Ontic and Epistemic Doubt B: Halakhic Aspects (Column 323)

Article Contents

With God’s help

Ontic and Epistemic Uncertainty II: Aspects of Jewish Law

In the previous column we examined the conceptual and scientific distinction between epistemic uncertainty and ontic uncertainty. Despite the confusion and incomprehension that the quantum description arouses in us, it turns out that commentators on the Talmud use this very same distinction. True, they do so in the legal sphere rather than the physical one, and as we shall now see, in the legal sphere it is very easy to understand and sharpen this confusing distinction. In this column I want to demonstrate this in two contexts of Jewish law: betrothal that does not permit intercourse (see my lectures on this here and in the twelfth volume of the Talmudic Logic series) and retroactive clarification (see the fourth volume in that same series). In the next column I will discuss the connection between these two issues, insofar as there is one, and we will see there an interesting implication for the phenomenon of quantum collapse (discussed in the previous column in the scientific context) in Jewish law.

Epistemic Uncertainty in Jewish Law

Jewish law deals quite a bit with situations of uncertainty. For example, a case in which two pieces of meat lie before me, one pork and the other kosher meat, and I must decide whether to eat one of them, and if so which one. This is a case of uncertainty, and it is decided according to the rules for adjudicating uncertainty. Even if I found a piece of meat lying on the sidewalk in the center of town, there is room for uncertainty as to whether it is kosher meat or not. One indication that may help us decide the uncertainty is a survey of all the stores that sell meat in that city: if most of them sell kosher meat, then strictly speaking it is permitted to eat the meat I found. This is an example of a decision rule for cases of uncertainty (the rule of following the majority).

In the Mishnah, Yevamot 118b, we find another case of uncertainty:

If a man betrothed one of five women and does not know which one he betrothed, and each says, ‘He betrothed me,’ he gives a bill of divorce to each and every one, and places the marriage settlement among them and withdraws; these are the words of Rabbi Tarfon. Rabbi Akiva says: This is not a way to remove oneself from transgression, until he gives a bill of divorce and the marriage settlement to each and every one.

If a man robbed one of five people and does not know from which of them he robbed, and each says, ‘He robbed me,’ he places the stolen item among them and withdraws; these are the words of Rabbi Tarfon. Rabbi Akiva says: This is not a way to remove oneself from transgression, until he pays each and every one.

Here we have a situation in which a man betrothed a certain woman but does not remember which of the five he betrothed. A similar case can arise when a man sends an agent to choose a wife for his son from a certain city (like Eliezer, Abraham’s servant, whom his master asked to seek a wife for his son Isaac in Haran), and the agent betrothed some woman and then died. For the sake of the discussion, let us assume that the woman’s father, who received the betrothal on her behalf, also died. Now the intended husband, and indeed no one else in the world, knows who his wife is. In such a case all the women are treated, out of uncertainty, as his wives, and he is forbidden to marry the sister of any one of them, and all of them require a bill of divorce before they may marry someone else. This too is a case of uncertainty, and the laws of uncertainty apply to it. The second case in that Mishnah deals with someone who robbed and does not remember whom he robbed. Here as well the situation is very similar. The tannaitic sages disagree as to whether he may place the object among them and withdraw, or whether he must return the stolen item to each of them in order to be certain.

These two situations are similar to the uncertainty regarding the pieces of meat. The factual state of affairs is single and fixed; we simply do not know it. God does know, because He knows everything, but human beings have limited information. Therefore, from our standpoint, it is either kosher meat or forbidden meat. Either it is a woman who was betrothed or it is not. All these cases reflect a lack of information on the part of the person, not in the world itself—that is, epistemic uncertainty.

Betrothal That Does Not Permit Intercourse: The Dispute between Abaye and Rava

By contrast, in the sugya of ‘betrothal that does not permit intercourse,’ Kiddushin 51a-b, there is a discussion about a man who comes to the father of two daughters and gives him money in order to betroth one of his two daughters, without specifying which one. Notice that here we are already dealing with sisters, not merely with two unrelated women. The problem that arises here is that if Rachel is the one betrothed to me, then Leah is my wife’s sister and I am forbidden to have relations with her. And if Leah is the one betrothed to me, then Rachel is my wife’s sister and I am forbidden to have relations with her. In short, after I have done this disgraceful thing, I find myself in a situation where I cannot have marital relations with either of them, since each is possibly my wife’s sister. This is the situation that Jewish law calls ‘betrothal that does not permit intercourse.’

