free web hosting | website hosting | Business Web Hosting | Free Website Submission | shopping cart | php hosting
From an African bestiary to universal science?

Main text Part I

by Wim van Binsbergen

homepage | Animal symbolism overview page | Part II

© 2002 Wim van Binsbergen[1]

1. Introduction: Universality and the extension of knowledge systems in space and time

On 23 November, 2001, I was invited by the Netherlands Society for the Philosophy of Science to deliver an address on ‘Knowledge and culture’. My discussion focussed on Sandra Harding’s work over the past decade in the philosophy of science.[2] While critical of the so-called ‘strong’[3] variant of relativism in the philosophy of science, she exhibits a weaker relativism; this makes her explore the social, political, economic and historical reasons which may have led to modern (i.e. Western, North Atlantic, or cosmopolitan science) to endow itself (rightly or wrongly) with the three crucial characteristics of rationality, universality and objectivity. Yet Harding clearly hopes that these external forces will not be all, and that there will turn out to be something in the internal nature and the special epistemology of modern science that would justify its claim to these three characteristics, regardless of the historical political package that has lend extra credibility to such a claim.

                        One of Harding’s arguments is that modern science appears universal, not only because

1.   it effectively applies wherever we can ascertain such applicability by means of sensory perception, but also because

2.   it is represented everywhere: there are no clear social, ethnic, linguistic or national limits to its actual application, its sphere of effectiveness in fact appears to extend endlessly; in fact, everywhere in the contemporary world there are recognisable centres of exchange for that science (in the form of universities, research institutes, schools,. book shops, museums, Internet sites, television documentaries, experts, etc.), both among professional scientists and between science and the wider society.

                        Without contesting the validity of Harding’s insight on this point, my familiarity, as an ethnographer, historian, and intercultural philosopher, with a number of other systems of knowledge than modern science makes her insight appear in a different light. In fact, many of these non-scientific systems of knowledge have a geographic extension of applicability that is far from local, showing an amazing continuity or convergence at a continental and even intercontinental scale. The point deserves to be developed in detail and with proper empirical backing, because if it can be shown to be true, it would help us take a relative view of the distinction between modern science and other systems of knowledge, and help us appreciate their nature, spread, and persistence over long stretches of time and space.

                        That a world-wide continuity in systems of knowledge is not only found in contemporary cosmopolitan science but is a long-established fact of cultural history, may be argued, in the first place, on the basis of the extremely wide spread of major mythological patterns like that of ‘hero fights monster’ (cf. Table 1), which we will summarise in the next section.

2. A near-universal theme in systems of mythological knowledge: ‘hero fights monster’

Bodies of mythological knowledge are among mankind’s oldest attested and (with important exceptions, see below) best studied systems of knowledge. The recognition of the similarity of mythological patterns as found in distinct linguistic and cultural tradition was already a fact in Antiquity, when it inspired the practice of the interpretatio graeca::[4] the projection of Greek mythological proper names and concepts onto the mythologies and ritual practices of the Egyptians, Scythian, Celts, etc. at the periphery of the Greek world — a practice well-known from the works of Herodotus and Plato. World-wide, the available mythological material is of an incredible wealth. To make, for the mere purpose of setting the introductory framework, the point of far-reaching continuity and convergence here, I prefer to select only one mytheme (i.e. the smallest meaningful unit of mythological narrative), that of ‘hero fights monster’, and to study it by reference to just one, highly reliable and authoritative, source: the account of Fontenrose’s explorations into the charter myth of the famous Delphic oracle in Ancient Greece. The mytheme involves two archetypal characters, the hero and the adversary, to which often a third is added: the usually passive heroine.

                        The table demonstrates the truly amazing, nearly universal distribution of this mytheme across world cultures.

Click here to open Table 1

Table 1. A near-universal theme of systems of mythological knowledge: ‘hero fights monster’

Table compiled on the basis of scattered information contained in: Fontenrose, J., 1980, Python: A study of Delphic myth and its origins, Berkeley etc.: University of California Press, reprint of the 1959 edition

                        Explaining such a near-universal distribution is another matter, and in this connection a number of hypotheses will be developed towards the end of the present paper. As far as our mytheme is concerned, the global distribution does not necessarily confirm a hypothesis (however obvious that would be) in terms of diffusion from a particular well-defined and limited geographical origin. Cross-cultural studies have always cherished the hypothesis of a constant, universal structure of the human mind and the human body, as a rival explanation for cross-cultural convergence of specific cultural and social-organisational traits, which might otherwise have to be explained in terms of diffusion from a specific origin. For the global distribution of our mytheme this means that it is quite possible that its central struggle merely reflects an internal struggle which occurs, time and time again, in the mind of every human being whenever and wherever, so that the expression of that struggle in myth would be a case of ever repeated parallel invention inspired by the universal characteristics and tendencies of the human mind — and not the gradual diffusion, from a specific origin in time and space, of a mytheme that was only invented once for all and, instead of being locally produced from scratch all the time by human minds, was transmitted by means of explicit cultural communication from one culture and period to the next, undergoing major changes in the process, but still retaining its basic mythematic structure. In his concluding pages, Fontenrose himself tends to an explanation in terms of the struggle that is part of the universal human experience: for him the mytheme sums up every human being’s life’s story in the face of inevitable death — the hero is simply Everyman. Personally I feel that here he was unduly yielding to the anti-diffusionist and pro-localising tendencies of cultural analysis in the middle of the twentieth century: the Everyman interpretation is disappointingly unspecific and blunt, in view of the many world-wide parallels, not just in the overall mythematic application but especially in the details of its elaboration and application — reference to some kind of historical process to explain these parallels by reference to the emergence and interactions of specific cultural formations would seem to be needed at least in addition to the all too predictable Everyman hypothesis.

                        It is however important to keep in mind that Table 1 merely shows a pattern of distribution in time and space, of one mytheme that has been defined in purely typological terms. The typological similarity between the mythemes found in the various cultures listed in Table 1, does not in itself allow us to take a further step and already take the typological similarity as evidence for a generic, historic relationship — as if we can already take for granted that ‘hero fights monster’ is in fact one and the same story told all over the world in only superficially differing ways. Methodologically, the step from distribution to historical explanation is a very major one, and it does not advance our insight if we would pretend that it is not.

                        One might even go further and question the assumption that Table 1 in its many cells lists in fact the recurrent occurrence of one and the same phenomenon here summarised by the concept of the mytheme. One could argue that as a result of the richness of narrative free variation in all the many different myth to which Table 1 refers, all the mythical characters listed are truly incomparable. This would make their reduction to the simple formula of ‘hero fight monster’ to an absurdity violating the literary contents and its value.

                        My response to such a challenge would be that the structuralist analysis of myth[5] has brought us to postulate that underneath the myths’ narrative surface structures (which certainly also need to be analysed in their own right) relatively simple schemes lurk, which are recurrent in space and time, and whose identification allows us to appreciate the structural unity underlying the surface diversity. This is not an appeal to any universal identity of the human experience or human mind, not to an idiosyncratic, intuitive method of literary hermeneutics, but to an body of theoretical viewpoints and analytical procedures (first formulated by Claude Levi-Strauss against the background of the linguistic and psychological structuralism emerging in the first half of the twentieth century) that allow contemporary academic analysts of myths, working within the continually developing intersubjective canons of their sub-discipline, to detect mythical infrastructures, to analyse individual surface myths as transformations of such an infrastructure, and by this procedure define, in considerable detail, the systematic correspondence and differences between surface myths, as found the same and different cultures and historical periods.

