Leon Rosenfeld - Session III

Kuhn:

I suggest, Professor Rosenfeld, that we go right back to where we were at the end of the last time. I had asked you then the question, “What was working with Bohr like?” And this, of course, is all sorts of questions, but I’d like you to talk about it just as personally as you can — hours of working, method of working, interplay of the two parties to the work, what determined when you worked on this and when you worked on that, and so on.

Rosenfeld:

Whenever one speaks of those things, one hears the opinion that Bohr needed an interlocutor, or rather, as we would say, a “Klangbord,” a resonator. There is something in it, though only a small part of the truth, because Bohr was very able to work alone; in fact he did most of his thinking alone, or perhaps not most, but at any rate a great deal. Strangely enough, much of his thinking was done during the night. How do I know that? Well, because it usually happened that when we left the thing in the evening, completely exhausted and usually in a mess —.

Kuhn:

About what hour would that often be?

Rosenfeld:

Oh, there was no definite time; it could be midnight, or it could be one o’clock in the morning, or earlier. Very often it went on, let us say, to midnight or one o’clock. Then we would leave the thing with the impression that we did not understand a bit of it, that the whole thing was a complete mess, and Bohr would conclude saying, “We shall sleep on it.” [R. repeats the phrase in Danish.] But “sleep on it” meant that he had a sleepless night about it; his thinking about it went on all night because the next morning, he said, “I thought of that, and it must be so.”

Kuhn:

How sure are you that in fact he did not sleep, because there are also all the reports of people who go to bed with the elements of a problem in their mind and wake up with a solution.

Rosenfeld:

No, it was conscious thinking. This unconscious work has been often written about; Poincare has a famous story about it, and probably it went on also in Bohr’s case, but he never mentioned anything like that. But whether there was conscious sleeplessness we might ask Fru Bohr occasionally: I think I remember that he sometimes complained that he did not sleep and she was a bit disquieted about his sleeplessness. He was certainly of a very nervous temperament, very excitable, and also the way in which he was physically tired showed that he must not have slept very much. Sometimes Margrethe intervened, saying, “You must cease, you must have some rest.” So that is my inference; that might be checked perhaps.

Kuhn:

When she intervened that way, was she generally successful? Would Bohr then stop?

Rosenfeld:

No, it was quite impossible. If Bohr had an idea, especially an unresolved problem in his head, it was quite impossible to put him to any other thing until he had come to some kind of conclusion. He never was satisfied with any solution or thing that he had decided until he had turned it over in his mind for days and weeks on end, in all possible directions; and he was always ready to abandon something which looked very nice and fine and that most people would have been happy about. He threw it out without the least regret whenever he saw any improvement, even the smallest, subtlest nuance. That was very impressive. And he had not the slightest respect for any authority or anything that was taken for granted. I remember at one stage, when we could not find any reason for this minus sign in the formalism, he repeatedly asked: “Is this minus sign a unique feature of the formalism? By a modification could one not get a plus sign?” Then I thought about it and I said: “No, it’s quite impossible; it’s straight from the conclusion, from the commutation rules; there’s no escape from it; it’s quite uniquely fixed by the formalism.” And he seemed to rest on that, but one day, I came to the office where he was working; the door was open, so I did not knock but came in directly. He was sitting at his table, and when I came in, he was startled and tried to conceal the book that was before him. I did not try ostentatiously to look at it, but I recognized it, of course. It was the volume of the Zeitschrift fur Physik, opened to the article of Jordan and Pauli where these commutation rules were derived, so he did not like to show me this! My positive and definite statements were taken with a grain of salt! He was not prepared to accept even a calculation by Pauli and Jordan if there were any indication from other considerations that there might be some doubt about it.

Kuhn:

How was the minus sign finally eliminated?

Rosenfeld:

It was eliminated when he had this thought that since this minus sign only gave trouble when there was a double communication between the two test bodies, that a signal could be sent from the one to the other and then the signal could be sent back from the second to the first within the, time interval. So it occurred to him that if that were possible, we might make use of this signaling in order to allow the second test body to tell something of its displacement to the first; then very ingeniously he arranged that. It had to be fully automatic, since the displacement was uncontrolled, but it could signal itself, could be controlled purely automatically; it turned out that this was sufficient just to produce this minus sign and nothing more. I mean, it was just exactly that. That was a most exciting discovery. Needless to say, those things were entirely his work; I couldn’t do anything of that kind — that was absolutely beyond me. All I could do was to make the calculations and to assure him that it was unique!

Kuhn:

Probably there was no such thing as a typical day, but how would the working day go? You would come to him?

Rosenfeld:

Yes, I would come to him in the morning —.

Kuhn:

About what time?

Rosenfeld:

It was different, because there were disturbances, or whatever you call them, so it was not very regular. The most regular was when we were in Tisvilde, when we started right after breakfast in the morning, around 9:00 o’clock, and then we went on until late in the night. However, there were interruptions, depending on how the work was going on. Sometimes, when it was very difficult and we got stuck, he would decide — at least if it was in Tisvilde — to take the bicycles and go to the woods, the fields, or somewhere around. I remember occasions when it was rather adventurous to cycle with him because, while cycling, he would go on thinking of the thing and then he would stop suddenly without warning and tell me, “Would you note down that sentence; it’s better than what we have.” But sometimes we would also talk of quite different things, for he had completely the word on all possible things.

Kuhn:

What was working with him like? How did the two of you interact?

Rosenfeld:

The usual method was his dictating whatever he thought about. He always insisted that all that he was dictating in that way was meant to be questions rather than statements, and in fact, it was just a trial; just as we jot down formulae to see how they look and try out various approaches to the problem, he would jot down sentences or bits of sentences and then look at them to see how they looked. Then he would try again. He worked with concepts, with words, as we work with formulae. He could also work with formulae, incidentally; one should not create the legend that he was unhandy mathematics. He was extremely clever in handling formulae and in getting the implications of a given formula or even solving; he would guess the form of the solution of a differential equation, though he would not, of course, write it down in the form of a Bessel function or that kind of thing. Perhaps Bessel function is just an unfortunate example because those things that he had used in his youth, such as the contents of Rayleigh’s Theory of Sound, he had at his fingertips. That is what he had used extensively in his early work, but all the fancy things, like Laguerre polynomials, he didn’t know; nevertheless, he could draw a curve representing the solution of the differential equation just from the equation, saying “now the second derivative goes up, then the maximum goes down,” and just from that he could construct a solution and leave it to me to work it out in formulae. It was rather trying to do this work because he was impatient to know the results, so I was always working them out under pressure. But he was also indulgent when it came out wrong and I came the next day, saying, “Now it’s different.”

