It has recently been found that some parts of the Caucasus have the highest percentage of J1 in the world (nearly 100% in some ethnic groups) and that the greatest genetic diversity of J1 was around Kurdistan and the Caucasus, not Arabia. This changes completely our understanding of J1's origins.
The Caucasus is actually a region of low YSTR diversity as far J1 is concerned and we can see quite clearly a bottleneck effect linked to endogamy which led to J1*'s high frequencies amongst North-east caucasian speakers.
The two places where diversity is found are the Zagros and Ethiopia [100% of Oromo J1 is J1* and 29% of Amharic J1 were J1*] (as pointed out by Chiaroni and Tofanelli's papers).
And there was no such thing as "Kurdistan" at the time...
I should also remind you that Oman has relative diversity too...
I totally agree with that. What I meant was that J1c3(d) was especially common among speakers of Semitic languages, but I do not doubt that the origin of Semitic languages lies within haplogroup E1b1b. Actually both Hebrew and Arabic originated in the southern Levant, near the Sinai, where the percentage of E1b1b is the highest in the Middle East.
There is only one E1b1b1 subclade which seems involved with the Semitic language family and that is E1b1b1c1 (M34+), a marker whose frequency follows neatly that of J1c3 (though being much less common, it peaks amongst Dead Sea Jordanians [bottleneck effect is likely to have played its part] and Ethiopian Semitic speakers while being found homogeneously amongst other Semitic populations)... Yet you forget one crucial element: Frequency doesn't provide a clue towards a marker's initial source.
As for Semitic being some sort of E1b1b1 product, you must also consider that languages are products not only of isolation but most certainly of interaction... And when addressing a language spoken by nomads, pastoralists and herder-hunters, you must take into consideration the fact that interaction is a fundamental part of its survival (isn't most of humanity speaking a nomad-derived language? Indo-european? Turkic? Semitic? Austroasiatic? With the notable exception of Chinese, most of the languages spoken nowadays were first spoken by nomads!), which is why Akkadian (first semitic language attested) also shares features with North-east Caucasian languages such a Chechen.
In proto-semitic vocabulary, we find many words for hills, mountains, bitumen and naphta (which are only found in the northern Levant)... A word for ice too, which suggests that PS probably has something to do with mountaineous areas... Like the Zagros where J-P58 and E-M34 have their greatest diversity.
Chiaroni cited Kitchen's Bayesian analysis of Semitic language and also mentionned E-M123 (E-M34's parent clade) in these terms: "Although J1e (J1c3) is one of the most frequent haplogroups in the region, haplogroup E-M123 also shows its highest frequency and haplotype diversity in regions of the Fertile Crescent, decreasing towards the Arabian Peninsula. This co-distribution pattern of Y-chromosome haplogroups J1e and E-M123 resembles mtDNA haplogroups J1b and (PreHV)1 distributions that also display low levels of diversity despite their high frequency in Saudi Arabia."
So why the need to speak only of E1b1b?
Why is everyone so unrealistic?
Wasn't J1 found amongst pre-hispanic Guanches (Fregel et al)?
Haven't you wondered why it has high YSTR diversity in Ethiopia?
The concept that E1b1b1 (M35) is the only marker linked to Afroasiatic is kind of shallow as you have to explain R1b1c (V88) and J1's frequencies amongst non-semitic Afroasiatic speakers too.
Not even that, other haplogroups such as T1a* or J2a4h could be linked with the spread of Semitic languages.