Sunday, April 09, 2006

Is Vendela distantly kin to Jaakkeli?   posted by Razib @ 4/09/2006 03:43:00 AM

I'm a bit confused. Offhand I stumbled upon this wikipedia and IMDB entry which stated that in 2004 Vendela Kirsebom (picture after the link) revealed her biological father is ethnically Turkish (Norwegian mother, born in Sweden). I had a hard time believing this, but I tracked down this article. It's in Swedish, but English speakers can make out the gist pretty easily, and if you put it through a translator it will become somewhat more intelligible. I note that Vendela is still listed as the "archetypical" "Hallstatt Nordic" on the Angelic Beauty: a tribute to Nordic women website, but I guess it's a geocities page, can't expect constant updates. Her looks haven't changed, and of course there are many fair Turks (when I was a kid there was a giant redhead troglodyte of a Turk that would terrorize the smaller kids at the mosque every Eid al Fitr). In any case, Vendela apparently believes in heritable differences in national characters, just like Greg and Henry, she says her personality is more Turkish than her appearance. No wonder no one guessed, who the hell cares about her personality???

Addendum: In response to the stupid reader who assumed that it is more likely that Vendela's father is part non-Turkish than that he was fair...I was going to go Bayesian, but I realized the illustration of my point didn't need the added precision and formality. My assertion, loosely stated, is that the majority of people with light brown hair in the nation of Ghana are either expats or the offspring of relationships between Europeans and native Ghanians. But, the majority of people in the nation of Turkey who are fair are Turkish...they are simply a rather small minority ("fair" normalized to Scandinavian fair). Nevertheless, they dwarf the number of fair halfbloods, it isn't that halfbloods aren't fair, there just aren't many non-Turks (or Kurds) in Turkey!

I have never been to Turkey, but my mosque as a child was heavy on Turks, and though most were swarthy, a large minority did not look very "Mediterranean," and this isn't because they looked Asiatic (I was frankly shocked to learn later that Turkish was a branch of an "eastern" language superfamily via a map in my dictionary). Additionally, most of the mixed offspring of Europeans and Middle-Easterners at my mosque didn't really look very "ethnic." Half-Turks1 like David Chokachi of Baywatch fame might be a bit extreme in favoring their Nordic side (his mother is Finnish, so it is language superfamily incest), but they aren't (in my experience) totally aberrant.2

In any case, I was a little surprised to see population-normal non-distributional thinking at work on the GNXP comment boards. Most of the traffic on this site is on search engines...but I assume that the core audience can accept that their prototypes are not Platonic ideals but central tendencies? Also, I have observed a cultural bias here in that southern Europeans and Middle-Easterners who are fair are characterized as almost certainly the product of exogenous genetic input, either recently or the past, but northern Europeans like Lena Olin are implicitly assumed to be part of the normal range of variation even if they lay at one tail (in her case, coloring, her jaw is all Swedish from what I can see).

Addendum II: Just to be explicit, this is what went through my head when I saw Reale's initial comment (see the follow up, he explains himself in detail, but I don't see that he explains the previous comment).

P(A|B) = [P(B|A)XP(A)]/P(B)

In the case of P(A|B) ~ Probability of blue eyes conditional on Turkish, we have empirical data, see this map. I was being conservative in my head and assuming that 1% of Turks have blue eyes.

Now, where

P(A) = foreign parentage and P(B) = blue eyes, you have probability of foreign parentage conditional on blue eyes being the product of the probability of blue eyes conditional on foreign parentage (assume European, so high) multiplied by the probability of foreign parentage (I assume very low) divided by the probability of blue eyes (evaluated within the Turkish national population). Even if you assume that only 1% of Turks have blue eyes, my understanding is that unlike Thailand or Philippines Turkey does not have a large expatriate community of whitesEuropeans who have settled and intermarried. Additionally, my understanding is that Turkey has a non-trivial indigenous blue-eyed population. In other words, I immediately inferred that the probability of Vendela's father being a blue-eyed Turk was far higher than the probability of her father being of partial foreign ancestry if he was truly a Turkish national. I could be uninformed, I guess it depends on the assumptions....

1 - In the mid-90s I watched David Chokachi tell Jay Leno that his father was from Iraq, so I assume that his father is from the Turkoman minority of northern Iraq.

2 - The variance has reasonable bounds, as I implied above. It is also a fair point that "Turks" are a genetically synthetic people who were generated via assimilation from the local substrate & enslaved peoples (Albanians & Slavs) to the Turkic elite. But that being said the ancestrally informative studies I've seen suggest only a minimal Asiatic tincture to the Turkish gene pool, and general continuity with the peoples to the north, west, east and south. Blonde Galatians and Russians are certainly part of the Turkish heritage, but the neutral marker data does, to my mind, support the contention that the fair end of the Turkish phenotype range is purely a function of admixture over the last 2,000 years with other peoples.