Philip W. Anderson - Session I

Kojevnikov:

Today is March 30, 1999. We are at Princeton Physics Department building recording an oral history interview with Philip Anderson. Perhaps we can start with the usual questions, if you could describe briefly your family background and specifically if there were any things — books or persons — in your early life and school years which could have influenced you towards choosing physics as a field, or could have encouraged your interests in becoming a scientist in general perhaps, and maybe physics in particular.

Anderson:

Well, of course, my father was a scientist. My grandfather actually was a mathematics professor, but it wouldn't really be very accurate to describe him as a scientist because he didn't have a Ph.D. He just was a math teacher at this small college.

Kojevnikov:

Could you give the names.

Anderson:

The Wabash College in Crawfordsville, Indiana.

Kojevnikov:

And the names. Names of your father and grandfather.

Anderson:

Oh, my grandfather was James Harvey Osborne. Known familiarly as Pat Osborne. My father was Harry Warren Anderson. He went to Wabash, where he met my mother, I guess. I never really knew where they met. He had been a farm boy, but he and his brother both went to Wabash and fell under the influence of a very good teacher of botany and general biology named Mason Thomas. And so both of them became fairly eminent plant pathologists. Each in fact was in charge of his own experiment station, one in Connecticut and one at the University of Illinois. My mother eventually earned a bachelor's degree, which was not that common in those days. And was a card carrying member of the American Association of University Women. She tended to be a card carrying member of any organization, and very active in a lot of organizations like the League of Women Voters the AAUW, etc. They have a chapter of her sorority at the University of Illinois, of which my sister and wife are members. Her brother, Pat's son, James Insley Osborne, also became a professor at Wabash. He had been a Rhodes' scholar and seemed to be heading for a very eminent career. He made a lot of very quite influential and important friends, both in graduate work at Columbia and as a Rhodes' scholar. Among others that I remember was Elmer Davis, the radio commentator, but there were several others. But he came down in his second year of graduate school with rheumatic fever, which weakened his heart. And for some reason the story is that he therefore returned to Crawfordsville and took up this professorship at Wabash College. Both he and Pat had enormous influence on the college. They were kind of the old guard ornaments of the university faculty. And there is a history of Wabash College, in fact there are a couple of historical volumes about Wabash College that feature the two of them. So my father was a professor at the University of Illinois.

Kojevnikov:

That's also where you were born?

Anderson:

No, I was not born in Urbana just because of a quirk—my mother insisted on having me born in Indiana because we lived back and forth between Crawfordsville and Urbana.

Kojevnikov:

How far away are they?

Anderson:

Seventy-five miles. It was essential that we return for Crawfordsville for all vacations and important occasions like the birth of a child and so on. For that reason I was born in Indianapolis, not in Illinois. It was almost inevitable that I would become a scientist because my father immersed me in things scientific, and I was fascinated by various parts—

Kojevnikov:

Were you the only child?

Anderson:

No, I had a sister. She was older—she is older. She too became a scientist. She earned a Ph.D. in biochemistry. Unfortunately, at the same time she earned a pair of twins and had to give up, at least for the time being, her science. Married a man who eventually became fairly high up in the Smith Klein French Laboratories, which are now, what is it? Something Smith Klein. It's one of the big drug companies. But when the twins were grown and the other child she had, she took up information science and became a historian of technology and a librarian. I think maybe she even earned another degree in library science, but I'm not sure. She was director of the Hagley Foundation, or acting director there for a while. She became rather an expert on trouble shooting in science libraries, so she rehabilitated the University of Pennsylvania Medical School Library, among other things she did. But then she and her husband took up tree farming and they now run a tree farm in northern Pennsylvania. They're probably the oldest active tree farmers in the country.

Kojevnikov:

Were there many science books at home?

Anderson:

Yeah, I read voraciously all kinds of books from very young.

Kojevnikov:

Anything in particular about physics you may remember?

Anderson:

Not much. I remember there was one— as a real child there was one book about astronomy that fascinated me. I forget exactly who wrote it, but I remember from that the main sequence of stars and things like that, and being very impressed by that. It wasn't really again until I got into high school that I began to read serious science. I very much enjoyed the Liebers' book on relativity and the Gamow's books that I picked up in the high school library where I lived. But I think at that time I already was directed very much toward math or physics just because of one influence or another.

Kojevnikov:

How much physics was there in high school at that time?

Anderson:

Well high school... I went to a very good high school. It's quite small. It still has classes of only about forty per year. It was a university high school. At that time it was not a laboratory school. It was just a high quality high school which was used as a training school for students in education at the University of Illinois. But all the teachers were professors, professorial status, and they were remarkably good teachers. The high school I think has as many Nobel Prize winners among it's graduates as any high school, perhaps anywhere in the world, but certainly in the United States. The same number as the Bronx High School of Science, which is much more famous. There's one economist, one molecular and myself. Three of us. Two out of the three, possibly the third, went through it in that period when it was not an experimental school as far as the subject matter was concerned. Strictly traditional in terms of subject matter, but it was just high quality. It had an opener approach, and you had closer contact with the faculty than you would at a normal high school. So there was a math teacher that was very influential on my future, and I almost, well, I intended to concentrate on math at Harvard when I went there. The other thing the school had was a very strong record in preparing people for the scholarship examinations at Harvard. Harvard was just beginning at that time to experiment with need-free scholarships. Essentially scholarships that would pay whatever a student needed to have to be reasonably comfortable. Not really comfortable, not rich, but to live reasonably at Harvard, live without scrimping. The first of these Nobel Prize winners, Jim Tobin, earned I think one in the very first year of these new scholarships. And he did fantastically well. And then the next year there was one from the high school who also I guess eventually became a member of the national academy and did very well at Harvard. So it was not unreasonable for me to apply for a scholarship to Harvard. And I got it. As a matter of fact, there were three of us that year who won scholarships to Harvard. I was the only one who got this completely need-free scholarship, which is now more or less the standard scholarship.

Kojevnikov:

How much were they at that time?

Anderson:

Well the total expenses for a year at Harvard (you won't believe this) was on the average thirteen hundred dollars: four hundred dollars in tuition and nine hundred dollars in living expenses. My scholarship was eight hundred dollars because I would have cost them five hundred dollars if I had gone anywhere else, and on a professor's salary, that was a lot.

Kojevnikov:

Were there any other choices but Harvard when you were thinking of applying to colleges?

Anderson:

Oh yes. I was making up my mind at that time between the University of Illinois, which of course would have been obvious and easy and free to a son of a professor. Or Wabash, which would have been equally easy for me. And I had some slight preference for Wabash because I had two cousins there. One of them was the cousin that I was closest to. Actually, the relative I was closest to in my generation. And an older cousin, and they both were in one of the most prominent fraternities on campus and I would have automatically gotten into that fraternity and a social community. So that seemed like a reasonable thought. But I won the scholarship to Harvard so I went there.

Kojevnikov:

And at that time you were concentrating on math?

Anderson:

Yeah, now the reason I thought of concentrating on math was that actually I had taken physics in high school and I hadn't liked it.

Kojevnikov:

What year was that?

Anderson:

I went to Harvard in 1940. At that point I was only sixteen. I entered grade school a few months early. My mother, as I implied earlier, was a rather determined woman, and she saw to it that I got into grade school, when I was still five, with my birthday, the class of my birthday. She did this with a slight wangle, which is irrelevant here. But then one of the things at Uni High, which I went to, was that you automatically skipped a year when you went there. There were two ways of skipping. The standard way was to take seventh and eight grade together, the so-called "sub freshmen year," and so I did that. I got another full year ahead. I took off a semester in my freshmen year and we went to Europe, which was no problem.

Kojevnikov:

When was this?

Anderson:

1937.

Kojevnikov:

And where did you go?

Anderson:

We went all over Europe. We actually ended up, spent two and a half months or something like that in England where my father had a lot of professional contacts. And then we spent three months touring the continent, and we drove our American car all the way to Sophia, Bulgaria, and back through Sarajevo and Dubrovnik and Triest and Fiume which was a very wild experience for someone coming from the deep middle West as I did. The whole experience was very exciting for me. I was just the right age, thirteen. Just extraordinarily curious about all possible kinds of things.

Kojevnikov:

How much mathematics did you know when graduating from the high school? Where did the mathematics education stop?

Anderson:

It was fairly standard. Even good prep schools didn't teach calculus in high school. It went up as far as advanced algebra.

Kojevnikov:

And in physics, what was not likable in physics?

Anderson:

The physics teacher was a famous teacher, as a matter of fact. He had earned a couple of awards for good physics teaching, and because he intrigued the students. He interested them. He was full of bangs and demonstrations and experiments and so on. He didn't explain to you any of the intellectual structure of the subject, so I didn't really come out knowing Newton's Laws. We played around a lot with vacuum pumps and things like that and I suppose I must've known Newton's Laws, but I don't remember in fact seeing the equation of motion of a particle or anything like that. There was no real physics. I didn't have this sense of physics as an explanatory science. It didn't interest me. Another influence on my attitude towards physics, and the reason I took freshman physics in college, was the head of the University of Illinois physics department who was a close personal friend of my parents, Wheeler Loomis. The laboratory at the University of Illinois is named after him. And he's the person who built the University of Illinois physics department into the major department that it has always been since that time. And he didn't say much, but he did say to my parents, "Make sure he takes physics the first year. You can never replace it if you miss it." And they listened and I listened, and I'm very happy that they did because that got me into the concentrator's course in physics. And from the first few days of that course I was hooked.

Kojevnikov:

Who was teaching?

Anderson:

Wendell Furry. The man who was later notorious for almost being fired during the McCarthy era. The second year was taught primarily by Street, who was a particle physicist. But the first year was marvelously taught. Furry was a very good teacher. And it really, really committed me. Although I stayed in math for a couple years, I left fairly soon and took applied physics. So the influences were really Wheeler Loomis and my parents' friends in the physics department, and this one math teacher, Miles Hartley, who was a superb math teacher.

Kojevnikov:

Now that was while the war was going on in Europe. Did war have any influence on the teaching and studies at Harvard at that time?

Anderson:

In the first place I think my family was much more conscious of the war than most. My mother, among the other organizations she was prominent in, was in the committee to do something or another by aiding the allies. This was an interventionist political movement at that time. She, I think because of her brother, had always been rather Anglophile, but I think we all were more conscious than most people, starting right back with the invasion of Manchuria and the Spanish Civil War, we were very worried about the events in Europe.

Kojevnikov:

Do you have any memories of this as far as the time of the travel through Europe?

Anderson:

Yes, very much so. Just the difference in atmosphere, which my parents noticed and I think even I noticed as we crossed the border from Holland to Germany. We had a very pleasant host, a colleague of my father's in Holland. She was talking about the horrible things going on in Germany, and well, we didn't see them that much, but you just had a feeling that people weren't talking to you. And occasionally since we were staying in Pensions that were recommended to us by contacts of my parents, occasionally the people we met would whisper things to us very confidentially about what life was like. Then we got to Austria, and we did in fact see demonstrations. We actually saw Schussing, saw the Austrian storm troopers "Heilborg Schussing. And saw, street demonstrations in Austria of various kinds. It wasn't obvious who was against whom. None of them were very nice people. So we saw some unrest in Austria.

Kojevnikov:

Was it during the time the Spanish Civil War or before that? The Civil War broke out in 1936.

Anderson:

The Civil War was winding down at that time. My sympathies were unquestioned, but we had no contacts with the Spanish Civil War. I knew about it in the papers.

Kojevnikov:

And by the time you got to college?

Anderson:

In college of course the war had already started. We all knew that we were in a hurry to get out of college before we got drafted. The draft existed already and you had an exemption in those days in college.

Kojevnikov:

An exemption for whom?

