Veda & Dharma‎ > ‎

Classifying the Universe

by Vidhyanath Rao

It is well-known that the Brahmana literature is full of classifications of all sorts of things, with analogies and equivalences drawn based on such classifications. It was often assumed that there was no pattern or logic to this. However, this view has been called into question recently. For example, has argued that the equivalences are systematic in the Kausitaki Brahmana.

Recently, Brian K, Smith has written a book [BC], claiming not only that there is a system behind the classification, but that the system is a heirarchy, based on the Varna scheme, and is common to all Vedic literature. In fact, Smith goes further and claims that one of the purposes, if not the principal one, behind vedic ritual and speculations is the justification of the Varna heirarchy, which is further qualified as fixed and hereditary. Given the remarks in the concluding chapter of this book and their potential use in the current political situation of India, it is desirable that Smith's arguments be examined in detail.

It will be helpful to spell out some of the possible positions on the pattern of analogies:

(1) The classifications and analogies are based only on the need to justify the rituals. There is no other logic. This is the classical position of the western indologists.

(2) The classifications and equivalences are based on a system. The ritual substitutions follow from this, rather than the other way around.

The latter can be split further into a "soft" position and a "hard" one:
(2h) The classifications are always based on the varna scheme. The varna assignments are (mostly) fixed and obviously follow the trifunctional scheme.

(2s) The classifications take into account the trifunctional scheme. However, justification of ritual takes precedence, and varna scheme enters only when it does not conflict with the need to explain the ritual. Furthermore, there are differences between sakhas, and these can be attributed to the differences in ritual between the sakhas.

Smith's position is (2h). In more detail, the claims are these: The Brahmin authors of the Vedic texts presented the whole universe, and any given category in it, as divided into a varna-like tripartite scheme. This was often done "surreptiously", without mentioning varna explicitly, but keeping it thinly veiled. The main conclusion drawn from this is that a rigid, hereditary heirarchial varna scheme is inextricably linked to Vedic religion. (We will discuss the supposed connection to latter jatis near the end.)

"Classifying the Universe" is meant as a 'proof' of these claims. As such, it is filled with quotations from Brahmanas which posit 'bandhus' or connections between different concepts or entities. These are organized by categories: Gods, space, time, flora, fauna and revelation. Each chapter is filled with an impressive number of quotations which, on a first reading, seem to support Smith's claim. Yet, detailed analysis raises serious questions, especially if one wishes to assert,as the blurb on the back cover does, that Smith has proved (2h).

Has the author succeeded in disproving (1)? Based on Smith's arguments, those presented in this book as well as elsewhere, the answer is no. To satisfy the skeptic, one would need an exhaustive discussion of all classificatory passages. While the number of quotations in this book is large, it is not exhaustive, or is that claimed. So it is not possible to do a quantative and qualitative analysis of the unexplained residue of classificatory passages, or even to be sure that the 'residue' is smaller than what is taken into account.

Adding to this problem is the evaluation of the quantative information: If we have ten quotations from the Satapata Brahmana, on each from Aitreya, Kausitaki, Pancavimsa, Jaimaniya, and the Brahmana portions of Taittiriya and Kathaka, how do we count it? As sixteen, or as seven, counting each brahmana once, or as four or as three, based on some theory of "family resemblences"?

This is not an idle question. Let us take the thirteen cosmogonies listed in Chapter 3 (all references are to [BC] unless otherwise specified). Four are based on quotations from Satapata Brahmana; another such is said to have a variant in Aitreya Brahmana, but the equivalences in the latter are much more limited. Yet another quotation from Satapatha has variants from Maitrayani, Kathaka and Taittiriya Samhitas and Taittiriya Brahmana. But Maitrayani and Kathaka (which in many respects are more closely related to each other than to Taittiriya) variants differ in a cruical respect from others (author's suggestions for emendations to fit his theories cannot be justified unless it can be shown that there are few other objections to (2h)). Two are based on quotations from Taittiriya Samhita and one from Maitrayani Samhita, with a variant from TS. Others are one each from Pancavimsa, Jaimaniya and Kausitaki Brahmanas and Maitrayani Upanishad.

As already noted, Smith's arguments require that all the vedic texts, taken as a whole and neglecting sakha distinctions, are consistent with each other. It is hard to believe that this is the case, and Smith does not do his job here. See for example the comments below on the quote from KB that Smith lists as Cosmogony XIII.

Given the impression prevelant among Vedists that Satapatha, the latest of all Brahmanas, is given to systematizing and combining the various, not-so-reconcilable views of its predecessors, prepondenance of quotes from that source is likely to to lead lingering doubts. Add to this the fact that Smith does not hesitate to quote Manava Dharmasastra to support his views, while quotes from RkSamhita are rare, and the doubts only grow.

To see how well, rather one should say how poorly, RkSamhita corrobarates Smith's claims, let us consider the Adityas, Agni and the Asvins.