What is to be done in such a case? Abaye and Rava disagree there in the Gemara:

It was stated: With regard to betrothal that does not permit intercourse, Abaye said: It is valid betrothal. Rava said: It is not valid betrothal. Rava said: Bar Ahina explained it to me: +Deut. 24+ ‘When a man takes a wife and has relations with her’—betrothal that permits intercourse is betrothal; betrothal that does not permit intercourse is not betrothal.

According to Abaye, this betrothal does take effect, and both women require a bill of divorce. Even according to his view, the man is forbidden to both of them, since each is his wife’s sister and relations with her would involve a possible prohibition punishable by karet. But in his view that does not prevent the betrothal from taking effect. Because it cannot be physically realized, the man must give both of these women a bill of divorce out of uncertainty and release them. According to Abaye, this situation is similar to the earlier case of one who betroths one of five women and forgets whom he betrothed. Here too we have a case of uncertainty, and the laws of uncertainty apply. But Rava holds that this is not betrothal at all, because it does not permit intercourse. Therefore, in his view, both women are permitted to others without any bill of divorce. I emphasize again that both agree that this betrothal does not permit intercourse, and their dispute is only over the question whether such a situation prevents the betrothal from taking effect. Let us note that this is one of the six exceptional cases in which the law is decided in accordance with Abaye (this is the q-case in the mnemonic YAL KGM).

What Is Betrothal That Does Not Permit Intercourse?

Before proceeding, let us clarify the concept of ‘betrothal that does not permit intercourse’ itself. At first glance, this means betrothal that cannot be realized in physical relations. That indeed seems to follow from Rashi’s words there, where he wrote:

‘When a man takes a wife’—that is betrothal by money, which we derive (above, Kiddushin 2a) through the verbal analogy of ‘taking’; and it is written ‘and has relations with her,’ from which we require that she be fit for intercourse.

From the verse we learn that betrothal must be fit for intercourse. But according to this, it follows that in Abaye’s view there is no such requirement. If so, it is not clear why betrothal does not take effect in unions prohibited on pain of karet (for example, one who betroths his sister or his mother). True, one cannot have relations with one’s mother or sister, but according to Abaye that should not prevent the betrothal from taking effect. Let me sharpen the point further. In this very sugya it is stated that if a man betroths Rachel and afterward her sister Leah, then according to everyone, including Abaye, Leah is not betrothed, because she is his wife’s sister and betrothal does not take effect in unions prohibited on pain of karet. The dispute concerns only a situation in which he betrothed one of the two, that is, both stood before him simultaneously. And even here, if he betrothed both together, everyone agrees that the betrothal does not take effect. The dispute is only in the case where he betroths one of the two without specifying which.

Tosafot there cites Rashi’s words and raises against him the opposite objection. Let us first note that in unions prohibited on pain of karet, betrothal does not take effect according to everyone, as we saw above. But with regard to unions prohibited merely by an ordinary negative commandment (lighter marital prohibitions outside the incest prohibitions—for example, a priest and a divorcée), the tannaitic sages disagreed whether betrothal can take effect, and the legal ruling is that it can. Tosafot, s.v. ‘Kiddushin,’ 51a, raises an objection to Rashi from here:

With regard to betrothal that does not permit intercourse, Rava said it is not betrothal—Rashi explained that this is because it is written ‘and has relations with her,’ so we require that they be fit for intercourse. But this is difficult, for those prohibited by an ordinary negative commandment are also not fit for intercourse, and yet betrothal does take effect with them.

Tosafot objects that according to Rashi’s approach, even in unions prohibited by an ordinary negative commandment the betrothal should not take effect according to Rava. After all, the physical ability to have relations exists both in more severe prohibitions and in ordinary negative-commandment prohibitions, so it is clear that the relevant condition is permission to have relations, not the mere physical possibility of doing so. If so, why is there a difference between unions prohibited on pain of karet and unions prohibited merely by a negative commandment? Both are forbidden by law, and therefore do not permit intercourse.

