Out of Africa, by Assumption

TL;DR

• Human ancestry supplies no automatic birthplace or date for a story.
• Witzel’s 2001 paper provides an early explicit program; the wider literature grows through distinct research projects, citations, shared data, and collaboration.
• Berezkin supplies migration-based direction; Le Quellec and d’Huy explicitly use genetic history to choose myth-tree roots.
• Newer network and genetic studies discover real patterns, but their oldest chronologies still require additional assumptions.
• The citation trail documents how the inference spread. Repeating the inference does not independently verify it.

A myth turns up in Africa, Australia, and the Americas. It concerns a serpent, the origin of death, or people emerging from underground. The distribution is remarkable. Then, somewhere between the map and the conclusion, the story acquires an African birthplace and a departure date tens of thousands of years before anyone wrote it down.

The passengers have been identified. Their luggage has been assumed.

In my critique of Julien d’Huy’s African roots, I examined a particularly clear instance: an African origin was supplied to a statistical tree, then the reconstruction was presented as a window into African prehistory. But d’Huy did not invent the larger move. A connected literature has spent decades making humanity’s migration history do double duty as the history of its stories.

Michael Witzel compares whole mythological systems. Yuri Berezkin maps thousands of recurring motifs. Jean-Loïc Le Quellec and d’Huy construct trees. More recent researchers bring information theory and ancient DNA. Their methods differ, sometimes substantially. The recurring temptation is the same: give a cultural resemblance the direction and age of a human migration before establishing that the resemblance records that migration.

This tradition has a history we can actually document. Its citations lead to printed pages, conference papers, databases, and explicit statements of method. The origin of the African-origin theory is considerably easier to establish than the origin of the myths.

A distribution is not a departure date #

Three questions keep getting compressed into one. Are two stories historically related? If so, where did their common material arise? When did it spread?

A peculiar bundle of shared details might make historical transmission persuasive. It still does not identify the inventor’s continent. A geographical pattern might favor one route over another. It still needs a reason to belong to the first settlement of those places rather than a later exchange.

The critical missing step is easy to state: why must the story have traveled with those particular ancestors?

A population’s descendants can adopt stories after their ancestors separate. People can transmit rituals across boundaries that impede marriage. New stories can displace old ones. A broad motif can also arise repeatedly: everyone encounters death, and everyone who can see the moon has something on which to project a figure. The more abstract the database category, the more different histories it can conceal.

This was already a methodological problem before anyone sequenced DNA. In 1916 Edward Sapir warned that geographical extent becomes useful for chronology only after examining how a particular cultural element travels. He also warned against treating broadly defined conceptual similarities as historical entities. That is an excellent description of the trouble that begins when somebody stole fire becomes a single prehistoric heirloom. Sapir, 1916, pp. 36–40.

Genetics gives comparative mythology an immensely valuable external history. It does not exempt mythology from establishing the connection. A dated road is not a dated journey by every traveler subsequently found at its ends.

Who started the modern Out-of-Africa reconstruction? #

The clearest early published anchor is Michael Witzel’s “Comparison and Reconstruction: Language and Mythology,” published in 2001. Its final mythology sections divide the material into large regional types, propose a still earlier common mythology, and explicitly connect the project to life before and after departure from Africa. This is the earliest explicit formulation verified in the source trail examined here.¹ Witzel, 2001, pp. 53–57.

Witzel’s own account explains how he arrived there. During a Japanese sabbatical in 1990, he developed a comparative scheme joining mythologies from Iceland through Egypt to Japan. He presented it at Kyoto on June 30. It was initially a Eurasian comparison. After returning to Boston, he encountered accounts of genetics and linguistics in late 1990 and early 1991 that divided human history into African, Out-of-Africa, and later Eurasian populations. These fitted the mythological arrangement he had begun. Witzel, Origins, foreword, pp. x–xi.

