Dharmaskandhah and Brahmasamsthah: a study of Chandogya Upanisad 2.23.1.

• Author: Olivelle, Patrick

I

1.1 Chandogya Upanisad 2.23.1 has been singled out for special attention by Indian theologians and exegetes, as well as by modern scholars, because they consider it to be the vedic basis for and the earliest evidence of the central Brahmanical institution of the four asramas. I give here the Sanskrit text together with Hume's English translation,(1) marking each phrase with a Roman numeral to facilitate reference to each:

A.i. There are three branches of duty.

ii. Sacrifice, study of the Vedas, and alms-giving - that is the first.

iii. Austerity, indeed, is the second.

iv. A student of sacred knowledge dwelling in the house of a teacher, settling himself permanently in the house of a teacher, is the third.

v. All these become possessors of meritorious worlds.

2. vi. He who stands firm in Brahma attains immortality.

   1.2 Arguing against the position of the "Jaiminiyas," that is, the followers of the Mimamsa tradition, Samkara (in his commentary on the Vedanta Sutras [VeS 3.4.17-20]) cites this passage as the vedic basis for the asrama system and for the legitimacy of the celibate modes of life, especially that of the wandering ascetic. Commenting on VeS 3.4.20, he presents two possible interpretations of our text:

1. The three branches of dharma include all four asramas - the first branch comprises the asrama of a householder, and the third, that of a vedic student, while the second refers in common to the asramas of both a forest hermit (vaikhanasa) and a wandering ascetic (parivrat). According to this hypothesis, the brahmasamsthah refers to a person in any of the four asramas who is firmly established in Brahman.(2) B) The three branches refer only to the first three asramas, the second branch comprising only the forest hermit. According to this hypothesis, the brahmasamsthah refers to the fourth asrama, that is, the wandering ascetic. Dismissing the first alternative, Samkara argues in favor of the second and concludes: "Therefore, those who belong to the first three asramas obtain worlds earned by merit. Only the remaining one, that is, the wandering ascetic, obtains immortality" (tasmat purve traya asraminah punyalokabhajah parisisyamanah parivrad evamrtatvabhak).

   In the commentary on the CU (2.23.1) ascribed to Samkara, however, the author expresses a very different view.(3) At the beginning of the commentary (p. 100), he says that the second branch of dharma is "a hermit or a wandering ascetic, although not an ascetic who is established in Brahman but only one who is established in just the dharma of his asrama, because the vedic passage assigns immortality to one who is established in Brahman" (tapasah parivrad va na brahmasamsthah asramadharmamatrasamsthah brahmasamsthasya tu amrtatvasravanat dvitiyah dharmaskandhah). Later in the commentary, however, he rejects the notion that the second branch could include a wandering ascetic, taking the latter to be simply the brahmasamsthah (p. 103). Although the text is somewhat unclear and Samkara can be seen as contradicting himself, he appears to make a distinction here between a wandering ascetic who belongs to the fourth asrama (see p. 106 of the commentary) and one who is brahtmasamsthah and therefore beyond all asramas. Indeed, he calls such a person by the technical term atyasramin (p. 105; see Sprockhoff 1981: 82; Olivelle 1993: 222-32). Thus, in his commentary on the CU 2.23.1, Samkara appears to include all four asramas in the three branches of dharma; in any case, he considers the brahmasamsthah as beyond the four asramas.

   Bhaskara, in his commentary on VeS 3.4.20, cites "the view of some," a view that is identical to Samkara's, and goes on to present a contrary interpretation similar to that of Samkara's opponent. Although he presents the latter as the view of "others," the arguments he puts forward in its favor make it clear that this second view is Bhaskara's own. Here he argues that the "three skandhas" comprise all the asramas and that the final statement regarding a brahmasamsthah applies to a person in any asrama. Ramanuja, in his Bhasya, as well as the Vedantasara (on VeS 3.4.18-19), basically follows Bhaskara's view.

   The native tradition, however, is divided on the interpretation of our text. The Mimamsists object strongly to the opinion of people such as Samkara and claim that at best this passage contains a "reference" to the asramas and not an injunction establishing them.(4) Indeed, an opinion presented anonymously by Ramanuja (on VeS 3.4.19), but probably representing the Mimamsist position, claims that all five virtues listed under the three dharmaskandhah - sacrifice, vedic recitation, gift giving, austerity, and brahmacarya - belong to the householder. Thus, the three would represent three different emphases in the life of a householder rather than three separate asramas.

