Other writings     Drawings gallery

Intelligent Design’s Abject Failure

I shall argue that Behe’s Irreducible Complexity fails to invalidate a proper understanding of Darwinian evolution by natural selection by considering three ways in which evolution might adequately explain seemingly irreducible complexity. I shall then argue that even granting Behe the falsity of evolution is insufficient to establish an Intelligent Designer. Lastly, I shall couple Behe with Dembski’s argument for reliable empirical indication of intelligent causation, and show this strongest version of Intelligent Design to be a fallacious argument from ignorance at worst or most charitably understood as an ultimately unwarranted inference to best explanation.


It should be noted that all ScienceIntelligent Design (ID) arguments are themselves components of an overarching argument for scientific legitimacy. That is, the aim of the ID movement is not only to provide arguments for the explanatory necessity of an Intelligent Designer but also to legitimise said arguments scientifically. As much can be said against the inclusion of ID, and as interesting demarcation in philosophy of science may be, I shall here solely concern myself with the philosophical validity and soundness of the former kind. Whether the arguments can be said to be scientific will take a backseat to whether they are right.

William A. DembskiAs per Dembski (1998) the two-pronged approach of the ID movement is the critique of Darwinism coupled with the provision of a positive alternative – i.e. Intelligent Design. If Dembski by ‘Darwinism’ does not mean ‘Darwinian evolution by natural selection’ (henceforth ‘evolution’) his allusions are utterly abstruse to me. However, this would imply that Dembski’s scientific literacy is wanting as attested by his claims that ‘[a]ccording to Darwinism, undirected natural causes are solely responsible for the origin and development of life’ and that ‘Darwinism [is] hopelessly entangled with naturalism.’ Neither, of course, is accurate. Evolution is concerned neither with the origin of life nor with ontology. Evolution merely states that those organisms or genes least likely to die in a given environment will therefore be the ones most likely to reproduce and eventually become quantitatively dominant. Evolution is silent on the matter of life’s origins and the existence of the supernatural alike.

Natural Selection GraphOf course, natural selection is a natural cause but it is hardly ‘undirected.’ The natural selection of random mutations in a population is very much directed by environmental pressures. It is likely that Dembski meant to say ‘undirected by any intelligence’ but a simple thought-experiment wherein we imagine life originating from a supernatural agent who occasionally enjoys tinkering with environmental pressures, thereby obliquely directing natural selection, should suffice as a counterexample to Darwinism ruling out ‘the possibility of a God or any guiding intelligence playing a role in life’s origin and development.’ Indeed, human beings ourselves constitute guiding intelligences playing a role in life’s development – granted mostly through artificial selection but also occasionally through altering environmental pressures and thereby affecting the course of natural selection.

Having a firm conceptual grasp on evolution we may now turn to the first prong of the ID approach, namely the critique of evolution as championed by Behe, and see how it fares. Behe argues for an Intelligent Designer from a supposed deficiency in evolution to explain how certain biochemical systems came about. Behe attributes such systems the property of Irreducible Complexity; defined as follows:

Michael BeheA single system which is composed of several interacting parts that contribute to the basic function, and where the removal of any one of the parts causes the system to effectively cease functioning.

(Behe, 2006, p. 257)

The idea is that since evolution requires every mutation selected for to be of gradually increasing advantageousness to survival, and if it can be shown that a system loses its advantageous function by the removal of any of its component parts, then that system is irreducibly complex and cannot have evolved gradually. The astute reader might rightly wonder how a disproof of evolution in any way is proof of a designer or proof against all possible natural explanations whatsoever. However, my charitability is here curtailed by my incomprehension of Behe’s position. He seemingly thinks that irreducible complexity somehow constitutes ‘a purposeful arrangement of parts that bespeaks design’ (2005, p. 87). I can, at the most charitable, only see it as an arrangement of parts unexplainable by evolution but possibly by some other explanation. It could be either natural or supernatural and surely purposiveness has yet to be demonstrated either way.

