Just addressing one point. Much of what you say I agree with, but once you are talking about medical implications, I think you too easily dismiss race.
First off, what is generally considered invalid is the idea that we can assign all humans to one of four or five discrete non-overlapping categories. All the the genetic and phenotypical…
Just addressing one point. Much of what you say I agree with, but once you are talking about medical implications, I think you too easily dismiss race.
First off, what is generally considered invalid is the idea that we can assign all humans to one of four or five discrete non-overlapping categories. All the the genetic and phenotypical traits tend to follow gradients rather than binary functions. That conception of "race" is too oversimplified at best, and I would be glad to discard it because it causes much harm and little or no good.
However, using multifactoral cluster analysis of human traits, one can discover fuzzy clusters, or density areas where population groups can be distinguished. This is more like distinguishing night from day - the boundaries are fuzzy and not clearly defined, but that doesn't mean that there is no meaningful difference between night time and day time. But the thing is - depending on how you set the parameters, your analysis might distinguish 3 or 14 or 25 statistical clusters. There is no reason to strongly highlight just 4 (or 5). But if you do choose parameter which yield 4 clusters, those clusters will tend to correlate with continents. That should not be surprising - there has been statistically more mixing within continents than between them.
What that means is that once you get rid of the concept of hard bounded categories and accept fuzzy statistical clusters, there could be useful medical distintions between clusters in some cases. Now there's no scientific reason to choose 4 clusters for this - probably it would be more accurate to know the effects of medication on 25 clusters of humanity. But on practical terms, probabalistic medical statistics may have been gathered based on self-identified "race" - the 4 cluster model. So that fuzzy proxy may be the best data we have.
Having defended that possibility of race as an unfortunate but sometimes useful proxy for fuzzy genetic cluster in some contexts, I return to assert that other than for medical or research purposes, I think it would be best to discard the obsolete concept of 4 (or 5) "races" as prescientific folklore. It misleads the mind more often than not.
Alas, getting past "race" is hard to do, especially after people have come to voluntarily intermingle it deeply with their concept of self, which must be defended at all cost. That is, race is used as a proxy for culture. Abolishing "race" as a concept then feels to them like their very identity and personhood is under attack. It's a tough problem - even if we can see that it's corrosive to keep reifying race, it has also become a sacred cow. A twisted love/hate thing.
Ah humans. A most amazing group of critters we are.
"but once you are talking about medical implications, I think you too easily dismiss race."
I'd actually be particularly tempted to dismiss race with regards medical implications precisely because when discussing medical implications we should be accurate. Race isn't.
If we talked about ancestry that would be far more useful medically speaking. If we talked about specific conditions (Chris and I touch on sickle cell anaemia/trait below) that would be useful. If we talked about family history, that would be useful. But if we talk about "black" people or even worse, "brown" people, we waste time and resources on people who have completely different risk factors and needs.
As you say, "race" is an enormously oversimplified way of looking at human complexity. And at the time it was conceived, sure, we couldn't do much better (we also hadn't figured out that bloodletting and phrenology weren't medically sound). But now we can.
I'm not suggesting we ignore our differences at all. But especially in a medical context, I think we should be as precise as possible. Ticking "Black" or "White" on a form is not precise.
"Ticking "Black" or "White" on a form is not precise."
I think we have a lot of agreement - it's definitely not precise.
Where we might differ is that I am suggesting that medicine often operates on correlations and probabilities which are not precise or absolute, and that the data which has been gathered often requires inexact proxies for pragmatic reasons.
So for example, one might notice that there are statistically significant different outcomes among 100,000 patients given some treatment, between those who checked white and black. We would agree that the researchers having a complete genotype of all patients would be much better; but in the real world, imprecise proxies are often all we have or can feasibly obtain, and so we need to nevertheless pay attention to any (fuzzy, not precise) signals that nevertheless rise above the noise floor. It's a lot more practical for researchers to acquire data which includes how a race box was checked, than to get a full genetic sequence of every subject (not just a few markers like 23andMe uses).
AND again, I am speaking of a limited context where pragmatically, race may be the best proxy we have for something which IS relevant. This does NOT generalize to most situations we encounter in our lives, and I am not urging more emphasis be placed on the damaging concept of race.
