The composition of the Maitreyi dialogue in the Brhadaranyaka Upanisad.

• Author: Brereton, Joel P.
• Position: Critical essay

0. INTRODUCTION

   In a familiar episode from the Brhadaranyaka Upanisad, as the sage Yajnavalkya is about to depart, he tells his wife Maitreyi that he wishes to divide his possessions between her and her co-wife, Katyayani. Instead of asking for an inheritance, however, Maitreyi asks him for his teaching. Yajnavalkya is very pleased that Maitreyi wishes knowledge rather than wealth and begins a conversation with her about the experience of death and the nature of the self.

   This narrative is unique in the Brhadaranyaka Upanisad because it occurs twice, once at BU 2.4 and again at 4.5. Furthermore, the Brhadaranyaka Upanisad and its embedding text, the Satapatha Brahmana, are transmitted in two closely related recensions deriving from two schools of the Vajasaneyin Yajurveda, the Madhyamdina (M) and the Kanva (K) schools. As a result, we have altogether four versions of the Maitreyi dialogue: two belonging to the Kanva recension and two to the Madhyamdina. Although the two recensions of BU 2.4 are very similar to one another, as are the two of 4.5, within each recension there are significant differences between the version at 2.4 and that at 4.5. These four versions provide a singular opportunity to analyze the composition of the dialogue and to refine the interpretation of the text. (1)

   Before examining the dialogue itself, let me situate the two schools and review the likely compositional history of the Satapatha Brahmana. In the middle and late Vedic periods, both schools were located near one another in eastern North India: the Kanvas in the Kosala area (eastern U.P.) and the Madhyamdinas further to the east in Videha (northern Bihar) (Witzel 1997a: 307). Not surprisingly, in view of the proximity of the schools, their texts show common sources and considerable interaction. The schools shared an original *Vajasaneyin Brahmana, of which only traces remain (Witzel 1997a: 314f.; Caland 1903: 508ff.). Each school revised and supplemented these materials and then subsequently influenced one another in creating the two recensions of the Satapatha Brahmana. Thus, while SBM 1-5 and SBK 1-7 are independently derived from the *Vajasaneyin Brahmana, the Madhyamdina recension significantly influenced the Kanva version of these chapters (Caland 1926: 90). SBM 6-10 / SBK 8-12, on the other hand, likely originated within a third tradition, that of the Sandilya Yajurvedins, who may have been centered north of Kasi (Witzel 1989: 115 n. 35, with reference to Jain 1947). If this was their location, these Sandilya Yajurvedins were geographically closer to the Kanvas than to the Madhyamdinas and therefore were more likely to have interacted with them. Thus, the Sandilya books of the SB may have first come to the Kanvas and then through the Kanvas to the Madhyamdinas. Finally, SBM 11-13 / SBK 13-15 is probably of Kanva origin (Caland 1926: 108), and the sources of the last book, containing the upanisad, are unclear. In schematic summary:

   (SBM 1-5 [right arrow] SBK 1-7) [left arrow] *Vajasaneyin Brahmana (SBM 6-10 [left arrow] SBK 8-12) [left arrow] *Sandilya Brahmana SBM 11-13 [left arrow] SBK 13-15 SBM 14 / SBK 16 (includes the Brhadaranyaka Upanisad) All these processes--independent variation of an original text, adoption of materials from other sources, and subsequent revision of each recension under the influence of the other--are visible in miniature within the Maitreyi dialogue.

1. AVRTAT

   One entry into the history of the dialogue is through the analysis of a specific passage, which can both illustrate the relationship between its different versions and provide a trail toward broader conclusions about the text. The passage I will examine is K/M 2.4.4 [approximately equal to] K/M 4.5.5, in which Yajnavalkya praises Maitreyi for choosing knowledge rather than possessions:

   M 4.5.5 priya khalu no bhavati sati priydm avrtat / hanta khalu bhavati te 'ham tad vaksyami / vyakhyasyami te / vacam tu me vyacaksanasya nididhyasasveti. Edgerton 1965: 166: "Of course you are dear to me, madam, and (naturally) you have acted in a manner dear to me. Well then, madam, I will explain and expound this to you. But do you weigh my words carefully while I am expounding it." K 4.5.5 priya vai khalu no bhavati sati priyam avrdhat / hanta tarhi bhavaty etad vyakhyasyami te / vyacaksanasya tu me nididhyasasveti. K 2.4.4 ([approximately equal to] M 2.4.4) priya batare nah bhavati sati priyam bhasase / ehy assva vyakhyasyami te / vyacaksanasya tu me nididhyasasveti. Olivelle 1988: 67: "You have always been very dear to me, and now you speak something very dear to me! Come and sit down, I will explain it to you. But while I am explaining, try to concentrate." In 4.5.5, the only difficulty in understanding the text is the verb avrtat in M, for which K has avrdhat. In 2.4.4, both M and K have bhasase 'you speak', which is a bland replacement for the problematic avrtat or avrdhat.

