Substack cometh, and lo it is good. (Pricing)

The Belgians did not invent the Hutu and Tutsi ethnic groups, who have different origins

Since I resurrected the analysis of Tutsi genotypes last year I’ve been getting a fair number of emails and messages from people. The issue is that periodically someone, usually, but not always, a white male, will explain that “actually, Tutsis and Hutus aren’t real ethnic groups, and were invented by the Belgian colonialists….” Many people from this region of the world are privately very skeptical of this viewpoint (they tell me so, but don’t want to get into a huge public spat with all-knowing-white-gods). After all, they are from this region, and Hutus are Tutsis are physically often quite distinctively different. They simply do not buy the social constructionist narrative as explaining everything that they see with their own eyes.

But we’ve seen this before, haven’t we? “Well actually, the Lombards weren’t ethnically different from the Romans, they were a Germanized group of mercenaries who created an identity de novo.” Also, “well actually, ‘caste’ is an ancient Indian concept but modern caste-jati groups were reified by the British in the 19th-century….” (genetics tells us both assertions were wrong).

Historiography of the early 21st-century will observe that many white semi-intellectuals took on the metaphorical role of Hamlet in world history, tortured and self-hating souls who put themselves at the center of every dramatic event. All roads lead back to Hamlet.

As it happens, I now have a single Hutu to compare the dozen or so Tutsis to.

Click to enlarge

On the PCA plot above you see the Hutu is near the Luhya and Bantu agriculturalists from Kenya. The Tutsis are shifted toward various Near Eastern populations. Nothing surprising.

Click to enlarge

The TreeMix plot to the right shows that the Tutsi seem to be similar to Ethiopian Jews. Basically, they have some Cushitic origins, even though they now speak a Bantu language (the same language at the Hutu).

We now know more about the process that may have led to this situation due to ancient DNA. Cushitic pastoralists were ancient inhabitants of eastern Africa, displacing and absorbing foragers. But later on, Bantu agriculturalists became demographic dominant. Perhaps the Tutsis were a group of pastoralists integrated into the Hutu social and cultural system?

That being said, there are clear genetic distinctions between the two groups. Ancient DNA Reveals a Multi-Step Spread of the First Herders into Sub-Saharan Africa has an explicit admixture model using ancient DNA (qpAdmin, not Admixture). I pulled the data and re-ran the analysis with Tutsis.

Below you can see one of the models from Ancient DNA Reveals a Multi-Step Spread of the First Herders into Sub-Saharan Africa (table S8). I followed the same outgroup populations as the supplemental methods. I included Luhya to make sure that my answer aligned with the paper. You can see the Tutsi/Hutu/Luhya results in bold. The rest are from the paper.

Test_PopDinkaIsrael_CMota
Tutsi0.6170.1940.189
Hutu0.793-0.0040.211
Luhya1.049-0.0510.002
Aari0.2160.1780.606
Agaw0.3770.5260.098
Amhara0.3430.470.187
Iraqw0.3970.3030.3
Kikuyu0.7830.1060.111
Luhya1.024-0.0530.029
Luo0.961-0.0460.085
Maasai0.6390.1860.175
Mursi0.7480.060.193
Ogiek0.5740.1920.234
Rendille0.5020.360.138
Sengwer0.6520.1270.221
Luxmanda (~3100 BP)0.1810.3710.447
Pemba (~600 BP)1.032-0.0760.044
Kenya PN (Elmenteitan)0.3950.4130.192
Kenya PN (other)0.3630.4020.235
Kenya PN (all)0.3740.4070.219
Kenya possible early pastoralists0.2270.4930.28
Kenya LSA-0.179-0.0491.227
Kenya PIA0.5740.2280.199
Kenya IA (all)0.6540.1660.18

The Tutsis and the Hutu individual are very distinct. The Tutsis are similar to the Maasai, in keeping with my earlier analyses. Here is the Wikipedia entry on Tutsi autosomal genetics:

In general, the Tutsi appear to share a close genetic kinship with neighboring Bantu populations, particularly the Hutus. However, it is unclear whether this similarity is primarily due to extensive genetic exchanges between these communities through intermarriage or whether it ultimately stems from common origins:

I think this is almost certainly wrong. Most people from this area of the world know it is wrong. But there is a concerted effort to impute all Tutsi-Hutu distinctions to the Belgians.

What next? Specifically, I need more Hutu genotypes. So I’m making the call for people to email me at contactgnxp -at- gmail -dot- com.

More generally, this should make readers cautious about experts on ethnographic history. They were wrong about a lot of things in the past generations (caste, the spread of farming and language to Europe, etc.). Right now in the case of the Tutsi-Hutu difference they have engaged a campaign of conscious or unconscious deception.

