Nick J. Allen

Massive similarities exist between ancient Greek and ancient Indian philosophy, and a massive study such as Thomas McEvilley’s was needed to assemble the materials and grapple with them.  One has to applaud the scope and ambition of the book.  Naturally, anyone who undertakes an interdisciplinary study on a grand scale risks making mistakes and misjudgments, and McEvilley sensibly anticipates criticism (p. xxxi) – which has not been lacking.  For instance, at the lowest level, the diacritics are appallingly slapdash, and the French accents are little better.[1]  However, my aim is not to criticize the book piecemeal, but rather to respond to it constructively at a level commensurate with the scale of its undertaking.  I shall address a systemic problem that is built into the initial assumptions.

                                     THE MISSING DIMENSION

McEvilley argues that the similarities between Greek and Indian philosophy are due to diffusion, i.e. to historical contacts.  More precisely, for the earlier period the choice lies between Mesopotamian ideas spreading both east and west, and Indian ideas spreading west via Greek-Indian encounters within the Achaemenid empire.  From Alexander to the Fall of Rome, the predominant flow was east-wards, whether via the Greek presence in the north-west of the subcontinent, Roman coastal trading ports, or Indian visits to the Mediterranean world.  Though not well equipped to judge them, I thought the evidence for contact in the later period was good.  However, for the earlier period the argument neglects an alternative explanation of the similarities, namely Indo-European common origin.

In principle, the difference between the two modes of explanation is sharp.  Since the very notion of Indo-European comes from linguistics, let us draw on that discipline.  If a word in one Indo-European language resembles a word in another, and the resemblance cannot be due to chance, then it may either be because one language has borrowed the word from the other (or both have borrowed from a third language, related or not), or because the words descend independently from the ancestral language.  In the former case the word was once foreign to the borrowers, in the latter it was never foreign to the speakers of either language.  In practice, the distinction may not always be easy even for the linguistic comparativist, and usually the latter faces an easier task than the cultural comparativist.  Nevertheless, if we wish to attribute similar philosophical ideas in Greece and India to diffusion, we need to be sure that the explanation by common origin is impossible.  Otherwise, the similar ideas could merely be continuations of an old idea that was current among those who spoke the common ancestor of Greek and Sanskrit.[2]

The shared Indo-European background of Greece and India does receive brief mention in the foreword (p. xxiv).  The trend emphasizing this background ‘culminated in the work of Georges Dumezil [sic] and others who have articulated parallelisms in the social structures of different Indo-European-speaking cultures’ (reference to Littleton 1982).  By ‘social structures’ McEvilley is thinking of the varna system in Manu, the classes in Plato’s Republic, and the social organization of early Latin peoples.  ‘But in order to account for striking comparative details such studies must be supplemented by postulates of historical influences.’  Neither Dumézil nor the relevant others are cited, and the topic scarcely surfaces again.

This treatment of the Indo-European background denatures Dumézil’s work, reducing it to just one of its components and oversimplifying even that.  Social structure was important in the genesis of Dumézil’s breakthrough in 1938, and it was the functions performed by the three twice-born varṇas that led to the label ‘trifunctionalist’, which is now attached to his work.  However, for Dumézil and those working in his tradition, the three functions form an ideology, a mental framework which patterns many social phenomena in addition to social structure – pantheons, rituals, legal procedures, narratives of various types (myths, epics, pseudohistory, tales).  Moreover, the reference to social structure is itself too simple.

A decisive step forward was taken on the day when I recognized, around 1950, that the ‘tripartite ideology’ is not necessarily accompanied, in the life of a society, by the real tripartition of society, as in the Indian model; that on the contrary, where it is to be found, it is possible that it is nothing but (is no longer anything but, perhaps never was anything but) an ideal and, at the same time, a means of analysing and interpreting the forces that ensure the course of the world and the life of men (Dumézil 1968: 15).[3]

Elsewhere the same recognition is expressed in different words.  Between 1938 and about 1950, overinfluenced by the role of the varṇas in his breakthrough, Dumézil had assumed that manifestations of the ideology, wherever they occurred, indicated the concurrent or earlier existence of a real division of society into distinct functional classes; but he now saw that it was not legitimate ‘to move from the ideology to conclusions about practice, from a philosophy to conclusions about social organization’ (1981: 338).