Anderson:

For college students in general until later in the War, the trick was to graduate by the time I was nineteen because then I would have a guaranteed exemption from the draft until I finished college, which seemed very sensible. I did try to join up with the ROTC, Reserve Officers Training Core, classes. They wouldn't take anyone with glasses. Their sole criteria was no academic or other criteria, just did you wear glasses or didn't you wear glasses. I think at one time I tried to join the Air Force as well, the officer's training school for the Air Force, and again the criterion was do you wear glasses or don't you wear glasses. So I was lucky. In that case I would have ended up a navigator and would be likely to have been killed, as a matter of fact. But I didn't know it. So instead of that I was— Harvard was fairly normal for the year 1940. Although, of course, we were very conscious of what was going on. I even spent a normal summer. A very odd kind of treading water summer between the first two academic years. Then I went back and they had changed the physics curriculum. You could either stay in physics or you could take electronic physics, and there was a great deal of pressure to take so called electronic physics.

Kojevnikov:

This was considered to be war related?

Anderson:

Yeah, this was war related. Of course the nuclear stuff was entirely in secret. So no one was supposed to know what was going on. And I didn't. I was very young, very naive. There were people who did know what was going on because they had contacts in the world of physics. One of my friends, Henry Sillsbee said, "No, I think I'll stay in physics." And he went off to Los Alamos. Another acquaintance, Ted Hall, stayed in physics and for the obvious reason that he was determined to get to Los Alamos and spy.

Kojevnikov:

Was he in the same class?

Anderson:

He was in the same class. Yes, I knew him. But he was not a close friend. He didn't stay with the group, as a matter of fact. He didn't graduate with us. He went early.

Kojevnikov:

He went during summers before graduation.

Anderson:

Yes. But Henry went for Los Alamos the same time I left for Naval Research Lab. Anyhow, we were told, of course, to help the war effort you've got to work on radar. So we learned this course called electronic physics.

Kojevnikov:

You chose the electronic?

Anderson:

Chose electronic physics, and had all kinds of courses on linear oscillators and antennas and things like that.

Kojevnikov:

And who was teaching this? Was it regular faculty?

Anderson:

We had no engineering faculty. We had an engineering sciences faculty. There was a man named Chaffe who was a well known electron tube man. There was a horrible little man named R.W. P. King who taught the antenna courses. Then somebody else taught us a rather nice microwave course. I learned a lot really. It was even relevant, but I hated the laboratories because they were engineering rather than physics laboratories. And I took some physics courses.

Kojevnikov:

How about quantum mechanics?

Anderson:

No quantum mechanics for an undergraduate. I mean even if you stayed in physics, you got no quantum mechanics as an undergraduate. It was a forbidden subject and was just considered too esoteric for us. Well, when I came back and learned quantum mechanics in graduate school the text we used was Pauli's handbook article in German. There was no good text in English.

Kojevnikov:

Well, Kemble's book.

Anderson:

There was Kemble's book, but then it was considered that wasn't up to the German. So I learned group theory from Wigner in German and quantum mechanics from Pauli in German.

Kojevnikov:

So when did you get into the grad school?

Anderson:

Immediately after the war years. So I graduated at 19, went directly to the Naval Research Laboratory, worked there on antennas for two years and a half. The Naval Research Laboratory was very sensible and graduated us quickly.

Kojevnikov:

Did you graduate in 1943?

Anderson:

1943 and worked there until October, 1945.

Kojevnikov:

In what position?

Anderson:

As just a member of staff. Just a dogs body really. I failed when asked to wire circuits, so I built antennas. And so I was put in the antenna group. Lived on the top floor in a very hot building in Anacostia, D.C. in a wooden shack with our antennas arrayed around us out on the roof. Well, it was a very interesting experience, but irrelevant to my further career.

Kojevnikov:

Were there any other former students from Harvard hired?

Anderson:

Yes, there were several.

Kojevnikov:

So it was a group of you?

Anderson:

Yes, Bob Houston— well, there was an unusual group of us actually in 1943 that graduated together. There were an extraordinary number of summa cum laudes and all of them were in physics. Among others there was Tom Kuhn, who was a friend in college. There was a young man named Bob Houston who went to NRL with me. There was Henry Sillsbee, went off to Los Alamos. Piere Noyes, my roommate, to RadLab. I don't remember all the others. For some reason, although I had been very relaxed in my college work after freshman year, I had a record of about an equal number of B's and A+'s. No A's at all. Well, I did get an "A-" in history. So I either had very good grades or very mediocre grades. Although I wasn't the academic wonder that Tom Kuhn was, they gave me highest honors, a summa cum laude.

Kojevnikov:

Was Kuhn an academic wonder at that time?

Anderson:

Oh yeah. He was everything. He was the biggest man on campus.

Kojevnikov:

Was he also in the same class as yours?

Anderson:

Yeah, same class. He was editor of the newspaper, The Crimson. He was junior marshal Phi Beta, was appointed to Phi Beta as a junior, and so on and so on. He never got anything but an A. Whenever I was in the same class with him I saw to it that my grade was a point or two higher, but that was what I considered the most amount of work that I would do.

Kojevnikov:

At the Naval Research Lab, was it a physics division or something that you were working for?

Anderson:

It is an engineering division. It was a division that had to do with radar — rather it was above radar. Not above, but more secretive than radar. It was Identification Friend or Foe (IFF) and jamming and radar counter measures. That sounds very fancy, but mostly what we did was to test equipment that was sent to us by the big manufacturers. I was telling you a little bit about this job we had of building jamming transmitters for the invasion fleet to prevent their being bombed by glide bombs, of which we had gotten hold of one and studied it.

Kojevnikov:

What was your exact assignment? Was it more experimental or testing?

Anderson:

My assignment was heavy plumbing. That's what microwave and low frequency antennas, are is pipes put together, and I learned how to put pipes together very well. I designed one antenna I thought was a little gem as a matter of fact. I don't know whether it was ever used, but it was certainly accepted by the powers that be for one of the 11414 identification beacons.

Kojevnikov:

Did you consider this being a temporary job? Just waiting until the end of the war.

Anderson:

Absolutely, just waiting to get back to graduate school.

Kojevnikov:

Were you committed to go back to graduate school?

Anderson:

In the Naval Research Laboratory, I guess it was Admiral Strauss, later of AEC, who pulled strings and saw to it that all of us temporaries were immediately released even though we were in uniform. I was by this time a Chief Specialist X. Had the most ridiculous looking uniform you can imagine. The girlfriend of a friend of mine said, "Oh, great. You look just like a street sweeper." And I did.

Kojevnikov:

There must have been a problem to find girlfriends at Naval.

Anderson:

Oh no, there were plenty of girls. Washington was full of girls. Washington was great. Harvard was not full of girls. There were no cars in those days and you had to go twenty miles. Either you had to compete for the tiny population of Radcliff or you had to go twenty miles to Wellesley, without a car. So Harvard was very celibate, but Washington was great. I was growing up— I graduated from Harvard at 19 and I was a very immature 19. Washington was actually fun. We worked very hard six days a week. Very hot, very miserable, but met lots of people, several wonderful girls. And had a good time.

Kojevnikov:

Did all of you go back to grad school in 1945?

Anderson:

Yeah, everyone went back. No, not all. Some of them stayed. We were beginning to play with missiles, with rockets toward the end of the War. We were designing antennas to avoid the ionized rocket exhaust. So we were trying to figure out how to get reliable communications to these things. And several of my friends stayed and were part of the very early development of the rocket age. I had a young Texan friend named Lagow who stayed in that group. I think who went up quite high in it. Many of the older people stayed.

Kojevnikov:

Did you know that you would go back to Harvard for grad school?

Anderson:

No, I chose— well, what I did was when I knew I was going to be released I went back to Harvard and went and saw Van Vleck, who was at that point chairman of the department. Said, "Can I come back?" And he said, "Why do you want to come back?" I said, "Because I didn't really study physics here as an undergraduate. I took all those strange courses." He decided that was a good idea. I had met him again during the War at one little meeting about propagation of radio waves. He was instrumental. Well, he wasn't instrumental. He failed to stop the military from using the so called K-Band, which he knew perfectly well was in the middle of a spectral line of water. He told them it won't propagate and they built the whole system and then it didn't propagate. I was at the meeting where he was trying to do this. He remembered me, and he seemed to want me to go back. So I didn't apply anywhere else. I just went back to Harvard.

Kojevnikov:

This was '45?

Anderson:

'45.

Kojevnikov:

In the fall?

Anderson:

Yes. Harvard was a wonderful place at that point. Again, I had grown up. I was no longer the wet behind the ears kid that had gone to Harvard. So I had a wonderful time in graduate school. Both in terms of learning and in terms of a social life and learning various other things. In terms of learning, Furry taught quantum mechanics that first year, but Schwinger came in the middle of the year and started a peculiar three semester course, which was essentially everything Schwinger knows. All about Green's functions, all about nuclear physics and so on. All the numerical tricks he had devised to solve quantum mechanical problems, nuclear physics problems. So we learned all about the Deuteron. But then also he was starting to build the machinery that was going to solve the quantum electrodynamics problems. We were treated to a lot of that machinery. And all of his Green's function works on wave guides. So for three semesters he just lectured an hour to an hour and a half twice a week, and we took scribbled notes as fast as we could. I was taking this simultaneously with quantum mechanics, so you can imagine it was fairly rapid, a fairly tough job to absorb that. I had some bad courses, but mostly they were marvelous courses, marvelously taught.

Kojevnikov:

Did Van Vleck teach?

Anderson:

Van taught us group theory, which was a weird experience because he's one of the most disorderly teachers you can imagine. But his mind was orderly even though his teaching style is disorderly. So you ended up knowing group theory, even though at any given time you wouldn't have the faintest idea what he was on about and on the board. But I had Wigner, and I puzzled it out between Van Vleck and Wigner, the German version of Wigner, which I still have heavily written in. We had wonderful summer courses. The first year they brought in Brillouin, and he taught us about periodic lattices, periodic systems.

Kojevnikov:

Was it the first solid state course that you had?

Anderson:

That's the only solid state course I ever had.

Kojevnikov:

Or was there something in the quantum mechanics course?

Anderson:

No, there was nothing about it in the quantum mechanics course. Van Vleck taught a course in solid state and I only audited. I didn't actually take it.

Kojevnikov:

How was his English, Brillouin?

Anderson:

Brillouin, oh it was fine. Very precise. Then we had Goudsmit . I think that was also the first year. Goudsmit came and taught us dynamics, or a course in dynamics. Absolutely inimitable style. It was wonderful. He had this pendulum with a bug and a wheel and a fly wheel, and the bug was crawling in a certain direction across the fly wheel and so on. All this in this wonderful Dutch accent. It was delightful, and we did learn a lot.

Kojevnikov:

How many grad students were in your class?

Anderson:

It must have been a big class. It must have been twenty or thirty. Three of them, we later on learned, were there because Van Vleck had specifically reached out and picked us by the scruff of the neck and said you belong in Harvard. There's no entrance exam or deliberate, careful deliberation, but it was a wonderful class. Extraordinary. Because of course there was all these people who came from everywhere, And thank God we were back in civilian life and back in graduate school.

Kojevnikov:

Who was in your class?

Anderson:

Tom Kuhn, Ken Watson, Roy Glauber, Walter Kohn, Ken Case, Fred De Hoffman, a lot of them. Most of them you've heard of.

Kojevnikov:

What was the time when you would have to think about choosing some narrow specialty? Between the theoretical and experimental physics or within different fields of physics?

Anderson:

Well I had to be a theoretical physicist because I had this experience during the War and I knew that I was only good at large pieces of plumbing. I knew I was not good at putting things together with my fingers.

Kojevnikov:

When did you have to make an official decision?

