If you want an article length rebuttal to Steven Pinker’s “decline of violence” thesis, you can’t do better than David Bentley Hart’s recent takedown in the latest issue of First Things. But if you only want an paragraph length rebuttal, you can’t do better than this from Timothy Snyder in Foreign Affairs:
Yet even if Pinker is right that the ratio of violent to peaceful deaths has improved over time (and he probably is), his metric of progress deserves a bit more attention than he gives it. His argument about decreasing violence is a relative one: not that more people were killed annually in the past than are killed in a given year of recent history but that more people were killed relative to the size of the overall human population, which is of course vastly larger today than in earlier eras. But ask yourself: Is it preferable for ten people in a group of 1,000 to die violent deaths or for ten million in a group of one billion? For Pinker, the two scenarios are exactly the same, since in both, an individual person has a 99 percent chance of dying peacefully. Yet in making a moral estimate about the two outcomes, one might also consider the extinction of more individual lives, one after another, and the grief of more families of mourners, one after another.
December 27th, 2011 | 6:08 pm
Joe
Where exactly is it that ten people in a group of 1,000 die? Why its probably a village or a small group of villages. But when the world had only one billion people in it where were they located? Why in various groups of 1,000. In those ‘small’ groups of 1,000 ‘only’ ten people were getting killed so by Snyder’s logic that’s no big deal. But in the whole wide world its ten million getting killed and that’s a big deal! Wait, this has nothing to do with how many people are getting killed but whether or not the observer is wants to pay attention to the entire world or just his small village!
December 28th, 2011 | 8:02 am
I thought that Stephen Colbert did a really good job of exposing how ridiculous this book was, and he did it to Pinker’s face as well!
“So,” Colbert said, “if you kill a million people in a country with 2 million, that’s pretty bad. But if you kill a million people but there are 40 million people in the country, that’s progress! Statistically. Because human life has suffered from inflation.”
December 28th, 2011 | 10:59 am
“so by Snyder’s logic that’s no big deal.”
No, Snyder’s argument is not “that’s no big deal.” It’s that it’s EVEN MORE of a big deal when you multiply both sides of the equation by 10,000.
“Wait, this has nothing to do with how many people are getting killed but whether or not the observer is wants to pay attention to the entire world or just his small village!”
It seems all altruistic to claim that it’s just as big a personal loss when people a thousand miles away die as your neighbor, but it’s really not. It would be a dysfunctional person indeed who either managed to cultivate a neighborly of loss for someone with whom he had no direct connection (think of the bizarre spectacle surrounding the death of the Princess of Wales), or had no more than a distant sense of loss when death struck next door (“after all, his death matters no more to the world than the child who died an hour ago in Afghanistan, so it shouldn’t to me.”)
So, yes, violent death spreading through a larger community and touching more people is actually worse, according to the actual way human beings process loss and experience the violation of violence happening close to home.
And I love Stephen Colbert’s remark.
December 28th, 2011 | 11:24 am
Pentamom–Great points as always.
Liam–it’s nice to hear that Colbert got something right for once (although I have only seen him a few times I seem to hear about him quite a bit).
What’s so great about Colbert’s comment is that it is both equally obvious and devastating. Once that point occurs to anyone, they have lost all reason to even consider the book. And yet Pinker spent how many hours/days/weeks/months/years(?) slaving over this book in order to crank out over 700 pages that makes him look like a fool the moment anyone thinks about what he’s saying for even a few moments.
No doubt Pinker thought he was on to some seminal takedown of Christianity. But what he accomplished reminds me of an episode of Freaks and Geeks where the lovable teenage Sam goes shopping for clothes on his own for the first time. The salesman convinces him that he’ll be a “super stud” when he walks into school wearing a bright turquoise jump suit. He spends the night before school fantasizing how cool he’ll be regarded the moment he walks in the door, only to discover that he’s an instant laughing stock. But Sam wasted only a day of his life.
Sometimes I think Satan lets his own disciples fail just for his personal entertainment.
December 28th, 2011 | 11:41 am
Pinker’s argument is simply an attempt by the secular-atheist crowd to avoid the fact that they presided over the bloodiest century in human history.
A related argument is the claim that modern weaponry is much more efficient at killing people, so the various ideologies that blighted the last century really aren’t more bloodthirsty than the animating forces of earlier centuries.
