Genetic Selection for War in Human Populations
Evolution takes place at the level of the gene. A particular gene increases in frequency if it improves the chance of the gene existing in the next generation.
We are familiar with behavior (ultimately from genes) where parents take awful risks, sometimes losing their lives to save their children. In an environment where such events were common, the gene(s) for doing so would become more common if (on average) the self-sacrifice of a parent saved three or more of their children. The parent has one copy; each of the kids has a 50% chance of having the same gene. On average more copies, 1.5, survive the event if a parent dies than if all three children die. This is an immense simplification; life is far more complex, but the idea should be clear as to the origin of this behavior trait.
If it is not, you might look up Hamilton’s rule,
“. . . a hypothetical gene that prompts behaviour which enhances the fitness of relatives but lowers that of the individual displaying the behaviour, may nonetheless increase in frequency, because relatives often carry the same gene.”
To analyze the spread of genes that lead to war behavior, we need to generate
a model from the “viewpoint” of such genes in a typical warrior 50,000 or 100,000 years ago. Tribes in those days were limited in population by the ability of the environment to provide food. On average, they were around the same size. If two tribes fought, each had an equal chance of prevailing. In this model, the winners typically killed all the adult losers and their male children. The winners incorporated the female children of the losers into the winner’s tribe as wives. (See The Book of Numbers, Chapter 31, verses 7-18, for an account of the aftermath of a war in Biblical times.)
Wars come about due to a resource crisis. For this model, we will assume 50% of the tribe will starve in the crisis, the alternative being to attack neighbors and try to take their resources. How often such events happened is not part of the model, but they probably averaged about once a generation.
Turning to the mathematical analysis, the warrior himself has one gene copy. He typically had six children, half males and half females. (From what we know, that’s about the minimum for a stable population in Stone Age times.) Each child has a 1/2 chance of carrying the gene(s) for war behavior. (The model is not very sensitive to the number of children.)
If war behavior is to be evolutionarily favored, the count of gene copies needs to be higher (on average) after the war vs. starving in place.
For the winners, the gene count for a warrior is four, one for himself plus 1/2 times 6 children for 4. Fifty percent starvation reduces this to two copies. This makes two gene copies the number to exceed if the behavior for war is to become more common than starving in place.
For the losers, the gene count is 1.5 from the female children that the winners incorporated into their tribe. That makes the average count of genes per warrior after a war (4+1.5)/2 (using a 50% chance of winning). Or 2.75 (for war)/2 (starvation) means that going to war is about 37% better from the gene’s viewpoint than starving (in this simple model, of course). That’s a big number, indicating strong selection if this model is close to reality.
The driver for this model is starvation due to a resource crisis, ultimately due to population growth and environmental variation (mostly weather). Does going to war without looming starvation make sense? No. Going to war gives an average gene copy remaining of 2.75 vs. 4 for no war, or 4 (no war) /2.75 (war) making the selection against going to war about 45% per event (or rather nonevent). That too is a big number, indicating strong selection for not going to war unless the alternative for genes is worse.
This places the detection of looming starvation under intense selection to get it right. (A challenging cognitive task.)
How does a tribe go from individual detection of a bleak future to mass attacking another tribe?
It’s obvious that attacking another tribe one at a time is a nearly sure way to be killed. We have an example: chimps make war on neighboring groups. When they do, all the males in the group attack at once, in fact, when they do boundary patrols they go in groups and kill any lone members they encounter from neighboring groups. “War mode” in chimps seems to be on all the time. It seems to never be on in bonobos (why is a good question).
For humans, “war mode” is on some of the time for the gene-based reasons discussed above. Most of the time human groups are on good terms with neighbors, swapping marriage partners with them. However, in times of stress (facing a bleak future, starvation) memes circulate that dehumanize the target tribe and make the warriors willing to attempt to destroy them. This last bit on memes is speculative, though there is a good deal of evidence that there is a connection between deteriorating economic conditions and xenophobic memes becoming common in a population.
Perhaps the most obvious recent historical example is the rise of Nazism in the economic disaster of Germany in the 1930.
Reference
“I don’t know anyone that wants to go to war with Canada.”
Some decades ago the Canadian were concerned enough about the possibility of the US invading that they trained people to infiltrate the US and engage in sabotage. I knew one of the people who trained for this role.
Although I feel confident about the model, you do have to be careful about extrapolating stone age evolution to the present day and conditions.
" Humans are susceptible to propaganda"
I.e., memes. In the build-up to war, particularly xenophobic memes. But how easy these memes spread depends greatly on how bleak the future seems.
