The Wizards of ID

In his article "Who's Got the Magic?" (Dembski, Metaviews 042, 2000), William Dembski discusses my book Tower of Babel: The Evidence Against the New Creationism, and defends Phillip Johnson's intelligent-design creationism (IDC(1)).  I had expected that Dembski would also respond to criticisms I had made of his own specific arguments, and was disappointed that he decided not to engage them. For instance, he does not address my criticisms of his "explanatory filter" or of his problematic, idiosyncratic uses of the key terms in his argument-"law", "chance", "information", "specified" and "design."  He does not confront examples I gave to show that the Darwinian mechanism can produce functional and novel complexity.  He ignores a counter-example that shows how a genetic algorithm can produce complex specified information (which he says is impossible to do), and that avoids what he claims to be a defect in examples Dawkins and Sober have given.  He does not engage my criticisms of his SETI analogy.  And so on.  Indeed, until now, Dembski has had nothing to say in response to my book and has only made a couple of brief ad hominem remarks on his web page. (Dembski, 1999b)

I will reserve for a forthcoming article detailed criticisms of what he has published since I finished Tower of Babel, and will here confine my reply to issues related to the one substantive question he does discuss in his current paper, which has to do with whether IDC or evolutionary theory is a form of magic.  With the exception of a passing remark in my book about Michael Behe's appeal to supernatural design as a scientific explanation, I have never discussed IDC explicitly in terms of magic, but Dembski's analogy is apt so I will form my reply along the lines that his analysis suggests.

Dembski distinguishes three kinds of magic.  The first is the familiar form of stage magic-"the art of illusion"-an entertainment he says is unobjectionable.  It is the second kind of magic-"to invoke the supernatural to explain a physical event"-that he objects to, at least when the label is applied to his and other IDCs' views. He is keen to distance his movement from the term "creationism" and suggests that it is wrong to link their notion of design with the supernatural.  There is no magic of the second kind in our design, Dembski appears to say.  It is really scientific naturalists who have the magic, he charges: a third type of magic which is the belief "that something can be gotten for nothing."

I shall argue that science is not guilty of this, but that IDC does involve both of these sorts of magic.  I'll show that IDCs practice the first form as well, which they use to try to divert attention from the other two.  Clever misdirection is what makes the stage show work.  Watch closely and see how the trick is done.


Dembski chides me for never using the term "intelligent design" without conjoining it to "creationism."  He implies (though never explicitly asserts) that he and others in his movement are not creationists and that it is incorrect to discuss them in such terms, suggesting that doing so is merely a rhetorical ploy to "rally the troops". (2)   Am I (and the many others who see Dembski's movement in the same way) misrepresenting their position?  The basic notion of creationism is the rejection of biological evolution in favor of special creation, where the latter is understood to be supernatural. Beyond this there is considerable variability.  Some creationists think the world is young while a fewer number accept that it is ancient.  (Among IDCs, Paul Nelson, Percival Davis, Sigfried Scherer and John Mark Reynolds, to name a few, are among those in the first camp, while Michael Behe is in the second.  Most IDCs hide their views on this and other relevant theses.) Creationists typically base their views on the Bible, but there are also, for instance, Native American creationists whose specific alternative creation stories stem from their own religious traditions.  So-called "scientific creationism" or "creation-science", which is our main concern here, is a form of apologetics that claims not to base its conclusions on a religious belief or text, but purports that these are supported by scientific research alone.  I have previously discussed many varieties of creationism in detail, showing how IDC compares to other forms (Pennock, 1999, Ch. 1), and will not repeat that discussion. Here I will focus mostly on Dembski's own views to document how they fit in the overall pattern.

Dembski's other complaint against me arises directly out of the first; he objects to having "design" yoked to the second form of magic that he defined, namely, invoking the supernatural to explain a physical event.  He says, "To call this magic is certainly a recent invention, since it makes most theists into magicians." (Dembski, 2000)  But Dembski surely knows that etymologists can trace the English word "magic" back through Latin and the Greek (magos) to the Old Persian word magus, which dictionaries define as a Zoroastrian priest, or a seer or wizard knowledgeable in sorcery, numerology and astrology.  The plural of magus is magi, and who were the Three Magi, if not theist astrologers who had read the sign of a star.  IDCs engage in their own form of numerology, believing that a few probability calculations will reveal signs of the divine.  Not all theists are magicians in this sense, but those who are should honestly acknowledge it.  Let me be clear that my aim was never to dissuade such theists of their faith in a supernatural designer-creator, but just to show it is not legitimate to accept the creationist contention that this is a scientific conclusion.  IDCs, like other creationists, do hold this view.  Again, I shall focus here mostly upon establishing that Dembski's position is as I have claimed.

In his response to me, Dembski cites variations of deism to show that God could have created without miraculous violations of natural laws. This response is puzzling, in that I had myself discussed the deist option in my book as one way that a person could accept evolution and scientific methodology while still retaining belief in God as Creator.  I gave this as one of several counter-examples to IDCs' rejection of such a possibility.  Johnson explicitly dismisses deist views throughout his writings.  Indeed, to try to set up the (false) dichotomy that he needs to legitimate purely negative argument, he goes much further and dismisses any form of theistic evolution. Dembski's response is all the more puzzling, since he adopts the same position.