21 thoughts on “The Belgians did not invent the Hutu and Tutsi ethnic groups, who have different origins

  1. The current Tutsi-run Rwandan government pushes the notion that Tutsi-Hutu differences were socially constructed by malign Europeans for divide and rule purposes. This is kind of like how the Mexican government in the 1920s started pushing the La Raza Cosmica view of Mexicans more or less all as being of the optimal mestizo ancestry.

    That worked out better for Mexico in keeping the peace than the more racially oriented worldview next door in Guatemala.

  2. I have to agree. I dunno if the Rwandan government can pull it off, but if they can, it’ll probably be for the best. Let the white liberals have this one.

  3. Hey Steve, you would have some sources about what you mentioned about Mexico, I really care and I would like if you could provide a source.

  4. I knew something was off when I first heard in school that Tutsi and Hutsi conflict was created by Belgians.

  5. The only distinguishing feature which binds the ‘expert’ opinion together is that it is openly and explicitly anti-white European.
    Without that viewpoint, it is nothing.

  6. Dear Razib,

    Can you point us to a step-by-step guide somewhere to making these PCA plots? I would love to play with them myself. (In this case, it looks like the Yoruba and especially the pygmies are the dominant cause of the 2nd principal component, which looks not so relevant to the story you want to tell. So I wonder if things would look more striking if you dropped those two populations and got a more interesting 2nd component.)

    Anyway, I suppose the key points are (i) where do you download the genomes, (ii) which software do you use, and (iii) how do you use it?

    Keep up the good work! I love it all!

  7. Related question, I’m not really up on the whole history of this, so may be wrong, but I can vaguely recall some waving about that Tutsis had a distinctive “Nilotic” ancestry related to Sudanese. May be wrongly recollecting.

    Does that come to anything, or are the Tutsis just on the Ethiopian->Luhya cline? That seems to be the case in the figure, but there’s no particular Maasai / Sudanese related source there.

    Part of the argument was this was associated with the Tutsis being tall and slim, unlike the Ethiopians who are fairly obviously not particularly tall set against Luhya and East African “Bantu” populations of similar economic development.

  8. tortured and self-hating souls who put themselves at the center of every dramatic event.

    This is just the White Man’s Burden in updated SJW form. The righteous SJW, usually white, points out the bad stuff (tribalism) that was created by bad whites and that the non-whites should do what the SJW whites know to be better and let it go by rejecting a tribalism that is alien and bad.

  9. Ah, just seen the previous post with ADMIXTURE chart of Tutsi, various Ethio-Somali groups, Luhya, and some Nilotic groups (Nuba, Nuer, Shilluk, Dinka), and PCA.

    Eyeballing, looks like closest fitting combination in the panel for Tutsi would be roughly two-way contribution from Somali/Oromo/Amhara+Luhya about 50:50, with no residual contribution from Nilotic groups?

  10. Considering that Tutsi and Maasai are so close on PCA in Razibs figure on the previous post, there must be some pretty interesting sociolinguistics story.

    Tutsi look like simple two-way Cushitic+Bantu, with really no further element at all needed of ancestry from Sudanese (Nilotic?) cluster. The Maasai look to me similar, with about 10-25% Sudanese.

    So makes sense that Tutsi end up speaking a Bantu language, but why and how did Maasai end up with a Nilotic language?

  11. Probably simplest solution would be if Nassau were Anuak like (40:60 Sudanese:Somali/Ethiopian like, speaking Nilotic language) then about 50:50 mixed down with Libya like people?

  12. The datog would be closer to tutsis than maasais. Maybe it’s due to lack of diverse cushitic samples!? Their history does corroborate these results because the barabaig amongst them seem to have been the true cushitic migrants. The Western part of Lake Victoria on both sides Tanzania and Kenya deserves a thorough study. The gambles caves, njoro river caves, gogo falls, Thimlich Ohinga,..In Rwanda the royal tutsi clan of banyiginya(Which is a runyankore word apparently!)are well known to have come from the North, very likely Nkore. I also have a feeling that tutsis are hima clans that migrated further south in Rwanda and Burundi. This can also explain why burundian bahima might have been a recent wave of pastoralists because tutsis marginalised them. So a hima sample analysis would be interesting. I(tutsi with no hima ancestry) also cluster very closely with a hema sample from gedmatch. Hemas are himas who live in Ituri RDC. Some article, years ago claimed that umututsi( tutsi in sing.) is a variation of omuntusi(An inhabitant of the ntusi earthmound). Could be but what is clear is that the population that lived in “Ntusi” and “Bigo bya Mugenyi”earthmounds had identical cultures with the western part of Uganda, Rwanda and Burundi.And herded Ankole cattle. Could they be the first cushitic-speaking folks to enter that region from Western Kenya? We know that the babito dynasty that still rules Kitara-Bunyoro and Tooro are off luo origins and married bahima wives, most likely from Ntusi and Bigo bya Mugenyi. Their monarchs still claim the earthmounds Regalia,artefacts as theirs which very likely belong to the famous bachwezi,who most probably were those bahima living in Ntusi\Bigo.

