Symposium on Eli Hirsch’s “Talmudic Destiny” 28 July–04 August

Talmudic Destiny

July 28 – August 04

Symposium on Eli Hirsch’s “Talmudic Destiny”

Participants include: John Hawthorne (Oxford), Rabbi Chanoch Waxman (Yeshivat Har Etzion), Jeff Russell (Oxford), Dean Zimmerman (Rutgers), and Jeremy Goodman (Oxford).

Click here for comments by John Hawthorne

Click here for comments by Jeremy Goodman

Click here for comments by Jeff Russell

Click here for comments by Chanoch Waxman

Click here and here for replies by Hirsch

12 Comments
  1. Aaron Segal

    Thank you all for this fascinating symposium and a special thanks to Eli for such a wonderful paper! There is much to absorb and discuss here, but I just had a couple of minor thoughts/questions:
    (1) Perhaps Eli addressed this somewhere and I overlooked it (if so, I apologize) but why couldn’t Rashi’s position be that there is simply a doubt as to whether the Aristotelian is right or the Ockhamist is right? It would be something like a sfeka d’dina (a legal, or in this case, philosophical quandary) — or at least it would not involve any ignorance of “matters of fact” (as it does in Hawthorne’s alternative interpretation) — and it would not involve any objective indeterminacy or muddle in the world. Is it Rashi’s particular formulation of the view that rules this interpretation out? If so, what is it about the formulation that does so? Or is there some more involved chain of reasoning that rules out such a view, and so makes such an interpretation uncharitable/implausible?
    (2) I didn’t quite understand Eli’s point about Jeff on Rashi-and-disquotation. I don’t think one needs to rely on an “intertemporal disquotation principle” in order to show that “It was true last week that A was going to be chosen” is equivalent to “Last week, A was going to be chosen”; all one needs is an “eternal disquotation principle,” that is, a principle that says “At every time t, p iff ‘p’ is true” (or to extend this in a way that Eli allows, “At every time t, p iff it is true that p.”) For an instance of that is “At every time t, A was going to be chosen iff it is true that A was going to be chosen”. Since last week is a time, an instance of that is: “As of last week, A was going to be chosen iff it was true that A was going to be chosen.” (I made some minor modifications to keep it sounding idiomatic.) And isn’t the following a valid inference schema: As of last week, p iff q; As of last week, p; therefore, As of last week, q? If so, we can conclude, “As of last week, A was going to be chosen” iff “As of last week, it was true that A was going to be chosen”. And isn’t the right hand side synonymous with: “It was true last week that A was going to be chosen” ?

    1. Dani Rabinowitz

      POSTED ON BEHALF OF ELI HIRSCH
      ————————————————————-
      Thanks, Aaron. I’ll try to respond.

      (1) This is an interesting possibility for understanding Rashi. I have three quick thoughts about it.

      (a) It seems that on Aaron’s proposal the proponents of ein bererah will not say, “I hold (according to the poverty of my understanding) that yesh bererah is a mistaken position.” Rather they will have to say, “I am unable (in the really depressing poverty of my understanding) to decide whether or not yesh bererah is a mistaken position; maybe it is and maybe it isn’t.” That seems to me an odd take on a Talmudic controversy. One really would expect the proponents of ein bererah to be claiming that the other position is mistaken, not just that maybe it is.

      (b) In my note 35 I try to present evidence that Rashi would be willing to apply memah nefshach reasoning in the context of bererah. I think that implies that, taking my example again, the safek is whether it is rauy for A to be chosen or rauy for B to be chosen (or perhaps rauy for neither to be chosen.) But on Aaron’s proposal the side of the safek in which Aristotole is correct would imply that neither of these options were rauy to happen, and I think that memah nefshach reasoning could not then apply.

      (c ) In note 31 I cite Rashi’s amazing statement that what was rauy to happen may not have happened. Hawthornes’ interpretation makes immediately good sense of this. And I try to make sense of it in terms of my interpretation. But I think it’s clear that on Aaron’s proposed interpretation we could make no sense of it.

      (2) I actually considered this question briefly in “Rashi’s View of the Open Future,” but I wonder if I’m missing something. Rashi holds (on my interpretation) that, since the sentence “(Last week) A was going to be chosen” was not determinately true last week, even now after A has been chosen, it is not determinate that the sentence was true last week. I suggested, however, that the sentence might be understood in a sense such that it is now determinate that the sentence is now true. We might take this sense to be simply “Last week was a week before the choice of A.” It’s obviously determinate now, after the choice of A, that this sentence is now true, thought Rashi would say that it was not determinate last week, and therefore it is not determinate now, that the sentence was true last week. Certainly that would not violate PPI. But I understood Jeff Russell to be saying that it would violate Rashi’s commitment to disquotation. That, I think, is not right. Rashi is committed to the standard formulation of the disquotation principle, which says “p iff ‘p’ is now true.” It does not say “p iff ‘p’ is true at time t”, for some arbitrary time t. Now what is Aaron’s response to this? I don’t see, first of all, why Aaron does not consider his “eternal” principle to be “intertemporal.” His principle implies “Last week will have been a week before A was chosen iff ‘Last week will have been a week before A was chosen’ was true last week.” What I say in “Rashi’s View” is that such a disquotation principle might be acceptable to B-theorist (who do not believe in “temporal passage” and do not take tenses to be fundamental) but could not be acceptable to an A-theorist like Rashi. So I would agree with Aaron that it is possible to formulate a version of the disquotation principle that would rule out the move I am allowing Rashi to make, but this version of the principle is not the standard version and it’s not a version that one should expect Rashi to accept.

