A theological lens on “random with respect to fitness”

In the last few posts, I’ve been trying to point out the epistemological limits of science (and how they are routinely transgressed). In particular, I’ve tried to show how “contingency” and “randomness” are, in effect, epistemological black boxes in science. To say something is random, in science, should mean nothing more than “we do not fully understand the causes, and cannot predict the effects.”

Today I want to focus on the randomness of mutations, which as I tangentially mentioned in the last post is properly understood, in evolutionary theory, to mean only “random with respect to fitness”. What that means, of course, is that, if one explores genetic mutations overall, there appears to be no direct correlation with fitness – mutations are not simply the addition of new adaptive features.

In fact, although I’m too lazy to look up the estimates just now, I believe it’s the case that the majority of mutations are highly deleterious and eliminated by “purifying selection” – which in English simply means the organism dies. Nearly all the rest are near-neutral – slightly deleterious or slightly advantageous. Since it is notoriously hard to measure such things, neutral theorists disagree on the proportions. Classically neutral theory worked on the basis that most non-selected alleles that become fixed are slightly deleterious, but some versions of the theory allow for a higher proportion of marginally advantageous ones. Remember, though, that these estimates regarding what is subject to selection are all based on modelling – and we should always remember statistician George Box’s adage that “All models are wrong, but some are useful”.

But all agree that significantly beneficial mutations are extremely rare – to the extent that it’s hard to find anyone willing to put a percentage on it. That’s why Darwinism has always had recourse to the extremes of geological time, and is always embarrassed, or ought to be embarrassed, by instances of rapidly occurring adaptive changes.

From Darwin’s time, when the question was just about variation of traits, to today’s genomic era, such observations have led immediately to the conclusion that variation or mutation occur “randomly”, meaning undirectedly, with natural selection bringing order to the process. But this conclusion is actually metaphysical or theological, not scientific, as I have been attempting to point out in these articles.

Actually, since scientifically speaking “random” can never mean more than “unpredictable because causes unknown”, “random with respect to fitness” means only “unpredictable with respect to fitness.” Theologically, other explanations are infinitely preferable to “ontologically random” or “undirected”, and I’ll explore just one.

The problem comes from science’s self-restriction to efficient causes: the issue is seen purely in terms of “mutations are the source of variation: do they all invariably produce useful variation?” To which the answer is, of course, “No.” But suppose one allows final causation into the picture, as one must in theistic evolution? The issue then becomes, “mutations occur for various reasons and purposes, including beneficial variation: which of them are intended for evolutionarily creative purposes?”

I was alerted to this line of thought by good old Asa Gray, Darwin’s correspondent and possibly the first theistic evolutionist, who in discussing variation wrote:

The origination of the improvements, and the successive adaptations to meet new conditions or subserve other ends, are what answer to the supernatural, and therefore remain inexplicable. As to bringing them into use, though wisdom foresees the result, the circumstances and the natural competition will take care of that, in the long-run. The old ones will go out of use fast enough, except where an old and simple machine remains still best adapted to a particular purpose or condition — as, for instance the old Newcomen engine for pumping out coal-pits.

In other words, he saw the variations which produce the innovations seen throughout biology as God’s creative (and contingent) acts – even as supernatural acts – and natural selection as merely a “market force” ensuring that the new, improved, models are what survives.

Here’s an analogy. Back in World War II my father spent a period at the RAF’s underground nerve centre at Uxbridge, receiving secret coded messages. He wasn’t an Alan Turing, sadly – he never understood any of them, being paid only to translate the morse code accurately and take the messages to others to de-encrypt. But you’ll be aware that at various stages in the war, armies of intelligence people were listening in to, and decoding, all Germany’s encoded radio traffic, with specific aims in view: for example, to look out for the vital messages that would portend the threatened invasion.

The vast majority of messages were utterly trivial. Requests for routine supplies, minor redeployments, and so on. The significant stuff, as far as the codebreakers were concerned, might constitute one in a million, or less, of the things they laboriously decoded, and discarded.

From a scientific point of view, it would be accurate to say that the whole body of messages was “random with respect to invasion.” But it would be quite wrong to say that invasion instructions were therefore fortuitous and undirected.

Here’s a second analogy (also in the warfare field, I’m afraid – I’m still affected by my former comparison of natural selection to history and by the old concept of evolution as a struggle for survival). Imagine current intelligence personnel controversially trawling through the bank records of hundreds of people they consider capable of Islamist radicalisation. Telephone bills, supermarket bills, and other “neutral” expenditure attract no interest. Even the large purchase of a secondhand car means nothing significant. Here’s a family that blows a lot of money on a known fraudulent organiser of pilgrimages to Mecca, which is very sad for them but of no interest to Homeland Security.

But what’s this? A couple with Saudi origins has just purchased two one-way tickets to Turkey. That single purchase is the one that indicates a potential security risk: the family has “evolved” (if you like) from mundane innocence to potential terrorism.

Applying these analogies, suppose that mutations occur not blindly, but for various reasons, ultimately within God’s teleological purposes (ignoring  any supposed internal teleology – I’m accepting conventional Neodarwinism for this discussion). Not all of these purposes are creative – some are more or less lawlike, which is why studies attempting to apply mutation commercially to plants foundered on the fact that the same old useless mutations turned up over and over again. Many may be “pathological”, just as every living thing is subject to failures and traumas under God’s providence.

Yet suppose that, on the rare occasions when God creatively decides to modify life, he introduces some variation which is entirely positively determined with respect to fitness, and which he providentially preserves sufficiently for it to be guaranteed to become fixed according to the statistical lawlikeness of population genetics.

Scientifically such a variation would be contingent, unpredictable and, therefore, “random” according to the limited definition of randomness permissible in science. But “random with respect to fitness” it most certainly would not be – that description would apply only to mutations considered en masse, which is all that our limited science of efficient causes can see.

Such considerations matter because of the persistent error of many Evolutionary Creation proponents in confusing epistemological randomness with the metaphysical assumption of ontological randomness assumed by so many biologists. They then seek to justify God’s creation of such randomness in terms of nature’s “freedom”, and therefore to treat randomness as a reason to deny, or at least doubt, that evolution meets God’s specific ends.

The fact that Job plainly proclaims the creative pride of God in the detailed nature of ostriches, or leviathan, or behemoth, or warhorses; the fact that Jesus attributes the form of the lilies of the field to God’s wisdom and beneficience, and so on is therefore ignored, and replaced (simply because scientific randomness is illegitimately interpreted as ontological randomness) by an agnostic unwillingness to affirm God as the sole Creator of “the world and everything in it.”

Such ideas don’t come from the Bible, nor from Christian tradition, and I have shown sufficiently, I think, that they cannot be legitimately reached from science. So where do such ideas come from – and why should anybody be expected to take them seriously?

Jon Garvey

About Jon Garvey

Training in medicine (which was my career), social psychology and theology. Interests in most things, but especially the science-faith interface. The rest of my time, though, is spent writing, playing and recording music.
This entry was posted in Creation, Science, Theology. Bookmark the permalink.

12 Responses to A theological lens on “random with respect to fitness”

  1. Sy Garte Sy Garte says:

    I like this very much. Lets try a thought experiment. Suppose we could somehow find a way to determine for any human population of a million or so people living in a defined area during the space of say one year, how many of the people in this sample who had been ill or blind or insane or suffering from some serious malady, suddenly and irreversibly improved or recovered from their illness. We might find a statistical proportion of such ill people who underwent rapid recovery, that would be roughly similar among different areas of the world at specific times. In modern times, it might be much higher in Western or advanced nations that it would be in primitive societies, or in ancient days and the explanation would be the advent of effective modern medicine.

    Suppose we did this experiment during the year 33 AD (I think that is correct) in the area of Galillee and Judea. We would find roughly the same rate as say in neighboring Egypt or Greece (perhaps on the order of 5% or whatever). Some of these cases would be spontaneous recoveries, some might be due to radical treatments, and some might be considered miraculous.

    The experiment is to test the notion that in Judea, but not in Egypt or Greece, many of these recoveries were in fact the miraculous work of Jesus Christ and that such an efficient cause could be determined from the data at hand. I submit the answer is no.

    In fact the additional number of such cases that resulted from the direct intervention of Christ (we dont know the number since the Gospels tell us it was larger than the number described in detail) might not have elevated the total much above the background seen elsewhere, and even if it did, it only lasted for a brief time, and would be lost in the noise of the measurements.

    So, my point is, even if we KNOW, (as we do in this case) that God’s divine providence worked miracles in the world of nature, we cannot see evidence for such intervention from a long view. And it is that kind of view we use when we look at mutations. Therefore it isnt at all surprising that mutations as a class appear to be ontologically random, since most of them probably are, but the few that count the most, could easily be miraculous, absolutely teleological, and beyond detection. Does that make sense?

    • Jon Garvey Jon Garvey says:

      Hi Sy

      Yes, that appears to me in line with the kind of comparison of contingency and reproducibility I’m trying to present.

      Let’s push the boat out a bit further. Reading Keener’s book on miracles (and surveying my own church informally when preaching) it seems that the healing miracles of Jesus continue at some level throughout the world even now. So whilst surveying our recoveries from diseases, the assumption that it was all due to medical advances might be less than secure. “Spontaneous remissions” could only be called “natural” if in each and every case prayer (or even common grace) were excluded… which is both methodologically and conceptually impossible.

      The bottom line is that one might well detect some more-or-less-constant low level pattern leading to a scientific conclusion that approximately 0.00whatever% of such cases get better – without realising that one was actually logging instances of unique miracles. So you detect reality being changed – but you’re totally unable to detect the cause, at least using science. Of course, if you heed patient testimony, it might be a very different story.

      Our suggestion that “the most useful mutations may be direct divine actions” is not, of course, a scientifc one: but that’s the whole point. It’s the kind of question that science ought not to make conclusions about, without any means to draw a distinction between the different classes of mutation.

      “A mutation is just a mutation” is, perhaps, the only working assumption one can make if ones epistemology is limited to material efficient causes. But one ought to be very clear what possibilities that approach excludes.

  2. GD GD says:

    “Scientifically such a variation would be contingent, unpredictable and, therefore, “random” according to the limited definition of randomness permissible in science..”

    I think a distinction needs to be made, and it is this lack of distinction that seems to me to be the cause of confusion and disagreement. I will do my best to make this clear, but I suspect my best effort may not be sufficient…. but here goes…

    (1) The ToE proposes that random mutations and some type of natural selection is the basis for the scientific credibility of the theory. Yet selection of any significance is either undetectable, or is inferred, or is misunderstood as an outmoded version of neo-Darwinism. If this summary is valid, scientists must object to the scientific basis for ToE – this is obvious as the science identifies a gross inadequacy at the most basic level.

    (2) Yet TE/EC claim that ToE is true, while some details “need to be worked out”, thus ToE is God ordained, or it is how God has done things.

    It should be obvious, I hope, that (2) cannot be consistent with (1) – if ToE is in doubt based on scientific non-detectability of whatever selection is proposed, than it cannot be true in the sense that “it is God-ordained”. Otherwise, the result of this would be the absurd statements(s) made against anti-evolutionists, in that God has decided to cause us to believe a lie, or to sustain a scientific untruth.

    I have tried to see this from other perspectives, but I cannot come to any other view in the end, in that people are conflating the mix of settled and speculative aspects of the sciences, with the notion that science reveals what God does. This is a classic case where accepting speculation and uncertainty in science can lead to theological error and absurdity. Trying to make it all appear “natural” as gravity or rain simply adds to the absurdity.

    This error makes anti-theists and materialists to be confident of (1), as they reject (2). and their rejection of (2) is embolden by the inconsistency between (1) and (2).

    • Jon Garvey Jon Garvey says:


      For all my skepticism, on (1) I think you may be slightly exaggerating the sidelining of natural selection, even in neutral theory. There certainly seem to be instances (such as lactase persistence) where it’s a significant factor. But you’re still right to say that its role is so fluid, and so subject to dispute and alternative interpretations, that to treat it as so settled that one builds it into ones theology as a key component is bizarre.

      In theistic evolution, the most benign version of this goes back to Kingsley et al: that unlimited variation and the sure and certain governance of natural selection can be seen as an elegant way for God to build his world towards perfection. And so it would be, if it were proven – but no more so, intrinsically, than the most literal form of special creation: the world is so wonderful that whatever means God used, or not, must be equally wonderful.

      But when, as is certainly the case in modern evolutionary theory, natural selection is relativised by neutral change and other mechanisms and accidents, and organisms are seen as jerry-built patchworks holding on to their survival simply because the competitors are worse; then that optimistic theological worldview won’t hold water, and the doctrine of the goodness and wisdom of creation becomes denied by what is still pretty provisional science (posing as the whole truth).

      It gets more malign when “randomness” (against which my argument has been chiefly made recently) is held up as a crucial principle of creation, re-branded as “freedom”, and used to completely subvert the historic doctrine of God.

      The result is the usual incoherence (and I note that having encore une fois pointed out this incoherence on BioLogos, two days later nobody has been prepared to defend it). There, “determinism” has been nonsensically opposed to nature’s “free will” – actually forgetting that free will is simply the determination of events by some other party than God, quite apart from the old question of the impossibility of “will” apart from “mind”.

      Evolutionary Creation, as a movement, has manifestly failed to learn the lesson of history of the difference between current science being compatible with the faith, and current science determining the faith. Why that should be is another matter, but it may have to do with the fact that it has developed largely amongst scientists, who are very vulnerable to the soft scientism that simply cannot perceive, in reality, that science is a limited tool.

      • Edward Robinson Edward Robinson says:

        Well said, Jon.

        The last sentence is of course one I agree with. When we look at the leaders of EC/TE in the ASA and on BioLogos (and in corresponding British organizations) over the past 20 years, there is a clear preponderance of people whose main (and sometimes only) formal intellectual training is in the natural sciences. Sure there are recent exceptions, such as Jim and Brad at BioLogos, and of course Ted Davis has a balanced education in science and the humanities, and George Murphy is well-versed in theology, and some well-rounded non-BioLogos types like Russell exist, but overall the movement in modern times has been led by scientists, and scientists of a conventional, non-maverick temperament, who are already professionally convinced that evolution (usually in neo-Darwinian garb) just has to be true, and therefore if Christianity seems to be in conflict with evolution, the right thing to do is change either traditional Biblical exegesis or traditional theology. The science part just isn’t negotiable.

        And this reflects the prejudice, particularly strong in America where the alleged distinction between “objective sciences” and “subjective arts subjects” is almost a fundamental religious dogma of university and popular culture. *Of course* the people who study the “soft” subjects (like theology and philosophy and history of Christian thought) ought to be the ones to bend, since their artsy knowledge is questionable to begin with, unlike scientific knowledge which is of course solid and reliable.

        Further, many of these scientists, in addition to having the typical American anti-arts prejudice, also are convinced that religion is purely about feelings, not about evidence, and so for them as long as one loves Jesus (or what one conceives Jesus to have been about), things like rigorous systematic theology or church tradition, not being about religious feelings, aren’t that important. One TE/EC who has recently come into prominence in the USA told me he disliked Discovery’s book on God and evolution, because it didn’t wax emotional about Jesus nearly enough for his religious taste. But for some of us, Christian theology doesn’t require waxing emotional about Jesus 24/7, and in fact emphasizing Jesus in certain ways (as in “kenotic” creation) actually obscures rational and empirical thinking about origins, without having any corresponding gain in theological clarity.

        The natural mediating body of thought between a hard-nosed science with subtle tendencies toward scientism and an emotionalist religion which has no use for systematic thought is the subject known as philosophy. But most of the BioLogos leaders have had very little use for philosophy, or even for the philosophical component of theology.

        • Jon Garvey Jon Garvey says:


          The late Austin Hughes, whose paper I offered at BioLogos, also wrote a very lucid essay on scientism here.

          His main target, of course, was the illogical positivists like Coyne and Dawkins, and he considered (I read elswehere) that ID was a scientifically naive enterprise (though I’m not sure he wrote anything that would enable one to see how much he’d studied it firsthand).

          But I note that he also has a swipe at the fawning attitude of many analytic philosophers to the inerrancy of science… he might also usefully have included a good number of theologians in that. He himself studied real philosophy, and knew his way round Descartes, Aristotle and Plato pretty well.

  3. pngarrison says:

    Jon, I just have to dispute one point. Sy may know better than I do, but every estimate I’ve seen of mutations in higher organisms with large genomes is that neutral and near neutral mutations represent the vast majority of mutations. Deleterious mutations constitute a smaller percentage on the order of a few % or so and of course mutations subject to positive selection are very rare.