                        Table 1 is only meant as an initial example of the kind of evidence we have for claims concerning the wider distribution in time and space of ancient, non-scientific systems of knowledge. We will not attempt here to subject the material presented there to further analysis. Meanwhile the table illustrates another thing: the paucity of African references in the context of this kind of analysis. Inevitably, ancient Greek material is over-represented in Fontenrose’s data base (he is primarily a classicist), and what little African references his book contains derives from ancient Greek sources. For the purpose of illustrating the world-wide distribution of the ‘hero fights monster’ mytheme this is immaterial. However, the problem is much wider: as compared to the wealth of academic knowledge production on kinship, social and political organisation, ritual, work on African myths is relatively rare, and whenever it exists it is usually in such an obscure and specifically African format and context that it is not available for intercontinental cross-cultural comparative studies by scholars who are not themselves Africanists.[6]

                        The anthropological study of myth has traditionally been coupled to that of a topic that captivated nineteenth and early twentieth century researchers but that has since sunk into virtual oblivion: totemism, by which is meant a system of social and natural classification in terms of which sets of people are named and otherwise associated with classes of phenomena in the natural world, especially with animal and plant species. Levi-Strauss revived this field in the middle of the twentieth century,[7] offering the structuralist framework needed to understand that the crux of totemism is not so much (as in the older works on the topic by Frazer, Van Gennep, Freud, etc.)[8] the individual association between one social group S and one natural class C, but the development of a productive relationship:

S1 : S2 : ...Si = C1 : C2 : ...Ci

In other words, totemism turned out to be an idiom to speak about the social world in terms of the natural world — the animal and vegetal world, and the relationships claimed to exist between the latter, providing the models of thought in terms of which everyday and ritual relations between groups could be articulated and manipulated.

                        Totemism, in which animal species feature overwhelmingly, thus appears as a particular form of a mode of thought that in the past has been called ‘irrational’, ‘pre-logical’, ‘peripheral’, ‘primitive’, and which Levi-Strauss’s work (especially in La pensee sauvage, 1964)[9] made us appreciate as obsessed with logic, rational, standard and common in all human societies past and present including everyday life and untutored thought and expression in the contemporary North Atlantic society — even though in the latter the influence of the institutionally and politically dominant forms of scientific thought filters through in the untamed everyday and ritual expressions, masking their wildness and creating embarrassment. In the light of the ubiquitous and ineradicable presence of ‘untamed thinking’ (for which Levi-Strauss coined the felicitous term ‘the science of the concrete’), contemporary scientific thought constitutes not the norm of human thought, but the exceptional case: one in which the conditions for the production of, and the assignment of truth to, verbal statements of a propositional form lies not only in their well-formedness and their referring to the natural world, but in the application of very elaborate, strict, intersubjective procedures stipulating the conditions under which such truth is assigned in an epistemologically valid and accountable manner.

                        We have now set the framework for the appreciation of animal symbolism as a very widespread form of untamed thinking, and indicated both its closeness and its distinction vis-a-vis contemporary scientific thought. Let us now return to Harding’s claim that it is the world-wide mediation of scientific knowledge which persuades us to attribute to such knowledge universality even regardless of whether science would be entitled to claim such universality on the basis of internal epistemological considerations. We have seen that there are mythemes (like the one of ‘hero fights monster’) that could claim practically world-wide mediation and representation. Let us now explore if the same applies to patterns of animal symbolism in non-scientific contexts world-wide.

3. The data set

In order to explore the extension in time and space, and the convergence, of various non-scientific systems of knowledge specifically those revolving on animal symbolism , in Table 2 I have brought together merely eleven series of animal symbolism, derived from widely differing locations (cf. Diagram 1) and periods. The series in this preliminary analysis have mainly been selected on the basis of their availability given the established context of my ongoing research in such fields as African and ancient history, Egyptology, African ethnography, and comparative religion and mythology (as part of a comprehensive historical and comparative analysis of African divination systems; of the applicability of the Black Athena thesis to sub-Saharan African; and of agency in precolonial African history).

                        The eleven series, while all hinging on the specific use of animal symbolism (often in combination with other conspicuous features of the natural world: celestial bodies, meteorological phenomena, the vegetal and mineral kingdoms, colours, and products of human creation) are highly diverse.

Series 1: Animal demons worldwide

The first series is that of animal demons, whose distribution in space and time largely coincides with that of the ‘hero fight monster’ mytheme as studied in Table 1, was identified as a by-product of Fontenrose’s exhaustive cross-cultural study of this mytheme: across the world’s mythologies, he was struck by the recurrence of a series of animal demons belonging to specific species.[10] Strictly speaking, it would be inappropriate to call this series’ distribution ‘world-wide’, since it is mainly attested for those parts of the world (largely coinciding with ancient literate civilisations) whose mythologies have been abundantly recorded and studied. Research currently initiated at the African Studies Centre, Leiden, seeks to bring to bear the African mythological material upon such world-wide comparisons.

Series 2: Egyptian nomes

The second series lists the names and symbolic associations of the 42 districts (‘nomes’, an ancient Greek rendering of the ancient Egyptian term sp3t) into which the Nile valley and delta were traditionally divided.[11] As is demonstrated by archaeological records notably the famous cosmetic palettes, the nomes’ nomenclature and symbolism go back to prehistoric times (which in Egypt ended with the establishment of the First Dynasty and the invention of writing, c. 3100 BCE), and its confusing complexity may be partly due to erosion in historic times when the underlying symbolic categories of pre-state local organisation were no longer properly understood.

Series 3: Egyptian gods

The third series lists the attributes, animal and otherwise, of the major ancient Egyptian gods of the historical period.[12]

Series 4: The Chokwe basket oracle

The fourth series lists the figurines as found in the basket oracle of the contemporary Chokwe people, dwelling in Angola, Zaire, and Zambia.[13] The oracle consists of a basket that contains dozens of man-made figurines carved out of wood, in addition to parts of animals, plants, and the mineral kingdom. During an oracular consultation, some of these items are caused to drop out of the basket, and the oracular response consists in an interpretation of the symbolic features of these items. The basket oracle, which is far from unique to the Chokwe people, is only one of a large family of African divination systems whose interpretation systems work along similar lines although the symbolic configurations to be interpreted are often generated in very different methods, by different random generators than a basket full of figurines. Important members of this family of African divinatory systems could be demonstrated[14] to be localising transformations of the Arabian divinatory system of Ailm al-raml, which was invented in Abbasid Mesopotamia by the end of the first millennium CE, on the basis of influences from Chinese Taoism (specifically I Ching), from astrology as formalised in Hellenistic and Imperial times on the basis of much older Mesopotamian and Egyptian divinatory astronomy, and possibly also independent influences from North and West African divination systems.

Series 5: Nkoya clans

The fifth series lists the nomenclature of clans (named human groups associated with a natural species or other natural phenomenon) among the contemporary Nkoya people of western central Zambia.[15] It was my sudden impression of a surprising parallelism between Nkoya clan nomenclature and Fontenrose’s world-wide list of animal demons which triggered the present analysis in the first place.

Series 6: Tswana clans:

The sixth series lists the very elaborate clan nomenclature among the Tswana people, a large ethnic and linguistic cluster in Botswana and South Africa.[16]

Series 7: The Chinese zodiac

The seventh series lists the nomenclature of constellations in the Chinese zodiac, which however contrary to most other zodiacs in the Old World represents not an annual cycle calibrating the Sun’s apparent progress along the ecliptic, but a twelve-yearly cycle.[17]

Series 8: Chinese lunar mansions

Our eighth series lists the rich nomenclature of Chinese lunar mansions. Throughout the Old World, ancient astronomies calibrate the Moon’s apparent progress along the ecliptic, on a (near-)monthly basis, by reference to a lunar zodiac, more commonly designated a series of lunar mansions or lunar houses, specifying for each day of the lunar month in which region of the sky (marked by a particular star or asterism) the Moon is to be found. The study of lunar mansions has formed a major topic in comparative historic astronomy ever since the early 19th century, when Colebrook, Biot, Weber and Burgess initiated the protracted scholarly discussion on the dependence or independent of the various Asian systems of lunar mansions vis-a-vis each other and vis-a-vis the ancient southwest Asian and Graeco-Roman astronomical and astrological tradition.[18] In the debate, the contradictions of European scholarship in the period of European colonial expansion came to the fore: on the one hand the scholarly perception of irreducible otherness of the Asian systems, and their fragmentation in terms of unconnected local systems, was in line with the underlying assumptions of European colonial domination, and its legitimation strategies; on the other hand, the contemplation of the sophistication of the Asian systems, and of their unmistakable similarity with the western astronomical and astrological tradition, conveyed a sense of respect and of Euro-Asian kinship (cf. the discovery of the Indo-European linguistic family a few decades earlier) in principle incompatible with colonial contempt. And the parallels were even stronger than scholar could realise at the time. Each of the 28 Chinese lunar mansions has both an animal association, and an association with a non-animal object. It is remarkable that also Hellenistic astrology (as documented in the Greek Magical Papyri there were to be discovered as from the end of the nineteenth century) designated the lunar mansions largely by animal names, of which there are 28 employed; the last seven of these also have non-animal names which would also eminently fit into our categories in Table 2: ‘Chimaera’, ‘Virgin’, ‘Lamp’, ‘Lightning’, ‘Wreath’, ‘Herald’s Stave’, ‘Boy’ and ‘Key’.[19] The ancient Greek system is in nomenclature rather similar to an Assyrian one from the seventh century BCE, with no more than 17 zoomorphic ‘lunar constellations’, that are in fact incipient mansions.[20] Time and again the idea crops up that the lunar mansions were profoundly associated with the invention of the alphabet — the number of letters and of mansions being virtually identical.[21] The idea is certainly attactive: by the time of the invention of the alphabet in the early second millennium BCE, the acrophonic principle of indicating a single phoneme by a symbol denoting a natural object whose name beings with that sound had already been available in Egyptian hieroglyphic writing for over a millennium,[22] and most of the earliest alphabetic signs unmistakable hark back to that source, but what was still needed was a fixed and culturally firmly supported framework of classification in the context of which one was led to rely on just over a score of different signs, instead of the hundreds that are needed for syllable writing, or the thousands for full hieroglyphic, cuneiform and Chinese writing. For many centuries, lunar mansions were in wide, daily use for calendrical purposes, and one may very well imagine how — against the background of the well attested constant influence of Egyptian hieroglyphic writing in Phoenicia / Syro-Palestine[23] — an initial system of merely designating these mansions by conventional signs ended up as an alphabet for the rendering of other words than the names of mansions, and finally for the rendering of all words.

