Safek Sefeika as Rov
Lecture #04: Safek Sefeika as Rov
By Rav Moshe Taragin
A previous shiur addressed the famed mechanism known as safek sefeika, a "doubt upon doubt." The shiur alluded to the Rashba's association between safek sefeika and rov. Here is a citation of that reference:
The Rashba, in one of his teshuvot, offers an interesting rationale for this halakha of safek sefeika. He equates safek sefeika with rov (probability), claiming that, essentially, we can view safek sefeika as a logical rov. [Chazal state that a woman who is accused by her husband of having sexual activity with someone other than her husband may remain married to her husband because of the concept of sefek sefeika.] Perhaps this is because most women who engage in premarital sexual activity do so in a manner that does not forbid them from continuing with their husbands (even though the act itself is forbidden). According to this approach, from a purely statistical standpoint, every safek sefeika yields a statistical "probability." A different way of viewing safek sefeika as rov may be to assess logical possibilities rather than statistical probabilities. Anytime we face two sefeikot, there exist four scenarios. Taking the case of the possibly disloyal woman, the following four cases are possible: a) she had voluntary contact prior to engagement; b) she had coerced contact prior to engagement; c) she had coerced contact subsequent to engagement; d) she had voluntary contact subsequent to engagement. Inasmuch as in alleged scenarios a-c she is permitted to her husband, we allow her to remain with her husband, since most suggested scenarios support this ruling. In any event, either articulation would firmly associate safek sefeika with rov.
This shiur will elaborate upon the two different manners of defining safek sefeika as rov.
On the surface, the equivalence between rov and safek sefeika is not an obvious one. The concept of rov is based upon assessing percentages and ruling according to the majority. Sometimes, the percentages arise from an immediate and defined population. For example, if nine stores on a street sell kosher meat and one sells non-kosher meat, we may assume - based upon rov – that a nondescript piece of meat originated from the kosher, "majority type" store. Often, however, a rov does not reflect a distinct and concrete population, but rather is based upon human behavior. Most children are born from their "recognized father" and not out of wedlock (rov be'ilot achar ha-ba'al). We can therefore execute someone who wounded his father with halakhic "certainty" that his recognized father is indeed his biological one. The former rov, pertaining to distinct items and their qualities, is known as a "ruba de-ita Kaman," whereas the latter rov, governing "general tendencies," is known as a "ruba de-leta kaman." Despite subtle differences, each relates to likelihoods of particular populations - universal populations or set and bound ones.
The concept of safek sefeika is neither type of rov would require "manipulation" in order to be likened to rov. For example, the gemara in Ketuvot (9a) discusses a woman who has been accused by her husband of committing adultery. In order for her to be prohibited to her husband, the suspected bi'ah would have to be consensual and subsequent to marriage. This dual requirement yields a safek sefeika: perhaps the bi'ah occurred PRIOR to marriage, and even if it occurred subsequent to it, perhaps it was involuntary (rape). The dual requirement necessary to create issur may yield a statistic: among the women who engage in non-marital sexual encounter, less than 50% commit willful adultery while married. Assuming 50% engage while married and 50% while not married, we may detract from the 50% who commit while married, the percentage of those who are raped, thereby reducing the overall figure of consenting adulterers to less than 50%. This produces a "logical" rov in favor of permitting a woman to remain with her husband.
A further debate may surround defining this type of rov as "ita kaman" or "leta kaman." It would seem, at first glance, to be a case of leta kaman since we are dealing with human nature and not with concrete identity; we are not considering ten pieces of meat whose identity has been established. On the other hand, we are not addressing a universal population but rather a distinct group of people (women who engage in non-marital sexual activity); this may render it a ruba de-ita Kaman. Either way, this engineered statistic may resemble classic mathematical rov.
Several comments of the Tosafot stipulate qualifications to safek sefeika which may imply a more abstract definition of safek sefeika as rov. For example, Tosafot in Ketuvot (9a) s.v. ve-I claim that a safek sefeika can only operate if each of the sefeikot is a balanced safek – known as a safek shakul. If one of the questions isn’t statistically balanced, the safek sefeika loses its impact.
If safek sefeika is a manufactured statistic, it should operate even if ONE of the safeikot is imbalanced. For example, even if MOST women who commit non-marital sexual activity do so voluntarily, and thus one case of safeik is tilted towards prohibition, the OVERALL safek sefeika may still be manipulated to a rov towards permissibility!! Since 50% of women who perform non-marital sexual activity do so BEFORE marriage, only the 50% who do so WITHIN marriage are even COURTING prohibition. If, among those 50% who commit within wedlock, even a small fraction do so involuntarily (thus still remaining permitted to their husbands), the OVERALL PERCENTAGE of women prohibited to their husbands dips below 50%. The percentage of women who engage in non-marital sexual activity voluntarily and within wedlock (the two necessary conditions to become prohibited) is less than 50% EVEN IF one of the pertinent safeikot is uneven. The position of Tosafot that both sefeikot must be balanced suggests that safek sefeika IS NOT merely a statistical rov.
