PDA

View Full Version : E-M35 and Afro-Asiatic proto-language



Farroukh
02-17-2019, 05:16 PM
Approximate divergence of subclades of the E-M35 and the main branches of the Afro-Asiatic macro family:
- Semitic E-PF1962
- Cushitic E-V32
- Egyptian E-V22
- Berbero-Libyan E-L19
- Omotic E-Y17750
- Chadic E-V5933

The separation is purely regional, some subclades are distributed outside one family. The ages of most subclades exceed 10 thousand years and it goes beyond the age of the Pre-Nostratic macro family.

All this leads to two different thoughts:
- Afro-Asiatic has an age of ~ 20 thousand years and its separation is associated with migrations oE-M35 subclades.
- the carriers of the E-M35 subclades settled over the Northern Africa in the Paleolithic times, but Afro-Asiatic languages ​​were adopted by them not earlier than 5 thousand years ago from other assimilator peoples.

Both theories have their own inconsistencies. Dispute is welcome.

Ruderico
02-17-2019, 08:54 PM
- Afro-Asiatic has an age of ~ 20 thousand years

This is highly unlikely, but I'd like to read what Aga has to say about it

Farroukh
02-19-2019, 01:24 PM
Proto-Afrasan dated as 8-10 Kya, too late for E-M35 subclades divergence. That is the question.

Shamayim
02-19-2019, 03:18 PM
You can't make sense of Afroasiatic without removing outliers i.e. Chadic speakers and most Semitic speakers. Once this is done, all Afroasiatic speakers belong to E-M35, formed circa 35k BC, and are roughly split into Libyco-Semitic speakers (E-Z827) and Cushito-Egypto-Omotic speakers (E-L539), both formed circa 24k BC. Only one culture matches this timeline: the Dabban culture of Libya.

The Dabban culture dates back to 38k BC and its industry also forms the "missing link" between the Iberomaurusian industry of Northwest Africa and the Halfan industry of Northeast Africa. It also includes Haua Fteah, an (the earliest) Iberomaurusian site dating back to 24k BC. PAA speakers may have originally lived in Libya and later split into Proto-Libyco-Semitic speakers who expanded westwards and Proto-Cushito-Egypto-Omotic speakers who expanded eastwards. Proto-Semitic speakers may have diverged from Proto-Libyc speakers and migrated eastwards into the Levant and created the Mushabian culture which shows affinities with North African backed bladelet industries.

Farroukh
02-19-2019, 03:56 PM
Dear Shamayim,
I can't be agree with you due to some linguistic reasons. PAA has some terms related to Neolithic reality, but we see no so old Neolithic cultures in our homeland Northern Africa. Also PAA has relations with Caucasian and other Eurasian languages, which is the evidence of it's early location in Eurasia.

Lank
02-19-2019, 05:07 PM
You can't make sense of Afroasiatic without removing outliers i.e. Chadic speakers and most Semitic speakers. Once this is done, all Afroasiatic speakers belong to E-M35, formed circa 35k BC, and are roughly split into Libyco-Semitic speakers (E-Z827) and Cushito-Egypto-Omotic speakers (E-L539), both formed circa 24k BC. Only one culture matches this timeline: the Dabban culture of Libya.

The Dabban culture dates back to 38k BC and its industry also forms the "missing link" between the Iberomaurusian industry of Northwest Africa and the Halfan industry of Northeast Africa. It also includes Haua Fteah, an (the earliest) Iberomaurusian site dating back to 24k BC. PAA speakers may have originally lived in Libya and later split into Proto-Libyco-Semitic speakers who expanded westwards and Proto-Cushito-Egypto-Omotic speakers who expanded eastwards. Proto-Semitic speakers may have diverged from Proto-Libyc speakers and migrated eastwards into the Levant and created the Mushabian culture which shows affinities with North African backed bladelet industries.
This scenario seems compelling for the formation and origin of M35 (or a spread from the Nile area). But I can't really find support for Haua Fteah having the earliest Iberomaursian site, these (known locally as Eastern Oranian) seem to be found later in Libya than in areas farther west. The first Iberomaurusian sites AFAIK are found in the western parts of the Maghreb.

Then again, it is plausible that the ultimate roots of the Iberomaurusian, along with E-M35, are to be found in early Upper Paleolithic industries of eastern North Africa, which include but are not limited to the Dabban, e.g. the Upper Paleolithic industries of the Nile Valley.

Megalophias
02-19-2019, 05:50 PM
I don't think there's much agreement on Afroasiatic's cultural vocabulary or external relationships - or time depth or internal branching or exact membership or much of anything else really. Still fun to speculate though.