Series 9: Ancient Babylonia’s oldest star catalogue

A famous list of ancient Babylonia’s 36 major stars records a very early phase in the standardisation of celestial description through asterisms and constellations, dating from before the fixing of the twelve zodiakal signs. Its symbolism, animal and otherwise, is listed as series (9).[24]

Series 10: Contemporary international scientific nomeclature of the constellation

The only series to derive from current cosmopolitan scientific practice is our series (10), which lists the current international nomenclature of the constellations.[25] It is the only one of our series which can claim universal, i.e. world-wide distribution.

Series 11: Ancient Greek gods

Finally, series (11) lists the symbolic associations, animal and otherwise, of the major ancient Greek gods.[26]

Diagram 1. Locations of eleven knowledge systems containing animal symbolism as included in the present analysis

(Click on the thumbnail to enlarge)


(Click here to open Table 2)

Table 2. Eleven systems of knowledge containing animal symbolism

4. Methodology

Without the slightest doubt, the present data set is extremely limited and one-sided in composition. One would like to see much more material included from Europe, South and Central Asia, Australia, Oceania, and the Americas, reflecting the animal symbolism of gods, asterisms, and such social groups as clans, but also extending to other fields of formalised culture. However, what little could be presented here, is already the fruit of months of strenuous collating and analysis, and could not readily be expanded without a further major research effort.

                        Meanwhile the most exciting research finding is already immediately clear from Table 2. For it turns out that systems of animal symbolism deriving from widely differing spatial and temporal contexts of cultural history may be conveniently collected in one large matrix, which brings out correspondences and formal continuities (not to speak of generic continuities) to a much greater extent than one would expect to be the case if all these local or regional systems of animal symbolism would have been invented totally independently from one another.

                        Collecting all these series into one comprehensive matrix requires a number of methodological choices.

                        In the first place, each series much be well-documented, but finite. One could spend the rest of one’s life investigating the ramifications of animal symbolism in Greek mythology, for instance, but for the kind of analysis encountered here, it is better to rely mainly — as I did — on one, comprehensive and authoritative source, and leave further details for a later phase in the analysis.

                        Another problem concerns aggregation. If one were to define a different category for each animal species and for each other kind of objects found in any of the series involved in our comparative analysis, the number of categories in the overall data set would be astronomical, defeating further analysis. Although the list of categories in Table 2 might look fairly exhaustive as a representation of the natural world,[27] the highly selective nature of natural species’ admittance to each of the local symbolic series can only be appreciated against an enumeration of the many taxonomic series (as distinguished by modern cosmopolitan biology) that objectively exist in each local natural environment.

                        While the species distinguished by today’s cosmopolitan biological science are more or less clear-cut and offer easy solutions for operationalisation (in other words, would make it easy to identify exemplars of the species in reality), we cannot assume that that kind of classification obtains or is meaningful in all the cultures featuring in our analysis. It may be advisable to subsume bats under birds, and marine mammals under fishes, because that it what many of the world’s cultures do. Antelopes are not universally distributed, and their northern complement would be the stag or deer, which may therefore be classified in the same category. The aggregate classification we end up with is a mere compromise. It will vary in the degree of specificity it observes with regard to certain types of animals. E.g. in Egyptian and Greek symbolism birds of prey are precisely distinguished and symbolically juxtaposed, so it would not do to lump vulture, eagle, hawk and falcon in one category, but in other cultures the taxonomic distinctions between these birds of prey may be less precise, or less precisely rendered in ethnographic or mythographic descriptions. Another reason for aggregation is that some kinds of animals (notably mammals and birds) are far more subjects of animal symbolism than others (gastropods, insects and other arthropods); thus in some cases a category used in Table 2 would amount to a taxonomic under-species in scientific biology, in other cases to an entire phylum. If we agree that animal classification is some form of inchoate science, it does not do to impose on any specific local systems the specific categorisation of another type of science notably that of cosmopolitan biology, but neither is it possible, in a comparative exercise, to do full justice to all the underlying local classifications. We would also tend to aggregate categories in the case that a specific category would be represented among only one or two series in our sample, unduly isolating it from all the other series. E.g., among the eleven series in our data set, the centipede is only specified in the series of the ancient Egyptian nomes, and in the series of the ancient Egyptian gods; it was found preferable to subsume the centipede under insects, although insects and centipedes constitute distinct sub-phyla within the phylum of Arthropods, to which nearly one million animal species belong, or about 75% of all animal species.[28]

                        An initially unforeseen feature of local systems of animal symbolism is the following: symbolism derived from the animal kingdom is often combined with symbolism based on the faunal and mineral kingdom, and on other aspects of the visible world, such as celestial and in general meteorological phenomena, abstract concepts, colours, etc. A major cluster of non-animal symbolism derives from man-made objects, which I have subsumed under one large heading ‘technology’. In some series the technological items are very numerous, even exceeding the faunal references in number. Since our emphasis is on animal symbolism here, I did not differentiate between the various ‘technological’ items.

                        The ensuing classification underlying (as the list of categories making up the extreme left-hand column) our comprehensive matrix in Table 2 is a mere compromise, and any results based on its analysis will have to be considered in a relative light: different classifications would be at least equally justified, and may have yielded different results. In order to allow a re-analysis in terms of slightly or entirely different categories and patterns of aggregation, I have always listed the original local category whenever I have listed a case under an aggregate category; e.g. when ‘goat’ as attribute of a specific god is listed under ‘ovines’— a category to which also sheep belong — , the name of that god appears in the box ‘ovines’ followed by ‘(goat)’ between parentheses.

                        Cluster analysis is the standard technique to bring out and underpin mathematically such clustering as one might intuitively perceive in the data listed in the above table. For this purpose, one assigns a numerical value to each cell, and ascertains whether, in the light of any of the usual mathematical linkage methods (average, centroid, complete, median, single or Ward’s),[29] certain series have more in common than others.

                        I assign to each cell the number of actual occurrences as listed in the data set; doubtful cases are counted for 0.5; species which feature symbolically elsewhere in the local society but not in the specific context as analysed (basket oracle, clans, etc.), will be treated as absent (0), since the same symbolic occurrences outside the specified context may also occur in all the other localities as analysed, without perspiring in the documentation. If an item matches more than one species or concept, it is listed twice and counted twice.

                        Since the data for at least one of our eleven series, that on animal demons per definition cannot include other aspects of the natural or man-made world than animals, one might decide to either

(a)  limit the analysis to those rows that actually concern animals, leaving humans, technologies, trees and plants, etc. out of the cluster analysis; or

(b)  extend the cluster analysis to all species and objects including non-animal ones.

