Human races do exist in nature
- Thread starter Hans
- Start date
A biconditional is an "if and only if" statement. They are very rarely used outside of philosophy and mathematics, as they tend to not hold up in the real world. Instead, when we speak of "rules" we really speak of probabilities within a range of values subject to our typical experience of something. In other words, when 95% of observations are within two std deviations of the mean, we tend to agree that our "rule" is good enough in practice to be considered a rule. Typical descriptions of oxygen (colorless, odorless gas, flammable) are true, until you supercool it and you get a magnetic liquid. Dogs have a certain distribution of behaviors that we characterize as "doglike". Just because one dog can ride a skateboard doesn't mean all dogs ride skateboards, that riding a skateboard is doglike, or that all our previous characterizations of doglike behavior are false.
Consider some drug treatments that only work half the time. Since they don't always treat x diseases, and sometimes treat y disease, does that mean they don't actually work? Just like we can generalize that people from Chicago tend to like the Bears more than people from Miami. Not all Chicagoites are Bears fans, nor are all Miamians non-Bear fans. Since I'm not constructing biconditionals, but working with statistical information, we're not making any logical fallacies.
Yeah, those genetic factors constitute 'race'. Or 'populations' if we want to avoid the racism hysterics that the word launches people into.
http://dceg.cancer.gov/files/genomicscourse/genes-112001.pdf
Race isn't necessarily skin-color, but it is there. Enough variation exists that we need to take it into account. This doesn't give people the license to hate, but ignoring reality in order to construct a fantasy-land of Leftist nonsense does people real harm.
There are several populations of black people that differ from each other much more than they each differ from white people as a population.roman said:
It might be possible to distinguish human races on some kind of genetic basis, as the OP claimed. The problem is that they will not correspond very well to the socially established "races" as recognized today.
Hence my question. The OP as presented for racial classification seems to imply either:
at least 7 or 8 races (dominant colors in the bar chart), be far from exhuastive (Africa and Asia under-represented, Australia missing), and classify Micronesians in a completely different race from any African, contrary to American custom ( and show geographic single allele, not multifactoral correspondence, "clustering" (which would be geographic "clustering", not genetic) - so we would be presumably up around a dozen "races" with this approach, and all of sub-Saharan Africa in one race - which is not reasonable, genetically - - -
or, by the tree diagram,
at the level at which the Chinese are one race, at least twenty different races with the French and Russians racially distinct.
Another way to see what is uninformative about this kind of cladistic approach based on alleles for "racial" classification is to work it backwards, collecting people into "races" at each level of branching as you go. At one point in the OP picture you would have the Yoruba, Bantu, and Mandenka in three separate races, with the French, all Chinese, Mayans, and Papuans, in one race.
Related question for any pros around: If a more reasonable multifactoral kind of Detrended Correspondence Analysis clustering were used to ID a "European" race, how many races would it ID overall?
This efficacy depends on genetic heritage and sociological circumstance. These are only correlated with the sociological races in limited areas among limited populations. If you assume that a drug that works a certain way on inner city black slave descendents in Atlanta will work the same way on Somalian refugees in New Jersey, you will be back to killing people.swivel said:
I never said race had anything to do with skin color. Look at the post above this one.
Race has to do with genetic relatedness. Skin color has a high correlation to race, but it isn't determinate.
The problem with race, subspecies or anything below the species level is that they are not denominators that correspond to a distinction in nature. Even the species classification that we so conveniently used is not directly reflected in nature, especially with respect to prokaryotes, but even in mammals it is problematic. That being said, making a classification that is even fuzzier than species.
That being said, we can create subpopulations according to a specific trait, however it must be clear that these are not taxonomic units, merely a distinction under a specific viewpoint.
For example, if we use sickle-cell anemia as a marker we will likely find that populations in heavily infected malaria regions will have a higher share of this particular trait. That means that concentrating on this trait will populations in malaria-regions to be classified together, even if they are genetically (overall) more distant to each other than respective populations that are near to the malaria regions but who are not under this particular selective pressure.
Like it or not, race implies a certain taxonomic classification with a precision that we do not have. As I mentioned earlier whole genome comparisons, might be able to give more information that may make stable classifications possible, but as the species concept is more and more collapsing, I kind of doubt that.
That being said, we can create subpopulations according to a specific trait, however it must be clear that these are not taxonomic units, merely a distinction under a specific viewpoint.
For example, if we use sickle-cell anemia as a marker we will likely find that populations in heavily infected malaria regions will have a higher share of this particular trait. That means that concentrating on this trait will populations in malaria-regions to be classified together, even if they are genetically (overall) more distant to each other than respective populations that are near to the malaria regions but who are not under this particular selective pressure.
Like it or not, race implies a certain taxonomic classification with a precision that we do not have. As I mentioned earlier whole genome comparisons, might be able to give more information that may make stable classifications possible, but as the species concept is more and more collapsing, I kind of doubt that.