IDCs see theistic evolution as, in Dembski's words, "an oxymoron." He writes "Design theorists are no friends of theistic evolution." (Dembski, 1995, p. 3.  Emphasis in original).  (I have quoted this line on several occasions, but in one article the last word was printed "evolutionists" instead of "evolution."  Let me assure Dembski that the mistake was unintentional and not "by design."  I do not know how the error was introduced, but I should have caught it during proofing and I apologize for missing it.  Dembski took offense at the error, and was quite correct when he complained that there is a huge difference in meaning between refusing friendship with a group of people rather than with their ideas.  More on that shortly.) Theistic evolutionists accept the truth of evolution as science has discovered it, but retain a belief in God, though obviously not in the kind of God that creationists-the large majority of whom are fundamentalist Christians-find worthy of the name.  Dembski tells us: "As far as design theorists are concerned, theistic evolution is American evangelicalism's ill-conceived accommodation to Darwinism." (Dembski, 1995, p. 3)  Like other IDCs, Dembski thinks that theistic evolution is nonsensical; he says it is equivalent to "purposeful purposelessness." ( Dembski, 1995, p. 3)

I recently attended a talk Dembski gave to a group of about fifteen high school and junior high school science teachers in which he tried to explain how they could incorporate intelligent design into their classes. (3)   Illustrating the concept of design by reference to violins and other musical instruments, Dembski noted that the early Christian fathers compared the world to a lute, and he advised that this is a better analogy than a watch.  Why?  A good watch doesn't need winding, but an instrument is meant to be played, he explained.

Anyone familiar with Paley's design argument would have understood exactly what he was implying.  If the world is like a watch that God created, then, the deist argues, we should expect that God got it exactly right from the start, and would not thereafter continually intervene to make adjustments.  But, as Johnson defines it, the creationist view he advocates holds that "a supernatural Creator not only initiated this process but in some meaningful sense controls it in furtherance of a purpose." (Johnson, 1991, p. 4)  Direct control in the form of supernatural intervention is at the very heart of the IDC program, which is aimed not just at overturning evolution, but scientific and metaphysical naturalism as well.  In this respect, IDCs are again right in line with the classic creationists, and Dembski is clearly in lockstep with this position.  He has written: "I don't believe in fully naturalistic evolution controlled solely by purposeless material processes". ( Dembski, 1995, p. 5)  Moreover, in his most recent and definitive discussion of ID, he says explicitly that biological forms were created with periodic "discrete insertions" of design and that the complex specified information purportedly required to affect such creation "transcends natural causes." (Dembski, 1999a, p. 171)

Look again now at how Dembski rebutted my challenge regarding IDCs' attempts to bring the supernatural into science.  He did so by citing the Stoics and saying that design "requires neither miracles nor a creator."  This is not to the point.  No one objects to archeologists, anthropologists, or even SETI researchers, who all make use of this completely ordinary and natural notion.  If that were the only sense of "design" that Dembski and company were talking about, then most of this discussion would be obviated.  IDCs often complain that the term "evolution" can be used in many ways and claim that scientists use it to "cover a multitude of sins," but the same can be said, with more justice, about their use of the term "design."

The trick can be seen by examining how Dembski's argument would look against just a slightly different stage setting.  For instance, it has become common in this New Age to hear purported explanations of all sorts of physical ailments in terms of blockage of the flow of the body's "energy."  Believers may hold that this is a spiritual form of energy that is somehow "infused" into the world, and they may reject the scientific explanation of the same ailments, deriding them as "materialist" and "reductionist".  They try to poke holes in medical theory by pointing to ailments that doctors don't know how to explain, and claiming that physicians maintain what is, after all, "just a theory" because of their "naturalistic blinders" and desire to hold on to their cultural prestige and power.  We may imagine a New Age Dembski who promotes "energy theory" and defends it by saying that energy need not involve the supernatural but is, as he puts it, "common, rational and objectifiable."  Well, perhaps so, but that natural notion is not what we were talking about, was it?

So, let us not be fooled by their stagecraft into thinking that IDCs constitute a real scientific movement. (4)  Replace "design theorists" in Dembski's statement about "ill-conceived accommodation to evolution" quoted above with "evolutionary biologists" or "physicists" or "complexity theorists" and one immediately sees the difference.  IDC is a theological movement crafted to win a particular political goal-initially, getting their form of special creation into the public school science classes-in what IDCs take to be the key strategic game in the "culture wars".

Dembski admits as much even in the title he chose for a course he prepared to teach recently at Trinity Graduate School and Trinity Evangelical Divinity School at Trinity International University: "Intelligent Design-The New Player in the Creation-Evolution Controversy."  In his syllabus for the course, Dembski describes ID as "an alternative to scientific creationism... that challenges Darwinian evolution and its naturalistic legacy." (5)

Dembski writes that ID "does not treat the biblical book of Genesis as a scientific text, but instead argues on general scientific and philosophical grounds that Darwinian evolution is a failed scientific research program."  This is exactly the strategy that classic creation-science followed in trying to get its views into the public schools-provide only a vague positive thesis and rely upon negative argumentation against evolution in the hope of winning by default. Indeed, as I elsewhere showed in detail (Pennock, 1999, p. 250-251), all the key elements of Dembski's argument, including the use of Yockey's and Orgel's concept of specified complexity, were previously articulated by Norman Geisler, who was one of the expert witnesses for the creationist side in the famous "balanced treatment of creation-science and evolution-science" case in Arkansas in the early 1980s.  Today the ambiguous term they use is "designer" instead of "creator," but otherwise the game plan of the "design theorists" is little different than that of the "creation scientists."