Thus Dumézil’s enterprise was an effort to recognize in the available facts the survival of the ideology or philosophy current among early Indo-European speakers.  Linguists sometimes argue about the ontology of the starred forms they reconstruct, and Dumézil was usually cautious in his wording, but in effect he reconstructs an Indo-European protophilosophy.  It follows that a comparative philosopher has three options.  The best is to take account of the Indo-European common origin of Greek and Indian tradition and incorporate the findings of the Dumézilian enterprise.  Another is explicitly to write off the enterprise, preferably by alleging serious reasons and not simply citing ‘authorities’.  The third, a compromise, is to try and drive a wedge between the reconstructed protophilosophy and the attested Greek and Indian philosophers.  This might be attempted by claiming that the former is not real philosophy (e.g., it is too limited in scope or too close to myth); or that, if it is philosophy, it does not connect historically with the attested philosophies; or that, if such connection does exist, it is too tenuous to account for striking similarities of detail.  McEvilley apparently espouses some form of the compromise argument, and we need to ask how valid it is.

In thinking about continuities and connections between early Indo-European speakers and attested texts, a problem to be confronted head-on is the relationship between the dating of texts and the dating of their contents.  The dating of texts (or of their stabilization within an oral tradition) is a topic for specialists and obviously worthwhile, albeit often intractable.  But it is all too easy to slide from this to the dating of contents.  Other things being equal, one assumes that form and contents go together, so that an earlier text contains ideas current at an earlier period, and a later text from the same tradition contains ideas that, in so far as they differ from the earlier ones, were developed later.  But once comparison enters the picture, other things are seldom equal.  When a later text contains an idea or theme that is strikingly similar to one in another branch of the same tradition, then the common origin explanation needs consideration even if the idea is absent from the earlier texts. The earlier absence can be explained in several ways.  The idea may have been known to the composers of the earlier text but passed over as inappropriate to their genre, or excluded as esoteric (a ‘Mystery’); or it may have been unknown to them but known to other social categories inhabiting the same area (e.g., unknown to priests but known to warriors); or it may have been known within the same social group but in another area.  Thus the absence may be ascribed to genre, social category or geography.  But whatever the explanation, ideas can bypass earlier texts to surface in later ones.  Because the motive for invoking such a bypass often comes from comparison, history as envisaged or written by comparativists is likely to differ from histories of the same cultures produced by non-comparativists; and since, for the present, comparativists are few in number and most of the history that is written is by non-comparativists, bypass phenomena will often be ignored.

These considerations bear directly on the a priori possibility that an Indo-European protophilosophy lies behind both Greek and Indian philosophizing.  If particular  philosophical ideas are absent from the Vedic hymns or the Brāhmaṇas, or from Homer and Hesiod, this does not prove that they were absent from society as it existed when those texts were stabilized; the ideas could have bypassed the earliest texts in one of the ways we mentioned.  For instance, a variant of the geographical explanation might envisage different waves of Indo-Aryan speaking migrants carrying different components of the tradition, as has been proposed for Vedic India by Asko Parpola (2004–2005: 27–28).  In the Greek case the dating of texts is usually more precise, but the bypass problem is no less real and applies even within the tradition.  It is usually assumed that the ideas of Socrates and his pupil Plato are later than those of the pre-Socratics, and it takes a certain effort to set aside so deeply rooted an assumption.  But although the texts are later, the ideas need not be.


Plato is particularly relevant here because, while the mature Dumézil wrote relatively little about Greece and even less about Greek philosophy, he was very aware of the Indo-European heritage in the Republic.  Already in 1941 (275) he raised the possibility of Plato’s ideal city being ‘in the strictest sense an Indo-European reminiscence’, and he reverted to the topic several times.  In 1982 (256 n.3) he talks of the Republic as containing ‘remarkable expositions of the tripartite ideology’.  The main account is in 1968 (493–96), in a discussion of the intelligence attributed in Ossetic folklore to the Nart hero Batraz: Dumézil remarks that the political psychology in play is close to the one that Plato expounds, ‘certainly on the basis of very old trifunctional speculations’.  Having summarized the correlations between the classes in the Republic (philosopher kings, warriors, and commoners – farmers and artisans being grouped together), the virtues (wisdom, courage and prudence), and the metals (gold, silver, then iron and bronze), Dumézil comments that, since Pythagoras and no doubt before him, Greek philosophers had speculated a lot on social tripartition; it was a concept they retained no doubt from the Indo-European past, even if Plato in Athens could observe a few survivals of the scheme (e.g., in the three archons) (for fuller discussion see Bodéüs 1972).

However, Dumézil does not here or elsewhere present a precise comparison between Plato and Indian philosophy.  On the one hand he is comparing the political psychology in Plato with that in the Caucasian folklore (the Ossetic language belongs to the Scythian branch of Indo-Iranian).  On the other, he is saying that Plato’s philosophical discourse about classes and virtues (and he also has in mind Plato’s three-soul doctrine) belongs to a tradition going back to the old Indo-European protophilosophy, or more precisely, to the application of that philosophy to the ideal organization of society.  It is the same source that lies behind the varna doctrine, itself so central to aspects ofdharma.  We are certainly dealing with continuity.