Therefore Tosafot himself explains the problem differently:

Therefore it seems that ‘betrothal that does not permit intercourse’ means that the prohibition of intercourse is brought about by the betrothal itself. That is the case when he betroths one of two sisters without specification: before the betrothal, each of them was permitted to him, and now, through the betrothal, both have become forbidden. But in unions prohibited by a negative commandment, the prohibition of intercourse was not brought about by the betrothal, for even beforehand they were forbidden to him for intercourse just as they are after the betrothal.

According to Tosafot, the problem is not that the betrothal cannot be realized, but that instead of permitting intercourse it forbids it.

And so too in the Rashba there, 51a, who wrote:

Betrothal that does not permit intercourse is not valid betrothal. And this is specifically where the betrothal completes the prohibition; but in unions prohibited by a negative commandment and those prohibited by a positive commandment, although they do not permit intercourse, they are valid betrothal.

Contrary to the simple understanding that emerges from Rashi’s words, Tosafot and the Rashba argue that the problem lies in the act of betrothal itself, not in its result. The resultant problem of a betrothal that cannot be realized in practice is another question altogether (discussed in connection with a wedding to a menstruant woman, and as a matter of law, according to most authorities the betrothal takes effect). Here the issue is fundamentally different.

It seems to me that according to Tosafot and the Rashba, ‘betrothal that does not permit intercourse’ is not a principle saying that betrothal must be consummated, but rather a principle that the act of betrothal must pave the way to marital relations and not block it. If the act of betrothal, instead of permitting the woman to her husband, brings it about that she is forbidden to him, then it is not an act of betrothal. One may compare this to divorce that lacks ‘severance.’ The Gemara states in several places that divorce that does not permit the woman to others is not divorce. For example, divorce on a condition that continues to accompany the woman forever: ‘This is your bill of divorce on condition that you never drink wine.’ Since in such a divorce the woman never becomes fully released from her husband (because there is always the possibility that she will drink wine, in which case it will turn out that she is still married to him), such a divorce is not a divorce. The explanation is that the point of divorce is to permit the woman and sever her bond with her husband. If it does not do that, it is not divorce. Here we see that betrothal too requires something opposite to severance: betrothal must connect a woman to her husband, and if it itself creates a rupture between them, it is not betrothal.

The practical difference appears when there is some prohibition unrelated to the act of betrothal that blocks the path to intercourse. In that situation, according to Tosafot and the Rashba, the betrothal takes effect even according to Rava (unless we are dealing with unions prohibited on pain of karet, where the problem is the prohibition itself and not the lack of ability to have relations). The reason is that there is no requirement that the betrothal be capable of being realized. The act of betrothal did not itself create a prohibition where there had been permission before (the prohibition was already there, independent of the act of betrothal). In such a case the betrothal takes effect.

What shall we say according to Rashi? Could he be denying the accepted rule that betrothal does not take effect in unions prohibited on pain of karet according to Abaye, or that betrothal does take effect in unions prohibited by a negative commandment according to Rava? That is unlikely. These are agreed laws. It may be that in his view incest prohibitions do belong to this category, whereas ordinary negative-commandment prohibitions do not. The reason is that prohibitions by ordinary negative commandment are lighter, and therefore this is not regarded as a woman with whom intercourse is impossible. It is merely forbidden; the obstacle is external. By contrast, in incest prohibitions the woman is altogether outside the sphere, not merely subject to a prohibition. The severity of the incest prohibition creates a situation in which it is as though the act of intercourse is not factually possible at all. But it is still not clear why Abaye does not disagree and maintain that betrothal takes effect even in unions prohibited on pain of karet. True, this does not permit intercourse, but on his view such a state does not prevent the betrothal. It may be that Rashi really meant what Tosafot and the Rashba say, but they did not understand him that way.

Betrothal That Does Not Permit Intercourse: Ontic or Epistemic Uncertainty?

