[EForEndeavour]

> You say that it's "easy" for you to determine whether somebody is "East Asian", just by looking at a photo. What are you looking for, though? Probably some sort of appearance affected by genetics. However, no East Asian can carry the complete gene pool from their region, so no East Asian can carry a set of genetic markers which is guaranteed to produce an appearance which is reliably recognizable.

It feels like this discussion is disproportionately weighted toward exceptions to demonstrably real correlations between a person's ancestral origin and their appearance and genes. Why would every member of a given cluster of related people be required to carry every one of the criteria that define the cluster? In my mind, that defeats the whole purpose of clustering in the first place, and seems especially out of place given that normal people in normal situations make fuzzy, heuristic classifications of people based on incomplete information.

Also, doesn't this "nobody carries the complete pool from their region" argument apply equally to any conceivable definition of culture? Not every rural Texan espouses exactly the same cultural values as each other, or cultural memes, or what have you. Despite the inherent fuzziness in defining what it means to be rural Texan, you can still make the definition meaningful. The ability to define a Sherpa culture seems equally plausible as the ability to characterize Sherpa genotypes and phenotypes. The fuzziness of a classification does not negate the fact that fuzzy classification is still possible.

> So, whatever you're seeing in that photo, it's not just their genes, but also at least some of your biases, which lead you to think that your classification rate is better than it actually is.

That makes sense to me.