Free Will and Choice – Lesson 16

[Rabbi Michael Abraham] Last time I began with a series of thought experiments whose purpose is to help us formulate a position regarding the issue of free choice, after we reached the conclusion that by philosophical means it is doubtful whether we can arrive at an answer, by scientific means we certainly cannot arrive at an answer, at least as of today, and then the question basically remains open and a person has to decide where he stands on it—to decide, yes, assuming that he really does decide—where he stands on it. And therefore this series of thought experiments is meant to give each of us tools to examine himself. That is, to check where he stands in relation to this issue. So I spoke about all those outrageous proposals, yes? That I’ll dismantle you, kill you, and afterward I’ll take your cells, or your molecules if you like, one by one, reassemble you, and I’ll also give you a million dollars. Are you willing to accept that proposal? Alternatively, I can kill you when in the next room I’ve already built you beforehand, don’t worry, everything’s fine, I don’t even need to build you. I’ll just kill you and everything’s fine, there’s already a perfect duplicate of you living with your family at home. I can kill you here and I don’t even need to do anything afterward, nothing will happen, you still exist. A series of proposals that, at least in my estimation, very few people—including extreme materialists—would be willing to accept. And basically what that expresses, I think, is that many people, although they declare themselves to be materialists, deep down have some implicit belief in dualism. That is, they really do think that there is something in a human being beyond the collection of molecules and cells that make him up. Of course, you can always hide behind the claim that I simply have some instinctive aversion—true, there’s no justification for it. But I have an instinctive aversion, and therefore I’m not willing to accept that proposal. But that’s just being disingenuous, because what do you mean by aversion? Then overcome the aversion—for a million dollars it’s worth overcoming an aversion. In other words, if you want I’ll give you a billion. That’s not it. Those are always the kinds of answers you get from a materialist when you show him that neither he nor I really assumes materialism, and then all kinds of formalistic answers begin that seem completely detached. That whole series of experiments I spoke about last time actually deals with the question of materialism, not determinism. It basically examines our conception of what a human being is—whether a person is a lump of matter, whether what there is in him is only matter, or whether there is in a person some additional spiritual, mental dimension beyond matter. So here those experiments are basically trying to show—or to give me a tool to realize—that I assume there is something in me besides cells and molecules. But that is about the question of materialism. I want to move to an experiment that will test the question of determinism, which is the question of whether we have free choice—not whether we have a soul and spirit—and I already said that these two questions are not the same question. There is overlap between them, but it’s not the same question. And before I move to the second part, I just want to complete a point I began last time, in order to sharpen this point further: what does it mean that there is something in a person beyond matter? I spoke a bit about the principle of the identity of indiscernibles; that is, I tried to show that philosophically this requires a somewhat more precise definition of whether there is something in the whole that goes beyond the parts. Look, I think there are a few—these are also thought experiments—that can help us on this matter. Just one second, there was something that slipped my mind. Okay, yes. I already spoke about the property of emergence, so let’s leave that for now. So there’s this: there is the Turing test, which I already described, which tries to provide a criterion for when a computer, or an artificial intelligence system, is something we can or should relate to as a human being, with all that that entails—rights and obligations and everything we attribute to a human being. And Turing proposed some practical test: once you put such a system and a human being in a closed room, and they communicate with me through, say, a computer, a keyboard—I don’t see who is speaking to me—and I won’t be able to distinguish which of my two conversation partners is the computer and which of the two is a human being. Once I can’t distinguish between them, apparently the computer has already reached human capabilities, and therefore it is essentially a human being, a human being in every respect. I think I mentioned the film Her, yes, where someone fell in love with a computer program he had, because he conducted all kinds of conversations with it through artificial intelligence, and it basically had intimate conversations with him, and he simply fell in love with it, with that software. So that’s an artistic expression, let’s call it that, of Turing’s experiment. It’s interesting to watch the film because suddenly you see that it really is—you really can enter into that kind of situation. When I say it here, it sounds detached, like what, falling in love with a computer sounds a little bizarre. But when you live alone with a computer and you communicate with it and it answers you, and suddenly you see that this is a very interesting person—or creature—with sensitivity, with all sorts of things of that kind, like when you communicate, say, online with someone you’ve never met in your life, it’s basically the same thing. You can fall in love with someone, a man or a woman, without seeing them, through online conversation. And in that sense it’s very similar, and therefore I think that when you watch the film you suddenly enter into that experience and understand that this talk is not quite so detached. Because you need a little capacity for abstraction in order to accept these Turing tests as real tests. Our initial instinct resists it, refuses to see a pile of metal as a human being. I’m not willing to either, but not because of an initial instinct; rather, I think it’s simply not true. So here in this context there are two more experiments that may perhaps sharpen the matter. One experiment is John Searle’s Chinese room, which basically proposes a fairly similar experiment. He says, let’s put a person in a closed room with an input window and an output window. And inside the room there are two big boxes with Chinese letters. Now this person is Israeli, yes, he doesn’t know Chinese, he has no idea what the Chinese letters mean, he does not know the Chinese language, doesn’t know how to read or write Chinese. Through one window come questions written in Chinese, and through the other window he is supposed to answer, with answers that are also in Chinese, to assemble from the letters he has in the boxes an answer and send it out through the output window. And every time his answer is irrelevant, he gets an electric shock. Painful. Okay. Searle’s claim is that over time, after an infinite amount of time—billions of years, let’s say, for the sake of discussion, with this person having eternal life, all right? So he’s like the hydra we talked about last time. So he has eternal life, and therefore after very, very many attempts and a very, very long time, the person develops some ability to give the correct answers. That is, he receives questions that he has no idea what they mean, and he assembles some answer, but he knows that if a question of this kind comes in, then I answer with an answer of that kind, because otherwise I get zapped. So he assembles letters for himself, and he checks empirically what will succeed in preventing the next shock. And through that he builds his answers. And let’s say that after billions upon billions upon billions upon billions of years, he succeeds in giving relevant answers to the questions he is asked. Okay. So now the question is whether he knows Chinese. Because when people talk to him, they ask him in Chinese, he answers appropriately in Chinese, the whole conversation goes perfectly well, yes. In the Turing test, if you put him together with a Chinese person in a closed room and talk to both of them, you won’t be able to tell the difference. That means both of them are essentially speaking to you, responding to you in Chinese, and carrying on a conversation with you like a person. In that sense it is really just like the Turing test, and the question is whether this man knows Chinese. And the claim is that no, he does not know Chinese. Why? Because one must distinguish between the question of input and output—yes, whether he functions like someone who knows Chinese—and the question of whether he knows Chinese. That is, a phenomenological description of his behavior does not necessarily indicate the essence. In other words, even if he communicates in Chinese just fine, that doesn’t mean he understands. Understands Chinese. Because understanding is not simply asking appropriately and answering properly. If you understand, then you can ask appropriately and answer properly. But if you ask appropriately and answer properly, that does not necessarily mean that you understand. The relation goes in one direction, not both directions. Here, for example, in the case of Searle’s experiment, after enough time you can reach a situation where you ask appropriately and answer properly, but you do not understand Chinese. Because understanding is some cognitive state, it is not the speech. Speech can express understanding, but speech is not itself understanding. To say that the man understands Chinese means that—logic distinguishes between semantics and syntax. That is, syntax is structure, and semantics is meaning. So one can analyze sentences in a language—or whatever things—one can analyze them linguistically at the syntactic, structural level, to see whether they are built correctly, and one can analyze them at the semantic level. At the semantic level the question is what this thing says, or what its meaning is for me. And in this syntactic sense, the man knows Chinese. That is, he builds sentences in a correct Chinese form. But semantically—that is, whether he understands the meaning of the sentence—that is already a question of cognitive processes, of processes of understanding in his brain, or in his mind actually, not in his brain. The brain is only an organ, as we said. That isn’t there, and therefore to say that this man understands Chinese is quite a misunderstanding. Someone who says that this man understands Chinese does not understand what it means to understand Chinese. That is, he has a problem in understanding; he is just saying something, he doesn’t understand it. And therefore, therefore it seems to me that this experiment describes why the Turing experiment, the Turing test, really cannot provide any criterion for the question of whether the thing standing before me is a human being. Because when the computer reaches such a level of human conversation, then at most it will have human syntax. That is, it will know how to build sentences that are syntactically correct and communicate with me, answer me appropriately, ask me appropriately, and everything is fine. But the understanding that exists in its consciousness, in its intellect, which usually produces the discourse—in the case of a human being, though not in the case of the computer—does not exist in the computer. And therefore to latch onto the symptoms or the derivatives and assume the existence of the basis that produces those derivatives is a logical error. It’s a bit strange to accuse Turing of a logical error, but it seems to me that here it is clearly a logical error. That is, one must distinguish between cause and effect. Understanding produces discourse, but that does not mean that when there is discourse, that necessarily indicates that there is understanding in the background. It is simply untrue. It’s hard for us to see how discourse could arise without understanding, but it turns out that there can be such ways, whether software programs of artificial software can produce it, or prolonged suffering over billions and billions of years in the Chinese room—this too can produce such a phenomenon of syntax, syntax without semantics. Just a remark, yes.