   Even though there are differences with regard to specific points, most modern scholars from Max Muller, in his early and influential translation of the CU (SBE I, 1879), and Paul Deussen, in his monumental translation of sixty upanisads in 1897, to R. E. Hume (1931) and P. V. Kane (1962-75) have, by and large, disregarded the Mimamsist views and followed the lead of Samkara in interpreting this passage as a reference to the asrama system.(5) These scholars, moreover, by and large ignore the problematic status of the brahmasamsthah highlighted by the native interpreters and take the expression as referring simply to the asrama of a wandering ascetic.

   Sprockhoff, in many of his writings (principally 1981), has vigorously rejected the traditional view that CU 2.23.1 is a reference to the asrama system. On the whole, he is right, and I will not repeat all his arguments here.(6) Even though many, including Sprockhoff, have referred to and frequently discussed this difficult passage, it has not been subjected to the close reading and analysis that it deserves. That is the aim of this paper.

   I.3 The native exegetical tradition and modern scholars alike have focused on phrases ii-iv and vi, considering them to be the most problematic. They have assumed that the meanings of the compounds dharmaskandhah and brahmasamsthah are clear and evident, even though the native tradition itself had raised some questions regarding the meaning and reference of brahmasamsthah (see IV.1). This confidence, I believe, is misplaced; the meanings of these compounds, especially of the former, are far from clear. An adequate understanding of their meanings, I believe, is the key to a proper assessment of the entire passage.

   Modern scholars, moreover, have discussed this passage almost exclusively as an independent and self-contained unit. Yet, it is embedded within the second chapter of the CU. In this regard, modern scholarship represents a step backward, because, at least in its interpretation of the compound brahmasamsthah, the native tradition has taken into account its context within the CU. Important light, I believe, can be thrown on the significance of this passage if we study it within the context of the Chandogya discussion on the udgitha and the saman (see IV.2).

   II

   II.1 Samkara (on CU 2.23.1, p. 100) presents a lucid explanation of the phrase trayo dharmaskandhah: "Its meaning is this: (there are) three - i.e., numbering three - skandhas of dharma, i.e., divisions of dharma" (trayah trisamkhyaka dharmasya skandhah dharmapravibhaga ity arthah). He thus takes skandha to mean "division" or "classification" (pravibhaga) and interprets the compound dharmaskandhah as a dependent determinative (tatpurusa). Ramanuja is not as explicit, but he too clearly takes skandha to mean a division (vibhaga, on VeS 3.4.19) and the compound as a tatpurusa.

   The received wisdom regarding these two points has never been questioned by modern scholarship. Max Muller (1879) translated the compound as "branches of the law." This rendering has been followed with remarkable faithfulness by subsequent translators.(7) Even Sprockhoff (1981: 80), who is usually cautious about accepting traditional interpretations, translates the compound "Zweiges des dharma."

   Following this interpretation, both the native commentators and modern scholars have seen the ordinal numbers in phrases ii-iv as referring to these "branches." Samkara is explicit, attaching the compound dharmaskandhah to each ordinal, e.g., iti esah prathamah dharmaskandhah ("This here is the first branch of dharma").

   But this interpretation runs into a serious problem in phrase v, which begins sarva ete ("all these"). Now, if we were to follow the traditional interpretation to its logical conclusion, "these" should refer back to the "three branches of dharma." Such an interpretation, however, does not sit well with the context; "branches" cannot come to possess or to attain worlds. All interpreters, ancient and modern alike, therefore, are forced to shift gears here and understand "these" as referring to people following the three branches of dharma. Samkara, as usual, is quite explicit and clear: sarva ete trayo 'py asraminah yathoktair dharmaih punyaloka bhavanti - "All these, i.e., all three asramins [those belonging to an asrama], by following the prescribed duties become possessors of worlds earned by merit." None of these interpreters has noticed or acknowledged this shift from branches to people.(8)

   Now, it is possible to argue that such a shift of reference in phrase v may have been influenced by phrase iv. Whereas in phrases ii and iii the "second" and "third" refer to impersonal acts such as sacrifice and austerity, in phrase iv the "third" is the vedic student (brahmacarin), that is, a person. This shift from acts to a person, one may argue, is carried over into phrase v, where "all these" would refer to people such as a vedic student and, by extension, to people who undertake the "first" and the "second" branches of dharma.

   Determining a shift here from impersonal branches of dharma to persons who follow those branches is further...