Singularity Summit 2012

A History of Violence


For more transcripts, videos and audio of Singularity Summit talks visit

Speaker: Stephen Pinker

Transcriber(s): Ethan Dickinson

Moderator: Professor Steven Pinker is next. He is the Johnstone Family Professor of Psychology at Harvard University. Dr. Pinker's research on visual cognition and psycholinguistics has won him many scientific accolades, including from the National Academy of Sciences, and the Royal Institution of Great Britain. In 2004 he was named one of TIME magazine's "100 Most Influential Scientists and Thinkers in the World." He is the chair of American Heritage Dictionary's Usage Panel, and he writes frequently for both The New Republic and The New York Times. He has twice been a finalist for the Pulitzer Prize, and Bill Gates has called his latest book, "The Better Angels of Our Nature," "A contribution not just to historical scholarship, but to the world." Please join me in extending a warm welcome to Professor Steven Pinker.

Steven Pinker: Thank you very much.

Believe it or not, and I know most people do not, violence has been in decline for long stretches of time, and today we may be living in the most peaceful era in our species' existence. The decline of violence has not been steady, it has not brought rates of violence down to zero, and it is not guaranteed to continue. But I hope to persuade you that it's a persistent historical development visible on scales from millennia to years, from wars and genocides to the treatment of children and animals.

I'm going to walk you through six major historical declines of violence, try to identify their immediate causes, that is particular historical events of the era, and then tie them together in terms of their ultimate causes, namely general historical forces interacting with constant human nature.

The first decline of violence I call the "Pacification Process." Until about 6,000 years ago, humans everywhere lived in anarchy, without central government. What was life like in this "state of nature?"

This is a question that thinkers have pondered for centuries. Thomas Hobbes in 1651 famously wrote that in the state of nature, "The life of man is solitary, poor, nasty, brutish, and short." A century later, Jean-Jacques Rousseau countered that, "Nothing can be more gentle than him in his primitive state."

Now these men were speculating from the armchair, they had no idea what life was like in a state of nature. But today we can do better, because there are two sources of evidence on rates of violence in nonstate societies. I'm going to mention only one of them, and that is forensic archaeology. Think of this as "CSI: Paleolithic." Namely, what proportion of prehistoric skeletons show signs of violent trauma, like bashed in skulls, decapitations, arrowheads embedded in bones, or mummies found with ropes around their necks.

Here I've assembled 19 estimates. They span quite a range, but they average out to 15 percent. 15 percent of prehistoric skeletons show signs of violent trauma. Just for the sake of comparison, let's pick some modern societies that are notorious for their high rates of violence, such as the 20th century Europe and the United States combined had a rate of war death of two-thirds of 1 percent. If you try to get an upper-bound estimate, the highest plausible estimate of the rate of violent death in the 20th century from war, genocide, and man-made famine, it comes in at about 3 percent. And here is the bar for the world in the 21st century so far. It is less than one pixel high, representing a rate of 3/100ths of 1 percent.

The immediate cause was the rise and expansion of states, giving rise to the "paxes" or peaces that history students are taught. The "Pax Romana," "Pax Islamica," "Pax Hispanica," and so on. When early kings and emperors established control over a territory, they tended to try to stamp out tribal raiding and feuding, not because they had a benevolent interest in the welfare of their citizens, but because internal squabbles are a nuisance to them. Just as a farmer has an incentive to prevent his cattle from killing each other, not because he necessarily cares about the cattle but he'd rather keep them alive for himself, so the early kings and emperors wanted to keep their subjects alive to supply them with soldiers and taxes and slaves.

The second decline of violence can be appreciated by examining this woodcut showing a day in the life in the Middle Ages. And the process that brought this degree of mayhem down has been called the "Civilizing Process."

In many parts of Europe, homicide statistics go back 800 years, and historical criminologists have plotted them over time. This graph goes from 1200 to the year 2000, and it plots homicide rates on a logarithmic scale from a tenth of a homicide per 100,000 per year, up to 100 per 100,000 per year. It shows there has been a massive decline in the rate of homicide in England, so that a contemporary Englishman has about 1/35 the chance of being murdered as his medieval ancestor.

