Doubt and Statistics – Lecture 5
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
Table of Contents
- Reasonableness versus probability and the space of events
- Rarity versus abnormality
- Doubt requires a reason, and Bertrand Russell
- The quorum example in Bnei Brak and presumptions in Jewish law
- “We burn and stone on the basis of presumptions” and the distinction between legal certainty and majority-based decision
- Rabbi, Rabbi Hiyya, and Rabbi Kook in Ein Aya
- One witness, splitting testimony, and Maimonides on lashes
- Rabbi Yehonatan Eybeschutz, “follow the majority,” and a kosher seal
- Modern evidence such as DNA and cameras
- Probability depends on information, and conditional probability
- The case of Sir Roy Meadow and Munchausen syndrome by proxy
- Rarity, comparison base, and statistical tests
- Representativeness errors, medical tests, and misunderstanding conditional probability
- Reasonableness versus probability, and continuation of the questions
Summary
General Overview
The lecture sharpens the distinction between reasonableness and probability, and shows how statistics and decision rules operate only after a situation has been defined as one of doubt. The claim is that not every lack of information creates doubt, and that sometimes there is no readiness to define a situation as doubtful at all; instead, possibilities that seem unreasonable are ignored unless there is some concrete indication. From there, the lecture presents basic principles of probability as dependent on information, conditional probability and its asymmetry, together with legal and medical examples that show common mistakes in statistical inference and in the use of evidence.
Reasonableness versus probability and the space of events
Probability is defined as a quantitative measure of reasonableness, and when there is no such measure, it is more correct to speak about reasonableness rather than probability. Quantifying reasonableness is possible when there is a defined space of events and a distribution that assigns each outcome a chance, as in rolling a fair die where each result has a one-sixth chance. Even with an unfair die, probability tools can still be used if the distribution is known; what is needed is not a uniform distribution but knowledge of the event space and the chances.
Rarity versus abnormality
If you roll a die one hundred times, every sequence of results is extremely rare in the sense that its probability is tiny, so it is expected in advance that a rare result will occur. Abnormality is different from rarity: a sequence like one hundred sixes is perceived as abnormal, whereas another random-looking sequence is just as rare but not abnormal. The conclusion is that rarity by itself teaches us nothing; only marked abnormality may carry significance.
Doubt requires a reason, and Bertrand Russell
Bertrand Russell argues that a statement like “God appeared to me and told me to do this” does not create doubt for him; he gives it essentially no weight, not a fifty-fifty status. Russell illustrates this with the “celestial teapot” orbiting Jupiter, arguing that not every claim that cannot be disproved because of lack of information must be treated as a real doubt. The claim is that in order to be in doubt, you need a reason, and that an initial judgment of reasonableness can remove possibilities from consideration even without evidentiary investigation.
The quorum example in Bnei Brak and presumptions in Jewish law
In the case of a boy who says he is thirteen, the natural response is not to be doubtful as long as there is no indication that he is lying, because doubt requires a concrete basis. Once a case is defined as doubtful, there are halakhic rules for dealing with it, such as majority, presumptions, and decision rules; but before that, the very definition of the case as doubtful requires justification. The claim is that this is a philosophical or legal distinction, not a probabilistic one, because the use of statistics assumes you have already entered a state of doubt.
“We burn and stone on the basis of presumptions” and the distinction between legal certainty and majority-based decision
The Talmud in Kiddushin states that “we burn and stone on the basis of presumptions,” meaning that in capital cases one may act on the basis of a presumption such as “most sexual relations are with the husband,” without treating the matter as a doubt resolved by majority. The argument is that if there were a real concrete doubt that was merely decided by majority, we would not execute on the basis of a decision that still leaves a minority possibility the other way; but when there is no reason to doubt, the case is defined as legally non-doubtful and treated as certain. The distinction is between a case decided by decision tools within doubt, and a case that from the outset is not recognized as doubtful despite some theoretical alternative possibility.
Rabbi, Rabbi Hiyya, and Rabbi Kook in Ein Aya
The Talmud in Shabbat 30 brings the contradiction in Proverbs between “Do not answer a fool according to his folly” and “Answer a fool according to his folly,” and resolves it by distinguishing between matters of Torah and ordinary worldly matters. Examples are brought in which a man says to Rabbi, “Your wife is my wife and your children are my children,” and to Rabbi Hiyya, “Your mother is my wife and you are my son,” and the response is, “Would you like to drink a cup of wine? Drink and burst,” together with Rabbi’s prayer, “May it be Your will, Lord our God, that You save me today from the brazen and from brazenness.” Rabbi Kook in Ein Aya explains that these are situations in which there is no reason to doubt, and therefore one ignores malicious talk and does not enter evidentiary inquiry unless a real doubt has arisen that requires discussion.
One witness, splitting testimony, and Maimonides on lashes
The claim is raised that in these cases one could have dealt with testimony and splitting testimony, but the response is that in a situation that is not defined as doubtful, one does not enter an evidentiary framework at all. An example is brought from Maimonides: one witness is believed in matters of prohibition, so if one witness testifies that the meat is pork, and afterward two witnesses testify that a person ate it after warning, he can receive lashes, because once the testimony of the one witness has been accepted, a halakhic fact of prohibition is established and no active state of doubt remains about the identity of the food. The claim is that discussion of evidentiary strength belongs to managing doubt, but there are cases in which, after a certain credibility mechanism takes effect, there is no active doubt.
Rabbi Yehonatan Eybeschutz, “follow the majority,” and a kosher seal
A story is told about a priest who argued for Christianity on the basis of “follow the majority,” and Rabbi Yehonatan Eybeschutz replied: “Where I have doubt, I follow the majority; but where I have no doubt, I do not follow the majority.” The example is given of meat in a city where most shops are non-kosher: meat found in the street is judged by the majority, but packaged meat bearing a high-standard kosher seal is not treated as a case of majority at all, but as a case without doubt, so there is no need to turn to the statistics of the shops. The claim is that the decision whether a situation counts as doubtful is not always a formal halakhic rule, but a prior determination that defines whether the laws of evidence are activated at all.
Modern evidence such as DNA and cameras
A possible route is suggested for explaining the use of modern evidence in Jewish law: instead of seeing it as a tool for deciding within doubt, which would require two witnesses, one can argue that it removes the case from the category of doubt altogether, so there is no need to enter the evidentiary framework of witnesses. The claim is that the question is not only “what is the status of this evidence in Jewish law,” but whether the case is defined as a doubtful situation requiring the laws of evidence at all.
Probability depends on information, and conditional probability
Probability depends on the information available, so new information changes the distribution function even if the experiment itself has not changed. In rolling a fair die, the chance of getting a two is one-sixth, but given the information that the result is even, the chance of a two is one-third; this is a formulation of conditional probability. It is shown that conditional probability is not symmetric, so P(A|B) is not equal to P(B|A), just as the chance of an even result given that it is two is one.
The case of Sir Roy Meadow and Munchausen syndrome by proxy
Munchausen syndrome and Munchausen syndrome by proxy are discussed, together with a case in England in which a mother whose two children died of crib death was accused of murder. Sir Roy Meadow argued that the probability of crib death is one in eight thousand, so two such deaths are one in sixty-four million, and from that he concluded that she murdered them; the woman was imprisoned. The argument is that multiplying probabilities requires independence, and there may be family or genetic dependence that invalidates the assumption of independence.
Rarity, comparison base, and statistical tests
The claim is made that even if an event is very rare, it may still be expected in a large population, and therefore rarity is not equivalent to guilt or abnormality, similar to winning the lottery, where each winner is rare but the existence of a winner is not abnormal. The need to compare alternatives is emphasized, because a mother murdering two children is also a rare event, and therefore conviction cannot rest only on the fact that the alternative of crib death seems rare. It is explained that the statistical significance of a test must be fitted to the rarity of the phenomenon, through the metaphor of the net and the fish, where the quality of the test must have sufficiently high resolution relative to the frequency of the phenomenon.
Representativeness errors, medical tests, and misunderstanding conditional probability
A case is presented of a rare disease in which a test that is 99% accurate gives a positive result, but given a prevalence of one in ten thousand, the chance of actually being sick after a positive result is only about one percent because of the large number of false positives. It is explained that the figure “99% reliable” is P(test positive | sick), whereas the relevant question is P(sick | test positive), and switching the directions creates a serious error. The claim is that practical diagnosis is done through “preliminary filters” such as symptoms or proximity to a crime scene, which raise the base probability before applying a test or a piece of evidence at the 99% level.
Reasonableness versus probability and continuation of the questions
It is clarified that the examples are given in terms of probability because there one can calculate, but the principle also applies to intuitive reasonableness when there is no quantitative measure. It is said that legal and medical cases are especially confusing because doctors and lawyers are usually not skilled in statistics. In conclusion, questions are raised about the responsibility of human judgment and the effect of qualitative information on statistical assessment, and a brief discussion is brought about providence and the laws of nature as a continuation that is not part of the core topic of doubt and probability.
