Simon says: Bounded rationality matters: Introduction and interview

Abstract

A brief discussion of Herbert A. Simon is presented, followed by a brief interview with the key figure in the history of 20th-century social sciences.

MEET THE PERSON

HERBERT A. SIMON: AN INTRODUCTION

Professor Herbert A. Simon is a key figure in the history of 20th-century social sciences, and he has made important contributions to organization theory, political science, computer science, psychology, management and economics, as well as other fields. The intellectual effects of his work are indisputable; he pioneered research on bounded rationality and satisficing in economics; helped establishing the field of organization theory; and greatly advanced our understanding of computers, psychological processes, and other fields. Despite working in such seemingly different fields, there is a remarkable continuity in Professor Simon's work; Simon was always interested in how limited rational human beings make decisions and solve problems. Particularly important to Simon has been the computational and informational limits to human rationality. From his collaboration with James March (and their classic work, Organizations [March & Simon, 1958]) and his important contributions to the behavioral analysis of rational choice (Simon, 1955, 1956), Simon appealed to psychology and cognitive science to understand how people make decisions under the influence of bounded rationality. He also helped establish the Graduate School of Industrial Administration at Carnegie Mellon University, which later became a model for U.S. business schools. As an introduction to the interview below, I will summarize some of Simon's significant ideas that are discussed in the interview.

Born in 1916 in Milwaukee, Wisconsin, Professor Simon was educated in political science at the University of Chicago (B.A., 1936; Ph.D., 1943). He has held research and faculty positions at the University of California (Berkeley); the Illinois Institute of Technology; and since 1949, Carnegie Mellon University, where he was Richard King Mellon University Professor of computer science and psychology until his recent death. Simon was the winner of many prestigious awards, such as the A. M. Turing Award (1975) and the Alfred Nobel Memorial Prize in economic sciences (1978). His most well-known writings include Administrative Behavior (Simon, 1947), Organizations (March & Simon, 1958), Human Problem Solving (Newell & Simon, 1972), Models of Bounded Rationality (Simon, 1982), and the autobiography Models of My Life (Simon, 1991). In addition to his research publications, Simon held important political positions in professional associations, which contributed to his impact on the social sciences. Simon was the first social scientist to be elected to the prestigious National Academy of Sciences (NAS), and from his work there, he became involved with the very important Committee on Science and Public Policy of the NAS (COSPUP) and the President's Science Advisory Committee (PSAC). He also contributed to professional societies and was, for example, a founding member and first vice president of the Institute of Management Sciences (TIMS). These efforts produced remarkable results for the advancement of a behavioral program of bounded rationality and decision making.

Although Simon graduated in political science (because he didn't want to take the accounting course, which was a requirement for graduating in economics [cf. Simon, 1991]), he acquired an interest in public administration, organizations, and management early on. He had also acquired a taste for econometrics and mathematical tools during his student years at the University of Chicago, where he studied with teachers such as Henry Schultz and Rudolph Carnap. From 1939 to 1942, Simon was working at the Bureau of Public Administration at the University of California, Berkeley, where he had gained insight on managers' attempts to create efficient and rational decision making in government organizations. This work planted the seeds for the ideas that Simon developed in his thesis, published as Administrative Behavior (Simon, 1947), on the importance of organizations in understanding the boundaries of human rationality (Simon, 1991, p. 74). After returning to the University of Chicago, Simon began attending the Cowles Commission seminars, where discussions on rational decision making and planning were prominent! This further stimulated Simon's interest in decision making and the limits to rationality, and when he came to Carnegie Mellon University (then Carnegie Tech) in 1949, he was ready to begin exploring the implications of his ideas for economics and organization theory. Simon had little faith in the idea from neoclassic economics that decision makers maximize utility and make rational choices, and he wanted to outline a new and behavioral model of understanding decision making. This was done in two very influential articles in the 1950s (Simon, 1955, 1956), which (a) pointed out the significance of the psychological limitations of knowledge and computational powers that the decision makers confront and (b) developed the idea of satisficing. The two articles were written during Simon's stay (in 1952-1953) at the RAND Corporation and were intended as summarizing Simon's reaction to rational choice theory and game theory that surrounded Simon at RAND.2 March (1978) has summarized Simon's findings in the following way:

It started from the proposition that all intendedly rational behavior is behavior within constraints. Simon added the idea that the list of technical constraints on choice should include some properties of human beings as processors of information and as problem solvers. The limitations were limitations of computational capability, the organization and utilization of memory, and the like. He suggested that human beings develop decision procedures that are sensible, given the constraints, even though they might not be sensible if the constraints were removed. As a short-hand label for such procedures, he coined the term "satisficing." (p. 590)

In the very interdisciplinary environment at Carnegie Mellon University in the 1950s and 1960s, Simon further developed these ideas. The Ford Foundation had decided to establish a program for research in behavioral science in the late 1940s, and much of their effort was centered on Simon and his colleagues' aims to set up their business school model. When Simon came to Carnegie in 1949, he believed that research in the areas of organizational psychology and sociology was necessary to understand human decision making, and he also had a strong interest in mathematics. He wanted to apply these interests to a major study of organizations. Thus, after hiring James March to come to Carnegie, the two began working on their book, Organizations (March & Simon, 1958). This work clearly showed that organization theory at Carnegie aimed to bring insights from social psychology, organization theory, and sociology, combined with a knowledge in economics and mathematical tools such as linear programming and operations research, to the understanding of decision making in organizations. Today, Organizations, together with A Behavioral Theory of the Firm (Cyert & March, 1963), represents the fundamental contributions to the behavioral perspective on organizations and organizational behavior. Simon also worked with John F Muth, Charles Holt, and Franco Modigliani on dynamic programming techniques, and during these studies, Simon began getting access to empirical and psychological data about human decision making. As a result, Simon's faith in rational decision making decreased further, and gradually, the psychological aspects in his work became more and more important because it was a way for him to understand human decision making. His work with Allen Newell on Human Problem Solving (Newell & Simon, 1972) thus represents a continuation of the basic insights of Simon's early work in economics and organization theory, applied to the psychological study of human problem solving through computer simulations of human thinking processes.

As Simon talks about below, this was a very important period, and the Graduate School of Industrial Administration at Carnegie Mellon University was a very important place. Simon and his colleagues made important contributions to the theory of bounded rationality, to the behavioral theory of the firm, to cognitive psychology, and to artificial intelligence. It was an interdisciplinary environment where economics, statistics, and mathematics were used and appealed to with as much admiration as psychology, sociology, and organization theory. The pioneering developments that took place at Carnegie Mellon were results of the belief that research questions mattered more than disciplinary boundaries did.

Simon himself was a leading figure in these developments and was driven by an overall vision of advancing a behavioral approach to the study of limited rational decision making. The inquiry into how people make decisions and solve problems as part of their daily behavior is at the center of his work. His use of bounded rationality considerations is more important than ever, sometimes reinforced by Simon by the use of the term procedural rationality (see below). In arguing for procedural rationality, Simon stresses the importance of rules as being significant for economizing the limited capacity for attention and information processing of the human mind. In doing so, he laid the foundation for subsequent developments about decision making, taking limited information and cognitive abilities into account. However, as mentioned in the following interview, not all developments that have used the concept of bounded rationality have built on Simon's insights. For example, Simon wrote in his letter to Thomas Sargent, who has been an important developer of the idea of rational expectations and who has recently begun using the term bounded rationality (but in a way different from Simon and closer to mainstream views of rationality): "I could complain and say: 'I invented it and have a right to decide how it should be defined/ but as I failed to apply for trademark rights, I guess I have no standing in court."3 Nevertheless, as Simon concludes in the following, bounded rationality and behavioral ideas have made major contributions to our understanding of human decision making and organizations. Herbert Simon has been very important in this development, and his ideas will continue to inspire, it is hoped, new insights into the understanding of limited rational decision making.

HERBERT A. SIMON: AN INTERVIEW

The following conversation took place between Herbert A. Simon and Mie Augier at Carnegie Mellon University on Tuesday, September 29, 1998.

Augier: You won the Nobel Prize in economics and have been working within different fields such as cognitive psychology, organization theory, political science, and management. You have mentioned that you didn't read Keynes until 1947, and you didn't graduate in economics because you didn't want to take the obligatory accountancy course. Thus, your main interest wasn't in economics in the beginning?

Simon: Actually, I think that I tried to read Keynes before that, but I didn't understand it until then, because that was when I saw the work of Franco Modigliani and John Hicks that I could understand their equations. But Keynes himself is not very penetrable, as you probably know. I think I bought my copy of Keynes's General Theory (Keynes, 1936) when I was still in Berkeley and that would have been before 1942. But I can't claim that I understood him in the beginning-or still do very much, except on "animal spirits"-I like that.

