Majority in Halacha and in General 2, Lesson 7

Table of Contents

• Not convicting on the basis of statistical evidence versus convicting on the basis of witnesses
• The prisoners and buses examples, and the claim of a categorical difference between kinds of evidence
• Three types of explanations in the world of Jewish law: probability, legal utility, and a sense of justice
• David Enoch: “knowledge,” the sensitivity criterion, and an attempt at a utilitarian explanation
• The speaker’s first explanation: “ninety-nine percent” as ignorance rather than statistics, and the connection to a majority present before us / not present before us
• The speaker’s second explanation: a majority in matters that depend on human choice is not a “rational majority”

Summary

General Overview

The text continues the discussion of statistical or probabilistic evidence and emphasizes that in law it is generally not accepted to convict on the basis of pure statistical probability, even though numerically it may sometimes seem stronger than eyewitness testimony, which is never completely certain. The text presents examples of a prisoners’ revolt and a “blue bus” accident to show that there is a categorical distinction between testimony and probabilistic calculation, and it maps out three possible types of explanation in the world of Jewish law: reliability / probability, legal utility and efficiency, and a judicial sense of justice that cannot be reduced to the first two. After examining David Enoch’s explanations concerning “knowledge” and the criterion of “sensitivity,” and the connection to deterrence, the text suggests that the real gap is probabilistic but in a tricky way: probability based on absence of information is not statistical knowledge and therefore is not a basis for conviction. It then adds another explanation, according to which when we are dealing with actions that depend on human choice and reasoning, statistics about “most people” do not determine what happened in the particular case.

Not convicting on the basis of statistical evidence versus convicting on the basis of witnesses

The text states that there is broad agreement that you do not convict a particular prisoner just because ninety-nine out of a hundred took part in the revolt, even though it seems there is a “ninety-nine percent” chance that he participated. It contrasts this with a case in which two eyewitnesses testify that the person participated, and even though witnesses can err by some percentage, we still convict on the basis of their testimony. The text asks why testimony is preferred over probabilistic calculation even when the probability supplied by testimony may be lower than the statistical calculation.

The prisoners and buses examples, and the claim of a categorical difference between kinds of evidence

The text brings David Enoch’s example of a city with two bus companies, where ninety percent of the buses are blue and ten percent are red, and a bus hit a car and disappeared. It notes that we do not hold the blue company liable just because most of the buses belong to it. The text argues that when there is testimony from someone who saw a blue bus, we do impose liability, even though there is still some possibility of mistaken identification. The text concludes that the difference is not determined by the percentages alone, but by a principled distinction between testimony and statistical probability.

Three types of explanations in the world of Jewish law: probability, legal utility, and a sense of justice

The text presents a simple probabilistic explanation according to which evidence is accepted when it is more reliable, and alongside it an explanation based on legal efficiency and utility that does not depend only on reliability. It illustrates this through the rule “the burden of proof rests on the one who seeks to extract property from another,” and argues that there is no basis for saying that statistically the current holder is more likely to be right than the claimant. Rather, the rule is needed in order to prevent a situation in which “everyone would pounce on everyone” and the system could not function, and also in order to “close loopholes” even against a minority of wrongdoers. The text adds a third kind of halakhic consideration that is neither utilitarian nor probabilistic, and calls it a sense of judicial justice or an intuition that “this is how one ought to act.” In this context it also places ideas such as migo as a kind of “strength of claim,” and the rule “a person does not make himself into a wicked person” as a normative determination that is not explained only by fear of torture or by lack of reliability.

David Enoch: “knowledge,” the sensitivity criterion, and an attempt at a utilitarian explanation

The text describes the proposal to distinguish between evidence that gives “knowledge” and probabilistic calculations that are not “knowledge,” but it presents Enoch’s criticism that without legal consequences this distinction does not explain why one kind of evidence should be accepted and another rejected when the chance of error is similar. The text brings the lottery example: before the results are published, a person does not say “I know” that he did not win, even though the chance of winning is one in a million; after the results are printed in the newspaper, he does say “I know” that he did not win, even though there is still some possibility of a printing mistake. The text presents Enoch’s “sensitivity” criterion in counterfactual terms, according to which a belief counts as “knowledge” when, if reality had been different, the belief would also have changed. It also presents Enoch’s attempt to explain that in law the issue is not “knowledge” but utility: when a verdict is not sensitive to the facts, punishment does not achieve deterrence because the suspect will be convicted “either way.” The text presents objections to this deterrence explanation, including the fact that it depends too heavily on the specific situation, and also the bus case, which is civil and not aimed at deterrence, and yet statistical evidence is not accepted there either.

The speaker’s first explanation: “ninety-nine percent” as ignorance rather than statistics, and the connection to a majority present before us / not present before us

The text argues that the use of “ninety-nine percent” in the prisoners’ revolt is “nonsense,” in the sense that there is no random process or measurement here that generates a genuine probability distribution regarding that specific prisoner; there is only a guess made in the absence of information. The text explains that the fact that ninety-nine participated and one did not is knowledge about the group, but from that there is no real probabilistic knowledge about the identity of the person behind the mask. The text distinguishes between a case in which the error rate of witnesses can be determined by experiment and testing and become a scientific generalization, and a case in which “statistics” is just a default assumption when there is no data. It emphasizes that in law “I do not gamble,” and if there is no information, then “I do not act.” The text connects this to the distinction between a majority present before us and a majority not present before us, as presented in the previous lecture, and argues that what looks like probabilistic data in cases such as the prisoners and the buses is really a made-up number that is not grounded in knowledge.

The speaker’s second explanation: a majority in matters that depend on human choice is not a “rational majority”

The text opens with an introduction from the Talmudic passage in Bava Kamma 46 about the dispute between Rav and Shmuel whether, in monetary law, one follows the majority in the case of “one who sells an ox to his fellow and it turns out to be a goring ox.” It presents the halakhic conclusion in accordance with Shmuel: “In monetary matters we do not follow the majority; rather, the burden of proof rests on the one who seeks to extract property from another.” The text cites the question raised by the medieval authorities (Rishonim): how does this fit with “follow the majority,” under which in a religious court we do follow the majority even in monetary cases? It presents that Tosafot offers several distinctions, including the claim that there is such a thing as an “inferior majority,” such as “most people buy for plowing.” The text brings, in the name of Rabbi Shimon Shkop, on the basis of Nachmanides, that such a majority is not a “rational majority,” because if someone belongs to the minority and buys for slaughter, that does not constitute any deviation from the laws of nature or from normal human conduct. Therefore, the mere fact that most people act differently does not negate his individual claim. The text illustrates this with the parable of a box of balls: if someone chooses with his eyes closed, the probability follows the ratio in the box, but if he chooses with his eyes open, the choice depends on the chooser’s preference and not on the distribution of balls. It concludes that when we are dealing with an action that results from decision and motive, statistics about the surrounding circumstances are not a decisive tool. It distinguishes this from testimony, where error is not the result of choice but a chance event relative to the capacity for perception and identification, so one can speak about it in probabilistic terms and rely on the majority. It concludes that the second explanation is similar to the first in the logic of not using statistics, but not identical to it, and it may even apply in a majority not present before us.