Kuhn:

Not like Pauli in that respect?

Rosenfeld:

No. I have already said that he was much less easily satisfied than I; for instance, I thought that when he had dictated a beautiful sentence expressing the conclusion of all the speculations that we had had, it was finished. But then he would sleep on it, and the next morning, the whole thing would be thrown out and taken up afresh. It was very seldom that one had to go back to an earlier version; it was continual improvement, polishing completely.

Kuhn:

Was the change in this respect usually a verbal change in the formulation for clarity, conciseness, and so on, or were there also substantive changes? Was this a dissatisfaction with the expression?

Rosenfeld:

Well, it depended. It could be both, or either, depending on the situation. With those fluctuations it was certainly substantive changes to the last because we had not really understood what those fluctuations were until the very last.

Heilbron:

Then the polishing wouldn’t just be a question of changing a word here or there, but he would begin at the beginning and dictate anew?

Rosenfeld:

It depended; it could be either also.

Heilbron:

Did you get annoyed on occasion with the unceasing rearrangements of things you considered adequate?

Rosenfeld:

No, not usually, because since all those things were preceded by discussion, I usually understood what the reasons for the attempted changes were. Sometimes what caused a bit of annoyance was when he came in the morning, and said, “Well, we shall start afresh,” when I thought that we had finished the thing. But even then, when I very soon saw that the new attempt was still better than the previous one, I reconciled myself to the situation.

Heilbron:

Were these subtle changes, which you were able to appreciate, generally known and felt? I mean, could one who was not so sensitive to Bohr’s moods, manners of expression, and so forth, see the reason for these detailed changes?

Rosenfeld:

I don’t think so. Well, experience shows that he is very little understood; in fact, only those people really understand his attitude who have had the occasion of conversing with him. Certainly, the written text is too condensed to —. It was quite typical that when he had a statement of an idea which I thought was very clear and unambiguous, you could bet anything that it would be changed and made less unambiguous and less clear, clear in the Cartesian sense. He was absolutely anti-Cartesian, and that is absolutely essential in his attitude. He could dictate that statement — that was his first idea — and it looked like a very definite statement, but when he saw it, probably what went on in his mind was to think of the complementary sides of it. So he felt the need to put this haze about it which would give an opening to the complementary side. When de Broglie say that Bohr shrouds himself in “le brume du Nord,” he is absolutely right because a Cartesian cannot appreciate this kind of thing at all. And you could never pin Bohr down to any statement; he would always give the impression of being evasive, and to an outsider who didn’t know him, he would make a very poor showing, because that’s not the usual way in which people discuss between themselves. But those who understood the background of the motivation found food for thought and a warning that every statement has its limitations.

Kuhn:

When you were working with him, you speak of his dictating sentences or scraps of sentences for consideration, to see how they looked. Were these always from the beginning sort of tentatively parts of a piece of written work?

Rosenfeld:

Yes. He was always starting. “Now we shall write a little paper about it.” When we started with Landau and Peierls, he said, “What do they mean actually,” and so on. And finally when he had seen, for instance, that you had to take mean values over finite domains, he was quite excited and he said, “We shall write a little note about that to set that right” So immediately we started a project of the small note. He always had to have a complete task which he decided himself. “Write something. Write a communication.” That is the theory of communication.

Kuhn:

But almost all of the actual research was done in the guise of actually preparing a paper or a note or something of this sort?

Rosenfeld:

Yes. It was not at all as most of us do, I suppose, with first scribblings, exploring, trying to find out, and then when one thinks that one understands, setting out to write it down. That was not at all his method; he started without even knowing where the thing would lead. Immediately he started writing a paper. It was a bit disconcerting, but that was his way. So I suppose in that paper there were many beginnings announcing quite different aims.

Kuhn:

As the paper developed, having started from a note and ultimately turning into something quite different, did one, as the goals were reformulated, always go back and start the paper again? Would he always have the beginning of whatever paper he was writing in hand before he would go on to the next stages of it?

Rosenfeld:

Yes. He felt the whole thing as a flow, you see. Of course one knew one could use scraps of the previous thing. He would never think of saying, “Well, we’ll leave that and come back to that later.” That never arose.

Kuhn:

That’s terribly interesting.

Heilbron:

His bigger papers, the review sort of things, like the Quantum Theory of Line 279 Spectra, and the long article on atomic structure —.

Rosenfeld:

I have no experience with those, but I imagine that they were written in exactly the same way. Whenever a question arose, for instance this paper that he read in Rome about the limitations of quantum theory and a paper later in Warsaw also, about the limitations in which he had only very general suggestions to make because he did not bother to go into the formalism, he would start by saying, “Let us try to summarize what we know.” Then he told me that, “to summarize what we know is the best way to start to think about a subject. That is how I did it in my early papers, trying to summarize what we know about the spectra. Then the unsolved problems and the weak points will occur in this attempt to summarize.” So he would start from there.

Heilbron:

And the summary would all be written down?

Rosenfeld:

Yes, in the same way, I suppose. That was then with Kramers who is no longer here to tell us, but I imagine it was this method. You see, his early manuscripts were written down by Margrethe. Of course, with Kramers, myself, or Casmir it was different in the sense that the man could react and discuss, whereas Margrethe was more passive, so in those early days, he was struggling all alone without any help, or without any resonance of any kind, but with the same method. Try to put down what one knows in order to see the possible harmony or lack of harmony between the various points. He had to have expressed it, not put it down. He very seldom read what he dictated; he had all that in his head. Whenever a thing was formulated, it was expressed in words, it was in his head. He could go alone into the woods with not a scrap of paper with him and then he could come back and say, “We shall change this work in another way.”

Heilbron:

These would be long sections that he would be able to consider in that way?

Rosenfeld:

Yes, he was able to have all that in his head.

Kuhn:

You’re reasonably clear that he would actually have the whole thing that he had dictated more or less verbatim in his head; it isn’t as if there would have been one thing that bothered him when he said it that he would hold on to?