Continuing in his syllabus, Dembski predicts that his movement will have earth-shaking consequences: "Because a vast naturalistic superstructure has been built on Darwinism, its impending collapse at the hands of intelligent design promises to be one of the great upcoming cultural convulsions (witness the recent furor in Kansas)." Here, of course, he is referring to the vote by creationists on the Kansas State Board of Education to expunge evolution, as well as any reference to the ancient age of the earth, from the state's science standards.  Creationists on the Board had also tried to insert the following statement into the science standards: "The design and complexity of the design of the cosmos requires an intelligent designer." (NYT, 1999)  That sentence was excised from the final approved version, but its initial inclusion is telling.  In the wake of the Board's decision, IDC leaders-Phillip Johnson, Jonathan Wells, Stephen Meyers, among others-flew to Kansas to take the creationist side in public debates, and to promote it in churches and in talks sponsored by campus Christian groups.

With these quoted passages as background, Dembski concludes by stating that the purpose of his course is to "examine intelligent design as a cultural and intellectual movement and show how Christian theology and apologetics stand to benefit from it."

One might occasionally get the impression that Dembski does not reject evolution, as when he says elsewhere, somewhat vaguely, that he accepts "that organisms have undergone some change in the course of natural history."  However, that impression is dispelled when he immediately qualifies this and asserts that "this change has occurred within strict limits."  Dembski's position is indistinguishable from that of Henry and John Morris and other creationists, who put the point in exactly the same terms.  Also like the Morrises, Dembski specifically insists that "human beings were specially created." (Dembski, 1995, p. 5)

We shall see further evidence of Dembski's view in subsequent sections, but this should be sufficient to establish the point. Judge for yourself whether it is unfair to call this position "creationism."

When IDCs use the word "design" as Dembski does, it is the equivalent of the conjurer who assures his audience "I hold in my hand an ordinary deck of cards."  Unlike scientists, who lay their cards on the table in peer-reviewed journals, the Wizards of ID do not want anyone to examine their props too closely, because it would then be obvious that their "design theory" is a façade.  This is one reason why IDCs are not forthright when asked to state and defend their positive views, and why their standard move is to try to shift the burden of proof.  Johnson and Plantinga are masters of this legerdemain and Dembski, their star apprentice, has learned their trick. (6)  Rather than confront the challenge that magic of the second kind poses, he first obfuscates the difficulty and then tries to reflect it back, charging that it is actually naturalists who are engaging in a form of magic.  In stagecraft this is known as a classic "smoke and mirrors" trick.


So, let us now turn to Dembski's charge that scientific naturalists are pulling their own magic stunt-the ultimate trick of "getting something from nothing."  In his article, he derides the scientific naturalist's method as "bargain hunting," and warns that you get what you pay for.  Dembski wishes us luck, but suggests that scientific naturalists will never pull a rabbit out of a hat.  (Michael Behe makes a similar complaint, claiming that Darwinians are illegitimately pulling so-called "irreducible complexity" out of the "black box" of the cell.)  Like the creationists on the Board of Education in Kansas, Dembski finds problems not just with biological evolution, but with cosmology and other sciences as well.  Here I will retain our focus upon the former.  Dembski makes his complaint at a very abstract level, but I will not take up the philosophical chestnut of how, metaphysically, something can come from nothing; that is a fascinating problem, but in this context it is just more smoke to distract from the empirical question of how biological species evolved.  Dembski believes his reflected complaint applies specifically to this issue.  It really is meant as a challenge to human evolution and to rabbit evolution as well, so let us examine it in that context.

This is obviously not the place to review the vast array of evidential support for evolution: we have abundant evidence that contemporary complex forms evolved from simple ancestral forms, and we know a fair amount about the pathways that that descent with modification took, as well as many of the causal processes that produced those changes. (7)  A review of the literature will show that evolutionary biology does not pull rabbits out of a hat. Rabbits (such as the European species Oryctolagus cuniculus) and hares (such as Lepus euopaeus) belong to the Leporidae family and are grouped with the pikas (of the Ochotonidae family) in the order Lagomorpha.  Such nested patterns of features among extant organisms are a potent source of evidence for the evolutionary hypothesis of common descent with modification.  Despite some creationists' claims that evolution is just "assumed", evolutionary hypotheses are open to empirical test and to disconfirmation.  For instance, J. B. S. Haldane used to point out how significantly we would have to change our understanding of evolutionary history if we were to discover a Precambrian rabbit.  But in fact we do not find rabbits where they do not belong; the patterns in the fossil record are another independent source of evidence that further supports the evolutionary picture. Primitive rabbits have been found in Paleocene and Eocene sediments and are abundant by the Oligocene.  Such fossil finds have helped sort out the evolutionary relationship of the lagomorphs to the rodents, to which they are closely related.  (Colbert and Morales 1991, pp. 301-302)

We also have molecular evidence to help refine the evolutionary picture.  For instance, scientists have recently investigated mannan-binding protein (MBP) in rabbits and incorporated this new biological and genetic information into what was already known about the genes that code for different types of this protein in humans and other animals.  Their analysis showed that one MBP gene was lost not only during the evolution of hominids, but also after the separation of birds and mammals.  (Kawai,  et al. 1998)  Molecular techniques now allow scientists to precisely compare the sequences of amino acids in proteins across species and to use this to help infer phylogenetic trees (i.e., the pathways of evolutionary divergence). As expected, when such phylogenies are constructed (as they can now be done routinely as an exercise in evolutionary biology lab courses), human beings turn out to be more closely related to rabbits than to frogs, fruit flies or yeast.  We retain other clear signs of that evolutionary history.  The human appendix, the small and often problematic worm-shaped structure that must often be snipped from the intersection of the large and small intestines, is a vestige of a large organ found in rabbits and other herbivorous mammals that aids their digestion of cellulose that has been imperfectly co-opted to a different function as part of our immune system.  One could easily extend this list of evidence indefinitely.

