The initial spread of herding into eastern Africa (ca. 3000 BC) coincided with the emergence of a distinctive monumental tradition centred around “pillar sites” built near Lake Turkana, Kenya. These monumental sites more likely served commemorative purposes similar to many of the previous Saharan ceremonial sites before it (see §V.1. Africa and the Levant), although they exhibit architecturally distinct elements. These construction changes coincided with the end of the African Humid Period (ca. 3500–3000 BC), which brought about profound changes in environment, economy, and material cultural expression, coupled with a major population collapse due to the decline in favourable climatic conditions (Brierley, Manning, and Maslin 2018).
In the Lake Turkana area, retreating shorelines ca. 3300–2000 BC disrupted fishing practices and exposed new habitats for herbivores, while exchange and/or herder immigration brought cattle and caprines into northwest Kenya, transforming economic strategies to include mobile herding. While previous fishers used local lithic raw materials, the new herders preferred obsidian from varied local, distant, and island sources, which point to extended exchange networks, including boat travel (Hildebrand et al. 2018).
Monumental mortuary expression appearing in the Sahara, Sahel, Nile, and Turkana at the same time as the dramatic environmental shifts—and the accompanying change to herding economy—marks probably a social change rather than the emergence of hierarchical social forms. Distinct forms of commemoration include cattle burials in the central Sahara, megaliths in the eastern Sahara, aggregate cemeteries in the southern Sahara and along the Nile, built mortuary spaces in the Red Sea Hills and around Lake Turkana, and cairn and cremation treatments linked to early pastoralism in central Kenya.
In Mesopotamia, rulers of the Early Dynastic periods (ca. 2900–2350 BC) were leaders of city–states (the so-called ‘theocratic temple-states’), where estates of the gods were possibly the property of the king and his ruling family, and the ruler was thus a protector of a city in the name of the city’s tutelary deity. Political centralisation dominated in Mesopotamia under the two dynasties of Akkad and Ur, when East Semitic-speaking elites would eventually dominate over the whole society after the rise of Sargon of Akkad. The Mesopotamian king became both a divine figure and a warrior and conqueror (Kristiansen and Larsson 2005).
In the Fertile Crescent, a specialised economy of sheep–goats already evident in the Late Chalcolithic reaches a full development in the mid–3rd millennium BC, perhaps because of commodification of textiles and wool production in Near Eastern polities. The elaborate networks of roads during this period, with a pastoral economic activity growing beyond settled areas and population, are probably the result of interactions between settlements, where individual farmers, labour supply, and flocks of caprines moved across the landscape and were drawn to cities from the surrounding villages or exogenous sources. Urban centres and larger settlements were particularly affected by this demand, with textual evidence from Ebla and Tell Beydar of an institutionalised and centralised pastoral economy, where massive flocks of caprines were directly managed by the palace (Altaweel and Palmisano 2018).
East Semitic languages probably entered Mesopotamia from the desert to the west of its core area before 2900 BC, since the first attestations come from Akkadian personal names in Sumerian texts about the 29th century BC. Similarly, the Kish Civilisation—encompassing Semitic states like Ebla and Mari in the north, or Abu Salabikh and Kish in central Mesopotamia—shows probably the first historical record of the language, in the 30th century. Both East Semitic migration events can then be related to the collapse of the Uruk period ca. 3100 BC.
The Bronze Age Levantine population from the site of ʿAyn Ghazal, Jordan (ca. 2490–2300 BC), can be modelled as Levantine Pre-Pottery Neolithic agriculturalists from Motza, Israel, and ʿAyn Ghazal, dated ca. 8300–6700 BC (ca. 58%), with contributions of a population similar to Iran Chalcolithic (ca. 42%), from a more recent period (Lazaridis et al. 2016). It has been suggested that samples from Sidon, Lebanon (ca. 1700 BC) can be modelled as a mixture of Levant Chalcolithic (ca. 48%) and Iran Neolithic or Late Neolithic-related ancestry (Haber et al. 2017; Harney et al. 2018).
The difference between both populations lies then in the Anatolian-related ancestry found in Levant Chalcolithic (ca. 36%), also found in the northern population, which suggests the reintroduction in the south of a population not affected by this Anatolian migration, potentially then from farther south. The lack of relationship between Levant Chalcolithic ancestry and present-day East African Levantine-related ancestry (Harney et al. 2018) further supports a migration of Proto-Semitic from north-east Africa into the southern Levant, and then a back-migration to the south into Arabia and East Africa.