According to Smith, the Adityas are typically Vaisya deities. Yet, when Adityas are named in the RkSamhita, they include such major deities as Varuna, Mitra and, in two cases, Indra. Smith quotes [MV, p.45] to say that Adityas are asked to bestow such "third varna boons" as long life and offspring. Yet Macdonell also points out that the Adityas are asked to forgive sin; they are said to be punishers of sin, kings, kshatriyas, pure and holy (rtavan). (These facts have led Dumezil to try to explain all Adityas as 'soverigns', belonging to his 'first function'.) Smith ignores all of this.

Next let us consider Agni. RS 1.26.1, 1.58.8, 1.96.3, 2.6.2 etc. refer to Agni as urjam pate, urjo napat, urjah putra and so on. Smith never once refers to this strong connection between Agni (brahmin as per Smith) and urj (vis according to Smith). Agni, like almost all gods in the RkSamhita, is asked to give multitude of sons. This too, is ignored, although the same trait is said to connect Adityas to vis.

There are just two places where Asvins are referred to. At one (p.286, n.97), Asvins are connected to the ass, a 'lower class animal'. At the other (p.115, n.53) it is suggested that Asvins are brahman, on the basis of RS 1.36.1, 3 and association with tejas and brahmavarcas in the brahmanas. Yet, in the RkSamhita, Asvins are known as physicians (the trait that is the most probable cause for their low status in Satapatha and Mahabharata), rescuers from the sea and givers of wife/husband. They are, like Adityas and Agni, prayed to for abundant progeny and long life (recall that Asvins' rejunavation of Cyavana is a recurring motif). Furthermore, in RS 1.92.17, 1.116.8, 1.118.7, 1.157.4 etc., Asvins are arked to bring urj or to give urj. So are Asvins lower class deities said to possess tejas and brahmavarcas, or are they brahmins with numerous vaisya qualities?

Of course, there are other things in the RkSamhita that go against Smith's theory: For example 10.181.2d says that "Bharadvaja brought Brhat from Agni". (10.181.1cd and 10.181.2cd also refer to Dhatr, Savitr and Visnu, connecting them to both Rathantara and Brhat.) As Bharadvaja means 'bearer of vaja', we have a single verse connecting Agni (brahman), Brhat saman (kshatra) and vaja (vis).

In many cases, etymological connections go against Smith's theory: Brhat means 'great' and comes from 'brh', the root also of brahman. Vasu means 'rich' or 'bountiful', while Aditya means unbounded, unrestricted.

Gonda has noted that the Vedas list a large number of triads which do not fit into a trifunctional pattern (see [GT], especially pp.137--143). Smith makes no reference to the exitence of such triads, their frequency, and their relevance to the question. If a trifunctional ideology was the only source of triads, we would expect such triads to be few. But not only are such triads frequent, but there also a few that directly contradict Smith's theory: SB 12.8.2.32 that associates Asvins to earth, Sarasvati to atmosphere and Indra to heaven; SB 12.9.1.7 that associates those three to the part of the body below the navel, between the navel and the head and the head respectively; (it should be noted that TB 3.6.13.1 connects Sarasvati with Rudras and atmosphere); SB 12.9.1.8 associates Indra, Savitr and Varuna to early, middle and last part of one's life; and TS 7.2.8.1 where gayatri is associated to word, trishtubh to food and jagati to sacrifical substance. What are we make of these texts which contradict the varna associations Smith imputes to the Vedas, but which are he ignores even while claiming to be careful to note all exceptions?

There is also the disquieting feeling that the quotations given are selective, rather than complete or random (whatever that may mean). Let us look at one example. According to Smith, the primary triadic classification of the 'worlds' is earth (bhu) = brahman, intermediate space (bhur) = kshatra and sky (svar) = vis. In order to make this conform to the usual hierarchy, it is emphasized that svar is distinct from svarga loka, and that the term 'yonder world' refers to svar and not to svarga, and that former occurs only when referring to material worlds and the latter in sorteriological parts. We are referred to [GL, p.91] for this.

Upon reading Gonda's monograph, one will come across many instances which contradict the inequality svar != svarga loka, taken with the associations to svar that are proposed: AB 3.46(7) equates vamadevya saman with yajamanaloka, amrtaloka and svargaloka [GL, p.49]; Smith classifies vamadevya as vaisya chant (p.301). KB 20.1 states that the loka of Varuna is adhidivam, and Mrtyuloka is pradivam [GL, p.56]; how does one fit this with the classification of Varuna? Gonda's discussion on p.80f points out the complex relation between 'yonder world' and svargaloka. SB 13.2.4.1 equates asau (loka) and devaloka [GL, p.84] (Smith notes that SB places the gods in the sky, but does not discuss it further). AA 2.6.6, AB 8.14.4 and 8.19.2 refer to 'amushmin svarge loke' [GL, p.88 and p.97]. PB 18.7.12 equates svarga loka with vaja [GL, p.97]; but, Smith equates vaja with vaisya. And there are those enigmatic verses from RkSamhita, 9.113.6--9, which refer to heaven as 'yasmil loke svar hitam'.

