Intelligent Design Icon Intelligent Design
News Media Icon News Media

Testing Darwinism and Design

In response to Michael Behe’s case for intelligent design in a recent edition of The New York Times, two letters to the paper’s editor charged design theory with being untestable. Design theory has failed to produce “statements that are susceptible to testing,” wrote Karen Rosenberg. Similarly, Donald Terndrup asserted , “Design will be a real science” when and only when “we have testable answers for these questions.”

But as philosopher of science Stephen Meyer has explained, the methodology used for intelligent design is strikingly similar to that used by Darwinists to argue for common descent. An argument against intelligent design “that appears frequently both in conversation and in print finds expression as follows: ‘Miracles are unscientific because they can not be studied empirically. Design invokes miraculous events; therefore design is unscientific. Moreover, since miraculous events can’t be studied empirically, they can’t be tested. Since scientific theories must be testable, design is, again, not scientific.’

Meyer continues:

Molecular biologist Fred Grinnell has argued, for example, that intelligent design can’t be a scientific concept because if something “can’t be measured, or counted, or photographed, it can’t be science.”

Gerald Skoog amplifies this concern: “The claim that life is the result of a design created by an intelligent cause can not be tested and is not within the realm of science.” This reasoning was recently invoked at San Francisco State University as a justification for removing Professor Dean Kenyon from his classroom. Kenyon is a biophysicist who has embraced intelligent design after years of work on chemical evolution. Some of his critics at SFSU argued that his theory fails to qualify as scientific because it refers to an unseen Designer that cannot be tested.

The essence of these arguments seems to be that the unobservable character of a designing agent renders it inaccessible to empirical investigation and thus precludes the possibility of testing any theory of design. Thus the criterion of demarcation employed here conjoins “observability and testability.” Both are asserted as necessary to scientific status, and the converse of one (unobservability) is asserted to preclude the possibility of the other (testability).

It turns out, however, that both parts of this formula fail. First, observability and testability are not both necessary to scientific status, because observability at least is not necessary to scientific status, as theoretical physics has abundantly demonstrated. Many entities and events cannot be directly observed or studied in practice or in principle. The postulation of such entities is no less the product of scientific inquiry for that. Many sciences are in fact directly charged with the job of inferring the unobservable from the observable. Forces, fields, atoms, quarks, past events, mental states, subsurface geological features, molecular biological structures all are unobservables inferred from observable phenomena.

Nevertheless, most are unambiguously the result of scientific inquiry.

Meyer adds:

Second, unobservability does not preclude testability: claims about unobservables are routinely tested in science indirectly against observable phenomena. That is, the existence of unobservable entities is established by testing the explanatory power that would result if a given hypothetical entity (i.e., an unobservable) were accepted as actual. This process usually involves some assessment of the established or theoretically plausible causal powers of a given unobservable entity. In any case, many scientific theories must be evaluated indirectly by comparing their explanatory power against competing hypotheses.

During the race to elucidate the structure of the genetic molecule, both a double helix and a triple helix were considered, since both could explain the photographic images produced via x-ray crystallography. While neither structure could be observed (even indirectly through a microscope), the double helix of Watson and Crick eventually won out because it could explain other observations that the triple helix could not. The inference to one unobservable structure‚ — the double helix‚ — was accepted because it was judged to possess a greater explanatory power than its competitors with respect to a variety of relevant observations. Such attempts to infer to the best explanation, where the explanation presupposes the reality of an unobservable entity, occur frequently in many fields already regarded as scientific, including physics, geology, geophysics, molecular biology, genetics, physical chemistry, cosmology, psychology and, of course, evolutionary biology.

The prevalence of unobservables in such fields raises difficulties for defenders of descent who would use observability criteria to disqualify design. Darwinists have long defended the apparently unfalsifiable nature of their theoretical claims by reminding critics that many of the creative processes to which they refer occur at rates too slow to observe. Further, the core historical commitment of evolutionary theory that present species are related by common ancestry has an epistemological character that is very similar to many present design theories. The transitional life forms that ostensibly occupy the nodes on Darwin’s branching tree of life are unobservable, just as the postulated past activity of a Designer is unobservable. Transitional life forms are theoretical postulations that make possible evolutionary accounts of present biological data. An unobservable designing agent is, similarly, postulated to explain features of life such as its information content and functional integration. Darwinian transitionals, neo-Darwinian mutational events, punctuationalism’s “rapid branching” events, the past action of a designing agent‚ — none of these are directly observable. With respect to direct observability, each of these theoretical entities is equivalent.