Full Transcript
[Rabbi Michael Abraham] We’re in the topic of doubt and statistics. In the previous lecture, after I talked about all kinds of situations of doubt, I started speaking a bit about probability. And today I want to continue a little further, where the first distinction I made was the distinction between reasonableness and probability. Meaning, probability is when I have some quantitative measure of reasonableness. And if I don’t have such a quantitative measure, then I can say more reasonable, less reasonable, intuitively it seems like this to me, intuitively it seems otherwise to me, but it’s not correct to speak in terms of probability; you have to speak in terms of reasonableness. And how do I create quantitative measures for reasonableness? If I have some event space—what can happen, what outcomes are possible, what is the chance of each outcome—for example, like rolling a die, then I have six outcomes, and if the die is fair then each outcome has a one-sixth chance, so I can quantify levels of reasonableness. What is the probability of getting an even number? It’s one-half, because I have three outcomes out of six. What is the probability of getting a number less than or equal to two? The answer is one-third, because it can be either one or two, which is two outcomes out of six, so that’s one-third, and so on. In other words, once I have a full map of the event space—what can happen—and the distribution, meaning the chance of each event in that space, then I can also quantify the level of reasonableness, and then I’m talking about probability. I said that suppose I have an unfair die, meaning the chance of each face is not equal; that still doesn’t mean I can’t deal with it using the tools of probability. You can deal with that too if you know the chance of each face. In other words, I need to know the distribution, but the distribution doesn’t have to be uniform, and the die doesn’t have to be fair. I just need to know the event space and the distribution. Another thing I spoke about was the relation between rarity and abnormality, and I said that suppose we roll a die a hundred times, then every sequence of one hundred results that we get has a very, very small probability. Six to the minus one hundred. But if I got the result six, six, six, six, six, a hundred times, which has exactly the same probability in the sense of rarity—it’s rare, and some other result is also rare—but in terms of abnormality there’s a difference. In other words, one hundred sixes is abnormal, while some other sequence of one hundred results is indeed rare but not abnormal. And the point is that if I don’t have an abnormal result, meaning the result isn’t marked out as something unique, then the fact that I got a rare result doesn’t mean anything. In rolling a die a hundred times, I can tell you in advance that a rare result will occur, because every result that occurs will be rare. A very, very small number—the probability of each possible sequence is very, very small. What is special is if an abnormal result is obtained, not if a rare result is obtained. And therefore that too is an important point. Okay, now I want to continue with this and maybe complete one more important point that touches on situations of doubt; after that I’ll continue a bit more with probability. But first of all I want to complete another point that I missed regarding situations of doubt. Bertrand Russell, a well-known twentieth-century philosopher, mathematician and philosopher, an atheist, said that if someone comes to him and says that God appeared to him and told him that one must do such-and-such or that one must not do such-and-such, then he wouldn’t relate to such a statement. Although seemingly you can’t know whether God really said it or not, so there are two possibilities. Assuming I have no information at all, then I give equal weight to the two possibilities, so the probability is fifty percent. Bertrand Russell says, what are you talking about? As far as I’m concerned the probability is negligible, I don’t relate to it at all. Why? Because—for me, let’s put it this way—not every situation of lack of information is defined by us as a situation of doubt. And the example he brought is the celestial teapot. He says, suppose someone comes to me and says, listen, there’s a little teapot orbiting the planet Jupiter. What, why don’t I see it? Well, you can’t see it—it’s small. Fine, so I have no way to know whether yes or no. So what do I do with such a statement? Fifty-fifty, right? Either there is or there isn’t; I don’t know, I have no information. Not at all. Obviously I’m not going to relate at all to the possibility that there is such a teapot, even though I have no information. Why? Because I assume that my interlocutor also has no information. Where could he possibly have gotten such a strange idea? I can invent millions of bizarre inventions and suggest them, and nobody will be able to refute them, but that doesn’t necessarily mean we have to take all those possibilities seriously. There is some kind of initial judgment of reasonableness that can remove possibilities from the field even without my having information about them. And therefore when Bertrand Russell has someone come to him and say, look, God appeared and said such-and-such, and from his point of view as an atheist there is no reason to assume that such an event could happen, then from his perspective he won’t see it as fifty-fifty. He’ll see it as some esoteric statement and wave the person away. That’s the celestial teapot—that’s the example, Bertrand Russell’s well-known example, and what it basically says is that in order to be in doubt you need a reason. For example, how do we know that—say, I was once in Bnei Brak, we went, I arrived at an evening prayer quorum, there were nine of us and there was another boy there. They asked him whether he was thirteen and he said yes. So someone there said, wait, but maybe he isn’t credible? Meaning, how can we believe him that he’s thirteen? What evidence is there? Does he have proof that he’s thirteen? So therefore there’s no quorum; we have to wait for a tenth. People waved him off, of course, and moved on to “And He, being compassionate.” What—why? What’s going on there? Because true, I don’t have clear information whether that boy is thirteen or not. And if the person says he’s thirteen, I have no reason to doubt it. Why assume that a person isn’t telling the truth? Now what do you mean, why assume? There are people who don’t tell the truth. True, but if I have no specific indication in this specific case that the person isn’t telling the truth, then I’m not in doubt. In order to be in doubt, you need a reason. Once I’m already in doubt, then there are rules for how I deal with situations of doubt. There is majority, and decisions, and decision rules, and all kinds of things like that, presumptions, all kinds of things of that sort. But all that comes after I’m already in a state of doubt. In order to define a situation as a situation of doubt, you need a reason. Not every situation in which I lack information is a doubtful situation.
[Speaker B] That’s basically the point, and Rabbi, Rabbi, Rabbi—can this seemingly metaphysical statement be turned into some kind of rule, that you need something in order to enter a state of doubt, so as to avoid this metaphysical statement? It should also be noted that here too it’s really all a matter of statistics. After all, when Bertrand Russell says that the statement that there’s a celestial teapot on Jupiter is an esoteric statement, that’s because really, what are the chances that an ordinary person on Earth would come, when he’s never been there and has no basis, and tell me such a thing, and then we’d investigate and find out it’s true? The probability is virtually zero. But if a person came whom I know somehow went—he was some emissary on some spaceship, he traveled there for years and came back from there and investigated every corner and was devoted to researching Jupiter—and he says, I was there in a spaceship and I saw it.
[Rabbi Michael Abraham] How can you determine the probability of such a situation?
[Speaker B] I’m not sure I’m going to determine it, but on the face of it statistically it seems impossible, extremely low.
[Rabbi Michael Abraham] What’s the probability that a person like—everything is correct except for one word: statistical. It has nothing to do with statistics. You think it’s not reasonable, and that’s exactly what Bertrand Russell says. What Bertrand Russell is saying is that I’m not in doubt.
[Speaker B] There is some doubt, but the probability on the side that he’s right, that there is a teapot, versus not right, is a probability that is much closer to zero.
[Rabbi Michael Abraham] Again, again—“there is some doubt,” I think that too is incorrect. There is some chance that he’s right. You’re moving here between reasonableness and probability. That’s why I made that introduction. The claim is that such a claim is very, very unreasonable in my eyes, okay? And therefore I don’t relate to it. I also have no indication where the person could possibly be getting this from. The person standing in front of me has no access to such information or such a claim. And therefore I don’t relate to it. So that’s what Bertrand Russell says. But leave statistics out of it—it’s not statistical, there’s no way to calculate statistically the probability of that. It has nothing to do with statistics—that’s exactly the point. Rather, I don’t relate to it because it sounds unreasonable to me, that’s all. Meaning, there are situations that I’m not prepared to define at all as situations of doubt. And once I use statistics, then I’ve already defined the situation as one of doubt, and the statistics, or the majority, or the presumption, or all these tools—those are tools that help me conduct myself within a state of doubt, to make decisions within a state of doubt.
[Speaker B] Like if Moses our teacher would say “there is a teapot,” or Isaiah the prophet would say it—even if he just casually said to you, “Listen, I think there’s a teapot”—then on the face of it he’s apparently not essentially different from the other person who said it.
[Rabbi Michael Abraham] Why not different?
[Speaker B] Because I think that if Moses our teacher would tell you, “Listen, I came down from Mount Sinai, there is Torah and there is a teapot.” So? Would you think that puts you into doubt, or that it’s just some empty statement?
[Rabbi Michael Abraham] Not only into doubt—I would assume it is actually true.
[Speaker B] So why—what changed?
[Rabbi Michael Abraham] Exactly—that’s the point. If Moses our teacher received it from the Holy One, blessed be He, then that’s information I’m prepared to accept. What’s the problem?
[Speaker B] So again, it’s because of the reasonableness that it’s true.
