The Peculiar Logic of Value | Ray Jackendoff | Talks at Google

Some of you may have followed the rise of the discipline of linguistics during the last half of the 20th century. Noam Chomsky's works, Syntactic Structures and Aspects of the Theory of Syntax ushered in a golden age of language investigation. In Jackendoff's words, for the first time ever it was possible to provide detailed descriptions of the syntax of natural languages. The following question naturally arose, given that even linguists who work hard to describe a natural language can't even get it right. How on earth does a child come to know such a complex system in a short amount of time? This question, and others akin to it, brought the field within the sphere of those interested in the study of the mind and the brain. Exciting stuff. But then, unfortunately, there was the founding of two houses, the Montagues and the Capulets. Well, there were many more houses than that. Psychologists and philosophers were, to use a vernacular, turned off by the squabbling linguists. In addition to their internal strife, linguists seemed to assume that linguistic institutions were sufficient to test. Hit it again. That linguistic intuitions were sufficient to test psychological reality of their theoretical claims. This didn't win votes of confidence from researchers who like to measure things. Where does this leave us linguists who feel we really do have something to offer researchers interested in the mind brain? Do we have anything to contribute? Ray Jackendoff is here to answer in the affirmative. Jackendoff is a professor of philosophy, Seth Marin Chair in the Humanities, and co director of the center for Cognitive Studies at Tufts University. And he is a professional musician. He has authored many books, among them Patterns in the Mind, Foundations of Language, and most recently Language Consciousness Culture, Essays on Mental Structures. And now, if I can turn the page, he is here today to talk about the peculiar logic of value. And he starts with the question, how do humans conceptualize values? Please join me in welcoming Ray Jackendoff.

