paul_j_steinhardt's picture
Albert Einstein Professor in Science, Departments of Physics and Astrophysical Sciences, Princeton University; Coauthor, Endless Universe
Albert Einstein Professor of Physics, Princeton University.

I believe that our universe is not accidental, but I cannot prove it. 

Historically, most physicists have shared this point-of-view. For centuries, most of us have believed that the universe is governed by a simple set of physical laws that are the same everywhere and that these laws derive from a simple unified theory. 

However, in the last few years, an increasing number of my most respected colleagues have become enamored with the anthropic principle—the idea that there is an enormous multiplicity of universes with widely different physical properties and the properties of our particular observable universe arise from pure accident. The only special feature of our universe is that its properties are compatible with the evolution of intelligent life. The change in attitude is motivated, in part, by the failure to date to find a unified theory that predicts our universe as the unique possibility. According to some recent calculations, the current best hope for a unified theory—superstring theory—allows an exponentially large number of different universes, most of which look nothing like our own. String theorists have turned to the anthropic principle for salvation. 

Frankly, I view this as an act of desperation. I don't have much patience for the anthropic principle. I think the concept is, at heart, non-scientific. A proper scientific theory is based on testable assumptions and is judged by its predictive power. The anthropic principle makes an enormous number of assumptions—regarding the existence of multiple universes, a random creation process, probability distributions that determine the likelihood of different features, etc.—none of which are testable because they entail hypothetical regions of spacetime that are forever beyond the reach of observation. As for predictions, there are very few, if any. In the case of string theory, the principle is invoked only to explain known observations, not to predict new ones. (In other versions of the anthropic principle where predictions are made, the predictions have proven to be wrong. Some physicists cite the recent evidence for a cosmological constant as having anticipated by anthropic argument; however, the observed value does not agree with the anthropically predicted value.)

I find the desperation especially unwarranted since I see no evidence that our universe arose by a random process. Quite the contrary, recent observations and experiments suggest that our universe is extremely simple. The distribution of matter and energy is remarkably uniform. The hierarchy of complex structures ranging from galaxy clusters to subnuclear particles can all be described in terms of a few dozen elementary constituents and less than a handful of forces, all related by simple symmetries. A simple universe demands a simple explanation. Why do we need to postulate an infinite number of universes with all sorts of different properties just to explain our one? 

Of course, my colleagues and I are anxious for further reductionism. But I view the current failure of string theory to find a unique universe simply as a sign that our understanding of string theory is still immature (or perhaps that string theory is wrong). Decades from now, I hope that physicists will be pursuing once again their dreams of a truly scientific "final theory" and will look back at the current anthropic craze as millennial madness.