  13. @Matt
    We can’t say Tutsis are a simple two way admixture. Nilotic/Sudanese admixture is present in them to an equal or greater proportion as Bantu.

    Razib’s Previous post says Tutsis are less Nilotic than Masai. This means they have Nilotic admixture.

    Commercial companies all point to presence of Nilotic admixture. Gedmatch and G-25 support presence of elevated Nilotic admixture. West African/Bantu admixture in Tutsis is somewhere between 10-30%.

    There is also presence of Great Lakes forager admixture in Tutsis that is often ignored. High level of forager haplos like L0f in Tutsis point to that. Even autosomal DNA show significant Hadza-like admixture.

    All that being said, imo, Tutsis are a 4-way admixed population.

  14. About the Belgians. I think we need to distinguish between european ethnographers and anthropologists from colonizers. IMO, Belgian colonial governments weaponized what was already and science of the time had backed up. It’s like today how some politicians are using tribes for their personal gains and cause tribal conflicts/some are becoming genocides while seating in power. Especially in D.R.Congo.
    That is to be condemned, using science to create division.

  15. The Maasai & Luhya populations are themselves heterogeneous. The book ‘Mountain Farmers’ by Thomas Spear gives an interesting example encounter of the interaction between a sedentary agriculturalist Bantu population the ‘Meru’ and migrating Nilo-Cushitic nomadic pastoralist Maasai and the subsequent formation of the ‘Arusha’ tribe who practice sedentary agriculture, speak a Nilotic language with heavy borrowing of Bantu vocabulary but retain Nilo-Cushitic cultural elements like the Age-Set system.
    There could be parallels between the Meru-Maasai/Arusha interaction and that of the Tutsi-Hutu though it would be interesting to know how the Tutsi lost all traces of a supposed Cushitic language and cultural practices like circumcision and Age-Sets.

  16. @Espoir, they have this component in admixture which is “Nilotic” peaking, but this component is not beyond two-way admix of the populations above, and seems likely to be mediated by Horn populations. Hence the question.

    35-40% of ultimate ancestry from West Africa seems about right judging by components and PCA position, 10% far too low. (About 25% of “West African” component in Razibs ADMIxtUre, which peaks about 60% in Yoruba. Ergo 25/0.6= 41% Yoruba like ancestry. Remaining ancestry pretty much close to Somali.).

  17. I read on Wikipedia that Tutsis have a lot of B-M112 / B2b (sorry I don’t know the nomenclature) which is highly pygmy / central African hunter-gatherer. I find it remarkable that on autosomal terms, they don’t seem very Twa-like.

  18. @Anonymous Coward

    I am afraid Luis and Tishkoff studies may not be representative of Rwandans/Burundians/Banyamulenge.

    E-M2 is most common according to the only study that was done on Rwandans by Luis 2004. The Tutsi dna study from Luis 2004 seems not complete. He doesn’t talk much about autosomal dna, wikipedia(anyone can edit on wiki) which is not reliable is generalizing without any proof. Also, it would be interesting to know where Luis 2004 got his sample, I doubt those sample were collected in Rwanda due to the fact that Hutu/Tutsi subject is a no go area in Rwanda. Noway the Gov of Rwanda would have allowed him to do dna study in the Rwanda. My suspicion is Luis may have collected sample from Rwandans from Uganda which is very open about dna. The other autosomal study is from Tishkoff which collected samples of 8 rwandans(residing in Cameroun)of mixed Hutu/Tutsi ancestries.

    Commercial dna companies like 23andme/ancestrydna, there are at least 50 Rwandans/Burundians/Bananyamulenge who calls themselves Tutsi who tested with 23andme. And it seems E-M293 is the most common haplogroup at ~41% followed closely by E-M2 and A-M13. Female haplogroups are overwhelmingly East africans with L0f being most common.
    Commercial companies are showing Tutsi autosomal they are predominently East Africans ~ 60-85% which matches the results of Gedmatch/G25. It is great to see that Razib Khan is confirming what is already known.

  19. There is an old colonial book that I read into whose author affjrms the local karagwe himas claimed they head lost their old language.The bahima in Ankole to this day still have a distinctive accent.

  20. In our region marrying your cross cousin was very common till recently,encouraged when wealthy or off higher status. Amongst royals even more so.I wonder the effects on genotype or phenotype!?

Comments are closed.