Dumézil also provides answers to the problem of ‘striking comparative details’.  In spite of the long stretches of time and space involved, he often shows that quite small details in attested materials go back to a common origin.  Thus numerous examples can be found in his comparison between one of the myths attached to Indra and the pseudo-history of the third king of Rome (Dumézil 1985a, followed up by Allen 2003).  But for the moment I leave Dumézil (who died in 1986) and take up more recent work, focusing first on the functions.

Evidence is accumulating that three functions are not enough.  Many of the hierarchized triads that Dumézil and followers have recognized are in fact situated within larger pentadic structures: the triad forms a coherent core, but this core is enclosed or bracketed by one element that is hierarchically superior and another that is inferior.  The triad ‘priests, warriors, producers’ is in many cases bracketed by the king at the top and the serf or slave at the bottom.[4]  To make a long story short, I think that the protophilosophy was pentadic.  For the core functions Dumézil’s definitions can stand, except that sovereignty should be excised from the first function.  The two extremes are covered by a fourth function, defined as relating to what is other, outside or beyond relative to the core; but the fourth function has two aspects, one valued positively, the other negatively.  The notion of positive and negative value will cover a variety of phenomena depending on context, but its application to king and slave is obvious.  In terms of ideology, whether or not an individual king happens to be good or bad, he represents the society qua totality, and is positively valued in the same sense as a whole relative to its parts. As for the slave, however valued he may be as an individual, in a traditional hierarchical society his status is so devalued that he is barely part of society.  In the religious domain the equivalents of king versus slave are Creator versus Devil, or Salvation versus Death/Destruction.  In the annual cycle they are the New Year (taken as a whole) and the Old Year or its closing phase.


The case for the pentadic schema depends primarily on the number, wide distribution and cultural importance of the contexts to which it relates, and on the rigor of the arguments supporting this relation.[5]  Let us for the sake of argument assume that the schema is well founded and ask how it might apply to Greek and Indian philosophy.  Since the schema has four functions and five slots, one obvious target is McEvilley’s chapter on the elements (pp. 300-309): both Greece and India recognize four elements, sometimes adding a fifth.

McEvilley’s tentative conclusions are as follows:

The doctrine of the four elements would seem to have arisen in a single source, perhaps in India, where the developmental sequence is clearer than in Greece, and to have entered Greece in different versions, partly conflated with the Doctrine of the Five Fires and Two Paths.  Some Near Eastern background, which can only be vaguely discerned, may have been in effect.  The doctrines of the fifth elements, ākāśaand aither [sic for aithēr] surely are cognate concepts… Most likely the Indian concept was imported into Greece in a later phase of the same general wave of Upaniṣadic influence which brought the transformations of [concepts of] matter (308–9).

In the present context a case study cannot be developed at length, but at least I hope to show that an Indo-European ancestral doctrine offers a rival hypothesis to diffusion.

The four elements are essentially the same in the two traditions: fire; air, wind or breath; water; earth.  This is the standard order in Greece (sometimes reversed), and seems to follow the order of the functions.  The standard order in India differs in that air regularly precedes fire (as also happens in Heraclitus).  For various reasons I draw mainly on Indian or Indo-Iranian data.

Fire (agni) provides a good starting point since in Vedic India Agni is the priest of the gods (more precisely, its hotar).  Though Agni shares priesthood with Bṛhaspati, fire is the only element to enjoy this status.  Since the attributes of any one Vedic deity tend to be shared with several others, passages can be cited that constitute exceptions to almost any theological statement; but the priestly role of Agni is such a standard feature of introductions to Vedic religion that citing details would be pedantic.  Moreover, the role makes good sense in that it is the fire on the altar, together with the priests around it, that links men to gods.  The association between fire and priesthood is equally clear in Zoroastrianism, with its sacred fires entrusted to fire-priests in fire temples, so it is surely at least Indo-Iranian.  Fire is thus a strong candidate for interpretation as first-functional (F1), provided that all the other elements can be linked to other functions.[6]

Air is represented in India by wind, vāyu or vāta.  Unlike Agni, Vāyu is a minor figure both in the Vedic and later Hindu pantheons, but he has one important role in the Mahābhārata.  He fathers Bhīma, the second of the five Pāṇḍava brothers, the one who for half a century has been associated by comparativists with the second function (F2), which pertains to physical force and war.[7]  Bhīma is indeed the largest and most muscular of the brothers: he once picks up the whole family and carries them with the speed and force of the wind (1.136.16-19; 1.137.23).  In his 1968 analysis of Bhīma Dumézil presents Vāyu as an old Indo-Iranian war god, who has largely bypassed the Vedas.  Moreover, the association between wind and force makes good sense – one has only to think of a hurricane.  Thus, as an element, wind is a reasonable candidate for F2.

