Portal de Periódicos da CAPES

“You can be a behaviorist and still talk about the mind – as long as you don't put it into a person's head”: An interview with Howard Rachlin 1

2022; Wiley; Volume: 119; Issue: 1 Linguagem: Inglês

10.1002/jeab.782

1938-3711

Carsta Simon,

Mental Health Research Topics

Dr. Howard Rachlin was Distinguished Research Professor of Psychology at Stony Brook University. He published more than 150 peer-reviewed papers, authored seven books and edited three. Before taking up his Ph.D. studies in B. F. Skinner's and R. J. Herrnstein's lab at Harvard, he earned a B.A. in Mechanical Engineering at The Cooper Union in New York and a master's degree in psychology from The New School for Social Research, New York City, a center of Gestalt psychology at that time. Dr. Rachlin studied patterning of choice behavior over time with the aim of understanding the psychological and economic basis of self-control, addiction, social cooperation, and altruism. One of his major contributions to the field of behavior analysis was his introduction of the teleological perspective. In this molar view, behavior is understood by the end it serves. The occurrence of an activity is explained by its purpose as part of a larger-scale activity. According to the conception of mind entailed in teleological behaviorism, mental terms are identical with extended patterns of overt behavior. In the following interview, conducted in 2012, Dr. Rachlin clarifies the meaning of the notion of teleological behaviorism by explicating what it implies about the direction of psychological investigation. He also reports on the contingencies that made him a teleological behaviorist and reflects upon the scientific community's reception of this final-cause framework. By explaining how self-control—behavior commonly explained in terms of inner causes—can be conceived in terms of temporally extended behavioral patterns, Dr. Rachlin exemplifies how taking a molar view allowed him to generate approaches to solving an applied problem. The idea of molar behaviorism comes from Edward C. Tolman. Tolman was a major figure in American psychology in the 1940s–1950s. He was an opponent of Clark L. Hull who advocated a very molecular system. Hull's thought was very connectionistic. This means that learning was based on connections between stimuli and responses, and these connections can be found somewhere in the head. He explained the behavior of rats in mazes with these little connectionistic terms of “SHR”, where “H” stood for habit strength, which builds up. “H” told the rat whether to go left or right in a maze. It is hard to imagine now, but in the postwar years the Hullian school was the major psychological school. Tolman reacted to that and started doing maze experiments that he felt could not be explained in terms of simple responses like “go right”. According to Tolman, the rat did not learn “go right”, “go left”, “go this way” or “that way”. Rather the rat developed what Tolman called a mental map of the maze and really learned to go towards the window or towards the door2, even if he had to turn left to get there. The turns occurred in the right direction, even though it was not the exact same response. The Hullians called themselves behaviorists, like Watson in the [19]20s. Tolman was the first to call himself a molar behaviorist. Tolman was influenced by the phenomenologically oriented European school of Gestalt psychology. The Gestaltists advocated learning fields instead of learning the Hullian connections; this characteristic very much appealed to Tolman. Especially early in his career, he was a behaviorist in the sense that the terms in his theories stood for observable acts. However, in his later writings, he started to internalize a lot of these concepts, like mental maps.3 You can talk about a mental map and still be a behaviorist. But once you put the mental map into the person´s head, then you are not a real behaviorist anymore, at least as I see it. Tolman did explicitly call himself a molar behaviorist. He did not believe in habit strength; he altered Hull's symbolism from SHR to S–O–R, where “O” stood for the organism as a whole, rather than the connection inside the organism. That put him a little bit closer to Skinner. Tolman's reaction to Skinner was that he should have put two levers in the box. When Herrnstein put two levers in the box (or rather two buttons since he was studying the pecking of pigeons rather than the lever-pressing of rats) and began to formulate the matching law, he knew very well that he was following the path of Tolman. Tolman himself had stopped with mazes. Skinner detested mazes. He believed that too many things were going on in a maze to develop simple laws of behavior. Tolman and his students did some very clever experiments. He tried to explain a lot of different behavioral patterns by mental maps and other kinds of concepts. They were molar type concepts—useful concepts, as long as you do not put them in the head. I came to Harvard with a master´s degree from the New School for Social Research in Manhattan, New York, which was a center of Gestalt psychology. At the beginning of World War II, a lot of the Gestalt psychologists had moved to the United States. Those professors were met at the boats and given jobs at the New School. They had established the graduate school mainly with these refugee Gestalt professors. So, they had a graduate faculty that was very distinguished, and it was, at that time, oriented towards Gestalt psychology. Even though I myself had not paid much attention to that. However, I did take some of these Gestalt courses, so I was familiar with Gestalt psychology as well as perceptual psychology and Gestalt thinking in general. So, when I came to Harvard, I came prepared for this Tolmanian cast of