Arjuna–Odysseus

Shared Heritage in Indian and Greek Epic

N. J. Allen

• Publisher: Routledge India
• Year: 2019
• Print ISBN: 9780367260477
• eBook ISBN: 9781000652000
• Edition: 1
• Topic: Literature
• Subtopic: Asian Literary Collections
• Index: Literature

Bringing together the study of the Greek classics and Indology, Arjuna–Odysseus provides a comparative analysis of the shared heritage of the Mah?bh?rata and early Greek traditions presented in the texts of Homer and Hesiod.

Building on the ethnographic theories of Durkheim, Mauss, and Dumont, the volume explores the convergences and rapprochements between the Mah?bh?rata and the Greek texts. In exploring the networks of similarities between the two epic traditions, it also reformulates the theory of Georges Dumézil regarding Indo-European cultural comparativism. It includes a detailed comparison between journeys undertaken by the two epic heroes – Odysseus and Arjuna – and more generally, it ranges across the philosophical ideas of these cultures, and the epic traditions, metaphors, and archetypes that define the cultural ideology of ancient Greece and India.

This book will be useful to scholars and researchers of Indo-European comparativism, social and cultural anthropology, classical literature, Indology, cultural and post-colonial studies, philosophy and religion, as well as to those who love the Indian and Greek epics.

1
A starting point

Since my purpose is to argue for a development or expansion of Dumézil’s theory, I have to begin with a statement of it. This has often been done before, but I try to do it in such a way as specifically to facilitate my own case.

Dumézil’s theory¹

The theory starts from the methods and findings of comparative linguistics. Comparison of the Indo-European (IE) languages shows similarities so numerous and so precise that the languages must go back to a common origin, a proto-language (or group of dialects) that was once spoken in a smallish area but which subsequently dispersed and differentiated. Presumably, the speakers shared other aspects of their culture, and in particular an ‘ideology’, a general view of the world and society; this would have spread and differentiated alongside the language, leaving equally little trace in the archaeological record. For evidence of it, one must look instead to the written sources (in general, the earlier ones), where it would manifest itself in a range of contexts: in institutions (particularly social structure, legal codes, the pantheon, ritual offices, acts and invocations) and in narratives (myth, legend, folklore). It is by comparative study of such material that Dumézil, since 1938, has been developing his theory, helped by a few others who can loosely be termed Dumézilians.²

According to this theory, the Indo-Europeans conceived of the world as dependent on the co-operation of three forces or principles or ‘functions’. These abstract principles are not usually attested explicitly, as such; what we find is rather their manifestations. Thus, the analyst must devise his own labels for the abstractions. Since the functions are usually manifested in a particular order, which also commonly represents their rank, the labels Dumézil uses are the ordinals ‘first, second, third’. For brevity and convenience, I shall usually use my own labels Fl, F2, F3. The first function, Fl, pertains to the sacred, to religious and magical power, to the juridical and to sovereignty; F2 to physical force and war; F3 to fecundity, abundance, and prosperity with their conditions and consequences: peace, nourishment, health, sensual pleasure, beauty, large number. When it comes to applying the theory, the full definitions are important, but for mnemonic purposes, one can say that Fl is sacred power, F2 warrior force, F3 fecundity/prosperity. A function can be envisaged either as a cluster of ideas or as a focal idea surrounded by a penumbra of associations or connotations; but in either case it must be envisaged as one element within a three-element structure.

The theory can scarcely be presented intelligibly, let alone plausibly, without examples. 1 refer, very briefly, to three. Hindu tradition conceives of society as consisting ideally of four strata or components called varṇas. The first three, the ‘twice-born’, are in several senses a group set apart from the fourth. Within the triad, highest rank goes to the Brahmins, who are priests dealing directly with the sacred, therefore Fl; then come the warrior Kṣatriya – F2, then the humbler Vaiśya, the producers of wealth – F3. Or take the Greek story of the origins of the Trojan War. Three goddesses compete before Paris for the prize of beauty. Hera offers sovereignty – Fl; Athena offers victory in war – F2; Aphrodite offers the beautiful Helen – F3 (Dumézil 1968: 582). Or again, among the most ancient ritual officiants in Rome were the three flamines maiores: the first served Dius Fidius, who was very closely linked with the sovereign Jupiter – Fl; the second served the warlike Mars – F2; the third served the humbler Quirinus, who is linked with peace and with festivals for agricultural prosperity – F3 (Dumézil 1974: 163–172).

Though it is not Dumézil’s habit, such examples can usefully be summarized in tabular form. In each row, one writes first the particular culture and context (institutional or narrative), then the representatives of each function in turn. The representatives will thus form three columns, each linked with one function (Table 1.1).