December 28th, 2011 | 12:56 pm
His statistical methods are shaky. He is neither an historian nor a statistician, and he uses metrics that do not actually measure what he wants them to measure. If the Western Allies killed more Nazis than the Nazis killed Western Allies, does that mean that Nazi nature had better angels? (The world population being the same for both.)
Surely what denominator matters if we are to consider how casually and readily humans have resorted to violence is the “potential victims within reach,” not the world population. What relevance has the number of people minding their own business in China and India to the bloodthirstiness of the Kickapoos and their allies when they wiped out the Illini tribe? They were never within reach of Kickapoo tomahawks.
Of course, Pinker was also capable of writing regarding religious violence:
“World War I, as I recall, was a war fought mostly by Christians against Christians.”
which surely indicates that his historical understanding rivals his statistical acumen.
There are a number of historical looks at his claims at Quodlibeta:
1. http://bedejournal.blogspot.com/2011/11/steven-pinker-and-an-lushan-revolt.html
2. http://bedejournal.blogspot.com/2011/11/pinker-tackles-albigensian-crusade.html
3. http://bedejournal.blogspot.com/2011/11/steven-pinkers-medieval-murder-rates.html
4. http://bedejournal.blogspot.com/2011/12/how-bad-were-mongols.html
5. http://bedejournal.blogspot.com/2011/12/this-is-bogus-statistic.html
December 28th, 2011 | 4:12 pm
Liam
“So,” Colbert said, “if you kill a million people in a country with 2 million, that’s pretty bad. But if you kill a million people but there are 40 million people in the country, that’s progress! Statistically. Because human life has suffered from inflation.”
Colbert is always good for a wicked laugh, but the fault in this logic is the world ‘you’. “YOU” are not killing one person, ten people, or 2 million people. We are talking about how many people are killed violently in a country with 2 million versus one with 40 million and its just silly to try to compare absolute numbers instead of some type of rate.
Imagine if we were comparing health care systems. Who has more absolute sick people, the US or North Korea? Why clearly the US which has many, many times more people than North Korea so therefore we must conclude that North Korea has better healthcare than the Us. To argue anything else, such as some metric like the % of deaths due to preventable illnesses, means you’re a moral ingrate who doesn’t hear the cries of the millions of Americans who are suffering with various illnesses!
So yes, a person who kills a million people is a million times worse than someone who kills one person, regardless of whether his murders happen in a country with a very large or very small population. That observation has nothing to do with Pinker’s thesis unless a single person happens to be responsible for all violence in any given year.
Pentamom
It seems all altruistic to claim that it’s just as big a personal loss when people a thousand miles away die as your neighbor, but it’s really not. It would be a dysfunctional person indeed who either managed to cultivate a neighborly of loss for someone with whom he had no direct connection (think of the bizarre spectacle surrounding the death of the Princess of Wales), or had no more than a distant sense of loss when death struck next door (“after all, his death matters no more to the world than the child who died an hour ago in Afghanistan, so it shouldn’t to me.”)
I think what’s getting missed here is that we are talking about whether overall rates of violence have changed through history. Not what type of violence has a larger or smaller emotional impact on particular observers. In real terms if I wanted to discuss which city was more violent, Detroit or New York, I would use some type of rate to adjust for different population sizes. Yes its true a Detroit father whose son is killed by a mugger while visiting NYC will not have this unbiased way of viewing the issue, but that’s not the proper context here. So if using a rate of violence is proper to compare NYC to Detroit why is it not proper to do the same thing to compare, say, the 20th century to the Bronze Age?
December 28th, 2011 | 4:21 pm
YOS
I briefly hit some of your links and I’m seeing a lot of quibbles but to what degree does this miss the forest from the trees? Or exactly getting so obsessed over the fact that what you thought was a tree happened to be a large bush that you neglect to note that yes you really still are inside a forest. Any attempt to compare rates of violence cross both the globe and historical epochs is almost certainly going to end up producing a lot of stats that will have to be ‘jerry rigged’ from numerous incompatitable sources. But what about the core issue, has there or has there not been a long term general decline in the rate of violence even counting the horrors of the 20th Century?