One would not find out about these "denying instances" (by which I mean the antonym of scientists' "confirming instances"), to Smith theory on vaja or vamadevya saman, by looking at the index of his book.

Smith quotes TB 1.1.4.8 (p.98) to connect the Rbhus to 'lower classes'; yet the same text connects Bramins to Adityas. Smith does not take this fact into account (at least in those places where, according to the index, the Adityas are referred to).

I believe that somewhere the author mentions Sankhyana Aranyaka 8.8 which associates vamadevya saman with atmosphere and brhat saman with sky. But this is missing from the index (and I failed to note down the page). The fact that Jaimaniya Brahmana exalts the brhat saman (kshatriya as per Smith) is noted (p.312, n.59); but this reference to brhat is missing from the index.

The associations listed on pp.291--292 ignore Aitreya Aranyaka 2.2.4 which connect consonants to body, sibilants to breath and vowels to Atman.

Another example along these lines is the fact that Aitreya Brahmana 8.17, as far I can see, is not referred to anywhere. The second triad of connections made here are as follows:

  • Visve devas - anushtub - 21 verse stoma - vairaja - vairajya
  • Maruts/Angirases - atichandas - 33 verse stoma - raivata - parameshtya
  • Sadhyas/Aptyas - Pankta - 27 verse stoma - Sakvara - (misc)
The first column is gods, second is meter, third is stoma, fourth is saman and the last is the purpose. This clearly the parallel passage to KB 22.1 etc., on which Smith's Cosmogony XIII is based, except for the fact that two brahmanas don't agree in the last two rows.

The associations gold-sky, atmosphere-silver and earth-iron, made mostly in the Tripura story, is interpreted to mean that the association is sky-brahmin etc (p.135). The texts mentioned do not make any further associations. Smith is putting in what he wants to read, rather than reading what is there. Of course, AVS 5.28.5, which makes the associations earth-yellow one (gold), Agni-iron and plants-silver is not mentioned in this context.

We are repeated told that the new year began in spring. Yet, Mahavrata, the 'mukha' of the year, is on winter equinox. There are also other references that support this view. The Agrahayana festival, whose name means 'of beginning of the year', takes place on Margasirsa full moon; this places it at the end of autumn (November, perhaps early December) for the period 800 BCE to 400 BCE. Smith does not take notice of these two facts (even though they might explain some of the unusual associations of AB).

We are also told that there is little evidence for association of asvattha tree to brahmanas in Vedic texts, and that it is usually connected to vaishyas and occasionally to kshatriyas (p.237, n.55). Yet the famous hymn RV 1.164 presents the world tree as an asvattha. The inverted asvattha and soma-bearing asvattha are well-known upanishadic motifs. We are also informed that the reason that asvattha is known as 'devasadana' is that Maruts are said to be seated there (p.225). The texts listed for this are RV 1.164.20-22, AV 5.4.3, RV 5.54.12, TS 4.2.6.2, TS 7.2.1.3, VS 35.4 and SB 13.8.3.1. Now RV 1.164.21 presents the asvattha as the seat of 'suparnas', and not Maruts. In RV 5.54.12, the Maruts are said to be shaking down the tree; we are not told if they are sitting in the tree or standing on the ground; it is not in the least clear why the former is more plausible than the latter. AV 5.4.3 does not refer to Maruts in particular in reference to 'asvattha devasadanas'; this is traceable to the idea of the world tree being an asvattha. TS 4.2.6.2 and VS 35.4 are variants of the same mantra, as are RV 10.97, KS 16.13 and MS 2.7.13; SB 13.8.3.1 is the brahmana on VS 35.4. This mantra is a hymn to the herbs ('oshadhis'). RgSamhita version says 'asvatthe vo nishadanam, parne vo vasatish krta' and other versions are almost the same. One wonders why the reference to parna tree has been supressed from the quotation. In any case, none of the variant the three cited refer to Maruts in particular. Thus, it turns out that Maruts sitting in the asvattha occurs in only in the SB.

The quotation from AB 3.27 (p.297) fails to note the context, or refer to all the parallel versions. The story actually begins at AB 3.25, where it is said that the meters Gayatri, Trishtubh and Jagati all had four syllables originally. Smith ignores this, and instead starts the next quote at 3.28, and, puts "originally" in brackets, leading the reader to assume that at the start, the number of syllables were 8, 3 and 1 respectively. While the versions at SB 4.3.2.7f and JB 1.287ff, are noted, the fact that the latter calls anushtubh the mother of Gayatri is ignored. The versions at TS 6.1.6 and KS 23.10 say that the original number of syllables were 4, 13 and 14 respectively; according to the the latter, the meters were the offspring of Kudru. Smith does not even mention either of these two.

Most serious along these lines is what Smith makes of Aitreya Brahmana 6.4 on pp.195--196. Part of the text is quoted, suggesting the associations morning-Agni-east and afternoon-Visve Devas-west. An extrapolation, noon-Kshatriya deities-south, is suggested. Now AB 6.4 has the following associations as well: South-Mitra and Varuna-morning, Center-Indra-morning, North-Indra and Agni-morning; these are followed by the two quoted associations. What is the need to extrapolate (except the desire to justify one's pet theory at all costs)?

