Ontic and Epistemic Doubt C: Quantum Collapse in Halakha (Column 324)

In column 322 we discussed the conceptual and scientific distinction between epistemic doubt and ontic doubt. In the previous column (323) we saw the similarity between a state of ontic doubt in quantum theory and in Halakha. It turns out that despite the confusion and lack of understanding evoked by the quantum description, Talmudic commentators use this very distinction. Moreover, in the legal sphere it’s quite easy to understand and feel this puzzling distinction, and I exemplified it in two halakhic contexts: “kiddushin not delivered to consummation” and bererah (retroactive specification).

What we saw there was the phenomenon of ontic doubt, or ambiguity. However, the collapse phenomenon that appears in quantum theory (see col. 322) did not show up in halakhic contexts. At first glance that’s no surprise, since there is no connection between collapse and the principles of uncertainty (ambiguity) in quantum theory. These are two entirely different phenomena: (1) In quantum dynamics, the wave function can be in a superposition of classical states (as in the formulas we saw above). (2) In quantum measurement theory, when a measurement is performed the wave function collapses to one of the pure states that compose it (with a probability determined by that state’s coefficient in the superposition). It’s commonly thought there is no connection between these two phenomena, and each is a novelty in its own right. Therefore, in Halakha, even if we saw phenomenon (1), we need not necessarily encounter phenomenon (2).

But, surprisingly, in this column we’ll see that at least according to some Rishonim, phenomenon (2) also appears in Halakha. The discussion begins at the point of contact between the two topics we met in the previous column: kiddushin not delivered to consummation and bererah.

The dispute between the Ritva and Tosafot Rid[1]

The Ritva on Kiddushin 51a (in the sugya of kiddushin not delivered to consummation), in the manuscript marginalia,[2] raises the following question on the Gemara there:

“Note, in the manuscript marginalia it is written: there is great room for inquiry, for if we say there is bererah, then these are kiddushin which are delivered to consummation from the outset.”

It seems his intent is to point out that according to the view that bererah applies even to biblical law, then in the case of one who betroths one of two sisters this should count as kiddushin delivered to consummation, since he can, the next day, choose one of them to be his wife, and that will clarify that she was his wife from the outset. In doing so he dispels the “quantum ambiguity” and selects a classical state (in which only one of them is his wife), much like collapse in quantum theory.

According to the Ritva’s question, if he subsequently chooses one of them, it is clarified retroactively that she is the one who was betrothed to him. If so, even before he chose, there was already a defined one who was his wife (the one who will be clarified tomorrow), only that now he does not yet know which. Therefore, these kiddushin should seemingly be delivered to consummation. In a sense we can say that on his view this is not ontic doubt but epistemic (for one of them is in fact betrothed, only I do not know which). The Ritva leaves the Gemara as a question, and indeed, on the view that there is bererah (though that is not the halakhic ruling) the Gemara here is not understood.

Tosafot Rid there raises this question as well, and resolves it as follows:

“All agree that where one said: ‘Forty out of eighty (loaves) shall be sanctified,’ they are sanctified. I have seen some ask: but Rabbi Yochanan holds everywhere that there is no bererah—and these are words of nonsense. For when we say that forty out of eighty are sanctified, is it because of bererah? This is similar to one who betroths one of five women and did not specify which—so that they all require a get. So too here, for each and every loaf one may say ‘this one he sanctified,’ and therefore they are all sanctified out of doubt and shall be eaten with the stringencies of thanksgiving loaves. Certainly, if he had said: ‘I will select forty of them and say: these shall be sacred and the rest shall be non-sacred,’ then it would be relevant to say there is bererah. But this we cannot say even according to the one who holds there is bererah, that we remove sanctity into non-sanctity by means of bererah, for there is nothing here to clarify; since he did not designate which he sanctified, they are all in doubt of being sanctified. And also with the five women, one should not rely on bererah and say: ‘This one I choose,’ and it will be clarified that she was the one he betrothed; since at the time of the kiddushin it was in doubt, what bererah is there to say, for he did not know which he betrothed? Certainly, if he had said: ‘Whichever I will want—today or tomorrow,’ and likewise with the thanksgiving loaves—there he could rely on bererah, since he made it dependent on his will. But where he initially betrothed in doubt and did not clarify which he betrothed, there is no bererah to be said.”