    Of course these estimates are for SNVs and they don’t hold necessarily for other types of mutations. Every human genome has some large copy number variations (CNVs) and complex rearrangements which are relatively few in number but affect more total nucleotides than point mutations do (SNVs.) Because they can be large, they are probably much more likely to have effects, positive or negative, since they are more likely to overlap genes and functional elements and change copy number of these things.

    There are also something approaching a million loci of so-called short tandem repeats, where each repeat can be from 2-8 or so nucleotides long, and the number of repeats in one STR locus shifts by one or two copies to give the highest mutation rates in the genome. The number of STR copy number changes in a generation may be larger than the number of new single nucleotide changes. These loci occur both in and between genes, and examples are known where they alter protein sequences and regulatory sequences and I don’t think I’ve ever seen any estimate of how they divide between neutral, negative or positive in their effects, but it’s probably different once again from other mutation types.

    In short, there’s a lot yet to learn, but at this point it seems like a reasonable guess that neutral mutations far outnumber selectable mutations.

    A high school son of friend once asked me what I thought about evolution and natural selection. My off hand response was, I generally thought it works but “maybe a few mutations aren’t really random.” I think I was close to your meaning there.

    As to the general philosophical point, you are making, I have long had a personal nomenclature which seemed to use common meanings to keep things straight. If something is predictable in principal, but unpredictable in practice, that means that either our knowledge or our computing power is not up to the task. That’s “uncertainty,” because it’s an admission that we don’t know, at least for now. Chaotic systems may be permanently in this class because of their intrinsic sensitivity to starting conditions, etc. The alternative is “indeterminacy,” in which no amount of knowledge of the system or amount of computing power could allow us to predict the outcome. I gather this is where some interpretations of quantum events fall. It doesn’t mean God is limited in knowing or affecting these things, only, to quote Paul Simon, ” the information’s unavailable to the mortal man.”

    • Jon Garvey Jon Garvey says:


      I’m happy to accept your dispute, since it doesn’t affect the point I’m making. My estimates were based on the memory of a paper I have somewhere, but the effort of finding it didn’t seem worthwhile. Since my estimates were also short on definition of what kind of mutations I meant, I wouldn’t die for them… except, of course, for the rarity of selectively favourable mutations.

      FWIW a quick internet search produced a couple of papers worth quoting:

      This one says:

      Analyses of DNA sequence data have also shed some light on
      the distribution of fitness effects of new mutations. It is evident
      from the highly conserved nature of most protein-coding sequences
      that most amino acid mutations are strongly deleterious.
      It has been estimated that approx 70% of all amino acid mutations
      have a deleterious effect of >2 X 10-5. It has also become
      apparent that there is a class of slightly deleterious mutations,
      mutations that are sufficiently weakly selected that they can
      contribute to polymorphism and occasionally become fixed.

      My idea may have come from a similar source, and of course it applies to amino acid mutations rather than all genes.

      The other one one says:

      We show that a gamma distribution with a shape parameter of 0.23 provides a good fit to the data and we estimate that >50% of mutations are likely to have mild effects, such that they reduce fitness by between one one-thousandth and one-tenth. We also infer that <15% of new mutations are likely to have strongly deleterious effects.

      Different subject, different methodology, and 15% deleterious – I wonder if that includes those so deleterious that they are never inherited? And, of course, a methodology that actually accounts only for 65% of the total. So hey-ho!

      Your terminology, I thought, was sounding reasonable (though the trouble with any tweaking of existing terminology is that it just gets added to the existing confusing terminology and then, in turn, gets tweaked by those who don’t get what you meant!)

      But in fact, though it distinguishes the commonest understandings of chaotic and quantum events, it misses out the key point of contention, which is the predictability of events to God, or perhaps better, events that do not find their ultimate origin in God and are therefore fully known to him. In this (which is where openness theology and the woollier versions of TE place their bets), God sets up a random-number generator to surprise himself, so that in effect such events are causing themselves according to no principle.

      That’s what I would call “ontological randomness”, in that it has no immediate cause at all. “Epistemological randomness” includes anything that’s knowable, that we don’t know: on that basis, whilst no computer could even in principle predict quantum events, asking God nicely enough could produce the answer, rather than the response, “Why ask me – it’s random, isn’t it?”

      On that understanding your “uncertainty” and “indeterminacy” would be subdivisions of epistemological randomness. But don’t tell that to commenter “Thanh_C” at BioLogos , who complains that even the division between the ontological and epistemological forms are too complicated to understand!