Among Levantine Bronze Age individuals, there is one sample (ca. 2400 BC) of J2b1-M205 lineage (formed ca. 13800 BC, TMRCA ca. 3300 BC), and another (ca. 2100 BC) of hg. J1a2b-Z1828, a haplogroup found previously in an Anatolian sample of the Bronze Age (see §v.1. Early Semites). Similarly, there is an individual from a Canaanite burial pit in Tel Shaddud (ca. 1250 BC) reported as of hg. J, which clusters—similar to another sample of hg. R1b1a1b-M269 (see below §viii.12. Greeks and Philistines)—among modern Levantine populations (van den Brink et al. 2017). The spread of early Semitic peoples was thus probably linked to communities of different local haplogroups, before the known Y-chromosome bottleneck of J1a2a1a2d2b2b-Z2331 lineages, particularly with the early expansion of Central Semitic dialects to the south.
The predominant East African genetic component of the Horn of Africa points to genetic homogeneity of all populations of the area, whether Afroasiatic (Omotic, Cushitic, Semitic) or Nilotic (van Dorp et al. 2015). They present a great degree of continuity with the ancient Mota individual (ca. 2520 BC), of hg. E1b1a2-M329 (Gallego Llorente et al. 2015). The Maasai from Kenia are also closely related, showing ca. 50% E1b1b-M215 (Pagani et al. 2012), and regional hunter-gatherers from the Horn of Africa—such as the Chabu, the Majang, and the Shekkacho—retain strong genetic affinities with the Mota individual (Gopalan et al. 2019), supporting the original association of this haplogroup and ancestry with indigenous Nilo-Saharan languages.
The Eurasian admixture events inferred from modern populations in Sudan and Ethiopia (Hollfelder et al. 2017) suggest the likely expansion of Egyptian speakers from the north through the Nile, and the likely expansion of Semitic in the Horn of Africa from Arabia is possibly associated with the “eastern”/Iranian farmer-related ancestry found among modern Afar and Somali (Skoglund et al. 2017).
On the basis of genetic data from Pastoralist Neolithic individuals, two phases of admixture can be detected, associated with the spread of pastoralism: the first one (likely ca. 4000–3000 BC) in north-eastern Africa, possibly (based on earlier herding dates from the region) through South Sudan and reaching the Turkana Basin first, although they may have moved via the Horn of Africa; the second one (ca. 2000 BC) between this admixed Early Neolithic Pastoralist group and eastern African foragers, both in the Turkana Basin and during the initial trickle of herding into the south-central Rift Valley (Prendergast et al. 2019).
Descendants of these Early Neolithic Pastoralist and local forager groups gave likely rise to the groups who developed the Pastoralist Neolithic culture traditions of southern Kenya and northern Tanzania. The earliest samples available to date are a man and a woman from Prettejohn’s Gully (ca. 2000 BC), whose ancestry profile suggest that they represent an initial (maybe one among many) limited dispersal of herder groups into the south-central Rift Valley that did not leave large numbers of descendants, while the group that gave rise to the Pastoral Neolithic cluster was much more demographically successful than the others. The lineage of this male is E2-M75 (xE2b-M98), while most Neolithic Pastoralists sampled belong to different E1b1b1-M35.1 subclades, apart from E2-M75, E1b1a-V38, and A-M13 lineages, further supporting the gradual and stepped admixture with local forager groups (Prendergast et al. 2019).
Whereas modern Omotic speakers—except those groups among Cushitic and Semitic groups—show strong affinities with the Mota individual, Cushites show continuity of an ancestry already found in a sample from Tanzania (ca. 1050 BC), with higher hunter-gatherer admixture than Bantus and mtDNA from the Rift Valley, suggesting that this Afro-Asiatic branch was widespread to the south before the more recent Bantu expansion, and thus likely to be associated with the Savannah Pastoralist Neolithic (Kießling, Mous, and Nurse 2008).
While ancient Afroasiatic languages were possibly introduced in the area by the expansion of R1b1b-V88 lineages and spread with pastoralism (see §iv.5. Late Afrasians), it remains unclear which lineages might have expanded Semitic languages to the region. Given the prevalence of haplogroup T1a1a1b2-CTS2214 near the Gulf of Aden, and the wide ancient distribution of this haplogroup through the Middle East (see §iv.4. Late Middle Easterners), the presence of some late subclades in the area probably reflects an acculturation of some East African or Middle Eastern group and subsequent Y-chromosome bottlenecks in the area. The various ‘western’ (sub-Saharan African admixture) and ‘eastern’ influences (south Asian and Levantine/European admixture) creating a west–east axis in the Arabian Peninsula support the nature of this region as a sink of different prehistoric migrations rather than a source (Cavadas et al. 2019).
Lexicological studies show that a Common Berber dialect continuum must have been still in close contact still around the foundation of Carthage (ca. 800 BC), based on early Phoenician/Punic loans found in most branches (Blažek 2014), which is compatible with the estimated dates for the disintegration of the language (Blažek 2010). This is supported in genetics by the incorporation and further expansion of this Afroasiatic branch under a bottleneck of E1b1b1b1a-M81 lineages, whose late successful expansion is coincident with that date (see above §iii.4. Northern Africans).