Augier: So, what was sparking your interest in economics? Was it the environment at Chicago or the Cowles Commission seminars that took place there?

Simon: No, actually not, although those seminars were quite important to me, particularly because of the group of people that were attending. You see, before I got to Chicago, I had been very much interested in the social sciences since high school. I had an uncle who graduated from [the] University of Wisconsin just before World War I, and Wisconsin was a very progressive place. There were a group of very progressive economists there, and my uncle was a student of John R. Commons. So, unlike most kids, who don't hear about the social sciences until they get to college, I was exposed to economics through my uncle's books, and I understood that was an area where you might do science. As a consequence, one of the ideas I had when I went to college was that it was perhaps a good idea to go into the social sciences and bring some mathematics to them. Make them 'real' scientists. And that was probably what led me to economics in the first place, as I remember it. The real competitor was biology. But I decided that perhaps color blindness was too big a handicap for studying biology. That may be the reason I ended up doing economics.

Augier: Did your years at the University of Chicago have an impact on you?

Simon: I went to Chicago as an undergraduate in 1933, and that was when I made these decisions about economics. But, even though I majored in political science, I continued to take economics courses. I took Henry Simons's intermediate price theory course-and I had already read Ely in preparation for high school debates (Ely, 1930) and Henry George (1882) on the "single tax." I was therefore ready to take an intermediate theory course. Actually, that course was the course that Frank Knight developed. So, I learned to do price theory, and when you can do price theory, then you can do a lot of classical economics.

The second thing I did was to take graduate courses while I was still an undergraduate from Henry Schultz, who was the principal econometrician at Chicago-he wrote a big volume on the theory of demand (Schultz, 1938) and did econometric work on it. I learned from him, not only Walrasian theory and the theory of demand but also econometrics. In sum, I got a very good classical economics education at Chicago.

So that was the Chicago department. Mostly socialists-it was in the depths of the depression. Henry Simons had a little Positive Theory of Laissez-Faire book at that time, but he also said in that book that if you can't have competition-if there is monopoly-then we must have state ownership. So, that was the kind of environment we had-certainty not Chicago today. In fact, there was very little of what we would now call Chicago economics at that time, except that when the people taught price theory, they taught it "straight." But a person such as Knight never did confuse that with real life. Never. And yet, he is regarded as the "father" of Chicago economics. But I think that it was only when [Milton] Friedman took it over and made a very narrow cult out of it that it became Chicago economics.

Augier: Who would you say have been the three most influential thinkers on your own early intellectual development?

Simon: That's very hard to say Henry Schultz certainly had a very large influence on me because he did really teach me a lot of classical economics, and he also really introduced me to the identification problem. So, he had a large influence on me. Then, I also took courses from Rudolf Carnap-and I was even probably more influenced by Principia Mathmatica by Whitehead and Russell (1935).

I was more probably more influenced by books than by people, actually. As I said, very early on I read Ely, which was a standard economics textbook-very solid. And I read Commons, and that had a big influence, partly in giving me an alternative framework to all this static classical theory in thinking of decision making.

Augier: You came here to Carnegie Mellon in 1947 to give a lecture and in 1949 to stay. You formed with Bill Cooper and Lee Bach the environment around the Graduate School of Industrial Administration (GSIA). That must have been an exciting period?

Simon: That was very exciting initially, because of all the operations research that was just beginning. Linear programming was just beginning. So we were teaching our master's students linear programming in their first year.

Now, this was in the business school. There was a lot of dissatisfaction back then with business schools-economists didn't really think very much of them. So, two separate foundations set up studies of business education-the Carnegie Corporation and the Ford Foundation. And when they saw the new things we were doing, trying to really base business education on science, economics, and operations research-bringing mathematics in, bringing organizational theory in through social psychology-they decided that was the right model of the business school. The common vision that brought us together was to build a business school in such a way as to bring the basic sciences and the practice of the "real world" together. So, we became sort of the stalking horse for the reform of business schools in this country. Of course, we were completely lucky on the timing of this.