Good morning. Good morning. Last time we started the topic of statistical or probabilistic evidence, and today I want to continue with that, and with the connection to the discussion we had earlier about a majority that is before us and a majority that is not before us, and the whole context of majority—we’ll see that in a moment, later on in the discussion. I’ll just briefly remind you what we already touched on. I spoke about the basic problem that in court we do not convict on the basis of statistical evidence. One example brought in an article I saw is a prison riot in which prisoners attack a guard. There are one hundred prisoners; ninety-nine of them took part, and one did not. Now they bring—no one knows which one out of the hundred was the one who didn’t participate. They bring him to trial in court and say there’s a ninety-nine percent chance that he took part, because ninety-nine out of one hundred did participate in this matter, and a ninety-nine percent chance is usually enough to convict, as it were, in a criminal trial. So ostensibly we should have convicted him. But there is, I think, sweeping agreement—at least as far as I’ve seen—that in such a case we do not convict. In contrast to that situation, take a case where there are two eyewitnesses who come and say that this fellow participated in the riot, in the attack on the guard. Witnesses too, as we know, are not one hundred percent reliable; there may be a few percentage points where they missed something, didn’t see clearly whether it was him or not him. In such a case, if more or less we understand that we heard eyewitness testimony, then despite the fact that, again, you can’t rule out the possibility that they didn’t see well or missed something, in that case we do convict. And the question is: what is the difference? Sometimes on the probabilistic level it’s even the reverse. Meaning, say in the first statistical case there is one percent that he did not participate in the attack. If we say that in such a situation the reliability of the witnesses is, say, a three percent error rate—ninety-seven percent reliability—then from a probabilistic standpoint the second piece of evidence is actually weaker, and still in the first case we do not convict, while in the second case we do. The question is why. What’s the difference? In the end, what matters to us is the chance of being right or the chance of making a mistake. Refutation? What? No, the criterion of being subject to refutation is a halakhic criterion; right now I’m talking about the statistical reliability of the evidence. Once you bring in the whole framework of probabilities for refutation, that’s already very remote. That each set of witnesses might be mistaken? Fine, that’s another discussion. Right now I’m comparing statistics to some other possibility of error. What about witnesses? In the first case there is no testimony. In the first case there is no testimony; we only know that one didn’t participate and ninety-nine did. We have statistical evidence; nobody saw anything. All we know is that ninety-nine took part and one did not. They’re all wearing masks, fine, but the cameras showed that only one didn’t participate. Now we don’t know who was behind the mask. But who is making the calculation—is it the religious court making the calculation? Let’s say. Okay, but in the second case, from the point of view of the court, there are witnesses, and there too someone is testifying to what… But the witnesses can be mistaken; there too it’s only ninety-nine or even fewer percent. I’m saying, let us suppose for the sake of discussion that in such a situation the witnesses are mistaken three percent of the time, and of course that can happen. There are many situations involving mistaken identification. The Talmud talks about identifying a face and how one identifies someone who died in order to permit his wife to remarry—it’s not so simple. So obviously there is some certain percentage of error even with witnesses; that’s obvious. And the question is why we prefer testimonial evidence over the first kind of evidence, when statistically speaking it’s more or less the same thing, or maybe the witnesses’ evidence is even worse than the statistical evidence. That’s one question. A second example, one that David Enoch brings in his article on this matter. He says, suppose in a city there are—I don’t remember if I brought this up then, so I’ll bring it now—suppose there are two bus companies in town, a blue company and a red company. Let’s say ninety percent of the buses belong to the blue company and ten percent to the red company. Now a bus hit a car and disappeared. We don’t know which bus it was. We want to sue the blue company because there is a ninety percent chance it was one of their buses. Ninety percent of the buses are theirs. Ostensibly this is very similar to our previous case, and in such a situation of course we do not convict. But if there is testimony from someone who saw the bus and claims it was blue, then we would convict, even though the witnesses may have been mistaken and it wasn’t blue but red. There is some percentage that they made a mistake. In such a case we won’t get excited about that; we’ll obligate the company to pay. But in the first case we won’t. And again, what determines the result is not the probability. Meaning, it’s not the percentages. Rather, there is some categorical difference between these two types of evidence, and the question is why. That’s basically what bothers many people; there is a great deal of legal writing on this issue, or legal philosophy on this issue. In principle, they distinguish there between several kinds of possible explanations. Let me first add one more point that I noted last time. When we look for an explanation of, say, an evidentiary principle of this kind, in Jewish law too we can think of several kinds of explanations—actually, it seems to me, more or less three. One kind is an explanation based on probability—that is, the reliability of the evidence is better or worse, and that determines whether it is accepted or not accepted. Okay, so it’s a probabilistic question. The question whether the evidence is good or not good—that’s of course the simple, natural explanation. There is another consideration, and that is what we might call legal efficiency or legal utility. Many times the probabilistic consideration may not be decisive, but there is legal logic in proceeding in this way. The legal utility is good. For example: the burden of proof rests on the claimant. I don’t think there is any basis for the claim that it is more likely that the object belongs to the one holding it than to the claimant. Why? Why assume that precisely the claimant is the liar and not the one holding the object? I think we discussed this—that what is written, that whatever is under a person’s control is his, is a presumption regarding objects in general, but not regarding objects that are in legal dispute. Objects that are in legal dispute—plainly both sides have a presumption of propriety. I have no reason to assume one of them is lying more than the other, and if I had no other data I would estimate it fifty-fifty. And still, legal logic says that the burden of proof rests on the claimant. Meaning, we place the burden of proof on the plaintiff. Why? You can explain it in all sorts of ways: otherwise you can pounce on anyone in the market and say to him, “Bring me my object.” Meaning, “give me the object—it’s mine.” He doesn’t know you, nothing. Fifty-fifty: either he’s lying or you’re lying. And I could grab anyone on the street that way. There would be no end to it; the legal system could not function like that. And that itself says that the probability is that the person holding the object… No, a presumption. No, wait, wait, I’ll tell you. No, it doesn’t say that, because the probability is created only after I establish the principle—we’ll see in a moment. Meaning, I can basically pounce on anyone in the street and tell him, “This object is mine.” Not if you’re a decent person and that’s the basis of your fifty-percent assumption—that just contradicts it. Not true. Why would I pounce on you if I… First of all, because there are indecent people in the world. Right? And one indecent person is enough to pounce on everyone; you don’t need to worry about fifty percent of the decent people. For this question one is enough. One such person can wreck everyone’s property, because he is one indecent person and he’ll pounce on everyone, even if all the others are decent. The concern about wicked people can never be based on statistics. Legal defenses, or closing legal loopholes against wicked people, are never based on statistics. And that’s fine; it can’t be otherwise. Therefore the legal logic in all legal systems is like this, not only in Jewish law: legal logic says that the burden of proof is placed on the claimant. No, we don’t let everyone pounce on someone else and say, “That object is mine.” And therefore we do this, even though in principle both sides have a presumption of propriety, and I don’t think you can assume one is lying more or less than the other. So indeed we do this for reasons of legal efficiency. As for Moshe’s earlier comment: if we had not established this rule, maybe you would be right. After we established this rule, it becomes self-reinforcing. Meaning, if we had not constructed this rule, then maybe if someone pounces