What the hell is The Anthropic Principle? This page is my attempt to explain it, and some of its applications, to a lay audience.
There is no single thing called The Anthropic Principle. There are probably dozens of "anthropic principles", all different, many confused, some totally absurd. However, I shall pretend that "The Anthropic Principle" refers only to anthropic reasoning as originally formulated by Brendon Carter back in 1974. Carter writes -
The gist of this is that conscious beings like us will only ever be found in those parts of reality that allow conscious beings to exist. Therefore, certain things may forever be unobservable, since no observer could ever be in a position to observe them. This is obvious enough. To say it another way, this is Carter's Weak Anthropic Principle or WAP -
I will now try and apply this type of reasoning to some problems in physics.
Many physicists assume something called the Mediocrity Principle. This is the idea that humans and the Earth are not in a special location in the universe. From this principle, scientists assume that the laws of nature we discover are applicable throughout the entire universe.
Yet it is obvious that humans are in a special location. We aren't in some random area of space, but instead are on a planet in a stable orbit around a rather stable star, at a distance where water exists in liquid form. This is a very special location. There is nothing mysterious about why we don't find ourselves floating in a random area of space: life couldn't exist in such a location.
On a grander scale, it's also possible (at least in principle) that we are in some special region of the universe. Suppose that (for whatever reason) 99% of the universe couldn't ever contain planets like Earth. We would naturally find ourselves in the 1% that could. Thus, it is entirely possible that we are, in fact, in a special region of the universe. There is, however, little or no evidence for such a belief. But I am only trying to show why it is not prima facie absurd. The Anthropic Principle reminds us that even if conditions suitable for life are rare, that is where we will be. We cannot assume that we are in a typical location.
There is a problem in physical cosmology called Fine Tuning. To understand Fine Tuning, you need to understand this:
There are about 20 constants in physics that cannot be deduced from first principles. They must instead be determined by experiment. There is no known reason why these constants should have the values they do. For many of these constants, if their value was slightly different, life wouldn't exist at all. Instead, the universe would be a cloud of hydrogen, or nothing but waves of radiation, or in any case not hospitable to life.
So, this looks like it can't be a coincidence. In fact, it seems like one of the strongest arguments for the existence of God. It appears that the universe has been designed - or fine tuned - for life. The constants of physics look like they were chosen by a designer.
But there's an alternative explanation. Suppose there is more than one universe. Suppose there are a great many. Most of these universes will have physical constants unsuitable for life, but a few will happen to have physical constants that permit conscious beings to exist. We naturally find ourselves in such a universe, because we couldn't be in a universe unsuitable for life.
This latter argument implicitly relies on The Anthropic Principle, which reminds us that our location as observers may well be special, and we cannot automatically rule out the presence of vast swathes of reality that are unsuitable for conscious beings to exist in.
I should note that some writers claim that this sort of argument for multiple universes commits a logical fallacy known as the Inverse Gambler's Fallacy. It doesn't, but the reasons why are sufficiently complex that I deal with it on a separate page.
A common complaint of Creationists is that abiogenesis (the formation of life from non-living matter) is so implausible that it must have had divine help. However, Creationists tend to forget just how big the universe is. The visible universe has about 30,000,000,000,000,000,000,000 stars in it. And those are just the stars close enough to be seen.
Even if the emergence of life on any one planet is ridiculously unlikely, the universe is big enough that the odds of it emerging somewhere are much better. The Anthropic Principle reminds us that, even if life will only be found on one planet in a trillion, those are the planets that observers will find themselves on. Thus, it's entirely reasonable that observers will exist in places where the chances of them existing were absurdly low - because the chances of them existing somewhere were not so low.
A natural existential risk is any natural phenomenon that could wipe out our civilisation. One argument against such concerns is that such risks must be unlikely, since such catastrophes have only rarely affected Earth.
Unfortunately, The Anthropic Principle tells us to be wary of this reasoning. It's conceivable that intelligent life can only develop given an unusually long run of good luck. Perhaps most planets suffer catastrophic extinction events every few million years, and it's only through extreme good fortune that intelligent life can evolve at all. In that case, all intelligent observers would exist on planets like ours where - through sheer chance - things went particularly smoothly. But there would be no reason to expect such a run of good luck to continue.
However, this idea could be ruled out if we could show that the Earth has indeed had its fair share of catastrophes; or if we could show that catastrophes have been uncommon due to some inherent fact about the Earth (its location in the galaxy, say) rather than mere luck.
I noted earlier that there are a great many "anthropic principles". Many people react strongly to the more dubious ones and dismiss talk of The Anthropic Principle as nonsense. This is unfortunate. Carter's Weak Anthropic Principle is so obvious that it is sometimes criticised for being a mere truism. Yet we've seen above that it has definite applications.
I think the reason anthropic principles provoke such a strong rejection from some people is that many are phrased in terms of "how the universe must be". The universe must be this, the universe must be that. The words "must be" are sufficiently ambiguous that some people interpret them to mean the universe couldn't have been any other way.
I particularly want to warn readers against some corrupt anthropic principles that are not pure logical truths the way Carter's WAP is, but rather theories about the world. Such theories may or may not be true, but they are certainly not provable through logic alone. Here is Barrow and Tipler's "Final Anthropic Principle" -
This is highly dubious. There is obviously no in-principle reason why life must come into being, nor is there any reason why it cannot go extinct later. I should note that Tipler in particular is infamous for trying to smuggle Christianity into physics. It's regrettable that such authors call their ideas "anthropic principles" when they are very far from the spirit of the real Anthropic Principle.
Nothing I've said is particularly original, and in fact I'm quite close to the discussion in chapter 3 of Nick Bostrom's book Anthropic Bias, which is an excellent but rather difficult work. In particular, I should warn you that halfway through the book, Bostrom temporarily accepts certain conclusions that he later re-examines and rejects. If you don't make it to the end you might therefore end up with certain dubious ideas.
I thought my section on natural existential risks was an original idea of mine. Then four days later my copy of Global Catastrophic Risks arrived, containing Observation selection effects and global catastrophic risks by Milan Cirkovic. Alas.
Brandon Carter's paper is called Large number coincidences and the anthropic principle in cosmology, and is found in Physical Cosmology and Philosophy edited by John Leslie.
Wikipedia's page on the subject might also be worth a look.
Updated: 2008-07-28
allancrossman.com