History Home | Book Catalog | International Catalog of Sources | Visual Archives | Contact Us

Oral History Transcript — Dr. Laszlo Tisza

This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.

This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.

Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.

Access form   |   Project support   |   How to cite   |   Print this page


See the catalog record for this interview and search for other interviews in our collection



Interview with Dr. Laszlo Tisza
By Kostas Gavroglou
At Tisza's office at MIT
April 4, 1988

 
open tab View abstract

Laszlo Tisza; April 4, 1988

ABSTRACT: The beginning of Tisza’s career, 1925-1937; recollections and comments from his two-year study at Kharkov with Lev Landau. Discussion of important work in low temperature physics, 1938-1952, especially superfluidity; recollections of Fritz London. Discussion of work in the 1950s in thermodynamics and statistical mechanics; involvement with the philosophy of physics; his recent work in applying algebraic methods to problems in particle physics.

Transcript

Session I | Session II | Session III

Gavroglu:

So maybe we start now with your work on superfluidity, and maybe you tell us a little of your work on superconductivity.

Tisza:

Yes. As a matter of fact, I would like to start a little differently, which I think in fact should please you as historians, because the leitmotif of my biographical material, up to this point, is interactions with senior theorists like Teller, Landau and London, and this is — at the end of our correspondence with London, this simply stops. That is no longer present. And interactions are always important, and I'm very conscious of to what an extent, and so I told you of the influence of university surroundings in Budapest, Gottingen, Leipzig, and then the surroundings of Paris and so forth. But personal contacts are very important, although I feel that I am using them in a very personalized fashion, and I will give you this manuscript I am just working on. You will see this personalized. Nevertheless, I do trace the influences in that. One of the aspects of it was that I didn't have a very easy time to sit down and write a paper. I had ideas and worked on them and got—made some progress, and somehow it was not easy to jell into a work, and that was taken up by junior people, by students, just at that time, so when it was finished, then I got an interaction with a number of young people, Joaquin Luttinger, Herbert Callen and Martin Klein, three very good graduate students, who did their theses with me in '48 or '51 or something like that, I think the dates are in the book, and actually the first of these papers was on dipole lattices, which I think it's a very nice paper really.

Gavroglu:

With whom was that?

Tisza:

With Luttinger. Originally the idea came to me, another sort of influence. I had a very friendly relationship here at MIT with Hans Mueller. I don't know whether you know his name. You do? Yes. Now, Hans Mueller, I learned a lot from him on ferroelectricity, on phase transitions and then later on polarized light, and we had a lot of contacts until he suddenly died of a heart attack. There were dipole lattices, he talked about for [???] circuit(?) and things of this sort. So I started to think on that, and put together things which gradually became quite interesting, but didn't quite end up doing enough, and then these young — Luttinger and Callen emerged in my life in 1943, in the summer in a class, when they joined MIT as juniors, I think, and I gave my thermodynamics class, and that was an enormously successful class, and they appreciated me, and flocked around me, and then the two of them did a thesis with me. And Luttinger had a very lucky additional idea about the dipole lattices, and it ended up quite a nice paper. Callen, with I suggested to work out in more detail the Onsager relations for thermoelectric effects, which he wrote up alone. We didn't publish it together. He then wrote into his textbook and which was also quite successful. And Martin Klein came a little later. All of these were solid state papers, so it was quite a project going into that direction. But the superconductivity, frankly that started in such a fashion that, with being through with helium, I was looking for something else which I could manage on similar grounds, and I wanted to pull off a quick symmetry relations or something of this sort, and that's why it came about. And I had the bad fortune that it came out exactly the time when Frolich and Bardeen made their first theories, although these were not the definitive ones as you know, but nevertheless, it aroused worldwide interest. And then I was invited into the Bureau of Standards in '51 for a meeting, sort of confrontation, I sort of tried to explain what I meant, without really pushing it anymore, and so that was that. So I think it was a try which I did not consider very highly, and maybe there are a few amusing ideas in it, but not good enough. It certainly didn't address the substance of the problem of interaction. I tried to skirt it, and I think it was not a — it was a less successful attempt of this sort.