At first glance, this too seems to be an ordinary case of uncertainty, since it is not known which of the two sisters was betrothed to him, and the laws of uncertainty should apply. But that seems incorrect on its face. First, in a case of three sisters one might ask why we should not nullify the one woman in the majority and permit all of them to him out of uncertainty. True, this question also arises regarding the sugya of one who betroths one of five women who are not sisters. Seemingly he should be able to release all of them without a bill of divorce. But perhaps there it would involve mutually contradictory rulings, like the case of the two paths. Moreover, in our case this leads to mutually contradictory rulings, for it would thereby permit two sisters to him simultaneously. But there still remains the possibility that he choose one and live with her, and not with both together, in which case there is no contradiction. Seemingly, this proves that we are dealing here with ontic rather than epistemic uncertainty, and therefore the ordinary laws of uncertainty do not apply. Below we will see this in greater detail, but before that I want to challenge the conception of the uncertainty here as epistemic from a different angle.

As is well known, the medieval authorities disagree regarding the rule that biblical uncertainty is treated stringently. The Rashba holds that this rule itself is biblical, whereas Maimonides holds that it is a rule instituted by the sages, that is, rabbinic rather than biblical. If so, according to Maimonides, in cases of biblical uncertainty one may be lenient on the biblical level, and only the sages imposed stringency. If that is so, then according to Maimonides it is not clear why, when a man betroths one of two sisters, this should count as betrothal that does not permit intercourse. After all, each of them is only possibly his wife’s sister, and according to Maimonides, in biblical uncertainty one may be lenient on the biblical level; that means that biblically he may have relations with her. According to his view, the prohibition of relations with her is only rabbinic, and therefore on the biblical level this should have been a betrothal that does permit intercourse. This difficulty exists not only according to Rava but also according to Abaye, for as I noted above, Abaye too agrees that he is forbidden to have relations with them, meaning that the betrothal indeed does not permit intercourse. He only argues, unlike Rava, that such betrothal nevertheless takes effect (though of course he may not have relations with them and he must give both of them a bill of divorce).

This is a first hint that the uncertainty here has a different character. Here even Maimonides would agree that it is biblically forbidden to have relations with either of them, and the ordinary laws of uncertainty do not apply. As we shall now see, the reason is that the uncertainty in this case is fundamentally different from the ordinary legal uncertainties described above. The ordinary uncertainties are epistemic uncertainties (absence of information), whereas here we are dealing with ontic uncertainty.

The one who formulated this clearly and sharply was R. Shimon Shkop, in __Sha’arei Yosher __Gate 3, chapter 22:

It seems to me, in my humble opinion, that in the case where one betrothed one of five women… and likewise where one consecrated a perutah coin within a purse, there is here a profound inquiry as to the manner in which the legal effect takes hold upon one of them.

For when one betroths one woman out of five, and neither the betrothing party nor the one being betrothed specified which of them was to be designated for betrothal, how is it possible to say that in actual reality one specific woman was designated for betrothal? By what reason or cause would Heaven single out one woman so that she should be the one betrothed and designated more than the other women?

Rather, it seems more reasonable to say that in such a case, even in actual reality, no specific woman was designated for betrothal. And in such a case it is not relevant to say, ‘if Elijah were to come.’ And when we say that each one is forbidden because of uncertainty, this is not like ordinary uncertainties in general, for the whole concept of uncertainty is that perhaps this is the definite reality; but here the point is that the entire legal effect exists because of the cause that brings it about.

When a man betroths one woman out of five and later forgets who she is, that is an ordinary epistemic uncertainty. There is one defined woman who is betrothed to him, and only he does not know who she is. God does know, because the information exists. But when he betroths one of five women and does not define at all who it is, in such a case one cannot say that there is one woman who is in truth betrothed to him, for Heaven did not designate any one of them to be betrothed to him. It is clear that such a case is ontic uncertainty, meaning that there is no specific woman who is betrothed to him. Even God Himself, if asked, would not be able to tell us which of them is betrothed. In the first kind of uncertainty there is one correct answer that God knows, only I do not know it, and therefore it is epistemic uncertainty. By contrast, in our case there is no correct answer at all. The problem here is not lack of information on the part of the person, as in all the ordinary uncertainties of Jewish law, but ambiguity or non-determinacy in reality itself. This is ontic uncertainty, or ambiguity.

