Philosophical-ish Objections to Intelligent Design: A Response to Paul Draper

Crossposted at Evolution News

Recently I was asked by several people whether I had ever responded to an old review of Darwin’s Black Box by Purdue University philosopher of religion Paul Draper. I had not done so, but will use the occasion to respond now and to clear up a couple of philosophical-ish objections that have been raised against intelligent design over the years.

In 2002 Draper — then on the faculty of Florida International University — published a paper in the journal Faith and Philosophy: Journal of the Society of Christian Philosophers, entitled “Irreducible complexity and Darwinian gradualism: a Reply to Michael J. Behe.”1 Draper wrote that “My goal in this paper will be to show that, while this challenge is both more original and, with a few modifications, more powerful than many of Behe’s critics realize, it is incomplete and for that reason does not refute Darwinism.”

I appreciate his recognition of the argument as a potent one and I’m grateful for Draper’s defense of it against some other objections. However, his own main line of criticism echoes one from the evolutionary biologist H. Allen Orr2 (Draper cites Orr) that I had already addressed,3 and Draper’s charge of incompleteness rests on neglecting differences between a philosophical argument and a scientific one.

I Say Mousetrap; He Responds … “A Then B”

Let’s start with Allen Orr. In 1996 H. Allen Orr reviewed Darwin’s Black Box for Boston Review. Orr correctly described the problem of irreducible complexity I had pointed out:4

You cannot, in other words, gradually improve a mousetrap by adding one part and then the next. A trap having half its parts doesn’t function half as well as a real trap; it doesn’t function at all. So Darwinism’s problem is obvious: it requires that each step in the evolution of a system be functional and adaptive.

And he agreed that such systems can’t be developed by recruiting parts from other systems:5

[W]e might think that some of the parts of an irreducibly complex system evolved step by step for some other purpose and were then recruited wholesale to a new function. But this is also unlikely. You may as well hope that half your car’s transmission will suddenly help out in the airbag department. Such things might happen very, very rarely, but they surely do not offer a general solution to irreducible complexity.

Orr then proposed a solution:6

Behe’s colossal mistake is that, in rejecting these possibilities, he concludes that no Darwinian solution remains. But one does. It is this: An irreducibly complex system can be built gradually by adding parts that, while initially just advantageous, become — because of later changes — essential. The logic is very simple. Some part (A) initially does some job (and not very well, perhaps). Another part (B) later gets added because it helps A. This new part isn’t essential, it merely improves things. But later on, A (or something else) may change in such a way that B now becomes indispensable. This process continues as further parts get folded into the system. And at the end of the day, many parts may all be required.

I pointed out in my response to his review that his proposed solution was quite vague and that he had dodged the critical question:7 “Some part initially does some job? Which part of the mousetrap is he talking about? A mouse has nothing to fear from a ‘trap’ that consists of just an unattached holding bar, or spring, or platform, with no other parts.” Yes, new ideas are often hazy, but in order to be considered scientific they must progress beyond hand-waving. Orr gives credit for the original “A then B” scenario to the early 2oth century geneticist Hermann Muller. Yet in 1918 — when, according to Orr, Muller first proposed it — no one knew what genes even were! The structures of proteins — the machinery of the cell — would not begin to be elucidated for another four decades. Since the 195os, however, biology has made tremendous progress in understanding the molecular basis of life. In the 21st century — more than a hundred years after Muller’s proposal — it is well beyond time to stop hiding behind letters. Exactly what are those magical parts “A” and “B”? And exactly what are they supposed to be doing?

Orr went on to suggest that the transformation of air bladders of fish into lungs of terrestrial vertebrates illustrated his point:8

The transformation of air bladders into lungs that allowed animals to breathe atmospheric oxygen was initially just advantageous: such beasts could explore open niches — like dry land — that were unavailable to their lungless peers. But as evolution built on this adaptation (modifying limbs for walking, for instance), we grew thoroughly terrestrial and lungs, consequently, are no longer luxuries — they are essential.

