David Dennison - Session III

Notice: We are in the process of migrating Oral History Interview metadata to this new version of our website.

During this migration, the following fields associated with interviews may be incomplete: Institutions, Additional Persons, and Subjects. Our Browse Subjects feature is also affected by this migration.

We encourage researchers to utilize the full-text search on this page to navigate our oral histories or to use our catalog to locate oral history interviews by keyword.

Please contact [email protected] with any feedback.

ORAL HISTORIES
Image not available
Interviewed by
Thomas S. Kuhn
Interview date
Location
Ann Arbor, Michigan
Disclaimer text

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.

In footnotes or endnotes please cite AIP interviews like this:

Interview of David Dennison by Thomas S. Kuhn on 1964 January 30,
Niels Bohr Library & Archives, American Institute of Physics,
College Park, MD USA,
www.aip.org/history-programs/niels-bohr-library/oral-histories/4570-3

For multiple citations, "AIP" is the preferred abbreviation for the location.

 

Part of the Archives for the History of Quantum Physics oral history collection, which includes tapes and transcripts of oral history interviews conducted with circa 100 atomic and quantum physicists. Subjects discuss their family backgrounds, how they became interested in physics, their educations, people who influenced them, their careers including social influences on the conditions of research, and the state of atomic, nuclear, and quantum physics during the period in which they worked. Discussions of scientific matters relate to work that was done between approximately 1900 and 1930, with an emphasis on the discovery and interpretations of quantum mechanics in the 1920s. Also prominently mentioned are: E. F. Barker, Niels Bjerrum, Niels Henrik David Bohr, Walter Colby, Paul Adrien Maurice Dirac, S. Dushman, Paul Ehrenfest, Kasimir Fajans, Fussell, Samuel Abraham Goudsmit, H.C. Hayes, Werner Heisenberg, Takeo Hori, A.W. Hull, Friedrich Hund, Elmer Samuel Imes, Oskar Benjamin Klein, Hendrik Anthony Kramers, Irving Langmuir, Hendrik Antoon Lorentz, R.W. Marriott, Matos, Yoshio Nishina, Wolfgang Pauli, Harrison McAllister Randall, Erwin Schrödinger, George Eugène Uhlenbeck; Kbnhavns Universitet, National Research Council, Swarthmore College, University of Michigan.

Transcript

Kuhn:

I thought I’d like to drop back to this list of developments that took place during the years before you went to Europe. [Outline, bottom of page 2 and top of 3]. Perhaps looking at that list will bring back to you topics that were of particular interest while you were in graduate school which you and Colby, you and Klein, may have talked of particularly or which may have gotten particular attention in seminar. There are just a whole group of developments rather combined together in there, and I’d be interested in knowing what seemed ‘hot’, what the subjects of greatest interest were.

Dennison:

Of course this period before I went abroad was a period when I was both learning and also not too far along. I suspect that the real point came let’s say in my second year while I was studying Sommerfeld’s book. This was of course in a period of development. But I at least had the feeling then that things were in moderately good shape, that this method of Sommerfeld’s was a good one and that it was indeed yielding interesting and valid results. I don’t think there was anyone here who was in a position to give me anything more than that. So one would say that it wasn’t that things were solved or that there weren’t many problems coming up, and many dark spots, but there wasn’t really, in my mind at least, a feeling that this wasn’t a legitimate way to proceed, and that from this one would essentially find everything one wanted. Now during the third year, when Klein came and brought then the Bohr viewpoint, then this, as I recall, changed my opinion. I began to feel that the Sommerfeld method was, so to say, in some respects a ‘gimmick’; it was a way of calculating that yielded results all right, but that it side-stepped all of the fundamental things. Bohr’s method, and even his fundamental work in 1913, did not rely on so to say ‘magic formulae’ but rather was always looking at the fundamental physics of the thing. For example, the way in which he used correspondence principle in order to find results. So then I’m sure that I took this viewpoint, that this was just a ‘black box’ method of calculation which really did not reveal anything of a fundamental nature. This certainly then was carried on as I went with Bohr. There were all these dark things, and the question of the nature of radiation and the relationship between light quanta and radiation began to really loom as very, fundamental questions at that point.

Kuhn:

The contrast that you’re drawing between Bohr and Sommerfeld is I think a very apt one and is certainly illustrated by their work. But you also say that Sommerfeld seemed to you not only a reasonable way to do physics — which later did not seem altogether reasonable –- but you also had the feeling that going on in this direction might very well produce the solutions to the problems that one was now trying to work out. The Bohr method clearly was a deeper method, but did you have the impression it too might very well develop and work out the problems that were there to be worked out, or did it seem clear that this now needed a fundamental new approach?

Dennison:

I think that as soon as I began getting Bohr’s viewpoint, and through talking with Klein and later with Bohr himself it appeared that something fundamental was needed. I’m not sure whether we would have said it in just that way. It was rather that the questions were deeper and darker than had appeared at first, and where the solution was to come one didn’t know — illustrated for example with the Slater and Bohr paper in which they were feeling for conservation of energy or something or other which would then really reveal a new approach to the problem. I think that’s right.

Kuhn:

I’ve been told that Bohr in these years, and perhaps for sometime before that, had really felt that the correspondence principle was perhaps a device for producing successive approximations which would lead gradually to a fall formulation, and this would fit with the attempts at an extended correspondence principle during the period you were there. I’d say the attitude at Gottingen in certain circles was rather different from this. It’s indicated in Born’s Atommechanik where he says ‘we’ll push this just as far as we can, but it’s all wrong; a step to something quite different is needed, that’s what I’ll do in Volume II’ — an attitude that I think would not have been found in Copenhagen, but I’m not clear about that.

Dennison:

It’s hard for me to be quite sure. I would more say it perhaps just repeating what I’ve already said, that, here were deep and dark problems. One used on these all the means at one’s disposal. The Sommerfeld, or the phase integral business was not a safe guide at all, although it would give correct results. But this was not really anything to be pushed. Correspondence principle was a much safer kind of thing, but again a means of trying to feel your way into these problems. I’m not sure that at that time, at least, I had any deep feeling, or that I picked up from people around me the feeling that there was a whole new approach that was necessary. It was more the question, here were these problems which one couldn’t understand; how to understand them? Well there were various means — the integral was one way, but not a very safe one; the correspondence principle, which was much closer to all the body of classical theory, was a better guide. One was trying to find the answers to these problems, the anomalous Zeeman effect for example, and things of this sort. But that it was to require an entirely new system, I at least was not aware of this, I think. And I don’t remember having picked up this viewpoint either from Klein or Bohr. That’s the way it looked to me.

I don’t think that there were many people that I came in contact with who understood the power of the correspondence principle or really what it was, with the exception of Klein, who was of course preaching it. I picked it up then from him, and began slowly to use it with greater certainty. But I didn’t feel that there were many people in the department who had very much knowledge about it. Well, this one we’ve spoken about [looking at outline] ‘the success of the Sommerfeld school’. My first impression was that this was indeed wonderful and the way to go about it. It was Klein who began sowing the seeds of doubt. [Reading] “How well did people in this country realize what was behind all of Bohr’s thinking in the analysis of the periodic table?” I’m sure that the people in this department at least didn’t, and these were the ones I came in contact with. And I probably didn’t until later through Klein and through Bohr. Of course by the time I was in Copenhagen this particular thing was on its way out; it was just cleaning up the last details. Very interesting, but still it was on its way out. Next you have the ‘definitive failure of semi-classical quantum theory with multi-body problems’. [Outline, p. 3] Here again are these two viewpoints that while I was studying Sommerfeld, I would undoubtedly have had the feeling that well it was just that one didn’t quite have the right ‘gimmick’ to do it — you were very close but you didn’t quite have it. Then later with Bohr, that no, this belonged to the class of the fundamental things which one really didn’t see how to go about doing, and that one could make some progress with correspondence principle but this is about where it ends up.