December 28th, 2011 | 6:34 pm
@Boonton
Doing any sort of statistical analysis on swags masquerading as data is like washing garbage. If Pinker wants to argue straightforwardly that this is his opinion or his impression, then that’s fine. But when he pretends it’s all “scientifical and statistic,” the statisticians can roll their eyes along with the historians.
I happen to think he’s right, but for the wrong reasons. Libertarianism is not the reason, and does not equate with being Bright.
December 28th, 2011 | 7:35 pm
Bontoon,
Thanks for the reply.
I don’t really think that’s what Colbert meant when he used the word “you.” At least, it never occurred to me that’s what he meant. I think he was being abstract, you could easily drop out the “you” and just put “if 1 mil were killed in a war.” I don’t recall him talking about mass murder as perpetrated by a single individual at that point.
I suppose the video is always around the internet…
December 28th, 2011 | 7:37 pm
I accidentally misspelled your name! Changing it to “toon” is not intended to be an insult, whoops!
December 28th, 2011 | 7:53 pm
“I think what’s getting missed here is that we are talking about whether overall rates of violence have changed through history. Not what type of violence has a larger or smaller emotional impact on particular observers. ”
If all Pinker is doing is relaying statistical information indicating that the percentage of people dying violently do or do not vary in a certain way, that would be true. But that’s not all he’s doing, is it? He’s arguing that humanity is “less violent” than formerly. In order for “less violent” to have a meaning, it has to have some subjective component relating to how we experience violence, or perhaps more properly, how people are more or less prone to violence. And for that to be meaningful, it has to relate somehow to what people experience.
December 29th, 2011 | 3:54 am
Pentamon
What he has to show, surely, in order to show that the world has become “less violent” is that fewer people are inflicting violence, rather than fewer people suffering it.
It is the number (or percentage) of perpetrators, not the number of victims that is relevant to his claim.
December 29th, 2011 | 7:18 am
YOS
Doing any sort of statistical analysis on swags masquerading as data is like washing garbage. If Pinker wants to argue straightforwardly that this is his opinion or his impression, then that’s fine. But when he pretends it’s all “scientifical and statistic,” the statisticians can roll their eyes along with the historians.
This is rather odd. You seem to be saying the job of the statistician is to sow extreme skepticism. I can say rather objectively Detroit is a more violent city than New York but comparing Rome of 50 AD to Rome of 1950 is just abstract painting! I think statisticians can do a bit better than simply saying the purpose of statistics is to warn people not to try to use statistics!
Liam
I don’t really think that’s what Colbert meant when he used the word “you.” At least, it never occurred to me that’s what he meant. I think he was being abstract, you could easily drop out the “you” and just put “if 1 mil were killed in a war.”
Pretend for a moment that these 1M people were killed by a single person whose last name was ‘War’. Bill War killed 1M people in a population of 50M. His sister Susan later on killed 2M people in a population of 100M. It’s perfectly fair to say that Susan is twice the murderer as Bill. But in both cases you can also say the odds that you would be killed by a member of the War family was 2% and that had not changed. N.P. on the name misspelling. I do enough violence to spelling myself!
pentamom
In order for “less violent” to have a meaning, it has to have some subjective component relating to how we experience violence, or perhaps more properly, how people are more or less prone to violence.
Why? Again return to the idea of comparing Detroit to New York. Why does one have to add a subjective element into such a comparision to have meaning? The FBI objectively reports the murder rate in each city. The father in Detroit whose son is murdered in NYC will subjectively feel worse about NYC’s violence, but that doesn’t alter the fact that objectively Detroit is a much more violent city.
How people are more or less prone to violence sounds less about the fact (violence rates have fallen) and more about the why (which could be for one or many reasons and will almost always be much more difficult to prove than the first).
December 29th, 2011 | 8:01 am
Here’s a quote from Pinker on murder rates — individual on individual — that is pretty compelling: “The criminologist Manuel Eisner has assembled hundreds of homicide estimates from Western European localities that kept records at some point between 1200 and the mid-1990s. In every country he analyzed, murder rates declined steeply—for example, from 24 homicides per 100,000 Englishmen in the fourteenth century to 0.6 per 100,000 by the early 1960s.” Now, I don’t know how accurate these figures are, but if they are accurate, it says something significant about medieval vs modern society, doesn’t it.