on another person in the street, maybe he is that one wicked person pouncing on everyone, and then perhaps some statistical calculation could be made. It may be that then there really is a majority in favor of the person in possession of the money; I don’t know. In any case, for our purposes, at least after they established this rule there is no reason to assume that one is more right than the other, but the legal logic is completely clear. The legal logic says there is some legal utility here—order, we want to achieve some kind of order. Therefore the burden of proof is on the claimant. You can offer lots of justifications for that. Meaning, if you hand your property over to someone else, then make sure it’s backed up. Meaning, you can take care of this matter, that you have evidence that it’s yours, or things of that sort. It doesn’t have to be utility. You can say the religious court has no evidence, so it does nothing here. No, but I think I discussed this on one of the earlier occasions—the question is whether this is only a default principle. According to what you’re presenting, it’s a default principle. But according to that, for example, a migo should be enough to extract. Because a migo is evidence, and then there is no default. But when there is a migo, it’s not a default. You’re only saying: well, if there’s nothing to do, I do nothing. Here, when there is a migo—if we say that a migo cannot extract, except for Nachmanides, but according to most medieval authorities (Rishonim), a migo cannot extract—then that means possession has evidentiary weight too. It’s not just, okay, I’m not doing anything, so let it stay as it is. It also has some evidentiary weight. In any case, that’s one consideration: legal utility. And that more or less seems to me to be the range of options in the legal literature. So there are two possibilities: either a consideration of reliability, or legal considerations—legal utility, order, legal certainty, all kinds of considerations that belong to the legal layer and not necessarily to the layer of evidentiary reliability. In the halakhic world there is another type of consideration. Some would call it a decree of Scripture, but a decree of Scripture isn’t a type of consideration. A decree of Scripture—we discussed this before—a decree of Scripture is a source, a verse that says it. But behind the fact that the verse says it, obviously some explanation is supposed to stand. Why does the verse say it? Just because it feels like it? Clearly there is some explanation behind it. What could that explanation be? My claim is that there may be a third kind of explanation, one that people usually ignore. There is some sense of justice that this is how it ought to be. Some kind of—I don’t know—something like a legal intuition that I can’t point to and say what damage will happen if I don’t do it. Meaning, it’s not that I’m bringing a consequentialist argument, saying: if I do this, then look, there will be some legal problem, some legal problem will arise, and that’s what I said last time. Rather, I don’t know, somehow this is the right way to act. I think even the burden of proof resting on the claimant, by the way, has that dimension too. Aside from the dimension of utility and order and preventing everyone from pouncing on everyone else, there is something that says, I don’t know, leave it with him. Meaning, what—you need evidence in order to move something. There is some consideration here—we might call it legal justice or something like that—but it’s not because of probability. Meaning, the probability may be fifty-fifty, at least in principle, and even in such a case the consideration still stands. Migo as argumentative force. Later authorities (Acharonim) discuss that the principle of migo—yes, that I could have lied with a better claim, and therefore believe me that I’m not lying—is the probabilistic explanation for migo. As if it is more plausible that I am right: if I had lied, I would have lied better. But there is also a concept of migo as argumentative force: the very fact that I could have prevailed with an alternative claim gives me an advantage even with this claim. Why? A decree of Scripture. What do you mean, a decree of Scripture? There is no such decree of Scripture. The verse nowhere says this. The Sages decided this, okay? This is Rabbi Breish, the Chelkat Yaakov. So he says: what is this thing? It’s like electricity. Meaning, we transfer force from that claim to this claim. But now he is making this claim, not that one. Why should the fact that he could have prevailed with that claim help here—where does that come from? And there is no verse either; there is no verse teaching migo. Migo is a logical inference. So how can that be? We see here that there is some other type of reasoning. It is reasoning whose goal is not legal utility, and it also doesn’t sit on a probabilistic calculation, on reliability, but on some sort of—I don’t know—sense of justice. This is right; and if it was in his hands and he could have prevailed, then fine, apparently he has an advantage in this situation, or things of that kind. Just some sense that this is how one ought to act legally, something like that, even though it is neither of the previous two explanations. So I think in the halakhic context you can see quite a few principles like that. Self-incrimination—“a person does not make himself wicked,” for example. There too, in the legal world they tend to look for consequentialist explanations. Meaning, if you allow conviction based on self-incrimination, then there will be torture in police investigations, they’ll extract confessions in problematic ways, and therefore we prohibit it. Why? Because there they always look for explanations in terms of either probability—there is some problem, if he confesses then something is questionable here, it’s not reliable—or in terms of legal utility. Meaning, if you recognize the admissibility of a confession, of self-incrimination, then you encourage the police to take problematic steps, to torture him. But in the halakhic context it is clear that it is neither this nor that. It is not concern about torture, and it is not a probabilistic problem either—except for Maimonides, where there is a contradiction in Maimonides, but there is also a contradiction within Maimonides himself. In any case Maimonides is puzzling; it straightforwardly goes against the Talmud. And the plain view is that here there is no question of reliability and no problem of legal utility, and still, at least to me—I don’t know, maybe I’m biased after… And maybe that too is enough. Meaning, there is something about what is fitting and unfitting, or legally right and wrong—not factually. Not because I suspect that this confession is unreliable, but I don’t know, there is some sense that this is the right way to proceed. But in the case of the prisoners, those judges who are trying the hundred prisoners are definitely convicting someone who is innocent. Wait, we’ll get there. We’ll get there—that’s a line of reasoning one can discuss. In a moment, in the continuation. I’m saying that there are basically two kinds, or two options, of explanations in the legal literature, and I think that in the halakhic world a third type is added. Because in the world… I’m saying, I assume maybe even in the legal world too there are certainly those who go in these directions, but in the articles I read it doesn’t come up, and it’s not typical there. It’s not typical because, what does that even mean? A decree of Scripture. Meaning, tell me either that it is not reliable, or that there is some utilitarian consideration because of which I don’t accept it. What else could there be? Yes, but there could be a difference regarding accepting or not accepting testimony that contains an element of doubt. That is very different from the question of where the doubt resides. When I accept a witness with ninety-seven percent reliability, I face two issues: first, whether to accept his testimony or not; and then he pointed to the conviction. What do you mean he pointed to the conviction? A person comes and says, “He is the murderer,” points to a person, and I know he has a three percent error rate. Once I accepted that he is a reliable witness within the bounds of legal reasonableness, then what he says is one hundred percent right. In contrast, in the case of the buses, I myself have to apply the probability in my own judgment. And that is what is invalid about the judging itself. A judge is not supposed to apply probability. Why not? In order to convict a person with certainty. When he decides that this witness is reliable, isn’t he also applying probability? He is also applying probability. In the process, not in the judging. I don’t see the difference; it’s a formal difference. I can present many formal differences between the two situations. I don’t see why that matters. I want to hit the truth—that’s what I want. What do I care whether the ninety-seven percent comes from here or from there? The question is what the chance is that I am right. Does it matter if the witness says, “I’m ninety percent sure that this is the murderer”? If the witness says… the witness himself, you could say, maybe that is actually similar to the buses case, perhaps; I don’t know, there is room to hesitate. But again, these explanations are formal explanations. It’s like refutation for witnesses—maybe for judges