Survival of the Fit EnoughI shall drop this for now. It is better dealt with in our later considerations of inferences to best explanation and in any case, as we shall see, Behe’s position can be strengthened by a symbiosis with Dembski’s. For now let us consider proposals as to how irreducible complexity might evolve after all. We said before that every selected mutation must be advantageous. This was a lie by omission. For ease of understanding evolution is often described in positive terms as selection for fitness. However, more accurate is the negative description of selection against unfitness, prompting the revision ‘survival of the fit enough.’ (Scott, p. 37) It is not ‘endorsement’ of advantage; only ‘censure’ of disadvantage and therefore mutations need only be fitness-neutral. Indeed, there is ample evidence that human beings have suffered neutral genetic drift resulting in losing the ability to produce vitamin C (as opposed to most other mammals) due to evolving in an environment with a ready dietary availability (Max, 2003).

This leads us to scaffolding. According to Michael Ruse the salient point ‘is not whether the parts now in place could not be removed without collapse, but whether they could have been put in place by natural selection.’ (2008) The oft-repeated analogy is that an arched stone-bridge is ‘irreducibly complex’ by Behe’s definition since the removal of any stone would collapse it. Yet we know it was not built instantaneously; rather onto supporting scaffolding later removed. Similarly a system might be irreducibly complex now, only due to past shedding of its reducible ‘scaffolding’ in neutral genetic drift.Arch Bridge

Above: The removal of any arch-stone would result in collapse.

Below: After the bridge is built the scaffolding (brown) is burnt away.

Building an Arch Bridge Another closely related evolutionary path to irreducible complexity is improbable but possible happenstance. Suppose an adaptive ability requires two (or more) mutations in order to function and either mutation alone is useless. As long as mutation A confers no disadvantage it could still be acquired through neutral drift without mutation B, priming the organism for advantage if B should turn up later. One might be inclined to scoff at the improbability. However, this possible scenario has already been proven actual in an ongoing experiment by bacteriologist Richard Lenski, in which e-coli bacteria evolved the metabolism to feed on citrate. The improbability of such an event fades into mere rarity by the law of large numbers but is still rare enough to demonstrate the requirement of more than one mutation (Dawkins, pp. 116-133).

HappenstanceThe final way in which evolution can give rise to irreducible complexity is cooption or exaptation. This is the commonest answer given to Behe’s favourite example; the bacterial flagellum. In the Dover Trial (Behe, 2005, pp. 80 – ) Eric Rothschild, attorney, continually mentions the type III secretory system as a possible precursor to the flagellum. We need not go into too much detail but the general gist is that even though the removal of any flagellum-part will cease locomotive function, partial flagella can still be put to some other functional use – e.g. secretion. Behe concedes as much yet nonetheless insists that even though the type III secretory system is indeed a subset of a flagellum – i.e. identical to a flagellum with parts missing – the flagellum is still irreducibly complex because said subset does not function as a flagellum.

TTSS Vs. Flagellum

Left: Type III secretory system. Right: Bacterial flagellum.

I make neither head nor tails of this insistence since evolution, to my understanding, implies no demand for specific functionality but as previously stated only selects against detriment. To make a silly but evolutionarily accurate allusion to Nietzsche: what does not kill you will make your descendants stronger if they can find a use for it – any use.

However, an argument for an intelligent designer – as opposed to against evolution – can hardly stand and fall at the failed or successive critique of evolution alone. From the supposed explanatory inadequacy of evolution regarding an unexplained phenomenon nothing else follows about what does explain said phenomenon. Conversely if an appeal to intelligence adequately explanatory of the phenomenon were produced nothing else follows about evolution.


QualiaSoup's explanation of what's wrong with arguments from ignorance.

Simply, evolution and intelligent design are neither exhaustive nor mutually exclusive. If we had two adequately explanatory hypotheses for the same phenomenon we could strengthen one at the expense of criticising the other but Behe has made no headway in producing neither arguments for – nor explanatory adequacy of – an intelligent designer. The case for one is therefore not strengthened by his critique. Let us then grant Behe the falsity of evolution for the sake of argument and seek assistance from Dembski’s proposed demarcation between design and non-design.

