|Wesley R. Elsberry
Joined: May 2002
|Wesley raised some very good points. Lets stick to his arguments and try to refrain from distracting from them through the use of strawmen.|
I was actually trying to steer discussion back around to Dembski's arguments given in the initial post. Perhaps I should have left Paul's interchange with John alone, but I thought I could clarify things there pretty quickly and move on. I may have been wrong about how quickly...
I appreciate the moderator letting the discussion continue in this thread at all, since it doesn't really seem to have a "brainstorming" style topic. It is a pretty straightforward response to criticism, and in this case the critics have chosen to get somewhat involved. (My involvement needs to be limited, as I'm getting acquainted with LaTeX for preparation of my dissertation and also have the usual daytime job. Well, perhaps not that usual.)
Anyway, I think that what should be followed are Dembski's arguments.
|(William A. Dembski:) Briefly, the claim that specified complexity is a reliable marker of design means that if an item genuinely instantiates specified complexity, then it was designed. As I argue and continue to maintain, no counterexamples to this claim are known.|
I was there when Ken Miller presented the Krebs cycle as a counterexample to Dembski on June 21st of this year. I think that Dembski should note that counterexamples have been proposed by Miller and also Rob Pennock. Now, it is a given that these have not been demonstrated to Dembski's personal satisfaction, but I think Dembski's phrasing of his claim is somewhat misleading to the reader.
Further, I think the claim doesn't mean much, anyway. Since 1996, Dembski has provided EF/DI calculations, in various degrees of completeness, for a total of four events.
- The Caputo case
- The Contact primes sequence
- Dawkins's METHINKS IT IS LIKE A WEASEL string
- The E. coli flagellum
(If I've missed an application of the EF/DI that comes with actual numbers and complies with more than two or three of the seven steps outlined on pages 72-73 of NFL, please let me know so I can expand the list.)
That's not much of an empirical base upon which to build such sweeping claims as the "no counterexamples" claim above.
Back in 2001 at Haverford College, I made the point to Dembski that collecting "confirming" cases does nothing to test his EF/DI. I suggested that he apply his EF/DI and perform calculations for a number of events that could be agreed have sufficient evidence of natural causation to provide real tests of his EF/DI. These included the Krebs cycle, since shown by Miller to be a real counterexample (well, he convinced me). I also suggested the mammalian middle ear impedance-matching system and "fairy rings" as good candidates for testing the EF/DI.
I think I brought up the point that the EF/DI should be applied to a broad range of biological phenomena at the June 21st get-together at the Fourth World Skeptics conference. Create a workbook style presentation of a series of EF/DI calculations starting with small-scale events that everyone can agree should not trigger a "design inference" and work up to larger-scale events that biologists have evidence for saying that natural causes are sufficient. Is the EF/DI a good guide to classifying biological phemomena? Until we see a series of real examples of complete application of it, I think that the issue is still wide open.
Well, I'll come clean. I expect that if such a workbook were attempted, that the EF/DI would find "design" at ludicrously small-scale events, ones that not even Bill Dembski would want to go on record as saying that they must be considered to be "due to design". I think that Dembski's statement at the end of TDI that a "design" conclusion is not easily reached via the EF/DI is simply false. A simple way to show me wrong is to actually produce such a compendium of example EF/DI calculations, where the EF/DI performs in a stable manner and produces expected (by ID advocates, natch) results.
The production of such a workbook would also do much to vitiate another criticism of mine, which is that the EF/DI framework is too unwieldy to be applied. Dembski says of Gell-Mann's "effective complexity" that it "resists detailed application to real-world problems" (I think that's verbatim, but I don't have NFL in front of me. Check around page 133.). I think Dembski's EF/DI very much "resists detailed application to real-world problems", and the fact that Dembski has offered so few EF/DI calculations (even including the only partially complete E. coli flagellum example) supports my view. Of the four examples, the Contact primes examples is plainly fictitious, neither the Caputo case nor the METHINKS string yield an improbability smaller than Dembski's "universal small probability", and the E. coli flagellum example suffers from a large number of defects. Does it really take a year-and-a-half, on average, to apply the EF/DI to any sort of problem, no matter how trivial or how many steps are skipped?
I'd be interested in hearing if any third party has attempted to apply or applied the seven-step process outlined on pages 72-73 of NFL. I no of no such examples yet.
"You can't teach an old dogma new tricks." - Dorothy Parker