Thank you, Ann. It's an adventure to be here and see how this half of the world lives. The stuff I'm going to talk about today really isn't. I don't think linguists would consider it linguistics. I don't know what anybody would consider it. But it's growing out of work that I've done for years on the semantics of natural language, trying to figure out how language hooks into the human conceptual system. And to some extent, I feel like I understand it well enough now that we can break away from some of the shackles of syntax and look at meaning at concepts more closely and in more detail. And so this. And here is. So the talk today is Coming from this book, language, consciousness, culture. I'm having as much trouble talking as you do. And the enterprise here is the study in a good part of the book is the study of social cognition, which is how humans understand interactions among people. Not just me to you, but you to you in the context of social institutions. And part of my methodology is coming from linguistics. I'm looking at expressions of social understanding among which are expressions of value. But at the same time, I'm trying to take evidence from anthropology and psychology and so forth seriously. If we're going to get at the conceptual system, we have to use every possible tool at our disposal and probably make up a few that nobody's thought of yet. Now, it's a pervasive element of social cognition. As soon as you start analyzing social concepts, you get to notions of value right away. And of course, these are fundamental to social understanding, as we know from, say, current political discourse. And the question is, what is a system of values? What does it mean to have values? And the way it's often asked is, what are values in the real world? This is the way moral philosophers will ask, what are moral values? As though that's something really out there. And I'm sort of sidestepping that question and asking rather, how do humans conceptualize values? That is, things are of value because we take them to be of value. So part of my methodology will be to use linguistic expressions of value and intuitions about inferences among sentences as clues for the organization of value concepts. So let me give you my general hypothesis first, and then we'll get to specifics. So the idea is that value is an abstract property which we conceptualize and attribute to our conceptions of objects, persons, and actions. And you'll see what that means in a minute. What I mean by abstract is that it's not perceptible. You can't see the value. But you know, something has value. And the value of an entity, an object or an action or a person plays a role in various rules of inference that affect the way we reason about these entities. Now, what do you do with abstract properties? I mean, how do you reason about them? How do you use them? Just having abstract properties that you can't perceive, how do you know to assign them? Well, what you need is you have to ground out this system, this abstract system, in concrete entities. Well, how do you do that? Well, there are three basic elements, I think, in a system of values. One is principles that say, if such and such is the case in the concrete world, assign this value to it. So I'll call those input rules. There are rules where the value is on the right hand side of the rule, right hand side of the inference. So you start assigning values to things. So what, how does that affect your action? Well, you also need rules that say here's what you do if such and such has a value, or here's how something having a value will affect how you interact with it. So I'll call those output rules. Those are rules that have values on the left hand side of the rule, but concrete actions on the right hand side of the rule. And finally, we need principles inside the system that say if this value is the case, then assign this other value to something else maybe or to the same thing. So ways that allow you to calculate within the value system. So you have input rules that get you into the system, you do these inferences within it, and then the output of that can affect your action. Now that may seem a little kind of abstract. The way I put it, a concrete way to think about it is like points in a game. So if you think about any old game, let's say it's soccer or something, an input rule is if you kick the ball into the goal, you get a point. And the point is an abstract property that's just sitting there, that belongs, the team has a point. What is it? A point, Just a point. The output rule is the team with the most points at the end wins. So winning is the concrete output, sort of concrete anyway. And the inference rules within the system are, you add up all the points to get the points at the end. So the inference rules just have to do with adding points, and they don't have to do with what's going on on the field. Now, values are more complex than points, but it's the same sort of idea. You're keeping track of these abstract entities. But values are not just sort of integers. They have a valence. Things are good or bad, and they can have an analog magnitude. So things can be better or worse. In addition, as we'll see, there are multiple types of value and some of them are affective value and resource value and so forth. I'll go through eight different kinds of value pretty soon. And so there are different kinds of value that have not been, to my knowledge, recognized very well in the literature. We'll see. They get all mixed up. And finally there is a further dimension. Sometimes values are taken to be objective, so such and such has a value. And sometimes they're taken to be subjective. Such and such has a value to you or to you or to me. So it's a much more complicated abstract system than keeping track of just points. Okay, here are the eight kinds of value and I'll go through them one by one. So that's just the list. Affective value means it feels good, utility is good for someone, benefit or cost, resource value is good to have, and so on. So let's start with affective value. Now the thing that's interesting is if you look at an anthology from symposia, we're having cross disciplinary discussions of value. The, the papers go past each other completely because these different kinds of value are studied by different kinds of people and they all call them value, but they don't have much to do with each other. So affective value is things like pleasure or pain. It feels good, it feels bad. So a tasty meal has positive affect value. An aesthetic experience has positive affective value. Being hungry, being in pain, being in danger. Those of negative affective value. And so these are the kinds of things that are studied by biologists looking at animal behavior and psychologists. This is Freud's lust and unlust perhaps. And so these belong to primitive biological sorts of things. So here's some expressions of affective value. Fred enjoys eating figs, or Joe dislikes being beaten up. These are subjective values, these are affects on particular people. Or we can do it objectively in saying eating figs is enjoyable. Being beaten up is unpleasant. Well, there's nothing can be enjoyable unless there's enjoyer. So who's the enjoyer in this case? Well, I think what this means is for anyone, for people. So there's this sort of hidden, I don't know if you can see it on the screen here. Eating figs is enjoyable to people or eating figs is enjoyable to ya, to y'. All. Something like that. So the objective one has this sort of hidden generic character that means anybody will enjoy it. Now notice this is still not. Even though it's stated as though it's objective. Figs are enjoyable. It's not value in the world independent of observers. It's really talking about a generic person observing it. And although but it says it doesn't matter who you are, it doesn't matter what you think, you will enjoy it. So it doesn't depend on a theory of other people's minds, what other people are thinking or feeling. It just says you'll enjoy it. Judgments of subjective value. You enjoy it and you don't depend on understanding the minds of different people. Theory of mind. What's the next one? Utility so utility is one that economists study a lot, to some extent animal ethologists, but biologists don't in general study this one. And the idea is calculating benefits or costs of things you're about to do, let's say. So here's some expressions of this. Eating this broccoli will be good for Bill, so it's good for, is the typical expression here. Or being overweight is bad for. That's the subjective version. The objective version is eating broccoli is good, or eating broccoli is good for you, being overweight is bad, or being overweight is bad for you. So it means for anybody, it doesn't matter who it is. And now with just this much, I can start to give you a feel for the overall character of the system. So an input rule says if event affects somebody positively or negatively, it's of positive or negative value to them, either affective or utility, depending on the event. So it might be an