Kuhn:

Can you pin that down enough to be clear, because the dates here are a little odd — whether you already saw this problem during your last year here with Klein or only after you got to Copenhagen? It’s a pretty firm establishment that the helium problem really went totally to the devil, somewhat of the same sort with H2+. This is a quite late realization, and I wonder whether Klein can have brought that with him.

Dennison:

I think Klein brought it with him, I think he was trying to knock this more narrow viewpoint out of my head during this period, and he succeeded in doing it. I think that’s really what it was; there was nothing to be done this way. And of course these attempts to extend the correspondence principle were going on while I was in Copenhagen — Kronig in particular was one of the people who was doing this. I’m not sure that it had a very profound effect on my work personally. I was very much interested in it but preferred to get the things through matrix theory and so on as this began to come out. I’ve forgotten just when Kronig was working — it must have been at the same time that the matrix theory was already coming out, wasn’t it?

Kuhn:

This comes with Kronig particularly through the intensity formulas, and those are during ‘24 and into ‘25. It may be pretty late in ‘24, there are a whole group of papers that come out — but Kronig certainly has published on the sum rules and intensities prior to the appearance of matrix mechanics.

Dennison:

It probably didn’t make too great an effect on me. I knew about them, was interested in them, but it wasn’t particularly anything that I myself was immediately following up. That’s really what it was.

Kuhn:

We’re virtually ready to go ahead to Zurich. I’d like to ask you just a little bit more, and in a fairly general way about Copenhagen before we do. Perhaps the thing to do is to ask you to tell me a little bit more fully what you did there. There were a couple of papers that you worked on. I’m not clear whether these took up almost all of your time during your period there, and just what the life of a physicist in Copenhagen was.

Dennison:

Well I wonder myself a little bit how it went, and I have to partly reconstruct and partly remember. Actually I didn’t publish many papers during this period. It was of the order of only one or perhaps two in a year, so it was not very rapid. On the other hand, this was a period when I certainly was learning all the things that were being done in Copenhagen, and in fact in Europe throughout this period. I think I was really moderately abreast of it so that I could use the new tools as they came out. I talked with Kramers a good deal, with some of the other people. I talked at times with Heisenberg — in fact a good many times with Heisenberg, and with Kuhn, Ray, and occasionally with Bohr, not as often perhaps as with these other people who were there. Kramers I suppose I talked with more than anyone else. But surely this was a period when I must have spent I think more than half of my time in trying to understand the new papers that were coming out — in the Zeitschrift fur Physik each one, you read all the important things in it, and they got pretty well dog-eared as a matter of fact during this period. Perhaps being married I may not have entered into the night life with the other boys as much as I might have done otherwise. There was a very interesting lady, Mrs. Maar, Frau Maar, who held an evening once a week.

There was some kind of protocol involved here, because neither Bohr nor Mrs. Bohr ever came to these. This was perfectly understandable — it was the lower echelon somehow or other. But all of the rest of us did go, and it was just an evening of conversation with a fairish amount of shop-talk in it — in fact quite a lot — and tea at one stage, and very frequently music. Heisenberg was one of the performers, I remember, and also Kramers. This was one sort of place where one did in fact meet and talk with people all the time too. My recollection is that during this particular period there were not so very many colloquia. There were a few, but not so very many. And there were no courses that I went to. Now I think most of the teaching was done in the university buildings and not at the Institute itself. You see when I first went there, for example, the Institute — well you know the old building — the Institute consisted only of the ground floor. The Bohrs lived, on the second floor, and the servants on the third floor. Then the additional part, which was to become the Bohr’s home was built probably during or at the end of the first year that I was there.

Then they moved into that, and the Institute spread out into the two top floors of the thing. Very good library — very fine informal library. I know when I came back here I was very much annoyed to find that the books here were put into various classifications, whereas in Copenhagen they’d all been by authors, and you always went exactly where you wanted to go and got the book and that was that. Here the system was much more complicated. I remember being very critical of the way things were done here. That’s the truth. We went out also to Frau Maar’s villa during the spring. She lived out there in the late spring and summer. I had my difficulties at one time. We had various troubles with places to live for one thing, and the first winter, along in February and March, the darkness and the general dismalness got into my nerves and I was really somewhat ill from it. It was suggested that we should go up to Norway to the sunshine up there. So we went to a very informal ski resort and stayed for several weeks, and this was a fine thing. It helped me a great deal. I did find the darkness though kind of hard. It was sort of like living in the bottom of a well in the wintertime. It’s hard for me to recall anything more than this of anything that’s helpful to you.

Kuhn:

We talked a little bit about the rotator paper [Paper No. 4] and there was one particular question I had wanted to ask you about. When you do that, you remember there’s a certain amount of arbitrariness left in the way you cut the downward progression off, in the attempt to find the ground state, and you don’t get a nice clean choice of half-integral or integral quantum numbers. In fact you could decide that they ought to be quarter integral, or anything in-between.

Dennison:

Yes, I think that’s probably right.

Kuhn:

I wondered how bothersome this was at this point, or what you tended to blame it on. Did it seem that this was the best matrix mechanics could do, did it seem that there might be a better use which would really set this thing nicely and cleanly?

Dennison:

Well, there was always I think a little suspicion in my mind that it was due to this method, using the constraints, and that —. Whether I hoped that there would be another method that would fasten it I don’t know. The other thing is, how soon, and at what point did I realize that in order to have really proper symmetry, one had to have either the whole integers or the half integers? It could have been around here or it could have been later. It’s hard for me to be quite sure.

Kuhn:

My recollection is that you don’t make any point of that sort here. You do suggest that anything except the half-integer or integer choices would be odd in these circumstances, but I don’t think you produce a real symmetry argument. Furthermore that’s a much more natural thing to come out with as soon as you do the thing with the Schrodinger equation rather than with the matrices.

Dennison:

Yes, I don’t really remember enough. I could look back and try to reconstruct it. But, my feeling was that it wasn’t really through Schrodinger that I came to this viewpoint, but rather through the notion of the various transition probabilities, the amplitudes — that one got into difficulty unless one had either the integer or the half integer; that things were not symmetrical up and down, for example. This was the only way out of the difficulty, either using integer or half integer — then it was all right. Now, whether I said that in this paper — I probably didn’t —.

Kuhn:

I’d have to look back at it myself to be sure, but I think you’re quite right that that idea certainly may be in there. In any case it’s something we can come back to and check. I do want to come to this paper, but I suspect that this paper is one that you send in December from Zurich, so it may be that the best thing to do now would be to proceed to your transition, your decision to go to Zurich, how that arrangement was made, and so on.

Dennison:

Well, it was simply that at the end of my second year at Copenhagen I broached the question to Bohr whether it would be a good idea to remain a third year and ask for more support from the International Education Board, and he had the feeling not. And he was I think quite right. I mean I had gotten what I could essentially get at that time. However, Professor Randall was not yet ready to get authorization for the new positions and he did want me then — he suggested it very strongly — to spend a third year abroad. Well the obvious thing to do was then to go to Schrodinger. Here was this new thing coming out, I had studied his works, so I very much wanted to go there. The funds were arranged as I said through what was the Engineering Research Institute, (now) doing applied work, and out of their overhead they created a few of these fellowships and I think again mine was the first one that had been created. And Randall was perhaps even instrumental in getting it created. So after the summer here we went back, probably in September, were located at Pension (Smelzberg) on (Smelzberg) Strasse, a very narrow steep street above the university.

