An interview with

H: Sometimes working very hard has two bad side effects. One is that you have solved the problem and there is nothing more to say. Two, it is so hard that nobody can follow it; it’s too hard for people to get into.

* * *

We were talking about various stations in your life. Besides City College, MIT, Princeton, and Hebrew University you have spent a significant amount of time over the years at other places: Yale, Stanford, CORE, and lately Stony Brook.

A: Perhaps the most significant of all those places is Stanford and, specifically, the IMSSS, the Institute for Mathematical Studies in the Social Sciences ­– Economics. This was run by Mordecai Kurz for twenty magnificent years between 1971 and 1990. The main activity of the IMSSS was the summer gatherings, which lasted for six to eight weeks. They brought together the best minds in economic theory. A lot of beautiful economic theory was created at the IMSSS. The meetings were relaxed, originally only on Tuesdays and Thursdays, with the whole morning devoted to one speaker; one or two speakers in the afternoon, not more. A little later, Wednesday mornings also became part of the official program. All the rest of the time was devoted to informal interaction between the participants. Kenneth Arrow was a fixture there. So was Frank Hahn. Of course, Mordecai. I came every year during that period.

It was an amazing place. Mordecai ran a very tight ship. One year he even posted guards at the doors of the seminar room to keep uninvited people out. But he himself realized that that was going a little far, so that lasted only that one summer.

Another anecdote from that period is this: the year after Arrow got the Nobel Prize, he was vacationing in Hawaii at the beginning of July, and did not turn up for the first session of the summer. Mordecai tracked him down, phoned him and said, Kenneth, what do you think you are doing? You are supposed to be here; get on the next plane and come down, or there will be trouble. The audacity of the request is sufficiently astounding, but even more so is that Arrow did it. He cancelled the rest of his vacation and came down and took his seat in the seminar.

The IMSSS was tremendously influential in the creation of economic theory over those two decades. And it was also very influential in my own career. Some of my best work was done during those two decades—much of it with very important input from the summer seminar at the IMSSS. Also, during those two decades I spent two full sabbaticals at Stanford, in ’75–’76 and in ’80–’81. This was a very important part of my life. My children used to say that California is their second home. Being there every summer for twenty years, and two winters as well, really enabled me to enjoy California to the fullest. Later on, in the nineties, we were again at Stanford for a few weeks in the summer. I told my wife there was a friend whom I hadn’t seen in a year. She said, who, and I said, the Sierra Nevada, the mountains. We had been there a few weeks and we hadn’t gone to the mountains yet. We went, and it was a beautiful day, as always. Many times during those years we would get up at 3 or 4 in the morning, drive to eastern California, to the beautiful Sierra mountains, spend the whole day there from 7 or 8 a.m. until 9 p.m., and then drive back and get to Palo Alto at 1 a.m.; exhausted, but deeply satisfied. We climbed, hiked, swam, skied.

The Sierra Nevada is really magnificent. I have traveled around the whole world, and never found a place like it, especially for its lakes. There are grander mountains, but the profusion and variety of mountain lakes in the Sierra is unbelievable. I just thought I would put that in, although it has nothing to do with game theory.

H: Getting back to the IMSSS summers: besides those who came every year, there were always a few dozen people, from the very young who were in the advanced stages of their doctoral studies, to very senior, established economists. People would present their work. There would be very exciting discussions. Another thing: every summer there were one or two one-day workshops, which were extremely well organized, usually by the very senior people like you; for example, you organized a workshop on repeated games in 1978 [42]. One would collect material, particularly material that was not available in print. One would prepare notes. They were duplicated and distributed to everybody there. They served for years afterwards as a basis for research in the area. I still have notes from those workshops; they were highly influential.

In all the presentations, you couldn’t just come and talk. You had to prepare meticulously, and distribute the papers and the references. The work was serious and intensive, and it was very exciting, because all the time new things were happening. It was a great place.

A: You are certainly right—I forgot to mention all the other people who were there, and who varied from year to year. Sometimes people came for two or three or four consecutive years. Sometimes people came and then didn’t come the next summer and then came again the following summer. But there was always a considerable group of people there who were contributing, aside from the three or four “fixtures.”

Another point is the intensity of the discussion. The discussion was very freewheeling, very open, often very, very aggressive. I remember one morning I was supposed to give a two-hour lecture. The lectures were from 10 to 11, then a half-hour break, and then 11:30 to 12:30. I rose to begin my presentation at 10 in the morning, and it wasn’t more than one minute before somebody interjected with a question or remark. Somebody else answered, and pandemonium broke loose. This lasted a full hour, from 10 to 11. After a few minutes I sat down and let the people argue with each other, though this was supposed to be my presentation. Then came the break. By 11:30 people had exhausted themselves, and I gave my presentation between 11:30 and 12:30. This was typical, though perhaps a little unusual in its intensity.

