Just Because We Can’t Define It Doesn’t Mean It’s Not Science

You may want to save this post at Uncommon Descent, in case it disappears down the memory hole.

If you’ve been following Intelligent Design, you’ve probably run across William Dembski’s notion of Complex Specified Information, or CSI. Basically, the argument is that if a system has CSI above a certain level, then it was intentionally designed (just as “Wherefore art thou Romeo” exhibits design, while “Mp YuMsAAVVa UU MbMZlPVJryn Viw MfHyNA FHh” doesn’t). Living beings (or their genomes) have sufficiently-high CSI, and were therefore designed. QED.

So the question from day one has been, “so how exactly does one calculate CSI and get an actual number?” From what I’ve seen, the standard answer is “go read Dembski’s book”. None of my local libraries have Dembski’s book, but from the reviews I’ve read, I gather that for all his talk about CSI, he never gets around to sitting down and describing how to calculate it.

And now for some reason, the people at Chez Dembski have invited someone going by the name of MathGrrl (whom I guess to be a frequent commenter; I stopped reading the comments there a long time ago, so I don’t know) to write a guest post. And not only that, but one in which she basically asks, “so anyway, how does one calculate CSI?”.

The first fifty comments consist mostly of “Yeah, well, evolution doesn’t explain it!” and handwaving, followed by a bunch of comments from MathGrrl to individual commmenters, all “Yes, but that doesn’t help me calculate CSI.”

Which is odd: you’d think that the first dozen or so comments would be links to FAQs, and maybe some Mathematica code to do the calculation. But no. And it’s not because they’re too busy to answer MathGrrl’s question, since a lot of them go on at length about how she’s not asking the right questions, or not using CSI correctly, or maybe some other measure of complexity would be better suited. But I’m not seeing a whole lot of anything that looks like math.

The thread looks, to me, like a gaggle of astrologers arguing about the proper way to calculate a horoscope.

So once again, getting information out of creationists is like pulling teeth.

Update, Mar. 25, 2011: The 200-comment mark has been reached, and no definition in sight. In fact, comment #201, by PaV, says:

Dear MathGrrl:

To provide a “rigorous definition” of CSI in the case of any of those programs would require analyzing the programs in depth so as to develop a “chance hypothesis”. This would require hours and hours of study, thought, and analysis.

You come here and just simply “ask” that someone do this. Why? You do it.

In other words, “Math is hard! Develop our theory for us!”

(Update, Aug. 4: Fixed typo.)

19 thoughts on “Just Because We Can’t Define It Doesn’t Mean It’s Not Science”

  1. UD also has a history of banning well-meaning true believers who start asking inconvenient questions. I kinda feel sorry for MathGrrl — she’ll probably be left wondering just what exactly she did wrong, and why all her Brothers and Sisters in Christ™ got so hostile all of a sudden.

  2. I think the pulling teeth analogy is inaccurate. More like constipation. No matter how hard you try or what you pull out, it’s likely to be crap.

  3. I think the pulling teeth analogy is inaccurate. More like constipation. No matter how hard you try or what you pull out, it’s likely to be crap.

    If this is a double post, is because I had browser problems.

  4. I believe that “mathgrrl” is not an ID supporter but a critic of ID who has occasionally made comments at UD, and (in a stunning departure for UD) been allowed to make a guest post there. This seems to have required a special posting by the administrator of UD, Barry Arrington, saying in effect “calm down, folks, we want to actually have a discussion”. I hope to give a perspective on CSI soon at Pandas Thumb.

  5. I went to Uncommon Design/faq just out of curiosity (I was BORN an atheist… well, almost…) and lo and behold, under the title “Introduction”, the second sentence reads: >Nevertheless, many critics mistakenly insist that ID, in spite of its well-defined purpose, its supporting evidence, and its mathematically precise paradigms, is not really a valid scientific theory.<

    How beautiful!!!! Nobody can inform MathGrrl just what the "mathematically precise paradigms" are!!!!

    Howcum nobody has pointed this out to her?????

  6. I tried, without success to make a post out at UD, but either WordPress or UD prevented my best efforts.

    So here it is.

    MathGrrl has set up a trick question. She wants an objective way to measure CSI. This, according to Dembski is not possible.

    “Specification depends on the knowledge of subjects. Is specification therefore subjective? Yes.” Page 66 of his “No Free Lunch”.

    The S in the specified is therefore subjective.

  7. Marconi Darwin

    She wants an objective way to measure CSI. This, according to Dembski is not possible.

    It would’ve been nice if this had been more widely publicized before I waded through his paper Specification: the Pattern that Signifies Intelligence.

    But of course this goes to the heart of the problem with Intelligent Design creationism: it’s based on a handful of intuitive notions like “it’s complicated, therefore it was designed”, but no one has managed to formalize these intuitive arguments.

  8. Marconi Darwin:
    That’s nice, but it’s hardly something that was pointed out by, say, ID supporters. Hell, even in the UD thread above, how long did it take Dembski’s defenders to get around to pointing out that CSI can’t be calculated objectively?

  9. How about this as a measure anything beyond 10^399? Isn’t there some upper limit that could make you doubt naturalism?

  10. How about this as a measure anything beyond 10^399?

    Isn’t there some upper limit that could make you doubt naturalism?

    How about 10^ 888888?

    If an event is that unlikely could it have a natural cause?

  11. bbbright:

    Isn’t there some upper limit that could make you doubt naturalism?

    Upper limit of what? How are you calculating that number?

    Be sure to show your work, and penmanship counts.

  12. arensb,

    I am simply pointing out that it is pointless to expect IDers to be honest in their attempts to respond. How long did it take Dembsky himself to finally admit that the designer is none other than Jebus?

    IDers should be ridiculed, not reasoned with

  13. bbbright asks :


    March 27th, 2011 at 10:45 pm
    How about this as a measure anything beyond 10^399?

    Isn’t there some upper limit that could make you doubt naturalism?

    How about 10^ 888888?

    If an event is that unlikely could it have a natural cause?

    Why not?

  14. bbbright Says:

    How about this as a measure anything beyond 10^399?

    Isn’t there some upper limit that could make you doubt naturalism?

    How about 10^ 888888?

    If an event is that unlikely could it have a natural cause?

    Considering the complete absence of any other data beyond a probability construct it’s meaningless. Correlation does not equal causation.

  15. Fez:

    any other data beyond a probability construct

    For that matter, it’s not even a probability: 10399 is a huge number (10888888 even more so). Probabilities are between 0 and 1.

    If, on the other hand, he meant to say 10-399, I should point out that ridiculously improbable things happen all the time. The probability of my sshd config generating the random 2048-bit key it did is 10-616. For 10-888888, I believe you need to go to 3-million-bit-long random sequences, which nobody would use, except maybe the NSA and people running tail /dev/random.

Comments are closed.