Anderson:

You had to take I guess an oral exam to be admitted to do a thesis. And Van said I only barely passed it, because although I'd had these marvelous courses on dynamics, I'd never actually had the dynamics of a rigid rotating body. The Euler equations. I didn't know the Euearian angles from a hole in the wall. That was the oral exam. Van essentially permitted me to pass, but he said, "You're only tentatively a theorist." I was confronted with something I didn't know anything about because of course our education had been very spotty. For instance, there was this fantastic course in complex analysis. It was taught by D.V. Witter of the math department. He managed to get us— well, he was bored with contour integrals so we did the whole of complex analysis without a single contour integral, which of course is totally disabling for a physicist because a physicist uses nothing out of complex analysis except contour integrals. I had that kind of education. I suppose everyone does, but it was particularly spotty up to this point. I chose theory because I had to. I chose condensed matter because basically I'm a contrarian. Everyone was eager to enter the popular field of nuclear physics, so I wanted to do something else.

Kojevnikov:

Like?

Anderson:

I thought it was crowded. I had seen a little bit of— I really was pretty naive at this point, but I had seen a little bit of the kind of work that was going on. A little of what Purcell was doing. Purcell and Pound. One of my fellow graduate students, Al Sachs actually. There's another name. He became a famous particle physicist at Nevons Lab at Columbia. He was working with Purcell and told me a little bit about what he was doing. I knew the spectroscopy work that Van Vleck and other people had been doing during the War. Furry actually and Pippard I met at that meeting during the War. That looked interesting to me and I just kind of gravitated toward it. I'm very curious. I was interested in how things worked. And it didn't seem to me the nucleus had much to say about how things work.

Kojevnikov:

Was this the time when the bomb was discussed everywhere and was on everyone's mind?

Anderson:

Once you learned about the bomb it was obvious how it worked. The principles of the bomb were nothing exciting. Maybe if you eventually find some new principles in nuclear physics. I really wanted to know how light propagated through gases.

Kojevnikov:

How about quantum electro-dynamics?

Anderson:

It was too hard for me. I'm not a formalist. I did listen to some of those lectures. I found the formalism forbidding. I only later came to understand that kind of formalism. I'm lazy. I'm mentally lazy. I believe in using only the tools that are necessary for the job. If I've got a serious problem and it's my problem that I have to solve, I'll go on inventing formalism until I find the answer. But in general I avoid the formalism if I possibly can.

Kojevnikov:

Among other physicists, who do you think is the closest to you in style?

Anderson:

Well some of them are experimentalists. I've always thought that Nicolaas Bloembergen and I were very similar. He's half a theorist, or say six tenths experimentalist and four tenths theorist, and I'm more like six tenths theorist and four tenths experimentalist. So that's fairly close. This is immodest of me, and I say this with all modesty, but I think Fermi worked the way I worked. He really was focused on the experimental question. And then he would do formalism if he had to solve it. And Van very much so. Van was no formalist. The reason I was rescued from complete ignorance of formalism was that I had these courses from Schwinger, and so I learned the Green's function formulas and some of that aspect of physics. But that was another thing about the nuclear physics. The Schwinger group of students and the few students who were working with Furry, they were focused on formalism. And I wanted to really explain experimental facts. So when Van gave me this problem—

Kojevnikov:

Was it his initiative or was it your initiative?

Anderson:

The problem was his initiative.

Kojevnikov:

No, just the choice.

Anderson:

My choice was I went to him.

Kojevnikov:

But that field at that time wasn't yet called condensed matter physics.

Anderson:

No, it was called chemical physics. The Journal of Chemical Physics had just been founded, and much of solid state physics was appearing in the JCP . But it would have been perfectly reasonable for me to put my work into the Journal of Chemical Physics.

Kojevnikov:

When you went to Van, was it before the oral exam?

Anderson:

After the oral exam. I said, "I would like a problem." And he gave me this problem, which although I didn't get very far with it immediately, it certainly fit very well into my own personal—

Kojevnikov:

What was it?

Anderson:

The new field of microwave spectroscopy had just been invented as a consequence of the war. And all over the world: there was Bleaney in Oxford and there was Charlie Townes of Columbia, actually at Bell Labs then. And Walter Gordy in North Carolina. They were running through all the gases with microwave spectroscopy. With optical spectroscopy you didn't have the resolution to really see the spectrum in detail. You look at the ammonia spectrum with microwave spectroscopy, you see a hundred lines. Every one of those lines has a different width. Van said, "Look, here's an enormous range of phenomena that have never been studied before at all. Let's find out what explains the widths. What explains the magnitude. What explains the variation from one line to another.

Kojevnikov:

Was it specifically ammonia, or was it just any gas, any kind of substance?

Anderson:

You could do any gas you like.

Kojevnikov:

But somehow ammonia was preferred at that time?

Anderson:

Ammonia is a marvelous gas for microwave spectroscopy because it's spectrum comes right at three centimeters. I don't remember. Anyhow, it's the preferred wavelength for War time radar and it has this wonderful inversion that gives you an inversion, but it also rotates and it vibrates. But particularly it rotates. So you have a lot of different rotational plus inversion lines. So it has a remarkably detailed spectrum.

Kojevnikov:

How much of this problem was residually connected to the military orientation of research?

Anderson:

None at all. Ammonia doesn't occur in the atmosphere. Yes, I mean the reason why— or not the reason why, but it had become an important issue during wartime because of these propagation problems. If the wavelength was too short it ran into spectro absorption lines in the atmosphere, in the water vapor in the atmosphere, and trace gases in the atmosphere for really long propagation. Or actually, there's an oxygen line around as well. So for really long distance propagation you needed to worry about the trace gases and so on. But no, essentially people had taken the spectrograms and the technology, like spectrographs and the techniques that they'd been using to make microwave measurements for engineering and just bought yourself a roomful of War surplus equipment.

Kojevnikov:

Did you have any of this microwave equipment at that time? Because most of this will come from the wartime.

Anderson:

Sure. Harvard had the radio research lab, MIT had a radiation lab.

Kojevnikov:

Did you have access to it?

Anderson:

No, I didn't.

Kojevnikov:

Were you and Van just working at that time without any connection with actual experiments?

Anderson:

Well the point was Van was the only person who knew this field. I went to one of the early American Physical Society meetings with him. I didn't get a chance to talk to him. I would say a few words and somebody else would come out of the crowd and say, "Van, I've been doing so and so and I've measured so and so," and he would show him the data. And Van would say, "Oh I suppose that's an asymmetric rotor molecule, such and such molecule. And you're seeing the lambda type doubling," and so on. So he was the communication nexus for all the data at that point in that field plus others. One other thing I should have said about graduate school. The third person who came was Cornelis Gorter who was head of the Chemo Kammerlingh Onnes Laboratory in Leyden, who had spent wartime trying to continue doing work. They had been able to do some low radio frequency work. If he hadn't been stopped by the War he probably would have discovered nuclear magnetic resonance. But they had all the preliminary ideas that people then applied to the new NMR and electronic spectroscopies. He taught a summer course about that.

Kojevnikov:

On what?

Anderson:

Really on radio frequency spectroscopy as they had done it in Leyden during the War. That intrigued me and influenced my choice because I liked that stuff. There were real problems here, but they were soluble type problems. They didn't overwhelm you with formalism.

Kojevnikov:

How many graduate students did Van have at that time?

Anderson:

I was one of two or three. He had Tom Kuhn.

Kojevnikov:

Oh, was Kuhn his student?

Anderson:

Yeah, Kuhn was his student. He did a very respectable thesis. And Arianna Wright, later became Mrs. Rosenbluth. And myself. So there were only three.

Kojevnikov:

And what was Van's style in dealing with grad students?

Anderson:

Just leave him alone. Give them a problem and let them do it. He was very much not— he didn't use you to work out his idea. He gave me three references. There was a man named Lindholm who'd done the only really serious effort to look at separate spectral lines. He gave me his thesis from some forest products laboratory in Sweden. There was an old paper by Weisskopf and I think that was about it. I don't remember. There may have been a third reference. So he said, "Look at these things and see what you can do."

Kojevnikov:

Was he approachable when you had to find him?

Anderson:

Sure, if he was around.

Kojevnikov:

Or would he approach you and kind of ask what's going on?

Anderson:

After I'd been kind of – Well, in the first place I was having a very good time. For example, Tom Leher had graduated from Harvard in the wartime generation and he had just entered graduate school at the same time as I did. Another one who had done that was David Robinson who's a long time executive of the Carnegie Foundation and a very close friend. I fell in with the group around Tom and Dave Robinson and a number of other people from a variety of departments. So I wasn't socialized with the physics department so much. I was socialized with a more general group. We played bridge a lot. I learned to play bridge from Dave Robinson, who is still a professional level bridge player. We sang Tom Lehrer's songs a great deal. We did puzzles together. Wrote doggerel. Another one in the group was Chan Davis who was notorious for being persecuted by the House Un-American Activities Committee and later became a fairly eminent mathematician. He and Tom Lehrer would write doggerel on the board for each other's math classes. But this was a very interesting life.

Kojevnikov:

Was it in any case a political active group?

Anderson:

No, we weren't politically active at all. Chan and I, for instance, were interested in science fiction, and I met some of the Astounding Science Fiction authors through Chan. And so on. So we had a very interesting, very nice life. But my thesis wasn't getting on very fast. So finally at one point Van called me in and said, "Where are you getting?" and I said, "Well, I've studied interactions a lot." He said write a little piece about it, and I did. And he seemed satisfied and I went back to work. That continued until the summer of 1947. Summer of 1947, I went home and I met a girl and we got married within a couple of months. And we had a child on the way within about four months. So I began to take my thesis more seriously starting summer of 1947. In the fall of 1747, I believe Van Vleck was away in any case so the things I was doing to solve the problem were essentially all my own ideas. By the summer of 1948 I had gotten the problem solved. I had gotten the basic ideas of how to solve this problem and really started writing the thesis. The child came in early 1948, spring of 1948. I remember writing my thesis half the time and half the time wheeling the child around the town of Belmont in her baby carriage. But that marvelously concentrates the mind, having a wife and child. I mean after all I had the GI bill. They gave me my national scholarship back so I had a little extra money. But we were spending roughly sixty percent of our entire income on rent, and we couldn't go on that way very long. So I finished in February 1949. But this bit about the social life I think has a lot to do with my choice of topic because I wasn't in this rather tight-knit group of physicists. I saw them as focused on Schwinger, but I do not think that's really fair. They were focused on their vision of what the right way to do physics was. And I really felt like a contrarian socially in the sense of I wanted to be broader than that, and a contrarian in the science in that I didn't really want to do this formalism along with a lot of other bright people who are probably just as bright as I was and would get to the answer at roughly the same time. And I didn't want to compete.

Kojevnikov:

And who were among those people? Of your classmates?

Anderson:

I should be able to come up with the names. Bryce Dewitt was one. Freddy De Hoffman was one. Glauber was one. Several did an excellent thesis. An important thesis. Lippman and Schwinger, the famous Lippman who likes to tell the jokes that at one time he managed to get Schwinger to say, "I'm the Schwinger of Lippman and Schwinger." It was that kind of thing. You were X and Schwinger. I didn't want to be X and Schwinger. Although, Julian is wonderful and great, and I learned a great deal from him. The methods I used in my thesis were methods I borrowed from Schwinger's course. They were called Shwingerian methods. They weren't as sophisticated as the QED, but they were very sophisticated in terms of really using the full operators of tensor algebra and so on. And using representation group theory in quite a sophisticated way. And so on.

Kojevnikov:

Did Kuhn finish his thesis?