too, only with witnesses the responsibility ultimately lies with the witness. If the witnesses are lying, then… The Talmud says, “Hearing should be no greater than seeing.” The Talmud says that if the religious court itself sees the act, it can convict. It can convict because hearing should be no greater than seeing. If through hearing from the witnesses we can convict, then if we ourselves saw it we can’t convict? So he is certain; the witness himself is one hundred percent sure. Right, the witness who is the judge is one hundred percent sure of himself. We assume the witnesses are one hundred percent sure of themselves. The judges know there is such a statistic. Right, but the responsibility is on the witness. So what? Then when you are the judge, I think maybe that is a somewhat weaker claim… His claim is fine, it’s healthy. I didn’t understand. The collective explanation? No, again, that somehow the Torah, or whatever, placed responsibility on… The judges cannot convict based on something that is only maybe; perhaps this is connected to the fact that there may be certainty of error, but if the witness is the one making the mistake, the responsibility is on the witness, not on the religious court. I’m saying it’s a formal distinction. I’m saying: there is a ninety-seven percent chance I am right—is that enough to convict or not? What do I care about all these formal calculations? I want to know what the chance is that I’m right or what the chance is that I’m mistaken. You keep repeating the same question all the time. Explanations like that are formal explanations. All I want is to render a true judgment; that’s what I want to do. The question is really about the dynamic. In truth, a trial is based on witnesses; “the hand of the witnesses shall be first against him.” The witnesses are really the ones here who carry out the judgment; the religious court only has to see that things are going in the right direction, that there are no contradictions, that there is no reason to disqualify the witness, who is not a judge. But really it is the witness who carries out the judicial act here, and not the judge. Okay, but the witness himself is also carrying out a problematic judicial act, because he has only a ninety-seven percent chance of being right. He says, “I’m sure.” He says—who is he to say? If he were saying to the religious court, no no, he says he is sure, but he is not right. We know that even a witness who is sure can be mistaken three percent of the time. What good does all this do? We also know there are lying witnesses, but include that too within the three percent—lying. Lying is not… lying and error together are three percent. I don’t care right now what it is; there is a three percent chance that the testimony is wrong. The question is why I don’t take that into account. But the religious court can also make mistakes, all the time. The religious court knows it can make mistakes. Okay, so how does it rule? Now the question is asked—granted, but what does that answer? So now you have another question. Right. We once spoke about a person’s testimony about himself, confession. Confession is not testimony. And confession is part of the process of repentance; it is not testimony. We are not informing the Holy One, blessed be He, that “we have sinned and betrayed”—He already knows. This is a psychological process of repentance, not testimony. In any case, I’m not disagreeing with the concept of court. A litigant’s admission is like a hundred witnesses in monetary cases. Right. But beyond that, in many cases a person calls the police and says: I killed so-and-so. Okay. And in Jewish law we say no. In Jewish law no; in secular law you need some corroboration. There are debates about the status of confession, of self-incrimination. If the victim is dead, is that corroboration? Corroboration maybe, I don’t know. The victim is dead; otherwise no trial would be opened. The question is who killed him. It’s like what we once discussed with Kishon in “We’re the Best.” You know, that labor-union deputy—what was his name, I forgot. David Dolniker, exactly. He arrives there in “We’re the Best,” builds a villa with horse stables and a magnificent swimming pool. So the investigator from the tax authority comes to him and says: Tell me, you’re some labor-union clerk—where did you get the money to build all this? So he says: I’ll tell you a story you won’t believe, but this is what happened. I dreamed one night that on the thirteenth of July I had to go north to a certain tree near some village, walk one hundred steps north, one hundred steps east, dig three meters into the ground, and I would find a treasure. I did it on the right date, dug, found a treasure, and with that treasure I built the whole villa. So he says to him: Wow, fantastic story. Do you have any proof of it? Documentation? There’s a villa here! What other proof do you need? That’s roughly what you’re saying, right? Okay. So that question now returns us to Yaakov, after David Enoch’s argument. David Enoch basically says that there are two explanatory tracks in the literature on this matter, or two possibilities that come up in the literature. One possibility is the probabilistic one, but as I said before, that doesn’t hold water. Meaning, even if the probability is the same probability, we can still prefer evidence of one type over evidence of another type. Therefore, he says, we need some other kind of explanation. So what is that other explanation? Either legal efficiency—meaning, I don’t know, maybe it leads to future problems or the system will suffer from it, and therefore we don’t… like the fruit of the poisonous tree, for example. If you bring me the result of an illegal wiretap, then I don’t accept the… In the United States at least, categorically I think, they don’t accept evidence of that kind. But that evidence is completely reliable. Listen to the recording; experts can check whether it matches the original speaker’s voice, and you can convict him—you know the facts. You still don’t accept it. Why? Because you know that then people will begin recording illegally, and that will project or create problems. That is exactly a consideration of legal efficiency, not of reliability. The reliability is fine; we know the truth, but we still won’t use the evidence. So he says: either considerations of legal efficiency or epistemic considerations, cognitive considerations. But the epistemic considerations in the sense of reliability—how much I really know that this is reality—in this context don’t work. Therefore he brings the explanation that the article I mentioned also gives; it appears in many, many places. The claim is that there is a difference between the two types of evidence in the following sense: if witnesses come and say that this is what happened, then I can say that I know that this is what happened. I have information that this is what happened. But if I rely on some probabilistic calculation, I cannot say, “I know that this is what happened”—some kind of explanation like that. Now, okay, I agree. After that everyone makes these gestures, like what—what does that even mean? So what? What difference does it make whether this is called knowledge or not called knowledge? In the end the question is what percentage I have of being right. What difference does it make whether it is called knowledge or not? Now, look: David Enoch then goes on to try to show—he wants to define what counts as knowledge. So he says that… or let me formulate it this way. He basically says that the problem with this explanation, that this is called knowledge and that is not called knowledge, is that it is not an explanation tied to any legal outcome, but rather an explanation in itself. I want the basis to be knowledge and not something else, and let’s say I managed to prove that this counts as knowledge and that does not count as knowledge. Quite apart from the question of probability, yes? It has no consequence, no legal consequence. Therefore he is unwilling to accept such an explanation. And then he proposes—he has a very nice move. Just explain that last sentence? What does it mean? He says: I’m not willing to accept this kind of explanation, that this is called knowledge and that is not called knowledge. If you show me a legal consequence, then I understand, because the probability is the same probability. So what do I care whether we call it “I know” or not “I know”? What is a legal consequence? A consequence like I said before—the fruit of the poisonous tree—some kind of utility, preventing damage or something like that. So he says the following. He makes a somewhat convoluted logical move. He basically wants to make the following claim: Let me… He says this to us: Let me show you that we make a distinction between knowledge and non-knowledge even in non-legal contexts. Meaning, I will show you that the consideration distinguishing knowledge from non-knowledge is not one that is the result of some legal consideration. Such a distinction exists in other contexts too. Then he gives an example, the following example: suppose a person buys a lottery ticket and the chance of winning is one in a million. Now the drawing has already taken place and he did not win, but he doesn’t know the results yet; still, he did not win. Can he say: I know that I didn’t win the lottery? He can say: I assume. He can say: most likely I didn’t win. He cannot say: I know that