radiation, the paradigm for all inflationary models since. But the exit came at a cost -- fine-tuning. The whole point of inflation was to get rid of fine-tuning – to explain features of the original big bang model that must be fine-tuned to match observations. The fact that we had to introduce one fine-tuning to remove another was worrisome. This problem has never been resolved.

But my concerns really grew when I discovered that, due to quantum fluctuation effects, inflation is generically eternal and (as others soon emphasized) this would lead to a multiverse. Inflation was introduced to produce a universe that looks smooth and flat everywhere and that has features everywhere that agree with what we observe. Instead, it turns out that, due to quantum effects, inflation produces a multitude of patches (universes) that span every physically conceivable outcome (flat and curved, smooth and not smooth, isotropic and not isotropic, scale-invariant spectra and not, etc.). Our observable universe would be just one possibility out of a continuous spectrum of outcomes. So, we have not explained any feature of the universe by introducing inflation after all. We have just shifted the problem of the original big bang model (how can we explain our simple universe when there is a nearly infinite variety of possibilities that could emerge from the big bang?) to the inflationary model (how can we explain our simple universe when there is a nearly infinite variety of possibilities could emerge in a multiverse?).

I have to admit that I did not take the multiverse problem seriously at first even though I had been involved in uncovering it. I thought someone would figure out a resolution once the problem was revealed. That was 1983. I was wrong. Unfortunately, what has happened since is that all attempts to resolve the multiverse problem have failed and, in the process, it has become clear that the problem is much stickier than originally imagined. In fact, at this point, some proponents of inflation have suggested that there can be no solution. We should cease bothering to look for one. Instead, we should simply take inflation and the multiverse as fact and accept the notion that the features of the observable universe are accidental: consequences of living in this particular region of the multiverse rather than another.

To me, the accidental universe idea is scientifically meaningless because it explains nothing and predicts nothing. Also, it misses the most salient fact we have learned about large-scale structure of the universe: its extraordinary simplicity when averaged over large scales. In order to explain the one simple universe we can see, the inflationary multiverse and accidental universe hypotheses posit an infinite variety of universes with arbitrary amounts of complexity that we cannot see. Variations on the accidental universe, such as those employing the anthropic principle, do nothing to help the situation.

Scientific ideas should be simple, explanatory, predictive. The inflationary multiverse as currently understood appears to have none of those properties.

These concerns and more, and the fact that we have made no progress in 30 years in addressing them, are what have made me skeptical about the inflationary picture.

Horgan: Edward Witten told me recently that "the evidence for inflation today is vastly greater" than it was 20 years ago. Do you agree?

Steinhardt: The story of inflation is very peculiar, and we cosmologists have obviously not done a very good job of explaining the situation. Here is what’s confusing. On the one hand, mounting observational evidence gathered over the last 30 years supports what we thought were the predictions of inflationary theory based on our understanding prior to 1983. (A notable exception is the pre-1983 prediction of gravitational waves, which was at a level that the WMAP and Planck satellites should have detected but did not; see my comments on BICEP2 below.) The irony is that our understanding of inflation has changed dramatically. We no longer believe that inflation makes any of those predictions so that none of the magnificent observations made over the last 30 years can be viewed as supporting inflation.

Since 1983, it has become clear that inflation is very flexible (parameters can be adjusted to give any result) and generically leads to a multiverse consisting of patches in which any outcome is possible. Imagine a scientific theory that was designed to explain and predict but ends up allowing literally any conceivable possibility without any rule about what is more likely. What good is it? It rules out nothing and can never be put to a real test.

I cannot blame people who wish to return to our days of inflationary innocence when the theory seemed predictive and in agreement with observations.

Unfortunately, as the leading proponents of inflation concede, we do not know how to return to this early vision.

Horgan: Do you think the BICEP2 observations may still turn out to support inflation?

Steinhardt: As just explained, it is not possible to find evidence to support or refute inflation because an inflationary multiverse includes patches with cosmic gravitational waves and without them. So I must reinterpret your question to mean: could the BICEP2 observations yet turn out to be evidence of primordial gravitational waves, as the experimental team originally claimed.

No. The BICEP2 observations have been completed and published. Based on what was known at the time when the results were announced, it is not possible to rule out the “null hypothesis.” In a scientific experiment, the null hypothesis is the claim that the data can be explained by factors already known to exist without introducing anything new. If an experiment cannot reject the null hypothesis, the scientists are not supposed to claim evidence of an alternative explanation. In the case of BICEP2, the null hypothesis is that their B-mode signal is due to known foregrounds: dust plus lensing plus synchrotron radiation plus systematics. Their published results are consistent with the null hypothesis. Hence, by the standard rules of scientific methodology, there can be no claim by BICEP2 of detecting primordial gravitational waves (an alternative hypothesis). In the wake of BICEP2, I hope we can return to good scientific practice.

A fresh race is now underway using more sensitive instruments to detect primordial gravitational waves or set stronger limits on their existence. The true story of primordial gravitational waves will be revealed by future experiments and should be credited to those experimental teams that mount them.

Also, even if we detect primordial gravitational waves, we should not rush to the conclusion that they are due to inflation. Better theories may come along that avoid the pitfalls of inflation and that nevertheless predict gravitational waves.

Horgan: In a recent essay on Edge.org, you criticized multiverse and string as well as inflationary theories. Can you summarize your concerns?

Steinhardt: My concern was that the multiverse is a ‘theory of anything’, a proposal that allows all possible cosmological outcomes (smooth or not smooth, curved or flat, etc.) and, consequently, is not subject to empirical tests. Some claim that superstring theory allows exponentially many (or perhaps infinitely many) possibilities for the fundamental laws (masses of particles, types of forces, etc.) and that there is no guiding principle to determine which set of physical laws is more probable. The sets of laws comprise what is called the “string landscape.”