psychology. I was not particularly interested in being a Skinnerian at all. It was just that the most interesting work was being done by Skinner and his students, and Richard Herrnstein. That was where people were arguing, talking, and discussing. It was very exciting to be in that lab. I did one experiment; it came out pretty well. So, I just kept doing experiments. Billy4 and I had to learn about Tolman. His name could be on the qualifying exam. We discussed Tolman with each other and with Herrnstein. Herrnstein was influenced by Tolman and I had my Gestalt background, so I was prepared—not that I could actually contribute anything—but I was prepared to hear this kind of thing. The Gestaltists were influential in American psychology at that time and we took them very seriously. But there is a story that Skinner tells in his autobiography: “The Gestaltists felt that they had a lot in common with me. So they invited me for lunch and tried to find a common ground. As we talked, we soon discovered all these irreconcilable differences. And that was that.” Then Herrnstein published his Matching Law paper5. Billy and I were of course very interested in that. Herrnstein and Phil Hineline started to do work on avoidance6. They had this avoidance procedure in which there was a negative correlation between lever pressing and shock, and that was all there was. The rats learned—and they learned better than if each individual response avoided an individual shock! They learned even better than if the shock was signaled and then the animal had to respond. So this is how far Herrnstein went, but Billy and I went further, as students often do. Herrnstein believed in matching. He believed that correlations had power, but he was still oriented to, what I consider to be, a sort of molecular view of matching. He explained matching with the notion of melioration. Melioration basically means: If an animal in a certain situation can improve things, he will. So, if you are responding on the left key and the rate of reinforcement becomes higher on the right key, you will switch. This does not happen due to an individual reinforcement, but yet the switch is in a sense reinforced by going up that step to the higher rate of reinforcement. That is a sort of molecular explanation of what is going on in matching7. Let me try to explain the connection between melioration and matching. Suppose you were a pigeon dividing your time between two sources of reinforcement both of which were continually supplied, but at different local rates (assuming a constant overall rate of pecking, local reinforcement rate = reinforcers per peck = Rf/P). Additionally, the more time you spent at either source, the more likely food would be available at the other source. According to melioration, you would start out at the source with the higher reinforcement rate, but as you spent more time at that source, the potential reinforcement rate at the other source would increase until it reached and then exceeded your current reinforcement rate. As soon as that happened, you would switch. If you kept switching whenever the reinforcement rate at the other source was above your current rate, the actually obtained rates of reinforcement would end up being equal from the two sources (Rf1/P1 = Rf2/P2). This strictly implies matching (P1/P2 = Rf1/Rf2). So, melioration is based on the notion that pigeons and people always choose the highest available local reinforcement rate. I once asked Herrnstein publicly; I raised my hand when he was giving a talk and asked, “Tonight I'm going to walk home instead of taking the bus even though the bus offers the higher local reinforcement rate. Why will I do that?” His response basically was: “You reconstruct the alternatives. You now choose the better temporally extended situation, which includes walking home and being a healthier person.” But why do we reconstruct? What I was trying to get him to say is that we often try to maximize reinforcement over a long period even when it goes against melioration8. But he did not say that. By then, both Billy and I called ourselves molar behaviorists. We took this term from Tolman (although our position was much more Skinnerian than Tolmanian) because it was deviating from Skinner in the direction of Tolman. One reason Billy gives is that “teleological” is sometimes confused with “tautological”, which is just by definition wrong. Secondly, teleology is often misunderstood. Exactly! If you tell them about final causes, they believe a final cause is causation in the opposite direction—that future things cause past things. They take teleological theories to argue that reinforcers work back in time and strengthen the response. Skinner avoided that by saying that the reinforcer does not work back; it just affects future responses. However, he talked as if the reinforcer had an effect on the previous response. That, of course, would have been wrong, and it would have been incorrectly labeled as teleological thinking. So, if I call my kind of psychology teleological behaviorism9 many people think that I am advocating backward causation—but I am not. A teleological cause is not like an efficient cause in reverse; it is really not like an efficient cause at all. The true meaning of “teleological” (that is, what Aristotle meant) is that it is a cause in the sense that it gives an answer to the question “why?” Often those answers would be something more molar, like, Q: Why are you hammering that nail? A: Because I am building a floor. Q: Why are you building a floor? A: Because I am building a house. Q: Why are you building a house? A: To provide shelter for my family. And so on. In order to build a house, you (usually) have to build a floor. So, the reason why you are doing a particular act is because that act is part of some larger, more