Of the many interesting individuals we encounter in the vedic literature, Svetaketu, the son of Uddalaka Aruni,(1) comes across as one of the most colorful and true-to-life characters, not least because he is frequently depicted as the vedic equivalent of a spoiled little brat. Although he appears with some frequency in vedic and later literature both as a young man and as a mature adult, his character is most fully developed and exploited for literary-cum-theological purposes in the story of young Svetaketu's(2) encounter with a king, a story that has become famous because it contains the important doctrines of "five fires" and the two paths along which the dead travel.
VERSIONS OF THE SVETAKETU STORY
We have three versions of the Svetaketu story in the upanisads: Brhadaranyaka 6.2. 1-8 ([B.sup.*]), Chandogya 5.3 ([C.sup.*]), and Kausitaki 1 ([K.sup.*]).(3) The aim of this paper is to examine the divergent ways in which the authors of these versions develop the character of young Svetaketu and to explore the possible theological and/or literary reasons for those divergences.
Of the three versions, [B.sup.*] and [C.sup.*] follow each other rather closely, while [K.sup.*] represents a distinctly different redaction. The king's name in the first two is Pravahana Jaivali and, in the latter, Citra Gangyayani (or Gargyayani). These versions have been studied repeatedly by scholars, whose principal, if not sole, aim has been either to establish which of the versions is the oldest and may have served as the archetype for the others, or to reconstruct a hypothetical archetype underlying all the versions.(4) Renou (1955) has rightly cast doubt on whether the priority of any of the existing versions can be established; indeed, it is highly doubtful that an analysis of these versions will ever provide us with a single clear archetype. Such archetypes are most easily constructed when, as in the case of manuscript transmissions, the changes introduced into the versions are unconscious and accidental, disclosing the genealogy of the manuscripts. The versions of the Svetaketu story, I will argue in this paper, are not accidental creations but deliberate literary inventions.
Although the archeology of texts has become somewhat unfashionable lately, my objection has less to do with its merits than with the fact that, as a result, a much more significant, interesting, and (most importantly) feasible project - namely, the literary study of these texts - has been ignored. Biblical scholars have taken a leadership role in exploiting the literary study of sacred texts; they have asked different types of questions and thereby obtained new insights into the literary and theological motives underlying the composition of biblical texts.(5) Close attention to language, style, narrative strategy, and choice of words helps us understand what the author is aiming to do, what message, subtle or otherwise, he is attempting to impart to his readers or listeners.(6) Scholars whose main goal is to uncover the most ancient versions of texts often tend to ignore later versions, even though it is these versions that provide insights into the religious, intellectual, and social history behind the texts. The story is told not just in the oldest but in the changes we can see from the older to the newer. Likewise, the literary study of texts can also become historically significant when we know the material the authors were working with. Historical and literary study of texts, therefore, need not be antagonistic to each other; they are interdependent and complementary.(7)
CONTEXT AND SOURCES
We have to address two issues at the outset. First, what were the sources at the disposal of the authors of [B.sup.*], [C.sup.*], and [K.sup.*] in composing their respective narratives? Second, what is the literary context within which these narratives are to be located and studied and which may shed light on the authors' theological and literary objectives? The second is related to the first in that a considerable part of the immediate context of the narratives is shared by [B.sup.*], [C.sup.*], and [K.sup.*] and is found also in other vedic texts (Jaiminiya Brahmana and Sankhayana Aranyaka), raising the possibility of tracing at least some of the source material (as opposed to a single archetype) used by the authors. The following is a schematic view of the literary context:
I Contest between faculties BU 6.1, CU 5.1.1 - 2.3, SA 9.1-7
II Mantha rite BU 6.3, CU 5.2.4-9, SA 9.8
III Svetaketu story BU 6.2.1-8, CU 5.3, KsU 1
IV Five fires BU 6.2.9-14, CU 5.4-9, JB 1.45-46 (first part), SB 18.104.22.168-10
V Paths after death: two versions
V.1 JB 1.46 (second part), 49-50, KsU 2-7
V.2 BU 6.2.15-16, CU 5.10
Since BU and CU follow each other closely, we are fortunate to have for each of these sections at least one other independent parallel which can serve as a check in uncovering possible sources. So, for example, in I, CU and SA list only five faculties and place II immediately after I, whereas BU lists semen as the sixth faculty and places II after V. We can, therefore, assume that these two features are innovations introduced by the author of BU, and we can ask what may have motivated him to do this (see below, [section]2.1.1). Likewise, the omission of IV by the author of KsU can be seen as an innovation, since IV is found in JB, as well as in BU and CU. It is, moreover, likely, as both Bodewitz (1973:113) and Schmithausen (1994) have noted, that the JB provides clues to the sources that may have been used by BU and CU, permitting us to see what innovations may have been introduced by the respective authors. It is also likely that V.2, the doctrine of the two paths, to gods and to ancestors - an innovation shared by BU and CU - goes back to a source they shared, while V. 1, the passage to heaven of JB, later recast in KsU, was probably the older sequel to the doctrine of five fires (Bodewitz 1973:113-14).
