The Beta Israel (Falasha): Ethiopian Jews, conversion or migration? The genetic verdict on Global25

The Beta Israel, the Jews of Ethiopia, present the mirror image of the Hittite paradox. The Hittites spoke the oldest written Indo-European language yet carried none of the steppe ancestry that defines every other Indo-European population. The Beta Israel have practised Judaism for many centuries, claim descent from the tribe of Dan or from the union of Solomon and the Queen of Sheba, yet carry essentially none of the Levantine ancestry that defines every other Jewish population on Earth. Where the Samaritans match Bronze Age Canaanite Ashkelon at a distance of 0.0201, and the Druze match the Phoenician coast, the Beta Israel match the Amhara, their Christian Ethiopian neighbours, at a distance of just 0.0116. The single closest Jewish population to them anywhere in the world, the Yemenite Jews, sits at 0.3096 on Global25, twenty seven times farther away than the Amhara next door. This article documents that result with direct figures and then runs the decisive test: when a Jewish or Levantine source is explicitly offered to the mixture algorithm alongside an Amhara source, the Jewish source receives a weight of zero, and the Amhara source absorbs the entire genome. The Beta Israel are, autosomally, an Ethiopian highland population that adopted Judaism, not a Levantine population that migrated to Ethiopia. The roughly 55 percent of Near Eastern ancestry they do carry is the ancient back-to-Africa flow that every northern highland Ethiopian shares in equal measure, not a Jewish inheritance. The verdict is as clean as the Hittite one, and points the same way: language and religion can travel without the genes that we expect to accompany them.

1. Who are the Beta Israel?

The Beta Israel, whose name means "House of Israel", are the Jewish community native to Ethiopia. For most of their documented history they lived in the highlands north of Lake Tana, in the regions of Gondar, Semien, Dembiya, and Qwara, as a distinct religious caste among a Christian and, in places, Muslim majority. They were farmers, weavers, potters, and smiths, often barred from owning land and regarded with suspicion by their neighbours, who called them Falasha, a word usually translated as "exiles" or "landless ones" and now avoided as a slur. They called themselves Beta Israel.

Their religion, known as Haymanot, is built on the Orit, the Ethiopic translation of the Torah, together with a body of para-biblical and monastic literature in Ge'ez, the liturgical language of Ethiopian Christianity. Like the Samaritans, the Beta Israel preserve a form of Israelite religion that predates, or developed in isolation from, the rabbinic Judaism of the Talmud. They observe the Sabbath strictly, practise male circumcision on the eighth day, keep dietary laws, and historically maintained ritual purity rules and a monastic priesthood. They had no knowledge of Hanukkah (a post-biblical festival) until contact with other Jews, a detail that tells us their tradition branched off, or was constructed, before the rabbinic period crystallised.

In the late twentieth century the existence of the community became widely known in the wider Jewish world, and after the Israeli rabbinate recognised them as Jews, the Israeli government organised two dramatic rescue operations during famine and civil war: Operation Moses in 1984, which brought thousands across the Sudanese border, and Operation Solomon in 1991, which airlifted more than fourteen thousand people from Addis Ababa in thirty six hours. Today the overwhelming majority of Beta Israel, perhaps 150,000 people, live in Israel, with a continuing community of Falash Mura (descendants who had converted to Christianity under pressure and have returned to Judaism) still arriving.

2. Two hypotheses, one question

The origin of the Beta Israel has been debated for more than a century, and the debate has always reduced to a single binary question: did Judaism arrive in Ethiopia with people, or with an idea?

3. The Beta Israel on Global25: an Ethiopian highland population

When every modern population in the Global25 panel is ranked by Euclidean distance from the average Beta Israel coordinate, the result is decisive and geographically tight. The closest populations are, in order, the northern Ethiopian and Eritrean highlanders: Amhara, Agaw, Afar, Tigray, Saho. Every Jewish population on Earth is more than 0.30 away, an order of magnitude farther. The contrast between the nearest neighbour (Amhara at 0.0116) and the nearest Jewish group (Yemenite at 0.3096) is the single cleanest statement of the result.

The same ranking against ancient populations tells the same story. The closest ancient samples to the Beta Israel are East African pastoralists carrying the back-to-Africa component (Kenya_EarlyPastoralN at 0.0317, then Sudanese Christian-period and Kenyan and Tanzanian pastoral Neolithic samples). The Levantine ancients that anchor the Samaritans and the Druze are nowhere near: Israel_Natufian sits at 0.2897, Israel_MLBA at 0.3493, Israel_IA at 0.3494, and Late Bronze Age Canaanite Ashkelon at 0.3553. The Beta Israel draw their nearest ancient affinities from the Horn of Africa, not the Levant.