  4. Jon Garvey Jon Garvey says:


    Since people here seem to have latched on to the concept I was seeking to present, I perhaps should add that it ought not to be restricted purely to adaptive selection, but (as I hinted in the OP) to any other purposes that God might have for a particular mutation.

    One example might be ORFan genes, which often, I understand, arise from sequences developing neutrally in non-coding areas of the genome. Therefore, although entirely de novo as genes, it’s argued that they are still the gradual work of random change rather than saltation.

    But of course, creating beneficial sequences gradually is no less wonderful and creative than doing it at once – especially when they build up without the guidance of adaptive selection. Whether or not the cells have their own physiological mechanisms for identifying and “genifying” such sequences, Asa Gray’s argument still holds: when beneficial variation occurs, it is far more likely to be the result of design than of chance.

  5. Sy Garte Sy Garte says:

    Preston is right about the frequencies (since most mutations either dont affect the amino acid choice or make a conservative substitution). And of course, the frequencies dont really matter that much. If a cell, or even an embryo doesnt survive a deleterious mutation it has little evolutionary significance. But a single mutation (and I am including the large ones, that Preston discussed) that produces a jump in fitness, a new body plan, a new fundamental function or structure, can have major consequences for the course of evolution. The question Jon raises, namely can we know if such “extraordinary” mutations are really providential or just good luck, may not be answerable. But there may be some clues.

    Last February, some here might recall that I posted a review of Michael Denton’s new book on Biologos. The ensuing comment thread took (as they usually do) several turns, and at one point I was asked to post links to some of my recent writing about new evolutionary ideas. I did so, including a link to a piece in God and Nature (the ASA ezine) which was a shorter version of a paper published in PSCF. A commenter suggested that I had misunderstood the literature I was reporting about, which showed that in some cases, bacteria undergo non random mutations. I then foolishly got involved in a heated exchange, which had some negative consequences. My point however is that the reason the exchange become so heated was that my opponents (there were two) refused to believe the message conveyed by a whole body of literature (my article cited 5 papers, if I remember correctly) and insisted that they didnt say what they in fact said. I can understand their agony, since the papers in question confirmed and extended an extraordinary finding by John Cairns in 1988 that when bacteria are placed under stress, they can indeed mutate the very gene that can alleviated the stress. This result was debunked, and then un debunked, and debated for decades, until it became clear that the phenomenon was real, and could be explained, not as acts of intelligence and planning on the part of the bacteria, nor by providential intervention, but by a very naturalistic mechanism involving the expression and unwinding of the chromatin at certain genes involved in the stress response, leaving them open to mutation and error prone repair.

    However, while there is no evidence for providence in the mechanism of what is now called “stress directed mutation” the fact is now more or less universally recognized (despite the strong opposition from some, including the two ill informed commenters on my post) that non random mutations are possible. Sometimes. In bacteria. So far.

    I believe this is a scientific opening of the door. Where that open door will lead we do not know. But with scientific evidence for the existence of non random mutations, the concept of providential mutation need not be restricted to purely extra scientific thought. If nothing else, it dispels the idea that God needs to “poof” such things to happen. Maybe God does “poof”. And maybe He uses much more natural means. I am not sure that matters. What counts is, we are here, and that to me is a miracle.

    • Jon Garvey Jon Garvey says:


      I’m so amused that the objection was that you’d misunderstood the paper… I wonder where I’ve heard that before, even this weekend!

      On the minor matter of mutation rates and “instant” purifying selection, I wonder if the high rate (maybe 25-50%) of early miscarriage (at least in humans) is relevant. They are known generally to be due to fetal, rather than maternal factors, and so maybe represent lethal mutations, at least sometimes. They wouldn’t be relevant to evolution, but would be highly relevant to raw mutation rates.

      But translating that into a comparison with te rate of near-neutrals would require a lot of maths, and we agree it doesn’t affect the point at issue.

      • Jon Garvey Jon Garvey says:

        Forgot to add, re this:

        And maybe He uses much more natural means.

        That perhaps we should regard “natural ” as a synonym for “lawlike”, as random is a synonym for “contingent”. Once we see both as God’s general and special providence, then it truly doesn’t matter which is afoot!

Comments are closed.