Dumézil himself (1973: 77; 1985a: 30; 2000: 121-138) connects water with the third function, for instance when writing on the Norse god Niord and his links with the sea.[8]  Water, whether from rain or irrigation, is essential to the fertility of the peasant’s fields, and fertility is an important component in the definition of this function (together with wealth, abundance, fecundity, large number, health, sexuality…).  It is of course needed not only for the growth of plants and crops but also for the well-being of herds (often a measure of wealth), not to mention the health of humans.  Fertility and well-being are more directly linked to water than to wind or fire, and one need hardly mention the common assimilation of water and semen (as when rain is viewed as heaven inseminating earth).  Water is intrinsically an excellent candidate for F3.

As for earth (pṛthivī), the first question is whether it stands apart from the other elements.  In mentioning that in the Atharva Veda fire, water and air or breath (or life-force, prāṇa) are each in different passages treated as the ultimate principles, McEvilley (p. 302) implies that earth is not.  Similarly those early Greek philosophers who derived everything from a single element made their choice from the same triad and (as Aristotle noted) never selected earth as an arkhē or ultimate principle.  However, in general the texts align at least the four elements, and the heterogeneity of earth, if it exists, must be sought among its taken-for-granted qualities or in narratives.  One obvious quality is immobility – in a traditional world view earth does not normally move, let alone spread, blow or flow.  Moreover, earth is the only element unambiguously personified not by a god but a goddess.  As regards cosmogony, Pherecydes of Syros derived fire, breath and water from the seed of Time, while Zeus, Time and Earth always existed (p. 306); and in epic, both Pṛthivī and Gaia complain of being overburdened and thereby initiate the Great War (e.g., Vielle 1996: 116). 

If earth stands apart from the other elements, is it in any way devalued?  Let us try to correlate elements (the division of matter) and human activities (the division of labor).  If the priest uses his sacred fires for rituals, the warrior emulates the speed and force of the wind, and the producer exploits the fertility of water, then who relates most intimately to earth?  Dumézil (e.g., 1985a: 30) combines earth with water under the third function, implicitly relating the element to the agriculturalist.  But from a four-functional viewpoint, a better answer is the miner or quarry worker, the blacksmith or stone-mason.  It is they who extract and use ‘the bones of the earth’, and in the caste system they belong among the Untouchables, the F4- component of society (Allen 2006).

As for the fifth element, the quintessence (which can be called ether), in Vaiśeṣika as in Aristotle, it is ‘kept carefully separate from the others’ (McE. 525).  It relates to matter so rarefied as to resemble mere space, and it ‘is characterized by sanctity.’  Its non-appearance in the earliest texts may reflect this very heterogeneity – it need not have been a later addition to the four.  In both traditions it (rather than any other element) is linked with the cosmos, i.e. the totality of things, and hence can represent F4+.

Entities representing the two aspects of the fourth function stand apart from those representing the core functions, and the ways in which they differ from the core may or may not themselves differ (in addition to contrasting as superior and inferior); but quite often representatives of the two aspects, taken by themselves, have points in common.  In this case, ether often shares in the effective immobility of earth, and in India at least it perhaps shares to some degree in earth’s femininity. Ākāśa is semantically close to dyu or Dyu (Heaven), with which Vedic pṛthivī or Pṛthivī is usually coupled; and dyu is feminine in about twenty passages, sometimes even when personified (Macdonell 1981: 22, 88).[9]

This cursory account of the elements in Greece and India can be complemented by a glance at the Zoroastrian Bounteous or Beneficent Immortals, the Aməša Spəntas.  By the post-Gathic period these six or seven spiritual beings came to constitute a more or less standardized list.  Each Immortal was correlated with a material entity, and among the entities are some of the familiar elements.  This theological structure was analysed trifunctionally by Dumézil (1977: 37-51; his tabulation in 1994: 60 is particularly neat).  Replacing the Avestan names with the English translations in Boyce (1975: 203), we can present the analysis as follows, underlining the correlations that are most relevant here. 

      Bounteous Immortal           function of B.I .    material correlate of B.I.

     I           Good Intention             F1                                bovine

     II          Best Righteousness                   F1                                fire

     III         Desirable Dominion                  F2                                metal

     IV         Bounteous Devotion                 F3                                earth

     V          Wholeness                                F3                                water

     VI         Life                                          F3                                plants

Although he cites the row V correlation in his analysis of Niord, Dumézil ignores the row II correlation when discussing Agni; but both accord with our functional analysis of the elements.  Similarly, the Immortals in rows V-VI, Haurvatāt (or Health) and Amərətāt (or Non-Death), have names that are similar both in morphology and semantics; the pair are convincingly compared by Dumézil with the twins so typical of the third function; and as we noted earlier, their material correlates fit well with the F3 notion of fertility.

