More than 2000 years ago, the myth of Pygmalion existed in North Africa

More than 2000 years ago, the myth of Pygmalion existed in North Africa
Translation of the research to Julien d’Huy
Based on a new set of mythemes, we reproduce in this paper the results published by d’Huy 2013a concerning the cladistics of the Pygmalion myths in Africa. We conducted a phylogenetic analysis using Mesquite 2.75, in order to obtain the most parsimonious dendrogram. We also used the NeighborNet algorithm, implemented in SplitsTree 4.12.8, to construct a phylogenetic network and assess the extent of borrowing in the construction of this tree. The results of the study largely confirm the original conclusions of d’Huy 2013a, i.e., a horizontal transmission of myths from one generation to the next. Moreover, it is possible to propose reconstructions of the following protomyth of Pygmalion, based on statistic calculations: “ A woman is carved out of a tree trunk by a man to lessen his solitude. A god gives life to the image, which turns into a beautiful young lady. She becomes the wife of her creator, even though another person also wants her to become his wife”. This tale was probably recounted over 3,000 years ago in the north of Africa. Finally, we propose a mechanism that could explain the exceptional longevity of myths throughout time.
Text integral.

1-The best-known version of the Pygmalion myth was proposed by Ovid in Book X of the Metamorphoses (v.243-297): Pygmalion was a sculptor from Cyprus who believed that women were nothing but vices, so he decided not to allow himself to be seduced by any of them. Nonetheless, for several months he worked on carving an ideally beautiful woman from ivory, and quickly fell in love with her: he adorned her with jewels and clothes, and lavished caresses and tenderness on her. Finally, he asked Venus to grant him a young woman who would resemble his work. She granted his wish, giving life and movement to the statue. Their union produced a son, Paphos, who gave his name to the place.

2This story, which is extremely well known today, is a founding myth of the West (Stoichiţă 2008). Critics generally agree that Philostephanos of Cyrene (c. 222-206 BCE) is one of Ovid’s sources. Today, however, it would seem that this founding myth of European culture has its origins in Africa.
Berber versions of Pygmalion predate the Greek version
3-Two stories of the pygmalionesque type were collected in the spring of 1914 by the German Africanist Léo Frobenius, during his stay in Kabylia. In the first story (Frobenius & Fetta 1997: 177-192), a woman offers her husband an engraving of herself, “painted in such a way that one really believed one was dealing with a beautiful, living creature” (Frobenius & Fetta 1997: 179). The wind blows the image out of the husband’s hands and reaches a village chief. He falls instantly in love with the image and isolates himself for many days, “expecting to see her open her mouth and speak”. Indeed, “he believed that this engraving was none other than the living form of a young woman who would soon metamorphose before his eyes” (Frobenius & Fetta 1997: 179). An old woman rouses the chief from his contemplation, makes him understand that it is an image, and offers to kidnap the woman who has served as a model; the chief accepts. The husband is neutralised by the old woman, who pulls out half a hair, and the young woman is abducted. But the faithful wife turns into an ogre and, although locked up, prevents anyone from approaching her. The husband is saved by his first wife, who equalizes the length of his hair, making them all the same length as the half-cut hair, and he manages to save his kidnapped beauty.

4-In the second story (Frobenius & Fetta 1997: 129-133), a man gets a mute woman to react by telling her a pygmalionesque story. The story goes like this. A craftsman carves a naked woman out of a tree trunk ‘so faithfully and precisely that you would have sworn it was a real woman in the flesh’ (Frobenius & Fetta 1997: 132). He then exhibited his work in front of the shop. Two rich merchants, passing by, dressed and adorned the statue. Then two other merchants painted its face and body. Finally, a prophet breathes life into it and the statue breathes, opens its eyes and begins to move. Who deserves to be her husband?

5-How old are these two stories? They are probably pre-Islamic. In Islam, images were suspected of leading to the worship of idols. Several Koranic verses denounce idolatry and the worship of statues (Koran XXI, 52; XXVI, 71; II, 256-257; IV, 51, 60, 76; XXXIX, 17). While no Koranic verse explicitly rejects the painted image, the rejection of idolatry naturally encompasses the idolised object, most often an image or sculpture (Ringgenberg 2006: 21). The hadīts, traditions relating to the deeds and words of Muhammad and his companions, corroborate this iconoclasm: for example, a hadīt reported by El-Bokhari (el-Buhārī) in the ninth century relates that Muhammad, seeing a cushion decorated with figurative representations, accused the artists, by giving life to beings they had created, of having wanted to make themselves the equals of God (Bukhari & Bousquet 1991: 134). The figurative painter thus violates the divine prerogative of creation. God alone has the right to grant existence and form to something that does not exist; any other creation is competition (Ringgenberg 2006: 21). As François Pouillon notes, “the Berbers of the Maghreb, who were brought to Islam late and with difficulty, were particularly radical on these questions of principle. But we would be just as quick to invoke ethnology on this point: Berber handicrafts are no richer in images. So we move from one colonial production to another: Roman statuary or mosaics on the one hand; Ottoman (or Venetian) art of the fixed under glass on the other”. (Pouillon 1996: 183). The two Kabyle stories of the pygmalionesque type, in which the creations are human, idolised and eventually come to life, could therefore only have appeared before the conversion of Kabylia to Islam. It should be added that from the fourth to the seventh century, on the eve of the arrival of Islam, many Berber tribes were Judaized (Hachid 2000: 306); the influence of Judaism, an iconoclastic religion, further delayed the emergence of Kabyle stories of the Pygmalionesque type.

