"Dances With Popper": An Examination of Dembski's Claims on Testability

Otto: Apes don’t read philosophy.

Wanda: Yes, they do, Otto, they just don’t understand it.

“A Fish Called Wanda”

In his new book, The Design Revolution, “intelligent design” advocate William A. Dembski invokes the late philosopher Sir Karl Popper as an authority on “testability” (ch. 39, pp.281-282). Perhaps Dembski has read Popper, perhaps he hasn’t. It’s certain, though, that Dembski does not understand Popper, and has a long history of not understanding Popper. Which is surprising, because Popper was an extraordinarily accessible philosopher.

Dembski bases his chapter on “Testability” in The Design Revolution (ch.39) on an essay he posted to the Internet in 2001. Between these two, Dembski switches from the term “falsifiability” to “refutability” instead. This is an odd thing for Dembski to do. It is explainable as a response to criticism that I made of his use of “falsifiability” in 2001, as I showed then that Dembski’s use of “falsifiability” differed markedly from that of Popper, who defined its usage in science and philosophy. The new version of Dembski’s argument shows a continuing misunderstanding of Popper and overlooks the fundamental flaws in Dembski’s argument.

Sir Karl Popper is justly famous as a philosopher of science. He proposed a demarcation criterion that, in his view, made the distinction between scientific theories and non-scientific conjectures. The basis of this criterion was what Popper called falsifiability. It should be noted that Popper’s proposal of a demarcation criterion has not been generally accepted in more recent treatments of philosophy of science. But the issue here is not over whether Popper’s falsifiability properly can be used as a demarcation between science and non-science. What is at issue here is whether William Dembski accurately conveys the concepts from Popper that Dembski cites.

Popper’s concern with testability focused on distinguishing between theories that are empirically testable and those that aren’t. This is the context into which Popper introduced the concept of “falsifiability”. “Falsifiability” refers to a deductive method of testing a theory: derive an entailed proposition from the theory that must be true if the theory is true, and attempt to determine the truth or falsity of the entailed proposition from empirical data. If the entailed proposition turns out to be false, one is justified in considering the theory that generated it false. Popper was explicit that “testability” and “refutability” meant the same thing as “falsifiability”, if they were to mean anything at all.

In order to be falsifiable, Popper asserted, a claim had to have the form of a universal statement. Only universal claims are susceptible to the application of modus tollens that underlies falsifiability. What about conjectures that come in the form of existential statements instead? Popper considered such “empirically irrefutable”.

Let’s examine what Popper said on these topics.

Some twenty five years ago I proposed to distinguish empirical or scientific theories from non-empirical or non-scientific ones precisely by defining the empirical theories as the refutable ones and the non-empirical theories as the irrefutable ones. My reasons for this proposal were as follows. Every serious test of a theory is an attempt to refute it. Testability is therefore the same as refutability, or falsifiability. And since we should call ‘empirical’ or ‘scientific’ only such theories as can be empirically tested, we may conclude that it is the possibility of an empirical refutation which distinguishes empirical or scientific theories.

If this ‘criterion of refutability’ is accepted, then we see at once that philosophical theories, or metaphysical theories, will be irrefutable by definition.

Popper, 1985, p.214.

And, of course, Popper held that strict or pure existential statements were empirically irrefutable.

With empirical irrefutability the situation is a little different. The simplest examples of empirically irrefutable statements are so-called strict or pure existential statements. Here is an example of a strict or pure existential statement: ‘There exists a pearl which is ten times larger than the next largest pearl.’ If in this statement we restrict the words ‘There exists’ to some finite region in space and time, then it may of course become a refutable statement. For example, the following statement is obviously empirically refutable: ‘At this moment and in this box here there exist at least two pearls one of which is ten times larger than the next largest pearl in this box.’ But then this statement is no longer a strict or pure existential statement: rather it is a restricted existential statement. A strict or pure existential statement applies to the whole universe, and it is irrefutable simply because there can be no method by which it could be refuted. For even if we were able to search our entire universe, the strict or pure existential statement would not be refuted by our failure to discover the required pearl, seeing that it might always be hiding in a place where we are not looking.

