Re: [asa] Re: ID without specifying the intelligence?

On 9/16/07, David Clounch <david.clounch@gmail.com> wrote:
>
> Iain,
> You are neither sanctimonious, nor a prig. BTW, what's a prig? :)

From the web:

*1. * A person who demonstrates an exaggerated conformity or propriety,
especially in an irritatingly arrogant or smug manner.

That is apparently how Michael sees me. It is a great puzzle to me that
Michael thinks it's perfectly acceptable to call someone a fool, even hiding
behind a scriptural quote to do so, but if I tell him he ought to grow up a
bit, I get called a sanctimonious prig. It is good to know that at least
some people appreciate my contributions.

>So I get a bit suspicious of ID because it would seem to be aiming to prove
> the existence of a Designer.
>
> Be suspicious of the IDM (ID Movement).
>
> IDT (ID Theory), on the other hand, need rigorous definitions of its
> components in order to even be evaluated.

Sorry about the imprecise terminology.

>
> And if one wants to discuss the IDM, one has to identify the various
> groups and beliefs. Its about people.
> IDT, on the other hand is about math.

Yes, in one sense you are right & I was initially attracted to it because of
my own academic area of statistical pattern recognition. Many of the
concepts are ones that are familiar to me - compressibility of data,
Kolmogorov complexity and so forth.

However, the deductions that are made from the math ( existence of a
Designer) are philosophical, and, for the reasons I've given recently about
Universal Explanatory Mechanisms, they are IMO unscientific.

Consider one of the alternatives. You come across something that looks
highly improbable, given your current state of knowledge. So you say "it's
just coincidence". Anything with a finite probability could happen, and
coincidence can't be ruled out. Even if the probability is below Dembski's
Universal Probability Bound of 10^(-150) you still can't rule out
coincidence. Toss a coin 1000 times and you've just witnessed an event with
probability around 10^(-300).

That's where Dembski's notion of Complex Specified Information. The problem
with the 1000 coin tosses is that it is "unspecified". However, if there
were a significant pattern in the data, then you have specification - the
data conforms to a pattern, and therefore only a subset of the possible
combinations will conform to the pattern. You get a realistic probability
by taking the number of sequences in the subset and divide by the total
number of sequences. If this is < 10^(-150) then you have observed
something of note that requires explanation.

There is a further subtlety - if the pattern is too simple (e.g. all heads),
then there is likely to be a simple natural explanation (like it was a
double headed coin). So, according to Dembski, the data has to be both
specified AND complex. An example he gives, I believe, is a Shakespeare
Sonnet. It doesn't admit to an extremely compact description (like "All
heads"), and is a sequence that is not predictable, but nonetheless shows
patterns (for example, Markov models can be used to learn the
characteristics of written text).

But where I run into a big problem with the whole thing is that given
something like DNA and observing that complex specified information is
present, that we must infer Intelligent Design, by analogy with what we know
of Shakespeare text. Just because the two have complex probability
distributions, doesn't mean they are both designed.

My reason for saying that is that I tend to espouse a Bayesian view of
probability. In order to make an inference, (that something was designed),
one has to have some idea of the prior probability of the existence of the
Designer. Now, we know that Shakespeare existed, so the prior probability
of the existence of the designer is 1. However, if we take the idea that
the patterns in DNA also indicate a designer, then we have to assign a prior
probability to the existence of that designer. Suppose the designer is a
sufficiently advanced and intelligent Extra-Terrestrial being. If Koonin's
calculations are right, then that prior probability is less than 10^(-1000)
& hence this is going to affect our ability to make a design inference. If
we say God is the Designer, then science gives us no idea at all how to
ascribe a probability to God existing, so we can't do the calculation at
all.

If we go back to Paley's Watchmaker, we run into the same problem. The
reason we can infer the design of the watch is that we know with probability
1 that watchmakers exist. But consider the situation where an alien being
lands in the middle of a desert and sees a watch & examines its intricate
structure. ET's form of life is absolutely nothing like ours - he's not
even carbon based & he's never seen a human or animal. Is he then justified
in making a Design inference, based on what he knows? I don't think so. I
think an equally reasonable conclusion would be like "Bones" in "Star Trek"
: "It's life, Jim, but not as we know it". It's possible that the only
way to discount the "life" inference would be to examine the watch carefully
and show that it had no way of reproducing itself.

So that's the problem I have with IDT (got it right that time!). It's not
the maths behind it. But all the maths can do is detect whether a pattern
exists in data. I do the same thing for a living. But the problem occurs
when the existence of the pattern (sufficiently "complex"), is taken to
imply the existence of the Intelligent Designer. I don't think that follows
at all, for the reasons given above.

Iain

----
To unsubscribe, send a message to majordomo@calvin.edu with
"unsubscribe asa" (no quotes) as the body of the message.