      1. Jeff Russell

        body { font-family: ‘GentiumPlus’; font-size: 25px; line-height: 1.250000em; }

        A quick reply to Eli Hirsch. First: what Aaron said. That was the reasoning I was implicitly relying on. In particular, what I was assuming wasn’t an “intertemporal” principle, but rather that the disquotation principle holds at all times. I took this to imply:

        Last week, (if A was going to be chosen then it was true that A was going to be chosen).

        This in turn I take to imply:

        If (last week A was going to be chosen) then (last week it was true that A was going to be chosen).

        A couple further notes about this. First, in abstract terms the principle I was relying on isn’t “For any time t, p iff it is true at t that p”. Rather it’s “For any time t, at t (p iff it is true that p)”. I.e., what I was assuming was that the principle (p iff it is true that p) holds at all times. You could deny this, and say that even though the disquotation principle (p iff it is true that p) holds now, that principle didn’t hold last week—but that would be surprising!

        (Is that what the Rashi view is meant to say? According to that view, am I then right to make the following speech? “Last week, A was going to be chosen. But it may be that last week, it wasn’t true that A was going to be chosen. Last week, it’s being true that A was going to be chosen required more than merely that A was going to be chosen.”)

        Second, I made a point of using the propositional “true that” form of the principle , rather than a sentence-truth principle. (“Redundancy” might a better label for this principle than “disquotation”.) Quotation is more delicate, because what propositions are expressed by what sentences is something that changes. But we can run things in those terms, too, if we’re careful about tenses. If the sentential-disquotation principle was true last week, then we can reason as follows.

        (1) Last week, “‘A is going to be chosen’ is true iff A is going to be chosen” was true.

        (2) So: last week, ‘A is going to be chosen’ was true iff A was going to be chosen.

        (3) Last week, A was going to be chosen.

        (4) So: last week, ‘A is going to be chosen’ was true.

        (The step from 1 to 2 is using a fact about what the truth conditions were last week for the long quoted sentence in 1. Spelling it out compositionally: Last week, the sentence “S is true” was true iff S was true. The sentence “A is going to be chosen” was true iff A was going to be chosen. And a sentence of the form “p iff q” was true iff: “p” was true iff “q” was true.)

        Is the view that Rashi would reject this reasoning?

        1. Yes, Rashi would reject the reasoning from (1) to (2). Certainly Rashi will accept (1); the disquotation principle must hold at all times. The issue in the move from (1) to (2) seems more apparent if we consider, instead of “A was going to be chosen”, a sentence like “Aug. 3 will have been a week before the choice of A,” where we suppose that last week was Aug. 3. Then I think the move from (1′) to (2′) does not seem obviously correct.
          (1′) Last week, “ ‘Aug. 3 will have been a week before the choice of A’ is (now) true iff Aug. 3 will have been a week before the choice of A” was true.
          (2′) So: last week, ‘Aug. 3 will have been a week before the choice of A’ was true iff Aug. 3 will have been a week before the choice of A.
          Maybe this is clearer if we elaborate the import of (1’) and (2’) as follows:
          (1’’) Last week, “It’s now indeterminate whether Aug. 3 will have been a week before the choice of A, and therefore it’s now indeterminate whether ‘Aug. 3 will have been a week before the choice of A’ is now true” was true.
          (2’’) It’s now determinate that Aug. 3 will have been a week before the choice of A, and therefore it’s now determinate that ‘Aug. 3 will have been a week before the choice of A’ was true last week.

          I’m suggesting that it’s consistent with Rashi’s commitment to disquotation that he can accept (1’’) and reject (2’’). It’s being determinately true now that Aug. 3 will have been a week before the choice of A, does not obviously entail (by disquotation or redundancy) that it is determinate now that it was true last week that Aug. 3 will have been a week before the choice of A. I suggested in my reply to Hawthorne that a principle that Rashi might have as a supplement to PPI is that, if the fact that p is a soft fact about t, then it can be determinate now that p is now true and not determinate now that p was true at t. That’s what blocks the move from (1′) to (2′). Whatever one thinks of this, I don’t think it can be refuted merely by appealing to disquotation or redundancy . In spelling out the move from (1) to (2) compositionally, Jeff says (changing the sentences again), “The sentence ‘Aug. 3 will have been a week before the choice of A’ was true [last week] iff Aug. 3 will have been a week before the choice of A.” But that in effect just denies PPI (as well, of course, as Aristotle’s position), doesn’t it? So I think the crux of the matter is whether PPI has any intuitive plausibility. As I’ve said, if we accept bivalence then it seems that we need PPI to understand a distinction between ignorance and indeterminateness, but I can certainly appreciate skepticism as to whether we do understand such a distinction.