When trying to link linguistic history with population history (and cultural history), what we're looking for is sociolinguistic conditions that line up (Malcolm Ross has written about this regarding Austronesian languages). When a population expands over a large area, that may correlate with the formation of a language family. When daughter populations remain next to each other and in contact, that may correlate with the development of a dialect continuum (linkage), with innovations spreading unevenly across the area. When daughter populations become geographically separated, that may correlate with a language breaking into distinct branches with unique innovations. When you have admixture, that may correlate with language convergence, or with language shift. Problem is there are a lot of relevant sociolinguistic factors beyond genetic relatedness. Maybe with a combination of detailed archaeological and genetic and linguistic information this is doable, but even Indo-European history is far from certain, and it is much more solid in all those aspects than Afroasiatic is.

Branching of a Y haplogroup need not, and usually does not, coincide with a population splitting up. Rapid growth, especially star-like, means something notable demographically.


Approximate divergence of subclades of the E-M35 and the main branches of the Afro-Asiatic macro family:
- Semitic E-PF1962
- Cushitic E-V32
- Egyptian E-V22
- Berbero-Libyan E-L19
- Omotic E-Y17750
- Chadic E-V5933
That would make Chadic, Omotic, and Cushitic all much more closely related to each other than the rest. Why not Berber with E-V65, Chadic with E-V259, Cushitic with E-V1515, etc? (BTW is there any Semitic-specific E-V22?)

Agamemnon
02-19-2019, 10:04 PM
Even though I agree that E-M35 and its branches are bound to be tied to Afroasiatic since the earliest stages, there are many issues with the scenarios mentioned here. I'll try to be as concise as possible:


Temporally, the estimates are simply too great. The caveats are numerous. To make things short, I'd say the most reliable estimates place Common Afroasiatic unity between 13,000 and 14,000 years BP.


It is generally assumed that the comparative method is valid only for time depths of up to ~10,000 years. I'd say Afroasiatic strengthens this assumption, indeed its validity is almost entirely based on morphological grounds, very few elements of Proto-Afroasiatic vocabulary can be reconstructed (apart from terms relating to body parts, bodily fluids, death, life, birth and other basic concepts) which is exactly what you'd expect from a language spoken more than 10,000 years ago. That being said, if PAA was spoken 20,000 years ago, it's doubtful AA morphology would be as uniform as it is (in other words, we wouldn't be able to establish the validity of an AA macrofamily of languages).


More importantly, we know that PAA was spoken by foragers and early grain collectors. The earliest AA dispersals were due to groups of hunter-gatherers, not farmers.


In terms of structure, your initial post infers a clean split between all branches at roughly the same stage. In reality, there is a good deal of evidence pointing to the existence of several nodes, with Omotic being the most basal split, Cushitic, Semitic and Libyco-Berber forming a node and Chadic and Egyptian being the last to branch off (I'd argue these two also form a node). These nodes share common innovations.


The homeland remains an open question. That being said, and you're free to disregard my opinion, I see no evidence of a high-level relationship (let alone contact) between PAA and Caucasian/Sumerian/Eurasian languages. At least not beyond some typological elements, which are also found in Niger-Congo (more on that below). There is however some evidence of contact between AA and Niger-Congo, and this is mainly morphological in nature (so it could even be used to argue in favour of a genetic relationship). So I favour the African shores of the Red Sea as the most likely contender for the PAA homeland, and view the Nile valley as a corridor through which the earliest AA dispersals took place, an area largely congruent with the green area on this map:

https://i.imgur.com/JGkfRcI.jpg

In my view, the potential archeological correlates would be the Qadan and Sebilian and probably even the Halfan cultures. The homeland debate is one of the areas where ancient genomics can provide very helpful hints. I could of course be completely wrong, Central Arabia might be the homeland for all I know, or even NW Africa. Personally, I think it makes little difference as Arabia and much of North Africa were essentially one ecozone 14,000 years ago, this should not be understood as a "Eurasia vs Africa" contest. Without Epipaleolithic data from NE Africa, I'll always get the impression we're missing something big.


I would argue that PAA had a complex noun class system that has left traces in several of the AA branches. This is similar both to Niger-Congo (Bantu in particular) and North Caucasian languages. The fossilised remnants of this system can probably prove useful in reconstructing PAA vocabulary, it would be in my view a good alternative to the comparative method. One of the markers seems to be the suffix -b which seems to denote dangerous animals (there are of course exceptions), compare for instance Proto-Semitic *ḏiʾb- "wolf" and Egyptian zꜣb https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_O34.png?63e15https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_G1.png?4d556https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_D58.png?12f84https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_E17.png?15ded "jackal". If we take sound changes into account, words such as PS *kalb- "dog" can be compared to Egyptian ṯzm https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_V13.png?2da65https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_O34.png?63e15https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_G17.png?3741ehttps://en.wiktionary.org/w/extensions/wikihiero/img/hiero_E14.png?2f8f7 "dog". There are other suffixes which fit the bill here (such as liquid consonants for non-threatening animals). This could explain the derivation of several triliteral roots from biliteral ones in several branches of PAA. There are, however, problems with this approach. The main problem as I see it is that the strongest evidence for this comes from Semitic, this is no mistake because this branch is both extremely well studied and happens to be phonologically-archaic in an AA context, another problem is that there is evidence of contact with a language that had noun classes in the early stages of Semitic (Pre-Proto-Semitic) and so this too is bound to have had an impact on the formation of triliteral roots.


