Has Darwin Met His Match? – Letters:

An Exchange Over ID

For more information about David Berlinski – his new books, video clips from interviews, and upcoming events – please visit his website at



Thanks to Commentary for posting the letters and responses to David Berlinski’s “Has Darwin Met His Match?” These are delightful to read and we encourage you to print them off, find a corner where you can laugh aloud and enjoy them to their fullest.

Below is the excerpt between Jonathan Wells, Michael Behe, William Dembski, Paul Nelson & David Berlinski – all fellows of Discovery’s Center for Science & Culture. It may seem odd that we should post this section regarding criticisms of ID and not Berlinski’s responses to the defenders of Darwinism, but it is because ultimately we wish to see ID fully debated on its crux issues. Our faithful critics often attempt to disqualify the debate over ID by demands that ID “is creationism,” that “it isn’t peer-reviewed,” and that “it isn’t science.” In contrast to these superficial (and incorrect) criticisms here is this refreshing exchange on the deep issues presented by ID.

The entire exchange, “Darwinism versus Intelligent Design: David Berlinski & Critics” is at:


David Berlinski’s “Has Darwin Met His Match?” is a breath of fresh air. Darwinism has become a sort of anti-religion, and defenders of the faith tend to demonize unbelievers rather than try to understand what they are actually saying. As a result, Darwinists typically distort intelligent-design theory—the latest and most powerful challenge to their orthodoxy—beyond recognition. Though a critic of intelligent-design theory, Mr. Berlinski is no Darwinist, and he is refreshingly fair-minded in his analysis.

I think Mr. Berlinski is mistaken, however, in characterizing the work that I and others have done as an attempt to resurrect William Paley’s natural theology. Paley argued that design proves the existence of the Christian deity, but I, for one, have never been convinced by his argument. Design necessarily entails a designer; but the Christian deity is much more than a designer, and additional premises and evidences must be adduced to prove His existence. I happen to believe in the God of Moses and Jesus, but my belief is based on much more than design.

The basic issue for intelligent-design theorists is not whether design gets us to God, but whether design is real. Darwinists contend that it is not. For example, Richard Dawkins, in The Blind Watchmaker (1986), argues that though organisms are “complicated things” and “give the appearance of having been designed…the evidence of evolution reveals a universe without design.” Dawkins’s claim is false, despite the Darwinists’ mantra that they have “overwhelming evidence” for their theory. I pointed this out in my book, Icons of Evolution (2000), and Mr. Berlinski agrees, writing that Darwinism is little more than a “fantastic extrapolation” in which the mechanism responsible for some minor changes within species “is read into the global record of life itself.”

The evidence of evolution does not reveal a universe without design, and it remains a possibility that some features of living things really are designed. Intelligent-design theorists like Michael Behe and William Dembski have proposed ways to determine which features are designed and which are not. Mr. Berlinski argues that they have not succeeded. In any case, though, their proposals for establishing criteria to detect design are not attempts to prove the existence of the Christian deity.

If intelligent-design theory really were a reincarnation of Paley’s natural theology, then Mr. Berlinski might be right that it is in danger of collapsing without a glimpse into the inscrutable mind of God. But it is not. It is an attempt to give a better explanation than chance and necessity for what our senses tell us is evidence for design in living things. In that light, it is poor Darwin who is (as Mr. Berlinski declares of Paley) “dead at last, or at least not very vigorously alive.”

Discovery Institute
Seattle, Washington


I always find David Berlinski’s writing delightful, and I agree with much that he says in ‘Has Darwin Met His Match?’ Specific claims about how life arose in the murky past should always be examined skeptically, especially if accompanied by grand philosophizing. On the other hand, the fact that much remains mysterious does not mean we cannot conclude anything at all with reasonable certainty.

On the general question of the sufficiency of unintelligent physical processes to produce the astonishing complexity of life, I think a negative answer is justified, for reasons I gave in my book, Darwin’s Black Box (1995). I quite agree with Mr. Berlinski that my argument against Darwinism does not add up to a logical proof. No argument that rests on empirical observations can have such force. Yet, despite my sloppy prose in suggesting that, “by definition,….irreducibly complex systems cannot be approached gradually” I intended the argument to be a scientific one, not a purely logical one. In a scientific argument, conclusions are tentative, based on the preponderance of the physical evidence, and potentially falsifiable.