Anderson:

About the same time as I did. He was made a Junior Fellow (a special honor) as a post-doc because he had been this incredible star all the time. As a Junior Fellow he was enticed into teaching the general education course and he was gone. He became interested in history and sociology of science. He had done this very good thesis with Van Vleck, it was a great exercise in how to do band theory, but it was just an improvement on the work Wigner's students had done. An improvement with a little tinkering. And he must have been bored a little bit by this thesis, although it was very brilliant and very intelligent. Compared to what I was doing I know it was boring, but also compared to what the Schwinger students were doing it was probably boring. And he found that this new world he opened into was more exciting. And I can understand that. Anyhow I solved this problem.

Kojevnikov:

Was it published?

Anderson:

Oh, yes. It's the basis for at least two review articles. In essence it's the last substantive advance in that field. And it was a substantive advance. There are kinds of physicists that I meet who say, "Oh you're the Anderson who did the pressure broadening thesis."

Kojevnikov:

The pressure broadening the microwave and infrared region?

Anderson:

Yeah. There are plenty of people who've heard about it and not the other work I've done. It's gone off into chemical physics.

Kojevnikov:

Yeah, because that field was closed pretty soon.

Anderson:

Well it's not so much closed as— There had been some further developments, but kind of the approach is— I really clobbered that problem and it was very useful. It was very useful, for instance, when Ali Javan, incidentally, was inventing the gas laser. He was at all times using this kind of sophisticated collision theory and relaxation theory for the molecules and using my methods for solving the scattering problems between molecules.

Kojevnikov:

Was there anybody else among faculty besides Van with whom you would contact or discuss things at work?

Anderson:

Yes. There was a group doing this kind of work under E. Bright Wilson, Jr., who was the father of Ken Wilson. In the chemical physics department. And he had a whole group of students who were doing this kind of work. So I could talk it over with him or with them. But the formalism very quickly got beyond the chemical physics level. Essentially— well, a couple of things. One was I invented some tricks as far as doing intermolecular collisions for understanding the quantum mechanics of two molecules colliding more or less classically and their internal states changing. So I found ways of expressing that. And then I found ways of putting that problem into the— using if you like the S-matrix. The time development matrices and S-matrices and so on. Not the ultimately sophisticated formulation of quantum scattering theory, but really using quantum scattering theory for molecular collisions. And then finding out how to average all the results of the scattering. And the other thing about it is it's perhaps the first real practical use of fluctuation-dissipation techniques. In other words, I calculated the correlation function of the moment of the entire gas rather than some of the moments of all of the molecules. And then from that calculated the absorption spectrum using the fluctuation dissipation theory, which hadn't been named yet. It hadn't even been written down. Callen and Welton were yet to write it down, but essentially you can do it if you just know Einstein's radiation coefficients. Callen and Welton did not do anything new that Einstein didn't know. I didn't do anything new that Einstein didn't know, but I used it in this practical way that amounted to the fluctuation dissipation theory. So I finished that. Then I started looking for a job. It was a terrible time to look for a job. Late 1948, early 1949, the wartime economy was deflating. The Korean War hadn't started. The Cold War hadn't started.

Kojevnikov:

The Cold War was already going on in a sense.

Anderson:

The people of the U.S. didn't know that the Cold War was going on. Certainly U.S. industry didn't know. At Bell Labs some people knew that something had happened, but most people didn't realize the transistor had happened. So I got two offers. Just from recruiters coming through the department. One of them was at Pullman State College in Washington and the other was from Westinghouse Research Labs. I had gone on a tour of various research labs, Westinghouse, GE, and Brookhaven. My mother, having lived on a professor's salary all her life, was determined that I should work at a research lab. And I thought, "Yes, if I had a wife to support maybe I'd better." So I didn't look for post-docs. So I got this offer from Westinghouse. I went to Westinghouse and talked to Ted Holstein. He's a marvelous physicist and understood exactly what I'd done and tried hard to get me a job in his group. Didn't succeed. I was supposed to sit in a large room with a box full of transistors that had been sent by Bell Labs and figure out how the hell the transistor works. And I wasn't going to do that. So I decided to go to the Pullman State College in the state of Washington, driving across the country and over the mountains in the dead of winter. So my parents bought us a car. And then Van Vleck, about three weeks before I left said, "Where do you want to go?" I said, "I want to go to Bell Labs." Bell Labs I'd gone to before I'd really finished solving my problem in late 1947, not in late 1948. And I gave a talk and they thought it was a very intelligent talk, but I obviously hadn't done anything so there was no job.

Kojevnikov:

What did you know about Bell Labs at that time?

Anderson:

Several things. One was I did know about the transistor. I did not know the names of the people there. Bardeen, Wannier, some others. Third, I had met Bell Labs people and Bell Labs equipment during the War and I'd been immensely impressed by the difference between Bell Labs and a typical supplier like Philco. We used Bell Labs tubes all the time. Doorknob tubes and Bell Labs magnetrons and so on. I knew that they were just more confident. So I had this great respect for their confidence. I had been to Brookhaven, and Sam Goudsmit later said he made a very bad mistake. He was in charge of the group there and he asked me if I going to carry on with my thesis work, and I said, "No, I solved that problem." He didn't believe me. He thought "if he's got something interesting going here he should want to continue it." He didn't realize how really a thorough job I'd done on my thesis, that I needed something else to work on and I wanted something else to work on. So he guessed wrong and guessed that I was just unimaginative. Later on he admitted that he was wrong. So I said I wanted to go to Bell Labs. Van Vleck, who loved trains, took the train to Bell Laboratories and went out to Murray Hill and talked to Shockley and Fisk and said you better hire this man. I don't think he had realized how good my thesis was actually until he had my thesis reviewed. At that time he must have been proud because he invited Weisskopf and Schwinger as the other two members of the committee. But I didn't realize what that meant. So he talked Bell Labs into hiring me. I had a long chat with Shockley. And Shockley was the person who hired me.

Kojevnikov:

And Shockley was in what position there?

Anderson:

He was a department head.

Kojevnikov:

Of the research department?

Anderson:

No, a piece of the research department. The organization then was in fact that a department was what is now a division. The present department then was called a sub department, but it was actually quite a big group. But no he was never high in administration at Bell Labs. That was the problem. Bell Labs knew better. Anyhow, Shockley said, "Well we'll take you on for a year." And I said, "It better be permanent." And he looked at me as though I was crazy and said, "Well, permanent if you like." And of course he didn't say, but it is true that the contracts at Bell Labs are annual. So he was not making any bets. In any case, so they hired me. At the last minute I gave up Washington State College and went to Bell Labs.

Kojevnikov:

What was the difference in salary at those times between the post-doc and the starting positions?

Anderson:

They were essentially identical. All of them were identical. Maybe a post-doc would have been a little less. A post-doc might have been four thousand, a permanent job or an assistant professorship, four hundred and fifty a month.

Kojevnikov:

No, I meant he difference between the academic and the industrial research.

Anderson:

None. I had the same offer from Westinghouse, from Pullman State College, and from Bell. Identical to the dollar.

Kojevnikov:

So why then would your mother be—

Anderson:

Well she didn't know the modern world had arrived and that scientific research was considered to be economically useful. So I went to Bell Labs. I later learned that there were two extra hires, there were two over what's called "nose count." Because at the last minute they'd also been able to attract Berndt Matthias, and they had hired already Jack Galt, and one other person which I forget. All of these were not hirings they regretted later, but they were over the limit. In principle they were expecting to fire two out of four of us. And in fact one year later I heard, (but not until much, much later), I heard that Shockley, for one, had strongly advocated firing me. But I didn't know it, fortunately, at the time. But Bell Labs personnel procedure at that time was much more secretive than it is now, so I didn't hear what went on in the management circles.

Kojevnikov:

Were you made part of the theory group?

Anderson:

There was no theory group. Well, the organization of Bell at that time was there was one thing that was very good about it which was that it didn't really have a physics department, a chemistry department, an engineering department. Every group contained some fraction of people from each of the relevant fields. At that time I guess the two main divisions of physics research were the solid state physics research and what was called physical electronics, which was what was left over from all their work on vacuum tubes that was going to wind down, although it wasn't obvious then. Each of these groups contained theorists and chemists and preparatory chemists and analytic chemists and various kinds of experimentalists.

Kojevnikov:

Which group were you in?

Anderson:

I was in solid state physics.

Kojevnikov:

How big was it?

Anderson:

Oh it must have been fifteen or twenty. There was already developing a separate— No, actually it didn't yet. There was no separate semiconductor research at that time. And Bardeen was still in the group. Charlie Kittel was still in the group. Conyers Herring was still in the physical electronics group. Wannier was in the physical electronics group. Wannier was the fourth hire. We came at the same time. It was a remarkable group of people. Wannier and I, being excess, sat in a room together and realized I really had no solid state physics. So the first thing I had to do was learn solid state physics. I learned solid state physics from Wannier; crystallography from Elizabeth Wood. Betty Wood was the crystallographer in one of these groups. A lot of solid state physics from Shockley, also he was very willing—

Kojevnikov:

Was he the head of (???)

Anderson:

He was the co-head. It was already known that he was a difficult person to get along with so there was a very, very mollifying, soft spoken, easy-going guy named Stanley Morgan who was the other co-head. And after people had been riled up by Shockley they would be calmed down by Morgan. It was a good guy–bad guy arrangement a good cop –bad cop arrangement and it worked very well.

Kojevnikov:

Were you assigned any topic right away?

Anderson:

Shockley had the idea that I would be his post-doc and that I would work on ferro electrics.

Kojevnikov:

Did Bell Labs have post-docs at that time?

Anderson:

No, they had none. He was going to treat me as a post-doc. That was his proposal and I said, "No, I don't want to be a post-doc. Industrial labs don't have post-docs. I want to be a regular employee." That's what he said yes to. He still was going to treat me as his post-doc. Now, as his post-doc I was given a lot of help and hospitable treatment. Joyce and I were very poor and we had just gotten this new car. We had a baby. So when we showed up at the door of the labs he said, "Well, I don't know where you're going to stay tonight. Why don't you come stay with us?" And we did stay with them a couple of days. They were very hospitable. He had a wonderful wife who he later got rid of, named Jean. She was very nice and he was very helpful with teaching me solid state. I learned a lot from him. Really a lot. Then on the other hand, he set me to work on ferro electrics and I was supposed to do what I hadn't done for Van—to take his ideas and calculate and use them on ferro-electrics. I played with them a little bit and I understood them and I liked them in principle, but I said, "Look, the science of chemistry and solid state physics isn't at a stage where this is a quantitated calculation. It's just too far from reality. I won't calculate with it. I will say quantitatively this is the explanation of ferro electricity. Yes, you're right, this is the explanation of ferro electricity. But no you can't calculate with it and demonstrate that." The Lorentz force that he was going to use was essentially a Lozentz catastrophe. I looked around and I thought about water and I thought about various things and I decided the Lorentz catastrophe was rather a mysterious question. Ionic crystals, like barium titinate, that was what I was supposed to work on. He had this idea that barium titinate was an important material for the future. That's why they hired Berndt Matthias because he could grow barium titinate crystals.

Kojevnikov:

Where did he come from?

Anderson:

He came from MIT where he'd been for a year, but before that he'd been at the ETH in Switzerland. He was a wheeler and a dealer and he came with these new kinds of crystal in his pocket.

Kojevnikov:

And why barium titinate was sought as promising?