molar (and more abstractly conceived) act. And that is the kind of explanation that we, both Billy and I, use. Yes, but a function is always a larger pattern of behavior. The behavior does not occur to attain a goal which is in the future. It occurs to be part of some larger more extensive pattern, as in providing shelter for your family. You are hammering the nail because that is part of some larger pattern. The rat presses the lever and gets food. There are two ways to look at why the rat presses the lever. One is, because that lever press is reinforced by the food. The food then is sort of acting backwards and reinforcing the lever press. Another way of looking at it is to say, the food and the lever press are one pattern of behavior. Pressing and eating is one thing; it is a pattern of behavior. You will press and you will eat. That pattern as a whole is more valuable than just sitting there in the cage. The contingency is there in the environment. The rat prefers the little package “press and eat” to the package “don't press and don't eat”. You can think of these behavioral patterns as packages. Q: Why does the rat press the lever? A: It prefers that package—which is the final cause —to the other available package. Yes. An economic utility function expresses utility as a function of a package of goods, for example, apples and oranges. Utility equals apples times three plus oranges times five to this power or that power. Give an animal the choice between apples and oranges, he will choose to behave so as to maximize that function. Not that there is a utility function engraved in his head and not that he knows what it is or that he is trying to actually obey it. It simply is the best description of his behavior. The utility function enables you to make predictions under various constraints. Let us say you constrain his behavior so that he has to buy five apples every time in order to get one orange. You can put those constraints into the utility function. He will still maximize the utility function as much as possible—given the constraints. This is standard microeconomic theory. I developed a theory with respect to behavior, instead of apples and oranges. The utility functions give you a distribution instead of a particular number. I found this way of talking to be very, very useful in understanding choice behavior. So, my work with utility functions was part of what made me a teleological behaviorist. This reason comes out of left field. At Harvard, I worked in Skinner's lab, although I never really called myself a Skinnerian. I was more of a Herrnsteinian. When I came to Stony Brook, people started attacking Skinner, asking me all those questions about him—and I refused to defend Skinner. I was an independent person; I had my own ideas, I believed. But as time went on, Skinnerian psychology started to be rejected by a lot of American academics, and I began to question what I really believed. Therefore, I began to read more philosophy. The more I read, the more convinced I became that the problem was not with behaviorism or Skinnerian behaviorism. The problem was that people were really interested in mental events. They were not objecting that behaviorism was not explaining behavior; they were objecting that it was not explaining the mind. Skinner did not have very much good to say about private events. He did not care for them and felt that they were an inferior way of looking at behavior. I started to read J. R. Kantor since I was concerned that what I was doing should have some relevance to what people were interested in—that is, the explanation of mental terms. It was Kantor who sort of gave me permission to understand mental terms as behavior of the whole organism over time. That was the way he saw it11. In his view, psychology is about the mind, but the mind is the behavior of the whole organism. Then, partly because of Kantor, I began to read Plato and Aristotle. Their writings began to solidify the way I thought about things; and I thought “Aristotle was a teleological behaviorist!” If Aristotle is a behaviorist and Aristotle is interested in explaining mental terms, then mental terms are capable of a behavioristic use! As Hank Schlinger says, “Consciousness is just a word.” And the question is “What is the best way to use that word?” That was Aristotle´s attitude. People who say that Aristotle was not a behaviorist mean that Aristotle was not a molecular behaviorist. But Aristotle was a behaviorist, a molar behaviorist. The same is true for Plato. The difficulty with Plato is that he wrote dialogs which are like little plays, and there is no one way to interpret them. I do not see Plato as a dualist but as a multilevel theorist, to use Phil Hineline's and Billy's terms. If you tell somebody I am interested in consciousness, they immediately think that cannot be behavioristic. The reason this is so difficult for people, is that they do not understand the Skinnerian notion that voluntary behavior is caused by the environment. That is the key. They think they carry their consciousness around in their heads. People are willing to accept quantum theory, even though they cannot envision it, because it is about the physical world and not about them. It is distant from them. They are willing to accept that the earth goes around the sun even though everyday talk is always about the sun moving. When it comes to our minds, it is still more difficult to abandon what seems to be absolutely certain. We find it almost impossible to accept a psychology where the explanations seem to go against our most fundamental perceptions. Consider pain, for instance. It is almost impossible to believe that pain is not inside us. Yet that is what I do believe12. Most scientists are monists. They believe that the world, including human beings, consists of