French Scholar Claims Computer Analysis Can Recover Paleolithic Mythology

I could have swornI talked before about French scholar Julien d’Huy’s claim that a computer program can prove that world mythology can be traced back to a single set of Paleolithic myths. But it turns out that I was mistaken. I actually covered the incredibly similar work of Jamie Tehrani of Durham University and Sara Graca Da Silva of Lisbon’s New University, who applied a computer program to try to use a linguistic analysis to determine the origins of myths and legends. D’Huy, who has followed a similar research program since 2012 and who proudly uses the same methodology, has a new article in Scientific American alleging that he can trace world mythology back to the Stone Age. 

I’m not buying it. At least not in total.
 
Now that’s not to say that myths and legends don’t have deep origins. Walter Burkert suggested that the story of Heracles might date back to the Stone Age, and we know that many myths in the Indo-European family must have derived from a common source before the diversification of Indo-European languages around 4500 BCE. I know that I should be more open to d’Huy’s claims, but the weaknesses in his argument jumped out at me before he had gotten halfway through his argument.
 
Nevertheless, let us try to parse out what d’Huy has done and why it probably doesn’t prove what he said it does.

Note: My objections are based on the information presented in the Scientific American article. Perhaps in some of d’Huy’s other published work, unread by me, he provides answers to these questions. I do not have immediate access to all of it to find out.
 
D’Huy’s methodology is to take a myth and break it down into its component parts. He then fed different versions of stories from around the world (or, more accurately, what look like different versions of the same story) into a computer, which calculated the degree of similarity between versions based on the component parts, or mythemes, present in each. By comparing the mythemes present in each version, d’Huy could therefore create a phylogenic tree showing the relationship of stories to one another. This, in turn, allowed him to calculate the date at which versions diverged by making some rather generous assumptions about rates and reasons for change, ultimately calculating that mythology originated in Africa in the Paleolithic period. He is using evolutionary biology to study folklore.
 
“My phylogenetic studies make use of the extra rigor of statistical and computer-modeling techniques from biology to elucidate how and why myths and folktales evolve,” he writes.
 
In essence, this method is no different than what the early comparative mythologists like Edwin Sidney Hartland and Andrew Lang did. Hartland, for example, found versions of what he thought were elements of the Perseus myth located everywhere on Earth and therefore concluded that they must date back to primal times. Lang thought that a Polynesian myth had a shocking similarity to the Greek myth of the Argonauts and therefore concluded that a lost group of mariners had spread the story around the world. A still earlier generation thought that the presence of flood myths around the world proved that the Noachian deluge really happened and that all accounts therefore derived from those of the three sons of Noah. On the other hand, such comparative mythology can produce important results, as when it identified family relationships among the Indo-European myths. D’Huy’s methodology may be computerized, but it is different only in degree, not in kind. Indeed, he refers to his studies as the “nouvelle mythologie comparée,” or “new comparative mythology.”
 
D’Huy’s best example is probably the story he rather grandiosely calls the “Cosmic Hunt.” It is based on the Greek myth of Callisto. Here is how Apollodorus gives the story in his Library 3.8.2:

But Eumelus and some others say that Lycaon had also a daughter Callisto; though Hesiod says she was one of the nymphs, Asius that she was a daughter of Nycteus, and Pherecydes that she was a daughter of Ceteus. She was a companion of Artemis in the chase, wore the same garb, and swore to her to remain a maid. Now Zeus loved her and, having assumed the likeness, as some say, of Artemis, or, as others say, of Apollo, he shared her bed against her will, and wishing to escape the notice of Hera, he turned her into a bear. But Hera persuaded Artemis to shoot her down as a wild beast. Some say, however, that Artemis shot her down because she did not keep her maidenhood. When Callisto perished, Zeus snatched the babe, named it Arcas, and gave it to Maia to bring up in Arcadia; and Callisto he turned into a star and called it the Bear. (trans. James Frazer)

D’Huy claims that the story can be found in close parallel around the world, and he cites a Native American story from the Iroquois that the same constellation, the Big Dipper, formed when three giant hunters chased and wounded a bear, whose blood spread on the autumn leaves as he climbed a mountain and jumped into the sky. The hunters followed. The three hunters became the three stars of the Big Dipper’s “handle,” and the bear became the four stars of the “dipper,” likely because the four bright stars recall the four paws of the bear. Thus, the hunters forever continue the chase.
 