This is true not just in England, but in every country for which data exists, such as Italy, the Netherlands, Germany, Switzerland and Scandinavia. The red line here shows the average of those five regions, and for the sake of comparison I've also plotted the rate for nonstate societies. So this gap here [points to nonstate societies average, then the early average of the five regions] is what I've been calling the Pacification Process, that further decline I've been calling the Civilizing Process.

The term comes from the classic book by the sociologist Norbert Elias, who argued that in the transition from the Middle Ages to modernity, there was consolidation of central states and kingdoms from the patchwork of baronies and principalities and duchies in feudal Europe. With it, criminal justice was nationalized, and the constant feuding, brigandage, and warlording of the medieval knights gave way to the king's justice.

Also during this transition, there was a growing infrastructure of commerce: financial instruments like money and contracts, and technological improvements such as better roads and carts and instruments of timekeeping. The result was that zero-sum plunder began to give way to positive-sum trade, a point that I'll return to later in the talk.

The third historical decline of violence can be appreciated by recalling some of the ways that the early kings and emperors kept law and order within their borders, namely extreme forms of physical punishment such as breaking on the wheel, burning at the stake, clawing with iron hooks, sawing in half, and impalement.

But in a process called the "Humanitarian Revolution," these forms of sadistic punishment were abolished in country after country. Here's a timeline of major countries with judicial torture, which shows that in the second half of the 18th century, there was an acceleration of the number of countries that eliminated sadistic punishment, including the United States in our famous prohibition of "cruel and unusual punishment" in the 8th Amendment to the Constitution.

Also abolished during this time was the profligate use of the death penalty for non-lethal crimes. In 18th-century England there were 222 capital offenses on the law books, including poaching, counterfeiting, robbing a rabbit warren, being in the company of gypsies –


Steven: – and "strong evidence of malice in a child 7 to 14 years of age."


Steven: By 1861, the number of capital crimes had been reduced to 4. Now, in every Western democracy but the United States, the death penalty has been abolished outright. The red line shows that most of the abolitions were in the 20th century, but the blue line, which shows the number of European countries that actually execute people, shows that well before the politicians got around to striking capital punishment from the law books, their countries had lost their taste for executing people, and on average 50 years elapsed between the last execution and the time that capital punishment was formally abolished.

I mentioned that the United States is an exception, or at least, 33 of the 50 states are exceptions, a number that is shrinking every year. But even in the United States, the death penalty is a shadow of its former self. This graph shows the rate of American executions on a per capita basis from colonial times, 1625, to the present. So nowadays, maybe 40 people get executed in this country every year, a country that has more than 17,000 homicides every year.

Also abolished during the Humanitarian Revolution were witch hunts, religious persecution, dueling, blood sports, debtors' prisons, and most famously of all, slavery. Slavery used to be legal everywhere on Earth. Every major civilization practiced it. No one seemed to think there was anything wrong with it. The Bible had no problem with it. Democratic Athens was a slave-holding society.

But then starting in the 18th century, there was a trickle of countries that began to abolish slavery, setting off a cascade that eventually enveloped the entire globe, culminating in 1980 when Mauritania became the last country to abolish legalized slavery. So we've been living through three decades that are unprecedented in human history in the fact that slavery is not legal anywhere on Earth.

What were the immediate causes of the Humanitarian Revolution? The most plausible, I argue in my book, is printing and literacy, the only technological change that antedated the Humanitarian Revolution. This graph shows that prior to the 18th century there was a 25-fold increase in the efficiency of manufacturing a book. That efficiency was put into practice so that during the 18th century there was an exponential rise in the number of books published per year, a kind of early version of Moore's Law, maybe you could even... [points to exponential increase on graph] I don't know, does that look like a singularity to you?


Steven: And for the first time in history, a majority of Englishmen could read those books. This was when literacy exceeded the 50-percent mark for the first time.

Why should literacy matter? Well there's another name for this era: The Enlightenment. It was the era in which knowledge began to replace superstition and ignorance. If you have a literate and educated populace, that is less likely to believe toxic notions such as that Jews poison wells, heretics go to hell, crop failures are caused by witches, children are possessed by the devil which has to be beaten out of them, Africans are brutish and fit only for slavery, and so on, it is bound to undermine many institutionalized forms of violence. As Voltaire said, "Those who can make you believe absurdities can make you commit atrocities."