Explanatory Filter ChartDembski’s argument is that the property he calls ‘complex specified information’ is a reliable empirical indicator for intelligent causation. To be honest no amount of reading has yet yielded me a clear understanding of Dembski’s concepts and I am not alone in my confusion. However, since mathematics and information theory are far out of my depth I defer deeper criticism to the equally puzzled Elsberry et. al. (2009). As far as I can gather we may strike ‘information’ from the property since, whatever it is, it follows from ‘specified complexity’ alone. According to Dembski Complexity ‘is a form of probability’ and specification is defined as ‘a match between an event and an independently given pattern.’ (Ruse, 2008) To the best of my understanding, it is a needlessly convoluted way of saying that non-arbitrary patterns are more indicative of design the greater their improbability of happening by chance.

At the risk of being uncharitable – and I make no pretension to correct exegesis – Dembski seems deliberately obscure for the sake of a cleverly disguised strawman. Evolution is not committal to but the rejection of improbable random chance in favour of probable non-random natural selection. Nobody is committed to ‘lucky chance’ and, as previously mentioned, the improbability of a specified outcome is indicative only of an explanation – not a specific intelligent one – even if we grant being ignorant of the explanation for the sake of argument. It should be clear now that design arguments are at worst fallacious arguments from ignorance and at best attempted inferences to best explanation.

Either they fallaciously argue to an intelligent cause of a phenomenon by appeal to our ignorance of its explanation, or – more charitably – they infer, from the premise that the design hypothesis would provide a “better” explanation for the evidence than would any other hypothesis, to the conclusion of its truth (adapted from Harman, 1965). However, such charity cannot stave off the resulting problems. First, the argument that theism cannot provide causal explanations (Le Poidevin, 1996, pp. 35-38) is easily adapted to ID since it scarcely meets the criteria of informativeness nor of generalised connection between cause and effect. ‘The flagellum was produced by a flagellum-producing intelligence’ explains nothing at all but merely asserts intelligence while said assertion was warranted only by its supposed explanatory adequacy. Moreover, there is no generalised connection between intelligence and flagellum.

Platonic AssemblySecond, enumerative inductive inferences from a particular instance to a universal affirmative proposition are reliable if and only if the possible counterexamples to the proposition are exhausted. ID arguments construed as attempted inferences to best explanation argue from the single instance of a failed natural explanation to the failure of all other possible natural and supernatural explanations bar one. The ID proponent only escapes the argument from ignorance by committing an equally fallacious false dichotomy of ‘either evolution or ID.’ Granted, we might be hard-pressed to come up with an alternative natural explanation to rival evolution but that might just attest to the success of Darwin or the failure of our imagination. It is easier by far to go supernatural; if I am allowed to posit a flagellum-producing intelligence why not an unintelligent flagellum-assembling machine constituted by the Platonic ideals of cogs and springs running on steaminess? I fail to see why the eternal existence of such a Platonic Assembler should be any more mysterious than that of an Intelligent Designer.


We have seen that properly understood evolution might produce seemingly irreducible complexity by way of scaffolding, possible happenstance by neutral genetic drift, or cooption – each of which is supported by ample evidence. Thus having met Behe’s objection we have seen that even granting the falsity of evolution makes no headway in establishing an intelligent designer since a disproof of the former in no way constitutes proof of the latter. We have then seen that Dembski’s empirical identification of intelligent causation by specified complexity verges on the incomprehensible and to the extent it can be understood and applied to Behe’s position makes ID arguments out to be arguments from ignorance at worst and attempted inferences to best explanation at best. Lastly, we have seen such an inference to best explanation as unwarranted since evolution and ID exhaust neither all possible natural nor supernatural explanations and thereby finally collapse into false dichotomy.


Share and Enjoy:
  • Facebook
  • Twitter
  • Digg
  • StumbleUpon
  • Reddit

2 Responses to “Intelligent Design’s Abject Failure”

  1. mcmasterp says:

    wow. very thorough. This will make great reference material thanks.

  2. Thank you so much. I’m only glad if it can be of use. You should have seen my background reference list. We’re only allowed to put something in the bibliography if we actually reference it in the paper, you see.

Leave a Reply