Review: The Better Angels of our Nature, by Steven Pinker

Every so often a book comes along that literally changes the way I think about the world. Thomas Kuhn's Structure of Scientific Revolutions was probably the <cough> paradigm case for me.

Steven Pinker's The Better Angels of Our Nature: Why Violence Has Declined is one of these books.

His core thesis, supported by a huge array of data and documentation, is that violence has declined dramatically over time — not always smoothly, not always consistently, past performance does not guarantee future results — but there's a clear downward trend.

The kernel of insight is one of those really-obvious-in-retrospect ideas that changes perspective on a huge amount of history: if you look at conflicts and categorize them not by how many people they killed, but by how many people they killed per capita (i.e. divided by the world population at the time), then generally speaking, the fraction of people who die in armed conflicts has been getting smaller over time. A lot smaller.

However horrible the wars of the 20th Century were, they (a) have not been repeated in the 67 years since the end of World War II; (b) did not kill as huge a fraction of humanity as earlier wars.

As an eerie bit of synchronicity, I had been looking up World War I while watching Downton Abbey in January, and when looking at the Wikipedia page for WWI, I was surprised to find the line "It was the sixth-deadliest conflict in world history". The embedded link is to Wikipedia's List of wars and anthropogenic disasters by death toll, which includes a table, sorted by absolute numbers. WWII is at the top, followed by the An Lushan Rebellion (755-763), the Mongol Conquests, the Qing Dynasty Conquest of the Ming Dynasty, the Taiping Rebellion, and then WWI.

You can choose the column by which to sort the table, and if you choose "Percentage of the world population", the An Lushan Rebellion comes out on top with 14-15% of the world population. (!!!) WWII drops to #5 on that table (1.7% - 3.1% depending on estimates).

That was simply stunning to me. I'd never even heard of the An Lushan Rebellion. And 8 out of the top 10 in that table (sorted by percentage of population) involve (or are entirely encompassed by) China.

Pinker includes a similar table (on p. 195 of the hardcover), which also includes famines, genocides, and the Mideast and Atlantic Slave Trades, with similar results. He writes

First of all: had you even heard of all of them? (I hadn't.) Second, did you know that there were five wars and four atrocities before World War I that killed more people than that war? I suspect many readers will also be surprised to learn that of the twenty-one worst things that people have ever done to each other (that we know of), fourteen were in centuries before the 20th. And all of this pertains to absolute numbers. When you scale by population size, only one of the 20th century's atrocities even makes the top ten.

 

Even if Pinker had stopped here, this would be a great service to the discussion of violence: what a lot of people think about the level of violence in the 20th century is simply false. But he does a lot more than that.

He marshals plenty of evidence from multiple disciplines showing that the death rate from violence in hunter-gatherer (and other non-state) societies was and is dramatically higher than those with a strong state. Hunter-gatherer societies don't go to all-out war as often (if at all), but they fight a lot, and when each fight leaves one or two dead, out of groups of 100-150, the per capita death toll is huge.

Basically, he argues that Hobbes was right, and Rousseau was wrong, at least as it pertains to the narrative underpinnings of their respective political philosophies. On top of everything else, I have a renewed appreciation for Hobbes' Leviathan.

Along the way, Pinker takes the reader on a tour of historical attitudes toward and practices of violence: duels, the original source for "cutting off one's nose to spite one's face" (cutting off people's noses for what we would consider trivial offenses was quite common in Europe at one time), tortures of the most gruesome sort, and conduct in battle and warfare that is almost literally unthinkable now. A lot of what he catalogues is not for the faint of heart.

He then follows through with sections titled "The Pacification Process", "The Civilizing Process" (which covers the decline in homicide in Europe — by a factor of 100), "The Humanitarian Revolution", "The Long Peace" (about the surprising lack of a Third World War), "The New Peace", "The Rights Revolutions", "Inner Demons", "Better Angels", and "On Angel's Wings". He brings in data and ideas from a wide variety of disciplines and a large array of researchers. Some ideas are more speculative and/or less convincing than others (but Pinker is careful to label which of these ideas he thinks are conclusive and which merely point to the need for further research).