4. The decisive test: offer a Jewish source, and watch it get rejected

A short distance to the Amhara is suggestive, but the rigorous test, exactly as in the Hittite case with the steppe source, is to offer the suspected source explicitly to the mixture algorithm and observe the weight it assigns. Here we model the Beta Israel as a two-way mixture of an Amhara source and a Jewish or Levantine source, and let NNLS decide how much of the latter is needed. If any genuine Jewish or Levantine ancestry distinguished the Beta Israel from the Amhara, the algorithm would seize on it. It does not.

5. The Near Eastern ancestry is shared with all highland Ethiopians

The Beta Israel are not purely African. Like all northern highland Ethiopians, they carry a large fraction of Near Eastern ancestry, the result of a well documented back-to-Africa migration that brought Levantine and Arabian-like ancestry into the Horn over the Neolithic and Bronze Age, long before any Jewish or Israelite history. The natural question is whether the Beta Israel carry more of this Near Eastern ancestry than their neighbours, which a Jewish admixture event would produce. They do not. Modelled as ancient East African (the 4500 year old Mota individual) plus a Levantine source, the Near Eastern fraction is essentially identical across the whole northern highland.

6. The inflow is real, but it is religion-blind and it is not Bronze Age Canaanite

Granting that uncertainty, two facts nonetheless constrain the question tightly, and both turn on dating rather than ideology. The first is that whatever the Iron Age Levantine source captures, it captures the same amount in everyone, regardless of faith. The second is that the Bronze Age Canaanite populations that define actual Jewish ancestry are rejected outright.

7. The Beta Israel and the Amhara are not even distinguishable

The strongest single statement of the result is a comparison of distances. The gap between the Beta Israel average and the Amhara average is 0.0116. But the internal scatter, the average distance of individuals from their own group centroid, is 0.0171 within the Beta Israel and 0.0203 within the Amhara. In other words, two random Beta Israel individuals are, on average, farther apart from each other than the Beta Israel average is from the Amhara average. The two populations do not form separable clusters on Global25; they form a single overlapping cloud. By the standard of every other Jewish community, which separates clearly from its host population, the Beta Israel simply do not separate from the Amhara.

8. The deep ancestry: a northern highland Ethiopian profile

Decomposed into deep sources (the 4500 year old Ethiopian Mota individual as the autochthonous East African base, plus the four canonical southwest Asian sources), the Beta Israel return 43.6 percent Mota, 48.9 percent Natufian, and small fractions of Anatolian Neolithic, Iranian Neolithic, and Caucasian hunter-gatherer. The Amhara model is the same to within noise (42.5 percent Mota, 47.4 percent Natufian). The dominant Near Eastern component is Natufian, the autochthonous Levantine hunter-gatherer ancestry, which is exactly what the back-to-Africa migration carried into the Horn. This is a northern highland Ethiopian profile in every detail.

9. The uniparental evidence: African fathers, African mothers

The autosomal verdict is confirmed on both uniparental lines. On the paternal (Y-chromosome) side, the Beta Israel are dominated by African lineages: haplogroup E1b1 and its sub-branches, which are the characteristic paternal lineages of the Horn and of much of Africa, together with haplogroup A, one of the deepest African lineages of all. The single lineage that has been raised as a possible Levantine connection is E-M34, present at about 13.6 percent (Cruciani et al. 2004). But E-M34 is distributed across the entire Horn of Africa and reaches its highest frequencies among Omotic speakers, so it is a regional Horn lineage, not a Jewish or Israelite marker.

The contrast with the Samaritans is the decisive point. The Samaritans carry a clear cluster of the Cohen modal haplotype, the priestly Israelite Y-chromosome signature, which ties them directly to the ancient Israelite paternal line (Shen et al. 2004). The Beta Israel carry no such cluster. If a founding group of Israelite priests or nobles had migrated to Ethiopia, we would expect at least a trace of this signature in the priestly lineages; there is none. On the maternal (mitochondrial) side the picture is the same: Beta Israel lineages are East African (haplogroups L0, L2, L3), with none of the Near Eastern maternal lineages that a migrant Israelite population would have brought. The mothers of the Beta Israel were Ethiopian women.

10. An honest nuance: the diluted founder

Rigour requires the same caveat we applied to the Hittites. The genetic data firmly excludes a substantial Israelite or Jewish migration: there is no population-scale Levantine signal, no Cohen cluster, no Near Eastern uniparental ancestry beyond the shared back-to-Africa layer. What the data cannot exclude is a very small founding group, perhaps a handful of Jewish or Judaised traders, missionaries, or refugees from South Arabia or the Levant, who carried the religion to Ethiopia and whose own distinctive ancestry was then diluted to genetic invisibility as the faith spread by conversion among the local Agaw and proto-Amhara population. This is the standard elite-dominance or founder-dilution pattern, and it is fully compatible with everything we see.

It is worth being clear about what this means. A minority view in the literature (associated with researchers such as Omer) argues for genuine ancient Jewish roots, emphasising that the loss of distinctive religious texts and the centuries of isolation could mask an original migrant core. The genetics neither proves nor disproves such a small founding seed. But it does settle the demographic question that the migration hypothesis was really making: the body of the Beta Israel people, the ancestors of the overwhelming majority of individuals alive today, were Ethiopian highlanders, not Levantine immigrants. The idea travelled; the people, in the main, were local.