experience. If you have a delicious meal or you have a disease, that would be something of affective value, positive or negative. Or it could be an action of some on somebody else's part. If somebody helps you, that's a positive value to you. And. And if somebody victimizes you, that's obviously of negative value to you, but so what? So it has values. This has to affect your behavior somehow to be of any use to you. So here's an output rule which is faced with a choice between two or more actions to take, do the one with the highest value to you, do what's better for you. And this is sort of a summing up of rational choice theory supplemented by heuristics along the lines of Kahneman and Tversky, or fast and frugal strategies along the lines of giga rents, or if these mean anything to you, if they don't, it doesn't matter. It's basically, there are a lot of people studying how you decide what's better for you and act on the basis of that. The subject aversion is saying, not just people do what's best for them. But faced with a choice between two actions, you in particular will do the one from my point of view, you'll do the one that you take to be best for you. So it's putting it in the mind of the observer, not just sort of the absolute values of these values as they are. So we'll get some more principles later. But you can see, given a choice between two things, with, you know, a tasty meal or weeding the garden, you'll go for the tasty meal. Except if weeding the garden has higher utility, in which case you might for whatever reason, but then you start calculating those values. Ok, we're going to get many, many, many more principles later, but I want to do some more kinds of value. So the next one is resource value. Resource value is something is good to have or is valuable. And this is one that's most pertinent to economics. So what does resource value mean? It's an object is good to have maybe because it has potential for an event with affective value. So food is good to have because it feels good to eat it. A house is good to have because it either feels good or is of utility to live in it. But there are other reasons things might be good to have. It could add to the esteem of your owner. And as we'll see, esteem is an important value. So something you want. So having a fancy painting might give you esteem, so it's a resource value. Or having a fancy car might bring you esteem or ladies or something like that, if that's what you want. So it's good to have a fancy car, that's why people want them. Or finally, something might have resource value because you can trade it for either another resource valued object or, or something with affective or utility value. And that's what money is. So it has resource value because of its potential for exchange. So what are the typical expressions? Objective, this piece of land is very valuable or it's worth a lot. Subjective. This piece of land is very valuable to Harry and this distinction is very important. My wife sells real estate and she always finds that the owners of the houses think their house is worth a lot more than anybody else does. Well, it's such a special house. Well, it's special to you, but not to this guy who wants to come and buy it. And that makes a huge difference in how much people are willing to pay. It's very hard to convince people this is what the market says the house is worth, not what you say the house is worth. So this distinction really plays a big role in our exchanges. Move right along. Another one is I'll do two in a row. Quality. So something like a good search engine and prowess, you're good at doing things. So quality means quality relative to other objects or events of the same type. So something like this is a good or a terrible computer or that was an excellent or a miserable back dive. Not too much to say beyond that about quality, but prowess is quality of your performance. So Harry is good at singing or Harry is a good singer and good at notice rather than good for which we had with utility. Now comes a really important one, which is normative value, which has to do with conformance to social norms. And there's a whole variety of these, as we'll see. And when you get to normative value, you're taught you're looking at people like ethicists and moral philosophers and anthropologists who look at this and don't look at economics and don't look at affective value. So the subdomains include things like moral and ethical value, religious value, legal codes, and then valuations according to standards of etiquette, like manners and politeness and custom and so on. Now, I want to make this breakdown partly because in the literature and in common life, people often confuse normative value with other kinds of value. If I may twit, one of the greats, Plato, is always going around asking, what is the good? What is virtue? And then he says, well, here's how we find out. What makes someone a good doctor? Well, he heals people. What makes someone a good pilot? Well, he gets ship safely into the harbor. What makes someone a good harpist? He pleases people's ears when he makes his music, and so on. So what's a good person? A good person is a person who does what he's supposed to do well, and what is that? And now he gets all tangled up and he ends up saying, well, I don't know what the good is. And you didn't know. You didn't know what it was. But now you know now, too. And that's what wisdom is about. And you know, this is the way Plato carries on, or, you know, Socrates in Plato's words. So I think the problem is what makes somebody a good doctor is that's a prowess kind of value, not a normative kind of value. What makes someone a good person is virtue. That's normative value, not prowess. So he's making a category error. I think I would like to know if somebody knows ancient Greek, whether the word is good for both kinds of value in Ancient Greek as well as in English. Another case is if you think about Cinderella, Cinderella is beautiful. That's quality. And virtuous, that's normative value. The stepsisters are ugly. That's quality and mean, that's negative normative value. And those are stereotypically aligned in all the fairytales, in all the novels, in all the literature and everything like that. And this is. Some people have called this Schiller's thesis. Evidently Friedrich Schiller actually thought those are really aligned. That beauty, which is a Q value, quality and normative value, virtue, moral value, are the same thing. Well, okay, too bad for Shiller then. But even. Okay, let's look inside normative value too. I mentioned all these different subdomains, ethical and legal and religious and so forth. And those don't always align. So you can have highly moral people with bad manners, and you can also have highly immoral people with good manners. Those are more worth watching out for. So morality and manners don't align. Moral code and religious codes don't always align. So think about desert traditions of hospitality, which I think are moral codes, but I don't think they have much to do with any particular religion. And religious codes for how you perform rituals don't necessarily concern moral or ethical domains. So to say, here's the religious code. You're supposed to take this thing and wave it this way and this way and this way and this way. That doesn't have anything to do with ethics. That's just the way you do it. But it's a religious code and you can have moral and legal codes mismatching. So, so stereotypical cases, the evil landlord in the melodrama who's foreclosing on the poor widow within his legal rights. So he's obeying the legal code, but we certainly consider him immoral. And in the other direction, we have civil disobedience along the lines of Martin Luther King or Gandhi, where they're doing something that is illegal but taken to reflect a higher morality. So we want to make the distinctions among all these domains in normative value. I think it's very important to do so. On the other hand, it's important also to keep them together. There is a distinction made by moral philosophers and social psychologists between morality and everything else. And they say we're interested in morality and the rest. The social conventions, well, they're not so important, okay? And so morality is this universal thing. It's timeless. Everybody will recognize it. And social conventions, well, you can alter them. Which side of the plate you put the fork on, who cares? This is the way we do it. But you could do it the other way. But killing people, no, you couldn't start killing people and change your morality that way. Well, the problem is that they're not as what's taken to be morality is also culture dependent. So if you think about sexual mores, those are generally taken to be morale, morality, issues of morality, at least in many cultures, including this one, but they are certainly culturally variable. The same thing with the status of slavery, which we take to be deeply immoral. But in many cultures, including that of, let's say the Old Testament, slavery was okay as long as you treat your