Rosenfeld:

Well, I cannot guarantee that he had everything verbatim, but anyhow, he was jolly good at remembering!

Kuhn:

You speak of the opportunity for a resonance with somebody who was technically competent or could comprehend the material he was taking down. How did that work?

Rosenfeld:

When I objected or put a question, even if it was a stupid one, he would not let it pass. I think that there was no question, interruption or remark which had not as a consequence a change, were it only of one word or a nuance in the sentence. He took account of everything.

Heilbron:

He would then respond directly to questions during the course of the writing of a paper?

Rosenfeld:

Yes. Sometimes he would say, “Wait a minute, and when I’m finished, you shall see better what I mean,” but sometimes he reacted immediately.

Heilbron:

I understand that in the course of a conversation he would not answer questions or would continue in his line of thought —.

Rosenfeld:

Well, sometimes he did, but then it was with good reason. He saw that the question was premature, that he had not finished his argument, and then he said so. “Wait until I have finished my argument and then you will see.”

Kuhn:

What sort of arguments, suggestions, changes, might originate directly from you or from someone else who was assisting him?

Rosenfeld:

It’s difficult to say. Klein once remarked that there was a noticeable change in his philosophical papers between the time when Klein was his helper and the time when I took over, and this may be true, I don’t know. That might be scrutinized. Klein surely has a certain tendency to, shall we say, idealism and even mysticism, and I am the opposite. Klein told me that he noticed the difference, and surely in the early papers, there were sentences which were a bit queer. My attention was called to that by Grunbaum. In the course of these discussions with Bohr I chanced to have some correspondence with Grunbaum who was very clever. Grunbaum has understood the point of Bohr’s theory better than many people, and he wrote to me once, “You tell me that Bohr’s fundamental attitude was not idealistic but rather materialistic.” Unconsciously, of course; Bohr would never hear of those words. Bohr regarded the question of the foundations of physics in the external world as something so trivial that it was not worth discussing. Grunbaum wrote, “Well, I believe you, since you tell me, but what do you make of the following sentences?” I should find the letter — I kept it preciously. He had a whole list of sentences taken from Bohr’s papers, and when you look at those sentences, you can understand, at least, that outsiders and casuists, like those pseudo-Marxists there, can accuse him of the worst kind of idealism and even solipsism. Those were just unhandy formulations, of course; he had not realized the implications of all those sentences.

Heilbron:

Do you remember any of the sort of things?

Rosenfeld:

Those were in fact in the early papers, before my time.

Heilbron:

They wouldn’t get by you.

Rosenfeld:

I was not responsible! Sometimes I could not prevent him from writing something which I did not like; that happened to me.

Kuhn:

Can you remember an example of that?

Rosenfeld:

No, I should find this letter of Grunbaum.

Kuhn:

Do you remember any discussion that you yourself had with him in which he put something in a way that you didn’t like but you were unable to get him to change it?

Rosenfeld:

Sometimes it was what I call his being too polite or too mild to other people. We know that he was very anti-clerical but he never expressed his anti-clericalism. In fact there are passages, even in his latest notes, when he mentions religion, which are unimpeachable if you scrutinize them, because they are formulations of the complementary aspects and he does not commit himself to anything. But they look very much like a sort of acknowledgement that there is something in it. He didn’t like to antagonize people, not because he was afraid of people — in fact, he was very firm when the occasion arose and when he saw that he had to be firm — but he felt that whenever you come with a definite statement about anything, you are betraying complementarity.

Kuhn:

You don’t remember any particularly lively discussions or give-and-take between the two of you? There must have been some.

Rosenfeld:

There must have been, yes, but we never had lively discussions, because there was never a violent antagonism and I was, of course, always reticent. I felt that it would be incongruous to oppose my opinions to his, and since finally he was the author of those papers — I did not sign those philosophical papers with him — it was his opinion which had to count, so I had to yield, if I didn’t like it. But he was very considerate; he never wrote anything before he at least had my consent to leave it at that. One should not insist too much on that because it was seldom that it arose. That’s another aspect. When he was invited to join the Accademia Pontificia I told him that he ought not to do that, that it was a trick of the Pope to enlist scientists in the service of the church. He just laughed and said to Margrethe, “Rosenfeld is such a violent enemy of the Pope that he suspects the Pope of the worst possible tricks.” He did not take it seriously and so he joined it all the same.

Kuhn:

You think he had no thought, in spite of his anti-clerical attitudes, of refusing this, that he never seriously considered that possibility?

Rosenfeld:

No. He was a bit naive in political questions. He was trying to assume that all those statesmen were just as honest as he was and that they were seriously analyzing the (epistemological) aspects of politics. He was very naive and inexperienced. In his youth he had lived in this euphoric atmosphere of bourgeois hegemony, in an environment of this ruling class which was very satisfied with itself. Especially in a small agricultural country like Denmark, what political experience could he have got? He was very much out of touch with the times. When this Nazi business came, it took him some time — I must say little time — well, it was violent enough — to understand what it was about. Then, of course, when he understood, and he saw that there were misery and human tragedies about, he was extremely energetic and he could take quick and very efficient action. He was also very practical-minded, for instance, in this committee for help for the Jews. He set that up with the help of his brother who had an eye for political questions. Harald Bohr did not ‘take his tongue in his pocket.’ No, you don’t say that in English. It’s a translation from the French, “sa langue en poche.” He was very open, even in my presence, about criticizing and even laughing at his brother and his naiveté in political affairs.

I always remember, and I think I can put that on this record since it is confidential, that in ‘33, shortly after Hitler’s taking power in Germany, I came to Copenhagen and I met Bohr on the ferry. He was coming back from Berlin or somewhere in Germany where he had just seen Heisenberg. So when we met there, the conversation started and he said, “Well, these events in Germany may bring peace and tranquility to Europe.” So I was rather shocked and I said, “Peace and tranquility by massacring and persecuting people?” Then he said, “Well, yes, but you must understand that in Germany with those Communists the situation was untenable,” and so on. Then I asked, “Who told you that?” “Well, I have just seen Heisenberg,” he said, “and you should have seen how happy Heisenberg was. Now we have at least order, an end is put to the unrest, and we have a strong hand governing Germany which will be to the good of Europe.” But [this naiveté) didn’t last. This is just to exemplify the naiveté, both of Heisenberg, incidentally, and of Bohr himself. Schrodinger was in Berlin at the time and he joked, “What is the difference between a Nazi and a “Sozi,” as the Social Democrats were called. He affected to regard that as infra dignitatem. Well, those people were not prepared; they had had no such previous experience.

