Re: [asa] Calculation of probability for life to originate on Earth unintelligently

Actually, I agree with you totally, but I think you're missing my point.

Furthermore, I think I can see where the confusion lies.

I agree totally that we can't compute the probability of life getting
started because we don't know the mechanism.

But what we can compute is a _conditional_ probability, P(X|H) which we'd
read as P of X given H. H is a hypothesis as to the mechanism of life
getting started, and X is the event that life gets started.

Koonin computes the conditional probability on the hypothesis that for life
to get started, a replication/translation unit has to come together by
chance, and once such an event happens, conventional evolution can take
over. With out the replication ability, conventional evolution is not
possible.

Given that hypothesis Koonin then computes P(X|H) and finds it to be
extremely small. There are two possible conclusions one may draw from this.

(1) H is an incorrect hypothesis. I think we both agree that this is the
conclusion we'd draw. We need to find a better hypothesis, a new mechanism
that makes it much more probable.
(2) Ours isn't the only universe - in most of them life doesn't get
started, but ours was the lucky jackpot winner, which is why we're here to
speculate on the question. (Anthropic principle).

But the point I'd make is that in making the deduction (1) or (2) we've
concluded that the probability 10^(-1018) is unacceptably low. So it is
valid to base an argument on low probabilities. That's why I think that to
dismiss arguments based on low probability based on your bridge hand example
is a silly example and one you should refrain from using, because it
involves specifying the target after the event. There is something quite
special about that 1 out of 10^1018 possibilities. It happens to be the
only one in which we're here to discuss the problem.

All scientific laws are deduced by observing that data conforms to a
pattern, and then making predictions by testing to see if future
measurements conform to the pattern.

Suppose I deal out 16 cards from the top of a pack, and you look at the
suits. You get the sequence:

Spade, Heart, Diamond, Club,
Spade, Heart, Diamond, Club,
Spade, Heart, Diamond, Club,
Spade, Heart, Diamond, Club

Then I stop and ask you what is the suit of the next card in the pack. Now
the probability of getting that sequence is 1 in 4^16. But the probability
of getting any other sequence is also 1 in 4^16 (apart from impossible
sequences where there are more than 13 of one suit).

So are you going to say that you haven't any idea what the suit of the 17th
card is, and that you don't suspect I've messed with the order? I suggest
you've formed the hypothesis that I've deliberately arranged the pack so the
sequence repeats, and given that hypothesis, the probability that the next
card is a spade is 1. When I deal the next card your hypothesis may have
more evidence, or it may be falsified. Of course it's not the only
hypothesis you could form. You might decide I'm a mean critter who
deliberately decided to trick you, in which the probability of a spade would
be zero under that hypothesis. But I think you'd be unlikely to say that
the probability is 1/4, given the unusual sequence that preceded it.

Iain

On Nov 12, 2007 6:39 PM, D. F. Siemens, Jr. <dfsiemensjr@juno.com> wrote:

> Iain,
> I think your missing my point. What we have is an end point with a
> probability of 1. We don't know the starting point. We don't know the
> process, but every little while we get a new insight that provides an
> additional mechanism. If we were computing probabilities on the basis of
> what was known about a century and a half ago, we'd have given the
> probability of evolution at ~0 or 10^(infinity), for it was held to be a
> logical impossibility. We now get someone who calculates ~10^1000. Want to
> extrapolate the rate of change to the next century or even decade?
> Dave (ASA)
>
> On Mon, 12 Nov 2007 14:44:25 +0000 "Iain Strachan" <igd.strachan@gmail.com>
> writes:
>
>
>
>
> On Nov 12, 2007 1:43 PM, Alexanian, Moorad <alexanian@uncw.edu> wrote:
>
> > I do not think anyone knows how to calculate possible outcomes from
> > given initial conditions. Unless one can know what the possible outcomes can
> > be, how is one to calculate probabilities for individual outcomes? One does
> > not know dynamics that well to make any sort of predications.
> >
> I think this is rather missing the point of Koonin's argument.
>
> His argument is to take the current front-runner theory of the origin of
> life (RNA - world), and the assumption that in order to kick start the
> process of evolution, we need to have translation and replication in
> process. This is a chance event that has to happen once only. Then, given
> generous estimates of RNA replication rates, and of the number of planets in
> the observable universe, ... under all these assumptions, then he finds that
> the probability of this chance event, for which he has done a conservative
> estimate, is vanishingly small for a single universe. And what you have is
> a conservative estimate, with RNA jiggling about on every square millimetre
> of every planet in the universe whether hospitable for life or not. The
> real probability (which we can't estimate) is likely much smaller, but we've
> at least had a shot at an upper bound.
>
> The reason I don't find this valid is probably the same as yours, David's
> and Michaels - ie that we don't know how it might have happened, and that
> therefore, if the current model gives a vanishingly small probability of it
> happening then the current model is wrong and we haven't solved the problem.
> A better theory will no doubt emerge in time as we get to know more, which
> gives a much higher probability of life starting up. The low probability
> argument proves neither the multiverse hypothesis, nor the Intelligent
> design hypothesis. All it proves is that we don't know yet how life got
> started up.
>
> What I was objecting to is the notion that you can airily dismiss it by
> saying "oh... well extremely unlikely events happen all the time". That's
> what I think is a silly argument. Supposed you were facing a firing squad,
> or marksmen whose guns were loaded with live or blank ammunition based on a
> coin-toss. 100 bangs go off and you walk away unscathed. Are you really
> going to walk away with a shrug of the shoulders and say "well, 100 heads is
> just as likely as any other combination, so I'm not surprised I escaped at
> all"? Of course not, because with any of the other 2^100-1 (equally likely)
> possibilities you are going to end up dead.
>
> Iain
>
>
>

-- 
-----------
After the game, the King and the pawn go back in the same box.
- Italian Proverb
-----------
To unsubscribe, send a message to majordomo@calvin.edu with
"unsubscribe asa" (no quotes) as the body of the message.