And so he goes on to explain there:

And the cause of the legal effect of betrothal is the act of betrothal—the giving of the money and the declaration. Since he gave the money and said, ‘One of these five women shall be betrothed to me,’ whereby it is fitting that one of them should attain the legal status of a married woman, by virtue of this each one is forbidden because of the act, and not because we are uncertain whether she is the betrothed one more than the four remaining women; and so too in every similar case…

What generates this situation is the giving of the betrothal money. Once the money has been given, a legal effect of betrothal has indeed been created here, but there is no specific woman with respect to whom that effect has taken hold. Therefore each of the five women is forbidden to marry someone else, because there is a concern that she is a married woman. And if those five are sisters, then both are forbidden to the one who betrothed them, because each is, so to speak, his wife’s sister.

But the term ‘uncertainty,’ which I used here, is not accurate with respect to our case. This is not uncertainty:

In every such case, the category of uncertainty does not mean that perhaps this is the definite reality, for in such a case there is no definite reality even in actual existence. Rather, for purposes of practical conduct we must treat each one as having the law of uncertainty because of the cause that generates the law—that one of them must be the one, though it is not determined who that one is, and all are fit to be that one.

In all these cases—that is, cases of ontic uncertainty—there is not really a state of uncertainty at all, because no information is missing. I possess all the relevant information, just as God does. And yet several possibilities remain. If so, this is not a lack of information but several possibilities in reality itself. Reality itself is ambiguous, that is, it is not in a sharp and fixed state. In yeshiva parlance this is called ‘certain uncertainty,’ as opposed to ‘uncertain certainty,’ which describes epistemic uncertainty. I am certain that reality itself is indeterminate, as opposed to epistemic uncertainty, where I am uncertain about a reality that is in itself fixed and clear.

Indeed, in conclusion it seems at first glance that the laws of uncertainty are applied here:

Therefore the law here is like the law of uncertainty, because even before Heaven it is not revealed that one is certainly forbidden and the rest certainly permitted, for in actual reality as well they are all equal and there is no difference between this one and that one.

But that is not quite precise. The prohibition upon the one who betrothed them with respect to each of the sisters is not doubtful but certain. Each of them is certainly his wife’s sister, except that she is the sister of his tenuous wife. Therefore with respect to each of them there is a certain but tenuous prohibition, not a doubtful prohibition. From this it is clear that even according to Maimonides there is no possibility of relations with either of them, for each is certainly forbidden to him, though by a tenuous prohibition.

Shkop too sensed this difference, and points to another distinction between these two kinds of uncertainty—uncertainty and ambiguity:

Accordingly, in such a case it is not relevant to speak of nullification by majority, for only where the minority is distinct from the majority did the Torah say, ‘incline after the many.’ But in such a case, where all are equal even in actual reality, there is here no minority and majority, for they are all forbidden because of the cause that generates the law—that one of them is fit to be forbidden, though it is not clarified who that one is. Which of them will you exclude and which will you include?

He argues that here too there is no law of nullification by majority (the first question I raised above). In an ordinary mixture of one forbidden piece among several permitted pieces, there is one forbidden piece that is nullified within the majority of permitted pieces. But in a case of ontic uncertainty there is no difference at all between one woman and another. There is no forbidden woman and permitted woman; rather, they are all in the same state. In such a situation one cannot speak of nullification by majority, because there is no majority and minority here at all. All the women here are in an identical state—not only from the standpoint of our information about them, but in reality itself. One may formulate this differently: nullification by majority is stated when there is uncertainty, as in following the majority. But in a case of certainty—and here with respect to each woman her status is entirely certain, for she is the betrother’s tenuous wife—the rule of nullification by majority was never stated.^[1]

The Connection to Quantum Theory

In the previous column we examined situations in quantum theory in which ambiguity appears, that is, ontic uncertainty (as opposed to chaos, where we have uncertainty rather than ambiguity, that is, epistemic uncertainty, or lack of information). It is interesting to note that Shkop wrote these words at roughly the same period in which quantum theory was coming into being in physics. I assume that he had never even heard of it, and certainly did not know it. Perhaps the Zeitgeist, the spirit of the age, led him to formulate a similar distinction in the Talmudic context, and as we shall now see, precisely there this distinction is much easier to digest, and it can even clarify for us something of the mystery of quantum theory.