So apparently he hadn’t grasped the concept of irreducible complexity after all. I had stressed in Darwin’s Black Box that it is the molecular level of life that must be examined to decide about irreducibility. Complex tissues such as air bladders and lungs are not “single [i.e., molecular] systems.” When one writes glibly about the evolution of whole organs (as Darwin himself did with the eye9), whose molecular components are many and mostly unknown, one quickly descends into fantasy. Notice that Orr doesn’t tell us in sufficient detail — or any detail, for that matter — how a Darwinian process could transform even an air bladder into a lung. Like most other evolutionary biologists, he simply assumes it can.

Six Years Later

Professor Draper’s line of reasoning from 2002, which is also an imaginary exercise of evolution-by-letter, misses the same point as does Orr’s.10

[A]n irreducibly complex two-part system AB that performs function F could evolve directly as follows. Originally, Z performs F, though perhaps not very well. (This is possible because, from the fact that AB cannot perform F without A or B, it doesn’t follow that Z cannot perform F by itself.) Then A is added to Z because it improves the function, though it is not necessary. B is also added for this reason. One now has a reducibly complex system composed of three parts, Z, A and B. Then Z drops out, leaving only A and B. And without Z, both A and B are required for the system to function. One might object that in this scenario AB is not really produced directly because, although the route from Z to AB involves no change in function, it must involve a change in mechanism since interaction between two distinct parts is essential to AB’s mechanism and so Z cannot function by the same mechanism as AB. If this is right, then no complex system of interacting parts, whether irreducibly complex or not, could evolve both gradually and directly from scratch. An indirect route would always be required in order to get started. But that’s no problem for this model. For I can simply stipulate that Z itself is composed of two parts (Z1 and Z2), which do perform AB’s function by the same mechanism, and then add that Z1 drops out first, followed by Z2.

In addition to being terminally fuzzy (like Orr’s), Draper’s proposal overlooks key problems of his own scenario. For example: 1) The fact that the necessary parts of Z (“Z1 and Z2″) may be attached to each other (that’s how I understand “composed”), as he stipulates, instead of physically separated does not obviate the problem, as he seems to think; rather, it makes the problem worse. The parts must be connected to each other in the correct way — more irreducible requirements. For example, all the parts of a mousetrap are attached in specific orientations (if, say, the spring were facing the wrong way, the trap would fail). Yet the unguided production of even that comparatively simple machine has resisted the efforts of Darwin defenders on the Internet for 25 years.11 2) If, as Draper concedes, Z is composed of two necessary parts, then of course one is starting with irreducible complexity, not explaining it. 3) Even if Z is already working (somehow) and the subsequent addition of A and then B improves it, as Draper posits, why in the world should we think A and B can later take over Z’s role? Improving the function of Z is a different task from performing the function of Z. Placing a drop of oil on the spring and a piece of cheese on the platform might improve a mousetrap’s function, but certainly neither alone nor both together would make the other parts dispensable. If A and B did take over Z’s role later, it would seem to be a gratuitous development, no different in kind or difficulty from the postulated irreducibly complex starting point.

Another putative route to irreducible complexity that Draper explores is also an imaginative exercise in evolution-by-letter.12

The sort of route I have in mind occurs when an irreducibly complex and irreducibly specific system S that serves function F evolves from a precursor S* that shares many of S’s parts but serves a different function F*. Notice that parts that S and S* share and that are required for S to perform F need not be required for S* to perform F* even if they contribute to F*, and parts that are irreducibly specific relative to F may be only reducibly specific relative to F*. Thus, both S* and the specificity of its parts may have been gradually produced via a direct evolutionary path. Then one or more additional parts are added to S*, resulting in a change of function from F* to F. And relative to F, the parts and their specificity, which had not been essential for F*, are now essential.

I’m afraid I find that scenario very unclear. Wouldn’t it be great if Professor Draper offered a concrete illustration of what he meant here? But he doesn’t, and I can only conclude that he can’t. Rather, the argument strikes me as little more than word play — just a riff on Hermann Muller’s hundred-year-old-idea from a time when no one knew much about the molecular foundation of life. The only physical illustration Draper cites is a standard mousetrap that has been degraded and purposely manipulated. In an endnote Draper explains:13

These claims are easy to demonstrate. Simply remove the catch from a trap and bend the holding bar so that (roughly) the middle of it can be placed just barely under the tip of the slightly curved end of the spring that extends under the hammer when it is armed. This will make it possible to arm the hammer without a catch, and the closer to the tip of the spring the holding bar is placed, the more insecurely the hammer will be armed. Additionally, one can make the trap even more sensitive to pressure on the platform by allowing the end of the holding bar to extend below the platform when the hammer is armed (so that the platform will not lie flat on the floor when the trap is set).

