Last week, we encountered hafka’at kiddushin, abrogation of betrothal, as proposed by Rav Ashi, a sixth-generation Amora and redactor of the Talmud. Let’s consider how this abrogation concept developed diachronically.

The concept appears in six sugyot. Of these, Yevamot 110a is primary, because it has the most named Amoraim (Rav Pappa, Rav Ashi, Ravina) and their interactions are geographically and chronologically plausible.

Beginning with Rav Ashi’s first statement, the labeled sugya reads:

I’ve added 3א based on manuscripts including Munich 95. Note that the justification in 1ג is Hebrew while 1ד is Aramaic. Sometimes such code-switching indicates something added by the Stamma¹.

The next quasi-primary sugya is Bava Batra 48b. If someone extorts another to sell, in many cases the sale is still valid. The Gemara (or likely Rava, fourth generation Pumpedita) asserts that if the extorter specified a particular field, so the seller didn’t exercise any selection, the sale is invalid. The Gemara dismisses this based on Ameimar, a sixth-generation Amora, whose teachers included Rava, whose students include Rav Ashi, and who reestablished and led the academy at Nehardea. Amemar declared that if a man extorted a woman to agree to betrothal, it is nevertheless valid (and the lack of choice is parallel to a particular field). Then:

In 1א, Rav Ashi is replaced with his seventh-generation son, also in Mata Mechasia academy (Sura). 1ב is an Aramaic reaction to the immediate context, but the rest of (1) is identical to what his father said. Both assert that where the man acted improperly in his betrothal, the Sages won’t let it stand and will “improperly” abrogate it. Thus, Mar bar Rav Ashi applies his father’s approach to combat injustice. (2) and (3) are likely copied/transferred from Yevamot 110a because of its relevance. First, why would Ravina and Rav Ashi have the same conversation again? Second, why should Ravina address Rav Ashi rather than Mar?

Next is Yevamot 90b. As discussed previously, Rav Chisda debates Rabba about whether rabbinic law has impact on a biblical level, bringing several proofs. Afterwards, it seems the Talmudic Narrator discusses other proofs, and one as follows. Biblically speaking, if a man sends a bill of divorce (get) to his wife via agent, he can cancel the get. They decreed that he shouldn’t do this. If he nevertheless canceled it, Rabban Shimon ben Gamliel disagrees with Rabbi and says that his cancellation isn’t effective, for otherwise, what’s the force of Bet Din good for? Thus, Biblically the get is canceled, yet rabbinically, we allow her to remarry? In answer, the gemara has:

Here in (1), we don’t have Rav Ashi saying it. Both justification (1ג) and application (1ד) is Aramaic. The justification differs—in betrothing, he acts based on the rabbis’ consent. Yes, he acted improperly here by ignoring the decree not to cancel, but we don’t explicitly note this. Also, his betrothal wasn’t the improper action. The application is different as well, and an extension—in order to effectuate the biblically canceled get, we cancel a perfectly fine betrothal. I believe this anonymous section is Stammaic.

The exchange between Ravina and Rav Ashi seems transferred. (Some manuscripts do have a 3א with אמר ליה.) Why discuss this again? Why discuss after the Talmudic Narrator? Also, the justification is 3ג that betrothal is contingent on rabbinic consent doesn’t leave the same ambiguity of possible hefker bet din hefker, and there shouldn’t be a plausible distinction of monetary vs. carnal betrothal. Therefore, I consider this a secondary sugya².

Gittin 33a is simply a parallel sugya to Yevamot 90b, where we strip the ability to cancel a sent get. It is standalone rather than brought as a Rav Chisda-proof. After the Gemara asks how something rabbinic could override biblical law, in which the get is canceled, we have:

Again, Rav Ashi doesn’t state (1), nor does Rabba (see footnote). A few manuscripts do have אמר ליה for 3א.

In Ketubot 3a, there are two variant sugyot brought in the Talmudic text. In the former, Rava maintains the אונס (unavoidable circumstances) have no legal standing in divorce. Thus, if he gives a get conditionally, that he doesn’t return within 30 days, then tries to return on day 30 but is separated by a river, the get is valid. After considering other reasons for Rava’s rule, the Talmudic Narrator suggests this is Rava’s innovative reasoning, because of modest women who would opt to stay chained and not remarry even where he willfully didn’t return, and because of immodest women who would improperly remarry even when the delay was unavoidable. Therefore, by rabbinic decree, she is legally unmarried. Can rabbinic law supersede biblical (especially within the Pumpeditan framework in which Rava operates)? Then, essentially the same text listed for Gittin 33a, starting with אִין (omitted here for space), that we nullify the original betrothal.

Here, no manuscript has attribution in 1א, because it’s the Talmudic Narrator. Also, far fewer manuscripts have אמר ליה in 3א. This is entirely derivative, and the Ravina/Rav Ashi exchange is transferred. Further evidence is that in the second variant, Rava asserts that אונס does have legal standing in divorce. Most of the sugya repeats, with sources in different attack/defend stance, but the Ravina/Rav Ashi exchange doesn’t appear. If original, the exchange shouldn’t disappear.

