March 6, 2014

Do dances reveal shared ancestry better than languages?

I've been puzzling over what other cultural links might be shared among the Indo-Europeans and Caucasus peoples. Language is the usual one that people discuss, and indeed there's a nice fit between what language predominates in a population and what genepool their ancestors came from.

You can use data to come up with a branching family tree of which language groups are siblings, which are descendants from a common ancestor, and where the forks are in the diagram. You can also do this for what variant there is at neutral sites across the genome, and tell which individuals come from the same group, and which groups are related to which others in what ways. Laying the cultural family tree over top of the genetic family tree shows a remarkable agreement between the two. Genetic variants and languages tend to get passed down together.

And yet there are striking exceptions to this rule, some small and some large. A small example: the Hungarians speak a non-Indo-European language that was introduced into the region by a nomadic group that has long since left, and without contributing much to the local genepool. The Hungarians are genetically peas in a pod with other Indo-European speakers from Central Europe. A large example: most of Latin America speaks either Spanish or Portuguese, despite most of them having a good deal of Amerindian and/or African ancestry.

Thus, while the correlation between linguistic membership and genetic membership is pretty good, we'd like to find some other cultural group marker that does better still.

First we need a quick reminder of why language can lead us astray when we infer who shares ancestry with who else. Quite simply, speaking a language is a highly utilitarian thing. One group might choose to adopt the language of another group if they want to trade with them, if they wanted to apprentice under The Other to learn some skill that cannot be taught without verbal instruction, if they want to serve in higher-status positions that are somehow controlled by The Other, and so on.

Maybe they couldn't care less about the other aspects of The Other's culture, but hey, are they going to turn down the possibility of more food on the table, better skills for themselves and their children, and higher status? There is an adaptive pressure to junk your current language and adopt that of The Other. It'll raise your odds of surviving and thriving.

That is why geneticists try to avoid looking at genes that could be under the pressure of natural selection. A variant that is favored by natural selection could independently pop up in two unrelated groups that face similar adaptive pressures. But if they're similar for a gene that doesn't appear to do anything, and across a large number of genes, that suggests that their shared idiosyncrasies are due to a common ancestor that bequeathed these variants to them, unaffected by selection (though still subject to random mutations that will make them subtly different). That's why it's better to look at neutral genetic variation.

So, which parts of culture are not under such strong selection that they could make two unrelated groups look related? After writing this post on a wedding ritual that is shared among Indo-Europeans and Caucasians, I picked up the scent of looking at dances.

Folk dances can be strikingly conservative. There's a clear sense of what the steps are, what the poses are, who participates, and with who else, etc. Sure, no two occasions of the "same" dance are exactly the same, but they're easy to identify as reflections of a single Platonic ideal dance form. You can't conserve something that doesn't have a strong sense of "this is how it is done." And you get personal, unmediated experience with how it is done over the course of your life -- it's not some body of doctrine that is taught to you and you parrot back, with little to make it stick.

Folk dances are always done in a communal setting, typically during a larger set of festivities, and they serve to glue the group together while they're working themselves up into high spirits. Dances therefore make excellent group membership markers.

And folk dances are not under much, if any, selective pressures to junk the ones you've got and adopt those of The Other. The more fragile areas of culture are those that are involved in the business of making a living -- and verbal communication sure plays a big role there. But how you dance at a wedding or on a holiday that your group celebrates? What difference does that make, beyond signalling group membership?

If you want to learn skills from The Other, trade with them, or move up their economic hierarchy, you don't have to adopt their dance styles. Think of Mexican immigrants in America -- they learn English to some degree, but they don't pay any mind to how Americans dance, and continue right on with their ranchero music and dance. They may even adopt a lot of personal appearance, clothing, grooming, and so on, in order to fit in when they're out in public.

But what about when they're just among themselves? They'll tune our ways out, and stick with their own. And dancing is one of the best examples of group activities that are only done in the company of other group members. You're just letting go too much of yourself, and that's only going to happen around people you really trust -- at least, people you trust not to look at you funny for how you dance, and who won't alienate you when it's their turn to dance.

Another major plus for doing historical research -- music from the past is more likely to be lost than dance forms, which can be described reasonably well in words from contemporary observers, as well as depicted in visual art. There are dancing scenes on pottery going back over 5,000 years.