Gavroglu:

Before you embarked on your work in thermodynamics, there was a publication of three papers by Feynman on fluid, liquid helium, 1953. What was your reaction to those papers?

Tisza:

Well, he was nice to me in the sense that in his first paper put my name in capitals. He gives me credit. I was very ambivalent, because I never quite understood what he meant. I think it was kind of an in-between thing, which is neither formal nor phenomenological, which didn't appeal to me, and I had the feeling that [???] had a very strange method of associating physics and mathematics, and Feynman of course in the path integral method achieved that, in a most successful and impressive fashion, but the way he seemed to say, this is superfluidity? I just didn't understand it, and so I remember, I was a little embarrassed once when there was a discussion and at MIT Weisskopf asked me, can you explain it? And I said, "I'm afraid not." I didn't think I could explain it. And since then I lost interest in the subject, and I still can say only that I don't understand.

Gavroglu:

Was Onsager at the time playing any role for the work in superfluidity?

Tisza:

Well, he advised the experimentalists at Yale, C.T. Lane and Fairbank, and he gave the idea of the measurement of the second sound. Oh yes, he continued with the vortex idea.

Gavroglu:

That's right, yes, earlier.

Tisza:

Earlier he worked with the vortex idea. And otherwise, he invited me once to Yale to talk about my theory. It was early, before the Landau controversy came up. And that was sort of pleasant, and I talked in my intuitive fashion, and he seemed nice, but there was no very strong interaction.

Gavroglu:

Now I think we can go to work in thermodynamics, which eventually was your book, GENERALIZED THERMODYNAMICS.

Tisza:

Well, this started, I'm not sure if I talked about it in one of the earlier ones. I started to give a course in thermodynamics in '41, and gave it probably practically every year once on the undergraduate level, and every time I, when I brought it up, on it, and as I say this group, Luttinger, Callen and others in the group, and of course that gave me a big boost, and the general clamor was to write it down, but at that time, I was practically unable to write in English, to write a paper, particularly notes, and then, something happened that I had a friend here, a mathematics instructor, William Whitmore who was in the circle of Norbert Wiener and Santillana. He was in their friendly circle, and he was a mathematician, and not very creative but very intelligent and very alive intellectually, and when he saw the situation he said, "I'll help you to write it down," and for two weeks we went down on Cape Cod and stayed there, and started to write notes, where he continued to put it in good English, and in fact, there was something of an article in [???] which he had to work up and put in a form which got into the notes, and so at this point, I gave what I suppose, maybe three chapters, and a little later followed by a fourth, which I handed out to students. In fact, people in other departments asked me for the same notes. It was very popular, and there was a big demand on me to make a book out of it, and then Callen suggested that we work on it together, which I accepted. I would have always liked to have a Lifshitz like Landau, but that really never came about. But for a while, we started to do work on it, but then he left MIT, and went to University of Pennsylvania. We corresponded. We occasionally meet, visit to each other, and after a while I was sort of rather critical, I had ideas of what I, of how I wanted to do it, and precise conceptualized, and then he said that he was getting tired of it, and he gives me some time ahead if I want to write a book, he won't do it another year or so, but if not, he will go ahead. And so that is actually what happened. So he went ahead, and he wrote me a nice thanks, a dedication in the introduction, but otherwise, I saw nothing of it actually. Well, how should I say? When my retirement — not that one, earlier in '73, when I retired from MIT, there was a big banquet and symposium, and he was here, and he sort of apologized and explained that basically he was good for me in the sense that he wrote it down and popularized it, and to some extent that was true, because if I hadn't done it at all, then that wouldn't have existed anywhere. And it definitely gained a certain popularity, so — And then, the chairman of the MIT press, Carol Bowen — very nice man, he left a long time ago, asked me very hard to write something for them, and I said that I really don't feel like writing a book. Also, I was retired at that time. I was interested already in the quantum mechanics project, and we made a compromise, a reprint of papers, and I write a special essay for the introduction, and I think that was a good compromise, and it was successful. I put in the historical introduction for which I would have no other use, and I wrote some introduction.[1] Then I developed two papers for ANNALS OF PHYSICS, one on phase transition, which was sort of rather well developed. And the second was part way. (or, with Paul Quay) This, I need some introduction. I had some interactions with Benoit Mandelbrot. First of all, I met him in the early fifties when he was here in the research laboratory of electronics, and we had just friendly personal contacts but not professional interests. Then in '56, there was here at [???] an international meeting on communication and information theory, and Benoit came here, I think at that time he was not at MIT but came here on this occasion, and we started to talk, and I showed him my paper with Manning, which was based on a Gaussian method, and he told me, this is all wrong, the Gaussian method only approximates and he was interested in more general statistical estimation method, so I came into interaction with him and a few years I spent studying a non-Gaussian estimation method, and on that basis, the method was originally developed by Szilard and it was his thesis in 1929 or 1930, and then Benoit Mandelbrot made a big statistical deal about it, the method of sufficient statistics, I don't know if you are familiar with this kind of method. So it is not probability method of statistical estimation, it is more used by statisticians and very few physicists use it, really. And so I learned this method, and then I had a student Paul Quay, a Jesuit, whom I got interested in that, and this is another paper which is in this GENERALIZED THERMODYNAMICS, and I think this is an excellent paper, because it really shows how you can derive the Gibbs distribution without any approximation, invalid approximation, that is, kinetic methods of Gaussian or anything. Really, very general. In other words, my feeling at that time that thermodynamics should be not limited to macroscopic method, but should apply to any method, and at the same time should not introduce a model method. And so with the statistical estimation method, I learned a technique which was enough to give a really satisfactory foundation to the [???] distribution. So I believe that this is a very good work. I don't think it received much appreciation, because people are not interested in this kind of method. They want to derive it from classical mechanics or from kinetic method. You know, I did the opposite. At one point, I remember I talked to Wigner and he asked me, "Why don't you write a book?" And I said I don't feel like writing a book. But then when I had the GENERALIZED THERMODYNAMICS, I sent it to him, a copy, and then I think he said something like, "Well, it is not what I expected, but it's good to have it anyway," or something like that. He clearly disliked it intensely. My class always brought to me students who very enthusiastically wanted to work with me. I didn't have an unlimited number of good thesis topics. Usually these papers I'm talking about, because I think they're all contributions in a certain sense, they contributed part of a method at least or something or other, I'm not talking about these because it's not worth talking.