[Speaker B] In schools where they teach little children Yiddish, they can answer and do everything, and afterward you ask them what there is in the verse and they don’t understand, but they know how to answer the teacher in Yiddish.

[Rabbi Michael Abraham] Very often, yes, I think very often students’ studying for a test is like that. That is, they learn what to answer, but they don’t really understand the material. In other words, they can pass the test successfully because they know what to answer. Yes, right, that’s a good example of the difference between semantics and syntax. In any case, Searle’s experiment demonstrates, I think, why the Turing test is not really a test that reflects—and I think I mentioned, I mentioned on one of the previous occasions—Mary’s room. Mary’s room also has a Wikipedia entry, for anyone who wants to look. I don’t remember anymore who invented that example, but it’s an example that says something quite similar. It says that Mary is a brilliant physicist, an expert in optics, the best optics expert in the world. She knows everything you could want in this context of the physics of light, of waves. And she knows what happens when red light meets this kind of crystal structure, whether it reflects. She knows everything. Now she goes outside—the room, her whole room that she lives in, is a black-and-white room. She has a computer, she has books, but everything is black and white, there are no colors there. And there she specializes in everything that happens, she knows what every color does and how it behaves under all conditions and in all circumstances. Now she goes out of the room and sees an anemone, a red flower. Has something new come to her? Does she now grasp something that she previously did not grasp, or did not know? So the answer is of course yes, because until now she had no idea what the color red is. That is, she knew how to describe what a red light ray does under all conditions and in all circumstances, but she did not know what a red light ray is, what red is. Because she had never encountered it in an immediate way. She knows how to describe how it behaves and what it does and what it is, like they say about Rabbi Chaim, Rabbi Chaim of Brisk, that he took the frying pans out of the kitchen. When people studied the laws of prohibition and permissibility until Rabbi Chaim, they were basically studying medieval physics or ancient physics—how frying pans absorb, what happens in frying, what happens in cooking, and basically you became a housewife when you dealt with those things. Now for Rabbi Chaim, the frying pan is not at all some concrete utensil made in a certain way and whatever, whatever you see before your eyes. For Rabbi Chaim, the frying pan is a theoretical construct. That is, it is some abstract vessel whose known function is that you put oil into it and it produces fried food. And he has no idea what it looks like; hypothetical Rabbi Chaim has never seen a frying pan in his life. The frying pan is some kind of concept or abstract object that has a collection of properties. When Rabbi Chaim enters the kitchen and encounters a frying pan, he won’t recognize it. Again, this is that difference between the thing itself and its properties or its implications. Very often we have some tendency to confuse the thing with its properties. And also in the halakhic and analytic context, by the way, there is room to wonder whether the Brisker abstractions help me arrive at the truth. Is that the right way to analyze the laws of prohibition and permissibility? Or should one really not engage in abstractions, but actually go into the kitchen and feel with one’s fingers the laws of prohibition and permissibility, and not build theoretical constructs that represent the pot, the frying pan, a first vessel, a second vessel, something sharp, a sharp knife, and all those laws of absorption and release, and so on.

