The Off Topic Topic

[garbon]

Beeb, as someone who stereotypes for a living, I think you're absolutely right in your reasoning, but you'll never win that argument (especially with the people you choose to argue with).  Ability to generalize might literally be the definition of intelligence, but it's also a definition of stupidity when virtue signaling is the name of the game.

Why would a person generalise about a specific individual they have had sustained contact with? Why would you judge them off one or two traits (likes theatre and has no girlfrieds, really likes sports) rather than viewing them as a whole individual? After all, you have the luxury of time to get to know them (observe them?) and judgments off a wealth of experience (data?).

You're right - I should never generalize about anyone.  Ever.

Now I'll just go sit in the box and feel shame.

Odd given DGuller understood what I was saying.

[Admiral Yi]

Odd given DGuller understood what I was saying.

DGuller was happy you legitimized generalization and distinguished it from stereotyping.  Beeb was unhappy (justifiably IMO) you denigrated his generalizations.

[Jacob]

So pattern recognition is a real thing and obviously very important in all kinds of contexts, including how we navigate our societies and evaluate the different people and situations we come across.

At the same time we operate in social contexts that influence how and when it's appropriate to apply our pattern recognition. And then we have to correct for confirmation bias, and for social influences. Am I taking someone else's word for this pattern existing? How important do I think the pattern is in actual practice? Do I emphasize (or de-emphasize) this pattern's existence to signal group membership or political allegiance? How do I choose to let this pattern influence my interaction with actual individuals? How do I talk about this pattern in different contexts (at work, with friends, with strangers, with people who are or are not included in the pattern)?

It's almost always nuanced, and sometimes contentious.

Perhaps a bit of an anodyne statement, but it's pretty clear that contention around different forms of pattern recognition (are these patterns real, do they matter, how do we apply them, how do we communicate about them) underpin some of the biggest hot topics of our current societies - so I think it's worthwhile to identify some of the dynamics in a relatively neutral form to maybe facilitate discussion.

...

... to offer some specific examples:

1: The different mortality rates - life expectancy - of men and women is a well established actuarial fact. Women tend to live longer than men, and when you're planning things like life insurance, pension payments, projecting health care costs, and so on that difference in mortality is salient. So looking at that fact and pension payments is it fair that men and women get the same total compensation over their lifetimes (so men get more on a monthly basis because, on average, they'll die first; therefore women have a lower standard of living than men), or is it fair that you get the same regardless of gender (so on average men receive a smaller total payout than women, but everyone gets the same monthly payout and therefore standard of living)?

2: You're a young woman walking home from the train station late one evening after a night out. You notice someone in your general vicinity approaching you. How anxious should you be? Do you take any actions for reasons of safety? What are appropriate actions to take? Does that change according to that someone's gender / apparent socio-economic status / age / ethnicity is (or whether they're a bear)? Quite probably, but how you discuss it may very well be different from your actual actions depending on who you are and on your audience.

3: You're a man walking somewhere on a Friday night. You see a gaggle of young people in the middle of their night out and you'll potentially cross paths. Does your level of stress and your actions change depending on whether they're all male, all female, a mixed group? Does it change based on apparent socio-economic status? Sub-cultural affiliation? Ethnicity? Nationality? Quite possibly, but most likely you'll talk about some of those factors differently depending on what they are and on an your audience.

4: Immigration is a major topic in the West (and it'll likely only get bigger). There are absolutely patterns you can recognize when it comes to different groups of immigrants in different places - but which ones to discuss and how to is massively fraught. Which patterns are real and significant, and which are overstated for political (or personal) reasons? Which patterns actually matter? What and how do we respond to patterns that are different than those prevalent in the country to which they've immigrated? What patterns are actually great, which are problematic (and why), and what do we do (if anything) about problematic patterns? Those conversations are a significant driver of the politics in a number of countries.

5: To connect to Beeb's post earlier - there are a handful of kids in my daughter's daycare who if in 10-15-20 years I learn that they've come out as trans and/or non-heterosexual, my response would be "yeah that's not really surprising at all" based on what I've seen from them so far. Please note, it's not predictive (kids do all kinds of things) and it doesn't matter to me in any real way (I only interact with the kids in passing and probably won't know them once my daughter leaves the daycare). Nonetheless, there are indicators of conforming to patterns that's established outside of that context that would lead me to be "not surprised at all". But as the conversation upthread has demonstrated, there are potential pitfalls in how (and with whom) I discuss those patterns (for the record, this is the first time I've brought it up with anyone).