Here is my thumbnail sketch of the modern design argument as I see it: either unintelligent processes can explain all of life or they cannot. Virtually everyone (including Darwinists) agrees that life appears to be intelligently designed. The only physical mechanism ever proposed that could plausibly mimic design is Darwinian natural selection. Yet, as I have argued, the irreducible complexity of biochemical systems is a barrier to direct evolutionary construction by natural selection, leaving Darwinists to hope for circuitous, indirect routes. No plausible indirect routes have been proposed, let alone experimentally demonstrated.

That leaves us with biological features that look designed, but only promissory notes for how they can be explained by unintelligent processes. What’s more, we know why the features we see in biological systems look designed: they are at once functional and very unlikely—exactly what William Dembski, whose work Mr. Berlinski also discusses, means by his phrase “specified complexity.” They look designed for the same reason that nonbiological artifacts like mousetraps look designed, and non-design explanations have turned out to be so much bluster.

It seems reasonable to me to conclude, while acknowledging our fallibility, that at least some features of life were really designed by an intelligent agent.

Lehigh University
Bethlehem, Pennsylvania


David Berlinski provides a clear and popular summary of my work on the theoretical basis for detecting design in nature. He also articulates several criticisms of my work. As it turns out, I specifically formulated my theory to meet the concerns that he raises. I am thus grateful for the opportunity to clarify several key points about my theory.

(1) As Mr. Berlinski explains, central to my theory of design detection are the twin notions of small probability and specification. I argue that highly improbable events that conform to independently given patterns are correctly attributed to intelligent design. Mr. Berlinski correctly points out, however, that some patterns are subjectively imposed upon events (or perceived in events) and do not justify inferring design. He is absolutely correct as far as he goes.

But he misses a critical distinction in my work. In The Design Inference (1998), I explain that there are artificially constructed patterns—I call them fabrications—that do not justify design inferences. There are other kinds of patterns that I call specifications, and these, in the presence of small-probability events, do justify design inferences. Moreover, I show that there is a clear way to distinguish specifications from fabrications. Specifications are patterns that, in the parlance of probabilists, are conditionally independent of the outcomes that they characterize and that, in the parlance of complexity theorists, exhibit a low minimum-description length (see Fabrications fail this test.

The distinction between specifications and fabrications is readily illustrated. Consider an archer who shoots at a target. If the target is fixed and the archer repeatedly hits the bull’s-eye, then we rightly draw a design inference by attributing skill to the archer. On the other hand, if the target is movable and always moves to where the arrow lands, then we may not draw a design inference by attributing skill to the archer. In the latter case, the target is a fabrication, in the former a specification.

(2) Mr. Berlinski is right that probabilities sometimes cannot be objectively assigned to various events. But sometimes they can be. And sometimes, when exact probabilities cannot be assigned, credible upper bounds can be. This suffices for a design inference. Sometimes probabilities can be determined on theoretical grounds. Sometimes they can be determined only on empirical grounds, as by running experiments or performing computer simulations. Assigning probabilities to biological systems to determine whether they are designed is an exciting area of research opened up by intelligent design.

(3) Throughout my writings I stress that the absence of specified improbability cannot rule out design, because a designing intelligence can act carelessly, or even deliberately, in ways that do not produce specified events of small probability. For instance, I might deliberately tip an inkwell so that the resulting ink stain is indistinguishable from a random accident. But with that same ink I might also spell a Shakespearean sonnet. In the latter case, the resulting ink “stain” would exhibit specified improbability and could not reasonably be referred to chance.

Thus, while I have always conceded that specified improbability is not a necessary condition for design, I have consistently argued that specified improbability is a sufficient condition for it. Mr. Berlinski argues against this, but his argument hinges on a failure to distinguish specifications from fabrications.

(4) According to Mr. Berlinski, highly improbable events happen “precisely as many times as one might expect, given their probabilities.” This claim is easily disproved by flipping a coin a thousand times. The probability of the sequence you get is around one in ten raised to the 300th power. How often should you have expected this sequence to occur? Not at all! With all the elementary particles in the universe furiously flipping coins for trillions of years, the expected waiting time for a given sequence places it well beyond the predicted heat death (or big crunch) of the universe.