Also, literacy is a technology of cosmopolitanism, the mixing of people and ideas. It's plausible that as people consume fiction, drama, history, and journalism, they start to inhabit the minds of people unlike themselves, which conceivably could expand their empathy and decrease their taste for cruelty, a point that I'll also return to.

The fourth decline of violence has been called "The Long Peace," and it speaks to the frequently made assertion that the 20th century was the most violent in history, an assertion that it turns out is quite questionable.

Let me show you trends in great power war over the last 500 years, these are the wars involving the biggest countries, whose deaths account for a majority of all deaths from all wars combined.

Here we have, from the year 1500 to 2000, the proportion of years that great powers fought each other. As you can see, several hundred years ago the biggest countries and empires on Earth always fought each other, that's just what great powers did, they fought other great powers. More recently, the great powers almost never fought each other. Here we have the duration of wars involving a great power, which is also showing a decline. History used to have things like the 30-Years War and the 80-Years War. The 20th century had the Six-Day War. Here we have the frequency of wars involving a great power, and that has gone down over the last 500 years.

There was one trend, however, that for most of this period went in the opposite direction. And that is, once a great power started a war, how many people was it able to kill per country per year? For most of this stretch of time, great powers got more and more efficient at killing people, until after the Second World War, when that figure did a U-turn, and that figure has been in decline. Meaning that for the last two-thirds of a century, we're living through a period in history in which the frequency of war, and the duration of war, and the deadliness of war, have all been going down in parallel.

If you multiply those numbers out so that you simply count out the total number of people killed in war, you get a zig-zaggy line, but one whose lowest point in the last 500 years corresponds to the last quarter of the 20th century. We can zoom in on the last hundred years and look at wars of all sizes. This is what the 20th century and the first decade of the 21st look like. It shows that there were two unmistakably horrific spikes of bloodletting, centered on the two World Wars, but contrary to predictions that this was only the start of an even deadlier progression, since then the line has been downward, and it has been hugging the floor.

This is the era that historians have called "The Long Peace," the fact that since 1946 there's been a historically unprecedented decline in inter-state war, wars between two countries. In fact the most interesting statistic of the post-war era has been "zero."

That's the number of wars between the two greatest powers of them all, the U.S. and the Soviet Union. Contrary to every expert prediction, World War Three never happened. No nuclear weapon has been used since Nagasaki, again contrary to every expert prediction that nuclear war was inevitable. There have been no wars between great powers of any size since 1953 with the end of the Korean war. No wars between Western European countries. Lest this seem a banal and obvious point, just recall that prior to 1945, Western European countries started two new wars a year for 600 years, as of 1945 that fell to zero. And there have been no wars between the 44 most-developed countries. We nowadays tend to think of wars as things that happen in remote, undeveloped, poor countries, but for most of human history it was the rich countries that were constantly at each other's throats, and because they could afford big powerful armies, they could do a lot of damage.

Well what about the rest of the world? In a process that I call "The New Peace," The Long Peace is starting to spread to the rest of the world. Here's in a nutshell what's happened since 1946. As I mentioned there were fewer inter-state wars. There were, however, more civil wars, both internal civil wars and internationalized civil wars, where some other country butted in to defend the government against insurgents. This happened as newly independent states with inept governments defended themselves against insurgent movements, both sides armed, financed, and egged on by the Cold War superpowers. But with the end of the Cold War, the number of civil wars started to decline as well.

The question now is, what kind of war kills more people? The old-fashioned wars between countries, or the more recent wars within countries? The answer is that the wars between countries did far more damage. There's nothing like a pair of advanced countries chucking artillery shells at each other, raining bombs on each other's cities, facing each other with thousands of tanks on the battlefield to rack up really high casualty figures.

So if you look at the total number of deaths from every kind of war, you get a stacked-layer graph that looks as follows. Here we have from 1946 to the present the rate of death in colonial wars, which tapered off to zero. The rate of death from inter-state wars, which shows a bumpy, but unmistakable, decline with peaks corresponding to, from left to right, Korea, Vietnam, and Iran-Iraq. Here we have the rate of death from pure civil wars and internationalized civil wars. The height of the entire stack corresponds to the entire rate of death, and what you see is that the overall trend has been downward, the first decade of the 21st century shows a thin laminate of layers, representing the low rate of death in war we're enjoying now.