He argues that the linkage between our sugya and the rule of bererah is mistaken. If the groom had said, “I betroth the one I will want tomorrow,” then of course, according to the view that there is bererah, he could come tomorrow and choose one of them and it would be clarified retroactively that she is the one betrothed to him and the other is not. Such kiddushin are indeed delivered to consummation (for this is epistemic doubt, since there is a specific one who is betrothed to him, even though at present it is not yet known whom he will choose). But in our sugya he betrothed “one of two sisters” (or one of five women) stam, without adding the definition that his intent is “whichever will be chosen tomorrow.” In such a case there is ontic ambiguity in the kiddushin, as we explained, and therefore one cannot come tomorrow and choose one of them. The ability to choose one of two women is conditioned on having stipulated in advance that one could do so. Plainly, after a halakhic status has been created, it cannot be altered at will. If a person betrothed a woman unconditionally, he cannot add a condition after she is already betrothed to him. He likens this to the impossibility of removing sanctity to non-sanctity after an item has already been consecrated.

Likewise in the examples we saw above, the cases in which bererah is applied are always ones in which the person states in advance (at the time of the act—divorce, or separating terumah) that he performs it on the basis of what will be chosen later. If the statement accompanies the act, it can qualify it and make it dependent on the future choice. But if a person applies an ambiguous status, then that is what takes effect, and what has already taken effect cannot simply be changed later.

At first glance it appears these two Rishonim disagree about whether bererah can be applied where there is ontic ambiguity—that is, where the actor did not explicitly tie his act to a future clarifying event. The Ritva assumes that nevertheless a mechanism of bererah can be applied here, while Tosafot Rid argues that it cannot. For our purposes, Tosafot Rid distinguishes between the sugya of kiddushin not delivered to consummation (where there was no prior stipulation) and the sugya of bererah (which speaks of cases where there was a prior stipulation) and sees them as two separate topics, whereas the Ritva connects them.

As I noted in the footnote above, the Ritva’s words in the Kiddushin sugya apparently do not appear in the original of his novellae and are an external addition from some source. But the Ritva himself, in his novellae to Yevamot 23b, raises this very question:

“If you will say: why are these not delivered to consummation from the outset? For since he said ‘one of them,’ he can clarify which, and it will be clarified that this is the one he betrothed retroactively—at least according to the one who holds there is bererah. And Abaye and Rava hold there is bererah in the chapter ‘Kol ha-Get’.”

But there he also offers three answers. The first:

“One can say that here we speak where he did not clarify even in the end, for example where they hung it on whomever the father or some other person will choose—and that person is not in the world to choose.”

This answer makes an ukimta, that the sugya deals with a case where the person ultimately decided not to choose. In other words, it assumes indeed that one can choose even without stipulating in advance (as we saw in the Ritva on Kiddushin, against Tosafot Rid).

The second answer:

“Alternatively, here it is where he betrothed stam and did not say ‘whichever I will want’ or ‘whichever so-and-so will want,’ and in such a case there is no rule of bererah.”

This is precisely Tosafot Rid’s answer cited above. According to this, he indeed accepts that the status is ambiguous, but he does not accept the possibility of collapse.

And the third answer:

“Alternatively—and this is the main one—that according to the one who holds that kiddushin which are not delivered to consummation are not kiddushin, he derives there from the verse that they must be delivered ‘from the outset’ at the very moment of acquisition.”

He argues here that even if one can choose one of them, it is still a case of kiddushin not delivered to consummation, because at the moment of kiddushin he has not yet chosen and therefore there is a doubt that prevents consummation. When there is no explicit stipulation, the clarification is not retroactive but only from that point forward (so writes the Avnei Miluim below as well, and we will explain this further there). Note that here the Ritva writes that this answer seems correct to him (“and this is the main one”).