Augier: When did you start working with Jim March? Simon: Well, when did he come to CMU? ... Let me see, I hired him. We were building up this faculty, so Lee Bach and I were doing most of the hiring. In those days, you didn't have those big committees, advertising jobs for 6 months and such nonsense. We went to schools where we thought that interesting things were happening and where interesting people were. And then we asked our friends about who were the good doctoral students. So someone gave me Jim March's name, and we had dinner, and I think I phoned Lee back that same night and told him that I was offering Jim a job. That simple it was then. He was tops.

Augier: So, how did the idea around Organizations begin to develop?

Simon: Well, I had, of course, in political science been interested in organization theory-my thesis was in organizational theory, and at Berkeley, I had been working with real administrations and organizations. So, when we came here, the first thing that we did-other that getting the operations research stuff study started-was to start studies in decision making in organizations. That attracted Jim. We had a whole group concerned with decision making. First thing we did was-speaking of my dislike of accounting-was to do a big field study in about four companies of how accounting data are or are not used by managers. That was how it started.

Augier: You noted in your autobiography that you and Jim later walked in diverging professional paths.

Simon: Yes. We share the basic behavioral interest. But Jim has a very large fascination (I have some) with even more extreme forms of irrationality than I do. So, in my organizations, there is still more rationality than there is in the garbage can. Thus, we would have to have a lot of discussions writing a book together again. On some of the things I would go quite a way with Jim, but not all of them. The book he did in 1994 (March, 1994) I like very much. That is a really good piece.

Augier: In 1970 you moved to the psychology department at Carnegie Mellon University. Can you explain a little bit the background for that move?

Simon: I moved into this office here about 1970, yes. But I moved into my activities in cognitive psychology and computer science about late 1950s, bul pretty much during the 60s. Newell came here in 1955. When I began to move in a more psychological direction, there were problems of reforming the psychology department here because "behaviorism" was dominant, and you were not allowed to talk about the mind or anything inside the head, only about what came out. So, I transferred my main political activities from reforming GSIA to reforming psychology. Reforming stimulus-response, or behaviorist psychology to cognitive psychology. As a matter of fact, I have lived with or against "behavioral" labels all my life.

Moving to the psychology department allowed me to continue my interest in decision making and to focus on the more psychological aspects of human problem solving; an interest I had developed when arguing against rational models of choice. But it was all part of the same thing: I wanted to understand how limited rational people behave.

Augier: Around the same time, GSIAbecame dominated by more neoclassical thinking and by rational models of choice. Was this a local phenomena, or do you think that it was a part of a general trend in economics?

Simon: I think it was part of a general trend. You know, when we started here, the world was still pretty Keynesian, and our faculty was pretty Keynesian. We had Franco Modigliani. So, this began to change in the 50s, and as we hired young people, we did it from good places, and Chicago is obviously a good place. And we were getting more and more people who had the "true religion"-the Chicago religion.

At GSIA, we tried to hire the brightest people we could find, and we tried not to enforce orthodoxy. But, when the balance tilted-and this is my personal version of it-there was not exactly the same amount of tolerance in the other direction. Because there is something about Chicago economists-they are "true believers." Then we began to have much more orthodoxy here.

Also, as part of the fact that I indicated to you earlier, I started out by thinking that I was going to bring more mathematics to the social sciences. But I got less and less enchanted by the degree of formalism and what seemed to me increasing neglect of hard empirical work-and disrespect for empirical work unless it was combined with formalism. You know, if you could do some fancy three-stages regression analysis, that was great. But if you mulled around with verbal "think-aloud" protocols, that wasn't regarded as so great.

So, that was the way I characterize the shift. Others will probably describe it in other ways.

Augier: It is quite interesting and perhaps ironic that bounded rationality and rational expectations theory came from the same environment.

Simon: Yes-and even from the same small set of people. Jack Muth worked with us closely. It really came out of the project that lead to the Holt-Modigliani-Muth-Simon book (Holt, Modigliani, Muth, & Simon, 1960).

The actual thing we started that study about was a factory, and we constructed these decision rules. And then Jack Muth, who was a graduate student, had the job of handling the prediction part of the project. Our main focus was on production rules, but you had to put into them expected values of sales in the future, and he got into the prediction literature, which was then dominated by frequency-analysis and time-series folks. At the same time, Franco Modigliani and Emile Gruenberg were doing something on the possibilities of public prediction: the fact that if you predicted, people might react to the prediction and falsify it. They wrote a paper on that topic, and it made people think.