Gavroglu:

In the paper that Martin Klein dedicated to you, having to do with the history of thermodynamics, he says that you had mentioned to him Gibbs’ obituary on Clausius.

Tisza:

Yes.

Gavroglu:

And that that was quite a — you know, made a change in his life and understanding of things.

Tisza:

Yes.

Gavroglu:

This is a paper that he has published with another paper.

Tisza:

Yes.

Gavroglu:

So also you have not only quite a lot of knowledge on the history of thermodynamics, and thermodynamical theory, but also you have particular preferences concerning, the way people approach the problems and the attitudes of the particular people. Maybe you can tell us some things on the thoughts you had on the history of that particular subject, which even today there isn't any consensus among historians and philosophers as to the —

Tisza:

— yes. I have a paper on that. Are you familiar with that? Which I presented at a meeting of the American Philosophical Association.[2] — it is certainly quoted in my Feschshrift. And I recommend it very much, because I think it's a rather good paper and summerizes the essential points.

Gavroglu:

This is "Apology for Physics"?

Tisza:

No. No, that is a short paper of — it's OK, but nothing very much. But the other one I mentioned, I worked on it quite hard, and — oh, "The Foundation of Statistical Mechanics," in Soppo and Asquith, editors, Philosophy of Science Association, 1976, pp. 585 and 617.

Gavroglu:

No, I don't have it.

Tisza:

Yes. I think — I have a reprint of it I will give you. I will give you a reprint, because before I go into great length, I think I must show that to you. But something's come to my mind, some thought that, I have recently talked about these matters. I was invited just for a luncheon meeting with the Einstein Project, and they asked me about the second volume, about Einstein's statistical papers, and in this connection, I gave them, I don't know how much exactly they used, but I talked to them, so if you like, I could elaborate on that.

Gavroglu:

Yes.