[Rabbi Michael Abraham] Exactly. And that’s what Bertrand Russell is saying. But there are situations where you have no statistical calculation—it’s not statistics and it’s not probability. Rather, it just seems unreasonable to you on its face, illogical, and therefore I say: I’m not in doubt, period. Once I start being in doubt, then you start doing statistical calculations whether yes or no; you make decisions within a state of doubt. There are situations in which I’m not prepared to define them at all as situations of doubt. And therefore my answer is not that there is such-and-such a probability that it is so or not so. Once I speak in terms of probabilities, I’m already speaking about situations of doubt. There are situations of doubt, and probability tells me how large the doubt is, how to make decisions within the doubtful situation. I’m claiming that there are situations that I’m not prepared to treat at all as doubtful situations. Now understand, for example, suppose a person comes to me and says Elijah the Prophet appeared to me, and he told me that there’s such a teapot. So that means—if Elijah the Prophet really appeared to him—then that information could be true, only I don’t think Elijah the Prophet appeared to him; it seems to me he hallucinated it, or he’s lying, whatever. But that’s how it seems to me. So can’t the person come and say, “Look, to you it seems one way, but I’m telling you Elijah the Prophet appeared to me, so it’s fifty-fifty”? No. It’s not fifty-fifty because I’m not prepared to define this situation as a situation of doubt. So this is an important point: the differences are philosophical or legal, and we’ll talk about them later, but not probabilistic. But it’s a significant difference, a significant difference in the way we relate to things, and we’ll see examples of this later. For example, the Talmud says in Kiddushin that “we burn and stone on the basis of presumptions.” What does that mean? Suppose, “One who strikes his father or mother shall surely be put to death.” So a child struck his father and mother; now the question is whether we execute him. One could say no, because maybe he’s not their child. Who said those are his parents? There is a presumption that most sexual relations are with the husband. But who says that’s true, a presumption? Very nice—but on that basis you’re going to kill? The answer is yes: we stone and burn on the basis of presumptions. Even though it could be that he’s not his son at all—these things happen, yes; it happens that a child is a mamzer, or it happens that a child is adopted, or a child is I don’t know what—he isn’t the son of the one presumed to be his father. How can you kill when I have some doubt? Usually, when there is doubt, of course we won’t take a drastic action like executing someone. But here yes—why? Because on the philosophical level, and also the halakhic one, we do not treat this situation as a state of doubt. Meaning, if I had a case where an actual doubt had genuinely arisen regarding this child for some reason, and then I would say, yes but there is a presumption, there is a majority—most sexual relations are with the husband—and then I would decide the doubt by saying that the child is the son of his father, in such a case my claim is that they would not execute him if he struck his father, because this is a state of doubt that you resolved on the basis of a majority. Fine, majority is fine, but there’s also a minority. You’re not going to kill a person when there is also some minority possibility that perhaps he is not liable to death. But if I have no reason to doubt this child, and only the general fact that sometimes a child is not the son of the presumed father—but I have nothing concrete about this child—then in such a situation I claim that the decision is not a probabilistic decision, or a decision where I go and ask myself throughout the world how many children are not children of their presumed father. Very few. That’s not the consideration. If that were the consideration, then they would not execute, because there is a minority who are not like that. So what if most are like that? So what? The consideration is much stronger. The consideration is that once a child is presumed to be the son of this father, I have no reason at all to cast doubt on the matter. If I have no reason at all to cast doubt on the matter, then the situation is not doubtful. It’s not that this is a doubtful situation that the majority resolved; as far as I’m concerned this is a situation that is not doubtful at all. If it’s a situation that is not doubtful, then if this child struck his father, he is liable to death. I do not concern myself with the minority of children who are not children of their presumed father. Okay? So this is a significant distinction legally as well. We’ll see examples in general law too, not only in Jewish law. There is a difference between situations where I am not in doubt, in which case I treat it as though it is certain—even though it is not certain, but philosophically or legally it is considered certain—versus situations where I have some doubt and I resolve it according to the majority. So true, I resolved it, but there is a limit to the confidence I will place in such a resolution; I’m not going to execute on the basis of such a resolution. There is a nice example of this, which Rabbi Kook brings in Ein Aya on tractate Shabbat. I think we spoke about this Talmudic passage—on page 30. Look at the Talmud here. The Talmud says as follows: It is written—regarding the book of Ecclesiastes, wait—they wanted to hide away the book of Proverbs because its words contradict one another. Right? Here at the beginning of the page. Its words contradict one another, so why didn’t they hide it away? Various reasons. What now? The Talmud asks: And in what way do its words contradict one another? What contradictions are there in Proverbs? So the Talmud says like this: It is written, “Do not answer a fool according to his folly,” and it is written, “Answer a fool according to his folly.” So should one answer a fool according to his folly or not answer? The Talmud says: There is no difficulty. This refers to words of Torah, and that refers to worldly matters. Okay? It says that answering a fool according to his folly is in matters of Torah; not answering a fool according to his folly is in ordinary worldly matters. Now the Talmud brings an example: Like that case of a man who came before Rabbi and said to him: Your wife is my wife and your children are my children. He said to him: Would you like to drink a cup of wine? He drank and burst. Yes, a man comes to Rabbi and says, you know, I had relations with your wife, and the children you thought were yours are my children; they are mamzerim. So instead of Rabbi panicking and saying okay, there are laws of doubt, what do we do now, and this and that, he says: Would you like to drink a cup of wine? He offers the man a cup of wine; he drank and died. Again, he probably died only in a literary sense, but the point is that he waved him off down all the stairs. He didn’t relate to his claim. Another case immediately after that: There was a man who came before Rabbi Hiyya and said to him: Your mother is my wife and you are my son. Yes, a similar case. He says: I had relations with your mother and you are my mamzer son. He said to him: Would you like to drink a cup of wine? He drank and burst. Rabbi Hiyya said: Rabbi’s prayer helped me, that they not make my sons into mamzerim. For when Rabbi would pray, he would say: May it be Your will, Lord our God, that You save me today from the brazen and from brazenness. So that brazen man did not really succeed in harming Rabbi, thanks to his prayer. What does this mean for our purposes? So Rabbi Kook asks here the obvious question. Someone comes as a witness and says to you that you are a mamzer or that your children are mamzerim. Now you cannot know with certainty that they are not. How can you know? It’s not even something that is in your control. Maybe your wife committed adultery. Relying on her is all very nice, but you can’t know; things happen. So how do Rabbi, or Rabbi Hiyya, ignore testimony like this that comes before them? So Rabbi Kook says exactly the distinction I mentioned before. There are situations in which I do not see any reason at all to doubt. Once I do not see any reason to doubt, I am not interested in evidence, or inquiries, or all kinds of things of that sort. I simply ignore completely—completely—this slander, this brazenness that the Talmud describes. That is the “do not answer a fool according to his folly,” right? That is basically the example of why you don’t need to answer fools. The person who said what was said here is defined by the Talmud as a fool. Meaning why? Because why on earth would one assume such a thing?
[Speaker D] But isn’t it proper at least to ask one question: do you have any proof of this?
[Rabbi Michael Abraham] The Talmud doesn’t describe all the details to me. It may be that Rabbi did make some preliminary inquiry first, I don’t know. But as the Talmud presents it, it sounds like with inquiries or without inquiries—I don’t know what happened there—but after everything that happened there was over, Rabbi remained with the fact that there is one witness here and he ignores him. Fine, that’s fine. And that means that from his perspective this is just empty slander. Meaning, he has no reason to doubt, and therefore he says, leave me alone already, so to speak. In other words, I have no reason at all to doubt and I am not—I am not relating at all to this claim. And Rabbi Kook says that if a person grows up in a house and is presumed to be the child of those two parents, the mother and father there in that house, no one can come and say that he is a mamzer, or maybe he is a mamzer, or I don’t know what. No, that’s not—if he is presumed to be such, then that means he is such. You need to bring me a good reason that will arouse doubt before I’ll even begin discussing it and looking for evidence. Not every person who comes and says so-and-so is a mamzer will immediately make the religious court sit down and start inquiries and checking whether yes or no. Now what exactly can arouse doubt? Good question. I assume that if this person had come as an eyewitness, or I don’t know exactly, or something like that, and the religious court or the people thought that maybe there was something to what he said, then perhaps they would indeed be concerned and conduct an inquiry. So I really don’t know everything that happened there in that story and exactly how it was. All in all, this is one witness. One witness is not something that can just be ignored. It requires an oath; in some cases one witness can impose an oath. One witness is not a trivial thing. In matters of prohibition, one witness is believed altogether. So it’s not simple. But what the Talmud here is saying is that at least in certain situations—and again, I don’t know exactly how to define them, the Talmud doesn’t go into definitions—at least in certain situations, even if evidence comes before me, I can ignore it. I ignore it because I have absolutely no indication that it’s true, and therefore I’m not in doubt. If I were in a state of doubt, now I would have to make inquiries. At that point it no longer helps to say, okay, I’m in a state of doubt but it seems to me not. No—once the situation is defined as doubtful, there is no “it seems to me”; now it requires inquiry. But it may well be that I can define a situation as one that is not doubtful. It’s perhaps a bit similar to what Maimonides writes—just occurs to me now. If—after all, one witness is believed in matters of prohibition. So in order to say that a certain piece of meat is kosher or not kosher, one witness is enough. But in order to give a person lashes, you need two witnesses that he committed a transgression, you need two witnesses. Now what happens if one witness testifies that this meat is pork? Now I come and eat that meat with two witnesses seeing that I ate it, and they warned me, and everything. Am I liable to lashes? After all, everything I know—that this meat is pork—is based on one witness. And therefore the strength of the evidence that I am liable to lashes for eating pork is only as strong as the weakest link in the chain. Right? The knowledge that there is pork here is based on one witness. Now that level of knowledge is not enough to give a person lashes. The fact that I have two witnesses that he ate it and that they warned me and everything—that’s all true—but if it isn’t pork, then what help are two witnesses who testify that I ate it? In order to know that it is pork, I seemingly need two witnesses; otherwise you can’t give lashes to the person who ate it. Maimonides says no. If one witness testifies that it is pork, then now you can give him lashes. If there are two witnesses who saw him eat it, you can give him lashes. Why? Because one witness is believed in matters of prohibition. Yes, but one witness is believed in matters of prohibition, not with regard to lashes. Correct—but after the witness testified, the fact that this thing is pork—it’s not that now there is doubt and it was resolved by one witness. Once the one witness came and one witness is believed in matters of prohibition, there is a fact here that it is pork. I am no longer in doubt. Once the fact is that it is pork, if someone eats it in the presence of two witnesses, he is liable to lashes. We do not go back and reopen the question whether it was pork or not pork. That question has been decided. It is no longer a state of doubt. We’ll see—we’ll also see examples of this later. So basically what I want to claim here.
[Speaker E] But sorry, you’re saying that what you’re saying, and what Rabbi Kook is saying, is that they ignore this person who came to say that he is the husband of Rabbi Hiyya’s wife—Rabbi’s wife.
[Rabbi Michael Abraham] In that he makes himself wicked, we apply splitting testimony. That’s a Talmudic passage in Sanhedrin 9. “So-and-so had relations with me willingly.” “So-and-so had relations with me willingly”—meaning, so-and-so had relations with me by force, and they say it was by force.
[Speaker E] But here, how do you split it here?
[Rabbi Michael Abraham] What’s the problem? Someone else had relations with her, not me, but it still obligates.
[Speaker F] If we were dealing here with the laws of forbidden sexual relations, that would be exactly the point.
