Not all men
Won’t somebody please think of the poor stigmatized men?
Lone male travellers should be allowed to sit next to unaccompanied children on passenger planes, a discrimination tribunal has ruled. The decision comes after a man was asked to swap seats with a woman by the cabin crew of a flight travelling from Oslo to Paris in October 2022.
They told him he could not be seated next to two children travelling alone, with the policy being enforced by Air France to prevent any possibility of predatory behaviour. But the man, Dominique Sellier, filed a complaint with Norway’s anti-discrimination tribunal, which ruled that the airline’s policy was discriminatory.
Yes it is but (sorry to repeat myself) sometimes discrimination is necessary. Sometimes it’s for bad reasons but sometimes it’s not. Life is complicated that way.
According to the tribunal’s ruling late last year, a copy of which was obtained by AFP, Air France’s policy stipulates that if a flight is fully booked, a woman should “preferably” be seated next to unaccompanied minors.
And why would that be? Because sexual predators are very rarely women.
In proceedings, a lawyer for Air France argued that the crew was merely following company policy, which was based on the argument that men account for 97.93 per cent of all suspected sex crimes.
There you go. 98 percent – that’s why you don’t want to seat a child next to a male stranger on a plane.
Mr Sellier said Air France generalisation about men went too far, adding: “How can we accept this kind of suspicion because we belong to the male gender?”
By being a god damn adult, that’s how. By realizing that it’s more important not to let children be creeped on than it is to coddle the vanity of men.
I wonder what the policy has to say about men who “identify” as women? After all, if governments consider it perfectly safe to put them into women’s prisons, then surely a mere airline should let them sit next to unaccompanied minors, amirite?
Yeah, as a man there is the slightest of rankles that I get tarred with the same brush as the worst of us. But then I get over myself. Many of these policies around harm prevention and risk management are rational and based on analysis of statistics (and, frankly, highly observable). So, you know, fair enough. Comply with good grace so you don’t look like an example of the problem.
It should be a mantra. “It’s just stats, it’s not personal.”
[...]a woman should “preferably” be seated next to unaccompanied minors.What if the unaccompanied minor is, say, a 17-year-old male? It is possible that seating a woman next to him would make it more likely for an offense to occur. Depending on age, there must be some tipping point between the likelihood of being a victim and that of being an assaulter.
This is an abuse of statistical reasoning.
There are numerous crimes that are predominantly committed by one group or another. Does that entail that we should reasonably expect that a member of one of those groups will commit one of those crimes? Suppose 100% of some crime is committed by Jews, but only 0.00001% of Jews commit the crime. Do we say nope, can’t trust the Jew? If 75% of some crime is committed by the majority group, but 20% is committed by a tiny minority that only makes up 3% of the total population, which does it make more sense to be wary of? That is, for which is group membership a more reliable predictor? 89.4% of under-18 DUI arrests were whites. Should we not allow white teenagers to drive?
I’m not necessarily saying that the policy is wrong, but the reason given does not per se show that it is. It needs additional premises.
NiV, as a statistics junkie, I will agree that you are close enough to correct in your analysis to assume it is near 100%. ;-)
The problem is, there is no way of knowing which men are the ones that will offend. As a scientist, I am aware that statistically, it is 99% likely that the man approaching me is not a serial rapist. As a woman, it is not really safe to take that risk. (Not to mention, a man doesn’t have to be a serial rapist to be a threat to women; he could be a one time rapist, or just someone with a horrible temper.)
I do think that it would be unlikely a man sitting next to a child in a full plane, in full sight of everyone, would be unlikely to offend, unless there was a way to do it without being seen or heard. Now, if said man offers to take said child to the bathroom, I think I would definitely find that a problem, even if he is being quite respectable. That is where you’re better to call a flight attendant, who can make sure the child gets to the bathroom safely.
@Nullius
This is a genuine problem that lies at the heart of the dispute between the genderists and the sex realists. I’ve had many arguments with genderists and this line of reasoning is almost invariably at the core of their belief system. The idea that we can make policies that rely on blanket generalizations about the behavioural differences between males and females strikes many as equivalent to making blanket generalizations about the behavioural differences between, say, Blacks and whites.
It’s a strong argument in the genderists’ favour, and we really do need to examine it further.
Personally, I address it in a way that many might find counterintuitive and unexpected coming from a liberal, and in a way that some feminists might find highly unorthodox: I make an appeal to the natural.