Hermann Weyl had an apartment just a street above. (Smelzberg) had been the old Pension where all of the old physicists had lived. As a matter of fact I think the rooms that we had, were the ones that had been occupied by (Abraham) for many many years. I’m not quite sure how I knew about it, or how I got reservations there, but probably through Werner Kuhn who by this time had come back to Zurich, and his family had a farm maybe ten miles out from Zurich. He took us there one time. Incidentally — it’s just amusing, this period of the times. (Ours was a) room with one of these great porcelain stoves, you know, that are just barely warm; but that’s what heated things. It was then that Werner [Kuhn] told me — I guess we were talking about drinking milk, and he brought up the viewpoint that milk was just absolute poison, that to drink raw milk was the same as drinking poison. And he was probably right, that all milk had to be boiled, it was just folly to think about drinking unboiled milk. This was a very interesting time in Zurich. I did go to some of Schrodinger’s lectures. I should have gone to some of Debye’s lectures since he was a very clear lecturer. I did go to a few, but I was very much interested in carrying on these research things. I puzzled about the problems of homopolar molecules, of equality, and what it meant. I took these puzzles to Schrodinger and talked to him about them. He was not particularly helpful. He was interested, but he didn’t seem to have any particular clues as to what to do with them. So I don’t think I got anything very directly from him. Of course just the general atmosphere of being there was well worthwhile.

Kuhn:

Were there a lot of people who, like yourself, came that year, particularly because of what had happened the preceding spring?

Dennison:

I don’t think so, no, no, no. I don’t really remember anyone else.

Kuhn:

Fues was there.

Dennison:

Fues, ja, that’s right. Now we must have met him in Copenhagen the year before, but had not met his wife. But he and his wife were there [in Zurich], that is correct. So he was one of those who must have come for this purpose. Well, there may have been more, but it was at any rate let’s say, very different from the collection of people in Copenhagen who had clearly come to see Bohr. In that sense it was quite different. Schrodinger left sometime around — I think in November — to come to this country, and actually we left Zurich before he returned I think. We left about New Year’s.

Kuhn:

Tell me again about the circumstances under which you got this paper, this earlier relativistic formulation.

Dennison:

This certainly came out from talking with Schrodinger about his work and what he was doing. He told me that he had these manuscripts — now just why he gave me the manuscripts of things that were already in print I’m not sure. Whether he had a shortage of reprints or what it was, whether be just wanted to get rid of them; any rate he gave me several of the manuscripts that were already printed. Then he told me that his first attempt, in which he used relativistic mechanics, had not turned out well at all. He had the manuscript of it, but he had never sent it in because it did not give the correct energy levels.

Kuhn:

Was it drawn up as a manuscript ready to be sent in?

Dennison:

Yes, it was just like the others.

Kuhn:

You had a typed carbon.

Dennison:

It was a typed carbon, that’s correct. So it existed. Now here again, perhaps if I spent enough time, I could reconstruct the thing. You probably know much better than I about this. In Schrodinger’s first paper, on the hydrogen atom — this was the one evidently in which he started by deriving or getting the wave equation which he then applied to the hydrogen atom — in the derivation the energy came in as an index of refraction.

Kuhn:

Not really in the first paper. The derivation in the first paper is a very odd one. The derivation in the “Zweite Mitteilung” where he does develop the analogy between the Hamilton-Jacobi equation and the wave equations, does bring in the index of refraction and the speed of the wave. In the first paper he does the thing by a variational principle that is really quite hard to understand, and nobody ever felt quite at home with why that derivation ought to work. So far as I know, at the time when that was done, he had not yet seen this full wave particle analogy through the Hamiltonian equation.

Dennison:

Well now the only remark that I can make on it, that would just give my view of it, was that in this derivation, or whatever derivation he used, there was a point in which the energy came in as an essential part. Now, if at that point, instead of putting in the energy nonrelativistically, you put in the relativistic term, that is essentially put in mc2 at this point, plus the other energy, then it does not lead to the same equation, or to the energy levels of the hydrogen atom. In my opinion that was the thing that he did. There was nothing profound in it. It was simply taking this derivation which he had, which you’ll find in the published paper, and instead of putting in the energy you put in the energy plus mc2 — and that extra term, mc2, somehow or other loused the thing up, and it would not come out then properly. Now I don’t even remember how it came out, but it just did not give the energy levels correctly. And if I were trying to reproduce it, this is where I would start, by taking these early papers and seeing where the energy comes in, and then just adding a relativistic term at this point and seeing what the difficulty is. There were the same questions when he went into this other type of derivation which gives the wrong velocity; the velocity is half the right velocity or something like that, isn’t that correct? You use index of refraction and at a certain point if you take the thing literally, and ask about the phase velocity of these things, you get v/2.

Kuhn:

That I don’t know. I’m terribly interested. I don’t think in the first case that he does get v/2, but there may very well be an earlier form in which he does.

Dennison:

Well, it was just a dark point which was then passed over, but it was in a sense an inconsistency of the whole thing. You would have thought it would have at this point just given you back the actual mechanical velocity, but it didn’t. O.K. I’m just pulling in little fragments, and if I’d looked at it properly I would perhaps be able to reconstruct. Maybe one day I’ll —

Kuhn:

Well I would say, if the paper doesn’t show up, then I would certainly be much interested — if you would —

Dennison:

Take a crack, see if I can reproduce what the trouble was.

Kuhn:

If you would do it. If the paper shows up, then it may not at this point make good sense to put as much time into it as would undoubtedly be required.

Dennison:

Well, maybe I could do it. But, as I say, I did not think that it was anything of a really fundamental nature, which I believe — well I haven’t recently read Dirac’s statement, but Dirac felt that it was perhaps a little bit more fundamental than I at this stage think it was.

Kuhn:

Dirac’s impression is that the only reason it gave the wrong energy levels was simply because it had no spin term, but he [Schrodinger] must have gotten the Klein-Gordon equation, but in the absence of spin, then got the wrong energy levels out of it. Now, that may be. But it’s one thing to go from the non-relativistic equation to the Klein-Gordon equation; it’s quite a different thing, without knowing the operator substitutions yet, to pick the relativistic Hamiltonian and then simply substitute the proper partial derivatives. So I’m not convinced that that is all that it was, and it’s that that makes this thing quite as important for the historical development of the equation as it is.

Dennison:

O.K. Well, we’ll try on that.

Kuhn:

Now, one other question about Zurich and Schrodinger in this period. You had come from Copenhagen, and had been much exposed to the Copenhagen approach. You may not have been there — I guess you probably were not there — in the summer of ‘26 when Schrodinger visited Copenhagen. At least already then, and in fact in Munich already somewhat before then in the interchange between Heisenberg and Schrodinger, there was strong disagreement on the subject of Schrodinger’s interpretation of the wave equation which had already come out. As you remember, Schrodinger introduces this as being, ‘the wave is the thing’. And he doesn’t like probability amplitude approaches, or anything of this sort. The cause of radiation is an actual interference of two real waves, so that the orbital frequency becomes an important frequency again, and the emitted frequency is simply a physical interference frequency. Now those issues must have been alive at this time with Schrodinger, or at least they may very well have been alive for you coming from Copenhagen.

Dennison:

Yes, I think they were. I think probably I wasn’t disposed to argue with Schrodinger about it. I felt that his approach was a very literal one. I don’t think that I was at all convinced that this was correct, but rather that he had stumbled on a method which worked. And to some extent I think Schrodinger himself had this feeling, that it was a method that worked. And then he took a literal interpretation of it, and this was the one that Bohr did not take, and neither did I feel that this was very correct. I didn’t really talk too much about this, at least trying to convince him about this point at all, but rather I learned what he could say about it. I will just re-emphasize this point, that I think it was essentially this very literal viewpoint, and that he hoped that perhaps as the thing developed, it would indeed reveal that it was just actually a wave for which you could in fact find proper interpretations. For example, the deep puzzles as to why for instance this implied that you could get fractional electrons out of anything. A piece of a wave was a piece of an electron. That this was indeed Schrodinger’s feeling about it at that time. The deepness of the puzzle I don’t think really occurred to him. That’s my feeling.

