Keep in mind that there are two very different questions one could ask. We could ask, on the one hand, whether "human nature" has improved, meaning that humans have become fundamentally more moral and genuinely more interested in the good of other human beings. We could also ask, on the other hand, whether institutions have evolved to constrain certain evil tendencies, thus producing a more benevolent societies. I suspect a positive answer to the latter question is far more likely than for the former.
Even so, even if private violence has decreased over time, this does not mean that violence in general has. I appreciated this review by Tyler Cowen, in which he makes this exact point:
Another hypothesis is to see modern violence as lower, especially in the private sphere, because the state is much more powerful. Could this book have been titled The Nationalization of Violence? But nationalization does not mean that violence goes away, especially at the most macro levels. In a variant on my point above, one way of describing the observed trend is “less frequent violent outbursts, but more deadlier outbursts when they come.” Both greater wealth (weapons are more destructive, and thus used less often, and there is a desire to preserve wealth) and the nationalization of violence point toward that pattern. That would help explain why the two World Wars, Stalin, Chairman Mao, and the Holocaust, all came not so long ago, despite a (supposed) trend toward greater peacefulness. Those are hard data points for Pinker to get around, no matter how he tries.And here is another very good point:
When doing the statistics, one key issue is how to measure violence. Pinker often favors “per capita” measures, but I am not so sure. I might prefer a weighted average of per capita and “absolute quantity of violence” measures. Killing six million Jews in the Holocaust is not, in my view, “half as violent” if global population is twice as high. Once you toss in the absolute measures with the per capita measures, the long-term trends are not nearly as favorable as Pinker suggests.To put it cynically, if I kill someone so that I can feed my children, it's murder; if America carpet bombs innocent civilians, it's national defense.
One more thought, which is sadly all too easy to dismiss by a great majority of the mainstream media: If not for the ideological claim advanced in the latter part of the 20th century that unborn children are not human, it would be very easy to demonstrate that violence has greatly increased in our generation. In our country alone, there have been 53,000,000 abortions since 1973. This cannot be counted as part of Pinker's data, because he cannot accept it as violence. But it's hard for me to accept that 53,000,000 bloody corpses are a sign of a more peaceful society.
Still, I think we do a disservice to society by always speaking in dreary terms, as if civilization never makes any moral progress. Slavery was once commonly accepted as a legitimate practice virtually everywhere around the world. Now it is accepted almost nowhere. That's progress.
The gate is narrow and the way is hard that leads to life. Moral progress is far from inevitable, but neither is it futile. It's good to be reminded of both sides; I think Pinker probably dwells on one more than the other. Still, it's probably a good read. Maybe I'll pick it up one of these days.