In the article by Elliot Sober I mentioned in my previous post, he describes a hypothetical experiment to show what evolutionists mean by “random mutation”, before going on to establish that, because the Neodarwinian explanation is causally incomplete, science can neither deny, affirm, nor even express agnosticism on whether those random mutations were directed by an agent such as God.
The metaphysical point is clear, but it’s good to step back and ask why the experiment is mentioned. It seems to me that in the end it’s to remind us why Sober himself is not a theist. We’re to understand that evolution is producing observably random results, but metaphysically that that doesn’t exclude the involvement of God in them. One might notice a similarity to the position of Darrel Falk, of BioLogos, in that if God acts it’s undetectable, rendering the metaphysical permission to admit God a rather Pyrrhic victory. Wow, maybe it’s God who tosses the coin.
The simplified experiment described consists of putting red organisms is blue or green environments, and seeing if advantageous mutations matching the colour of the new environment occur more often than disadvantageous ones. It’s fictional, of course, but Sober points to similar, more sophisticated experiments that always show mutations to be random with respect to fitness. QED.
But is that how things really work? For a start, James Shapiro has given numerous examples suggesting that organisms initiate and even direct their own mutations for advantage. Those counter-examples alone show that the kind of experiments Sober describes are not conclusive.
Then again consider what real experiments could be managed as he describes.
(1) They are purely mutation based, which means restricting the organisms unnaturally to a cloned line – Lenski’s E. coli studies come to mind – and not much else. Otherwise you could never spot what is a new mutation. But mutation is only one aspect of variation.
A pseudonymous poster on BioLogos known successively as John, S. cerevisiae and Melanogaster is such a champion of population genetics that he frequently says (with little indication of hyperbole) that evolution has no need of mutations at all. Melanogaster’s posts, incidentally, are so sociopathic that it’s hard to resist the conviction that BioLogos keeps him as a kind of rotweiler on a chain to intimidate critics, as the usually irenic Bilbo suggested recently to Dennis Venema, whose bizarre defence of Melanogaster’s accusations was along the lines of “Yes, I know he’s got his jaws round your genitals, but you should listen to what he says.” It’s an interesting exercise to check the archives to see how many of his interlocuters have lost their patience with his insults and been banned, whilst Melanogaster himself pushes even the most sympathetic enquirers towards Creationism. That’s a digression, but it’s one that the leadership of BioLogos would heed if they had any sense, or honour.
Be that as it may, Melanogaster’s observation that real evolution is a population effect is valid. To watch out for individual mutations and their direct effects is not to watch evolution as she is spoke. Most evolution happens by the selection of alleles of genes within a heterogenous population. And that’s too complex to see in such lab experiments.
(2) Similarly, the need to observe every mutation, and its effect on the organism, bears little relationship to reality. Even Sober’s fictional experiment does not actually observe random mutations as he suggests, but random phenotypic effects assumed to be due to single random mutations. Of course, you need an unnatural cloned population to be able to know that any variation is due to a mutation.
In real life phenotypic changes are usually polygenic, and mainly occur through the combination of existing genes, not mutation. Are those combinations random? How do you control the variable enough to tell? Millions of varying genes and introns, simultaneous epigenetic effects, mutations popping up here and there in the population, varying phenotypes varying because of mutations or maybe the other factors – and someone can tell that the mutations were random? I don’t think so.
(3) Sober’s experiment deals with easily seen adaptive mutations. Gene mutates, colour changes, organism survives according to the match with environment. In real life, an environment is not just a colour, as the now infamous peppered moth research shows. In the real world neither “selection” nor “fitness” are possible to define. More variables to hinder the assessment of randomness.
(4) Another fallacy in Sober’s approach of “all other things being equal, each mutation is seen immediately to be adaptive or not” is that most mutations are near-neutral. They do not lead to selection. The theory goes that mutations accumulate under the radar of natural selection until, suddenly, you have a new functional protein, or a flagellum, or a DNA replication system. So crude fitness measures like colour match, or indeed any measures of fitness, can’t be used to gauge the randomness of whole sequences of near-neutral mutations leading, eventually, to selectable traits. It’s a different ball game deciding if a few hundred mutations that built up free of selection’s supervision into a functional system occurred randomly.
(5) Real evolution is a big picture, and lab experiments like Sober’s are inevitably studying the very small picture. Lenski’s work, despite the vast number of E. coli generations involved, has tracked a handful of minor mutations in one line of one species isolated in lab conditions, for just a few years. Evolution, thought, ought to be about the origin of the species. But the random pattern of coloured dots in a square mm of digital image says nothing about the ordered picture the photographer has created. The question is not whether a mutation is random, but whether the whole pattern of mutations is. No number of red organisms can show that.
(6) Lastly, such experiments in the context of the question of divine governance suffer the same shortcomings as lab studies of prayer. You divide cancer patients into a study and a control group, get a priest to pray for the former, and measure survival. Your results don’t show whether God exists, or whether he answers prayer, but merely whether he’s willing to co-operate with your silly games. Similarly, lab experiments like Sober’s hypothetical one or Lenski’s real one can’t deal with God’s potential objection: “I didn’t actually want to evolve your red organisms in any particular direction, whereas I had every reason to evolve mankind from a hominid template, and that’s too complex a situation for you to detect the non-randomness of my mutations.
So whilst I applaud Sober’s clear differentiation of the scientific from the metaphysical, I still have grave doubts that evolution, or the mutations contributing to it, can really be scientifically shown as random with respect to fitness. I believe that, too, to be a metaphysical assumption arising from Darwin’s theory rather than from the data as a whole, even though it can be demonstrated in laboratory experiments at scales to small to be generalisable. In other words, there may be more room for the detection of the effects of God’s action than Sober’s article allows.