slaves sort of decently. Okay? So that's taken to be. Those are cultural variables that are taken by some cultures to be part of morality. And in fact, social conventions or moral and cultural values are often intertwined. So if you look at the Ten Commandments, right, The sixth commandment is don't murder. The fourth commandment is keep the Sabbath. Well, you'd say, don't murder. That's morality. Keep the Sabbath. Well, that's a cultural convention. But if you keep reading on a little bit in the Book of Exodus, you find that the penalty for violating either of them is the same, namely, you get put to death. So in the world of the Old Testament, keeping the Sabbath and not killing people have exactly the same status. So even from our point of view, one is social convention, one is morality. They're really part of the same world there. So you can't sort of extract the moral of this is. You can't extract the universal, timeless morality from all the rest of the culture. You have to look at them together. Okay, so let's look at some expressions of normative value. This is one that focuses on the action. It's good of Harry. Notice of this time, good of Harry to wash the dishes. Taking cookies without asking is bad of Harry. We can also turn it around grammatically and say, Harry is good to wash the dishes without being asked. Okay, so we can focus on the person. But notice now, normative value can only be ascribed when there's responsibility. So there's a difference between normative and utility value. Washing the dishes is good of, but being overweight isn't good of Harry sounds funny because Harry can't be responsible for it. Well, I have known people who did conflate these and thought that things that you couldn't help that were of bad utility were your fault. So being poor was a moral failing and being rich was a moral failing. All kinds of things this happens. Those are other kinds of conflations that happen. But we do say being overweight isn't good for Harry. That's utility value. We can also bleach out the actor and just saying, it's good to wash the dishes without being asked. It's good of who? Well, it's good of people. It's good of you if you wash the dishes without being asked or it's bad to kill people. It's really, it's bad of people to kill people. So again, we have this object diversion of normative value. Now we're at a point where we can talk about some important interactions between the kinds of value. These are inference rules internal to the system. Basically, the idea is if I act in a way that is bad for you, affectively or utility bad for you, then that action is normatively good or bad of me. So an action that's affectively bad or good for someone else is a normatively good or bad action for the actor. Okay? It's good to be nice to people and bad to be mean to people. Normative good this time. Okay. And there's another interaction that's kind of interesting, which is if you do something of, let's say, good normative value, if I do a good deed, that may make me feel good. If I do a bad deed, that may make me feel bad, okay? So this is an interaction from my normative goodness to feeling affectively good or bad in the other case. And we call that conscience or guilt. Let's see where I'm. Okay, so this is maybe being. It's sort of being good for its own sake. It feels good to do good, but not exactly, because there are also the possibility of doing good out of duty, which might feel bad. So there are other reasons to do it as well. Now, this rule, actually the first of these rules, is the one that. That has been focused the most on in the recent work on intuitions of morality. If you've seen Mark Hauser's book, what's it called, Moral Minds, he has all of these moral dilemmas that people face that have to do with what do you do if it involves hurting people? And that would be bad. How do you balance off killing one person in order to save five, and under what conditions? And there's a huge literature on this now, and it's all about this one rule up here now. But it isn't the case that all normatively good things have to help somebody else. So religious rules can be normatively good, but affectively neutral. Illicit sex is affectively good. It feels great, but it's normatively bad. That's why we call it illicit. Right? So, okay, so that's normative value. A little bit about that. And that in turn feeds into personal normative value, making you a good person. So now we get to the starting to get to the heart of virtue. So if we say Harry is good, what does that mean? It means he does normatively good things generically. Harry is a good person. Subjective normative value would be to Joe. Harry's a good guy, but not to me. I don't think his actions are as great as Joe thinks they are. Well, how do you acquire personal normative value then? Well, by doing normatively good things or normatively bad things. And the problem is now we get to a huge, what do I want to say, computational problem. Because normative values aren't mathematically additive. Think about these two scenarios. One is, even though what you just did was horrible, I'm not going to hold it against you. So that's saying you did something of bad normative value, but I'm going to keep your personal normative value positive. But here's the very same scenario. That one horrible thing you just did has wiped out my whole good opinion of you. So that's a normatively bad thing that changes personal normative value, changes the valence. Poof. Just like that. It's the same scenario. And who knows what that depends on? So I don't know how to calculate that in general, but that seems to be the way we work, which is why I call this the peculiar logic of value. The last one, the last one is esteem or prestige or respect or good reputation. And this seems to be a composite of a bunch of other values. One is your personal norm of value. How good a person you are, how good you are at things, your prowess, your status in the dominance hierarchy, whether you're a leader or a follower. It might depend on your wealth, your accumulation of R value, it might depend on your attractiveness, your Q value, and so on. And they add up in some again way that I don't know how to specify. So we have expressions like Harry is prestigious or well respected. Harry has a good reputation. That's the object aversion. Subject aversion is Joe respects Harry or Joe holds Harry in esteem. Now this is interesting because it interacts with your group membership, which is a whole social thing that I haven't talked about today. But it's very fundamental. And the group membership could be ethnic groups or professional groups or whatever kind of groups you want a clique in high school for that matter. And the idea is, we're great, you stink. Okay? That means my group gets, members of my group get more respect than members of other groups. But in addition, that adds up with also, I'm in the low status group, we're great, you stink. You know the high up groups, well, you still show them respect even though, you know. So there's this sort of mixed effect. The two are in conflict with Each other. Whereas the high status groups say we're great. You really stink down there. Okay, so we have these two different factors involved. And in addition, there are exceptions for wealth and prowess. So somebody of a really low status group may get a lot of respect because he's very dominant or very good at something. So a good example is Al Capone, right, Who is certainly a member of a low status group, but held in a great deal of respect. The term usually used is notorious. What does notorious mean? It means great deal of respect, high esteem, low personal normative value. That's what notorious means. I think. So that's esteem. And now we get to the big problem for philosophy and evolutionary psychology, which is why be good. And Socrates is thinking about this, right? Well, being good is achieving a high personal normative value, which in turn makes a contribution to your esteem, where esteem is a function of all kinds of things, but the personal normative value is one of the ones you can do something about. So what? So you have high esteem. So. Well, if you have high esteem, I really want to hang out with you and I really want to do things with you and I really want to make projects where we can both make out pretty well. Okay? So if you have high esteem, you're in a position to get a lot more goods because people want you to work with them and cooperate with them. So enhanced esteem is a sort of resource value. Respect is a good thing to have. Now you might say, well, that's really cynical. The only reason you want to be a good person is so that people will be nice to you and you can then profit thereby. Well, if you start reading the literature on evolutionary psychology these days, that's the way it seems to come out. They say, why would we even worry about normative systems? Why would we evolve those? Well, this is one of the reasons. So I don't know whether it's cynical or realistic. And there may be other reasons as well, but they're ones that, well, you