4 ms·

As a concept "privilege" is useless except as rhetoric to get what you want against your supposedly more privileged enemies. I think your defending language is

by Jach 10y ago

As a concept "privilege" is useless except as rhetoric to get what you want against your supposedly more privileged enemies. I think your defending language is sloppy here: ""Male privilege" doesn't mean "your life must be great", it means on average you have more and better opportunities that may or may not result in a better life." That's not what it is supposed to mean at all, it's supposed to mean men on average (not you as a man on average) have better opportunities etc., and that on observing that you are a man, knowing nothing else about you, I can infer in a bayesian way that you are more likely than not to have better opportunities etc. There are many things I can infer depending on what my priors are and what I learn about you. But when dealing with individuals, priors on such broad factors like race/gender/religion become useless very quickly since you learn so many more specific details about the individual that are far more important than those broad attributes. Additionally not all attributes have the same distribution. So what if the average man has more opportunities etc. if when you look at the distribution no individual man fits the average? Distributions matter: https://en.wikipedia.org/wiki/Anscombe's_quartet https://en.wikipedia.org/wiki/Anscombe's_quartet Drilling into the concept, "privilege" is nothing more than "attributes" we happen to find external positive correlations for. There are many, many such attributes, people are multi-dimensional complex beings, and there are many attribute combinations that interfere both constructively and destructively. What annoys me more than "privilege" advocates focusing so much on broad attributes like gender (and trying to instill feelings of guilt -- you might not do it but it is very common) is ignoring the equally valid concept of "detriment" (except sometimes in the special case of having "not this privilege"), i.e. attributes for which you can find negative correlations. Note that many attributes have both! Being tall is nice because you can reach high things, but sucks because flying anywhere is basically mini-torture. And there are many more positive and negative correlations I can list, just for being tall. This gives me prior information for inference when I observe a tall person, but my inferences should not be very strong because of all the other attributes I don't know about or am not considering.

burkaman 10y ago
I agree with pretty much everything in this comment, except the word "enemies". We're discussing emergent trends in society that can't be blamed on any group of (living) individuals, it doesn't have to be framed as a conflict. I guess I don't really see where we disagree. Of course there's more to life than your gender, of course other things matter, of course these various attributes are complicated and aren't purely beneficial. None of this means we have to stop talking about gender, it's not like nobody talks about class and education and culture and whatever else. All of it matters, we should talk about all of it. Can you expand on this point: > So what if the average man has more opportunities etc. if when you look at the distribution no individual man fits the average? I haven't really done any statistics since high school so I'm probably missing your point.
- Jach 10y ago
  Where we disagree is probably in using the language of "privilege" and related concepts. I see no point to it, and it's probably harmful in the long run. It's imprecise (typically dealing only with vague averages), ignores detriments, ignores privileges when they imply the existence of detriments (i.e. there is certainly female privilege, but it is often ignored because that implies that some male privilege could be counter-balanced to some extent by some male detriment, i.e. not being a female), applies poorly to individuals (because they are so multi-dimensional), ignores complicated combinations in favor of simple ones that seem only to infer positive correlation (white cis male etc...), and is championed most loudly by means of pathos appeals. Why bother with that framework? I don't think we should stop talking about gender, or class, or race, but we ought to be more careful about it, especially when we try to assign causes, and especially when we talk about individuals. Focus more on ethos and logos. Use full data sets, talk about positive and negative, talk about interference both constructive and destructive. Make predictions and update when your predictions are falsified. Make predictions about data before you look at the data to see if your priors are well-calibrated. Talk about consequences more than feelings. For the last point, it's mostly a complaint about using a (mean) average for everything without considering whether the underlying distribution makes such a statistic worthwhile to use or not. Consider a block of houses valued roughly the same amount, then some rich dude comes in, knocks one down, builds a towering mega-house that hits the limits of what is legally allowed, and then sells it to some other rich dude. His house has bumped the (mean) average house value of the neighborhood up considerably, but no house in the neighborhood is actually worth the average. Talking about the (mean) average is then pretty useless, and no one is (literally) going to buy into one of the normal houses being worth the average of the neighborhood. The median average is much more useful, at least for property values. But even then you'll want to look at the full distribution to say whether the median is really more useful or not, and when you're buying or selling a specific individual house, you're going to want even more details about that specific house that could drastically change the valuation one way or another.
  - burkaman 10y ago
    Do you have any suggestions for how to communicate this stuff better? I certainly understand objections to a slogan like "equal pay for equal work" or "check your privilege" or whatever, but I also understand that there are real issues affecting millions of people and discussing statistical nuances is probably not the most efficient way to solve them. I'm not trying to be snarky, I'm honestly wondering if you've seen any more effective approaches or have any ideas.

by Jach 10y ago

burkaman 10y ago

I agree with pretty much everything in this comment, except the word "enemies". We're discussing emergent trends in society that can't be blamed on any group of (living) individuals, it doesn't have to be framed as a conflict. I guess I don't really see where we disagree. Of course there's more to life than your gender, of course other things matter, of course these various attributes are complicated and aren't purely beneficial. None of this means we have to stop talking about gender, it's not like nobody talks about class and education and culture and whatever else. All of it matters, we should talk about all of it. Can you expand on this point: > So what if the average man has more opportunities etc. if when you look at the distribution no individual man fits the average? I haven't really done any statistics since high school so I'm probably missing your point.

Jach 10y ago
Where we disagree is probably in using the language of "privilege" and related concepts. I see no point to it, and it's probably harmful in the long run. It's imprecise (typically dealing only with vague averages), ignores detriments, ignores privileges when they imply the existence of detriments (i.e. there is certainly female privilege, but it is often ignored because that implies that some male privilege could be counter-balanced to some extent by some male detriment, i.e. not being a female), applies poorly to individuals (because they are so multi-dimensional), ignores complicated combinations in favor of simple ones that seem only to infer positive correlation (white cis male etc...), and is championed most loudly by means of pathos appeals. Why bother with that framework? I don't think we should stop talking about gender, or class, or race, but we ought to be more careful about it, especially when we try to assign causes, and especially when we talk about individuals. Focus more on ethos and logos. Use full data sets, talk about positive and negative, talk about interference both constructive and destructive. Make predictions and update when your predictions are falsified. Make predictions about data before you look at the data to see if your priors are well-calibrated. Talk about consequences more than feelings. For the last point, it's mostly a complaint about using a (mean) average for everything without considering whether the underlying distribution makes such a statistic worthwhile to use or not. Consider a block of houses valued roughly the same amount, then some rich dude comes in, knocks one down, builds a towering mega-house that hits the limits of what is legally allowed, and then sells it to some other rich dude. His house has bumped the (mean) average house value of the neighborhood up considerably, but no house in the neighborhood is actually worth the average. Talking about the (mean) average is then pretty useless, and no one is (literally) going to buy into one of the normal houses being worth the average of the neighborhood. The median average is much more useful, at least for property values. But even then you'll want to look at the full distribution to say whether the median is really more useful or not, and when you're buying or selling a specific individual house, you're going to want even more details about that specific house that could drastically change the valuation one way or another.
- burkaman 10y ago
  Do you have any suggestions for how to communicate this stuff better? I certainly understand objections to a slogan like "equal pay for equal work" or "check your privilege" or whatever, but I also understand that there are real issues affecting millions of people and discussing statistical nuances is probably not the most efficient way to solve them. I'm not trying to be snarky, I'm honestly wondering if you've seen any more effective approaches or have any ideas.