Is the biological evidence for ‘irreducible complexity’ best understood in terms of a design argument or design discourse?

Michael Behe has suggested that many structures found in biological systems are irreducibly complex, that is, complex in such a way that they could not have arisen via indirect, chance development but bear clear evidence of having been designed. This evidence of design comes from the fact that these systems have a number of parts without which the mechanism in question simply does not function—not with degraded or partial function, but no function at all. He offers up several prominent examples of this phenomena, among them the human eye and the bacterial flagellum. With each, he shows how problematic a materialist evolutionary account for their development must be, charting six steps in “design space” which any such account must fulfill. Take the human eye: (1) Development begins with a light-sensitive spot. (2) All steps which finally result in a human eye occur via heritable genetic variation. (3) At no point does a designer intervene; all steps occur solely by random mutation. (4) Each step along the way must be either adaptive, conferring greater survivability and reproductive success on the organism, or at least not maladaptive, making the organism less likely to survive. (5) No step may be exceptionally improbable relative to the others in the sequence. (6) The final step in the sequence must be the actual, functioning human eye. Behe argues variously that such a naturalistic evolutionary path is either impossible or at least incredibly improbable for many biological systems. As such, he thinks he has evidence or proof that the systems did not develop (and perhaps could not have been developed) via purely naturalistic means.

Several major lines of criticism have been advanced against Behe’s argument, notably by Paul Draper. Draper first notes that while Behe establishes a challenge for evolutionary explanations of these biological, he by no means proves them irreducibly complex. Asserting that they are is not proving that they are. Second, and closely related, many evolutionary biologists deny that these systems, complex as they are, are irreducibly complex. A number of potential explanations have been offered for how the systems Behe describes might have developed in a way that satisfies his criteria. Third, he notes that Behe struggles to define clearly what constitutes a “part” in such systems. But since a “part” is an integral element of his scheme of irreducible complexity, it seems a definition should be forthcoming. Finally, Draper suggests that there are evolutionary phenomena capable of explaining the phenomena Behe outlines. If an element S has a function F at time T, and later at time T ʹ develops into S ʹ, which at time Tʹʹ is integrated into a larger system Q which confers functionality Fʹ, then Sʹ is not an irreducibly complex part of Fʹ. It arose independently, even though it was incorporated into the later function Fʹ.

A partial rejoinder to Draper is in order. Though he is right to note that Behe has not proven these structures to be irreducibly complex, neither have evolutionary biologists offered any actual path through the design space as hard evidence against Behe. Granted that it may not be possible for them to do so; still, telling a just-so story does not constitute a particularly strong claim. Moreover, that such a story can be told in no way makes it probable. This might defeat Behe’s stronger claim (irreducibly complex structures cannot arise by purely natural evolutionary means), but not the weaker form (it is extremely unlikely that irreducibly complex structures could arise by purely natural evolutionary means). Draper’s points do weaken Behe’s case, however.

Alvin Plantinga offers two further criticisms. First, assume Behe is right and that these developments are extraordinarily improbable. So what? Many extraordinarily improbable things have happened. If humans do inhabit a materialist universe in which no intelligent designer exists, then this extraordinarily improbable thing is precisely what happened. Second, and perhaps more serious—not only for Behe but for other Intelligent Design theorists, such as William Dembski—the prior probabilities of these outcomes are simply unknowable. One might devise some set of numbers in an attempt to represent the problem space, but such an attempt would always be fundamentally just guesswork.

As an alternative option, Plantinga suggests that Behe’s approach be taken not as an inductive argument for these systems’ being designed, but as a sort of design discourse. That is, Behe’s points are not something like an inference to the best explanation; rather, they are a normal, rational part of human cognition, and as such may be well-warranted even if a sustained argument for them cannot be mounted. Plantinga notes that many rational beliefs cannot be argued: the reliability of one’s senses, or memories, or the existence of other minds. The existence of other minds is particularly relevant, because when an explorer finds something and attributes design to it, that is just what she is doing: attributing the existence of some artifact to another mind. Such behavior is clearly rational in many situations: if astronauts on the moon discover a very large obsidian monolith, they are not obliged to chalk it up to some purely mechanistic and hitherto unknown lunar process. They immediately, and rightly attribute it to another intelligence.

Behe’s discussion seems to be just along these lines. When humans look at complex biological systems, they give every appearance of having been designed—and not only to the theists who would presumably have a bias in that direction. This appearance of design has led eminent atheists to offer up any number of comments explaining away that appearance, Sir Francis Crick and Richard Dawkins both famously among them. When atheist writers are wont to say things like “Biology is the study of things which appear to have been designed, but were not,” they strongly suggest that Plantinga’s reading of Behe may be on the right track.

Beliefs formed basically are not, of course, unassailable. A person suffering from amnesia might come to distrust her memories. If someone is informed he is under the influence of a hallucinogen, he would have reason to doubt his observations of the world. Basic beliefs can be defeated by being undercut, or by being flat-out rebutted. In the case of a design discourse, for example, the belief could be totally rebutted if one could offer proof that natural, totally unguided evolution had produced one of these biological systems. This, however, is impossible: science itself is incapable of showing that the evolutionary process was not directed by God. Even if the mutations that led to the development of the system were totally random, they could nonetheless have been planned by God in the creation of the whole universe from the beginning. Such a claim is therefore inherently metaphysical, and cannot be demonstrated empirically. It would be easier to undercut the design belief, by showing it both possible and not profoundly unlikely that an evolutionary path could have given rise to the phenomena in question. If such a path could be shown, it would indeed serve to weaken one’s confidence in a design assessment. For one thing, though, such paths do seem to be rather improbable (even if the actual probabilities are not knowable). For another, even if such an undercutting defeater were granted entirely, it might not totally defeat the basic belief, because there can be deflectors for such defeaters. If one has many other reasons (philosophical, empirical, experiential, and so on) to affirm Christianity, for example, then even a reasonable claim for the viability of some naturalist evolutionary path might not defeat the design assessment—only weaken it.

In short, Behe’s analysis qua argument has a number of weaknesses. It may serve as one part of a broader inference toward theism, but on its own merits, it does seem to fall prey to some of the critiques Draper in particular offers—not least, because Behe advances a very strong claim about the irreducible complexity of the structures in question. It falls to him to defend that claim. But taking his analysis as exemplary of a basic, and therefore well-warranted, recognition of design in biological systems seem rather more secure. Taking the two together may be the best of all. It may be possible to see and immediately recognize design, and to reason inferentially about designed mechanisms, and for both to have a confirmatory status in one’s affirmation of Christian theism.