[Rabbi Michael Abraham] If there were doubt here and I needed evidence, then that evidence is good evidence—that’s splitting testimony. But the claim is, yes, in this situation I’m not defining it at all as a state of doubt. So I don’t enter into questions of testimony and splitting testimony—no, there isn’t, there is no doubt. This is not at all a state of doubt, so I don’t even begin a discussion about evidence. Okay? There is.
[Speaker E] So a decent person—a person presumed to be righteous—if he said this, we also wouldn’t relate to him in the same way? We wouldn’t relate to him?
[Rabbi Michael Abraham] A righteous person who said what?
[Speaker E] “I had relations with your wife and this son is your son,” and the whole story.
[Rabbi Michael Abraham] I can’t really tell you, because if you ask me in terms of credibility, then maybe I would apply splitting testimony, but he is still only one witness. One witness is not enough for this. The question is whether I believe him. It could be that I would need to separate from my wife, because in the end if I believe him that is enough to obligate me to separate from her—like in a column I wrote not long ago about the wife of a kohen who was raped. Yes? If the kohen believes her that she was raped—and usually there isn’t testimony about a rape—then how does it happen that the wife of a kohen who is raped separates from her husband? She comes and says, I was raped. But there aren’t two witnesses for that. Now if the kohen believes her, he has to separate from her. Okay, so in terms of the laws of evidence, this is not a simple question. The question is whether I enter here at all into an evidentiary discussion. So I don’t know how to answer. Meaning, if there were a trustworthy person here, righteous, I don’t know exactly what, who came and said this—then regarding lashes or death, I don’t think a religious court would give lashes or death on that basis; this is one witness. But regarding my obligation to separate from the woman because she committed adultery, that does exist. If I take him seriously and believe him, I have to separate from her. Okay, so it depends what issue we’re talking about. In any case, there is a nice story in this context, a well-known story about Rabbi Yehonatan Eybeschutz. All the stories about Rabbi Yehonatan Eybeschutz—so this one too. And the story says that a priest came to Rabbi Yehonatan and said to him: Why don’t you follow us Christians? After all, we are the majority. The Torah says, “follow the majority.” We are the majority, so the Jewish law should follow the majority. So Rabbi Yehonatan Eybeschutz said to him: Where I have doubt, I follow the majority, but where I have no doubt, I do not follow the majority. What does that mean? It sounds like a clever quip; usually it’s told as a joke, but the truth is that it’s an entirely serious answer. What does it mean? Suppose we find a piece of meat in the street, and the basic rule is that we follow the majority of the shops in that city. If most of the shops are kosher, the meat is considered kosher; if most of the shops are not kosher, then the meat is presumed non-kosher. Now let’s say that most of the shops in the city are not kosher. Okay? A city of gentiles with one Jewish shop, say, and nine gentile shops, okay, where the meat is not kosher. I find a piece of meat in the street. So according to the law, the meat is non-kosher, right? Now what happens if I find a piece of meat that is packaged with a seal of high-standard kosher supervision? Now I ask myself whether this meat is kosher or not. There are nine non-kosher shops in the city; we should assume the meat is non-kosher, we follow the majority. Reasonable or not reasonable? Not reasonable. Why? Because there is a seal that says this meat is kosher. So what? Where in Jewish law do we find the law of a seal? What law is that? I know there is a law of majority. What is the law of the seal?
[Speaker E] That there is no doubt. Exactly.
[Rabbi Michael Abraham] Meaning, I follow the majority if I define the situation as one of doubt, and then I begin statistics, majority, I make decisions, presumptions of one kind or another. But if I find a piece of meat with a seal, I am not in doubt, so now I’m not going to start doing statistics about what meat shops there are in the city. Exactly. I follow the majority if I define the situation as one of doubt. And then I begin statistics, majority, I make decisions, presumptions of one kind or another. But if I find a piece of meat with a seal, I’m not in doubt. So now I’m not going to start doing statistics about what meat shops there are in the city. I will do that statistical analysis if I define this situation as a doubtful one. But if I found a piece of meat with a seal, it’s not because of some “seal law” in Jewish law. I don’t need Jewish law at all. I go looking for the laws of evidence in Jewish law after I’ve defined the situation as a doubtful one. But if I don’t define it as doubtful, I’m not going to ask a halakhic decisor or do statistics.
[Speaker G] Rabbi, once again, if the Rabbi really had a doubt…
[Rabbi Michael Abraham] Wait, wait, wait. If the piece of meat—if the definition, whether this situation is a doubtful situation or not—is a definition entrusted to me, then this is not about the rules of Jewish law. There are no rules of Jewish law about seals. Rather, if I see a seal, then I’m not in doubt. The moment it’s my mandate to decide whether I’m in doubt or not, then after I decide that I’m not in doubt, I don’t need the evidentiary rules of Jewish law. So here I don’t need to ask a halakhic decisor. It’s not that the decisor tells me, “It’s fine, you can eat.” I don’t go to a decisor at all. Because first I have to define whether this situation is doubtful. If it is, then I’ll go to a decisor who will guide me what to do. But if I don’t define the situation as doubtful, then no. That’s why the seal changes the situation—it’s not a law from the laws of evidence. Majority is a law from the laws of evidence. The seal says that the situation is not a doubtful situation at all, and therefore I don’t need the laws of evidence. Yes.
[Speaker G] Rabbi, what does it mean that the situation is not a doubtful situation? If in fact the Rabbi was in doubt… he was in doubt in the case of the seal because for some reason all the rabbis, all the kashrut agencies, suddenly become fraudulent kashrut agencies. Then you can say a thousand times that we’re not in doubt, but we are in doubt. So what is this statement, that there are situations where I’m not in doubt? If you—it doesn’t depend on you. It’s like what the Rabbi says about truth.
[Rabbi Michael Abraham] There are situations where I am in doubt, I know.
[Speaker G] No, but it’s not because there’s some metaphysical determination like that. The question is whether there really is doubt or not. In a case where there is kashrut certification, then what do you mean not doubtful? Initially, when I approach and examine the certification, I am in doubt. But when I examine it and say, wait a second, it says here that this is under the supervision of the Rabbinate of Rehovot, and that’s a serious rabbinate, then I say: given the doubt, the chance that this is false, that it’s forged, is very small, and therefore I decide that it’s kosher. It’s not because of some metaphysical determination.
[Rabbi Michael Abraham] No, no, no, not because of that. You present it as a decision based on majority, and I claim no. It’s not a decision. The moment I see a seal of the Rabbinate of Rehovot, as far as I’m concerned the meat is kosher; I have no question. I don’t go ask, do statistics, check. That’s exactly the difference. Now again, my claim is not a probabilistic claim. You can say that even in the Rabbinate of Rehovot there’s one percent forgery, or I don’t know exactly what, or mistakes. Obviously. But I don’t care, because I don’t enter into statistics if the situation is not defined for me as a doubtful situation. And that has nothing to do with metaphysics.
[Speaker G] But it doesn’t depend on you whether to define it as doubt. It’s not something that… reality is that reality is doubtful; it’s imposed on us, just as truth is imposed on us, just as…
[Speaker H] There isn’t here, there isn’t here any reality of doubt at all. You walk down the street, you see meat, you see on it a kosher stamp. Then you look around and say, but there are six more non-kosher stores here. But you found meat with a kosher stamp, you take it, go home, and eat. That’s all.
[Speaker G] And if you look at the stamp and see that it’s a stamp of Tiv Ta’am, the kosher supervision of some chain… that’s something else. So do you understand?
[Rabbi Michael Abraham] So in cases where there is doubt, what does that prove about cases where there is no doubt? Right. What does that have to do with it? Obviously.
[Speaker G] No, I’m saying, the statement that there is no doubt is an empty statement, because it doesn’t depend on us; it’s imposed on us. In the end it’s a matter of likelihood. Whether it’s likely or not is… the fact that we make some pious wish that we not be in doubt won’t change reality.
[Rabbi Michael Abraham] And again we come back to this point. It’s not metaphysics; it’s common sense. What you may mean to say—and with this I agree, I said it before too—is that it’s not a question of probability. There can be a situation where the probability is ninety-ten or ninety-five-five, but I am in doubt and resolve it with ninety-five against five. There can be a situation where the odds are still ninety-five versus five, but I don’t define that situation as a doubtful one. If that’s so, then I don’t need statistics. As far as I’m concerned, the situation is not doubtful even though it’s the same statistics, ninety-five versus five. I said that the difference is not a probabilistic difference; it’s a philosophical or legal difference, however you want to define it. But still, philosophically and legally, we see these situations as situations that are not situations of doubt. Even though there’s one percent that maybe it’s not true, or five percent that maybe it’s not true. It doesn’t matter. There’s also five percent that maybe this child is a mamzer. There are people, there are children who are mamzerim. But Rabbi Yehuda HaNasi says no—it doesn’t interest me, there is no indication, I do not relate to that possibility. And it exists; it is a real possibility. I’m not claiming that statistically it doesn’t exist at all in my event space, that the state of a mamzer doesn’t exist. It exists. But in a legal or philosophical decision, or whatever it may be, I determine that such a situation is not doubtful. And if I determine that this situation is not doubtful, then I don’t need to resort to the rules of Jewish law. By the way, this is one possibility, for example, for dealing with various kinds of modern evidence. Say DNA evidence in court, right? So in Jewish law it’s not entirely clear what the status of such evidence is, because for certain purposes at least you need two witnesses as evidence. And evidence of a very, very high level, but which is not two witnesses—there are those who will say we don’t accept it because it still isn’t two witnesses; there is a scriptural decree. So how, nonetheless, do decisors follow evidence like DNA evidence, or a camera, or things of that sort? One possible route is to say: look, the moment I saw this, I’m not in doubt. If I’m not in doubt, I don’t need witnesses; it didn’t resolve my doubt. Witnesses resolve doubt, but here if I saw DNA or a camera or whatever it may be, I’m not in doubt. If I’m not in doubt, I don’t need to look for witnesses or evidence or whatever, and the definition of whether I am in doubt or not is entrusted to me; it doesn’t belong to Jewish law. Therefore there is nothing here to discuss in terms of what force DNA evidence has from a halakhic perspective, or to what extent Jewish law recognizes such evidence. Rather, I don’t enter the whole discussion of evidence at all, because I say: this is a non-doubtful situation.