A good liberal operates on the assumption that whatever statistical behavioural differences are observed with, say, Blacks vis-à-vis whites, or Jews vis-à-vis non-Jews, or Newfoundlanders vis-à-vis mainlanders, or Poles or Mexicans or Chinese or gays or left-handeds or name your subgroup that carries with it some kind of prejudiced stereotype, that these differences are products of circumstance rather than inbuilt differences. It’s absolutely true that there’s a higher rate of convicted criminals among the African American population versus the white American population, but that’s because of the circumstances surrounding being a Black American, not because Black Americans are innately more criminal. It’s been drilled into all of us that to ascribe innate criminality to Blacks is racist. Because it is!
If the higher criminal conviction rate among Blacks is a product of circumstance rather than innate difference, then it’s a societal problem that we as a society bear collective blame for, and it’s one that we as a society have a collective obligation to fix. This collectivist interpretation of humanism is at the heart of the liberal mindset. isn’t it? (That’s certainly how I see it. It’s what I’ve always understood Western liberal progressivism to mean.)
The behavioural differences between males and females, on the other hand, are intuitively understood to be much more deep-seated. I sense that this is a point of friction and tension within a lot of feminist discourse: there’s a risk that acknowledging the deep-seated, natural differences between male and female human behaviour can also be used to justify blanket rules that deny women their rightful equality in society. Which is why some feminists still adhere to long-discredited “blank-slate” frameworks to explain the fact that sex-biased behavioural differences are consistently observed across cultures and times.
But “blank-slatism” is a double-edged sword. It has in fact contributed to the explosion of gender woo among so many feminists: if womanhood and manhood are just products of our social upbringing, then why can’t a male be a woman or a female be a man?
It’s the fact that male behaviour is fundamentally and consistently different than female behaviour that we allow for laws to openly generalize between men and women in situations that involve child safety and women’s safety. And I should add: it’s male sexual behaviour we’re talking about. A little boy or girl left alone next to a man isn’t likely to get punched or harassed by him — the child is likely to be sexually assaulted by him.
And yet, once we start acknowledging natural differences, a new problem crops up: the one of rationalizing and excuse-making. If men are simply more naturally inclined to let their inner sexual demons out and assault children if given half the chance, that comes awful close to fatalism. It implies that boys will be boys. It implies that rapism and sexual abuse are just a natural part of the human experience. They’re not something we as a society can transcend through the normal channels of human enlightenment. We can’t educate or socialize these ills away, because they weren’t socialized to begin with: they’re inbuilt flaws of the human condition, not social anomalies.
Think about that for more than a few seconds and blank slatism becomes extremely appealing all over again. Don’t we wish male criminal sexual behaviour was just a side effect of bad upbringing, and that if we as a society simply cracked the code of how to rear male children better, in a healthier society, that it would all just melt away?
It’s a conundrum.
Here’s how I see it: yes, male behaviour is innately far more dangerous than female behaviour, in terms of a propensity to violence and a propensity to sexual assault. There is a kind of “original sin” element to male behaviour. No, it’s not socialized. But yes, we can socialize these powerful and dangerous instincts out of men. The whole criminal justice system — a massive edifice that takes up a huge chunk of every single society on Earth’s public funds, all those courthouses and jailhouses and police stations and law firms — is basically built around controlling men’s destructive animal instincts that they are failing to control themselves. It stems from male brains’ limbic systems, which is to say, our animal instincts taking charge over our cerebral reasioning centres. The project of socializing men is to strengthen their cerebral reasoning so that they can override their inbuilt destructive animal instincts.
To go back to Nullius’s comment:
This is the additional premise that I propose. The statistical difference between rates of male and female sexual abuse of minors is not the same as the statistical difference between (say) Black and white criminal behaviour. The difference is that one is indeed inborn and natural and society’s job is to curb it, and the other is not at all inborn and in fact it’s society’s failings that have created it.
I’m fully onboard with that as a hard determinist… Mind you, you could theoretically create a “race” of people that was disproportionately criminal by nature but that’d require an captive breeding program and quite a long period of time.
What gave him away? Was it the pronoun badge? How were they able to identify a “woman” to take his place? Was it the pink dress? The stilettos? The makeup? The head tilt?
It was the head tilt, right?