Actually, it is worse. On p.103, AB 6.4 is quoted as saying that both Indra and Agni act as warriors by protecting the north against demons. Yet when Agni does the same, he is brahmana, not kshatriya. On p.167, n.98, AB 6.4, positioning Mitra-Varuna in the south is dismissed as a "typical myth about the attack of the demons". Apparently, the value of AB 6.4 varies depending on how well it agrees with Smith's theories.

Given that the above instances were based on limited knowledge of original texts and a few secondary sources (mostly [GL] and [GT]), one gets the feeling that the number of exceptions not noted by Smith must be quite large. This does not inspire confidence in his arguments.

The extreme to which Smith goes is demostrated on pp.30--31: On p.31, we are told that Sudra, when he appears, is soley depicted in terms of service and dependcy on others. Yet, the preceding quotation (SB 13.6.2.10) says "Sudra is tapas". This 'anomoly' passes without comment. We see this pattern repeated time and again.

Smith buttresses the association of vaja to vaisyas by referring to Gonda who insists that vaja is a 'generative power'. This has not been accepted by any of the lexiographers. It should be noted that Gonda's argument in [GV] are based on certain assumptions which run counter to Smith's theory: According to Gonda, Indra is a god of fertilization (see also [GO, p.]). Vajra is connected to fertility because it is mostly mentioned in the RkSamhita in connection with Indra winning cows, waters and the Sun. Yet, if Indra represents the power of fertilization, trifunctional ideology falls on its face.

Now, as Indra's main feat in the RkSamhita is the defeat of Vrtra and the winning of cows, waters and the Sun, it is not surprising that mention of vajra is often connected to those three. Vaja can be translated as vigor or vitality in all the places that Gonda mentions. Of course, it would include the power to fertilize, to grow etc., but would not be limited to that. That food is vigor can be explained by the obvious need of food to be vigorous and active. But not much more can be read into that.

It should be noted that Gonda also refers to ojas, a kshatriya power according to Smith, as "creative power" and connects it to organic growth [GO, p.], [GT p.142]. Gonda's arguments for this are similar to those for his interpretation of vaja. Smith gives no reasons to explain why he accepts only part of the argument.

This brings up another instance of suppressed evidence: Smith notes that Pushan is called vajanam pati in RS 10.26.7 (p.99). That term is used seven times in RS, five times refering to Indra (1.11.1, 1.29.2, 6.45.10, 8.24.18, 8.92.30), once to Indu (Soma) (9.31.2) and once to Pushan. In addition, the term vajasya pati is used once of Indra (10.23.3) and once (1.145.1) appears in a hymn to Agni; the verse itself has no names, but terms used strongly suggest Agni. Agni is often referred to in terms that evoke the same idea: RS 2.1.10, vajasya ... isishe etc. So are Indra and Agni vaisya in the RkSamhita?

Given these problems, it is a moot point as to whether Smith has succeded in disproving (1). Asking if he has proved (2h) rater than (2s) opens a much bigger can of worms.

A charge that has been levelled against Dumezil can also be made against Smith: This is that varna scheme is teased out of any triadic partition, on whatever basis will serve that purpose for that situation. For example, if horse is brahman due to its association to Prajapati, the origin of everything in the universe (as brahman is said to be for other two varnas), and horse if ksatra due to its association with chariots and war, and also lower class due to it being produced from the feet of Prajapati and acting as a bearer, does tripartition mean anything? This is another thing that has to be kept in mind as we analyze the vedic texts. If this charge sticks, then (1) has not been disproved. If the charge is not completely refuted, we have to content with (2s) rather than (2h). If horse can switch varnas, why not a human?

Smith does not hesitate to combine associations made in different texts. The trouble is that only combinations that are made are those which support (2h). But, if we make other combinations, we do get other conclusions. It is not clear why we should not do so. To avoid combinations simply because they run counter to (2h) does nothing to prove that (2h) is correct.

A partial list of such examples is given below. All the equivalences and associations are what Smith imputes to the Vedas. The purpose of this list is to question the claim that Vedas support fixed, heirarchial varna scheme, made for and by Brahmins, and not a fluid, non-heirarchial classification scheme meant mainly to explain and justify the rituals.

In Table 3.2, horse is equated to Sudra. Yet the association of horse to royalty is undeniable in the earlier texts.

In Table 3.3, north is equated to Sudra, and to Mitra-Varuna or to Soma. Neither deity is lower class in RV nor are they in bulk of later texts.

In Tables 3.2 and 3.3, anushtubh, 21-verse hymn and autumn are equated to lower classes (Sudra especially). But when we move to the six-part classification displayed in Table 3.4, the same list is said to correspond to brahman, because the associated deity is Brhaspati.

