Most languages of North Africa, East Africa and the Near East are descended from a common ancestor known as proto-Afro-Asiatic. According to linguists this proto-language was spoken no earlier than 9500 years before present, which makes it one of the oldest identified proto-languages. The original homeland of proto-Afro-Asiatic has not yet been identified with certainty. In order to do this we need to know how the various branches of Afro-Asiatic languages spread, which is difficult because these developments mostly took place in prehistoric times. Moreover, because the proto-language is so old the extant branches, like Cushitic, Chadic, Semitic, Berber and Egyptian, are highly divergent. This makes it difficult to group these branches into subgroups. To illustrate the problems I will give an overview of the scholarly concensus on this issue.
The subgroups
Afro-Asiatic is a macrofamily that includes several branches, including Cushitic, Chadic, Semitic, Berber, Egyptian, and possibly Omotic and Ongota. These branches are distinct from each other, but internally they have many common features. Of the surviving branches proto-Cushitic was probably spoken in eastern Ethiopia and Eritrea, proto-Chadic around lake Chad, proto-Semitic in the Levant, proto-Berber in northern Libya and proto-Egyptian in the Nile Valley. Note that ancient Egyptian was just one language that likely belonged to a larger branch that has left us no other recorded languages. Omotic and Ongota are both spoken in western Ethiopia, but their inclusion within the Afro-Asiatic macrofamily is controversial. If they are indeed part of this macrofamily, they must have split off early on. The position of these languages is comparable to that of Hittite in Indo-European studies. If we are to reconstruct the diffusion of Afro-Asiatic, we need to determine the relation between these subgroups.
Subdivision according to Ehret
According to Christopher Ehret the first branch to split off was Omotic. The remaining Afro-Asiatic languages then split up into a northern group, named Boreafrasian, and a southern group, named Cushitic. Later the Chadic branch split of from the Boreafrasian group and the remaining Boreafrasian languages spread north along the Nile Valley. Somewhere in present day Egypt these languages split into a western (Berber), a central (Egyptian) and an eastern (Semitic) branch. In conclusion, Ehret considers Omotic and Cushitic to be the oldest branches, both of which are found in Ethiopia. The Berber, Egyptian and Semitic languages are descended from one northern branch that split off relatively late. It is therefore no suprise that Ehret considers the proto-Afro-Asiatic homeland to be in Ethiopia and sees the northern distribution of Afro-Asiatic languages as a relatively recent event.
Subdivision according to Fleming
Harold Fleming agrees with Christopher Ehret on many issues. He agrees that the Omotic branch split off first, but he also includes Ongota, a now endangered language spoken in western Ethiopia, among the early split-offs. This strengthens the case for an Ethiopian homeland for Afro-Asiatic. Fleming follows Ehret in identifying a major split between the Cushitic languages and the northern Afro-Asiatic languages, but he classifies Beja, a major language spoken along the coast of Sudan, as a northern Afro-Asiatic rather than a Cushitic language. As for the remaining branches (Chadic, Berber, Egyptian and Semitic) the subdivision is the same. The minor disagreements between Fleming and Ehret demonstrate some of the problems that come with grouping languages. For example, Fleming recognizes Ongota as an Afro-Asiatic language, but Ehret does not. Also, Fleming draws attention to the fact that Beja is apparently highly divergent from other Cushitic languages. Fleming draws the conclusion that Beja is therefore is not a Cushitic language, but the divergence could also be illustrative of the variation within the Cushitic subgroup, which again hints at its antiquity.
Subdivision according to Bender
So far, all the evidence seems to point towards an Ethiopian homeland for Afro-Asiatic. However, different classifications will lead to different conclusions. Whereas Ehret and Fleming agree that the Berber and Semitic languages belong to a relatively new northern branch, Lionel Bender, in his article ‘Upside-down Afrasian’ (1997), changes the whole structure of the Afro-Asiatic family tree and groups Semitic, Berber and Cushitic together. According to him, Semitic and Berber split off much earlier then is commonly assumed, with Berber moving into the ‘green Sahara’ and Semitic moving into ‘green Arabia’, crossing from Ethiopia to Yemen. Chadic and possibly also Egyptian split off even earlier, not long after Omotic. Bender places the proto-Afro-Asiatic in northern Sudan. Since I am not a linguist and because I do not know that many Afro-Asiatic languages, I cannot judge the arguments used to justify this classification. However, the fact that such divergent conclusions can be reached tells us a lot about the difficulty that comes with classifying these Afro-Asiatic languages.
Subdivision according to Militarev
Although most scholars still place the proto-Afro-Asiatic homeland in Africa, the idea of a homeland in the Levant is gaining support. One of the leading proponents, Alexander Militarev, claims that proto-Afro-Asiatic civilization is represented by the Natufian culture. The Natufians flourished in the Levant between 12.500 and 9500 BC and may have been the ones who first made the transition to agriculture. Much of the Levantine hypothesis is based on the assumption that agricultural terms in various branches in Afro-Asiatic can be traced back to a proto-Afro-Asiatic root. This is controversial, because many linguists date the proto-Afro-Asiatic language before the invention of agriculture. Moreover, not all branches of Afro-Asiatic share the same words for these agricultural terms. In fact, there are very few words that are shared by all branches of Afro-Asiatic. As for Militarev’s classification, he sees the first split as one between a northern group and a southern group. The northern group stayed put in the Levant, while the southern group, which includes both Omotic and Cushitic (note that Militarev does not consider Omotic to be the earliest branch to split off) moved towards the Horn of Africa; either by moving through the Arabian peninsula or by moving south up the Nile Valley. An argument in support of the Arabian route is that there appears to be a Cushitic substratum in the Modern South Arabian languages. The northern group eventually split up in proto-Semitic, which again stayed put in the Levant, and an African North-Afrasian group. The African North-Afrasian group then split up into Egyptian and Chado-Berber. Egyptian stayed put in the Nile Valley, while Chadic and Berber spread across the green Sahara.
Conclusions
As I said before, I am not an expert on this topic. Still, I will try to draw some preliminary conclusions. I tend to agree with the traditional view that regards Omotic as the earliest split-off and Cushitic as one of the most ancient groups. According to the vast majority of linguists Omotic is the furthest removed from all other branches and the Cushitic branch is highly diverse. Some even go as far as to regard Cushitic as consisting of several smaller branches that influenced each other through a Sprachbund. This all speaks for the antiquity of Cushitic. I also tend to agree with the idea of a relatively new northern group, although I do not know the subdivision within this northern group. Chadic, Berber, Egyptian and Semitic are, despite their similarities, still highly divergent. Considering the Levantine hypothesis, I do not want to rule out the possibility, but the case for a common agricultural vocabulary for the Afro-Asiatic languages should be made stronger. Moreover, the idea that Semitic is one of the oldest branches does not have much support among other linguists, so that case needs to be strengthened as well. Finally, the idea that Omotic and Cushitic, two very archaic branches, apparently ‘decided’ to move all the way to Ethiopia and started to diversify only after arriving there seems a bit far-fetched to me. Even if Militarev’s classification turns out to be true, an Ethiopian homeland still works better than a homeland in the Levant.