Friday, December 5, 2008

Dembski Has Dispensed With The EF. Big Deal!

This is a couple of days old, but this comment by Dr. William Dembski at UD has some tickled pink:

I’ve pretty much dispensed with the EF. It suggests that chance, necessity, and design are mutually exclusive. They are not. Straight CSI is clearer as a criterion for design detection.

For me, it's no big deal. I've said EF has had its shortcomings here. And apparently even hard-core ID defender Dave Scott saw it, too.

The asylum is open. Keep it clean.


  1. Interesting.

    Now if only someone could provide some evidence - ANY experimental evidence - indicating that Dembski's concept of CSI was relevant to biology.

  2. Um, aghunt is Art. The new comment interface outsmarted me.

  3. No problemo, Art. I hope no one else is having problems with the comments section.

    What I would like to know is how Dembski defines CSI without use of the EF. Last time I checked (Design Revolution), he needed the EF to get to CSI.

  4. I suspect that one could, if so inclined, dig out lots of statements from DaveScot that indicated he bought the EF nonsense hook line and sinker. I suspect that he is probably engaged in scrubbing those comments from UD, if his past behavior is any guide. His belated conversion to EF-skepticism immediately after Dr. Dr. D. disowned it is, if I may say so, pathetic.

    I agree with Art. The problem with the EF, besides its bad assumptions, was that it was never shown to have any relevance to biology. The fascination with math by much of the ID pantheon only masks a remarkable ignorance of basic biology, which is allegedly the focus of their "research". CSI suffers from the same fatal disease; it has never been shown to be a useful, or even definable, concept for biological research.

    So I predict that CSI will suffer the same fate as the EF. As long as Dembski keeps obfuscating about it, the ID acolytes will keep trying to understand and defend it. But someday it will be abandoned in favor of a new obfuscation, and the acolytes will pretend that they knew all along that it was bogus.

  5. At least for proteins (and I presume nucleic acids, both of which may be represented as linear arrays of symbols), I think Dembski defines CSI using a two-step process. (At least that is what I gather from Ch. 4 of NFL, where he discusses an example I introduced to the ev/cre lexicon.) Basically, he estimates the specified information as being related to the proportion of functional to total sequences. He then asserts (based on arguments spelled out in TDI) that, if this proportion is less than one in 10^150, then the specified information is complex.

  6. I would add in passing (even if a few days later - time flies and all ...) that my description here is similar, as far as I can tell, to the approach Kirk Durston uses for studying CSI.