          1. Jeff Russell

            I think this is helping me get a clearer idea of what’s going on.

            (I’m going to keep using the “is going to” version, because I just have a hard time parsing and correctly past-shifting the “will have been before” sentence. For what it’s worth, these “was going to” sentences sound idiomatic to me, though sometimes a little odd out of context.)

            It looks like a key issue is this claim I assumed about past truth conditions:

            (PastT) Last week: “A is going to be chosen” was true iff A was going to be chosen.

            As Eli points out, Aristotle denies this claim—since Aristotle has unusual views about truth conditions in general. That is, Aristotle also rejects this:

            (NowT) Now: “A is going to be chosen” is true iff A is going to be chosen.

            I didn’t understand before that even though Rashi accepts (NowT), he sides with Aristotle on (PastT). That puts the disagreements in a different place than I expected to find them. (What I thought before was that Rashi accepted PastT, and rejected as indeterminate the claim: Last week, A was going to be chosen. And though that rejection sounds odd, I thought that Rashi maintained it as a consequence of something like PPI.)

            Relatedly, I was thinking that (PastT) is just what it would take for the principle (NowT) to have held last week—we’ve just uniformly shifted “is” to “was”. But I see now there was a kind of circularity in the argument I offered for that. (Even so—it still looks to me like plausible semantics.)

            One question I still have is, according to Rashi, what were the truth conditions of “A is going to be chosen” last week? How would things have had to be last week for this sentence to have been true? I think the answer must be: that A was definitely going to chosen. In that case, though, why doesn’t the truth of the sentence “A is going to be chosen” today require that A is definitely going to be chosen? This looks to me like an odd asymmetry.

            To sum up, if I’ve got things straight then the following claims are inconsistent:

            Last week, it was indeterminate whether A was going to be chosen.
            Last week it was determinate that (“A is going to be chosen” was true iff A was going to be chosen).
            If last week it was indeterminate whether “A is going to be chosen” was true, then it is now indeterminate whether last week “A is going to be chosen” was true.
            It is now determinate that last week, A was going to be chosen.

            As I understand things now, Rashi accepts 1, 3, and 4 and rejects 2. (And my original mis-reading was that Rashi accepted 1, 2, and 3 and rejected 4.)

        2. I think Jeff and I are having a problem getting past a certain point that may be largely terminology, but I’m not sure. I think that if you assert “A was going to be chosen” this can be understood in a sense that commits you to a claim about past truths or facts, i.e., to the claim “It was true that A was going to be chosen.” Now Rashi (after A has been chosen) will not assert “It was true that A was going to be chosen,” so Rashi will not assert “A was going to be chosen” in that sense. I suggested, however, that Rashi might go along with what I took to be a standard Aristotelian move of understanding “A was going to be chosen” in a sense in which it does not mean “It was true that A was going to be chosen”, and I said that Rashi (like Aristotle) will accept “A was going to be chosen” in that truth-free sense. I think that Jeff assumes that we are dealing with such a truth-free sense. Okay so far. But now what about “It was determinate that A was going to be chosen”? I think Rashi will reject this because I don’t think there is any way to accept it without also accepting “It was determinate that it was true (a fact) that A was going to be chosen, ” and Rashi must certainly reject the latter. I hope that Jeff agrees with this, and therefore when he says at the end of his last comment that Rashi rejects

          Last week it was determinate that (“A is going to be chosen” was true iff A was going to be chosen).

          he really meant that Rashi would reject

          Last week it was determinate that (“A is going to be chosen” was true) iff A was going to be chosen

          Let me touch on something else. Jeff raises a hard question that I’ve also been thinking about:

          One question I still have is, according to Rashi, what were the truth conditions of “A is going to be chosen” last week? How would things have had to be last week for this sentence to have been true? I think the answer must be: that A was definitely going to chosen.

          I don’t see how that could be the answer, since Rashi wants to say that’s it’s indefinite whether that sentence was true last week but definite that it was not definite last week that A was going to be chosen.

          The only suggestion that I can make at this point is that any A-theorist, even those who do not subscribe to views like Aristotle’s or Rashi’s, will be presented with the question Jeff asks. Take the proposition P that JFK’s death is in the past. How would things have had to be in 1960 for that proposition to have been true then? The answer can’t be “Things would have had to be that JFK dies before 1960.” By bringing 1960 into the truth-conditions of the proposition one is in effect giving in to the B-theorist (this is, I think, essentially Mellor’s argument against the A-theory). But that’s all I can say right now.

          Thanks for pressing me on these points, Jeff.