[Speaker B] And why is that different from quantum theory, where there are all kinds of mathematical methods and nobody in the world really understands what is going on there.

[Rabbi Michael Abraham] That’s a good question. I’m not sure it’s different. If we had some way to encounter the electron directly, maybe we wouldn’t need these strange abstractions of quantum theory. The problem is that there we don’t have that. We don’t see electrons; they’re too small. We have no way to encounter them directly. Wherever we have no way to encounter something directly, we have no choice but to rely on mathematical models. But that doesn’t really mean that the mathematical model gives me some kind of direct grasp or understanding of the thing itself. In many cases we really are in the same situation as Mary in her room: we know everything about the thing, but we still don’t understand what the thing itself is. And to a large extent, I think that’s the root of this sense of puzzlement about quantum theory. Our puzzlement is that we don’t really understand what’s going on there. We have the mathematics, we know how to do the calculations, I can maybe tell you what will happen in any given situation, but what does this all mean? I can’t manage to understand what is really happening there. And that’s exactly what I’m describing here, except that there we’re forced to be inside a Turing test or Mary’s room or the Chinese room or whatever you want, because we have no choice—we can’t get out of the room. We can’t encounter the thing directly, so we use descriptions of it, models for it, languages we developed in order to talk about it and describe it, but we don’t understand what the thing itself is. No physicist knows how to tell you what an electron really is, just like what Mary’s red color is. The physicist knows how to tell you that an electron behaves like a solution to the Schrödinger equation in such-and-such situations. Fine, I understand. That’s like saying that red light hits water and it refracts this way or that way, or bends around, or interferes, or all kinds of things of that sort. But what is red light? Same thing. The physicist doesn’t know how to answer what an electron is. He doesn’t have the tools to answer that. He knows how to tell you what the electron will do under such circumstances or other circumstances because he knows how to describe it with an equation. It’s really just like—I know, just as Mary knew optics. But what is an electron? To encounter it, to understand it in direct perception—of course direct perception means sight—but sight at that resolution is irrelevant, so there’s no such thing. It’s doubtful to what extent one can even conceptually talk about direct perception of an electron, because in vision things depend on wavelength. Vision feeds on light; light has a certain wavelength. Things that are smaller than the wavelength of light—it’s not even clear to what extent vision can be defined with respect to them. So it’s not only our inability to see; it may be that the very concept of sight is not defined for things that small. That already brings us to other conceptual questions. In any case, the claim is that very often we confuse the functions and practical expressions with the thing itself. And the Turing test does the same thing, and I mentioned, I think, those articles about the mathematical ability of sperm cells. As if sperm cells do calculations, solve differential equations, in order to move toward the egg and fertilize it. And once in Haaretz I saw some article that was tremendously impressed by the mathematical ability of sperm cells, and I said that this is simply a misunderstanding; once again, it’s the same confusion. Sperm cells do not have mathematical ability. Mathematical ability belongs to the mathematician or physicist who describes the mode of operation of sperm cells—because it’s usually not a biologist, so I say mathematician or physicist—who describes the way a sperm cell behaves because he uses mathematics. A sperm cell does not solve differential equations. A sperm cell simply moves. It moves as its nature causes it to move. I spoke about the mathematical ability of water and the Navier-Stokes equations and so on. Well, all of this basically shows us that these thought experiments are really coming to say: notice that everything you know about yourselves—that is, cells and molecules—are expressions of something that is somewhere behind all that. And you must not confuse the thing itself with its practical expressions. There is, in fact, me myself. Who is this me myself? Me myself is not the collection of my cells or my mental traits. Me myself is the possessor of the traits, the one whose traits these are. Therefore it seems to me that I once mentioned that I saw in an article by a man named Aharon Rabinovitch, from psychology—he was in psychology at Bar-Ilan, maybe still is, I don’t know, I don’t know him—that I once saw an article of his in Bedad, where he tries to locate the self of a person on the psychoanalytic map. Where is the self located? And I thought to myself that before even looking at his answers, his search is built on a mistake. His search is built on a mistake because the self is not supposed to be found inside the psychoanalytic map. The self is that thing of which the map is the map. Meaning, everything the psychoanalytic map describes is really functions, traits, connections, of what I call the self. It is the possessor of the traits. You cannot describe it on the map of traits. Which trait is the self? It’s like asking which of the traits is the table. Is it the fact that it’s made of wood? That it has legs? That it’s one meter high? What exactly is the table? That’s nonsense. The table is that object of which all these are the traits. In the list of traits you won’t be able to locate who the table itself is, because every list of traits is the traits of the table. The table is the possessor of the traits; it will not be found on the map of the traits themselves. Yes, it’s like looking on a topographic map, or a geographic map, sorry, of the Earth, for where the Earth is. The Earth is the object of which this map is the map. You won’t find the Earth on the map. It isn’t at this point on the map or that point on the map. The map describes the Earth. The Earth is the entity that the map describes. It is not on the map or inside the map. So there is a whole set of mistakes here when we relate to these things, some of which stem from a kind of begging the question. We basically assume that there is no such thing as a self, that there is only our body with its derivatives, and then we ask, okay, so where does this illusion of the self come from, this consciousness, this self-consciousness that a person has. But we assume that it is an illusion, and therefore we begin to look for where it is on the map. But it is not correct to assume that, and therefore it is also not correct to look for it on the map. I may speak about another important point connected to the relation between the thing and its components. Now we