Early folklorists were struck that the Greeks (but only a few other Europeans before the Middle Ages) and some Native Americans (the Zuni among them) identified the Big Dipper as a bear. And yet many more groups did not: The Egyptians thought it a thigh, the Arabs a coffin, the Germans and some Celts a wagon, etc. But enough thought it a bear that it is an interesting comparison. D’Huy’s analysis concluded that if the Greeks and Iroquois both thought it a bear, then the story must have emerged before both groups separated, which is to say, more than 15,000 years ago.
 
Other versions don’t have hunters, or feature other animals, or even other constellations.
 
To support this conclusion, he boils down the myth to a group of mythemes: “a man or an animal pursues or kills one or more animals, and the creatures are changed into constellations.” Not ever version will possess all of these, but they are the abstract of the story. Yet surely the tale has been so simplified as to be meaningless. Every culture that recognizes constellations has some sort of myth for how they got there. Given the widespread presence of hunting in human culture, it seems ridiculous to suggest that there is no chance for two cultures to have independently struck upon so simple a story. Indeed, the Greeks themselves have another myth that fits the same criteria. Orion the hunter pursued Scorpius the scorpion, and Zeus placed them both among the stars. This myth meets all the same criteria D’Huy laid out in the article. 

So-called "well scene" at Lascaux.

Indeed, when we drill down into the 93 specific mythemes D’Huy uses (sample: the animal is “big,” the animal is “alive”), the only ones that are significant and widespread are the simplest and most obvious; the other traits are quite clearly arranged into culturally specific groupings. Nevertheless, he is committed enough to this idea that he looks at the Lascaux Caves and interprets the famous “well scene,” featuring a bison and a bird-beaked humanoid above what seems to be an arrow or spear, as this Cosmic Hunt. He even reads the black marks beneath the bison as part of the Iroquois myth! “The black stains on the ground under the bison suggest the bloodstained autumnal leaves of the hunted animal.” No, they really, really don’t. That isn’t even part of the story except in eastern North America, and nor would it be in Paleolithic Europe. European fall foliage doesn’t turn as red as it does here in the northeastern U.S.
 
This problem is even more obvious in his second example, the tale of Pygmalion. Most of us know the story of the sculptor who falls in love with his sculpture, which comes to life for him. D’Huy finds versions of the “living doll” myth around the world, and he similarly concludes that they are Paleolithic in origin. But surely any culture where sculpture was an art form must have had individuals who wondered what would happen should a realistic artwork come to life?
 
D’Huy assumes that similarity equals connection, but he provides no way to falsify this to distinguish coincidence or convergent evolution from actual connection. To put it in his favored evolutionary terms, he does not distinguish between analogy and homology.
 
This does not even begin to get into the problems of contamination, cross-pollination, and other ways that myths influenced one another in historic times. This is to say, when Europeans colonized the Americas, they brought their stories with them, and the traces of those European tales folded into Native American oral traditions (best seen in Noah’s Ark stories taken into Native mythology) quite clearly do not imply that Native Americans had those European tales in the Stone Age. Similarly, in Italy in the nineteenth century there were old women telling a folktale that is very clearly the Classical story of Perseus under a different name. Does that mean that Italians and Greeks both inherited the story from an earlier source? No, of course not. The Italian version is a derivative of Renaissance retellings of Classical myth. Would his computer modeling be able to distinguish between archaic Roman myths and the remodeled versions they took on from Greece?
 
