I don’t think it’s possible to do it with 100% certainty, but it’s certainly possible to do it with odds that are better than what we’d expect from chance. There definitely is a difference between a “French” face and an “English” face, for example, although I don’t believe I could personally describe it.
I don’t see why people are so resistant to the idea that there might be a distinctive French or German or Japanese or Korean “look”. Most people in a country are probably going to marry and have children with somebody else in that same country - it seems reasonable that there is going to be a bunch of “national” genes floating around.
This isn’t saying that every Lithuanian or Thai is going to look the same. But there probably are appearance traits that are more common in Italians or Filipinos than in non-Italians or non-Filipinos.
This is the site where the original composites were made: https://pmsol3.wordpress.com/
They were taken without his permission. Whoever borrowed them labeled the Argentine composite as South African. (Mr. Dibble, you’re correct, not South African at all.)
About the maker: About Me | The Postnational Monitor
I’ve interacted with him online for several years. He’s a thirty something black American with a background in international development, who currently works in IT in Switzerland. He has lived and worked in China, speaks Chinese, and has an ongoing interest in East Asian affairs. He also has a serious interest in population genetics, especially those of Africa and black America.
So, even sven, he knows perfectly well that all of Africa isn’t one country. He went through thousands of photos to put together the composites, and he used what he could get.
He doesn’t make any claims that the composites are rigorously scientific. They’re just the best that he could do with the time and resources available.
The best part of this is that Roissy picked up on the composites without knowing that the maker was a black American. (Roissy wishes fervently for an all white USA). When he found out, he was not pleased.
What loveliness. :rolleyes:
Also, he’s the kind of high functioning sociopath that lots of good looking young white women find very attractive.
Specific nationality in Europeans is tougher than Chinese/Korean/Japanese, I think.
Still, I bet you and I both could still score well making binary distinctions, at least. Imagine getting a series of paired photos and being asked which one is nationality X, or a series of single photos and the choice of nationalities Y and Z. If you’ve seen a large enough sample of people from most populations, I think you’ve picked up a lot more clues about the ranges and patterns of appearance than you probably think. The key is having the sample to learn from in the first place (in RL or photos)–being able to look at fifty people that you know are all French, another fifty you know are all German, or whatever; after that, I think most of us could do pretty well assigning fifty more one way or the other. You wouldn’t expect to get them all, but you’d get a lot, I think.
It would be fun to try, anyway.
Here’s a quiz to try for the three Asian sets.
I just got 12/18, which is a little worse than I’d hoped, but the average score is only 7. (Random chance would be 6.)
Thanks for the link.
This page shows a number of averaged men as well as averaged women. I’m curious if the same effect exists. Do women (or gay men) find the averaged men generally more attractive?
And from those links, here’s the average white American male and white American female.
I don’t know, they all look the same to me.
You mean white South Africans, or women?
Yeah, blah blah blah, it just seems needlessly dismissive of a continent. You’d be kind of 
if a Chinese artist put together a similar project and carefully labeled every country and even made fine definitions within countries, and then lumped the US, Canada, Mexico and the Caribbean as “North America.” It’s easy to not really care about Africa, but the way this project presents Africa is lazy and content in it’s ignorance.
He’ll find your notion that he’s dismissive of Africa highly entertaining.
What’s interesting is that lots of people from different parts of the world have similar complaints: “You didn’t get right the part of the world that I think is most important!”
The West African composite was motivated by the fact that this is where the African ancestors of most black Americans came from. He was primarily interested in a comparison between West Africans and black Americans.
He made regional melanges because the African component of our ancestry is a regional melange. He’s not particularly interested in differences between modern African political states, because those are relatively recent developments, and don’t necessarily correspond to ethnicity anyway. Many of the major ethnic groups are spread across several countries, as you know.
Finding enough Akan or Mende face front portraits to do composites may not be impossible, but it’s definitely more time consuming and difficult than doing a West African one.
And the software is available. Everyone can do their own composites on the peoples of their choosing.
Really? I wouldn’t care.
I mean, this guy didn’t do America at all.
He did.
It’s on his website.
You’re just reading way too much into it. It wouldn’t make sense to do Nigeria, or Cameroon, or Gabon either, considering the diversity of ethnic groups within each of these countries.
I just wanted to address the confusion about the word average. There’s more than one type of average. When we use average in casual conversation about non-numerical quantities, we usually mean that mode; however, this system is using the mean. When we discuss the average person from such-and-such country, we are thinking of the traits we see in the majority of the people we have seen from the country. This software isn’t doing that. It’s doing hundreds of numerical calculations, averaging the measurements of the picture given.
Responses:
Well, he did warn you by saying he was a white supremacist. Racism and sexism tend to be correlated.
I signed up and was going to use the software to do this, but
You don’t hear people claim they can do it, but people do try to do this, all the time. I’ve had people guess at what my ethnic background is on many occasions - often in relation to them trying to puzzle out my last name, which is hyphenated. People tend not to be shy about it, and I’ve been quizzed and told I do/don’t look like my ethnic background on numerous occasions.
ETA: btw, I think the reason all of these women look so attractive is that the software has made their faces symmetrical, which, as I understand it, we register as “attractive”.
Weird. I like the girls where I live now, but none of the East-Asian women did anything special for me. Maybe Aomori really does have awesomely beautiful girls.
She reminded me of the time in the American Southwest that she was mistaken for a Navajo, too.
I MUST go to Peru. Yowzah!