*Codes of the Underworld: How Criminals Communicate*

That's the new book by Diego Gambetta and it is the best applied book on signaling theory to date.  Gambetta's task is well summarized by a single sentence:

Given these propensities, one wonders how criminals ever manage to do anything together.

The signaling problems faced by criminals are unusual in the following regard.  On one hand they wish to signal a certain untrustworthiness, namely that they are criminals in the first place.  This is useful for both meeting other criminals and also for intimidating potential victims.  On the other hand, the criminals wish to signal that they are potentially cooperative, for the purpose of working with other criminals.  Sending these dual signals isn't easy and Gambetta well understands the complexity of the task at hand.  As Henry points out, facial tattoos are one particular effective method of signaling that one is a criminal for life.

Here is a passage which I found striking:

…Women are significantly less violent than men in the outside world and less lethal when they are violent.  This holds in all times and places for which relevant data exist.  And yet in prison this universal fact is overturned: women become at least as violent and often more prone to violence than men are.  Although women in prison rarely commit homicide, a large study of Texas prisons by Tischler and Marquart showed that there was no difference between women and men in the incidence of violent episodes.  Table 4.2, based on comprehensive statistics for England and Wales, shows that the gender pattern is even reversed; women assault each other twice as much as men do, and they fight one and half times as much as men do, a result that disconfirms the testosterone hypothesis.

Generally, women are convicted of proportionally fewer violent offenses than men are and have shorter criminal histories, two circumstances that rule out some of the possible selection effects that could explain away the high rates of female prison violence…

Gambetta wonders whether women in prison resort to violence so frequently because they have fewer alternative credible means of signaling toughness.


Interesting. How much time does Gambetta devote to legislation?

I always thought the disproportionate violence of women in prison was due to the fact that women didn't get sent to prison unless they were the worst of the worst.

In prison a lot of men turn into women. Perhaps similar doesn't hold for women in prison--and thus they are more unhappy?

Isn't a more plausible explanation for the higher observed level of violence among female prisoners simply that male prison violence is under reported?

We see the pattern one would expect to see if this were the case: lower incidence of homicide for women (which would always be reported) but a higher incidence of minor crimes (which wouldn't be reported for men).

Thanks Rob

I read somewhere that women start as many domestic fights as men. Men finish them.

Speaking from my own experience, I agree with that.

But the way I ended my ex-wife's verbal and physical assaults was separation and divorce, not more violence.

It's been a pet theory of mine that the reason women display less physical aggression has more to do with people being hard-wired to use violence when they're generally bigger and stronger than those around them and to avoid it when they're generally smaller and weaker than those around them than with either sex having hard-wiring for or against aggression (note how much less violent men get as they age past their physical prime). Perhaps, in line with that, the reason women are so violent in prison is because prisons are segregated by sex, so women in prison are surrounded by other women, and no longer in a situation where there's usually many bigger and stronger people about.

Why grasp at weird explanations involving signalling when simple conditional probability sufficies? Call the violence level that gets a person into prison V = 1.0. Clearly there are many more men than women in the general population with V > 1.0. But once we look at the prison population we have already selected for V > 1.0. There isn't a lot of range in the scale above that level (we only imprison a few percent of the population), so it should come as no surprise that the average V for men and women in that population is very close.

At Steve Sailer, you're right in general, but wrong in particular. The Chicago Outfit is, famously, open to membership for or cooperation with several ethnicities and has been since it was led by Torrio. And the alliances between Jewish and Italian mobsters are longstanding from the 1920s as well. So, *sometimes* the Hollywood fantasy is real.

Call the violence level that gets a person into prison V = 1.0.

One big problem with that theory is that violence is not the only thing that gets you into prison. In fact, it isn't even the *primary* thing that gets you into prison. A majority of the prison population has not been convicted of any violent crime. This is what happens when you treat drug users the same way you treat armed robbers and murderers.

Now, it would hardly surprise me if the population of non-violent offender inmates were more violent than the population at large (in fact, I'd expect it), but they aren't actually *selected* for violence. They are selected for legal disobedience, and the probability with which they are selected *given* said disobedience has a lot to do with race, class, gender, mental illness, etc.

It's not simply about violence in the way you imply.

An incredibly interesting page of information. I found these facts very fascinating. Thanks. - Spencer

Comments for this post are closed