Good Math, Bad Math

Bad News for Uncommon Descent

In my ongoing search for bad math, I periodically check out Uncommon Descent, which is Bill Dembski’s
blog dedicated to babbling about intelligent design. I went to check them today, and *wow* did I hit the jackpot.

Dembski doesn’t want to bother with the day-to-day work of running a blog. So he has a bunch of bozos
who do it for him. Among them is Salvador Cordova, who can almost always be counted on to say
something stupid – generally taking some press story about science, and trumpeting how it proves
intelligent design using some pathetic misrepresentation of information theory. [That's exactly what
he's up to this time.](http://www.uncommondescent.com/archives/17816)


He starts off with a typically clueless statement:

>In information science, it is empirically and theoretically shown that noise destroys specified
>complexity, but cannot create it. Natural selection acting on noise cannot create specified
>complexity. Thus, information science refutes Darwinian evolution. The following is a great article
>that illustrates the insufficiency of natural selection to create design.

This is manifestly *not* true. In information science, it has *never* been shown that there exists *any* quantity that can be referred to as “specified complexity”. They keep trying to foist the idea on readers as if it’s an accepted part of the mathematical theory of information, but they even
*agree* on what it is, much less present an actual mathematical definition.

It’s rather hard to show both empirically and theoretically that something that isn’t defined and can’t be measured can only be affected in particular ways by noise.

Except of course that Sal’s definition of “specified complexity” is roughly something like “stuff that seems really complex, but which I can describe imprecisely using very few words”. Since that’s
mathematically gibberish, he can claim anything he wants about it, but it won’t say *anything* about reality other than “Sal Cordova doesn’t know what he’s talking about.”

He continues by quoting an article about Zebrafish and their regenerative abilities:

>>”Interestingly, some species have the ability to regenerate
>>appendages, while even fairly closely related species do not,” Poss
>>added. “This leads us to believe that during the course of evolution,
>>regeneration is something that has been lost by some species, rather
>>than an ability that has been gained by other species. The key is to
>>find a way to ‘turn on’ this regenerative ability.”
>
>If the ability to regenerate major organs is hardly visible for natural selection to preserve, how
>in the world will natural selection be able to even create the ability to regenerate major organs in
>the first place?
>
>Natural Selection does not trade in the currency of design (ala Allen Orr). I have also argued here
>why contingency designs are almost invisible to natural selection. The ability to regenerate major
>organs is an example of a contingency design.
>
>The discovery by these researchers again illustrates the ID’s Law of Conservation of CSI and ID’s
>formulation of the 4th law of thermodynamics.

This is a combination of nonsense, cluelessness, and non-sequiter.

**The nonsense**: “if regeneration is hard for natural section to preserve, how will natural selection create regeneration?” Who said anything was hard for natural selection to preserve? The only statement is that *some species lost* the ability to regenerate. That is an *entirely* different statement from “Regeneration is hard for natural selection to preserve”. Sal is just making it up by *pretending* that there is something in the statement which *is not there*, and then arguing that the thing that isn’t there supports him.

**The cluelessness**: More of the above. The idea that some species lost X means that X is hard to preserve is nonsense. Change happens. Genes get co-opted into doing other things. Sometimes, under some conditions, the benefits of some trait might be outweighed by some disadvantage of that trait. (For example, we know that there are many genes that work in various ways to fight cancer. What if the ability to regenerate organs is directly tied to an increase in the likelihood of developing cancer?) Other times, an individual that has both a highly beneficial mutation and a negative mutation *together* can pass on its genes to a population, where the benefit outweighs the loss. (Cats can’t taste sweets. We know the [specific mutation in the feline genome](http://scienceblogs.com/pharyngula/2006/09/cats_candy_and_evolution.php) that causes this; and we know that tigers have the same mutation. Does *not* being able to taste sugar have some advantage? Probably not. But probably some ancestor of the cat family had that mutation tied to some other trait that was selected for. And there are other possibilities as well. Sal’s explanation of “regeneration is hard to preserve” as the only possible explanation is a demonstration of the limits of Sal’s imagination, not the limits of evolution or natural selection.

**And More Bonus Cluelessness**: the fact that a trait was lost by a population doesn’t mean that it’s necessarily gone forever. It’s not at all uncommon to see some gene “turned off” by a simple mutation or developmental glitch, but to have the gene preserved. (Just look at the number of
virus genes carried in our genome!) If the gene is there, but not expressed, it’s not difficult for
*another* mutation to re-activate it.