On the other hand, row IV runs counter to the idea that earth represents F4-, and Dumézil’s F3 interpretation has to be questioned.  Since V and VI gain their feminine gender from the abstract-forming suffix –tāt and were probably originally male (Dumézil 1977: 45-46), Bounteous Devotion (Spənta Ārmaiti) is the only Immortal who is straightforwardly female, which sets her apart from the other Immortals and suggests comparison with Pṛthivī and Gaia.[10]  Moreover, when the list contains seven Immortals, it may open with Bounteous Spirit (Spənta Mainyu, correlating with man) and put Bounteous Devotion last (Varenne 1981: 582b).   Spənta Mainyu is conceptually close to and sometimes identified with Ahura Mazdā, the sovereign Creator, so one suspects that Ahura Mazdā and Spənta Mainyu represent F4+, and Spənta Ārmaiti and the Evil Spirit (Angra Mainyu) represent F4-.

Critics of Dumézil often object that the evidence for the three functions in the Vedas, and especially the hymns, is less plentiful than one might expect.  According to Dumézil, the typical or canonical expression of the trifunctional pattern in the Indo-Iranian pantheon is Mitra-Varuṇa (F1), Indra (F2), the Aśvin twins (F3), the pairing of the first two and last two gods being reflected in the pairing of the first two and the last two Immortals.  This pattern can be found both in the hymns (Dumézil 1977: Appendix 1) and the Brāhmaṇas, but it cannot be said to dominate them; Agni and Soma (both of them visible entities as well as gods) are more prominent than Mitra-Varuṇa and the Asvins, as Dumézil recognizes (1968: 58).  But if Agni represents F1, as we have argued, Soma could represent F3.  Like Agni, Soma has a large and complicated dossier, but if Agni is saliently a priest, Soma is saliently king of plants or herbs and lord of waters – the same conjunction as we found for the correlates of the paired F3 Immortals.

In this connection one might think of the triadic classification of Vedic ritual based on the main oblation: between offerings into the fire of dairy products or cereals and libations of soma, one findspaśubandha, the sacrifice of animal victims, where ‘control of the breath [of the victim] is of paramount importance’ (McClymond 2003: 235).  Since breath or life-force, like air, relates to F2, the classification seems trifunctional.  If so, it supports both the general Dumézilian argument for the pervasive presence of the core functions, and the link proposed here between the elements and the four functions.[11]

Of course, even if one rejects their link with functions, the elements are not unrelated entities that just happen to be juxtaposed in a certain order.  However one understands the ordering of air and fire, the overall sequence from ether to earth is one of condensation and descent, while its reversal implies rarefaction and ascent.  Both traditions used these unifying principles in their cosmogonic speculations, but also in their eschatologies, when describing the path of the reincarnating soul which dies and is reborn (McE. 41).  If the protophilosophy did indeed associate functions and elements, perhaps it too used the association to think about changes or transformations of macrocosm and microcosm alike.

The above argument needs to rebut at least two counterproposals.  Firstly, the similarities might indeed be due to use of the four-functional ideology to analyse materiality, but the application might have been carried out independently in East and West.  However, even if the old ideology was sufficiently alive in sixth century Miletus, it is unlikely that the results of the application within two such different cultures would be so similar.  Secondly, if the application was made only once, but at some point in the history of the Indian tradition, this does not rule out the east-west diffusion envisaged by McEvilley.  To answer this objection we must broaden the discussion and return to the question of whether common origin can account for striking similarities of detail.


Apart for the idea of a split fourth function, another development since the death of Dumézil has been the idea of an early Indo-European protoepic or protonarrative, features of which can be reconstructed by Greek-Indian comparison (e.g., Vielle 1996).  It is now hard to doubt that a substantial epic tradition bypassed the Vedas to surface around the turn of the eras in theMahābhārata.  Whatever may be the scope for comparison between Greek and Vedic traditions, the scope for comparison between the two epic traditions is immense.  Such comparisons may involve macro-structures such as the five phases of the wars at Troy and Kurukṣetra, but they may also relate to tiny details.  One instance is the use in cognate contexts of the masculine accusative singular of the present active participle of a verb that is cognate in the two languages (Allen 1999: 164).