Kuhn:

You’ve told me from time to time that you had now and then made notes of things that Bohr had told you about his own earlier development and that sort of thing. I think you said that you had notes, among other things, on the question of the Riemann surfaces.

Rosenfeld:

Yes. The content of this particular thing I have put in a paper in Washington. Did I give you a copy? … It will be published in Physics Today. But there I have put all that I know of those things.

Kuhn:

There’s one thing that I particularly wanted to check with you about. You remember that in that very last talk we had with him, Bohr made that remark about reading James under the urging of his psychologist friend Rubin, and you said when you looked over the transcript that you were perfectly certain that had happened only much later. You had read it together in ‘32. I don’t know quite what to think about this, both Bohr himself indicating that the influence of James was quite early, and several other people have mentioned to me his talking about James, although they’ve not been perfectly sure that he was attributing this early. But from this whole story of Rubin’s saying to him, “You have to say that’s the way you talked before,” one knows that he was a longstanding friend of Rubin. It seems so likely that this suggestion about reading James would have occurred earlier, through Rubin, and that that influence, if it was an influence, would have begun at a much earlier stage.

Rosenfeld:

No, I don’t think so, because Rubin was a friend of Bohr’s at the university, or at least at that age. And at that age it was not sure at all that Rubin knew already about James. He was only a student, but he might have.

Kuhn:

That was a great big book, The Principles of Psychology, and his other writings. If Rubin had been interested in psychology he’d very likely have known about this, and Bohr was obviously talking in this period about the sorts of problems of language.

Rosenfeld:

When Rubin went into his career as a psychologist, they saw each other occasionally, but they were not such close friends as that, so I can very well imagine that it is only when Bohr came out in public with complementarity, which happened only in 1927, that it struck Rubin that these were his old ideas. Then of course Rubin was well prepared to think of James, but he might not necessarily have seen the similarity at this early stage, so I don’t see any incongruence there. All I can say is that I can remember very distinctly Bohr’s enthusiasm. Obviously he read James for the first time then; there was no doubt about it.

Kuhn:

Did you also read it along with him then?

Rosenfeld:

Yes. I had read some small articles by James before, but not his Principles of Psychology; I don’t remember quite what it was, some small work on pragmatism, a popular exposition of pragmatism.

Kuhn:

But the two of you at this point read James’ Principles of Psychology?

Rosenfeld:

Yes; well, at least parts of it. Bohr could read even large books very quickly, and he could also spend a night reading a book when he felt like it. I remember that he had to write a page about Zeeman, about some celebration of Zeeman, which one I don’t remember exactly and he spent the previous night reading a whole book containing Zeeman’s papers.

Kuhn:

Was he then able to write the page quite quickly?

Rosenfeld:

Yes.

Kuhn:

That was not something that he had to do over and over again in successive drafts?

Rosenfeld:

No. But it was always retouched, and the proof was also.

Heilbron:

This is the paper in the Zeeman Festschrift?

Rosenfeld:

Yes, I think so. It’s only a page or so.

Heilbron:

It doesn’t talk at all about Zeeman, just about the Zeeman effect!

Rosenfeld:

No, it’s just a page of generalities, but when you look at such a page you say, “Well, this man got a letter from the Duchess asking for some words, and then he jotted down a few sentences.” That’s the impression that he gives, but he certainly spent several days thinking of nothing else but Zeeman and reading all his papers again, talking all the time of the troubles they had with understanding the anomalous Zeeman effect and how important it was. Once I came to Copenhagen in order to work with Bohr, and I found him reading a novel. He was very embarrassed and he said: “Well, I feel like a criminal, but, you see, I got that novel as a present from the author and now I must thank him. It’s overdue, and so I have to read it before I can thank him.”

Heilbron:

Did he then read very few novels and stories?

Rosenfeld:

Very few. He had no time, of course, although I heard when we were interviewing Bohr that he read detective stories, but I don’t know that from my own experience. I only know that sometimes when we left very late in the night, he took a Wodehouse or something like that, saying, “That’s what I need to get to sleep.” He was very interested in literature, of course, and we had conversations about literature, but it was always Victorian literature, the things he had read in his youth: Meredith, Dickens, and so on, as well as the Danish classics. But he did not follow modern literature, except occasionally when he got a novel that he had to thank the author for. Then he felt compelled to read it.

Heilbron:

Did he read other things, for instance, histories, just occasionally?

Rosenfeld:

Just occasionally, yes, when he got the book from somebody.

Kuhn:

By the time you were working with him, was the further internal development of the notion of complementarity still very much on his mind, or was that pretty well fixed by ‘32?

Rosenfeld:

No, it was very much on his mind: he was always very attentive when he talked to anyone about it, about the reactions, and then he thought of new ways of modifying the formulation. Then when he had found such a thing, he was always very excited. And surely one could trace those modifications in the successive papers. Of course, at that time, new things came out; it was a time when he especially developed the applications to other activities biology, psychology, human cultures, and so on — and every time he was very excited about the points that he made.

Kuhn:

Do you remember particular things that excited him at the time you worked with him or saw him? Things that he felt as new realizations, or —.

Rosenfeld:

Well, the points that were put down in the papers: for instance, in biology, this idea that when you try to analyze an organism you kill it, and that kind of thing.

Kuhn:

Did you work with him on that paper?

Rosenfeld:

Light and Life, yes.

Kuhn:

Was there any sign at that point?

Rosenfeld:

I was not particularly happy with that. I was never quite happy with this idea, and I always came back, occasionally in discussions, saying that, after all, genes are molecules and nothing else; we don’t see anything else in them. And I had the impression that he was himself always a bit uncertain; he never was completely at ease with this question of life. When he wrote and Life and Life he was very definitely anti-mechanistic; surely he remained that, that’s clear, but he had a tendency to exaggerate the thesis of mechanicism. He took it, as, I suppose, he had received it from his father, in the nineteenth century way, and it is only in later years, when be was confronted with those new discoveries, that he considered this physical and chemical explanation in a really modern light. He was always returning in conversation to the difference between an organism and a clock, and that kind of thing — a completely antiquated stage of the discussion.

