Reasons and Rationales for the Commandments, Lesson 6

This transcript was produced automatically using artificial intelligence. There may be inaccuracies in the transcribed content and in speaker identification.

🔗 Link to the original lecture

🔗 Link to the transcript on Sofer.AI

[Rabbi Michael Abraham] Eighteen, period. No distinction—righteous women, wicked women, everything is eighteen, that’s it. Does Rabbi Shimon agree with that? No. Rabbi Shimon says if there’s even one wicked woman, you don’t take her, because she turns his heart away. There’s another commandment: “and his heart shall not turn away.” Is that an additional commandment? Well, it’s not really another commandment; it explains the previous commandment. So the question is really prior to the question. First of all, when there’s a dispute between Rabbi Shimon and Rabbi Yehuda—and that’s how all the later authorities routinely explain it—we’re used to saying that the Jewish law follows Rabbi Yehuda, because we do not derive law from the reason for the verse. The truth is that in Maimonides that’s not correct. The Jewish law follows neither Rabbi Shimon nor Rabbi Yehuda; it follows the first tanna. That makes a difference, because Rabbi Yehuda says that when the reason is written explicitly in the Torah, then he does derive law from the reason for the verse. Maimonides rules like the first tanna, that even when the reason is written explicitly in the Torah, we do not derive law from it. The big question is why.

So wait, let’s try to see what is written. Wait, same thing—that’s the same question. Why is it written there, and therefore why don’t we derive law from it? Yes, that’s the same question. As I said earlier, it seems to me that this connects to the difficulty in the standard explanation. The standard explanation apparently fits Rabbi Yehuda’s view. The explanation that says we do not derive law from the reason for the verse out of concern for error—then it would follow that if the Torah itself states the reason, there’s no concern for error, so there we do derive law from it. That is Rabbi Yehuda’s claim. Therefore Nachmanides and all those who read this Mishnah as though it contains only two opinions—Rabbi Shimon and Rabbi Yehuda—they also explain the idea of “we do not derive law from the reason for the verse” in accordance with Rabbi Yehuda, because they think there are only those two. Either Rabbi Shimon or Rabbi Yehuda; the Jewish law follows Rabbi Yehuda, and that’s it. And then what comes out is that the problem with deriving law from the reason for the verse is the concern for error. And if the reason is written explicitly in the Torah, then there’s no such concern, so yes, you do derive law from it. Then there’s no problem.

But Maimonides says we do not derive law from it even when the reason is written explicitly. What does that mean? It means that the fact that we do not derive law from the reason for the verse is not because of concern for error. Because if it were because of concern for error, then when the reason is written explicitly there is no such concern. So why not? First, just one more remark before that. I already said that concern for error is a bad explanation for why we don’t derive law from the reason for the verse, because not deriving law from the reason for the verse is even more likely to lead to error—that’s what I said at the beginning. Okay, so that connects to what I said earlier, and maybe that itself is what leads Maimonides—or first of all the first tanna, and afterward Maimonides who follows him—to say that this is not what “we do not derive law from the reason for the verse” means. It’s not about concern for error. So what is it? Why don’t we derive law from the reason for the verse even when it is written?

It seems to me that a plausible explanation, at least, is this: when do you derive law from the reason for the verse? When the verse itself is written in a certain way, and purposive interpretation leads you to a different conclusion. Here the verse says, “You shall not take a widow’s garment as collateral,” which sounds like any widow; the verse doesn’t qualify it. And then you say: yes, but the logic of the law exists only regarding a poor widow, therefore only a poor widow is included, not a wealthy widow. So wherever purposive analysis—deriving from the reason—leads to a different conclusion from what is written, that is where the question of deriving law from the reason for the verse arises. Right? Only then do we discuss it, because then the question is whether to follow the reason or not.

Maimonides says: wherever there is a contradiction between purposive interpretation and what is written in the Torah, apparently you are not doing the purposive interpretation correctly. Because there is no reason to assume that the Torah expresses itself imprecisely. At least on the assumption that the Holy One, blessed be He, wrote the Torah, then He knows how to write, and what you are really trying to say is that He wrote it imprecisely. Here—we do purposive interpretation and see that the wording is not precise, so we have to interpret it differently from what the wording implies. So the first tanna, and following him Maimonides, will apparently say: no. Our assumption is that the wording is precise. If it says, “You shall not take a widow’s garment as collateral,” it means any widow. Ah—but the reason? Rabbi Yehuda also agrees that the reason is what Rabbi Shimon says. There is no dispute that this is the reason. The whole question is only whether to derive law from it.