An alternative proposal for explaining the Ritva’s words

In principle, one could explain that the Ritva does not intend to dispute Tosafot Rid’s distinction, but that in his view, someone who applies an ambiguous status (as in kiddushin not delivered to consummation) implicitly intends to leave for himself the option to choose one of the women—as if he had explicitly said, “I betroth whichever of them I will want.” On this reading, the disagreement between him and Tosafot Rid would be only about assessing the groom’s intent in such cases; but both would agree that without an initial stipulation one cannot later choose.

But the Ritva’s wording does not sound that way, for in his question he assumes the possibility of choosing as self-evident. If his assumption were an umdana (inference) about the groom’s intent, it does not seem enough to challenge the Gemara on the basis of the logic of yesh bererah. The Gemara could simply assume there is no such umdana, end of story. Indeed, several Acharonim understood that the Ritva holds that even if the groom did not stipulate explicitly or implicitly, the simple assumption is that he has the ability to choose one of them. If so, there is a substantive disagreement between the Ritva and Tosafot Rid. In the column after next I will return to this point and offer a logical argument proving that such an interpretation of the Ritva’s words is impossible.

Comparison to collapse in quantum theory

Above I explained that the ambiguous status created by betrothing one of several women parallels ontic ambiguity in quantum theory. Accordingly, one can continue and argue that the future choice which the Ritva assumes corresponds to collapse due to measurement in quantum theory. Choosing one of the women turns the state of ontic ambiguous status into a sharp (classical) status—i.e., kiddushin that apply to a specific woman. On this view, the groom’s choice of one of the two sisters causes the “wave function” of ambiguity to collapse into a classical state (in which the one chosen is betrothed and the other is the sister of his wife).

What is fascinating here is that according to the Ritva, the very existence of an ambiguous status necessitates a collapse mechanism. Not every source assumes as self-evident that there can be no ambiguous status without a mechanism of collapse to a sharp ontic state. With him this is so strong that he even challenges the Gemara on the strength of this assumption. In his view, even without the actor explicitly stipulating so, the very creation of an ontically ambiguous status requires that there be a collapse mechanism. The coexistence of two possibilities side by side requires the possibility of deciding in favor of one of them. He is unwilling to accept ontic ambiguity as a self-standing state.

Note that this of course stands in contrast to what I described regarding quantum theory. There we saw that collapse is an additional assumption (of measurement theory) independent of quantum dynamics and of ontic ambiguity itself. In physics, the conclusion that there is a collapse mechanism arises from experiments that forced physicists to accept it. It is not a necessary consequence of ontic ambiguity per se. By contrast, according to the Ritva this is an a priori argument. In his view, collapse follows from the mere existence of an ambiguous status. If there is a status that applies to one of two women, that necessarily means that the groom can choose one of them. A vague state, by virtue of being such, is subject to human choice. It seems that on his view we should also have expected in quantum theory (even before experiments showed it) that if there is an ambiguous state, there is likely also a collapse mechanism.

A logical-philosophical note: where exactly is the quantum ambiguity?

It may be that the Ritva’s argument is connected to understanding the state of superposition. In earlier columns I explained that it should not be understood as a simple conjunction; that is, it is not correct to say that “both sisters are betrothed to me (weakly),” or that “every log of the wine is terumah (weakly).” The reason is that those states are impossible: it cannot be that two sisters are both betrothed to me, and likewise it cannot be that all the wine is terumah (we require ‘reshit such that its remainder is noticeable’; see Mishnah Terumot 4:5 and elsewhere). I therefore argued that the more precise formulation is that the general state (the wave function) is a combination of two classical (pure) states, each of which is built in the ordinary (classical) way. Each pure state is a state of the entire system: in terumah there are two log of terumah and the rest is non-sacred; in kiddushin there is one woman who is my wife and the other is my wife’s sister. The overall state is a weighted sum of all the possible pure states.