need to work them out. Also, there may be more than one reason, but notice this is all now calculations, internal the system. The inference rule then is if you respect somebody a lot, then it's good utility, good to associate with them. So that's another rule inside the system. Now there are all kinds of. So far the only ways I've gotten into the system in terms of creating values are something feels good or it's good for you. But there are lots of other rules that are sort of culture bound. So here are some that get you into the normative system. Right away. It's normatively bad to dress such and such a way. It's normally bad to eat such and such, normally bad to show attraction to people of the same sex and so on and so forth. There are millions of these and we have to learn them as part of our culture or fight against them if we don't like them and you know, form a counterculture or something like that. So there are huge numbers of these and these are part of learning your own particular culture. I'm going to skip a whole section here and although it's important, I think, but this stuff is kind of more fascinating. So I want to talk about a couple of other systems, social systems that invoke the value system in their working out. So let's look at fairness. So I think the basic idea behind fairness is that an action is fair if it's equally good or bad for everyone. Well, so we can state that within the value sort of calculus to say Y acts fairly towards a bunch of people, if for each one the value of Y's act to that person equals the value of the act to the other person. Okay, that's just sort of making it a little more formal looking, but it's saying the same thing. Now, there are a lot of variance in this is why doing something that impinges on everybody at once, let's say distributing a piece of meat that I brought in from the hunt or something like that to make it sound evolutionary, or I might be doing the same thing for each person on succession. So now's your turn. For your birthday present you get this, and for your birthday present you get a present of equal value. And for your birthday you get a present of equal value and so forth. That also conforms to this. If we take the giving of birthday presents sort of collectively as wise action, this could be beneficial. I could be giving out things. I could be giving out meat or birthday presents, or I could be creating burdens. That is, it could be negative value here rather than positive value. So it could be assigning tasks. You have to wash the dishes and you have to sweep the floor. No, that's not fair. Sweeping the floor is harder. So it could be assigning tasks or it could be levying taxes. And in fact we can get a little more sophisticated and say, well, is this calculated objectively, that is equal sub actions by Y or subjectively, that is taking into account the different needs and abilities of the X's. So a simple case would be a flat tax, which is just saying equal objectively, and a progressive tax which is trying to at least simulate the fact that people's abilities to give it will hurt. So making the tax hurt each person about the same amount. That is giving, making poor people pay less tax and making rich people pay more and more and more tax, which seems to be dropping out of our cultural values in our society these days anyway. So that's fairness. Now, there's an alternative to fairness that we also see all over the place, which is distribution by rank. So rank is some inequality either in personal normative value or in prowess or in esteem or in dominance or in wealth or whatever. And the idea is, if you're the winner in first place, you get a bigger prize than if you're in second place, you get the second prize. It isn't like, was it WC Field, two other contests. First prize is a week in Philadelphia, and second prize is two weeks in Philadelphia. Right. Well, the implication of that is two weeks in Philadelphia is worse than one week in Philadelphia. Right. So this is rank based distribution. So if you're of high rank, you get more resources and interestingly, get less punishment. That's very important. Which we also see in terms of rank based distribution in today's society. Not more punishment. If you're higher up, less punishment. And in general, high rank gets a better house, better pay, better mate, less difficult work, and less stringent punishment. And this is broadly applicable and present in social animals like chimpanzees and I think, wolves as well. They're things like rank based distribution. Well, now you say, well, there are two different ways of distributing things. Right. How do we decide on any occasion whether to do it by fairness or by ranking. And Alan Fiske, an anthropologist, has proposed that there are culture dependent normative principles that say, here's, you know, under these circumstances, you use fairness. Under these circumstances, you use rank. And you have to learn in each society which circumstances go in which list. He actually proposes four different ways of which these are two. So that's something else you have to learn when you're learning your culture. Now, I want to do a slightly more complex subsystem that has to do with reciprocity. How am I doing on time? Are we all right? Yeah. Okay. I think I'm good. Yeah. I'll try to speed it up. Okay, so this is action in return for an action. So Fred cooked Lois dinner for fixing his computer, or negative version, for insulting his sister. He slashed her tires. Right. And this turns out to involve value, affective or utility value, I think. And they have to match in valence. So we can't say. Fred cooked Lois dinner for insulting his sister, which seems to imply that cooking Lois dinner was either he wanted her to insult his sister or something. Right. Fred slashed Lois tires for fixing his computer. He really was averse to using that computer. That's what it suggests. And they have to match in magnitude. Fred cooked Lois dinner for saying hello to him. He's really overreacting. Or Fred cooked Lois dinner for rescuing all his relatives from certain death. He's sort of not doing enough to make up for it. And the same with the negative case. The overall principle is that the values and the valence have to match up pretty much. Now, the positive version of this is reciprocal altruism. I'll scratch your back because you scratched mine, or one good turn deserves another. The negative version is reciprocal retribution or retaliation. The punishment fits the crime. Right. But there's a third thing that goes with these, which is restitution. So if I do something bad to you, I might do something good to make up for it. So Fred cooked Lois dinner to make up for having embarrassed her in public. And again, the values have to match in magnitude. Fred gave Lois his vast fortune to make up for forgetting her birthday or brought Lois flowers to make up for killing her whole family. They just don't cut it. So the overall principle here is in restitution, again, the magnitudes have to add up, but with opposite valence. And notice this one is not neutral for valence, like the reciprocity rule. There is no if I do something good to you, I have to do something bad to make up for it. So Fred slashed Lois's tires to make up for having cooked their dinner. That just doesn't cut it again. And I think that, you know, that's a. Such an obvious inference. But where do we get that system from? That's a question. So the overall principle. So the overall normative principles are, then it's good if somebody. If you do something good for me, it's normatively good for me to reciprocate to do something good back to you. If I do something bad to you, it's normatively good for me to make up for it by doing something good for you. What's missing is the negative case. So if I do something of negative value to you, then. Well, turns out there are three possibilities and they're culturally determined. One is it's a good thing for you to retaliate. Another one is it's okay for you to retaliate. Nobody is going to stop you. And the third one is it's bad for you to retaliate, that's turned the other cheek. So we're seeing all these little aphorisms cropping up in this logic, which I kind of like. Now. It turns out there's another system that has exactly these same properties which has to do with offering expressions of respect. So Joe praised Sue for saving the drowning child, or sue scolded Bill for his bad manners. Okay? Now it's offering expression of esteem for doing something of normative value or showing prowess or something like that. Okay? And you might say, well, what happens if I do something bad to you? Can I make an expression of esteem? Yes, I can say my esteem is lower, my self esteem is lower. And that's what an apology is. I'm sorry, I'm a schmuck. Right? So Bill apologized to sue for insulting her. And there's the for again, that's showing its reciprocity. So the general normative principles are it's good to express respect for somebody having done something good. And it's. It's good to apologize for things that you've done that are bad.

