Bayesian Investor Blog

Book review: The Better Angels of Our Nature: Why Violence Has Declined, by Steven Pinker.

Pinker provides a cautiously optimistic view of the dramatic reduction in violence over the past few centuries. He has tied together a wide variety of violent behavior (from genocide to cruelty to animals) into one broad pattern of people becoming more civilized.

He mostly covers the west, and I wished for more about the trends in the mideast and the more peaceful parts of asia. And why does he avoid commenting on the reported homicide rate for Somalia that’s less than the US – does he have reject the data as inaccurate, or does he want to ignore evidence that’s inconsistent with his claim that the rise of leviathan reduced crime?

The substance of the book is less reassuring than a casual glance would suggest. In particular, the section on the power law distribution of war deaths shows that the model he uses says the expected number of war deaths is infinite, apparently due to small probabilities of really destructive wars. It’s easy to overlook this disturbing implication if you aren’t reading carefully. I’m disappointed that Pinker didn’t talk about this more – how is it altered if we try to incorporate the finiteness of the available population? Should we focus a lot of our attention on avoiding unlikely megawars?

He cautiously speculates about possible causes (increased global communications, trade, democracy, self-control, reason, feminization, education, and intelligence via the Flynn effect).

If you’re an ideologue looking for an excuse to be offended, you’re fairly sure to find one in this book.

Pinker exaggerates the evidence for violence among hunter-gatherers (see the book Sex at Dawn for the other side of this debate). And the evidence Pinker does present is quite consistent with the hypothesis that violence increased when people switched from hunter-gatherers to stationary farmers.

The book is very long, and that’s not due to a simple desire to be thorough. Much of the evidence seems selected more for vividness and memorability (e.g. his evidence from fairy tales – not completely frivolous, but insignificant enough evidence that he wouldn’t have put it in a peer-reviewed article).

I didn’t learn enough from it to justify the time spent, but the ideas in it deserve to be known more widely.