                        Alternatively, considering the extremely selective way in which the series were constructed out of an enormous available literature, one may well doubt whether the number of recorded occurrences of a particular trait in each of the cells of Table 2 is a reliable and valid representation of the relative weight of this trait in the actual material, if it could be known and taken into account in its entirety. Therefore there is something to be said for a dichotomisation of the data, basing the analysis on the simple fact of whether a cell in the column is empty (= 0) or non-empty ( = 1), without taking into account the actual number of occurrences recorded per non-empty cell; such dichotomisation moreover has the advantage that a stronger, parametric distance metric may be used in cluster analysis: the Pearson correlation coefficient.

                        This yields four analytical approaches:

          animals and non-animal items, number of actual occurrences                                                                                                     (Analysis 1all categories, actual occurrences)

          animals and non-animal items, dichotomised (Analysis 1all categories, dichotomised)

          animals only, number of actual occurrences (Analysis 2animals only, actual occurrences)

          animals only, dichotomised (Analysis 2animals only, dichotomised)

Since the data in Table 2 are based on a very limited selection of the available literature, we can make no assumptions as to the underlying probability distribution, and therefore prefer a non-parametric distance metric: normalised percent disagreement; as variance method we prefer Ward’s, which in comparative assessments has often turned out to be both subtle and reliable.

5. Cluster analysis

Using all series and all categories as in Table 2, and taking the number of actually listed incidences per cell as the cell score to be entered into cluster analysis, the following cluster analysis is produced as Analysis (1):

Diagram 2. Cluster Analysis 1all categories, actual occurrences.

Distance metric is normalised percent disagreement; Ward minimum variance method


This cluster structure has a few features which recommend it as convincing and illuminating up to a point. The two Egyptian series (nomes and major gods), which are relatively close in space and time as compared to the other series in our data set, do cluster together; they also cluster together with the Greek mythological series, which recent research has emphasised to have much in common with the Egyptian material.[30] Series (2), (3) and (11) thus constitute cluster (A), which is opposed to cluster (B) which comprises all the other African and Asian series. Cluster (B) falls apart in two sub-clusters (C) and (D), both of which invite systematic interpretation. Sub-cluster (C) displays a certain spatial and temporal consistency in that it comprises two African societies with elaborate animal symbolism in their clan nomenclature and divination system respectively; however, one is surprised to see the Chinese system of lunar mansions to cluster with Tswana and Chokwe, while Nkoya (the third African society in our data set) and the Chinese zodiac appear as clustering in sub-cluster (D) of the same branch (B). Systems of lunar mansions are found in all Asian major civilisations, to begin with ancient Mesopotamia;[31] we may therefore postulate that the Chinese version of lunar mansions has a considerable antiquity.[32] We have no direct way of ascertaining the antiquity of African clan systems and divination systems, and therefore cannot gauge the time distance between the African and the Chinese material in cluster branch B; but whatever the time dimension, undeniably this material encompasses huge distances in space. That yet the African and the Chinese material clusters together, and in close association with the Babylonian material, suggests an unsuspected formal, and perhaps even generic, kinship to exist between these series. That a genuine, historical relationship is involved here is suggested by the fact that cluster (D) encompasses both the 36 Babylonian stars, and the modern constellations.

                        The considerable convergence in the delineation and even the naming of some major constellations across societies throughout the Old and the New World suggests a Palaeolithic origin, whose details are extremely difficult to reconstruct. Nonetheless, already in the early twentieth century the possibilities of an astronomical interpretation of Upper Palaeolithic signs, cupmarks, rock art was attempted, with more sophistication than recognition, in the work of the French prehistorian Baudouin.[33] In more recent decades, Marshack’s work on the possible interpretation of scratch patterns as found on Upper Palaeolithic mobile artefacts has revived this concern.[34] Meanwhile the professional astronomer Ovenden has suggested an astronomical method to solve the problem of the origin of the constellations, based on the following question: should we not simply ask at which place in the Ancient Near East and the eastern Mediterranean basin were the earliest attested constellations visible and during which period?[35] Background of this approach is that precession of the equinoxes causes many stars except the circumpolar ones to be alternatingly visible and invisible during certain periods of the c. 26.000 years out of which a full precession cycle consists; nedless to say, only visible constellations can be named and made into an astronomical system, and this brought Ovenden to situate the emergence of the constellations in the Early Bronze Age and the eastern Mediterranean basin — well in line with converging scholarly views about the increase of maritime contacts in that period, which (if they had to include open-sea crossings, e.g. from Crete directly to Egypt; which is far from certain for that period) had to involve sailing by night, and therefore navigation on the stars (contrary to the established Phoenician practice of day-light hopping from factory to factory across ditances of 25-30 km).

                        But Ovenden’s approach, though illuminating, does not take into account the virtually world-wide recognition of certain asterisms (the Pleiades, the Great Bear, Orion’s Belt), which if it is to be attributed to diffusion rather than to parallel cultural invention, would seem to imply a time scale for the earliest definition of these near-universal asterisms far more extensive than the few millennia which Ovenden’s approach would grant us. For certain constellations meanwhile the specific cultural origin (and in those cases far more recent than the Palaeolithic) has been authoritatively reconstructed by astronomically informed specialist in Ancient Near Eastern studies.[36] This does not mean that all constellations dat back to historical times: rather, a picture emerges according to which only a few constellations, heavy with animal symbolism, were discerned in the sky, leaving large stretches of the sky unnamed and unstructured, until the drive at scientific consistency and systematics, in the context of increasingly complex and state-based systems of knowledge, prediction, and control, finally caused the entire sky to be mapped and named, through still largely in tems of animal symbolism.

                        The first attestations of constellations in written and archaeological evidence derive from ancient Mesopotamia. There is a well-established intellectual continuity[37] between Babylonian astronomy (including the first attested constellations), subsequently Greek, Hellenistic, Roman, Indian and Arabic astronomy, and modern scientific astronomy; of the latter the nomenclature of the constellations still forms a modest part. Moreover, there is detailed evidence to suggest that Chinese astronomy owes a considerable debt to Babylonian astronomy.[38] Also the more or less world-wide (cf. Diagram 1) series of animal demons situates itself in this sub-cluster; Fontenrose’s formulation of this series of animal demons was based on a close inspection of the world’s recorded mythologies in the light of the ‘hero fights monster’ mytheme. Therefore sub-cluster (D) includes series between which genuine historical relationships exist, despite their mutual remoteness in both space and time. This suggests that also the appearance of the African material in culture (D) and (C) is not an artefact of blind statistical procedure, but equally reveals some genuine historical relationship, which we shall explore below.

                        For the methodological considerations given in section (4), we might be persuaded to base our cluster analysis not on the actual occurrences per cell, but on dichotomised data instead. For the entire data set of both animal and non-animal symbolism, this yields the following cluster Analysis 1dichotomised:

Diagram 3. Cluster Analysis 1all categories, dichotomised

Distance metric is normalised percent disagreement; Ward minimum variance method

legend: see Diagram 2.

Under dichotomisation, the clusters (A), (C) and (D) as identified in Analysis 1all categories, actual occurrences remain practically unaltered, although their linkage with one another is affected: (A) and (C) now cluster more closely together, instead of (C) and (D) as in the original Analysis (1).

                        Since we are working with dichotomised data, we are allowed to use the Pearson correlation coefficient as our distance metric. This yields the following Analysis (1)dichotomised, Pearson:

Diagram 4. Cluster Analysis 1all categories, dichotomised, Pearson

Distance metric is 1-Pearson correlation coefficient; Ward minimum variance method

legend: see Diagram 2.

Use of the Pearson correlation coefficient does not substantially affect the cluster analysis: the clusters remain intact but their interlinkage again shifts (now it is (A) and (C) which cluster more closely together than (D)); moreover the two Chinese series now cluster together within cluster C, which is somewhat more convincing by analogy with the two Egyptian series).

                        If, in the light of the methodological considerations in section 4, the analysis is limited to animals only, a clustering pattern emerges as Analysis 2animals only, actual occurrences in section (3):

Diagram 5. Cluster Analysis 2animals only, actual occurrences

Distance metric is normalised percent disagreement; Ward minimum variance method

legend: see Diagram 2.

Analysis (2) yields results very similar to those we considered above under Analysis (1). In Analysis (2), the basic clusters (A), (C) and (D) re-appear, with only two modifications: the two Chinese series now cluster together (as was to be expected, by analogy with the two Egyptian series); and whereas in the first analysis (C) and (D) clustered together to form (B), now (C) and (A) cluster together instead of (D).