6-The second Kabyle story (Frobenius & Fetta 1997: 129-133) takes us even further back in time to the origin of the motif in Berber lands. This story is very close to the Ovidian version. But it probably predates the Greek version: the Ovidian myth did not really develop until the European Middle Ages (Stoichiţă 2008: 41), and until then was not sufficiently famous to be adopted in North Africa through Roman influence. The only periods in which the myth was transferred through regular contact, direct or otherwise, between Greek culture and the Berbers were Cyrenaica and the Carthaginian civilisation; this period, between the seventh and first centuries BCE, preceded the writing of the Metamorphoses: the myth of Pygmalion would therefore have come from Africa (d’Huy 2011b).

7-Corroborating this hypothesis, the first Kabyle story reverses a passage from the Tale of the Two Brothers, an Egyptian story dating from the reign of Sety II, between 1200 and 1194 BC. The inversions and structural transformations between the two stories (see table below) are such that their similarity cannot be denied: KABYLE TALE
A very beautiful woman has a portrait of herself
(made by the hand of a man).
She offers this image to her husband.
Her husband hunts in the desert while the wife stays
at home.
During one of the husband’s hunts, the wind carries away the
wife’s portrait.
The
wind carries this portrait very far, to a shepherd,
who in turn presents him to his master, the village chief.
At this sight, and on the advice of an old woman, the chief
of the village decides to kidnap and marry the woman. The
kidnapping succeeds by trickery, with the help of the old woman.
The woman involuntarily betrays her husband and reveals
the secret of his immortality.
The husband is temporarily neutralized by using his only weak
point: a hair which, as soon as removed, would lead to his death.
But the old woman only pulls out half of it,
and the hunter’s first wife (first home) manages to save
him by cutting his hair in such a way that they
are all the size of the hair of which half had been torn out.
EGYPTIAN TALE
A very beautiful woman is made by the gods.
As an image, it is given to her husband.
Her husband hunts in the desert while the wife
stays at home.
During one of the husband’s hunts, the Nile carries
away the wife’s hair.
The Nile carries this hair far away, to the chief
of Pharaoh’s washers,
who in turn presents him to his master: Pharaoh. In
view of this, and on the advice of scribes and scholars,
the Pharaoh decides to kidnap and marry the woman. The
kidnapping succeeded by trickery, with the help of a woman
and soldiers.
The woman voluntarily betrays her husband, and reveals
the secret of his immortality.
The husband is temporarily neutralized by using his only weak point:
the pine tree hiding his heart which, as soon as it is knocked
down, leads to his death.
But the heart is preserved in the form of seeds,
and the man’s brother (first home) manages to
save him by plunging these seeds into cool water.

8-By applying the structural rules highlighted by Claude Lévi-Strauss, it is possible to show that the motif originated in Libya over 3,000 years ago (d’Huy 2012c). Indeed, according to the famous anthropologist, the inversions observed between Kabyle and Egyptian stories are often associated with the crossing of linguistic or cultural borders (Lévi-Strauss 1996: 305). In fact, there is a cultural and linguistic difference between the Berber civilisation, of which the Kabyles are representatives, and the Egyptian kingdom (although there may still be passages; see Le Quellec & de Flers 2005, d’Huy 2009a and b; d’Huy & Le Quellec 2009). Moreover, the Egyptians and eastern Libyans have been fighting each other since prehistoric times (Hachid 2000: 92-100). The origin of the motif is undoubtedly Berber, as the Egyptian story has a strong historical basis (Manniche 1975), emphasises distributive justice (the good are rewarded, the bad lost) and replaces the literal expressions of the Berber tale with their poetic equivalents: the engraving made of a woman becomes a woman made by the Gods, the portrait blown away by the wind is transformed into hair carried by the Nile, the thin hair cut off becomes a massive tree that is felled; as Claude Lévi-Strauss notes, referring to a similar Amerindian example, “all these characters show that (…) a decisive passage is made from a hitherto mythical formula to a novelistic formula” (Lévi-Strauss 1996: 310-311).
9-To sum up: of the two Kabyle versions collected, one predates the Muslim conquest and the Ovidian version; the second is said to be based on an Egyptian story dating back more than 3,000 years. This antiquity seems to be corroborated by a fear of animation that may have existed among the ancient Libyans and may have led them, in rock art, not to depict dangerous animals, or to depict them with arrows (d’Huy 2009a, d’Huy & Le Quellec 2009, Le Quellec 2012).

Pygmalion’s road to Africa: or how myths can be followed in their tracks
Preamble

To reconstruct the genealogy of species, biologists have developed complex computer programmes. Even if only a few taxa are studied, the number of possible permutations in a phylogenetic tree increases exponentially; for n taxa, the number of possible trees is 1 x 3 x 5 x 7 x 9 x … x (2n -3). Five taxa therefore yield 105 trees, and ten taxa yield 34,459,425 trees! To choose between the different possible trees, biologists use the method of parsimony, or economy of hypotheses. Of all the possible trees, they select only the one that requires the fewest hypotheses of independent changes, in other words, the one that minimises the number of characters that change several times independently along a phylogenetic tree.

12-Myths – and other elements of culture – are also transmitted by modified descent, which has enabled phylogenetic algorithms to be successfully applied to them (see references in d’Huy & Le Quellec 2014). More specifically, we know that two versions of the same myth tend to diverge progressively over time (Lévi-Strauss 1971: 585) and distance (Ross et al. 2013). Moreover, the evolution of a myth usually takes place very slowly, which makes it possible to recognise, throughout the world, similar complex narratives that spread at the same time as the first migrations of humanity (Witzel 2012). Some Greek stories, such as the myth of Callisto or Polyphemus, can be traced back to at least the Upper Palaeolithic (Berezkin 2006, 2007, d’Huy 2012b and c, 2013b and d, Le Quellec & d’Huy n.p.). Let us add that the idea of applying phylogenetic laws to myths is not a new thing (d’Huy & Le Quellec 2014): Carl von Sydow (Sydow(von) 1948) already considered that tales, like living beings, should be studied from the angle of their ‘biology’, including their formation processes, variations, decline(s) and rebirth(s). However, we were the first, to our knowledge, to apply the algorithms developed by biologists to families of myths (d’Huy 2012a, b and c, 2013a, b, d, e and f).

The 2013 study

With this preamble in mind, let us return to the myth of Pygmalion. We recently drew up a cladogram grouping together almost all the versions of Pygmalion recorded in Africa by Sicard (Sicard (von) 1965), adding to the corpus one version collected by Frobenius (Frobenius 1997: 129-133) and discarding another (Frobenius & Fetta 1997: 179), too far removed from the type, although studied elsewhere (d’Huy 2012a).

14To reconstruct Pygmalion’s tree, we have broken down each version into several myths. Remember that myths are the smallest logical units of a myth, presented in the form of a simple sentence (Lévi-Strauss 1958: 233). The choice of myths was made in order to express the variability specific to each version as widely as possible. However, the statistical translation of our data makes it impossible to take account of inversions, which are then translated in the form of two different syntagms that are not explicitly linked.

15This notion of myth is very important, because it allows myths to be translated into a language that can be used by phylogenetic algorithms. Myths, like articulated language and the genetic code, are made up of “a finite set of discrete units, chemical bases or phonemes, which are themselves devoid of meaning but which, variously combined into higher-ranking units – words of language or triplets of nucleotides – specify either a meaning or a specific chemical substance” (Lévi-Strauss 1971: 612). Language, the genetic code and myths are therefore all replication systems, based on discrete heritable units (phonemes/nucleotides/myths) that are transmitted from one generation to the next. This property allows each of these systems to replicate faithfully from one generation to the next. Moreover, the reduction of myths into myths makes it possible to measure the changes that take place from version to version, with certain myths altering, falling away or appearing progressively over time (Lévi-Strauss 1971: 603-604).

16Once we had established a set of myths reflecting the main variations in our corpus of myths, we coded them, according to their presence or absence, by 1 or 0. We thus obtained as many binary definitional chains as there were versions, which enabled us to apply phylogenetic algorithms to the corpus. These algorithms are normally used to calculate the relationships between species according to their degree of similarity, and then to create a tree summarising this information.

17The tree was created using Mesquite 2.75 (Maddison & Maddison 2011) based on the principle of parsimony. In other words, out of all possible trees, the tree selected was the one that required the least evolutionary change. We used the SPR (subtree pruning and regrafting) method to test this tree, swapping its parts in order to obtain the best possible result.

18The tree we obtained was then rooted on both the Greek version and a Malagasy version, the two most geographically distant versions (d’Huy 2013a). There are other arguments in favour of this rooting: not only is the Ovidian myth the oldest Pygmalion story known today, but this double rooting also places the Kabyle version at the base of the tree: now, as we saw above, it would seem that in certain geographical areas of the prehistoric Sahara, men were afraid of animated images, which can be compared with the animation of the sculpture in the corpus studied. We should add that, as we have also seen, the myth of Pygmalion is very old in the Berber domain.

19Our tree has been constructed on the reasonable presupposition that two versions with very similar characteristics should have diverged recently and that the evolution of the myths was sufficiently slow to allow us, by studying them, to go back relatively far into the past. It should be noted that only the idea of modified descent is necessary for this type of analysis, and not that of competition and natural selection, which is also present in Darwin’s work. This defeats one of Alain Testart’s main arguments (Testart 2011) against the phylogenetic approach to culture.

20-The results obtained were very interesting, since the various versions of the Pygmalion myth were found to be organised in a geographically coherent manner in a gradient running from North to South Africa via the East (d’Huy 2012a), which could be compared with a 2000-year-old migration (Henn et al. 2008). In addition, phylogenetic algorithms enabled us to reconstruct, from the myths selected for our analysis, those that had the highest probability of existing in the proto-myth. We were thus able to reconstruct the following narrative (d’Huy 2013a):
    a man makes a wooden sculpture from a tree trunk; he, or another man,         dresses it. The sculpture is perceived as a living person and one person falls in    love with it. The statue becomes alive, thanks to a third person.

Problems raised by the 2013 study and reassessment

The establishment of the myth base (d’Huy 2013) is problematic, however, because some myths are causally linked to it, which can lead to the exaggeration of certain features and produce undue similarities (d’Huy 2012c, d’Huy & Le Quellec 2013). For example, we had proposed the following myths:
1. The statue is animated by an external individual.
1.1. God animates the statue.
1.2 A divine representative animates the statue.
1.2.1. A teacher of the Koran or a prophet animates the statue.
1.2.2. The medicine man animates the statue.

22-Answering positively to myth 1.2.2. necessarily leads to the validation of myth 1.2. and myth 1, giving them a weighting that can bias the results, which would lead to versions that would not otherwise have been brought together being reconciled. We have therefore recreated a new database of myths, largely different from the previous one, in an attempt to reproduce our results while limiting the causal dependency links as much as possible. We built the database on the basis of Frobenius (Frobenius & Fetta 1997) and Sicard (Sicard (von) 1965), but discarded two versions: one did not provide enough information (Ntumbi), and the other (Rotse) was heavily influenced by royal traditions (and thus reshaped to such an extent that the phylogenetic signal was no longer perceptible; this type of rapid evolution for identity reasons has been studied in d’Huy 2013d). As far as we know, our analysis therefore covers almost all the versions recorded in Africa.

23-The sources do not appear to be very open to question. Sicard’s article brings together different versions collected directly in the field by various ethnologists writing in French, English or German. As for Léo Frobenius, his accounts were first translated into French, probably by Kabyle teachers (Touderti 1998: 359), then translated and published in German before being re-translated into French by Mokran Fetta. But while the various translations may alter the versions of a myth superficially, they cannot touch the heart of it, because the substance of a story lies not in the style, syntax or mode of narration, but in the story that is told. As Claude Lévi-Strauss wrote, “myth could be defined as that mode of discourse in which the value of the formula traduttore, traditore tends almost to zero” (Lévi-Strauss 1974: 240).
The procedure followed for data analysis was similar to that of Julien d’Huy (d’Huy 2013a). One hundred trees were established using Mesquite 2.75 based on the parsimony principle and were tested using the SPR method, then a synthesis of these trees was carried out, in order to obtain the most likely cladogram, known as the consensus tree. The rooting of the Malagasy and Greek versions was retained for the reasons given above. The resulting tree can be seen in Figure 1.
1. Phylogenetic dendrogram of the Pygmalion myth in Africa. The graph has been rooted in both the Greek and Malagasy versions.

25-The consistency index of the tree is 0.61. This index is used to calculate the homoplasic contribution in the constitution of a tree, in other words, the character similarities between taxa that do not come from a common ancestor. It corresponds to the formula R / L, where R is the number of character state changes required to observe all the character states observed and L is the effective number of transformations on the tree obtained. If the CI is 1, there is no homoplasy. However, the CI tends towards 1 if the number of autapomorphies (derived characters specific to a taxon) in the terminal groups increases. Another tool, the retention index (RI), turns out to be lower: 0.44. The RI is the result of the formula: (G – L) / (G – R) where G is the maximum possible number of transformations for a character in relation to the taxa analysed that possess it. It therefore represents the ratio between the number of observable homoplasies and the number of observed homoplasies. It should be as high as possible. A high CI and RI (for example, greater than 0.60), while not measuring the same thing, usually indicate low horizontal transmission (including total or partial borrowing from neighbouring societies) and an essentially vertical direction of transmission, from one generation to the next (Nunn et al. 2010). We would therefore be dealing here with a phylogenetic signal that can be identified by algorithms, but tempered by numerous borrowings. The weakness of the IR can also be explained by the disappearance of intermediate versions, which would have made it possible to establish more solidly the filiation between the versions collected today. It should be noted, however, that the IR obtained from the African versions of Pygmalion remains higher than in many biological datasets that can easily be assumed to have been structured by evolution (Collard et al. 2006).

26-As a myth evolves, it may incorporate elements of another myth, or at least be influenced by it. In order to check the existence of a phylogenetic signal and to assess the extent to which our data can be combined in the form of a tree, we completed our analysis using the Neighbor-Net algorithm, implemented in SplitsTree4.12.8 (Huson & Bryant 2006). The rectangles and diamonds in the graph obtained (fig. 2) from the same data used to create tree 1 indicate conflicting signals, and make it possible to assess the degree of similarity between versions as well as horizontal transfers – in other words, borrowings – from one version to another (Bryant & Moulton 2004, Gray et al. 2010). We note that the versions are grouped by geographical area: this indicates a high degree of diachronic stability in the patterns. Furthermore, the shape of the graph does not resemble a star – which would indicate the absence of a phylogenetic signal – and seems to be organised around a progression across the continent, from north to south or vice versa.


2. Phylogenetic network drawn up from the database used to construct the tree in Figure 1. Note the geographical consistency of the mythemes.
H0.01
LOZI 2 AFRICA_AUSTRAL
LOZI_1_AFRICA_AUSTRAL
SOLD SOUTH AFRICA
MAKUA SOUTHERN AFRICA
CHAGAS EAST_AFRICA
LENGE SOUTHERN AFRICA
KABYLE NORTH_AFRICA
CHICHEWA SOUTHERN AFRICA
BARA MADAGASCAR
BIINE AFRICA_AUSTRAL
PYGMALION_GRECE_ANTIQUE
MAKONDE_AFRIQUE_DE_L_EST
SAFWA_EST_AFRICA
KONO_WEST_AFRICA
SWAHILI EAST AFRICA
The results obtained in previous articles (d’Huy 2012a, 2013a), which highlighted a coherent organisation of versions from north to south, are therefore confirmed by Figures 1 and 2. It should be noted, however, that a late ‘counter-migration’ may have taken place in the opposite direction, from south to north (cf. the place of the Swahilli and Afwa versions), unless the Chichewa and Bwine versions were early and isolated pushes from East Africa to Southern Africa. This latter hypothesis seems the most likely: in fact, an analysis in principal coordinates (c=4) and a non-metric Multidimensional Positioning using a Jaccard distance (Figures 3 and 4) from the same matrix (using Past 3.0 software; Hammer et al. 2001) show a relatively uniform geographical progression. These results are maintained when the distances used are varied (Simpson, correlation and cosine distances). In the images, the red lines separate the different regions, while the black lines indicate the migration cluster, taking North Africa as the point of origin.
3. Principal coordinate analysis using Jaccard distance

Furthermore, it is unlikely that these stories were transmitted independently of a migration. We know that ethnolinguistic barriers prevent the passage of folk artefacts more reliably than that of genes (Ross et al. 2013), which makes self-migration without population displacement (Krohn 1922: 21) less likely than the reverse. Moreover, the spread of the versions seems to be consistent with a 2000-year-old migration of populations (Henn et al. 2008).
4. Multidimensional non-metric positioning using a Jaccard distance

The proto-myth of Pygmalion and its permanence through time

Reconstruction of the proto-myth of Pygmalion

29-Reconstructing the proto-stories that gave rise to today’s major families of myths poses numerous problems that have occupied specialists since the eighteenth century. To solve these problems, mythologists have developed complex methods, comparing the many versions of the same story today in order to reconstruct, in a hypothetical form, a proto-story. These reconstructions shed useful light on several aspects of our past, such as migrations, levels of technology and beliefs that have disappeared. Comparing several reconstructions of proto-myths at different nodes of the tree also highlights the way in which myths evolve and how certain features tend to be preserved – or not – over time.

30-There are often no written records of the earliest versions of a myth. Fortunately, by drawing on numerous versions of the same complex myth and using computational methods borrowed from the evolutionary sciences, it now seems possible to reconstruct mythological ‘trees’ (see above) and identify the genealogical relationships that unite its various versions, in order to reconstruct proto-narratives (d’Huy 2012c, d’Huy 2013a, b, e and f, Le Quellec & d’Huy n.p.).

31-Once the tree has been established, we can estimate the probability that a particular trait existed in the ancestors common to two or more stories. In this way, we can try to establish which ancestral, inherited traits are present in the current versions of the myth, and which derived, more recent traits are sometimes specific to a particular version.

32-The following summary is obtained from the tree tracing the story of Pygmalion. We used two distinct methods (maximum likelihood and parsimony). Passages not in brackets indicate a probability greater than 75% that the myth is present in the proto-myth. Sentences in brackets have a lower probability, between 50 and 75%:

a woman is carved out of a tree trunk by a man to fill his loneliness. (A man who is not the one who made the sculpture dresses the work. A god is asked to bring the sculpture to life). The god gives life to the image, which is transformed into a beautiful young woman. She becomes the wife of her creator, although another person wants her to be his companion too.

33-The differences with the previously obtained account (d’Huy 2013a) are minimal, and essentially due to the redistribution of the myths. It would therefore appear that logical links between myths have a lesser effect on phylogenetic analyses than previously estimated (d’Huy 2012 b and c, 2013b and e, d’Huy & Le Quellec 2013). In order to avoid the rooting bias, which affects the reconstruction of the narrative, we also reconstructed the proto-myth by rooting the tree with the Lozi 1 version:

A woman is carved out of a tree trunk (by a hare, a lunar animal) to furnish her solitude and comes to life. The woman becomes the wife of her creator, although another individual wants her to become his companion too.

Someone comes in the husband’s absence. The chief is informed of the woman’s existence by a servant, and she is taken away by this man of high social standing. (The hero sings. The woman returns alone. The sculpture is thrown to the ground). The woman becomes a tree again.

34-Then with the Venda version, the most southern :

the woman is sculpted by a man (from a veranda post or tree trunk, to furnish his solitude) and comes to life. At first, she becomes the companion of her creator. However, another man wanted her as his wife. (But the sculpture is thrown to the ground and) the woman becomes a tree again.

35-The myths with the highest probability in the first version are found in the second (with the exception of divine intervention and the human status of the creator) and in the third (except for divine intervention and the woman’s directly vegetal origin). It should be remembered, however, that the rootedness that led to the first reconstruction is more credible – for the reasons already explained – than a rootedness with the lozi or venda version.
A proposed mechanism to explain, at least in part, the persistence of the myth

36-It remains to be explained why the myth of Pygmalion has endured over time. This can be explained very simply by the way our brains work: it would seem that we remember better stories in which the images come to life because our own minds unconsciously tend to animate these images, so that we find this type of story credible (d’Huy 2013c) and memorise it more easily. This principle would explain the universality of the motif of the image that comes to life, without accounting for the particular form adopted each time by this motif, which is a cultural fact. However, this neurobiological bias does not explain the permanence of the Pygmalion motif, only the stability of the motif of the image that comes to life.

37-We can see that the narratives reconstructed by the phylogenetic method, while differing according to the points at which they are rooted, always display a great economy of means. Perhaps this is another explanation for the myth’s long continuity. The substance of each version of Pygmalion, like all myths, lies above all “in the story that is told” (Lévi-Strauss 1974: 240). Each local particularisation of the myth takes the form of a narrative, in other words a set of facts linked by causal or, failing that, temporal relationships (Warren et al. 1979, Black & Bower 1980). The causal structure of a story influences its memorability. Black and Bower (Black & Bower 1980) have shown that any action belonging to the chain linking the beginning to the end of a story – in other words, any action that is important, completed and moves the action forward – will be remembered better than other actions that are secondary, unfinished, without effect or useless to the smooth running of the story. The more important an event is in the causal chain, the more important it is judged to be in absolute terms, and the better it is retained (Omanson 1982). This explains the persistence of a structure common to all African versions of Pygmalion:

a woman is carved out of wood and becomes the companion of her creator; but another individual wants her to become his wife, hence the various twists and turns.

38- This is the minimal core of the Pygmalion myth, made up solely of the essential actions and over which wear and tear has little influence, since this core can be found on both sides of the continent, as well as in every reconstruction of the proto-myth. Moreover, as we saw above, this structure supports a social desire to preserve the myth unchanged (Lévi-Strauss 1971: 585).

39-Supporting our hypothesis, the role of causal structure in the memorization of a story has also been amply demonstrated. The use of causal connections – preferably logical ones – between the elements of a text (sentences or actions) predicts how easy it is to remember (Ackerman 1993, Cain et al. 2001, 2004, Myer et al. 1987, Trabasso & Van den Broek 1985, Trabasso & Suh 1993), the importance coefficient attributed to it (Trabasso & Sperry 1985, Trabasso & Van den Broek 1985, Van den Broek 1988) and the probability that certain facts will be retained in a summary of the text (Van den Broek & Trabasso 1986). The idea of a minimal core for each version of a myth, backed up by a desire to be consistent in its transmission, would therefore allow the story to be permanent. This essential core would be made up of the minimum causal structure of the story, to which would be added what each ethnic group considers to be causally essential to the proper conduct of the story.

40- On a day-to-day basis, the permanence of the myth would be based on a twofold movement, constantly repeated, of simplification and then development. Each time it is transmitted, the story is mentally memorised by the receiver in its simplest causal form (minimal core), which is easily memorised and therefore preserved over time; then the story is fleshed out again when the time comes to retell it. Thus, it is the superfluous details of the main theme of a story that first disappear when a story is transmitted orally (Krohn 1971). Similarly, when a subject has to transcribe a story he has just told, his version will be shorter and denser than when it is told orally (Tannen 1982: 8), since he is essentially relying on the logical core of the story. This could be one of the keys to the ability of certain myths to survive over several millennia.

41-The phylogenetic analysis of myths is promising, opening up an extraordinary window onto the way our ancestors perceived the world around them. It makes it possible to reconstruct ancient migrations, to establish the proportion of borrowings and horizontal transmissions for families of myths, to establish their general coherence on the basis of a multitude of features (and not just a few, as has been the case until now) based on the existence of a common ancestor, and to reconstruct prehistoric narratives by calculating the probability that a particular myth was included in them. What’s more, it would seem that in some cases this process allows us to go back in time to the Upper Palaeolithic (d’Huy 2012b and c, d’Huy 2013b, e and f, Le Quellec & d’Huy n.p.). The way is now open to a more detailed analysis of mythological corpuses, from a synchronic and diachronic point of view, in order to reconstruct narratives that were thought to have disappeared forever.

References

Ackerman 1993, ACKERMAN B., Children’s Understanding of the Relation between Referential Knowledge and Referential Behavior, Journal of Experimental Child Psychology, Amsterdam, 56, 1993, p. 385-411.

Berezkin 2006, BEREZKIN Y.E., The cosmic hunt: variants of a Siberian–North-American myth, Folklore (Tartu), 31, 2006, p. 79-100.
DOI : 10.7592/FEJF2005.31.berezkin

Berezkin 2007, BEREZKIN Y.E., Dwarfs and Cranes. Baltic-Finnish Mythologies in Eurasian and American Perspective (70 years after Yrjö Toivonen), Folklore (Tartu), 36, 2007, p. 75-96.

Black & Bower 1980, BLACK J.B., BOWER G.H., Story understanding as problem solving, Poetics, Amsterdam, 9, 1980, p. 223-250.

Bryant & Moulton 2004, BRYANT D., MOULTON V., Neighbor-Net: An Agglomerative Method for the Construction of Phylogenetic Networks, Molecular Biology and Evolution, Lawrence, 21, 2, 2004, p. 255-265.

Bukhari & Bousquet 1991, BUKHARI M.I.I., BOUSQUET G.-H., L’authentique tradition musulmane : choix de h’adîths [traduction, introduction et notes par G.H. Bousquet.], 2ème édition, Paris, Sindbad, 1991, 291 p. (Petite bibliothèque de Sindbad).

Cain et al. 2001, CAIN K., OAKHILL J.V., BARNES M.A., BRYANT P.E., Comprehension skill, inference making ability, and their relation to knowledge, Memory and Cognition, New York, 29, 6, 2001, p. 850-859.

Cain et al. 2004, CAIN K., OAKHILL J.V., BRYANT P.E., Children’s reading comprehension ability: Concurrent prediction by working memory, verbal ability, and component skills, Journal of Educational Psychology, Washington, 96, 2004, p. 31-42.
DOI : 10.1037/0022-0663.96.1.31

Collard et al. 2006, COLLARD M., SHENNAN S.J., TEHRANI J.J., Branching, blending, and the evolution of cultural similarities and differences among human populations, Evolution and Human Behavior, New York, 27, 2006, p. 169-184.
DOI : 10.1016/j.evolhumbehav.2005.07.003

Darlu & Tassy 1993, DARLU P., TASSY P., La reconstitution phylogénétique, Paris, Masson, 1993, 245 p. [disponible sur http://sfs.snv.jussieu.fr/resources/darlu_tassy.pdf]

Frobenius & Fetta 1997, FROBENIUS L., FETTA M., Contes kabyles. Tome 3 : Le fabuleux [recueillis par Leo Frobenius ; trad. des textes allemands par Mokran Fetta], Aix-en-Provence, Edisud, 1997, 222 p.

Gray et al. 2010, GRAY R.D., BRYANT D., GREENHILL S.J., On the shape and fabric of human history, Philosophical Transactions of the Royal Society of London B: Biological Sciences, London, 365, 1559, 2010, p. 3923-3933.
DOI : 10.1098/rstb.2010.0162

Hachid 2000, HACHID M., Les Premiers Berbères : entre Méditerranée, Tassili et Nil, Aix-en-Provence / Alger, Edisud / Ina-Yas, 2000, 317 p.

Hammer et al. 2001, HAMMER O., HARPER D. A. T., RYAN P. D., PAST: Paleontological statistics software package for education and data analysis, Palaeontologia Electronica, 2001, 4(1), 9 p.  http://palaeo-electronica.org/2001_1/past/issue1_01.htm

Henn et al. 2008, HENN B.M., GIGNOUX C., LIN A.A., OEFNER P.J., SHEN P., SCOZZARI R., CRUCIANI F., TISHKOFF S.A., MOUNTAIN J.L., UNDERHILL P.A., Y-chromosomal evidence of a pastoralist migration through Tanzania to southern Africa. PNAS, 2008,105(31), p. 10693-10698.
DOI : 10.1073/pnas.0801184105

Huson & Bryant 2006, HUSON D.H., BRYANT D., Application of Phylogenetic Networks in Evolutionary Studies, Molecular Biology and Evolution, Lawrence, 23, 2, 2006, p. 254-267.
DOI : 10.1093/molbev/msj030

d’Huy 2009a, d’HUY J., New evidence for a closeness between the Abu Râ’s shelter (Eastern Sahara) and Egyptian beliefs, Sahara. Preistoria e Storia del Sahara, Segrate, 20, 2009, p. 125-126.

d’Huy 2009b, d’HUY J., Une volonté de s’approprier symboliquement les qualités de l’autruche serait à l’origine de certains jeux graphiques dans l’art rupestre du Sahara oriental, Cahiers de l’Association des Amis de l’Art rupestre saharien, St-Lizier, 13, 2009, p. 81-84.

d’Huy & Le Quellec 2009, d’HUY J., LE QUELLEC J.-L., Du Sahara au Nil : la faible représentation d’animaux dangereux dans l’art rupestre du désert Libyque pourrait être liée à la crainte de leur animation, Cahiers de l’Association des Amis de l’Art rupestre saharien, St-Lizier, 13, 2009, p. 85-98.

d’Huy 2011a, d’HUY J., Le récit du “Chasseur adroit” : un mythe kabyle à remonter le temps ?, Almogaren, Wien, 42, 2011, p. 37-42.

d’Huy 2011b, d’HUY J., Le mythe ovidien de Pygmalion trouverait l’une de ses origines en Berbérie préhistorique, Cahiers de l’Association des Amis de l’Art rupestre saharien, St-Lizier, 15, 2011, p. 19-25.

d’Huy 2012a, d’HUY J., Le motif de Pygmalion : origine afrasienne et diffusion en Afrique, Sahara. Preistoria e Storia del Sahara, Segrate, 23, 2012, p. 49-58.

d’Huy 2012b, d’HUY J., Le conte-type de Polyphème: essai de reconstitution phylogénétique, Bulletin de la Société de Mythologie Française, Paris, 248, 2012, p. 47-59.

d’Huy 2012c, d’HUY J., Un ours dans les étoiles : recherche phylogénétique sur un mythe préhistorique, Préhistoire du Sud-Ouest, Cressensac, 20, 1, 2012, p. 91-106.

d’Huy & Le Quellec 2013, d’HUY J., LE QUELLEC J.-L., Comment on: Robert M. Ross, Simon J. Greenhill and Quentin D. Atkinson (2013), Population structure and cultural geography of a folktale in Europe, Proceedings of the Royal Society B. Biological Sciences, vol. 280 no. 1756., Nouvelle Mythologie comparée, Paris, 26/06/2013, [en ligne] http://nouvellemythologiecomparee.hautetfort.com/archive/2013/06/26/julien-d-huy-and-jean-loic-le-quellec-comments-on-ross-green.html, 2013.

d’Huy 2013a, d’HUY J., A phylogenetic approach of mythology and its archaeological consequences, Rock Art Research, Caulfield South, 30, 1, 2013, p. 115-118.

d’Huy 2013b, d’HUY J., Polyphemus (Aa. Th. 1137). A phylogenetic reconstruction of a prehistoric tale, Nouvelle Mythologie comparée, Paris, 20/01/2013, [en ligne] http://nouvellemythologiecomparee.hautetfort.com/archive/2013/01/20/julien-d-huy-polyphemus-aa-th-1137.html, 2013.

d’Huy 2013c, d’HUY J., Neural correlates of myths in which an image becomes alive, Leonardo, Oxford, 46, 2, 2013, p. 145-150.

d’Huy 2013d, d’HUY J., Les mythes évolueraient par ponctuations, Mythologie française, Grasse, 252, 2013, p. 8-12.

d’Huy 2013e, d’HUY J., A Cosmic Hunt in the Berber sky: A phylogenetic reconstruction of a Palaeolithic mythology, Cahiers de l’Association des Amis de l’Art rupestre saharien, St-Lizier, 16, 2013, p. 93-106.