give a computer an exercise—say, do one plus two—and it gives us that the answer is three. Basically, is the computer performing addition? Is the computer really adding one plus two and reaching the conclusion that it is three? Of course not. The computer is simply moving electrons from place to place. There are billions and billions and billions and billions and billions of electrons moving around there inside the computer to all kinds of places. That’s what the computer does. And therefore one can describe the computer at several levels of integration. One can describe the computer at the level of elementary particles, or if you like a higher level, at the level of electrons and atoms. Okay? That’s the level of electric currents and electric fields, the physical level. On top of that one can describe the computer at the level of logic gates. Logic gates are units in the computer that carry out one logical operation or another. That’s a higher level of integration. And on top of that one can describe the computer through functions—what it does. Meaning, it knows how to calculate, it knows how to calculate addition, it knows how to calculate exponents, produce logarithms, or things of that sort—what it knows how to do at the functional level. That’s a third level of integration. When I look at the calculating operation the computer performs, one plus two equals three, you have to understand: the computer is not performing an addition operation at all. It is simply a misunderstanding to think that the computer is performing addition. What the computer is doing is moving electrons from place to place according to its structure and the software I put into it. Now what happens is that I, as the one who built the computer, built it in such a way that when I type on the keyboard one, and then I type plus, and then I type two, and then I type equals, I built it so that it gives me three. How do I do that? I construct some arrangement of logic gates, and then the beam—say, in the past it was an electron beam hitting the screen and creating a numeral there, right? So I move the beam in such a way, and I make sure that such-and-such a shape is formed on the screen. Now the computer doesn’t know that it is now writing three. The computer knows that it does this dot dot dot dot dot dot dot dot and then again a half-circle at the bottom and that’s how the numeral three is formed. But when I built it in that way, from my point of view what the computer did was an operation of one plus two and it gave me the result three. But from the computer’s point of view those are geometric shapes—three—and not even geometric shapes, really, but simply some sequence of operations by which it strikes the screen in a certain pattern at very specific points that create the shape three. This has no connection whatsoever to the calculation one plus two equals three. The computer does not do that. Okay? Where does the meaning of the operation performed by the computer reside? Where? Only in my head. Meaning, if another person came along—say, a person who doesn’t know, doesn’t think like us, has a different way of thinking, an alien, not a human being, it doesn’t matter, doesn’t think like us, doesn’t know the form of writing our numbers and letters—he would start typing on the computer, the computer would produce all kinds of things for him, and he’d have no idea what the computer is doing. He would not be able to know that the computer is doing one plus two equals three, because the meaning of the computer’s operation is not found in the computer itself. If you analyze the computer itself, you will not be able in any way to understand that it is now carrying out the operation one plus two equals three. You know how to decode it only because the one who built the computer built it so that when it outputs the results of its arbitrary operations, it will make one plus two equals three appear on the screen. You simply see these shapes on the screen. It’s typography. Meaning, it’s simply a form drawn on the screen, and we give it the meaning of an arithmetic calculation. But the fact that an arithmetic calculation was done here is only our interpretation of the physical processes happening there. There is nothing there apart from the physical processes. We build—say, a person who speaks English and a person who speaks Hebrew—their computer will look different because it needs to build on the screen, or at least partly different, because it needs to build on the screen different forms that express one plus two equals three, once in Hebrew and once in English. Well, here it’s the same because these are numerals, but I mean at the conceptual level. Therefore this identification of the computer, the calculating operations that the computer performs, with the ability to think, is simply a misunderstanding. The computer does not think at all; the computer moves electrons in every direction. The meaning of what the computer does exists only in my head. There is no meaning at all to the operations the computer performs. The operations of the computer in themselves have no meaning. The meaning is created in the connection between the builder of the computer, who was a human being, and me as the consumer of the computer. Because we both use the same language, we both know how to add and subtract numbers, and then he builds an automatic device such that when I see it, I can understand from that what the result of one plus two equals three is. It’s somewhat similar to—before the age of digital computers, they used to make analog computers. Analog computers are computers that output the result not by means of the digital numerical calculation we know today, but through a physical process that simulates the result. For example, if I want to differentiate a function, then I apply voltage to a coil, and the derivative of the voltage function gives me the current function. Okay? That’s the property of the coil. The coil differentiates the voltage across it and outputs a current that is the derivative of the voltage function. So if I put in a voltage function that is sine, then the coil will output the derivative of sine. Does the coil differentiate? The coil does not differentiate. The coil is simply a physical object that operates as it operates. I mean, just as water does not solve the Navier-Stokes equation, and just as a sperm cell does not solve a differential equation. Rather, when I see this, I understand that as a mathematical model of what was done here, the fitting mathematical model is a derivative. What was done here is differentiation. Okay? Therefore all these meanings are located in me. It is so mistaken, really at a childish level, to think that these things actually perform mathematical operations and thinking operations. And that’s why I say that the Turing test is so absurd that it’s hard to believe there are serious people who really believe in it. Meaning, who think that the moment a computer reaches levels of conversation and response and interaction like a human being, then it really will be a human being.