We also had a big conference here at that time on prediction and expectations in economics. So, there was a lot of talk here about uncertainty. You have to remember that at that time, almost the whole literature on uncertainty in economics that a graduate student would be exposed to was possibly Schumpeter. That was exciting, but nothing really that you could put your teeth into. There was no real model. Or, of course, Frank Knight's (1921) book on Risk, Uncertainty and Profit. That was the literature on risk and uncertainty. So, we had this big conference here on uncertainty with all the people who were working on it, and there was a lot of talk here about how you deal with prediction and uncertainty. And that I think was what started Jack Muth thinking about this. But he came out in a very different place than we did. That's fine. It is really a crime that they didn't include him in that Nobel Prize. It was his idea-it wasn't Lucas's ideas, it was Muth's idea. Lucas was the first one who really made something of it-but the ideas were all there in Muth's (1961) first Econometrica article.

That is why I think we were talking about uncertainty and expectations here a lot. And of course, these issues are still very important.

Augier: Ward Edwards (1954) did a paper in 1954 which is pretty close to bounded-rationality thinking. Others had developed similar ideas too around that time.

Simon: Well, yes. There were a whole bunch of interesting decision theory types at the RAND Corporation. It was really exciting. And again, that was part of the atmosphere that created that conference and the environment at GSIA-the RAND connection and the Cowles connection. Both Tjalling Koopmans and Jacha Marchak were here. Those two I knew very well, of course, from my RAND and my Cowles connections. But the most important thing was when the Ford Foundation set up their program in "behavioral science"; that really got things started, and we developed our ideas on bounded rationality and behavioral sciences.

Augier: Brian Loasby told me that George Shackle had told him that he met you once and that he expected to be crunched up-but that he found you delightful.

Simon: Well-I must have been in a very good mood that day. No, I liked him-his theory was something we took very seriously at that time. But one doesn't keep up personal relations with everyone one meets at a professional meeting. Especially not when one is going to more professional meetings than one ought to, which I was doing at that time. Oh, Georgescu-Roegen was there too, of course. He was really also doing behavioral economics-- and so was Shackle, I think.

Augier: He has this concept about bounded imagination, which may seem in some ways similar to your concept of bounded rationality.

Simon: Well, you have to realize about the bounded rationality terminology that I began to use this as a label for the things that economists needed to pay attention to-and were not. It was never intended as a theory in any sense. And they needed to pay attention to the variables that we have to put into a decision theory when we build it. So, don't expect to find any precise characterization of it. Typically, when I define it, I would say, "Well, what are the bounds." The bounds are the bounds on knowledge, bounds on calculation, multiple objectives, or competing objectives if you have a group of people competing to make the decision. But the terminology... one really has to distinguish between the sort of guidelines that we use as directions for research and what we would call a theory. GPS [The General Problem Solver] is a theory of human problem solving, but bounded rationality is not a theory in any precise sense. Allen Newell and I once talked about Laws of Qualitative Structure (in our Turing Award Address). We asked, What kind of a theory is the germ of disease? Well, the germ theory of disease is little more than the suggestion that if somebody is sick, look around and see if you can find a microorganism. Sometimes you do and sometimes you don't. So, the germ theory of disease is just as vague as bounded rationality and it serves the same purpose: To point research in a particular direction. So, I didn't intend to exclude or embrace any particular thing-Shackle may be using one kind of bounded rationality too.

Augier: What will you say would be the differences between bounded and procedural rationality?

Simon: Well, the term bounded only enters the picture-limits on rationality enter the picture-because you are trying to take seriously the available processes in decision procedures. If you are only interested in whether something is, in fact, rational-whether it, in fact, meets the conditions of the environment or not-then you don't have to concern yourself with process. So, there is a linkage here. If you are concerned with processes, then you have to be concerned with the limits on rationality because process doesn't allow you do to this magical optimization in any way. If you are only interested in the product, then you don't have to worry about bounded rationality. If the environment is simple enough, the solutions that people reach are close enough to the solutions people reach by climbing the highest mountain. That is why those two are connected.

Augier: After your article in the Latsis book (Simon, 1976), I think you have focused more on procedural rather than bounded rationality. Am I wrong here?