His endnote identifies the bridge: popular articles on genes and language, including Luigi Luca Cavalli-Sforza’s 1991 “Genes, Peoples and Languages.” That article juxtaposes human population and language trees and associates their branching with migrations from Africa. Mythology would become a third history fitted to that arrangement. The geneticist had not dated the myths; the comparative mythologist supplied the extension.² Cavalli-Sforza, 1991, p. 104.

That sequence matters. His comparative observations came first; the demographic interpretation gave them a much deeper history. By 2006 the program had acquired the wonderfully confident title “Out of Africa: The Journey of the Oldest Tales of Humankind.” The paper placed its mythological families alongside population movements and genetic classifications. Witzel, 2006, pp. 21–24, 41–43, 56–57.

Older diffusionists had already imagined immense journeys for cultural material. But an ancient Egyptian or Near Eastern center exporting civilization is a different proposition from Paleolithic humans exporting ancestral myths. In Witzel’s own historiography, Frobenius and Baumann represent older models of spreading cultural complexes; his proposal moves the decisive ancestry much farther back, into the branching history of modern humans. The genetic map supplies a new timescale and a new reason to privilege Africa. Witzel, Origins, pp. 8–10, 15–16.

Witzel turns different mythologies into stages of ancestry #

Witzel’s best-known distinction is between Laurasia and Gondwana. The geological names are labels for mythological groupings, not a proposal that people told stories before the continents separated. Laurasia embraces a large northern complex with an overarching sequence from cosmic beginnings through generations of gods and humans toward destruction or renewal. Gondwana groups southern traditions that, in his classification, lack that same encompassing sequence. Shared material behind both becomes Pan-Gaean. Witzel, 2001, pp. 54–57.

The argumentative hinge is the conversion of difference into succession. A tradition lacking the northern sequence is read as preserving an earlier arrangement. A motif shared across the large divisions becomes a remnant from before they separated. Human population history then helps locate the older common inheritance.

But a classification is not yet a genealogy. A missing overarching narrative could reflect loss, different local arrangements, or the kinds of texts collectors recorded. A widely shared episode could have crossed between groups after their populations diverged. Neither possibility disappears because a typology resembles a migration map.

Witzel’s flood revision shows the machinery particularly well. His 2001 sketch treated the absence of flood myths as a Gondwana characteristic. In a later study he acknowledged that this was wrong: southern flood traditions pushed the flood farther back, into the proposed Pan-Gaean inheritance. The correction improves the distributional account. It does not independently date the flood story. Expanded geographical coverage has been translated into deeper ancestry within the same framework. Witzel, 2010, especially pp. 225–226 and 239.

The crucial evidence would connect the myth to the ancestral population. The arrangement of the theory keeps making that connection available in advance.

Berezkin supplies the map—and borrows the arrow #

Berezkin’s catalogue is the great empirical resource of this literature. It turns scattered ethnographic records into comparable distributions of narrative motifs. His own historical arguments are often more specific than the grand reconstruction of whole mythological systems.

His publication trail also needs to be kept distinct. In his 2021 retrospective, Berezkin traces his first African candidates—origin-of-death motifs—to his own 2007 and 2009 papers. Witzel’s earlier general proposal does not establish that Berezkin borrowed his initial argument from it. The directly inspected synthesis is the 2010 paper; the earlier dates come from Berezkin’s retrospective and bibliography.³ Berezkin, 2021, pp. 91–93.

In “From Africa and Back” (2010), he argues that origin-of-death motifs spread from Africa about 60,000 years ago. He also considers an Asian origin followed by transmission to Africa and the Americas. Africa wins in his account because of the motifs’ African prominence and diversity, interpreted alongside archaeology and genetics. Then comes the sentence that should be printed underneath every resulting map:

“the folklore data itself does not help us to define the direction of diffusion.”

That is Berezkin’s statement, not an accusation from a critic. Immediately afterward, human migration history supplies the direction. The mapped contrasts are his mythological evidence; the movement out of Africa is the external historical argument that orients them. Berezkin, 2010, p. 39.

