Barefoot in the wilderness
in search of understanding

The ID debate (VI)

(Following on from The ID debate I, II, III, IV and V, here is my final contribution.)

Peter said:

For a defence of premise (a) of the argument for ID you really should read some of the work by William A. Dembski on ‘specified complexity’.

I’ve read some of what Dembski has written and I’m not completely impressed. His logic is often good, as far as it goes, but when applied to the actual biology there are some serious issues. The first is whether biology is “complex” in the way he needs it to be.

The initial problem here is how we calculate the probability – from what previous state is the allegedly complex state produced? In “NFL”, Dembski makes an astounding error by trying to calculate the probability of a flagellum spontaneously assembling from proteins. There are in fact two errors here. First, evolutionary biologists do not assert that the flagellum appeared spontaneously – they assert that it developed from prior forms. This is crucial and renders the calculation moot. Second, though, the question isn’t at all to do with the assembly of proteins – evolution works on DNA, not protein! The calculation that must be done is as follows: What set of DNA alterations must be gone through to generate a flagellum from a putative prior functional DNA sequence?

A crucial point to note here is that we can only assess the probability of (for example) the flagellum given a specific prior state. The probability will be radically different given different prior states. (Dembski actually touches on this in his criticism of Dawkins’ equally terrible attempt at justfying the generation of complexity, but sadly doesn’t seem to follow up on it.) This makes establishing whether the complexity of a biological system crosses the threshold rather difficult; one would have to establish all possible prior states and also balance the relative probabilities of those states in order to arrive at a final assessment of this value.

The fact that Dembski even tried the calculation based on protein assembly rather than DNA sequence (let alone published a book containing it) does not give me the impression of someone who understands the biological issues involved.

The idea of specification is also questionable here, because natural selection is a generator of specification, and its influence must be accounted for. To establish Design, we don’t just need specification but “more specification than natural selection could achieve”.

From what I’ve read, though, Dembski sometimes stops short of saying that biology does have SC – he seems to want to establish SC as an independent set of criteria. However, it seems to me that we have problems even if biology does show SC. Dembski claims that we have only one known source for SC, which is intelligence. However, this is assuming precisely what ID is attempting to prove! If biology does show SC, the question is whether this necessarily implies intelligence at its source; we cannot therefore simply claim that the only other examples we’ve seen of SC were designed by humans and therefore that evolution must also have been designed by an intelligence. An inductive chain of reasoning like this must be suspect. If biology does show SC, it is a type of SC that we’ve never seen elsewhere, which makes induction even more problematic; it is assuredly not designed by humans and hence is outside the category in which we have previously seen SC.

The big problem, though, is that even if we accept that SC can produce a “design detector”, it is not itself that detector – it’s only a toolkit that promises to let us build one. Determining whether and, if so, how SC applies to biology is a task that still needs to be done. The only actual attempt I’ve seen (irreducible complexity) seems to me as a biologist to be fatally flawed. It cannot reasonably be maintained that any of the examples so far produced is actually irreducibly complex, let alone incapable of generation via an evolutionary process. And even if it was, it is not certain that irreducible complexity itself meets the criteria for SC (Dembski, for example, keeps adding new criteria in various places).

Would you claim that, contra Dembski, either in principle or in practice, there is no scientific way to rule in ‘intelligent design’ as the best explanation for anything?

The scientific method involves finding the most parsimonious explanation for a phenomenon. Generally, this means using intelligent intervention as the final possible resort. Which, incidentally, is partly why SETI is such a difficult task – ruling out any possible natural cause for a putative signal is required for it to be ruled to be of intelligent origin.

That is, the scientific method will never “rule in” intelligent design – only fail to rule it out. This is the very nature of the scientific process, and one reason for such strong objections to ID theory from the scientific community.

And what about cryptography, forensics, etc?

I’m not sure what you mean here. Both these are human investigations of human phenomena (if you like, they are engineering, not science – the application of knowledge rather than the discovery of knowledge). Although it’s interesting in this context that, in cryptography, a perfect cipher would be indistinguishable from random noise unless you possessed the key to reading it. That is, the more intelligent it is, the more random (and less specified) it appears.

