Kenneth Miller has made his name in a number of areas. A professor of biology at Brown University, Miller is the Author of popular high school and college textbooks in his field. He is a professed Christian and an avid Darwinist, and many people know of his earlier book, Finding Darwin‘s God, which attempts to show the compatibility of Christian faith with Darwinian evolution. More than that, in a series of school board hearings, Miller has served as the leading expert witness for the side that believes there is no scientific controversy over evolution and hence no need for biology curricula to include problems for the Darwinian account.
In Only a Theory, Miller gives his take on the 2005 hearings in Kansas (at which I was an expert witness on the opposing side) and Pennsylvania, and explains why he rejects the criticisms of intelligent design theory (ID). Not confining himself to rebutting the scientific and philosophical claims of ID, he further argues that ID is a danger to America’s “scientific soul,” claiming that ID threatens to politicizescience and undermine its objectivity.
An Unsympathetic Reader
Unlike many scientists, Miller is a gifted public communicator. He writes clearly and well, and with all of the rhetorical flair of a trial lawyer. But he fails at the level of scholarship. An effective scholarly critic of a sophisticated theory must do careful and sympathetic exegesis of the strongest versions of the theory, and engage those. While Miller does cite both of Michael Behe’s books, he only cites the popular and dated work Intelligent Design (1999) and a magazine article by William Dembski. In an attempt to disarm his critics, Miller proposes to test ID by “embracing it,” claiming that “the most sincere compliment anyone can pay to a scientific idea is to take it seriously” (p. 44). But Miller does not take ID seriously, repeatedly saddling it with a picture of the designer that is easy to ridicule, but to which ID is not committed. Looking at the fossil data for the horse family and the patterns of appearance and extinction, Miller concludes, “Our designer doesn’t just design; he does it again and again and his designs don’t last… In other words, intelligent design is actually a hypothesis of progressive creationism” (52). But while a proponent of ID might believe that the designer works in this way, he need not.
Miller mistakenly supposes that inferring design from a phenomenon implies that the proximal cause of that phenomenon is the miraculous, creative work of the designer. But this involves two mistakes. First, a designer can work through unintelligent means. For example, if I see a computer printout of the architect’s plans for a new cathedral, I immediately infer that it was designed, but not that the computer (or its program) was the designer: the real designer, the architect, lies further back in the causal chain. Likewise, inferring design from the complex specified information in DNA implies nothing, by itself, about when that information was added to the system. It might have been specially added at a certain point or it might have been “front-loaded,” built into nature from the beginning. It is particularly odd that Miller ignores the latter possibility, because he spends chapters 5 and 6 arguing for the fine-tuning of the laws of nature, and for the nonrandom constraints on evolution imposed by the master
control genes known as Hox genes-”a tool kit for generating body form” (129)-and by convergent evolution into similar “adaptive spaces,” claiming that “the universe, in a certain sense, had us in mind from the very beginning” (161). One could argue this makes Miller a proponent of ID; indeed he writes: “There really is a design to life, but it is not the clumsy, interventionist one in which life is an artificial injection into nature” (134). Miller seems to oppose ID mainly because he wrongly thinks it must subscribe to a “clumsy, interventionist” picture of the designer that he rejects.
A second and closely related point is that inferring design is not the same as inferring a miracle. Miller makes this mistake because he misunderstands ID’s opposition to methodological naturalism (MN), the centerpiece of the minority report’s 2005 argument in Kansas. Miller thinks that ID’s agenda is “theistic science,” and supposes that by rejecting the idea that science must infer natural causes (MN), ID is calling for a science of the supernatural. Thus he confesses, “Try as I might, I couldn’t think of a single nonnaturalistic explanation of anything that didn’t involve a supernatural agent” (188). This is amazing given the fact that proponents of ID have repeatedly made it clear that inferring design says nothing about the ultimate metaphysical character of the designer. Citing William Dembski, I said in my written testimony at Kansas in 2005,
“[TJhe contrast between natural and supernatural causes is the wrong contrast. The proper contrast is between undirected natural causes on the one hand and intelligent causes on the other. “1… [WJe can investigate whether nature manifests signs of intelligence without settling the question of whether the designer is supernatural, although there may be independent evidence for or against this further conclusion.2
This distinction explains why there are agnostic supporters of ID, like David Berlinski and Michael Denton, and why James Barham testified in Kansas that although he did not believe in the supernatural, he did believe that irreducible teleology (design) was part of nature. ID as science might find additional evidence (such as the fine-tuning of the cosmos) that is better explained by a supernatural agent than by an impersonal immanent teleology, but a design inference is not by itself an argument to the supernatural. Fallaciously claiming that it is, of course, is very convenient if one seeks to argue that discussing any evidence for design in nature in the public schools is a violation of the First Amendment. The truth is in fact just the opposite: MN means that students will hear only the evidence for undirected causes in nature, and that unconstitutionally favors some religious beliefs, including secular humanism and theistic beliefs that deny the natural knowledge of God.
Similar problems arise for Miller’s discussion of the empirical evidence for ID. Using the analogy of a standard fivepart mousetrap, Behe argued that many biological systems, including the bacterial flagellum, the blood-clotting cascade, and the immune system exhibit “irreducible complexity” (IC) because they have several well-matched parts, all of which are necessary for the system to function. Miller counters that a subset of the system’s parts might have some other function, which was co-opted,
noting that three of the mousetrap’s parts make a good catapult, and arguing that a protein pump called the Type III secretory system (TTSS) contains ten of the over thirty proteins in the flagellum. Miller does not claim that the flagellum evolved from the TTSS (a good thing, since several authorities think the TTSS devolved from the flagellum), but does think that the existence of the TTSS refutes IC, because it shows that parts of the flagellum can function by themselves (60).
This, however, is a serious mistake. Behe defines IC as a property of the whole system, not of its parts. He never denied one could make something else functional out of a subsystem’s parts, but this is irrelevant. As Casey Luskin argues, the existence of the unicycle does not show that the bicycle does not need both wheels to function. Secondly, Behe’s challenge is that the flagellum cannot be built gradually by Darwinian means. Noting the existence of the TTSS as a possible island on the way does not show that it could: as Dembski argues, this is like claiming we can walk from Los Angeles to Tokyo because we have found the Hawaiian Islands. More fundamentally, it does not show how all of the parts of the flagellum were made available at the same time and place, how the relevant interfaces were all compatible, or how they were properly coordinated by their assembly instructions.3 Indeed, Miller admits, “the existence of the TTSS today doesn’t answer the question of how the flagellum actually evolved” (61).
Miller also argues that the blood-clotting cascade is not irreducibly complex, because, “The genome of the fugu, or puffer fish, lacks three of the factors—and its blood clots just fine” (63). However, Behe’s actual argument concerned a core system, which is the same in puffer fish and other creatures.4 And Miller inaccurately recounts that during the Dover hearings, when presented with fifty-eight peer reviewed articles on the immune system, Behe merely said that this was not “good enough” (73). Actually, the words belonged to Eric Rothschild, who was questioning Behe, but Judge Jones attributed them to Behe.5 Behe actually said, “It’s not that they aren’t good enough. It’s simply that they are addressed to a different subject” since none of them contained data supporting a “step-by-step, mutation by mutation analysis.”6
Miller realizes that the fundamental question is whether Darwinian processes can generate the information necessary to specify new biological structures. He considers Dembski’s argument for a Law of Conservation of Information (LCI), according to which unintelligent processes can shuffle around complex specified information, but cannot produce it. As a counterexample to LCI, Miller cites Thomas Schneider’s program “ev,” which simulates the parallel Darwinian evolution of populations of digital organisms over many generations. Miller claims that ev “unequivocally shows that this process leads to an objective and quantifiable gain in information” (76). But, on inspection, ev gets this result only because the algorithm is not Darwinian but teleological: it relies on a measure of the distance between a digital organism and its future target called the Hamming distance, analogous to an Easter egg hunt in which children are assisted in their search by hints of “warmer” and “colder.”7 When Miller turns to real-world examples the results are the usual unimpressive examples of microevolution, for example antifreeze in fish. As Behe points out, “The job of the antifreeze protein is a very simple one, and it is relatively easy to improve the protein incrementally … The antifreeze protein is not so much a molecular machine as it is a blood additive.”8
Miller does have a strong conceptual objection to design, which ID theory needs to address. He rightly notes that a completely unspecified designer is scientifically useless, “since arbitrary design … could be consistent with anything” (65). What this shows, however, is that the bare idea of design must be supplemented with more specific models that attempt to capture testable design principles. Miller provides no good reason to think that scientists should not try to do just that.
Angus Menuge is professor of philosophy at Concordia University, Wisconsin. He is the author of Agents under Fire and many articles on philosophy of mind, intelligent design, and apologetics.