And notice, here there is a very interesting point. The first tanna—or Rabbi Yehuda—will say to us: so what are you saying? You are basically saying that it is permissible to marry thirty women if they are as righteous as Abigail, right—more than eighteen? “He shall not take many wives” means eighteen, but it would be permissible to marry thirty women if they are as righteous as Abigail? Because the whole problem is the concern lest his heart turn away. So the answer is no: the rationale is not merely the concern lest his heart turn away from wicked women; the rationale is the concern lest his heart turn away, period. Even thirty righteous women will turn his heart away. The application of the reason—I agree with the reason you are saying, Rabbi Shimon. The reason is lest his heart turn away. That is obvious; it is written in the Torah. I’m not disputing a reason written in the Torah. I am only claiming that if you arrive at a dissonance between the reason and what is written in the verse, then no—you have not applied the reason correctly. You think that turning the heart away happens only because of unrighteous women. Not true. If he has thirty women in his palace, then even if they are very righteous, they will turn his heart away—he’ll spend all his time dealing with all their little dramas instead of dealing with important matters. And therefore even thirty women will turn his heart away. We don’t have the wife who usually sits here now, so maybe not thirty. Yes, right.

[Speaker C] But then why is the reason written at all?

[Rabbi Michael Abraham] So he says—Maimonides says this explicitly—the reason is written in order to explain the law to you.

[Speaker C] And why there and not elsewhere?

[Rabbi Michael Abraham] Because there they wanted to explain the law to you. He didn’t give you a criterion for when yes and when no. And he writes explicitly that when the Torah states the reason, it is in order to explain the law to you, not in order to create a new law. It is to explain the existing law.

[Speaker C] That’s the whole issue of his fifth root.

[Rabbi Michael Abraham] And now notice—the point is an interesting one, because what comes out of this is really what I said earlier. This was the question we were actually dealing with. The question I’m starting from is that the standard explanation for why we do not derive law from the reason for the verse is not logical. Because that explanation says: concern for error. And I say, what do you mean? If you don’t derive law from the reason for the verse, then there’s no concern for error? The concern for error is even greater. So he says no—what is that based on? It is based on the fact that the reason suggested by Rabbi Shimon is accepted by everyone. They do not dispute that this is the reason. The only question is whether to use it in interpretation; there is no dispute that this is the reason.

So I said: fine, if that is the reason, then the likelihood of making a mistake is greater if you don’t use the reason. He says no—now notice this example is a wonderful one, because here the reason is written, and everyone agrees that this is the reason, and still there can be a mistake—not in the reason, but in the application. You understood the reason correctly, but you think that the reason, “lest his heart turn away,” exists only where the women are unfit. The Torah tells you: no, you are mistaken in the application. The reason is correct. The reason is that his heart should not turn away. It is forbidden to multiply wives. But your application is mistaken, because there can be a turning away of the heart even from righteous women.

[Speaker C] If it’s one wicked woman, then would he permit it?

[Rabbi Michael Abraham] What? Say that again.

[Speaker C] By that same principle, would Maimonides permit a king to marry one wicked woman?

[Rabbi Michael Abraham] Because—

[Speaker C] I’m not understanding correctly? One wicked woman won’t turn his heart away?

[Rabbi Michael Abraham] Or at least not enough for the Torah to prohibit it. Meaning: with one wicked woman I can cope; with thirty righteous ones, good luck to him. Yes. Anyway, I’m bringing this as an example of a mechanism that on the one hand accepts Rabbi Shimon’s interpretation of the law. In other words, the reason Rabbi Shimon proposes is a correct reason; there is no dispute that this is the reason—or at least it is plausible that it is correct. You can always make a mistake, but that is not the level of the disagreement. Rabbi Yehuda and the first tanna do not argue with Rabbi Shimon because they think he is mistaken in interpretation. They say that even if you are right in the interpretation, you do not make use of it in halakhic interpretation. Write it as a commentary on the Torah if you want, but don’t use it in halakhic interpretation.

And why not? Rabbi Yehuda says: why? Because maybe you are mistaken. But then I say: what do you mean maybe I’m mistaken? This is the plausible reason—why assume I’m mistaken? If I don’t use the reason, it is even more plausible that I’ll be mistaken. And the first tanna says: no, no—because you are mistaken in the application. The reason is correct; I have no problem with that. When the reason is written in the Torah it is certainly correct. You are mistaken in the application. You think that turning the heart away comes only from wicked women, and that’s not true. Even from a multitude of righteous women there is turning away of the heart. So I think that is the most compelling plausible explanation for the first tanna and for Maimonides.

