**4. Textual Passages**

There are two sources in the ancient scriptures that are regularly referred to in the debates on the practice of *sat¯ı* (Sharma et al. 1988; Datta 1990; Hawley 1994; Major 2007; Brick 2010). The one mentioned above (*Atharvaveda* 18.3.1-8), if partially read, can easily be considered as advocating 'this ancient custom' (*dharmam pura¯n. am*), whereas the larger part (2–8) gives the impression that 'lying with the deceased husband' is not at all 'dying with the deceased husband' but merely a symbolic gesture. In some commentaries, it is said that her brother should be the one to help her rise, in some other comments it is posed—more logically, in a patrilinear joint family system—that it was the brother-in-law who rightfully took her hand and accepted her as a co-wife. The reference to wealth and offspring may indeed make more sense this way. Even more cryptic is the parallel passage from *R. gveda* 10.18.7-8:

Let these [accompanying] women, who are not widows, and who do have good husbands, sit down, [their eyes] shining with clarified butter used as collyrium;

May the wives, whose eyes are tearless, who are free from disease, and who are well-decorated, occupy the front seat [respectively: step into the fire].16

Oh, woman! Rise, up towards the world of the living; you lie down near this departed [husband]; Come, the state of being a wife to the husband who held your hand, and whose love for you has now been fulfilled, [is over].

Should we read this passage, being part of a Vedic funeral hymn, as advocating *sat¯ı* or not? The first part was to be repeated when the wife was made to sit near the corpse, as in any cremation ceremony, and the second half was employed when she was instructed to rise. While some have read this as the remnants of ancient customs (providing the dead man with weapons, ornaments, clothes, food and possibly his widows as accompanying him to the hereafter), others have pointed at the dynamics between the two half-passages: the widow should first sit or lie down, then rise and return to the world of the living, and possibly to a new married life with her brother-in-law. Over the course of time the middle passage has been understood in either of two ways, reading *agne* ([sit, or step] in(to) the fire) instead of *agre* ([sit] in front)!

It is often stated that such ambiguous Vedic verses had been incorporated into *sat¯ı* ritual around the sixteenth century to accord Vedic-Brahmanic legitimacy to the practice, such as by Raghunandana, a *pan. d. it* from Varanasi. This probability redirects us to text genres such as the *Dharma´sastras ¯* <sup>17</sup> and *Dharmanibandhas*<sup>18</sup> for providing a more reliable setting in which we may find a clue for the emergence of (or at least the ideology about) *sat¯ı* as an actual and lawful practice. The *Dharma´sastras ¯* , an extensive

<sup>16</sup> As we will see below, any translation of this part is problematic. There have been variations such as "let them occupy the seat in front (first)", but also alternative renderings such as "let them first go up into the dwelling' or even 'let them first mount the womb", or, more dramatically and meaningfully: "let them step into the fire". The controversy centers around the

meaning of the word *yon. <sup>i</sup>* here (womb, or seat) as well as about reading either *agre* (first, or in front) or *agne* (in(to) the fire). <sup>17</sup> *Dharma´sastras ¯* (roughly the first millennium CE) are texts rooted in the *Dharmasutras ¯* , which are dated to the second half of the first millennium BCE. They are both law books on *Dharma*, containing guidelines for individual and social behavior, including ethics and personal, civil, and criminal law. The later *Dharma´sastras ¯* consititute commentaries on the earlier *sutras ¯* , as well as elaborate treatises on duties, and thus form an expansion to the *Dharmasutras ¯* .

<sup>18</sup> *Dharmanibandhas* are one further step in the *Dharma* literature: digests and commentaries written from the early twelfth century onwards. For historical accuracy it is crucial to note that *Dharma* texts are prescriptive, not descriptive. Yet, as a genre that spans some 2.5 millennia, one can guess that they at least reflect something of the evolving regional, ethical, ideological, cultural, and legal practices (Olivelle 1999, pp. 175–8; 2006, pp. 184–5). The genre of *Dharma* texts gained an entirely new status, when in the early colonial period they were declared to be 'the law of the land in South Asia' by the British administrators.

compilation of sociomoral prescriptions produced by Brahmanic traditions, ranged from one or two centuries BCE to probably the end of the first millennium CE. In them, there was a widespread tendency to value the widow as a person who best honors her husband's memory not by self-immolation on his pyre, but by living a life of chastity, frugality, and commemorative rites in support of her husband, especially memorial rites to keep him fulfilled in the hereafter. It is only later that the topic of a wife's co-cremation began to enter Dharmashastric discourse, possibly under influence of royalty and certain martial classes. One explicit example from the *Vis.n. usmr.ti-Dharmasutra ¯* (25: 14) still presents a middle position:

When a woman's husband has died, she should either practice ascetic celibacy or ascend [the funeral pyre] after him.