December 29th, 2011 | 9:43 am
Siger,
Good point. You can quibble with the actual calculation of murder rates between 14th century Europe and today. In fact, because the figures were kept using inconsistent methods different statisticians will almost certainly employ different methods to try to collect the data into national figures and that will almost certainly produce different numbers, but the real question is does it produce a different trend? If England’s homicide rate was 20, 25, 30 or 15 per 100,000 in the 14th Century that doesn’t alter the fact 0.6 per 100,000 in the early 1960′s represents a huge decrease. Unless you can say the methods used to calculate the homicide rate was so off that in the 14th Century it was actually closer to 0.6 per 100,000 (or the homicide rate in the 1960′s was improperly measured and was really more like 24), you’re just nitpicking.
To Pentamom’s point, the fact that a parent who lost their child to violence in 1960′s England subjectively doesn’t feel that England was a non-violent place shouldn’t really enter into this discussion. It’s quite possible the 14th Century Englishman didn’t feel his time was very violent because he was numb to the high rate of homicides. I had an experience like this some time ago when a relative pulled out some old newspapers she had saved from the moon landings. Deep in the back pages was a brief article about a man who had killed his wife and 5 of his 6 children who had refused a visit in jail by his remaining son being prepared to go to trial. What was amazing was that if a crime like that happened in present day northern NJ, it would be front page news in screaming fonts. Subjectively we may ‘experience’ violence much more when it becomes much more rare because we become ‘used to it’ when its common. I think this is only a valid point in this discussion if you want to argue that we should be concerned about our subjective emotions so therefore embrace the much higher homicide rate so that way any particular homicide won’t feel as shocking to us.
December 29th, 2011 | 10:37 am
“Why? Again return to the idea of comparing Detroit to New York. Why does one have to add a subjective element into such a comparision to have meaning? ”
Because Pinker didn’t just throw a bunch of statistical comparisons about violent death on the table and then walk away. He’s trying to make a point about human nature, and that point is only important insofar as we experience it. If you want to defend the idea that Pinker is making an absolutely objective but trivial point, or is dealing in a cold hard fact with no ramifications, I guess that’s your right, but I do believe Pinker sees himself as doing something a bit more existential than that.
December 29th, 2011 | 10:39 am
“I think this is only a valid point in this discussion if you want to argue that we should be concerned about our subjective emotions so therefore embrace the much higher homicide rate so that way any particular homicide won’t feel as shocking to us.”
I find it fascinating how often you take the opposite point from the one I’m arguing. I’ll blame it on my own communication skills.
December 29th, 2011 | 11:27 am
“Now, I don’t know how accurate these figures are, but if they are accurate, it says something significant about medieval vs modern society, doesn’t it.”
We are not just talking about homicide rates, although Pinker would like to keep the focus there. That way, he can avoid inconvenient facts like two world wars, the Holocaust, and the various communist mass murders in the last century.
December 29th, 2011 | 11:42 am
Brian
We are not just talking about homicide rates, although Pinker would like to keep the focus there. That way, he can avoid inconvenient facts like two world wars, the Holocaust, and the various communist mass murders in the last century.
Unfortunately for your argument he does account for them as well.
Homicide rates, though, are useful because we typically view war an exceptional whereas homicide is typically a fact of life in human societies. But from what I’ve read here and elsewhere, Pinker does try to account for the rate of wars and the intensity of wars as well and even incorporating all the horrors of the 20th century, it seems it still trumps the 14th and earlier periods.
pentamom
Because Pinker didn’t just throw a bunch of statistical comparisons about violent death on the table and then walk away. He’s trying to make a point about human nature, and that point is only important insofar as we experience it.
I’m still not quite getting this. Let’s use a different possible argument about some more simple aspect of human nature, biology. Suppose Pinker’s statistics showed that over the last few thousand years humans have gotten taller on average. He goes further to attribute that to improvements in diet, medicine and hygene. That’s clearly an argument about human biological nature, but doesn’t require any discussion of how we ‘experience it’. If, even though we are all taller, the ‘tallest person in class’ doesn’t happen to feel especially taller than us whereas it did to the typical 14th century peasent viewing the members of the nobel familes who lived in the castle, well that’s interesting but not very important to the discussion of whether human nature, in regards to height, has been changing over the long sweep of history.
