American Scientist

The Fabric of the Cosmos: Space, Time, and the Texture of Reality. Brian Greene. xii + 569 pp. Knopf, 2004. $28.95.

Unlike the anonymous peer reviews in which scientists evaluate one another's papers and grant proposals, book reviews are not required in principle to be unbiased. They are usually signed, which helps ensure some degree of fairness. But there is also a great tradition of bitchy book reviews written by rival authors, spurned lovers—and even mothers: When The Early Universe, a book on cosmology by Rocky Kolb and Michael Turner, was published, a pair of humorous reviews appeared that were attributed to their mothers. "Mrs. Dorothy B. Kolb" kvetches about the drab illustrations:

My son told me that it is not a "trade" book, but a serious book intended for graduate students and professional scientists. "Big deal," I told him. "You think you are too smart to put in a few pretty pictures."

Book reviewers must, of course, fully disclose their possible conflicts of interest. So I should confess that I am the author of two books that might be seen as competing with Brian Greene's The Fabric of the Cosmos (one has a similar title and the other covers similar territory). I am also one of the inventors of a theory, loop quantum gravity, that is generally seen as the main rival to string theory, which Greene advocates in this and a previous book. Even worse, I get mail from readers who complain that I am not as good-looking as Greene, even though I write better. (Fortunately, I am reassured by she-who-matters that the opposite is true.) I should hasten to add that I know Greene and like him personally, and I feel a great solidarity with him as a fellow native New Yorker trying to make it in the harsh world of science.

Having said all this, let me begin by noting that this is a wonderful book for the lay reader who wants to get a glimpse of what we theoretical physicists are thinking about. It is extremely well written. To prepare the reader to understand the theories that currently animate the frontiers, Greene gives an introduction to the main ideas of 20th-century physics—relativity, cosmology, quantum theory and particle physics—and he does it very well. As someone who, like Greene, has struggled—twice—with the problem of how to write a fresh and compelling book that introduces the reader to these topics, I commend him for his success here. He is a master expositor and popularizer, and these parts of the book really shine.

The Fabric of the Cosmos also has a feature that is missing from too many popular books in this area: It is for the most part fair and evenhanded. Although Greene has an agenda—he is, after all, a string theorist—he does give space both to the current mainstream theories and to their rivals. Although he indicates his own preferences and hunches, he does this openly, with a tone that makes it clear he respects those who disagree. His chapters on quantum mechanics introduce the various competing views as to the meaning and the adequacy of quantum theory. Each view is presented sympathetically, but its weak points are also mentioned. One comes away understanding why it is that smart people have reached no consensus on the problem.

Similarly, his discussion of inflation, a hypothesis about the very early universe, strikes exactly the right tone. Greene brings the reader to marvel at how closely recent observations fit the predictions of the inflation theory. But he also contrasts it with a rival theory put forth by cosmologists Neil Turok and Paul Steinhardt, pointing out the strengths and weaknesses of both theories.

When it comes to my own area of quantum gravity, I find far less to gripe about than I had expected after reading an unsigned review in The Economist, which claims that the book gives "short, if civil, shrift" to loop quantum gravity. In fact, Greene gives pride of place to loop quantum gravity: In his closing section he praises it and proposes that the next revolution in string theory will come about when it incorporates loop quantum gravity. Although I might put it the other way around, I can't complain much, as Greene's vision is not substantially different from the future I argued for in the closing of my own book on the subject. Moreover, he clearly embraces the main ideas behind loop quantum gravity: that space and time are relational rather than absolute and that all properties of space and time are the result of dynamical evolution. As Greene says, this is the main lesson of Einstein's theory of general relativity, and it is realized in loop quantum gravity but not in string theory.

There are, however, two grounds on which I believe the book falls short: It is far too uncritical of Greene's own subject, string theory. And it offers little for the reader who remembers that science is based on experiment and who may therefore wonder how it is that all these beautiful, exciting ideas are to be tested against the harsh light of reality.

A careful reader may wonder what is behind Greene's admission, late in the book, that all is not well in stringland:

Even today, more than three decades after its initial articulation, most string practitioners believe we still don't have a comprehensive answer to the rudimentary question, What is string theory? . . . [M]ost researchers feel that our current formulation of string theory still lacks the kind of core principle we find at the heart of other major advances.

Indeed, string theory has been "promising" for more than 30 years. And like those perpetual students one meets in the cafes of any college town, it has yet to grow up and earn a serious living. That is to say, it has yet to make any unambiguous prediction that could be tested up or down in doable experiments. It is fun to speculate about many unseen dimensions and particles, but they have remained just that: unseen. There is, as Greene discusses, a small possibility that string theorists are wrong about the sizes of strings, in which case perhaps they will be observed in the Large Hadron Collider now under construction at CERN, which will be the world's most powerful particle accelerator when it is switched on in 2007. But the most likely scenario is that strings, if they exist, are more than 10 orders of magnitude too small to be seen directly by any accelerator built with known technology.