I didn’t win the lottery. A different story: he didn’t win the lottery, and a notice was published in the newspaper saying who won the lottery, and he sees that he didn’t win. Now of course there is some chance of a printing error in the newspaper, right? It happens sometimes. Okay, sometimes the errors are just typographical mistakes; often the errors are put in there deliberately. But there could be an error, right? So now, after reading in the newspaper that he didn’t win, can he say: I didn’t win—I know that I didn’t win? The intuition is yes. Okay, even though again, as far as the probability calculation goes, let us assume for the sake of discussion that it’s similar, and there is no problem constructing situations in which the probability calculation comes out similar. Why here do I say “I know,” and there not? Here this is already not a legal issue. It’s unrelated; there’s no legal matter at stake. It’s simply how I relate to the situation. You see that this distinction between a situation in which I can say “I know” and a situation in which I cannot say “I know,” even though the probability is the same, is a distinction that exists in other philosophical contexts too, not only in legal contexts. That means there is something here beyond legal utility. And now the question is: what is that thing? He wants to claim, in order not to give up on the legal route—meaning, he says that because in the legal world the explanation can only be one of those two I mentioned earlier, he proposes the following. He says: why really is the first thing called knowledge and the second not called knowledge? He calls it the sensitivity criterion. Let us suppose the reality had been different—hypothetically, a counterfactual, yes? Some… I’m positing a non-actual state of affairs, a situation that did not occur. Let us suppose the situation had been different. If the situation had been different. Say I had won the lottery. In the first case, would I think differently? In the case where I am the one in a million who won, but I don’t know the result of the lottery—would anything in what I say change? Nothing, right? The probability is the same probability, one in a million. Before too I thought I didn’t win, and now too. It doesn’t matter what actually happened. Meaning, what actually happened—what I think is not sensitive to what actually happened. What actually happened does not affect what I think, right? That is called something I cannot say I know. Even though the probability is one in a million, I still cannot say I know, because it has no connection to what happened. It’s a kind of a priori consideration, a probabilistic calculation. What happens in the second case? When a report is published in the newspaper. If I had won, then almost certainly the newspaper would have published that I had won. Right? Because the chance of error is one in a thousand, I don’t know, one in a hundred thousand, whatever it is. Usually what is written in the newspaper—about lottery winners, at least—is correct. Okay. Meaning, if I had indeed won—the counterfactual, if I had won—then the newspaper would have said that I won, and I would then have said that I had indeed won, even though the newspaper may have been mistaken. Right? Because the newspaper still could have been mistaken, but here I would indeed have said that I won. That means reality affects what I think, or rather what I think has some interaction with reality. Therefore here I can say that I know. I know—why? Because what happened in reality determines what I think. It’s not a matter of chance; it’s a matter of whether what I think is drawn from something I know about reality or from some general calculation that has nothing to do with what actually happened. Even though again, the chance of error is the same in both cases. But here I know, and there I do not. And now he says: okay, now we have defined the concept of knowledge. But the question still stands: the fact that we defined the difference still does not explain why here I would accept it as evidence and there I would not accept it as evidence, because ultimately what matters to me is being right in judgment. What do I care whether this is called knowledge and that not called knowledge? So far we’ve only defined what it means to know. Then he makes the following claim. He says: true, that itself really does not matter. But for some reason the sensitivity criterion affects both contexts. Meaning, when I ask myself whether I know whether I won the lottery or not, that’s philosophy. It has nothing to do with law and nothing to do with anything. It’s a question of how a person relates to such a situation. So in the first case I say: it is not correct to say that I know I didn’t win; and in the second case, it is correct to say that I know. Meaning, if the philosophical question of when to call something knowledge is indeed determined by the sensitivity criterion—and that’s true, legal scholars don’t care, it doesn’t matter—but for philosophy it is true. Okay. In the legal world too, the sensitivity criterion has an effect, but not because this is knowledge and that is not; rather, it is a utilitarian explanation. Even though it’s the same criterion, the sensitivity criterion. Therefore he says that when people—one second, I’ll explain the utilitarian explanation—but because the sensitivity criterion operates in both places, people get confused and think that in the legal world too what matters is whether it is knowledge or not knowledge. That is not correct. What matters is the sensitivity criterion. In the philosophical context, the sensitivity criterion defines what counts as knowledge and what does not count as knowledge. In the legal context, the sensitivity criterion says that this is the proper way to behave. Why? He says as follows. In a place where the outcome is not sensitive to the situation—understand?—for example, in the bus case or the prisoners case. The outcome is not sensitive, right? Whether the prisoner participated or did not participate, my judgment about him will be the same. One out of a hundred. I don’t know whether he participated or not. Whether the truth is that he participated or the truth is that he did not participate, my judgment will be the same. Sensitivity is not present there. Right? Same with the buses. Whether the truth is that it was a bus from the blue company or a bus from the red company makes no difference at all. I would say the same thing, because most buses are from the blue company. The sensitivity criterion is absent. Therefore people somehow say: ah, so apparently if it isn’t called knowledge, then you cannot judge on its basis. He says that’s wrong, nonsense. The sensitivity criterion matters because of legal utility. Why? He claims that if we judge in a place where there is no sensitivity, then the law will not achieve its purpose. You want, after all, to deter a person. Now you tell him: look, I’ll judge you this way or that way; what difference does it make whether you did it or didn’t do it? You’ll get the same verdict. So what is the point of punishment? You punish a person in order to create deterrence. But if no matter what he does he gets the same verdict, then the punishment will not deter him from doing it. Take the prisoner. The prisoner is deciding whether to participate in the riot or not. Okay, now he knows that ninety-nine percent participated. So when he stands trial, if we assume we convict, he will be convicted in any case. So the punishment will not deter him from participating in the riot. Consequently, because of that, he argues that it’s not because this is knowledge and that is not knowledge, but because the sensitivity criterion shows us that there is legal utility in this policy. One second. Because the sensitivity criterion operates in both of these fields—in the legal field it shows where there is utility and where there isn’t, and in the philosophical field it defines what counts as knowledge and what does not—people get confused and say: ah, the difference between knowledge and non-knowledge is what determines the legal criterion. But that’s not true. Instead of calling it knowledge and non-knowledge, call it sensitivity and non-sensitivity. That’s what he says. That’s what he says. Don’t call it knowledge and non-knowledge, because that’s not right. Sensitivity and non-sensitivity. Right, that’s exactly what he says. He says: you’re mistaken in calling it knowledge and non-knowledge. Knowledge and non-knowledge do not matter. But underlying the distinction between knowledge and non-knowledge is sensitivity, and that’s what you should be talking about—that’s what matters. But what about deterrence? Maybe deterrence isn’t the purpose of law. Maybe it’s justice. Okay. There are many, many objections one can raise here; I don’t accept this explanation. There’s another objection. I’m now among the hundred prisoners, and I see one guy sitting off to the side, so I’ll go and join the lynching. The legal utility depends very much on the situation. I didn’t understand. You said that if the punishment is that they kill everyone, then I have no incentive not to join the lynching. The ninety-nine, not the hundred. I am now one of the ninety-nine, and I see one person not participating. I’ll go and cut off his hand. I know that the law, as we understand it, does not convict—one non-participant is enough. This law encourages me to participate. Right. You’re