What were the immediate causes of The Long Peace and The New Peace? Well, Immanuel Kant suggested three of them several hundred years ago in his essay "Perpetual Peace," in which he argued that democracy, trade, and an international community all disincentivized leaders from waging war. More recently Bruce Russett and John Oneal, have tested Kant's hypotheses quantitatively, showing that all three factors increased in the second half of the 20th century, and all of them are statistical predictors of peace holding all other factors constant.

Just yesterday, the Nobel Peace Prize was given to the European Union, which originally was the European Common Market, and before that the European Coal and Steel Community. This is a recognition of the role of trade and an international community in bringing peace to Europe.

Finally there are the "Rights Revolutions," the targeting of violence on smaller scales directed against vulnerable sectors of the population, like racial minorities, women, children, and animals.

The Civil Rights movement saw an end to the practice of lynching, which used to take place, at the end of the 19th century, at a rate of 150 a year, that's 3 a week, by the 1950s that was brought down to 0. Hate crimes against blacks have been in decline since the FBI first started keeping track of them.

The Women's Rights movement has seen an 80 percent decline in the rate of rape since statistics were first kept in the 1970s. Similarly dramatic decrease in domestic violence. And in the most extreme form of domestic violence of all, namely uxoricide, the murder of wives and mariticide, the murder of husbands.

The Children's Rights revolution has seen a steady decline in the number of American states that allow corporal punishment in school, like paddling and strapping. In every public opinion poll, approval of spanking as a method of disciplining children has been in decline in the West. And rates of both physical and sexual abuse of children have been in decline.

The Animal Rights movement has brought about a decline in hunting, an increase in vegetarianism, both in the U.K. and the U.S., and a sharp decline in the number of motion pictures in which animals were harmed.

Now the question is why has violence declined on so many scales of time and magnitude?

One possibility is that human nature has changed, and somehow people have had their inclinations toward violence bred out of them. I'm very skeptical of this possibility. For one thing we continue to see violence in our children, our two-year-olds still kick, bite, and hit, our boys still are involved in play-fighting and rough-and-tumble play. We continue to enjoy vicarious violence, such as murder mysteries, Greek tragedies, Shakespearean dramas, video games, hockey –


Steven: – and movies starring a certain ex-governor of this very state.


Steven: And in surveys of homicidal fantasies, the question "Have you ever fantasized about killing someone you don't like?", 75 percent of men admit to at least occasionally having fantasized about killing someone. What does this say about human nature? It says that 25 percent of men are liars.


Steven: A more likely possibility is that human nature is extraordinarily complex, and it has always comprised both inclinations toward violence and inclinations that counteract them, what Abraham Lincoln called "The Better Angels of our Nature," and that historical circumstances have historically favored our peaceable inclinations.

What historical developments brought out our better angels and stayed our hands before they could result in acts of bloodshed? One possibility is that Hobbes got it right when he identified "The Leviathan," a government with a monopoly on the legitimate use of violence as a force for peace.

A judicial system with a monopoly on violence can neutralize the incentive for exploitative attack by penalizing aggression. That also reduces the need for pre-emptive attack by those fearful of being attacked first. It reduces the need to maintain a belligerent stance of deterrence, and it eliminates the need for revenge after the fact.

In particular, it circumvents the self-serving biases that humans are prone to where both sides to a dispute always believe that their opponent's attacks are naked unprovoked aggression out of the blue, but their own attacks are justified retaliation after the fact. When you have both sides believing that, you can have endless cycles of vendetta and blood feuds, which outsourcing revenge to a disinterested third party can circumvent.

The second possibility is a theory of "Gentle Commerce," that plunder is a zero-sum game, the gain by the victor equals the loss to the victim, whereas trade is a positive-sum game, one where everybody wins. And as improving technology allows the trade of goods and ideas over longer distances, among larger groups of people, and at lower cost, it becomes cheaper to buy things than to steal them, and more and more of the rest of the world becomes more valuable to you alive than dead.

The third possibility has been called "The Expanding Circle," and it starts from the observation that humans are equipped with a sense of empathy, but unfortunately by default we apply it only to a narrow circle of blood relatives, close allies, and cute little fuzzy animals. But over the course of history, one can see the circle of empathy expanding to embrace the village, the clan, the tribe, the nation, other races, both sexes, children, and perhaps eventually other species.