Heilbron:

The problem of recognizing, of course, when you have a complementary situation, what the complementary aspects of the situation are, is exceedingly difficult. I wonder if you recall other applications that he had might have made but had never published other attempts to find the complementary aspects?

Rosenfeld:

I think he published practically everything that he thought about, or, at least, all that I heard of this complementarity between charity and justice that he was mentioning in later years has also been inserted somewhere. I think he has published practically everything. He always insisted, “It’s only a sentence, you see.” It was never enlarged upon, but that, he insisted, ought not to be done. He said about freedom of will, “I have nothing else to say about freedom of will except those few sentences that I have put down.” He could only spoil the thing by saying more about it, because that was the essential epistemological aspect and there was nothing more to say about it, because those words were used in an unanalyzed manner: such a concept as freedom in connection with will was a primitive concept, according to him, and so there was nothing more to say. You do not discuss the meaning of integers because they are the beginning of mathematics; there is nothing more to say about it. That was the kind of attitude that he had; this epistemology should not be a pretext for discursive chatter which would only confuse the issue. Rather, it was to point out that you use those primitive concepts to apply to various situations and to cover the situation adequately.

Kuhn:

Did he at this time intend to do a big systematic account of the philosophy? Many people have taken it that he did; I have been told again and again that he was working on it.

Rosenfeld:

Personally, I urged him not to write a systematic account; first of all, he would never have been able to do it, it would have been impossible. That would spoil complementarity to put it in a system. It’s not a system; it’s a method. But I urged him to put down in writing a discussion of this analysis of measurements in quantum theory, because there are a lot of small remarks that have never been published. I’m rather afraid that we shall never be able to quite reconstruct those things with all the finesse that he had in them.

Kuhn:

Did he want to do that?

Rosenfeld:

Yes, at least, he professed so. He acknowledged that it would be a fine thing to do, but he certainly had inhibitions. He always found something else to do rather than to tackle that.

Heilbron:

What would the reason for that have been?

Rosenfeld:

Probably because he may have thought that those things are not for systematic exposition. He felt the difficulty, which is a real one, that if you begin to enunciate principles and put them into systems, you fossilize the thing, you make it too rigid.

Heilbron:

But a more detailed discussion of examples, of the application of methods, I’m sure, would be legitimate.

Rosenfeld:

Yes. He made a great effort in that line in his paper with the discussion with Einstein. But he always had to have some occasion which forced him to do that. To set out writing a book without any concrete aim in view or concrete occasion in view was apparently impossible.

Heilbron:

When he would enunciate something ambiguous like charity and justice or other complementary positions, would you then discuss the thing more thoroughly with him or would you just let it go?

Rosenfeld:

Well, there was little else to do but let it go because he would not enlarge it. He would say that that’s an obvious thing. Mostly, those remarks about things being complementary meant that all the discussion that had been going on for ages about those questions was futile, there was nothing more to say about them, and there was no real conflict. This complementarity was just a harmonious relationship and that settled the matter. All these books that philosophers had filled with considerations of freedom of will were just rubbish; there was nothing in it except the use of the words to describe complementary situations.

Kuhn:

Did he himself feel at all impelled to investigate the nature of the complementarity more? To decide that p and q were complementary does not end what you can say about p and q. Did he feel any —?

Rosenfeld:

First of all, he asked, “Why are those Russians dissatisfied?” And I tried to explain, “They accuse you of being a positivist,” and so on. Then he said, “Is that it? But those things are so trivial; they are not of interest to physicists. Physicists are beyond that point and that is not the thing that we are interested in. We are struggling with real problems, not with those trivial statements about our living in an external world.” You see, he refused to look at those problems probably because there was no complementarity in them; they were such trivial things. He did not feel that there was a real thing for discussion. It’s not quite clear to me whether that was entirely it, or whether he just did not like to commit himself to points which were regarded as taking sides or being labeled one thing or another. He felt that all points of view have their justifications, or at least a partial justification; but this was not in the spirit of compromise; he hated compromise. The few times when I saw him really irritated and rather hostile to a person were when the person would suggest that ‘after all, complementarity is so nice because it is a compromise between opposite views, it establishes harmony by compromising.’

Kuhn:

That made him very angry?

Rosenfeld:

That made him angry, yes.

Kuhn:

Doesn’t it seem to you that sometimes this was an appropriate thing to say, that there were occasions, in fact, when complementarity in certain of its non-physical applications really became a doctrine of at least co-existence?

Rosenfeld:

Co-existence, yes, but co-existence of sharply defined antagonistic statements.

Kuhn:

But why was there so little effort to sharply define? To say “justice” and “charity” are complementary really scarcely begins to have content until you really say more about what justice and charity are, and it’s just at this point that Bohr did not take the steps that might have been forthcoming. And it’s to some extent within the analytic philosophy tradition that they really are being done.

Rosenfeld:

Yes, you may very well say that. He felt that those concepts are defined by referring to a familiar situation which was known to everybody and which everybody recognized and thus was not in need of definition. At least, that is my impression; of course, it is debatable, but surely, it is the worst misunderstanding of complementarity to interpret it as an attempt at compromise. That is quite definite.

Kuhn:

Yes, purely intellectually, that is all wrong. To what extent were you yourself a participant in the several rounds of the Einstein problems? The first of them were before your time.

Rosenfeld:

The first recollection I have there is this famous box problem in 1930. I just happened to see Bohr immediately after the paradox had been formulated by Einstein. That occurred at the Solvay Conference, and though I was not at all involved in that conference in any way, I was then in Liege and, knowing that Bohr was in Brussels, I came to Brussels to see him and discuss things with him. I arrived in the late afternoon or evening and went to the club, La Fondation Universitaire, where I knew Bohr was living, and when I came in, all the participants were just coming from the session. I saw Einstein followed by a court of “lesser fry” and he was beaming. He sat down with a circle around him and with intense satisfaction he was repeating the argument before all those admiring people. Bohr was absolutely like a dog who has received a thrashing, with hanging head, and when I came to him, he immediately started saying, “But, you see, that is quite, impossible.” I did not know what it was all about, but I suspected that there had been some argument from the scene with Einstein which I had seen first before coming to Bohr, but I did not know at all. He did not even realize that I had not been present. At that time, he had no answer, of course, but he was just trying to say that it is quite impossible that such a contradiction could occur. Then I started asking him, “What is it all about? What kind of contradiction?” and he explained to me. We had. dinner together and there were other people at the same table, among them Manneback; during the entire dinner Bohr was addressing in turn all the people, trying to let them agree with him that it was quite impossible that such a contradiction could occur. He was terribly, terribly excited and he had no real argument; he said, “If that were true, then the whole of quantum theory would be absolutely impossible, so what would we do with all the evidence that supports it?” But he had not seen —.