And then what this really means is that the problem is not the interpretation of the verses as such, but that wherever there is dissonance between the interpretation and the wording of the verse, there I am concerned that perhaps I am applying the interpretation incorrectly. So again, we still remain with the idea that one can interpret the Torah and write that as a Torah commentary. But one cannot use that on the halakhic plane. The rationale is a different one—not because the reason is incorrect, but because the reason is not being applied correctly. And that is the point, against the background assumption that the Torah is presumably formulated correctly and precisely. If the Holy One, blessed be He, wrote something, then He wrote what He intended.

I’ll just note that this actually connects to something else. I said in the previous class—we were talking about Passover—I said it was just in honor of Passover, but it is related to this issue, because there I derived law from the reason for the prohibition of leavened food. Just in honor of Passover. I’m only saying this here to get us to the stage where we are now, and I’ll return to it later. Because what I did there was derive law from the reason for the verse—not I, rather I claimed that Maimonides did that. I’m allowed, because Maimonides did it, not me. So Maimonides derived law from the reason for the verse, and the question is how that fits with the claim that we do not derive law from the reason for the verse. That’s what I was dealing with. I tried to show that Maimonides consistently derives law from the reason for the prohibition of leavened food. And that raises the question: how can that be? And to that question are joined many other examples in which we see that purposive interpretation is indeed used—what the Jewish law is. We do derive law from the reason for the verse in various places.

And then I’ll have to present several limitations on this matter of “we do not derive law from the reason for the verse.” In truth it is a very narrow rule. I mean, it looks like a sweeping rule: we do not derive law from the reason for the verse; whatever is written in the Torah we do as is. But between ourselves, there is not one verse that we actually do exactly as is. We are always interpreting. So the question is how this squares with this sweeping rule that we do not derive law from the reason for the verse. It’s not sweeping at all. What? Did I speak about the general rule? Yes, I spoke about the general rule. Here, this rule does not merely have a few exceptions—it itself is one big exception. So I’ll get to that. There are several very significant limitations on this matter of not deriving law from the reason for the verse, limitations that narrow its application considerably.

But just one more remark on what I said about Maimonides. What Maimonides is really telling us—or the first tanna, according to how Maimonides explains him—is that if we want to know the law, we must use linguistic interpretation and not purposive interpretation. That is, we need to take the wording of the verse and understand what follows from it, and not get into the question of why and what and what the Torah wants to achieve—what is called purposive interpretation. Now the claim is that usually—not usually, but if I do the purposive interpretation correctly—then it will always come out parallel. Meaning, the result will always be the same result I get from interpreting the verse itself. Therefore, where that is not what happens, and then the question arises whether to derive law from the reason for the verse or not, apparently you made a mistake in the purposive interpretation. That is basically what the first tanna says. The assumption is that purposive interpretation ought to fit linguistic interpretation. And that is a very interesting point.

I think we spoke about this once before—I no longer remember on what topic or in what context—but what Maimonides and the first tanna are really saying here is that we can relate to the same thing on several different planes, and on each of them carry out interpretive processes that seem, on the face of it, completely different, yet ultimately arrive at the same result. There is a correlation between linguistic interpretation and purposive interpretation. Right? That is basically the claim. And if there is no correlation, then you simply made a mistake in the purposive interpretation. Meaning: if I am precise with the language of the Torah, I arrive at halakhic conclusion X. If I do purposive interpretation—that is, I explain what the Torah intended to achieve—I will also arrive at conclusion X. It is the same conclusion. That is the assumption. Therefore we do not derive law from the reason for the verse, because they are really supposed to run in parallel. And if they do not run in parallel, then apparently there is a problem with the application of the purposive interpretation.

This is an example of such parallels. Again, I don’t remember in what context we discussed it, but many times in logic people speak about semantics and syntax. That is, we can perform some formal calculation and reach some conclusion, reach some result; or we can enter into the meanings and think about what they say and reach the same conclusion—and it will be the same conclusion. In logic there is even a theorem about this, what in Hebrew is called the theorem of adequacy and completeness. I don’t know how to say it in English. Soundness and completeness theorem. Okay. In any event, the claim is that if logic did not maintain this fit between the content—between the semantics, the meaning, and the syntax—then there would be no point in engaging in logic. This is how we do logic: we symbolize, we give form to the philosophical argument, and then we work with X, Y, Z, P, these things, and arrive at a conclusion. And our assumption is that the conclusion is correct. Why is it correct? Because the claim is that the calculation runs fully parallel to the philosophical thinking we would have done, and we would reach the same results.