This just begs the question of what expanded the circle, and technologies of cosmopolitanism are a plausible answer, that as people increasingly travel, read history and literature and drama and journalism, they begin to see the world through the eyes of others and have more trouble demonizing them. Indeed there's experimental evidence that when people are encouraged to adopt the perspective of a real or fictitious individual, they not only sympathize with that person, but with the entire category that that individual represents.

Finally there's "The Escalator of Reason," the possibility that the growth of literacy, education, and public discourse have encouraged people to think more abstractly and more universally. They rise above their parochial vantage point, which makes it harder to privilege their own interests over others'. They are encouraged to stand back and recognize the futility of cycles of violence, and increasingly see violence as a problem to be solved, rather than as a contest to be won.

Some historical evidence is that abstract reasoning abilities, as measured by IQ tests, increased over the course of the 20th century, the so-called "Flynn Effect." And other studies have shown that people with higher levels of education and intelligence are less likely to get into violent disputes, and that countries that have higher levels of education and intelligence are more receptive to democracy.

The final question I'll ask is why do all these forces seem to be pushing in the same direction? I think the answer is that violence is what game theorists call a "social dilemma." It's always tempting to an aggressor to exploit a victim. Of course it's ruinous to the victim. Since over the long run, victims become aggressors and vice versa, all parties would be better off if they could all somehow agree to avoid violence simultaneously. The dilemma is, how do you get the other guy to renounce violence at the same time as you do, because if you beat your swords into plowshares, but the other guy keeps his as swords, you could find yourself at the wrong end of one of those swords.

Over history one could well imagine that human experience and human ingenuity have gradually chipped away at this problem, just like we've dealt with other scourges of the human condition like pestilence and hunger. The common denominator among those four historical forces is that all of them have increased the material, emotional, or cognitive incentives of all parties to avoid violence simultaneously.

Regardless of what the best explanation of the decline of violence turns out to be, I think its implications are profound. For one thing, it calls for a reorientation of our efforts towards violence reduction from a moralistic mindset to an empirical mindset. Instead of lamenting, "Why is there war?" perhaps we should ask, "Why is there peace?" Not just, "What are we doing wrong?" but, "What have we been doing right?" Because we have been doing something right, and it sure would be good to figure out what exactly it is.

Also, the decline of violence I believe calls for a reassessment of modernity, of the centuries-long trend that has eroded family, tribe, tradition, and religion, which have given way to individual rights, cosmopolitanism, reason, and science. Everyone acknowledges that modernity has brought us many gifts: longer and healthier lives, less ignorance and superstition, richer experiences. But there's always been a current of romanticism and nostalgia that has questioned the price. Is it worth is if we have to live in the shadow of terrorism, genocide, world wars, and nuclear weapons?

If, though, despite casual impressions, the long-term trend, though halting and incomplete, is that violence of all kinds is decreasing, it seems to me that calls for a rehabilitation of the ideals of modernity and progress, and it's a cause for gratitude for the institutions of civilization and enlightenment that have made it possible.

Thank you very much.


Moderator: Thank you Professor Pinker. We have time for a couple of questions. I had one backstage myself, which is, clearly this sort of peace post-Second World War has been incredible. One major factor that seems stable so far is the mutually assured destruction relationship that the nuclear powers have with one another. Is that what you think could be the biggest weakness in the decline in violence or the biggest risk to that trend reversing? Or do you have other risks that you see as more likely?

Steven: Yes, one political scientist suggested that the nuclear bomb be given the Nobel Peace Prize.


Steven: But in "The Better Angels of Our Nature," I actually come out as skeptical that the "nuclear peace" theory is correct, that is that it's responsible for The Long Peace. I think certainly nuclear bombs deter an all-out, existential threat to a country, but there are a number of reasons I think it wasn't responsible for the postwar decline of war.

One of them is that the deterrence in the form of conventional destruction was probably plenty, that with Europe in rubble in 1945, no one was looking for a rematch any time soon. Even without nukes, the conventional destructive power of artillery and bombing and tanks had been shown to be so great that it's almost a difference between falling out of a five-story window and a hundred-story window. The hundred-story-window fall is much worse, but the five-story-window is enough to deter you. It's an analogy that comes from John Mueller.