Kuhn:

This might be a good point to talk about this piece in Nature which I had missed. [“Wave Mechanics and Rotation of Homopolar Molecules” Nature (26 Feb. 1927)]. This was submitted from Zurich, and it’s the one that raises this very interesting point about the possibility of solving some of the molecular band problems by using ψψ* as the thing which repeats.

Dennison:

Well it was a confused era, and this is a confused paper. It was trying in a sense to understand getting just alternate energy levels for the homopolar and yet at the same time not destroying what came out for the heteropolar. And I must say I found it difficult to really appreciate what the differences were between these two. But that’s what I was trying to do, just to put those two things together. As I say, I certainly had the notion here that perhaps one way out of it was this question that it was the single-valuedness of the density function rather than of the wave function which was important. I’m sure I must have been influenced in this somewhat by Heisenberg’s viewpoint that only observable things were really to be counted, and puzzled as to why it was that this was not a more accepted view.

Kuhn:

Did you talk at all with Schrodinger about this substitution of the density function?

Dennison:

Yes, I certainly did. I talked about all of these things with him. Again, my impression was that he was interested, he thought they were puzzles, but he didn’t have anything actually to contribute to it. That was really the feeling I had.

Kuhn:

You do make this remark which puzzles me a good deal, that you think Heisenberg had also perhaps suggested this use of the density function rather than the wave function in choosing quantum numbers, and I can’t —

Dennison:

Oh no, no, no — that was a misunderstanding. Did I actually say it there?

Kuhn:

Well, you say, [reading] “A way out of these difficulties may be found by employing a new postulate in the wave mechanics, namely that given the wave equation of Schrodinger we shall seek such solutions to fit it for which the function “ψψ*” -– it’s ψ‾ψ here — “rather than ψ alone shall be regular single valued functions, where ‾ψ is the conjugate to ψ. This assumption, which we believe has also been suggested by Heisenberg, has much to recommend itself.”

Dennison:

Gee, I really don’t know.

Kuhn:

Now there certainly is a kinship to the Heisenberg approach there, as you suggest. Could that have been added in the proofs as a misunderstanding of the correspondence with Klein and Bohr? [See letter from Klein to Dennison, 22 Jan. 1927]…

Dennison:

I think what they did was to keep it there, so evidently this must have been —. I don’t think it came back to me.

Kuhn:

Well Klein suggests that you may want to make a change when, you see the thing in proof.

Dennison:

Yes, Klein makes that suggestion, but actually it doesn’t look as though I did. It would be quite normal, perhaps from Klein’s letter to think that I would have put in some such thing as this. Whether this could have been slipped in in proof — it’s possible that it could have been slipped in in proof without my actually putting anything in. The publications in those days would sometimes do that for you, make a very short change. That may have been that I was just leaning on what —

Kuhn:

Ja. It would still be a misunderstanding, because I think it’s clear that in that paper Heisenberg does not suggest anything of this sort. What about the idea of the solutions that does not communicate with each other, which comes out here and which is clearly the point that ties it to the Heisenberg paper that’s still to appear? You talk about two sets of complete solutions, one with even values of j and one with odd values of j, but complete sets which do not combine. And this notion of non-combining terms in this way is a quite odd one. You don’t talk about their being symmetric and anti-symmetric.

Dennison:

No. I would guess that what I had in mind here was thinking very strongly about the matrices and the matrix elements connecting the different states. And this fact that you can have a complete set with the integers, in which you have all the appropriate combinations and the appropriate symmetries, that you can have a complete set with half-integers, and that there are no matrix elements that connect these two, I think that was what I was really getting at. So it wasn’t in a sense deeply from the notion of the equality of particles — which I wish I had had — but it was probably more from just looking at the matrices themselves. I think that’s probably where it came in.

Kuhn:

You say at the end of this paper that you’re going to print a much fuller version with the computations. So far as I know you never did.

Dennison:

No, no, that’s right. I suppose one of the reasons was that I felt that probably the Copenhagen school with Heisenberg had really gone much further with this and had begun to discuss symmetries and things of this sort and that consequently this had now already been explored. That was probably the way it went. Well, then there was an interval. I left Zurich. We went to Munich and were there about six weeks as I recall. I went to the Institute a few times. I met Houston there, also the Paulings, Linus Pauling was there. I can recall very well that while I was there he had received the invitation from Cal Tech to come back as an Instructor. He was very much pleased with this. That had just come at that time. They were very nice to us. I had forgotten about it, but Helen reminded me the other night, that while I was sick there in the Pension he came to see me several times. He brought news and flowers and things. I had the misfortune to have a very bad case of flu there — we would have been scared to death if any of my family had as bad a case now. I have a personal recollection that a very nice German doctor came who had been a prisoner of war of the Americans during the First World War. He was very nice to me, came for a while every day to examine my chest. He didn’t have very much money though.

His equipment consisted of a small block of wood, and this was his stethoscope. He’d put the block down on my chest and then would listen at the other end, and kept on moving this all over. Principally he prescribed some wine and things of this sort — there just wasn’t very much. But he did keep fairly close tab on me during this period. So actually this time in Munich was not very fruitful and I didn’t get to talk to Sommerfeld too much at that time. From there we went to Copenhagen and must have been in Copenhagen probably in late February and March. This was the time when Heisenberg gave his colloquium on the uncertainty principle, and this was just absolutely wonderful. Sometimes when a great discovery is made, the people who listen to it don’t quite appreciate that it was. Certainly for us we did appreciate it. Suddenly the thing just opened way out, we saw all kinds of implications and ideas that had not existed up to this time. It was really one of these wonderful experiences to have heard that.

Kuhn:

How much had those problems been on your mind previously?

Dennison:

Well, this is hard to tell. I certainly was up on all the developments that were going on, and of course very much concerned with the problems of particles and waves. We had seen the Heisenberg matrices developed and based in many ways on Heisenberg’s contention that it had to be observable quantities which came in, and then we had seen the Schrodinger wave equation with its waves, but still completely ignoring the fundamental physics. It was a way of getting energy values and matrix elements and so on, but it quite ignored, I think, the physics. And suddenly this was that wonderful uniting thing which brought it together; then one suddenly saw it. One could live with this dualistic approach to it.

Kuhn:

Were there while you were there any further arguments and discussions on this point? Prior to this at least there had been a good deal of friction between Bohr and Heisenberg about this paper and the approach.

Dennison:

I was not too aware of this. I know that there was so to say — I would have thought what you would say would be only a small, minor kind of friction which showed itself only in that some of Heisenberg’s arguments were not completely water-tight. There were some things in the gamma ray microscope which Heisenberg had ignored, for example the aperture of the microscope. And this was according to Bohr — and in fact is — a fundamental point in it. Heisenberg of course was an extremely able guy, pretty young, inclined to be a little bit careless on some points of rigor, and Bohr was much more deeply fundamental, seeing so to say the fundamental implications of this, Heisenberg seeing a little bit just the machinery of it. Now this could be in a way the origin of the friction that you’re talking about.

Kuhn:

One now looks back and tends to think that complementarity and the Heisenberg uncertainty principle are parts and parcel of the same thing, but historically they were not…

Dennison:

Yes. I wish I had more to add to it than I have. All I really remember on it was just this viewpoint that Heisenberg’s was more applying it to this particular problem so to say, or this class of problems, and Bohr’s feeling that there was something deeper behind it. My recollection is that the discussions during those days already on Bohr’s part had this groping for some wider way of looking at it, and that Heisenberg was perhaps a little bit impatient with this, thinking that the real thing was what he had done and not this deeper way of looking at it. But I don’t remember any details of the way in which these two interacted; this was just the impression I had of the feeling of the two. I was continuing at this time to work on molecular problems and beginning to go back over the problems of intensity and how one could understand the broadening of spectral lines and what this meant experimentally. This came out as a paper which was, I believe, sent in after I’d returned to this country.