[Speaker B] So maybe one is a subset of the other.

[Rabbi Michael Abraham] Fine, I’m not—

[Speaker B] Going into precision. Maybe syntax is a subset, and then you have a problem.

[Rabbi Michael Abraham] Fine, and then there are problems. But I’m saying here the assumption is that this—

[Speaker B] Is not a contradiction but an expansion.

[Rabbi Michael Abraham] Doesn’t matter. It is still a problem, because there are places where you won’t be able to do it. Fine. Here the assumption is that this does not exist. Meanwhile I said there is richness here too, and maybe we can return to that remark. So here, regarding this matter, I’ll maybe bring another example. It is commonly thought that there are four planes on which we interpret the Torah: plain meaning, hint, homiletic exposition, and secret—let’s say—which is of course only an acronym, and I don’t even know its sources, pardes. And each of these is of course divided into many shades; there are many kinds of hint, many kinds of homiletics—interpretations—but four types of interpretation. Right?

Now apparently, wherever we want to know what the Jewish law is, we ought to do the calculation on all four planes of reference: do the calculation according to the plain meaning, do the calculation according to homiletics, do the calculation according to the secret. Hint may not really be one of them, but at least according to plain meaning, homiletics, and secret. We don’t do that. Usually we do—

[Speaker C] Does each plane have the same weight?

[Rabbi Michael Abraham] No, it’s not a matter of weight, I’m saying. Is it a level of importance? No, that’s not the issue. That’s why I say, we don’t need to get into that question, because they all have to be parallel. If they are all correct, why don’t we do the calculation on all the planes? Not because one is more important than another, but because they are supposed to give us the same result. That is, the assumption is that if we do the calculation in the language of the secret, in the world of the secret, we will arrive at the conclusion that the law is such-and-such, and if we do the calculation in the language of the plain meaning, we will arrive at the conclusion that the law is such-and-such—those conclusions are supposed to coincide.

Therefore there is no need always to check with the tools of the secret if we already did the calculation with the tools of the plain meaning—even though there are people of that sort. For example, they say this about Rabbi Chaim of Sanz, that he never issued a ruling before checking whether it also fit the matters of the secret. But in practice there are many halakhic decisors who are not versed in the secret at all and are still considered extremely important decisors, and people heed them, and no one—

[Speaker C] Are there also disputes because of the secret? Rabbi Eliyahu rules according to the secret.

[Speaker B] That’s more problematic, because then it means there isn’t overlap; then it means there isn’t overlap.

[Rabbi Michael Abraham] And that’s true. Obviously this is some sort of hypothetical claim, saying there is overlap, and therefore I need not get excited—I can do the calculation on the revealed level and everything is fine. In practice, though, there are differences. There is the Magen Avraham in section 32, there in the laws of phylacteries. The Magen Avraham says there that according to the secret one gets a different result—I don’t even remember which law it is talking about—and then he says: but when there is a dispute between the revealed and the hidden, one rules according to the revealed. That is what the Magen Avraham says. Of course, the kabbalists rule according to the hidden, and the Magen Avraham himself was a man of the revealed, so they rule according to the revealed. And on that itself there would be the same dispute between the kabbalists and the non-kabbalistic decisors—the common denominator is to bake it, okay. In any case, yes, you do see contradictions. The Vilna Gaon, for example, in many places tries to reconcile what comes out of the plain meaning—for example, with the directions of beds, for those who know. There there is a difference between the plain meaning and the secret, and the Vilna Gaon tries to reconcile it.

[Speaker B] But if all of them are absolute, then it can’t be—if all of them were absolute, then they would all have to—but they are all according to the—

[Rabbi Michael Abraham] You are absolutely right; that was my point. This conception, as if they are supposed to be completely parallel, would be correct if each of them were mathematics. Meaning, if I could do the calculation of the plain meaning and arrive at an agreed, clear result with logical certainty—as if there were no doubt at all that this is the conclusion—and the same in the secret, then maybe—I don’t know, that is the thesis at least—I would arrive at the same place with both systems of tools. But that is not the situation. We know that even in the revealed there are many disputes. So what does it mean that the secret accords with the revealed? The secret accords with the revealed—which side of the revealed? There are disputes.

[Speaker C] Even the revealed doesn’t fit the revealed.