Tisza:

What I liked in Gibbs’ obituary of Clausius, and Martin liked also, it is certain statements that Gibbs made which had a marvelous view of how mathematical physics should be defined, the point of view from which the phenomena appear the simplest, and then Clausius had this fine sense of discrimination, which separated truth from falsehood. So that was very interesting. I was always very much interested in analyzing the work of people into various ingredients, many of which are often in contradiction. Of course, historians should think about it. They designate theories and sets of ideas in terms of people, but that is sort of designating chemical substances in terms of the geological sites, rather than according to chemical composition. Now, the geological site of materials is interesting. It's interesting how it comes up. But one should be able to separate it into constituents. So Gibbs, to me, was an enormously important person. At the same time, I felt that he was limited in a strange fashion, in a mechanical view. Now, I have expressed this matter in detail, and I call your attention to it, in GENERALIZED THERMODYNAMICS. There is a place on composite systems, and — well, at any rate, you can find an article on composite systems, where I describe this whole situation, that Gibbs develops both his thermodynamics and his statistical mechanics in terms of a single system, and not a composite system, that gives an enormous amount of confusion, because it is practically impossible to say what is the work and heat contributed to the system if you don't have clearly a composite. A composite system. Now, I know what the origin of that is, because if you really take a mechanistic view, then you have to have one single phase space and everything is a process in that phase space. And if you have a calorimeter and if you open a valve, or if you unite two systems into one, then you go from one phase space to another, and that is not a mechanical process. Now, I think that is the greatest catastrophe of mathematical physics, both for the foundation of statistical mechanics and quantum mechanics. In the sense that the famous problem of entropy increase would really be a non-problem, because if you realize the theory of a system and you have two systems, and with the systems combined into one box and not into the other, and you open a valve, then you change the phase space, and that causes irreversibility and there is no mystery about it at all, and there is no difficulty of explaining it mechanically. But the process of opening it is not a mechanical process. In other words, the problem has not been solved. And then, many people like Pauli and Fierz and von Neumann tried to solve it in a single system, as if that would help, if they solved the problem in a single system, but it doesn't correspond to the experimental facts, because the experimentalists manipulate the system and this manipulation of systems is part of the story. If you have a chemical factory you have this manipulation. If you have a living organism and you have the DNA and you manipulate systems — in other words, the fact that we have thermodynamic operations when we manipulate the system, and then we have molecular process which takes place under the conditions, that's something else, and then we again change manipulation, there is an interplay which gives rise to very complex phenomena. And these mathematical physicists who pretend that they are fundamental, they think that they should not use that, because things then become too simple. It is a challenge to solve it without it, and I was in a terrible fight against people who sort of tried to make a big issue out of that solution and looked down on me. The fact is, of course, the way I solved such a problem, there was not much to it, because I eliminated the problem. What I was trying to do was to eliminate this cottage industry, that you can do always a paper on the increase of entropy. You cannot go wrong if you publish things of this sort. But this is the type of thing. Now, this has an important role in the paper which I will give you, which is now being xeroxed, in the quantum mechanics of magnetic moment and spin. I don't want to get into that now, but the same thing comes in on a very deep level in particle physics, and it all comes from this strange sort of — I don't like to call it metaphysics, because I think that metaphysics is an honorable activity, — call it rather mythology. It is mythology when you insist on explaining the world in terms of your familiar experiences, whether it applies or not. Now, that may be naive and artistic, like Greek mythology, or it may be complicated and ugly, as the mechanistic mythology which appears here. So that is basically what thermodynamics meant for me.

Gavroglu:

Later on, you had this paper on the methodological aspects of physics.[3]

Tisza:

Yes.

Gavroglu:

Which was a REVIEWS OF MODERN PHYSICS paper.

Tisza:

Yes.

Gavroglu:

Which at that time, of course, was totally unusual, to have those kinds of papers. Maybe you can tell us a few things, not so much on your views, which are extremely clear in this paper and the subsequent paper in this volume, but how the physics community reacted to that kind of a problem, I think, of a person who up to then was one of the community, was the "normal" physicist?