This leaves us with III, the story of Svetaketu, which forms the preamble to IV and V. 2 in BU and CU, and to V.1 in KsU, but which is missing in the parallel passage of JB. In her pioneering and detailed study of this episode, Sohnen (1981) has analyzed all three versions, paying close attention to the language, style, and selection of words. Hers is in some ways a literary study of this story, but her analysis is aimed at discovering the historical priority of the respective versions. That aim sometimes biases her judgments, as when she takes brevity or "logical consistency" as an indicator of historical priority (1981: 199). Sohnen takes [K.sup.*] to be the oldest version and the probable source of [B.sup.*] and [C.sup.*], and in many areas she thinks [C.sup.*] has preserved an older version than [B.sup.*]. When a passage of [B.sup.*] or [C.sup.*] is in agreement with [K.sup.*] we can readily accept that it probably goes back to an original source and that the author of the other version has introduced something new and ask why he may have done so. I am, however, not convinced that there is compelling evidence to claim that [K.sup.*] is either the oldest version or the model for [B.sup.*] and [C.sup.*]. Sohnen has shown that [K.sup.*] is brief and its narrative structure is logical and simple. But does that necessarily make it older? Simplicity and logic can be imposed on a rambling story by a narrator just as, or even more, easily than a simple and logical narrative can be turned into a disjointed one. If, as seems likely, the author of KsU omitted IV, though found in his sources, then he might well have made other drastic changes to the narrative sequence that he deemed necessary for his own literary or theological purposes. What I propose to show is that each version has its own narrative logic from the viewpoint of the respective author, and the additions, subtractions, and modifications can be viewed as part of the narrative strategy of each author.(8)
It appears likely that of the five text fragments I have isolated above, the fragments I and II existed as a separate unit (which I will call [I-II.sup.*]) as evidenced by SA, and likewise the fragments IV and V form a unit (which I will call IV-[V.sup.*]) as evidenced by JB, a unit which may have contained other material.(9) It also seems likely that in this unit the path after death was at first represented by V.1, since it is found in both JB and KsU. At some point IV-[V.sub.*] was recast with an introductory story containing three protagonists: a royal person, Svetaketu, and his father.(10) This recast unit (which I will call III-[V.sup.*]) was the source of the KsU version. The recast unit appears to have been further modified by replacing V.1 with V.2 and by combining it with [I-II.sup.*]. Now, it is possible that this last version (which I will call [I-V.sup.*]) was the work of the author of either BU or CU,(11) in which case we must assume that the one borrowed this version from the other. Given the discrepancies between the two versions, and the partial agreement of each with other versions of these fragments, especially with [K.sup.*] in fragment III, it appears more probable that the BU and CU versions are modeled on a version of I-[V.sup.*] that is now lost. Let me present this hypothetical relationship and derivations of the five text fragments:
[Unknown Text Omitted]
2.1 Theological and Literary Intent
In analyzing their theology and the narrative strategy, I find that the author of BU intends to teach a theology of sexual intercourse as a fire sacrifice, while the author of CU pursues a theology of the fire sacrifice offered to one's breath (pranagnihotra). The clue to the literary intents of these authors, I believe, is found in the concluding sections that they have appended to I-[V.sup.*], sections that deal with sexuality and offering food to the breaths, respectively. The intent of the author of the KsU is more difficult to determine; it appears that his purpose was...