Each is roughly equivalent with respect to testability as well. Origins theories generally must make assertions about what happened in the past to cause present features of the universe (or the universe itself) to arise. They must reconstruct unobservable causal events from present clues or evidences. Positivistic methods of testing, therefore, that depend upon direct verification or repeated observation of cause-effect relationships have little relevance to origins theories, as Darwin himself understood. Though he complained repeatedly about the creationist failure to meet the vera causa criterion a nineteenth-century methodological principle that favored theories postulating observed causes, he chafed at the application of rigid positivistic standards to his own theory. As he complained to Joseph Hooker: “I am actually weary of telling people that I do not pretend to adduce direct evidence of one species changing into another, but that I believe that this view in the main is correct because so many phenomena can be thus grouped and explained” (emphasis added).

Indeed, Darwin insisted that direct modes of testing were wholly irrelevant to evaluating theories of origins. Nevertheless, he did believe that critical tests could be achieved via indirect means. As he stated elsewhere: “This hypothesis [common descent] must be tested . . . by trying to see whether it explains several large and independent classes of facts; such as the geological succession of organic beings, their distribution in past and present times, and their mutual affinities and homologies.” For Darwin the unobservability of past events and processes did not mean that origins theories are untestable. Instead, such theories may be evaluated and tested indirectly by the assessment of their explanatory power with respect to a variety of relevant data or “classes of facts.”

Nevertheless, if this is so, it is difficult to see why the unobservability of a Designer would necessarily preclude the testability of such a postulation. Though Darwin would not have agreed, the basis of his methodological defense of descent seems to imply the possibility of a testable theory of design, since the past action of an unobservable agent could have empirical consequences in the present just as an unobservable genealogical connection between organisms does. Indeed, Darwin himself tacitly acknowledged the testability of design by his own attempts to expose the empirical inadequacy of competing creationist theories. Though Darwin rejected many creationist explanations as unscientific in principle, he attempted to show that others were incapable of explaining certain facts of biology. Thus sometimes he treated creationism as a serious scientific competitor lacking explanatory power; at other times he dismissed it as unscientific by definition.

Recent evolutionary demarcationists have contradicted themselves in the same way. The quotation cited earlier from Gerald Skoog (“The claim that life is the result of a design created by an intelligent cause can not be tested and is not within the realm of science”) was followed in the same paragraph by the statement “Observations of the natural world also make these dicta [concerning the theory of intelligent design] suspect.” Yet clearly something cannot be both untestable in principle and subject to refutation by empirical observations.

The preceding considerations suggest that neither evolutionary descent with modification nor intelligent design is ultimately untestable. Instead, both theories seem testable indirectly, as Darwin explained of descent, by a comparison of their explanatory power with that of their competitors. As Philip Kitcher no friend of creationism has acknowledged, the presence of unobservable elements in theories, even ones involving an unobservable Designer, does not mean that such theories cannot be evaluated empirically. He writes, “Even postulating an unobserved Creator need be no more unscientific than postulating unobserved particles. What matters is the character of the proposals and the ways in which they are articulated and defended.”

Thus an unexpected equivalence emerges when design and descent are evaluated against their ability to meet specific demarcation criteria. The demand that the theoretical entities necessary to origins theories must be directly observable if they are to be considered testable and scientific would, if applied universally and disinterestedly, require the exclusion not only of design but also of descent. Those who insist on the joint criteria of observability and testability, conceived in a positivistic sense, promulgate a definition of correct science that evolutionary theory manifestly cannot meet. If, however, a less severe standard of testability is allowed, the original reason for excluding design evaporates.

Jonathan Witt

Executive Editor, Discovery Institute Press and Senior Fellow, Center for Science and Culture
Jonathan Witt, PhD, is Executive Editor of Discovery Institute Press and a senior fellow and senior project manager with Discovery Institute’s Center for Science and Culture. His latest book is Heretic: One Scientist’s Journey from Darwin to Design (DI Press, 2018) written with Finnish bioengineer Matti Leisola. Witt has also authored co-authored Intelligent Design Uncensored, A Meaningful World: How the Arts and Sciences Reveal the Genius of Nature, and The Hobbit Party: The Vision of Freedom That Tolkien Got, and the West Forgot. Witt is the lead writer and associate producer for Poverty, Inc., winner of the $100,000 Templeton Freedom Award and recipient of over 50 international film festival honors.