Shiur #11a: Understanding "Sefeik Sefeika"
Translated by David Silverberg
The Gemara (9a) discusses the halakha of Rabbi Elazar that if a groom claims to have discovered on his wedding night that his bride was not a virgin, this claim suffices for her to become halakhically forbidden to him. Given the possibility that she had betrayed him during their period of eirusin (betrothal), in which case they may no longer live together, his claim is sufficient grounds to trigger this prohibition. The Gemara questions this ruling in light of the concept of "sefeik sefeika," or "compounded doubt." Forbidding a bride to her husband in this case requires two assumptions: that her previous relations occurred after the betrothal, and, secondly, that she engaged in relations willingly and not coercively. Therefore, the Gemara claims, since we have two doubts to resolve before conclusively considering her forbidden, she should remain permitted. The Gemara suggests two cases where there is only one safeik. First, this halakha applies specifically to a case where a woman was betrothed - through the agency of her father - younger than the age of three. Since no significant sexual act can be performed before the age of three, this bride must have engaged in relations during the betrothal period. Thus only a single doubt remains - whether or not she committed the act willingly - and she is therefore forbidden. Alternatively, the Gemara suggests that we deal here with a kohen's wife, who becomes forbidden to him even if she is raped after her betrothal.
It emerges, then, that whereas we must rule stringently where a single point of doubt exists, in a case of a sefeik sefeika we may rule leniently. Other rules that we follow in situations of doubt, namely, "rov" (following the majority) and "chazaka" (following the status quo), are derived by the Gemara from Biblical verses (see Chulin 10b-11a). The Gemara does not, however, supply an explicit source for this rule of sefeik sefeika. In this shiur we will try to determine the origin and nature of this halakha.
As mentioned, our sugya concludes that a newlywed husband's claim of "petach patu'ach" renders his wife forbidden to him in only two cases: if the betrothal occurred before she was three years old, and if he was a kohen. The Yerushalmi, however, rules otherwise. According to the Yerushalmi, since word quickly spreads when a rape is committed, lacking those rumors we cannot entertain the possibility that the woman lost her virginity as a result of rape. We are therefore left with but a single safeik - whether she engaged in relations before or after the betrothal - and "safeik de-orayta le-chumra" (in cases of doubt concerning Torah law, we must rule stringently).
Seemingly, the Talmud Bavli felt that even if there were no rumors we nevertheless can entertain the possibility that she had been raped. We therefore have a situation of sefeik sefeika, except in the two specific cases mentioned. Tosefot (s.v. ve-i ba'it eima be-eishet yisrael), however, understand our Gemara differently. They ask why we do not employ the rule of chazaka to permit the wife to her husband. A girl before her third birthday was certainly permitted to her groom at the time when her father accepted kiddushin on her behalf. Why, then, may we not presume the maintenance of this status in the case of petach patu'ach, and assume that the woman had been raped? Tosefot answer by invoking the aforementioned Yerushalmi: "The news of a rape spreads, as we say in the Yerushalmi; since there was no news [regarding this woman's having been raped], the possibility [that she committed the act] willfully constitutes the majority, whereas [the possibility] of rape is the minority. And [where we have a conflict] between rov and chazaka, we follow the rov." Tosefot must certainly acknowledge that the Bavli cannot fully accept the Yerushalmi's position, for, after all, the Bavli considers the possibility of rape in order to establish a sefeik sefeika in standard cases (see Tosafot HaRosh). Tosefot therefore proceed to cite the following enigmatic comments of the Ri: "This rov [by which we assume that it occurred] willingly is not an outright rov, but rather mi-derabbanan; therefore, in a situation of sefeik sefeika we allow it, whereas in a situation of a single safeik the rabbis consider [the possibility that the relations were engaged in] willingly as the majority with respect to [the chance that it was] rape." The solution of the Ri requires explanation. What is a "rov de-rabbanan" (a rov instituted by Chazal), and why does it supersede a chazaka but does not override a sefeik sefeika?
A Form of Rov
In order to understand this passage, let us take a step back and examine how a sefeik sefeika works. The Rashba writes in one of his teshuvot (401) that sefeik sefeika operates on the principle of rov. After all, if both sfeikot are more or less equal, there is approximately a seventy-five percent chance that the woman is permitted to her husband. The Rivash, by contrast, appears to have understood differently. In one of his teshuvot (372), the Rivash questions Tosefot's position in light of a Gemara in Masekhet Bekhorot (20b). The Gemara there establishes that only a mi'ut (minority) of animals give birth to a halakhic bekhor - a male firstborn. Fifty percent of animals born are females, and a minority of fetuses are stillborn. The combination between the female young and the small number of stillborn animals yields a statistical minority of animals who deliver male firstborns. The Rivash questions why a similar calculation is not made in our sugya. In a standard case of petach patu'ach, we have an equal possibility that the woman had engaged in relations before or after betrothal, and due to lack of rumors a minority likelihood that she had been raped. The net result, therefore, is a majority likelihood that the woman may continue living with her husband. The Rivash answers that the chance of a male firstborn and that of a female are statistically equivalent; thus, adding the consideration that a minority of pregnancies end in stillbirth generates a majority of animals who do not give birth to male firstborns. In our case, however, the two possibilities of relations having occurred before and after betrothal are not statistically equivalent. Each possibility is plausible, and we have no means by which to determine which is the correct one. Nevertheless we cannot claim a statistical equivalence between them. Therefore, even when we add the minority possibility that the woman had been raped, we do not come up with a rov. Thus, according to the Rivash, we cannot allow the woman to remain with her husband on the grounds of rov, but we may do so only because of a sefeik sefeika. Seemingly, then, the Rivash did not accept the approach of the Rashba, who viewed sefeik sefeika as a rov.