arguing against the conclusion, not against the initial premise. Meaning, if we don’t accept such evidence, then true, granted. The legal utility is very context-dependent. There are many stories in which I can show that the legal utility goes the other way. More than that: in the bus case, unlike the prisoner case, the bus case is civil law, not criminal law. In civil law, when I sue because I want you to pay, it is not in order to deter. You have to pay because you caused damage. Meaning, restore the damage—what does that have to do with deterrence? And in civil law too they don’t accept this evidence. So why not? In short, I don’t accept this explanation. So the question is, all right, what is the explanation? There is simply a difference between what is needed for conviction and what needs testimony. Words. What does it mean “I need certainty”? What is testimony? The question is how one reaches certainty. Testimony is an active act. Statistics are something passive. Therefore, to convict I need someone… Why? Why? I understand—that’s the definition. But now I’m asking why. I’m not asking what the definition is; I know the definition. The question is why this is so. If it’s ninety-nine percent here and ninety-nine percent there, that’s what I need. I want there to be a true judgment—that’s what I want. What do I care about all these formal definitions? Look, I’ll offer you three explanations. There is a connection among them, but it’s not necessarily exactly the same thing. The first explanation depends on what we discussed in the previous lesson. Let me briefly remind you: I made a distinction between a majority that is before us and a majority that is not before us. The Talmud makes that distinction. And there we asked, regarding the position of the Chinukh, and Rabbi Shimon Shkop asked about the Chinukh’s position, why the majority in a religious court is considered a majority that is before us. The Talmud says it is a majority that is before us. But according to the Chinukh’s explanation, why do I rely on the majority in a court? Because in most cases that occurred, when there was a dispute of majority versus minority, the majority was right. So the majority that the case before me belongs to is the majority of all adjudications that took place in all places and times. It is not a judge from among a set of judges present before me, and therefore it should be a majority that is not before us. That is Rabbi Shimon Shkop’s question. There I explained as follows. I wanted to claim, and I said that maybe this is what he means, because he says that a majority that is before us is not statistics. And everyone says that’s nonsense—what do you mean? It’s the opposite; it’s even more statistics. And I tried to explain why he is right, why this is not based on… he doesn’t really say this, but I really agree with his bottom line. It is not statistics. And I explained that if, say, I find a piece of meat in the street and there are nine kosher shops and one non-kosher one, there is no way to perform a controlled experiment that would show the chance that this happened. So in fact, when I decide that the chance is ninety percent, that is not a result of statistics at all; it is the result of an a priori reasoning. Not the result of measurement or of an experiment, a generalization based on a sample as we do in science. Rather, I say: there are ten possibilities, nine of them are like this and one is like that. In the absence of any other information, the assumption is this. And the assumption is that it probably came from a kosher shop—ninety percent. But the phrase “ninety percent” is misleading. It has no connection to ninety percent. No one can show me that it really is ninety percent. Nobody can show me that. It is not really ninety percent. The assumption that it came from the majority shops is an a priori assumption; it is not the result of a calculation. So it’s not only that the ninety percent is uncertain—it simply is not ninety percent. Rather, in the absence of information, I assume it came from the majority shops and not from the minority shops. That is an a priori reasoning. And I think that this is the story in the legal context as well. When I see one hundred prisoners, ninety-nine of whom attack a guard and one of whom does not, and I say there is a one percent chance that this prisoner is innocent and a ninety-nine percent chance that he is guilty—that is nonsense. It is nonsense. There is no random process here taking place within some distribution that tells me this is one case out of a hundred. It is simply not true. What, did somebody randomly sample the prisoners? Exactly. Did someone check the prisoners in some random way? If you chose one prisoner at random from the hundred, then the chance that a particular prisoner would be selected is one in a hundred—that I’m willing to accept. If you choose randomly, if you know the selection is truly random—and even that is a question, how do you know? But here that is not the case. Some prisoners attacked, some didn’t; each one chose what he chose. One didn’t attack. Now what? There are no distributions here, no random process, nothing. The chance is not one in a hundred; that’s simply a mistake. Rather, in the absence of information, if I have no other information, then if you asked me what I would bet on, I would bet ninety-nine to one that he did participate. I don’t know—if I am forced to bet, on the one hand, and on the other hand I have no information. Right. But that does not mean there is really a ninety-nine percent chance here. That is simply not true. And therefore my claim is that the difference between the witness case regarding the riot and the statistical case regarding the riot is that in the witness case, it really is a statistical determination that they are mistaken three percent of the time. Because I can conduct experiments and see whether people see well or not. I simply ask them what they saw, check whether what they saw was correct, and I’ll find, say, that ninety-eight percent of the time they saw correctly and two percent of the time they did not. So on the basis of scientific testing I can generalize and conclude that they see correctly ninety-eight percent of the time, like a law of nature. Then that thing really is ninety-eight percent. The ninety-eight percent in that case is the result of knowledge. The ninety-eight percent in the case of… the ninety-nine percent in the case of the prisoner who didn’t participate—the statistic—there I have no knowledge at all. The ninety-nine percent in the case of the prisoner who didn’t participate—the statistic—is not knowledge at all that I have ninety-nine percent; it is the result of lack of knowledge. I have no knowledge, not “knowledge” in the sense I said earlier, as in “I know,” but in the sense of lack of information. For example, if I toss a die here and you know nothing about it. You don’t know whether it’s fair or not. With a fair die, the chance on each face is one-sixth, right? But if there is a die here and you have no idea what it is—maybe it’s fair, maybe not—and I ask you what the chance is that it lands on two, one-sixth. You would say one-sixth, right? I would say one-sixth too. Why? Because when I have no information, if the die is fair—I know it’s fair—then one-sixth is a probability based on a calculation. Meaning, I make a calculation and conclude that the probability is one-sixth, and that’s perfectly fine. Not certain that it will happen, but I know that that’s the probability. But if I know nothing about the die, then I’ll still assume the probability is one-sixth, but that’s just an a priori assumption. It’s an assumption that stems from ignorance. From lack of knowledge. Not ignorance in a pejorative sense. I have no knowledge. Do a test on most dice. I can do a test, but let’s say I have no way of doing a test. The die was here and they threw it into the sea. Now I ask what it was. Fine, it doesn’t matter. But you can do a test on most dice. What? How many of the dice. Fine, then again that will be a majority that is not before us. The question is how you chose this die. But that won’t help; it won’t solve the problem. Like how did you choose those witnesses? Fine, exactly. But you don’t know how you chose, and you don’t know how this process occurs, what the distribution is. There is no way to know. Now I ask: these witnesses before me—granted, ninety-eight percent of witnesses tell the truth—but I have no idea whether these witnesses are telling the truth or not. Here too I always assume—and this is what I discussed when I spoke about a majority that is not before us—I said that with a majority that is not before us we make a scientific generalization. Now a scientific generalization always assumes that the sample before me is representative. That’s always an assumption. You can’t know. There may be something special about this population, right? But that is an assumption within science. Meaning, we say: if I don’t see indication otherwise, then positively I say that this is a representative sample. Not that in the absence of information I have no choice. So I agree, here too there are some assumptions. These are the assumptions we use also in scientific contexts, and they probably work. But here it is simply absence of information. The truth is that it is not ninety-nine percent. It’s not that it is ninety-nine percent but you don’t know how