The two-dimensional display corresponds to Dumézil’s two rules for trifunctional analysis (1977: 224, 1979: 77, 1982: 113). Looking horizontally, the terms in a row must be distinct, solidary, homogeneous, and exhaustive. Thus, the varṇas are non-overlapping, and mention of one implies the others. They are homogeneous qua categories of twice-born, and they and they alone make up the twice-born as a whole. Looking vertically, each entry in the column unambiguously relates to the cluster of ideas defining the column’s function; it need not relate to all the individual ideas in the cluster (especially in the case of F3), but it must be shown convincingly to belong in its own column and not in any of the others. Thus, the Vaiśya are not associated with sensual pleasure or beauty (as Aphrodite is), but their place in the F3 column is earned by their association with wealth, and there are no arguments for putting them in the Fl or F2 columns.

Table 1.1Tabulation of results of three well-triangular known trifunctional analyses

Culture	Context	F1	F2	F3
Hinduism	theoretical social strata	Brahmins	Kṣatriya	Vaiśya
Greece	deities in episode of epic	Hera	Athena	Aphrodite
Rome	category of priest	fl. dialis	fl. dialis	fl. quirinalis

Note: Making the functions head columns rather than rows is not only convenient typographically but also conventional in that structuralists regularly show syntagmatic relations horizontally and paradigmatic ones vertically.

The simplest possible manifestations of the ideology have one representative per function, as in the examples cited so far, but a single function may have more than one representative: for instance, in a single context, F3 is sometimes represented by twins. Moreover, such multiple representation of a single function may be sporadic and apparently unsystematic, but it may also be so common and consistent as to justify recognition of a subordinate level of structure. The best-known example relates to Fl, which not uncommonly shows an internal binary split between a friendly aspect named after the Vedic god Mitra and an alien anxiety-provoking one named after his associate Varuṇa. Dumézil’s theory is concerned with three functions, not with triplicity per se; and it can readily accommodate more than three entries per row and even the subdivision of columns.

I have introduced the theory as though the functions came first, the examples second, as if the columns had headings before they had contents. In fact, Dumézil first saw the analogy between the varṇas and the flamens (1938), and only subsequently, in light of other examples, abstracted the definitions of the functions. Similarly, the theory that the three functions together constituted the ideology was not an a priori postulate; it was a judgement formed as examples of manifestations accumulated and proved to be surprisingly pervasive. By pervasiveness I mean not so much the number of Indo-European societies in which manifestations can be recognized (this naturally depends on the date and character of the available sources) but rather the frequency of examples, the variety of types of context in which they occur, and the importance of the manifestations within the cultures. It seems that the ideology is most pervasive among the Celto-Italics and Indo-Iranians, that is, at the two geographical extremes of the IE world, which are thought to have been more conservative than the centre.

This macrohistorical point brings us to two broad issues that have not received sufficient attention: what is the relationship of the trifunctional ideology, on the one hand, to contemporary ideologies in the IE world, and on the other hand, to ideologies elsewhere?

IE macrohistory: from segmentary ideology to non-segmentary ideologies

The IE-speakers of today, dispersed over all the continents, call on various and competing ideologies including Christian, Islamic, Marxist, liberal, and mixed. Opinions would differ on how to set about analyzing them, but no one suggests that trifunctional analysis offers the key. One feels intuitively that modern ideologies are qualitatively different from the sort of thing that Dumézil envisages (indeed this may account for part of the hostility his theory has encountered). It is not merely that modern ideologies are consciously in competition, while the trifunctional ideology was presumably taken for granted. The main point is that modern ideologies do not consist of discrete clusters of ideas such as could form the headings for columns containing lists of their manifestations. Let us use the term ‘segmentary’ to distinguish ideologies that are susceptible to display in columns from non-segmentary ones that are not. A segmentary ideology is one that organizes the major ideas that it handles into a small number of sharply contrasting clusters.³

To stress the typological contrast is not to deny the historical continuity. The legacy of the old ideology may be detectable today in various forms, ranging from straightforward punctate manifestations in folkloristic contexts to diffuse attitudes. One thinks of the recent discovery of a trifunctional triad of saints in a festival still performed at Gubbio in Italy (Dumézil 1979: 123–143) and of the enduring importance of the varṇas in India. But these cannot be recent inventions. They are clearly survivals, creations of earlier times. In broad terms, the trifunctional ideology is ancient history, dead and gone, or at most transmuted beyond recognition.

From this point of view, the history of the Indo-Europeans consists in a transition from an early period, when the segmentary ideology was creative and pervasive, to the contemporary period, when it has, to all intents and purposes, given way to non-segmentary ideologies. However, the idea of the earlier period lacks precision. The problem is not only practical – the patchiness of the evidence from early times, the difficulty of reconstructing prehistory – but also conceptual. Continua are often difficult to handle, and an ideology can very easily be only partly alive. It might be alive for the masses, dead for the elite, or vice versa. It might be strong enough to effect adaptations, too weak to inspire genuine innovation. It might be alive in narrative, dead so far as institutions were concerned. And so on. I imagine that even the earliest of our sources date from periods when the ideology was no longer as alive as it had been when the proto-Indo-European (PIE)-speakers were beginning to disperse. However, it would be arbitrary to assume that that was the time when the segmentary ideology was at its strongest. The process of decline could well have begun earlier – one can say nothing about dates. Nevertheless, either the notion of a long-term global loss of segmentation from the ideology is incorrect or one is forced to hypothesize a period when the segmentariness was at its maximum. To conceptualize this period, it is helpful to construct an ideal type, a model society in which the trifunctional ideology was as alive and creative and pervasive as an ideology could be.