  2. Chanoch Waxman

    First and foremost, I would like to thank Eli Hirsch for his serious response to my comments. My apologies for not clarifying things sufficiently. In my attempt to spare the general reader the jargon of conceptual Talmudic analysis, or at least to translate things into more philosophical terminology, something was apparently lost. I will try to clarify the three questions Hirsch raised in response to my comments by means of a general sketch of the various elements involved in a halakhic transaction or status change and the conventional wisdom regarding two possible models that emerge from such a sketch. In addition, although this is just a response to a response, I would like to address Hirsch’s theory regarding Rashi’s position and the possibility of a third alternative other than Hirsch’s presented position and the new Hawthornian position sketched out in John Hawthorne’s comment and Hirsch’s response.
    In the tradition of conceptual analysis adopted by most modern practitioners of Talmud study, the Halakhic transaction or status change (kinyan) can be viewed as consisting of three elements: i) The mental state, the da’at, of the agent or agents that effect the transaction, ii) the action, the ma’aseh kinyan, that effects the transaction, transfer or change and iii) the result, the chalut, of the said transaction. As a matter of theoretical interest, but also possible practical difference, we may raise the question as to the exact role of factors (i) and (ii) in generating the eventual result (iii). On a broad plane, without differentiating between particular types of transactions or status changes at this point in time, the conventional wisdom outlines two possible poles as theoretical models. In model one, which we may call for now the da’at model, it is the mental state of the agent or agents that plays the fundamentally primary role in generating (iii), the chalut. Any requirement of (ii), such as the writing of a contract, exchange of money, verbal statement, or even the physical movement of some object should be viewed as no more than an “expression” of the mental state, the da’at that actually accomplishes the transaction. In model two, things are the reverse if not an actual mirror image. Calling this for now, the “ma’aseh kinyan” model, here it is the action, not the mental state that carries out the fundamentally primary role of generating (iii), the chalut. The role of the mental state is no more than to either legally allow the transfer and transaction carried out by the “ma’aseh” or perhaps to define the said action on a certain formal level.
    Now when I referred to the “idealist” model of Halakhic transactions I meant something like the “da’at” model of Halakhic transactions sketched above and when I referred to the “physicalist” model I meant something like the “ma’aseh kinyan” model sketched above. Perhaps this will help clear things up. On the “idealist” or “da’at” model or prejudice I accused Hirsch of adopting, it is the mental state, the da’at that does the work, the actual transfer and generation of the “chalut”. The action taken upon the object to be transferred or have its status changed whether a verbal or even physical action is no more than an expression of a given mental state. For this perspective, I agree with Hirsch that the “minimal criteria of reference theory” is certainly correct and there is no problem in breirah cases provided one is an Ockhamist. As long as one refers successfully to the object to be transferred at time “t”, the time of the transaction, we can say that there is sufficient “da’at” to carry out the transaction at time “t”. One can not raise a “ma’aseh” based objection to breirah cases as the ma’aseh is no more than an expression of da’at, already assumed to be sufficient on the “minimal theory of reference” and Ockhamian account.
    As a parenthetical note, it is important to point out that this model is often attributed to Rabbi Chaim Soloveitchik, known as Rabbi Chaim of Brisk. It was explicitly asserted by Rabbi Avraham Karelitz, often referred to as the Chazon Ish and may certainly have influenced Rabbi Aharon Kotler, one of Hirsch’s central sources for his treatment of breirah. In my opinion, there exists near definitive evidence that this model originates in Rashi and Tosfot on numerous Talmudic topics including that of breirah. Hirsch’s intuitions are right on.