So where does this leave us? Well, first off I'd say the Proto-Afroasiatic speech community already carried several distinct branches of E-M35, and not just one. There's also the distinct possibility that early para-AA dialects spread out and reached Northwest Africa and SW Asia, I'm thinking about the Iberomaurusians and the Natufians here in particular. We could be looking at a very ancient occurrence of dialect levelling too. So while an association between, say, E-M81 + E-V65 and Libyco-Berber, or between E-V22 and Egyptian is fairly sound, some of the lineages might have had nothing to do with the initial spread of AA and merely adopted the language (to be sure, the language they would have spoken before would've been fairly similar). What is also clear is that this isn't going to be as simple as "E-M35 = AA". Some branches of Y-DNA haplogroup T (and possibly E1a) are likely to have been present during the early stages of AA. And for at least two of the branches, Semitic and Chadic, lineages other than E-M35 were present at proto-stage and had an important role in the spread of these branches.

NetNomad
02-19-2019, 10:14 PM
@Farroukh

E-V1515 and some subtypes of E-V22 and E-V12 are also definitely Cushitic-associated.

Just like Indo-Europeans had both R1b and R1a, different E-M35s types can exist in the same language family.

Farroukh
02-20-2019, 01:51 PM
Sure, there is the mixture of subclades. I just noted basic prevailed directions.

Tz85
02-20-2019, 05:44 PM
There is also a distinction between E-V12 and E-V12*

NetNomad
02-21-2019, 02:40 PM
Sure, there is the mixture of subclades. I just noted basic prevailed directions.

I think a yet-unknown subtype of E-M34 and E-M329 (overall) is more strongly Omotic-associated. E-Y17750 is likely Ethiopian Cushitic instead.

See this thread:
https://anthrogenica.com/showthread.php?10734-mtDNA-amp-Y-DNA-E-predictions-for-Plaster-et-al-Horn-Africans-by-Passa

However, E-M329 is pre-Afroasiatic by age (formed 39 KYA) and is likely from Ethiopian foragers absorbed by incoming Afroasiatics (E-M34?) from North Sudan/Red Sea region.

Also, J1 is common in some Northern Omotics (Shekkacho). However, I am not sure if it is pre-Ethiosemitic or not.

Agamemnon
02-21-2019, 08:07 PM
I think a yet-unknown subtype of E-M34 and E-M329 (overall) is more strongly Omotic-associated. E-Y17750 is likely Ethiopian Cushitic instead.

See this thread:
https://anthrogenica.com/showthread.php?10734-mtDNA-amp-Y-DNA-E-predictions-for-Plaster-et-al-Horn-Africans-by-Passa

However, E-M329 is pre-Afroasiatic by age (formed 39 KYA) and is likely from Ethiopian foragers absorbed by incoming Afroasiatics (E-M34?) from North Sudan/Red Sea region.

Also, J1 is common in some Northern Omotics (Shekkacho). However, I am not sure if it is pre-Ethiosemitic or not.

I very much agree regarding E-M34, reducing it to a simple association with Semitic speech does nothing to explain its presence among non-Semitic AA speaking-groups in the Horn.

As far as J1 is of concern, I find any association with AA at its earliest stages doubtful and, barring some kind of groundbreaking shift in Paleolithic results from the Near East, simply unconvincing. To give you some perspective, remember that Satsurblia (who carried a very rare and basal branch of J1) was alive when Proto-Afroasiatic still was a spoken language. It's hard to picture this lineage being simultaneously anywhere near East Africa at the same time. While Omotic is the most basal branch of AA, the presence of J1 in the Shekecho, Keffa and Yem does not even remotely suffice to assert that the lineage was present in the earliest AA-speaking groups. Most of it is probably P56, while an early foray during the Neolithic is possible, odds are this branch arrived with the first Ethiosemitic speakers (the TMRCA estimates fit perfectly with this scenario), J1 is going to be strictly associated with Semitic-speaking groups (and more generally with nomadic pastoralists who took part in the peopling of deep desert environments since the Chalcolithic) IMO and this counts for the more obscure lineages as well.