Upon further reflection, however, even the Rivash can agree to the Rashba's view, that sefeik sefeika operates on the principle of rov. The Rashba clearly did not view a sefeik sefeika as a statistical rov generated by the combination of two sefeikot; he rather referred to a majority of theoretical possibilities. In our case, we face the possibility that the woman engaged in relations willingly on the one hand, and the possibility that she was raped, on the other. And within the possibility that she engaged willfully, there is still a possibility that she remains permitted to her husband - if this occurred before her betrothal. In other words, combining the two sefeikot results in four different possibilities, whereby three of which allow for the woman to remain with her husband. The prohibition takes effect only according to one of the four possibilities: if she willfully betrayed her husband after their betrothal.
In light of this understanding of the Rashba's view, we may accept the Rivash's approach, as well. Sefeik sefeika works not according to a numerical majority, but rather based on a majority of possibilities. This is precisely what the Rivash claims, that only regarding bekhorot we can make this calculation, of combining the minority of stillborn animals to the female births to come up with a majority that are not bekhorot. In our case, however, we do not have a numerical equivalence between the possibilities of the relations having taken place before or after the couple's betrothal. We may therefore implement the sefeik sefeika leniency even without a numerical majority, since we have a majority of possibilities.
According to what we have said, we can also explain the Ri's distinction between a "rov de-orayta" and a "rov de-rabbanan." Before we show how, let us raise another question, regarding the halakha of rov. Does this provision resolve the doubt at hand, such that once we follow the rov we are left with no doubt whatsoever, or does it simply instruct us how to proceed when a doubt arises, without resolving it conclusively? The Tanna'im argue as to what happens in a situation where a rov opposes a chazaka. According to Rabbi Meir, we follow the chazaka, since we can add to it the minority. The Chakhamim, however, rule that we follow the rov. Quite possibly, Rabbi Meir holds that a rov does not resolve the doubt; were it to resolve the doubt, then we would have no right to add the minority possibility to the chazaka. After all, once the rov resolves the safeik, we have no mi'ut at all! The fact that we may add the minority possibility as a consideration onto the chazaka shows that our doubt remains even after applying the rov. Therefore, Rabbi Meir holds that we cannot override the status quo (chazaka) based on the rov.
With this in mind we can return to the distinction drawn by Tosefot between a "rov de-rabbanan" and a "rov de-orayta." Perhaps a "rov de-orayta" resolves the safeik, whereas a "rov de-rabbanan" serves merely as a guideline, telling us how to proceed in a situation of safeik. Meaning, a "rov de-rabbanan" does not have the capacity to resolve the doubt; it allows us to follow the majority despite the fact that our uncertainty remains. Therefore, if we accept the view of the Chakhamim, that rov supersedes chazaka, then despite the woman's chazaka - status quo - as permitted to her husband, the rov that she had willingly engaged in relations is sufficient to render her forbidden. However, in a case of sefeik sefeika, this "rov de-rabbanan" will be of no consequence: since it does not have the strength to resolve our doubt, we follow the majority of possibilities, and she thus remains permitted. In a case of a "rov de-orayta," by contrast, we have effectively resolved our doubt and are then left with only a single safeik.
We find among the Rishonim an additional approach to understanding the concept of sefeik sefeika. The Ra'a, in his Bedek Ha-bayit (his critique of the Rashba's Torat Ha-bayit), claims that sefeik sefeika works only when the first safeik yields only a rabbinic prohibition, and not a Torah prohibition. When we add the second safeik we then have a situation of safeik de-rabbanan - a doubt concerning a rabbinic enactment, in which case we generally rule leniently. For example, if a piece of meat that had been sacrificed for idolatry - which renders it forbidden - falls into a group of kosher pieces of meat, Torah law permits the consumption of the pieces of meat (through the institution of bittul). Chazal, however, forbade their consumption. Thus, if unidentified piece from this group then falls into another group of pieces of meat, the second group is permissible for consumption due to the principle of sefeik sefeika. In our sugya, the question of whether the woman had relations before or after betrothal yields a rabbinic prohibition; according to Torah law, we can apply her chazaka and assume that she remains permitted to her husband, but she is prohibited mi-derabbanan. But by adding the second safeik, we arrive at a situation of safeik de-rabbanan, and we therefore rule leniently.
Tosefot clearly did not accept this approach, for they in fact asked why our sugya does not take the woman's chazaka into account. They even conclude that in this case we cannot rely on this chazaka. According to the Ra'a, we indeed implement the chazaka and specifically as a result of doing so we encounter a situation of sefeik sefeika. We should note that the Ra'a represents the minority view. Most Acharonim follow the Rashba's approach and view sefeik sefeika as a manifestation of rov. (See, for example, the Shev Shemateta.)
We began by questioning why we find no source for sefeik sefeika. According to the Ra'a, sefeik sefeika does not describe a new halakha, but is merely an application of an existing principle safeik de-rabbanan le-kula. Similarly, according to the Rashba, sefeik sefeika is a form of rov and as such can be derived from the source the gemara suggest for rov.
To be continued in next week's shiur.