Anderson:

It was a ferro electric. That is, it spontaneously develops electric polarization. It was very useful for a number of things. And Shockley and some of the other people in the administration thought that if it wasn't useful for a memory device it might be useful for non-linear electricity and things like that. It's in fact still— in fact I think recently it's been revived as a possible form of electronic memory, but it was very premature to be thinking of it in those days. Nonetheless, there was a lot of physics to be done and the question was why is it ferro electric? And I was supposed to take Shockley's ideas and put them down on paper and do appropriate calculations. And I did them and they didn't come out anything like reality. But the principle was fine. But he was very annoyed that I didn't— John Slater wrote down at the same time identical ideas in a paper which has been fortunately forgotten because it didn't make any sense. He was over optimistic. I discovered that even the alkali halides weren't even perfect ionic crystals. His idea was ions just move as a whole and the sodium ion carries charge plus e and the chlorine ion carries charge minus e and they move relative to each other and that leads to an electric dipole movement. I studied the alkali halides and I discovered that wasn't the case. The amount of charge that moved was much smaller. I had an idea for calculation of that effect and used it to explain the frequencies of alkali halide vibrations in ionic crystals. Which I looked at recently and I realized it was right, but that the in the wave functions I was using then, they were too primitive, weren't right. But the paper's right. It actually explains the effect that it's supposed to explain. Although I never published it because the numbers didn't come out right. I did a little statistics on Ferro-electrics. I think that was what saved me and I did at least give one meeting talk on ferroelectricity. I did one other paper on ferroelectricity, a review paper in which I talked about Shockley's theory qualitatively and showed that the experimental data supported it, but a different kind of qualitative experimental data. So I thought I had done my assignment reasonably well, but Shockley didn't and he tried to fire me. In the meantime, actually I happened to be looking at my notebook and realized this happened very quickly. Charles Kittel was at Bell Labs at that time. It's hard to realize how many of the great figures of condensed matter physics were there — the people who were later to become the great figures were in Bell Labs at that time. Bardeen was there. Kittel was there. Herring was there. Wannier had just arrived. Suhl was in a separate group working with John Pierce in an engineering department. I must be missing some.

Kojevnikov:

Pines came later?

Kojevnikov:

Pines came later.

Anderson:

No, Pines was never at Bell Labs. That's about it, but we also had consultants. Peter Debye came and consulted once a month and we reported what we were doing to him. Slater had consulted with them for the previous two years. I guess that's really it, but you know there weren't that many condensed matter physicists. That was about half of the condensed matter theorists in the country. Certainly in terms of quality it was over half of the condensed matter theorists in the country. The only comparable groups were in Europe. It was an incredibly powerful group. And everything that happened remotely having to do with condensed matter was passed through. So one of the things that passed through was Cliff Shull with the first measurements of anti-ferromagnetism using neutron defraction.. Charlie Kittel showed up at the door of the office that Wannier and I were sharing and said, "Shouldn't we think about anti-ferromagnetism line missed (see edited man.) in that he redid the Onsanger solution of the Ising model for the triangular anti-ferro magnet which was the first exact derivation of a zero point entropy. And I played with that a lot, but then I got to thinking about things like the experimental data about anti-ferromagnetism. And Neel's observations and other people's observations. And I realized that a couple of things were true. One was that there was a big difference between the local field as measured by the susceptibility and the local field as measured by the face transition temperature, and that I could explain that by some structure of the exchange interactions. And this was really a trivial thing, but it's a nice, simple, trivial thing that nobody had thought of and it really worked a lot of things. For one thing it told you how to think about the exchange interactions in these new materials. They were not irrelevant new materials because at the same time we in the department were working on ferro magnets, on pharites and later on garnets, both of which are anti-Ferro magnets with uncompensated spins. Those are practically very important. So then Charlie again suggested a couple of references, and I looked around and we discovered old papers by Kramers on what he called super exchange or long range exchange interactions. And that was one of the first papers I wrote. The first thing was the "Generalizations of the Weiss Molecular Field Theory," and the next was "Anti-Ferro Magnetic Theory of Super Exchange Interaction."

Kojevnikov:

Now it seems if it was completely your and Kittel's choice to do this?

Anderson:

That's a good point. It wasn't, but I didn't know it wasn't. Fortunately there were already great strains going on in the physics department, strains between Shockley, who wanted us all to work on semiconductors—his ideas on semiconductors and a side interest in Ferro electrics—and Bardeen and Kittel, who had their own ideas on semiconductors and were also interested in broader ranges of kinds of physics, and thought we should work on magnetism and this and that. So Kittel was not in the management. He was not privy to any of the decisions. Nonetheless, he talked Stan Morgan into defending my choice to do this anti-Ferro magnetism work.

Kojevnikov:

So how did you have to justify or how did you have to defend your choice?

Anderson:

I didn't. It was defended for me. We had a thing called case write-ups. Once a year we had to write up a summary of the work we had done. Later on these became very elaborate documents, but in those days you just did exactly that—you summarized the work you had done for the year. This went into a big book that disappeared. It was a write only memory—nobody read it. It was a write only memory. And of course there were the annual merit reviews, which was a serious business where committees of the management got together and decided on changes in salary. At the annual merit review I gathered Shockley thought I should be fired because I had taken the freedom to work on Charlie's stuff rather than his stuff. I'd thought I'd done very well with his stuff, but he didn't. He was humiliated by the fact that he'd taken me out to lunch with Slater, and he and Slater had argued over these ideas and he said after that lunch, "Well you ought to get to work and get something published because Slater's obviously going to publish." And I said— I didn't say. I just didn't because I didn't think it was worth publishing. So it was just stupid. I just didn't know. I was too insensitive to know what danger I was in. But apparently I was in great danger because I had not done what Shockley intended, and I was working on other people's work. And I just damn well did. The other thing which happened was that Van was interested in this work on anti-Ferro magnetism that I'd done. And so he or someone saw to it that in the March meeting 1951, I think...no, it must have been 1950. Whoever is really interested can look it up. It was a meeting of APS in Oak Ridge, the solid state meeting. I gave an invited paper. And once you give an invited paper at that age you're probably not touchable at the Bell Laboratories. I mean, invited papers are the currency with which we measure our success. And we didn't get that many invited papers at the major meetings of the Physical Society in those days. We did better later, but that was a feather in the Lab's cap. Then Van came to Oak Ridge and sat with me and really absorbed what I had done and he used it as a basis for one of his major talks at the International Low Temperature meeting in Leyden that year. So after that Shockley couldn't touch me. That's the Bell Lab system, or became the Bell Lab system. You can do what you like, but it better succeed. If you do what you like and it doesn't succeed, you're out.

Kojevnikov:

But in general, how would you describe the difference between the industrial setting and the academic setting at that time? So your personal experience.

Anderson:

At that time, the academic setting, the industrial setting— both were changing very rapidly. Let's put it that way. Industry in general was still the big room of eight desks in Westinghouse and everybody trying to figure out how the transistor works, and nobody doing his own research. Or it was GE where there were a few people doing their own research, but they weren't really the high quality of Bell Labs. Then Bell Labs was fairly unique. IBM was to become an imitator. Various things were to become imitators. But Bell Labs was unique. With the attitude at least at this point that you had a lot of freedom, they were just beginning to test the limits of that freedom, and at this point I really couldn't have taken off for somewhere else, and I certainly wasn't free to travel at will.

Kojevnikov:

So what types of limits?

Anderson:

The limit was— it became. I think I set the limit by doing it this way and succeeding. I set the limit that if you succeeded whatever you did was all right, so long as it wasn't obviously totally irrelevant and so long as you did some fraction of your work consulting for people who were being useful to Bell Labs, which I did at all times. I was talking at the same time as I did this feral electrics work. I was talking to the relevant experimentalist. Particularly to— Well there was a third experimentalist, a third person that was hired. Walter Merz from Switzerland came to do experiments on feral electrics. And all the time I was working on barium titinate I was also looking over the shoulder of Walter Merz and telling him what the experiments that he was doing meant. And that was the third place where I published my ideas on Ferro electrics, which was simply as an appendix in Walter Merz's paper. He didn't offer me co-authorship. He should have. By modern standards he absolutely had to. But I was naive then and I thought it was enough just to be of service to an experimentalist. Some of the most interesting parts of my work appeared in this experimental paper. So I was consulting with people all the time. And there was a magnetism group. Jack Galt was doing experiments on single crystals of ferrites. I was consulting with him at all times and with the magnetic resonance group. So I was keeping up my consulting work. Informally, strictly— it's all informal, but you do talk with people and it is noticed if you are being useful or if not.

Kojevnikov:

What other things that were (???) in the lab? Were there seminars, for example?

Anderson:

Oh yes, lots of seminars.

Kojevnikov:

So how did the change in the system happen?

Anderson:

I guess Gregory Wannier was one of the younger Turks that I told you about.

 

Kojevnikov:

Was he older than you?

Anderson:

He was quite a bit older, but he'd been at True Industrial Laboratory. An oil company lab in south Jersey. He was found there and brought to Bell Labs and worked on gaseous electronics. He was interested in things about collision theory. We did a lot of thinking about Wannier functions and the Wannier representation, which was a very important part of my education in solid state physics. An important thing that I learned was reverence for Lars Onsager, which Gregory Wannier had a great deal of.

Kojevnikov:

So you said he was one of the young Turks. What was the goal?

Anderson:

He was one of the younger Turks in that he pushed the envelope a little bit too. In just personal habits. They were strange. One time he decided that he belonged on a twenty-three hour day. So he came to the lab on a twenty-three hour clock instead of a twenty-four hour clock.

Kojevnikov:

What does this mean?

Anderson:

It means that he rotated an hour in his working hours every day. So he was arriving at the labs at midnight or later. Things like that.

Kojevnikov:

Were you supposed to be at the lab for a certain period of time?

Anderson:

In principle, but in practice, not at all. There was an opening hour and a closing hour. But in practice, especially the experimentalists didn't pay really much attention to it. I did because I like regular hours. On the other hand, in those days, especially after I began to feel quite stable and safe, I tended to work at home for one day a week. Let's see, I guess after '53, I think, I began doing that very regularly. So many of my discoveries were made while working at home.

Kojevnikov:

Did many people do this?

Anderson:

Yeah, a fair number of people would casually take time off.

Anderson:

Now, what was happening in academia was the grant system was just developing. It was very easy to get grants in those days. They were usually military. We just depended on the appropriate person in the appropriate office saying yes or no. There was very little peer review. NSF had not really become formalized. So many of our academic friends were making a lot of money consulting with us and with other people in the summers, had their own grants, so they could take off on military air transport planes and go anywhere at any time. So the summer school business, the summer school gang, Les Houches and Vannevar and Erice and all these places, the places that Jeremy Bernstein referred to as the leisure of the theory classes, was just beginning for them. So we felt very envious of our academic friends because they had much more freedom than even we did to do things like these junkets. This of course came later, but it was at this time that we began to stretch the limits. And it was because I succeeded, Matthias succeeded very much because he discovered a whole sequence of other ferro electrics. Materials which are still important as non-linear optic materials like lithium niobate [?]. And we did a lot of interesting experiments on barium titanate, even though it did not become the memory material of the future. So he succeeded. And he brought in— well, that's another story. So where are we? [Break] I was just saying some things about comparing the academic world at that time with the industrial world, and that in a way the academic world was very free, but we were gradually acquiring an equal degree of freedom. And what we didn't realize was that within a decade or so the tables were going to be turned and we were going to be considerably freer than our academic colleagues in many ways. But perhaps I should talk about Berndt Matthias. This marvelous, charismatic character that is unforgettable to everyone who has ever known him. He had considerable success in finding interesting ferro electric materials, but then he was attracted away to the University of Chicago. I think at that time Chicago opened it's Franck Institute that focuses on solid state physics. All around the country people were beginning to pay attention to solid state physics because the transistor had happened.

Kojevnikov:

That was what year?

Anderson:

That would be '51. Anyhow, he was attracted away, but he took leave from Bell Labs and went and worked with another man, John Hume, on superconductivity. And his specialty in superconductivity was just what he had done in Ferro electrics—discovering substances. He was widely known as an alchemist or widely described as an alchemist. I think that's an oversimplification, but this is not about Berndt Matthias. Anyhow, the labs gave him unpaid leave to go and learn superconductivity, but they decided that he was important enough and exciting enough that they wanted to attract him back. And he said, "That's fine. I will come back, but you have to begin research in really low temperature physics. Get yourself a liquid helium machine and start doing liquid helium physics."

