ID’S IRREDUCIBLE INCONSISTENCY REVISITED

By Mark Perakh

January 30, 2005

I normally do not read posts on Access Research Network (ARN). Sometimes, though, some of my colleagues point to certain curious posts there and in such rare cases I briefly look them up.  Such was the occasion a few days ago when a contributor to ARN advised about a funny exchange of posts on ARN with the participation of William Dembski.

            That consistency is not William Dembski’s forte is not news.  However, once in a while this inordinately prolific propagandist for intelligent design offers notions that are so obviously contrary to others of his own notions that one wonders whether Dembski is serious or his opuses are spoofs designed mainly to attract attention to his voluminous output.  Also, it is funny that there is a bunch of Dembski’s admirers (such as, for example, Salvador Cordova) always ready to spin his notions in a positive light, often doing that in a truly acrobatic manner.

            The case in point is the recent article by Dembski titled “Irreducible Complexity Revisited” at http://www.iscid.org/papers/Dembski_IrreducibleComplexityRevisited_011404.pdf which invoked the mentioned discussion on ARN.  In this paper, Dembski endeavors to redefine (once again) Behe’s irreducible complexity (IC). Additionally, Dembski posted several messages on the ARN discussion board in response to critical comments signed by RBH (see http://www.arn.org/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=13;t=001884;p=0 ).  

            In this discussion the contributor to ARN who signs his posts as RBH offered   critique of Dembski’s paper. In particular, RBH has pointed out that Dembski’s new definition of IC means in fact “the death” of IC because it adds an impossible condition for a system to be recognized as IC. This condition does in fact require proving a universal negative: to be IC, the system, according to Dembski’s new definition, must be such that its function cannot be performed by any other simpler system. With this requirement no system can ever be asserted to be IC because it is impossible to assert that there is no other, simpler system anywhere in any form that can perform the job, even at a lower level of fitness. If we can’t point to such a simpler system, it does not mean it cannot exist; the possibility that such a system exists but we simply don’t know about it can never be excluded.

            Although, to my mind, RBH’s critique of Dembski’s new definition of IC is very pointed and logical, there are some additional points related to Dembski’s paper in question which seem worth mentioning.

            Recall that in his book No Free Lunch Dembski devoted many pages to a justification of Behe’s irreducible complexity concept. In particular, Dembski claimed there that Behe’s definition of IC was “neither exactly correct nor wrong.” He suggested a “salvaged” definition differing from Behe’s original definition in some respects; for example, he introduced the sub-concept of an “irreducible core” of the system; he also slightly modified the view of Behe’s concept of a “basic function” which is lost if a system’s irreducible core loses at least one of its components.

            As I argued in my book Unintelligent Design, Dembski’s “salvaging” modifications of Behe’s original definition have no principal significance and are essentially a casuistic attempt to get out from under the fire coming from critics who denounced Behe’s idea.

            Before reviewing Dembski’s newest rendition of IC, let us note that if a system is IC this means it lacks self-compensatory mechanisms. This is a serious shortcoming. A well designed system normally is expected to have reserved paths to be taken if the regular path becomes unavailable. Being IC means that any accidental damage to the system makes it unusable. Cars normally carry a spare tire to be used if a regular tire blows up. IC systems by definition have no spare parts – if a part is missing, the system ceases to function – that is how Behe defined an IC system.

            If such a system has evolved naturally, there is nobody to blame for its serious drawback. If such a system is a product of design, then it testifies to the designer being stupid, inept, or deliberately making fools of his believers.  I have discussed this point at length in my book. (BTW, neither Dembski nor Behe have ever responded to my critique in any form, shape, or manner).

            A relevant point regarding Dembski’s newest discussion of IC is how, in his book, he tries to defend Behe’s endlessly recycled model of an IC system  - the mousetrap. Behe maintains that a mousetrap is an example of an irreducibly complex system. If any of its five parts is missing, says Behe, the mousetrap becomes dysfunctional.

            Behe’s model has been shown to be in fact readily reducible without losing its function, albeit with a lower fitness. There are several such demonstrations, one of the most spectacular suggested by Professor of biology John H. McDonald

( http://udel.edu/~mcdonald/mousetrap.html ) who has shown (with animation) how parts of the mousetrap can be removed on by one, the remaining contraption preserving its ability to catch mice, even if not as successful as the full five-part mousetrap. 

            Any scientist who adheres to the common rules of intellectual honesty, if confronted with such a spectacular show of having insufficiently thought through his example, would promptly admit its fallacy. Instead, Behe persists in defending his example, inventing various ad-hoc explanations (for example, emphasizing the requirement that parts of an IC system must be “well-matched,” otherwise it is not an IC system; unfortunately, Behe did not suggest criteria distinguishing “well-matched” parts from “not-so-well-matched”).  So Dembski came to Behe’s rescue, devoting a lengthy discussion in his book to an attempt to dismiss McDonald‘s demonstration of the inadequacy of Behe’s example.

