Valentine Pontifex

William Dembski has replied to Richard Wein's review of No Free Lunch. It is evasive and mean-spirited to say the least. Dembski has probably unintendedly painted himself in a very bad light.

<a href="http://www.talkorigins.org/design/faqs/nfl/" target="_blank">Wein's review of NFL at The Talk.Origins Archive</a>

<a href="http://www.iscid.org/ubb/ultimatebb.php?ubb=get_topic;f=10;t=000022" target="_blank">Dembski's reply</a>

Happy Wonderer

Erm, any statisticans out there have a clue what his interpretation of "degrees of freedom" is about? The term "degrees of freedom" refers to how many independent observations you have relative to a particular statistical parameter. It is part of formulas that say "more observations give you a higher confidence than fewer do." (Oh Heck that is a lame definition, go to <a href="http://seamonkey.ed.asu.edu/~alex/computer/sas/df.html" target="_blank">to a good page on degrees of freedom</a>)

What that has to do with a combination lock is beyond me. (Are the possible combinations supposed to be an analogy for observations?) How this limits our understanding of physical processes is even further beyond me.

(Here is an abbreviated version of the quote so that you can find it in the PDF file, read the original 'cause I may be skipping something important.)

Huh? (Any typos are mine, and I believe this is 'fair use' of the quote.) Is there a kernel of deep insight here? He apparently has a Ph.D in statistics but this isn't making any sense to me. (I have some graduate background in math-stat but it has been many years...)

-HW

(Even at a very basic level his analogy fails. A combination lock has one correct combination, which certainly mandates one combination to the exclusion of all others. WHAT IS HE TRYING TO SAY?)

[ May 10, 2002: Message edited by: Happy Wonderer ]

[ May 10, 2002: Message edited by: Happy Wonderer ]</p>

hezekiah jones

Richard Wein explains William Dembski's ideas better than William Dembski does. That's probably why Dembski spends half his rebuttal crying like a baby.

hezekiah jones

I found this on another discussion board, quotes from No Free Lunch:

A naturalistic scenario for the evolution of the bacterial flagellum, centerpiece of "design theory," would not invalidate design:

But wait. Such a naturalistic scenario would deny its "specified complexity," and therefore its "design":

Not only that but a naturalistic explanation for such a mechanism would "falsify" ID in its entirety:

[Emphasis added.]

No wonder nobody knows what the hell he is talking about. It almost sounds as if he doesn't either.

theyeti

Don't worry man, it's not just you. Dembski's writings are needlessly jargon laden and obfuscatory. However, when you get right down to it, what he's actually claiming is really quite simple, and quite obviously wrong. But by putting it in coded language, he makes it difficult for anyone to know just what it is that he's getting at, and he puts others in the position of having to take the blame for his own equivocations. He's not alone in that last respect; Jonathan Wells vicously accuses others of dishonesty when they call him out on his own dishonest ambiguities. Being intentionally vauge yet chock-full of innuendo is a trademark of the ID movement.

Dembski's followers just lap it up. They don't know what he's talking about either, but it sounds impressive and it's what they want to hear, so they shower him with praise, giving him absurd titles like "The Issac Newton of Information Theory". This sort of thing just tends to feed what is obviously an already over-inflated ego -- I think a lot of creationists and pseudo-creationists are driven by the fact that the mere message that they espouse gives them a huge base of adoring fans.

The beauty of Richard Wein is that he actually has the time and the patience to cut through Dembski's BS and get to the real issues. What it comes down to is that Dembski's major claim is circular: How to detect desgin? Specified complexity! How do we detect specified complexity? Specified complexity is that which must have been designed! Now add to this the fact that Dembski equivocates the defintion of SC, making it seem as if it's something that's detectable by means of its physical properties on one hand, and then defining it as something that has a probablistic and question-begging quality on the other. As Wein points out, Dembski should drop the whole SC business altogether, because all it does is confuse what is otherwise a straight-forward and very weak argument. In his response, Dembski doesn't even address this issue, and instead goes about with his usual equivocations, labeling this and that as SC and essentially saying that it must have been designed because it must have been designed.