[Rabbi Michael Abraham] Exactly. So to which of the revealed answers is the secret supposed to correspond? Now notice, this is trickier than what I just said, because one could say that this still isn’t relevant. Why? Because if there is a correlation between the secret and the revealed, then the person who did the calculation in the revealed—when he himself does the calculation in the secret—will arrive at the same answer. And if someone disagrees with him, no problem: he will disagree with him both in the revealed and in the hidden, because these calculations are basically supposed to correspond. But that of course assumes some sort of naïve fit that does not exist—what I once called a local fit. A local fit means that every principle in the revealed has a parallel principle in the hidden, and then really you are just talking—like speaking Hebrew and English. Fine, you say the same thing in two languages. That is not the situation there. The fit there is more interesting.

There is a book by Godel Escher Bach—Hofstadter’s book—and there he gives some very nice examples of this matter. He presents a puzzle there, yes, the MI puzzle, M-I. He says there: we start with MU, M-U. There is a language with three letters, M, U, and I. Okay. He defines a puzzle—M, U, and I. The one legal word known to us is MU. Fine. Now there are, I think, four production rules—meaning: if something ends with U, you can duplicate everything after the M; if there are two I’s, you can delete both of them; and so on, things of that sort—rules for how you move from one word to another. Meaning, if one word is given, then by means of these rules you can derive other legal words from it, and thus you can define a language: you define the set of legal words in the language as the set of words that can be derived from the initial word by these derivation rules.

And then he asked—ultimately, okay, who asked this?

[Speaker B] Ultimately he—

[Rabbi Michael Abraham] Showed, he showed that a certain word is not legal. Meaning, he proved from the rules that that word is not legal. And afterward he said: I can formulate all of this in a completely different language. Meaning, I can say, for example, let M be three, I be one, and U be zero. To represent them with numbers instead of letters—that’s allowed, right? It is just another notation. But now notice that the first legal word is thirty-one. Right? If M is three and U is one—thirty-one. And now every derivation rule can be formulated in mathematical language. For example, if something ends with one, then you can multiply it by ten and add one. That’s really three hundred and eleven—duplicating the one. Right? So he translates it into arithmetic, basically. And now you can work with a computer, meaning you can work with mathematical tools, and see whether you can get to the word thirty from the word thirty-one. And maybe it is much easier that way. In any case, when you do it in other languages, a different world of rules opens up before you. But it is absolutely clear that there is a correlation. Meaning, if you prove it with algebraic tools, you will also arrive at the correct answer with the typographic tools, with the M, I, and U. Therefore these are really two different languages but there is a complete fit between them. That is exactly what I called a local fit, and therefore it is obvious that if, say, there were a disagreement in the arithmetic language, there would also be one in the typographic language. So the fact that there are disagreements in the revealed does not mean there won’t be the same disagreements in the hidden, and there could still be a fit.

But now I want to use this example one step further. Suppose that in the end there is some set of words that are the legal words in the language. Okay, let’s say we found them all—it is probably an infinite set, doesn’t matter—but suppose we found them all. Now I take this set of words and want to define it using a different grammar. I start, say, from the word MMMU, let’s say that is one of the legal words, and I define, say, ten completely different derivation rules—not related to those rules—which just happen to produce only legal words when applied. Completely different rules; no connection to the first ones. But what they generate—the set of words you get—is the same set of words. Two different ways of defining the same thing. There is no principled objection to such a thing, right? There’s no problem with that.

[Speaker B] So then there is a one-to-one mapping between the rules?

[Rabbi Michael Abraham] No. No, that’s not correct. There is no mapping between the rules. The results—exactly, that is the difference. There is a mapping in the results, not in the rules. I think that the fit between the revealed and the hidden, if such a fit exists at all, is of that kind—not a fit between the rules. And therefore where will the implication be? What we earlier called a decree of Scripture. Right? Where there is a local fit, there can never be an exception to it. Right? If the rules are parallel—typographic versus arithmetic—every word that is legal here will be legal there, and vice versa.

[Speaker B] Two languages are really the same thing, just described differently.

[Rabbi Michael Abraham] Exactly. Like Hebrew and English, or whatever—some kind of translation. But where the fit is global, then theoretically it can be there, but one should expect exceptions. That is, one should expect this set of rules to generate a certain set, and that set of rules to generate a set that is almost the same but maybe has two or three words that are not, or vice versa.

[Speaker B] And that is what happened.

[Rabbi Michael Abraham] When you have an infinite set, I understand.