On the website I was asked where exactly ontic ambiguity appears in quantum theory. I agreed with the questioner that superposition as such does not express ontic ambiguity. Ontic ambiguity is a state in which there is a particle that goes through slit A and also through slit B. But at the superposition stage the object is not a particle but a wave, and there is no problem for a wave to pass through two slits together (no one is bothered by a light wave passing through two slits). If we are not speaking of a particle, there is no room to speak of ambiguity. There is no doubt or ambiguity in such a case, and the object in question is not “maybe wave, maybe particle.” It is a wave, period. So how does ontic ambiguity arise? When a detector is placed, an electron in a state of superposition collapses and becomes an ordinary particle, and then it can be measured as passing through slit A or slit B, and there is no way to know in advance which path it will take (unlike a particle in classical physics). Here—and only here—there is ontic ambiguity: not between particle and wave, but between a particle passing through slit A and a particle passing through slit B. But note that this ambiguity exists only before the detector fires, for when it fires it shows us that the particle passed through a defined slit, just like a classical particle; there is no ambiguity or doubt. That is, when we place a detector we decree that the wave will collapse into a particle, but before the detector indicates where the particle passed there is ontic ambiguity about the particle’s path (since it can pass through either slit). Thus, the ambiguity exists only in the presence of a detector, but only in the state before the detector actually measures. Consequently, it is hard to separate ontic ambiguity from collapse. Without collapse, the superposition would have no meaning of ontic ambiguity (it would just be an ordinary wave). This perhaps reflects the Ritva’s notion, which sees a connection between the two phenomena.

In the halakhic context as well, one can speak of ambiguity only if we relate to the woman as a defined object (a particle, not a wave that is a combination of the two women), about whose status there is doubt (is she the betrothed one or her sister?). So too with the wine. The combination of states (the superposition) means that from the perspective of halakhic status there is an ambiguous state of the woman (each of the two), since it is composed of a sum of two possible pure states. We are dealing with an object that can be consecrated or not, terumah or not, but not both together. Therefore, only the possibility that the wave function collapses into one of the pure states makes the pre-collapse state into ontic ambiguity—precisely as the Ritva assumes. Remember: he is essentially saying that we cannot define a state of ontic ambiguity without speaking of collapse possibilities.

This can also be seen from another angle—via the question whether we are dealing with a conjunction of “and” or a disjunction of “or.” At first glance, superposition is a conjunction of “and,” since the two states exist in parallel (and not only one of them). On the other hand, collapse happens to one of the pure states, and the superposition expresses the possibility that collapse will be “to here or to there.” The initial “and” is what enables the “or” that happens in the end. In light of what is going to happen after the collapse, it is clear that the pre-collapse “and” expresses a combination of “or” (the function is destined to collapse to state [A> or to state [B>). Before collapse we have a combination that can be seen as ontic ambiguity. That combination requires that there be a possibility to collapse to each of the two pure states, and that is precisely the Ritva’s argument: one cannot speak of ontic ambiguity without there being, prior to collapse, an “and”-combination that permits collapse in both directions.[3]

The Avnei Miluim’s distinction

Let us now return to the content of the Ritva’s words. In Shulchan Aruch, Even ha-Ezer 41:2, the law is ruled like Abaye, that kiddushin not delivered to consummation are nonetheless kiddushin:

“If one betroths two women at once, whom it is forbidden to marry together due to a forbidden relation, they are not betrothed. How so? If he betrothed a woman and her daughter, or two sisters at once, saying: ‘Behold, both of you are betrothed to me,’ they are not betrothed and none of them needs a get. But if he betrothed only one of them and did not specify which—e.g. he said to a father: ‘One of your daughters is betrothed to me,’ and the father accepted the kiddushin—or if two or three sisters appointed one of them as an agent for her fellow to receive kiddushin, and he put it in her hand and said to her: ‘One of you is betrothed to me’—they all require a get from him, and it is forbidden to have relations with any one of them, lest she be the sister of his wife.”

The author of Avnei Miluim there (sk. 2) cites the words of Tosafot Rid and the Ritva:

“‘And it is forbidden to have relations with any one of them.’—The Ritva in his novellae (Kiddushin 51a) wrote: ‘There is a great difficulty, for if we say there is bererah, these are kiddushin delivered to consummation from the outset.’ And Tosafot Rid (there) wrote: ‘Certainly, if he had said, “whichever I will want today or tomorrow,” in that he could rely on bererah, since he made it dependent on his will; but where he initially betrothed in doubt and did not clarify which he betrothed, there is no bererah to be said.’”