[Speaker I] But this, may I ask a question? Yes. Is this parallel, like in the series of classes on faith, where the Rabbi defined that there are three states: absolute certainty, absolute skepticism, and the middle one, if I remember correctly, is defined as intuition?
[Rabbi Michael Abraham] Maybe. I don’t remember that I defined it that way, but yes, okay.
[Speaker I] So I’m saying, meaning, we also in a case, say, with the kashrut stamp, we don’t have certainty.
[Rabbi Michael Abraham] There is no state of absolute certainty. There simply is no such thing.
[Speaker I] Right, so it’s not that we need certainty, but on the other hand it’s not skepticism, so…
[Rabbi Michael Abraham] Correct, that’s what I’m saying. No, skepticism means fifty-fifty, but there are situations where I say it’s not fifty-fifty, it’s ninety-ten. Again, I’m talking about plausibility, not probability, so the numbers here are only for illustration. Okay? So it’s ninety-ten, but that same ninety-ten can sometimes be defined as doubt because I have a reason to doubt, and the ninety-ten resolves the doubt, because the majority of ninety overcomes the ten and tells me: true, you have a doubt, go after the ninety and not after the ten. That is a situation that is doubtful, only I resolved it באמצעות majority. By contrast, there are situations where, if I have no indication whatsoever to assume that a child is a mamzer, and let’s say that three percent of children in the world are mamzerim just for the sake of discussion, okay, but I have no indication about a particular child that he is a mamzer, I am not saying he is not a mamzer because there is a majority of non-mamzerim, but because I have no reason to doubt. Even if the rule were, say, that in mamzer status we would not follow the majority, here I still would follow the majority because for me this is not called following the majority; it’s a situation about which I am not in doubt. The moment there is positive evidence about this child—not certain evidence, because then he would be a mamzer, but evidence that raises doubt—now I ask myself, okay, but ninety-seven percent of people in the world are not mamzerim, so I’ll decide that this child is not a mamzer, but that is already the resolution of a doubtful situation through majority. But if I have no evidence that raises a specific doubt about this child, then I say he is not a mamzer not because there are ninety-seven percent who are not mamzerim, but because I have no reason to doubt. Now, of course, in the background sits this whole fact that ninety-seven percent are not mamzerim. I am not claiming that the statistics are different; the philosophical and legal status of the claim is different, not its force. Its statistical force is ninety-seven versus three, but the legal and philosophical situation is different. It’s a claim about which I am not in doubt, not a claim that was decided on the basis of majority. Okay?
[Speaker G] What is this—an example where we determine reality? Our consciousness basically determines our relation to reality, not the statistics of reality itself, which is a very surprising statement for a conservative approach.
[Speaker I] But it works the other way too. It also works the other way, because even when we treat a product as non-kosher or a person as a mamzer, you can still raise all kinds of very, very remote considerations or hypotheses as to why our identification is mistaken, and nevertheless we treat it as a given.
[Rabbi Michael Abraham] Okay, yes, we…
[Speaker I] So that’s exactly…
[Rabbi Michael Abraham] the claim. Yes,
[Speaker I] not only on the side of permitting, but also on the side of disqualifying.
[Rabbi Michael Abraham] They asked me on the website how I relate to conspiracy theories. Now when I thought about it, I told him I don’t understand the question and I don’t understand what I’m supposed to answer. But in truth, it may be that he meant something like what I’m saying now. Conspiracy theories are exactly like saying this child is a mamzer. Why is that a conspiracy theory? It’s a conspiracy theory because I have no real basis to doubt regarding this child. Now that doesn’t mean it’s absurd; a conspiracy theory can, after all, be true. There are some mamzer children in the world; it’s not that the phenomenon doesn’t exist, okay? But if I have no indication. Take Rabin’s assassination. Why is it a conspiracy theory that the Shin Bet planted the bullets? It could happen. There have been conspiracies in the world. Intelligence organizations sometimes do all kinds of deeds of this sort. Why, why is this defined as a conspiracy theory? Because we decide that we have no doubt regarding this case. Even though there is one percent, or I don’t know how much, that intelligence organizations do things like this. It can happen. In other words, I’m showing you that in life—our life, in philosophy, in law, but also in everyday life—there are situations where we are not willing even to consider the possibility, even though it cannot be ruled out. There are two percent that it exists. But I won’t relate to it if I have no positive indication about this specific case that there is room for concern regarding it. Okay? The same thing with a person’s presumption of fitness. What is a person’s presumption of fitness? A person’s presumption of fitness is not a presumption. Because if it were a presumption, then how can it be that we execute a person on the basis of two witnesses? Who says those witnesses are fit? They have a presumption of fitness. So what? Do you execute a person on the basis of a presumption? You can’t execute a person on the basis of a presumption. Presumption of fitness means he is definitely fit; for me this is not defined as a doubt. If there were a witness about whom doubt arose and I resolved the doubt on the basis of a presumption, I would not execute a person on his basis. Let me give you an example of such a case. Suppose a person lied in religious court. Two witnesses were found or they discredited him; he lied. Now there is testimony he gave a week earlier. So now I don’t know whether that testimony was true or not. Now he is a liar. From now on he can no longer testify. He is disqualified as a witness. Okay. Now I ask myself: did he lie there too? In such a situation we still hold him as fit until it becomes clear in the religious court that he lied, but here there is already a positive reason to suspect him. It’s not just some empty statement. There is a positive reason to suspect this specific witness. Okay? Therefore here, even if I define him as a fit witness regarding the earlier testimony, it matters that I validated him because of a presumption of fitness. So there it is already fitness on the basis of a presumption. It’s not that the witness is fit because I have no doubt about him. I do have doubt about him. The rule is that I do indeed have doubt about him, but we follow his presumption of fitness until we reach the moment when it is clarified that he lied. Before that moment he has the status of presumed fit. But this is presumed fit regarding a witness about whom there is already real doubt. So if, say, he had given testimony on the basis of which I would have had to execute a person, I would not execute the person. Because his testimony has the force of a presumption, not the force of a witness. I have doubt about him, and the presumption resolved the doubt. Okay? By contrast, if a witness comes and tells me that this person must be executed—he murdered—and I have no indication, he is a fully fit witness, still this is a presumption of fitness, and there are unfit witnesses in the world. How can you rely on him to execute a person? The answer is that if I have no indication that he is a liar, then he is a fit witness, period. And I have no doubt—not because of the presumption that most people are fit. Okay? It’s a subtle point, but you can’t deny it. Every legal system assumes it, and in our everyday life we assume it. Okay, so that’s regarding this supplement I said—that in order to doubt you need a reason. Not every situation of lack of information is defined by us as a situation of doubt. Another point. Now this is just a supplement regarding situations of doubt and when I start using statistics. Now I return to the issue of probability and statistics. So we talked about the fact that you need an event space and you need a distribution, a distribution function, that tells me what the probability is for each of the events. Another point that is important to know regarding probability calculations and probability functions is that probability depends on information. I use probability where I have lack of information, but what information I lack, or how much information I have, can change the probability function. For example, the probability—say I roll a die. Or no, let’s say I rolled a die yesterday, and now a person asks himself which face the die landed on. What is the probability that it landed on two? The answer is one-sixth, right? Because it’s a fair die. The answer is one-sixth. Good. If I tell him, listen, the die landed on an even result, and now the person asks himself what is the probability that it landed on two, the answer is one-third. Why? Nothing changed in terms of rolling the die; the die is fair, I’m asking the same question, yes? What is the probability it landed on two? So what did change between the two cases? The information he has. In the first case he had no information; in the absence of information the probability is one-sixth. In the second case he has additional information; the information is that the result was even. That can change the distribution. The moment I have additional information, the distribution function changes. Now there is a one-third chance that it is two, a one-third chance that it is four, and a one-third chance that it is six, right? Because I have only three even outcomes. Two is one of the three even outcomes. When I ask myself what is the probability that the result was two, the answer now is one-third. What changed? I performed the same experiment, I ask the same question, the difference is the information in my possession. The information in my possession can change the probabilities and the distribution. Therefore probability always depends on information. Prior information is important for determining probability. The formal way to do this is when we define conditional probability. What is conditional probability? Probability—when I ask what is the probability that the die landed on two, that is ordinary probability. The probability is one-sixth. But now I ask: what is the probability that the die landed on two given the fact that the result was even? Here the probability is one-third. Right? Two is one of three even results. I denote this as conditional probability: the probability that two came out on condition that it is known that the result was even. It is conditioned on a certain piece of information. Okay? So conditional probability, the probability of event A given information B, is called conditional probability, P of A given B, okay? Conditional probability is different from ordinary probability when I have—usually, not usually, almost—it is either greater than or equal to ordinary probability. Right? Because it can only be more, it can’t be less than one-sixth, the chance that two came out, unless every additional piece of information only increases the chance, okay? No—that is, actually that’s not precise. If I know that the result is odd, then it reduces the chance to zero, from one-sixth to zero. Okay, that’s true actually, so it’s not always like that. In any case, usually additional information gives me more indications, so the probabilities will generally be greater, okay? So that’s regarding conditional probability. Now notice a point, the next point. I said before: what is the probability that the result was two given that the result was even? The answer is one-third. Now I ask the opposite question: what is the probability that the result is even given the fact that two came out?
[Speaker C] What do you say?
[Speaker J] The same thing, no? Of course, of course.
[Rabbi Michael Abraham] One. The probability is one.
[Speaker J] Did you hear me?
