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.
Julien d'Huy* and Jean-Loïc Le Quellec**
*Institut Marcel Mauss (UMR 8178), Equipe LIAS, Ecole des Hautes Etudes en Sciences Sociales (France).
**Centre d'Etudes des Mondes Africains, CNRS, UMR 8171 - Honorary Fellow, School of Geography, Archaeology and Environmental Studies - University of the Witwatersrand, Johannesburg.
The paper by Ross and al. is based on contrasting features of the tale AT 480 as presented by Roberts (1958). In their use of these data, the authors carried out problematic, non-explicit operations, and they made factual errors when counting the occurrences. These problems are indicative of a lack of rigor and call into question the relevance of their work.
I – Logical dependencies between the features.
One matter of concern in this paper is the logical dependencies between the features (see Supplementary Table S2). The authors are obviously aware that a number of these features are hierarchically structured, ant they illustrate the structure by levels of indent in the table. For instance:
VI B1 Helper advises heroine
VI B1a Helper advises heroine to demand objects one at a time
VI B1b Helper advises which reward to choose
VI B1c Helper advises heroine to demand objects one at a time
VI B1d Door (variable created by Ross and all)
VI B1d1 Not to open the door
VI B1d2 Helper advise heroine to open door
The problem is that a "yes" on item B1d1 or B1d2 implies a "yes" on items B1d and B1. This can affect the results of the analysis by exaggerating the clustering. Some clusters are created artificially by the logical structure of the questionnaire rather than by the historical processes the authors are trying to detect.
II - The trait VII D.
In their Supplementary Table S1, Ross et al. provide these informations :
VII D1 10 Herd of horses
VII D2 7 Herd of cows
VII D4 2 Sheep
These data include and adapt Roberts’ data (1958: 94), that are as follows :
"D. Animals
D1. Herd of horse
FE 4, 6, FH 1, SR 27, 30, SRF 3, 11, SP 6, 9, SC 6, RPAm 2, RR 2, Afr 1, 2, 4, 20, 22, 23, 26, 30.
D2 Herd of cows
FF (81), SRF 11, SP 9, 41, SC 6, RPAm 2, RR 2, RI (11), Afr 1, 3, 4, 20, 22, 23, 31, 36, 40, 41, 47, Am Neg 10.
D3 Goats
Latv 5, RPAm 2, Afr 3, 22, 25, 26, 31, Am Neg 10.
D4 Sheep
SC 6, RPAm 2, RR 2, Afr 3, 4, 26, 31, Am Neg 10.
D5 Other
SRF 11, SP 5, Ind 1, Born 1, Afr 4, 20, 22-26, 31, 32, 39, 40, 47."
The table below compares original Roberts’ data with their use in
Ross et al. This table includes, on the first hand, Roberts’ abbreviations and comments (1958: 12) and, on the second hand, the number of versions that check the features from VII D1 to D5. We used the T columns to calculate the total number of versions found by Roberts, while the Eu columns show the number of European versions selected by Ross et al. ("In total, our analysis included 700 folktale variants drawn from 31 European ethnolinguistic populations, with a mean of 23 folktale variants per population (Armenian, 3; Basque, 2; Bulgarian, 8; Czech, 11; Danish, 48; English, 8; Estonian, 16; Finnish, 83; Swedish in Finland, 25; Flemish, 6; French, 16; German, 61; Greek, 11; Icelandic, 11; Irish, 22; Italian, 33; Latvian, 13; Norwegian, 48; Polish, 45; Portuguese, 2; Romanian, 4; Russian, 32; Finno-Ugric in Russia, 23; Scottish, 3; Slovenian, 6; Spanish, 11; Swedish, 101; Swiss German, 3; Turkish, 32; Walloon, 3; Yugoslavian, 13)." p. 3).