Lank
02-21-2019, 10:45 PM
I very much agree regarding E-M34, reducing it to a simple association with Semitic speech does nothing to explain its presence among non-Semitic AA speaking-groups in the Horn.

As far as J1 is of concern, I find any association with AA at its earliest stages doubtful and, barring some kind of groundbreaking shift in Paleolithic results from the Near East, simply unconvincing. To give you some perspective, remember that Satsurblia (who carried a very rare and basal branch of J1) was alive when Proto-Afroasiatic still was a spoken language. It's hard to picture this lineage being simultaneously anywhere near East Africa at the same time. While Omotic is the most basal branch of AA, the presence of J1 in the Shekecho, Keffa and Yem does not even remotely suffice to assert that the lineage was present in the earliest AA-speaking groups. Most of it is probably P56, while an early foray during the Neolithic is possible, odds are this branch arrived with the first Ethiosemitic speakers (the TMRCA estimates fit perfectly with this scenario), J1 is going to be strictly associated with Semitic-speaking groups (and more generally with nomadic pastoralists who took part in the peopling of deep desert environments since the Chalcolithic). IMO and this counts for the more obscure lineages as well.
J1-P56 is fascinating because as you say, a Neolithic arrival is not entirely ruled out. Despite the ancient DNA record from the Levant showing so clearly that J1 along with CHG/Iran ancestry represents post-Neolithic migration, African P56 sticks out as a lineage with an 11 kya divergence date that is also rare outside East Africa. If we didn't have ancient DNA from the Levant demonstrating the arrival of J1 and Iran/CHG since the Chalcolithic, I would argue more confidently that J1-P56 is a Neolithic lineage in East Africa. But it is interesting that the current TMRCA for P56 is such a neat fit with the arrival of D'mt Ethiosemites. If that holds up with future P56 samples, then a recent spread with Ethiosemites is pretty much confirmed. But if P56 did spread with Ethiosemites, I guess it may still trace back to pre-Semitic Arabians, considering its paucity in other Semites?

Also would be interesting to find out where the minor J(xJ1,J2) in Maale (probably present in other Omotics too) is from, even if it's rare. Is it somehow related to J* lineages in Soqotra? Migration reaching Arabia from Iran, followed by low-level gene flow across the Red Sea?

RCO
02-22-2019, 12:35 AM
Pontus Skoglund

That is exactly right. Present-day Cushitic speakers in the Horn of Africa have the "eastern"/Iranian farmer ancestry though very clearly.
https://twitter.com/pontus_skoglund/status/1098503263886946304

I presume J1 brought that component to the Horn of Africa

Agamemnon
02-22-2019, 12:40 AM
J1-P56 is fascinating because as you say, a Neolithic arrival is not entirely ruled out. Despite the ancient DNA record from the Levant showing so clearly that J1 along with CHG/Iran ancestry represents post-Neolithic migration, African P56 sticks out as a lineage with an 11 kya divergence date that is also rare outside East Africa. If we didn't have ancient DNA from the Levant demonstrating the arrival of J1 and Iran/CHG since the Chalcolithic, I would argue more confidently that J1-P56 is a Neolithic lineage in East Africa. But it is interesting that the current TMRCA for P56 is such a neat fit with the arrival of D'mt Ethiosemites. If that holds up with future P56 samples, then a recent spread with Ethiosemites is pretty much confirmed. But if P56 did spread with Ethiosemites, I guess it may still trace back to pre-Semitic Arabians, considering its paucity in other Semites?

Also would be interesting to find out where the minor J(xJ1,J2) in Maale (probably present in other Omotics too) is from, even if it's rare. Is it somehow related to J* lineages in Soqotra? Migration reaching Arabia from Iran, followed by low-level gene flow across the Red Sea?

Indeed, the presence of relatively basal and isolated clades of J1 not only in the Horn but in South Arabia is a pretty fascinating subject that deserves a thread of its own quite frankly (I sincerely do not wish to highjack this thread). As you said, a Neolithic arrival cannot be ruled out, even if the odds are not exactly in favour of such an arrival.

I think you're asking the important questions here: Is an early spread of J lineages in SW Asia prior to the Pottery Neolithic-Chalcolithic plausible? While it is IMO clear that the most important branches of J1 (under P58) are closely tied to Iran_ChL-type ancestry, I'd say this largely depends on what was in the region prior to the arrival of ANA and E-M35 lineages at the onset of the Kebaran which would've arrived from North Africa (a Eurasian origin for E-M35 isn't convincing). While an origin of J1 in the Iranian plateau is definitely a possibility, I strongly suspect Mesopotamia might still prove to be a better contender, still there isn't much of a difference between this or the Caspian/Iran/Caucasus. The only real constraint in time and space as I see it is J-M304's relationship to I-M170 via IJ-M429.