Kuhn:

But you were already working on it in Copenhagen?

Dennison:

I was already working on it in Copenhagen and had a good deal of it worked out. As a matter of fact I think all the essential things were worked out at that time. Incidentally, this again is one of those puzzles — why this particular paper did not lead to more immediate consequences in the development of molecular theory, because it hit a lot of interesting things, things which have in a way been only rediscovered in the last l5 years. Now it may have been partly the experimental difficulty of intensities which were very great in those days. But lots of the formulae there, things which I guess — well, astronomers use the term (‘curve of growth’) which I had never heard of of course, and later had lots of arguments with (Leo Goldberg) until we finally found out that the astronomers again had rediscovered this and were doing it just the way I had really done it. This was a good paper but it didn’t ever amount to anything. Nothing much came out from it. This sort of thing happens. I might digress for one moment. Later, when I was here, this was in the early or middle 30’s, I was working on the ammonia problem and predicted that there should exist a microwave line, not called microwave, but a line at about one wave per centimeter. And Professor [N.H.] Williams here and one of his students, [C.E.] Cleeton, constructed a magnetron which made waves down into the millimeter region and actually found this line where it had been predicted. Now this was, as far as I know, the real beginning of microwave spectroscopy. But simply nothing happened after that. Williams and another one of his students did a little bit more with somewhat longer wave lengths, but just nothing happened. He bad the apparatus set up here in the building with these big parabolic reflectors; you could walk in front of the beam — he had this microwave modulated with a record or something — and the thing would go down. You could put pieces of wood in; when you pub the grain of the wood along the electric vector one thing happened and when you put it crosswise something else happened. He really had a lot of stuff to go ahead with. But then nothing.

Kuhn:

That is odd!

Dennison:

Isn’t that funny? Until the war came, and of course then with the development of radar and radar plumbing and all those things, then suddenly after the war everybody took this up, starting again with the ammonia line, but now with high dispersion, getting all kinds of things.

Kuhn:

When was this work of yours and Williams done?

Dennison:

Well, it was certainly about in the middle 30’s. Well, I was working on this problem, and I think that Fowler at Cambridge had suggested that I should go there.

Kuhn:

Had you met him in Copenhagen?

Dennison:

Yes, I had met him and talked with him a good deal in Copenhagen. Of course he was always interested in molecular problems. So I went there, probably in April and May. We talked of course a good deal about molecular problems. He was giving a course, probably the background, or associated with the book he was writing in statistical mechanics, and he suggested that I give three lectures. The first of these, as I remember, was on the general things of simple molecules of water and ammonia and methane, and in fact just what vibration rotation spectra were like. The third one, actually, as I gave it, was on the shapes of the spectral lines and the intensities. This was the work which later was published from this country. Incidentally, it gave measures of the size of molecules. This was really using the Lorentz broadening formula which then gives the Lorentz form, and then I could calculate the number of collisions per second and the size of the molecules. This shows one of my personal characteristics. I blush to say that actually at that time I did not really know of the Lorentz formula. I didn’t know what Lorentz broadening was. I had never actually been exposed to it, and so I was myself rediscovering it. It was a couple of years later before I found out that this was one of the more or less well known things in physics. Well, this is characteristic. I enjoy always doing things myself, checking with other people later, but doing it first if I can. The middle lecture I didn’t have enough material for, and I was running around trying to find something that was interesting and new. It was then that I hit on the problem of the specific heat of hydrogen, of relooking at this. I suspect what I may have done was to look again at Hund’s work and at the various trials that had been made to give a correct specific heat and what was wrong with them.

Kuhn:

Now you had actually also referred — and this was one of the reasons I was interested in this paper in Nature. You do refer to the specific heat of water in that paper.

Dennison:

Yes. Well I was interested in those problems all right. As I remember I was a little bit glib here in thinking that the agreement wasn’t too bad using existing ways of doing it. For this lecture I evidently was turning over all possibilities and this one occurred to me and I was very pleased to see that it did come out very well. So I presented it.

Kuhn:

Did you work on that lecture in Copenhagen?

Dennison:

No. This was in Cambridge. I was probably getting a little nervous that the time was getting close for it and I didn’t have enough material there.

Kuhn:

Now how does it happen that you had immediately available to you then if you were in Cambridge, these measurements of Hori? I notice that you particularly point to them to say that the Hund two to one has to be wrong. This suggests that the problem at least may have arisen for you while you were still in Copenhagen.

Dennison:

Oh yes, I certainly had thought about the specific heat much earlier, and how it would be related to the spectra and to the moment of inertia. Hori was one of the Japanese physicists, a very young man, who was studying there in Copenhagen. Now I think his experimental work was done there also, and that he had come up with this value for the moment of inertia. I really had all of these in hand and in fact had all of the measurements on specific heat — various people who had measured it –- so all the material was there to go ahead with. As you say, it was a problem I had been thinking about. I’m sure there were many things that I worked on — some came out and some didn’t, and this was one that hadn’t come through. But fortunately for this lecture, it did, and this was the right notion. Fowler was very much interested in it. When I finished he asked whether I had published it or was going to publish it. I said I just hadn’t thought about it; it had just been done for this lecture. I showed him the curve that I had made, that gave the predicted specific heat, and the I, and the measurements. He was very much taken with it and wanted me to leave the curve with him. He urged me to publish it as soon as I could because he wanted to use it in the book. I did write it up and I sent it on, but through Professor Bohr, who sent it in to the Proceedings of the Royal Society. Bohr wrote that very nice letter to me about it.

Kuhn:

I think it’s actually communicated by Fowler to the Proceedings.

Dennison:

Oh, was it? Maybe so. Well, if that is it, then that’s right. Let’s see, I will have it here someplace.

Kuhn:

Ja. [Reading] “Communicated by R. H. Fowler.”

Dennison:

Then that would be right. But I certainly must have sent a copy to Bohr and got this very nice letter back. Now here’s of course this difficult thing of picking out what I really thought of at that time. I have a very strong impression, for whatever memory is worth, that I was perfectly aware of the meaning of the three to one and the two different classes, that these were just the natural result of the spin of the proton, extending to the proton the same thing that had been done with the electron. And it had seemed in fact so absolutely self-evident that in writing it up I just wasn’t going to push this point, it was too clearly there. It just didn’t occur to me to do it. And when Bohr wrote back that one should probably do this, I agreed that this was all right to do and maybe did improve it somewhat. So when I got the proof I added this in proof. Now it wasn’t in any sense, one should say, anything like a discovery. It was just one of those things that was so completely in the air in those days that you said, “Well this is just naturally so.”

Kuhn:

I thought when you first told me the story that I would push you a bit on it. As I read that paper, it isn’t simply that you haven’t pushed the point previously, but when you criticize Hund, among the reasons for being skeptical about Hund’s two to one result, one of the things you mention is that it isn’t what one would expect from the statistics and there isn’t any mention of it at all until the note. So that it comes in as a brand new idea in the note. You’re very likely right, but it certainly would never have occurred to me reading it. And it is an early date to make this sort of statement, that the thing was perfectly clear to you, that this meant that the proton must have a spin, that if it does this is just what we should get. But it’s certainly not the first time anybody’s mentioned the possibility of proton spin. Hund mentions it too.

Dennison:

Ja.

Kuhn:

But the very fact that Hund doesn’t know what to do with it and that it ought to give three to one is itself leading here.

Dennison:

Ja. Why did Hund put in two? Just because the thing worked?