(click on the image to enlarge)
We found two serious problems here:
1. There are 11 (and not 10) versions that check the feature D1 as indicated in the column D1(Eu).
2. The disparition of the feature D5 ("other") in the authors’s tables (cf. Supplementary Table S1) can be explained by their explicit choice to exclude the traits coded as "other" ("traits coded as ‘other’ were excluded because it is a catchall category"); yet that does not justify the suppression of the trait D3 "goats" that is found in a Latvian version. In this table, one can see that none of the traits selected by Ross and al. are <2 ; this suggest that the unique trait were systematically eliminated.
III - The trait II A2.
The Supplementary Table S1 gives the following data for this trait :
II A2 40 The main actors are sent for water
II A2a 13 To a spring
II A2b 16 To a well
II A2c 2 To a river
This is based on Roberts (1958: 76) :
"A2 The main actors are sent for water
2a To a spring
RILit 1A, RFLit 1, GSLit, GS 3 A, 6 A, GeAM 7, 8, Am Ind 1A, CWGy 1, CI 8, RF 1, 4-6, 8, (11), RFWal 1A, RI 11, Burm 4, Afr 48, Am Neg 1.
2b To a well
GELit, GN 6, (30), GS 11, 9A, GG 34, 51, GF 1, GE 5, GEAm 12, CS 3, CI 4A-6A, SB 5, RI 4A, Gre 1, Am Neg 1.
2c To a river
GN 1A, RSAm 10, 14, RS 3, Burm 1, Am Neg 15, 22.
2d Other
GS 12, GSF 2A, 3A, GD 15, 4A, GG 4, 3A, SP (21), RFAm 1, RI 12, Afr 4, 34, Am Neg 10, (16)."
The following table is organised like the previous one:
(click on the image to enlarge)
The data coded by Ross et al. are problematic :
1. There are 14 occurrences of trait II A2a, not 13 as indicated by Ross et al.
2. As seen above, the authors explain that they exclude the traits coded as "other" ("traits coded as ‘other’ were excluded because it is a catchall category such that a shared presence of ‘other’ does not represent similarity"); yet it can be seen that the trait 2d "other" was included in the total calculation of the versions that verifies the proposal A2 ("The main actors are sent for water") whose value 40 corresponds to the sum A2a-d (13+16+2+9). This operation could be justified, because "other" refers here to a place where there is water that is not "spring", "well" and "river"; however, this is in direct contradiction to the premise: "traits coded as 'other' were excluded".
3. The introduction of a numeric value for the generic trait II A2 is also problematic, compared to the treatment used for other traits. In the case of the previously seen VII D trait, the authors excluded VII D 5 "other". However, the generic trait VII D being "animal", VII D5 means "other animal", that is to say neither "herd of horse" (D1), nor "herd of cows" (D2), nor "goats" (D3), nor "sheeps" (D4). Therefore, the authors could have given a numeric value for VII D by summing D1-D5, as they did for II A2 by summing A2a-d. One cannot address the problem of the "groups of traits" by using different criteria in the same analysis ; this precludes any positive result.
IV – The origin of the problem
Ross et al. are an integral part of a long tradition in mythological studies, tradition which they do not seem to be aware of. Yet it is important to be conscious of the theorical origins of this work.
According to Carl Wilhelm von Sydow (who himself was following a long tradition, see Hafstein 2005), folktales are like biological beings. They tend to adapt to their environment and they evolve by means of natural selection. This explains why so many variants differ from general tale-types. As soon as 1909, Arnold van Gennep stated that folkloric elements should be studied ''comparatively, with the aid of the biological method'' (van Gennep 1909: 84). Starting in 1927, Von Sydow introduced the concept of oicotypification into folkloristics, that is to say "a certain unification of the variants within one and the same linguistic or cultural area on account of isolation from other areas" (von Sydow 1948: 238). He says that "changes brought about by various kinds of mutation, and by oicotypification, are naturally of the greatest weight for a scientific study of folklore" (von Sydow 1948: 239). In line with this approach, some recent works have then tried to apply phylogenetics to mythology (d'Huy 2012 a, b, c ; 2013 a, b), but Ross et al. do not mention their predecessors.