Simon: No, that is absolutely correct. You have to understand that science is a process of trying to find new knowledge, trying to establish new knowledge, and trying to convince people that it is new knowledge. And, in my frustration with economists-and believe me, I've had a lot of frustration with economists-I have been constantly seeking, over these years, the right way to convince them that you simply can't go on with this optimization nonsense. And the bounded rationality label with the little description of it that I mentioned a few moments ago seemed a good way, and I think it has survived very well-everybody is now a "bounded rationalist" as it were, even those who don't know what the word means.

Augier: Many scholars closer to mainstream economics have taken notice of bounded rationality. People such as David Kreps and Robert Aumann are using the term. Oliver Williamson, of course (who is not mainstream) as well.

Simon: Yes, but .... Well, look at Williamson. He is one of our students and a good friend. But nevertheless, we argue when we get together. He still isn't happy about anything until he shows that really there is an implicit contract and people are really maximizing utility within that contract. So, what does that really buy you except the feeling that, "gee-you are still safe in the womb of rationality." So, Olly and I disagree on that particular point, although I have recently come to realize that he is more behavioral in his later work than I had pictured him before. I was recently rereading his 1985 book (Williamson, 1985). As you know, I have been accusing him, from time to time, of pouring the new bounded rationality wine into the old maximizing bottles (see Simon, 1997). But after rereading his book, I see that he takes a bounded rationality position, and I did not (in that book at least) find any claim that "best" solutions were found by transaction cost. He also used nice empirical evidence, especially historical evidence and case studies, to support his basic arguments. So, I did find his position very much closer to the one I hold than I expected to, although I still disagree with him on issues such as the significance of "opportunism." I have expressed these concerns in my work on organizational identification and altruism (also discussed in Simon, 1997).

However, it is true that many mainstream economists have borrowed "bounded rationality" and have introduced all kinds of assumptions as bounds on their theories, but it never occurs to them that the real issue is what people in fact do. "Armchair economics" cannot say anything about that. Certainly, you won't say that Sargent is saying anything about what people are really doing. And much less Lucas. Or, for that matter, Gary Becker with his impressionistic empirical studies. If you look at the way that he does empirical studies, you will think that he is not serious. Did you see the critique in the symposium that we had on the Chicago School (Simon, 1987)? I picked on him there. His empirical studies are not good. They are as bad as Friedman's monetary history. At no point does that volume on monetary history, as I recall, ever worry about the identification problem. This is their way of doing economics, but it isn't the right way to do science.

Augier: You have written about economics and its "sister science" psychology, as you call it. The works from Kahneman and Tversky seem very congenial to the bounded rationality notion. Do you think these works are consistent with your view on bounded rationality?

Simon: Oh, yes, very much indeed. And I think that they had a very big impact. Although it was extremely difficult to get published in economics journals at first-in the beginning, they were just turned down. Their first publication was in Econometrica, I believe. And I'm not sure-- but I think I was a referee on that. Anyway-they had a terrible time getting in economics journals at first.

They are certainly important for the future of behavioral economics.

Augier: Would you still think that the metaphor of chess is an appropriate metaphor of decision making? I mean, by using this metaphor, one indicates that there is indeed a structure "out there"-the agents are just not clever enough to find it.

Simon: Well, I will put in a slightly different way. I would say that chess, as a task, as a decision-making task, has turned out to be an excellent laboratory for studying the psychology of decision making. There is a large literature, not just of chess machines-that is only one part of it-of theories of decision making applied to chess, some of which has been modeled by machines. It has had a very large impact on theories of perception and memory. It started with Adriaan DeGroot, who is a psychologist who did a famous experiment in 1946 in which he had chess players reproduce a chessboard they had only seen for 5 seconds. That line of thinking led to a whole line of research, some of it here and some of it in the Netherlands and other places. It has had a big impact on theories of perception and memory.

Augier: But in a chess game-that is in principle a fully defined system. I guess decision situations are not?

Simon: Well, bounded rationality types wouldn't worry about that "in principle" because obviously it has nothing to do with human behavior. Human beings don't calculate all the possibilities. You see, the relevance of that, and that is really the point, is that chess is a good approximation, not an optimization of anything. But no-decision situations are not fully defined systems when you take limited human computation into account. The chess metaphor just happens to be a good approximation and illustration.

Augier: How did you come to the term behavioral economics?

Simon: Perhaps because when we invented that (if we did) in the early 50s, that was when the Ford Foundation was trying to bring the social sciences together with the label "behavioral science." And that was a good label to use-- behavioral was just a more or less natural term to use at that time.