**The non-sequiter**: the claim that this is in *any way* connected to information theory or complex
specified information. There is *nothing* in the original quote, or in Sal’s expansion on it to
support the idea that there is a measurable quantity of CSI in the genes for regenerative abilities;
nor is there anything in the article or Sal’s text to support the idea that there was *any* loss of
complex specified information involved in the loss/deactivation of regenerative abilities. It just
*does not follow* from the article. œ(Not to mention that the so-called “fourth law of
thermodynamics” is in no way, shape, or form a valid law of thermodynamics.) This invocation of
Dembski’s conservation of information nonsense is a complete non-sequiter.

Comments

  1. #1 John Lynch
    November 16, 2006

    And mysteriously, Cordova’s post has disappeared. Whodathunkit.

  2. #2 Jens Ayton
    November 16, 2006

    “they even agree on what it is” -> “they can’t even agree on what it is”
    “non-sequiter” -> “non-sequitur” (x2)
    “the article. œ(Not to mention” -> “the article. (Not to mention”

  3. #3 Jens Ayton
    November 16, 2006

    It strikes me that this whole “can’t create complexity” thing goes from unsupported to downright laughable in the face of genetic algorithms.

  4. #4 The Ridger
    November 16, 2006

    Ha-hah!

    Does not being able to taste sugar have some advantage? Probably not.

    Of course it does! A specific advantage designed into them for their role as pets! Hahahahaha … it’s more proof that God likes cats better! As if the whole “no dogs in heaven” thing wasn’t enough – cats won’t drink anti-freeze!

    Is that enough exclamation points? ;-)

  5. #5 Blake Stacey
    November 16, 2006

    Cats will so drink anti-freeze, if you put tuna in it.

    (I’m gonna regret posting this. . . .)

  6. #6 Blake Stacey
    November 16, 2006

    @John Lynch:

    Cordova’s post didn’t go away; the link was just in error. Try the following:

    http://www.uncommondescent.com/archives/1781

    (without the “6″ at the end).

  7. #7 The Ridger
    November 16, 2006

    Cats will so drink anti-freeze, if you put tuna in it.

    (I’m gonna regret posting this. . . .)

    Hah! They will drink tuna, but not antifreeze – or at least, not antifreeze ’cause it’s sweet and they stumbled upon it. Only antifreeze if you deliberately subvert God’s plan to save his favoritest creatures ever!

    You will regret it – God’s gonna get you!

    Mwhahahahahahaha

  8. #8 Troublesome Frog
    November 16, 2006

    It’s interesting to watch the conversations on UD. One would think that Dembski would clarify some of the issues they folks over there are fighting about. As is typical over there, the conversations always degenerate into nitpicky arguments over the meanings of passages of The Sacred Text, No Free Lunch. It’s like watching a theological debate raging over the consumption of shellfish when God Himself is sitting there without chiming in.

    Maybe it’s just the fact that anybody who asks revealing questions (like, “How do I measure CSI?”) is banned immediately, but it seems like absolutely nothing of substance goes on in those threads.

  9. #9 Ichthyic
    November 16, 2006

    As is typical over there, the conversations always degenerate into nitpicky arguments over the meanings of passages of The Sacred Text, No Free Lunch.

    you seriously can’t imagine a good reason WD40 doesn’t intervene?

    I can think of several, including the most obvious:

    it’s not worth the time to clarify, given the nature of the material to begin with.

    followed closely by:

    controversy sells, so why bother trying to curb it?

    the only reason he maintains UD is to sell his books.

  10. #10 Boo
    November 16, 2006

    So, is “complex specified information” actually a real concept, or is it just like something Dembski made up?

  11. #11 Zeno
    November 16, 2006

    It’s something Dembski made up. But even he can’t define it. Very sad.

    Learn more at talkorigins.

  12. #12 Millimeter Wave
    November 16, 2006

    @Blake Stacey:

    I think you need to postfix your statement with the word “accidentally”. That should fix it…

    @Boo:

    What’s your best guess here? ;-).

    Since I’m a communications guy (rather than a computer science guy), I tend to use the Shannon definition of “information”, rather than the Kolmogorov definition of “complexity”, which Dembski prefers. They amount to very much the same thing, but are measured in different ways. I find it kind of fascinating that he chooses to mix both concepts in his nomenclature. I honestly have no idea what it should mean.

    However, if you take the phrase “specified information”, it roughly means “that which is neither random nor deterministic”, so adding the “complex” qualifier would seem to indicate that it means “the complex empty set”. Or something. I’ve read quite a few of Dembski’s papers, and I haven’t seen anything that persuades me they’re anything more than gibberish intended to sound impressive to the uninitiated.