Now he challenges Tosafot Rid:

“His words are not clarified in this, for certainly according to the one who holds there is bererah, even where he did not make it dependent on his will we also say bererah.”

He claims that according to the view that there is bererah, one can choose one of the women even if the groom said nothing at the time of kiddushin—like the Ritva. It is quite clear that he understands the Ritva as I explained above: that he holds there is bererah even where there was no stipulation (not dependent on intent or implicit bererah).

He brings a proof from the sugya in Yoma:

“As we say in Yoma (55a) concerning: they purchased their bird-offerings in partnership, and one of them died—let us take four zuz and throw them, and the rest will be permitted—on account of bererah.”

A zavah woman must bring a pair of birds (ken) for her offering. The Gemara there discusses a case where two zavot purchased four birds (two pairs) in partnership, without defining which pair belongs to which woman. One of the women died, and now only one pair should be offered, and the other pair is chullin. The Gemara says: take any two birds and redeem the two remaining birds (belonging to the deceased woman) and thereby they are permitted. We see from here that choosing two birds for the living woman clarifies retroactively that from the outset those were her two birds and the others belonged to the one who died—even though there was no explicit stipulation that made it dependent on a future choice. The original plan was that all the birds would be offered together for the two women without distinguishing them at all (i.e., that the ontic ambiguity would remain forever). If so, we see that where there is an ambiguous status there must be the possibility of future bererah (collapse), even where this was not stated explicitly. And this is an explicit statement in the Gemara like the Ritva. Moreover, here it is very hard to interpret this as an implicit stipulation, since the women who purchased the birds certainly did not anticipate that one would die, and therefore did not implicitly stipulate that in such a case they could distinguish the pairs.

He then brings a question from Tosafot in tractate Temurah:

“And in Tosafot, Temurah 30a s.v. ‘Ve-Idach Le’ishteri’, they ask: if so, when an issur fell into heter (a forbidden piece fell into permitted ones), let us cast one aside and permit the rest, etc.”

Tosafot ask: why, in a case where a forbidden piece of meat falls into a mixture of permitted pieces (a case where there is no rule of nullification, of course, and then prohibition falls upon the whole mixture), can we not take one piece out and thereby clarify retroactively that this is the forbidden piece and discard it, and the rest will be permitted? If indeed there is bererah even where it was not stipulated in advance (as we see in Yoma), then in such a mixture there should be this way out as well.

Tosafot answer:

“And my teacher, the RMR, says a great general rule in this matter: certainly, any case where the prohibition was clarified at the outset and afterward it became mixed with the permitted—we do not rely [on bererah], for the mixture is with an issur. But those mixtures where the mixture was initially in the permitted—since the issur was not clarified before the mixture, and after the mixture the issur was generated—we rely on bererah. And if so, likewise here in ‘he betrothed one of two sisters’: since the issur was not known at first but only through the mixture—let us say bererah.”

The RMR distinguishes between a piece that was known and only afterward got mixed in, and a case where the doubt arose ab initio when there was already a state of mixture (as with the bird-pairs). It is quite reasonable that his intent is to explain that if we are dealing with a piece that was already prohibited earlier and only thereafter got mixed in, this is epistemic doubt and not ontic (like a woman who was betrothed and afterward we forgot who she is—i.e., the betrothed woman, from our perspective, is mixed among the others). And in epistemic doubt there is no possibility of bererah. But with the bird-pairs not only did the problem begin in a state of mixture, but it is a case of ontic ambiguity and not epistemic doubt (no pair truly “belongs” to a specific woman; it is not our lack of knowledge, but ambiguity in reality itself), and therefore there one can clarify by a future choice.

Thus, both in Yoma and in Tosafot in Temurah we see like the Ritva: whenever an ontically ambiguous status arises, there must be a mechanism of retroactive clarification even if the person did not explicitly stipulate it.