To understand this, note that the situation in betrothal that does not permit intercourse is completely identical to the situation in the double-slit experiment described in the previous column. There we described a case in which the particle passes through both slits, and we explained that this is not meant literally. The state of the particle is a sum, a superposition, of two ‘pure’ classical states, in one of which it passes through slit A and in the other through slit B. We described this formally as follows:

For our purposes, state [A> is the state in which Rachel is my wife and Leah is my wife’s sister (the particle passed through slit A and not through B), and state [B> is the state in which Leah is my wife and Rachel is my wife’s sister (the particle passed through slit B and not through A). The state after the betrothal is the state , that is, the superposition between these two states. Note that here it is already self-evident that the meaning of this state is not that both women are married to me, since it is impossible for a man to be married to two sisters simultaneously, not even in tenuous marriage. The combination being made here is not between the women but between the states. The overall state of me and the two women, in terms of our personal legal status, is a combination of two classical personal-status states, in each of which only one of the sisters is married to me. Each classical, pure state is itself a state of the three of us: state [A> describes me as married to Rachel and Leah as unmarried, though she is my wife’s sister. State [B> describes me as married to Leah and Rachel as unmarried, though she is my wife’s sister. The overall state of the three of us is a sum of two states, each of which is itself a state of the three of us.

Therefore it is not correct to say that each of them is married to me in a tenuous way, for then both of them would be married to me together, and that is impossible. The more correct formulation is that each is married to me with probability one-half, that is, each of the two classical states has probability one-half. But note that here the probability is not the chance that this is the correct state, because there is no one ‘correct’ state. Rather, it is the weight that this state has in the superposition that constitutes the general state. This is ontic rather than epistemic probability, exactly as in quantum theory (in the case of betrothal that does not permit intercourse, it is clear that the coefficients α and β are equal).

It is worth noticing that in the legal-halakhic case it is very easy to understand this state and its implications, although in quantum theory this very same description is highly confusing. It is very difficult to explain to people or to students the difference between quantum uncertainty or indeterminacy, which is ontic ambiguity, and a state of classical uncertainty, which is epistemic uncertainty. Thus, for example, in the previous column it was hard for me to explain why it is not correct to say that the particle passed through both slits together. I said there that its wave function is composed of functions that pass through the two slits, and the wave function describes the ontic probability. Here, precisely because we are dealing not with physics but with a normative-legal world, there is no difficulty explaining it or grasping it.^[2] It is obvious to everyone that it is impossible for both women to be married to me, and therefore it must be that what we have here is a superposition, and that it is not identical to a state of tenuous marriage to both of them. If you try now to read the previous column again, I suspect it will be easier to understand.

Collapse of the Wave Function

Interestingly, as we saw in the previous column, in quantum theory the probabilities α and β represented in the wave function have yet another meaning. First, they describe the composition of the general wave function, as can be seen from the formula above. But we saw that when one places a detector near one of the slits, in such a case the wave function collapses: the particle behaves in a civilized, classical manner and passes only through one slit. In that situation too, the probabilities have meaning. These coefficients describe the chance of obtaining a given outcome in the experiment. If these coefficients are equal, then the chance that the detector will signal that the particle passed through slit A, and not through B, is equal to the chance that it did not pass there, but rather through B.

Thus these coefficients determine both the composition of the compound wave function and the chance that we will obtain each result in a measurement, that is, the chance that the wave function will collapse into each of its two classical components. In this sense, if we had the ability to measure which of the two women is married to the one who betrothed them, then the coefficients of the superposition would determine which of them would be married to him. In our case the coefficients are equal, and therefore we have a state of balanced ontic uncertainty—more precisely, uniform ambiguity.^[3] In the next column we will deal with the phenomenon of collapse and its appearance in Jewish law. But before that I must preface it with a discussion of the second sugya in which we see a state of ontic uncertainty: the sugya of ‘retroactive clarification.’