    What I tried to point out in interpreting Ramban and through counterexamples is that there exists the possibility of the second model and this carries implications for both i) the criteria for transactions and ii) the question of breirah. While not necessarily the case, in my opinion the “ma’aseh kinyan” model just sketched should be viewed as reflecting the “physicalist” model I outlined. Precisely because the world of Halakhic transactions is viewed on the model of physical events, a “ma’aseh kinyan”, interpreted as a real world action on a particular object, not just as an expression of a mental state, is the key factor in effecting a transaction. It would seem to me that such a version of transactions would lead to a stronger criteria for transactions than the theory of “minimal criteria of reference” and have specific implications for breirah. To perform a real word action upon a real word object requires more than just clarity of mental state, referring to the object and coherent speech. It requires, pointing to the object in a way that allows the “connection” of action “x” the ma’aseh kinyan performed on object “y” to the said object at time “t”, the time of action. To utilize a poor analogy, consider the billiard ball model often utilized to illustrate the conceptual schema of Newtonian mechanics. A cue ball cannot impact or effect the movement of some ball upon the table without actually touching it. So too, the ma’aseh kinyan must “connect” in some strong sense to the object it impacts.
    This may be clearer if we return to the “breirah” issue. At time ‘t’ we can talk of “the house that my wife will choose”, but we cannot pick out, point to, or even “strongly” refer to as a hard fact the house that my will choose. After all that has not yet been determined and on the theory of the “ein breirist” cannot be determined retroactively. Hence there is no specific house that we can view as being subject to a “physicalist” type of “ma’aseh kinyan”. I take this as being the import of Ramban’s comments that in “breirah” the what of the matter is unclear at time ‘t’, the time of the transaction. From a “physicalist” perspective one cannot perform an action at time ‘t’ upon an unclear “what of the matter”. Again, admittedly, such is not the case from an “idealist” perspective where the “ma’aseh kinyan” is no more than an expression of mental clarity regarding some object described by the sentence “the house that my wife will choose”.
    Regarding, Hirsch’s second question to me, how would the “physicalist model” differentiate between “breirah” and “tenai”, I believe that I have already more or less addressed this in my original comments. Ramban himself views these two categories as fundamentally different, what vs. whether. I believe this distinction is sustainable in light of my clarification above. In breirah the lack of clarity as to what at time ‘t’ generates an intractable problem of lack of “connection” between the physicalist’s ma’aseh kinyan and the supposed entity effected at time ‘t’. In “tenai” there no such lack of clarity as to what exists at time ‘t’ and consequently no such problem of ma’aseh kinyan at time ‘t’. At the time of the accomplishment of the condition the outstanding logical problem of whether is resolved and the transaction is considered valid as of time ‘t’. While one may still object that this requires some sort of retroactivity and we are here dealing with the position of the “ein breirist” this objection can be handled relatively easily by moving “tenai” to the realm of intent, i.e. claiming that according to the “ein breirist” one can retrospectively clarify intent, the question of whether, but not occurrences in the real world, i.e. what. Alternatively, one can relay upon the view that tenai milta achriti hi, that tenai is a minor matter having no real bearing on the transaction itself and therefore once clarified at time ‘t-1’, the actual transaction itself performed at time ‘t’ is valid as of time ‘t’. While the details of any of these latter two claims still need to be worked out, I would maintain that the breira – tenai distinction can be defended according to the “physicalist” model.
    Regarding Hirsch’s third question, the attempted counterexample of the “deep pool” case; great case! But I am not really sure why this is a problem. It seems to me that this is not a “breirah” case at all. In this case, which of the pools is the deeper pool is an objective hard fact as of time ‘t’. It is already “settled”. It is just not known by the agents of the transaction which of them is in fact deeper at time ‘t’ until later on at time ‘t-1’. This would seem to be a matter of retrospectively knowing at time ‘t-1’ something which could in principle have been know at ‘t’ but for some reason or another was not. On the assumption that temporary ignorance does not affect the meaning of terms or reference, talking about the “deeper pool” as of time ‘t’ would seem to be sufficient for the “stronger” type of reference, the one that singles out a particular object in the world at time ‘t’ required by the “physicalist” model. Or maybe not.