Augier: Where should people go if they would do research in behavioral economics today?

Simon: Well, if I were to recommend, I would say, "Go to a business school," not a department of economics. And in some business schools, you will find people doing this, for example, at the University of Minnesota and at the Wharton School. You still have, of course, to pick your people within the school, but you will find some.

Augier: Did you ever talk with Ronald Coase about bounded rationality? He had an interesting note in 1993 (Coase, 1993) where he said that he has "reservations about his concept [bounded rationality] as I do to any concept that includes the word `rational"' (p. 98).

Simon: No, I didn't talk with him about that. From the beginning, I ran across Coases's early piece (Coase, 1937) about why there are organizations. I cannot place it anymore, but I probably ran into it before I went to California, that is, before 1939 and before I even started on my thesis. I always assumed that there was a reference to it in my thesis, but that certainly isn't in Administrative Behavior. And I have been sort of promising myself that at some occasion, I would try to track that down a little more. I don't even have a copy of my original thesis-oh yes, I do have the microfilm of it somewhere, I guess. But somewhere very early on, I knew about that article, and why I never referred to it I don't know. You see, it is a funny piece. You can interpret it either as a clear-cut utility-maximizing piece or you can interpret it as concession to limited knowledge. I think he was really considering the limits on human beings, but I never had any correspondence with him about it or talked with him about it, because he didn't publish much more about that. But I am sure that I knew that paper very early on, and I am sure that I also knew his second paper (Coase, 1960).

Augier: But you knew Frank Knight pretty well?

Simon: Yes, I got to know him in Chicago, actually because we went to the same church. He was a very nice man. The interesting thing about him as the father of the Chicago school is that he was almost a cynic about human affairs. He really was not a believer in progress, for example. And if you go back and read his book on Risk, Uncertainty and Profit, you will find that most sentences would end with a question mark. This is a man wrestling with a problem-not a man who pretends to know all the answers. It is a man who struggles, isn't it? And in the end you get that kind of sad conclusion that there really aren't profits, except by mistake. There is return on investment and there is return on your labor, but profits are just mistakes. And that doesn't sound like a modem Chicago economist.

Augier: What would be the unifying theme for you in all these years and different areas?

Simon: To try to understand how people make decisions. Very simple. There are a lot of decisions in economics, there are a lot of decisions in political science, and there are a lot of decisions in psychology. In sum, there are a lot of decisions in doing science. It is all the same subject. My version of the science of everything.

My ideas have implications in different disciplines, but I've always been motivated by research questions rather than disciplinary boundaries. I found that different disciplines such as operations research, linear programming, organization theory, economics, and psychology were converging upon a shared set of questions and methods which were all related to the study of decision making in complex social systems (such as organizations). And that was my research question!

Augier: What to you think is the most important article or work of yours?

Simon, Well, in economics, the most important paper was the one from 1955 (Simon, 1955). But I have done other things in economics; my 1952 paper on causality (Simon, 1952) is also important. But I don't think of any one of them of special importance for bounded rationality, except the 1955 article. I must add that the 1956 paper (Simon, 1956), the companion piece, I think was important for developing the argument. Those two papers really go together.

Augier: What are your hopes for the economics profession? Simon: That it will take seriously the foundations of economic theory and the methods that are necessary to deal with bounded rationality and real-world decision making. To do that, they have to reform the graduate training. We are bound by our tools. Until economists are good at dealing with such issues, economics is not going to get very far. It isn't going to go anywhere. Shackle had at least a process theory of what was going on.

Augier: Professor Simon, thank you so much for your time and patience.

Footnote

NOTES

Footnote

1. The Cowles Commission was a group of mathematical economists doing pioneering research in econometrics, linear and dynamic programming, and decision theory (among other things). The economists connected to the Cowles Commission included such well-known names as Kenneth Arrow, Jacob Marshack, Tjalling Koopmans, and Gerard Debreu, and they held regular seminars to discuss their research.

2. Like the Cowles Commission, research at the RAND Corporation centered mostly on econometrics, game theory, linear programming, and other rational tools. Later, organization theory also became a research topic for RAND, and RAND was important in establishing the early contact between Simon and Newell, leading to a large part of Simon's later research on computer science and cognitive psychology (Simon, 1991, p. 167-168).