But also because nuclear weapons are so insanely destructive, it's almost that they're useless as weapons of actual war, and so in the conduct of actual wars, countries have behaved as if they don't exist. That's why you have lots of pipsqueak non-nuclear countries threatening nuclear countries – Argentina taking over the Falklands from Britain, for example, confident that Britain was not going to reduce Buenos Aires to a radioactive crater. And therefore the nuclear threat, paradoxically, is so destructive that it's something of a bluff.

And finally there is also an increase in wars among non-nuclear powers, where nuclear deterrence doesn't factor in one way or another.

[next question]

Man 1: In this age of Facebook and Twitter and social networks broadcasting information more quickly, on one hand they can increase everyone's compassion and empathy and gratitude more quickly, but also due to the anonymity of the Internet, studies have also shown especially that there's more violence when you're wearing the mask. So how do we make sure that our social networks which we have created increase compassion, empathy, and gratitude, rather than increase violence and passive-aggressiveness?

Steven: Well certainly the era of social networks – which it's really too early to say what the long-term effect is, since they're really only about five years old as a global phenomenon – but it is an era in which, domestically, the rate of violent crime has continued to sink downward. And of course they have been given credit – we don't know whether it's really deserved or not, too early to tell – for the color revolutions and the Arab Spring.

It seems that by and large the effect seem to be positive, which would not be the first time in history, because the Humanitarian Revolution, as I mentioned, was preceded by really an earlier version of social networking, namely the republic of letters, the fact that printed communication in the form of correspondence, pamphlets, books that were quickly translated and went viral, exploded. And for that matter the Long Peace and the New Peace and the rights revolutions took place in the era of the so-called "electronic global village."

So while we don't know for sure, it looks like opening up the channels of communication, though you're right that it can have negatives as well as positives, it looks like the positives by far outweigh the negatives.

[next question]

Man 2: Fantastic talk, Professor Pinker. To what extent are these improvements irreversible? To what extent is it possible that the trends could be undone by growing tensions over resources, whether it's water or food or oil, or conflicts between Israel and Iran over fears of nuclear... to what extent might there be a recurrence of the reversal which happened in the Second World War? That was a reversal to the trends, could there be a worse reversal in the near future?

Steven: It's certainly possible, and it would be foolish to say that it can't happen. However, all of the risk factors for war are moving in the encouraging direction. So everything that's been correlated with peace in the past is on an upward trajectory. Again, this doesn't rule out catastrophes, and if there was a mammoth collapse of civilization, we could go back. But short of that, the signs are more or less positive.

The resource wars, I suspect, are not going to be a major threat to peace. For one thing, very few major wars in the last century have been fought over resources. They're fought over pride and fear and cosmic justice and ideology and grandeur and honor and all kinds of intangible psychological qualities, because one thing about resources are, that's something you can actually split. You can't split whether the holy land is going to belong to one religion or another, but you can divide up a supply of water. Studies that have looked empirically at correlations between resource stress and war, most of them find that resource stress at time-1 does not predict an increase in war at time-2. Again, it would be rash to rule it out, but that's not generally the cause of most organized armed conflict.

[next question]

Man 3: I had a question about, in the singularity where we have the power of let's say a terrorist organization to inflict great harm – you mentioned briefly the fact that the great powers had more ability to inflict great harm in certain wars. There's a possibility of either some sort of technical warfare, or also biological warfare, that a few people, or a small group, can actually inflict great, great harm. What's your take on that whole... ?

Steven: Well, it's not as obvious as it's sometimes said. I know that Kevin Kelly, for example, has argued that there's been virtually no increase in the amount of destructive power that an individual can wreak. Probably the best way that an individual can do a lot of damage is by taking over a state. But short of that, I think most nuclear security experts think that the window of vulnerability for stealing finished nuclear weapons is more or less closed. Fissile material is still a threat, but there is more and more effort to lock them down. It turns out probably not so easy to make a nuclear bomb in your garage. There are an awful lot of things that could go wrong, even if terrorists wanted to do it. Chemical weapons, most of them would do much less damage than good old-fashioned explosives, and that's probably true of homemade biological weapons as well.

So I still consider the greatest threat to be states, taken over by individuals, with big organized armies. Not that it's impossible that someone could steal a loose nuke or make one in their garage, but as best we can tell the odds are pretty small.

Moderator: Professor Steven Pinker, thank you so much. Excellent talk.