It is, then, worrying that many of the claims made in the book for string theory are exaggerated. Greene describes clearly the beautiful results that inspire hope among those who study string theory. But a presentation for the general public of a speculative and untested theory should be honest about which problems remain open. There are a number of cases in which conjectures widely believed by string theorists have remained unproven despite many years of hard work by very smart people. In several key cases Greene nevertheless presents those conjectures as facts.

For example, he clearly describes how strings, by having a finite size, suppress the quantum fluctuations of the gravitational field, perhaps leading to a consistent unification of gravity and quantum theory. As Greene explains, the result is that the theory gives finite answers, an obviously necessary condition, but one hard to satisfy. What he does not say is that the suppression has been demonstrated  only in a certain low-order approximation. In fact, determined efforts by many physicists and mathematicians over a period of more than 20 years have failed to produce a proof of the finiteness or consistency of string theory. The failure has been so complete that I am aware of only two people, out of a community of more than a thousand string theorists worldwide, who are still trying to solve this problem.

This is the kind of thing that a truthful exposition for the general public should include. It is, unfortunately, not the only instance of important information going unmentioned. One of the strongest motivations for believing in string theory, originally, was that it was thought that there could be only one theory that unified all the forces. If string theory did that, it had to be the one right theory. Thus it was worrying when initially five different string theories were discovered. These were followed by many more, corresponding to hundreds of thousands of different ways to curl up the unobserved dimensions. Understandably, there was a great sense of relief when in 1995 Edward Witten conjectured that these theories were part of a single theory. Here is how Greene describes this:

Witten . . . uncovered a hidden unity that tied all five string theories together. Witten showed that rather than being distinct, the five theories are actually just five different ways of analyzing a single theory. . . . [A] single master theory links all five string formulations. The unifying master theory has tentatively been called M-theory. . . . String theory, Witten demonstrated . . ., is a single theory.

Greene's exposition fully communicates the excitement that greeted the announcement of Witten's conjecture. But one does not learn from Greene that it was—and almost 10 years later still is—an unproven conjecture. In fact, that "single," "unifying master" theory has never been written down. No one has ever produced a convincing proposal for its fundamental principles or its mathematical formulation. What Witten and others did was to uncover mathematical evidence that the different theories might be related and to assert that this might be explained by the existence of a single unified theory. This is important, but it is not what Greene says. Conjecture is not proof, and conjectures about the existence of unknown theories are not theories.

There are several other examples of the same kind of exaggeration, in which conjectures are reported as truths, and the phrase "convincingly argued" is used to imply that something is proven. This overstatement greatly weakens the value of the book for the public. It also raises questions. Brian Greene knows that the proofs of these assertions are incomplete. At the same time, his fairness in other matters shows that he is a person of integrity. So I am sure that there is no intent here to deceive. What is happening, I suspect, is a phenomenon much more troubling. I'm afraid that it is simply inconceivable to him, as it is to many members of the string theory community, that these conjectures could be wrong.

Among the uncomfortable developments not mentioned by Greene is that 18 months ago estimates for the number of different string theories increased, to incredible numbers such as 10¹⁰⁰. And none of these appear to be involved in the conjectured unification of the five original theories proposed by Witten under the name of M-theory. If this is right, then there may never be an experiment that could disprove string theory, for there will be a vast number of string theories that all agree with any possible result of any future experiment.

What some of us worry about is the possibility that, in the absence of experimental tests, a community of even the smartest and best-educated people in the world can convince themselves of the truth of a theory, to a degree far beyond what an objective reading of the actual evidence would justify. We have recently seen this type of self-delusion in other domains. For example, it appears that a vast majority of members of the international intelligence community were convinced that Iraq had weapons of mass destruction. Similarly, although greed certainly played a role in the accounting scandals of recent years, I suspect that many of the people involved, having gotten to the top of their professions by hard work and superior intelligence, were simply confident that their version of economic reality would prevail.

It is because of the tendency of even brilliant, well-educated people to fool themselves—especially collectively—that the progress of science has always required an interplay of theory and experiment. What is most troubling about Greene's book is that there is no evidence that he sees the "experimental correction" coming. Indeed, the last few years have brought several surprising experimental developments that have the potential to dramatically challenge our understanding of fundamental physics. It turns out to be simply false that the Planck scale, where gravity and the quantum meet, is inaccessible to experiment. New experimental results, in cosmic-ray physics and in observations of gamma-ray bursts, and new discoveries concerning the properties of dark matter and energy appear to offer unexpected windows into the Planck scale. It may be that Greene does not mention these experiments because they are so new, and it is certainly too soon to draw definitive conclusions from them. But they are, very possibly, the beginning of a new chapter in fundamental physics, one driven by surprises forced on us by experiment rather than by the fantasies of theorists. Perhaps they will be the subject of Greene's next book.