If, on the other hand, there are ways of ruling in design as the best explanation, why is it in principle illegitimate to apply these methods to the things studied in physics/biology?

The problem isn’t that it’s illegitimate. Rather the issue is twofold: (i) a weak “design detector” is used to make strong conclusions; (ii) design should always be the last answer, not the first one. Science always reaches for design last because doing otherwise cripples the investigation process – once we decide it’s Designed, the investigation stops because no further explanation is possible. Indeed, one could say that the entire point of science is trying to remove the claim that “it’s designed by God” and instead say that events are controlled by invariant natural laws and principles. This is not anti-theistic, or need not be. Rather, it is the attempt to investigate how the world is made. If we eventually find that we cannot rule out Design, that will be a profound scientific discovery – but it would also be tentative. For all scientific “truth” is tentative. As knowledge advances, so our old answers are surpassed. So, until our knowledge is complete (which it can never be), science will never “prove” that the Universe is Designed – although it may fail to prove otherwise. (That is, there may be things that science never succeeds in explaining properly; however, such failure can never prove that no explanation is possible.)

So, in terms of your two premises, I still have not seen a detector that passes criterion (a), which makes point (b) moot. If we were to construct a solid design detector, I would be as fascinated as anyone to see whether anything actually passed it! Given the nature of science, though, as I described above, I’d be surprised if we really can construct a viable design detector – not that I think it to be impossible in principle; rather, I think that the contraints that would have to apply would make a generalised design detector practically unattainable.

on the face of it your theological critique seems to be quite different from fellow theistic evolutionist Denis Alexander’s theological critique

Wouldn’t surprise me – I’ve not read anything by him. These are my own thoughts, for better or worse.

What is wrong with positing two modes of divine action, one of which is describable in terms of secondary causes [snip] and the other of which is not describable in the same way?

If we assert that God can work through secondary causes, as ID tries to and theistic evolutionists presumably do, we are saying that evolution as observed is capable of producing variation, novelty and adaptation. So far, so good.

Then, we try to introduce a second category of events that somehow God cannot produce by the above mechanism. This immediately suggests the application of Ockham’s razor and demands that we ask why God cannot use the usual tools. For the evolutionist (of whatever flavour), the answer is that God can, in fact, use those tools because they are sufficient to produce any variation required – given enough time. The IDer must produce a convincing theological (and scientific) explanation for why this is needed (the onus is on ID here because the proposal goes against Ockham).

This is the point where dualism is a danger, and something that I’ve seen in various places when this is discussed. Inevitably, it seems, we have to describe evolution as lacking somehow, that God cannot work through it but must instead get involved directly. This relegates evolution to being a shoddy tool, unworthy of being wielded in the divine hands – it even leads to evolution being regarded as not truly being wielded by God at all, being instead “blind accident”. When God really wants to get something done, God gets stuck right in! Which is tantamount to saying that, unless God interfered grossly in Creation, God’s plans wouldn’t come to fruition.

This is the point where ID, it seems to me, is not taking seriously enough God’s position as Creator and sustainer of the world. For we now have two radically different kinds of activities – one in which God’s plans are carried out (direct intervention) and one that limits God’s plans (evolution). ID looks too much like a search for God in science – because the searchers cannot see God already there. There seems to be a need for gross evidence of the divine handiwork, because the ongoing humdrum activity of Creation is somehow too mean to be where God is seen. But, if God is Sustainer, the humdrum is precisely where we ought to expect to see Him.

This ties in with the nature of “miracles”, because I can see that some people might start gnashing their teeth at the above, thinking that I’m ruling out “miraculous” intervention in the Universe! However, I’m not doing that at all. The issue isn’t “supernatural power” but how God uses that power. Whenever we see miracles, the issue isn’t the exercise of power but the sign that is made. Indeed, in almost every case, “miracle” would be better translated “sign” (or so I am assured). That being so, I think that it’s clear that putative Design events hardly qualify as signs in the biblical sense; for even if they were found in the scientific evidence, they would have no meaning
beyond themselves, no divine message to convey.

pax et bonum