Irreducible Complexity and the Multiverse: An African Cosmological Perspective — Epoche B2
Irreducible Complexity: The Right Objection and Two Wrong Ones What Behe claims In Darwin's Black Box: The Biochemical Challenge to Evolution (1996 [1] ), the biochemist Michael Behe introduced a concept he called irreducible complexity: a single system composed of several well-matched, interacting parts that contribute to the basic function, wherein the removal of any one of the parts causes the system to effectively cease functioning His standard example is the bacterial flagellum, a rotary motor with a driveshaft, bushings, a hook and a filament, built from dozens of distinct proteins. Take away the motor and the filament does not turn; take away the filament and the motor drives nothing. Behe's argument is that Darwinian evolution builds by slight successive modifications, each of which must confer an advantage; a system that fails entirely when a part is missing cannot have been built that way, since every intermediate would be useless. He concludes that such systems are better explained as the products of intelligent design. This column argues three things. Behe's argument is a design argument and inherits a defect that David Hume identified in design arguments generally [2] . But the reply most often reached for — the multiverse — does not work against it and should be dropped. And the cosmological ideas of another philosophical tradition are worth bringing in, though not for the reason usually given. Why it is a design argument Behe presents his case as biology rather than natural theology, and objects to being classed with William Paley [3] . The structure is nonetheless Paley's. In Natural Theology (1802) Paley asks us to imagine finding a watch on a heath. Its parts are arranged so as to produce motion that would not occur without that arrangement; the inference to a maker is irresistible; organisms exhibit the same feature more so; therefore they too were made. Behe's flagellum is Paley's watch relocated to the bacterial cell wall, with the inference sharpened by a claim about what gradual processes cannot do. Both arguments are, at bottom, probabilistic. Neither claims that a functional arrangement is logically impossible without a designer. Each claims that it is so unlikely as to make design the better explanation. That is the structure, and it is where the trouble lies. The first wrong objection: the multiverse The tempting reply runs: if there are enormously many universes, then even events of vanishingly small probability occur in some of them, and we should not be surprised to find ourselves in one where the improbable thing happened. This is the standard response to cosmological fine-tuning. Applied to Behe it fails, for two reasons. The first is that it solves a problem Behe does not pose. Fine-tuning concerns the values of physical constants, which are fixed once per universe; there is no process within a universe that could sample alternatives, so extra universes are the only available source of trials. Behe's claim concerns events inside biological history [4] , and here the trials are abundantly available in our own universe. The number of bacterial cells alive at any moment is estimated in the region of ten to the thirtieth power, and they have been dividing for some three billion years. Whatever else is wrong with an improbability calculation about flagella, it is not that our universe offers too few opportunities. Importing other universes to supply what this one already has is a category error. The second reason is more decisive. A multiverse explains an improbable feature only in combination with an observation-selection effect: we could not have found ourselves in a universe that did not permit observers, so the life-permitting values are unsurprising given that anyone is around to notice them. That filter selects for observers, and nothing else. Bacterial flagella are not required for observers to exist. So a multiverse plus observer selection gives no reason at all to expect flagella, and the reply, applied here, has no explanatory work to do. It changes the subject and sounds like an answer. Saying so matters, because a bad objection strengthens the position it attacks. If the best reply to Behe were an unobserved ensemble of universes, his argument would be in much better shape than it is. The objection that works The real answer is available in Behe's own formulation. He argues that an irreducibly complex system could not have been produced directly by successive modifications of a precursor that worked by the same mechanism. The qualification is his, and it is fatal, because nobody claims evolution proceeds only by direct routes. The alternative route is co-option: parts that arose and were maintained for one function are recruited to another, and the system becomes irreducible only afterwards, as the parts adapt to each other and their earlier independent roles are lost. The geneticist H. J. Muller described this pattern in 1939, arguing that evolution should be expected to produce systems whose components have become mutually indispensable — his term was interlocking complexity — precisely because later additions become entrenched. On this account, irreducibility is a predicted consequence of gradual evolution, not evidence against it. That reverses the argument completely. There is concrete evidence in the flagellum's own case. A subset of flagellar proteins is homologous to the type III secretion system, a syringe by which some bacteria inject proteins into host cells — a functioning system built largely from a proper subset of the parts. Kenneth Miller pressed this point in 'The Flagellum Unspun' (2004 [5] ). When Behe's claims were examined at length in the 2005 Kitzmiller v. Dover trial in Pennsylvania, the same structure of reply was applied to his other examples, including blood clotting and the vertebrate immune system. The Humean defect underneath Behind the biology lies the defect that this collection's original thesis was reaching for. A