[DGuller]

I think you're getting at the most challenging issue in this conversation.  How do deal with generalizations which are statistically valid but are unfortunate, on a societal level?  Do you mandate, by cultural norms or otherwise, that you don't mention them even if you're thinking about them?  Or do you make it okay to mention them, if you can back it up, but not act on them?  Or is okay to act on them, if it's a proper generalization and not an over-generalization?

I think it's a tough question.  On the one hand, philosophically I believe that very rarely do good things come out of suppressing speech (by any means, not just legal).  It's a toxic environment where people have one thing on their mind, but are mandated to say something different.  On the other hand, generalizations can be self-reinforcing:  they may become even more true, or continue being true, precisely because they are believed and acted on.  By not censoring the expression of generalizations, even those that are true and expressed at the right level of certainty, you are perpetuating them.

It's an interesting conversation to have, as long as you have it with people who have it in good faith and are motivated purely by wanting to discover where the best solution may lie.

[crazy canuck]

The problem is of course when people claim their views are valid observations when they are demonstrably false.

Like "Lesbians like sports more than straight women - on average"

[Jacob]

I think you're getting at the most challenging issue in this conversation.  How do deal with generalizations which are statistically valid but are unfortunate, on a societal level?  Do you mandate, by cultural norms or otherwise, that you don't mention them even if you're thinking about them?  Or do you make it okay to mention them, if you can back it up, but not act on them?  Or is okay to act on them, if it's a proper generalization and not an over-generalization?

I think it's a tough question.  On the one hand, philosophically I believe that very rarely do good things come out of suppressing speech (by any means, not just legal).  It's a toxic environment where people have one thing on their mind, but are mandated to say something different.  On the other hand, generalizations can be self-reinforcing:  they may become even more true, or continue being true, precisely because they are believed and acted on.  By not censoring the expression of generalizations, even those that are true and expressed at the right level of certainty, you are perpetuating them.

It's an interesting conversation to have, as long as you have it with people who have it in good faith and are motivated purely by wanting to discover where the best solution may lie.

Yeah definitely interesting and IMO relevant, but also fraught.

A few things:

I think you're right that the (a) big challenge is what to do when something is statistically significant but also unfortunate on a social level, BUT... IMO there's still a massive space to be negotiated between "statistically significant" and "appropriate action" (whether on an individual level, policy level, or informal but social level).

Like say (and I'm no statistician here, so please change the numbers to make things "statistically significant" if my example numbers do not). Say there's a [problematic thing] that has a prevalence of 0.1% in the general population, and 1% among a [specific group] that's statistically significant (it's 10 times greater), but what actions does that suggest? How do we talk about it. I imagine that continually linking [specific group] with [problematic thing] in the public discourse is going to be tedious in the extreme for the 99% of [specific group] who don't engage in it. Same thing if the numbers were 30% among the general population vs 45% among [specific group] - statistically significant, but to what degree does it justify radically different approach in how we discuss it? Because in my experience, with that sort of difference it's not infrequent at all that the problem in the general population is basically ignored ("of course it's a problem, but look at those people, they're the actual problem..."), while it's framed as an important problem to solve for the [specific group] and treated as basically inherent to them. That can easily drive alienation, polarization, become self-fulfilling, and/ or spark undesirable reaction.

RE: censorship... I have a number of thoughts re: the exercise of power and the framing of conversations, but I'm not 100% ready to articulate them right now. I will say, however, that I don't think it's a binary censored vs uncensored situation, but much more about how and when to talk about things. I think we can all trivially think of facts (and "facts") which have been used as cudgels (politically, interpersonally) widely out of proportion to their actual substance if you drilled into them dispassionately (or even with a bit of good will).

EDIT: ... and this is, of course, assuming that the underlying statistic is not somehow accidentally (or deliberately) biased.

[Josquius]

The big problem with making assumptions based on stats is often people will grasp the wrong stat and make the wrong assumptions.