That's why population geneticist use neutral markers.
This is the only area of population genetics where we "don't know enough." Somehow we know enough to do population genetics on leopards, martens, owls, geese, salmon and whales, but not humans? What makes humans an exception? Why can't we look at human populations the same way we do with other animals? Because it might offend someone's sensibilities?
No it doesn't. Race is a social construct. Population differences, genetic or otherwise, are called just that: population differences. Certain population differences are used in the construction of some racial categories, but they need not be and, regardless, the overwhelming majority of genetic differences are completely ignored in the construction of races.
This stuff is obvious when you look at how the construction of race varies over time and place, without any corresponding variance in the genetics of the populations.
Enmos
Valued Senior Member
In other words, it doesn't really exist. Thank you..
We can do this with all species. But the the relevance is confounded to specific markers in every case. There are no truly neutral markers, only markers from which we do not know under what kind of selective pressures there are. Even using neutral markers we find a lot of divergence in mutation rates. Also if we categorize animals according to certain markers, these distinction are not absolute, as I already said, even the once-believed true distinctions on the species levels are being contested right now.
So can we categorize populations? Sure we can. In some cases it makes a lot of sense, like if you correlate specific markers with susceptibility of certain drugs (AKA pharmakogenomics or personalized medicine).
But you should not confuse it with a true taxonomic distinction. This are just denominators that we come up with, which is not necessarily reflected by nature.
Again, to see whether those marker based distinctions are truly correct one would need to analyze a big chunk of the whole genome in try to correlate that with individual markers. There are other approaches (e.g. DNA barcoding) but they would only work on the species level, not below.
See, what I think you are confusing is population genetics with taxonomy.
Population genetics deal with variations between and among populations but is not directly involved in building up taxonomic entities as molecular phylogenetics.
Just to reiterate, everything below the species level (and quite possibly the species itself) is a construct. It does not mean that is is useless, though.
Last edited:
Well if no one is talking about the sociological races based on skin color, we're back to the question of what difference groups we are talking about, in the OP, and why we would want to call them "racial" classifications.swivel said:
You were talking about racial differences in drug efficacy (noting the warnings by first Hercules succinctly and then less pointedly me, about the tricky nature of such things), so you must have some racial classification method in mind - how does it work?
One question yet unanswered, and possibly unanswerable here, is how many races these cladistic - or, as would be my preference, ordination- methods, applied to human genetic data, would typically yield at the resolution necessary to put "European" people in one race by themselves.
Last edited:
The answer is surprisingly easy: it depends what you use as observations in your clustering algorithm.
Also with regards to K (that is the number of clusters) this is a value is at least partially chosen. Normally you would cluster your data (or observations) according to a number of Ks (e.g. 2-20). After that there are several approaches to estimate the "true" K. One is a comparison the probability of each outcome and find the point where the relative change maximizes, for instance.
The disadvantages of this approach are two-fold. First, it is computationally expensive, so that genome-wide studies are problematic, and it requires the explicit assumptions.
What is kind of annoying is that I cannot track down the original paper. The first half of the sentence of this OP makes sense
But what comes after that is not something I would expect to find in a normal publication. Has anyone had any success in finding the publication in question?
Edit: figures, two minutes after I posted this I come across the publication from which the figures were ripped:
Rosenberg et al. 2002 Science
As expected the word "race" did not appear once in the article. Also note the lack of diagnostic genotypes (in the main text) as well as the inter-population vs. intrapopulation differences.
Also with regards to K (that is the number of clusters) this is a value is at least partially chosen. Normally you would cluster your data (or observations) according to a number of Ks (e.g. 2-20). After that there are several approaches to estimate the "true" K. One is a comparison the probability of each outcome and find the point where the relative change maximizes, for instance.
The disadvantages of this approach are two-fold. First, it is computationally expensive, so that genome-wide studies are problematic, and it requires the explicit assumptions.
What is kind of annoying is that I cannot track down the original paper. The first half of the sentence of this OP makes sense
, as well as the graphs.
But what comes after that is not something I would expect to find in a normal publication. Has anyone had any success in finding the publication in question?
Edit: figures, two minutes after I posted this I come across the publication from which the figures were ripped:
Rosenberg et al. 2002 Science
As expected the word "race" did not appear once in the article. Also note the lack of diagnostic genotypes (in the main text) as well as the inter-population vs. intrapopulation differences.
Last edited:
No, race most certainly exists. Just ask the nearest minority.
But if you mean "race is a social construct with no meaningful genetic basis," then, yeah, I agree.
Enmos
Valued Senior Member
You mean prejudice.
That's what I meant
And that is invalid usage of the word..
Sure, it's no problem to apply population genetics to humans. Scientists do that all the time, and learn interesting things.
The problem is when people confuse this with race, which is totally different, and only tangentially related.
No, I mean race. You have to have different races in order for (race-based) prejudice to exist in the first place.
Enmos
Valued Senior Member
No, all you need is differences in phenotypes.