History of Violence

I have my annual health check tomorrow morning, followed by a two-day project meeting in Kyoto so I'll be rather busy until the end of the week¹. But just so you won't feel a lack of long-winded reading, here is an excellent (and long!) essay by Steven Pinker on A History of Violence.

If you don't have the entire morning to set aside for the essay you can get the gist simply by looking at the collection of graphs scattered throughout the piece. Just be careful that you check the scales and units of measurement. Many graphs are logarithmic; it lets you see variation at both the low and high ends of the range, but it tends to understate the magnitude of the overall change. Also, he frequently plots rates — incidents of violence per 100 000 people — not absolute number of incidents. They reflect your risk of violence, which is probably what you're mostly interested in, but says nothing about the absolute number of events.

Pinker argues, quite persuasively, that levels of violence around the world is at a historical low, no matter of criterion for violence you take, and almost no matter what timeframe. Long time scale or short, we're less violent to each other than ever before. To illustrate, let me show you the data from one of his sources, Manuel Eisner² about overall level of homicide in Europe from early middle ages to today. Instead of a log plot like Pinker does, let's plot it on a linear scale to clearly see what the overall trend looks like:

Homicide rates in Europe from the 1300's to the 20th century. Rate is in victims per 100 000 incidents, with one data point per century. Data from Eisner (2003).

The data is very sketchy for the distant past of course, but overall the rate of homicide has dropped some 40-fold over time. The risk, today, for us to be killed by somebody else is only a fracťion of the historical average. And Pinker argues that the same thing is happening at shorter timescales as well as longer, and not just for murder but for war deaths, rape, child abuse, racially motivated violence and so on and so on. 

Then why do people feel so unsafe today? And why do Japanese — living in what is arguably one of the least violent societies in history — seem to be at least as afraid of violence as people in far less settled areas? Perhaps people feel unsafe precisely because violence is becoming so rare. As it becomes rarer, each incident becomes more newsworthy, grabs more media attention and focuses peoples minds more on it than before.

It's the same as with car accidents and airplane accidents: You're in much more danger in a car than in an airplane, and there's many, many more deaths from car accidents than plane crashes. But car crashes are so common that they don't get much attention, while the rare airplane accident gets big headlines. We estimate the risk to ourselves by the number of incidents we hear about in the news, not by the number of incidents that actually happen. So, paradoxically, we end up feeling airplanes are less safe than cars, in part because the reality is the opposite.

It's not just the rarity, though. We hear about more violent incidents than before simply because we get much more news about anything than we did before. If a man got beaten and robbed in a bar in Yokohama a century ago it's unlikely anybody in the rest of the country would have heard of it. If it happens today there will be a news item about it in every news outlet in Japan. Robberies where much more commonplace and less newsworthy a century ago, but news coverage at the time was nothing like what we have today.

And the goal posts have moved. We accept much less violence than we used to. Some type of violence that we rightly condemn today — beating your spouse or children for instance — were legal some decades ago; they were approved and even recommended in some cases. We're safer from violence overall, so our tolerance for violence is decreasing in turn.

So, we take acts of violence more seriously than before; we hear about violent crimes from a far larger area than we used to; and the very fact that violence is becoming rarer means each violent act gets more media coverage. These factors all mean that people become more worried about being victims of violence, and believe violence has been increasing, even as the reality is the complete opposite.

--

#1 If I'm so busy, why do I spend most of my lunch hour writing about this essay rather than hurry along with my project? I do need an occasional break of course; also, for some reason other people's research always seem more interesting to me than my own.

#2 Eisner (2003), "Long-Term Historical Trends in Violent Crime", Crime & Just. 83 (2003) - webpage here, and direct link to the PDF from the author here. It's a very readable paper, even for a complete non-specialist like myself.