Unless there's very strong assortative mating...

Wouldn't assortative mating that is sufficiently strong to overcome the 50% dilution per generation amount to consanguinous mating?

▲

dash2 7 days ago | parent [-]

Not necessarily. You only have to overcome the dilution on specific dimensions. Suppose everyone mates only on intelligence, and marries a person of exactly equal genetic intelligence to themselves. Then there's no dilution per generation on intelligence, though there is on e.g. height.

	▲	pcrh 5 days ago \| parent [-]
		The question here is why names are so highly predictive. It seems highly unlikely that each generation would be selecting on precisely the same genetic features of a particularly prominent family name. This is because that which provides social prominence is not consistent over time.