With his proposed very careful rearranging of the functioning standard mousetrap, the intelligent agent Paul Draper runs squarely into a problem that I had flagged in Darwin’s Black Box:14

To feel the full force of the conclusion that a system is irreducibly complex and therefore has no functional precursors we need to distinguish between a physical precursor and a conceptual precursor. The trap described above is not the only system that can immobilize a mouse. On other occasions my family has used a glue trap. In theory at least, one can use a box propped open with a stick that could be tripped. Or one can simply shoot the mouse with a BB gun. However, these are not physical precursors to the standard mousetrap since they cannot be transformed, step-by-Darwinian-step, into a trap with a base, hammer, spring, catch, and holding bar.

Perhaps because of professional inclinations, philosopher Draper does not even try to argue that his imagined mousetrap is a physical (rather than just a conceptual) precursor to the standard mousetrap I discussed. I responded to a similar intelligently guided mousetrap-scenario by Professor John McDonald of the University of Delaware years before Draper’s paper was published.15 Mousetraps can of course be built in many different ways. But McDonald’s and Draper’s examples are not Darwinian, single-random-step precursors to a standard mousetrap. Rather, they are intelligently re-engineered systems that resemble the trap I pictured only because that is the goal they are intended to reach. As I wrote in my response to McDonald, those may be conceptual precursors to a standard mousetrap, but they are not physical precursors and therefore not Darwinian precursors either.16

It’s a lot easier to imagine that, through a long series of unexplained steps, “A” could morph into “B” than that, say, a glue trap could morph into a standard mechanical mousetrap. I think that is one crucial difference between philosophical approaches and scientific ones. If intermediate steps are not enumerated and tested for a physical pathway, then a theorist is in serious danger of getting stuck in a fantasy world. Recall that evolutionary biologist Allen Orr agreed that borrowing parts used for something else is quite unlikely to work for irreducibly complex systems: “You may as well hope that half your car’s transmission will suddenly help out in the airbag department.” Draper’s schemes seem to be built entirely on such hopes. Yet if Draper and others can’t account for even a humble mousetrap (even if we supposed, for the sake of argument, that mousetraps reproduced and were subject to small mutations), why should we think Darwin’s theory can account for the sophisticated machinery of the cell?

In Darwin’s Black Box I argued that for all practical purposes irreducibly complex biochemical systems met Darwin’s challenge that “If it could be demonstrated that any complex organ existed which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down.”17 Twenty five years later there is still no reason to think otherwise.

The State of the Literature Circa 2020

Near the end of his review from 2002 Professor Draper writes:18

A defender of Behe might respond that the improbability of all Darwinian paths is established by the silence of the scientific literature on the issue of how Behe’s systems evolved. … Should we conclude, then, that this silence provides substantial indirect evidence for Behe’s position that Darwinism cannot explain the development of these systems? In other words, is it likely that the literature would be much noisier if Behe’s position were false? I am inclined to give negative answers to these questions for two reasons. First, the discipline of biochemistry is very young.

The discipline of biochemistry is now nearly 25 years older than when Darwin’s Black Box was published. Yet, as I document in my latest book, Darwin Devolves, the state of the literature regarding irreducibly complex systems is unchanged — even as problems for Darwin have multiplied.19 Although the science of life at the molecular level has advanced by leaps and bounds, that new work conspicuously does not include explanations for how Darwinian processes could produce irreducibly complex systems. Despite the intervening years, despite the immense progress of science, and despite the intense dislike of intelligent design to motivate many very smart scientists, even the examples I highlighted in 1996 have gone completely unexplained. That in itself is strong evidence that Darwinists have been barking up the wrong tree.