Tisza:

OK. I can tell you one definite statement on that effect. You see, in 1962-63 was an interesting year for me.[4] I was on a sabbatical year in Paris, and just before leaving, I published this paper in the REVIEWS OF MODERN PHYSICS, and I think from that I published a paper with Quay, so I got some important things out, and this paper came out. I visited during that time also in Copenhagen, and met there my old student Luttinger, who originally considered me sort of, thought I was great, but at this time, we were in the Tivoli Garden and had dinner in Copenhagen, and sort of a nice summer evening, and he said, "Hey, what are you doing? Everyone says you are crazy." So that is what I heard. Now, my secretary got maybe six, seven hundred requests for reprints. I didn't have enough. I mean, I couldn't fill all the requests for reprints. So I got a large number of people who asked for reprints, but never the people who counted. In other words, the leading theorists considered it was crazy. And there were many people who had a common sense understanding for the problems, and tried to understand what I had to say. But they don't quite understand it but it appeals to them. So that is what happened. So it was of great interest among not very high class physicists, but a complete rejection by high class physicists.

Gavroglu:

What about philosophers of science?

Tisza:

Well, that is another question.[5] Actually, it started first, because I published — I gave a talk at the BU colloquium in December, '61, so two years before I gave the paper first there. Now, I can tell you, that was sort of an amusing thing. I felt that my paper is to some extent an attack against a rigid positivist view. Now, the talk in this colloquium, Philip Frank was there, and with Philip Frank I was on very friendly terms, and Philip Frank, was not a doctrinaire positivist in any way, and so I had a little bit of the nagging feeling that I must be very careful, and I should not offend him. Therefore in my phrasing, I sort of tried to lessen the contrast to positivism, in contrast to Kuhn , who ironically attacks empiricists, I tried to be nice and sort of smoothed out, and I said basically it is consistent to say it is only a little different, and so I gave the paper, and Philip Frank was there. Bridgman died shortly before.

Gavroglu:

I don't know.

Tisza:

Well, it is printed here, in this volume, also the ones who were present and participating in the discussion, so you can see but I think Bridgman was not there anymore. So I felt at that point that I wanted to loosen this deadly boring grip of empiricist positivism, and introduce a kind of metaphysics in this metatheoretical basis, and I tried to present it in an innocuous fashion. The talk was well received. There was a lot of discussion and a great deal of interest. I suppose the most critical part was sort of a feeling, "but say, you still want to have some kind of unification," or something of this sort, with a grand system, and I really ducked the issue at that point. I had no answer, didn't know — I was sort of presenting what I had. So there was no difficulty with the philosophers, but I don't think anyone paid any attention to it. Cohen once told me that there were some reviews, and he promised to send it to me and never did. But whatever it is, there's no question about it that there were no repercussions.

Gavroglu:

Has there been any change in your views between the Boston colloquium paper and the REVIEWS OF MODERN PHYSICS paper, since you have a change in title? One is (crosstalk) and the other has a conceptual foundation.

Tisza:

I think it was just an artificial thing which I easily made, I just wanted to give it a different title, to make a distinction. And in the REVIEWS OF MODERN PHYSICS paper, I wanted to address it more to physicists, and that's why I put in physics details. It was where I showed how quantum mechanics can contribute to a unification and organization of physics, so that was what I added that was not there in the first version, so I just added it in order to be more finished material. I don't remember any change of view. One thing I may add to the reception — how is it possible that it was published? And that I happen to know the answer. Because Walter Elsasser as I said, I told you once, you know?

Gavroglu:

I know ... yes...

Tisza:

Yes. There are three places where he preceded me and I came after him. Finally I caught up with him in Philadelphia, when Callen invited me to give a talk, and there I actually met him, so we were sort of feeling like old friends. And we have a real common good friend Egon Orowan, who is a Hungarian applied physics here, a professor at MIT. And we met again, and then he told me what happened. When I submitted this paper to the REVIEWS, Condon showed it to him and said, "Look at this paper, what should I do with it?" And he said, "Publish it." So I guess that is what happened, I sneaked by in such a fashion. But there was no repercussion, no influence, neither physicists nor philosophers paid any attention to it. Now, Abner Shimony [???] likes it very much, and of course in his fashion he tried to popularize it and I think he said he tried to convince philosophers to pay attention. At one time, in connection with the [???] I met with Joseph Agassiz at BU. What do you think of him?