A second 2010 paper shows the argument in action. For stories of people emerging from underground, Berezkin concedes that African and Asian occurrences alone would leave either homeland about equally probable. Australia tips his judgment: its long isolation makes later Eurasian borrowing less plausible, so the Africa–Australia distribution is assigned to the first coastal dispersal of humans. This is a substantive argument, but its crucial bridge is still historical: why must the shared story belong to the first settlers, rather than to another episode of transmission? The population route identifies a possible journey; it does not establish that this story made it. Berezkin, 2010, pp. 118–119.

His 2021 “African Heritage in Mythology” states the dependency again. Comparative mythology cannot internally determine the dates or direction of the material; an African-origin hypothesis therefore requires external support. He also explicitly observes that genes do not carry culture. Those are sound distinctions. They make the missing demonstration especially visible: the external population movement must be connected to the particular motifs, rather than merely sharing their geography. Berezkin, 2021, pp. 93–94.

Berezkin does not place every widely distributed myth in Africa. His 2018 synthesis gives later formative roles to Southeast Asia, southern Eurasia, and northern Eurasia. His African monograph devotes substantial attention to stories moving back into Africa. These distinctions sharpen the criticism. If myths can spread after the first settlers, cross population boundaries, and enter Africa from elsewhere, an African presence cannot by itself certify an African ancestor. Berezkin, 2018, pp. 149, 155–156; 2013, introduction and conclusion.

This remains a live method. In 2026 Berezkin again explains that folklore lacks an internal chronology and uses the peopling of regions as the clock. He treats the African homeland of animal-trickster narratives as less certain than his other proposed inheritances. The clock still comes from the passengers. Berezkin, 2026, pp. 210–212, 223–224.

When population history chooses the root #

Le Quellec’s 2015 study of human-origin stories gives the transfer from genetic ancestry to myth ancestry a literal location on the page.

After assembling and comparing emergence narratives, he constructs a consensus tree and roots it on a !Kung version. His reason is that this version should reflect an ancient state of the myth, judged from genetic studies placing the population’s ancestral divergence deep in human history. The passage explicitly invokes a divergence at least 100,000 years ago. Le Quellec, 2015, pp. 252–253.

An early population divergence cannot establish an early divergence of whatever story its descendants now tell. A community with a deep genetic ancestry can learn a new narrative tomorrow. Its neighbors can preserve an old one. The date of the population split and the date of the story split are different quantities.

The computational reconstruction may still be useful. Given that root, it estimates which features belong at ancestral nodes. But the root’s genetic justification is an additional historical premise. Numerical precision downstream cannot supply the missing link upstream.

D’Huy’s 2017 matriarchy paper makes the same choice even more explicitly. It invokes the African origin of modern humans and the hypothesis that the mythology accompanied their departure; accordingly, “the tree was rooted in East Africa.” Its conclusion then describes agreement among mythological reconstructions and genetic history as very strong evidence. D’Huy, 2017, pp. 161, 165.

The problem is the change in evidentiary status. An African root can be a legitimate hypothesis to explore. The resulting ancestral reconstruction is conditional on it. Presenting the reconstruction as fresh support for the migration story spends the same historical premise twice.

D’Huy had stated the distinction himself in 2013. In his dragon study, he explains that the genetic migrations chosen to date the myth are those that best match the reconstructed tree. They can therefore supply a proposed chronology, but cannot confirm that tree. That is exactly the standard required here: a history selected for its fit cannot be counted again as an independent witness. D’Huy, 2013, p. 204, “Datation du mythe.”

The earlier d’Huy article examines this problem in serpent trees. Here the larger point is that the move travels between researchers and subjects. Change the motif from serpent to emergence to primordial women, and the ancestry of modern humans remains ready to orient its history.

How the assumption acquired a bibliography #

The citation trail is unusually instructive because the authors often tell us exactly what they borrowed.