to substantiate that claim. No—I’m not even saying it’s ninety-nine percent. It is not true that it is ninety-nine percent. You know nothing about it. But in the absence of information, if I have to bet, then I’ll bet on ninety-nine against one. But that’s just—it’s not… it’s lack of knowledge. Is that in the nine-tenths case? Yes. But didn’t you say there was a camera that showed that one person sat on the side? Doesn’t that count as knowledge or something? No, that counts as knowledge that one person sat on the side and ninety-nine participated. But the question is whether the one standing before me is that one or not. And the question I am asking is about the specific prisoner before me. In this case I claim that this difference is the first kind of difference, the simple one—a probabilistic difference. It is a probabilistic difference. I cannot convict on the basis of statistical evidence. Why? Because what they call statistical evidence means evidence based on lack of information. If you want to make assumptions and gamble on dice, do what you want. You cannot send a person to prison on the basis of that. You know nothing about him. You don’t even know that there is a ninety-nine percent chance he is guilty; you don’t know even that. If I had to bet and had no choice, I would say, okay, ninety-nine percent, I’ll bet on that. Just because I have no choice and I have to bet on something. I have no alternative. I don’t know what else to do. But here, in court, I am not gambling. And if I have no information, then I don’t act. Therefore I think that quite rightly we do not judge on the basis of such evidence. It is simply the distinction between a majority that is before us and one that is not before us. That’s all. The blue buses and red buses case is the same thing. Before me there is a distribution—ninety-nine percent, say, blue buses, one percent red buses—the same as the prisoners. You don’t know there are ninety-nine; you don’t know the distribution of buses that hit people. But if you have no information at all, then you say: well, if most buses are blue, I’ll bet ninety-nine to one. Exactly like the prisoners. You can’t convict on the basis of that. Not in civil law either, all the more so not in criminal law. You can’t, because there really is no ninety-nine percent here. That’s just a figure of speech. It’s not statistics. I didn’t understand. Regarding the meat shops, you said that if there had been cameras there, then you would say yes. Right. Here too I’d say the same thing. Here we have a camera. No, that’s the same question as before. No, I don’t have a camera telling me which prisoner is standing before me, where he came from. Do a controlled experiment. Do you know when this would become a majority that is not before us? If you performed an experiment on one hundred riots, okay? And checked somehow how you manage to isolate a prisoner in a certain way and verify that he really belongs ninety-nine percent to the majority population. These are just words; there is no way to do that. So you can say: the prisoner before me probably belongs to the majority population. The fact that ninety-nine prisoners participated and one did not—that is obvious, it’s a fact, I’m not disputing it. That’s clear. I’m only asking whether that means that the prisoner standing before me has a ninety-nine percent chance of belonging to that population. About that I don’t know how to perform an experiment. If someone manages to perform such an experiment, granted, I’ll accept it. That would be a majority that is not before us. So this is an interesting point, because at this point what we see is that, surprisingly, although everyone clearly thinks that this option certainly cannot be right—the probabilistic option—and that the answer must be the utilitarian option or halakhic decrees of Scripture of the kind I spoke about before—not so. In my view the explanation here is a probabilistic explanation. Only this probability is a little tricky. When the probability is based on lack of knowledge, you can’t rely on it. It is an assumption you make if you have no choice. But in court you have a choice: you can refrain from convicting. A majority that is before us. What? Yes, a majority that is before us, exactly. A majority that is before us is a probability based on knowledge. It is a scientific law. You are relying on knowledge. That is the knowledge we have. Therefore there you can… We’re returning a bit to the difference between something I know and something I don’t know, but of course from a completely different angle. It’s not a philosophical difference; it’s a probabilistic difference. It’s simply probability, that difference. And that’s an interesting point, because in both sources we use the expression ninety-nine percent, and still I claim that there is a probabilistic difference here and not something else, because the ninety-nine percent here is just a dummy number; it is not a real distribution. That’s the first explanation. The second explanation requires a short introduction. There is a passage in the Talmud in Bava Batra—in Bava Kamma, sorry—on page 46. The Talmud brings a dispute between Rav and Shmuel over whether in monetary law we follow the majority. It says there: it was stated, if one sells an ox to another and it turns out to be a goring ox, Rav said this is a mistaken sale, and Shmuel said he can say to him: I sold it to you for slaughter. Someone sold another person an ox, and it turns out the ox is a goring ox. I can’t use it for plowing. So I come to him and say: return the money, take back your ox, return the money—I bought under mistake. So the seller says: what? I sold it to you for slaughter. For slaughter an ox can gore as much as it wants; you slaughter it and that’s it. An ox for plowing—you can’t use a goring ox. But an ox for slaughter—it makes no difference. So the buyer says: no, I bought it for plowing, not for slaughter. So there is a dispute whether the transaction was for the purpose of plowing or slaughter. The Talmud’s conclusion—I’m skipping ahead—is: Rav said this is a mistaken sale, follow the majority, and most people buy for plowing. Meaning, Rav claims that we follow the majority, and most people buy for plowing, not for slaughter. That’s the reality. And Shmuel says he can tell him: I sold it to you for slaughter, and we do not follow the majority. For we follow the majority in matters of prohibition, but in monetary matters we do not follow the majority; rather, the burden of proof rests on the claimant. Shmuel says we do not follow the majority in monetary cases. Therefore I don’t care that the majority buy for plowing. So the principle—and that is how we rule—is that it belongs under the burden of proof on the claimant. There you must prove that he sold it for slaughter and not… Yes. Now the seller is holding the money. Yes, the seller has the money, so the buyer is the one seeking to extract from him, to have the money returned. So the claimant must prove what the seller meant, what was clarified at the time of the transaction, what the seller intended in the question more… what happened between them in the deal. Was it clear in the deal that this was an ox for slaughter? For slaughter. So the claim, basically, is that in monetary law we do not follow the majority—a dispute of Rav and Shmuel in monetary law. We rule like Shmuel. The Jewish law is like Shmuel in monetary matters and like Rav in matters of prohibition. So basically we do not follow the majority in monetary law. But there are other passages where we see that we do follow the majority in monetary law. For example, in a religious court: when there is a majority among the judges, two against one, we follow the majority there too, even in monetary cases, and derive it from “follow the majority.” And we have already seen that following the majority is a principle learned from the verse “follow the majority.” But the source that teaches you to follow the majority is dealing with monetary law. So how can you say that in monetary law we do not follow the majority? That is what Tosafot asks, and so do other medieval authorities (Rishonim). Tosafot in Sanhedrin gives one answer; Tosafot in Bava Kamma page 27 and Sanhedrin 3 gives another answer. Tosafot in Bava Kamma says that the majority in a religious court is a special kind of majority—it doesn’t matter for our purposes—but the principle is that indeed in monetary law we do not follow the majority. Only in a religious court do we follow the majority because it is a special principle; the minority is considered as if it isn’t there; there is some other issue there. Rabbi Chaim has a famous discussion of this. But Tosafot in Sanhedrin says the opposite. He says that Shmuel too agrees that we follow the majority in monetary law, but the majority for plowing is a special kind of majority; it is a weak majority, and there Shmuel says that we do not follow the majority. But Tosafot doesn’t explain. He says: why is that a weak majority? So Rabbi Shimon Shkop explains this based on Nachmanides. There is a Nachmanides in Milchamot in Kiddushin, and there the case is of a man who betrothed a woman with less than a perutah, and afterward he sent her gifts. Or not afterward—we see that he sent her gifts. The question is whether those gifts were intended to fix the betrothal and make up a perutah for betrothal, in which case she is betrothed, or not. The Talmud there