Kuhn:

Did you work with him any further?

Rosenfeld:

No. Well, there was no need for it. My last vision of that evening is of Bohr trotting along with Einstein and also trying to explain to Einstein that this was impossible, Einstein hardly listening and going his way. Then I saw Bohr again the next morning and he was beaming — he had the solution. As he said, “This is the most beautiful harmony between —.” And then he started saying, “How is it possible that Einstein could believe that such beautiful theories as his own general relativity and quantum theories, to both of which he has contributed so essentially, could be in conflict with each other?”

Kuhn:

He was deeply bothered, was he not, by the fact that Einstein was really never persuaded?

Rosenfeld:

Oh yes, he was very unhappy about it, and, of course, a bit spiteful too.

Kuhn:

Can you remember things he said about it?

Rosenfeld:

He never said that he was spiteful, of course, but I had that impression several times. He always said that he was unhappy, but he said that he was unhappy that ‘Einstein had left physics’; that was how he expressed it. Einstein had abandoned physics after 1920.

Kuhn:

He had no sympathy for Einstein’s work on the general field theory?

Rosenfeld:

No.

Kuhn:

He was himself really never very deeply interested in relativity at all, or was he? The special theory he certainly could and did use, but —.

Rosenfeld:

No; he had understood very deeply the essence of general relativity, but then it was like justice and charity. He said: “Well, there we are; the last word has been said about that. The principle of equivalence is there and there is nothing more to say. There it is.” He felt that that was just the adequate description of the situation and then you had only to unfold the consequences. So that was perfectly all right. Then when he had reached such a stage he gave the impression of not being interested, because, for him, it was settled, but it did not mean that he did not appreciate it; no, no — on the contrary, he had the greatest admiration for it and he knew extremely well what it meant. He could answer all those tricky questions that you can ask about coordinate systems and he was very clear about them.

Kuhn:

Can you think of other things, of episodes in your own experience with him or things that he told you of his earlier life that ought now to be gathered up?

Rosenfeld:

There are definite stories which are already well known, for instance, the whole story of the spin and the rest of this question. In fact, in thinking of this work of editing the papers, the worst job that I see in front of me is precisely this period from 1914, let us say, to 1924, those ten years. There is very little concrete information that I have heard about him, aside from the published papers; well, I have heard, of course, such general statements as that things went too fast for him. That explains why he did not publish the fourth part of the big series on the line spectra, because, as he said, most of the content had been published by other people, although less thoroughly and less pointedly than what he had written — that kind, of general statement, you see. Also that Kramers had helped him very much by his knowledge of the Hamiltonian theory and celestial mechanics and that kind of thing, but that when he had gotten the mathematical solution from Kramers, then usually together they found some simpler presentation, for instance, his treatment of the Stark effect, which is so beautifully simple. That is typically the Bohrian way; he was not satisfied when he saw a beautiful derivation starting from Hamiltonian equations, nor did it impress him particularly.

He wanted to understand the situation in more detail by some simple arguments and simple principles. Then I know that he said, that the greatest trouble that he had to overcome was how to treat degenerate systems; that was the greatest handicap, and then that all that was connected with the spin was a great mystery until the idea of Uhlenbeck and Goudsmit came up. All the anomalous Zeeman effect business — that was the crux. But those very scrappy things never went into any detail. I tried once to give a course of lectures in Utrecht about the older quantum theory, and so I got involved in those conditionally periodic systems; I told Bohr that I had found it cumbersome to present the idea of conditionally periodic systems to the students because you have to go to those Stackel conditions and that kind of thing, or at least, I thought so. He was very surprised, and he said, “But that’s a very simple notion; the way to do it is presented in those papers about the line spectra.” When you look at those papers you see that he presents the notion without analyzing it, but just simply as something given, as systems with that structure; then of course he introduces the action variables, so then it looks very simple, but that is typical. The fact that this is not linked up with the fundamental dynamical equations, with d’Alembert’s principle, does not bother him, and that is a very fundamental feature in his attitude. He takes things as they are and then the task is, “What can we do with it, how can we use it in order to analyze the phenomena?” The same thing applies to statistical mechanics; he was enthusiastic about Gibbs and when I tried to talk to him about the problem of ergodic theory, he did not appreciate this way of putting the problem at all. He said, “Well, I like Gibbs so much because he starts from the notion of temperature that we have and he shows how it can be put into a statistical description and what we can do with it.” But the idea of connecting this modulus with the behavior of a mechanical system didn’t impress him as an interesting problem. He said, “We know that it works, we know that we can account for the thermodynamical behavior in that way and so that’s it. That’s a property of the mechanical systems, and there we are.”

Kuhn:

But didn’t he also have the feeling, which is perhaps not incompatible with what you say but at least is not implied by what you say, that there was some sense in which Gibbs had a very much more fundamental approach to the problem than the Boltzmann approach?

Rosenfeld:

No. Well, he said that Boltzmann had spoiled things just by insisting on the properties of mechanical systems and that he had raised this problem of reversibility and irreversibility which was not a real problem. Because if you started directly, as Gibbs does, from a given statistical distribution, the irreversibility was in it, somehow. Then, quite naturally, the complementarity idea came in there; those are two complementary aspects.

Kuhn:

Which you couldn’t say about the Boltzmann approach.