3. Letter from Herbert Simon to Tom Sargent, July 11, 1995, p. 1. Herbert Simon Papers, Carnegie Mellon Univer

Footnote

sity Library. I am indebted to Simon for giving me a copy of this letter.

References

SELECTED WORKS OF HERBERT A. SIMON

References

Holt, C. C., Modigliani, F., Muth, J. F., & Simon, H. A. (1960). Planning, production, inventories, and work force. Englewood Cliffs, NJ: Prentice Hall.

March, J. G., & Simon, H. A. (1958). Organizations. New York: John Wiley.

Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice Hall.

Ridley, C., & Simon, H. A. (1943). Measuring municipal. Chicago: International City Managers' Association. (Original work published 1938)

Simon, H. (1947). Administrative behavior. New York: Free Press.

Simon, H. A. (1951). A formal theory of the employment relationship. Econometrica, 19, 293-305.

Simon, H. A. (1952). On the definition of the causal relation. Journal of Philosophy, 49, 517-528.

Simon, H. A. (1953). Causal ordering and identifiability. In W. C. Hood & J. C. Koopmans (Eds.), Studies in econometric method (pp. 49-74). New York: John Wiley.

Simon, H. A. (1955). A behavioral model of rational choice. Quarterly Journal of Economics, 69, 99-118.

Simon, H. A. (1956). Rational choice and the structure of the environment. Psychological Review, 63,129-138.

Simon, H. (1957). Models of man. New York: John Wiley

Simon, H. A. (1959). Theories of decision making in economics and behavioral science. American Economic Review, 49, 253-283.

Simon, H. (1965). The shape of automation for men and management. New York: Harper & Row.

Simon, H. (1962). The science of the artificial. Cambridge: MIT Press.

Simon, H. A. (1976). From substantive to procedural rationality. In S. J. Latsis (Ed.), Method and appraisal in economics (pp. 129-148). Cambridge, UK: Cambridge University Press.

Simon, H. A. (1978). Rationality as process and as product of thought. American Economic Review, 68,1-16.

Simon, H. (1982). Models of bounded rationality (2 vols.). Cambridge: MIT Press.

References

Simon, H. A. (1987). Rationality in psychology and economics. In R. M. Hogard & M. W. Reder (Eds.), Rational choice: The contrast between economics and psychology (pp. 25-40). Chicago: University of Chicago Press.

Simon, H. A. (1991). Models of my life. New York: Basic Books.

Simon, H. A. (1992). Economics, bounded rationality and the cognitive revolution. Aldershot, UK: Edward Elgar.

Simon, H. A. (1997). An empirically based microeconomics. Cambridge, UK: Cambridge University Press.

References

OTHER REFERENCES

References

Coase, R. H. (1937). The nature of the firm. Economics, 4, 386405.

Coase, R. H. (1960). The problem of social cost. Journal of Law and Econmics, 3 (October), 1-44.

Coase, R. (1993). Coase on Posner on Coase. Journal of Institutional and Theoretical Economics, 149, 96-98.

Cyert, R., & March, J. G. (1963). A behavioral theory of the firm. Oxford, UK: Basil Blackwell.

Edwards, W (1954). The theory of decision making. Psychological Bulletin, 51, 380-417.

Ely, R. (1930). Outlines of economics. London: Macmillan.

George, H. (1882). Progress and poverty. Garden City, NY: Doubleday.

Keynes, J. M. (1936). The general theory of employment, interest, and money. London: Macmillan.

Knight, F H. (1921). Risk, uncertainty and profit. Boston: Houghton Mifflin.

March, J. G. (1978). Bounded rationality, ambiguity, and the engigeering of choice. Bell Journal of Economics, 9, 587-608.

March, J. G. (1994). A primer on decision making. New York: Free Press.

Muth, J. (1961). Rational expectations and the theory of price movements. Econometrica, 29,315-335.

Sargent, T. (1993). Bounded rationality in macroeconomics. Oxford, UK: Oxford University Press.

Schultz, H. (1938). The theory and measurement of demand. Chicago: University of Chicago Press.

Simon, H. (1934). A positive program for laissez-faire. Chicago: University of Chicago Press

Whitehead, A. N., & Russell, B. (1935). Principia mathematics. Cambridge, UK: Cambridge University Press.

Williamson, 0. E. (1985). The economic institutions of capitalism. New York: Free Press.