    Regarding Rashi and the central thesis of Hirsch’s paper, I refrained from commenting in my original response for two reasons. First and foremost, I was not sure that I had much to add to his great ideas. Second, Hirsch follows the conventional wisdom regarding the position of Rashi as expressed in Mishnat Rebbe Aharon to Gittin 25a. I have my doubts whether this conventional wisdom is wholly accurate. As my two statements seem to contradict, I will try to explain and from there proceed to what seems to me like a variation on some of Hirsch’s ideas.
    Ironically, none of the three texts of Rashi cited by Hirsch in footnote (1) dictate the position of “permanent safek” as elaborated by Hirsch. However, there is a fourth comment of Rashi found at Meila 22b s.v. rabi yehuda which certainly does. Moreover, the comments of Mahri found in Tosfot Eiruvin 37b s.v. elah referenced by Hirsch in footnote 35 certainly indicate that such a position exists in the Ashkenazic tradition. Nevertheless, the scope of this position may be crucial for interpreting it, and hence some brief explanation is in order. We may distinguish between what I will call “normal” cases of “permanent safek” and what I will call “peculiar” cases of “permanent safek”. In “normal” cases, the “safek” is generated by some understandable mechanism. The relationship to “breirah” consists of the fact that the “ein breirist” rejects retroactive clarification and hence the safek remains intact. The three cases cited by Hirsch in footnote (1) all conform to this model. I will go in Hirsch’s order. Case (1): Rashi Gittin 25a s.v. lekuchot hen comments on the case of inheritance where the mechanism of inheritance leaves a doubt as to whether portion ‘a’ of the father’s land goes to brother ‘x’ with ‘b’ going to brother ‘y’ or the reverse. Given the position of “ein breirah”, this doubt\apportionment cannot be clarified retroactively and the division of the land later performed by the brothers has a status of a sale reversible by the jubilee year. Here the position of “ein breira” is no more than a necessary condition for the persistence of this “normal safek”. Case (2): Rashi Gittin 74a s.v. rabi yosi comments on the case of where Reuven attempts to divorce his wife while using ambiguous terminology regarding the timing of the divorce. According to Rabbi Yosi the terminology of the Mishnah should be interpreted as meaning “the day before I die” and her status is that of safek, or “divorced and not divorced” each and every day. Rashi comments that Rabbi Yosi holds “ein breirah” and hence Reuven living or dying the next day after any given day does not resolve the safek regarding the previous day. Again this conforms to our previous model. The safek is generated by the fact that on any given day ‘d’, Reuven may die the next day ‘d-1’. Given that “ein breirah”, the safek cannot be retroactively clarified. Again, “ein breirah” is no more than a necessary condition for the persistence of a “normal safek”. Case (3): Rashi Eruvin 37b s.v. Va’chachamim omrim comments on the case of a “wise man”, one trusted on matters of tithes who purchases vegetables for himself and a “man of the people”, one considered suspect on the matter of tithes. The original purchase leaves a safek as to which portion of the vegetables is intended for which party. Given the position of “ein breirah”, this doubt cannot be retroactively clarified and hence the transfer to the “man of the people” at time ‘t-1’ is viewed as a sale rather than a transfer of an already determined portion. This requires the “wise man” to tithe before the resell in accord with the rules of “demai”. Once again, we face a “normal safek” generated by the original purchase. The role of “ein breirah” is no more than necessary condition that leaves the original safek intact.
    It seems to me, that nothing about these cases, nor Rashi’s comments to them says anything about “destiny”, “thinking about the future” or the like in any deep or interesting way. Rashi simply notes that the “ein breirist” rejects retroactive clarification and hence certain cases of safek, generated in a perfectly understandable fashion persist. Such is not the case in our next case, not quoted by Hirsch.
    Case (4): Rashi to Meila 22b s.v. rabi yehuda comments on one of the classic cases of breirah, the case of “two measures”. In this case, Reuven claims that he is currently tithing the “two measures” that he will actually separate at some later point in time ‘t-1’. While according to the “yesh breirist” it is permissible to drink of the formerly untithed wine immediately as of time ‘t’, according to the “ein breirist” it is forbidden. Obviously, this is the Talmudic forerunner of Hirsch’s “the house that Reuven’s wife will choose”. Now according to the normal rules of things, we may expect that according to the “ein breirist” the batch of wine which had been untithed has a status of “tevel” and cannot be consumed by anyone, even one of priestly descent. The actual act of tithing that happens at time ‘t-1’ cannot be retroactively moved to time ‘t’ and one would expect that the wine remains unaffected, i.e. tevel at time ‘t’. But Rashi comments that the wine is forbidden in consumption as “each and every cup may be the one’s tithed”. In other words there is a “safek” regarding which “two measures” were tithed at time ‘t’ and this doubt cannot be retroactively clarified by the actual act of tithing at time ‘t-1’. While this need not concern us for now, the Halakhic bottom line is that one of priestly descent can consume of the wine as it has been tithed. What need concern us is that somehow, even according to the “ein breirist” some kind of act of tithing has occurred at time ‘t-1’ which generates a safek. I view this as the “peculiar” case of “permanent safek”. Nothing obvious has happened at time ‘t’ that explains the generation of the safek at time ‘t’. There is no obvious act of tithing that occurs at time ‘t’ by talking of that which will happen at time ‘t-1’. It would seem that perhaps there is something essential to the position of the “ein breirist” that helps generate the safek and the role of “ein breirah” seems to be something more than just leaving a previous “normal” safek intact.