His argument in many of these sections is that in many cases, people became less violent due to a self-reinforcing civilizing process: violence and murder became seen as vulgar and declassé; reformers (abolitionists, prison reformers, Charles Dickens writing about poor houses, and so on) pushed society to change. It didn't happen overnight, but it happened, and happened independently of most technological changes that many people have suggested were the cause of the declines in violence. Basically, people can change, and people can, collectively, increase self-control, de-emphasize being quick to respond to slights on their honor, and change their views of violence. Culture is not written in stone, and changes over time, and has changed in ways that has reduced violence by several orders of magnitude.

There were many topics that Pinker raises, about which I felt compelled to think further, or to do more research about because they were so interesting. A big example is the legitimacy of state/police power (and its monopoly on violence). This is especially interesting with regard to the correlation between levels of violence and a "culture of honor" — and how that seems to match up with locations where state power is not legitimized (either because it has not reached there [e.g. Appalachia in the 19th century] or has been deliberately neglectful [e.g. poor/minority areas of cities from the 1970s to early 1990s and up to the current time], or is corrupt, or outright tyrannical). Connections with the successes of Community Policing seem fairly direct to me.

One other thought: this book made yet even more happy that I don't live in a medieval fantasy world. (I like to watch Game of Thrones; I don't want to live there.)

Even when Pinker infurates me, he's always interesting. A few of his riding-his-hobbyhorse moments seem to have slipped through the editing process, but overall he's actually fairly restrained. I don't buy everything he says, but he's not just making stuff up. (And, honestly, I think the sheer scale of historical cruelty and horror, [e.g. the circumstances which inspired so many social movements] made him re-evaluate some of his hobby-horses in the light of historical perspective.)

Overall, I think this is a hugely important book. Almost everything I've read since I finished it a couple of months ago has been interacting with it in my mind; being interpreted in the light of the ideas it stirred up.

I found it very easy to read (other than the gruesome parts). His future-looking sections at the end of the book are somewhat less engrossing than the earlier, more data-driven sections, but still quite interesting.

Bottom line: highly recommended.

Places to purchase The Better Angels of Our Nature: Amazon BN Powell's Harvard Bookstore

---

Addendum: I'm leaving this one gripe at the bottom since it doesn't really fit above.

The biggest single problem I have with the book is his discussion of the surge in violence in the 1960s and 1970s in the United States, and the tail-off of that surge in the 1990s (looked at on a graph from 1950 to 2010, it's a big spike — but in the larger historical trends, it is a smallish blip). This discussion is (perhaps by necessity) more nebulous than a lot of the rest of the book, but there are several ideas which he simply doesn't mention, which I had encountered simply by being a reasonably-well-read reader. The one whose absence surprised me the most — given Pinker's background — is lead (especially in leaded gasoline).

Lead and lead compounds are potent neurotoxins in and of themselves, but they are also causes of more subtle developmental problems, especially in the brain and central nervous system. One reasonably obvious effect is to delay or inhibit full development of the prefontal cortex, the part of the brain that performs "executive function" -- impulse control, long-range planning, and basically everything that keeps people from being idiots. Given how much of the rest of the book is about how strengthening impulse control helps create a less violent society, it seems obvious to talk about how lead and lead compounds would affect this.

Tetraethyl lead was introduced as a gasoline anti-"knock" additive in the mid-1920s, and was finally banned in the US in the mid-1990s. Its use in the US had been on the wane since the early 1980s, though, as alternative anti-knock formulations came into use, and as catalytic converters (to meet emission standards) became common. The rise and decline in violence in the US, from the 1960s through the 1990s roughly mirrors the use of leaded gasoline and lead paint. Even if it isn't all of the story, it is one with a clear biological basis, and is one that people have suggested and documented. It deeply surprises me that Pinker didn't mention this, even in passing.

Examples: "Reduced cognitive abilities in lead-exposed men"; "Scientist taught world to get the lead out"; and two articles about tetraethyl lead by Deborah Blum, based on her research for The Poisoner's Handbook.