There's a fair chance (though I personally wouldn't bet on it) that Dzudzuana-like populations were common throughout the region prior to the Kebaran period, Bar Yosef basically infers that human occupation at the site was relatively recent and attributed it to foreign foragers, it's also likely that Dzudzuana has something to do with the "UHG" signal that has often been discussed here. A possible (albeit extremely speculative) scenario for the J* lineages in Soqotra and in the Maale would see the early spread of this type of ancestry from a more southern culture, for example Baradostian, with a direct migration from the Iranian plateau towards East Africa through the Gulf (which was a large oasis during the late Pleistocene). There are of course many possible caveats, a more in-depth analysis of the J* lineages is desirable.

There's a pretty strange situation going on with MSA-speaking groups in general, these different populations tend to carry very rare and peculiar branches of J1, such as L93 (common among the Mahra) or L860. In this context, the J* in Soqotra just follows the pattern. Here too though, the TMRCA estimates suggest an arrival with Semitic speakers.

IronHorse
02-22-2019, 09:58 AM
Indeed, the presence of relatively basal and isolated clades of J1 not only in the Horn but in South Arabia is a pretty fascinating subject that deserves a thread of its own quite frankly (I sincerely do not wish to highjack this thread). As you said, a Neolithic arrival cannot be ruled out, even if the odds are not exactly in favour of such an arrival.

I think you're asking the important questions here: Is an early spread of J lineages in SW Asia prior to the Pottery Neolithic-Chalcolithic plausible? While it is IMO clear that the most important branches of J1 (under P58) are closely tied to Iran_ChL-type ancestry, I'd say this largely depends on what was in the region prior to the arrival of ANA and E-M35 lineages at the onset of the Kebaran which would've arrived from North Africa (a Eurasian origin for E-M35 isn't convincing). While an origin of J1 in the Iranian plateau is definitely a possibility, I strongly suspect Mesopotamia might still prove to be a better contender, still there isn't much of a difference between this or the Caspian/Iran/Caucasus. The only real constraint in time and space as I see it is J-M304's relationship to I-M170 via IJ-M429.

There's a fair chance (though I personally wouldn't bet on it) that Dzudzuana-like populations were common throughout the region prior to the Kebaran period, Bar Yosef basically infers that human occupation at the site was relatively recent and attributed it to foreign foragers, it's also likely that Dzudzuana has something to do with the "UHG" signal that has often been discussed here. A possible (albeit extremely speculative) scenario for the J* lineages in Soqotra and in the Maale would see the early spread of this type of ancestry from a more southern culture, for example Baradostian, with a direct migration from the Iranian plateau towards East Africa through the Gulf (which was a large oasis during the late Pleistocene). There are of course many possible caveats, a more in-depth analysis of the J* lineages is desirable.

There's a pretty strange situation going on with MSA-speaking groups in general, these different populations tend to carry very rare and peculiar branches of J1, such as L93 (common among the Mahra) or L860. In this context, the J* in Soqotra just follows the pattern. Here too though, the TMRCA estimates suggest an arrival with Semitic speakers.

If haplogroup I is associated autosomally with Villabruna/WHG ancestry, then it's reasonable to suggest that the non-Basal ancestry of Dzudzuana is the autosomal counterpart of haplogroup J, assuming this:

I started a thread once about a strange result in a PCA of Europeans/West Asians/North Africans, but it seems I was "uncharismatic" in my persuasion. https://anthrogenica.com/showthread.php?16359-Saami-and-Finns-Southwest-Asian-and-the-fourth-dimension

the fourth component reveals an Arabian like component that North Africans and Caucasians seem to lack, but Northeast Europeans like Finns are shifted slightly to the right, autosomal results of Saamis also confirm that they have Southwest Asian/Middle East ancestry higher than their neighboring populations, I connected this with Dzudzuana ancestry in Baltic/Scandinavian foragers, this Dzudzuana seems to be non-Basal Eurasian, as suggested by Anatolian hunter-gatherer paper.

so I assumed that there is a ghost population that drives this fourth component, I tried to look for mt-DNA connection between modern Arabian and mesolithic Baltic samples, but all is U2/U5 which is lacking in Arabian peninsula .. but Y-DNA.

some J (J1 ?) was found in mesolithic North East Europe, right ? assuming that this non-basal Dzudzuana is the origin of the fourth component in PCA, then Arabians should have extra Basal-less Dzudzuana .. this is consistent with deep subclades of J1 in South Arabia, that means they were there all along.

the only problem with this hypothesis is the association of Iran_N/CHG ancestry with J1 in BA Levant.