[Speaker B] The question is whether, with computers that now learn by themselves, they can do things that I didn’t plan, that I don’t understand; it knows more than I know. That’s not…

[Rabbi Michael Abraham] Right. What we concluded in the previous lesson was that water also knows that better than you do. It solves the Navier-Stokes equation that you don’t know how to solve. That doesn’t mean anything. You programmed the computer, you put into it this kind of artificial intelligence that has some sort of ability—again, you used the term “learn,” I wouldn’t use that term, because the computer does not learn. Rather, the computer responds in the way a person who was learning would respond. But when you refer to this process as learning, that’s like referring to the coil as differentiating. The coil does not differentiate. The coil simply outputs current according to its structure. I know that in the mathematical model that describes the operation of the coil, that’s a derivative. Same thing here: I know that the model describing the operation of artificial intelligence is what is called thinking, but that does not mean that the artificial intelligence really thinks. It doesn’t think. It does things, it acts and reacts, it receives input and produces output. I, in the meaning I assign when I look at this, suddenly understand that in fact what happened here is that it solved some problem that I don’t know how to solve, because in the meaning I give to things, a problem was solved here. From the computer’s point of view, the computer solved no problem and learned nothing. It does not learn and does not solve and nothing. I, when I give meaning to what the computer does, see it as solving, as learning, as remembering—all the functions that we define. And therefore there is here—this really is a very, very fundamental mistake that exists. And again, with artificial intelligence people I have no issue, because they are not supposed to make philosophical claims. They are professionals, and in their field I assume there are talented and good people there, and surely they know their field better than I do, and therefore I have no argument with them in the professional domain. I have an argument with them when they move on to the philosophical implications of the matter. When they speak about a computer becoming a human being, they are already stepping outside their professional role; they are making a philosophical claim, and here, in my view, that philosophical claim is really a childish mistake.

[Speaker C] Rabbi, can we say that it computes and doesn’t learn? Is that the difference between a computer and a learner?

[Rabbi Michael Abraham] Yes, but I would go even further and claim that it doesn’t even compute. Any physical lump like a computer, or like any other machine, is simply a causal system that receives input and does with it what its nature causes it to do with it. When I look at the computer, I call that the computer computing. But the computer does not compute. The computer does not perform the operation one plus two. The computer moves electrons, opens transistors, and does all kinds of things—that’s all. It does none of the things that I put into it. I look at it and give it some meaning because from my point of view—from my point of view it has meaning, from my point of view it simulates an act of calculation. It isn’t even an act of calculation, but rather from my point of view it is a simulation of an act of calculation, and therefore I can make use of this resemblance, this analogy, between what the computer does and what…

[Speaker D] But if that’s circular logic—meaning, whoever really thinks that way also thinks that our brain works exactly that way, that it’s neurons, that that’s how it works. So you’re saying exactly what they’re saying.

[Rabbi Michael Abraham] No, but my claim is that it’s like in the room—that’s why I brought the examples of the Chinese room, for instance. So what, in the Chinese room will he say that the man understands Chinese?

[Speaker D] No, but they’ll say, what do you mean by understanding Chinese? They really do say that, that’s what they argue. What we do, what we do—we learn, when we’re children, we learn the language somehow from these rules, we understand it, or maybe it’s even the hardware inside our brain and that’s how we work. There isn’t—this is what they claim. I’m not saying it’s true, I’m just saying—but it’s very…

[Speaker B] A determinist will say that a human being is just a computer.

[Rabbi Michael Abraham] Right—a materialist, not a determinist, but a materialist.

[Speaker B] A materialist will say that a human being is just a computer. He…

[Rabbi Michael Abraham] No, but I want to clarify something. Look, that’s why I brought the example of the Chinese room. In the Chinese room, basically, what Danny wants to say is that from the point of view… from the point of view of the artificial-intelligence person—I don’t know if it’s an artificial-intelligence person, because an artificial-intelligence person could think like me, but the philosopher of artificial intelligence, not the professional practitioner of artificial intelligence—would basically argue that a person who understands Chinese also operates that way. But here that’s simply a mistake, because with a person who understands Chinese, we know what it means to understand a language. That we know from ourselves. When I talk about something—when I understand Hebrew, that doesn’t mean that when you ask a question I answer relevantly. That is an expression of the fact that I understand Hebrew. But there are processes of understanding within me that I am well aware of regarding myself. I don’t know what happens inside you; I know what happens inside me, because each person knows only himself from within. And therefore there I have no doubt that what is happening here is not what happens in the computer, or at least there is no indication that this is what happens in the computer. It could be that by chance things like that also happen in the computer, but I’m saying at the conceptual level there is no reason to assume that this happens. I built an electrical circuit and it does exactly what I told it to do, so there is no reason to assume that behind it there is semantics; it is only syntax. But in myself I know there is semantics. After all, I know that when I’m asked a question, I don’t just mechanically assemble an answer, but rather I think about the meaning of the question and I choose how to construct the answer to it. So to deny that, you can of course deny reality and remain very consistent, but that’s nonsense. It is obvious that it is not true.

[Speaker B] But again, the materialist—that’s the whole idea of the materialist—he says you have no choice, you have to do it, and there’s nothing more than that.

[Rabbi Michael Abraham] Materialist and determinist—don’t mix them up. You’re talking about a materialist, not a determinist. A materialist. The “I have no choice” part is the determinist. But the point is—I want to say again—it’s not like that. I brought up in one of the previous lessons, when I was talking about materialism and determinism, Professor Yosef Neuman from Tel Aviv University, who passed away some time ago. I spoke with him—he was a materialist—and I spoke with him on the phone after he read my book on evolution, where he is also mentioned. And we had a very interesting discussion. He was a fascinating man, and during it he said to me—and this was even before I wrote The Science of Freedom on free choice—he said, listen, I can’t understand my fellow materialists. They think that human beings have no mental dimensions at all. Meaning, that a human being is only biology. Now I, as a materialist, know they are talking nonsense, because it’s obvious that we have… what, do you deny that you love, want, think, understand, remember? You can’t deny such a thing. We have such mental functions. Except that he, as a materialist, claims that these mental functions do not occur within the human psyche, because in his view there is no such thing as a psyche, but rather they are an emergent result of the material whole. Meaning, the material whole produces these functions; there is no need to posit another substance, yes, another kind of thing existing within us, a psyche or soul—rather, it’s all matter. So an enlightened materialist is not a materialist who denies the existence of thoughts and insights and perceptions or desires or emotions, all the mental dimensions. That’s a stupid materialist. An enlightened materialist is a materialist who says: I understand that all this exists—after all, he experiences it exactly as you and I do—but he says that this does not necessarily require positing that there is a soul in the world. It could be that there is only a body, and the bodily whole produces these phenomena. We have often spoken about this, and also John Searle in that same book, about water molecules: when a collection of molecules succeeds in producing liquidity, something that with respect to each molecule by itself you cannot define whether it is liquid or not liquid. Liquidity is a property of the aggregate. Therefore he says there are certain properties of the material whole that somehow give rise to mental phenomena, and there is no need to posit the existence of a soul for that. That is an enlightened materialist. I disagree with him, but it is certainly a position that can be stated. But the materialist who says no, there’s no such thing at all, emotions are a collection of electrical currents—he’s simply babbling. It may be that the electrical currents produce the emotions, but the emotions themselves are not electrical currents. Just as yellow light is not an electromagnetic wave—I spoke about this. Yellow light is not an electromagnetic wave of such-and-such a wavelength. That is simply nonsense. An electromagnetic field of a certain wavelength hitting the retina produces in us an awareness of yellow light. That’s cause and effect, but it is not identity. Yellow light is not an electromagnetic wave. Yellow light is the sensation in my consciousness that is created when the electromagnetic wave hits my retina. You can’t identify them; that is a mistake.