I still hope to persuade Mr. Berlinski that his concerns about detecting design can be (or have already been) satisfactorily addressed. In any case, he has identified several key issues raised by my theory, and I look forward to the critical conversation that his piece will engender in the design-theoretic research community.

Baylor University
Waco, Texas


As an admirer of David Berlinski’s intellectual stubbornness and independence, I welcome his critical scrutiny of the theory of intelligent design. No theory was ever improved by being coddled.

Still, when Mr. Berlinski writes that design theorists “underestimate the enduring intellectual force behind [Jacques] Monod’s claim that the categories of chance and necessity are mutually exclusive and jointly exhaustive,” I must note that Monod himself did not rely solely on these categories. No sane human being does. The argument of Monod’s masterpiece, Chance and Necessity (1970), leans heavily on the notion of “teleonomy,” which he defines as a “characteristic property” of organisms as “objects endowed with a purpose or project.” In other words: objects marked by teleology—or, if you will, by design.

But renaming a property to take away its metaphysical sting should fool no one. Monod claimed that he could dissolve teleonomy—the unmistakable designedness of organisms—into chance and necessity. Well, he did not, and if design theorists are right, he could not. Chance and Necessity is one long dance around the problem, ending with Monod’s leaping into the arms of “a unique occurrence”: the causally inexplicable origin of life on earth, an event, Monod observed, whose “probability was virtually zero.”

Chance abused in this way can “explain” anything. As a philosophical naturalist, Monod was of course being true to his principles. But let us not give this sort of reasoning the good name of science. It is a philosophy—one contender among many—and not much of a contender at that.

The great theme that unites the intelligent-design community is the falsity of naturalism as a philosophy of explanation. Chance and necessity do not exhaust the causes that we know. The task facing design theorists is to turn this intuition into knowledge, by showing that their theory provides understanding and discoveries not forthcoming within a strictly naturalistic framework.

Mr. Berlinski doubts that this is possible. Time will tell. Paley is dead; so is Darwin; so, too, is Monod. Let the dead bury the dead.

Discovery Institute
Seattle, Washington

David Berlinski Responds:

As Jonathan Wells writes in his admirably clear letter, the gravamen of the intelligent-design movement is the inference to design, an inference that carries us from the observed properties of biological systems to the existence of a designer. The identification of the designer with the Christian deity requires a separate inference, one that design theorists need not make. I accept the point. In my essay, I tried to suggest as much, at least implicitly, by incorporating Fred Hoyle’s supercalculating intellect into the class of possible designers. Hoyle was a life-long atheist, and in reading him I find no reason to believe that in fundamental matters he ever changed his mind.

Still, I admit to a tickle of discontent at Mr. Wells’s attempt to dissociate the designer’s existence from the designer’s identity. It may be helpful to put the matter in deductive form. If there is design in nature, then there is a designer: that is Mr. Wells’s leading premise. A second premise follows, embodying a factual claim: there is design in nature. The conclusion that there is a designer follows as a matter of logic.

But, articulated in this way, the conclusion seems suspiciously trivial. If the designer’s nature is not known, then in what sense does his existence tell us anything that is not expressed in the second premise? To be sure, the twin propositions that there is design in nature and that there is a designer do not say quite the same thing; but, like certain houses in Paris, they have a tendency to collapse toward one another. Which designer? The one handling the design. Which design? The one handled by the designer. In the end, we are left with the fact that certain properties of living systems have not been explained. On this point Mr. Wells and I agree completely.

In characterizing his own work, Michael J. Behe is entirely too unassuming. By demonstrating that irreducibly complex systems cannot arise along one Darwinian path—the assembly-line model of construction—he has indeed provided a logical argument against Darwin’s theory. This is an important achievement. My point was only that his argument has not been generalized to include all Darwinian paths. This is a modest criticism.

As long as I find myself explaining Mr. Behe’s achievements to Mr. Behe, let me go a bit further. It seems to me that Darwin’s Black Box is destined to play the same role in evolutionary thought that Karl Lashley’s “The Serial Order of Behavior” played in behavioral psychology over fifty years ago (cf. L.A. Jefress, editor, Cerebral Mechanisms in Behavior, 1951). Although a behaviorist by training—and a student of John Watson, the founder of the field—Lashley came to understand that a certain class of familiar psychological acts involve a complicated serial order. In formulating a sentence, we typically see to the end of the sentence before venturing on its beginning, adjusting our stream of speech accordingly to account for grammatical relations of subordination and deferred placement. Serial order, Lashley realized, could not be explained by any system of associative “chaining” in which each act in a behavioral repertoire is explained by an act that has already taken place.