V – Folktale and popular litterature
According to the authors, most of the folktale variants they studied were collected during the late nineteenth and early twentieth centuries, before communication technology and air travel transformed the way ideas and people spread. Yet they forget the influence of popular litterature.
Versions of the "The tale of the kind and the unkind girls" have played a part in literary tradition: they are included in Basile's Pentamerone ("Le tre fate" ie "The Three Fairies", 1634), Peele's The Old Wives Tale and Shakespeare's The Merchant of Venice (for the motif of the choice among casket); Perrault's Contes de ma mère l'Oye ("Les fées" i.e. "The Fairies", 1697); Grimm's Children's and Household Tales ("Die drei Männlein im Walde", i.e. "The tree little men in the wood" and "Frau Holle", i.e. "Mother Holle", 1812), Bechstein's ("Goldmaria und Pechmaria", i.e. "Gold-Mary and Pich-Mary", 1845), Naubert's ("Der kurze Mantel", i.e. "The Cloak", 1789).
Yet during the 17th and the 18th centuries, a lot of chapbooks circulated first in European cities and then throughout rural areas around. These books borrowed their material from folk culture as well as from literary classics, and they touched whole social group, like farmers, artisans, and peasants.
For instance, in France, the easily available, crudely illustrated volumes that made up la bibliothèque bleue were often read during long winter evenings. Those who could not read listened to the readers and memorized the tales, while the contents were copied down by those who could write (Malarte-Feldman 2008: 120). After the first publication of Perrault's book, La Bibliothèque Bleue also republished his stories (Soriano 1958, Mandrou 1985). Yet the texts were often modified and simplified to make them more attractive and easier to read, though they were still published under Perrault’s name. Literary fairy tales finally returned to traditional folklore (Jean 2007: 281-282). Perrault’s and other popular authors’ influence on folklore is very important, as shown by the case of The Sleeping Beauty that has been folklorised, although the story was originally a literary tale (Jean 2007: 281-282). Into another linguistic area, the earlier oral versions of some Irish oral legends were eclipsed by a mid-eighteenth-century Irish chapbook by James Cosgrave, 2008: 181). Grimms' Fairy Tales were also very popular in their written form. On the other hand, some written version of the tale of The Kind and the Unkind Girls, as the Peele's one, remained little known and had no influence on the oral form of the tale (Roberts 1958: 121),
The flaw here is that litterary publications may allow a deeper structuration of a folktale in a specific linguistic area: each text was written in a certain language and it is most likely to spread where this language is spoken.
VI – The slow evolution of myths
Ross et al. suggest that language differences between cultures create stronger significant barriers to folktale diffusion than those for the exchange of genes. Folktale and language histories are decoupled, they say, because "the folktales spread much later than the spread of languages across Europe or because any legacy of deep cultural ancestry inherited down language lineages has been obscured by subsequent folktale evolution and geographical diffusion."
The reported data, however, could also suggest that folktale preserves themes and motifs that are more ancient than language. One of us (d'Huy 2013a) suggests that myths can show little evolutionary change for most of their history, evolving significantly only during rare and rapid events of branching speciation.
In agreement with this slow evolution, mythologies preserve themes and motifs that are very ancient. Claude Lévi-Strauss revealed that all the mythologies throughout the entire American continent ultimately derive from a single myth that the populations have gradually transformed through some mysterious impulse (1981: 563), and this is why a large part of the Amerindian mythology remained almost the same since the first settlement(s) ― albeit through many variants and inversions. For instance, the dance choreography of Brazilian Krahô-speakers today is similar to rock art images painted by their ancestors, 12000 years ago (Guidon and Buco 2006: 128). Creation myths of the northern part of the Meso-American area show a strong relationship with similar stories from the Pacific islands, and from East and South Asia; their distribution suggest an ancient channel of cultural contact between Eurasia and America (Hatt 1949a, Rooth 1957). The fact that many complex Eurasian and American myths are highly similar, as well as the congruence of many motifs in North America and North-East Asia, show that these stories could only spread when a former land bridge joined present-day Alaska and eastern Siberia during the Pleistocene ice ages (Berezkin 2004). These myths should be palaeolithic (Berezkin 2005a, b; Witzel 2013). Consequently, mythologies are stable and can conserve deeper signals than language.