[Speaker E] Rabbi, Rabbi, excuse me—isn’t it simple to say that when a person is asked a question, today he can answer one way and tomorrow another way? In a computer there’s no such thing. It will always be the same answer; or the same question to two people, each one will answer something completely different.

[Rabbi Michael Abraham] No, that’s not an indication. You can build a computer that gives you a different answer every day; put a random system into it.

[Speaker E] But you’d have to rebuild it.

[Rabbi Michael Abraham] Here, the same per—

[Speaker E] A person, with his mood, with his knowledge…

[Rabbi Michael Abraham] No, no, no, no, no. In a computer you can build it with a random component; it will give you a different answer every time. What’s the problem? The changed answer proves nothing. The question is what changes the answer. Is what changes the answer some kind of random mechanical process, some sort of lottery, or some kind of judgment, of thought, of understanding, such that because your understanding changed, you answer differently? Therefore the change itself proves nothing. The question is what stands at the basis of the change, what creates the change. That’s the important point. Okay.

[Speaker E] But the person is the same person without change, and he gives…

[Rabbi Michael Abraham] What? I didn’t understand. I didn’t understand.

[Speaker E] I’m saying the person is the same person, exactly the same person, and he gives a different answer,

[Rabbi Michael Abraham] He can give different answers.

[Speaker E] The computer can’t; for that you need to change it.

