One claim made by Craig that we have not yet addressed in detail is this: that population genetics models have been shown to significantly overestimate the ancestral sizes of real populations where their real genetic history is known. The main example that Craig proffers in support of this claim concerns a scientific study examining a population of invasive sheep on a remote island in the southern Indian Ocean.
It’s little wonder that this study caught the attention of those wanting to hold to humanity descending uniquely from a single ancestral couple. In the 1950s, a population of domestic mouflon sheep was founded on Haute Island in the Kerguelen archipelago with a single ram and a single ewe, both introduced as lambs. The extreme remoteness of the Kerguelen archipelago (over 3000 kilometers to the next inhabited island!), coupled with the continuous monitoring the islands have had by French scientists, ensures that no other sheep have been introduced since the population was founded. As such, this population genuinely has the sheep equivalent of a genetic “Adam and Eve” in the sense that Craig maintains for humanity. Even more exciting, though, was the finding that the sheep on this island have higher levels of heterozygosity – which we will discuss in detail later but for now can think of as one measure of “genetic diversity” – than a certain mathematical model predicts. For an apologetics-minded approach, it seems tailor-made: population genetics models overestimate the heterozygosity of a population known to be founded by only two individuals – and thus measurements of human population sizes might similarly be overestimates, and we can hold on to a literal Adam and Eve who were our sole genetic progenitors! This basic argument first appears in the Christian anti-evolutionary literature in 2010, in the work of Reasons to Believe (RTB). Since then, it has appeared in numerous places online, though Craig is basing his argument primarily on RTB sources. As we will see, his argument is not a valid one. It is based on a significant misunderstanding of the study in question, and population genetics in general – but it will take some effort to explain why this is so.
Based on this study, Craig makes the following four claims:
- The study shows that the Kerguelen sheep have higher heterozygosity than expected, and show an increase in heterozygosity over time.
This claim is true, but Craig misunderstands why it is true, as his second claim shows:
- The observed increase in heterozygosity is attributable to an increased mutation rate driven by natural selection.
This (erroneous) claim is repeated several times as Craig discusses this study. Some examples include:
So natural selection actually accelerates the rise of genetic diversity because it has survival value in the struggle for survival.
In other words, had they not known that there were originally only two sheep placed on that island, looking at the genetic diversity exhibited by the present sheep using the mathematical models they would have over estimated the minimal size that that population would have had at any time in the past because the models did not take account of the accelerated rates of genetic mutation that were driven by natural selection.
This claim is simply false: natural selection does not, and cannot, increase the rate of mutation, however beneficial an increase in mutation frequency might be. Natural selection acts only on genetic variation that already exists in a population. Craig has confused an increase in heterozygosity over time with an increase inmutation rate over time. In fact, as we will see, an increase in heterozygosity over time does not require any new mutations at all.
- Natural selection therefore may have driven an increased mutation rate in the human lineage, and an appeal to divine intervention for increasing human genetic diversity may not be necessary.
- If so, then population genetics models may underestimate human ancestral population sizes because they fail to account for natural selection. If so, then the “traditional view” that we descend uniquely from an ancestral couple remains defensible.
You might recall that Craig has been attempting to argue for an increased mutation rate in the human lineage in order to explain how humans are so genetically diverse today (leaving aside our discussion that a mere increase in mutation rate is not adequate to explain the pattern of variation we observe in humans). He clearly has this issue in mind when invoking the mouflon study:
If it is the case that natural selection can drive the increase in genetic diversity then that calls into question the assumption that the mutation rates have been constant over time for humanity, and hence it calls into question the population estimates based on that assumption.
The Kerguelen sheep study thus does double duty for Craig: it offers, in his mind, a natural explanation for increased mutation rates, and simultaneously throws human population genetics models into doubt. Since alternative views that accept that humans descending from a population raise theological challenges for him, he sees this evidence as reason not to be “forced” to accept such a view:
… I am just saying that realize that all of these alternatives have really interesting and unsettling theological reverberations that we need to be aware of. So that is an alternative that some people have suggested. What I am arguing right now is I am not sure we are forced to that alternative because I am not convinced that the evidence is inconsistent with there being an original historical human pair. The bottleneck got so small that it was just two people – Adam and Eve – and if you then imagine accelerated rates of mutation which would be either by divine intervention or just naturally, like the sheep on Haute Island, then that is entirely consistent with the evidence that we have today.
In order to understand why Craig is mistaken, we will have to understand exactly what heterozygosity is, how it relates to genetic diversity, and how natural selection has shaped it over time in the Kerguelen mouflon population. Next, we will start with these key issues.