But now we face a difficulty. For if the Ritva is indeed correct and every ambiguous status must have a clarification mechanism as we saw, then it is not clear why, in one who betroths one of two sisters, these are kiddushin not delivered to consummation. After all, he can choose one of them and thus clarify that she was betrothed to him from the outset, and thereby solve the problem. That was the Ritva’s question on our sugya, which he did not resolve (though we saw that in his novellae to Yevamot he does offer two answers).

The author of the Avnei Miluim resolves the difficulty thus:

“It seems to me that since he did not say ‘whichever I will want today or tomorrow,’ and he initially betrothed in doubt—then at the moment of kiddushin there were no kiddushin, since they were not delivered to consummation. And since it did not take effect at the time of kiddushin, even though afterward he clarifies one of them, we no longer say this is the same kiddushin he performed, since they were not kiddushin in their time. But if he says ‘whichever I will want’—the kiddushin will be according to whichever he will want—then at the time of kiddushin he made the matter dependent on his will; thus there are kiddushin.”

He distinguishes between two kinds of yesh bererah. Where one says explicitly that the status is dependent on a future clarifying act, there the clarification is retroactive. That is, if one explicitly says he is betrothing whichever sister he will choose tomorrow, then the mechanism of bererah causes the future choice to determine that she was his wife from the outset (i.e., already from the moment of the kiddushin themselves), and therefore this is essentially a case in which a specific woman is the one betrothed to him. It is true that at the time of kiddushin he has not yet chosen anyone, but this is already epistemic doubt, and in epistemic doubt such kiddushin are delivered to consummation. By contrast, if one betroths one of two sisters without explicitly saying that he makes it dependent on a future choice, then indeed there is bererah (as he proved above), but the bererah does not operate retroactively. When he later chooses one of the two sisters, she will indeed be betrothed to him and the other will be clarified to be the sister of his wife—but this happens only at the time of his choice and not retroactively from the time of kiddushin. But if it does not occur retroactively from the time of kiddushin, then at the time of kiddushin they were still not delivered to consummation (and consequently, now it is clear he cannot choose any of them later). This is exactly the Ritva’s third answer in Yevamot cited above.

He adds that the same holds in the case of the log of wine for terumah:

“Similarly we wrote in Ketzot ha-Choshen 61:3 concerning ‘two log that I am about to separate’—that there the remainder is noticeable at the time of separation, since he made the terumah dependent on what he would separate, and what he will separate will have a noticeable remainder. But if he says ‘two log within it are terumah,’ although he is permitted to drink relying on bererah because the prohibition was not recognized (as Tosafot write in Temurah 30a that in such a case we say bererah)—in any case, at the time of separation, since he did not say ‘that I am about to separate,’ it does not take effect, since at the moment of separation there was no noticeable remainder; and afterward bererah no longer helps.”

That is, such a distinction also exists in the case of separating terumah. If he explicitly says that the terumah will be the two log that remain at the end of the drinking (this is the case discussed in the sugya there), then the clarification indeed takes place retroactively, since he explicitly stipulated it. Once the clarification occurs retroactively, it means that the terumah in the barrel has a noticeable remainder—i.e., there are two specific log, even if we do not know them. In our terms: this is epistemic, not ontic, doubt. In such a case the clarification is retroactive. But if he said “two log within the barrel shall be terumah” without defining a future event as clarifying, then here there is ontic ambiguity, and therefore the two log that remain at the end might be the terumah, but this is terumah whose remainder is not noticeable, since it does not take effect retroactively (because there was no explicit stipulation). Therefore one cannot, in such a case, drink the wine relying on that separation. Consequently, one never reaches a situation where two log remain, and the entire process fails—just like with kiddushin.

At the end he returns to the sugya of kiddushin not delivered to consummation:

“And similarly here: since they do not take effect at the time of kiddushin—because they were not delivered to consummation—bererah no longer helps afterward. According to this, for us—who rule that kiddushin not delivered to consummation are kiddushin, only that relations are forbidden with any of them due to the doubt of ‘the sister of one’s wife’—bererah does help as to the prohibition, as Tosafot write in Temurah, that where the prohibition was not recognized, bererah helps. But in biblical matters we hold that there is no bererah. Ponder this.”