This is not to say that evolutionary science is close to being able to answer all questions.  Dembski's criticism of the speculative evolutionary accounts some scientists have given of phenomena like consciousness does contain a kernel of truth; indeed, that criticism and the very term he used-"just-so stories"-was first made by evolutionary biologists against others who drew specific conclusions that went beyond the available evidence.  However, these are far from representative and, in any case, it is wrong to dismiss such accounts wholesale, since many may be regarded as what philosophers of science call "how-possibly explanations."  Even these do not come "from nothing" but are based upon known types of causal processes.  As such, they are constrained and perfectly legitimate, serving a useful purpose of illustrating the explanatory range of a theory and suggesting lines for further investigation.

What about design explanations?  Here there are fewer constraints, but in certain contexts, if we stick to our ordinary, natural notion of intentional design, we can still make some headway; when archeologists pick out something as an artifact or suggest possible purposes for some unfamiliar object they have excavated they can do so because they already have some knowledge of the causal processes involved and have some sense of the range of purposes that could be relevant.  It gets more difficult to work with the concept when speaking of extraterrestrial intelligence, and harder still when considering the possibility of animal or machine intelligence.  But once one tries to move from natural to supernatural agents and powers as creationists desire, "design" loses any connection to reality as we know it or can know it scientifically.

In his response, Dembski again declines to confront this problem, and simply claims that design does not explain everything: "There's no reason to invoke design to explain a random inkblot; but a Duerer woodcut is something else altogether."  But here, as elsewhere, Dembski's examples are not to the point.  His goal is to overturn scientific naturalism, but his examples of inferred design always involve natural objects, natural agents and natural causal processes. Also notice that by specifying that the inkblot is "random" Dembski is simply begging the question about its possible design.  The relevant issue is whether and how, given an observation of a particular inkblot, we may infer its design.  It is all too easy to read in design (purposeful intention) to a random design (pattern), as in Rorschach ink blots.  Recently, an oil stain on a glass office building supposedly looked like an image of the Virgin Mary, and was interpreted by hundreds of believers as a divine sign.  Moreover, while an inkblot on a sheet of paper may have occurred by chance, it might instead have been drawn by a doodler or an artist.  Or perhaps it only appears to be a real inkblot, but is actually a plastic inkblot left by a practical joker.  Given a naturalistic method we can often find evidence to distinguish among such possibilities, but how can we tell whether it was designed if we abandon this method and admit supernatural interventions, as IDCs want science to do?  Except for a few hints from Johnson about use of "sacred books" and "mystical states of mind," IDCs have proposed no new method to replace the naturalistic one they claim is dogmatically biased. (Pennock, 1999, p. 197)

Dembski's example of a Duerer print is also problematic. Again, no one disputes the identification of intelligent design in Duerer's renderings of horrifying scenes of the Four Horsemen of the Apocalypse or sublime pictures of nature.  Duerer's charming print of the "Young Hare" is a human artifact, and our inference to that conclusion is entirely naturalistic.  That inference, however, is not so simple as Dembski would have us believe.  Dembski does not explain how his design inference is supposed to work to identify a Duerer print as having been designed, but it must be because in looking at a particular print we supposedly can see that it contains complex specified information (CSI).  Presumably, the image of the hare is complex and specified in the relevant technical sense that licenses the design inference.  (There are problems with Dembski's technical conditions and how one tells when they apply, but I will reserve criticisms of those issues for another occasion.  In this article I am working under the assumption that the specific examples he gives count as CSI as he defines it, and that it is on that basis alone that we are to infer that they were designed.)  To make a woodcut print or a painting we know that the artist had to carve or paint every line by hand.  Could Dembski inadvertently be smuggling in such background information?  He claims that the design inference does not make any reference to such causal knowledge, so to avoid that feature of the example, let us consider a photographic print instead.

A professional photographer may set up in detail all the elements of the photograph and so have "designed" the picture in every particular.  An amateur, on the other hand, may simply point and click, and the resulting picture may be intended, that is, be "by design" in a weak sense, but not designed in the first, strong sense. A novice, unfamiliar with the workings of a camera, may touch the shutter release accidentally and take a picture quite by chance.  A robot camera or an automatic "McGregor's garden web-cam", say, may take pictures using a motion detector, or at regular or random intervals.  We may suppose for the sake of argument that all of these cases resulted in a photographic image of a young hare; they vary, of course, but are all comparable to Duerer's image in the relevant sense.  Each picture "contains CSI" to the appropriate degree.  Yet each picture was produced by a different combination of law, chance and design (or lack thereof), and only the first was designed in the strong sense that Dembski claims his inference rule reliably detects. (Dembski actually goes even farther, and says that once his design inference picks out something as designed that designation can never be undermined by further evidence.)

Note that it will not do to dismiss these counter-examples on the grounds that the camera was intelligently designed; the same is true of Duerer's tools, of course, and we are not now asking the question of how these acquired their own complex specificity, but that of the images, which is completely different and which Dembski singled out in his example.  Furthermore, Dembski's CSI criterion supposedly identifies design without needing to know anything about the causal processes that produced it, so the design of the camera should be irrelevant.  Nor may one object that the CSI in the photo was already in the real hare and thus is not a counter-example because it was not new information, but "just moved around"; the same relation holds for Duerer's print, and I am taking Dembski at his word that his criterion does license the inference to design in that case.  Moreover, if he were to retract that inference and say that in both cases the CSI came from some prior intelligence (an unspecified "hare designer"), this would beg the question.

