Author and cosmologist Paul Davies has got a new book out, The Goldilocks Enigma: Why is the Universe Just Right for Life? Like other recent books, it touches on the Anthropic Principle, the idea that we can learn something deep about the universe from the simple observation that we — sentient, water- and carbon-based life — exist. To some authors like Peter Woit and Lee Smolin, that some aspects of modern physics may lead us to the Anthropic Principle blights the whole fabric of current research. Davies takes it as an opportunity, in ways that I’m not comfortable with, even if I don’t want to damn the whole program along with the curmudgeonly Woit and Smolin. I’ve reviewed Davies’ book for Physics World, but I am vain enough to prefer the following unedited version.

Paul Davies, The Goldilocks Enigma

In their 1990 text, The Early Universe, Rocky Kolb and Mike Turner write “It is unclear to one of the authors how a concept as lame as the ‘anthropic idea’ was ever elevated to the status of a principle.” Much to that author’s chagrin, we are now seeing a renaissance in discussions of that idea, lame or otherwise.

The anthropic principle (pace Rocky and/or Mike), as originally expressed by Brandon Carter in 1974, states that “what we can expect to observe must be restricted by the conditions necessary for our presence as observers.” That is, given that we — life in general, an intelligent species in particular — are here, we already know that the Universe must have had properties (values of coupling constants, number of spacetime dimensions, magnitude of the cosmological constant, …) such that [intelligent] life could evolve. Is this a vacuous statement, a tautology? Is it at least descriptive? Or is it predictive — can we get a better handle on fundamental theories armed with this knowledge?

Paul Davies’ new book, The Goldilocks Enigma, starts by inverting the question: why are those properties so perfectly suited to life? In the canonical example, Hoyle pointed out that the formation of heavy elements in stars depends exquisitely upon thermonuclear reaction rates — tiny deviations in either direction and stars wouldn’t be able to burn helium into the carbon of which all life (with the usual “as we know it” caveat that lurks behind these arguments) is made. But these rates aren’t written in any obvious way into the fabric of the standard model of particle physics. They just emerge in an as-yet incalculable way from the more fundamental parameters of the model, or the even more fundamental ones lurking underneath in string theory or whatever.

These questions have come to the fore over the last decade as string theory has morphed into M theory with its landscape of 10²⁰⁰ possible vacua, each with its own fundamental parameters. Is the real world somehow uniquely determined or are we simply a random sample from that vast set of possibilities? Or are each of those vacua somehow actualized in a gargantuan multiverse? (Multiverses also arise in even more fanciful theories such as the many-worlds interpretation of quantum mechanics, although I’m pretty sure that the word first gained currency in 1970s comic books.)

The first half of the book is a workmanlike introduction to the standard models of particle physics and of cosmology. This material has probably been treated as least as well elsewhere, but is necessary background for the discussion of the anthropic argument itself.

The anthropic argument, however, is entirely devoted to distinctions without differences: choosing between theories that, by definition, have identical observational consequences (they all can at least possibly produce our world as we’ve currently been able to observe it). The upshot of these ongoing discussions of the anthropic principle is that we’re still too ignorant to make any definitive statements. In the Bayesian parlance implicit in most such discussion, we can’t even arrive at an appropriate prior to encode the knowledge that we do have (although Radford Neal, in a recent paper, “Puzzles of Anthropic Resolved Using Full Non-Indexical Conditioning”, math.ST/0608592, gives it a stab).

Still, Davies himself is convinced that these arguments have force. Rather than a unique, “just-so” universe that will only be understandable in the light of a truly final theory, or an “absurd” universe that just happens to have its properties randomly dialled-in for no underlying reason at all, he prefers what he proudly refers to as teleological explanations: a so-called “life principle” arising out of physical laws that only become definitely determined as the universe evolves, or the even more mysterious “closed explanatory loop” adapted from J.A. Wheeler that somehow brings itself into existence with meaning as well as physics: “the universe explains observers, and observers explain the universe.”

Davies says, “nobody would deny that atoms, stars and galaxies are fundamental features of the Universe.” He needs to be careful of his use of “fundamental”; clearly these things are emergent properties of the universe – and it is exactly this emergence out of quarks, or strings, or whatever lies at the bottom that gives force to anthropic arguments at all. He goes on “it seems clear that life (and mind… and culture, too) is an equally significant step on the path of cosmic evolution.” That’s certainly true, at least if you are one of the living beings, with a mind, inside the only culture we know of. But just as the Universe doesn’t need an “atom principle” to ensure that atoms are formed, or even a “carbon principle” to enforce the details of stellar nucleosynthesis worried about by Hoyle, it’s not clear that we should elevate the existence of life to something in a need of a proximate cause separate from the laws of physics themselves, the same laws responsible for those atoms and nuclei. Just because we’re not yet able to understand why the universe supports life, doesn’t mean we need some external “life principle” behind it. Indeed, this is just the criticism correctly levelled by Davies himself against the proponents of Intelligent Design: just because we’re not yet able to understand the evolution of (say) the mechanism of the flagellum, doesn’t mean it didn’t evolve.

Still, these ideas are entertaining, if only barely coherent. And along the way, Davies highlights some of their weirder repercussions. Chief among them is something that philosopher Nick Bostrom has called the “Presumptuous Philosopher Problem”: if you take anthropic arguments at face value, they give you one factor in a “prior probability” for different theories of the Universe, proportional to the number of intelligent species (or possibly human-like species) that are predicted in such a theory. A realistic calculation is impossible with current knowledge, but the rub is that we should a priori prefer theories that predict more intelligences, rather than fewer — which probably means that all else being equal, we should prefer vast universes to tiny ones, perhaps by odds that no observational evidence could possibly overcome. In particular, Bostrom (and Davies) uses this to point out that we should probably take seriously the possibility that we are living, Matrix-like, in a computer simulation, since (if they are possible at all — another hotly-debated philosophical topic which I shall ignore) simulated minds could be more common than real ones.

Getting to this point, however, we realize that we’ve circled round to something philosophers have been worrying about for longer than physicists: how do I know I’m not a ‘brain in a vat’? This strain of scepticism started with Descartes and was only quashed by Wittgenstein and his followers in the last century. Wittgenstein famously said, “Whereof one cannot speak, thereof one must be silent” and perhaps, in the absence of better theories and data to support them, that’s how physicists should treat anthropic questions.