Yet another stipulation of Tosafot concerns the classification of two safeikot as integrated - known as "mishem echad." For example, Tosafot consider a safek sefeika as follows: perhaps the bi'ah was involuntary (permitted to husband); even if it were voluntary (and potentially impermissible), perhaps it occurred while she was a minor and is therefore considered rape, which does not prohibit her to her husband. Tosafot reject this safek sefeika, since both questions or both scenarios concern a common basis for permissibility: the possibility of ones. Whether actual rape or voluntary activity while a minor - which is considered "statutory rape." Occurred, the permissibility is based on classifying the bi'ah as rape. A safek sefeika must be constructed by joining two different questions, each necessary to yield a halakhic state of issur. If the questions sound halakhically similar, even though they are empirically different, they can not constitute a safek sefeka.
Why should "nomenclature" impact the dynamic of a statistical safek sefeika? If only 50% of women who engage in non-marital activity do so voluntarily and OF THOSE 50% a percentage are minors (and still permitted to their husbands), why can't we state statistically that less than 50% of non-maritally sexually active women commit an act which renders them prohibited to their husbands?
It would appear that Tosafot assert a very different type of rov as the basis for safek sefeika. If a halakhic predicament yields multiple halakhic conclusions and a majority of those halakhic positions point toward one halakhic ruling, we may rule according to that option based on the majority of OPTIONS. Unlike empirical rov based upon human behavior or the identity of a sample population, this type of rov is based upon the majority of halakhic options. Considering an adulterous women, we must consider four scenarios: voluntary and within wedlock, voluntary and out of wedlock, involuntary and out of wedlock, and finally involuntary and out of wedlock. Out of these four options, the latter three all warrant maintaining permissible status to her husband. The only option that suggests prohibition is the first; since most options suggest permissibility, halakha demands that we tilt in that direction in the absence of hard evidence.
Given this perspective, we may more easily adopt the aforementioned qualifications of Tosafot. If one of the participating sefekot is not "even," the question may no longer present dual halakhic options. We may employ a natural rov to resolve and essentially eliminate one safek. Again, taking the base example, most women committing extra-marital sexual activity do so voluntarily. Consequentially, this question is no longer active since we assume, based upon rov, that the bi'ah was consensual. As this question is no longer "uncertain," the only question that obtains is whether the bi'ah occurred within wedlock. As there are only two options - one to permit her and one to forbid her - we cannot conclude a pesak based upon the "majority of halakhic options." Safek sefeika based upon a majority of opinions may indeed collapse if one safek is lopsided.
Similarly, the failure of safek sefeika if the two questions can be categorized as one (mishem echad) may also be understood if safek sefeika is based upon a majority of opinions. If both questions surround the same halachik issue, we ultimately enjoy only two halakhic options. In the absence of a majority of opinions, safek sefeika fails. Again, using the example of Tosafot, both the question of whether the bi'ah was voluntary as well as the question of whether it occurred while she was a minor yield an option to permit her based upon the act being considered halakhic rape. This is considered one option and not two, and therefore we no longer enjoy a majority of "heter" options.
A third qualification of safek sefeika may emerge from the comments of the Ri Hazaken, cited in detail by the Rashba in several locations. Would safek sefeika apply if a questionable piece of meat fell into a mixture with less than the requisite shiur for bitul? Can we articulate a safek sefeika that reads: maybe the original piece of meat in question was permissible to eat; even if it were forbidden, the piece that the person now selects from the mixture into which this meat fell may or may not be that original problematic piece? Safek sefeika should allow us to eat the selected piece of meat!!
The Ri Hazaken claims that a safek sefeika does not apply in this instance. Although his logic is a bit unclear, many have offered the following explanation: since the two sefekot do not present simultaneously, we cannot apply safek sefeika. Typically we face the two logically dependent questions at once when attempting to resolve a halakhic situation. When considering a woman who may have committed adultery and is forbidden to her husband, we simultaneously face the question of timing of the bi'ah as well as whether it occurred voluntarily. In the case of a questionable piece of meat which fell into a mixture, we were presumably aware of the halakhic question surrounding the meat BEFORE it fell into the mixture. This prior awareness prevents the two sefekot from combining into a safek sefeika.
If safek sefeika were merely a manufactured statistic, why would the "presentations" of the two sefeikot be critical? If, however, safek sefeika entails a procedure for permitting items when the majority of halakhic opinions yield that ruling, we may adopt this timing principle. Once one of the sefeikot was addressed prior to and independent of, the second safek, the sefeikot can no longer combine to form one halakhic query about which multiple opinions exist with a clear majority of opinions. Essentially, the two cases must be addressed autonomously and we are facing two DIFFERENT halakhic queries - one about the original nature of the meat and a second independent query about which piece was ultimately selected out of the mixture. The inability to incorporate the two questions into one halakhic query prevents the construction of a safek sefeika, which asserts a majority of halakhic options. Again if safek sefeika were merely an engineered statistical rov we may build a safek, sefeika model even when the two uncertainties do not 'present' simultaneously.