Wim van Binsbergen’s 2005 “Mythical Archaeology” explicitly takes up Witzel’s 2001 long-range program. He recodes African cosmogonies into narrative complexes and places them against genetic, linguistic, archaeological, and other proposed chronologies. He rejects Witzel’s sharp southern–northern division and allows extensive Asian innovation and return movement into Africa, while retaining a small pre-Out-of-Africa package. This is a revision inside the broader project, not another independent discovery that worldwide myths began in Africa. Van Binsbergen, 2005, introduction, pp. 14–15 and 32–33.

The revised version published in 2006 makes the progression unusually visible. Out-of-Africa is its declared point of departure; section 5.2 then announces broad confirmation of Witzel’s proposal through agreement among genetic, linguistic, and mythological trajectories. The framework is doing work on both sides of the claimed agreement. Van Binsbergen, revised manuscript, p. 1 and §5.2, p. 10.

Conferences gave the project a shared venue. The 2006 Beijing program includes Witzel’s Out-of-Africa synthesis; the Ravenstein proceedings bring Witzel, Berezkin, and van Binsbergen into the same comparative conversation. Participation establishes contact, not agreement. The stronger evidence of transmission is what later papers actually cite and use. Beijing abstracts, 2006; Witzel’s Ravenstein chapter, 2010.

Witzel’s book makes the exchange concrete: he thanks Berezkin for papers and maps, and dates his discussions with van Binsbergen to their meeting at Leiden in December 2003. The literature grew through an interacting network, with material moving between its participants. Witzel, Origins, foreword, p. xvii.

D’Huy’s matriarchy paper supplies an almost diagrammatic example. On page 160 it quotes Berezkin’s 2013 reconstruction of an African ritual complex involving opposition between the sexes. It then uses Berezkin’s catalogue. On page 161 it supplies the African root. On page 165 it appeals to agreement with Berezkin and Le Quellec. A predecessor’s proposal has become an input, then returned as an apparent corroborating result. D’Huy, 2017, pp. 160–165.

The connection is also personal and methodological: d’Huy and Berezkin coauthored the 2017 Pleiades study. Its reconstructions compare East African and Bantu roots, while the Orion analysis uses other African rooting choices. Comparing two African roots can test sensitivity within the African premise. It does not test Africa against the rest of the world. D’Huy and Berezkin, 2017, pp. 113–118.

By 2018, Marc Thuillard, Le Quellec, d’Huy, and Berezkin were coauthors of a large-scale study using the catalogue. Their paper explicitly declines to present a global phylogenetic network because the data fit it inadequately, especially in Eurasia. The collaboration is documented; the paper itself supplies a warning against turning every geographical cluster into a branch of ancestry. Thuillard et al., 2018, pp. 409–411.

The papers are not interchangeable, and shared data do not make new analyses worthless. They do mean that agreement needs to be unpacked. A new algorithm applied to the same catalogue, interpreted with the same migration history, supplies less independent historical confirmation than a bibliography full of different titles suggests.

The first five rows are documented above; the last two show how the argument is developing now.

More computation does not automatically produce a clock #

Hyunuk Kim, Marcus Hamilton, Woo-Sung Jung, and Hyejin Youn’s 2024 preprint finds a nested structure in Berezkin’s data without feeding geography into the clustering algorithm. That is a real result: the recorded motifs contain geographical information. The study also identifies three extremely widespread motifs involving canine tricksters, in-laws’ tasks, and figures on the moon. Kim et al., 2024, pp. 6–10, 20–22.

But common occurrence is not temporal precedence. Something can sit near the top of a statistical hierarchy because it is generic, recurrent, or readily combined with other themes. That position is not a fossil layer.

The authors acknowledge that their analyses cannot distinguish African ancestral inheritance from independent convergence for those motifs. They nevertheless describe a Neighbor-Net pattern through an Out-of-Africa sequence. The network measures similarities; it has no independently estimated historical root or clock. Its fit to geography is a different result from a demonstration that stories flowed out of Africa. Kim et al., 2024, p. 22 and supplement, section II.B.