Rosenfeld:

No, surely not. Those people, Boltzmann and Clausius also, had a sort of right feeling, I think. I think they had seen the problem rather clearly, Boltzmann especially, but they had not the words to express it, nor the mathematical basis. Even Gibbs lacked the mathematical basis. After all, now we have mathematical methods, with the Lebesque integrals and so on, to tackle those problems and to formulate them in a rigorous way; that did not exist in Boltzmann’s time nor in Gibbs’ time. This does not concern Bohr, but if we may take it as an aside, I feel that Boltzmann has been badly treated. I think that Boltzmann had everything and even more than Gibbs, and Gibbs himself was very modest because Gibbs did not mean anything more than an exposition of Boltzmann’s, ideas, but with a proviso. Gibbs had the ambition of only putting in his books those statements which he was able to prove rigorously, and therefore he avoided the ergodic problem, or at least, he hid it in the 14th Chapter, I think it is, instead of putting it in the first. In this 14th chapter there is a very deep formulation of the ergodic problem, but it is hidden in a very cryptic way because Gibbs did not feel that he could prove those statements, so he was very reluctant to commit himself; that was his attitude.

Kuhn:

Coming back for a minute to Bohr’s view here, am I right in my feeling that when Bohr spoke of the mechanical and the statistical treatment as being complementary, he was thinking of something that he could find in the Gibbs approach and would really not be there for him in the Boltzmann approach?

Rosenfeld:

Yes. He made a distinction somehow, and he was not the only one. It is a sort of error, in my view, which has crept into the tradition of teaching statistical mechanics, that Gibbs and Boltzmann are opposed to each other, whereas, in fact, they are the same, in my view, with only that nuance. But certainly Bohr accepted from teaching (???). I don’t think that he had ever read Boltzmann, or at any rate not Boltzmann’s papers. He had read, I think, and studied very carefully Boltzmann’s lecture on gas theory, which is not a good book and in which Boltzmann has not succeeded in expressing his ideas as clearly as in his papers. So when you read that book you have only a very incomplete and distorted view of Boltzmann’s ideas. So Bohr knew Boltzmann’s gas theory, but then he had studied Gibbs very carefully and he had been enthusiastic about Gibbs’ presentation, which is much more elegant, surely, than Boltzmann’s. Then this complementarity was his own way of analyzing the problem; he pointed out, quite rightly, that when you have a modulus, when you have this statistical distribution in energy, then you have no definition of the energy of the system and this is complementary to the mechanical situation in which you have an isolated system. Incidentally, this is neither Gibbs nor Boltzmann because you know that Gibbs introduced not only the canonical distribution, but also the micro-canonical one, and that he argued that they were equivalent for practical purposes; that is to say that both of them could give what he called the thermodynamic analogies. So when Bohr says that they are complementary, he does not follow Gibbs. This question of introducing complementarity here came in ‘31 when he was writing his Faraday lecture; it may have occurred to him before, but that was the occasion on which he had to formulate it completely, and this was something which was neither Gibbs nor Boltzmann but was Bohr. He knew statistical mechanics from the time that he was writing his thesis on metals and there he had found the great advantage of using a canonical distribution for the temperature distribution of the electrons and so he had not bothered about this distinction, canonical-micro-canonical; that was not his main concern. So he took his point of departure from Gibbs, but then he dominated the situation without attaching himself to anyone.

Kuhn:

Do you know if he had ever argued or discussed these questions in any detail with Ehrenfest? Ehrenfest seems to have been a strong Boltzmann enthusiast, but had very little use at all for Gibbs.

Rosenfeld:

No, I don’t know anything about that. When I asked Pauli what he thought about this ergodic problem, he said, “Look up the Ehrenfest article,” but that was not very helpful. Although Pauli has written this paper, with Fierz about von Neumann, he was more concerned, I think, with the purely technical details of the proofs. The proofs seemed to him distastefully complicated and he tried to simplify them, but I don’t think that he was very deep in the logical formulation. The only things that might be of interest to put on record, although they are also to be treated confidentially, I think — it is just as well that those things which are not published should be on the record — are the events which occurred with the discovery of fission and Bohr’s contribution to it in the States. I described part of it in this little article, [“Nogle minder om Niels Bohr,” Fysisk: Tidsskrift 1963] but there is one thing which of course cannot be published, at least not now, and that is a conflict that he had with Fermi at that time, which was a pretty bitter one. I might just as well start right at the beginning.

When we went to Princeton, on the boat Bohr immediately told me that he had received this note from Frisch suggesting first of all the interpretation of Hahn and Strassmann’s experiments as being of fission, and even suggesting the name. Frisch had probably spoken with a biologist who had suggested this name, ‘fission.’ Further, the note gave Bohr the plan of the experiments that he was intending to do in Copenhagen to see the fission fragments in the cloud chamber. Bohr was very concerned about the question of how this fission could occur with such a large probability, and so all the time during the sailing, which was about five or six days, I think, we were working in the usual way on that problem. Then it occurred to him that it was just a consequence of this droplet model of the nucleus. The compound system had also this mode of vibration that was a degree of freedom just as well as was the expulsion of the neutron, and therefore there was equapartition between all the degrees of freedom and therefore the two were equally probable. We11, that was very rough, but it was in that line.

So when he arrived in Princeton, he was reconciled to the idea that this was the right explanation and that it fitted in with the general conception of nuclear reactions. It was also clear that the threshold for excitation could be very small for the heavy nuclei because of this instability. All that time, then, we had discussed the problem quite freely and I had the impression that Frisch’s note was already published or would appear in the next number of Nature. Then in New York Wheeler came to meet us and Bohr had some business — I don’t remember what it was — but at any rate Bohr remained in New York and I travelled with Wheeler to Princeton on the same evening. I still remember that we had a pair of Bohr’s skis that we took with us so as not to encumber him, and there was a Negro porter who saw us looking for the train. He asked us, “Where are you going? To Atlantic City?” Because of those skis! Anyhow, we came to Princeton and it must have been the next day that there was a (Jonal) club meeting in the evening at the Institute and Wheeler asked me to come and perhaps say something. So it was quite natural and quite obvious that I should talk about this note of Frisch and what we had done on the boat about it. That created quite a sensation, of course, and I suppose that even the same evening people were putting through long distance calls to California and other places.

Anyhow, in the next few days, all the American laboratories which had oscillographs were trying to produce fission. When Bohr came and heard about that, he was very annoyed because, he said, “I did not want to talk about this Frisch business before Frisch’s note had appeared in Nature. Now you will see that those Americans will be the first to publish it, and so it happened. It was quite impossible to prevent people —. Bohr tried to intervene at the Physical Review to stop the notes that were pouring in, but that was not quite possible. The Americans were pretty tough, you see; some were nice, but others said, “Well, if I do it, that’s independent work and it must count as independent work.” But that was only part of the story. Then Placzek arrived — it must have been in the beginning of February, I think — and that is the conversation that I have described in my notes. Bohr said, in order to say something nice, “Well, now we are at least rid of this business of the isomers,” the transuranium isomers. And Placzek said “Not at all. That’s just as obscure as ever because you can produce transuranium by neutron capture; there is a resonance neutron capture in uranium, and this is a phenomenon which is not connected at all with fission, since the cross section for fission does not show any resonance there.