But.

But now what do you do if somebody does something bad? Is it good to scold them? Or is it good to just sort of let it slip? And again, that's culturally dependent, and it's a very delicate issue. Okay, so let me sum up the idea that a value system, as I've talked about it, is a multidimensional, abstract calculating system that helps us govern our actions. And the dimensions include these eight kinds of values and the subtypes, especially in normative value, plus the dimension of objective versus subjective and the values assigned to individuals. Their prowess, their quality, their personal normative value, their esteem are crucial in deciding how to interact with them. And that learning a culture is in part acquiring a whole lot of special rules for assigning values to various actions and people. And what I hope I've shown you is the complexity of our value judgments and our reasoning involving value, and also the connection between the value system and the linguistic expression of value system. There are these huge ambiguities in good, which you can then pick up by these uses of prepositions like add and of and for that help resolve them. And this sort of work is parallel to the work that's gone on for years and years and years in looking at spatial language, which I can't even sum up here. And now we can ask. We can, I hope, sharpen some of the questions that have been around. Like children's acquisition of values. There's A literature on this going back to Piaget. How do children learn what's good and what's bad, what's fair and what's not? And there are people working on it since then. To what extent is a particular value system in a culture going to benefit the culture as a whole? Unlike language, where language is pretty much neutral, any old language would do in any culture within bounds. It might be the case that certain sets of value systems set a culture on sort of a disaster course. And they might be easily wiped out, or they might wipe themselves out, or we might wonder if we're collectively on that today. I don't know. So are value systems culturally contingent? How do they change over historical time? And finally the evolutionary question, how much of this value system is innate? That is the potential for different kinds of value and which particular rules might be universal across humans? And also how much is unique to humans and how much of it do we share with other social mammals, let's say. And we can see fragments of this, I think, in the chimpanzee societies. So what I hope I've talked about today helps sharpen all these lines of inquiry. And so I'll stop here, thank you very much.