Kuhn:

He gets two, or approximately two because this is the way the thing fits best. As I remember it, the paper does include one or two remarks about the possibility of this being related to the proton’s having spin, but I’m not — I’m sure he says something about the proton’s having spin, and I’m equally sure that he doesn’t see that if that’s right then one ought to have a three to one ratio.

Dennison:

Or he probably tried the three to one and it doesn’t work. It has a bump in it where it shouldn’t have. But just to corroborate in some way what I’m saying, that this was all in the air: I make a remark here that [reading] “The coupling of the nuclear spin with the spin of the molecule which determines these transitions will indeed be very small, much smaller than the coupling forces between the electronic spins in the orbits which give rise to the very weak transitions between ortho and parahelium.” So I was simply thinking always that it’s this same sort of thing. Here are two particles, spin one half, the transitions with the orbital, which in this case is the rotational, motion is going to be very small. But it’s just like the corresponding thing with electrons. So I probably had the feeling, when you say it’s something like this, that here are these spins and they are just like the helium one but the transitions will be very much weaker, and so you make the assumption that they don’t occur during the time of the experiment and then it begins to come out all right. No. It was not a discovery of this, but it was just one of those things that was so self-evidently in the air that when it came out, ‘well, yes, this is what one expects.’ It had already been done for helium, with electrons. Here is exactly the same problem of the three to one; it comes out exactly as before. And also it’s the right one; the odd states are the ones with the right three to the one, or whichever it is. So this is the way it seemed to me. I must also say that in this, as in many of the other things, sometimes as I look back and try to think what I was thinking, I remember the stupid things but don’t remember the clever ones. I’m just saying that this is true, because when someone suddenly brings up something else that I wrote at that period which shows that I did in fact know what was happening, even though when I look back on it now I have a feeling of puzzling and that I didn’t really know and didn’t know until later what was up. But I knew a lot more, and was doing a lot more than I actually remember. I’m sure this is true, one always remembers the stupidities, but the things that were a little clever you forget about.

Kuhn:

Ja, I think that is very often true. Well now how long were you in Cambridge then?

Dennison:

It was probably about two months.

Kuhn:

Getting there, what, April?

Dennison:

It was probably April and May. And then we left there and went for a very short time to Leiden, and I gave a colloquium on this for Ehrenfest there. Ehrenfest was moderately skeptical. He did not immediately say ‘well this is evidently the answer to it.’ It seemed a little odd to him somehow or other.

Kuhn:

Do you know what it was that seemed odd to him about it?

Dennison:

I don’t know. I suppose just this notion that you could have two sets of states which you knew had to be able to combine, and yet that in the time of an experiment they didn’t combine. All molecular and atomic transitions are relatively short, and here was one which was really quite long. I think this seemed very odd. He was very nice to us. We went to his house; his first remark to both Helen and myself was that we didn’t eat enough, we ought to get fatter. [Chuckle] I expect we were pretty lean in those days. Then he suggested that I should go out and visit Lorentz who was retired at that time and lived in a neighboring town. We went there by train and Ehrenfest, always full of advice for people, said, “Now when you go there don’t spend all your time telling about this specific heat business, but try and learn something from Lorentz, because he is a great man and it’s much better to do that. Otherwise you’ll just have told one more person and this doesn’t matter.” So I did. Lorentz was very nice. We had tea — he and his sister I think it was. He was of course very much interested in quantum theory, but so to say not thoroughly convinced of it. In particular the problems of the corpuscular nature of light versus the wave theory — I don’t think that he was understanding very much, or at least believing, what was happening to bring this as a proper dualistic theory. He still was in a way struggling between these two. Lorentz told me an example which I always use then I come to the appropriate place in Maxwell theory, which is quite interesting. Namely, it’s one of the problems in which you have light which impinges on a surface which is partly conducting and partly has an index of refraction. Now if you ask the question as to what are the forces on a very small layer which is right at the surface, and you write down the Maxwell stress tensor and calculate the forces that are right at the surface you find to your astonishment that these forces are not in the direction of the impinging light, but are in the reverse direction. This is correct. He said, “Now how can this thing possibly be given by a corpuscular theory in which you have particles which are coming in and a fewer number are being reflected back out again? How can it be that the forces given by these particles coming in could be in the reverse direction for this thin layer on the surface?” He was very sure that the Maxwell theory was correct, that that answer was correct. And how would the corpuscular theory ever deal with such a problem as this? [chuckle]

Kuhn:

That’s a nice point!

Dennison:

Quite a cute problem! Now me with my feeling of always using correspondence principle and things like this, I agree with him thoroughly. The answer of Maxwell is undoubtedly right. If you do this a la quantum theory you will also find, the same answer I’m sure. But I’ve never actually carried this one out. Well that was very nice. Maybe just another slightly personal remembrance. At the time when I was there there were two young men who had come to Leiden and they were just starting on a European tour, I think probably just for that summer. One of them took me aside and said, “Well now you’ve been here in Europe for a long time, give us the low-down of all the things we ought to do and how we’re supposed to get along.” These two guys were Jesse Beams and Ernest Lawrence. [chuckle] And of course they were talking about rotation problems but no cyclotrons yet.

Kuhn:

You’re still in Europe in the spring of ‘27, when the Wigner group theory papers begin to come out. Did you see those there do you suppose?

Dennison:

If so, they did not make an impression on me. They did not make an impression on me.

Kuhn:

Then do you begin actually in your own work to see that as a relevant tool? I notice that up to ‘30 you really make literally no use of group theory, though you do, as Hund does, some problems that might be handled by group theory, by more straightforward techniques.

Dennison:

This is correct. And it is again one of the weaknesses of the way in which I do things. Instead of spending the time to really learn group theory, at this period at any rate, and apply it, I find that the things that I want to do in molecular theory I can do without the formalism of group theory. Now I use, so to say, the underlying spirit of group theory, that there are things, for example selection rules and all kinds of things that can be found in molecules, using the group approach. I took as the principle point that it existed, that there were these properties which depended only upon the symmetry and not upon the magnitude of the forces. Then the next step was, if this is so, then by using ridiculously simple and unphysical forces, which however have the appropriate symmetry, and observing what are common properties with a couple of these, I can infer what must be the symmetry properties of the thing. Since this worked so well in all the things that I tried it on, for me it became a very powerful method. It never was of any use to anyone else for some reason or other. No one ever used this particular thing. But in many respects it is still a very powerful method. If you appreciate that there exist only these properties, that to find them you can use any kinds of forces that you want that are so simple you can calculate on your fingers, then you will find all that is enclosed in group theory really.

Kuhn:

Have you really always gone on using that method?

Dennison:

And I’m afraid that I have mainly really used that. At one stage I did learn group theory but for the most part forgot it again because I didn’t use it directly. And I know enough of the notation so I can now write papers which will sound as though I know it but in fact I don’t get the results that way.

Kuhn:

That’s very interesting.

Dennison:

It’s just a trick of my own approach. I might say that had I been cleverer I would have fastened on this very much earlier. The reason is that when I was originally working on methane, for some reason or other I got a hold of a paper by Brester on crystal structure which essentially gave the kinds of things that I was finding. I recognized that there was a similarity between the two, and puzzled a good deal on this paper, but I didn’t know enough to go back to the fundamental mathematics and learn group theory so I could read Brester’s paper and from that point on use it.

Kuhn:

…That’s a paper I don’t know. That is very interesting.

Dennison:

It’s a very early paper. It’s not exactly on this, but it was so close that I could have really inferred how things went. Nevertheless this other method has really worked very well. There are an awful lot of properties that you can get out by this very simple-minded approach.

Kuhn:

Good. Well you then came back here, what, in June, July?