    I have often tried to explain this Rashi, albeit with minimal success, with the vague theory that i) intent is sufficient for an act of tithing, and ii) tithing can affect a batch on a general plane, leaving it a matter of doubt which parts of a batch have been defined as the tithed part. My thinking about Hirsch’s enlightening comments in his Talmudic Destiny and his definition of the PPI principle in his response to Hawthorne’s comments now make me think of Rashi’s position in the following fashion: i) It is a matter of soft fact that as of time ‘t’ that there are “two measures that Reuven will separate” ii) Whether as due to a particular local rule regarding tithing or a general “idealist” perspective regarding Halakhic transactions and status changes, the theory of “minimum criteria of reference” is sufficient to generate a status\result of tithing. iii) Consequently, Reuven’s act of tithing at time ‘t’ is valid and the “two measures that Reuven will separate” have the status of tithed measures as of time ‘t’. iv) But in the world of real objects and real world actors it is not know at points between time ‘t’ and ‘t-1’ which of the set of all two measures are those two, i.e. there is a safek on each cup. v) Due to PPI, or something like it, which I take as a nice philosophical formulation of “ein breirah” this safek\indefiniteness as of time ‘t’ cannot be retroactively resolved, nor even retrospectively viewed as anything other than a safek. vi) Consequently, we are left with Rashi’s famed position of “permanent safek”.
    Now it is not clear to me, whether my distinction between “normal” cases of “permanent safek” and the “peculiar” case of “permanent safek” was really necessary to make this argument for Rashi. Nevertheless, I think the distinction highlights the fact that the resources created by Hirsch’s analysis are only necessary to explain the Rashi on Meila 25a and that the distinction thereby limits the scope of the supposed difficult position of Rashi. On the other hand, if we think about it, I have simply used the resources provided by Hirsch to reconvert the “peculiar” case into the logical equivalent of the “normal” case, where the position of “ein breira” in fact has nothing to do with the creation of the safek. It is merely, a necessary condition for the persistence of the permanent safek. I have in fact worked hard to render the distinction mute and generate a general position for Rashi.
    What is clear to me is that what I have outlined here is certainly different than the Hawthornian position, the claim that there are “hard facts”, “something in the cards” as of time ‘t’ that facilitates the transaction but somehow may have changed or worked out differently by time ‘t-1’. This claim strikes me as counterintuitive as an explanation of the safek. Moreover, I am not sure how we can maintain a “hard fact” as of time ‘t’ that posits a certain future while simultaneously maintaining that that future is not the one that eventually works out at time ‘t-1’.
    To get to the heart of the matter, I am not really sure whether the position I have outlined is the same or different than Hirsch’s thesis. But on further thought, it seems somewhat different. It seems to me that Hirsch’s claims about Rashi’s position regarding any given two measures are correct. For any given two measures it is “indefinitely true” and simultaneously “indefinitely false” that they are “the two measures Reuven will choose”. But I am not sure that this has anything to do with “objective muddle” or “the superposition of contrary states” on analogy to quantum mechanics. Rather it is simply the result of i) the validity of the tithing at time ‘t’ of whatever two measures will be chosen at ‘t-1’, ii) the safek at time ‘t’ regarding which two measures those are, and iii) the fact of “ein breira”, the impossibility of retroactively clarifying certain kinds of indefinite situations. To put this a little bit differently, “objective muddle” might be the result of Rashi’s position on “ein breira” but there is no theory or metaphysical fact of “objective muddle that constitutes the cause of Rashi’s position on “ein breira”. Or yet in another way, let us consider R. Aharon Kotler’s claim, viewed as incomprehensible by Hirsch that “only prophecy” can resolve “breirah” cases according to the “ein breirist” of Rashi. While Hirsch’s “metaphysical objective muddle as cause” theory cannot “make sense” of this comment, it strikes me as eminently reasonable. If we know for certain as of time ‘t’, not by some probabilistic evidentiary rule such as “rov” or “chazaka”, that “x” are the two measures Reuven will choose then the muddle is never created at time ‘t’. As such, we need not be troubled by “ein breirah”. Despite the impossibility of retroactive resolution, one can drink of the wine, provided one does not drink those two measures picked out by the prophetic revelation.
    Two last comments regarding “objective muddle” and “breirah”. First, although I have denied the “objective muddle” or “superposition of contrary states” as the grounds for explaining Rashi and “breirah”, there is certainly much room for this kind of analogy in explaining Talmudic sugyot. Rabbi Joseph Soloveitchik often explained the topic of “bein ha’shmashot” the halakhic category of twilight, termed “safek yom safek laila” not as a case of doubt whether the time period of twilight is either day or night but as a case of both, in other words an “objective muddle”, or “superposition of contrary states” simultaneously possessing properties of both. I am unaware as to whether R. Soloveitchik ever drew the analogy to the Copenhagen interpretation of quantum mechanics but I am quite certain that is what he had in mind. On a similar note, in many of his meta-halakhic writings he claimed that the Halakha operates with “non-standard” logic and even went so far as to state that Halakha denied the law of the excluded middle. I also think that Hirsch’s claim in footnote 36 regarding cases of “safek” marriage in certain types of cases is but the tip of the iceberg regarding similar cases and indicates that numerous Talmudic topics may be amenable to this kind of treatment.
    Second, while Hirsch has devoted the lion’s share of his efforts to working through the prism of Rashi, I have tried to show that Rashi is not that interesting. Although I am not sure that this is an argument, except for the brief exception of Tosfot at Gittin 74b, Rashi’s supposedly difficult position is never mentioned or attacked in the entire medieval literature on “breirah”. In fact, Tosfot at Eruvin 37b quotes “Rashi’s position” in the name of Mahri approvingly. In contrast, it seems to me that the actual position of “ein breirah” as defined by Hirsch and for some reason attributed specifically to Rashi, his PPI, is the truly interesting claim.

