Questions about Universal Ancestry

[Swamidass (Peaceful Science)]

@Jay313

God has surprised us many times! In this case, one could claim that God could arrange it so that Adam was the MRCA, which bumps up everything substantially. But if one is going to claim divine intervention to arrange a perfect MRCA, then all bets are off anyway, aren’t they?

We are not talking about MRCA or MRUGA. We are talking about UGA, which is what many people believe is taught in scripture. There is no claim about MRCA or MRUGA, but there is a claim (some people think) about UGA.

@Jay313

There are a lot of assumptions built into the model, of course. I think Rohde did yeoman’s work in building it, but he lays out numerous qualifications, especially on the migration numbers. Their lifespan was a problem, too, being based on early 20th-century figures (51.4 median age at death, as I recall). Median lifespans began to climb in the 20th century. Rohde should’ve chosen something from the 19th century, at the latest. (Median lifespans were fairly steady from antiquity until the 20th century. We take it for granted these days that we should live to at least 70.) A function of the math that I frankly don’t completely understand is that the MRCA fluctuated mainly with migration and wasn’t affected too much by other factors, but the IAP seemed to be affected by every variable. At least, that’s how I read it. Maybe someone with more math skills than me can decipher the prose …

Edit: I meant migration numbers, not patterns.

It is critical to understand the form of this argument they are making. They are arguing for a very difficult to believe conclusions, so they built the simulation with a whole range of barriers against recent UGA to (1) make it match the theory closer, (2) head off criticism that making the simulation more accurate would increase time to UGA (in fact it woudl decrease time), and (3) test the dependence on several variables. They find…

1. The only important variable is migration levels, nothing else they tweaked with had a strong effect.
2. Several features they did not model are all expected to decrease time to UGA.
3. IAP has high variance (so they cannot really find strong relationships to variables here), but MRUGA has lower variance, and nearly IAP has lowest variance. We care most about nearly IAP (and I explain why in the paper)

Now regarding the migration (and other details), they intentionally used a model designed to increase the time to UGA, so their final conclusion would be most strong.

1. Ony move per person per lifetime. No offspring before move.
2. Lifespans are high, just as you noted (which increases time to UGA), and has the effect of reducing migration (see #1)
3. Migration (except when demanded) was as individuals, no populations), which dramatically decreases intermixing.
4. Unrealistic high limits to migration between continents (e.g. only about 100 individuals / generation between europe and asia) so simulation would match their theoretical analysis, and to support their argument that more realistic migration would continue to support their conclusion (making UGAs more recent).
5. Ignoring known large scale migrations and trade routes (e.g. the Silk Road)
6. Assuming all couples are monogamous (unless absolutely required not to do so). So prostitution, infidelity, polygamy is all ignored, which would all dramatically reduce time to UGA (but also be much harder to model).
7. Ignored outbreeding (anti-incest taboo) cultural push entirely.
8. Used very low levels of migration, and it is only in even lower unrealistically low migration that MRUGA increases.
9. It ignored the effects of large scale population interactions like refugees and war which has an effect of totally mixing populations over long ranges.
10. It ignored entirely modern transportation like global travel in airplanes. (@Chris_Falter, does this help with your question?)

So each of these simplifications pushes the UGA estimate farther back in time. That is why this result has stood the test of time (13 years now). A more realistic simulation is possible, but because they biased their simulation so heavily against recent UGAs, a more realistic simulation is expected to reduced time to UGA. That is why no one has even attempted to challenge their results, and this is accepted by population geneticists.

Whatever skepticism you may have, I am certain they have faced more from more trained people. Getting a paper like this into Nature is no joke. It is very difficult.