December 29th, 2011 | 2:21 pm
“Pinker does try to account for the rate of wars and the intensity of wars as well and even incorporating all the horrors of the 20th century, it seems it still trumps the 14th and earlier periods.”
No he doesn’t. Read some of the reviews of the book, including the one linked to in the article.
And do you really believe that there was more violence in the 14th Century than there was in the 20th? Seriously?
December 29th, 2011 | 3:07 pm
Brian,
Two links were provided in this article, the first is to another First Things piece dismissing Pinker but not mentioning WWI & WWII but making an argument against using a rate of violence metric because that detracts from the importance of the actual number of lives lost. The article doesn’t say so directly but one gets the impression its trying to counter Pinker’s analysis of WWI/II. That’s not about Pinker failing to account for WWII in his analsys, that’s disagreeing with the way Pinker accounts for WWII.
The Foreign Affair’s piece extensively cites Pinker’s treatment of Nazi Germany not because the author thinks Stalin and Hitler actually put the modern world’s homicide rate on a par with the 14th century one, but because the author thinks Nazi Germany’s existence questions Pinker’s thesis that the decline in violence is caused by the world becoming more literate and more pacified by states forcing individuals away from direct violence and towards market based interactions.
Look I said over on the “The Precious Steven Pinker” thread that I think there are areas where his thesis is very vulnerable to attack but mostly in the cause of the decline in violence. Pinker is a pretty smart guy, ‘forgetting to include World War Two’ when comparing rates of violence between the 20th century and previous ones is probably a mistake he is not going to make. If you think this is your big ‘gotcha now Pinker!’ moment, you should really double check yourself.
December 29th, 2011 | 4:38 pm
“Pinker is a pretty smart guy, ‘forgetting to include World War Two’ when comparing rates of violence between the 20th century and previous ones is probably a mistake he is not going to make. If you think this is your big ‘gotcha now Pinker!’ moment, you should really double check yourself.”
I never said he forget about it. I said he claims it is irrelevant to the point he is making, which is nonsense. Removing things like wars from your consideration of the level of violence is absurd.
And once again, do you seriously believe the world was more violent in the 13th Century than it was in the 20th?
December 29th, 2011 | 4:39 pm
[I]I think statisticians can do a bit better than simply saying the purpose of statistics is to warn people not to try to use statistics![/I]
Or at least warn amateurs not to try them at home. They can say not to use statistical methods developed for measured data and apply them to swags. Or not to use conflicting and non-operational definitions. Etc. IOW, statisticians can say that if you’re going to do statistics, you should do them right.
For example, Manuel Eisner cites a chart by a previous researcher in which each point on the chart represents an “estimated homicide rate for a city or county for periods ranging from several years to several decades.” Even a non-statistician can see why you should not plot yesterday’s apples and the past year’s barrels of apples on the same chart. What does a trend line mean if one point comes from Bristol and the next comes from London? In statistical language, we are dealing with two different processes (B) and (L), each of which may have inherently different process means.
Here’s an example: Prior to the establishment of state police apparatus, like the London bobbies, a great deal of justice was meted out by the free enterprise system. That is, the victim’s relatives would take matters into their own hands. But then this counts as another homicide. So the post-judicial figures ought to include the number of executions that would otherwise have been vengeance killings. And Pinker himself has mentioned that much of the decline in homicide rates is due to abortion getting rid of future killers before the fact. So, abortions also should count, no?
+ + +
[I]“Now, I don’t know how accurate these figures are, but if they are accurate, it says something significant”[/I]
Well, the accuracy of the figures is precisely what is at issue.
For example, Eisner writes:
(M)any historians have questioned whether counting homicides documented in premodern records and comparing respective rates over eight centuries is scientifically sound. This problem has a methodological side, including issues of how complete premodern records are, what the historical ‘‘dark figure’’ of homicide might look like, what the relationships among homicide and other forms of interpersonal violence are, what impact changing medical technologies have, and how accurate population estimates are. But some historians of crime also raise a more substantive problem since the comparison of numbers over time assumes some comparability of the underlying substantive phenomenon, namely, interpersonal physical violence.
Eisner’s paper is here:
http://videojuegosycultura.files.wordpress.com/2008/01/long-term-historical-trends-of-violent-crime.pdf
His more extensive database shows estimated homicides declining logistically as Europe became steadily Christianized. What Pinker apparently overlooked when he ascribed the decline to the Enlightenment and libertarianism is that Eisner plotted a decline that began *before* the Enlightenment. Effects do not precede causes in that manner.
