There are certain decisive principles, כללי הוראה, which halachic decisors, פוסקים, utilize when analyzing a Talmudic passage and establish the halacha. Two such principles are הלכתא כבתראי, the law is like the Amora of later scholastic generation; and הלכה כלישנא בתרא, the law is like the latter of two textual variants laid out in the Gemara. Both come into play in the Rosh’s analysis of our sugya, on Gittin 74b . Both relate to continuous concerns of this column—relationships between Amoraim and textual variants.
As background, if a house in a walled city was sold, and not redeemed by the seller within 12 months, the purchaser keeps it forever. A Mishna in Arachin states that initially, because of this, the purchaser would go into hiding on the final day of those 12 months so that the seller couldn’t redeem it. Hillel the Elder therefore instituted that the seller could deposit the money in the Temple treasury, then bust into the house and retake possession.
In the first variant, called a lishna kamma, Rava (fourth-generation) deduced from Hillel’s ordinance that, generally speaking, giving against the recipient’s will doesn’t work. Therefore, if a husband gives a get, divorce document, on condition that she give him 200 zuz, the get is only valid if she pays it with his consent. To this, someone objects. The objector may be Rava’s fifth-generation student, Rav Pappa. Alternatively (וְאִיתֵּימָא), it’s Rav Shimi bar Ashi, Rav Pappa’s contemporary who frequented Abaye’s household but isn’t precisely reckoned among Abaye and Rava’s students. This objector says that perhaps (וְדִלְמָא) the aspect that Hillel’s ordinance addressed was that the giving wasn’t before him, since the seller had gone into hiding, but giving works generally even against his will. (I’d note that this objection seems so obvious that it’s hard to imagine Rava overlooking it. See also Tosafot’s issue with the version—why did he go into hiding?2)
In the second variant (lishna batra), Rava deduces from Hillel’s ordinance that generally speaking, money must be given in the recipient’s presence; but in his presence, it can even be given against his will. Therefore, if the husband stipulates the monetary condition, it will work even against his will. (This is the reverse of the first variant.) To this, Rav Pappa or Rav Shimi bar Ashi objects that “perhaps” Hillel instituted what was necessary (since the seller went into hiding), but generally speaking, it also wouldn’t work against his will even in his presence.
Picking First or Last
When presented with two Talmudic variants, and not ones that we discover ourselves by examining manuscript evidence, but ones which were incorporated into the standard Talmudic text by the redactors themselves, or perhaps some scribes along the way, which wins? This is a major dispute among Rishonim.
In Avoda Zara 7a, a brayta discusses what happens when two Sages sit together and disagree whether an item is pure/impure, or whether it is permitted/forbidden. You follow the Sage who is greater in wisdom or number. If neither is greater, you follow the stringent position. Rabbi Yehoshua ben Korcha disagrees. Rather, if in biblical law, you follow the stringent position; if in rabbinic law, you follow the lenient position. Rav Yosef, the Amora, endorses Rabbi Yehoshua ben Korcha’s approach.
Tosafot ad loc. survey approaches to אִיכָּא דְּאָמְרִי: For biblical law, Rashi followed the stringent approach; for rabbinic law, he followed the last listed variant (lishna batra). The Riva (Rabbi Yitzchak ben Asher, Rashi’s student) considered the first variant (lishna kamma) primary and any variant listed after it (lishna batra) ancillary. Rabbeinu Tam applies Rabbi Yehoshua ben Korcha’s approach to אִיכָּא דְּאָמְרִי, treating each listed Talmudic variant as its own Sage, adopting the stringent reading for biblical law and the lenient for rabbinic law. Finally, Rabbeinu Shimshon introduced an extra twist. In some instances of אִיכָּא דְּאָמְרִי, if one examines other Talmudic sugyot, one could see that the Talmud maintains like one of the variants, in which case we follow that version3.
We might imagine other positions, such as always following the last (lishna batra) even in biblical law (perhaps Rif and Rambam); or weighing the relative strength/plausibility of each reading. Further, this is a rule-of-thumb, occasionally overruled by other factors or decisive principles.
Picking Later Generations
Another decisive principle, similarly named, is hilcheta kevatrai, the law is like the later scholastic generation. This is typically applied from the fourth scholastic generation (Abaye and Rava) and on4. Otherwise, we might give greater weight to the earlier Amoraim. A countervailing principle would be to select the teacher over the student—so typically Rabbi Yochanan over Reish Lakish. After all, דברי הרב ודברי התלמיד דברי מי שומעין, who will you listen to, the teacher or the student? This is then a practical reason for knowing Talmudic biography, thereby knowing an Amora’s scholastic generation and who his teacher is.