The authors would like to thank Vincent Hirtzel for his help and his useful comments.
Berezkin YE. 2004 Southern Siberian-North american links in mythology. Acta Americana 12(1), 5-27.
Berezkin YE. 2005 Continental Eurasian and Pacific links in American mythologies and their possible time-depth. Latin American Indian Literatures Journal 21(2), 99-115.
Berezkin YE. 2005 The assessment of the probable age of Eurasian-American mythological links. Archaeology, Ethnology & Anthropology of Eurasia 1(21), 146-151.
Guidon N., Buco C. 2006 Zone 3 : Brésil ― Nordeste - États du Piaui, Pernambuco, Rio Grande do Norte et Paraiba. in : Rock Art of Latin America and the Caribbean. Paris, ICOMOS, 122-137.
Hafstein VTR. 2005 Biological metaphors in folklore theory: an essay in the history of ideas. in: A. Dundes (ed) Folklore: critical concepts in literary and cultural studies. London / New York: Routledge, 407-435.
d’Huy J. 2012a Un ours dans les étoiles: recherche phylogénétique sur un mythe préhistorique. Préhistoire du Sud-Ouest, 20(1), 91-106.
d’Huy J. 2012b Le motif de Pygmalion: origine afrasienne et diffusion en Afrique. Sahara 23, 49-58.
d’Huy J. 2012c Mythes, langues et génétique. Mythologie française 247, 25-26.
d’Huy J. 2012d. Le conte-type de Polyphème: essai de reconstitution phylogénétique. Mythologie française 248, 47-59.
d'Huy J. 2013a Polyphemus (Aa. Th. 1137): a phylogenetic reconstruction of a prehistoric tale. Nouvelle Mythologie Comparée (New Comparative Mythology) 1.
d'Huy J. 2013b A phylogenetic approach of mythology and its archaeological consequences. Rock Art Research 30(1), 115-118..
Jean L. 2007 Charles Perrault's paradox: how aristocratic fairy tales became synonymous with folklore conservation. Trames 11(61/56), 3, 276-283.
Lévi-Strauss C. 1981. The Naked Man, London: Jonathan Cape.
Malarte-Feldman CL. 2008 La Bibliothèque bleue. In: Haase [ed] The Greenwood encyclopedia of folktales and fairy tales. Westport, conn: Greenwood Publishing Group, p.120.
Mandrou R. 1985 De la culture populaire aux XVIIe et XVIIIe siècles: la Bibliothèque bleue de Troyes. Paris: Imago.
Ó hÓgán Dáithií 2008 Chapbook. In: Haase [ed] The Greenwood encyclopedia of folktales and fairy tales. Westport, conn: Greenwood Publishing Group, p.180-181.
Roberts WE. 1958 The tale of the kind and the unkind girls: Aa-Th 480 and related tales. Berlin, Germany: De Gruyter.
Soriano M. 1968 Les contes de Perrault: culture savante et traditions populaires. Paris: Gallimard.
Sydow, CW. von. 1932 Om traditionsspridning. Scandia: Tidskrift för historik forskning, Nov., 322-344.
Sydow, CW. Von. 1948 Folktale studies and philology: some points of view. In: A. Dundes [ed.] The Study of Folklore. Englewood Cliffs: Prentice Hall, N. J. 1965, 219-242.
Sydow CW. Von. 1927 Folksagoforskningen. Folkminnen och Folktankar 14, 105-137.
van Gennep A. 1909 Religions, mœurs et légendes: essais d’ethnographie et de linguistique (deuxième série). Paris: Mercure de France.
Witzel EJM. 2013 The origins of the world's mythologies. Oxford: OUP.