Are there any questions? Okay, can you pass this back to this guy back there?

Well, the simple question is, so do people ever do good for the sake of doing good? Because you didn't seem to leave much room for that. You seem to suggest that when people do good, it's for the sake of esteem from others.

Yeah, I mean, here is the argument that the evolutionary psychologists have made. They said if a gene developed in a social organism that said that sort of made them prone to doing more altruistic deeds, those individuals would be, it's always at a cost to do something good for someone else as opposed to for you. So those individuals would not be around very long. They wouldn't leave as many offspring on average as the people who got helped. So you are prone to cheating and being exploited if you're a do gooder. And so the question that is being asked is, is there any reason why such a gene should develop? And one of the reasons that seems to be argued is you want to. Well, it's quite complicated. Let's see if I can run through the argument. You want altruism, you want a society that rewards altruism and punishes non cooperation. And the punishment has to be inflicted by anyone, not just the person who is not cooperated with. And this will have some effects on promoting the survival of that particular group as opposed to other groups. Now there's a whole dispute about group selection that that feeds into that I can't get to. So the question is why do you need a lot of cooperation in groups? You know, why is that a plus for groups? And a guy named Richard Alexander has argued that in early hominid evolution, basically the main predator on humans was other humans. So it paid to be part of a big group because you could defend yourself against other groups better that way. And that's why there was a selection for greater cooperation. So it's a long convoluted story. Now there's another piece of this. Let's see if I can reconstruct it. People argue that it is probably better for the self presentation of an altruist if they look like they're doing it altruistically. And they will rather than selfishly. And they will look more altruistic if they think they're being altruistic. So this would provide a sort of an impetus towards self deception. Now I know this all sounds weird, but this is the kind of dispute that's going on now. How could we evolve being good just for the sake of being good or the feeling that that's a positive? Yeah.