Once again, how could they possibly know about all these people’s ”inner sense of self”? What test did they do to establish that 97,93 % of suspected sex crimes were perpetrated by people who think/feel/identify/”present”/etc. the same way as ”Elliot” Page? The good news is that those who don’t accept the premises of gender ideology (hence making labels like ”gender identity”, ”man”, ”woman” etc. non-applicable), along with those who identify as ”woman”, ”non-binary” etc., are only responsible for 2,07 % of all suspected sex crimes. Of course this puts the supposed moral superiority of those who do accept gender ideology in a rather awkward light…
Nullius, bear in mind the reasoning you criticise is substantially the same as that which is used to justify separated male and female amenities. The additional element used to justify the separation becoming a matter of law, rather than just the non-binding policy of a company, is that toilets etc. involve various states of undress and have an extra vulnerability to invasions of privacy.
There’s no abuse of statistical logic involved here. The airline manager thinks in terms of risk management. That means likelihood X cost. If you seat children only next to women, you’ve reduced the likelihood of sexual abuse to 1/25 of what it would have been if you ignored the sex of the passenger you seated them next to. Risk mitigated, job done.
Of course, that doesn’t consider the likelihood and cost of hurt fee-fees of big boys who don’t want to accept that they belong to the violent sex.
Nullius, what if the crime only committed by Jews is assassinating Nazi war criminals. Doesn’t that make it rational for Nazi war criminals to simply avoid Jews? Apologies for the fact that in this analogy women have somehow ended up equivalent to Nazis but the additional premises should be obvious enough (and were before we engaged in an act of collective cultural amnesia). Most members of Group A are vulnerable to some members of group B. We don’t know which members of Group B represent the real threat but there is a minimal-real-world-cost method of eliminating the threat. (I don’t consider hurt feelings a serious cost but I’m fully aware that in that I’m out of step with current fashion). The actual threat may be statistically small but the method is effective, proportional, and limited in scope. The only objection I can see is the issue of symbolism. (It’s not like herding all French men called Dominique into some sort of male only concentration camp – yeah, I’ve already invoked Godwin so I might as well go there.)
Nullius in Verba,
I could not reasonably use “89.4% of under-18 DUI arrests were whites” as an argument to convince you that DUI arrests must be drastically reduced. But, if for some reason I wanted to reduce DUI arrests, that 89.4% figure does tell me that I could achieve my goal by barring white teenagers from driving. The issue, then, is to figure out whether my goal is legitimate and matters more than allowing white teenagers to drive.
Suppose that the policy’s intent was to achieve a significant reduction in sex crimes. If so, we would certainly need other statistics to show that the policy would be effective: for instance, what are the odds that a given sex crime is committed by an adult on an unaccompanied minor next to whom he/she was seated in a fully booked flight from Air France? I don’t know for sure, but I’d guess it’s very close to zero, making the policy accordingly inept. The percentage of sex crimes commited by men would be irrelevant here.
However, according to the article, the point of the policy is “to prevent any possibility of predatory behaviour.” Given this, the 98% figure seems like a sufficient rationale for the policy’s content. Of course, it still does not tell us whether it should be followed through or not. What matters more, avoiding a likely very small amount of sex crimes, or not requiring likely very few men to sit in a slightly different location on the basis of their sex? More statistics might be useful here. But even if we had data showing that a hundred percent of men, given the chance, would commit such crimes against unaccompanied minors, we would still have to decide whether preventing the sex crimes is more important than not discriminating. We end up needing to consider both the likelihood of an offense occuring (which is a statistical matter) and the severity of the offense compared to that of the remedy (which is an ethical matter). Air France seems to have gone off the assumption that these sex crimes should always be avoided, hence making the particular rate of offense irrelevant provided that it is nonzero. To illustrate this, let’s suppose that there’s a crime which we know will only be committed by National Geographic subscribers (I’d rather avoid using Jews for this example). What if the crime is to launch nuclear warheads on Russia? We really don’t want a nuclear war to start, so we might decide that we have to discriminate against the subscribers even if 99.9998% never commit the crime.
Relying only on the 97.93% figure isn’t necessarily an abuse of statistical reasoning, although we’d still need to decide whether the policy’s aim is fair or not.