[Rabbi Michael Abraham] If you look inside his brain, then you’ll see that he is not the same person. Every time he gives a different answer, his brain has a slightly different structure. Otherwise the answer would not be different. We operate with our brain. Our brain ultimately generates speech. So if the speech is different, then the brain-state is also different. That’s obvious; even a libertarian dualist like me admits that. That’s not the issue. And the computer too, by the way—if it gives a different answer, then it is a different computer, because electrical currents are flowing in it differently. Meaning, you would see different physical processes inside the computer. Again, in my opinion the change says nothing at all. The question is what generates the change. Okay, up to now I’ve actually spoken more about materialism than about determinism, as I said before—that is, the question whether there is something in us beyond matter. And I’ve already spoken in the past about the connection between the two questions. They are not identical. One can be a materialist and advocate free choice, and one can be a dualist and advocate determinism, in principle. But when you look at the properties of physics, the laws of physics, and so on, then I showed that nevertheless they usually go together. Materialism and determinism usually go together, and vice versa. Now I want to add another thought experiment that will help us examine ourselves not with respect to dualism, to the existence of spirit beyond matter, but with respect to libertarianism. And here I want to speak about Buridan’s donkey. Buridan was the rector of the University of Paris in the fifteenth century, I think, or the fourteenth, something like that, and he dealt quite a bit with what rational thinking is, with rationality. And he built a very famous example that is meant to illustrate that rationality is not always what we think. He basically says the following. He says: let’s say you place a donkey between two troughs. Let’s look at this—I’ll share here, I took this from a caricature in a newspaper. When Congress was deliberating over the Panama claim, whether to dig the Panama Canal, and the question was where—in Nicaragua or in Panama—where to make this canal, there was a caricature in an American newspaper in 1900, which was criticizing the endless indecision of the American Congress about what to do with this canal, and they drew there Buridan’s donkey. You see, this is Congress, the Capitol, yes, and here there is a donkey, and here there is Panama and Nicaragua as two troughs, yes, the donkey’s food. Now the donkey stands in the middle between the two troughs, and stands there. At some point it begins to get hungry and wants to eat. Now it has to decide whether to turn to the trough of Panama or the trough of Nicaragua—yes, let’s call them that according to the drawing—the right trough or the left trough. Now the situation is completely symmetrical. For the sake of the discussion, this is a thought experiment, okay? In a thought experiment I say: these are two completely identical troughs, located completely symmetrically around the donkey, and the donkey too is completely symmetrical. All right? For the sake of discussion, we are setting up here—of course, as I already mentioned, when physicists want to deal with a donkey they begin: let’s start with a point donkey, and see what a point donkey does, and then afterward we’ll see whether it has legs, whether it has a tail, whether it has ears, slowly we’ll build up the problem. But let’s start with the simplest problem, this kind of toy model. Right? So let’s begin: a point donkey sits between two identical point troughs located on either side of it in a completely symmetrical way. Now it is hungry, it needs to decide from which of them to eat. It has no rational way to decide whether to turn to Panama or to Nicaragua, right? They are both identical troughs, they will both satisfy it, and therefore the question is what to do. So Buridan claims that the rational donkey will die of hunger, because the concept of rationality he uses—I called it in the book causal rationality—rationality in his sense means that what you say, what you do, has a good reason why you do it. Okay? There is a good reason why you do it. A rational person or creature does things only if it has a good reason to do them. Now here, if you go to eat from the Panama trough, you won’t be able to present me with a reason why you turned דווקא to the Panama trough and not the Nicaragua trough, so basically that isn’t rational. Why did you turn to Panama and not Nicaragua, or vice versa? Therefore a rational donkey will basically die of hunger. It will be stuck between two identical troughs loaded with all good things, but it will not be able to decide from which of them to eat, and therefore it will die of hunger. What should a donkey that is really rational do—not causally rational but teleologically rational? It should basically draw lots, right? It should draw lots and decide on one of the two troughs, because there is no reason to die of hunger—eat from one of them. You have no reason, but precisely rational thought says: never mind, I’ll do something without a reason because I want to achieve the result, to stay alive. And in that sense this is a much more rational rationality than causal rationality. This is teleological rationality, and it is a rationality that says: I will do the action needed in order to achieve the results I want to achieve; I will not do the action because of the reasons that caused me to do the action. Action out of reasons will kill me here by starvation; action for the sake of results will save me. And basically Buridan tried to present here that rational thought is not always an action that can be justified in terms of reasons; sometimes precisely acting without a sufficient reason is the rational way to act. That was Buridan’s discussion. Okay, I’ll add a few more remarks for people who know this topic a bit. We also spoke about this in the context of chaos. In practice, a non-theoretical donkey, a real donkey, will not die of hunger. Why? First, because it itself is not symmetrical; second, because the two troughs are not at exactly equal distances on both sides—there is always some difference, and therefore that difference can provide a reason why you turn to one of them or the other. It doesn’t matter; that is, the symmetry of the problem is not complete, and therefore the donkey can turn to one with justification. It will turn, say, to the trough closest to it because that is the shortest way. So if there is a difference in the path length to the two troughs, then there is a rational consideration for turning to one of the troughs. But even in a completely hypothetical experiment in which I place the donkey at exactly equal distances from two completely identical troughs, with no difference at all, what will usually happen is what physicists call spontaneous symmetry breaking. Spontaneous symmetry breaking means that some light breeze will come from one side. Something will break the symmetry here, because perfect symmetry does not occur in nature, as the Talmud says: there is no circle in nature. In nature there are no perfect, completely symmetrical forms; in nature there are always deviations from symmetry. And therefore even if you managed to create a picture completely symmetrical between right and left, something will break that symmetry. Some light breeze will come from there, some leaf will suddenly fly by and break the symmetry, and then the donkey will be able to decide where to turn. Yes, think about when we place a little ball—we spoke about this when I spoke about chaos—when we place a little ball on top of a basketball, yes, we place a marble on top of the basketball. In principle the marble is supposed to stand there and not deviate to either side, right? But you understand that this will never happen. It will never happen because I won’t be able to place the marble exactly at the point that has right-left symmetry, although in principle such a point exists. But I won’t be able to place it there; I can’t be that precise. But let’s say I did manage to be that precise and I put the marble exactly at that point. Still, in the grooves of the basketball there is some difference between right and left. Let’s say the basketball is also completely symmetrical—still, there could be some light breeze that suddenly comes from this side or that side; something will always break the symmetry spontaneously without my initiating it. And therefore symmetries are always broken in nature, and this is a very important phenomenon in the context of phase transitions and all sorts of other things in thermodynamics—spontaneous symmetry breaking; there are also spontaneous symmetry breakings in field theory, and physicists deal with them a lot. But I am now dealing with a completely hypothetical experiment: there is no symmetry breaking, the whole thing is perfectly symmetrical. Everything is perfect, perfect symmetry.

[Speaker B] A small remark: in my work, if you run a simulation of this on a computer, you have to put in something that breaks the symmetry, because in a computer it will be perfectly symmetrical.