   Formally, K's avrdhat is unproblematic since it is a regular thematic aorist of avrdh 'increase'. Its only drawback is that it does not make much sense. The strain it produces is apparent in Edgerton's additions in his translation of K (1965: 166 n.): "you have added to what is dear (or pleasing, sc. to me in you)." Its formal clarity and semantic obscurity make it likely that avrdhat is a largely unsuccessful effort to interpret avrtat, the reading in M. Translators' attempts to understand avrtat reflect the challenge that this form poses. Edgerton has "you have acted in a manner dear to me" and Bohtlingk (1889: 73) "... hat sie [die Verehrte] uns Liebes erwiesen." Both assume that avrtat is a thematic aorist of [square root of vrt], which is formally unobjectionable (cf. Hoffmann Aufs. I: 248 with n. 9). The problem is the meaning of this line, and the two scholars attempt to account for [square root of vrt] in different ways. Edgerton takes priyam as an adverb with [square root of vrt], although the expected idiom for 'act in a manner dear' is priyam [square root of kr] (cf. priyamkara and priyamkarana, PWIV, 1163), and Bohtlingk posits a unique sense for [square root of vrt]. (2)

   More likely, the form reflects an underlying root aorist avrta 's/he has chosen' or 'has preferred'. This 3 sg. medial root aorist of [square root of vr] occurs in early Vedic, and the 2 sg. avrthah is still attested in later texts (SB 10.3.4.1, JB 3.188, TS 2.5.9.5).3 Following a suggestion of Stanley Insler (pers. comm.), we can then account for the form avrtat by haplology. That is, avrtat represents avrta tat, with omission of the repeated syllable -ta-. This derivation of avrtat from avrta tat yields an unproblematic translation: "You, my lady, are, of course, cherished by us, (and) being so, you have preferred that cherished (by us)." Another consideration in favor of this interpretation is that it produces an idiom that appears elsewhere in Vedic. The notion of "choosing" or "preferring" something spoken occurs in RV 9.101. 13ab, where we even have the phrase vrta tat: pra sunvandsyandhaso, marto na vrta tad vdcah "Like a mortal, [the dog] prefers that speech of the (soma-)plant as it is pressed." (4) An even better parallel to this passage--and an upanisadic one at that--provides a variant of the expression priya + [square root of vr]. This is KaU 2.1 hiyate 'rthad ya u preyo vrnite "But he misses his aim, who prefers the more pleasant." Despite the currency of the idiom priya + [square root of vr], the avrtat attested in M 4.5 was unrecognized by the other three versions of the Maitreyi dialogue, which either chose to replace it by a familiar, if inappropriate form (K 4.5) or substituted their understanding of what Yajnavalkya meant to say (M, K 2.4).

2. RHYTHMIC PROSE

   Recognition of the haplology in M 4.5.5 has an unexpected consequence. Restoring the missing syllable helps reveal a distinctive style of composition underlying this passage. M 4.5.5 was originally composed in a succession of approximately eight-syllable lines, in which individual lines normally corresponded to syntactic units. These are not verse lines, for there is no consistent pattern of syllable length, and therefore we might describe this style as a form of rhythmic prose. Nor is this the method of composition of the transmitted text in any of its versions, for in the development of the text the old rhythmic prose passage was expanded and transformed into ordinary prose.

   Let me illustrate this style and how it can be uncovered in M 4.5.5. With some gentle massage, Yajnavalkya's answer to Maitreyi falls into a sequence of generally eight-syllable lines:

   sa hovaca yajnavalkyah a. priya khalu no bhdvati b. sati priyam avrtad (< *avrta *tad) c. hanta [khalu bhavati] te 'ham tad vaksyami [vydkhyasyami te] d. vacam tu me [vyacaksanasya] nididhyasasva[-iti] (9 syllables) bravitu bhagavan iti Yajnavalkya said, "You my lady are, of course, cherished by us, (and) being so, you have preferred that cherished (by us). Come, [my lady; of course] I shall tell that to you. [I will explain it to you.] But you must try to reflect on my words [as I explain it to you]." "Speak, sir." In this and other reconstructions, I have bracketed words I believe to have been added to the core rhythmic text when it was recomposed in non-rhythmic prose, and I have set to the left passages that lie outside of the rhythmic text. I have placed in parentheses possible modifications of the text or ways of reciting it that produce approximately eight-syllable, verse-like lines. I realize that there is a circularity to my method here: it is because I anticipate eight-syllable lines (based on lines like ab) that I modify the text (in cd for example) to show eight-syllable or approximately eight-syllable lines. But my choices for alteration and excision have additional justification. In line c, for example, I omit khalu bhavati because it is unnecessary to both syntax and sense and because it echoes khalu no bhavati in line a, from which it was...