The recent spate of race related threads sees one point being repeated reasonably often by the camp opposed to the idea of race - that it makes no sense to divide people on the basis of ‘race’(defined, largely, by skin colour) - as the genetic differences between these ‘races’ are minimal. I have no dog in that fight. But skimming through those threads has made me curious to know if there are sub-populations into which humanity can be sensibly sub-divided. If so, what are they?
Yes, they are called populations. Here’s a wiki article that discusses one such classification scheme - Human genetic clustering - Wikipedia. It’s an inexact science, since boundaries between populations are somewhat arbitrary and are mutable.
One distinct difference across large populations is the fact that homo sapiens interbred with Neanderthals after they migrated out of sub-saharan Africa. Sub-saharan populations with pure ancestry from that area don’t have any Neanderthal DNA while other human populations are hybrids.
The first chart on that page from Cavalli-Sforza is highly misleading. It starts with the common subdivision and works backwards to get highly questionable groupings of divergence.
The two most divergent populations of humans are the Hadza of Tanzania and the Khoi-San peoples of SW Africa. (Interestingly, they both have click languages, which suggests their ancient common ancestor language was also a click language.)
Any chart that throws those groups into one group and separates Asians/North Asians/Amerinds is useless for just about any discussion. Especially the one of interest to the OP.
If you want to classify humans into ~5-6 populations, it’s going to be weird to most people. 5 or so sub-Saharan African groups and then everybody else. The next step separates the New Guinea/Australian folk from the rest, etc. But adding more subdivisions in Eurasia is going to be more than matched by more subdivisions in Africa. If you want to classify into hundreds of populations, knock yourself out. Lots of ways of doing that.
The fundamental problem with discussions of “race” is the problem too many people have of trying to lump all sub-Saharan people into one group. Any reasonable Scientific method based on genes, language, etc., that ends with all those people in one group will have to include everyone.
You would get 5 or so sub-Saharah African populations and then everyone else plus a few Sub-Sahara African groups.
If you look at the Rosenberg K2-K6 graph in the lower right of Telemark’s link, it would appear that you can divide into;
Africans - mostly orange across the board
Eurasians (Europeans, central and southern Asians) - Blue K3-K6, Orange/Purple K2.
East Asians - Pink K4-K6, Purple/blue K3, Purple/Orange K2.
Oceana - Mostly green K5, K6, Pink K4, Purple/blue K3, Purple/Orange K2
Americas 1 - Yellow K6, Purple K2 - K5.
Americas 2 - Purple K2-K6.
Would tend to indicate that Americas 1 is a mutation/subset of Americas 2 and that Oceana is a mutation/subset of East Asian.
Of course, I have no information on how accurate the data is.
Yes, I think the key quote from the Wikipedia article referenced is:
" “Lewontin’s Fallacy” that: “It is not true, as Nature claimed, that ‘two random individuals from any one group are almost as different as any two random individuals from the entire world’” and Risch et al. (2002)"
When you look at clustering of genetic variation, it does seem like humans break out into groups that roughly correspond to old notions of “race.”
And then the recent discovery of non-homosapien genetic material in non-Africans adds an interesting twist, with some people having up to 10% of their genome come from non-homosapiens (the rest is partly from neanderthals and partly from the short-in-stature humanoids whose name I forget)
Homo floresiensis. But there’s no known evidence that homo floresiensis ever cross-bred with sapiens. In fact, I don’t think anybody even envisioned this hypothesis, let alone tried to test it.
Neanderthals are the only ones for whom we have evidence of cross-breeding, and even in this case, it’s a very recent discovery.
Yes. Both the clading diagrams cited upthread suffer to some degree, I think, from postulating clusters based on ethnicity or geographic location and then measuring genetic disparity.
When this topic came up previously I linked to an excellent paper (pdf) on this topic. (Free registration may be required.) Figure 2A shows the result of PCA clustering on a global sample; Figure 2B is the result for just Africans.