makes it depend on the question whether in that place they customarily send gifts from groom to bride before betrothal. If they customarily send them before the betrothal, then these are not betrothal gifts—meaning, these are gifts not intended to effect betrothal. If they send them after betrothal, then there is concern—not concern, rather the thought—that he intended through those gifts to betroth her, to repair the deficient betrothal, and therefore it is valid betrothal. There is a dispute among the medieval authorities (Rishonim) how to understand the Talmud there; not so important for our purposes. But the Talmud there says that we follow the majority. There Nachmanides asks: why are we not concerned for the minority? He says there—presumption, not important—but what matters is Nachmanides’ reasoning. Nachmanides claims that in such situations we do not follow the majority. Why? Let me read you Rabbi Shimon Shkop’s wording. What emerges from his words is that by Torah law such a majority helps even in a matter as serious as a married woman; only the Sages were stringent. And according to what we wrote, the idea is that such a majority is a Torah majority and not an intellectual majority. It is effective by Torah law, only the Sages were stringent. Therefore they concluded that “people buy for plowing” is not an intellectual majority. If this person needed it for slaughter, that does not deviate at all from the laws of nature or common practice. Suppose a person wants to buy an ox for slaughter. A minority of people buy an ox for slaughter, right? But some do buy one. And there is nothing strange about a person buying an ox for slaughter. Some people do buy an ox for slaughter. A person comes and says: I bought—or sold, it doesn’t matter—an ox for slaughter. So what’s the problem? Is there anything problematic about that? The fact that most people don’t do this—so what? In that sense the majority carries no weight at all. What difference does it make that most people don’t do it? I’m telling you that I belong to the minority. Just like if I said: I’m one meter ninety-five tall. Okay? And they tell me: that can’t be, most people are not one meter ninety-five tall. So what? I am. Some people are that height. What does that have to do with the majority? This is a bit related to what I said earlier, if you notice. But the point is that there are really two possible ways to understand this. In principle I’ll explain it, at least for now, in one way. Think of a case where a person receives a box full of balls, and in this box there are ninety-five blue balls and five red balls. Now he sticks in his hand, lifts one ball with his eyes closed, and I ask: what is the chance he picked a blue ball? And it’s well mixed, okay? So what is the chance? People say ninety-five percent, right? Now a person with his eyes open, okay, reaches in and picks a ball. What is the chance he picked a blue ball? Fifty—no, I would say fifty percent if that. Why? The question is whether he likes blue balls or red balls. If he likes blue balls he’ll pick a blue one; he’ll look until he finds a blue one. If he likes red balls he’ll look for a red ball, right? Now again, I don’t know the distribution in the population of people who like blue balls and red balls—is it fifty-fifty? But that’s what will determine it, not the number of balls in the container, right? This means that in actions a person performs out of decision, out of his own preference, out of his own choice, there is no point bringing in statistics about the circumstances. What does that have to do with it? If I decided to send gifts to a woman—true, in this place people usually do that only after the betrothal—but I felt like doing it earlier. So you tell me: no, no, that can’t be, because most people do it afterward. I like doing it before. What does the question of what most people do have to do with this? Same thing here: I bought an ox for slaughter. Most people buy an ox for plowing. Sorry—for slaughter? Most people buy an ox for plowing. Okay, so what? Here it’s actually the seller, but it doesn’t matter. The seller says: you bought an ox for slaughter because that is what was agreed between us, and there is no reason to say that because most people buy for plowing, this can’t be true. This is an action—it is somewhat related to what I said before, because it is not a random action that you can analyze by distribution. It is an action in which you decide what to do. And when you decide what to do, it depends on what you want to do. Wait, didn’t you ask him what he wanted to do? The dispute is about what was between them when they made the transaction. What do you mean, ask him now? You ask him now. As you said—why decide by majority or by mistake? If he isn’t here, then you need to decide. No, but they have a dispute. The seller says: I sold it for slaughter, and the buyer says: I bought it for plowing. And these gifts—whether she is betrothed or not betrothed. So if he died… No, but maybe he is biased, maybe he went abroad, it doesn’t matter right now, or he has an interest and we can’t accept his account now. And we want to judge what happened there. It may well be that the seller intended slaughter and the buyer intended plowing. No, the dispute is about what was agreed between them, not what they wanted. What they wanted doesn’t matter. The question is what they were supposed to assume. If you do not explicitly state your intention, then one has to assume something about what you wanted. And I claim that for assumptions about what you want, you cannot rely on the majority that most people buy for plowing, because I wanted it for slaughter. That’s a matter of will. This is related to that Mishnah about one who hires workers and doesn’t stipulate with them what… No, there the question is what the default is. Here too. Maybe if the majority says this is that… No, there it is a constitutive determination. It is not the question of what really was, but what the law is. The law is that it doesn’t matter what you intended. The custom of the locale determines what is correct. It’s not that the local custom is an indication of what you intended. No, but here too maybe if you did not spell it out explicitly, even if you thought it, you did not reveal the matter. But the Talmud doesn’t say that. The Talmud doesn’t say here everything follows local custom; it brings a principle of majority. And then one can discuss why. And then the question is why. Why not follow the majority? We do follow the majority. So in fact, understand, this is pretty closely connected to what I said earlier. Because basically I’m saying something very similar. I’m saying that once the thing depends on a person’s action, you can’t work with statistical tools. Again, if I had to bet whether this person bought for slaughter or for plowing, I would bet ninety-five against five, because most people buy for plowing and not for slaughter. That’s true. But when a person comes and says: look, that’s my preference, I want it for slaughter—what do you want from me? Then you can’t say to him: no, no, there’s a ninety-five percent chance not. There’s no such thing. This is an action he decides upon. But with witnesses too—isn’t it that we accept them because of what he thinks? Now let us return to our issue, whether one relies on witnesses or not. In a trial, it’s not that the religious court doesn’t believe that a person really thinks this. It’s not like what you were saying regarding slaughter, what he really intended. No, on the contrary—therefore with witnesses we do follow the majority. Since witnesses are usually right, we believe them. Exactly because of that. That is exactly the point. Witnesses are in fact a random process. It is random because these mistakes are mistakes that occur not under their control. The witnesses think they saw correctly. Right—they come in good faith, let us assume for the sake of discussion. We know that sometimes witnesses are mistaken. So in that case it really is statistics. You need to know what the chance is that these witnesses think they were right even though in truth they were not right. Three percent? No problem—on that I can rely. But when there is one prisoner who chooses whether to riot or not to riot and says: friends, I’m a moderate person, I don’t riot and I don’t attack guards. I’m sitting here for adultery, not murder. So I’m a moderate person—what do you want from me? Since that is so, you can’t tell him: but ninety-nine percent rioted. This is a matter of decision and taste and a person’s character. There is no connection here; the majority—you can’t decide by majority. So that is the second explanation. Okay, I’ll have to stop, so maybe we’ll complete this next time. But it is connected to the first explanation, with one difference. The explanation I just gave is one that could in principle also be true of a majority that is not before us. Right. Meaning, that’s why I say the logic is similar. I’m showing that in both cases we don’t use statistics. But it’s not exactly the same explanation. It’s a different explanation. Okay? The first explanation distinguishes between a majority that is before us and one that is not before us. The second explanation says that even with a majority that is not before us—it’s true that “most people buy for plowing” is a majority that is not before us—but I can still say, okay, but I belong to the minority. All right.