  13. #13 June
    November 16, 2006

    All ID arguments for proving that X was designed are forced to follow the pattern: “We cannot see any way that X could have evolved. Therefore, X was designed.”

    This form of argument started 5000 years ago, when shepherds found sea shells and fish fossils on mountain tops: “There is obviously no way that sea creatures can get on top of mountains. Therefore there must have been a big flood.”

  14. #14 sparc
    November 16, 2006

    Sal is not even right about regeneration. Indeed wound healing that even he must have had experienced is a form of regeneration. In mammalian embryos it even works without scar formation. In addition, mammalian liver can regenerate quite well. Limited capacity in adult mammals may be restricted by epigenetic effects. Indeed, when you reprogram a fibroblast nucleus in a denucleised oocyte it is well possible to regenerate (clone) a complete animal.

  15. #15 RBH
    November 16, 2006

    Millimeter Wave wrote

    Since I’m a communications guy (rather than a computer science guy), I tend to use the Shannon definition of “information”, rather than the Kolmogorov definition of “complexity”, which Dembski prefers.

    Dembski talks about K-C complexity, but in fact for purposes of “Complex Specified Information”, Dembski defines “complexity” as improbability. Period. That ain’t K-C complexity.

  16. #16 Millimeter Wave
    November 17, 2006

    @RBH:

    yes, that does seem to be the general pattern. He starts off saying that he’s using some particular body of established theory, but then proceeds to make it up as he goes along.

    As I alluded to earlier, it’s “Star Trek science”. It’s supposed to sound impressive, but it doesn’t actually mean anything.

  17. #17 Barron
    November 17, 2006

    Popping over to check the “this post mysteriously disappeared” point I see happy praise for a new ID book: “Darwin Strikes Back: Defending the Science of Intelligent Design” by Tom Woodward

    Whenever I see something about a new ID book the first thing I do is check the publisher. In this case the publisher is “Baker Books” which seems pretty average until you check their website and find their goal is “… building up the body of Christ at every level with books that are relevant, intelligent, and engaging”. Their science and faith books include “Genesis Record” By Henry Morris (if ICR fame, “Bones of Contention” by Marvin Lubenow and “The World That Perished” by John Whitcomb (another ICR alum).

    But wait, one might complain, isn’t this guilt by association? Well, yeah. If this was the only example. But if you look at the books the DI recommends you’ll find a vast majority are some religious, evangelical publishing houses. Yet another way of showing that Judge Jones was dead on in Kitzmiller. Again and again the supporters, publishers, popularizers of the ID dogma are repurposed “creation scientists” of yesteryear.

  18. #18 valhar2000
    November 17, 2006

    It strikes me that this whole “can’t create complexity” thing goes from unsupported to downright laughable in the face of genetic algorithms.

    That is why the Uncommonly Dense people at Uncommon Descent used to spend quite a lot of time arguing that Genetic Algorithms haven’t done the things they actually have done, or that the answer was codified into them at the beginning, without the programmers having an inkling that this was so.

    By the way, I propose the first Law of Tontodynamics: the Stupidity of a closed system of creationists cannot decrease.

  19. #19 Jud
    November 17, 2006

    Mark CC: “Sal’s explanation of ‘regeneration is hard to preserve’ as the only possible explanation is a demonstration of the limits of Sal’s imagination, not the limits of evolution or natural selection.”

    The whole of ID can pretty well be summed up as “If I can’t imagine how it happened, then it couldn’t have happened.” Or to put it another way, “God ain’t no smarter’n I am.” For a religiously motivated movement, ID sounds awfully heretical to me.

  20. #20 Torbjörn Larsson
    November 17, 2006

    I saw a reference to Cordova’s piece the other day, and recovering from some ‘shit happens, that’s life’ I thought that he was on to something for once. But this really ripped the pants of him.

    The point I had trouble with isn’t visible because Mark has been a little unconventional. (In addition to this remark, and the link trouble, I also offer “they even agree on what it is” which probably should be ‘even disagree’.)

    Cordova says “If the ability to regenerate major organs is hardly visible for natural selection to preserve, how in the world will natural selection be able to even create the ability to regenerate major organs in the first place?” In Mark’s simplification that becomes ‘ “”if regeneration is hard for natural section to preserve, how will natural selection create regeneration?”” ‘. (Inner citation mark is Mark’s use of ” instead of modifier ‘, middle is mine, outer is mine modifier to compensate for the earlier lack. Clear? :-)

    “Hardly visible” is visibly not ‘hard’, and googling it one finds some use for “visible for natural selection” in papers. Though after some pondering initiated by Mark’s simplification I can see that here it amounts to nearly the same thing. Cordova seems to think that since “fairly closely related species” may loose regeneration, it is hardly visible for evolution and so hard to keep or get. Okay, that is nonsense as explained.