Popper, 1985, pp.212-213.

Let’s examine what Dembski says about “intelligent design” in light of Popper’s statement:

The fundamental claim of intelligent design is straightforward and easily intelligible: namely, there exist natural systems that cannot be adequately explained in terms of undirected natural causes and that exhibit features which in any other circumstances we would attribute to intelligence.

Dembski, 2004, p.45.

Dembski delivers a clear “strict or pure existential statement” here. One doesn’t have to accept Popper’s notion of falsifiability as a demarcation criterion to recognize that Popper’s argument for considering pure existential statements as being empirically irrefutable is still sound. But nowhere within Dembski’s chapter on “testability” does Dembski confront and attempt to rebut Popper’s argument. The chapter reads as if Dembski were completely unaware or ignorant of Popper’s statements in this regard.

It is useful to point out the provenance of Dembski’s chapter 39 on “testability” in The Design Revolution. It is derived mostly from an earlier essay posted to the Metanexus MetaViews email list and web site on January 24th, 2001 and entitled, “Is Intelligent Design Testable?” (IIDT hereafter for short. A copy is available at ARN.) Within this essay, one will note the absence of any reference to or use of the term “refutability”. What one does find is reference to “falsifiability”:

Dembski in IIDT wrote:

In relation to science testability is a very broad notion. It certainly includes Karl Popper’s notion of falsifiability, but it is hardly coextensive with it and can apply even if falsifiability does not obtain. Testability as well covers confirmation, predicability, and explanatory power. At the heart of testability is the idea that our scientific theories must make contact with and be sensitive to what’s happening in nature. What’s happening in nature must be able to affect our scientific theories not only in form and content but also in the degree of credence we attach to or withhold from them. For a theory to be immune to evidence from nature is a sure sign that we’re not dealing with a scientific theory.

What then are we to make of the testability of both intelligent design and Darwinism taken not in a generic abstract sense but concretely? What are the specific tests for intelligent design? What are the specific tests for Darwinism? And how do the two theories compare in terms of testability? To answer these questions, let’s run through several aspects of testability, beginning with falsifiability.

FALSIFIABILITY: Is intelligent design falsifiable? Is Darwinism falsifiable? Yes to the first question, no to the second. Intelligent design is eminently falsifiable. Specified complexity in general and irreducible complexity in biology are within the theory of intelligent design the key markers of intelligent agency. If it could be shown that biological systems like the bacterial flagellum that are wonderfully complex, elegant, and integrated could have been formed by a gradual Darwinian process (which by definition is non-telic), then intelligent design would be falsified on the general grounds that one doesn’t invoke intelligent causes when purely natural causes will do. In that case Occam’s razor finishes off intelligent design quite nicely.

Dembski, 2001.

One will note that Dembski’s deployment of “falsifiability” is unrecognizable as any sort of usage that could be said to be derived from Popper. Demsbki does not proceed from some “theory of intelligent design” and find a proposition that is an entailed consequence and test its empirical validity, as Popper required for his “falsifiability”. Dembski asserts that an essentially unrelated proposition, whether some phenomenon can be explained sufficiently well by reference to a completely unrelated theory, somehow has implications for the truth value of the conjecture of interest. This has no corresponding construct in Popper’s framework, perhaps for the simple reason that it is an obviously invalid approach that Popper wouldn’t have touched with a ten foot pole. (See below for more.) It was this clearly erroneous deployment of “falsifiability” that I strongly critiqued in my presentation on June 17th, 2001 at the CTNS/AAAS “Interpreting Evolution” conference at Haverford College with William Dembski and Michael Behe in attendance (see slides 23-25).