Agamemnon
02-22-2019, 10:36 PM
Yes, J1 was found among the Mesolithic Karelians from Popovo and Yuzhnyy Oleni Ostrov. These samples probably carried the same branch as Satsurblia (which is found in Finland nowadays).

Govan
02-25-2019, 02:47 PM
You can't make sense of Afroasiatic without removing outliers i.e. Chadic speakers and most Semitic speakers. Once this is done, all Afroasiatic speakers belong to E-M35, formed circa 35k BC, and are roughly split into Libyco-Semitic speakers (E-Z827) and Cushito-Egypto-Omotic speakers (E-L539), both formed circa 24k BC. Only one culture matches this timeline: the Dabban culture of Libya.

The Dabban culture dates back to 38k BC and its industry also forms the "missing link" between the Iberomaurusian industry of Northwest Africa and the Halfan industry of Northeast Africa. It also includes Haua Fteah, an (the earliest) Iberomaurusian site dating back to 24k BC. PAA speakers may have originally lived in Libya and later split into Proto-Libyco-Semitic speakers who expanded westwards and Proto-Cushito-Egypto-Omotic speakers who expanded eastwards. Proto-Semitic speakers may have diverged from Proto-Libyc speakers and migrated eastwards into the Levant and created the Mushabian culture which shows affinities with North African backed bladelet industries.

Not necessarily..
A proto Berber at Iberomaurusian stage would be quite challenging. Not impossible but very challenging.

On the other hand, the hypothesis of Roger Blench could be closest to the truth : a small but powerful group of tribes from Neolithic Egypt expanded west and Berberized a much larger Neolithic Maghreb ( KEB.. ) population, itself the results of Iberian-related Neolithics absorbing local Iberomaurusian foragers. In this case Proto-Berbers would have contributed a fraction (15%) of extra-Natufian-like admixture in most Maghreb population (probably a double more in Eastern Berbers towards Libya).

It would be a scenario very different from East Africa where proto - Cushitics also from Upper Nile Valley largely remplaced and dimished the local East African foragers (Omotics and others). Luxamanda pastoralist from Tanzania looks very much Somali-like in terms of DNA profile, also its West Eurasian profile fit very much Neolithic Egypt as how we would imagine it going with archeology alone : a mixture of Neolithic Levantines, Mesolithic Egyptians and East African Dinka-like populations.

Govan
02-25-2019, 02:58 PM
Even though I agree that E-M35 and its branches are bound to be tied to Afroasiatic since the earliest stages, there are many issues with the scenarios mentioned here. I'll try to be as concise as possible:


Temporally, the estimates are simply too great. The caveats are numerous. To make things short, I'd say the most reliable estimates place Common Afroasiatic unity between 13,000 and 14,000 years BP.


It is generally assumed that the comparative method is valid only for time depths of up to ~10,000 years. I'd say Afroasiatic strengthens this assumption, indeed its validity is almost entirely based on morphological grounds, very few elements of Proto-Afroasiatic vocabulary can be reconstructed (apart from terms relating to body parts, bodily fluids, death, life, birth and other basic concepts) which is exactly what you'd expect from a language spoken more than 10,000 years ago. That being said, if PAA was spoken 20,000 years ago, it's doubtful AA morphology would be as uniform as it is (in other words, we wouldn't be able to establish the validity of an AA macrofamily of languages).


More importantly, we know that PAA was spoken by foragers and early grain collectors. The earliest AA dispersals were due to groups of hunter-gatherers, not farmers.


In terms of structure, your initial post infers a clean split between all branches at roughly the same stage. In reality, there is a good deal of evidence pointing to the existence of several nodes, with Omotic being the most basal split, Cushitic, Semitic and Libyco-Berber forming a node and Chadic and Egyptian being the last to branch off (I'd argue these two also form a node). These nodes share common innovations.


The homeland remains an open question. That being said, and you're free to disregard my opinion, I see no evidence of a high-level relationship (let alone contact) between PAA and Caucasian/Sumerian/Eurasian languages. At least not beyond some typological elements, which are also found in Niger-Congo (more on that below). There is however some evidence of contact between AA and Niger-Congo, and this is mainly morphological in nature (so it could even be used to argue in favour of a genetic relationship). So I favour the African shores of the Red Sea as the most likely contender for the PAA homeland, and view the Nile valley as a corridor through which the earliest AA dispersals took place, an area largely congruent with the green area on this map:

https://i.imgur.com/JGkfRcI.jpg

In my view, the potential archeological correlates would be the Qadan and Sebilian and probably even the Halfan cultures. The homeland debate is one of the areas where ancient genomics can provide very helpful hints. I could of course be completely wrong, Central Arabia might be the homeland for all I know, or even NW Africa. Personally, I think it makes little difference as Arabia and much of North Africa were essentially one ecozone 14,000 years ago, this should not be understood as a "Eurasia vs Africa" contest. Without Epipaleolithic data from NE Africa, I'll always get the impression we're missing something big.


