I didn't mean any disregard for Fretless' account, which is focussed on the movements of groups of people. And Malcolm is right to deride notions of "race," which are mostly bankrupt from a scientific point of view, and have had no positive effects from a social or philosophical standpoint. But there's nonetheless the matter of cultural affinity, and specific usages do migrate along with individuals and the social units to which they belong. The linguistic evidence is pretty clear that a variety of groups denominated themselves, as distinct from others, by a term which consisted of an initial consonant, a vowel ("accidents happen to vowels" is a historical linguist's proverb)and "L." The discussion was about the "correct" pronunciation of the initial consonant, and the evidence is, the pattern of development was complex but regular. The earlier question was a good one -- about whether the pronunciation was derived from the people's self-designation or Graeco-Roman ethnography. If I were asked to assert my own opinion, I'd say that in the Anglophone world, the "S" version of the initial consonant is a spelling pronunciation which has gained authority from being taught (along with spelling) in school, like putting a "t" in "often" or "L" in "palm," "calm" and "psalm." Now it's for others to decide which authority they recognize: schoolbooks, linguistic history, the prestige of those who have or claim the right to speak for the people in question, newspaper and TV editors (increasingly decisive), the social status of those favoring a given usage, or sheer weight of numbers. All these things contribute to the not-at-all-obvious issue of "usage."

My recommendation is to do what feels right and let others do the same. Language is a chameleon behavior, and we all eventually adopt usages we regard as wrong if enough people around us do them with sufficient frequency.

Best,

Adam