Dembski describes his complexity-specification criterion as a net for catching design.  He admits that some things that are designed will "occasionally slip past the net", giving false negatives, but claims that it makes no errors in the other direction and catches only things that were indeed designed in the strong sense.  He writes "if things end up in the net that are not designed, the criterion will be worthless." (Dembski, 1999a, p. 141)  As we have seen, his criterion does give such false positives.

Returning now to Dembski's challenge to naturalists, we already have part of the answer.  Where did the information that makes up Y come from?  It came by being copied with slight "errors" from some ancestor X.  Copying... replication... reproduction... and always with new, chance variations: this is the sort of thing that rabbits (and other biological organisms) do very well.  This point holds even with inkblots.  Additional sheets of  paper applied to the blot on the first sheet will pick up a copy of the blot's shape: exact mirror copies or ones with random variations depending upon the conditions (e.g., how wet the ink remains, whether one uses blotting paper or not).  Rorschach inkblots have both chance and lawful features. (Note that any of these blots might also have been blotted "by design"-the categories of law, chance and design are not mutually exclusive categories, though Dembski defines them as such in his technical treatment).  Suppose now that whether and how often inkblots get to be blotted depended upon their shape.  If certain shapes conferred a functional advantage to some inkblots over others in this competition, then we would have the elements of the Darwinian process, and natural selection would kick in to refine the shapes to improve their adaptations.  With regard to inkblots this is just a thought-experiment, but for organisms function is indeed dependent on shape.  This is so for the organism as a whole, but also at lower levels of organization including organs, cells and even bio-molecules.  A protein's function is determined in large measure by its physical shape (the tertiary structure that is caused in part by the primary structure of its sequence of amino acids).  We shall come back to this in a moment.

In the question period following Dembski's talk in Lubbock, one teacher asked for an example of CSI she could use to illustrate the design inference to students.  Dembski concurred when someone suggested the sentence "John loves Mary" written in the sand-the example used in Of Pandas and People, the IDC school textbook.  (Such examples are significant in that they show that Dembski is not strict in holding to the arbitrary 10-150 probability bound that he previously set for the design inference.)  Naturally, there are some significant differences we have to keep in mind when using a linguistic case as a model for the biological-the former can occur by "Lamarckian" as well as "Darwinian" processes-but for the present purpose it works very well.  Since IDCs do use linguistic strings as an example, I shall first respond to Dembski's challenge with that analogy and then move to a real biological case.  I'll use the accidental misquotation I mentioned earlier, since Dembski's comments about it illustrate an important aspect of his misunderstanding. Here, again, are the two sentences, which we are now to consider (like IDCs' Pandas example) on analogy with genetic "information" as examples of "specified complexity".

      Design theorists are no friends of theistic evolution.      
      Design theorists are no friends of theistic evolutionists.

Of my error Dembski wrote: "Perhaps as an evolutionist himself, Pennock thinks the evolution of 'evolution' into 'evolutionIST' represents a minor adaptive change. I don't. I think it represents shoddy scholarship." (Dembski, 1999b)  It is worth noting that Dembski himself here accidentally misquotes my accidental misquotation of him, leaving out the final 's'.  This would give a third variation of the sentence.  I do not attribute his error to shoddy scholarship or intentional "mischief making" as he does of mine, but take it as a nice, inadvertent illustration of the kind of small, random error that occurs without intelligent design, as in Darwinian evolution.

All three sentences are cases of CSI, so according to Dembski's inference rule each had to have been designed in the strong sense. (If Dembski were now to insist that his formal probability threshold be met, we could just extend to the adjoining sentence about "ill-conceived accommodation" quoted above, or however much more he might require.)  This kind of example also undermines his inference rule.  Dembski is correct that the improbability of these two (almost) matching sequences leads us to disconfirm the hypothesis that they both arose independently simply by chance-that is just a standard statistical inference in which we reject the null hypothesis-but he is wrong to leap to design.  It was not chance alone, but chance acting in concert with a copying mechanism that produced the result.  Are those gaps of but one to four letters really impossible to bridge naturally?  Why couldn't one sequence have arisen from the other or both from a third?  Note that we are not assuming that either was specified as a target.  There are also a variety of other sequences with fewer differences that might have occurred instead (or served as "intermediates"), including, but not limited to:

      Design theorists are no friends of theistic evolutions.      
      Design theorists are no friends of theistic evolutionism.

To fill out the analogy further, we would also include the force of natural selection, which selects for adaptive functions in specific environments, but I won't take the time to do that again now.  When IDCs use examples of a sting of letters to illustrate the supposed impossibility of evolution by natural processes they ignore natural selection in an environment.  They also invariably focus upon a single specific case in isolation (like "John loves Mary" written in the sand) and fail to make the analogy relevant to the situation in the biological case.  Once the existence of multiple copies and variations is included, it is easy to see how a series of small changes can be shaped by natural selection to evolve significant changes in shape and function.