Eisner also states that:
[I]“By around 1950, most European countries experienced their lowest historically known levels of homicide rates. Since then, an increasing trend has prevailed.”[/I]
Which may indicate that some causal factor has been intensifying since 1950. Perhaps the lesser angels have been taking hold of human natures since secularization gathered steam in Europe.
December 29th, 2011 | 5:42 pm
Brian English –
Are you sure he’s only making one point?
December 29th, 2011 | 10:11 pm
Here’s an example: Prior to the establishment of state police apparatus, like the London bobbies, a great deal of justice was meted out by the free enterprise system. That is, the victim’s relatives would take matters into their own hands. But then this counts as another homicide.
OK, but we are talking about a rate of 24 per 100,000 in the 14th Century to 0.6 in 1960. Let’s say that half the homicides in the 14th Century were really just applications of the death penalty by ‘private law’. If you want to back that out then you’re only down to 12 per 100,000 which is still pretty big compared to 0.6 per 100,000.
And Pinker himself has mentioned that much of the decline in homicide rates is due to abortion getting rid of future killers before the fact.
Was abortion a big factor in London in 1960? In order for it to be according to the Levitt hypothesis (i.e. legalized abortion reduces future crimes because mothers in difficult situations opt to abort) the rise in abortion would need to happen in 1940 and before, not 1960 in order to impact 1960 crime rates.
Likewise regarding the chart where some points come from London, some from Bristol, some represent a few years and some a few decades. Let’s just say for the sake of argument that the homicide rate is totally random and oscillates between 0.6 and 24 per 100,000. 1301 just happened to be a bad year at 24 and 1960 happened to be a good one at 0.6. Errors caused by ‘crappy data’ should roughly even themselves out. The odds of drawing a ‘fat point’ is no more likely during a good patch of years than a bad one.
On the other hand, assume the trend is real, 1301 was the worst year at 24 and 1960 the best at 0.6 and every year inbetween was a steady march more or less downwards. Even if some of your data points are inconsistent, you’re still going to see a trend line.
His more extensive database shows estimated homicides declining logistically as Europe became steadily Christianized. What Pinker apparently overlooked when he ascribed the decline to the Enlightenment and libertarianism is that Eisner plotted a decline that began *before* the Enlightenment. Effects do not precede causes in that manner.
I’m glancing at what appears to be his trend line on page 3 of the pdf file, I’m noticing that the most dramatic decrease he finds is between roughly 1500 and 1700 where it looks like Engand went from about 15 down to maybe 5. England began to be converted sometime around 600AD and was more or less Christianized by 1,000 AD. The trend line appears to be ‘bottomed out’ after 1800….even though that’s the point where you can probably at least make a case for increasing deChristianization among some of the upper classes. Unless you think getting rid of Roman Catholics is what makes a country enjoy a low murder rate, it doesn’t seem as good a match. But I will grant you England’s decline in homicide rates does appear to have been well underway before the Enlightenment. The data for the Netherlands likewise shows a rather dramatic fall between about 1400 and 1700….not quite Enlightenment periods but also hardly periods of Christian conversions.
Brian English
And once again, do you seriously believe the world was more violent in the 13th Century than it was in the 20th?
Yes I do. Let me put it this way, suppose I told you that I was to randomly deposit you in a ‘lifetime’ of a person from the 20th century or the 13th. Even given the fact that hundreds of millions in the 20th century suffered under communism and the two World Wars, the fact remains it appears the odds of your lifetime NOT ending in some type of man made violence would be better in the 20th century.
On the level of crime I think you’ve already admitted this without explicitly saying so. It’s very plausible to believe the odds of getting smashed in the head while walking home late at night in London was higher in 1350 than 2005. This is why you’re pushing to include the World Wars, hoping the body count goes up enough to make us even with the 13th century. But the problem is that while the World Wars claimed many lives relative to previous wars, the world’s population increased by billions while wars and oppression claimed millions.
Let me ask YOS and others, do you all seriously believe the world is not less violent than it was in the 13th century or are you really just disagreeing with Pinker’s explanations for why it is?
Links
Blogs
Find Us
Contact