The Better Angels of Our Nature: Why Violence Has Declined

Steven Pinker’s The Better Angels of Our Nature is one of the most ambitious books in recent memory. Encyclopedic in scope, it sets out to demonstrate the bold and counterintuitive thesis that violence has declined over the millennia. Although Pinker has scientific roots in evolutionary psychology, this is not a book arguing that the decline is the result of natural selection.

Pinker identifies and summarizes six trends away from violence: the pacification process from warlike anarchy of hunter-gatherers to agricultural settlements, the civilizing process of organized states (“the Leviathan”) with centralized authority, the “humanitarian revolution” against socially sanctioned violence and cruelty, “the long peace” after World War II to the present, “the new peace” with a decline in large-scale armed violence since the end of the Cold War, and the “rights revolution” now extending past humans to animal rights.

Documenting these trends takes Pinker on an intellectual adventure far beyond his already broad and deep academic expertise in psychology and social science. The scale of his superhuman enterprise reminds me of my favorite undergraduate professor, the great Pitirim Sorokin, the Russian-born Harvard sociologist whose “empirical” work—like Pinker’s—crossed all academic, geographic, and historic boundaries. The problem with Sorokin’s monumental efforts, as with Pinker’s, is that the empiricism on which their theories are based is an amalgamation of other people’s data and other people’s catalogs of data, some of which seem unreliable or worse. It would not be completely unfair to Pinker to suggest that he prefers bad data to no data.

I have spent weeks reading and rereading The Better Angels of Our Nature. It covers every theory of violence I have studied, researched, and taught in my classes for more than four decades. I disagree with Pinker on some matters, typically because he accepts the findings of contemporary authorities in several areas that I reject, such as current theories about violence based on primatology, forensic archeology, evolutionary psychology, and cross-cultural anthropology. In all these areas I believe the limited and controversial evidence does not sustain the conclusions drawn from them. But in each of our disagreements I would emphasize that Pinker represents the prevailing views of mainstream academia, and I am the skeptical outlier. Nor does Pinker really need the supposed evidence about the lethal raids of chimpanzees or the warlike behavior of the Yanomano tribe to make his Hobbesian point that in the early millennia, “life was . . . nasty, brutish and short.”

There is much on these pages that is intellectually awe inspiring: the reach of Pinker’s mind, his mastery of so many disciplines, and the clarity and grace of his writing. Bill Gates of Microsoft describes it as one of the best books he has ever read. It seems to combine hope and social science. Better Angels is the 21st century’s optimistic answer to Freud’s fatalistic Civilization and Its Discontents.

Nevertheless, I have certain misgivings that I want to share with readers. Let me begin with a quibble about the title The Better Angels of Our Nature. The phrase is from the last words of Abraham Lincoln’s first inaugural address. It may well be Lincoln’s most politically disappointing speech. In it he pledges not to interfere with the institution of slavery and beseeches the Southern states not to secede. His hope is not that our “better angels” will end slavery but that they will prevent secession.

Beyond that quibble I have certain reservations about the certitude of Pinker’s thesis. He offers us dozens of data sets, all of which are indicative of the decline of violence over the centuries. Many are compelling, such as concern the decline of homicides in Europe from 1100 to 1900. However, a few are dubious. The first data set that raised my hackles is his list of atrocities. Pinker reports that “The worst atrocity of all time was the An Lushan Revolt and Civil War” during China’s Tang Dynasty in the eighth century. Pinker retrieved the data from a blog, and other academics have apparently relied on it. Unfortunately, after Pinker published his list, the blogger decided to downsize that atrocity by more than 60%. According to Pinker’s account, the blogger, who has no academic credentials, takes the median of “the figures cited in a large number of histories” to come up with an estimate. But no such calculation can explain how this change came about.

I must emphasize that the blogger’s numbers were exactly as Pinker quoted them when he accessed the Web site. But that makes my point that Pinker has relied on an unreliable data source to construct this data set. There is for me a huge question mark about the size of all the atrocities on this list. These death tolls “include not just deaths on the battlefield but indirect deaths of civilians from starvation and disease.” Who was counting deaths by starvation and disease in China in the eighth century? Authorities still argue about the death toll in the 20th century’s First World War and the flu epidemic that followed it. For me most of the atrocity data are pseudo-statistics—bad data being preferred to no data. Pinker alludes to all of these problems of reliability, including the record keeping of innumerate cultures, but then forges ahead with the blogger.

And there is a systematic correction in all of Pinker’s historical calculations that is also problematic. He scales “death tolls by the world population at the time.” To use the blogger example again, the site estimates that eight million people died in the fall of Rome; when Pinker scales that number to the world population, it becomes 105 million and the fifth worst atrocity in the history of the world. All of this manipulation of dubious data is meant to convince us that violence is going down. Pinker does not need these pseudo-statistics; most of the data sets—at least on my reading—seem to be convincing. Despite minor misgivings, I find it impossible to reach any conclusion other than to urge mental health professionals to read this book. Pinker is so knowledgeable and so charming that one ends up hoping that he is right.