Faith & Science
Answering Swamidass on Theistic Evolution: Sketchy Science, and a Swerve into Metaphysics
Dr. S. Joshua Swamidass of Washington University has written a review of the book Theistic Evolution: A Scientific, Philosophical and Theological Critique, for which I served as an editor of the science section. I in turn have been evaluating his review (see here and here).
Swamidass talks about God’s providential guidance of evolution but his subtext is intelligent design, as the alert reader may have noted. I have used his language about God throughout these several posts because we have been dealing with theistic evolution, which does make claims about God’s action in the world. Intelligent design theory, however, stops short of the theological leap to God. We can infer a designer philosophically, but we go no further.
This next section of Swamidass’s review puzzles me. To start off, Swamidass dismisses the book’s scientific arguments by stating:
The scientific arguments, however, are not convincing. They talk past the mainstream account of our origins without ever actually engaging it (pp. 503–22). Those hoping to understand the strengths and weaknesses of evolutionary science should look elsewhere.
I’ll have to look elsewhere to understand the strengths and weaknesses of Swamidass’s argument. He says we don’t engage the mainstream account of our origins, but he provides no specifics and shows no lack.
In fact, all the chapters in the science section engage mainstream science: the origin-of-life problem, the problem of information, the difficulty of engineering or evolving new proteins, the intricacies of developmental biology and their challenge to Darwinian explanation, the many kinds of codes in the cell that go beyond DNA and that cannot be accounted for by a mutation/selection mechanism, the fossil record and the problem of the sudden appearance of innovation, the evidence and arguments against common descent, why frontloading or evo devo, emergent properties, or other “extended synthesis” ideas won’t work, and the problem of accounting for human origins. I submit that it is Swamidass who hasn’t engaged our arguments, simply dismissing them with a rhetorical wave of the hand.
The only place Swamidass does engage the book’s scientific arguments is regarding the chapter on human genetics, “Evidence for Human Uniqueness.” (By way of disclosure, I was the primary author of this chapter.) Much of what he says needs some unpacking, so bear with me.
Swamidass begins with a challenge based on the neutral theory of molecular evolution. He sees this as profound evidence for common descent. The neutral theory claims that most evolutionary change is driven by genetic drift rather than positive selection. Genetic drift is a random process, where luck determines which mutations are preserved and which are lost. We usually speak of evolution “selecting for” beneficial mutations and “selecting against” harmful mutations. But in fact, selection is only one mechanism of evolutionary change, and most common mutations we find in genomes today are thought to be nearly neutral, neither good nor bad. They have become common or universal (fixed) in the population because every generation some mutations decrease and some increase in the population, for random reasons having nothing to do with whether they are beneficial or harmful. Some are lost entirely and some advance toward fixation. The net effect is an accumulation of mutations at a more or less steady rate. From this a number of interesting mathematical measures can be derived.
But Swamidass aims at nothing so sophisticated. He simply wants to point out that the longer two species have been separated the more different they are (assuming common descent). In his critique Swamidass says:
[T]here are ten times less differences between humans and chimpanzee genomes than there are between mice and rat genomes (e.g., see The Chimpanzee Sequencing and Analysis Consortium, “Initial Sequence of the Chimpanzee Genome and Comparison with the Human Genome,” Nature 437 : 69). Even with different measures, mice and rats are much more different than chimpanzees and humans. Why? Humans and chimpanzees mutate slower and diverged more recently. [Emphasis added.]
This is actually a trivial statement, given the neutral theory. It can be formulated as
R x 2T = D
R (mutation rate) x 2T (total divergence time) = D (genetic distance). One can use D and T to calculate R, or D and R to calculate T. But there’s a danger: you can’t use this relationship to “confirm” common descent, unless R, T, and D are determined independently of one another, because the equation depends on the assumption of common descent.
Swamidass cites the chimpanzee genome paper mentioned above as the source of his number “10 times.” He uses data on the number of synonymous differences among aligned genes to arrive at his estimate of ten times the genetic distance between murids (rats and mice) and primates (humans and chimpanzees). These measures give me a value of 15 times the genetic distance.
I also found an estimate of the number of substitutions (genetic distance) and the divergence time for rats and mice in the rat genome paper, which says that “the rat genome has accumulated 8–10% substitutions since the speciation from mouse 12–24 Myr ago…” For humans and chimpanzees the genetic distance is 1-2% depending on sources, and the divergence time is 6 Myr. According to this source, Swamidass is right that the genetic distance between rats and mice is about 10 times that of chimpanzees and humans.
But what’s the point? Although the above implies humans and chimpanzees diverged more recently than mouse and rat, within a framework of common ancestry, it does nothing to validate the framework itself. The former pair is simply more similar than the latter pair. To say the fact that rats and mice are 10 times more different genetically than we are from chimpanzees does not prove common descent nor does it disprove it. All it says is that mutations accumulate at a measurable rate, a rate that is by definition consistent with the proposed time of divergence. Swamidass likes the way the numbers work, saying they support common descent. If common descent is true, this steady accumulation of mutations is consistent with it. However, to justify the statement that common descent is true because the relationship D = 2T x R is true, he must demonstrate it is true independently of any methodological assumption of common descent.
But there is still a bridge to be crossed. The neutral theory is not sufficient to account for us as we are. Some mutations must have been guided to account for the differences between us and chimpanzees. Random neutrally accumulating mutations cannot generate in 6 million years the kinds of coordinated functional changes to anatomy, physiology, intellect, language ability, and sociality (among other things) that must be explained. These are things Swamidass does not address. I point out the kinds of changes that might be involved, namely changes in gene regulatory elements, splicing, RNA editing, non-coding regulation of chromosomal behavior, and more. But they would have to be guided, as my chapter discusses.
I have every hope that we will continue to learn more about how we differ from chimpanzees. It’s not all in the genome. One fact that should already astound us: as Josh says, “Humans and chimpanzees mutate slower and diverged more recently” than rats and mice, and so should be more similar to each other than rats and mice are. Instead, we differ a great deal more phenotypically than rats and mice do. We have language, abstract thought, music, art, physics, humor, ethics and morality, and religion. Chimpanzees don’t. Something big must have happened.
The Y-Chromosome Mystery
Swamidass then focuses on one sentence from my chapter: “In addition, the Y-chromosomes of chimpanzee and human males are very different from each other; the Y-chromosome represents 1.8% of the genome.” He says:
Likewise, human and chimpanzee Y-chromosomes are more different than the rest of the genome. Why? For the same reason; Y-chromosomes mutate more quickly than the rest of the genome. Remarkably, the increased divergence of Y-chromosomes is presented as evidence against common descent, rather than what it is: clear and quantitative evidence for common descent.
First, let me acknowledge that Y-chromosomes do have a higher mutation rate than autosomes (non-sex chromosomes), and that males have a higher germline mutation rate (they pass on more mutations to their children) than females do. I suppose that is what Swamidass is saying in his statement, “Y-chromosomes mutate more quickly than the rest of the genome” (the rest of the genome meaning, I think, the rest of the human genome for the human Y and the rest of the chimpanzee genome for the chimpanzee Y). It is reasonable to conclude that “human and chimpanzee Y-chromosomes will be more different from each other than the human and chimpanzee autosomes will be from each other.” The question is how much more different.
The initial paper from the chimpanzee genome project reported that the Y-chromosome had diverged 1.9 percent, compared to 1.06 percent over the whole genome, and 0.8 percent on the X-chromosome. The authors acknowledged they had difficulty aligning and masking repetitive sequences, and had insufficient data for comparison, all of which would potentially affect their results.
What data do we have now about the Y-chromosome? Human mutation rate estimates for the Y-chromosome have gotten progressively better as more data has become available. A 2009 paper that sequenced Y-chromosomes from two Chinese men that were separated by 13 generations reported a mutation rate of 3 x 10^-8 mutations per nucleotide per generation (MPNPG). A 2015 study sequenced 753 Y-chromosomes from Icelandic men related by descent over 1300 generations, and found a mutation rate in humans of about 3 x 10^-8 MPNPG also, this over 21.3 Mb of the male-specific Y (MSY). Calculating the expected divergence yields 1.8 percent mutational divergence between chimpanzees and humans, assuming they have the same mutation rates and MSY sequence size, and assuming common descent. That’s a pretty good match with the value reported in the chimpanzee genome paper, 1.9 percent divergence.
But this neglects the Y-chromosome structure. In 2010 the chimpanzee Y was carefully sequenced and compared to the human Y. The surprising result was that the non-orthologous MSY regions of the human and chimpanzee differ by 30 percent. This was not predicted. The authors of the paper describing the Y comparison said at the time that everyone expected the Y to be stagnant because it had no partner to recombine with.
Jennifer F. Hughes et al. continue:
…[C]himpanzee and human MSYs should have changed little since the separation of these two lineages just 6 million years ago. To test this prediction, we aligned and compared the nucleotide sequences of the chimpanzee and human MSYs. As expected, we found that the degree of similarity between orthologous chimpanzee and human MSY sequences (98.3% nucleotide identity) differs only modestly from that reported when comparing the rest of the chimpanzee and human genomes (98.8%). Surprisingly, however, >30% of chimpanzee MSY sequence has no homologous, alignable counterpart in the human MSY, and vice versa. In this respect, the MSY differs radically from the remainder of the genome, where <2% of chimpanzee euchromatic sequence lacks a homologous, alignable counterpart in humans, and vice versa.