    PPI: If it is indeterminate at time t whether a proposition p is true at t, then it is always indeterminate whether p is true at t. In terms of facts, the principle says that if it is indeterminate at time t whether a certain fact obtains at t, then it is always indeterminate whether that fact obtains at t.

    This is fascinating. In my opinion, what is says about the impossibility of retroactivity in general, to say nothing of its perspective on time and knowledge remains to be worked out on both the Talmudic and philosophical levels.

  3. Jeremy Goodman

    I’d be interested to hear more about how Hirsch conceives of the connection between Rashi’s view and tense logic. Does he take it to be a view that we can formulate in a language with tense operators and a determinacy operator (as I think John, Jeff and I were assuming in our comments)? If so, I’d be interested to know which of the principles I mentioned in my comments Hirsch thinks Rashi would accept and which he would reject. Or does Hirsch think that Rashi’s view resists expression in such a language? If so, should we think of it as a view that leaves standard (i.e., determinacy-operator-free) tense logic as it is (as I take it John was assuming in his argument that the view seemed committed to the assertability of claims of the form “p, and not determinately p”), or should we think of it as a view that presupposes that there is something fundamentally misguided about the whole tense logical enterprise?

    Also, it appears that there’s a whole volume on Temporal Logic in the Talmud focussing on these issues! http://www.collegepublications.co.uk/stl/?00005 I haven’t been able to get a hold of it, but it might be of interest.

  4. I’ve tried to suggest in my responses to the comments of John, Jeremy, and Jeff that there is a possible ambiguity that is screwing everything up. Let’s consider the principle that Jeremy cites in his initial comments as an axiom central to tense logic:
    p → PFp
    The translation of this suggested by Jeremy’s remarks is:
    (i) If p then it was once the case that it will be the case that p.
    I’ve suggested that there is a potential ambiguity in (i) as between the following two propositions.
    (a) If p then it was once true (a fact) that it will be true that p.
    (b) If p then, if a prediction was once made that it will be true that p, then that prediction has come true.
    Or instead of (b) you could have something simpler like:
    (c ) If p then there was once a time prior to the time at which it’s true that p.

    Rejecting (a) as a determinate assertable truth is the central idea of Rashi’s position. My point is that Rashi can certainly accept propositions like (b) and (c) while rejecting (a). I don’t know all that much about tense logic but my initial bet would be that we could probably do all of tense logic interpreting Jeremy’s axiom in terms of propositions like (b) or (c) instead of (a). Might Rashi’s position imply that there is “something fundamentally misguided about the whole tense logical enterprise”? I don’t see how that could be. Even if the move that I’ve just suggested can’t work out, I suppose we could always introduce some technical symbolism to regiment Rashi’s position in one way or another. (I know that Rich Thomason had a model some years ago of what was supposed to be an Aristotelian system, but I think he messed things up with regard to disquotation and bivalence.) As for why John thought that Rashi’s view seemed committed to the assertability of claims of the form “p, and not determinately p”, I think this is because I didn’t make it clear enough that I was appealing to an ambiguity. Rashi, I’m suggesting, will assert (i) in the sense of (b) or (c), and will deny the determinate truth of (i) in the sense of (a). But, I pray, Rashi will not assert a sentence in a certain sense and also say that the sentence in that sense is not determinately true.

    Thanks, Jeremy, for that great find of a volume on Temporal Logic in the Talmud. The blurb for the book says that to understand the Talmudic material “we need multi-dimensional temporal models with backward causation and parallel histories.” God, I hope not.

  5. Jeremy Goodman

    Thanks Eli, that’s very helpful. One thing that seems to be particularly distinctive about Rashi’s view is that it entails that a proposition p can be true at a time t without it being true at all times t’ that p is true at t. Here’s the idea.

    Let’s assume that there is always exactly one present time, that times are linearly ordered, and that the order and composition of the time series is a determinate and unchanging matter. Let’s also assume that tense operators are equivalent to claims about what is true at appropriately related times. Let p be the proposition that Reuven’s wife chose house A. On Eli’s interpretation, Rashi thinks we should (now) deny p → it has always been the case that it will sometime be the case that p – at least on its standard tense-logical interpretation, as in his (a). Given our assumptions, this is equivalent to: p, and not all times t before the present time are such that it true at t that (at some time t’ after the present time, p is true at t’). The first conjunct entails that it is true at the present time that p (by the principle that everything that is true is true at the present time), from which it follows given our assumptions that, all times t before the present time are such that, at some time t’ after t, p is true at t’. On the other hand, the second conjunct entails that not all times t before the present time are such that it is true at t that (at some time t’ after t, p is true at t’) – by the principle that “t” is intersubstitutable with “the present time” within the immediate scope of “at t”, since, at t, the present time is t. Given what should be uncontroversial principles about how “at t” interacts with quantifiers over times, it follows that, although the present time is such that, at it, p, the present time is not such that, for all times t, it is true at t that, at it, p. So a proposition can be true at a time without it being true at all other times that it is true at that time.

    If this is right, then I think Eli is right to formulate Rashi’s view in terms of what is true at which times, rather than using tense operators, since it helps to lay bare this heterodox feature of the proposal.

    1. Yes, that seems right. But the idea that Jeremy says is particularly distinctive of Rashi’s view( i.e., the idea that a proposition p can be true at a time t without it being true at all times t’ that p is true at t) also holds for Aristotle’s view, if I have that idea right.

      1. Jeremy Goodman

        Good point. Second try: what is distinctive about Rashi’s view is that (like Aristotle) he thinks a proposition p can be true at a time t without it being true at all times t’ that p is true at t, while (unlike Aristotle) he thinks that “at t” commutes with “not”.

Leave a Reply

Event Details
  • Start Date
    July 28, 2013
  • End Date
    August 28, 2013
  • Location
  • Category