Steven Pinker's Medieval Murder Rates

In a highly problematic passage from Steven Pinker’s ‘The Better Angels of our Nature’ the professor highlights the fact that homicide rates have plummeted across Europe since the 13th century. He does this with reference to the work of sociologists and historians such as Ted Robert Gurr and Carl I Hammer which show that murder rates dropped sharply across the centuries – 14th century England was about 95% more violent than the present era. What conclusions does Pinker draw from this? He seems to be pushing some theory by Norbert Elias which states that a civilizing process occurred. Medieval people were boorish, animalistic and lacking in habits of refinement. According to Pinker:

‘over a span of several centuries, beginning in the 11th or 12th and maturing in the 17th and 18th, Europeans increasingly inhibited their impulses, anticipated the long-term consequences of their actions, and took other people’s thoughts and feelings into consideration. A culture of honor—the readiness to take revenge—gave way to a culture of dignity—the readiness to control one’s emotions.’

By contrast according to Pinker

‘The people of the Middle Ages were, in a word, gross. A number of the advisories in the etiquette books deal with eliminating bodily effluvia: Don’t foul the staircases, corridors, closets, or wall hangings with urine or other filth. • Don’t relieve yourself in front of ladies, or before doors or windows of court chambers. • Don’t slide back and forth on your chair as if you’re trying to pass gas. • Don’t touch your private parts under your clothes with your bare hands. • Don’t greet someone while they are urinating or defecating. • Don’t make noise when you pass gas. • Don’t undo your clothes in front of other people in preparation for defecating, or do them up afterwards…..In the European Middle Ages, sexual activity too was less discreet. People were publicly naked more often, and couples took only perfunctory measures to keep their coitus private. Prostitutes offered their services openly; in many English towns, the red-light district was called Gropecunt Lane. Men would discuss their sexual exploits with their children, and a man’s illegitimate offspring would mix with his legitimate ones’

Now at this point once again I have to jump to the defence of the poor benighted medievals. Unlike Steven Pinker I am a regular watcher of the ‘Maury Povich’ show in the United States (for UK readers the immediate point of reference is the ‘Jeremy Kyle Show’) and I have been on numerous pub crawls in UK city centres. All of the gross practices highlighted by Pinker are in evidence – one might say omnipresent - in modern society so it makes little sense to rat on our ancestors for displaying them. Perhaps Pinker needs to spend less like in the urbane, sophisticated environment of Cambridge Massachusetts and more time somewhere like Calton Glasgow. Then he might not have as much confidence in the voodoo like properties of Peter Singer's ‘empathy circle’.

So what to make of Pinker’s historical data? Well, from the start I would expect to see a drop in homicide rates across the centuries for four reasons. Firstly societies have gradually increased centralised power in the state and established a monopoly on violence. Secondly courts of law have become more effective as venues for settling disputes, thereby making the use of violence unnecessary. Thirdly schooling and education have introduced a greater civility – perhaps this counts as a ‘civilising process’? Fourthly, it is now much harder to kill people due to modern medicine and the emergency services. Wounds which would previously have been fatal and resulted in homicide now result in grievous bodily harm*. A Saturday night in Newcastle which in previous centuries might have resulted in a bloodbath now simply results in the A&E being clogged with aggressive drunks. It would therefore not be surprising if homicide rates were higher before these variables developed – what would be surprising is if they were lower.

Before looking at Pinker’s figures I should point out how homicide rates are calculated, as n per 100,000 of population per annum. Basically you take the number of murders and divide it by the population size (of say Medieval Norwich). You then multiply this by 100,000 to give you the murder rate. Pinker has some figures from Gurr which show the murder rate in Medieval London as having homicide rates from of around 50 per 100,000 during the 14th and 15th centuries (the present figure is more like 1.8 per 100,000). He quotes a figure from Carl Hammer showing that the murder rate in 14th century Oxford was 110 per 100,000 which is astonishingly high given how sleepy and civilised the place is today (this murder rate - calculated based on 36 cases of homicide between 1342 and 1348 - is akin to that of cartel ridden Ciudad Juarez in Mexico).

Are the figures accurate? Here we run into a number of problems. You might have noticed that the homicide rates are highly dependent on the population statistics. Michael Prestwich discusses this in Plantagenet England 1225-1360 (p507-508). One estimate he quotes is that London in the first half of the fourteenth century had a homicide rate of between 5.2 and 3,6 cases per 10,000 (equivalent to 52 per 100,000 and 36 per 100,000 meaning London was as violent as present day New Orleans). However this estimate was based on the population of London being 35,000 to 50,000. It’s become increasingly clear that these estimates are wrong. For example it’s clear that building densities around Cheapside were extensive by the end of the 14th century – at levels not reached again until 1600 when the population was 100,000-200,000 including suburbs. According to Prestwich estimates of the city's population now reach as high as 107,900 to 176,000. At a population of 100,000 the murder rate would be 1.8 per 10,000 (18 per 100,000). This would make London’s murder rate equivalent to present day Atlanta or Pittsburgh. A slightly higher population estimate would make the murder rate equivalent to present day Boston across the Charles river from Stephen Pinker’s office – which seems unlikely. If that were correct then the question we would have to ask is why our present day cities are more dangerous than their equivalents in an age of comparative lawlessness** ?

What of the Mexican murder rate for Oxford? Prestwich says that the high figure may be explained by the fact Hammer used coroners records to come up with his statistics. Unlike the present day these report the circumstances of a mortality and do not distinguish between murder, manslaughter or accidental death – hence you end up with an extremely wide range of possible rates***. Given the paucity of data – Pinker seems to have gone for the highest one in order to massage his thesis. Furthermore such records only cover a period of a few years and might reflect a one off crime wave ****.

Any conclusions based on what little statistics we have must therefore be provisional and potentially unsafe. For example, according to Prestwich, the records show that there were ‘only three larcenies in Norwich in 1313, as against 703 in Bedford, Indiana (a town of similar size), in 1975’. It would be ill-advisable to read that statistic and go on to write a book called ‘The terrible daemons of our nature’ showing the slide into criminality of Western Culture – especially since 14th century crime reporting probably less quite a lot to be desired.

*This is perhaps the most important point. For example Randolph Roth author of American Homicide argues that given modern medicine—emergency response, trauma surgery, antibiotics, and wound care—three out of every four people murdered before 1850 would probably survive today.

**The issue of how violent Medieval society was is seriously hampered by lack of evidence. Alternative interpretations exist such as Phillipa Maddern’s ‘Violence and Social Order: East Anglia 1422-1442’ which argued that the allegedly violent landscape of East Anglia (then the most urbanised area of England) was in fact, remarkably free of criminal violence and that this model could be applied to the rest of the country.

***As an example of the difficulties with this approach the only surviving run of coroners’ records for England’s 2nd largest city Norwich are from 1263 to 1268. These document 36 cases, 14 seem to be accidental death or theft. In 5 the conclusion is more ambiguous – either the jury swore the death was accidental or the suspect was cleared by compurgation. That leaves 17 possible instances of murder over 5 years – a proportion of which could classed as manslaughter. If these were all murders the average rate per year given a population of 17,000 would have been 20 per 100,000 – a rate akin to Philadelphia in 2010. If half were murders the rate would be 10 which is slightly less than Boston.

****Oxford was undoubtedly a violent place in the Middle Ages. Of 29 coroners’ reports that have been preserved for the period 1297-1322, 13 are murders committed by scholars. Attacks on townspeople were sometimes countenanced and even led by officials of the university. For example in 1526 a Procter organised a riot in which many citizens were attacked and their houses looted. In 1355 in what became known as the ‘St Scholastic’s Day riot’ an argument in a tavern became a pub brawl which went on for the next 3 days. It began when a group of students at an inn near Carfax disapproved of the wine they were served. The inn-keeper having given them ‘stubborn and saucy language’ the clerks ‘threw the wine and vessel at his head’. The townspeople then seized the opportunity to arm themselves with bows and arrows and attack scholars. Gangs of academics and citizens clashed in the streets and academic halls were burned. Six students and scholars were killed.

Discuss this post at the Quodlibeta Forum