                        Also for Analysis (2)animals only, we may proceed to a dichotomised approach.

Diagram 6. Cluster Analysis 2animals only, dichotomised

Distance metric is normalised percent disagreement; Ward minimum variance method

legend: see Diagram 2.

While the basic clustering pattern of the three clusters (A), (C) and (D) is maintained under dichotomisation, remarkable shifts occur: the modern constellations series leaves the proximity of the Babylonian series and instead joins cluster (A) (not without historical basis, for also Egyptian astronomy contributed to the definition of contemporary constellation, while many of their names refer to episodes in Greek mythology); and the Chokwe divination basket series (originally in (C)) comes to straddle the boundary between (A) and (C).

                        Again, the Pearson correlation is admissible as the distance metric for dichotomised data, yielding the following cluster Analysis (2)dichotomised, Pearson:

Diagram 7. Cluster Analysis 2animals only, dichotomised, Pearson

Distance metric is 1-Pearson correlation coefficient; Ward minimum variance method

legend: see Diagram 2.

With use of the Pearson correlation coefficient appears the by now well-known cluster structure (A), (C) and (D) re-appears, albeit somewhat blurred in that the Greek mythology series dissociates itself from (A) and joins (D). Perhaps this is an artefact of the Pearson approach, where the Greek series as by far the most elaborately documented one may behave somewhat oddly. But I prefer to see here a systematic reason: Fontenrose’s delineation of the ‘animal demons’ series was largely based on his analysis of Greek mythological patterns, subsequently enriched with an extensive cross-cultural comparison involving the mythologies of Rome, Egypt, Canaan, Syria, Anatolia, Mesopotamia, India, China, Japan, and the Americas. The Babylonian central mytheme of the famous Enuma Elish cuneiform series[39] (‘Marduk fights Tiamat’) in part inspired both the ‘animal demons’ series and the Babylonian series of major stars and asterisms, with their symbolic associations. As stated above, the compilation of Table 2, and the present analysis in general, was instigated, in the first place, by my intuitive observation of the apparently close parallels between the species after which Nkoya clans are named, and Fontenrose’s ‘animal demons’ series; this observation is borne out by all the above cluster analyses, including Analysis (2)dichotomised, Pearson.

                        Considering the high degree of aggregation in the non-animal categories (especially in ‘humans and gods’, ‘arthropods’, ‘sky etc.’, ‘technology’, and ‘trees and plants’), as compared to the far greater degree of precision in the delineation of our animal categories, it is Analysis (2) which should guide us when seeking to formulate conclusions about patterns of animal symbolism in our data set. But once again, the difference between Analysis (1) and (2) is slight.

                        Cluster analysis is a blind statistical procedure. It is often contentious,[40] since, depending on the choice of distance metrics very different results may be produced that are highly manipulable and full of mere procedural artefacts, even if we use a method that was found to be relatively reliable, such as Ward’s. Our analysis however shows signs of considerable consistency, both formally (the same triple cluster structure (A), (B) and (C) appearing time and again, no matter how we vary the composition of the data set (with or without non-animal elements, and using either actual number of incidence or dichotomised data), and empirically (the Egyptian pair of series, the Chinese pair of series, and two of the three African societies, clustering each consistently together). Therefore is would be foolish to dismiss the outcomes of our cluster analysis as mere accidental or as a mere artefact of cluster analysis. Instead we have to look for a convincing explanation of what we may take to be a genuine, empirically well etablished pattern.

[1]              I am indebted to the African Studies Centre, Leiden, and to the Netherlands Institute for Advanced Studies in the Social Sciences and Humanities, Wassenaar, for greatly facilitating the research on which the present paper is based; and to Henk Visser and the Netherlands Association for the Philosophy of Science, for creating the context in which I was brought to write the present paper, as a by-product of my work on the more specifically philosophy-of-science argument: van Binsbergen, W.M.J., 2001, ‘Noordatlantische wetenschap als etno-wetenschap: Een intercultureel-filosofische reflectie op Sandra Harding’, paper read at the seminar on ‘Kennis en Cultuur’ (Knowledge and culture), Annual Meeting, Netherlands Association for the Philosophy of Science, Utrecht, 23 November, 2001; English version in preparation; soon available at: .

[2]              Harding, S., 1991, Whose science? Whose knowledge? : Thinking from women’s lives, Ithaca, NY: Cornell University Press; Harding, S., 1992, ‘After the neutrality ideal: science, politics and ‘‘strong objectivity’’ ‘, Social Research, 59: 567-587; Harding, S., 1993, ed., The ‘racial’ economy of science: Toward a democratic future, Bloomington: Indiana University Press; Harding, S., 1994, ‘Is science multicultural? Challenges, opportunities, uncertainties’, Configurations, 2, 2, reprinted in: Goldberg, D.T., 1994, ed. , Multiculturalism: A Reader, London: Blackwell; Harding, S., 1997, ‘Is modern science an ethnoscience? Rethinking epistemological assumptions’, in: Eze, E.C., ed., Postcolonial African philosophy: A critical reader, Oxford: Blackwell, pp. 45-70. My argument below is largely based on the latter article but against the background of Harding’s other publication as cited.

[3]              Philosophers and historians of ideas often differentiate between ‘strong’ and ‘weak’ versions of a particular theoretical position, such as relativism, empiricism, falsificationalism, materialism etc. The stronger version consistently takes the theory to its ultimate consequences, often at variance with conventional wisdom; the weaker position is less extreme and consistent, and humours conventional wisdom to some extent. Cf. Bloor, D., 1993, ‘Strong programme’, in: Dancy, J., & Sosa, E., eds., A companion to epistemology, Oxford/ Cambridge (Mass.): Blackwell’s, first published 1992, where the strong variant of social constructivism is described as holding the view that also true knowledge is nevertheless socially determined; incidentally, such a position also underlies Harding’s as discussed here: she is looking for (a) social determinants of science’s claims to objectivity, rationality and universality, but does so without (b) excluding the possibility that there could also be internal epistamological grounds for such a claim. Goldman, A.I., 1988, ‘Strong and weak justification’, Philosophical Perspectives, 2: 51-71, who describes strong and weak versions of subjectivism; and Harding, S., ‘After the neutrality ideal‘, o.c., where varieties of objectivity as a scientific ideal are explored.

[4]              Cf. Griffiths, J.G., 1980, ‘Interpretatio graeca’, in: Helck, W., & Otto, E., eds., Lexikon der Agyptologie, Wiesbaden, Harrassowitz, vol. III, cols. 167-172.

[5]              Cf. Levi-Strauss, C., 1960, ‘Four Winnebago myths: A structural sketch.’ In: Diamond, S., ed., Culture and history, New York: Columbia University Press, pp. 351-362; Levi-Strauss, C., 1968, ‘The story of Asdiwal’, in: Leach, E.R., ed., The structural study of myth and totemism, London, 2nd impr., pp. 1-47, first published 1967; Levi-Strauss, C., 1969-78, Introduction to a science of mythology, 4 vols., trans. by J. Weightman & D. Weightman, Harmondsworth: Penguin / Chicago: Chicago University Press, original French edition: Mythologiques I: Le cru et le cuit, 1964; II: Du miel aux cendres, 1966; III: Origines des manieres de table, 1968; IV: L’homme nu, 1971, all at Paris: Plon; Levi-Strauss, C., 1971, ‘Rapports de symetrie entre rites et mythes de peuples voisins’, in: Beidelman, T.O., ed., The translation of culture, London: Tavistock, pp. 161-177; Levi-Strauss, C., 1973, ‘La structure des mythes’, in: Levi-Strauss, C., Anthropologie structurale I, Paris: Plon, 1973, pp. 227-255; Levi-Strauss, C., 1979, Myth and meaning, London: Routledge & Kegan Paul; Leach, E.R., 1967, ed., The structural study of myth and totemism, London: Tavistock; Leach, E., & Aycock, D.A., 1983, Structuralist interpretations of biblical myth, Cambridge: Cambridge University Press; van Binsbergen, W.M.J., 1985, ‘The historical interpretation of myth in the context of popular Islam’ in: van Binsbergen, W.M.J., & Schoffeleers, J.M., eds., Theoretical explorations in African religion, London/ Boston: Kegan Paul, pp. 189-224; also at .