[Speaker B] Now I—

[Rabbi Michael Abraham] Am saying that usually that won’t be the situation. Usually the situation will be that there is a global fit in principle, but there are a few exceptions. In a local fit that cannot happen. Now how do I know that the fit between the secret and the plain meaning is of the second type and not the first? I have a proof. Rabbi Eliyahu? Maybe, but let’s say he is simply mistaken. The proof is the concept of a decree of Scripture. If there were a fit between the revealed and the hidden, then that fit would never break. That is, everything that has an explanation in the revealed would have an explanation in the hidden and vice versa. A decree of Scripture means something that has no explanation, or no full explanation, in the revealed, but apparently does have one in the hidden—otherwise why would it have been said? It has some explanation, but not an explanation in the revealed, not in the system of tools we have.

Where there is a deviation from that fit, it means the fit is not local, because in a local fit by definition there is a one-to-one correspondence in the results. Therefore—yes, you know, when I spoke about this once I brought the example of Cartesian and polar coordinates. You know, when you define a point. Now polar and Cartesian are still almost a local fit, because I can express X in terms of r and theta and Y in terms of r and theta, and vice versa. But I cannot express X by theta and Y by r; it is not trivial in that sense. And the point is that in a fit of that sort there is indeed one point at which there is no fit: the origin. At the origin, look, X equals zero and Y equals zero, but in r and theta that is a whole line—all the thetas. R equals zero and every theta.

[Speaker B] R zero and any theta.

[Rabbi Michael Abraham] That’s what I said—all the thetas, r equals zero and all the thetas. Meaning, it is a line, okay? So the correlation between the two descriptions breaks at a certain point. It breaks, and there topological defects appear, for those who know. That break stems only from the fact that the fit is not a local fit. In a fit like, I don’t know, a rotation or a translation—

[Speaker B] You won’t have such a thing,

[Rabbi Michael Abraham] Because in a rotation and translation it really is something local, and therefore there will never be a break of this kind. So if I return to our matter, the claim is that the fit between the revealed and the hidden is a global fit. And the problem is that when it is a global fit, then there can indeed be places—things—where there is an explanation in the revealed but not in the hidden, or vice versa. Meaning, a decree of Scripture means there is an explanation in the hidden and no explanation in the revealed. There can also be opposite situations, where there is something the revealed says but the hidden rejects—that is the Magen Avraham—and then once again we are in some kind of conflict and must decide which prevails.

So when I speak about two planes of reference that are supposed to fit, that is a fairly general statement. What I said about Maimonides—that we do not derive law from the reason for the verse because the linguistic analysis and the purposive analysis are basically supposed to give the same results, even though we use completely different tools—here we use tools of, I don’t know, textual understanding, and there we use conceptual tools, meaning what is right to do and what is not right to do and all kinds of things of that sort, which are completely different tools. But we are supposed to reach the same results.

Now it is naïve to think that we really always reach the same results—despite all due respect to what the topic here is doing, it is obviously not true. Therefore there is no need to be surprised, as I said earlier, that in many places there are situations where we do use purposive interpretation in Jewish law despite the rule that we do not derive law from the reason for the verse. And why? Because the fit is a global fit. When the fit is a global fit, there are many places where it breaks—the topological defects.

[Speaker C] If there is a mismatch here, why don’t we go by the wording?

[Rabbi Michael Abraham] Wait, I’ll get to that. I’ll have to look at every mismatch and discuss it. Okay, so I’ll bring one or two examples just to give us—give us—yes, you know what, maybe I’ll say one more sentence about the fit. You know the example of John Searle, the American philosopher—the Chinese room. In the Chinese room there sits a man, a person who speaks Hebrew, yes, and no other language. He sits in a room with two windows, input and output. Through one window, questions in Chinese come in to him, and through the other he has to answer in Chinese. He has no clue about Chinese; he knows only Hebrew, okay? He has a box full of Chinese letters, infinitely many, and they feed him questions, and every time he answers incorrectly he gets an electric shock. An ethics committee has been appointed—they give him electric shocks until slowly he starts learning to give the right answers to the questions. Machine learning. Yes. So the final state, after infinite time, when he already knows how to give the correct answers to all the questions—the question is whether he knows Chinese. Linguistically yes, purposively no, right? He has no clue; he does not know Chinese. What he knows is that for combinations like this he should answer with combination like that, and for combinations like that he should answer with another combination.

By the way, since you mentioned machine learning—many people really make this jump and say: if there are learning machines, then apparently they also have consciousness or will have consciousness. That is not so; nothing requires that. This is exactly it: machine learning gives you the efficiency of use, the relation between input and output, but it does not cross into the cognitive dimension of understanding, which is what exists in human beings.