Now, by examination of Dembski’s chapter on “testability” in The Design Revolution, it appears that Dembski did get the message that his deployment of “falsifiability” was flawed. But rather than fix the underlying problem, Dembski chose simply to introduce another term with which to replace “falsifiability” that he could redefine to suit his already existing text. Unfortunately, Dembski again ties the new term of choice, “refutability”, to Sir Karl Popper. Here is Dembski’s justification for invoking Popper on “refutability”:

The main point of Popper’s criterion of falsifiability is not so much that scientific claims must have the possibility of being demonstrably false as that they must have the possibility of being eliminated as the result of new evidence. To underscore this point Popper even wrote a book titled Conjectures and Refutations. That is the point of refutability.

Dembski, 2004, p.281.

The aphorism about judging a book by its cover leaps to mind. Examination of the book in question, though, leads to an understanding that Popper treated testability and refutability as synonyms for falsifiability (see pp. 37, 39, 197, 219, 256, and 258. See p. 279 for discussion of Carnap, who makes a similar error to that of Dembski.). In other words, the point of refutability is, according to Popper, quite unlike what Dembski has represented in his book.

So much for invoking the authority of Popper as a prop for Dembski’s version of “refutability”. But does Dembski’s formulation have any merits of its own? Let’s have a look.

Refutability comes in degrees. Theories become more refutable to the degree that new evidence could render them unacceptable. Note that refutability asks to what degree theories could be refuted, not to what degree they actually have been refuted. Thus refutability gauges how sensitive theories are to refutation in principle rather than on the basis of any particular evidence. The more sensitive to evidence generally, the more refutable the theory. According to Popper, one mark of a good scientific theory is that it is highly refutable in principle while consistently unrefuted by the evidence in practice. Better yet are those theories on which scientists have expended tremendous diligence to refute them, only to have their efforts come to nothing. Within Popper’s scheme of scientific rationality, theories are corroborated to the degree that they resist refutation.

Let’s now ask, Is intelligent design refutable? Is Darwinism refutable? Yes to the first question, no to the second. Intelligent design could in principle be readily refuted. Specified complexity in general and irreducible complexity in biology are, within the theory of intelligent design, key markers of intelligent agency. If it could be shown that biological systems that are wonderfully complex, elegant and integrated – such as the bacterial flagellum – could have been formed by a gradual Darwinian process (and thus that their specified complexity is an illusion), then intelligent design would be refuted on the general grounds that one does not invoke intelligent causes when undirected natural causes will do. In that case Occam’s razor would finish off intelligent design quite nicely.

Dembski, 2004, pp.281-282.

Here Dembski’s “refutability” runs head-on into Popper’s argument concerning the empirical irrefutability of strict or pure existential statements, such as the fundamental claim of intelligent design quoted above. The result is fatal for Dembski’s “refutability” and the claims he makes for it. No matter how many systems ID advocates assert might have specified complexity or irreducible complexity and later have them overturned by empirical inquiry finding that directed natural causes, such as natural selection, are perfectly capable of explaining them, the ID advocates can always propose yet another system as a candidate. (Dembski’s phrasing of “undirected natural causes” excludes natural selection, since natural selection is constrained and thus guided by local environmental conditions and factors like co-evolution. If Dembski wishes to redefine “guided” as “guided by an intelligent agent”, he needs to do so explicitly.) The cycle is endless, as Popper quite clearly saw with his example of search for the ten-times-larger-pearl. We already see the beginning of this, as ID advocates used to be very keen on using the human blood clotting cascade as a model system showing “intelligent design”. Good responses to the ID claims on blood clotting have made this system less tenable as an illustration, but ID advocates do not thereby say that the “fundamental claim of intelligent design” is thereby to that degree refuted. To the contrary, they simply have picked up and moved on to another system to serve as a poster-type example, in this case the flagellum of E. coli bacteria. We can already observe that the “sensitivity” of “intelligent design” conjectures to empirical evidence appears to be “none whatsoever”. On Dembski’s own criteria, as well as Popper’s, “intelligent design” is irrefutable.