[6]              Within the research group on Agency in Africa of the African Studies Centre, Leiden, established 2002, I am now initiating research intended to cover this relative blind-spot. The present paper could be counted as the first product of that research. The intercontinental continuity of myths including African myths also plays an important role in my forthcoming bookGlobal Bee Flight, o.c.

[7]              Levi-Strauss, C., 1962, Le totemisme aujourd’hui, Paris: Presses Universitaires de France.

[8]                Ankermann, B., 1915, ‘Verbreitung und Formen des Totemismus in Afrika’, Zeitschrift fur Ethnologie; Armstrong, W.E., 1961, ‘Totemism’, in: Ashmore, H.S., ed., Encyclopaedia Brittanica: A new survey of universal knowledge, Chicago / London / Toronto: Encyclopaedia Brittanica, XXII: 317-320 (in fact an antiquated account reflecting scholarly views in the early twentieth century); Durkheim, E., 1912, Les formes elementaires de la vie religieuse: Le systeme totemique en Australie, Paris: Felix Alcan.; Frazer, J.C., 1887, Totemism, Edinburgh: Adams & Charles; Freud, S., 1918, Totem and taboo, New York: Random House, English translation of German edition, Totem und Tabu, first published 1913; Hartland, E.S., 1915, ‘Totemism’, in: Hastings, J., with Selbie, J.A., & Gray, L.H., eds., Encyclopedia of Religion and Ethics, Edinburgh: Clark / New York: Scribner, XII: 393-407; Lubbock, J., 1870, The origin of civilization and the primitive condition of man: Mental and social condition of savages, London: Longmans, Green; Mclennan, J.F., 1865, Primitive Marriage, Edinburgh: Adam & Charles Black.; van Gennep, A., 1904, Tabou et totemisme a Madagascar: Etude descriptive et theorique, Paris: Leroux. For a recent re-consideration of the issue of totemism, cf. Comaroff, J., & Comaroff, J.L., 1992, ‘Totemism and ethnicity’, in: Comaroff, J., & Comaroff, J.L., eds., Ethnography and the historical imagination, Boulder, CO: Westview Press.

[9]              Levi-Strauss, C., 1962, La pensee sauvage, Paris: Plon; Engl. translation The savage mind, 1973, Chicago: University of Chicago Press/ London: Weidenfeld & Nicholson, first published 1966.

[10]                Fontenrose, J., 1980, Python: A study of Delphic myth and its origins, Berkeley etc.: University of California Press; paperback edition, reprint of the 1959 first edition.

                                The footnotes to the filled cells in this column specify page, and motif (numerical code preceded by letter) where this type of animal demon is discussed in Fontenrose’s book Python, o.c.

[11]            Bernal has persuaded us to recognise in the ancient Greek toponym of Sparta; Bernal, M., in press, ‘Review of ‘‘Word games: The linguistic evidence in Black Athena’, Jay H. Jasanoff & Alan Nussbaum’, typescript in my possession, now published in: Black Athena Writes Back, Durham: Duke University Press.

[12]            Also based on Vergote, J., 1974, De Egyptenaren en hun godsdienst, Bussum: De Haan, second impr., first ed. 1971; Gardiner, A.H., 1994, Egyptian grammar: Being an introduction to the study of hieroglyphs, rev. 3rd ed., Oxford: Griffith Institute/ Ashmolean Museum, this edition first published 1957, first edition published 1927; Bonnet, H., 1971, Reallexikon der agyptischen Religionsgeschichte, Berlin: de Gruyter, first published 1952.

[13]                Rodrigues de Areia, M.L., 1985, Les symboles divinatoires: Analyse socio-culturelle d’une technique de divination des Cokwe de l’Angola ( ngombo ya cisuka), Coimbra: Universidade de Coimbra.

[14]            Skinner, S., 1986, The oracle of geomancy: Divination by earth, Bridport (Dorset)/ San Leandro (Cal.): Prism, first published 1977; Trautmann, R., 1939-1940, ‘La divination a la Cote des Esclaves et a [ check: a la ] Madagascar: Le Vodou Fa — le Sikidy’, Memoires de l’Institut Francais d’Afrique Noire, 1, Paris: Larose; Maupoil, B., 1943, La geomancie a l’ancienne Cote des Esclaves, Paris: Institut de l’Ethnologie; Maupoil, B., 1943, ‘Contribution a l’origine musulmane de la geomancie dans le Bas-Dahomey’, Journal de la Societe des Africanistes, 13. ; van Binsbergen, W.M.J., 1995, ‘Four-tablet divination as trans-regional medical technology in Southern Africa’, Journal of Religion in Africa, 25, 2: 114-14, also at ; van Binsbergen, W.M.J., 1996, ‘Transregional and historical connections of four-tablet divination in Southern Africa’, Journal of Religion in Africa, 26, 1: 2-29, also at ; van Binsbergen, W.M.J., 1996, ‘The astrological origin of Islamic geomancy’, paper read at ‘The SSIPS [ Society for the Study of Islamic Philosophy and Science ] / SAGP [ Society of Ancient Greek Philosophy ] 15th Annual Conference: ‘‘Global and Multicultural Dimensions of Ancient and Medieval Philosophy and Social Thought: Africana, Christian, Greek, Islamic, Jewish, Indigenous and Asian Traditions, Binghamton University’’, Department of Philosophy/ Center for Medieval and Renaissance studies (CEMERS).

[15]            van Binsbergen, W.M.J., 1992, Tears of Rain: Ethnicity and history in central western Zambia, London/ Boston: Kegan Paul International; and author’s fieldnotes. An extensive discussion on Nkoya clans is forthcoming in my Global Bee Flight, o.c.

[16]                Schapera, I., 1952, The ethnic composition of Tswana tribes, London: London School of Economics and Political Science, Monographs on Social Anthropology no. 11. In addition to those listed, Schapera mentions two totems whose meaning he cannot explain: mokowe and mphareng; these words, or the roots from which they might be derived, are not listed in the standard Tswana dictionary either: Matumo, Z.I., 1993, Setswana English Setswana dictionary, Macmillan/ Boleswa/ Botswana Book Centre, revised version of the 1875 edition of Tom Brown’s Setswana dictionary. I suggest mokowe relates to the colour white.

[17]            Walters, D., 1989, Chinese astrologie: Het interpreteren van de openbaringen van de boodschappers des hemels, Katwijk aan Zee: Servire, p. 77; Dutch translation of D. Walters, 1987, Chinese astrology, Wellingborough: The Aquarian Press.

[18]            Cf. Colebrooke, H.T, 1807, ‘On the Indian and Arabian divisions of the zodiac’, AR, 9: 323-376; printed in: Colebrooke, H.T., 1837, Miscellaneous essays, vol. 2, pp. 321-373; Biot, J.B., 1840, ‘Sur les nacshatras des Hindous: Et les mansions lunaires des Arabes’, Journal des savants, 1840: 264-279; Weber, A., 1850-1853, ‘Ueber den Taittitiya-Veda, astronomische Data aus beiden Yajus und eine Stelle des Taittiriya-Brahmana uber die Mondhauser’, Indische Studien, 1: 68-100, 2: 390-392; Weber, A., 1865, ‘Zur Frage uber die Nakshatra’, Indische Studien, 9: 424-459. Burgess, E., 1866, ‘On the origin of the lunar division of the zodiac’, Journal of the American Oriental Society, 8: 309-334. For the Chinese lunar mansions (sieou) listed here, cf. Walters, o.c. Other sources on the sieou include: Schlegel, G., 1875, Uranographie chinoise, 3 vols, Leiden: Brill; Whitney, W.D., 1874, ‘On the lunar zodiac of India, Arabia and China’, Oriental and Linguistic Studies, 2nd series, article 13: 340-421; Boll, C., 1912, ‘Der ostasiatische Tierzyklus im Hellenismus: Vortrag gehalten am 9 April 1912 auf dem XVI. Internationalen Orientalisten-Kongress zu Athen’, T’oung Pao, 13: 699-718; Hentze, C., 1933, Mythes et symboles lunaires (Chine ancienne, civilisations anciennes de l’Asie, peuple limitrophes du Pacifique), Antwerpen: De Sikkel; Petri, W., 1966, ‘Uighur and Tibetan lists of the Indian lunar mansions’, Indian Journal of the History of Science, 1: 83-90; Mostaert, A., ‘Introduction’, in: Cleaves, F.W., ed., 1969, Manual of Mongolian astrology: With a critical introduction by The Rev. A. Mostaert CICM, Schilde, Belgium, Cambridge (Mass.): Harvard University Press, pp.1-65; Oldenburg, H., 1909, ‘Naksatra and Sieou’, Nachrichten von der Koniglichen Gesellschaft der Wissenschaften zu Gottingen, 1909, pp. 544-572, reprinted in: Oldenburg, H., 1967, Kleine Schriften, ed. Janert, K.L., Wiesbaden: Glasenapp-Stiftung, vol. 2, pp. 1352-1380.