A common one in the UK that fascists like to repeat that white working class men have it hardest of all. Statistically their school performance is lowest.... But what the stats actually show is crappy small towns, which for obvious reasons have never gotten much immigration, tend to perform worst. The white working class part is irrelevant. When you get a minority in those areas they are just as poorly served.

The problem is of course when people claim their views are valid observations when they are demonstrably false.

Like "Lesbians like sports more than straight women - on average"

I'm not sure this is demonstrably false. Gay women do tend to be pretty over represented in sports as compared to their numbers in the general population.

Does this mean if a girl is into sport she's probably gay? Of course not. The majority are still straight.
But the gay percentage does seem to be bigger in sporting circles - with men it being quite the opposite.

The reasons for this are where the interesting things to think about come. To what extent is it due to something "natural" to gay people vs the stereotypes themselves and over the years spaces associated with sports becoming attractive to lesbians and unattractive to gay guys.
It is my theory that the reason you don't see many gay male  pro footballers is at a grass roots homophobic viewpoints are too common so any who might have talent are just turned off.

[Sheilbh]

The big problem with making assumptions based on stats is often people will grasp the wrong stat and make the wrong assumptions.

A common one in the UK that fascists like to repeat that white working class men have it hardest of all. Statistically their school performance is lowest.... But what the stats actually show is crappy small towns, which for obvious reasons have never gotten much immigration, tend to perform worst. The white working class part is irrelevant. When you get a minority in those areas they are just as poorly served.

But doesn't that also just reflect a perhaps peculiarly British view that you can separate out class from ethnicity and place.

And to move it the other way of generalisations based on data that can be harmful. It is also still very contextual. I've read Stephen Bush, a black British journalist with an interest in education, and Sam Freedman, an education wonk, both describe talking to American audiences of education specialists. They have simply not been believed for stating the fact that black British kids are, on average, more likely to leave school with good qualifications and more likely to go to university than white British kids (though, they note, there is a big divergence between African and Caribbean heritage). That's because it is different in the US but also the meaning that fact assumes is different.

Facts are facts but they're also meaningless in their own way (this is one of the reasons I'm dubious about the misinformation/disinformation/epistemological diagnosis of our problems). They acquire meaning through interpretation - and we all operate on both levels the factual and meaning-making.

[Barrister]

I don't know why you guys are bothering to continue on this topic.

You clearly can't assume anything about anyone.  It's so obvious to me now.  Generalizations are evil.

[Syt]

Statistical generalizations can be useful information when looking at sufficiebtly large groups of people, if you correlate enough data (e.g. is group x more prone to y intrinsically or because of z which is a more general indicator for y, and if so, why is group x more related to z?), but are a crapshoot when dealing with individuals, esp. when they step out of "their group".

[DGuller]

They can't be a complete crapshoot at individual level, or else they wouldn't aggregate to a useful information at a group level.

[Syt]

It only gives you a probability about a person, but not facts about them.

If 75% of all people with blonde hair and green eyes like eggs for breakfast, then there's a better than even chance a random person with blonde hair and green eyes that you meet will like eggs for breakfast, but you won't know until you ask them.

[Admiral Yi]

It only gives you a probability about a person, but not facts about them.

If 75% of all people with blonde hair and green eyes like eggs for breakfast, then there's a better than even chance a random person with blonde hair and green eyes that you meet will like eggs for breakfast, but you won't know until you ask them.

Probabilities are useful.  Most of the decision we make in life are probabilistic.

[Valmy]

It only gives you a probability about a person, but not facts about them.

If 75% of all people with blonde hair and green eyes like eggs for breakfast, then there's a better than even chance a random person with blonde hair and green eyes that you meet will like eggs for breakfast, but you won't know until you ask them.

Probabilities are useful.  Most of the decision we make in life are probabilistic.

Yeah but are the generalities actually based on facts on the ground? For decades most American generalities about Europe were based on ones that came back with the GIs from World War II. Ridiculously outdated and inaccurate. I just don't think many of our generalities are actually fact checked with actual probabilities. They rather have a "truthiness" and just sort of feel right, hence they are prone to lead us in the wrong direction.

[The Brain]

Society will grind to a halt if people have to ascertain absolute knowledge about other individuals when dealing with them. Do I know if the airline pilot knows what they're doing, or should I cancel my flight? Or should I assume that they do, and relax?