As criticism of ID arguments about the E. coli flagellum accumulate, one can see that ID responses are tending to insulate against empirical refutation. One class of ID responses claims that the information needed to make flagella was “front-loaded” into some ancestral strain of bacteria. Another is that the “intelligent designer” acted at the quantum level to produce the flagellum. And a third class of response requires video-camera certainty concerning every step of proposed natural pathways to development of bacterial flagella. (It is useful to note here that “intelligent design” advocates select examples where knowledge concerning their historical origins is sketchy to non-existent. If “intelligent design” were more than a “bare possibility”, the ID advocates should be able to use as examples biological systems whose historical origins are well-known, but which remain unexplained by various evolutionary hypotheses or mechanisms. Instead, whenever there is sufficient evidence of the origin of a biological system, it uniformly is explained by some evolutionary hypothesis or mechanism. By the sort of inductive process invoked by Dembski elsewhere (e.g., Dembski, 2004, pp.95-96), “intelligent design” advocates should concede that this will continue to be the case for all future examples.)

On a further note, Dembski’s claim that “Darwinism” is irrefutable is clearly a mistake. If one credits Dembski’s formulation of “refutability”, it is clear that he has misapplied it in his haste to say something negative about “Darwinism”. Dembski’s entire program of finding “specified complexity” in biological systems is dependent upon his “generic chance elimination argument” (GCEA) being able to consider – and eliminate – evolutionary hypotheses for the origin of some event. If Darwinian hypotheses were actually irrefutable, as Dembski claims, then his GCEA would get nowhere in considering biological systems. Dembski cannot “have his cake and eat it, too” in this instance, since “Polite society frowns on such obvious bad taste.” (See also my essay on Huxley and the “typing monkeys” metaphor.)

As I have noted before elsewhere, falsifying tests for natural selection date back to Darwin.

Natural selection cannot possibly produce any modification in any one species exclusively for the good of another species; though throughout nature one species incessantly takes advantage of, and profits by, the structure of another. But natural selection can and does often produce structures for the direct injury of other species, as we see in the fang of the adder, and in the ovipositor of the ichneumon, by which its eggs are deposited in the living bodies of other insects. If it could be proved that any part of the structure of any one species had been formed for the exclusive good of another species, it would annihilate my theory, for such could not have been produced through natural selection.

Darwin, 1859, ch. 6

Famously, Popper himself had a go at an opinion on the status of “Darwinism”, which he originally critiqued as being “almost tautological” and thus relegated its status to that of a useful “metaphysical research program”. Popper recanted his earlier stance in an article published in Dialectica in 1978, saying that natural selection could be formulated in a way that was far from tautological and also testable. Dembski, predictably, also fails to learn this lesson from reading Popper. Perhaps understanding of Popper will one day come to Dembski. Until then, we’ll know to check the original sources when Dembski makes a claim about Popper.

  • Darwin, Charles R. 1859. On The Origin of Species by Means of Natural Selection, or The Preservation of Favoured Races in the Struggle for Life, First Edition. (http://www.talkorigins.org/faqs/origin.html, last accessed 2004/05/03.)
  • Dembski, William A. 2001. “Is Intelligent Design Testable?” MetaViews. (http://www.arn.org/docs/dembski/wd_isidtestable.htm, last accessed 2004/05/03.)
  • Dembski, William A. 2004. The Design Revolution. Downer’s Grove, IL: InterVarsity Press.
  • Popper, Sir Karl. 1978. “Natural Selection and the Emergence of Mind,” Dialectica 32:339-355.
  • Popper, Sir Karl. 1985. “Metaphysics and criticizability.” In: Popper Selections, David Miller (ed.). Princeton, NJ: Princeton University Press. Originally published in 1958.
  • Popper, Sir Karl. 1992. “Conjectures and Refutations.” Routledge; 5th edition.