And by the way, I appreciate your way of looking at things in historical context rather than removing them from context to evaluate them in isolation and under today's light.
Just addressing one point. Much of what you say I agree with, but once you are talking about medical implications, I think you too easily dismiss race.
First off, what is generally considered invalid is the idea that we can assign all humans to one of four or five discrete non-overlapping categories. All the the genetic and phenotypical traits tend to follow gradients rather than binary functions. That conception of "race" is too oversimplified at best, and I would be glad to discard it because it causes much harm and little or no good.
However, using multifactoral cluster analysis of human traits, one can discover fuzzy clusters, or density areas where population groups can be distinguished. This is more like distinguishing night from day - the boundaries are fuzzy and not clearly defined, but that doesn't mean that there is no meaningful difference between night time and day time. But the thing is - depending on how you set the parameters, your analysis might distinguish 3 or 14 or 25 statistical clusters. There is no reason to strongly highlight just 4 (or 5). But if you do choose parameter which yield 4 clusters, those clusters will tend to correlate with continents. That should not be surprising - there has been statistically more mixing within continents than between them.
What that means is that once you get rid of the concept of hard bounded categories and accept fuzzy statistical clusters, there could be useful medical distintions between clusters in some cases. Now there's no scientific reason to choose 4 clusters for this - probably it would be more accurate to know the effects of medication on 25 clusters of humanity. But on practical terms, probabalistic medical statistics may have been gathered based on self-identified "race" - the 4 cluster model. So that fuzzy proxy may be the best data we have.
Having defended that possibility of race as an unfortunate but sometimes useful proxy for fuzzy genetic cluster in some contexts, I return to assert that other than for medical or research purposes, I think it would be best to discard the obsolete concept of 4 (or 5) "races" as prescientific folklore. It misleads the mind more often than not.
Alas, getting past "race" is hard to do, especially after people have come to voluntarily intermingle it deeply with their concept of self, which must be defended at all cost. That is, race is used as a proxy for culture. Abolishing "race" as a concept then feels to them like their very identity and personhood is under attack. It's a tough problem - even if we can see that it's corrosive to keep reifying race, it has also become a sacred cow. A twisted love/hate thing.
Ah humans. A most amazing group of critters we are.
"but once you are talking about medical implications, I think you too easily dismiss race."
I'd actually be particularly tempted to dismiss race with regards medical implications precisely because when discussing medical implications we should be accurate. Race isn't.
If we talked about ancestry that would be far more useful medically speaking. If we talked about specific conditions (Chris and I touch on sickle cell anaemia/trait below) that would be useful. If we talked about family history, that would be useful. But if we talk about "black" people or even worse, "brown" people, we waste time and resources on people who have completely different risk factors and needs.
As you say, "race" is an enormously oversimplified way of looking at human complexity. And at the time it was conceived, sure, we couldn't do much better (we also hadn't figured out that bloodletting and phrenology weren't medically sound). But now we can.
I'm not suggesting we ignore our differences at all. But especially in a medical context, I think we should be as precise as possible. Ticking "Black" or "White" on a form is not precise.
"Ticking "Black" or "White" on a form is not precise."
I think we have a lot of agreement - it's definitely not precise.
Where we might differ is that I am suggesting that medicine often operates on correlations and probabilities which are not precise or absolute, and that the data which has been gathered often requires inexact proxies for pragmatic reasons.
So for example, one might notice that there are statistically significant different outcomes among 100,000 patients given some treatment, between those who checked white and black. We would agree that the researchers having a complete genotype of all patients would be much better; but in the real world, imprecise proxies are often all we have or can feasibly obtain, and so we need to nevertheless pay attention to any (fuzzy, not precise) signals that nevertheless rise above the noise floor. It's a lot more practical for researchers to acquire data which includes how a race box was checked, than to get a full genetic sequence of every subject (not just a few markers like 23andMe uses).
AND again, I am speaking of a limited context where pragmatically, race may be the best proxy we have for something which IS relevant. This does NOT generalize to most situations we encounter in our lives, and I am not urging more emphasis be placed on the damaging concept of race.
And by the way, I appreciate your way of looking at things in historical context rather than removing them from context to evaluate them in isolation and under today's light.