Well, I'm going to come back just briefly. What it seems that can't account for is many of our most say notorious do gooders over history, People like William Wilberforce and the abolition of slavery, or Gandhi in India, or Martin Luther King in this country. And you can find many examples of people who have believed very firmly that they were doing good even though the society around them persecuted them for it. How can you account for that?

I can't. And I think that my answer to you is if we can find that reason, it can be stated within this system. Okay. My goal at this point was not to find the reasons, but to state the reasons that people have found and justified and find all these sort of common sense observations in the linkages. So I don't want to take on the burden of the evolutionary psychologists. I've just told you the story I'm they're telling, but I don't. I would love to believe what you would like to believe in too. And everybody would. And yet nobody can find a reason why we would have evolved such a trait yet. I'm hoping we will.

So I just have a follow up comment on the evolutionary psychology. I'm not one obviously, but I think one of the defenses they make about this, doing something honestly out of good, I think that can be answered a little bit by Considering who carries out this calculus, is it carried out in our brain or was it carried out by evolution? Right. Maybe it was carried out most of it by evolution. And then in our brain, we don't carry out this calculus. It's just there, sort of in hardware, to use our language.

I don't know, but I'm not quite sure I understand the comment. I mean, it's in our brain as a possibility. And now the question is, how did it get there? Is it an accident that it got there? There are lots of evolutionary accidents, but the evolutionary psychologists are saying that would be selected against. Well, maybe it's interesting that such people are extremely rare. If you start looking at how the evolutionary psychology works, it's always sort of. You're looking at population distributions of a gene. So it might be that the proper distribution has 1% of such people or something like that. I don't know. There was a question from New York. Can you hear me? Yes. Okay.

I'm not a Platonist and I'm not a classicist, rump to rump to slavery to slaver, but I.

I do think.

That when Plato talked about virtue, he was talking about P value and possibly Q value as well.

He was trying to develop a general.

Theory of Q value, you think?

Well, in that case, I would say that if he's trying to develop a theory of Q value, he wasn't really observing the distinction between, let's say, prowess at. At a profession and normative value or personal normative value.

I don't think he's talking about normative value at all. The word he uses, arete, definitely means excellence, you know? Well, the arete of a stove is to cook. Well, that's Q value.

Right. So it's like, good in English, it's polysimus. I think.

When I was talking with an Italian last week that came up against this because he used the word virtue, he was talking in English, and I had to explain that virtue is not virtu. Virtu is excellence. Okay, of course, that's exactly the same point.

Okay, yes. So I guess my argument in that case is the word is actually polysemus. And we see it in these sort of secondary grammatical properties, like good of versus good at. And so his argument was sort of, in a sense, a play on words. An unintended play on words. Yeah, something like that.

Okay, one last question. Oh, sorry.

Are we. Okay.

So in our business, there is partially a wisecrack, partially a rule of thumb. If you have a function with eight parameters, you've forgotten one So I have the same question with your eight kinds of value. Do you believe that there are eight and only eight and do you have a logical argument for that or are these just eight kinds that you've noticed?

There are eight kinds that I've noticed. I had six in the first draft of the paper and somebody started giving me grief about something and I realized that I had to split A and U into 2. Although they behave very much the same in the logic, they have sort of different computational properties I think in how you assign them. So you know it. And the way I've divided up normative value into four or so subdomains with the etiquette one having three subdomains, that's a little bit arbitrary and certainly subject to further research. I wouldn't say there's any essentialness. Say if we go on doing this intensively for 50 years and we find three more and no more than that, that's probably pretty good.