Since I’ve recently encountered a few who made this mistake, I will stress that hilcheta kevatrai only refers to being later chronologically, not to being the final person/position listed in a sugya. I can see how this mistake can crop up, since the Talmudic Redactors will regularly describe the discourse and the relative positions in chronological order, with the chronologically last also being the textually last. Also, in many Talmudic arguments even among people of the same generation, either the Amora, or the Talmudic Narrator on his behalf, often states a reason for rejecting the preceding proposal, with the last listed person/position standing with no objection. In such cases, we rule like them, but not because they were listed in the final position.
I’ve seen them misattribute this position to the Rosh, misunderstanding the Rosh Bava Kamma 1:5. There, the Rosh notes that second-generation Rav Huna appears last in the sugya so maybe we should rule like him. However, he explicitly says that this is because fifth-generation Amoraim, Rav Pappa and Rav Huna b. Rav Yehoshua appear textually earlier, and this inversion of (the Talmudic Redactor) Rav Ashi’s typical following of the chronological order may indicate Rav Ashi’s endorsement of the last position. The Rosh then appeals to a Talmudic variant (“amar Rav Huna” instead of “Rav Huna amar”) to have Rav Huna appear in a separate sugya commenting on the Mishna, so the chronological inversion isn’t present. Regardless, the Rosh only proposes this for cases of chronological inversion in the text, not generally for all textually final positions in the sugya.
So too, hilcheta kelishna batra, as used by Rishonim, is only invoked for אִיכָּא דְּאָמְרִי variants5, which are relatively few. Lishna here doesn’t mean position but formulation/variant. I’ve never seen it applied to typical Talmudic disputes between Amoraim, which appear on most pages of Talmud. In any dispute between two Amoraim—Rav Nachman and Rav Sheshet, Abaye and Rava, Rav and Shmuel—someone has to appear last! Order of appearance isn’t necessarily indicative of the Talmudic Redactor’s legal judgment.
Our Sugya
In our sugya in Gittin, the Rosh quotes the Baal HaIttur (Rabbi Yitzchak ben Abba Mari of Marseilles) who quotes Rav Hai Gaon that we rule like Rava of the lishna batra—so if the husband stipulated a monetary condition for the get, she can pay it even against his will. What about Rav Pappa’s objection? After all, he’s a batrai, chronologically later! Well, he doesn’t come to argue with Rava, but to give-and-take (נשא ונתן) within Rava’s words, like a student before his teacher. Further, Rav Pappa uses the terminology of “perhaps”, so the matter wasn’t clear to him. The Rosh also quotes the Rambam6 that she is doubtfully divorced. Therefore, it is good to be stringent. However, primarily, Rosh holds like Rav Hai Gaon.
I’m not entirely persuaded. This teacher / student relationship may work if dealing with Rav Pappa, who is Rava’s talmid muvhak, distinguished student. However, if it’s Rav Shimi bar Ashi, he isn’t reckoned so much as a student. Also, וְדִלְמָא is appropriate for the מַתְקֵיף לַהּ objection because Rava tried to make a clear deduction to deduce new law, so Rav Shimi bar Ashi only needs to show its not the only possible conclusion from Hillel’s decree. He does not and cannot argue that the Halacha is otherwise, just that Rava’s deduction is not proven.
Rabbi Dr. Joshua Waxman teaches computer science at Stern College for Women, and his research includes programmatically finding scholars and scholastic relationships in the Babylonian Talmud.
1 copied from the primary sugya in Arachin 31b
2 Also, in Arachin, in this first variant, the printed text has Rav Ashi. However, the manuscripts all have Rav Shimi bar Ashi.
3 An example of this would be Ketubot 3a, where the two variants are either that Rava maintains there is, or there is not, ones begittin, a valid claim of duress in a conditional for divorce. Meanwhile, Gittin 34a operates based on Rava maintaining that there’s no such valid claim, so we should follow that variant.
4 Though this might also be true even when the later generation is fourth-generation, I’ve seen it applied in the Rosh to selecting Abaye over Rabba.
5 or ikka dematni areisha / aseifa variants.
6 This was probably an error for “Ramban”, since Rambam, Hilchot Geirushin 8:21 calls it a גֶּט פָּסוּל.