Combine the inflationary multiverse with the string landscape, and now one has a ‘supertheory of anything’: both the cosmological properties and the microphysical properties of the universe are accidental and unpredictable.

As we understand superstring theory better, I truly hope we find that there are sound reasons why the physical laws we observe are naturally selected. Superstring theory, combined with an improved cosmological picture, may then lead to a powerfully explanatory and predictive theory.

Horgan: Have you gotten any blowback for your criticism?

Steinhardt: For the most part, the discussion has been civil and intellectual. It has been fascinating to hear the variety of views.

Horgan: Witten thinks string theory is still "on track" and represents physicists' best hope for a unified theory. Comment?

Steinhardt: I share Edward’s view that string theory represents our best hope at present for a unified theory. However, I think success requires that the string landscape issue be resolved and that we find some empirical evidence for supersymmetry.

Horgan: What is your cyclic model of the universe? Is it falsifiable?

Steinhardt: The cyclic model emerged when my collaborators and I asked the question: is there any way of explaining the smoothness and flatness of the universe and small ripples in density without inflation? The answer was yes: the key is to have a universe in which the big bang is replaced by a big bounce. In this picture, the present period of expansion and cooling is preceded before the bounce by an epoch of contraction, and the important events that shape the large-scale structure of the universe (smoothing, flattening and generating fluctuations) occur before the bounce during a period of slow contraction. There is no high-energy inflation phase – the universe goes straight from the bounce into a period of slow expansion and cooling. Inflation is not needed to smooth and flatten the universe. Consequently, there is no multiverse.

The bounces can repeat at regular intervals resulting in a cyclic universe. In some versions, the theory is geodesically complete (existing infinitely into the past), unlike inflation, which requires a beginning and special initial conditions.

The cyclic theory makes one generic model-independent prediction: no detectable primordial gravitational waves. Hence, if BICEP2 had been correct or if primordial gravitational waves are observed in future experiments, all cyclic models based on smoothing via slow contraction will be entirely ruled out.

Therefore, yes, the cyclic theory is definitely falsifiable!

Horgan: Why do you continue to work on quasicrystals? Are they in any way linked to your cosmological interests?

Steinhardt: Quasicrystals fascinate me because they have repeatedly led me to the kinds of thrills of discovery that attracted me to science in the first place. The underlying mathematics is elementary geometry and algebra. The underlying physics is the fact that solids are composed of atoms that are packed closely together, like children’s building blocks. From these simple ingredients emerges a major surprise. For over one hundred years, scientists thought they understood all the possible ways atoms can arrange themselves in solids … but they were wrong. As first shown by my student, Dov Levine (now at the Technion) and me, there are an infinite number of arrangements with astounding beautiful symmetry that had been missed. We called these quasicrystals. These possibilities include grains with perfect pentagonal faces, once considered to be mathematically and physically impossible. The first examples were synthesized in the laboratory by Dan Shechtman and collaborators in the 1980s (for which Shechtman was awarded the 2011 Nobel Prize in Chemistry), and now over 100 different synthetic examples are known.

Even after thirty years, the subject is still in its infancy with new opportunities for discovery emerging all the time. For example, until 2009, all known quasicrystals were synthesized in the laboratory, and we did not know of any quasicrystals formed in nature. For my collaborators and me to find the first example of a natural quasicrystal required a detective adventure worthy of a novel and a geological expedition to the Russian Far East. Through a strange, circuitous path, the pursuit of natural quasicrystals has connected me to new questions about the conditions during the formation of the solar system. And there continue to be surprises, both theoretical and experimental. There is no direct connection to my cosmological interests, at least at present.

Horgan: In the 1980s, Stephen Hawking suggested that physics might someday provide such satisfying answers that the field would end. Do you--or did you ever--share this hope?

Steinhardt: No. I tend to feel that every answer generates more questions, and my hope is that this continues.

Horgan: Are you religious? Can you be a physicist and also believe in God?

Steinhardt: I never answer the first question because I consider religion to be a private matter. My scientific views stand on their own and I would like them to be evaluated independent of my private views about religion.

In answer to your second question, it is a demonstrated fact that successful physicists can believe in God.

Further Reading (all except first two by Horgan):

Endless Universe: Beyond the Big Bang(Broadway, 2008), by Paul Steinhardt and Neil Turok.

"The Inflation Debate," by Paul Steinhardt. https://www.scientificamerican.com/article/the-inflation-summer/

“Why I Still Doubt Inflation, in Spite of Gravitational Wave Findings.”https://blogs.scientificamerican.com/cross-check/2014/03/17/why-i-still-doubt-inflation-in-spite-of-gravity-wave-findings/

"My 1992 Profile of Cosmic Trickster and Inflation Pioneer Andrei Linde." https://blogs.scientificamerican.com/cross-check/2014/03/18/my-1992-profile-of-cosmic-trickster-and-inflation-pioneer-andrei-linde/

“Hey Physics, Get Real!” http://chronicle.com/article/Hey-Physics-Get-Real-/126662/

“If You Want More Higgs Hype, Don’t Read This Column.” https://blogs.scientificamerican.com/cross-check/2012/07/04/if-you-want-more-higgs-hype-dont-read-this-column/

"Cosmic Clowning: Stephen Hawking’s “new” theory of everything is the same old CRAP." https://blogs.scientificamerican.com/cross-check/2010/09/13/cosmic-clowning-stephen-hawkings-new-theory-of-everything-is-the-same-old-crap/