Ch. 13: The evolution of altruism
Aaron Swartz
networks of landlords, religious orders, leagues of pirates, warrior
bands, and many other forms of governance" (Charles Tilly) by World
War I, only the nation-state remained. Why? Because nation-states are
the form of governance most conducive to collecting taxes and taxes
allow you to raise an army without the support of the population. Thus
nation-states won the wars.
As a result, the nation-state conquered the world. Not because it was
better for its subjects, but because it as better for its rulers. They
promoted norms like tax compliance, national loyalty, respect for
property rights, and monogamy, which further aided their projects.
These norms may have aided the group but were less beneficial for the
individual.
Group conflict seems to be the usual way norms evolve and spread. How
does it work?
In chapter 7 we saw how tit-for-tat could allow cooperation to be a
mutual best response. But this hardly is enough to explain all the
altruism we see in humans. Humans are altruistic even when they know
they won't see the other party again -- early humans could ensure this
since before agriculture their band of foragers could simply run away.
The Folk Theorem shows that all sorts of results are equilibria, not
just altruism. And tit-for-tat doesn't really work in large groups,
since small chance events can cause the whole group's trust to
unravel.
So why are we so nice to strangers?
Darwin argued the solution was _group selection_: "a tribe possessing
[courage, sympathy, and unselfishness] in a high degree would spread
and be victorious over other tribes." But modern Darwinians insisted
evolution only operated on the individual level and that group
selection could never outweigh individual selection. But we can show
otherwise.
Imagine an altruistic trait A, say bravery in defense of the group,
which costs the individual c and confers a benefit b on a random
member of the group. These are defined so a group composed entirely of
altruists produces b - c more descendants than a group with no
altruists. b - c > 0 so altruism is group-beneficial. But by the same
token, it's damaging for the individuals.
If the group is entirely homogenous, obviously only between-group
selection is at work and the trait will reproduce. If there's only one
group, obviously only within-group selection operates and the trait
will die out. In between, then, it's a matter of degree: it depends on
how beneficial the trait is and how homogenous the group is. Imagine
one group with 1/4 altruists and one group with 3/4 altruists. If the
latter outcompetes the former, the number of altruists increases as a
whole, even if it's decreasing within every group.
But if altruism is decreasing within each group, how did we end up
with a group of 3/4 altruists to begin with? Well, human bands of
foragers are quite small and frequently divide, so it's quite possible
random chance will bring a large number of altruists together.
In simulations, even if you start with no altruists (they're only
produced by mutation), we see that the majority of the population
eventually becomes altruistic. Of course, altruism in this model only
lasts as long as there is war -- an outbreak of peace leads to a sharp
reduction.
In real life altruistic tribes are less likely to break up and when
they do break up, the altruists tend to clump together, making
high-altruist groups even more likely to occur. And while good data is
hard to find, experts suggest that increasing population density led
to increased conflict between human bands in the Pleistocene. There
was also probably significant migration. All of this suggests our
simulation is a plausible explanation of the human story.
Next time: We begin part 4, Conclusion, with a discussion of economic
governance.