Gavroglu:

He's an important figure. His contributions have been important. But I don't know whether they will be lasting. But at that time, he was a dominant figure in the trade.

Tisza:

Yes. He is a colorful person. I think he's sort of flaky. So at one time, I remember I had luncheon with him and Feyerabend. Feyerabend was very amusing, he had an interesting style, much more than Agassiz. I tried to explain to them. I had a sense that they didn't catch it at all. Also, some time ago, maybe a couple of years ago, I read a Feyerabend article which I liked, and I asked him if he'd send me a reprint... So I don't know if there were difficulties. But there was no real contact. In other words, the — what I published with Irving Manning was with correlations, but the Gaussian Markov process basis, and I didn't find it possible to deepen all these ideas into a synthesis. When I went, started out on my sabbatical year in the summer of '62, then I picked up a book in the Harvard Coop by De Groot and Mazur on irreversible thermodynamics, and I thought that it was the type of thing I will work on, and I took the book home for a day, turned the pages in it, and I felt so repelled by the inelegance, and it just had absolutely no appeal. And at the same time in the Coop I found another book by Roman on elementary particles, and you know it? And I did toy with elementary particle ideas, but I found that the whole literature on it is above me. I just couldn't connect. I couldn't find any point of contact. Roman's book was somewhat different, because he started in an algebraic fashion which appealed to me and which connected already. I'm sorry, no, he didn't connect, it just appealed to me, I thought that was the type of thing I would like to do, and I went to the Coop, I exchanged the De Groot book for the Roman book. I took that along to Paris then and in Paris then I started to work in this direction. So that is now a new chapter which I — but the transition occurred in a rather dramatic fashion in a single day.

Gavroglu:

OK.

[1]The replacement of the classical mechanical foundation with a postulation operating in the thermodynamic context opened up new vistas and I farmed out my problems to graduate students who often had a better mathematical background than I had. The more successful of these ventures are reprinted in my book, Generalized Thermodynamics. There were two major themes to emerge from these interactions. The first of these is the circumstance that in Landau’s group it was taken for granted that Landau’s inspiration was Gibbsian, both in the phenomenological and the statistical versions. It took me some time spent building up my course to realize that Landau was primarily influenced by Einstein’s statistical thermodynamics as it is implicit in research papers rather than in any text. Einstein’s approach had a more intuitive appeal, but it did not reach the generality of Gibbs’ austere postulation of more powerful results. Among the papers reprinted in GTD there is one in which Einstein’s intuitive ideas are fused with Gibb’s formal generality. The key to this fusion is a method of statistical estimation to which I was introduced by Benoit Mandelbrot who also called my attention to the fact that Leo Szilard used this method to connect thermodynamics and statistics. In his theory these disciplines were in the versions of Carnot cycles and Boltzmann statistics respectively. My idea was to translate all these into a Gibbsian format and show that the method of statistical estimation that experts call sufficient statistic was a tool to connect the phenomenological and statistical versions of Gibbs, which are singularly unconnected in their original versions. I suggested the topic to a student of mine, Paul Quay, for his PhD thesis. Our paper (Stat. Thermodynamics of Equilibrium) is reprinted in GTD. I think it is a fine paper, but it is not widely appreciated because its fundamental acceptance of statistics is in conflict with the Boltzmannian tradition of reducing statistics to deterministic mechanics. This tradition had powerful champions, thus Einstein’s insistence that “God does not play dice” and Schrodinger’s attempt to transfigure the random aspect of the cat’s death into a deterministic superposition of life and death. By contrast the Tisza-Quay theory holds that the certainty of death is consistent with the uncertainty of its instant. The last two papers of the book make a case in favor of changing obsolete traditions.