Kojevnikov:

And that was as early as '51?

Anderson:

That would have been about '51 or '52. And so we needed liquid helium in fact for several reasons, because basically you can't do respectable condensed matter physics without liquid helium, and the new Collins liquefier had meant that all of our competing universities and laboratories would have liquid helium. So we were glad enough for him to build up a liquid helium facility at the Bell Labs. That meant that he was to work on superconductivity. It's not evident how you were ever going to use superconductivity in the telephone system, and no one has to this day, but that seemed to be acceptable to our masters. That was one of the first of the breakthroughs.

Kojevnikov:

Speaking about masters, we have never mentioned the names. Were they all faceless or were there some personalities?

Anderson:

By no means faceless.

Kojevnikov:

Who would the crucial people would be?

Anderson:

Well, one crucial person is Jim Fisk who is the person who set up the original quantum solids group with Shockley and so on. Stan Morgan was of course always important as this moderating character, but also a very open-minded man. One person who was terribly important was Addison White, a very little known man of enormous competence, who started out as head of the physical electronics group, but then graduated. I think was promoted to the next level of management. Another important person was W.O. Baker who eventually ended up— well, both Fisk and Baker ended up as successive presidents of Bell Labs. Baker was a very profound believer in the value of basic research, and possibly basic research got a little bit too far ingrained into the Bell Lab system by the end of Baker's presidency. Two others were Bruce Harman who followed Baker up the ladder starting in semiconductor chemistry; and Joe Burton.

Kojevnikov:

How would you describe? To what extent they would buy into the Vannevar Bush thesis the basic research would always find applicable uses later, and one shouldn't worry at the very beginning of the uses?

Anderson:

We were always— I think throughout this entire period there was a feeling that we were not going to go off too far and get too far from the telephone system. That we would support certain kinds of research, partially because we thought that the person involved in this research was valuable to us in other ways. For instance, we helped in the end, we were helping some of the astrophysical experiments. Of course there's the cosmic background radiation, but that's directly related to the Bell Labs. John Klauder was one of our theorists who eventually became very much a mathematical physicist. One of the country's eminent mathematical physicists. But he also always had an engineering interest and was a useful consultant engineer. Not always did we support these people, and sometimes they eventually dropped off, but on the whole there was very little that we wouldn't touch unless it just conflicted with the Bell Labs culture. We can not do particle physics because particle physics involves enormous collaborations, which we have no structure to work with. We couldn't do conventional molecular biology, because again there was a large group, many post-docs. Our overhead was just not structured in such a way that we could afford to do that. Too few results per million dollars in overhead, and so on. There are things that we can not do. But if it's an individual piece of research and that researcher does other things that are useful to us, we in the end had no serious difficulty with that.

Kojevnikov:

I was wondering about whether there were any specific units for let's say exchange of ideas and information. Presumably, the Bell Labs were dealing with patents and with possibly commercially useful information. How was a decision made about what gets into printed form? What can be openly exchanged? What goes into more protected form?

Anderson:

The era that I'm talking about the lab was in an anomalous position with regard to patents in that we were either forced or volunteered to be very open with our patents. We used patents primarily as quid pro quo and by barter arrangements with other major manufacturers. We had the transistor patented. We used it only to make sure that no one else had it exclusively. That the integrated circuit patent that someone else had would not be enforced on us and not interfere with our business if we didn't interfere with their business by enforcing the transistor patent. And so on. So openness was in general encouraged. There was, when I first went, a long procedure of patent review before you could publish. A time consuming procedure. It became less and less time consuming and in the end essentially broke down completely. Yes there were patent reviews. They became proforma.

Kojevnikov:

Was there any review procedures for papers when you wanted to publish a regular paper?

Anderson:

Yes, absolutely. I think that lasted for many years. But from being a very time consuming procedure in my first years there, by the 1970s it was just proforma and essentially you would have to say, "I would like the patent people to have a look at this." Rather than the patent people insisting on looking at it before you were allowed to publish.

Kojevnikov:

And how about going to conferences and discussing things there? Were there any review procedures?

Anderson:

No, if you could get to the conference there was complete freedom on that. In fact, usually there would be more concern with the scientific quality and veracity in general: whether you were taking the right point of view. Were you likely to get yourself in trouble? than they would be with any patent consideration. I remember one case which we may get to. Where management was very concerned about press coverage of a certain— not because it was of any patent reasons, but because it contradicted something that someone else from the Bell Labs had said. It was right, but they didn't like the appearance of it. It wasn't even a contradiction in appearance. It was the appearance of negating something someone thought was important.

Kojevnikov:

Who do you consider as a major competitor in institutional terms?

Anderson:

At first GE was the traditional competitor. The Laboratory of Langmuir and Steinmetz and so on. GE didn't really make it in the condensed matter business. RCA was a minor competitor. Then IBM began to grow up and toward the end of my stay there IBM considered themselves our major competitor, although they never were more than half the size and strength. Of course there were a large number of imitators that appeared and then disappeared. Like Xerox Parc Laboratory, which was very good for a while, but then didn't really make it into the big time in terms of basic research. IBM was serious. GE was serious in the early days. The labs always had some kind of formal liaison with Philips Laboratories in Holland. Although, Philips is much smaller. Nonetheless, we had a great deal of respect for them. Siemens Laboratory I don't think was ever that serious. Although again, we kind of had formal, courteous relationships with them. When administrators, when bosses would junked they would take the trouble to visit Philips or Siemens.

Kojevnikov:

Okay, so maybe if we return to Matthias's case and how he had managed to establish low temperature research. You mentioned earlier that Bardeen had left because he didn't feel like being able to do this kind of research at Bell Labs.

Anderson:

Well, Bardeen left for a mixture of reasons. There was the great difficulties with Shockley that are historical record by now. There was also the fact that he very much wanted to work on superconductivity, and felt that full time superconductivity was not the right thing and at the time that he left wouldn't have been. Then there's the simple fact that he left for more money. The labs got behind in it's salary scale because of it's conservative—

Kojevnikov:

Behind the academics?

Anderson:

Behind the academics in salary scale. Academics found it too easy to get very good consulting contracts, as well as to some big salaries for stars like Bardeen. Later on, for instance, there were the Einstein professorships, very political at the time.

Kojevnikov:

You were not eligible to apply for grants, right?

Anderson:

No, we were not eligible. Not until much, much later did labs people apply for grants. We didn't have grants. We didn't have independent funding, so we couldn't essentially take off to a meeting on our own funds. We always had to ask for money to go somewhere or do something.

Kojevnikov:

When did you personally start seeing yourself as being totally independent and what topics you would deal with and choose and how you would deal with it? When did your status in the lab, maybe formal maybe informal, improved to the degree that you wouldn't have to worry about whether your topic—

Anderson:

Well this is a little bit of a long story, and maybe I should tell it, which is the story of the formation of the theory group. It started with Bardeen. Bardeen left in 1951 already. We had this marvelous theory group which contained the best theorists in the world.

Kojevnikov:

How big was it?

Anderson:

It wasn't so terribly big. It was in the order of ten if you added, but it wasn't a group, you see. It was a theorist here, a theorist there, working in different departments. As I said, Suhl and Walker were working—

Kojevnikov:

So it wasn't a formal theory group.

Anderson:

It wasn't a formal theory group. And all of us felt a little bit that we were judged very much— I didn't mind. We were judged very much on our consulting ability, which kind of measure I could have gotten along very well because the whole field of magnetic resonance was growing by leaps and bounds at that time and I was the only theorist around who could consult effectively on magnetic resonance. Let me answer your question first. Then the next loss was Kittel, and that hurt quite a bit. Kittel left to go to the University of California in Berkeley. The University of California Berkeley had a lot of open slots because they had great losses during the oath controversy, the big controversy about the loyalty oath. And so Charlie went, and he was the senior theorist in the place. So he had a lot of freedom to build up his own group.

Kojevnikov:

He didn't consider the oath a problem?

Anderson:

He didn't have to sign the oath. He came after the oath.

Kojevnikov:

So they didn't require it from him?

Anderson:

They didn't require it from him, or at least so he said. Then the semiconductor people, during the summers, had gotten in the habit of bringing in two theorists from academia. One was Walter Kohn, and the other was Quin Luttinger. Very able people who essentially built up the dynamical theory of electrons in semiconductors. Solved the problem how to take spin orbit into account and these complicated bands, the many minima in the conduction band of silicon and so on. And did a lot of other work too. We had always hoped that we could hire at least Walter Kohn. Quin seemed to be more of a committed academic type. But then Walter went off to the University of California in San Diego, which was just building itself up. And he became the department chairman there. And what's more, he began attracting people from the Bell Labs to the University of California in San Diego. Gregory Wannier left for the University of Oregon about then. We had another physicist, interestingly also involved with the oath, named Harold Lewis, who had been an oath refuser who had been at the Institute for Advanced Study, but had then decided that he would come to the Bell Laboratories and work on superconductivity with Matthias. That never worked out. The two of them were as different as— absolutely incompatible. Jack Sprat and his wife or whatever. Harold, although he hadn't signed the oath, throughout the entire oath controversy maintained his secret clearance and later on became a very considerable hawk and the chairman of the Jason Study Group. But we were the first place that would hire him, and he came quite willingly and was very compatible with the group. Added to our strength no end. He left and went off to UCSD as well. Harry Suhl was threatening to go there. So the Bell Labs began to be worried would there be any of us left. So there was Conyers Herring and there was I myself, and there was a young man names Wolff. And so they called us in and Addison White and Baker between them said, "What do you want?" Interestingly, the person who designed the theory group was the one who left. Namely Harold Lewis. He was a good organizer. He designed the group and said here are the things you should want, you should demand. And we demanded them. We demanded freedom to travel. We demanded freedom to have post-docs.

Kojevnikov:

Official post-docs?

Anderson:

To have official post-docs. We demanded freedom to work on what we pleased and they said yes. In 1956 this was. That I think is the point at which officially— I mean I would not pay any attention to what anyone told me to do before, but only through stubbornness and stupidity. Now I knew I had the right to work on what I pleased. This was an official decision, actually, of these two people. Baker by now was head of the entire research department, and Addison White was the second level of administration above our department chairman. So it was a decision to hold some kind of theory group together.

Kojevnikov:

This means creating a separate group institutionalized?

Anderson:

It was an institutionalized group. Later on this group became the model for a number of other things at Bell Laboratories. The post-doc idea was borrowed liberally and freedom to travel within budgetary limits was. That was the first, and this was the true younger Turks rebellion. Peter Wolff and myself and to some extent Conyers Herring. Harold Lewis, who then left. For the first two years we had a wonderful time with our new group. The first two years we had an even more eminent set of visitors. We had Keith Brueckner, the nuclear physicist. The man who discovered the N-star I believe. We had Philippe Nozieres who was then working with David Pines. We had David Pines himself. Bob Schneffer was there both years, both the 1956 and 1957 year. Walter Kohn and Quin Luttinger were of course there. We had these incredibly fertile two years of having anybody in that we wanted. Some particle theorists, J.G. Tayler, John Ward. And we hoped that some of them would stick. Mel Lax did come and stayed with us. We also had two post-docs who were pretty successful too. The first post-doc was Jim Phillips, who essentially invented the pseudo potential method. The second post-doc is John Hopfield, who did marvelous work in solid state physics, but he's really known for his biophysics work, which is really spectacular and eminent. So there was a specific time and it was a specific act of rebellion or at least a request on our part saying we all can get jobs and we can go where we like. Do you want to keep a theory group or not? What we wanted was freedom to do what we liked—what we thought was important in theory. We were still going to consult, to discuss experiments, and we did very well with that over the years. But the basic idea was that the criterion for quality was theoretical physics rather than otherwise.