I would argue that PAA had a complex noun class system that has left traces in several of the AA branches. This is similar both to Niger-Congo (Bantu in particular) and North Caucasian languages. The fossilised remnants of this system can probably prove useful in reconstructing PAA vocabulary, it would be in my view a good alternative to the comparative method. One of the markers seems to be the suffix -b which seems to denote dangerous animals (there are of course exceptions), compare for instance Proto-Semitic *ḏiʾb- "wolf" and Egyptian zꜣb https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_O34.png?63e15https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_G1.png?4d556https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_D58.png?12f84https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_E17.png?15ded "jackal". If we take sound changes into account, words such as PS *kalb- "dog" can be compared to Egyptian ṯzm https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_V13.png?2da65https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_O34.png?63e15https://en.wiktionary.org/w/extensions/wikihiero/img/hiero_G17.png?3741ehttps://en.wiktionary.org/w/extensions/wikihiero/img/hiero_E14.png?2f8f7 "dog". There are other suffixes which fit the bill here (such as liquid consonants for non-threatening animals). This could explain the derivation of several triliteral roots from biliteral ones in several branches of PAA. There are, however, problems with this approach. The main problem as I see it is that the strongest evidence for this comes from Semitic, this is no mistake because this branch is both extremely well studied and happens to be phonologically-archaic in an AA context, another problem is that there is evidence of contact with a language that had noun classes in the early stages of Semitic (Pre-Proto-Semitic) and so this too is bound to have had an impact on the formation of triliteral roots.


So where does this leave us? Well, first off I'd say the Proto-Afroasiatic speech community already carried several distinct branches of E-M35, and not just one. There's also the distinct possibility that early para-AA dialects spread out and reached Northwest Africa and SW Asia, I'm thinking about the Iberomaurusians and the Natufians here in particular. We could be looking at a very ancient occurrence of dialect levelling too. So while an association between, say, E-M81 + E-V65 and Libyco-Berber, or between E-V22 and Egyptian is fairly sound, some of the lineages might have had nothing to do with the initial spread of AA and merely adopted the language (to be sure, the language they would have spoken before would've been fairly similar). What is also clear is that this isn't going to be as simple as "E-M35 = AA". Some branches of Y-DNA haplogroup T (and possibly E1a) are likely to have been present during the early stages of AA. And for at least two of the branches, Semitic and Chadic, lineages other than E-M35 were present at proto-stage and had an important role in the spread of these branches.

I think we should give to Caesar what belongs to Caesar.

Sure many Middle Easterners and Arabians would like the Proto Afro Asiatic urheimat to be in the Middle East or Arabia.

Question of pride.

Now I don't care much about their pride. The most important is the truth and indeed everything suggests a closest PAA in the Nile Valley by the late Mesolithic.

Now, sure Neolithic North Africans of the Nile Valley also got Levantine Neolithic admixture going on. But it doesn't look like there was a population remplacement as archeology shows the population density of Mesolithic Egyptians was strong, and a large cultural continuity is observable in the Nile Valley from Mesolithic to Neolithic time.

So my 'guess' would be that very North Egyptians could have adopted the language of Neolithic Levantines but not the rest of the Nile Valley which kept of their languages (Proto-Cushitic, Chadic, Semitic). Northern Egyptians would have been Egyptian-ized during the late predynastic era with the expansion of the Naqada culture.

NetNomad
02-28-2019, 11:53 AM
It would be a scenario very different from East Africa where proto - Cushitics also from Upper Nile Valley largely remplaced and dimished the local East African foragers (Omotics and others). Luxamanda pastoralist from Tanzania looks very much Somali-like in terms of DNA profile, also its West Eurasian profile fit very much Neolithic Egypt as how we would imagine it going with archeology alone : a mixture of Neolithic Levantines, Mesolithic Egyptians and East African Dinka-like populations.

Dinka-like autosomal admixture historically mostly started in North Sudan rather than Egypt proper. Especially given the autosomal affinities of Upper Egyptian Copts (AEs). The Nubians and Ababda in Upper Egypt likely originated from more Southern regions.

Govan
03-02-2019, 05:01 PM
Another modeling of an Algerian with G25.

Anatolia_BA,35.2
Iberomaurusian,25.2
Beaker_Iberia,15.6
Natufian,11.4
Yoruba,7.6
Iberia_MN,5

[1] "distance%=3.2509"


Tunisian,51.4
Algerian,48.6


[1] "distance%=1.7791"


Mozabite,77.4
Moroccan_Jew,22.6

[1] "distance%=2.1539"


Algerian43A34,65
Algerian,19
Libyan,16


Iberomaurusian looks about right. IberianBellBeaker is needed to fit the bill for the additional steppe layer, a clear presence of an extra-Natufian layer as expected . Anatolian BA largely represent Barcin N/Anatolian N ancestry, but the pick up of Anatolian BA suggest the presence of an additional layer of CHG, Levant N.


