104
and Greenland, spread across 6000
kilometers.
They found that linguistic relatedness and
geographic proximity were independent
factors
in
predicting
shared
folktales,
calculating that geographical distance is
related to 25.50% of variance in the
inventories. In Lankford’s data, geographical
distance appears related to only 3.09% of the
variance in the first dataset (p=0.004), and 9%
(p=0.002) in the second (Figure 1a–b). In other
words, geographical distribution does not seem
to be a significant factor in which groups share
motifs. This result is far lower than in Ross and
Atkinson’s study, which is a potential indicator
that the distribution of relevant motifs is an
outcome of their longer history rather than
attributable to recent contacts between
adjacent populations and their networks. This
preliminary finding is a potential indicator of
the greater endurance of star-related beliefs
among hunter-gatherer cultures. Similar
conclusions have been reached in studies of
d’Huy (2015a) and da Silva and Tehrani
(2016), although these must be regarded with
caution because their datasets were based on
the Aarne–Thompson–Uther tale-type index,
which is inadequate for such analyses (see
Berezkin 2015; d’Huy et al. 2017). The small
dataset and the large geographic distances
between the groups serve as a reminder that
caution is needed when attempting to interpret
this data. Nevertheless, the analysis of data by
George Lankford suggests strong connections
in star-beliefs across North American cultures
that would at least be consistent with deep
historical roots in these traditions.
Further perspective on this data is provided
by a NeighborNet (Jaccard; Bryant & Moulton
2004) analysis constructed with
Splitstree
4.14.4. (Huson & Bryant 2006) in order to
visualize
specific
relationships
among
traditions of the linguistic-cultural groups
(Figure 2a–b).
The network shows a low mean
delta score (0.32 for the first database; 0.23 for
the second). In general the closer to zero the
delta score is, the more the data will exhibit a
straightforward stemmatic tree. These scores
can be compared to the mean delta-score of the
principal language families of the world. These
language families are generally accepted as
analysable in tree-like relations, and the mean
delta score within these language families has
been calculated by Søren Wichmann et al.
(2001) as 0.31. Comparison with the scores
reflected in Figure 2a–b suggests that the
transmission of these mythological traditions is
more tree-like than for many language
families. Moreover, the NeighborNet analysis
shows a low correlation between the
mythologies and language family. Such low
correlation
presents
the
possibility
of
mythological
substrate
influence
which
Figure 2a: NeighborNet based on the Jaccard folktale distance matrix for the 24 ethnic groups in Lankford’s dataset
exhibiting three or more motifs. Box-like structures indicate a conflicting signal (i.e. suggesting similar independant
inventions, borrowings).
105
antedates the spread of the language families
(cf. also Frog, this issue, on the spread of
Proto-Sámi). Geographical distance appears as
a strikingly low factor in the Mantel tests,
suggesting that contacts between groups have
not been significant
in the distribution of these
traditions. This observation, coupled with the
common ‘Amerindian’ nature of the myths,
makes a hypothesis of recent changes owing to
contact
networks
improbable.
Although
further research is needed, the low correlation
of motifs with language family make it seem
more probable that these elements of star-
related mythology have largely survived
historical changes in culture and language,
potentially going back to the first inhabitants
of the continent. Although these findings
remain tentative owing to limitations of the
corpus
analysed,
mythology of the stars seems
to be particularly well suited to phylogenetic
analysis for long-term perspectives on the
history of motifs, potentially extending back
into the Palaeolithic.
Material
The databases used in this paper were built
from a database developed by Yuri Berezkin.
This database, available in Russian on http://
ruthenia.ru/folklore/berezkin, was consulted in
October 2017. The Analytical Catalogue of
World Mythology and Folklore consists of ca.
55,000
summaries
of
narratives
and
descriptions of mythological ideas among ca.
1500 large and small ethnic groups combined
into almost 1000 traditions, mostly on the basis
of language. Most of the texts in the database
were recorded between 1850 and 1980.
Such narratives easily pass the test of a
certain degree of distortion due to translation,
which affects their linguistic surface but not
their structure. Indeed, according to Lévi-
Strauss (1958: 232), the signification of a text
is preserved even through the worst translation.
The basic content of stories is easily translated,
for which elementary command of the
corresponding language is usually sufficient.
1
What is impossible to translate – people’s
attitude towards the stories and their feelings
when they hear them – are studied by another
discipline: cultural anthropology. Furthermore,
Yuri Berezkin classifies narratives at a high
level of abstraction, which reduces the
probably of impacts from translation on the
encoding of mythological motifs. Such a
classification system avoids dealing with
elements that may be deformed, such as
surface details of narration. Within a corpus of
this size and at such a level of abstraction,
issues of the ‘quality’ of individual sources and
their translation do not present significant
methodological problems in quantitative
analysis.
The mythological motifs selected in
Berezkin’s database are cultural elements
subject to replication. There is no evidence in
the database that all the motifs studied have the
same history. On the contrary, each motif
seems to have a distribution area of its own.
Figure 2b: NeighborNet based on the Jaccard folktale distance matrix for the 15 ethnic groups in Lankford’s dataset
exhibiting four or more motifs. Box-like structures indicate a conflicting signal (i.e. suggesting similar independant
inventions, borrowings).