About the question of violence in the usa there is an enormous literature - as well as a lot of popular perceptions which themselves incorporate some myths.

Contrary to public perception, historical criminologists now agree that the long-term trend in violence in Western societies is downwards. In England, longer-term data is available than anywhere else, and in a celebrated piece of research Gurr (1981) showed that the chances of getting murdered were about 40 times greater in thirteenth-century Oxford than in the mid twentieth cen-tury. The decline was not a smooth curve: there are shorter-term ?uctuations.

For instance, most countries experienced an upturn in violence from about 1960, and a renewed downturn since the 1990s. Trends in homicide in the usa run very much parallel to those in western Europe and European coun-tries overseas such as Australia and New Zealand. Data compiled by Eisner (2005) show this very strikingly.

But it is necessary to distinguish between the trend and the level of violence (Mennell 2007: 122-54). What is distinctive about the usa is the level: there are simply more homicides there than in comparable countries. Although other forms of violence ought really to be considered separately, homicides per annum per 100,000 population are used as a general index of violence because a homicide is a relatively unambiguous crime, and thus the measurement of homicide tends to yield comparable ?gures from one country to another, whereas lesser forms of violence are greatly affected by differences in legal de?nitions and indeed by legal changes over time. By this measure, the usa has something like four times as many murders as comparable coun-tries.

Here again, another popular myth comes into play: the phrase ‘crime and violence', current among politicians and the general public, implies that the two are almost synonymous. Yet, as Zimring and Hawkins (1997) have shown, in the usa ‘crime is not the problem'. One is considerably more likely to have one's house burgled in London than in New York. Zimring and Hawkins showed that homicides in America are not highly correlated with ordinary crimes such as burglary, theft and so on. Murders were not to any exceptional extent committed in the course of instrumental crimes - those associated with the pursuit of money in particular. What accounts for the unusually high incidence of homicide in the usa is the high incidence of affective violence - that is murders committed impulsively, under the in?uence of strong emotions.

Why should Americans be less able to control their murderous emotional impulses than Europeans? One answer is that they are not, but that a ?st ?ght outside a pub after closing time (or a domestic dispute) that generally results in cuts and bruises is more likely to end up with someone dead in a society awash with handguns. While there is most likely some truth in that, it does not tell the whole story. For in this case it is more than usually dif?cult to talk about ‘Americans' in general. There are very marked geographical variations in the incidence of homicide.