Dennison:

July probably, or the first part of August. I began as an Instructor, and, as the letter from Randall shows, with a salary of $3000. I received the first promotion I think after just one year, and then promotions came fast enough so in ‘35 I became full professor and went on from there. I was very much interested in molecular problems always, and of course this was a center where a lot of experimental work was going on. My own work has been mainly in the details of the simpler molecules rather than in the more complicated ones which now are of more interest to chemists and biochemists. In the talk that I gave in Copenhagen this summer I was slated to talk on molecular theory. What I elected to do was to stress the fact that molecules, at least leaving aside the electronic part, form, in a way, I think, the most beautiful and complete examples of classical quantum theory. And that in a way this played a certain role in the development of quantum theory because at that time there were relatively few such nice examples. This was a way, in a sense, of validating it with many further examples of how well the theory worked.

I think to a great extent this has been my own feeling about it, that it’s never been the application, but it’s been the fact that it’s been a very beautiful kind of example of how this non-relativistic quantum theory will in fact predict so many beautiful things about molecules. I gave then just a few of the simple examples, the ammonia inversion, the asymmetric rotator as shown by the water molecule, and finally the molecule that has internal angular momentum in the methyl alcohol in which the theory predicts simply a beautifully regular array of energy levels inter-twining with proper periodicities — very beautiful. But when you come to actually write down the transitions these then lead to an array of lines which are about as chaotic as you can imagine. And this business of taking this very chaotic array of lines, hundreds of them, and then reconstructing what the levels are, and finding that the levels have this beautiful smooth periodicity coming out of this array of lines, this gave me a great deal of pleasure. This was in fact the spirit of what I was saying this summer.

Kuhn:

You never do really get yourself involved with electronic levels?

Dennison:

Actually no, no I never have. I think it’s probably just for the same reason that while this is an interesting sort of problem, for me it does not have the beauty that the other ones do. At this moment it just doesn’t. I mean there are no solutions that are close enough to being exact solutions, no way of really dealing with these things, and to me the notion of putting a problem on a computer and getting out an answer that fits or nearly fits is not particularly interesting.

Kuhn:

There is one other development that we haven’t talked about, which again opens up before you get back from Europe, though it’s main impact is after you’re back here; it’s basically the Heitler-London theory for homopolar binding, but of course what goes with this is a continuing debate that goes on at least into the early 30’s, that I know very little about but keep running into, which is between the adherents of this approach and the adherents of a molecular orbital approach which for a while seemed to people to be in conflict with each other.

Dennison:

Well I think the Heitler-London is beautiful — in the same sense as the things that I have liked to do. Well if you now ask any molecular spectroscopist who works with electronic wave functions, “Do you have any other examples that are in any way as nice as hydrogen, where you can actually compute things and they fit beautifully, and it all goes together very nicely?” — the answer is really no. Except, say, the hydrogen ion or something like that. But with the more complicated ones there are always so many hedges on it.

Kuhn:

You’re telling me now a contemporary attitude which I’m sure is right. I wonder how this seemed at the time? There were fights over this one.

Dennison:

[Yes].

Kuhn:

Were you involved in any of them?

Dennison:

No, I wasn’t really, because I didn’t work in the electronic wave functions at all. But I was very much pleased with the Heitler-London paper. I think I was particularly pleased with the concept of it, with the method that was employed. Now with the later calculations, and which particular functions you should use, and the mixing and so on, this was only moderately interesting to me. So that’s why I would not have gotten into the controversy.

Kuhn:

What about this paper you do after you’re back on the nature of light, the proposed experiment? What is the background for that?

Dennison:

Really very little. I think to a large extent among the people with whom I came in contact, except perhaps Uhlenbeck, perhaps Goudsmit, I felt that there wasn’t very much deeply understood about the problem of the dual nature of light. To illustrate this, here was a kind of problem which was so clear when you thought about it — what the thought-experiment was — but if the thought-experiment was not appealing then actually here was something that could be done. But it was more I think just my pleasure in thinking of the thought-experiment… That you thought of a diffraction grating and the fact of having different spots of light off on one side or another was so intimately connected with Maxwell theory and the interference of light — now just how would this really go when you also considered the individual events? It was more taking a specific example and thinking of this one. So I’m not sure that I really ever thought that anyone was going to do it or that it was even necessary to do, but it was perhaps amusing to point out that if you had any doubts about it, here was an experiment that in fact could be really done.

Kuhn:

What did you find? Did people have doubts about it?

Dennison:

No, I don’t think as any result of this paper at all as far as I know. Perhaps other people reading it had the same feeling that it helped a little bit towards the general understanding of what light must be like, and it was not really crucially necessary to do it. Perhaps events kept moving so rapidly then that it became less and less necessary to do it. I don’t know.

Kuhn:

When you get back here, clearly at this point Michigan has a larger collection of people who know what’s going on in modern physics than any place else in the country. Is that right?

Dennison:

Well, I think it was at least one of the very particular centers. This was of course Randall. Randall began to have the notion that theoretical physics would pay off very well indeed, and he had first thought that he would want me to return and that then he would get some European physicist to come over, also a young man and a theorist. He sent Colby on a recruiting expedition, and I believe it was Ehrenfest who pointed out Goudsmit and Uhlenbeck. Ehrenfest advised him very strongly that he should take the two of them, that he had a much better chance. You had to have several people together, and it was much better to have two than it was just to have one of them. And that was quite a step you know. Here was a small department with perhaps not more than a dozen people, probably less, in it. And then to suddenly put on three new people, it was quite a step I think he must have had to do quite a selling job with the administration to get it. But he was quite right, quite right. This was the thing that was going to pay off. It went very well. Goudsmit, Uhlenbeck, and I immediately organized theoretical colloquia that hadn’t existed up to this time, and got students interested.

Kuhn:

Had there been no colloquium previously?

Dennison:

Yes, yes, there had been a general colloquium once a week.

Kuhn:

And that went on?

Dennison:

That went on simultaneously, but it turned out that before too many years it was the theoretical colloquium that got to be the live one, and the other one just gradually died out. It lasted for another ten years probably, but then ours became the colloquium. That’s the way it went.

Kuhn:

How sharply were the lines defined between the ‘new-guard’ and the ‘old-guard’ in the department?

Dennison:

[chuckles] Oh Randall must have had his difficulties! He must have had his difficulties I can appreciate them more and more, how (close) it must have been. Here were these young people who had come in; in order to hold them he had to advance them pretty rapidly. I know he told me that in talking with other people around the country, other chairmen, he asked them to keep their hands off for a period of I think two years. And I think they did. Then he did make these advancements. People who were already here then didn’t get advanced as rapidly, and there certainly must have been some friction. Randall was so clever about it though that it was not very evident; it really went all very nicely.

Kuhn:

Did this older group at least appreciate the physics, or were they skeptical about this whole thing, modern physics, modern mechanics?

Dennison:

I expect they were moderately skeptical. I am now a very quiet person and take everything that comes along, and it was somewhat of a shock to read in one of the papers, the letters that Helen dug up. This couldn’t have been one that I showed you, but this was one in which I was writing I think to my mother and was describing the colloquium that I’d given here. This must have been the first fall. I had all the exasperating impatience that I sometimes see in young people now and have to kind of try and play down, you see! But I said something that — well. Of the people that were there, there were 60% I think that didn’t understand a word that I said, and that promptly went to sleep. Then there were about 40% that kept awake but didn’t understand. Then there were only about 10% that did understand it, and while I was very pleased to have done it for the 10%, it did seem a waste of time for the other people to go there who might just as well have stayed away. So I was just as impatient as anybody else at this period. I can remember one of the older physicists, not from here, but at one of the Physical Society meetings, sort of standing off and looking me up and down, and saying, “Now, are you one of these young newer physicists?” or something like that [laughter]. It was quite clear that there was a good deal of potential friction.

Kuhn:

When people drew this sort of distinction between the younger generation and the older one, was it that the younger generation was so theoretically and mathematically inclined, was it quantum mechanics with sort of the substantive beliefs about Nature that were involved?