Psychologists were slow to appreciate the force of Lashley’s critique. But when Noam Chomsky, writing ten years later, pointed to certain features of natural language that were inaccessible to finite-state mechanisms or even Markov processes, he was sharpening and extending Lashley’s insight. Thereafter, a shrewd insight became a movement in thought: the so-called cognitive revolution.

It is curious that so few biologists appreciate the formal similarities between Darwin’s theory of evolution and behavioral theories in psychology. But the same basic ideais at work, with “reinforcement” in psychology called “natural selection” in biology. And here is an odd point. Although behaviorism is widely thought to have been stabbed through the heart, most especially by Chomsky’s criticism of B.F. Skinner, the fact of the matter is that in the end, Skinner has proved more durable than anyone might have guessed, and with bat-wings flapping has burst buoyantly from his crypt.

We know that behavioral psychology provides an explanation for a limited class of experimental results; beyond those results, it fails. What an individual does in acquiring a natural language cannot be explained in terms of any schedule of reinforcement. Indeed, it cannot be explained in terms of the environment at all; reference must be made to the individual’s innate endowment. But linguists who came early to scoff at Skinner—Chomsky now included—are involved in making Skinner’s argument on the level of the species. For what are random variation and natural selection but a form of behavioral biology?

The intellectual history of the past half-century is not without its ironical aspects.

In his graceful letter, William A. Dembski argues that in some respects I have misunderstood his views, and in other respects I have insufficiently appreciated their force. It is certainly possible; the issues that Mr. Dembski raises are subtle, and his letter and my response should both be regarded as efforts to get things a little clearer.

Under what conditions may we rationally eliminate chance as the explanation for certain events? This is Mr. Dembski’s question. Two things, he argues, are necessary: small probabilities and reasonably tight specifications. When these criteria are met, we are entitled to strike off chance; if the event in question is also not explicable by natural laws, design emerges as the only plausible alternative.

In developing his argument, Mr. Dembski has a certain model in mind. The design comes first, expressed perhaps as a blueprint, agenda, schedule, or even a system of thought. Next comes the designed event or object. “How a designer,” he writes in No Free Lunch, “gets from a thought to a thing is, at least in broad strokes, straight-forward: (1) A designer conceives a purpose. (2) To accomplish that purpose, the designer forms a plan. (3) To execute that plan, the designer specifies the building materials and assembly instructions. (4) Finally, the designer or some surrogate applies the assembly instructions to the building materials. What emerges is a designed object, and the designer is successful to the degree that the object fulfills the designer’s purposes.”

This is executive design, to coin a phrase and mark a distinction. In my essay, I suggested that there is a range of states, acts, or processes that are clearly intentional—they are brought about by intelligent agency—and yet share none of the features of executive design. The design of a painting is very often revealed in its execution and not before. Design in this sense might well be called immanent. The painter Francis Bacon often stressed just this point in commenting on his own work (see Francis Bacon, 1975), and the distinction between executive and immanent design appears as well in Nelson Goodman’s Languages of Art, a book that design theorists might study with profit. With respect to immanent design, there are no prior purposes, no plans, and no application of assembly instructions to building materials. For this class of artifacts, probabilities are not relevant and specifications are inapplicable.

I am well aware of Mr. Dembski’s distinction between a specification and a fabrication, the more so since I discussed the distinction at length in my book Black Mischief: Language, Life, Logic & Luck (1986), where Mr. Dembski’s “fabrications” appear as “retroactive specifications.” But this distinction has nothing to do with the distinction between executive and immanent design. Both specifications and fabrications apply, and apply only, to events capable of possessing an executive design. Velasquez’s painting of the royal court, which I cited in my Commentary essay, lacks both a specification and a fabrication. A painting is not like a point in archery. It cannot be specified before it exists, and specifying it after it exists is useless, voiding the intended contrast between a specification and a fabrication.

