In his recent work Darwin and Intelligent Design, evolutionary geneticist and philosopher Francisco J. Ayala argues that (i) intelligent design theory (ID) is “bad” science and “bad” theology, and (ii) religion and science are epistemologically independent of one another. The present work will focus on showing that if ID is scientifically “bad,” then it is not theologically “bad”; and vice versa.
Perhaps the most modest belief in design is deistic, the view that a supernatural entity created the cosmos and its laws of nature (without necessarily sustaining them), including the laws of evolutionary biology, which eventually gave rise to complex biological entities (e.g., the vertebrate eye). But this kind of cosmic design is not usually what ID proponents have in mind. Their concern is typically organismic design, the idea that a supernatural entity directly created at least some complex biological mechanism.
Ayala’s principal theological complaint is that ID “implies attributes of God that are incompatible with Christianity and other monotheistic religions.” These include qualities of perfection, such as omniscience, omnipotence, and moral perfection. However, Ayala points out: “The design of organisms is not intelligent, but imperfect and riddled with dysfunctionalities.” If the design of organisms implicates a dim or morally reprehensible designer, then so much the worse for the ID program; poor or cruel “designs” are not the work of God.
Schematically, Ayala’s argument can be given as follows:
(P1) If an omnipotent, omniscient, and morally perfect God created organisms, then we would expect to find no dysfunctions among organisms.
(P2) But we do find dysfunctions among organisms.
(C) Thus, an omnipotent, omniscient, and morally perfect God probably did not create organisms.
Ayala clearly doesn’t offer this argument as a defense of atheism. Rather, he intends to show that if ID is true, then the designer cannot be God. (This is a stronger claim than the common objection that even if ID were true, it would not entail that the designer was God.) In short, the ID proponents’ preoccupation with undermining evolutionary mechanisms yields no fruit for theism, as it is duplicitous to suggest that evolutionary biology can account for the blemishes of biological life, but where perfection rests, we can reinstate God as its creator.
ID fares no better from a scientific point of view. The idea of testability has been central to the philosophy of science, and even though some philosophers have wished to do away with the notion, it has been well argued that this is a mistake. A full treatment of testability is beyond the scope of the present discussion, but the general idea is that the evaluation of a given hypothesis and the evidence relevant to it requires that the hypothesis have suitable auxiliary assumptions. For instance, it is generally accepted that Einstein’s general theory of relativity can be tested by recording the degree that light bends onto the silhouette of the moon during a solar eclipse. This testing situation presumes auxiliary assumptions about the absolute position of the light source once the eclipse takes place, and the reliability of the instruments used to gather the evidence. Such auxiliary assumptions are suitable when they have been independently established as veridical, not invented; and only then can they be used to collect evidence to discriminate between rival hypotheses in a testing situation.
Ayala concludes: “The evidence and arguments of ID proponents are bad science; they have no scientific cogency whatsoever.” He justifies this assessment by pointing out the absence of suitable auxiliary assumptions in testing ID, such as independently established propositions regarding the putative designer’s goals and abilities. ID proponent Michael Behe even explicitly states: “The reasons that a designer would or would not do anything are virtually impossible to know unless the designer tells you specifically what those reasons are.” Ayala’s retort is that this concession “destroys intelligent design as a scientific hypothesis,” for “it provides it [ID] with an empirically impenetrable shield against predictions of how ‘intelligent’ or ‘perfect’ a design will be.” In other words, without independent attestation of the putative designer’s goals and abilities, ID fails to predict anything about the features of the biological world. And this is certainly a sign of a pseudoscientific proposition and article of faith.
Knowledge of the goals and abilities of a putative designer is necessary to predict the observable effects of its unobserved behaviors; and those predictions can then be tested against biological observations. Consider the following illustration of testability. Suppose that you are presented with a clock while walking on the shore of a beach. You examine it, and notice that it is composed of many different parts, all of which work to produce the effect of measuring time. If one of the parts were taken from the clock, it would hardly be able to fulfill its time-keeping task. So you infer from these features that it is more likely that the clock was made by an intelligent designer (a human) than that it is the product of mindless chance. Your inference, though, is grounded in your knowledge of the goals of at least some human beings (to tell time more precisely) and their abilities (to construct measuring devices). If you were not aware of these things, the inference from the features of the clock to its creation by human intelligence would wither away. This might sound odd only because you are aware, under ordinary circumstances, of a sufficient number of goals and abilities of many human beings. The inference that the clock was produced by some intelligent designer is testable only if auxiliary assumptions about the relevant goals and abilities of the putative designer are included.
ID is scientifically vacuous, then, because the putative designer’s goals are simply “impossible to know unless the designer tells you specifically what those reasons are,” as Behe puts it. And there is, of course, no independently established testimony from any designer about his reasons for creating organisms.
One of Ayala’s most important theses is that “science and religion cannot be incompatible, because they concern nonoverlapping domains of knowledge.” Indeed, his primary message is that religious persons need not appeal to any scientific arguments at all to defend their views, as “faith” alone is sufficient justification for Christian belief.
I take Ayala’s separation of religion and science to be an epistemological thesis about the different ways in which one obtains scientific knowledge on the one hand, and religious “knowledge” on the other. Ayala maintains that it is rational to keep their methods completely disjoint. Both religious and scientific statements are testable, but each has its own distinct methodology, and justifications offered for either kind of statement are incomparable.
Ayala uses existing arguments to shore up this perspective, concluding that ID is bad theology because empirical results are contrary to God’s attributes, and that ID is bad science because it is not scientifically testable. The solution, therefore, is to assign them different epistemological environments, where they can thrive on their own accord without the bewilderment caused by conflating their environments.
There is a contradiction in Ayala’s views here, one that can be made clear by looking more deeply at the concept of testability. If an observation is more likely to occur under one hypothesis compared to another, then each hypothesis is testable. More formally:
(T): A hypothesis H1 is testable against H2 if and only if there is an observed datum O and suitable auxiliary assumptions A such that Pr(O given H1 & A) is unequal to Pr(O given H2 & A).
Now consider Ayala’s reason for believing that ID is bad theology—that organisms’ dysfunctions count against ID because ID supposes that the putative designer is omnipotent, omniscient, and morally perfect. But this argument only works if both evolutionary biology (EB) and ID are testable. That is, if my characterization of Ayala’s position on the separation of science and religion is accurate, then P2 of Ayala’s earlier argument can be legitimately framed as:
Pr(dysfunctions given EB & A) > Pr(dysfunctions given ID & A).
In other words, the probability that organisms have dysfunctions given evolutionary biology and suitable auxiliary assumptions is greater than the probability that organisms have dysfunctions given ID and suitable auxiliary assumptions. Although Ayala’s overall argument says much more than this, his conclusion that ID is bad theology requires it. Thus if ID is bad theology, then ID is testable against EB. But this a troubling result since, according to Ayala, (i) ID is bad science because it is not testable and (ii) theological assertions (e.g., that God exists) are epistemologically independent from the assertions of evolutionary biology, as religious hypotheses cannot be tested against scientific ones.
Ayala has only one way out of this conundrum that does not create any inconsistencies, and both retains as many of his arguments as possible, and sustains the importance of those arguments to his primary aim. And that is to settle upon the views, first, that science and religion are epistemologically independent (fideism), second, that ID is bad science, and third, that ID is not bad theology because it creates a conflict between God’s predicates and the observed data. There is good reason to suppose that ID is not a genuine scientific hypothesis; and fideism, though perhaps unfashionable, is not logically inconsistent or obviously false.
But if Ayala insists upon retaining the idea that ID is bad theology for the reasons given in Darwin and Intelligent Design, then to be consistent he must abandon his other two theses. I suspect that Ayala would conservatively opt for the former alternative and abandon altogether his reasons for thinking ID is simultaneously bad theology and bad science.
In any case, Ayala cannot uphold his three theses consistently, and that is a daunting state of affairs. However, a solution is available that retains much of Ayala’s other important work, and only at the cost of jettisoning a single, relatively insignificant conclusion for his overall program—a program to show why evolutionary biology is good science and ID is a farce.
 Elliott Sober, “The Design Argument.” In The Blackwell Guide to Philosophy of Religion ed. William E. Mann (Oxford, UK: Blackwell Publishing, 2005).
 A more modest form of ID can be formulated which maintains that some entity (natural or supernatural) created some organic mechanism (such as the vertebrate eye). However, if irreducible complexity of such mechanisms demands a designer, and the universe is finitely old, then there must have been a supernatural entity that brought about the first irreducibly complex natural being. Otherwise no designer would be necessary to account for irreducible complexity. See Elliot Sober’s “Intelligent Design and the Supernatural—The ‘God or Extraterrestrials’ Reply” in Faith and Philosophy Vol. 24, No. 1 (January 2007): 72-82.
 Elliott Sober, “Testability.” Proceedings and Addresses of the American Philosophical Association 73 (1999): 47-76.
Copyright ©2009 Jeffrey T. Allen. The electronic version is copyright ©2009 by Internet Infidels, Inc. with the written permission of Jeffrey T. Allen. All rights reserved.