[Speaker B] Learning is deriving rules from—

[Rabbi Michael Abraham] Fine, I don’t want to get into that whole topic now. I’m only saying that even in machine learning you really see this tension, and some people are not sufficiently aware of it. Now there too it is an example of the same thing, because what is he doing? He is not deriving law from the reason for the verse. The fellow sitting in the Chinese room. Right? What does he do? He receives a question formulated in Chinese and gives an answer formulated in Chinese. He has no clue what any of it means. What it says. But he knows how to match an answer to a question without it passing through his understanding. And therefore he is operating on the—let’s call it the linguistic plane, although it isn’t even linguistic, it is just technical manipulations—but let’s say he is operating on the linguistic or syntactic plane without entering semantics, meaning without entering into meaning.

[Speaker C] Meaning it’s possible—that’s a very strong assumption of his. What? If the database of questions is infinite, apparently that’s not true.

[Rabbi Michael Abraham] Could be. Could be. But let’s say for a finite set of questions it is true, so from my perspective that is enough to show there is a possibility of operating on the syntactic plane rather than the semantic one. Okay, so—

[Speaker B] It feels a bit artificial, doesn’t it?

[Rabbi Michael Abraham] He understands that very well. Or maybe he doesn’t understand—even that, by the way, one could say: he doesn’t understand. It may be that he just reacts to it. Exactly. But that does not necessarily mean it passes through his cognition. Fine. So that is another example showing that if, say, we do linguistic interpretation of the Torah, someone might say: wait a second, you’re not addressing the question of what this is trying to achieve, the purposive interpretation, the reason for the thing—how will you reach the correct result? The answer is that I will reach the correct result. Meaning, or at least there is a chance I will reach the correct result. Because many times using linguistic tools in interpreting the verse, in understanding its language, without entering into the question of what it seeks to achieve, will bring me to the correct result even without referring to the meaning of the matter or the purposive interpretation.

I’ll maybe finish with two examples where we see deviations, and that will open the next chapter, which is deviations from this rule that we do not derive law from the reason for the verse. The first thing is tooth and foot in the public domain. The Rif at the beginning of tractate Bava Kamma—because we know there are four primary categories of damages: ox, pit, maveh, and fire. Within ox there are three primary categories—three or four—there are horn, tooth, and foot. Horn, whether innocuous or forewarned, and tooth and foot. There are different laws distinguishing them. For horn, the innocuous one pays half damages; after three times, full damages. Tooth and foot are exempt in the public domain. Horn is liable in the public domain, and the forewarned one pays full damages from the outset; tooth and foot pay full damages from the first damage, but not in the public domain.

Now the Rif at the beginning of Bava Kamma writes as follows: “And tooth and foot in the public domain are exempt, because that is their normal way,” because it is their way to walk there in the public domain. So on that the Rosh brings the Rif’s words and comments: “It is puzzling to me—why did he need to explain the reason for their exemption as being because it is their normal way? For it is written, ‘and it consumes in another man’s field,’ and we interpret this as ‘in another man’s field’ and not in the public domain.” It says, “and it consumes in another man’s field,” and from here we understand: in someone else’s field, not in the public domain. So from that we derive the exemption of tooth and foot in the public domain. Why does the Rif need to give me the reason, that it is their normal way? Or in other words, the Yam Shel Shlomo formulates it more sharply: the Rif here is deriving law from the reason for the verse.

Now I’ll tell you more than that. The Rosh then says this: “And it is possible that he came to explain the reason of the verse—why the Torah exempted it in the public domain—because it is its normal way to go in the public domain, and it is impossible for the owners always to be behind them.” You can’t constantly follow your animal around; animals go in the public domain. Someone who puts fruit in the public domain should take into account that animals are walking around there. I can’t walk after my animal all the time. Don’t leave your fruit in the public domain. So the Rif is only explaining the Torah’s law that tooth and foot are exempt in the public domain. But horn is liable in the public domain even though it is its way to go there, because once it is forewarned—he knows it gores—its owner ought to guard it. Never mind. And this is also so with an innocuous horn, though that’s a bit more difficult, but let’s leave that aside; there is also a remark on that.

And there is a practical difference based on this reason—now this is surprising. I would have expected him to stop here, because if the Rif came to explain the law learned from the Torah, “in another man’s field” and not in the public domain, okay, so the Rif explained the reason; he wrote a Torah commentary. But no. The Rosh says—and the Rif himself did not say this, the Rosh says it regarding the Rif. At first the Rosh asks: why do I need a reason, since it is explicit in the Torah? Then he says: I need the reason in order to explain why the Torah says that tooth and foot are exempt in the public domain. But not “why” in the sense of understanding the verse—not as biblical commentary—but “why” in the sense of making a halakhic decision now. And that is what he says.