Although both Mr. Dembski and I agree that specified improbabilities are not necessary to trigger a design inference, we do so for different reasons. On the question of whether they are sufficient, we part company altogether. In “Has Darwin Met His Match?” I argued that specifications are not necessary to trigger a sense that an unlikely event has occurred, and that if specified improbabilities are not necessary, they are also not sufficient to trigger an inference to design, since plainly improbability by itself tells us nothing about design.

The issue, then, turns on a tight circle. My argument for the irrelevance of Mr. Dembski’s specifications begins with the obvious. A specification is a human gesture, a bit of descriptive apparatus. It does not change a system of probabilities, and so it has nothing to do with expectations based on those probabilities. Highly improbable specified events happen precisely as many times as one would expect, given their probabilities, but so, for that matter, do highly improbable unspecified events.

It is this last claim that Mr. Dembski would deny. His counterexample is a fair coin flipped 1,000 times. The particular sequence of heads and tails that results has the probability of roughly one in ten to the 300th power. “How often should [I] have expected this sequence to occur?” Mr. Dembski asks. Not at all, he responds cheerfully. But the sequence has occurred, and the relevant record of heads and tails is there in plain sight. This persuades Mr. Dembski that, if left unspecified, unlikely events tend to occur more often than any assignment of probabilities would suggest.

I understand Mr. Dembski’s reasoning, but I reject it as unwholesome. He is stuck, after all, with an expectation—a particular sequence should not occur at all—that is in plain contradiction to the facts—a particular sequence has indeed occurred. This is not necessarily a fatal flaw, but it does suggest that something has to give. What has to give is the nice coincidence between what the theory of probability affirms and what the facts reveal. Here Mr. Dembski’s intuitions seem to be making certain clanging noises while emitting a good deal of smoke.

My own intuitions, by contrast, are smoke-free and purr like a charm. Is the sequence that has just been revealed highly improbable? It is. How often should it have occurred? Precisely as many times as one might expect, given its probability. In a sequence of identical and independent experiments, this particular outcome would not be realized again until the heat death of the universe or the rehabilitation of Senator Trent Lott, whichever comes first. The fact that the sequence has already occurred is of no significance. Nothing in the theory of probability prevents an extremely unlikely event from being realized in the very first experiment, just as nothing in the same theory prevents a man from winning the lottery after purchasing his very first ticket. I see no reason to be swayed from my conclusion that unlikely events happen as many times as one might expect, given their probability.

With the distinction between specified and unspecified improbabilities wiped out, the inference to design sputters. Improbabilities are not by themselves sufficient to trigger that inference, and, a fortiori, neither are specified improbabilities.

Paul A. Nelson reminds me that, in Chance and Necessity, Jacques Monod began by acknowledging the fact that living creatures are driven by a sense of their purpose in life, and then persuaded himself that this obvious property was not so obvious after all. What seems to be their design, Monod concluded, is an illusion, or an artifact; there is necessity, and there is chance, and there is nothing more. Richard Dawkins has made almost the same argument in The Blind Watchmaker, and Mr. Nelson is right to remark that there is in all this something nutty. Monod’s efforts to explain away the obvious did not succeed.

Of course, there is a distinction between saying that Monod did not succeed and saying that he could not have succeeded. Drawing that distinction, Mr. Nelson places his hopes for the latter possibility on the prospects for intelligent design. But to the extent that design theorists have overestimated their own arguments, to that extent—precisely—have they underestimated the enduring force of Monod’s claim that necessity and chance exhaust our powers of explanation. In my essay, I went no further than this, and I am not prepared to go further now.

Whatever the issue between Mr. Nelson and Jacques Monod, it is (as he says) naturalism that is the real target of the intelligent-design movement, and, I presume, his target as well. If I cheer him on only weakly, that is because it is not entirely clear to me what the target is or how to fight it. From what he writes, I gather that naturalism is embodied in Monod’s disjunction itself. But an inference to design is consistent with Monod’s disjunction. A designer may have no choice in his design, a point I make in the concluding paragraphs of my essay. But, equally, a denial of Monod’s disjunction is consistent with the failure of an inference to design. There may be facts in nature that are the result of neither necessity, chance, nor design, and we must accept the possibility, at least, that the most obvious facts about living systems—their existence and their nature—may have no deeper explanation than that this is the way things are.

Something more is at issue in the assessment of naturalism because something more must be at issue. I encourage Mr. Nelson to make that something clearer.

See all the letters and responses at: