Content uploaded by Julien d'Huy
Author content
All content in this area was uploaded by Julien d'Huy on Feb 03, 2015
Content may be subject to copyright.
115
Rock Art Research 2013 - Volume 30, Number 1.
BRIEF REPORTS
A phylogenetic approach
of mythology and its
archaeological consequences
By JULIEN D’HUY
Mythology has provided apparently solid and re-
liable clues to the meaning of Saharan rock art (see e.g.
Le Quellec 2004; Le Quellec et al. 2005). Yet to reveal
the mythological meaning concealed in pre-Historic
images may represent the versions of a myth which
are either unknown or diverge substantially from the
be representative of versions belonging to previous,
more archaic stages of development (Devlet and
Devlet 2005: 233). This is why we suggest a new and
alternative method to study ancient mythology.
There are many parallels between the process of
biological and mythological evolution (d’Huy 2012a,
2012b, 2012c). For instance, like genes, mythemes are
discrete heritable units, pass from one generation to the
next and change slowly. Like biological species, myths
parallel features exist, e.g. natural selection vs social
selection and trends; biological mechanism of replication
vs teaching, learning and imitation; genetic mutations
vs innovations and structural transformations; and
horizontal gene transfer vs borrowing. Based on simi-
lar parallelisms between genetic and other cultural
areas, tools from evolutionary biology are being im-
ported to analyse linguistic and cultural phenomena
damental questions in evolutionary science involve
reconstructing ancestral states (Nunn 2011). However,
despite these parallels and common goals, prehistorians
have not used the quantitative phylogenetic methods
that have revolutionised evolutionary biology to
reconstruct ancient mythology that can explain rock
art.We used recently a phylogenetic algorithm, Bio-
Neighbor-Joining (implemented in the program Splits
falls in love with an image of a woman — often a woo-
den doll. The man believes that it is a real woman; it
becomes alive and marries the master. We constructed
a database including the typological variations of the
Pygmalion versions (d’Huy 2012a). The presence or
absence of mythemes for each version was coded as
1 or 0, respectively, to produce a binary matrix of 58
mythemes in 13 versions. Our results implied that the
prep.), we postulate that the more two myths diverge,
the more their ‘genetic’ relationship is distant, geo-
graphically and temporally. The same matrix was re-
analysed in the phylogenetic package Mesquite 2.75,
using a simple model to calculate the parsimony tree-
length of the tree and matrix. Character matrices were
by sub-tree pruning and regrafting. Finally, we rooted
(d’Huy 2009, 2011a, 2011b; d’Huy and Le Quellec
2009; Le Quellec 2012) agreed that Sahara (Berbers)
was a good homeland for a myth that tells us that
explain why wild animals that constitute a threat,
such as felines, elephants and crocodiles, were often
represented incomplete in the Libyan rock art: they
also were less dangerous and could no longer come
alive. One other procedure that constantly appears is
that dangerous animals were represented pierced with
arrows or simply not represented (d’Huy 2009; d’Huy
and Le Quellec 2009; Le Quellec 2012). Moreover,
the Berber myths of Pygmalion seem to be much older
than a Muslim iconoclasm (d’Huy 2011a) and the
ally, a Kabyle version seems to inverse the Egyptian
Tale of two brothers that dates from the reign of Seti
II, who ruled from 1200 to 1194 according to the
structural method (Lévi-Strauss 1983), the Egyptian
tale may be the product of an old Egyptian borrowing
to their Berber close neigbours (d’Huy 2012a). Yet
rooting the tree with the Kabyle (Berber community
the analysis in favour of our own theory. We thus
rooted the tree with two versions of the mythological
family: the Greek and the Bara. Both versions were
closely related (d’Huy 2012a) but also were the more
on an island that did not allow for great population
Rock Art Research 2013 - Volume 30, Number 1.
116
the Greek version has been borrowed by the Greeks
from the Berbers (d’Huy 2011b) and remained isolated
Greek and Bara versions probably preserve one of
shown in Figure 1, and its topology is in agreement
with the tree obtained by the Bio-Neighbor-Joining
algorithm and a Berber origin of the myth (d’Huy
index of 0.60 and retention index of 0.52. The CI is
commonly used to measure the extent of homoplasy
(resemblance not due to inheritance from a common
ancestry, convergence) and the RI is used to measure
synapomorphy (derived states shared by two or more
taxa and their most recent common ancestor) in the
data. High CI and RI values (for example, greater
than 0.60) are usually indicative of low horizontal
transmission (Nunn et al. 2010), yet a RIs corpus
for biological data sets usually ranges from 0.35 to
0.94 (Lycett et al. 2009). Consequently, horizontal
transmissions (borrowings between neighbouring
tribes) did not seem a problem for the phylogenetic
mythological comparative method (d’Huy in prep.),
because it appeared that the majority of mythemes
populations and are consequently relatively conserva-
The orderly and geographically consistent phylo-
genetic signal shows that phylogenetically analysed
consistent with old human migrations. The results
the geographic distribution of the E3b1f haplogroup.
The genetic data suggest an expansion through Tanza-
migration of Bantu-speaking peoples along a similar
(Henn et al. 2008). Whereas
memes can travel without migration, myths seem to
be related to the people’s history (d’Huy 2012a, 2012c,
in prep.). Consequently, the study of myths can be
useful to reconstruct large and ancient movements of
populations. To be accepted, our results imply that
the past populations did not get very far, but this is in
microsatellite data.
What was the ancestral state of the Pygmalion
family by applying to each mytheme of this family
maximum likelihood or parsimony reconstructions
(i.e. phylogenetic reconstruction methods). We only
(probability of more than 75% with the maximum
likelihood method and of 100% with the parsimony
method). Figures 2 (maximum likelihood) and 3
(parsimony) provide an example of how we did this;
the analysed mytheme was ‘the sculptor is a human
Figure 1. Phylogenetic tree of Pygmalion mythological family inferred with Mesquite 2,75.
Figure 2. Ancestral reconstruction of the mytheme ‘the sculptor is a human being’ using maximum likelihood.
117
Rock Art Research 2013 - Volume 30, Number 1.
being’. For each mytheme, both maximum likelihood
and parsimony reconstructions were similar. Maxi-
mum likelihood reconstruction with model Mk1 was
the following tale: a man makes a wooden statue from
a tree-trunk; he or another man clothes it; the statue is
seen as a real person and it becomes alive, thanks to
parsimony reconstruction: a man falls in love with the
version of the descendant versions. Of course, it is im-
portant to remember that the protomyth is likely to be
as rich in complexity as the versions upon which the
reconstruction is based.
The protomyth can itself be used to make inferences
about the behaviour of its Saharan speakers. It in-
forms us about what they were communicating and
documents evidence for a strong belief in the possibility
for an image to come to life (as do-cumented by
archaeological evidences: d’Huy 2009; d’Huy
and Le Quellec 2009; Le Quellec 2012).
The phylogenetic model used in this paper
integrates archaeological, mythological and ge-
netic data. It allowed us to make inferences
about human migrations; test the impact of my-
thological borrowings between neighbouring
tribes; and reconstruct ancestral states of a my-
dence and implication in the rock art inter-
pretation.
Julien d’Huy
École des Hautes Études en Sciences Sociales
75013 Paris
France
dhuy.julien@yahoo.fr
hp://ehess.academia.edu/JuliendHuy
REFERENCES
2005. Myths in stone.
World of rock art in Russia (in Russian)
Moscow.
2008. Y-chromosomal
evidence of a pastoralist migration through Tanzania
PNAS 105(31): 10693–10698.
, J. 2009. New evidence for a closeness between the
Sahara 20: 125–126.
, J. 2011a. Le récit du ‘Chasseur adroit’: un mythe
Kabyle à remonter le temps? Almogaren 42: 37–42.
, J. 2011b. Le mythe ovidien de Pygmalion trouverait
l’une de ses origines dans la Berbérie préhistorique. Les
Cahiers de l’AARS 15: 19–26.
, J. 2012a. Le motif de Pygmalion: origine afrasienne et
Sahara 23: 49–58.
, J. 2012b. Mythes, langues et génétique. Mythologie
française 247: 25–26.
, J. 2012c. Le conte-type de Polyphème: essai de
reconstitution phylogénétique. Mythologie française 248:
47–59.
Figure 3. Ancestral reconstruction of the mytheme ‘the sculptor is a human being’ using parsimony.
Figure 4. Phylogeography of the Pygmalion myths expansion.
Locations of cultures are connected with the phylogenetic tree.
Rock Art Research 2013 - Volume 30, Number 1.
118
, J. in prep. Un ours dans les étoiles: recherche phylo-
génétique sur un mythe préhistorique. Préhistoire du
Sud-Ouest.
, J. and J.-L. 2009. Du Sahara au Nil: la
faible représentation d’animaux dangereux dans l’art
rupestre du Sahara oriental pourrait être liée à la crainte
de leur animation. Les Cahiers de l’AARS 13: 85–98.
, J.-L. 2004. Rock art in Africa: mythology and legend.
Flammarion, Paris.
, J.-L. 2012. Iconoclasties rupestres au Sahara.
Sahara 23: 59–74.
, J.-L., P. and P. 2005. Du Sahara
au Nil. Peintures et gravures d’avant les Pharaons. Collège
de France / Fayard / Soleb, Paris.
C. 1983. How myths die? In Structural anthro-
pology, vol. 2, pp. 257–268. The University of Chicago
Press, Chicago.
and 2009. Cladistic
analyses of behavioural variation in wild Pan troglodytes:
exploring the chimpanzee culture hypothesis. Journal of
Human Evolution 57: 337–349.
, C. L. 2011. The comparative approach in evolutionary
anthropology and biology. The University of Chicago Press,
Chicago.
and M. M.
2010. Simulating trait evolution for cross-cultural com-
parison. Philosophical Transactions of the Royal Society 365
(1559): 3807–3819.
RAR 30-1075
