Gender-based violence and apophenia

Earlier today, my friend @kelltrill said

https://twitter.com/kelltrill/status/434275566293090304

and this led to a little bit of to-and-fro between her and some others who seemed to think it somehow obvious that if Oscar Pistorius had intentionally killed Reeva Steenkamp, it would have to be classified as gender-based violence. Now, that might be typical usage of the phrase gender-based violence. But if it is, I’d like to suggest to you that it’s wrong, and lazy, to speak of cases like this (i.e. a man killing a woman) as axiomatically gender-based.

Steenkamp & PistoriusNone of what I say here is intended to minimise or trivialise the fact that women are overwhelmingly more likely to be the victims of domestic assault by their partners than men are. There are hundreds of things I could link you to, but the evidence is so overwhelming that there’s no need – you can easily find something yourself. (And in case any MRA’s happen to wander past here, no, I’m not saying that men aren’t sometimes victims of various forms of discrimination themselves.)

Furthermore, I’m quite happy to regard this case as at least in part an instance of gender-based violence (on the assumption, for the sake of argument, that Pistorius intended to shoot Steenkamp). I’m happy to do so because Pistorius fits a classic alpha-male stereotype – proud, strong, with a history of short-temperedness and violence. The stereotype might not fit or be fair, but I’m disclosing it to wall it off, in that this case in particular is not my focus – I want to instead address the use of that generalisation (gender-based violence), with the case as a springboard for doing so.

The mere fact that a victim is female (or whatever) does not mean that the violence can be described as whatever-based. If Pistorius knew that he was shooting Steenkamp, then – obviously – the most fitting label for this action is Steenkamp-based violence, where Steenkamp is also a woman.

Even if it’s true (as it is) that more men abuse and kill their female partners than vice-versa, Pistorius can’t be known to have been more likely to shoot Steenkamp than he would be to shoot anyone else who he was ill-disposed to, or where he could benefit from doing so.

If a person had a history of violence against a certain sex, race, nationality or whatever, the generalisation has more merit – but before establishing whether those facts hold, we shouldn’t jump from a) the existence of a general culture of violence against X to b) the conclusion that a particular instance of violence against X fits that pattern.

I’ve argued something similar in a post about “Satanic” killings, where while it’s easy to generalise, doing so can obscure important details about motivation and how we should respond (for example, that psychiatrists might be more useful commentators than ghostbusters like Kobus Jonker).

The same danger of over-generalising in a confounding sort of way could occur with a murder or assault that is perpetrated across races – in South Africa, entrenched distrust between races could (more in some parts of the country than others) explain the motivations behind a murder, but they can’t be assumed to do so.

Take Eugene Terre’Blanche as an example: yes, he was a white supremacist, but the farmworkers who murdered him might have done so because he was also an abusive employer, or a rapist (as the murderers alleged). So while you could call that an instance of race-based violence, doing so would (or, could) distract from more pertinent details.

In short, what I’m arguing is that we should be careful of affixing convenient labels to events or people, even if they are often true. Harriet Hall has a review of an interesting-looking new book on critical thinking on Science-based Medicine, where I was introduced to a useful idea I hadn’t encountered before. It’s called apophenia, and

It means the spontaneous perception of connections and meaningfulness of unrelated phenomena, the tendency to find personal information in noise, seeing patterns where there are none, the kind of subjective validation that cold reading exploits.

To recap: I don’t dispute that gender-based violence is a real thing, and a real problem. But to call every instance of violence across genders (usually male on female) an example of gender-based violence is hyperbolic, in that it might be a judgement that claims more than what the evidence tells us.

This, in turn, could be problematic, not only because it’s a simple instance of laziness in not making fine discriminations regarding what data can tell us, but also because the more things you fit into a category, the more diluted that category might become.

It’s precisely because gender-based violence is such a real thing, and is such a problem, that we might want to be more cautious about affixing that label to cases that it might not fit.