The following poll suggests that a majority of my Twitter followers think war will decline; in the next 80 years we won’t see a 15 year period with a war death rate above the median level we’ve see over the last four centuries:

What is the highest rate of war deaths (averaged over 15yr period, as in graph from link above) the world will see by 2100? Below yr 2000 level, median level seen over last 400 yrs, world wars level, or even higher?

— Robin Hanson (@robinhanson) July 24, 2019

To predict a big deviation from the simple historical trend, one needs some sort of basis in theory. Alas, the theory arguments that I’ve heard re war optimism seem quite inadequate. I thus suspect much wishful thinking here.

For example, some say the world economy today is too interdependent for war. But interdependent economies have long gone to war. Consider the world wars in Europe, or the American civil war. Some say that we don’t risk war because it is very destructive of complex fragile physical capital and infrastructure. But while such capital was indeed destroyed during the world wars, the places most hurt rebounded quickly, as they had good institutional and human capital.

Some note that international alliances make war less likely between alliance partners. But they make war more likely between alliances. Some suggest that better info tells us more about rivals today, and so we are less likely to misjudge rival abilities and motives. But there still seems plenty of room for errors here as “brinkmanship” is a key dynamic. Also, this doesn’t prevent powers from investing in war abilities to gain advantages via credible threats of war.

Some point to a reduced willingness by winners to gain concrete advantages via the ancient strategies of raping and enslaving losers, and demanding great tribute. But we still manage to find many other motives for war, and there’s no fundamental obstacles to reviving ancient strategies; tribute is still quite feasible, as is slavery. Also, the peak war periods so far have been associated with ideology battles, and we still have plenty of those.

Some say nuclear weapons have made small wars harder. But that is only between pairs of nations both of which have nukes, which isn’t most nation pairs. Pairs of nations with nukes can still fight big wars, there are more such pairs today than before, over 80 years there’s plenty of time for some pair to pick a fight, and nuke wars casualties may be enormous.

I suspect that many are relying on modern propaganda on our moral superiority over our ancestors. But while we mostly count humans of the mid twentieth century as morally superior to humans from prior centuries, that was the period of peak war mortality.

I also suspect that many are drawing conclusions about war from long term trends regarding other forms of violence, as in slavery, crime, and personal relations, as well as from apparently lower public tolerance for war deaths and overall apparent disapproval and reluctance regarding war. But just before World War I we had also seen such trends:

Then, as now, Europe had lived through a long period of relative peace, … rapid progress … had given humanity a sense of shared interests that precluded war, … world leaders scarcely believed a global conflagration was possible. (more)

The world is vast, eighty years is a long time, and the number of possible global social & diplomatic scenarios over such period is vast. So it seems crazy to base predictions on future war rates on inside view calculations from particular current stances, deals, or inclinations. The raw historical record, and its large long-term fluctuations, should weigh heavily on our minds.