According to Cosmogony IV, sky is associated to fame (yasas). Is fame kshatriya or vis? The author opts for the former (as does Hiltebitel, quite convincingly, [HR]) on p.38 and elsewhere; but switches, without any comment, to yasas=vaishya when that suits him (p.64 where yasas is the 'elemental quality of vis', and p.292).

We are presented with the equations saman=svar=vaisya (pp130--131). But samans were quite exalted, from vedic times down to the Bhagavadgita. This fact is noted, but Smith does not explain why this does not contradict his theory.

The equation sky=vaisya rightly troubles the author as contradicting the hierarchy. One explation that he gives is that sky displays multitude in the form of stars and solar rays. Yet the passage he quotes on p.133 also says that earth displays multiplicity in the form of plants and atmosphere in the form of birds. Earlier (p.127), Satapata was quoted, placing solar rays in the atmosphere. But on p.133, the author ignores this, and instead quotes Chandogya Upanishad placing light rays in the sky. Again the discrepancy is not noted.

What of the chain of equations earth = rk = king = woman, sky = saman = purohita = man? The author relegates this to an end-note (p.159, n.32). Incidentally, this is not listed in the index either under Sama Veda or under Rg Veda.

Combining the association of cardinal directions to varnas and of locations to varnas we end up with west = earth = brahman = east = sky = vis!

Savitr is vis on p.119 (n.11), ksatra on p.90 and p.95, and brahman on p.154. We may escape this impasse by positing that the ideal king was expected to have trifunctional virtues (as do Dumezil, Hiltebitel etc.). But Smith's insistance on kingship = kshatra < brahman would not allow us to do so.

On p.155 we find out that anushtubh is "all encompassing". Equally weighty evidence was presented earlier to suggest that anushtubh is connected to Sudras.

The quintessentially braman god Brhaspati turns out to be connected to afternoon and autumn, both of which, we have been told, are vaisya.

Sudra, associated to Prajapati's feet, his "firm foundation", is considered lower class for that reason. Yet, when braman is associated to earth, the "firm foundation" of the universe, that is said to be a positive attribute (p.130).

It is said that the vedic texts exalted brahmans as "like mothers" to other varnas. Yet the passage quoted equates as follows: brahman = rk = mother, kshatra = yajur = father, vis = saman = child. This would make brahman the wife of kshatra, not mother.

On p.187, earth = east and sky = west. On pp.141--142, east = up = world of gods or svargaloka and on pp.144--145, west = world of snakes. Why should we not combine these associations?

In Table 7.1, bilva is connected to prosperity, offspring and animals, same features that make Adityas into vaishyas. But bilva turns out to be brahmana varna, according to Smith (Table 7.2).

Smith does not discuss in detail, in Ch.5, the "problematic case" of matching the varnas and directions in pendatic and six-part schemes, preferring to leave it aside (p.160, n.36). But the same reluctance was missing in Ch.3

We are informed that Agnidhra and Subrahmanya priests are connected to Brahmins, via either name or goat (p.281, n.43). Agnidhra connected to spring is, presumably, also tied to this (p.203, n.49). Yet, Agnidhra may originally been an assistant to the Adhvaryu (connected to YajurVeda) and in later texts was associated to the Brahman priest and thus to AtharvaVeda; recall that the Brahman priest is connected to autumn (p.203, n.49). Furthermore, the Agnidhriya fire is connected to bhuvar and YajurVeda (p.63). The Subrahmanya priest is an assistant of Udgatr.

Garhapatya is connected to brahman varna, according to Smith. Yet, in his earlier book [SR], garhapatya was described as "weak" version of ahavaniya, and simple extension of the "feminine" grhya fire. If both are correct, how can Smith claim that the vedas glorify the brahman varna?

While Smith does not hesitate to quote Manava Dharma Sastra to support his arguments, he fails to note that both it and the Mahabharata ascribe tejas to kings.

On p.238, n.68, only AB 7.32 and AsvGS 1.19.13 are said to connect the udumbara tree to Ksatriyas. Yet, on the next page, it is admitted that MS 4.4.2, TB 1.7.8.7, ApSS 18.16.3, 5, KSS 15.5.30 explicitly connect udumbara to Kshatriyas.

It is also not clear why the use of asvattha as the churner to produce fire is considered associating it to vaisyas, while the use of the acacia (sami) as the bottom part does not. Is it because acacia is strongly associated to Agni, the "brahmin god" (cf p.224)?

It is possible to continue this list further. But this much should give an idea of why (1) was dogma for so long, and how much needs to be explained before we can accept (2s), let alone (2h).

There are other simplifications as well. For example, Smith consistently replaces rajanya by kshatriya. While Smith claims to have been careful in noting all exceptions and differences, he fails to note that AB 7.20 distinguishes rajanya and ksatriya.

There are instances of anachronistic translations: In the quotation of SB 6.4.4.3, purisha is translated as 'excrement'. In the vedas, purisha has only either the generlized meaning of 'fill' or specialized meaning of 'soil', 'clay'. In particular Agneh purishavahana cannot be disassociated from SB 6.3.3.3,4 which refers to the sacrificial procession going to fetch Agni purishya, 'Agni in the soil'. It is hard to escape the impression that the mistranslation is solely due to Smith's attempt to impute certain value judgement to that passage.