Of course, as in philosophy, the hypothesis of diffusion raises its head, but here too it faces numerous difficulties, which are discussed in several of my papers.  In brief, the following seem to be the main problems.  (1) The difficulty of envisaging a context for the encounters (where, when, in what language).  (2) The fact that in some respects the Mahābhārata parallels closely not only Greek epic but also other Indo-European traditions such as Roman pseudo-history (Allen in press b).  (3) The deep embedding of each epic within its local religious and cultural traditions (much deeper than is usual with folktales); neither ‘feels’ like a borrowing.  (4)  In world-historical perspective, the correlational style of thinking manifested in the Indo-European ideology  (a ‘primitive classification’) has been losing ground over the millennia, yet its patterning effect is apparent in both epic traditions.  It is easier to suppose that it operated on a protonarrative than that it operated twice, independently, in the two branches of the tradition.

The more one accepts the idea of a protonarrative lying behind the epics, the less reason there is to resist the idea of a protophilosophy, for the two genres turn out to be much less separate than one might anticipate.  The last part of this paper introduces three philosophical topics that can be studied at least in part via epic.

McEvilley gives plenty of attention to yoga, suggesting tentatively that an early Mesopotamian doctrine diffused in both directions, and that the Indian variety, after elaboration, spread back west to Greece in the sixth century bc (p. 287).  But the network of similarities between Arjuna’s ascent to heaven in Mahābhārata Book 3, Odysseus’ passage to Scheria in Odyssey Book 5, and the yogin’s undertaking in Śvetāśvatara Upaniad and Patañjali (Allen 1998a) suggests that the protonarrative told of a cosmic/shamanic journey, presumably relating to shamanic practice, that somehow fed into yoga.  It is not clear at what stage or stages the shamanic tradition underwent the interiorization that characterizes yoga, where this occurred, or whether it was a process that essentially occurred just once or one that occurred in parallel in different branches of the tradition.  Even so, discussion of the shamanism-yoga complex needs to take account of the Indo-European common origin hypothesis which, here again, can relate to striking details.  For instance, the references to thistles and chaff in the Odyssey passage are cognate with Patañjali’s references to thorns or cotton fibers (ibid.: 13).

If yoga is treated repeatedly by McEvilley, Sāṃkhya, the darśana with which yoga is traditionally paired, is virtually ignored – despite its pervasiveness in Hindu culture (including theMahābhārata), and despite the various tempting comparisons with Greek thought.  One such comparison concerns the use of numbers in Sāṃkhya and Pythagoreanism, particularly the use of five (for the Greek see, e.g., Mattéi 1996: 108-117).  Arguably (Allen 1998b), in Sāṃkhya the emanational sequence of twenty-five tattvas or principles opens with a set consisting of Puruṣa (associated with cosmogony), the three gunas (sattva, rajas, tamas), and ahaṃkara.  The gunas, which have been described as intelligence stuff, energy stuff and mass stuff respectively (i.e. F1,2,3 – cf. Sergent 1995: 339), constitute prakṛti (‘primal nature’), and the overall one-three-one structure might recall the fragmentary text of Pherecydes cited above.  If the analysis is right, the initial pentad incorporates a manifestation of the four functions; and the final pentad in the sequence consists simply of the five gross elements (mahābhūtas) – our familiar set, in the standard Indian order ending with earth.[12]

Between the first five principles and the last five come three other pentads – the five sense-capacities, the five action-capacities, and the five subtle elements.  These three ‘core’ pentads show a degree of correlation: thus the first member of each is respectively hearing, speaking, sound, while the second is feeling, grasping, touch.  But rather than pursuing the details and asking if or how these individual principles relate to the functions, we can view the pentads as units.  The first core pentad, the buddhīndriyas, relates to the cognitive domain, while the second (karmendriyas) relates to action; and these domains qualify for F1 and F2 (action being a philosophic substitute for dynamism or force).  The label for the third, tanmātras (‘only so much or little, rudimentary, trifle’ – Larson 1979: 187) suggests its lower standing, but the main reason for taking it as F3 lies in its origin.  Whereas the two sets of indriyas emerge from ahaṃkara in its sattva mode, the subtle elements emerge from it in its tamas (F3) mode (the structure F1+2 versus F3 is familiar in the trifunctionalist tradition).  While the core pentads are held together by their origin from ahaṃkara, the final pentad has a separate origin, viz. from the subtle elements.  I take it to be devalued by virtue of its materiality, as well as its position.

In other words, Sāṃkhya appears to manifest the four functions in at least three ways.  On the global level, the pentads themselves show the chacteristic one-three-one hierarchy.  Only the first pentad is linked with creation, and within it we can probably recognize another but interrupted manifestation – compare the anomalous ordering in most lists of Bounteous Immortals.  Finally, the last pentad presents the functionally organized elements.[13]

Since enlightenment and related ideas are so important in both Indian and Greek philosophies, McEvilley naturally returns to them frequently, and one might think that here at least is a purely cognitive topic, to which studies of epic can hardly contribute.  However, if one compares the biographies of Arjuna, Odysseus, Cúchulainn and the Buddha (Allen in press a), in each case one finds, following a period of privation or asceticism, some sort of breakthrough to another and better world, whether it is conceived in terms of cosmology, poetic geography or soteriology.  Arguably, all these events are cognate and go back to a single story in a protonarrative.