[2]Yes, I have a paper in the Journal of the American Philosophical Association, 1976, eds. F. Suppe and P. D. Asquith, pp. 585-608, 617-619. It is an easier summary of my rather heavy paper with Paul Quay. My Generalized Thermodynamics was undoubtedly the realization of my resolve to work along the line of Landau’s pursuit of thermodynamics. However, I had also my more specific agenda: to establish thermodynamics in a central role in the overall structure of physics. I was rather successful in connecting statistical estimation with the canonical distribution function. I was successful in the sense of satisfying my own standards. I did not succeed to convince the community that there is an objective criterion to influence a change of standards. This task is still ahead of me. I note that Gibbs’ first great discovery was the fundamental equation which has different versions connected by Legendre transformations. This is the phenomenological theory. The second Gibbsian discovery: in statistical mechanics a closed thermodynamic system can be represented as a mechanical system of its molecules. Mechanical processes in closed systems do not involve entropy increase. How to account for entropy increases in real-life processes? According to a tradition established by Boltzmann and Ehrenfest there are mechanical processes that increase entropy. By contrast, according to Tisza and Quay entropy increase in composite systems in which the relaxation of internal constraints leads to a new system, an operation that is not a mechanical process in a closed system. I was even less successful in dealing with the statistical aspects of quantum mechanics.

[3]Gavroglu was familiar with my paper in the REVIEWS OF MODERN PHYSICS on this matter and he asked me for some details about its publication history. He was aware of the nonconformist nature of the paper and wondered how it got published and how was it eventually accepted. It is easier to answer the first question. The key is my late friend Walter Elsasser. He was a Berlin physicist, slightly older than myself. He is best known for being the first theorist to pick out de Broglie’s thesis for attention. He also claimed that Davisson had already an experimental confirmation. He was at that time in Gottingen, where I followed him a few years later. This became a pattern, I followed him in Kharkov and then in Paris. I eventually caught up with him in Philadelphia where I was invited to give a seminar talk. Ever since we have been friends and ran into each other occasionally. He told me that Ed Condon, the editor of Review of Modern Physics, asked his advice as to my unusual paper. Walter suggested publication, so I slipped by somewhat by accident. I visited at that time in Copenhagen and was quite unsuccessful to drum up interest. My friend and former student Joaquin Luttinger told me: everyone in Copenhagen says you are crazy. Meanwhile my secretary at MIT mailed out the almost thousand available reprints in answer to requests. All those who counted, were against it, but many who did not count were intrigued.

[4]Here the first step was to work my way out of a personal misconception. For a long time I was under the impression that I need irreversible thermodynamics to connect with quantum mechanics. On leaving for my sabbatical leave for Paris in 1962, it was a moment of decision. De Groot and Mazur came out with a new book on irreversible thermodynamics. I bought the book with the intention of getting deeply involved in it. However, at the same time I became aware of Paul Roman’s book on particle physics that was centered on group theory and the spinorial representation of the Lorentz group. Whereas I could not consider myself a particle physicist, the algebraic slant taken by Roman seemed to me less forbidding than a more typical text on particles. During my year in Paris I pursued a double orientation. My official connection was towards Michel Magat’s group working on physical chemistry and thermodynamics, but I also participated in the seminar of de Broglie where my main contact was, J. P. Vigier. Although I did not find here a theory ready to accept, I started to think in terms of the algebra of the Lorentz group which I think led me toward a sound long-range program that is beyond the scope of this interview. There is a general feature that characterizes my methodological difference with the community. The world is full of diversity. The traditional reductionism tends to reduce the apparent diversity to a deeper unity. By contrast, I tend to make it more acceptable to live with a fundamental diversity.

[5]On a few occasions I addressed philosophers; primarily in December 1961 at the Boston Studies of the Philosophy of Science, Vol. 2. My main concern was that I emphasized the importance of concepts and was contradicting the positivists’ dismissal of concepts as evil “metaphysics.” I had very friendly relations with Philipp Frank who was not a doctrinaire positivist and I wanted to avoid offending him. Another friendly opponent was Bridgman, died earlier this year. My talk was well received, but there was no lasting impression among philosophers. I was friends with the late Richard Martin, a logician, who once invited me to talk at the Metaphysical Society. And, of course, the discussion of the foundation of statistical thermodynamics discussed in footnote 2.

Session I | Session II | Session III