(ETA: sciencemag.org/content/324/5930/1035.full.html may be a better, html version.)
It is interesting, as ftg points out, that Hadza and Bushmen are so distant genetically yet speak similar languages. Pygmies seem close to, yet more divergent than, Bushmen in Figure 2B. (AFAIK, nothing is known of Pygmy’s original languages, extincted by the Bantu expansion.)
It is also thought that there was some crossbreeding with the little known Denisovans in Asia in addition to Neanderthals. Markers show up in certain Austronesian populations indicating hybridization with early modern humans as they left Africa.
I have no interest in force-fitting anybody into one group. What are the differences between the Hadza of Tanzania and the Khoi-San of SW Africa? Are these differences mainly genetic? If so, how do these manifest themselves in terms of more…macro-characteristics (for lack of knowledge of a better term) like physical appearance, cognitive/reasoning ability etc?
The terminology you’re looking for is how does Genotype affect Phenotype.
I’ve always wondered why humankind’s long existence in Africa hasn’t allowed for “smearing” of genetic differences through interbreeding, making the populations less distinct?
For that matter, are there any populations, like Icelanders and Tasmanians, who are known to be unable to interbreed?
Awesome, that was exactly the terminology I was looking for. Thanks! 
Now to wait for someone to come along with an answer.
The problem you’re going to have is finding a “pure” sub-Saharan population - consider the level of penetration of non-“pure African” genes into the hinterland e.g. the Jewish ancestry of the Lemba of Southern Africa. Possibly the Khoi-San and other marginal groups are unaffected, but otherwise, there’s not been any genetic isolates in SSAfrica - Arab traders penetrated down to southern Africa on the Indian Ocean side, and ditto into the Sahel/West Africa.
Basically, people are not salamanders - mountains, deserts, oceans - none of these have isolated any human populations, completely enough, for long enough, for there to be anything approaching subspecies. Sub-populations, yes, sure, but lots of those. In the upper hundreds if not thousands. I’d propose that language groups would serve as a good analogue for ancestral sub-populations - look at the variety there. Then notice how they don’t map to traditional races.
Denisovans having been discovered only one year ago or so (or more exactly a couple bones have been discovered), I at first didn’t believe that with a so recent discovery and so few material, evidences of hybridization could have already been found. But according to Wikipedia, you’re right. I’m amazed.
No.
They were discovered much more than one year ago, but you’re right in that very few bones have been discovered. We know very little about them, but all you need is a little bit of DNA to do a lot.
So, you’re saying there ain’t no mountain high enough? Ain’t no river wide enough?
There’s a lot more outward differences in appearance among African peoples than most Americans know about. For example, the non-Bantu people of southern Africa tend to have a lot of folk with reddish hair. Very wide differences in height, of course. So the tall Tutsi live among the short Twa people near the great African lakes. But these are all quite superficial differences.
Lumping all these people together as one, separate, group is very 18th century. Some of these populations were originally quite separate from the others for tens of thousands of years. (The Great Bantu Expansion has blurred the lines quite a bit.)
As to “cognitive/reasoning” abilities. They’re humans. Just as smart as anyone else. Not separate long enough for anything noticeable to arise in the first place and mixed well enough since to blur out anything that might have arisen since. The rise and spread of modern humans 200,000 to 40,000 years ago is far too recent and overwhelming to produce any differences that should matter to anyone.
The problem isn’t that humans can’t be subdivided genetically, it’s that there are an arbitrarily large number of ways to do it. If we assume that people pretty much stay in the same place (see below for when we don’t assume that), then you can distinguish the Scots from the English, the English from the French, the French from the Spanish… In fact, you can distinguish almost any geographic area from another. So, which are the “right” ones?
Now, the assumption I made above is not a good assumption. Just look at All the countries south of the US. The people living there are a complex admixture of European, Native American and African populations. And you have lots of people with every kind of mix you can imagine.