Moreover, we need not assume that every change is a single "point mutation", which seems to be how IDCs conceive that the calculations must be done.  A specific genetic string often has utility in multiple settings, and may appear in a new setting by recombination. In our analogy, the suffix "ists" functions like that, since it serves in many settings in a similar fashion; attach it to "material" or "creation" or "clarinet" and so on.  Moreover, we need not assume that this specific form is the only possible one that could serve that "purpose".  There can be multiple ways to achieve the same effect; the suffixes "ers" and "ians", for instance, can serve as well as "ists" to shift the referent from the "object" to the persons who "use" it.  It may simply be a historical accident that English speakers attach "ist" to "clarinet" and "er" to "trumpet" rather than the other way round.  Nor need we assume that the sentence must be perfectly formed to be understood-it often can function even if it contains some spelling and grammatical errors.  All these processes, and more, operate at the biological level.

So much for the supposed "irreducible complexity" that IDCs claim undermines Darwinian evolution.  Over the decades, creationists have gradually been forced to accept microevolution, but they now try to draw a sharp line between it and macroevolution and continue to reject that latter, claiming that small changes within a species has nothing to do with speciation itself.  IDCs also take this view, and Dembski was aligning himself with this position in saying that he accepts some natural change but only within "strict limits."  I have previously discussed their error in trying to draw a sharp distinction in this way. (Pennock, 1999, pp. 156 - 157)  Now it should be clear how this error is connected to the so-called "information problem".

On the introductory page of his web page article, Dembski wrote: "Small syntactic changes can introduce big semantic changes-Robert Pennock's misquote of me does just that." (Dembski, 1999b)  In grammar, syntax has to do with the specific arrangement of letters and words, whereas semantics has to do with their resulting meaning. Syntax and semantics correspond to two of several different senses of "information" that IDCs regularly conflate or confuse.  (Both are different from technical notions of information defined in various ways in information theory.)  The meaning of a string of letters depends not just upon their order, but also their context; for instance, the same string can be meaningful in one language, but have a different meaning or be totally meaningless in another, similar to the way that biological function and fitness varies with environment. The salient point for us here, however, is that a small error that leads to a big change in meaning can be introduced not by design but by accident, as was actually what happened in this case.  It is ironic that in his complaint, Dembski has inadvertently articulated the very feature of the Darwinian mechanism-that small chance variations can produce novel functional information-that he claims "design theory" rules out in principle.

The confusions in Dembski's information argument deserve more detailed treatment, but this will have to suffice for the moment. Let me conclude this discussion with a brief illustration that shows how the points made in our analogy apply in a real biological case. Rather than strings of letters and words in English that in specific orders produce sentences with different meanings, what we are really after has to do with nucleotides in DNA and sequences of amino acids that form proteins with different functions-molecular differences that make for the differences between species.

Let us look at an important protein-the beta subunit of the hemoglobin molecule-as an example.  The specific sequence of amino acids for beta-globin has been determined not only for human beings, but for a variety of other organisms.  Here we see an alignment of the human and the rabbit sequences:




7Exact matches between amino acids in the two sequences are connected by a bar: the human and rabbit sequences are 91% identical.  (The sequences are 92% similar, in that the amino acids connected by dots though not identical, share important chemical properties such as charge or hydrophobicity.)  By way of comparison, the beta-globin sequence of cows is 85% identical to the human sequence, that of chickens is 69% identical, and that of carp is 53% identical.  On the other hand, that of gorillas is over 99% identical to the human sequence-indeed, it differs by just a single amino acid. (8)   The simplest explanation of this pattern of information is that human beings and gorillas descended from a recent common ancestor (compared to less recent common ancestors with the other species) and that the original sequence was modified in one line by a single mutation. Processes of natural law together with chance variation produced the CSI of both.  Where did the specific, singularly improbable polymer come from that IDCs cite as the evidence of biological design?  It descended with modification from another sequence that perhaps differed by just one or a few elements, and that from another that differed perhaps a little more.  Viewed in the opposite temporal direction, we see the mutations, insertions, deletions, recombinations and duplications, shaped by natural selection, adding and molding information, until it is transformed from one species into another.  Small "syntactic" changes can indeed produce big "semantic" changes.

IDCs have yet to give any evidence to support their claim that such small microevolution changes cannot produce macroevolutionary differences.  Even if we were to agree, for argument's sake, that Dembski's formal design inference was without flaws, for him to show that it applied in a biological case, he would have to provide an example of an actual genetic sequence of the requisite length that not only has no variations in other extant organisms, but also none in recent organic history.  Neither Dembski nor any other IDC has proposed such a case and it is hard to imagine how they ever could. And yet Dembski continues to insist that it is impossible to produce CSI except by intelligent design, and that human beings were specially created separately from gorillas and other organisms.

Evolutionary biology does not pull a rabbit out of a hat, nor does it try.  Rather it shows how rabbits (and human beings as well) are a twig on the branch of the mammals, which itself is part of the deeply rooted tree of life.  It does not say that rabbits came "from nothing", but shows some of the processes by which such complex forms evolve from simpler ones.

Dembski seems to acknowledge this when he briefly qualifies his criticism about naturalists wanting to get something from nothing, noting in passing that "[t]he 'nothing' here need not be an absolute nothing."  But then he immediately returns to using the term "nothing" without qualification in the remainder of his complaint. As we have seen, this characterization is far from the truth.  This is one more example of how IDCs tend to use ordinary terms in "creative" ways for rhetorical effect, to make a difficult empirical problem appear to be a profound philosophical one.  Despite Dembski's attempt to reflect back the criticism, it is not evolutionary biologists, but IDCs who want to get something from nothing.  They utter their magic word "design," but offer no account of the processes by which that purported creative activity is supposedly accomplished.  No surprise, since the supernatural infusions and insertions of design that creationists want to wedge into science are literally a creation of something from nothing-Creation ex nihilo. IDCs hope to pull a rabbit not out of a hat, but out of thin air.