Draper adds:20

Second, any indirect evidence that the silence of the literature provides in support of Behe’s position is offset by indirect evidence against Behe’s position. For in assessing the probability that some system evolved gradually, one cannot just examine the specifics of that system. One must also consider what we know about how other systems evolved. And Behe admits, as well he should, that much evolutionary change is both gradual and driven by Darwinian mechanisms. These Darwinian success stories raise the probability that Darwinian evolution produced Behe’s biochemical systems, even if we cannot yet specify step by step exactly how that happened.

Time has not been kind to Draper’s position. As I discuss in Darwin Devolves, we are now in a much better position than we were in 1996 to see how Darwinian processes work in nature, rather than in people’s imaginations. Turns out there is indeed much evolutionary change that is both gradual and driven by Darwinian mechanisms, as Draper noted. However, the surprise — made possible by recent advances in DNA sequencing techniques — is that the large majority of even helpfulbeneficial mutations swept through a population by natural selection are degradative. That is, they break or damage pre-existing genes. One can call examples of evolutionary degradation “success stories” if one wishes. Degradative processes, however, are prohibitively unlikely to have initially constructed the elegant biochemical machinery modern science has discovered in the cell.

The Conflict Lies Elsewhere

In his 2011 book, Where the Conflict Really Lies: Science, Religion, and Naturalism, the Christian über-philosopher Alvin Plantinga discusses my Darwin’s Black Box in connection with Paul Draper’s objections and says explicitly that he thinks Draper is right. Yet, in support of his agreement, Professor Plantinga quotes the same fuzzy paragraph of Draper’s about S, F, S*, and F* that I quoted here. Plantinga then emphasizes a critical difference between philosophical and scientific arguments:21

It’s important to note that the possibilities Draper suggests are merely abstract possibilities. Draper doesn’t argue or even venture the opinion that in fact there are routes of these kinds that are not prohibitively improbable; he simply points out that Behe has not eliminated them. And of course this is quite proper, inasmuch as Draper is doing no more than evaluating Behe’s argument. All he is trying to show is that Behe’s conclusion doesn’t deductively follow from his premises.

But no scientific argument simply follows deductively from its premises. An argument about nature necessarily concerns detailed empirical facts — evidence — so it cannot be a completely deductive one. Although scientific thinking of course involves logic and deduction, all scientific arguments — including the argument for the irreducible complexity of the molecular machinery of life — rest most heavily on empirical data. No scientific explanation has ever ruled out all possible rival explanations by dint of deductive logic alone. Draper used a wholly inappropriate standard to evaluate my argument, one that is not applied to any other scientific theory.

Plantinga goes on to discuss my argument from The Edge of Evolution that empirical results — especially data on the development of resistance to the antimalarial drug chloroquine — show there is a real limit to what Darwinian processes can explain in life. He then wonders whether this makes it likely that life was designed:22

How does the argument go? One possibility: the main alternative to intelligent design is unguided evolution, but the probability that unguided evolution should produce these protein machines is so low that we must conclude that it is false. Is this right? Not clearly. First, exceedingly improbable things do happen, and happen all the time. Consider a deal in a hand of bridge. If you distribute the fifty-two cards into four groups of thirteen cards each, there are some 1028 possible combinations. Therefore the probability that the cards should be dealt just as they are dealt is in the neighborhood of 10-28. So consider a rubber of bridge that takes four deals: the probability that the cards should be dealt precisely as they are, for those four deals, is about 10-112.

Since improbable-yet-fair bridge hands are dealt all the time, he implies, low probability alone can’t be the criterion for design.

True enough, as far as it goes. But that doesn’t mean we can’t confidently recognize design in low probability events. This line of thought was explored by William Dembski in his 1998 The Design Inference. He explained that an inference to design required not only an event of small probability, but also one that was specified — that is, one that hit a pre-determined or retrospectively surprising target. For example, there are of course very large numbers of possible bridge hands, but if a person at a table kept getting hands with 13 cards of the same suit, eyebrows would be raised. The “lots of things are improbable” tack to explain away the elegant structure of the machinery of life is simply not a serious response.