[Rabbi Michael Abraham] Because if two came out, two is even. So if two came out and you ask me what is the probability that the result is even—the answer is one. Obviously it is even because I know that the result was two. Therefore conditional probability is not symmetric. P of A given B is not equal to P of B given A. This is a very important point because we very often mix these things up; very often we switch these directions. And let me give you an example that I already mentioned before: Munchausen syndrome by proxy. Yes, a controversial psychiatric syndrome. Munchausen syndrome is a person who basically tells lies in order to get attention. Yes, Baron Munchausen is of course the source of the name, the tales of Baron Munchausen, a person who tells Arabian Nights stories, fairy tales, and so on, in order to get attention from his surroundings. So this is a phenomenon we know: a person lies or exaggerates or tells all kinds of stories that never happened, so that people will relate to him, listen to him, give him attention. That is called Munchausen syndrome. What is Munchausen syndrome by proxy? It is when a person tells stories that never happened about someone else in order to get attention. For example, a woman who sometimes not only tells stories but harms her son so that her son will be terribly miserable and then everyone will pity him and her, how beautifully she is taking care of him. Okay? That is called Munchausen syndrome by proxy. There was the starving mother, yes, there was such a case a few years ago, the starving mother who starved her children, and the claim was that this was in fact Munchausen syndrome by proxy. She starved her children so that everyone would pity the family and give her attention, basically. She has children in very, very difficult condition and she’s taking care of them and so on and so on, in order to get that attention. She did something to someone else, to another party. She didn’t just tell fairy tales about herself, but actually told about someone else or did something to someone else in order to get attention. That is called Munchausen syndrome by proxy. Now, one of the events that actually helped bring this syndrome to public attention was a story in England. There was a mother whose two children died of crib death. Not one—two. Now they took her to court on a murder charge. The claim was that she murdered the two children. Why? So some British doctor came, Sir Roy Meadow by his title then—the knighthood was taken from him later—but he came there and argued before the judge that the probability of crib death is one in eight thousand children. The probability that two children die of crib death is one in eight thousand squared, meaning one in sixty-four million. It has almost no chance, and therefore it’s obvious that she murdered them. If it’s not crib death, then apparently she murdered them. And on that basis this woman was put in prison, on that argument. And Roy Meadow’s claim was basically that this woman had committed an act of Munchausen by proxy. She murdered two children so that people would pity her for having lost two children. It’s a classic case of Munchausen syndrome by proxy. And she was put in prison for that. And I don’t know after how long—she sat in prison quite a while, I don’t know how long it took until some statistician came along and put that idiot judge in his place.
[Speaker C] And he basically told the judge an argument that can be phrased in several ways, so I’m thinking which wording to choose. The claim basically is the following: how many mothers are there in England?
[Rabbi Michael Abraham] I don’t know, several good millions, right? So let’s say for the sake of discussion ten million mothers. So once there is a probability—first of all, maybe this is also an important point in our introduction—if the probability of crib death is one in eight thousand, the probability of crib death of two children is not one over eight thousand squared. Because you have to assume that the two events are independent in order to multiply the probabilities. Meaning, if the events are dependent—say, for example, if there is something in the family, we talked about a kind of factor, remember from the previous series? If there is something that causes crib death in the genetics of this couple, of these children, then the fact that one died of crib death is not independent of the other one dying of crib death. It may be that the same factor causes both, or at least with some probability causes both of them to die of crib death. Therefore it is not correct to multiply the probabilities and say that the first dying of crib death is one in eight thousand and the second is another one in eight thousand, so the chance that both will die is one in eight thousand squared. You multiply probabilities only if the events are independent. But if they have a common cause—say if there is a reason why all the children in this family will die of crib death, some genetic disease or another—then you understand that the probability that two children will die is the same probability as one will die, simply because the same cause will kill all the children there, right? Only where it is an independent event—meaning it just happens statistically independently and twice it struck the same family—only there is the probability one in eight thousand squared, because it’s independent. But if there is dependence between the events, it is not correct to multiply the probabilities. Okay? So first of all it’s not one in eight thousand squared. But beyond that, let’s say it’s not one in sixty-four million, let’s say it’s one in a million, okay? Second claim: but there are ten million mothers in England. How many of them had two children die of crib death? Apparently only this one, because in fact only she was sued. So if the probability is one in a million and there are ten million mothers, then obviously there will be one mother here—or a few mothers—whose two children die of crib death. So here, this is that mother. How can you put her in prison when there is such a tiny probability? Notice, this is somewhat related to rarity versus exceptionality. It’s like saying Reuven won the lottery. What is the probability that he will win the lottery? Very small, one over the number of—I don’t know what—ticket buyers. Let’s say one in a hundred thousand. So apparently he lied, because it’s one in a hundred thousand that he would win, and in fact he won, so apparently he lied. What do you mean? But for anyone who wins it’s one in a hundred thousand. It’s a rare event but not an exceptional one. Here too, it’s a rare event that two children die to this mother, but it’s not an exceptional event if there are ten million mothers and the probability is one in a million that this will happen; then there will be several mothers to whom this will happen. So how can you put such a woman in prison?
[Speaker K] Can you also put against this a statistic of what the probability is that she would murder her two children?
[Rabbi Michael Abraham] Exactly. Another formulation of the problem—I don’t know whether another formulation or actually another problem—it’s another problem, not just another formulation of the problem—is that opposite the probability of what is the chance that two children die to the same mother of crib death, put the probability: what is the chance that a mother murders her two children? I don’t think that phenomenon is any less rare. Now you always have to compare, as Sherlock Holmes says in The Sign of Four, if after you have eliminated the impossible, whatever remains, however improbable, is probably the correct option. Okay? Meaning you always have to compare between alternatives. The fact that something is improbable does not mean it won’t happen. If the alternative is impossible, then this possibility, although it is improbable, is probably what happened, because the alternative is impossible. Okay? Now here compare the probabilities. You have the probability that a woman kills two children, and the probability that two children die of crib death. I don’t know which of these two events is the one with the very, very small probability. So how can you decide in favor of her having murdered them just because the probability that her two children die is a terribly, terribly small probability? But the probability that a woman murders her two children is also a very, very small probability. Therefore the fact that something is rare is not a reason to assume that it won’t happen or that it couldn’t have happened. If the alternative is also rare, then you cannot convict a person on the basis of such an argument. A third way to present the problem—and this is already connected a little more to us—is that when we use a statistical test, the statistical significance of the test has to be relative to the rarity of the phenomenon we are examining. Yes, the analogy I like to give for this is: when I use a net to catch fish, it’s a good idea to use a net with very, very small holes, right? But that’s all assuming I’m catching fish that are larger than the size of the holes in the net. But if the fish I’m catching are smaller than the holes in the net, then it won’t help me that the holes in the net are small. If the fish are even smaller, then even if the holes in the net are small I won’t be able to say, ah, this is a net with small holes, surely it will catch the fish. No. It depends on the size of the fish. Meaning, the net I use to catch the fish, the size of the holes there has to be smaller than the size of the fish, otherwise it’s irrelevant, and it doesn’t help to say that the holes in this net are very small and therefore it’s an excellent net. It’s an excellent net for certain fish, but if the fish are smaller than the holes of the net, then what do I care that the holes in the net are small? Yes, they need to be small relative to the fish you’re catching, not small in some objective sense. Okay? Meaning that the test I use to capture a certain result, to prove a certain result, has to be a test that is good enough on the scale of the phenomenon. Therefore what? The moment I have a woman whom I think murdered two children, the chance that a woman murdered two children is very, very small. Right? And now I ask myself what test would be good enough to prove that this very rare event happened? You understand that the test has to be a test with very high resolution in those terms as well. Meaning, if the chance that the woman murdered her two children is one in a million—that one woman in a million murders her two children—then the statistical test that proves she is a murderer has to be something with significance of one in a hundred million. Because significance of one in a million is not good enough, since the phenomenon itself is of a magnitude of one in a million. So a test whose significance is one in a million is not a good test for such phenomena. A beautiful example of this is Daniel Kahneman’s representativeness fallacies. The phenomenon, in the last formulation—the phenomenon I can demonstrate to you—is the following. Suppose you went to a doctor and the doctor wants to check whether you have some disease, some rare disease. He sends you for a test and the reliability of the test is ninety-nine percent. Meaning in one percent of cases it misses. In ninety-nine percent of the cases the test is good. Now in both directions. Meaning it misses sick people one percent of the time and healthy people one percent of the time. Okay? Meaning the error is in both directions—false positive and false negative, both are one percent. Okay? Now I went, the doctor sent me, I went for the test. The test showed that I’m sick. What is the probability that I am sick?
[Speaker C] Ninety-nine percent. Ninety-nine percent.