H: That was typical, exactly. There was no such thing as a twenty-minute grace period. There was no grace whatsoever. On the other hand, the discussions were really to the point. People were trying to understand. It was really useful. It clarified things. If you take those twenty years, probably a significant part of the work in economic theory in those years can be directly connected to the Stanford summer seminar. It originated there. It was discussed there. It was developed there in many different directions. There was nothing happening in economic theory that didn’t go through Stanford, or was at least presented there.

A: We should move on perhaps to CORE, the Center for Operations Research and Econometrics at the Catholic University of Louvain, an ancient university, about seven or eight hundred years old. CORE was established chiefly through the initiative of Jacques Drèze. I was there three or four times for periods of several months, and also for many shorter visits. This, too, is a remarkable research institution. Unlike the IMSSS, it is really most active during the academic year. It is a great center for work in economic theory and also in game theory. The person I worked with most closely throughout the years—and with whom I wrote several joint papers—is Jacques Drèze. Another person at CORE who has had a tremendous influence on game theory, by himself and with his students, is Jean-François Mertens. Mertens has done some of the deepest work in the discipline, some of it in collaboration with Israelis like my students Kohlberg, Neyman, and Zamir; he established a Belgian school of mathematical game theory that is marked by its beauty, depth, and sophistication.

Another institution with which I have been associated in the last ten or fifteen years is the Center for Game Theory at Stony Brook. The focus of this center is the summer program, which lasts two or three weeks, and consists of a large week-long international conference that covers all of game theory, and specialized workshops in various special areas—mostly quite applied, but sometimes also in special theoretical areas. The workshops are for smaller groups of people, and each one is three days, four days, two days, whatever. This program, which is extremely successful and has had a very important effect on game theory, has been run by Yair Tauman ever since its inception in ’91. In the past I also spent several periods of several months each during the academic year teaching game theory or doing research in game theory there with a small group of top researchers and a small group of graduate students; that’s another institution with which I’ve been associated.

I should also mention Yale, where I spent the ’64–’65 academic year on sabbatical. This was after publication of the work with Frank Anscombe, “A Definition of Subjective Probability” [14]; Frank was the chairman of the statistics department at Yale. At that time I was also associated with the Cowles Foundation; Herb Scarf and Martin Shubik were there. A very unique experience was the personal friendship that I struck up with Jimmy Savage during that year. I don’t know how many people know this, but he was almost totally blind. Almost—not quite. He could read with great difficulty, and tremendous enlargement. Looking at his work there is no hint of this. I again spent about six weeks at Yale in the late eighties at the Cowles Foundation, giving a series of lectures on interactive epistemology.

One more place that influenced me was Berkeley, where I spent the summer of ’64 and the spring of ’72. There the main contact was Gerard Debreu, who was a remarkable personality. Other people there were John Harsanyi and Roy Radner. In addition to his greatness as a scientist, Gerard was also well known as a gourmet. His wife Françoise was a terrific cook. Once in a while they would invite us to dinner; Françoise would go out of her way to prepare something kosher. Occasionally we would invite them. It was his practice at a meal to praise at most one dish. Sometimes he praised nothing; sometimes, one dish. That totally transformed a compliment from Gerard from something trivial to something sublime. Nowadays, I myself cook and give dinners; when a guest leaves saying everything was wonderful, it means nothing. Though I allow myself to be kidded, it really means nothing. But when a guest leaves and says, the soup was the most delicious soup I ever had, that says something. He doesn’t talk about the meat and not about the fish and not about the salad and not about the dessert, just the soup. Or somebody else says, this was a wonderful trout mousse. One dish gets praised. Then you know it’s meaningful.

I also spent a month at NYU, in February of 1997. It was interesting. But for me, the attractions of New York City overwhelmed the academic activity. Perhaps Esther and I took the city a little too seriously. This was a very beautiful time for us, but what surrounds NYU was more important to us than the academic activity.

* * *

H: Maybe it’s a good point to ask you, in retrospect, who are the people who have most influenced your life?

A: First of all my family: parents, brother, wife, children, grandchildren. My great-grandchild has not yet had a specific important influence on me; he is all of one and a half. But that will come also. My students have influenced me greatly. You have influenced me. All my teachers. Beyond that, to pick out one person in the family, just one: my mother, who was an extraordinary person. She got a bachelor’s degree in England in 1914, at a time when that was very unusual for women. She was a medal-winning long-distance swimmer, sang Shubert lieder while accompanying herself on the piano, introduced us children to nature, music, reading. We would walk the streets and she would teach us the names of the trees. At night we looked at the sky and she taught us the names of the constellations. When I was about twelve, we started reading Dickens’s A Tale of Two Cities together—until the book gripped me and I raced ahead alone. From then on, I read voraciously. She even introduced me to interactive epistemology; look at the “folk ditty” in [66]. She always encouraged, always pushed us along, gently, unobtrusively, always allowed us to make our own decisions. Of course parents always have an influence, but she was unusual.