One could start by assembling from all over the IE world all the types of context in which manifestations have been discovered, and one could try to devise others, and for each type of context there would be row after row of manifestations. The members of the society would experience as obvious the unity of a function, the homogeneity of the entries in a column. The ideology would not only pattern the social structure, institutions, and narratives but also dominate speculation on the macrocosm and the microcosm and guide the interpretation of events and the assimilation or development of cultural innovations.

Such a model is intended simply as an aid for reflection. It is not necessary to imagine that it was ever realized in all its simplicity and consistency, or that the subsequent history of the ideology consisted of nothing but decline in segmentariness. Global recrudescences seem implausible, but one cannot exclude occasional bursts of creativity in particular types of context. Even so, the overall trend has to be decline, and to conceptualize such a process, we need a starting point.

IE and non-IE: segmentary ideologies viewed anthropologically

Dumézil sometimes presents his own comparative method as a transposition to a larger scale of the comparisons routinely undertaken on a smaller scale, for example by Hellenists who compare Dorians, Ionians, and so forth (Dumézil 1973: 11). But he is well aware (ibid: 21–22) that something should be gained by moving to a wider scale again, beyond the range of individual language families. This means abandoning the well-tested family-tree approach to comparison but opens up other sorts of questions. In particular, do we know of other societies possessing segmentary ideologies that are as systematic as the one we have attributed to our model society?

In fact we do, though they have usually been discussed in different analytical vocabulary, in terms not of ‘ideologies’ but of totemism, symbolic classification or (more recently) isomorphisms. Some of the clearest examples come from Australia, though similar phenomena are not uncommon in the tribal world. The locus classicus remains the 1903 essay by Durkheim and Mauss on primitive classification. This rich paper, often misunderstood [Allen 2000a: Ch. 2], would certainly be important in any exploration of the intellectual roots of Dumézil. He himself has often attributed the genesis of his theory to the influence of Mauss and, above all, Granet (e.g. in Dumézil 1981a: 21), and it would be easy to show Granet’s debt to the paper by Mauss and his uncle.⁴

After their introduction, Durkheim and Mauss start off by collecting reports from aboriginal Australia to the effect that all the members of the social world (the tribe) and all the contents of the natural world (animate or not) are parcelled out into a small number of discrete classes. Thus, a single class will contain not only humans of a certain category but also particular species of plant and animal (‘totems’) and sometimes meteorological phenomena, heavenly bodies, ‘elements’ such as fire, and so on. How similar is such a classification to the ideology of our model society? The small fixed number of classes corresponds to the small fixed number of columns, but one can perhaps recognize two main differences.

First, the Australian material appears to put more emphasis than Dumézil does on the totalities structured by the classification. We shall return to this issue.

Second, the Australian classifications, as reported before 1903, primarily concern the concrete and visible rather than spiritual forces or principles. It is not clear what abstractions, if any, give coherence to the classes and provide a rationale for the allocation to them of particular contents; in other words, there is no obvious equivalent for the functions that define and label the columns in the IE case.⁵

For our purpose, it is not necessary to pursue these differences by recourse to further sources, old or new. The point is that they do not suffice to make the term ‘ideology’ inappropriate to the ethnographic material (nor, conversely, to make ‘classification’ ina...

Table of contents

• Cover
• Half Title
• Title
• Copyright
• Dedication
• Contents
• List of figures
• List of tables
• Acknowledgements
• List of abbreviations
• Signs
• Introduction
• 1 A starting point
• 2 Five relationships
• 3 Homer’s simile
• 4 Hero and horse
• 5 Yoga
• 6 Crocodiles and nymphs
• 7 Monkey and dog
• 8 Durgā and Athena
• 9 Draupadī and Penelope
• 10 Bhīṣma and Sarpedon
• 11 Hesiod’s Succession Myth
• 12 Five elements
• 13 Rings and rotations
• 14 Achilles’ shield
• 15 Dumézil and Dumont
• 16 Yudhiṣṭhira and Agamemnon
• 17 Kauravas and suitors
• 18 Hanging over abyss
• 19 Gods descend to battlefield
• 20 Heroes and supercategories
• 21 Cyavana and Prometheus
• 22 Telemachy
• 23 Droṇa and Chryses
• 24 Aśvatthāman and the Wooden Horse
• Bibliography
• Index