More could be said, even on the basis of comparative work already published (e.g. on cosmic time), and it is only when the approach has been taken further, and its limitations emerge more clearly, that we shall be well placed to see how and where to call on the other cultural inputs referred to by McEvilley – Egypt, Mesopotamia, Indus Valley, Dravidian substrate.  One way ahead might be to take the ideological framework, as elucidated so far, and speculate on possible applications to philosophical topics.  For instance, one can reflect on dualisms:

          F4+      The Absolute, Ultimate Being, Totality, Infinity…

    F4-      (the opposite) non-being, unreality, nothingness, the  infinitesimal…

Such contrasts between the aspects of the fourth function may, somehow, underlie the Vedantic Brahman-Ātman equation.  But one can also play with triads: the transcendent One (F4+), the realm of structure (F1–3), meaningless multiplicity, or devalued oneness (F4-); or again with pentads, or the ‘vertical’ scales that underlie them.  This proposal is of course only a heuristic, a way of generating ideas for testing.

My main aim has been to identify a gap in the problematic within which McEvilley works – and not only McEvilley, for the gap is equally obvious in many other writers.  Practitioners of Indo-European cultural comparativism have a vast task ahead if they are to modify the academic landscape to the point where their contribution to historical understanding is taken seriously across the disciplines.  The task is not only intellectual but also concerns the sociology of academe.  Anticipating opposition to his approach, McEvilley talks briefly of ‘issues of turf’, and it is true that neither classicists not Indologists are likely to welcome being told that certain questions relating to their fields can only be answered by going outside it.  But I would rather end on a positive note.  Those like McEvilley who are willing to invest the time and take the risks can enter an exciting and sparsely cultivated comparative field where, despite the pessimists, scientific progress is possible.


[1] Another detail: among the 750 or so references admirers of Marcel Mauss will miss Mauss 1927.

[2] According to one estimate, the common ancestor of Greek and Sanskrit was spoken before 2300 bc, the common ancestor of Mature (‘Brugmannian’) Indo-European was spoken before 2800, and the common ancestor of Proto-IE proper (i.e. including Hittite) was spoken before 3500 (West 2005).

[3] My translation.  This well known passage is cited again in Dumézil (1985b: 321n).

[4] The complications of the Indian case are treated by Allen (2006).  As for Plato, the king has been effectively merged with the philosophers, and arguably the artisans from the bottom have been merged with the farmers.  To put it briefly, the three functions are the bundles of associations attaching respectively to priests, warriors and producers, though there is more to the third function than productivity.

[5] A list of my IE comparativist papers can be found at

[6] Though Dumézil mentions the priesthood of Agni (1968: 204), he does not emphasize it: ‘Among the individual male gods Agni is the one who is most obviously trifunctional and the only one who is constantly so’ (ibid.: 119).  But he is not examining the elements as such.

[7] Trifunctionalists maintain that Arjuna too represents F2, but if so, it is in a less straightforward sense thanBhīma: as son of Indra, king of the gods, Arjuna, though not himself a king, is better seen as representing F4+ (Allen 2006).

[8] When connecting waters with F3 in 1985 he writes of them as ‘fécondantes, nourricières, guérisseuses, nettoyeuses’ (providing fecundity, nourishment, healing and cleansing, p. 30).

[9] Since Dyu is etymologically cognate with Zeus, and Zeus is sometimes equated with aithēr, it is interesting that aither too can be either masculine or feminine.  For some ramifications of the Zeus-Dyu comparison see Allen 2004.

[10] The enlargement of Spənta Ārmaiti (= the earth) in Vidēvdāt 2 is cognate with the enlargement of the earthen container of Matsya in the Indian Deluge story (Allen 2000: 292).

[11] Conceivably the agnicayana, with its bricks, relates to earth, but the whole topic needs deeper study.

[12] I use the translations from the convenient diagram in Larson (1979: 236), who appears in McEvilley as Lawson.  Is there not a remote connection between the genealogical framework of Sāṃkhya and that of Hesiod’s Theogony?

[13] Curiously, when discussing the gross elements, Frauwallner (1973: 228) presents them in the standard Greek order, which accords with none of the three Mahabharata texts he is citing.

References cited

Allen, Nicholas J. 1998a. The Indo-European prehistory of yoga.  International journal of Hindu studies 2: 1–20.