[Rabbi Michael Abraham] Do you agree? If it were like that, I wouldn’t ask. The answer is no. In fact, if the disease is a rare disease, there is no chance that I am sick. A test that is ninety-nine percent accurate—there is no chance that I am sick. I can go home happy and cheerful. For example, if the disease occurs in one out of ten thousand in the population—not all that rare, but rare. Okay? One in ten thousand in the population. And I went and took the test and I came out sick. The probability that I’m sick is one percent. Because it’s one hundred versus one in ten thousand. We said it needs to be the same ratio. Let’s do the calculation. I’ll show it to you very simply. Let’s say there are a million people in the country. Fine? The disease is one in ten thousand. That means I have a hundred sick people. Right? Out of the million, there are a hundred sick people because one in ten thousand is the prevalence of the disease. Okay? Now there is a test that diagnoses this disease and its reliability is ninety-nine percent. I put all one million people through this test. How many people will come out sick? We have one million minus one hundred healthy people and one hundred sick people. Right? So let’s do the arithmetic. Out of the hundred sick people, the test will miss one. Right? Meaning one will come out healthy and ninety-nine will come out sick. Out of the million healthy people, the test will also miss one percent. Meaning ten thousand will come out sick and nine hundred ninety thousand will come out healthy. Right? So I have ten thousand healthy people who came out sick, plus ninety-nine sick people who came out sick. So all in all this test gives me about ten thousand—ten thousand and a hundred came out sick out of the whole population that underwent the test. How many of them are really sick? A hundred. Meaning that if ten thousand one hundred come out sick and only a hundred are truly sick, then if I came out sick—what is the chance that I, having come out sick on the test, am really sick? One hundred divided by ten thousand. Meaning one percent. Because most of the people who came out sick are actually healthy people whom the test misclassified. And the rarer the disease, the more healthy people there are and the more healthy people the test will misclassify, because there is one percent that it misclassifies. Why is this the same as what I said before? Because once the disease is rare at the level of one in ten thousand, that’s the size of the fish. The size of the fish is one in ten thousand. The quality of the test is the size of the holes in the net. That’s the net trying to catch the fish, trying to distinguish in the test. Now if the size of the net is one in a hundred, but the size of the fish is one in ten thousand, then the net won’t catch the fish. Even though one in a hundred is a good test. Yes, the fact that the false positive is one in a hundred is perfectly fine, that’s a good test. But there is no chance that I am sick. If the disease is really rare—say one in a million—then there is certainly no chance that I am sick. So the chance that I am sick is—I came out sick—the chance that I am sick is one in ten thousand. A test that is ninety-nine percent reliable came out saying I’m sick. The chance that I am really sick is one in ten thousand, meaning there’s no point at all even checking further. I can simply—having come out sick—I can go home happy and cheerful; there’s no point in continuing to test at all. There’s no chance that I’m sick. In other words, there’s no point in using this test. For a disease with rarity on the level of one in a million—or even one in ten thousand—there is absolutely no point in using a test whose reliability is ninety-nine percent. That test is worth nothing. And I know—I spoke with medical students and discussed these matters with them a bit—so they told me that today they already learn this, at least in Jerusalem, those were the students in Jerusalem. So they told me that today they already learn it. But until not very long ago, doctors didn’t learn this. I assume that even those who learned still don’t know it, but doctors didn’t learn this and didn’t know it. Therefore a doctor can send you for a test whose reliability is ninety-nine percent and tell you, sir, you need to start taking medication or undergo such-and-such treatment, when the probability that you are sick is negligible. Because the test is ninety-nine percent reliable, so the doctor says: what do you mean, it’s an excellent test. It says you are sick, the overwhelming majority of chances are that you are sick. Not true. The same thing, by the way, with legal evidence. You have legal evidence that is ninety-nine percent reliable; that’s very good evidence. Think about a murderer. How many murderers are there in the population? I don’t know, one in a million. Fine. Now you have evidence with ninety-nine percent accuracy that I am a murderer. What is the chance that I really am a murderer? One in ten thousand. There is no chance that I am the murderer. So what is evidence of ninety-nine percent worth for a relatively rare crime like murder? Nothing, it is worth nothing. Now at first glance this is a bit discouraging. Why? Because if that’s so, then there is no way to diagnose diseases, at least rare diseases, and also no way to identify murderers. So what do we do in court, or what does a doctor do when making a diagnosis? The answer is that there is a way. Why? Because usually when they bring me to trial, it’s because I had motive, opportunity—yes, all these criteria in criminal law—and I was in the area. I had motive, I had opportunity, and I had the means. Okay, so in fact how many people who had motive, opportunity, and ability are there in the whole population? Five. How many people who were at the murder scene at that time and also had the possibility to do it and the motivation to do it? Five. Now when you bring evidence with ninety-nine percent reliability, that’s good evidence. If there were a million candidates to be murderers, then evidence of ninety-nine percent is worth nothing. But if I have some preliminary filters—that is, things that reduce the potential murderers to five or ten or three or I don’t know what—and now you’ve brought evidence about someone that he is the murderer with quality of ninety-nine percent, that’s good evidence. The same thing in medical diagnosis. If I have symptoms of the disease, but they are not conclusive—symptoms—but these symptoms already say that there is at least a twenty percent chance that I am sick, now a ninety-nine percent test will be good. If you just send a person with no indication whatsoever for a test for a rare disease and the test is of ninety-nine percent quality, it’s worth nothing. But if I have symptoms that arouse suspicion and now they did the test, then that is perfectly fine. But one has to be aware that the symptoms have to narrow the population sufficiently so that the quality—the quality of the test will already be relevant to the size of the population or to the chance that I am really sick. And this is a very, very confusing business and very non-trivial. And lawyers and doctors generally don’t really understand statistics, so one has to be a little careful. Why is this connected to us? Because with disease it demonstrates—but all the examples are the same thing. I’ll demonstrate it with disease. What does it mean that the quality of the test is ninety-nine percent? The quality of the test being ninety-nine percent means: given that I am sick, what is the probability that the test will reveal it? That it will say I’m sick. That determines the reliability of the test. Or given that I am healthy, what is the probability that the test will say I’m healthy. Okay? That is the measure we define as the quality of the test or the reliability of the test. But the question we ask is the opposite question. Given that the test says I’m sick, what is the probability that I’m sick? This is conditional probability, but it’s not P of A given B, it’s P of B given A. Again, the reliability—and here I close the circle, that’s why I brought this example—the probability, remember what I said about the result two? If the result was even, what is the chance that it is two? One-third. If the result was two, what is the chance that it is even? One. If it is two then obviously it is even. Meaning P of A given B and P of B given A are not the same thing. In our case, if I am sick, what is the probability that the test will say I’m sick? Ninety-nine percent. Fine? So the probability that the test says I’m sick given that I’m sick is the ninety-nine percent; that’s the data of the test. But what question do I ask the doctor? I ask the opposite question. The test said I’m sick—what is the probability that I’m really sick? That is P of B given A, not P of A given B. And it is not the same number. And if the test is for something rare, then it could be that P of A given B is 0.99, ninety-nine percent, but P of B given A is nothing—one in ten thousand. Therefore it is very important, when we talk about conditional probability, to distinguish between the probability of A given B and the probability of B given A. And next time I’ll do this a bit more systematically, but these examples were only meant to clarify this point of conditional probability. The direction of the conditioning is very important. Okay, I’ll stop here. Anyone who has comments or questions?
[Speaker I] Rabbi, in the cases of the murderer and legal cases, is it not a matter of probability—is that the tool? Is probability, statistics, the tool? Isn’t it more plausibility?
[Rabbi Michael Abraham] Plausibility—it doesn’t matter—but it’s still the same thing. The fact that I don’t have a quantitative measure isn’t important, so give me an intuitive measure: one percent. Fine. Ninety-nine percent that he’s the murderer—I’m convinced at the level of ninety-nine percent. Not because I have some probabilistic calculation, but that’s the level of plausibility I feel. Yes, but if there’s a murderer at a rate of one in a million, then that kind of plausibility doesn’t help. I illustrated it with probability because there I know how to do the calculation, but the idea is also true regarding plausibility.
[Speaker G] Thank you. Rabbi, doesn’t this place a lot of responsibility on us in understanding that when we relate to reality, our judgment, our agendas, really have a big influence? For example, that mother of the two children—again, as the Rabbi said, it’s very rare—but suppose we had known her and seen that, for a long time, the family doctor and the schoolteachers had the impression that she wasn’t really a good mother, maybe even a very bad mother. Her attitude toward the children was very alienated and distant, and she didn’t take care of them and scolded them—not something violent, but something that’s not what a normal mother is like. That’s our judgment, so then all the statistics would already look a bit different. Maybe the judge should have said, wait a second.
[Rabbi Michael Abraham] You could even quantify it and ask ourselves: among mothers who are at that level of motherhood, how many mothers might murder two children? Here the answer would no longer be one in a million. Here the answer would already be one in a hundred.
[Speaker G] Exactly, but I’m saying, “a mother like that,” a statement like that, “a mother like that”—that’s already something we, to a large extent, create. It could be that if we went to some culture where all the mothers are like that, where mothers are very alienated—
[Rabbi Michael Abraham] What I’m demonstrating right now is exactly the point that we don’t create it. Look, I translated it into factual statistics. That’s exactly the point. Meaning, after all, I know, say, that a mother who murders two children is one in a million. I checked all the cases; it’s one in a million among mothers in Britain, for the sake of discussion. Okay? But I know—come on, let’s gather all the bad mothers. That mother who murdered two children is probably one of the bad mothers, so let’s assume I found a hundred bad mothers and that mother is one of them. If I know she’s a bad mother, then the chance that she’s the murderer is already one in a hundred, not one in a million.
[Speaker G] So here the answer already—
[Rabbi Michael Abraham] —would be one in a hundred.
[Speaker G] Right, but suppose we call her childhood friend, her soul-friend, that same mother’s closest friend, who has known her since birth and loves her with all her heart and soul, and she says: she has a good soul. She looks like a bad mother, but inside she has a heart of gold.
[Rabbi Michael Abraham] Here there’s no calculation anymore. The question is whether you believe her or not.
[Speaker G] That’s why I’m saying this casts doubt on everything we explained here—that even the probability with which we approach reality is, to a large extent, something we create.
[Rabbi Michael Abraham] There’s the story of the data from what happened in the United States after DNA was accepted as admissible evidence in criminal trials—there were catastrophes there. They tested convicted murderers who were sitting in prison, and very high percentages came out showing that they were proven not to be the murderers, even though they had been convicted in court. And it turned out that a very high percentage were not the murderers. Who knows, among those they executed, how many people actually weren’t murderers at all. But what happened? They didn’t have the tool of DNA at that time.
[Speaker C] Okay, thank you very much.
[Speaker J] Sabbath peace.