In fact, Dembski ignores most of Wein's arguments. Instead he spends the first 20% of his response with tangenital issues, basically invoking the argument from authority (I got my supporters who have good credentials to put blurbs on the dust jacket, therefore my book must be good), attacking Wein's credentials (you've just got a BS in statistics, so you can't be right), and defending his claim that TDI was peer-reviewed, apparently missing Wein's point that statisticians and other mathematicians had not been highly receptive of it as Dembski had previously claimed.

Knowing Wein, I'm sure that he'll respond before long, and when he does, it will be a skewering. I don't expect that to make a difference to any of the Dembski faithfull. In fact, I doubt any of them will read Wein's writings.

theyeti

Undercurrent

Just as a side note. Bill Macready (of NFL theorem fame) is one of my coworkers (used to be my boss.) I didn't realize that the IDers were using his theorem to pretend to disprove evolution. He'll be thrilled to find out.

m.

hezekiah jones

Get back to us on that. His comments would be pure gold.

Happy Wonderer

Thanks! His books must be a real treat In my experience, the true R. A. Fischers, Newtons and Turings of the world can explain their ideas in a way that makes sense if you are willing to put the work into it.

Dare I say it? My reading is that from a statistical standpoint he is not just obscure, but is spouting gibberish. Why, I don't know, unless it is to scare people off from looking at his arguments closely. Statistical methods can only be used to interpret empirical data, so they add nothing to his (data-less) general argument about design or complexity. So perhaps he is using terms like "degrees of freedom" that define concepts which many people have heard of, but which few really (care to!) understand?

"Every relevant probability distribution?" Sounds impressive, but is totally meaningless. I'm trying to come up with possible interpretations for what he could mean, but it hurts my brain too much. A parameter has one probability distribution. A parameter can't have both a normal distribution (say) and an F distribution. In other words probability that N% of the observations will be within two standard deviations of the mean can't be both %90 and %20!<a href="http://www.itl.nist.gov/div898/handbook/eda/section3/eda361.htm" target="_blank">Here for a definition of probability distributions.</a>

"Stochastic process" btw just means a process that can be analyzed using the rules of probability. So he is just saying that his probabilistic methods will be applied to any process that can be analyzed by probabilistic methods. That's useful... It is pretty hard to come up with something that can be analyzed by probabilistic methods that doesn't involve "chance" in the equation. (That is the whole point of rejecting the null hypothesis... oh the heck with it, my brain hurts!)

HW

Dr.GH

I see one major math problem with the degrees of freedom argument (DF doesn’t apply at
all to the example given) that would take a bit too much time right now. (BTW thanks for
that link Happy Wanderer, I haven’t taught statistics since 1983). But there is an stopper,
and that is that the combination lock is only mysterious if we cannot open the back up
and study the tumblers. The paragraph just before Wild Bill blows smoke about degrees
of freedom he wrote:
Material mechanisms do most certainly determine the combination of the lock; does
Dembski propose a non-mechanical or magical lock is involved? (Actually he does but
that is called creationism and he wants to deny that label). And this is why combination
locks can be opened at all, and can be opened in a number of ways even if one does not
know the combination a priori. There is the brute force method of trying all
possible combinations, there is using a sensitive measuring device that can detect when a
channel or tumbler is properly aligned, and there is the practical approach of opening up
the damn thing to determine the specific properties of the system/lock under study. As
far as Dembski’s analogy, brute force is only a matter of time, and biological systems and
the Universe itself have an abundance of time. There are any number of possible
instrumental means to determine the state of the lock as it is manipulated. And I see that
opening the lock up for examination is just what we attempt in science.

[ May 11, 2002: Message edited by: Dr.GH ]</p>

hezekiah jones

<a href="http://www.talkreason.org/articles/dem_nfl.cfm" target="_blank">Free Lunch in a Mousetrap, By Mark Perakh</a>

Another gigantic deconstruction of No Free Lunch.

[ May 11, 2002: Message edited by: hezekiahjones ]</p>