Kojevnikov:

Before the jobs became plentiful, at the time when theorists were as you said, dispersed over several departments and groups, how was the information exchange organized besides informal talking to each other?

Anderson:

Well the first thing Wannier did and possibly the best thing he did for the Bell Labs was to institute a theorists tea. Another institution which developed in the course of these years was Conyers Herring's journal club. Conyers Herring's journal club still takes place, but it's in Stanford, not here.

Kojevnikov:

What was the format of this?

Anderson:

The format of the journal club was to spend a morning discussing papers.

Kojevnikov:

Is it once a week?

Anderson:

Once a week. One morning once a week. Take an hour and a half or two hours, discuss not one paper, but half a dozen to ten papers at one go.

Kojevnikov:

Who is presenting or how is it?

Anderson:

Anyone. Experimentalists, theorists, whoever Conyers chose.

Kojevnikov:

So he chose the papers.

Anderson:

He covered the entire literature. He read Russian. He covered the Russian literature particularly.

Kojevnikov:

Where did he get Russian from?

Anderson:

I don't know. He just knew it. I don't know how he learned his Russian. But that's how we learned about all the Landau group's paper. But also there was lots of very interesting Russian experiment. The first electron drop. The first mesoscopic measurements by Keldysh. Not Keldysh, but— of course there's the Keldysh work, but Sharvin and Sharvin's work we heard of from Herring. We heard of Azbel and Kaner, Azbel's work from Herring. He would take for himself three or four of what he called quickies and sometimes he'd get to them and sometimes not. But we had a critical discussion of papers all across the board of our interests. The discussion was critical. It wasn't, "Here's what he says." It was, "Here's what he says and he's full of beans." Or, "Gee this looks interesting." Or whatever.

Kojevnikov:

How would people follow Herring? Would you accept his suggestion to review an article?

Anderson:

Generally accepted his suggestion because it was usually well chosen. If you didn't there was no pressure.

Kojevnikov:

How many people would attend?

Anderson:

It varied. People wandered in and out. There would usually be twenty or thirty in the room at any given time. There was a schedule more or less ahead of time so that you would know more or less what was going to be talked about.

Kojevnikov:

And the theorist's tea, what was that meeting for?

Anderson:

It became both coffee and tea. Just that everybody met together at more or less accepted times.

Kojevnikov:

But that was without any presentation?

Anderson:

Without any presentation.

Kojevnikov:

Was there any discussion of your own work in a formal way? Like presenting at a seminar?

Anderson:

There was a solid state seminar and colloquiums. A standard thing. There were a lot of talks. Do I have a copy? There still is to this day. There's a research calendar that's fairly full—as full as a typical university department. There were lots of talks. I think there was a regular solid state seminar and if you had something you presented it.

Kojevnikov:

In case you would have some ideas which you would like experimental physicists to test?

Anderson:

You would go and see the guy.

Kojevnikov:

And so what would their reaction be? Was it easy? Was it difficult? Was it any problem?

Anderson:

Well it depends on whether it's a good idea, whether he could do it. Of course he's much freer to do what you want him to do at Bell Labs because he's not tied to an enormous amount of equipment that he's got from some grant. Equipment is first quality. It's usually bought, but he doesn't have to— no, it's not capitalized at a great cost. Okay, someone like Doug Osheroff has a sub milli-Kelvin cryostat and he's only going to do things that involve the sub milli-Kelvin cryostat. But someone like George Feher, he has an electron magnetic resonance spectrometer. He'll stick anything in there and he'll do other experiments if he's interested. When we heard from Nicolaas Bloembergen of the first possibility of building a three level maser, George Feher, Derrick Scoville, and one other person set to work. And in three weeks they had a three level maser operating. That's the kind of response time that we had. Several experiments went like that. When you heard that someone had done something— Well of course when we heard that what's his name in California had managed to make a laser out of ruby, Art Schawlow had one working the next morning. It was very quick response. And the point is there's no heavy capitalization, no investment in ten graduate students who have to be found work, no time delay while the graduate student learns what the hell you're up to. You just do it. That was wonderful. I had an idea, it could be worked on tomorrow. Only that didn't happen terribly often, but it did happen occasionally.

Kojevnikov:

Is it time to go back to your papers and to work ?

Anderson:

Sure, let's go back. Yeah, maybe this is a very good time to stop talking about the laboratory and what it was like to work there, which was wonderful. Particularly, no grant applications. You went to your boss and said I need equipment for this or I need to go see so and so and learn something from him. You went.

Kojevnikov:

What would be the acceptable limits of demand, so to say? In any lab there is some understanding of course.

Anderson:

Well, of course. It has to be an amount of equipment that can fit, say on a desk top or something.

Kojevnikov:

So basically it was equipment for one room?

Anderson:

Yeah, one room. Well a new experimentalist goes in, his laboratory is automatically equipped with what he asks for. So then I got interested. I continued of course with odds and ends from the theory of line widths in gases. So it was a natural thing for me to be interested in paramagnetic resonance line widths. And one of the things experimentalists were studying at the labs at the moment was paramagnetic resonance lines. And so in exchange in paramagnetic resonance, that's one of the two things that I then did that were major projects rather than cute little neat things.

Kojevnikov:

So these were starting to experiment—

Anderson:

Ferro electrics was a major project too. Ferro electrics it was conceptual rather than— I was saying there's two kinds of phase transitions. There's an orientational kind of phase transition where a magnet pins turn over, then there's displacing kind of phase transition where some susceptibility just gradually goes through zero to negative and becomes an instability. And that's a conceptual thing. But I also was interested in the general problem of spectral line widths and relaxation in materials. There was known to be this phenomenon of exchange narrowing, but there was no formal theory of it. There was a numerical estimate using the so called moment method by Van Vleck and Gorter. If nothing else I showed really how formally the exchange causes narrowing of the spectral lines and introduced the concept of an exchange reservoir and a spin reservoir, a lot of the concepts that later on were made more formal by Abragam and his ideas about negative temperatures and so on. And this I published with Peter Weiss at Rutgers. Peter Weiss really didn't contribute much to it, but he had been writing a paper that was kind of a sub-bit of this paper and was independent. So I thought it much neater to invite him in and write the single paper. The other thing I did at this time is this thing called an approximate quantum theory of the anti-feral magnetic ground state. And that may be the first paper in which I did something really, really important in the sense of reaching into the fundamentals of physics. The whole question of why anti-ferromagnetism occurs had always been an interesting one for theoretical physicists. Bethe did the solution of the one dimensional anti-ferromagnetic chain way back in 1931. And the one dimensional anti-ferromagnetic chain with anti magnetic exchange integrals is not anti- ferromagnetic. It doesn't exhibit order—it doesn't make anti-ferromagnetic order. A lot of very famous physicists had played with this problem of is there anti-ferromagnetic order in principle?Ê It was discovered in practice after the second war as soon as neutron defraction became possible. The Oak Ridge group found anti-ferromagnetic order. So now the question was experimentally solved. But why was it possible? Why shouldn't it be possible? It's because to put it in theoretical terms (Landau, incidentally was among the unbelievers.) Landau, although he was one of the discoverers of the phenomenon of anti-ferromagnetism, the co-discoverer with Neel in the 1930s, he justified it by having ferromagnetic layers that then just happened to be opposite to each other. But he didn't really believe that a true three dimensional anti-ferro magnet could happen. Kramers had worked on it a great deal. So many of the great figures had worked on it and thought about it and found it puzzling. The reason is that the ground state of the anti-ferro magnet is clearly not the nominal low temperature state, which is not an eigenstate, (it's not the grand state). The ground state of the ideal anti-ferro magnet can be proved to be a singlet and they all knew the proof goes back to Bethe I believe. But therefore, it can not have a preferred orientation in space. And the question is why does it have a preferred orientation in space? And this is the core of the phenomenon that I named later, much later. I named it broken symmetry.And what happens is that we call the ground state of such a system quasi degenerate. There are a great many "ground" states, all or almost the same energy. Since many ground states or many low states can be formed into a white packet, what I was able to do in this particular case was to show that wave packet takes a very long time to break up. I made an estimate that even the very ideal case would take thee years to break up. And if you apply any weak constraint on it, it just never breaks up. I didn't realize quite how important this was. I was very young. I thought this must be known to everyone. I somehow thought that there were all these wise people around, brilliant people who had invented quantum mechanics and they must know that this principle exists. But it turned out they did not.

Kojevnikov:

Can you mention is there maybe a specific area where the idea could come from?

Anderson:

The idea how to do the math actually was in the literature. You'll be interested to know that it was in the thesis of M.J. Klein, another historian of science, who then left physics. But he was only dealing with Ferro magnets. And so this leads to trivial—

Kojevnikov:

And where was he?

Anderson:

I don't know where he was at that time. This leads to trivial results for the Ferro magnet. But for the anti-Ferro magnet the same method leads to these very interesting results. The formal theory I didn't even know until twenty years later. The formal theory turns out to be something called the non linear sigma model. But the basic physics is this idea of the quasi-degeneracy of the ground state and the long living wave packet that is easily manipulated. And I recognized that this same phenomenon was really responsible for phonons in solids and so on. We got a spin wave spectrum and the spin wave spectrum also had pre existed due to semi-classical theory. There were semi-classical studies of it, but nobody had ever really done the quantum theory properly. This told you why two and three dimensional anti-ferromagnets could exist, but one dimensional could not. In fact, two dimensional isn't thermally stable as is now known. (I think I knew it at the time). This one together with exchange narrowing called the attention of Ryogo Kubo [?] particularly to me. And I met Kubo who was at Chicago at that time. And we discussed this and we discussed things having to do with fluctuation dissipation and it's use and in spectra. He did not speak very good English and I didn't speak any Japanese at all, but we somehow communicated. He had tried to do this same job and not succeeded. Not really completely succeeded, had not gotten a complete answer. So this seems to have impressed him, so he got me invited to Japan through the Fulbright Foundation. I was at that time twenty-nine and three years post Ph.D. and very little known. So it was very strange that the very first Fulbright physicist ever to go to Japan should be me. So Kubo: claims much later he told me, "After all, I discovered you." I said, "But I thought I discovered you." But anyhow, he said, "Could you come for a year starting in 1952?" and I said, "No. No way. I just bought a house." And he said could you come for half a year starting in 1953?" and I couldn't say no to that because I had said if it was less and later maybe I could do it. I was not sufficiently definite in refusing, so I went. Bell Labs let me go.

Kojevnikov:

What did he hope to achieve?

Anderson:

Well, they achieved what they hoped to achieve, which was to train a group of young physicists in the theory of magnetism. And for me to work with Kubo. Both happened.

Kojevnikov:

Which was where? What institution?

Anderson:

At the University of Tokyo. Also they were having their very first post-War meeting. They had an international congress of theoretical physics. It didn't have everybody. They were very disappointed they didn't get any Nobel Prize winners. All they did is get all the future Nobel Prize winners. They had essentially everyone who was to become anyone. Like Yang, myself, and Onsager who had not yet got his Nobel Prize. I don't know. The list went on and on and on. I once counted twenty future Nobel Prize winners at that meeting. No present ones. Van Vleck of course. It had the effect that I met everyone— or not everyone, but a great many of people of the people who were to become important people in theoretical physics. That was important. Contacts were terribly important I think in the course of history. I also met Per-Olof Lowdin who in the end was the chairman of the committee that got me the Nobel Prize. But I don't think that had anything to do with it. I had actually corresponded with Per-Olof before on my work on ionic crystals, as a matter of fact. But that was important. And then working with the Japanese group. And there were a number of very important physicists who essentially ran Japanese solid state physics for twenty or thirty years. Kubo, Moriya, Yoshida, Kanamori, Kondo, were all in my seminar. The whole of Japanese solid state physics. So here I was this twenty-nine year old who couldn't speak Japanese and they couldn't understand what I was saying. But they copied out my notes and published them, and they all learned magnetism from that. I think I influenced the direction. Maybe not for the better, but I influenced the direction of Japanese physics.