And then he says: “And there is a practical difference based on this reason: if there were a long piece of wood lying partly in the public domain and partly in private property, and the animal stepped on it in the public domain and it broke vessels in the private property—since it is its normal way to walk and tread there, they are exempt.” Meaning, if there were a long board, one end in the public domain and the other in private property, and the animal is walking in the public domain, steps on that board, the board moves into the private property and knocks over and breaks some vessel there—the Rosh says that according to the Rif he is exempt. Because if you read the verses as is and do not derive law from the reason for the verse—what? Then the exemption is only in the public domain; one is liable for damage in private property, not exempt there. And here the damage happened in private property, so he should be liable.

[Speaker C] Does that count as the animal’s normal way of walking? What? The animal didn’t enter.

[Rabbi Michael Abraham] His claim is that the damage has to be “in another man’s field”—I’m just following him, the specific issue itself isn’t important to me right now. But according to the Rif’s explanation, that the basis of the exemption is that it is normal for it to walk in the public domain, here too it walked in the public domain in its ordinary way. What do you want—that the owner walk after it? The owner cannot constantly walk after it. So if you left some long board there, the owner isn’t supposed to trail after the animal, and therefore he is exempt. But then it really comes out that when the Rosh understands the Rif as explaining the exemption of tooth and foot in the public domain, he does not claim that the Rif is merely writing Torah commentary. Rather, he claims that the Rif is saying this as practical Jewish law—that one can derive a halakhic conclusion from it. In other words, he is deriving law from the reason for the verse.

[Speaker D] The Rosh—the Rosh infers from the Talmud that this is an example. Wait, wait, okay, not only from the Talmud—that this is, this is an example. So then already, you can see in the Talmud its definition…

[Rabbi Michael Abraham] But about the Talmud / Talmudic text itself, I can ask: how does the Talmud / Talmudic text do this? That’s my next example. So my next example is—really, you’re right—that it’s basically just a continuation of this example. The Talmud / Talmudic text there, at the beginning of tractate Bava Kamma, when it discusses that Mishnah about the primary categories of damages, talks about tooth and foot, and horn and pit, and all those matters. It discusses there whether their derivative cases are like them or not like them. And then the Talmud / Talmudic text starts getting into the definition of each primary category of damage. So we would have thought: the Torah says tooth and foot, so tooth and foot mean what the animal eats, or what it tramples with its foot as it walks. But the Talmud / Talmudic text says no. For example, goring, biting, crouching, and kicking are derivatives of horn. Even though goring is done with the foot—or kicking, that’s done with the foot. Goring is with the body. But kicking is with the foot. So why is that considered horn? So the Talmud / Talmudic text says that horn, tooth, and foot are conceptual primary categories of damages. They are not literally categories of damage based on the animal’s organs. So horn is damage done intentionally to cause harm, and it is unusual. Tooth and foot are things whose normal way is to go and cause damage. Okay? So there’s some kind of—there is benefit from its damaging, that’s tooth, okay, there are all kinds of characteristics of each of the primary categories of damages. So in fact, the ideas of the Rif already appear in one form or another in the Talmud / Talmudic text itself. Because the Talmud / Talmudic text itself talks about tooth and foot, and then says: since their normal way is to go and cause damage. True, in the Talmud / Talmudic text it’s less prominent, because in the Talmud / Talmudic text they are basically just explaining the concepts that appear in the Torah. The Torah has tooth and foot; now, for a moment, we interpret that as all these kinds of damages whose normal way is to go and cause damage. No, we are not directly deriving from that some halakhic ruling, even though halakhic rulings do come out of it. Various conclusions come out of it—if there are damages that are not done with the foot, if they are done in their normal way, then you would be liable for full damages from the first time, because it resembles foot and not horn. That has halakhic implications, but it is less conspicuous in the Talmud / Talmudic text. In the Rif it is more conspicuous, and that is why they discuss the Rif, because the Rif takes this and turns it into a halakhic ruling. So that whole discussion there in Bava Kamma about the definition of the primary categories of damages also really raises the question of deriving the reason for the verse. Because, for example, the Torah speaks about “if an ox gores the ox of his fellow” or “and it grazes in another’s field,” so it speaks about certain specific cases, and suddenly we turn that into these abstract damagers: whose normal way is to go and cause damage, unusual, intention to cause damage, there is benefit from its damaging, its initial making was for damage. All kinds of characteristics of that sort—which of course the Amoraim or the Tannaim are bringing with them from home. Meaning, this is basically interpretation—seemingly purposive interpretation, the reason for the verse—and the question is: how do they do that? Okay? So we’ll stop here.