So, Afroasiatic... is it really legit?

[Pabappa]

just a passing comment ..... Wikipedia suggests that proto-Nilo-Saharan might be even older than proto-Afro-Asiatic, in that it was spoken in the Upper Paleolithic, which began around 50000 BC and ended around 20000 BC.* that's a REALLY old proto-language. However I wish they had just given a year instead of saying that it was the Upper Paleolithic.

*Im doing my best here, because Wikipedia gives the date of the end of the UP as 20000 BC in one place and 10000 BC in another. I think what they mean by this is that some places made the transition earlier than others, and I think it would be fair to assume that the Sahara region was near the locus of human civilization at the time, so it would be closer to the earlier date than the later date.

[WeepingElf]

"Proto-Nilo-Saharan" is not a thing. It hasn't been reconstructed yet, and many linguists doubt that "Nilo-Saharan" is a family at all. It's just a cover term for all those African languages which show no evidence of relationship to either Semitic or Bantu, and do not have clicks It is a matter of sociology of science, it seems, that Africanists are readier to accept long-range language relationship proposals than Eurasianists.

And what regards the time of the Upper Paleolithic, this depends on the region. In Europe, the Upper Paleolithic begins with the arrival of Homo sapiens about 45,000 years ago, and ends with the climate change about 9600 BC; I don't know about when it is considered to have begun and ended in Africa.

[Richard W]

"Proto-Nilo-Saharan" is not a thing. It hasn't been reconstructed yet, and many linguists doubt that "Nilo-Saharan" is a family at all. It's just a cover term for all those African languages which show no evidence of relationship to either Semitic or Bantu, and do not have clicks

Wouldn't that definition exclude Nilo-Saharan? Roger Blench has been pushing the idea that Niger-Congo is Nilo-Saharan.
Still, interesting to learn that Austro-Asiatic does not exit. (It hasn't been reconstructed yet, either.)

[zompist]

It is a matter of sociology of science, it seems, that Africanists are readier to accept long-range language relationship proposals than Eurasianists.

Basically, I think, because Greenberg's rework, though entirely based on mass comparison rather than the comparative method, was way better than what came before.

R.M.W. Dixon was so appalled by Greenberg's work on the Americas that he went back to look at his African work, and thought it was just as bad.

[bradrn]

My understanding of the current state of African ‘families’ is as follows:

• Niger-Congo sensu lato is somewhat doubtful. No-one argues that Atlantic-Congo at least is valid, but there’s questions over whether to include some of the peripheral families, especially Mande, Ijọ and Dogon.
• Afroasiatic is mostly held together by consistent morphological patterns, as well as some cognates in basic vocabulary, but it’s too old to easily reconstruct.
• Nilo-Saharan is extremely doubtful; there’s only about five or six morphological correspondences.
• Khoisan is simply incorrect, though some lower-level connections might be valid.

[WeepingElf]

My understanding of the current state of African ‘families’ is as follows:

• Niger-Congo sensu lato is somewhat doubtful. No-one argues that Atlantic-Congo at least is valid, but there’s questions over whether to include some of the peripheral families, especially Mande, Ijọ and Dogon.
• Afroasiatic is mostly held together by consistent morphological patterns, as well as some cognates in basic vocabulary, but it’s too old to easily reconstruct.
• Nilo-Saharan is extremely doubtful; there’s only about five or six morphological correspondences.
• Khoisan is simply incorrect, though some lower-level connections might be valid.

Yep - that's precisely what I think about these matters.

At this point it must be pointed out that a successful reconstruction of the latest common ancestor is not necessary for the acceptance of a language family. There are many well-accepted families without reconstructed ancestors. It is often unclear which languages are conservative and which have innovated, especially if there are only two units to be compared. For instance, with Early PIE (the common ancestor of Late PIE - the common ancestor of the non-Anatolian IE languages - and Anatolian), it is unclear which of the two daughters has innovated where they differ. Did Early PIE have three genders, or two? Did it have the tripartite verb aspect system, or not? It seems as if Anatolian was the more conservative daughter, since there is not much in Anatolian that's not in Late PIE, while there is quite much in Late PIE that's not in Anatolian, but we can't be sure about that. So we cannot reconstruct the Early PIE state of affairs. Yet, there is so much Late PIE and Anatolian have in common - the ablaut system and other phonological traits, much of the nominal inflection, the verbal personal endings, and of course hundreds of words - that there is no doubt that these two are quite closely related to each other.