[19]            Cf. Gundel, W., 1936, Dekane und Dekansternbilder: Ein Beitrag zur Geschichte der Sternbilder der Kulturvolker: Mit einer Untersuchung uber die Agyptischen Sternbilder und Gottheiter der Dekane von S. Schott, Studien der Bibliothek Wartburg, Bd 19, reprint 1969, Darmstadt: Wissenschaftliche Buchgesellschaft, p. 223. Inclusion, in Table 2, of these Greek Magical Papyri and early Babylonian series in Table 2 should be considered in the course of further analysis along the lines developed in the present paper.

[20]            Cf. Parpola, S., 1983, Letters from Assyrian Scholars to the Kings Esarhaddon and Assurbanipal, Part II, Commentary and appendices, Alter Orient und Altes Testament; Veroffentlichungen zur Kultur und Geschichte des Alten Orient und des Alten Testaments, Band 5/2, Neukirchen-Vluyn: Neukirchener Verlag/ Kevelaer: Butzon & Bercker, appendix B, ‘Lunar constellations’, pp. 385-38.

[21]            Cf. Hommel, F., 1891, ‘Uber den Ursprung und das Alter der arabischen Sternnamen und insbesordere der Mondstationen’, Zeitschrift des deutsches morgenlandischen Gesellschaft, 45: 592-619. Stucken, E., 1913, Der Ursprung des Alphabets und die Mondstationen, Leipzig — which despite the wild suggestion contained in its title is a very thorough and authoritative study; Kelley, D.B., 1992, ‘The twenty-eight lunar mansions of China: Part 2: A Possible Relationship with Semitic Alphabets’, Reports of Liberal Arts, Hamamatsu University School of Medicine, No. 6, 1992; Burckhardt, T., 1974, Cle spirituelle de l’astrologie musulmane d’apres Mohyiddin ibn Arabi, Milano: Arche, Bibliotheque de l’Unicorne.

[22]            Gardiner, A.H., 1942, ‘Writing and Literature’, in Glanville, S.R.A., ed., The legacy of Egypt, Oxford: Clarendon Press, pp. 53-78; Cerny, J., 1971, ‘Language and writing’, in: Harris, J.R., ed., The legacy of Egypt, 2nd ed., Oxford: Clarendon, pp. 197-219.

[23]                Dussaud, R., 1946-1948, ‘L’origine de l’alphabet et son evolution premiere d’apres les decouvertes de Byblos’, Syria, 25: 36-52; Redford, D.B, 1992, Egypt, Canaan, and Israel in ancient times, Princeton: Princeton University Press; and extensive references cited there.

[24]            Walker, C.B.F., & Hunger, H., 1977, ‘Zwolfmaldrei’, Mitteilungen der Deutschen Orient-Gesellschaft (Berlin), 109: 27-34; cf. van der Waerden, B.L., 1949, ‘Babylonian astronomy: II. The thirty-six stars’, Journal of Near Eastern Studies, 8: 6-26. The relatively early texts to which these scholarly articles refer, by no means offer a full account of the complete series of constellations as recognised in ancient Mesopotamia; cf. Weidner, E.F., 1924, ‘Ein babylonisches Kompendium der Himmelskunde’, American Journal of Semitic Languages and Literatures, 40: 186-206; Hunger, H., & Pingree, D., 1989, MUL.APIN: An astronomical Compendium in Cuneiform, Horn (Austria): Verlag Ferd. Berger & S. Gesellschaft; Pingree, D., & Walker, C., 1988, ‘A Babylonian star catalogue: BM 78161’, in: Leichty, E., Ellis, M deJ., & Gerardi, P., eds., A scientific humanist: Studies in memory of Abraham Sachs, Philadelphia: Occasional Publications of the Samuel Noah Kramer Fund, 9, pp. 313-322. For the general background to science in the Ancient Near East, cf. Neugebauer, O., 1969, The exact sciences in Antiquity, New York: Dover, 2nd edition, first published 1957, Providence (R.I.): Brown University Press. For the magical, especially divinatory use of astronomy in the ancient Babylonian context, cf. Reiner, E., 1995, Astral magic in Babylonia, Transactions of the American Philosophical Society, 85, 4 , Philadelphia: American Philosophical Society; van Binsbergen, W.M.J., & Wiggermann, F.A.M., 2000, ‘Magic in history: A theoretical perspective, and its application to Ancient Mesopotamia’, in: Abusch, T., & van der Toorn, K., eds., Magic in the Ancient Near East, Groningen: Styx, pp. 3-34, also at: . 

[25]            Moore, P., 1984, The new atlas of the universe, London: Beazley, p. 203.

[26]            The number of sources available for Greek mythology and its animal associations is overwhelming. In order to keep the size of this series within manageable limits, I have greatly limited myself, drawing the data mainly from Robert Graves’ extensive and authoritative collection: Graves, R., 1964, The Greek myths, 2 vols., Harmondworth: Penguin, first published 1955. This book is very elaborate on the epithets of major gods, and translates every proper name, although the etymologies given are often admittedly doubtful. Additional data were drawn mainly from: The New Larousse Encyclopedia of mythology, introduction R. Graves, London/ New York/ Sidney/ Toronto: Hamlyn, 11th edition, especially the contribution there by Guirand: Guirand, F., 1975, ‘Greek mythology’, in: New Larousse Encyclopedia of mythology, o.c., pp. 85-198; Long, C.H., 1993, ‘Mythology’, in: The New Grolier Multimedia Encyclopedia, Release 6, 1993; and Criss, P.J., n.d. [ 2000 ] , ‘Animals as represented in mythology and folklore,``grandpa/animals.html .

[27]            For a fairly exhaustive enumeration of current African mammal species, cf. Haltenorth, T., & Diller, H., 1988, A field guide to the mammals of Africa, London: Collins, this edition first published 1980, Engl. translation of Saugetiere Afrikas und Madagaskars, Munchen: BLV, 1977, with extensive references. It is remarkable that only a limited selection of the hundreds of species listed there found their way into the systems of animal symbolism as treated in the present argument. The same applies, a fortiori, to the birds and other phyla of the animal kingdom, whose extremely rich ramifications may be gleaned from any standard encyclopeadia.

[28]                Cloudsley-Thompson, J.L., 1993, ‘arthropod’, in: Grolier Encyclopedia, o.c.; Ewing, H.E., 1961, ‘Anthropoda’, in: Encyclopaedia Brittanica, o.c., II: 456-459.

[29]            For a discussion of some of these methods and their merits, cf. Anderberg, M.R., 1973, Cluster analysis for applications, New York: Academic Press; Everitt, B., 1974, Cluster analysis, London etc: Heinemann, pp. 69ff.

[30]            Bernal, M., 1987, Black Athena: The Afroasiatic roots of classical civilization, Vol. I, The fabrication of Ancient Greece 1787-1987, London: Free Association Books/ New Brunswick: Rutgers University Press; Bernal, M., 1991, Black Athena: The Afro-asiatic roots of classical civilization, II, The archaeological and documentary evidence, London: Free Association Books; New Brunswick, N.J.: Rutgers University Press; Davison, J.M., 1987, ‘Egyptian influence on the Greek Legend of Io’, paper given to the Society for Biblical Literature; Berard, J., 1952, ‘Les Hyksos et la legende d’Io: Recherches sur la periode pre-mycenienne’, Syria, 29: 1-43; Lambropoulou, A., 1988, ‘Erechtheus, Boutes, Itys and Xouthos: Notes on Egyptian presence in early Athens’, The Ancient World, 18: 77-86. I earlier objected in print to the idea of close continuities between Egyptian and Greek myths (van Binsbergen, W.M.J., 1997, ‘Alternative models of intercontinental interaction towards the earliest Cretan script’, in: van Binsbergen, W.M.J., ed., Black Athena: Ten Years After, Hoofddorp: Dutch Archaeological and Historical Society, special issue, Talanta: Proceedings of the Dutch Archaeological and Historical Society, vols. 28-29, 1996-97, pp. 131-148; also at: ) but have meanwhile accepted this idea wholeheartedly (cf. my forthcoming Global bee flight, o.c.; and the greatly expanded and revised reprint of Black Athena Ten Years After, under the title: Black Athena Alive, Hamburg/ Munster: LIT and New York: Transaction Press, in press.