One thing that is not clear is what texts are included by Smith under the rubric of 'Veda'. We have already noted that Manava Dharma Sastra is given more prominence than the RkSamhita. We need to know the answer, never given explicitely, to this question if we are to discuss the assertion that "the Veda represented itslef---and not represented later---as absolutely authoritative ...". As far as I know, RkSamhita does not represent itself as authoritative. The Brahmanas usually treat only the samhita as revelation.

Smith sometimes fails to answer the objections that have been raised to his assumptions, in the very references that he uses: For example, he disputes the connection between 'vis' and 'every', ignoring the detailed arguments in [GT, pp.137--140]. Nor does he note the often repeated fact that men of all varnas took part in war (see, for example, [GT, pp.142--143]), and even the dharmasastras give fighting as the first alternative livelihood for brahmins who are unable to live by teaching, sacrificing and receiving gifts.

Another problem is that the author fails to prove his claim that the varnas in the Vedas were seperate and hereditary. Instead, he assumes from the start that brahman, ksatra and vis are near synomyms for brahmana, ksatriya and vaisya. But that merely begs the question. What needs to be proven is that the varnas were endogamous, that one who once acted as a warrior or king could not become a priest and so on. Hypergamy is common in brahmanas and even in the Mahabharata. As the authors of the brahmanas never talk about who married their daughters, we do not know if hypogamy was present or not. To make such assumptions without answering all the questions is not the right way to proceed.

It should be noted that (2h) is neccessary for the conclusions that Smith draws in the concluding chapter. By itself, (2s) is insufficeint. If the horse can be brahmin, kshatriya or 'lower class' depending on context, that cannot be used to claim that varnas should be hereditary and fixed for ever.

Smith notes that an opposing view has been presented, in a persuavise fashion, by Heesterman. Yet, the examples given by him are not discussed here, nor are we given any references that refute Heesterman. The author's review of Heesterman's essays also makes only similar dogmatic assertions, without any supporting evidence. (Heesterman has made a good case that the diksita, at certain point leads a raid (or a mock raid) and after his return, begins to act like a brahmana. As the various sacrifices are said to follow each other, such changes in role would be common, till the sacrifical spiral is ended. This possibility is not considered either. It may not be out of place to note that the 'feasts of merit/confirmation' among the 'kafirs' of Hindukush are rather similar to the picture presented by Heesterman. These feasts are in two parallel sequences, one for warriors and another for those aspiring to priestly purity. Those engaged in the latter sequence may still take part in raids ([JR, pp.27--28, p.88], [RK, p.140] on the Uta]).

One also gets the feeling that the conclusions drawn by Smith depend on certain unspoken assumptions. In particular, he seems to assume that vedic religion originally dealt with deified powers of nature and that the rituals were developed later. How else are we interpret the "sic" after 'supernatural' on p.210, or the comments on rituals in chapter 10? This is not the place to discuss the objections to 'nature religion' (see [CQ], ch.4), or beat the dead horse of Max Mueller's theory on the origin and growth of rituals.

At this point, we must take time consider an earlier publication of Smith. In [SR], he analysed the use of resemblences in vedic ritual theory and practice. There the claims on the origins of rituals are made more explicitely. Since these claims seem to underlie some Smith's concluding remarks, let us study them in some detail.

In [SR], the following comments are made:

(1). It is, at any rate, just as possible to posit an origin within vedic ritualism for the domestic sacrifice as it is to imagine it "absorbed" by vedic ritualism; but the question reamins unresolved. (p.144)

(2). The 'feminization' of domestic sacrifice ... had as one of its purposes to denigrate it in comparison to the srauta ritual. (p.155)

(3). pakayajna was conceived and represented as a diminutive, condensed form its larger model, the srauta ritual. (p.160).

(4). Jan Gonda demostrated that upanayana is both conceived as and modelled on the initiation and consecration of the sacrificer. (p.207)

The use of the word 'conceived', and the talk of 'origins' are at best misguided and at worst misleading. Gonda, as others before him, has not failed to note that the vedic 'rites of passage' are similar to those of other Indo-European cultures: See the general remarks in [GR, pp.546--547] and about upanayana in particular in [GC, p. 364]. While Gonda talks about the similarities between upanayana and initiation/consecration rituals, he wisely refrains from talking about their 'conception'. We should follow his example.

Smith attempts to butteress his arguments with the following quote from [GR, p.549]:

There is no denying that the domestic cult as described in the grhya manuals was to a considerable extent influenced by and even modelled on, the srauta rites which in some cases run parallel."
(see [SR, p.170]). However Smith fails to note that Gonda goes on to say
Since, however, fundemental features of the the solemn rites must have been based on or borrowed from an ancient simple and 'general' or undifferentiated ritual, it is very difficult to make certain about the question as to how far those who first compiled grhya manuals codified the ancient and genuine rites and traditions of their communities and how far they tried to adapt part of the rites to principles and prescripts prevailing in the srauta ceremonies, how far they elaborated the domestic rites hither to performed in simple unadorned fashion by means of the abundant liturgical material of the solemn sacrifices.
In particular, we do not know if the srauta rituals of agnihotra, darsapurnamasa etc. are elaborated forms of the (earlier forms of the) domestic versions, or if the domestic versions are simplifications of the srauta rituals 'conceived' by priests.