only physical events. But at the same time, they believe that our pains are inside of us—that they are essentially private events. If pain is a private event, and you want to be a monist, then pain must be a neural event or a pattern of neural events. That is what philosophers call neural identity theory, that all of mental life is just neural behavior. It simply does not make sense. A lot of people say that pain is an action in the central nervous system. There are problems with that. For example, why should it be pain when the neural message goes from point A to point B in the brain, but not pain when the neural message goes from point C to point D? How can a particular electric current in your brain itself be pain? In answer you would have to say that particular neural events are not themselves pain. At this point there are two ways to go. One is dualism; the neural events are not themselves pain but they cause pain in our nonphysical (spiritual) selves. However, then you would have to posit a little homunculus in your brain who is actually feeling the pain. The other way, teleological behaviorism, says that the pain exists in the temporally extended and abstract pattern of your overt behavior—extended in time (the past and future) rather than in space (the interior of your head). Even if we all find it convenient in everyday life to speak as if our minds were in our heads (just as we find it convenient to speak as if the sun revolved around the earth), neuroscientists should not use pain or any other mental term to refer to neural events. They may say that these neural events and networks are efficient causes of pain. But as long as they remain monists, neuroscientists must recognize that the effects of those neural events cannot occur in another world. Then what is the pain itself? Teleological behaviorism provides an answer. Pain is a certain pattern of overt behavior correlated with aversive events—and no less horrible for that. Such a correlation, as in the Herrnstein–Hineline avoidance experiment, is not 1:1 but is nevertheless statistically significant. Thus, a given person will at times exhibit pain behavior unaccompanied by aversive stimulation and vice-versa. When these occur, it is tempting to say that the pain is in our heads. But that is a temptation scientists (when speaking as scientists) should avoid. Billy and Phil Hineline share my view. We understand each other perfectly. We could argue about certain points but we are all in tune. My students have difficulties. I do not insist on this viewpoint. I let them have their viewpoints. All I care about is that they do something interesting. They do the experiments, write them up and interpret them. Actually, mostly I do the writing and the interpreting. Very few of my students have really absorbed my view of the mind. It is sad in way. On the other hand, I give a talk and people do understand up to a point. It is just that, when I tell them really what I am saying, they cannot believe it. In a way, I do not blame them. Who would accept quantum theory if the atomic bomb was not exploding? You cannot understand this unless there is a development of application. Then you force people to understand it and to believe it. The problem is you can interpret that work with a more molecular viewpoint. However, what generated it was a molar viewpoint. So, if my work has an effect on people, that is good; but I do not have the ambition to make it have an effect. I am not indifferent. I do care about what people think, but it is a question of time allocation. If I can devote X hours to psychology rather than just sitting and enjoying myself, I would rather spend them on furthering my own understanding, doing experiments, interpreting them, and writing them up, than on popularizing my position. I'd rather do my little thing, than try to advertise it. Nevertheless, I have advertised to some extent. I give talks. I have written books. One is called Behaviorism in Everyday Life (1980). Another is called The Escape of The Mind (2014). This is as far as I go. I will not go around the world, down to South America like Fred Keller went, to popularize my views. Yes, for example, group-selection of behavior and learning altruism. I do not think you can be a behaviorist and explain how a person can learn to be altruistic unless you have this kind of molar view. By definition, altruistic acts are not reinforced. If reinforcement of individual acts is your only means of explanation, how do you explain learning of something that is not reinforced? Individual altruistic acts are never reinforced. Being a kind and good person is either reinforcing in itself or it is reinforced because people will generally be nicer to you. But no individual act will be reinforced. In my book The Science of Self-Control (2000), there is this anecdote about an American comedian14 who met a British guy15 backstage. The British guy was smoking a cigarette, and when the American started commenting on the health risks of smoking, he said, “I know that smoking will kill me, but I do not know that this particular cigarette is going to kill me.” That is the problem: Refusing or putting out that cigarette is never going to be reinforced. It is not reinforced with delay three weeks later by waking up a healthier person. There is positive reinforcement for smoking that cigarette, and there is no negative reinforcement. There is nothing negative about it. Smoking that particular cigarette is never punished. However, if you are a real mentalist, you can argue that refusing to smoke is self-reinforced or that you reward yourself by internally petting yourself on the back. The absurd idea that you could internally reinforce yourself16 was another motivating force for me to develop an alternative explanation for problems of self-control.