            The main argument offered by Dembski was that MacDonald’s reducible mousetraps are an invalid illustration because in each step of removing the mousetrap’s parts one by one, the mousetrap’s functionality is preserved by modifying the shapes of the remaining parts. According to Dembski, the definition of an IC system implied the requirement that the removal of any part of an IC system results in losing its function only if the shape of the remaining parts is not altered.

            In fact, no such requirement was indicated anywhere in Behe’s original definition. Moreover, such a requirement would make the entire concept of IC irrelevant for biology.

Evolutionary biology recognizes evolutionary processes wherein at every step of evolution, adaptation and/or excaptation take place wherein changes of various characteristics of evolving system, including changes of its parts’ shapes, occur routinely.

Dembski’s argument in his book was an ad-hoc quasi-argument employed to save not just the ID concept but also Behe’s unsuccessful example, perhaps because admitting the fallaciousness of a highly propagandized example would be a major embarrassment.

            So, let us summarize the above story: In his book No Free Lunch Dembski maintains that a system which, upon losing a part, cannot preserve its function unless the shape of the remaining parts is altered, is IC.  Otherwise it is not IC.  In particular, a mousetrap is IC.   

            Switch now to Dembski new discourse, titled “Irreducible Complexity Revisited.”

            This time Dembski approaches the matter from a different angle and offers an example. This example is a three-legged stool.  Its function, says Dembski, is to provide an elevated platform as a seat. OK. Let us accept this notion. If any of the three legs is removed, the stool loses its function.  (As RBH has noted, besides the legs, also the seat itself is an indispensable part of a stool whose removal renders the system dysfunctional as well; Dembski forgot about the seat). Each of the four parts (the three legs and the seat) is indispensable. Obviously, according to Behe’s original definition, such a stool is IC.

            Not so, says Dembski in his new article.

            Before discussing Dembski’s new example further, it seems proper to point out that the example with chairs having various numbers of legs was suggested in my book (in chapter 2 which is about Behe’s ideas).  Of course, Dembski’s forgetfulness of whatever his predecessors may have discussed is not news either.  He seems to have lost his memory on many occasions, forgetting both what his co-travelers and what his detractors have written before. Moreover, in this case it seems worth mentioning that an example with chairs (or stools) was suggested, before my book, by Dembski’s colleague in the anti-evolution crusade, David Berlinski (to whom I referred when discussing the example with chairs).  Apparently it makes no difference for Dembski whether he ignores a predecessor who is sympathetic to his ideas or somebody who has been critical of his ideas. The main point for him seems to claim as many notions for his own as possible. He does not stop if asserting his superiority entails diminishing the status of a colleague (such as Behe).

            So, what is his new thesis? Now he asserts that a three-legged stool is not IC. Why? Because there are other, simpler systems which can take over the stool’s function. For example, a concrete slab can serve as a seat. 

            What a fine idea! And what then about mousetraps? A mousetrap, according to Dembski, is IC because its function can be preserved upon removing a part only if the remaining parts change their shape. Only if the function could be preserved without changing these shapes, would the mousetrap be deemed not IC.

            Where is the difference between the reduced mousetrap and a concrete slab serving as a seat? Indeed, remove three out of the four parts of the three-legged stool (two legs and the platform). To preserve function, change the shape of the remaining leg, making it thicker so it would stand on its own and its upper butt end would serve as a seat.  Now, according to Dembski’s new idea, the stool is not IC because there is a different, simpler system with the same function. This system, however, is obtained from the original stool by removing three of its four parts and then altering the shape of the remaining part (and also, if desired, changing its material from wood to concrete, although it is optional).  Exactly the same procedure was performed by MacDonald with mousetraps.

            Moreover, if a stool can be replaced by a concrete slab, which is a much simpler system, likewise a mousetrap can be replaced by a much simpler system – a small spot on the floor covered with glue with a piece of cheese in its middle.  However, with Dembski’s habitual irreducible inconsistency, the mousetrap is claimed to be IC but a stool is not.

            In this case Dembski seems to have disproved his own thesis with charming simplicity. 

            Of course, there are many more possible ways to state mutually incompatible notions, and with Dembski’s ingenuity it would be intriguing to see the next variation of his assertions negating his own earlier statements. Perhaps we can expect many more hilarious events stemming from William Dembski’s inventive philosophical flexibility. 