11. What the Beta Israel teach us

The Beta Israel complete a triptych of Levantine and Jewish genetic studies. The Samaritans are the southern Levant frozen at Ashkelon; the Druze are the northern Levant frozen on the Phoenician coast; and the Beta Israel are the demonstration that a Jewish community can exist with no Levantine ancestry at all. Set beside the Yemenite Jews, who genuinely descend in large part from Arabian converts but still cluster in the Near East, the Beta Israel show the most complete decoupling of religion from ancestry in the entire Jewish world. They are Jews by faith, practice, and self-understanding, and Ethiopians by descent.

This is the same lesson the Hittites taught about language. We are tempted to assume that a cultural inheritance, a language or a religion, must travel inside a population, leaving a genetic trail we can follow. Sometimes it does, as with the steppe Indo-Europeans or the Yemenite Jews. But sometimes the inheritance moves faster and lighter than the genes, carried by a few people or by ideas alone, and adopted by populations who share none of the ancestry of its originators. The Hittites are Indo-European without the steppe. The Beta Israel are Jewish without the Levant. In both cases the absence of the expected genetic signal is not a problem to be explained away; it is the most interesting fact about them, and a reminder that the history of peoples and the history of their cultures are two different histories that only sometimes run together.

12. The Beta Israel Global25 coordinate

For readers who wish to reproduce the analysis in Vahaduo or any Global25 tool, the scaled average Beta Israel coordinate used throughout this article is given below.

13. Myth versus reality

14. Conclusion

The Beta Israel were long the subject of a romantic question: were they a lost fragment of ancient Israel, carried to the Ethiopian highlands in the mists of the Solomonic age? Palaeogenetics and modern population genetics answer plainly. They are an Ethiopian people, indistinguishable on Global25 from their Christian Amhara neighbours, twenty seven times closer to those neighbours than to any Jewish population on Earth, with African paternal and maternal lineages and no priestly Israelite signature. The Near Eastern ancestry they carry is the same ancient back-to-Africa inheritance that runs through the whole northern highland, Christian and Muslim and Jewish alike. When a Jewish source is offered to the mixture algorithm, it is refused outright. The Beta Israel did not migrate from the Levant; Judaism migrated to them, and they made it their own, much as the Hittites made an Indo-European language their own without ever setting foot on the steppe. Their genome is a monument not to an ancient exodus but to the power of an idea to cross a sea and a desert on its own, and to be embraced by a people who owed it nothing by blood and everything by faith.

References

1. Behar, D. M., Yunusbayev, B., Metspalu, M., et al. (2010). The genome-wide structure of the Jewish people. Nature, 466(7303), 238 to 242. DOI: 10.1038/nature09103 Genome-wide Jews
2. Tishkoff, S. A., Reed, F. A., Friedlaender, F. R., et al. (2009). The genetic structure and history of Africans and African Americans. Science, 324(5930), 1035 to 1044. DOI: 10.1126/science.1172257 African structure
3. Cruciani, F., La Fratta, R., Santolamazza, P., et al. (2004). Phylogeographic analysis of haplogroup E3b (E-M215) Y chromosomes reveals multiple migratory events within and out of Africa. American Journal of Human Genetics, 74(5), 1014 to 1022. DOI: 10.1086/386294 Y-DNA E-M215
4. Shen, P., Lavi, T., Kivisild, T., et al. (2004). Reconstruction of patrilineages and matrilineages of Samaritans and other Israeli populations from Y-chromosome and mitochondrial DNA sequence variation. Human Mutation, 24(3), 248 to 260. DOI: 10.1002/humu.20077 Y-DNA, mtDNA
5. Non, A. L., Al-Meeri, A., Raaum, R. L., et al. (2011). Mitochondrial DNA reveals distinct evolutionary histories for Jewish populations in Yemen and Ethiopia. American Journal of Physical Anthropology, 144(1), 1 to 10. DOI: 10.1002/ajpa.21360 mtDNA Yemen Ethiopia
6. Kaplan, S. (1992). The Beta Israel (Falasha) in Ethiopia: From Earliest Times to the Twentieth Century. New York University Press. History
7. Quirin, J. (1992). The Evolution of the Ethiopian Jews: A History of the Beta Israel (Falasha) to 1920. University of Pennsylvania Press. History
8. Pagani, L., Kivisild, T., Tarekegn, A., et al. (2012). Ethiopian genetic diversity reveals linguistic stratification and complex influences on the Ethiopian gene pool. American Journal of Human Genetics, 91(1), 83 to 96. DOI: 10.1016/j.ajhg.2012.05.015 Ethiopian gene pool
9. Davidski (Eurogenes). Global25 scaled coordinates and the Vahaduo Global25 tools. NNLS modelling performed for this article in Python (scipy) on the scaled coordinates. G25 method