    June:
    Doh! I have noted the common flood myths, but I sloppily attributed them to inflated memories of local flood events or tsunamis. That is a much better explanation for the commonness.

  21. #21 Torbjörn Larsson
    November 17, 2006

    Umm, Cordova’s text is here too. Sorry about the confusion. I guess I haven’t got my brain in gear already.

  22. #22 Boo
    November 17, 2006

    If the ability to regenerate major organs is hardly visible for natural selection to preserve, how in the world will natural selection be able to even create the ability to regenerate major organs in the first place?

    I thought natural selection was a process that didn’t “create” anything at all. Or am I stupid? (Let me guess, it’s not me, it’s Sal.)

  23. #23 Craig Helfgott
    November 17, 2006

    And here I thought that “conservation of CSI” was talking about how, if the network stations aren’t showing CSI on a given night, TBS has to show 4 hours that day to make up for it.

  24. #24 Will Von Wizzlepig
    November 17, 2006

    And, as usual, the clueless footsoldiers of the wing-nut right keep us busy refuting their nonsense while their special ops people are busy elsewhere, lobbying, legislating, and planning further mayhem. Are we destined to forever be the fractured left?

  25. #25 Mark C. Chu-Carroll
    November 17, 2006

    Boo:

    In the writings of many IDists, they use the term “natural selection” to mean “evolution not directed through intervention by a supernatural entity”. So when someone like Sal writes “natural selection”, he doesn’t just mean the selection process, but the entire evolutionary process including mutations and other kinds of genetic change.

  26. #26 Boo
    November 17, 2006

    It just does not follow from the article. œ(Not to mention that the so-called “fourth law of thermodynamics” is in no way, shape, or form a valid law of thermodynamics.) This invocation of Dembski’s conservation of information nonsense is a complete non-sequiter.

    Well, if Dembski’s “fourth law” is something along the lines of “the complex specified information of a closed system cannot increase” (I think I read that somewhere), and if there is no such thing as complex specified information, then technically Dembski’s law is 100% correct, right? That which does not exist by definition cannot increase?

    So how does Dembski explain people getting smarter? Do people’s brains, like, leach the complex information from the environment thereby making the environment less informatively complex or something? (Forgive all the questions please, I just have time on my hands today)

  27. #27 Coin
    November 17, 2006

    Boo, read some Dembski and you’ll very quickly realize that literally everything he writes implicitly or explicitly assumes that the human mind exists outside of physical reality and is not subject to the laws of physics, thermodynamics or “information”. He usually tries not to state it openly like that, but he doesn’t usually do a very good job of hiding it either. Whereever the word “Intelligence” appears in Dembski’s writing, it is a code word for the immortal soul.

  28. #28 Bronze Dog
    November 18, 2006

    I had a weird dream this morning. I was at some kind of Intelligent Design event at a park. I asked them to define “complexity”. They offered an attempt, followed by me asking them to define it mathematically: What numbers do you stick into an equation. They put some up on a chalk board, not telling me what the numbers were: Just a big numeric calculation with an end result of something like several trillion units of complexity on Earth, calculated down to two decimal places.

    I asked them what the plugged numbers were, expecting some Texas Sharpshooter probability to be involved.

    It was even sillier:

    The first term in the equation was the number of Electabuzzes in the world. Apparently the IDers in my dream thought there were 860.

  29. #29 pwe
    November 26, 2006

    Hello Mark CC;

    You wrote in the OP:

    This is manifestly not true. In information science, it has never been shown that there exists any quantity that can be referred to as “specified complexity”. They keep trying to foist the idea on readers as if it’s an accepted part of the mathematical theory of information, but they even agree on what it is, much less present an actual mathematical definition.

    Just for your convenience I spent a couple of hours yesterday to write a mathematical defininition of “specified complexity”.

    Actually Dembski has already provided one in his Specification paper; but I have elaborated a few bits on it and given some critique.

    You can read the result of my efforts at Decomposing specified complexity. Especially the section Is specified complexity information? should interest you.

    Please note that I’m not an IDist myself; I have only done this, so you need not make any more threads asking for a mathematical definition of specified complexity :-)

    best regards
    - pwe

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.