[Rabbi Michael Abraham] Exactly. So the computer is a good tool for doing thought experiments, because whatever you put into a computer is what’s in it. If you put symmetry into it, it will be symmetric. Of course not on the physical level; on the physical level there are always thermal effects on a computer, and there are fluctuations you didn’t plan for in the computer. But on the software level, on the level of logic, it’s completely symmetric. That’s why they make the one and the zero in a computer differ by, I don’t know, five volts or ten volts, because one volt this way or that can change with temperature, but five volts won’t jump around. So the logic remains sound. In any case, for our purposes, this is Buridan. Now I want to ask: what would happen with a person who was in that situation instead of the donkey? So each of us would say, I assume, that the person certainly would not starve to death. The person would make a random choice and decide on one of the troughs, take one of them, and eat from it; he would not die of hunger. Now notice what that means. Buridan dealt with this thought experiment in order to clarify the concept of rationality. I want to look at it on a completely different plane. I want to look at it as clarifying the question of determinism, not the question of rationality. And here I want to make the following claim. In other words, basically, under the assumption that the donkey—I’ll formulate it this way—a point-like donkey, or a perfectly symmetric donkey, standing in a perfectly symmetric situation, would in my view really starve to death. Truly, that’s what would happen to it. Why? Under the assumption that the donkey has no free will, under the assumption that a donkey is a deterministic system. And why? Because there are theorems in mathematics—yes, Noether’s theorems, symmetry theorems in mathematics—but their basis, as far as I understand it at least, is really some property of mathematical equations, overall a fairly simple property: given a certain mathematical equation, say a differential equation whose solution is a function, yes, not a number—so I have a certain equation, and the solution will have at least the same symmetry as the equation or the problem. The solution to the problem will have at least the same symmetry, or even more, than the symmetry of the problem. That’s a theorem in mathematics. Okay? Now let’s look at this donkey as a physical system. If it’s a physical system, then I can describe the physical laws governing its behavior, and basically those are laws described by some set of equations. Okay? Now if I build a situation that is perfectly symmetric, then mathematically it turns out that the donkey cannot move from its place. This is not a psychological feature of a donkey; it’s not a question of the donkey’s rationality. It’s a logical constraint. In other words, a physical system will starve to death in this situation. And why? Because think now about the situation in the diagram here. It’s symmetric between right and left, and the donkey itself is also completely symmetric. Now, if the donkey moves toward one side, that means the solution—the function that comes out of this problem—does not preserve the symmetry of the problem, because it turned right even though the problem is symmetric between right and left. But that can’t happen. A theorem in mathematics says that can’t happen. If it turns right, that does not preserve the symmetry of the problem; it breaks the symmetry of the problem. Therefore the donkey can either stand in place or move straight forward. It cannot deviate to one side if the problem is symmetric, and therefore it is a mathematical theorem that the donkey will starve to death in such a situation. Now what does that actually mean? It means that if I place a person in the same situation, and I view him as a completely physical creature whose decisions are decisions arrived at by solving mathematical laws, physical laws—in other words, solving differential equations—then a person too would starve to death in this situation. No question about it. In other words, the materialist or the physicalist has to admit that in the hypothetical situation—of course this is hypothetical—in which the symmetry is perfect, both of the person and of the tables loaded with food on both sides of him, on his right and on his left, the person standing in the middle would starve to death. Because the decision he makes is basically an expression of processes that are laws of physics, and in such a situation the laws of physics must yield a solution that is symmetric between right and left. And so people don’t notice this, but Buridan’s donkey is a mathematical phenomenon, not a psychological one. On the mathematical level, a system sitting in such a situation cannot turn right or left. It can move only on the axis of symmetry. That’s all. Or stand still. Standing still is moving at zero speed on the axis of symmetry, so it’s the same thing. So the claim, basically, is that if a human being is a physical system, then he will starve to death in this situation. It’s true that in practice this won’t happen, because in practice there is always something that will break the symmetry. It won’t always be exactly the same on the right and on the left, and maybe there will be some wind, and the person himself is also not completely symmetric. His heart leans a little to the left, and between his right and his left there is no perfect symmetry. By the way, artists have already pointed out that human aesthetics require a certain asymmetry. In other words, a person with perfect right-left symmetry is something unaesthetic. But that’s just a parenthetical remark. In any case, for our purposes, this means that this thought experiment is really a practical implication, or a touchstone, a litmus test, for the question of whether you are a determinist or not. Because the determinist can say to me—and indeed there were people who told me this when I asked them—you’re right, we really do think that a person in such a situation would starve to death. Correct! In the hypothetical case where everything is perfectly symmetric, of course, yes, we really do believe that a person in such a case would starve to death. Because they think that even when a person chooses to go right or left, he doesn’t really choose; rather, there was something that broke the symmetry and caused his calculation to take him right or to take him left. But if there is nothing at all breaking the symmetry, then the person will starve to death in the middle. That’s what they say. Now I wholeheartedly believe that this is wrong. I think that a person in such a situation would make a random choice and turn right or left. But understand that if such a thing happens, it means by definition, necessarily, that the person is not a deterministic system. That the person is not a physical system solving differential equations, but that there is something more in him that can go beyond his physicalism, because otherwise he cannot make such a random choice. Understand that determinism—and I spoke about this in one of the first classes—determinism, just as it contradicts free will, also contradicts randomness. A random draw is also illegitimate under determinism. When you make a random choice and decide whether to go right or go left, the random draw is something that can take you right or left—in other words, it breaks the symmetry. A deterministic view is not willing to accept a random draw either, not only free will. And therefore, basically, the determinist should conclude here that a person in such a situation would starve to death.

[Speaker D] That’s what he should conclude. I didn’t understand—why can’t a rational person define for himself that he knows he’ll die if he doesn’t make a decision? He understands this problem, he’s reflective, and then he’ll decide to make a random choice?

[Rabbi Michael Abraham] But he has no process for making a random choice. Because that random choice, basically, breaks symmetry. How is he ultimately going to move to the right? After all, in order to move to the right he has to activate—

[Speaker B] If instead of a person there were a computer there, the computer would make the decision and say, I know that if I do nothing I’ll die, and therefore it will decide something.

[Rabbi Michael Abraham] First of all, it doesn’t know, okay? It will only make a random choice.

[Speaker B] Okay, use another word.

[Rabbi Michael Abraham] And beyond that, that would only be because the person who built it put into it the possibility of going right even in a symmetric situation, but then that means they put something asymmetric into it, something broke the symmetry. Understand: a physical system—you’re describing things on the plane of cognition, on the plane of thought. I’m looking at this on the plane of mechanics. I’m asking how, mechanically, the person will begin moving to the right. What force will move him to the right when the whole situation is symmetric? There can’t be such a thing. So this is not—I agree with you that cognitively it is completely obvious that the person won’t starve to death and will choose one of the sides. I’m only asking the determinist: how will you explain that? After all, you need to give some physical explanation, and physically there can be no explanation for it. Now, the problem with thought experiments is that thought experiments won’t convince anyone. In other words, if the materialist insists, he’ll say: correct, a person in such a situation would starve to death, and I haven’t proved anything to him. Therefore I define the— I explained what a thought experiment is. A thought experiment’s role is to help you make decisions—in other words, to discover what you yourself think. So now each of you has to examine himself. If you think that in this hypothetical situation the person really would starve to death, no problem, I can’t tell you anything. But if you too, like me, are convinced that the person would ultimately turn to one of the sides, then that means you are not a physicalist. Fine. So this is the thought experiment of free choice; with that I’ve more or less finished the move. Next time I’ll do—next time I’ll finish the— I said this would take me one or two more times. I’ll finish the move of the book, and then maybe one more time we’ll talk about the meaning of choice, the spiritual meaning of choice, the Torah-based spiritual meaning of choice, and that’s it, and we’ll finish this series, which really has become long. That’s it, up to here.

[Speaker C] Long and enjoyable.

[Rabbi Michael Abraham] Okay, I hope so. Goodbye, have a peaceful Sabbath, happy Hanukkah.

[Speaker E] More power to you.