The 2025 preprint “Worldwide patterns in mythology echo the human expansion out of Africa” goes further. Hélios Delbrassine and colleagues—including Berezkin, Eugenio Bortolini, Jamshid Tehrani, and Luca Pagani—compare mythological distances with geography and genetic distances. They also introduce ancient-DNA time slices. This brings new evidence to the question. Their overall result gives substantial weight to cultural diffusion, and several previously proposed motif histories show weak genetic association after spatial control. Delbrassine et al., 2025, pp. 4–10.

Its deepest chronology still requires several extra steps. Motifs shared between Africa and South America or Oceania are selected as potentially old; present-day myth differences are compared with reconstructed Eurasian genetic landscapes; Ethiopia serves as an exit-point proxy in another analysis. The association with a preglacial population structure is then extended toward an Out-of-Africa interpretation. Delbrassine et al., 2025, pp. 7–15.

Ancient DNA anchors parts of this genetic time series; present-day samples and assumptions of population continuity fill other parts. The corresponding myths all come from much later records. Even a perfectly sampled 38,000-year-old genetic landscape would not directly date those stories. An association with that landscape is not an observation of a 38,000-year-old telling, much less an African birth certificate. To establish that history requires showing why later transmission would not produce the surviving association. Delbrassine et al., 2025, pp. 7, 12.

These newer studies should be read as different contributions, with different strengths. They also demonstrate why the distinction matters. Discovering geographical structure, detecting an association with demography, identifying a source, and dating a narrative are four achievements. Success at one cannot silently stand in for all four.

Every story has to earn its own history #

The African-origin framework is attractive because it solves an enormous problem with a history that has already been written. Human origins supply the homeland. Migration routes supply the arrows. Population splits supply the dates. A catalogue of recent traditions then seems to recover the stories people carried between those points.

But the decisive historical work lies in connecting the story to the journey. That connection cannot simply be inherited from the geneticists.

There are ways to make a cultural chronology answer to evidence outside its own distribution: dated texts, identifiable borrowings, material associations, independently constrained environmental events, and comparisons that genuinely let different routes and origins compete. The bullroarer complex, for example, invites comparison among instruments, ritual institutions, and narratives. The Bullroarer Atlas makes the physical and ethnographic records available for that work. An instrument’s age still has to be distinguished from the age of its accompanying story. Keeping the objects separate is how an argument gains traction.

None of this requires abandoning deep history. It requires reconstructing it rather than assigning it. Africa can be the right answer for a particular myth. So can somewhere else. The question must remain open long enough for the relevant evidence to choose.

The academic citation system has preserved much of the trail. We can watch Witzel’s comparison meet genetics, Berezkin’s distributions acquire direction, Le Quellec’s population history choose a root, and d’Huy’s conditional reconstructions return as corroboration. What the citations establish most securely is the transmission of an argument among scholars.

An assumption does not become an ancestor because it has descendants in the bibliography.