Dembski quotes with approval Richard Dawkins' comment that "biology is the study of complicated things that give the appearance of having been designed for a purpose."  I do not dispute that biologists sometimes speak in design terms.  Consider virologist Linda Stannard's comment that "After many, many years of peering at virus particles through the electron microscope, I have still not ceased to be amazed and excited by the precision and intricacy of design in something so very, very small." (Kaiser, 2000, p. 923)  However, although the world (and not just the biological world) exhibits all manner of useful arrangements of parts, we cannot simply read off intentional design from such natural objects.  On the contrary, the striking thing about genuine artifacts is that the kind of order they exhibit stands out in such stark contrast from the kind of order we observe in the biological world and the rest of nature.  There is, of course, a perfectly reasonable sense of evolutionary design and function that biologists legitimately refer to, but the appearance of agency that others see in nature is probably the result of a tendency to anthropomorphize the world.  As with inkblots, clouds and ceiling cracks, it is all too easy for us to read in a human aspect and to see our own image apparently reflected in natural design.

If the design argument really did allow us to infer agency in the creation of organisms, as IDCs claim, then what else would follow? Taking Stannard's observation together with Dembski's inference rule, it would seem that we must conclude that the intelligent designer lavished far more attention upon the design of the multitude of micro-organisms than upon less fortunate creatures such as ourselves, who suffer as their hosts.  Perhaps the theodicy problem is not greater in aggregate for Dembski's view than for other theist views, but it is certainly far more pointed, for we would have to attribute purposeful insertions of design to account for that precise and intricate specified complexity of every different kind of virus and bacterium.

Of course, IDCs will often say that all they are claiming is that biological organisms had to have been created by some "designing intelligence," but that they draw no conclusions about who that agent or agents might be.  Perhaps extraterrestrials, they occasionally suggest.  In his closing remarks to the science teachers, Dembski called our attention to a significant fact: "Notice that I never used the G-word in this whole talk."

So, what is this name that they dare not speak?  Who is this ghostly apparition who supposedly created all the complex specified information of the biological world?  In other settings IDCs do not play dumb and are more forthright about the identity of the designer. As one put it with a wink and a grin when introducing another IDC to a campus Christian group, intelligent design is the "politically correct way to talk about God."  Well, who ever thought it was otherwise?  Even Dembski, when writing for theologians, will sometimes fill in the blank and say that he believes that it is "God [who] created the world with a purpose in mind". (Dembski, 1995, p. 5)

In such discussions, we find further evidence that the term "creationist" is fitting to describe his position.  When Dembski says that life could not have arisen by purposeless material processes and required insertions of design, he is not giving us the conclusions of IDCs' scientific research, if such could be found.  When he tells us that biological change has occurred only within strict limits and specifies that human beings were specially created, he is doing just what the classic creationists have always done-he is starting with his conclusions already in place.  Perhaps as a teleologist himself, Dembski thinks that starting with one's end-point in hand and designing one's theory so it fits one's preferred conclusions is a reasonable method.  I don't.  I think it represents ideological bias.

Dembski's theistic science begins and ends with a Biblical assumption, read under a particular hermeneutic.  He writes: "If we take seriously the word-flesh Christology of Chalcedon (i.e. the doctrine that Christ is fully human and fully divine) and view Christ as the telos toward which God is drawing the whole of creation, then any view of the sciences that leaves Christ out of the picture must be seen as fundamentally deficient."  (Dembski, 1999a, p. 206)  IDCs' theistic science is inherently sectarian, but Dembski tries to assure us that privileging this specific Christian view is epistemically beneficial rather than harmful.

Why won't "privileging Christology", as he puts it, undermine the integrity of science?  Because, says Dembski, Christ is not an "addendum" to science, but a "completion" of it. (Dembski, 1999a, p. 207)  According to Dembski, "[T]he validity of the scientist's insights can never be divorced from Christ, who through the incarnation enters, takes on and transforms the world and thus cannot help but pervade the scientist's domain of inquiry."  (Dembski 1999a, p. 209)  Using a mathematical analogy, he assures us that "Christ, as the completion of our scientific theories, maintains the conceptual soundness of those theories even as the real numbers maintain the conceptual soundness of the applied mathematician's calculations. Christ has assumed the fullness of our humanity and entered every aspect of our reality.  He thereby renders all our studies the study of himself."  (Dembski 1999a, p. 210)

According to Dembski, the key to all this and the reason that the disciplines will be "taken seriously" and supposedly not lose their integrity is that since on the Chalcedonian doctrine Christ is both fully divine and fully human, "Christ can never be less than human." (Dembski 1999a, p. 206).  But being not less than human is not the problem.  The problem is being more than human: being also superhuman.  Dembski still has not solved the problem we began with, but just concealed it in the kind of obscurantist reasoning that gave the Sophists a bad name.

Metaphysically, the doctrine that Christ is at once both fully divine and fully human may be taken as a theological paradox or religious mystery.  However, if one were to view it as an epistemic principle it is simply a contradiction (unless, of course, "human" and "divine" are synonymous), and to advocate it as such would be to promote an absurdity that undermines reason itself, for in deductive logic anything whatsoever follows from a contradiction.  Similarly, scientific inductive reasoning collapses once one scraps the methodological constraint of lawful causal regularities and tries to introduce supernatural interventions as an epistemic principle, as IDCs do indeed desire in their "theistic science."