He explains that according to Rava—who holds that kiddushin not delivered to consummation are not kiddushin—then without explicit wording there is no retroactive clarification, and therefore at the time of kiddushin they are not delivered to consummation and do not take effect; consequently, even if he later chooses one of them it will not help. Whereas if he says so explicitly in advance, then these are kiddushin delivered to consummation as we explained.

But we must remember that the halakha is ruled like Abaye, and according to him such kiddushin are valid kiddushin, only that it is forbidden to have relations with any of the sisters because each is possibly the sister of his wife (this does not prevent the kiddushin from taking effect, and therefore he must give a get to each of them). If so, according to Abaye, a future clarification will be effective even without explicit wording at the time of kiddushin, though without explicit wording the clarification does not occur retroactively (as we saw). Why? Because from the moment he chooses one of them she becomes his wife (not retroactively, but from then on), and the other is her sister. Accordingly, from then on it is permitted to have relations with the chosen one, and her sister is forbidden as the sister of one’s wife. Therefore, he need not divorce either of them; he can live with the one he chose, who thereby becomes his wife, and the other becomes the sister of his wife—no get required.

It emerges from his words that when there is explicit wording at the time the ambiguous status is applied, then—according to the view that there is bererah—there is a collapse mechanism retroactively (and if there is no bererah, as we rule in practice, then not). But when there is no explicit wording, then bererah helps only from then on; that is, a collapse occurs, but in such a case it does not operate retroactively. Note that on the Avnei Miluim’s view it comes out that if one betroths one of two sisters, then according to Abaye he does not need to give a get to both. He can choose one of them and live with her from now on. Recall that the Rambam and the Shulchan Aruch (whose language we quoted above) explicitly ruled that he must give a get to both and that relations are forbidden with any of them, since each may be the sister of his wife. It may be that all the Avnei Miluim means is only according to the view that there is bererah; but if there is no bererah, then bererah does not help even from now on—though it is commonly said that bererah “from now on” exists according to all views (even according to those who hold there is no bererah).[4]

[1] The discussion here is taken from chapter 3 of the book Vague Logic in the Talmud, the twelfth book in the series Talmudic Logic.

[2] In the Mossad HaRav Kook edition, note 24 says that it seems this passage did not come from the Ritva himself but was inserted here by the printer as a reference to the words of Tosafot Rid that will follow. However, we will see below that the Ritva himself raises this question in his novellae to Yevamot 23b and even offers three answers to it (the second of which is the answer of Tosafot Rid cited here). Note also that the Pnei Yehoshua in his novellae here (51a) elaborates greatly with various versions of this question.

[3] In Bava Metzia 94b and parallels there is a dispute between R. Yonatan and R. Yoshiya regarding the vav of conjunction in the Torah:

“As it was taught: ‘A man who curses his father and his mother’—I have only [when he curses] his father and his mother [together]; how do I know [he is liable for] his father without his mother, and his mother without his father? Scripture says: ‘His father and his mother he cursed’—‘his father he cursed, his mother he cursed’—the words of R. Yoshiya. R. Yonatan says: it implies both together, and it implies each one by itself, until Scripture specifies ‘together.’”

According to R. Yonatan, when two things are written in the Torah with a vav of conjunction, the meaning is: either both together or each separately. According to R. Yoshiya, it means only both together (unless there is another derivation that divides them; see there). The difference between “or” and “and” is often very subtle. When you say “his father and his mother,” your intent is to a combination of father and mother, but the combination can also take the form of “or” (the verse intends both—that is, either this or that). In our case we see a special structure that explains the relationship between the two: when there is a wave function conjoining them in the mode of “and,” it enables collapse to each of them in the mode of “or.” The “and” contains within it the possibility of each one separately.

[4] In Responsa Rashba II (Toldot Adam) §82, he writes that the discussion about bererah applies only where retroactive clarification is required (see Beit Yishai §63 and the beginning of “Ma‘arechet ha-Kinyanim” at the end of Sha‘arei Yosher, who discuss this response).