Sources #

1. Sapir, Edward. Time Perspective in Aboriginal American Culture: A Study in Method. Geological Survey of Canada, Memoir 90, 1916. Especially pp. 36–40.
2. Witzel, Michael. “Comparison and Reconstruction: Language and Mythology.” Mother Tongue VI (2001): 45–62.
3. Witzel, Michael. “Out of Africa: The Journey of the Oldest Tales of Humankind.” Generalized Science of Humanity I. Tokyo: ILCAA, 2006, pp. 21–65.
4. Witzel, Michael. “Pan-Gaean Flood Myths: Gondwana Myths—and Beyond.” In Wim van Binsbergen and Eric Venbrux, eds., New Perspectives on Myth, 2010, pp. 225–242.
5. Witzel, E. J. Michael. The Origins of the World’s Mythologies. Oxford University Press, 2012; online edition 2013. Official preview, including the foreword and opening methodological discussion.
6. Van Binsbergen, Wim M. J. “Mythical Archaeology: Situating Sub-Saharan African Cosmogonic Myths within a Long-Range Intercontinental Comparative Perspective.” Kyoto conference paper, 2005. Revised as “Mythological Archaeology” in Toshiki Osada and Noriko Hase, eds., Proceedings of the Pre-Symposium of RIHN and 7th ESCA Harvard–Kyoto Roundtable, 2006, pp. 319–349. Author-hosted revised manuscript; its own pagination is used above.
7. International Conference on Comparative Mythology: Beijing, May 11–13, 2006. Abstracts. Harvard and Peking University.
8. Berezkin, Yuri E. “From Africa and Back: Some Areal Patterns of Mythological Motifs.” Mother Tongue XV (2010): 1–67.
9. Berezkin, Yuri E. Africa, Migrations, Mythology: Areal Patterns of Folklore Motifs in Historical Perspective. St Petersburg: Nauka, 2013. In Russian.
10. Berezkin, Yuri E. “Manifestation of Worldviews in Traditional Narratives.” Archaeology, Ethnology & Anthropology of Eurasia 46(2) (2018): 149–157.
11. Berezkin, Yuri E. “African Heritage in Mythology.” Antropologicheskij forum 48 (2021): 91–114. In Russian. doi:10.31250/1815-8870-2021-17-48-91-114.
12. Berezkin, Yuri E. “The Spread of Motifs Characteristic of Oral Narratives Featuring Zoomorphic Characters in Sub-Saharan Africa and the Circum-Pacific World: The Problem of Dating.” Etnografia 1(31) (2026): 210–227.
13. Le Quellec, Jean-Loïc. “Peut-on retrouver les mythes préhistoriques ? L’exemple des récits anthropogoniques.” Comptes rendus de l’Académie des Inscriptions et Belles-Lettres 159(1) (2015): 235–266. Rooting argument, pp. 252–253.
14. D’Huy, Julien. “Matriarchy and Prehistory: A Statistical Method for Testing an Old Theory.” Les Cahiers de l’AARS 19 (2016–2017): 159–170. Preprint.
15. D’Huy, Julien, and Yuri E. Berezkin. “How Did the First Humans Perceive the Starry Night? — On the Pleiades.” RMN Newsletter 12–13 (2016–2017): 100–122.
16. Kim, Hyunuk, Marcus J. Hamilton, Woo-Sung Jung, and Hyejin Youn. “Deeply Nested Structure of Mythological Traditions Worldwide.” arXiv:2408.07300v1, 2024. Preprint; version inspected September 19, 2026.
17. Delbrassine, Hélios, Massimo Mezzavilla, Leonardo Vallini, Yuri Berezkin, Eugenio Bortolini, Jamshid Tehrani, and Luca Pagani. “Worldwide Patterns in Mythology Echo the Human Expansion out of Africa.” bioRxiv, January 25, 2025, version 1. Preprint and supplementary information; inspected September 19, 2026.
18. Cavalli-Sforza, L. L. “Genes, Peoples and Languages.” Scientific American 265(5) (November 1991): 104–110. Teaching copy of the original article.
19. Berezkin, Yuri. “The Emergence of the First People from the Underworld: Another Cosmogonic Myth of a Possible African Origin.” In Wim van Binsbergen and Eric Venbrux, eds., New Perspectives on Myth, 2010, pp. 109–126.
20. D’Huy, Julien. “Le motif du dragon serait paléolithique : mythologie et archéologie.” Préhistoire du Sud-Ouest 21(2) (2013): 195–215. Author-uploaded full text, especially p. 204.
21. Thuillard, Marc, Jean-Loïc Le Quellec, Julien d’Huy, and Yuri Berezkin. “A Large-Scale Study of World Myths.” Trames 22(72/67), no. 4 (2018): 407–424.