I’ve already mentioned my math teacher in high school—“Joey” Gansler. On the Jewish side, the high school teacher who influenced me most was Rabbi Shmuel Warshavchik. He had spent the years of the Second World War with the Mir Yeshiva in China, having escaped from the Nazis; after the war he made his way to the United States. He had a tremendous influence on me. He attracted me to the beauty of Talmudic study and the beauty of religious observance. He was, of course, khareidi, a term that is difficult to translate. Many people call them ultra-orthodox, but that has a pejorative flavor that I dislike. Literally, khareidi means worried, scared, concerned. It refers to trying to live the proper life and being very concerned about doing things right, about one’s obligations to G-d and man. Warshavchik’s enthusiasm and intensity—the fire in his eyes—lit a fire in me also. He eventually came to Israel, and died a few years ago in Haifa.

The next person who had quite an extraordinary influence on me was a young philosophy instructor at City College called Harry Tarter. I took from him courses called Philo 12 and 13—logic, the propositional calculus, a little set theory.

H: So your work in interactive epistemology had a good basis.

A: It was grounded in Philo 12 and Philo 13, where I learned about Russell’s paradox and so on. We struck up a personal relationship that went far beyond the lecture hall, and is probably not very usual between an undergraduate and a university teacher. Later, my wife and children and I visited him in the Adirondacks, where he had a rustic home on the shores of a lake. When in Israel, he was our guest for the Passover Seder. What was most striking about him is that he would always question. He would always take something that appears self-evident and say, why is that so? At the Seder he asked a lot of questions. His wife tried to shush him; she said, Harry, let them go on. But I said, no, these questions are welcome. He was a remarkable person.

Another person who influenced me greatly was Jack W. Smith, whom I met in my post-doc period at Princeton, when working on the Naval Electronics Project. Let me describe this project briefly. One day we got a frantic phone call from Washington. Jack Smith was on the line. He was responsible for reallocating used naval equipment from decommissioned ships to active duty ships. These were very expensive items: radar, sonar, radio transmitters and receivers—large, expensive equipment, sometimes worth half a million 1955 dollars for each item. It was a lot of money. All this equipment was assigned to Jack Smith, who had to assign it to these ships. He tried to work out some kind of systematic way of doing it. The naval officers would come stomping into his office and pull out their revolvers and threaten to shoot him or otherwise use verbal violence. He was distraught. He called us up and said, I don’t care how you do this, but give me some way of doing it, so I can say, “The computer did this.”

Now this is a classical assignment problem, which is a kind of linear programming problem. The constraints are entirely clear. There is only one small problem, namely, what’s the objective function? Joe Kruskal and I solved the problem one way or another [3], and our solution was implemented. It is perhaps one of the more important pieces of my work, although it doesn’t have many citations (it does have some). At that opportunity we formed a friendship with Jack Smith, his wife Annie and his five children, which lasted for many, many years. He was a remarkable individual. He had contracted polio as a child, so he limped. But nevertheless the energy of this guy was really amazing. The energy, the intellectual curiosity, and the intellectual breadth were outstanding. A beautiful family, beautiful people. He made a real mark on me.

Let’s go back to graduate days. Of course my advisor, George Whitehead, had an important influence on me. He was sort of dry—not in spirit, but in the meticulousness of his approach to mathematics. We had weekly meetings, in which I would explain my ideas. I would talk about covering spaces and wave my hands around. He would say, Aumann, that’s a very nice idea, but it’s not mathematics. In mathematics we may discuss three-dimensional objects, but our proofs must be one-dimensional. You must write it down one word after another, and it’s got to be coherent. This has stayed with me for many years.

Picture 7. At the 1994 Morgenstern Lecture, Jerusalem:

Bob Aumann (front row), Don Patinkin, Mike Maschler, Ken Arrow (second row, left to right), Tom Schelling (third row, second from left); also Marshall Sarnat, Jonathan Shalev, Michael Beenstock, Dieter Balkenborg, Eytan Sheshinski, Edna Ullmann-Margalit, Maya Bar-Hillel, Gershon Ben-Shakhar, Benjamin Weiss, Reuben Gronau, Motty Perry, Menahem Yaari, Zur Shapira, David Budescu, Gary Bornstein

We’ve already discussed Morgenstern, who promoted my career tremendously, and to whom I owe a big debt of gratitude.