So it is quite unconnected, and you are just as you were before. Those are two different phenomena.” When Bohr heard that, he could not stay any longer; Placzek: was absolutely forgotten, and he rushed to the office with me following him. When he arrived in the office he said, “Now you see I have it already,” and he drew those two pictures [Reproduced in Rosenfeld’s “Nogle minder om Niels Bohr”] showing the similarity between thorium and uranium 238 and the quite different behavior of uranium 235. Immediately afterwards he said: “That is quite certain; those are simply facts, or inferences from the facts, but absolutely unique. But why is it that U235 has this different count, is so fissionable with slow neutrons? That must be because this threshold is much lower.” But that was also quite understandable because one knew that the last neutron of an even number is much more strongly bound than the odd neutron, so the whole thing was done within an hour. Then came the big surprise. Placzek turned up a bit later, and Bethe turned up, and Bohr, full of enthusiasm, of course, explained the whole thing to then. But both of then remained very, very reticent, saying, “Yes, well, perhaps, but now we must do experiments at least; we must try to study fission of the separated isotopes, see if we can get hold of them, and so on.” Then surely as it is told in this book, The New World, Dunning accepted Bohr’s view and decided to start work on measuring the cross sections and so on on Bohr’s hypothesis. To be quite blunt, this was not because Dunning was more penetrating than the others, but perhaps just the opposite. Dunning was ready to accept anything that Bohr said, but that was all right.

Then there was this meeting in Washington to which Bohr came and explained all that, and Fermi was absolutely closed to the idea. He did not understand the argument which seemed so clear. Fermi said that that could not be so and he would do the experiments and show exactly how it was quite impossible. That was only part of the story; what followed was more annoying yet. I still remember going to Washington. Tuve was quite excited because he had produced some fission; he had a Van de Graaff and an oscillograph and so he produced those fission bursts. The Van de Graaff was then at the Institute for Terrestrial Magnetism, of all places, and we went there in the evening; Teller was with us and Tuve and Roberts were both there and very excited, and we had to sit there and look at those things on the oscillograph. Then the telephone was ringing and Tuve rushed to the telephone; it was the newspapers and he painted then a picture: ‘Professor Bohr is looking at the fission produced by the Institute for Terrestrial Magnetism —.’ It was a bit —. But Tuve was nice enough, I must say, when he published his note, because of course he had to publish it, to mention that he had heard from Bohr that Frisch had done the same experiment and that that it would shortly be published. Tuve was on the side of the nice people! Buy anyhow, that is only to picture the atmosphere. Then Fermi gave a talk on the radio about this fission, and although Bohr had of course told him every detail, he did not mention Frisch; he mentioned all the others, but he didn’t mention Frisch.

So Bohr was furious. I never saw him angry except on that occasion. I don’t know where Fermi was at that time, but that you can find out; anyhow, Bohr decided to go to see Fermi and thrash it out with him. I accompanied him, but I had nothing to do with the matter. I think Fermi was at that time in a university in Washington, but I am not absolutely sure. I did not witness the interview; I just sat in the library all the time that they were closed up in a little room. I saw only their faces when they came out — it was a long time -— and both of them were quite pale, harassed, quite exhausted. I thought surely that it had been a very harsh thing because Bohr told me afterwards that he had reproached Fermi quite openly, that it was not fair, and so on. Of course, as you would imagine, Bohr had started gently and as friendly as possible, but to his surprise Fermi had not accepted any reproach. Fermi had defended the thesis, that when work is published, it has to be mentioned, and that when work is not published, it has not to be mentioned, or something like that, and that all the thinking has been done independently and without any influence from Frisch. So they left it at that. Very queer thing.

The other point, that Fermi did not understand Bohr’s argument, is in Fermi’s character. Fermi did not understand what other people were doing because he was such a strong mind himself; he went his own way, had his way of doing things, and, his contribution to quantum electrodynamics is also of that type. He had heard about the difficulties of quantizing the electromagnetic potentials and so on, and then he had thought independently about the problem, without —. In fact it is quite understandable that he had this idea of people’s thinking out a problem independently and then publishing it and not acknowledging anything from others, because that was his way of working exactly. He had only heard there was a difficulty there that people were embarrassed, but he did not bother to inquire what the embarrassment was or how they were doing it. He just took up the problem, thought about it independently, and produced this solution, which was quite different from the line of thought of the others. That was his way of thinking and that was how he tackled all the problems that he so successfully dealt with. But in this case it had rather unpleasant consequences. Finally, of course, the whole thing settled down because eventua1ly Frisch’s note appeared, but it has not been purified completely, because in this story in The New World, which is a kind of more or less official story, or at any rate a story that people would like to see pass as an official record, it is told quite wrongly. Well, that was the only time in which I saw Bohr really angry and quite burning with passion, and that was because he was defending another.

Kuhn:

Did he reestablish cordial relations with Fermi? How were their relations later?

Rosenfeld:

I have no direct experience with that. He must have seen Fermi in Los Alamos. Well, he told me something about that and I don’t think it was very good; he told me that he had another clash with Fermi at Los Alamos about this question of implosion. They were afraid that the bomb would go to pieces before the fission, the chain reaction had developed. In order to prevent that one had to avoid big pieces. One way to break it up into small pieces was to have the surface of the cavity quite irregular, that is, not to have a regular spherical cavity but a quite irregular one. That solution was, I think, finally pushed through by Bethe, with Bohr’s support. Fermi apparently opposed it violently while Bohr advocated it, so there was another clash of the same kind because Fermi was again incapable of going into another person’s argument. He had his own ideas and he was not to be diverted from them. But that is only a vague account; I don’t remember quite exactly what happened, but there is this evidence of another clash. Apart from that, Bohr expressed great admiration for Fermi’s ingenuity in this work in general; it was only this incident. He did not give me any details, but he said, “Fermi has been incredibly ingenious in all this work.”