Professor Plantinga then raises an objection first put forward by the philosopher of science Elliott Sober, that we need to have a positive reason to think a designer would want to make cellular machinery to conclude that it did.23 Plantinga seems to agree — a designer who didn’t want to design machinery probably wouldn’t do so.24

Well, should we instead compare P(protein machines/unguided evolution), the probability of the existence of these protein machines given unguided evolution, with P(protein machines/intelligent designer), the probability of the existence of these protein machines given the existence of an intelligent designer? Here, again, the problem is that we don’t have a very good grasp of either of those probabilities. … Suppose there is an intelligent designer: how likely is it that he or she (or it) would design and cause to come to be just these protein machines? … An intelligent designer that really hated life, or proteins, or protein machines, would be very unlikely to design protein machines.

Try as I might, and as admiring as I am of other work of both Plantinga and Sober, I see absolutely no force to this argument, and it is quite susceptible to reductio ad absurdum: Sure, a designer who didn’t want to design something probably didn’t design it. But let’s not stop there. After all, a designer who really did want to design machinery, but got distracted and forgot, wouldn’t have designed it either. Nor would a designer who wanted to design but went on vacation instead. Or one that came down with the flu. Or was stopped by someone else. Or got its head stuck in a vase. Do we really have to individually eliminate all those and many other possibilities before coming to a conclusion of design?

I don’t think so. Rather, the rejection of all those hypotheticals is baked into the positiveaffirmative design argument. The problem with the Sober approach, I think, is that it seems (perhaps unconsciously) focused on a particular candidate for designer — maybe God, maybe a space alien, whatever — and imagines that the proper question to ask is, how do we know that this particular candidate has the motivation to design something? Wrong question. The intelligent design argument rightly starts with the physical system itself and asks, Does this system exhibit a strongly purposeful arrangement of parts?

(As an aside, the major focus of my books on demonstrating the stark inability of Darwinian processes to achieve what proponents claim for them is of course necessary in order to dispel that common misconception. But a conclusion of design is not, as the caricature frames it, simply a default judgment rendered in the absence of some other plausible explanation. Rather, as I have explained in a number of places,25 design — the work of a mind — is positively perceived in a purposeful arrangement of parts.)

It would be silly to examine, say, Mount Rushmore, or the monolith found on the Moon in the movie 2001: A Space Odyssey, and say no inference about their possible design could be made because we didn’t know whether a designer wanted to make either of them. Rather, we conclude from observation of a system itself that it was designed. Indeed, interesting secondary questions (such as who designed it, whenhowwhy, and so on) can’t even be posed unless we already suspect that a system was designed.

Photo: Michael Behe and the iconic mousetrap, in a scene from Episode 2 of Secrets of the Cell with Michael Behe.


  1. Draper, Paul. 2002. “Irreducible complexity and Darwinian gradualism: a Reply to Michael J. Behe.” Faith and Philosophy 19:3-21.
  2. Orr, H.A. 1996. “Darwin v. intelligent design (again).” Boston Review, 1996, 28-31.
  3. Behe, Michael J. 1997. “The sterility of Darwinism.” Boston Review, 1997, 24.
  4. Orr, 1996. 
  5. Orr, 1996.
  6. Orr, 1996.
  7. Behe, 1997.
  8. Orr, 1996.
  9. Darwin, C. 1859. The Origin of Species. New York, Bantam Books (1999), pp. 155-158.
  10. Draper, 2002.
  11. Behe, M. J. 2019. Darwin Devolves: the New Science about DNA That Challenges Evolution. New York, NY: HarperOne, Chapter 9.
  12. Draper, 2002.
  13. Draper, 2002.
  14. Behe, M. J. 1996. Darwin’s Black Box: The Biochemical Challenge to Evolution. New York: The Free Press, pp. 43-45.
  15. McDonald, J. H. 2000. A reducibly complex mousetrap.
  16. Behe, M. J. 2000. A Mousetrap Defended.
  17. Darwin, 1859, p. 158.
  18. Draper, 2002.
  19. Behe, 2019, pp. 283-301. 
  20. Draper, 2002.
  21. Plantinga, A. 2011. Where the Conflict Really Lies: Science, Religion, and Naturalism. New York: Oxford University Press, p. 231.
  22. Plantinga, 2011, pp. 235.
  23. Sober, E. 2002. “Intelligent Design and Probability Reasoning.” International Journal for Philosophy of Religion 52:65-80.
  24. Plantinga, 2011, pp. 235-236.
  25. Most recently in Behe, 2019, Ch. 10.