The people with whom you interact also influence you. Among the people who definitely had an influence on me was Herb Scarf. I got the idea for the paper on markets with a continuum of traders by listening to Scarf; we became very good friends. Arrow also influenced me. I have had a very close friendship with Ken Arrow for many, many years. He did not have all that much direct scientific influence on my work, but his personality is certainly overpowering, and the indirect influence is enormous. Certainly Harsanyi’s ideas about incomplete information had an important influence. As far as reading is concerned, the book of Luce and Raiffa, Games and Decisions, had a big influence.

Another important influence is Shapley. The work on “Markets with a Continuum of Traders” was created in my mind by putting together the paper of Shapley and Milnor on Oceanic Games and Scarf’s presentation at the ’61 games conference. And then there was our joint book, and all my work on non-transferable utility values, on which Shapley had a tremendous influence.

* * *

H: Let’s go now to a combination of things that are not really related to one another, a potpourri of topics. They form a part of your worldview. We’ll start with judicial discretion and restraint, a much disputed issue here in Israel.

A: There are two views of how a court should operate, especially a supreme court. One calls for judicial restraint, the other for judicial activism. The view of judicial restraint is that courts are for applying the laws of the land, not making them; the legislature is for making laws, the executive for administering them, and the courts for adjudicating disputes in accordance with them.

The view of judicial activism is that the courts actually have a much wider mandate. They may decide which activities are reasonable, and which not; what is “just,” and what is not. They apply their own judgment rather than written laws, saying this is or isn’t “reasonable,” or “acceptable,” or “fair.” First and foremost this applies to activities of government agencies; the court may say, this is an unreasonable activity for a government agency. But it also applies to things like enforcing contracts; a judicially active court will say, this contract, to which both sides agreed, is not “reasonable,” and therefore we will not enforce it. These are opposite approaches to the judicial function.

In Israel it is conceded all around that the courts, and specifically the Supreme Court, are extremely activist, much more so than on the Continent or even in the United States. In fact, the chief justice of the Israeli Supreme Court, Aharon Barak, and I were once both present at a lecture where the speaker claimed that the Supreme Court justifiably takes on legislative functions, that it is a legislative body as well as a judicial body. Afterwards, I expressed to Mr. Barak my amazement at this pronouncement. He said, what’s wrong with it? The lecturer is perfectly right. We are like the Sages of the Talmud, who also took on legislative as well as judicial functions.

H: Do you agree with that statement about the Talmud?

A: Yes, it is absolutely correct.

There are two major problems with judicial activism. One is that the judiciary is the least democratically constituted body in the government. In Israel, it is to a large extent a self-perpetuating body. Three of the nine members of the committee that appoints judges are themselves Supreme Court judges. Others are members of the bar who are strongly influenced by judges. A minority, only four out of the nine, are elected people—members of the Knesset. Moreover, there are various ways in which this committee works to overcome the influence of the elected representatives. For example, the Supreme Court judges on the committee always vote as a bloc, which greatly increases their power, as we know from Shapley value analyses.

In short, the way that the judiciary is constituted is very far from democratic. Therefore, to have the judiciary act in a legislative role is in violation of the principles of democracy. The principles of democracy are well based in game-theoretic considerations; see, for example, my paper with Kurz called “Power and Taxes” [37], which discusses the relation between power and democracy. In order that no one group should usurp the political power in the country, and also the physical wealth of the country, it is important to spread power evenly and thinly. Whereas I do not cast any aspersions now on the basic honesty of the judges of the Israeli Supreme Court, nevertheless, an institution where so much power is concentrated in the hands of so few undemocratically selected people is a great danger. This is one item.

H: The court not being democratically elected is not the issue, so long as the mandate of the court is just to interpret the law. It becomes an issue when the judicial branch creates the law.

A: Precisely. What is dangerous is a largely self-appointed oligarchy of people who make the laws. It is the combination of judicial activism with an undemocratically appointed court that is dangerous.

The second problem with judicial activism is that of uncertainty. If a person considering a contract does not know whether it will be upheld in court, he will be unwilling to sign it. Activism creates uncertainty: maybe the contract will be upheld, maybe not. Most decision-makers are generally assumed to be risk-averse, and they will shy away from agreements in an activist atmosphere. So there will be many potential agreements that will be discarded, and the result will be distinctly suboptimal.

H: But incomplete contracts may have advantages. Not knowing in advance what the court will decide—isn’t that a form of incompleteness of the contract?

A: Incomplete contracts may indeed sometimes be useful, but that is not the issue here. The issue is a contract on which the sides have explicitly agreed, but that may be thrown out by the court. Ex ante, that cannot possibly be beneficial to the parties to the contract. It might conceivably be beneficial to society, if indeed you don’t want that contract to be carried out. A contract to steal a car should be unenforceable, because car theft should be discouraged. But we don’t want to discourage legitimate economic activity, and judicial activism does exactly that.