Here a modeling of a Berber Kabyle woman. If you we had to focus on an actual Berber - speaker.

Barcin_N,45.4
Iberomaurusian,31.2
Armenia_EBA,11.6
Yoruba,4
Yamnaya_Samara,3.2
WHG,2.2
Ganj_Dareh_N,1.4
Ethiopia_4500BP,1


Again very same pattern, instead she needs Armenia BA to fit the bill with that seems to be extra-layer of steppe, CHG and Levant N in North Africa.

Iron Age Phoenician are probably needed and could explain the layer of CHG here.

I would say by this stage, a very logic attempt would be to believe that Proto-Berbers expanded by the Late Neolithic from western Egypt alongside Proto-Chadic to fit the given date stage of initial Proto-Afro-Asiatic. Proto-Chadic moved south. It could be evasive to assume how much, but it looks like Proto-Chadic are the Afro-Asiatics that left the least ancestry out of all the 5 main Afro-Asiatic tribes (ProtoBerber, ProtoSemitic, ProtoEgyptian, Proto Cushitic) with the core of current Chadic speakers owning very little to no autosomal admixture from the Proto-Chadic. However some of their uniparental libeages could be potentially linked with.

IronHorse
03-22-2019, 09:22 AM
If haplogroup I is associated autosomally with Villabruna/WHG ancestry, then it's reasonable to suggest that the non-Basal ancestry of Dzudzuana is the autosomal counterpart of haplogroup J, assuming this:<br>
<br>I started a thread once about a strange result in a PCA of Europeans/West Asians/North Africans, but it seems I was "uncharismatic" in my persuasion. <a href="https://anthrogenica.com/showthread.php?16359-Saami-and-Finns-Southwest-Asian-and-the-fourth-dimension" target="_blank">https://anthrogenica.com/showthread.php?16359-Saami-and-Finns-Southwest-Asian-and-the-fourth-dimension</a>

the fourth component reveals an Arabian like component that North Africans and Caucasians seem to lack, but Northeast Europeans like Finns are shifted slightly to the right, autosomal results of Saamis also confirm that they have Southwest Asian/Middle East ancestry higher than their neighboring populations, I connected this with Dzudzuana ancestry in Baltic/Scandinavian foragers, this Dzudzuana seems to be non-Basal Eurasian, as suggested by Anatolian hunter-gatherer paper.<br>

so I assumed that there is a ghost population that drives this fourth component, I tried to look for mt-DNA connection between modern Arabian and mesolithic Baltic samples, but all is U2/U5 which is lacking in Arabian peninsula .. but Y-DNA.

some J (J1 ?) was found in mesolithic North East Europe, right ? assuming that this non-basal Dzudzuana is the origin of the fourth component in PCA, then Arabians should have extra Basal-less Dzudzuana .. this is consistent with deep subclades of J1 in South Arabia, that means they were there all along.

the only problem with this hypothesis is the association of Iran_N/CHG ancestry with J1 in BA Levant.

Maybe I'm reading too much into this, but ADMIXTURE analysis from the recent Sardinian paper (https://www.biorxiv.org/content/10.1101/583104v1.supplementary-material) may have hinted at my conclusion, which is Southwest Asian component harbours ancestry related to non-basal Dzudzuana present in Arabians and NE-foragers and their Saami and Finnish descendents.

at K=4 &amp; K=5; Bedouins score more of the WHG component than Druze and Turkish (almost null), with Palestinians following in second place.

at K=7; EHG gets ots own component (seperate from Steppe_EBA), again the previous WHG like in Bedouins and Palestinians is replaced with this component, consistent with Dzudzuana ancestry in NE-Euro foragers.

at K=8-11: a component highest in Bedouins (a.k.a Southwest Asian) eats up the previous ancestry and is proportional to it.


https://anthrogenica.com/attachment.php?attachmentid=29485&d=1553246229
https://anthrogenica.com/attachment.php?attachmentid=29486&d=1553246257

Farroukh
02-05-2021, 07:03 PM
Militarev is the leading linguistic proponent of the Levantine origin for Afrasian (“Levant theory”, opposed to Afroasiatic Urheimat in Northeast Africa theory proposed by Christopher Ehret, Roger Blench and others) linking the proto-Afrasian speakers to the Levantine Natufian culture.

https://en.m.wikipedia.org/wiki/Alexander_Militarev