[31]            Cf. Gundel, Dekane, o.c.; Parpola, Letters from Assyrian scholars, o.c.

[32]                However, this assumption may have to be revised in the light of suggestions of Western nineteenth-century CE borrowings into East Asian astrology; Cf. Gundel, Dekane, o.c., p. 216; I come back to this in a footnote below.

[33]                Baudouin, M., 1916, ‘La prehistoire des etoiles au Paleolithique: Les Pleiades a l’epoque aurignacienne et le culte stello-solaire typique au solutreen’, Bulletin et Memoires de la Societe d’Anthropologie de Paris, ser. 6, 7: 274-317; Baudouin, M., 1926, La prehistoire par les etoiles: Un chronometre prehistorique, Paris: Maloine.

[34]                Marshack, A., 1972, The roots of civilization: The cognitive beginnings of man’s first art, symbol and notation, London: Weidenfeld & Nicholson/New York: McGraw-Hill. A reconstruction of the earliest astral science will be attempted in: van Binsbergen, W.M.J., in preparation, Cupmarks, stellar maps, and mankala board-games: An archaeoastronomical and Africanist excursion into Palaeolithic world-views (for a preview, see: .

[35]            Cf. Ovenden, M.W., 1966, ‘The origins of the constellations’, The Philosophical Journal [ Transactions of the Royal Philosophical Society of Glasgow ] , 3: 1-18.

[36]            Cf. Porada, E., 1987, ‘On the origins of ‘‘Aquarius’’ ‘, in: Rochberg-Halton, F., ed., Language, literature and history: Philological and historical studies presented to Erica Reiner, New Haven (Conn.): American Oriental Society, pp. 279-291; Hartner, W., 1965, ‘The earliest history of the constellations in the Near East and the motif of the lion-bull combat’, Journal of Near Eastern Studies, 24: 1-16; van der Waerden, B.L., 1952-53, ‘History of the zodiac’, Archiv fur Orientforschung, 16: 216-230; Lewy, H., 1965, ‘I?tar-„ad and the Bow Star’, in: Guterbock, H.G., & Jacobsen, T., eds., Studies in honour of Benno Landsberger on his seventy-fifth birthday, April 21, 1965, Chicago: University of Chicago Press for Oriental Institute of the University of Chicago, pp. 273-282; Miller, R.A, 1988, ‘Pleiades perceived: MUL.MUL to Subaru’, Journal of the American Oriental Society, 108: 1-25; Borger, R., 1972-1975, ‘Himmelsstier’, in: Edzard, D.O., ed., Reallexikon der Assyriologie und Vorderasiatische Archaologie, Berlin/ New York: de Gruyter, 4. Band, p. 413-414; and (non vidi) Gleadow, R., 1968, The origin of the zodiac, London. A wealth of information on ancient Mesopotamiam astronomy also to be found in: Parpola, S., 1983, Letters from Assyrian Scholars, o.c.

[37]            Boll, F.J., 1903, Sphaera, Leipzig: Teubner; Bezold, C., & Boll, F.J., 1911, Reflexe astrologischer Keilinschriften bei griechischen Schiftstellern, Heidelberg, Akademie der Wissenschaften Philosophisch-historische Klasse, no. 7: 1-54; Barton, T., 1994, Ancient astrology, London: Routledge; Tester, S.J., 1989, A history of western astrology, New York: Ballantine, reprint of the 1987 first edition; Pingree, D., 1973, ‘Astrology’, in: Wiener, P.P., ed., Dictionary of the history of ideas: Studies of selected pivotal ideas, I, New York: Scribner, pp. 118-126; Baigent, M., 1994, From the omens of Babylon: Astrology and Ancient Mesopotamia, Harmondsworth: Arkana/ Penguin Books; Berthelot, R., 1938, La pensee de l’Asie et l’astrobiologie, Paris: Payot; Nilsson, M.P., 1943, The rise of astrology in the Hellenistic age, Meddelande fran Lunds Astronomiska Observatorium, Ser. ii, nr. iii, Historical notes and papers, no. 18. In recent decades, the fundamental continuity underlying astronomy and astrology in major civilisation of Antiquity, the Ancient Near East, South Asia, the Arab world, and pre-modern Europe, has been studied with greatly impressive scholarship by David Pingree: Kennedy, E.S, & D. Pingree, 1971, The astrological history of M?sh?®all?h, Cambridge (Mass.): Harvard University Press; Pingree, D., 1959, ‘A Greek linear planetary text in India’, Journal of the American Oriental Society, 79: 282-284. Pingree, D., 1963-64, ‘Indian influence on early Sassanian and Arabic astronomy’, Journal of Oriental Research (Madras), 33: 1-8; Pingree, D., 1971, ‘On the Greek origin of the Indian planetary model employing a double epicycle’, Journal of the History of Astronomy, 2: 80-85; Pingree, D., 1973, ‘Greek influence on Islamic astronomy’, Journal of the American Oriental Society, 93, 1: 32-43. Pingree, D., 1973, ‘The Mesopotamian origin of early Indian mathematical astronomy’, Journal of the American Oriental Society, 93: 32-43. Pingree, D., 1978, The Yavanajataka of Sphujidhvaja, Harvard Oriental Series 48, 2 vols, Cambridge (Mass.)/ London: Harvard University Press (which contains, among much else of great value, a complete cross-cultural history of astrology); Pingree, D., 1979, ‘Ilm al-hay’a’, in: Lewis, B., Menage, V.L., Pellat, C., Schacht, J., eds., Encyclopaedia of Islam, new edition, Leiden: Brill, pp. III, 1135-1138; Pingree, D., 1989, ‘MUL.APIN and Vedic astronomy’, in: Behrens, H., Loding, D., & Roth, M.T., eds., DUMU-E2-DUB-BA-A: Studies in honor of Ake W. Sjoberg, Philadelphia: S.N. Kramer Fund, pp. 439-445.

[38]            Bezold, C., 1919, ‘Sze Ma Ts’ien und die babylonische Astrologie’, Ostasiatische Zeitschrift, 8: 42-49; Ungnad, A., 1932-, ‘China und Babylonien’, in: Ebeling, E., & Meissner, B., eds., 1932-, Reallexikon der Assyriologie, Berlin: de Gruyter, II, 91-93. On the other hand, Kugler advanced an astronomical detail (reference to the longest day lasting 14 hours 24 minutes, as in Honan, China, 35° N, but not as in Babylon at 32° 30’ N) which might suggest an influence from East Asia upon ancient Babylonia; cf. Kugler, F.X., 1900, Die Babylonische Mondrechnung, Fribourg/ Brisgau, pp. 79f. 

[39]                Pritchard, A.B., ed. 1969, Ancient Near Eastern texts relating to the Old Testament, Princeton: Princeton University Press, first published 1950.

[40]            Cf. the controversy over the cluster analysis of the world-wide variation in mitochondrial DNA. It was on the basis of this cluster analysis that the ‘African Eve’ hypothesis was first formulated; Cann, R.L., Stoneking, M., & WilsonA.C., 1987. ‘Mitochondrial DNA and human evolution’, Nature, 325: 31-36; Templeton, A.R., 1997, ‘Testing the out-of-Africa replacement hypothesis with mitochondrial DNA data’, in: Clark, G.A., & Willermet, C., eds., Conceptual issues in modern human origins research, Amsterdam: Aldine de Gruyter, pp. 329-360; Shreeve, The Neandertal enigma?, o.c.; Laine, A., 2000, ‘Eve africaine ? De l’origine des races au racisme de l’origine’, in: Fauvelle-Aymar, F.-X., Chretien, J.-P., & Perrot, C.-H., Afrocentrismes: L’histoire des Africains entre Egypte et Amerique, Paris: Karthala, pp. 103f .

homepage | Animal symbolism overview page | Part II

page last modified: 25-03-02 18:07:20