[Speaker L] Rabbi, can I ask a question that isn’t related to the lesson? Yes. Sorry—regarding, well, people have asked the Rabbi this many times, regarding what the Rabbi says about providence. Is it true that providence—yes, I see the Rabbi is used to this—that the Holy One, blessed be He’s providence is not—well, not illogical—that there is, on the one hand, providence, and on the other hand, laws of nature, right?
[Rabbi Michael Abraham] What do you mean, not illogical? If there is involvement by the Holy One, blessed be He, that involvement breaks the laws of nature. There is no involvement within the framework of the laws of nature.
[Speaker L] Right, but after all, the laws of nature are ultimately deterministic.
[Rabbi Michael Abraham] He created them; He can also freeze them.
[Speaker L] Certainly, but the laws of nature are deterministic, right? Absent His intervention. So if from the outset He arranged reality in such a way that things would happen—that there would be an earthquake in Turkey in who knows how many billions of years—and that itself is providence, because it can also be fitted to a person’s deeds, after all every deed a person does, then that is providence, basically.
[Rabbi Michael Abraham] What kind of providence is that? Influence from a distance—but for that He’d have to wait and see what the Turks do.
[Speaker L] No, that things are affected, after all. Things are affected. Look, one time I was debating whether to go to prayer or not go to prayer; I didn’t have the strength to get out of bed. I got out of bed, went to prayer, and found a 200-shekel bill on the floor. So my actions caused it, in the sense that thanks to that, yes, I found the bill.
[Rabbi Michael Abraham] He’s asking how He determined there would be an earthquake in Turkey when He didn’t know—maybe the Turks would become the greatest lovers of Israel?
[Speaker L] Yes, but after all, I don’t know how it works with earthquakes specifically, but in principle—the butterfly effect.
[Rabbi Michael Abraham] That a person, for argument’s sake, in Japan—forget the butterfly—
[Speaker L] The butterfly—
[Rabbi Michael Abraham] —that’s completely deterministic. What do butterflies have to do with this?
[Speaker L] Not actual butterflies specifically, but because someone in Japan saw that Israel declared a few new settlements in Judea and Samaria, and in frustration he made some gesture with his hand on the sofa, and that caused a butterfly effect leading to a tsunami. Doesn’t matter, just an example—but every action a person takes can have effects.
[Rabbi Michael Abraham] But why did those tsunami victims die? What are they guilty of? I don’t understand.
[Speaker L] Doesn’t matter with the tsunami—every action everyone takes has some effects. Through the butterfly effect there can also be enormous effects. So if from the outset the Holy One, blessed be He, arranges reality so that every action a person takes—such that, in a fitted way, whatever each person does will cause the desired effects, so that it really will be providence tailored to each individual. Providence from a distance, admittedly, but tailored to each person.
[Rabbi Michael Abraham] So you want to say that the earthquake in Turkey struck them because they deserved it, and this was planned in advance? Because I want to bring you down to concrete examples, not stay with slogans.
[Speaker L] No, I’m asking about the principle that you—right, you say there isn’t—it isn’t logical, the laws of nature don’t fit with providence? So I—I’m not even saying there is providence. All I’m saying is that the laws of nature don’t contradict the fact of providence.
[Rabbi Michael Abraham] Of course they contradict it. What do you mean they don’t contradict it? I have free choice. The moment I have free choice, then the apple will fall—fall on me from the tree because the force of gravity pulls it, regardless of whether I sinned or didn’t sin.
[Speaker L] But it could be that because I sinned, I’ll come to that tree in a bad mood, and because I come in a bad mood, I’ll come—God will arrange things so that I choose a different tree instead of that tree, and not the tree from which the apple fell.
[Rabbi Michael Abraham] Again, so in any case He intervenes—so He always intervenes.
[Speaker L] So He’ll arrange it. No—from the start, no—from the start He arranged it in a deterministic way. After all, from the start He established the laws of nature in a certain way. From the start, the laws of nature would cause—there were two possibilities. If the person commits the sin, then what would happen deterministically is that the person would not—Newton, not Newton—would not get to that tree. If he didn’t commit—
[Rabbi Michael Abraham] —the sin, then that means the laws of nature create the punishments; the laws of nature create punishments for whoever deserves them.
[Speaker L] Yes, but not specifically—not generally in the way the Maharal says that if a person eats pork, then tomorrow his leg gets stuck in a table, because it’s a law of nature that pork always causes your leg to get stuck in a table, but rather in a way—rather, each time it’s fitted, because from the outset—
[Rabbi Michael Abraham] But then that’s why I’m asking—so I’m returning you to the example I gave earlier. So the earthquake in Turkey basically struck only people who really deserved to die? That’s what you’re claiming.
[Speaker L] I even claimed that maybe yes, even though it’s the laws of nature. And the fact that it’s the laws of nature doesn’t contradict the answer being yes.
[Rabbi Michael Abraham] And why do babies deserve to die?
[Speaker L] I don’t know. I wasn’t discussing the question of— I said that it doesn’t contradict the laws of nature. The Rabbi says, wait, according to what you’re saying, it comes out that an earthquake in nature was not unfair. Maybe. I don’t know exactly what’s going on there. Collective punishment is another thing that seems very difficult to me. Right now I don’t know whether there was any punishment there at all. Maybe—maybe there are major natural disasters that happen without any connection to providence. I don’t know. That’s not the point I’m making. All I’m coming to say is that there can be providence tailored to the deeds of each and every person, even though everything is the laws of nature, because from the outset everything happened deterministically.
[Speaker M] But we’re skipping over a question here. Who are you, anyway—who are you to decide, who are you to decide when God decides to punish or not punish?
[Rabbi Michael Abraham] No, simply, He doesn’t decide.
[Speaker M] What, it’s not discretion? He can also decide not to punish.
[Rabbi Michael Abraham] He doesn’t decide. He sets up a possible mechanism. Meaning, if the claim is that God punishes people for their sins, you can reconcile that with natural determinism. Whether it happens or doesn’t happen—we’ll talk about that later, and how it happens—but the question is whether it can be reconciled with that.
[Speaker L] Exactly, so it can be reconciled, right?
[Rabbi Michael Abraham] But what, on the principled level?
[Speaker B] So then no—you can’t say that this is Maimonides’ intention when he speaks about general providence and individual providence. General providence is exactly what this fellow just said: that in principle, obviously the Holy One, blessed be He, knew fourteen billion years ago, within determinism, what would happen another fourteen billion years later. He didn’t know the person’s free choice, that’s true. But it could be that He created the—
[Rabbi Michael Abraham] His claim is that He doesn’t need to know. His claim is that He doesn’t need to know. He leaves the mechanism in such a way that sins will cause punishments—or something else to appear in the end—and the laws of nature do it.
[Speaker B] Fine, but from our perspective He knew it. If He had run the divine supercomputer, He would have seen—look, exactly—
[Rabbi Michael Abraham] Rabbi Abraham will do an act of kindness, and in the accident it will turn out to help Reuven exactly.
[Speaker B] No, no, even without His knowing.
[Rabbi Michael Abraham] He doesn’t know.
[Speaker B] What do you mean, knew? Why should I care whether He knew? What does it even mean that God knew? God—there’s no reason He shouldn’t know. Is He busy? What’s the problem with His knowing?
[Rabbi Michael Abraham] Not that He’s busy—because it isn’t deterministic.
[Speaker B] But the fellow asked that it is deterministic. Fixed laws of nature in advance—that’s deterministic.
[Rabbi Michael Abraham] You didn’t understand the question. The claim was that even though sins are not deterministic, it still could be that punishments can come only through the laws of nature.
[Speaker B] Well, I agree with that. What’s wrong with it?
[Rabbi Michael Abraham] No claim—he isn’t claiming determinism. He’s claiming that you don’t—
[Speaker B] —need to arrive at determinism in order to do that. Fine, but I understood him to be saying that this doesn’t count as saying there’s no providence, since the laws of nature, if the Holy One, blessed be He—
[Rabbi Michael Abraham] —made them deterministically, then supposedly He already knows everything.
[Speaker B] Right, so if, say, God created different deterministic options such that the laws of nature would operate according to human actions, then supposedly that too is providence.
[Rabbi Michael Abraham] Correct, that’s what he’s claiming.
[Speaker B] Then I agree.
[Rabbi Michael Abraham] Then I agree too.
[Speaker L] Okay. So I’m saying that now, from the Hebrew Bible (Tanakh), after all we see that even in places where there are no miracles—that is, the Book of Esther, or whatever, the story of the brothers
[Rabbi Michael Abraham] and Joseph—wait, there’s some echo here.
[Speaker L] Is it okay now?
[Rabbi Michael Abraham] Aside from the person speaking, maybe everyone else could mute? Yes, let’s try.
[Speaker L] Yes, I’m saying that the laws of nature that—I think from the Hebrew Bible (Tanakh) we see that even in situations that look natural, even in places that look natural, yes—the Book of Esther, now the month of Adar, whatever, the story of the brothers and Joseph—it’s clear that even things that are completely within the laws of nature, the Torah relates to as providence from God, it seems to me. So it seems that the Hebrew Bible (Tanakh) teaches that there is providence even apart from miracles—even indeed without miracles.
[Rabbi Michael Abraham] Listen, I have no problem with that thesis, no problem. Fine—so providence of that kind, as long as it’s done through the laws of nature and the laws of nature organize everything, everything is fine, no problem. Only of course we still haven’t solved the problem of evil and everything we talked about earlier, but yes, granted. As long as you’re not telling me that He is involved in a specific case and changes reality because I sinned or something like that, but rather the laws of nature do it in one way or another, granted—I’m willing to accept that. So excellent.
[Speaker N] Thank you very much. Thank you very much. Okay.
[Rabbi Michael Abraham] All right, have a peaceful Sabbath, and goodbye.