            We also can expect that Dembski’s devout Sancho (sorry, Salvador) will promptly spin Dembski’s blunder to the delight of his fans.

            I also was amused seeing in the posts to ARN by Dembski’s supporters, such as Salvador Cordova, references to “excessive complexity” which Salvador construes as a marker of intelligent design.  Recall that neither in Behe’s book Darwin's Black Box where  irreducible complexity was suggested as a sign of intelligent design, nor in Dembski’s book No Free Lunch wherein the definition of irreducible complexity was “salvaged” was excessive complexity ever mentioned. So why does this concept suddenly emerge now in posts by Dembski’s supporters? I take the liberty of providing a plausible explanation.

            Excessive complexity was discussed in my book Unintelligent Design which appeared at the end of 2003. In fact, however, I introduced this term and the concept it relates to in my essay critical of Behe’s work posted to the web back in 1999 ( www.talkreason.org/articles/behe2.cfm ). I believe this term was first used there (although I may not know of its prior use by somebody else; it is hard to suggest a term that was never before suggested by somebody else, perhaps in another context).  In a little different way a similar concept of “redundant complexity” was discussed (also in 1999) independently by Niall Shanks and Karl H. Joplin (Philosophy of Science, 66, 1999, no 2, 268-77).

            Both Dembski and Behe ignored my essay, possibly because they were unaware of it, or for any other reason. However now, when my book has been published and dozens of its reviews have appeared both in print and on the web, the ID advocates have apparently decided that it would be more gainful for them to appropriate the concept in question and to adjust it to their goals rather than to continue ignoring it. In their usual manner, though, they do not deem it useful to refer to the work of a predecessor, and more so when the predecessor, like myself, was their adversary.

           This all is good and nice, but one should remember that excessive (or redundant) complexity (EC) is the opposite of irreducible complexity (IC). However, Salvador construes both EC and IC as markers of design. Well, there is nothing surprising in that because it is not uncommon for universal conjectures like ID being so flexible that notions contradicting each other can be equally deemed supporting it. Such contradictory notions, however, hardly carry evidentiary value  (see also www.talkreason.org/articles/Popper.cfm by Elsberry where this item is discussed from the standpoint of Popper’s falsifiability criterion).   

          Now I’d like to add a few words about another case wherein, though, it was not Dembski himself who offered a strong critique of some of his notions. In this case it was Henry Morris of the Institute of Creation Research, the well known young earth creationist - see

http://www.icr.org/pubs/btg-a/btg-194a.htm (see also Wesley Elsberry’s post at http://www.pandasthumb.org/pt-archives/000766.html#more where Morris’s essay is discussed). In his essay Morris expressed appreciation of intelligent design ideas but also his strong opposition to the ID advocates’ attempts to hide their religious roots.

            For example, Morris quotes Dembski:

“Intelligent design is a strictly scientific theory devoid of religious commitments. Whereas the creator underlying scientific creationism conforms to a strict, literalist interpretation of the Bible, the designer underlying intelligent design need not even be a deity.³ “

Then Morris concludes:

“Dembski himself may not believe such nonsense…. “

Isn’t this quote entertaining?  It seems to show that sometimes (although not very often) even YEC advocates may utter a correct statement (of course, if it does not cast shadow on their own ideas).

            Then Morris writes that the main concepts of ID championed by Dembski and his colleagues are in fact not new at all, as they were suggested by YEC advocates, including Morris himself, a long time ago. One such notion is the “specified complexity” vigorously promoted by Dembski as one of his most important discoveries and as such one of the main components of his theory.  Morris asserts that he had already suggested this concept many years ago under the label of “organized complexity,” which, in Morris’s view, was a better rendition of the concept in question. Morris also points out that he suggested a long time ago a quantity which is an exact analog of Dembski’s “universal probability bound” of about 10^-150 with the sole difference that his suggested value was 10^-110.  Morris also says that he and his fellow YECists have suggested  bacterial flagellum as an example of design a generation earlier than Dembski.

            However contrary to science Morris’s biblical literalism is, there is no reason to doubt the veracity of the above claims.  It is just one more confirmation of what we know about anyway: Dembski habitually plays at Alzheimer regarding his predecessors, not only those who are critical of ID, but often also those who share his views.

            As to Morris’s characterization of Dembski’s position as nonsense, it jibes with what we have known anyway – ID advocates’ pretension that their ideas have no religious roots but are rather purely scientific, is a propagandistic trick which hardly can deceive anybody who has at least a rudimentary knowledge of the real situation. 

            While both YEC and ID “theories” have all the features of crank science, ID advocates, unlike YEC people, additionally are trying to hide their real religiously motivated agenda.  

            It all would be funny if it were not really sad.