Kojevnikov:

So how long did it take at that time to start from not knowing solid state physics at all into becoming an expert in at least some field?

Anderson:

Well solid state physics had barely existed before. I was an expert in magnetism. Magnetic resonance hadn't existed when I started out, for practical purposes, so it wasn't hard to become an expert. That's the story of my life. I'm always in a field which I don't have to learn because it doesn't exist until I've been in it for a while. That's not quite as arrogant as it sounds, but it does mean that I'm not a good person to ask for references, for instance, because I very seldom can give you the references because I learned it before the textbook was written. Kubo and I worked on these things like stochastic methods and line broadening problem. And I published in Japan.

Kojevnikov:

Was Japanese culture much different from the American?

Anderson:

Popular culture was very different and yet Westernized in some ways. There is a book by an Italian called "Meeting with Japan" which describes it perfectly at that time. Many traditional aspects survived, we stayed in Japanese inns, were served occasionally by geishas at restaurants. There were many traditional handcrafts available, etc. The present corporate culture however, had not so in some ways it was freer than now. It became more corporate and more collective. Kubo was very western. Is very western.

Kojevnikov:

So did he have any training?

Anderson:

Not really. He spent some time as a post-doc in Chicago. But no, not really. He just was. He was very independent and very free wheeling. Of course the particle and ion energy physicists have always been much more like westerners than the other physicists. I guess it's changing a little bit now, but there were a lot of social customs that were strange. We had them all at the department over for a cocktail party. We had a little bell outside the door on the garden gate. We hear the bell ring again and again and again. Apparently all the people had to arrive and order themselves properly for coming in the door. The senior came first and then all the others in a chain. Then they all sat around the room and we served cocktails and hors d'ourves and nobody spoke very much unless I spoke or unless my wife spoke. And then after three rounds of cocktails the senior man said, "I'm drunk." Then the next senior man said, "I'm drunk too." And it went all the way down the row and then they got up and put their shoes on and left. So socially they were very formal.

Kojevnikov:

How about academic life?

Anderson:

No problem. I worked with some students, some post-docs. Worked actually together with them and it was perfectly good to and fro with Hasegawa for instance, with whom I wrote a paper, with Kubo with whom I wrote a paper. I had a very close relationship, and became very, very fond of Kubo. Saved his life once. Everything was heated by these awful, soft, rubber tubes of gas. His heater tube broke and he was bright red lying in his office when I realized what had happened and dragged him out.

Kojevnikov:

You said that was kind of test for the Bell Lab to let you go?

Anderson:

They did not give me a leave with pay. It was not a sabbatical. It was definitely leave. I think the next year they gave Conyers Herring a leave with pay to come to the Institute. They gave him an actual sabbatical. And thereafter these arrangements were mostly on a sabbatical basis. One of the things we had insisted upon was formal sabbaticals in the 1956 rebellion. And we all took them. So when we got back to Bell Labs, it turned out that now the problem was salaries. The starting salary level had risen faster than the laboratory system of raises, so new people were being hired at salaries higher than those who had been there longer. The longer you had been there the less money you made. So we had a rebellion about that. Bob Shulman led that rebellion. In particular, George Feher had been hired at a salary that was I think a hundred dollars a month more than I had left with and twenty-five dollars a month than I came home with.

Kojevnikov:

How did you know?

Anderson:

There was great secrecy, enormous secrecy. Bob Shulman fixed that. He invented the Shulman list. He wrote his own salary on the top of the list and said to the next person, "I'll show you the list with my salary on it if you will put your salary on it." And then he went to the next person and the next person. In the end we had all the salaries, and they were all the same essentially. They were all equal to the starting salary or slightly less. And so the Bell Labs changed that. That was actually the first rebellion that happened before. The first year I think Conyers Herring got a seventy-five percent raise and the next year I got a seventy-five percent raise and so on. So they started to spread the scale out. And then we were as well off as our academic colleagues. Almost as well off as our academic colleagues. So there was that whole— That too with the younger Turks. Bob Shulman did his younger Turk bit. Then there was the oath and the yellow signs, but those were political.

Kojevnikov:

How is this related to Bell Labs?

Anderson:

Well, it was in the McCarthy period. There wasn't an oath. There was a questionnaire that you were supposed to sign saying, "Do you belong to this organization, that organization, or any organization?"

Kojevnikov:

And at the Bell Labs this was also signed at this time?

Anderson:

We all signed it except Gregory Wannier and Alan Holden and myself. And people watched with interest and nothing happened. I think all of us sent the questionnaire back as a questionnaire, but we didn't sign it formally. We chose not to become legally liable for this questionnaire, but we answered it truthfully in each case.

Kojevnikov:

What is it that you mentioned? Yellow signs?

Anderson:

The yellow signs. Never mind. That was nonsense. They kept yellow signs saying that this is a defense facility. "No person with any subversive intent can ever pass this yellow sign." We were already having regular visitors from the Soviet Union, so it was clear that communists were regularly passing those signs.

Kojevnikov:

Who would come first from the Soviet Union. Who were those visitors?

Anderson:

I went in December 1958, and in a return visit Vul, Vonsovsky – Three of them. The guy who's the nephew of the biologist who was executed.

Kojevnikov:

Vavilov?

Anderson:

Vavilov. The three of them came. There's some funny stories about that, but you don't need to know them. Like poor Vavilov refused to eat anything but soda crackers and water straight from the tap because he was the commissar for that visit. The next visit he wasn't the commissar, so he got drunk and insisted upon playing footsie with Berndt Matthias' wife under the table. On the other hand, Vul on the first visit got quite drunk and pounded our piano to great effect. So it depended on who was the commissar and who was not. But the whole thing was ridiculous. Anyhow, Bob Shulman set to work once and took down all the yellow signs. We started to take down all the yellow signs.

Kojevnikov:

When was that?

Anderson:

Oh I don't know. That must have been 1960 or so. They were the last remnant of a law which every other part of had been declared unconstitutional. We were stopped and our phones were tapped. Large security people moved us away from there. It was very funny, but it didn't mean anything.

Kojevnikov:

So what was the relationship with the real defense work and the work done at Bell Labs?

Anderson:

It was real defense work, but it was done almost all at a different lab. Whippany [?] Lab. There was some defense work here and there, but it was carefully segregated. Bell Labs gave up most defense work. It kept on as chief contractor for a Safeguard Ballistic Missile Defense, but then we even ran a protest movement against Safeguard. At one point when the academic world was doing the March 4 protest and so on and running campaigns against antiballistic missiles, we had an anti-antiballistic missile group organization at Bell Labs which actually published an ad with lots of names in it and so on. We didn't really behave differently from academics even though our laboratory was the prime contractor for that operation. We were allowed political freedom. We were not allowed to sign the institutional affiliation, but that got sneaked into the text and there was a fuss about that. Back to science: well, I came back from Japan and kind of spun my wheels for a while. I did a lot of different kinds of work. Work on ferromagnetic resonance. Harry Suhl's first work on non-linear effects in ferromagnets I joined, but really he was the important person on that. And all this time I was thinking very hard working regularly with George Feher. He came to the labs while I was away in Japan. He was a new generation in terms of the spectroscopy magnetic resonance business, in terms of automation of his laboratory, and used liquid helium as a matter of course. He had the appropriate computerization at that stage of development. He was state of the art in that respect, and he was extremely brilliant. He became a biophysicist. I believe he should have shared the Nobel Prize on the chlorophyll structure, but he didn't. He certainly did very good biophysics. Anyhow, I just went in and watched over his work. He was working on these fantastic samples of semiconductors that we were being given by the semiconductor preparation people at the Bell Laboratories. There had been early work by Fletcher and other people on the magnetic resonance. And while I told them how to interpret it in my consulting role, I didn't appear on any of the papers. But I told them what it all meant. But I was fascinated by George's quality and the fantastic detail. And the new method that he developed called Endor after the witch of Endor. He was an Israeli who had fought in the War of 1948. I became fascinated by this Endor method and how it worked. He was studying magnetic resonance as a function of concentration of impurities. So the problem becomes is there an impurity band, or are the impurities, the wave functions of the impurities really separate? This again was a deep conceptual problem and it fascinated me more and more. We kind of wrote, or he and Meyer Weger who was working with him, wrote various papers on the whole question of how these various cases, what we call passage cases. What happens when you change the frequency and you dig a hole in the line and watch the hole repair. There was an atmosphere or a manifold of impurities of slightly different frequencies. Slightly different energies. In the semiconductor, by changing the concentration, you make them interact weakly or strongly. You can watch what happens when they interact weakly or strongly. And none of it made any sense in terms of the thinking of solid state physics as it then was. So although I worked, I spent most of two years sitting in that laboratory watching his data come out... not really most of two years, but at least one day a week, sitting and watching his data come out. Talking over the data with him. Helping various people. I helped Kohn and Luttinger who interpreted the frequencies of all these lines that we saw. Elihu Abrahams and David Pines came one summer and talked with me about these problems. Essentially I puzzled about it and then I eventually realized that what was happening was localization or what later became localization. And so I invented a model. Nothing like the real thing, but it was a model which I could solve and told me that wave functions can be either localized or extended. That's the absence of diffusion, which was the Nobel Prize paper.

Kojevnikov:

What was the problem with conventional quantum theory?

Anderson:

Well conventional ideas were that when two wave functions overlap they mix. When an infinite number of wave functions overlap, they all mix. And the wave functions must just be extended over the whole sample. In other words, if you have a continuum of energy levels it's a true continuum. I showed that you could have a continuum of energy levels which locally is not a continuum, but is discrete. I showed this strange fact of discreteness to happen on a local basis, even though the total spectrum is a continuum. And well there's several ways of thinking of it. Basically the way I approached it was also somewhat new. I used what I called the locator series as opposed to a propagator series. Kind of a set of Feinman diagrams in terms of Green's functions on local sites. And treating the hopping between sites as a perturbation, and doing a perturbation theory in that. Showing that it can converge or diverge. It converges in the localized case.

Kojevnikov:

What was the response to this paper? I presume it was—

Anderson:

Zero. I talked about it when I finally got it done. I had a lot of help. Peter Wolff particularly. I was basically very naive about perturbation theory in those days. Not naive about using it. Naive about the formalism of perturbation. Peter talked to me about that. Various people taught me a lot of things. I talked about it, and the two people I respected most aside from Conyers, who kept his own counsel, were Walter Kohn and Quinn Luttinger. And Quinn Luttinger said that's trivial, and Walter Kohn said it's wrong. So I knew I had found something because I knew it wasn't wrong. It was not obvious and it wasn't wrong. So it's really something reasonably subtle. It did turn out to be right. But for five years at least it made no splash. I published it. I was already thinking these thoughts in 1956. They had a rerun of Tokyo, of Kyoto actually, of the Japanese meeting. They had a big meeting in Seattle to which I was invited because I knew all the people and had been in Tokyo. It was invitational and some of my friends were quite bitter that I was invited and others weren't. I went, of course. And I tried to talk about this localization problem. I did not have it really clearly in mind. It lit with a complete thud and nobody paid any attention to it. But I did at least talk about it. That was already in 1956. It took me another year to write it up and then another year to get it published. So it appeared in 1958, although the work was really done in 1956. And then George never forgave me, but I went with Larry Walker and Harry Suhl to Princeton to listen to David Pines, who had come from Illinois to give the colloquium on superconductivity, and I changed fields...