This is not meant as an attack upon Dembski's faith or that of others who hold those same specific theological views; rather it is a defense against his attempt to wedge them into science and to his simultaneous denial of the same.  Like Johnson, Dembski is attacking ontological naturalism and he regularly papers over the distinction between it and the epistemological position that I and others are defending as the basis of scientific reasoning.  To be sure, some scientists also fail to recognize the distinction, as Lewontin does in the passage Dembski quoted.  But while we may excuse a scientist for not being aware of even such significant distinctions in a subject outside his field, we should expect more from someone like Dembski who is trained as a philosopher and is promoting a philosophical position.

Dembski began his article by quoting from Tower of Babel a part of my challenge to Johnson: If you expect science to abandon its naturalistic methods and reintroduce appeals to supernatural interventions, show us first in your own field of the law how to do it.  I did not ask Johnson to delete design, in its ordinary natural sense, from the law.  No one doubts that criminals may leave a trail of evidence and that careful detective work can sometimes uncover their nefarious designs.  But the empirical reasoning involved in such sleuthing relies upon a thoroughly naturalistic method, and without it the notion of empirical evidence itself would collapse. Dembski at first seemed that he was going to address this worry, but with a little sleight-of-hand he avoided confronting the problem entirely.

Despite its denials, IDC does incorporate supernatural intervention and it does so as a central belief, right alongside a belief in creation from nothing.  The Wizards of ID practice both of these forms of magic but, as we have also seen, they try to obscure this fact in an elaborate smoke and mirrors show.  IDCs are performing a magic act of the first kind, hoping to fool the diverse audiences to which they are playing.  Dembski's own definition of this form of magic applies very well; it is "the art of illusion, where appearance is carefully crafted to distort reality" (Dembski, 2000).  Thoughtful seekers after truth would do well not to be deceived by their act and, when the unnamed apparition is conjured on stage, to pay close attention to the men behind the curtain.

Robert T. Pennock
Michigan State University


1. For simplicity, I'll also use this acronym as an abbreviation for "intelligent-design creationist".  The context will make it clear which is intended.

2. It is interesting to note how often Dembski draws such conclusions about his critics' supposed designs and intentions, though without ever explaining how he uses his "design inference" to infer them.

3. Dembski gave his talk at the annual CAST meetings in Lubbock, Texas, October 29, 1999.

4. In a letter in Books and Culture (1999, Vol. 5, No. 6, pp. 5), IDC John G. West challenged my contention that IDC is top-heavy with philosophers and he listed half a dozen scientists whom he said were leading the research.  I wrote to each of these asking for reprints or references to their scientific publications on intelligent design. No one will be surprised to hear that not one could provide me with any.

5. All the quotations and information about the course come from the syllabus, which Dembski kindly sent me of his own accord.  The syllabus also stated that thirty percent of the student's grade would come from class participation and the remaining seventy percent from a "5,000 to 6,000 word critical review of Rob Pennock's Tower of Babel" It continued, "We will not read Pennock's book for class. Pennock's book is supposed to be the "state-of-the-art" refutation of intelligent design. Your task is to use what you've learned in this course to assess and answer Pennock's critique of intelligent design. If (per impossibile) you happen to agree with Pennock, then you will need to write a joint critical review of my book [i.e. Intelligent Design] and Pennock's, showing why intelligent design is a failed intellectual project."  I thank Dembski for this friendly warning shot across the bow.

6. On Johnson's attempts to shift the burden of proof see (Pennock 1999, Chapter 4); on Plantinga's see (Parsons 1990).

7. I have discussed a bit of this evidence in (Pennock 1999, Chapter 2).

8. Learning to analyze such comparisons to help infer evolutionary relationships is becoming a standard laboratory course exercise now that nucleotide and protein sequence data is easily available on line.  The comparisons given come from a typical exercise <> .  To explore further, one can make any number of sequence similarity comparisons for a wide variety of species using NCBI's BLAST sequence search tool <>.


Colbert, Edwin H. and Michael Morales (1991). Evolution of the Vertebrates. New York, Wiley & Sons.

Dembski, William A. (1995). "What Every Theologian Should Know about Creation, Evolution, and Design." Center for Interdisciplinary Studies Transactions 3(2): 1-8.

Dembski, William A. (1999a). Intelligent Design: The Bridge Between Science & Theology. Downers Grove, InterVarsity Press.

Dembski, William A. (1999b) "Pennock's Convenient Distortion." Published on Dembski's personal home page. <>.  Expanded from original letter to the editor, Books and Culture Vol. 5 No. 6, Nov/Dec 1999, pp. 4 - 5.

Dembski, William A. (2000). "Who's Got the Magic?" META(042).

Johnson, Phillip E.  (1991). Darwin on Trial. Washington, D.C., Regnery Gateway

Kaiser, Jocelyn (2000). "Virus Portraits." Science 288(5468): 923.

Kawai, Takao, et al. (1998). "Molecular and biological characterization of rabbit mannan-binding protein (MBP)." Glycobiology 8(3): 237-244

NYT (1999). "Kansas School Officials Monkey With Evolution: Concept deleted from state science curriculum." New York Times. August 12.

Parsons, Keith (1990). God and the Burden of Proof: Plantinga, Swinburne, and the Analytic Defense of Theism. Buffalo, New York, Prometheus Books.

Pennock, Robert T. (1999). Tower of Babel: The Evidence against the New Creationism. Cambridge, MA, MIT Press. --


Join Metanexus Today

Metanexus fosters a growing international network of individuals and groups exploring the dynamic interface between cosmos, nature and culture. Membership is open to all. Join Now!