Dennison:

Well, I think that it was the substantive beliefs about nature plus the fact that we did know quantum theory, and they didn’t. This is in fact true of many really very clever people, that it was awfully hard for anyone I think who was fifty or older to accept these notions and to really be able to go ahead and use them. There have been examples of people that I know very well, whom I am very fond of and are very clever people, but they just — at the most they would be able to read and say ‘I understand each step of this,’ but to have any deep feeling about it, so as to be able to use it, or deeply believe it, this they were not capable of doing.

Kuhn:

Fifty seems quite old for you to draw that line. I’d have supposed it might have been drawn even younger.

Dennison:

It might very well have been younger. I was going to say forty, and then I was thinking of the actual ages of the two or three people that I was talking about, who must have been fifty at this time.

Kuhn:

Were they people who had previously been doing old quantum mechanics, or not even that?

Dennison:

Some old quantum mechanics, yes. But not having come up through it in the way that we did. To have known the old, to have felt the impact of the passing of the old and the fact that you had to take on these new ones and then to have really learned that stage of it was our great advantage.

Kuhn:

It would interest me, though I obviously don’t want to press you to do it, if you would put names to the people you have in mind largely because this phenomenon of having the field change under you, and then not being able to go on with it, and its impact on one’s own career is something that needs I think more investigation than it has had. To find, people on whom one could try to investigate this, by looking at their work would be of real interest. On the other hand I appreciate that you may not want to put names to this. I can’t imagine its doing any real harm.

Dennison:

Well, I’ll give you just one name, and he was in a slightly different field, that’s Professor Fajans who was here. Really a very fine scientist and one who did extraordinarily fine work with the isotopes and things of this sort. Hevesy was the one who actually got the credit for it, but if you look back, they were very closely together at one earlier stage. Very much interested in molecules, but never really understanding what quantum theory was doing for it; always a struggle with him. So in a way he made up his own variety of quantum mechanics which doesn’t fit with the existing ways of doing things. But a very good scientist. Just one of those things. After you get to a certain age it’s really very hard to pick these up. I’m sure I’d have the same difficulty now to be able to really work in modern theoretical physics. I listen to it, I understand a fairish amount of it, but to be able to be creative in it — I’d love to, I wish I could, but I’m afraid it isn’t there. I’ll always say it’s because I haven’t time to work on it, but the truth is it’s probably just not there. [chuckle] That’s it.

Kuhn:

About the early 30’s there are, as you must know, certain people who say, ‘Now look, quantum mechanics isn’t interesting anymore’; and make the transition over to nuclear physics. ‘I think that’s where we’re going to really learn things from now on.’ Did you see much of that, did you have any such feeling yourself?

Dennison:

To some extent, yes. It was clear that this was going to be a very exciting, interesting thing. At times I’ve made the attempt to learn some of it but have not really ever worked — I’ve written one or two papers on nuclei, but always from the point of view of molecules really. I think it was because here were still so many problems that I found just fascinating. It was nice if anybody else enjoyed them, enjoyed what I found, but I don’t think it mattered very much to me. It was just that it pleased me to be able to understand some of these spectra and to be able to relate them to what seemed to me a very nice quantum problem.

Kuhn:

The fact that, at least probably after 1930 or thereabouts, it seemed reasonably clear that nothing fundamentally new beyond the problem itself was likely to come out of this work was not something that made a great deal of difference to you? This is not meant to be critical or anything of the sort. For some people the fact that these were now just puzzles was bothersome — for other people the pleasure of the puzzle was all that was required.

Dennison:

Right. I don’t think it bothered me particularly. I just enjoyed what I was doing, and found it exciting and beautiful, and that was it. Molecular theory, molecular structure problems were in a somewhat better position than atomic spectra. Atomic spectra in a certain sense appeared to come to an end sooner than the molecular ones did.

Kuhn:

Why is that?

Dennison:

I’m not quite sure why it is.

Kuhn:

At least if one stays away from molecular electronic spectra.

Dennison:

Yes, you would think it would. Well, there were still new phenomena that were occurring you see. For instance, well, this problem of inversion. While one realized it quite early in the ‘30’s, still the real working it out took certainly into the middle of the ‘30’s anyway. Then there were these problems of the internal, so-called hindered, rotation which just occurred in about the middle 30’s. We worked on them until the beginning of the war. Then there were problems that came out from the microwave work which then furnished very interesting problems that hadn’t existed before. So one could say that there were still quite exciting things. Also, many of the problems that we felt could be done had never been thoroughly worked out. This question of the forces between atoms in a molecule, how they vary from molecule to molecule. For instance, [P.M.] Morse had this very amusing notion of the Morse function for a diatomic molecule which works very well and has lots of implications. Now, similar sorts of things for polyatomic molecules I feel sure must exist. I’ve tried it at times, and gotten a little ways along, but never too far.

So there were these things that were developing — one that Fermi worked on, the so-called Fermi resonance, first in carbon dioxide, that explains something that no one had understood before. So they were coming along. Whereas in spectra, somewhere in the early 30’s you began to see that it was just one more complicated one, then another and another, but actually there was nothing really new that was coming out it seemed. Well, Goudsmit was awfully good in this field, and very productive. He did a lot of things, and this extended up to, as I say, about the middle 30’s. At that time he had the feeling that the atomic spectra and their problems were going to be of no particular interest, and the nuclear ones were. So he in a sense almost artificially changed his field at that time. To some extent I followed along; at least we read a lot of things, and had colloquia and so on. But I kept on working my molecular problems. He stopped working the atomic ones. Now, although he’s done some things with nuclei, I’m sure that he would feel that he had never made anything like the kind of contribution or the volume of work there that he did with the atomic ones.

Kuhn:

That certainly is true.

Dennison:

And I’m just not sure that it wouldn’t have been the same with me, that if I’d gone into nuclear work I wouldn’t really have ever gotten too far with it. Just too old! Maybe at 35 or 40 you begin to stop having that —.

Kuhn:

When would you say the department here began to pay really significant attention to nuclear problems?

Dennison:

Probably about the mid-thirties. At that time [J.M.] Cork had the notion that he wanted to build, first a Van de Graaff — which he did. Then he wanted to build a cyclotron. We had what was the world’s largest cyclotron for about 6 months here. It was one of the very early ones; did very well. He wrote dozens and dozens of papers, never quite got all the pay-off on it because of perhaps a slight superficiality in places on it. But he at least had the quality that he could go into new things getting later on in life. He must have been forty at that time. And Dick [H.R.] Crane came towards the end of the thirties; then of course all our summer symposia began to have all these nuclear problems… You did ask one question, what happened to the second volume of this [Paper No. l0]; …this first one, the review article of infra-red spectra, was in the summer of 1930. I wanted a nice place to go, and there was a Physical Society meeting in June at Cornell, so I went there and inquired around a little bit whether I could find a job to pay my way during the summer, and there weren’t any available. This was a very good thing. I think I had to borrow $200. I wrote this in the morning and we walked around in the afternoon and I’d sort of composed my thoughts, and then come back and worked all the next morning, and it was a very good paper, well worth doing. The sort of thing that nobody will do now, of course, to so to say pay their own way, and I had thought that I would write the second part of it rather soon, but —

Kuhn:

Did one actually get paid for these papers to the Reviews of Mod. Phys.?

Dennison:

No, these were not paid for. But then I had a sabbatical in 1939 and 40, and it was at that time that I wrote the second part… I was very sorry it took ten years to do it, but I was busy and writing a lot of things in that period too. Well, I think you’ve gotten me pretty well dry of all the history (I went through)!

Kuhn:

It’s been immensely helpful and I’m very very grateful to you.

Dennison:

Well it’s been very nice indeed.