09 January, 2006

Intelligent Design and probability

בְּרֵאשִׁית, בָּרָא אֱלֹהִים, אֵת הַשָּׁמַיִם, וְאֵת הָאָרֶץ.
In the beginning God created the heavens and the earth.
— Genesis 1:1

Never get so attached to a poem
you forget truth that lacks lyricism
Joanna Newsom

A long post this morning. For those uninterested in the topics, I highly recommend forwarding to the * for a nihilist giggle.

I hesitate to dip my toes into the turbid puddle of today's religion / science contretemps; but alas, it has become difficult to browse or prowl the web without turning up prosy brawls over Intelligent Design. Furthermore, after perusing a copy of the Skeptic magazine lying around my in-laws' Fairfax residence, I came fascinatedly to realise the extent to which Creationists will go in defence of their tenets. Some of the pseudo-statistical claptrap put forward by Bill Dembski, for instance, is rather engaging.

But unlike the liberal humanist brigades and free-thought societies, I am not disgusted by the very fact of this debate's existence. Evolutionary biology is taken on faith by most atheists (myself included) who haven't the time or the interest to get to the bottom of the hard theory and empirical data involved in the science. As at so many times in history, the conflict of old and new ideologies has brought fresh ideas to both sides, and an awareness more acute of the relevant materials by the general populace—which can only be beneficial to genuine free thought and thus to the education of our youth. Dogmatism, even dogmatism about the scientific endeavour, is a poor attitude for anyone: it is the reasoning process, not its result, which is most important to teach. Added to which, debates over ID have disseminated interesting arguments in a number of fields, including epistemology and probability theory.

Probability is a subject which has always intrigued me. I recall asking one of my A-Level maths teachers how there could be a 50% chance each of head and tail, when there were so many things a coin could (theoretically) do other than yield either result: any number of finitely-improbable events involving bizarre atomic motions consistent with the laws of physics. The lesson from this conundrum, which I confess that I don't fully understand, is that probability has no a priori, objective existence: it is merely constructed subjectively by humans to model real problems.

Where does that leave those who argue for ID from the relative probabilities of Design and Chance as explanations for what we see around us? This rude mechanical has heaped up quotations by famous scientists—including Richard Dawkins!—to show that it isn't just religious nutnuts who see intentional design behind the 'incredible' unlikeliness of Life. But was the probability of that first spark of life, and of its subsequent evolutions, really so small? Careful articles by Sober and Ikeda / Jefferys have addressed the subtleties of this and other questions of probability which are raised by Intelligent Design.

As Sober's article points out, there are several ID arguments from probability. The one which most caught my eye claims Design as the only possible (or probable) explanation of the fact that the physical constants of the universe, such as the strength of fundamental forces, are exactly right ('fine-tuned') for the existence of life. Man, in other words, has a privileged place in existence, and thus it is more probable that the universe was designed, and designed for man, than that it came about by chance. Such a teleological argument is immediately odd. First of all, what makes life (at least as we know it, Jim) special or noteworthy as a phenomenon? Although any religious value-system will privilege living over non-living entities, there is nothing objective about such a stance—no scientific reason to regard life as cosmologically important. Second, if there is an infinite amount of time, whether in many universes or in one, any event, no matter how unlikely, is bound to occur. We are accustomed to think of time in human terms, as a duration over which we wait for a given event; but we cannot conceive pre-human time in this way, as it has no determinable beginning or end. In fairness, I should mention that Sober rejects a similar argument, though I am not entirely convinced by his refutation. Third, any particular permutation of values in a set of independent variables is equally probable: if we flip a coin twenty times,


is just as probable as


The first seems to be a coincidence, the second merely a random incidence of coins; but the Markovian nature of coin-flipping ensures that both specific outcomes have the same probability, namely 0.00000095367431640625 (0.520). There can be no privileging by blind chance of systems which look chaotic (2) over those which look ordered (1). Similarly, an arrangement of particles, forces and constants which produces life (such as this one) is no less likely than any other given arrangement. As Ikeda / Jefferys put it, 'Most actual outcomes are, in fact, highly improbable, but it does not follow that the hypotheses that they are conditioned upon [ie. the undesigned origin of the universe] are themselves highly improbable'.

Fourth, and most importantly, the ID reasoning is backwards. It has made claims about the probability of what we know to be the case, without taking into account the (weak) anthropic principle, namely the fact of our existence—an epistemic frame which skews our ability to make probability-claims on the subject. Sober treats this at length, invoking Eddington's principle of the OSE (observational selection effect). Naturally, his article is much more sophisticated than what I can offer here, but his basic argument is as follows:

X: The constants of the laws of physics are exactly right for life.
D: The universe is the product of intelligent design.
C: The universe is the product of blind chance.
A: We exist, and if we exist the constants must be right (the anthropic principle).

The ID-theorist seeks to assert

(1): P(X • D) > P(X • C)

The probability of X given D is greater than the probability of X given C.

But since A, the fact of our existence, must be taken into account, (1) is true but not explanatory, ie. not a valuable model. Instead we propose

(2): P(X • D & A) = P(X • C & A) = 1

The probability of X given both D and A is the same as the probability of X given both C and A, both being equal to one. Both scenarios yield (and must yield) what we know to be the case, namely that we exist. We have no basis, on this argument, for deciding between D and C, between design and chance. Ikeda / Jefferys argue along similar lines; according to their paper, the ID-theorist has moved from the assertion (taken to be true) that 'P(X • C) is small' to the false assertion that 'P(C • X) is small', ie. that the probability of an undesigned universe, given the exactness of its physical constants, is small. As the authors observe, this move is an 'elementary blunder in probability theory'.


Finally, an entertaining argument against Intelligent Design, which also happens to be the saddest joke ever told. It is an anecdote recounted in Samuel Beckett's Endgame:

An Englishman, needing a pair of striped trousers in a hurry for the New Year festivities, goes to his tailor who takes his measurements. "That's the lot, come back in four days, I'll have it ready." Good.

Four days later. "So sorry, come back in a week, I've made a mess of the seat." Good, that's all right, a neat seat can be very ticklish.

A week later. "Frightfully sorry, come back in ten days, I've made a hash of the crotch." Good, can't be helped, a snug crotch is always a teaser.

Ten days later. "Dreadfully sorry, come back in a fortnight, I've made a balls of the fly." Good, at a pinch, a smart fly is a stiff proposition.

(I never told it worse. . . I tell this story worse and worse.)

Well, to make it short, the bluebells are blowing and he ballockses the buttonholes. "God damn you to hell, Sir, no, it's indecent, there are limits! In six days, do you hear me, six days, God made the world. Yes Sir, no less Sir, the WORLD! And you are not bloody well capable of making me a pair of trousers in three months!"

"But my dear Sir, my dear Sir, look—

—at the world—

—and look—

—at my TROUSERS!"

1 comment:

Anonymous said...

"Second, if there is an infinite amount of time, whether in many universes or in one, any event, no matter how unlikely, is bound to occur..."

First off, this is untrue. I think you fail to understand some of the probability arguments. For example, the arguments stemming from the information imbedded in DNA. You point out that all outcomes are equally probable, which is true. But it is infintely improbable that chance would produce the outcome that held enormous amounts of information. At a certain point, we can find facts to support whatever we wish to believe, but when you step back, it seems evident on many levels that there is design in nature, indicating a designer. I would recommend reading Darwin's Black Box as well, which deals with biological systems that could not concievably have come about from strict Darwinin (or neo-Darwinian) evolution.