… Indeed, at 6 million years of separation, the difference in MSY gene content in chimpanzee and human is more comparable to the difference in autosomal gene content in chicken and human, at 310 million years of separation. [Emphasis added.]
Clearly, this was not predicted by the neutral theory. 30 percent change in 6 million years for the Y compared to <2 percent for autosomes is a big surprise. Logically, it cannot be used as an argument for common descent. Certainly, mechanisms can be proposed, and they have been. But that does not constitute an answer. The 30 percent divergence can be used to suggest a unique origin. Ultimately, we don’t know why the Y is so different in chimpanzees and humans, except that it is.
While we are talking about arguments for and against common descent, it’s good to remember that common descent cannot explain biology, even if it should prove to be historically true. It is not a mechanism of change, it is a statement of relatedness. Two organisms can be similar for a number of reasons. There are two chapters in the Theistic Evolution book on the subject of common descent that Swamidass does not engage, three if you count Bechly and Meyer’s chapter on fossils, each with LOTS of exceptions to the apparent relatedness of biological things.
Swamidass favorably mentions another chapter on human origins that I contributed to, because it has to do with the population genetics of human origins, which is a particular interest we share.
A Swerve into Metaphysics
Having given us his analysis of the chapter on human genetics (and leaving much of it out!), in the next section Swamidass makes a move into metaphysics and a new mode of speaking. He has moved away from scientific argumentation and into philosophy and theology. He says:
If God providentially governs all things, including the random cast of lots, he also providentially governs evolution. Still, science remains silent about God’s action. This silence, however, is because science is limited. [Emphasis added.]
First, if I grant that God providentially guides evolution (by whatever mechanism it happens), that does not mean science is silent about it. Swamidass is making a metaphysical move here. Several philosophical and theological theses have probably already been written about this kind of move. What Swamidass is doing is defining science as strictly limited to material and efficient causes. Science can make no inferences to immaterial causes by that definition. That is a move into philosophy. The chain of reasoning is continuous though: science provides the evidence, philosophy makes the inference as to cause, and theology indicates the nature of the cause. To say science is silent on God’s action is true if you mean science by itself. Yes, science is limited in that sense. But partnered with philosophy and theology, there is a great deal to be said.
Another metaphysical point: If God providentially guides evolution, that does not mean his guidance must be undetectable. The results of his guidance may be accessible to science. To say that they are not accessible is begging the question. That’s what the dispute is about — whether science can detect evidence of his guidance (with the above distinctions understood about the chain of reasoning involved).
Philosophy and Theology
Unfortunately, Swamidass’s engagement with the philosophy section is lacking, ignoring several chapters about methodological naturalism and taking a casual swipe at J.P. Moreland. (I am sure he would not have done so if he knew more about J.P. Moreland’s work.). He favorably mentions the chapter by C. John Collins, whose work he does know, as a segue to a discussion of his own genealogical Adam model. His genealogical Adam is also the reason he gives for approving Wayne Grudem’s introduction to the theology section, because Grudem advocates for the historicity of Adam. Grudem “aptly explains how poorly historical Adam theology was represented in a recent book by theistic evolutionists (p. 793),” says Swamidass. That book would be Adam and the Genome by Dennis Venema and Scot McKnight. Swamidass particularly disagrees with that book, as do I, but for different reasons.
I will end my review of Swamidass’s review by noting that it is quite spotty and selective, probably more than it should be, given the book’s size and scope. In particular, his choice of targets for criticism and for praise have a lot more to do with his particular agenda than the defects or merits of those whom he critiques.
Fortunately there are other reviews available. For science, I encourage reading David Snoke’s review in the same issue of Themelios, the online journal of the Gospel Coalition, where Swamidass’s review appears. Snoke, a physicist at the University of Pittsburgh, says “Theistic Evolution, no matter what its detractors may say, is clearly an important book, one that everyone with a serious interest in the evolution/intelligent design debate should read.”
For theology, I recommend the review by theologian Robert W. Yarbrough, of Covenant Theological Seminary, also in this issue, who says “As far as its biblical and theological coverage, this book must be adjudged a notable success in analyzing what is at stake, and what Scripture and Christian teaching continue to affirm, in the face of an important current debate.”
Image credit: DasWortgewand, via Pixabay.