Friday, July 20, 2007

McGrath on God's improbability

At one point in the interview posted below (at 9mins 15secs - 9mins 55secs) McGrath (see, I can spell it correctly) says something like: the issue of God's improbability is not really the issue. The question is, does he exist. After all, our existence is also extremely improbable (what are the odds on my parents meeting, exactly that sperm fertilizing that egg, etc.), yet we know (can be quite sure) we exist.

What is going on here? Seems to me there's some sleight of hand going on with the notion of probability. In fact he's muddling objective and epistemic improbability. But what has gone wrong exactly? Comments?

4 comments:

ajn said...

I think you've got it right. In "real world" terms, without getting lost in some Cartesian demonic brain-in-a-vat philosophical thought experiment, I have certain knowledge that I exist, that you exist, that there is a woman in Ethiopia taller than six feet, that there is an albino man in India, etc. I have no personal knowledge of the latter two, but I would be astounded if they didn't exist given my knowledge of the incidence of albinism and the variations in human height.

There's probably a man in Mauritius who has read "Tristram Shandy" - I don't know for sure, but it seems probable given that there are well over a million people there and it's a reasonably popular book. I would be surprised if I was wrong, but not utterly astonished.

As far as God's concerned, in the absence of evidence many of us find his existence far, far more improbable than the Mauritian Shandean, to the point where unless we're being anally pedantic we say the probability is zero and he doesn't exist.

That is (as you say) a different kind of improbability than the "what are the odds of that happening" improbability. In that sense, it is improbable that I exist (because there were however many million or billion sperm which could have fertilised my mother's egg, and any of the others would have resulted in a different person). In the former sense, it's not at all improbable that I exist.

Incidentally, McGrath has a different concept of "evidence" to most of us. I had a longish discussion with a Christian friend who thought "Dawkins' God" was excellent - we came to the conclusion that although Dawkins bases a lot of his arguments on a selective interpretation of the word "faith" which has very little to do with the way Christians use the word, McGrath uses a particularly stupid form of circular reasoning where his faith is based on evidence, and the evidence is anything that gives him faith (i.e. it is evidence for him, not evidence which anyone else would necessarily accept).

ajn said...

To expand on McGrath's "evidence" - what he's saying in DG, as far as I can tell is:

1. Dawkins says "faith" means "belief in the absence of evidence".
2. This is not the only meaning of the word "faith".
3. Christians normally, or often, mean one of the other senses of the word (for example trust, certainty, firmness of belief).
4. My (i.e. McGrath's) faith is based on good evidence.
5. I know the evidence is good, because it gives me faith.

I think 1-3 are correct, 4 and 5 are the circular argument. The problem that Dawkins (and others) have when arguing with religious believers is that believers think that they have good evidence for their beliefs, and that their faith is based on evidence, so the two sides are arguing at cross-purposes. The argument ought to be about what constitutes good evidence for belief, not on the definition of the word "faith" - simply saying "faith means belief in the absence of evidence and you admit you have faith so you admit there is no evidence" is no more convincing to a theist than "the proof that God exists is that I believe in him" is convincing to an atheist.

Anonymous said...

Stephen -

You have it exactly right in my opinion.

(A) P(I exist) is indeed very small as it is dependent on so many contingent events as described.

However, (B) P(I exist | My parents met, fell in love, got married, weren't infertile...plus 1000 other factors) is very high. Hence, I'm not surprised about my existence, because all the things on the right are known background events.

And personally, I would start my chain of Bayesian reasoning from B, and if I want to work out (A), I back out the required probabilities.

It gets trickier (as always) when we turn to deities:

(C) P(G exists) is clearly an epistemic probability and if you condition on nothing at all, then you can only justify it being high on the grounds of divine simplicity (itself controversial)

The thing is, if you're going to go Bayesian, you start by conditioning on *everything* you know already.

Hence

(D) P(G exists | every piece of evidence that supports naturalism, neo-darwinism, Problem of evil, aruments against moral naturalism, euthyphro dilemma, divine hiddenness, failure of ontological arguments etc....)

I have my doubts that (D) is high, and I'm not alone in thinking that. But it's D that's the right probability to consider. If I want to apply Bayes in reverse and back out (C), then I can do so. But (C) simply doesn't apply to the epistemic situation I'm in.

(Obviously, everything I'm conditioning on in (D) can be debated. But that's not the point of this comment...)

snafu

onthesideoftheangels said...

I have major problems with this line of thought:

Primarily the 'probability of life' calculations are grounded in speculation, reverse induction, the fallacies of potentia cum/ab/ex/post hoc ergo propter [potentia] hoc [i.e presuming possible alternate causes; complementary or supplementary promoters/inhibitors ;effects;offshoots; ignorance of affectors/actuators etc.] I also believe it involves what I call 'fallacies in the momentum of causation'; but few experts would concur.

One of the best way of explaining it is by resorting to works of past genius:

www.online-literature.com/chesterton/2573

To suggest this even encroaches upon objective probability
is specious and frankly erroneous. Part of the problem is what
I call the iguanadon horn fallacy - forcing individually coherent aspects into a 'pleasing/conducive' whole similar to other shared-predicate phenomena; without provisions for holistic cohesion.
Life exists - therefore any 'thinking backwards' for probability will be contaminated by this existence.
Chemistry, let alone biochemistry, is hardly conducive to mathematics - there are almost miraculous seemingly unrelated 'conditionals' that come into play - statistical probability would have no problems with aluminium carbonate: earth physics does! Stats would think of one colour for oxidised chromium ions rather than the whole spectrum in reality. Adding sulphuric acid to an acyl chloride should mathematically have but one result- not the three utterly different results dependant upon temperature. See where I'm coming from ? Probability of Life is a worthy exercise , a bit of fun where we can hone and clarify certain aspects of it - but as a whole it is far from objective because we are utterly ignorant of too many factors and conditionals.

Probability of God : A futile endeavour ; because due to the nature of the argument 'God' becomes an 'Apeironic' [i.e. external to the cosmos] collation of disputed/contested super-predicates - a polymorphic cognomen !
Because of the proposition, inherant within the 'nature of the beast' there is an ironic axiom which leads to one thing - ultracontrariety of predication.
Either everything proves the existence of God, or everything proves the non-existence of God. There is no other possibility.

Because of the Apeironic nature of the epistemological construct - nothing cosmic can either prove or disprove the proposition - the only thing that can be negated is the nature of the proposition and the authenticity of the collective predicates involved within it.
It's a neat paradox of perspective where even though there is ostensibly debate and argument; in reality it's like black and white arguing that green's 'varying shadows of greyness' is a priori self-evidential proof of its absolute whiteness or blackness.

So , due to the nature of the argument, any probability always collapses to evens. Dawkins can speak of extreme remote probabiltity verging on the point of negation; but the only reason he can do such is by denying the nature of the argument and inventing his own through the distortion of 'antinomies' e.g. 'faith is belief in the absence of evidence' - this is incredibly inane ; for by the very nature of faith the only thing that cannot be absent is evidence - it's what the evidence is 'evidential of' that is the problem ; and that is invariably dependant upon the predilections of the discerner.