There is a different explanation of the state of affairs in the ritual manuals: The rituals can be divided into three groups: (i) 'rites of passage' such as upanayana, marriage and funerals; (ii) acts of worship obligatory to all 'twice born' such as agnihotra and darsapurnamasa; and (iii )elaborate sacrifices that double as 'status displays', such as vajapeya, rajasuya and asvamedha.

Group (i) have parallels in other cultures, especially IE ones. While some of the mantras used in them may be new, taken from the samhitas, the meaning can be quite old; in effect, the phrases used have changed, influenced by te Samhitas. Group (ii) has parallels in other IE cultures as well, although not as striking and the materials used change quite a bit. For example, simple animal sacrifices made as prayers, and with the help of one or two priests are attested among Zorostrians. Regular offerings into a permenantly kept fires are known, though they may have been restrcted to a whole group rather than be required for each family seperately. The Agrahayani rituals (together with a part of the Mahavrata) have parallels in the new year festival of the Kalash Kafirs. Bali offerings have parallels in Germany, for example.

One may claim, with more plausibility, that the srauta versions developed from the domestic versions. To claim that the domestic versions were 'conceived' as simplified versions of their srauta counterparts, while ignoring parallels from other IE cultures, especially the Iranians, is absurd. Existence of parallels suggests that the hypothesis of seperate evolution from 'undifferentiated ritual' is more likely. That is to say, there existed rituals corresponding to Group (ii) that evolved into the grhya versions, while the srauta versions were developed as more ostentatious versions of the same.

Smith claims that the ekagni is put on par with the wife of an ahitagni. This is based on the statement in some kalpa sutras that the wife of an ahitagni is told to tend to the domestic fire, which will then be used to cremate her at her death. Manava Srauta Sutra explicitly notes that one famous authority, Sakalya, contradicts this. In addition, there is no evidence that the widow of an ahitagni could do the rites that all heads of families, whether ahitagni or ekagni, were required to do, namely the rites of Group (ii). The widow seems to be simply preserving a fire to be used for her eventual funeral. In effect, widows of srauta sacrificers have a privilege that widows of ekagni lack, but do not carry out the duties that head of a family is required to, using a single fire.

Strangely enough, Group (iii) may not be new either: Some similarities between the asvamedha on one hand and October Equus and an annual (northern) Irish rite on the other have been pointed out several times ([DA], [Pu]). Parts of the rajasuya have parallels in other IE royal rituals ([HA]).

Smith seems to read too much in to the fact that the grhya manuals refer to srauta manuals. Srauta sutras would have come into existence earlier for the simple reason that srauta rituals were more complicated and rarer, and consequently, future priests would be less likely to learn them by observation. A need for codification of the grhya rituals was not felt till the srauta sutras became common. The reason the grhya sutras refer to srauta sutras is keep the former as short as possible, not because the grhya rituals were 'conceived' as 'simplifications' of the srauta rituals.

One conclusion to be drawn from the long digression above is that the various rituals were evolving in the vedic period, and the sutras are simply the codifications of the rituals as they existed in the autors' milieu. The speculations of the brahmanas are attempts to explain the ritual practices of the time, rather than attempts to justify the rituals as they were 'conceived'. We cannot assume, as Smith apparently does, that theory preceeded practice.

One more topic must be addressed, namely the relationship between varna and jati. Smith's position seems to be that varna system was the 'origin' of the jatis. In the opening chapter, where such views as asserted, we are promised an in-depth discussion in the concluding chapter. But, there we find nothing but a summary of various views, with no attempt at justifying Smith's position. In the meantime, Rathakara is glossed as 'offspring of vaisya father and sudra mother', even in the index! This ignores the fact that in the Yajnavalkya Smrti, a more complicated origin is posited for Rathakaras, and the quoted view is that of Baudhayana Dharma Sutra. We are never told why one is preferrable to the other. While [MR] is quoted, Smith ignores the former's conclusion that Rathakaras do not form an endogamous group in the earliest texts. The much-questioned attempts of Dharma Sutras to derive jatis out of varnas is simply assumed by Smith, yet this is exactly what he promises to defend eventually. (The names of jatis in the dharma sutras, some of which, like Rathakara or Karmara, are occupational, some such as Vaideha are names of regions, and yet others seem to be names of tribes, suggest manifold origins of the castes.) Without any such defence, it is not possible to claim that "caste is indeed Vedic in its origins" (p.320).