The Sugya of ‘Retroactive Clarification’

There is a whole collection of Talmudic sugyot that deal with the law of ‘retroactive clarification,’ in which, at least according to the interpretation of some of the medieval authorities, one can see the same phenomenon. We may demonstrate this through an example that recurs in many discussions throughout the Talmud (Gittin 25a, Hullin 14a, and elsewhere), dealing with the separation of the priestly portion:

For it was taught: One who buys wine from the Samaritans says, ‘The two log that I am going to separate are hereby the priestly portion, ten are first tithe, nine are second tithe,’ and he redeems it and drinks immediately—these are the words of Rabbi Meir. Rabbi Yehuda, Rabbi Yosei, and Rabbi Shimon prohibit it.

The case is that of a person who bought a barrel containing one hundred log of wine from which the priestly portions and tithes had not yet been separated. He wants to drink the wine without separating them, for example because it is the Sabbath and one may not separate priestly portions and tithes on the Sabbath. In the case before us, that person says on Sabbath morning that the two log that will remain on Saturday night, one-fiftieth, which is the standard average measure, will be the priestly portion, and all the rest will be ordinary wine, and he wants to drink the wine immediately now on the Sabbath on the basis of that separation.

The tannaitic sages disagreed over whether he may drink the wine on the basis of the future separation—this is the position that there is retroactive clarification, meaning that the future can clarify the present state—or not. This latter is the position that there is no retroactive clarification, meaning that Jewish law does not recognize such retroactive determination. As a matter of law, nearly all legal decisors rule that in biblical law there is no retroactive clarification, which means that it is forbidden to drink the wine on the Sabbath itself. The question is why.

Defining the Situation According to the View That There Is No Retroactive Clarification

At first glance, according to the view that rejects retroactive clarification, the priestly portion has not been separated until Saturday night arrives, and therefore now all the wine is untithed produce forbidden for drinking. But Rashi here, and in several parallel sugyot, writes in explaining the prohibiting opinion, that is, the one that rejects retroactive clarification:

They prohibit—thus Rabbi Yehuda does not accept retroactive clarification, and he is concerned lest he drink the priestly portion.

According to his approach, Rabbi Yehuda prohibits drinking the wine out of concern that the person may drink the priestly portion, and not because of the prohibition of untithed produce. How can one say that, according to the view that the separation was not made at all and never took effect? Clearly, according to him, separation made in this way does take effect, but on the Sabbath itself it is not determined which specific two log are the priestly portion. There is here a state of uncertainty regarding which two log of the wine in the barrel bear the designation of priestly portion, and therefore it is forbidden to drink any of the wine in it. Other medieval authorities held this way as well. Some of the medieval authorities held that this really is a prohibition of untithed produce and not a doubtful prohibition of the priestly portion, but our concern here is to clarify the position of Rashi and his school.

Incidentally, the practical implication of this dispute is that according to Rashi and his school, a priest would be able to drink all the wine from this barrel, since he is permitted to drink wine of the priestly portion, whereas according to the other medieval authorities, of course not, for untithed produce is forbidden even to priests.^[4]

Ontic or Epistemic Uncertainty

One could have interpreted this view as involving epistemic uncertainty, and that is the simple sense of Rashi’s wording: ‘and he is concerned lest he drank the priestly portion.’ After all, at present we do not know which two log will remain on Saturday night. Any pair of log might be the two log of the priestly portion, and therefore, out of uncertainty, we are forbidden to drink any of the wine. But there are various indications that the intent here is different. What we have here is ontic uncertainty: there is a tenuous designation of priestly portion upon every pair of log in this mixture, and therefore there is a certain but tenuous prohibition upon every pair of log. Rashi’s expression ‘he is concerned lest’ is only a conventional turn of phrase. In light of this we can understand, for example, why the priestly portion is not nullified in the majority of ordinary wine. Rabbinically, the priestly portion is nullified only in one hundred parts, but on the basic biblical level it is nullified by a majority. Exactly as we saw above with regard to the women, here too there is no rule of nullification by majority, because in ontic uncertainty there is no difference between the majority and the minority, and in fact there is no majority and minority here at all.

According to this description, every pair of log is in the same status as the others, and the overall state is a superposition among a great many ‘pure,’ classical states, each of which is a state of all the wine in the barrel. In each of those states, a certain pair of log is the priestly portion and all the rest is ordinary wine.^[5] For those familiar with mathematical notation, one could write here a formula similar to the one I wrote above, which describes the state as a superposition of n classical states (see note 3):

where the state [i> describes the i-th pair of log, that is, a certain set of molecules מתוך the totality of molecules, as the priestly portion, and the rest as ordinary wine.