There is a problem in that d’Huy is inputting into his model only those stories that he prejudges as to be part of his scheme. This means that there is a potential for his judgment of what stories to analyze to bias the results by selecting only those versions of stories that most closely conform to the model he wants to find. The Callisto story, for example, has more than a dozen variants in the Greek and Latin writers. Forgive me for copying the list from James Frazer rather than trying to look it all up myself: “Paus. 1.25.1; Paus. 8.3.6ff.; Eratosthenes, Cat. 1; Libanius, in Westermann's Mythographi Graeci, Appendix Narrationum, 34, p. 374; Tzetzes, Scholiast on Lycophron 481; Hyginus, Fab. 155, 176, and 177; Ov. Met. 2.409-507; Serv. Verg. G. 1.138; Lactantius Placidus on Statius, Theb. iii.685; Scholia in Caesaris Germanici Aratea, p. 381, ed. F. Eyssenhardt (in his edition of Martianus Capella); Scriptores rerum mythicarum Latini, ed. Bode, i. p. 5; First Vatican Mythographer 17; vol. ii. p. 94, Second Vatican Mythographer 58.” How do you pick which version to declare official? How much more complex is it to do so when all that survives are European missionaries’ culturally biased accounts of non-Western stories?
 
Yet another problem is the issue of mythemes. By breaking the stories he explores down into very simple scenarios and situations, D’Huy accidentally exposes the weakness of his argument: How can we distinguish between the movement of myths and the movement of mythemes? I have no doubt that Paleolithic people had myths and legends, and these could have included elements—“strong hero,” “angry monster,” “evil snake,” etc.—that traveled with the people who migrated out of Africa. But how do we distinguish between a myth that moved wholesale and one that was reconstructed out of very simple mythemes that are universally shared? For example, “hero hunter,” “big animal,” and “animal in sky” are simple elements that might could be recombined without the Callisto story going with them.
 
It did not surprise me to see that d’Huy similarly worked himself back to simplistic (but not universal) ideas that he also considers Paleolithic in origin: the sun as a “big mammal” and women as keepers of sacred knowledge.
 
D’Huy, though, is so confident in his methodology that he believes—get this!—that he can reconstruct early Homo sapiens’ interactions with other hominid species from his computer analysis!

Eventually I hope to go back even further in time and identify mythical stories that may shed light on interactions during the Paleolithic period between early H. sapiens and human species that went extinct. Evolutionary biologists have identified possible interbreeding with Neandertals, Denisovans and perhaps other archaic humans. Material exchanges, as well as language and mythological borrowings, may have also occurred.

This is less than a stone’s throw away from the common fringe idea that Bigfoot is a “racial memory” of Neanderthals or Gigantopithecus or whatever, or the weird idea that the Nephilim were memories of Neanderthals mating with anatomically modern humans.
 
Given how stories change almost beyond recognition even over a few generations, it seems extremely difficult to imagine that d’Huy’s method will uncover genuine Neanderthal encounters. Consider, for example, another piece of the Perseus story: In the 1300s, the author writing as Sir John Mandeville recorded the bizarre tale that was all that remained of the once widespread story of how Perseus rescued the princess Andromeda from where she had been chained to a rock at Joppa for a sea monster to swallow: “And yet there sheweth in the rock, there as the iron chains were fastened, that Andromeda, a great giant, was bounden with, and put in prison before Noah’s flood, of the which giant, is a rib of his side that is forty foot long” (Chapter 5). Mandeville’s Andromeda was a male, and a Nephilim-Watcher bound beneath the Earth (as in 1 Enoch), and had been killed. The only element of the story that remained the same was binding in iron chains. Given how easily mythemes can recombine at will—and Mandeville’s tale is a perfect example of Nephilim mythemes recombining promiscuously—how could we possibly expect to discover stories unchanged from Neanderthal times? How could we distinguish between a genuine Paleolithic myth and one that simply recycled in unrecognizable form some mythemes that might have been Paleolithic?