It should also be pointed out that Dumont's views on caste were endorsed woleheartedly in [SR], and are used implicitely in the concluding chapter. But, while objections to Dumont are recognized, they are not replied to. Dumont's theory has been criticised from anthropological viewpoint often enough and there is no need to rehash them here (see [QI], for example). Even from textual data, Dumont makes quite a few mistakes. For example, the duties and privileges of the king in relation to dharma (see, [GA] for an overview) do not support Dumont's claims about the subordination of kings. The 'purity-impurity axis' of Dumont is not without its weaknesses.

Let us now return to 'Classifying the Universe'. In the concluding chapter, we are advised that scholars of religion should ask "says who?", that is to say, "to whose advantage?". One must wonder if the same question should be asked of scholars of religion as well. Smith leaves no doubt that he believes that brahmanas were written by brahmins to reinforce their position and authority, and that the equivalences he lists are proof of that. Are we entitled to ask "To whose advantage is it if Smith's theory is accepted, without detailed examination of all the evidence?" (rther than selected quotes). When combined with incomplete quotes and strange lacunae in the index, this question could indeed be quite uncomfortable.

Seen from that angle, certain phrases which might seem as harmless conceits take on a more sinister tone. For example, Maruts are referred to as "storm troopers" (p.97 etc). Given the association that term has today, is its use deliberate? The same question arises when we see 'paroksha' translated as "surreptious(ly)" on pp.222--223 whereas most (including Smith, elsewhere in this book) use "secretly" (literaly, it is 'paras+aksha', "beyond the eye"). The anachronistic translation of purisha has already been noted.

In conclusion, we can only say that Smith as failed to prove his case, that the evidence he presents is selective, that he has ignored conflicting evidence even when he should be aware of it and that he has let his prejudices overrule his reasoning ability. We can only hope that non-specialists will become aware of the serious problems with this book and that they will not be misled by the overwhelming number of quotations into thinking that the evidence presented is exhaustive rather than being carefully selected.

References Quoted:

  • [BC] B. K. Smith, "Classifying the Universe", Oxford University Press, London, New York, 1994.
  • [BR] B. K. Smith, "Reflections on resemblence, ritual and religion",
  • [CQ] Chase, "Quest for myth", Lousiana State U Press, Baton Rouge, 1949.
  • [DA] P-E. Dumont, "L'Asvamedha", P. Guenther, Paris, 1927.
  • [GA] J. Gonda, "Aspects of early Visnuism", Oosthoek, Urtrecht, 1954. (Reprinted by Motilal Banarasidas, Delhi, 1969).
  • [GC] J. Gonda, "Change and continuity in Indian religion", Munshiram Manoharlal, New Delhi, 1985 (Ist Indian edition). (originally published by Mouton 1965, Disputationes Rheno-Trajectinae 9)
  • [GK] J. Gonda, 'Ancient Indian kingship from the religious point of view", E. J. Brill, Leiden 1969 (reprinted from Numen 3, 4).
  • [GL] J. Gonda, "Loka. World and heaven in the Veda", North Holland, Amsterdam, 1966/1967. Verhandle,ingen der Koninklijke Nederlandse Akademie van Wetenschappen, Afd. Letterkunde, Niewe reeks, d.73, n.1)
  • [GO] J. Gonda, 'Ancient Indian ojas, Latin *augos and the Indo-European nouns in -es-/-os", A. Oosthoek, 1952.
  • [GR] J. Gonda, "Ritual sutras", Harrassowitz, Wiesbaden, 1977. (History of Indian Literature)
  • [GT] J. Gonda, "Triads in the Veda", North Holland, Amsterdam, 1976. Verhandle,ingen der Koninklijke Nederlandse Akademie van Wetenschappen, Afd. Letterkunde, Niewe reeks, d.73, n.1)
  • [HA] J. H. C. Heesterman, "Ancient Indian royal consecration", Mouton, The Hauge, 1957 (Disputationes Rheno-Trajectinae, 2).
  • [HR] A. Hiltebeitel, "Ritual of battle: Krishna in the Mahabharata", Cornell University Press, Ithaca, NY, 1976.
  • [JR] K. Jettmar, "Religions of the Hindu Kush", (translated by Adam Nayyar) Aris and Phillips, Warminster, Wiltshire, England, 1986.
  • [MI] K. Mylius, Die vedischen Identificationen am Beispiel des Kausitaki-Brahmana, "Klio", 58(1976) 145--66. (see also Altorientalische Forschungen 5(1977) 237--244)
  • [MR] Minkowski, Rathakara's eligiblity to sacrifice, Indo-Iranian Journal, 32(1989) 177--194.
  • [MV] A. A. Macdonnel, "Vedic mythology", Motilal Banarasidas, Delhi, 1974 (Originally published in 1898 at Strassburg).
  • [Pu], J. Puhvel
  • [QI] D. Quigley, "Interpretations of caste", Clarendon Press, Oxford, New York, 1993.
  • [RK] Robertson, "Kafirs of the Hindu Kush", Johnson Reprint Corp, New York, 1970. (reprint of 1896 edition by Lawrence and Bullen at London; (another reprint by Oxford University Press, Karachi, New York, 1974; there is another edition dated 1900).