We thus learn that, at least according to Rashi, as a matter of law in such a situation the separation of the priestly portion does indeed take effect and is effective already now, except that the designation of priestly portion does not take hold on one specific pair of log, but on all pairs of log, exactly as in betrothal that does not permit intercourse.^[6]

According to Rashi, this state is really a state of ontic uncertainty, or ambiguity, and not uncertainty. In reality itself there are no two specific log of wine that are the priestly portion; rather, every pair is a tenuous priestly portion. Again, it is not correct to say that all of them are a tenuous priestly portion; rather, the overall state is a superposition of classical states, in the sense that in each of them there are two specific log of the priestly portion and the rest are ordinary wine. Even in Heaven they do not know which pair are the two log of the priestly portion. And even if they know which pair will remain on Saturday night, that does not matter, because according to the view that there is no retroactive clarification, at present that is not the pair of log of the priestly portion.

In closing, let me note that in the sugya of retroactive clarification in Gittin 25a they discuss a case of a man who wrote a bill of divorce for one of his two wives, namely, the one who would go out first through the doorway. According to Jewish law, a bill of divorce must be written for the sake of the woman being divorced as a condition of its validity. This man has two wives, and he decided to divorce one of them without determining which one of the two. He instructs the scribe to write a bill of divorce for the sake of that woman who will be chosen in a future event, namely, whichever one will go out first through the doorway of the house tomorrow. The Gemara makes the validity of such a bill of divorce depend on the dispute regarding retroactive clarification. Here one can already see even more clearly the similarity to the sugya of one who betroths one of two women discussed above. In the next column we will discuss the connection between the two sugyot and its implications for the phenomenon of quantum collapse in Jewish law.

1.

Footnotes

1. I should note that Shkop’s great disciple, Rabbi Shmuel Rozovsky, does not accept his teacher’s theory on this issue, and in his view this is epistemic uncertainty. In the third chapter of my book, Fuzzy Logic in the Talmud (volume 12 in the Talmudic Logic series), I discussed his words and explained Shkop’s position.
2. In column 33 we dealt with time travel, and there too we saw a similar phenomenon: in the legal world it is easier to understand than in the physical world.
3. Incidentally, even if the coefficients were not equal, I still do not think it would be correct to decide according to the majority and permit one of them. As Shkop wrote in the passage cited above, in ontic uncertainty one does not follow the majority.
4. One could analyze this further in light of the dispute among the medieval authorities, Rashi and Tosafot on Yevamot 86a, discussed in __Atvan DeOraita __sec. 2, whether the prohibition of untithed produce is due to the priestly portion mixed into it or whether it is an independent prohibition. But in the straightforward sense it is an independent prohibition, and that is what I am assuming here.
5. Since this is liquid mixed with liquid, there are infinitely many possibilities for choosing the pair of log, and not merely fifty, of course. But on intuitive grounds it seems that in every molecule that I drink there is a prohibition in a weakened measure of 1 in 50, because it belongs to very many pairs among the totality of possible pairs.
6. One may wonder why we do not say that the priestly portion should take effect upon that very pair of log that will remain on Saturday night. After all, if I drink the wine on the Sabbath itself, then with regard to every sip that I drank it is obvious that it does not belong to the two log of the priestly portion, since it was drunk and therefore certainly will not be part of the two log that remain on Saturday night. But we must remember that because it is forbidden to drink the wine, in practice this will not really happen. One may debate what the law would be if someone were to come and drink anyway, despite the prohibition. In such a case there do remain two specific log, and everything that was drunk does not belong to them, and then there would be room to say that he drank ordinary wine and that therefore it had been permissible to drink. One must say that this is precisely the principle of no retroactive clarification, which says that one may not rely on such a later determination. Even if I did drink and on Saturday night two specific log remain, it is still not correct to say that these were the two log of the priestly portion from the outset. Because there is no retroactive clarification, the state on the Sabbath itself is that the designation of priestly portion exists equally on every pair of log in the barrel.