Allen, Nicholas J. 1998b. The category of substance: a Maussian theme revisited.  In W. James and N. J. Allen, eds., Marcel Mauss: a centenary tribute, 175–191. Oxford: Berghahn.

Allen, Nicholas J. 1999. Les crocodiles qui se transforment en nymphes. Ollodagos 13: 151–167.

Allen, Nicholas J. 2000. Imra, pentads and catastrophes. Ollodagos 14: 278–308.

Allen, Nicholas J. 2003. The Indra-Tullus comparison.  In B. Drinka & J. Salmons (eds) Indo-European Language and Culture: Essays in Memory of Edgar C. Polomé, part I General Linguistics 40: 148–171.

Allen, Nicholas J. 2004. Dyaus and Bhīṣma, Zeus and Sarpedon: towards a history of the Indo-European sky god. Gaia 8: 29–36.

Allen, Nicholas J. 2006. The close and the distant: a long-term perspective.  In

      Georg Pfeffer, ed., Periphery and Centre: Norms, Values, Categories, 273–290. Delhi, Manohar.

Allen, Nicholas J. in press a. Asceticism in some Indo-European traditions. Studia Indo-Europaea 2:

Allen, Nicholas J. in press b. Structures qui s’entrecroisent.  Anthropologie et sociétés 29, 2:

Bodéüs, Richard. 1972. Société athénienne, sagesse grecque et idéal indo-européen. L’antiquité classique 41: 455–486.

Boyce, Mary. 1975. A history of Zoroastrianism, Vol.1,  the Early period. Leiden/Köln: Brill.

Dumézil, Georges. 1941, Jupiter, Mars, Quirinus, Paris: Gallimard.

Dumézil, Georges. 1968, Mythe et épopée, vol. 1. Paris: Gallimard.

Dumézil, Georges. 1981 [1973]. Mythe et épopée, vol. III. Paris: Gallimard.

Dumézil, Georges. 1982 [1971]. Mythe et épopée, vol. II. Paris: Gallimard.

Dumézil, Georges. 1977. Les dieux souverains des Indo-européens. Paris: Gallimard.

Dumézil, Georges. 1985a [1969]. Heur et malheur du guerrier.  Paris: Flammarion.

Dumézil, Georges. 1985b. L’oubli de l’homme et l’honneur des dieux: esquisses de mythologie. Paris: Gallimard.

Dumézil, Georges. 1994. Le roman des jumeaux: esquisses de mythologie. Paris: Gallimard.

Dumézil, Georges. 2000. Mythes et dieux de la Scandinavie ancienne (ed. François-Xavier Dillmann). Paris: Gallimard.

Frauwallner, Erich. 1973. History of Indian Philosophy, vol.1: the philosophy of the Veda and of the epic – the Buddha and the Jina – the Samkhya and the classical yogo-system, (trans. V.M. Bedekar). Delhi: Motilal Banarsidass.

Larson, Gerald James. 1979 [1969]. Classical Samkhya: an interpretation of its history and meaning. Delhi: Motilal Banarsidass.

Littleton, C. Scott.  1982 [1966]. The new comparative mythology: an anthropological assessment of the theories of Georges Dumézil. Berkeley: University of California Press.

Macdonell, A.A. 1981 [1898]. Vedic mythology. Delhi: Motilal Banarsidass.

Mattéi, Jean-François. 1996 [1993]. Pythagore et les pythagoriciens. Paris: Presses universitaires de France.

Mauss, Marcel. 1969 [1927]. Rapports historiques entre la mystique hindoue et la mystique occidentale. Oeuvres, vol. 2, 556–560. Paris: Minuit.

McEvilley, Thomas. 2002. The shape of ancient thought: comparative studies in Greek and Indian philosophies. New York: Allworth Press.

McClymond, Kathryn. 2003. Death be not proud: reevaluating the role of killing in sacrifice. International Journal of Hindu Studies 6,3 (December 2002): 221–242.

Parpola, Asko. 20045. The Nāsatyas, the chariot and Proto-Aryan religion. Journal of Indological Studies 16-17: 5–67.

Sergent, Bernard. 1995. Les Indo-européens: histoire, langues, mythes. Paris: Payot.

Varenne, Jean. 1981. Iran préislamique. In Yves Bonnefoy, ed., Dictionnaire des mythologies et des religions des sociétés traditionnelles et du monde antique, A-J, 574–586. Paris: Flammarion.

Vielle, Christophe. 1996. Le mytho-cycle héroïque dans l’aire indo-européenne: correspondances et transformations helléno-aryennes. Louvain-la-neuve: Institut Orientaliste.

West, Martin. 2005 (20 January). The Indo-European hero.  Lecture at Classics Centre, Oxford.