Francis Boyle: P(A|B) [i.e., the probability of A given B] cannot be assumed to be similar to P(B|A). For example, P(S is a Jew given that S is a Nazi assassin) is not necessarily close to P(S is a Nazi assassin given that S is a Jew). It’s the latter probability that determines whether it’s rational for Nazi war criminals to avoid Jews. P(J|N) = 1 in this case, but what is P(N|J)? There are currently around 15.7 million Jews worldwide. How many are assassinating Nazis? Whatever number, it’s going to be divided by 15.7 million, which is probably going to be a whole lot less than 1. (If we say there are a hundred active Jewish assassins of Nazi war criminals, we actually five 9s of reliability that any given Jew isn’t one.) The Nazi might then compare P(S is a Nazi assassin OR S is a murderer GIVEN S is a Jew) to P(S is a murderer GIVEN S is not a Jew). If the Jewish murder rate is less than other populations, Jews may actually still be less of a threat to Nazis despite the fact that 100% of Nazi war criminal assassins are Jews.
“[group] commits [percentage] of [crime]” doesn’t get us to “avoid [group]”. Not on its own. It needs additional premises.
iknklast:
Risk tolerance is a subjective thing. If you objectively assess the odds and subjectively evaluate the resulting probability to be unacceptable, that’s entirely up to you. We all have to set our own thresholds for the degree of risk we’ll allow.
Arty:
I don’t think it is, honestly. We can and should make policy based on generalizable descriptions of behavior—if-and-only-if we actually do good statistics. That means, for example, that any hasty generalizations and base rate fallacies are right out. It also means no mixing up correlation and causation. The fact that eating ice cream is tightly correlated with increased probability of drowning isn’t a good reason for Hawaii to ban ice cream.
I’m reminded of “born this way” rhetoric. What does being born gay or straight have to do with anything? Surely what should matter is that being gay in no way infringes on anyone’s rights and doesn’t make anyone else’s life worse. Justifying policy by appeal to the cause of difference isn’t something I’m comfortable with. (Apologies if you aren’t arguing for that.)
Whether we curb or correct, we are in either case making “policies that rely on blanket generalizations about the behavioral differences between males and females”, or blacks and whites, or young and old. To do nothing, because the differences are natural, is to commit some form of naturalistic fallacy. To insist that there are no natural differences, because natural differences are permanent and immutable, is to transitively make the same error. The important thing is that our analysis must actually correspond to reality. If 99% of a particular group is violently criminal, whether we act shouldn’t depend on whether that criminality is innate or acquired. That the trait is innate or not determines the nature of the policy, not its justification.
There are certainly circumstances that are the fault of society as a whole. There are also circumstances that are the fault of society in part, of a group of individuals, of nature, and of happenstance. We bear collective blame only for those circumstances for which we’re collectively culpable. We very well may bear moral obligation to remedy circumstances of some other origin, but obligation and blame are two very different things, wouldn’t you say?
I was criticizing the fact that there are suppressed premises, not the use of statistical reasoning about regarding populations to determine policy. As I said, “I’m not necessarily saying that the policy is wrong, but the reason given does not per se show that it is. It needs additional premises.”
@Nullius
I agree that the relevant probability here is P(N|J). I also agree that the premises are important. But the premises are given by the context. Implicitly, given that there can’t be many Nazi war criminals still alive, the context in my example is something like Europe in the late 1940s when Jews were understandably doing quite a bit of assassinating Nazis, (P(N|J) was not insignificant though not exactly large). That gives us something equivalent to the threat men pose to women and children. Does Borman flee to Israel or to Argentina? What he is interested in is P(N|R), where R is anyone he might happen to meet. And if P(N|J) is high enough he will choose the easier option of minimising P(J|R), the probability of meeting a Jew.
Here we want to minimise P(A|R) the probability of a child sitting beside an abuser. Since P(A|M) the profitability that a random man is an abuser is large enough that we can’t ignored it I think it’s reasonable to set P(M|R) to zero, when, as it is here, trivially easy to do.
Well, we can’t reliably get the suppressed premises from context, even in your toy example, because I apparently inferred a different context and therefore different premises.
My objection is pretty basic: don’t leave important premises implicit. Doing so encourages poor reasoning. In this case, it encourages the assumption that P(A|B)=P(B|A), an assumption that fuels prejudicial, incorrect stereotyping. The opposing position is what? Do leave important premises implicit? Unless mindreading becomes a commonplace ability, that seems like a bad idea.