Well, the proof of the pudding is in the eating. I plan to cover more extensive relationships among groups, as shown by their shared dances, in a little bit. To give one good example now, though, consider the Hungarians again. Looking at language gives the wrong answer about their genetic ancestry. How about a distinctive dance of theirs?

Here is a video of a men's dance called a legenyes, (it starts around 4:30). It's from Hungarians in Transylvania, but in a region that is heavily ethnic Hungarian. It looks like Celtic stepdancing (e.g., the Riverdance video here). By around 5:20, it looks like another famous step dance from nearby, the schuhplattler of the rowdy Alpine Germans (see this video). Looking at dance forms gives us the right answer about who they're related to -- not the nomadic people from the Ural Mountains who dropped by over a millennium ago.

It turns out that stepdancing is practiced much more widely -- among any group with a decently high level of Indo-European or Caucasian genetic ancestry. And hardly or not at all among their southern neighbors, the first farmers of the Fertile Crescent, the Levant, and the Mediterranean -- who of course have their own identifying dance style. That's a story for another time, though. The purpose here is to raise the issues and take a look at one clear example where dance outperforms language as a predictor of shared ancestry.


  1. if you never saw it, here is a film of Romani gypsy song and dance from across the world that may support your thesis or not. worth seeing just for being great

  2. Interesting point about the Hungarians, who speak Magyar. Another unique aspect of them is that they have the highest per-capita of people with green eyes in the world. That is largely from being overrun by the Mongols. All people who have green eyes are of mixed ethnic heritage. But maybe that's another topic for another time.

  3. The Gypsies are going to make an appearance in the post about stepdancing among Indo-Europeans, namely flamenco. It looks strikingly like the classical Indian dance style called kathak.

    Green eyes, as far as I remember, are a mixture of dark and light eyes from the parents. Genetically, Hungarians are a mix of the Northern Euros (mostly Slavic, I believe) and Mediterraneans. So their green eyes could be a result of light eyes from their northern ancestors and dark eyes from their southern ones.

  4. 45% of Brazil is Caucasoid white, as in immigrants from Portugal, Italy, Germany, Ukraine, Poland, Lebanon, Syria. 45% of Brazil is mixed raced. 7% is black African and 0.5% is Amerindian. 2% is East Asian, from Japan and Taiwan Chinese.

    The whites in Brazil average from 80 to 100% in genetics since the influx was mostly white immigration after slavery was banned and during WWI and WWII. The mixed-raced Brazilians on average range from 50% white to around 75% white, and just 50% to 25% black. They're quadroon to octoroons on average.

    Brazil isn't the USA. There was a lot of sexual relations going on between the original colonizers, Portuguese men, who had babies with both female Amerindian slaves and black female slaves. By the time slavery was abolished, mixed-raced children, outnumbered the blacks and Amerindian.

    When the whites arrived, some of those biracials mixed with whites, creating quadroons to octoroons. It was a huge policy in Brazil for the "whitening", or "lightening" of the population, to create "whiter" mixed-raced people.

    They succeeded in part, because the average Brazilian woman looks like Camilla Alves.

    Of course, Brazil is more about "looks" than genetics, so the Camilla Alves look and the Adriana Lima look, is the mixed-raced pardo look. Of course, mulattos are also mixed-raced (think Tais Araujo).

    It's like a caste system. Most mixed-raced people only married each other and "preserved" the bloodlines and looks. Few married "down" (most avoided darker people). It's also quite colorist so people want lighter-skinned grandchildren.

  5. The biggest irony is that Brazil might be the country in South America where the mixed-raced people have high percentage of white blood (averaging 75% to 80% white) thanks to excessive miscegenation during colonial times and more mixtures with white immigrants during WWI and WWII. It's great "blackness/Amerindianess" is exaggerated and nothing more than propaganda.

  6. eye color genetics3/9/14, 10:03 PM

    The blue-green-brown range of eye color involves at least two genes, OCA2 and gey. Possessing the dominant allele for OCA2 confers brown eyes regardless of the what gey allele is possessed. A person with two recessive alleles at OCA2 but one or two dominant gey allele will have green eyes. A person has to have all recessive alleles at both loci to have blue eyes. More genes are certainly involved as this model doesn't explain hazel, grey or violet eyes or varying shades of blue, green and brown. The link below gives a fuller explanation;


You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."