Finally, Gittin 73a. A man, deathly ill, gives his wife a get, but then recovers. Rav Huna (72b) declares that there was an implicit condition, so the get is nullified. The Gemara (73a) notes that Rava and Rabba disagree with Rav Huna. They declare the get valid, since otherwise people observing the get canceled would mistakenly think that, if he didn’t recover, the get would take effect after death. There is therefore a rabbinic decree that it is valid regardless. The Gemara asks, can rabbinic law supersede biblical law and render her free? Then, essentially the same text lasted for Gittin 33a, starting with אִין (omitted here for space), though our printed text also has אמר ליה for 3א. Thus, we effectuate the divorce by nullifying the original betrothal. Again, this is entirely the Talmudic narrative, including the transferred Ravina/Rav Ashi exchange.

Taking stock, the original idea was expressed by a name Amora, Rav Ashi, to explain an earlier court ruling. The snatcher betrothed inappropriately, so the Sages can act “inappropriately” and abrogate that betrothal. Rav Ashi was explaining an earlier court ruling, but (contra Rav Pappa) thereby extended it past that one incident. His son, Mar, was faced with an unfair system, where as enunciated by Ameimar, a woman who agreed to betrothal under duress was nevertheless married. Mar extended his father’s principle and said that, here too, the Sages would undo the improper betrothal.

We cannot necessarily extrapolate from Rav Ashi/Mar bar Rav Ashi to agunah cases at large today, where the betrothal was finely conducted, and the husband acts improperly by withholding the get. However, a while back, a husband tried extorting his wife, in divorce proceedings, by claiming he’d accepted a betrothal from an unnamed party on behalf of their minor daughter. He’d reveal who, if the wife caved in to his demands in their divorce proceedings. Abrogating that daughter’s purported betrothal seems like a relatively straightforward extension.

The remaining three cases are Stammaic. In these three cases, the husband issued a get which was in his wife’s possession, but it’s somehow biblically invalid. Named Tanaim/Amoraim nevertheless declare her unmarried. The Talmudic Narrator then channels Rav Ashi anonymously—that we effectuate it by invalidating the initial perfectly fine betrothal, substituting conditioning on rabbinic approval in place of improper action, then quoting the Ravina/Rav Ashi exchange. The pathway was probably first to Yevamot 90b, from there to its Gittin 33a parallel and, because of אין, from there to Gittin 73a and Ketubot 3a.

While a novel extension by the Stamma, I’d note two restrictive aspects. The Talmudic Narrator is always justifying earlier named positions, not applying it to new cases. Also, there’s always a get in the woman’s possession. (Indeed, Rashi to Ketubot 3a says they remove it via that get: שיהיו קיימין קידושין לפי דברי חכמים ויהיו בטילים לפי דברי חכמים על ידי גיטין שהכשירו חכמים.) We cannot necessarily extend from this closed set to new cases, where the husband betrothed properly and there is no get extant. However, in some modern cases, the husband has deposited the get with a deviant bet din so he can remarry, and they’ll grant her the get if she appears before them and caves to his demands. There’s both an extant get and improper action, turning Jewish law into a weapon of oppression. Perhaps a legitimate bet din could legally transfer the get to her possession and apply abrogation. Ask your global orthodox rabbi.

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 Compare Beit Shammai’s retort on Yevamot 107b: פִּישׁוֹן הַגַּמָּל בְּמִדָּה כְּפוּשָׁה מָדַד, לְפִיכָךְ מָדְדוּ לוֹ בְּמִדָּה כְּפוּשָׁה.

2 As discussed last column, several manuscripts, including Munich 95, have אמר ליה for 1א. This wouldn’t be Rav Ashi, or the Talmudic Narrator. It would be voiced by Rabba, in response to Rav Chisda. And indeed, Yerushalmi Gittin 1:4 understands this working as a Suran / Rav Chisda manner. If so, we’d be compelled to reverse the diachronic development, and Yevamot 90b would be earliest, with Rav Ashi taking the approach from Rabba. Indeed, I discovered, after writing all this, that Shamma Friedman, in HaIsha Rabba, argues that all seven proofs (a typological number) are from Rav Chisda, pointing to Eruvin 43a with seven teachings (shev shmayta) from Rav Chisda to Rabba. Accordingly, Rabba did respond here, and the Aramaic portion of Rav Ashi’s statement would be quoting Rabba, with the Hebrew / Aramaic shift revealing this. Personally, I disagree. I don’t think אמר ליה is original, but consider it an easy scribal insertion. The Gittin 33a parallel lacks it. Also, at this point in the proofs, we’ve transitioned to Stamma. Also, the application seems like a forced extension of Rav Ashi’s idea.