Specht, Robert. Date: June 26, 1989. Interviewer: Martin Collins. Auspices: RAND. Length: 2 hrs.~ 30 pp. Use restriction: Public.

Specht (b. 1920) initially reviews his upbringing and education. He then discusses ~1hy he accepted a position in the Mathematics Department of RAND in 1949, John William's leadership of the department, the usa of consultants like Warren Weaver and , his involvement with the development of systems analysis, and the evolution of the Systems Development Division. Specht next describes his impressions of Frank Collbohm, other projects the Mathematics Department worked on, and the department's contributions to RAND.

TAPE 1, SIDE 1 1-3 Early family background and history 3-4 Initial Lttraction to RAND; recommendations from Quade and Germond 4-5 Responsibilities and activities at RAND~ John Williams' approach in Mathematics Department 5-7 Description of RAND interdisciplinary projects during the first fifteen years RAND's use of consultants; Warren Weaver's contact with the Mathematics Department 7-8 Von Neumann's contributions to 8-9 Developm:nt at RAND

TAPE 1, SIDE 2 9-11 summer studies at RAND 11-12 Contact with Quade and Paxson during their development of systems analysis 12 Specht's involvement in and sensitivity to Systems analysis issues 12-13 Striking a balance between research and administrative activities 13-16 Evolution of the Systems Development Division Characterization of Frank Collbohm's positive and negative prejudices 16-17 Project briefings and presentations Investigations into dynamic and linear 18 Programm.:l.ng

TAPE 2, SIDE 1 18-20 Sam Genesky's creation of the Center for the Partially - sighted 20-21 Intellect.nal framework of early RAND studies 21-23 Notion o.~· the Kriegspiel as an integral aspect of RAND culture 23-28 Mathematics Departments' contributions to RAND efforts 28-30 Closing c.::>mments ORAL HISTORY INTERVIEW

Interviewee: Mr. Robert Specht

Interviewer: Mr. Martin Collins

Date: June 29, 1989

Location: Santa Monica,

TAPE 1, SIDE 1

Mr. Collins: Just as an introduction into your experiences, it might be useful to sketch out your background, where and when you were born, your parents and your educational background.

Specht: Background, what I'm qualified to do--there's an ancient story: a mother says to her son, "By me you're a lawyer, by you you're a lawyer, but by a lawyer, are you a lawyer? When I told the joke in Germany the punch line is " ... aber bei un Advokat bist du ein Advokat?" That's about the extent of my German. I have the union card of a mathematician, but by mathematicians, I'm a cellist, and vice versa. Neither group will admit to knowing me. In any case, I was born in Seattle, Washington, 1913, moved to Florida as an infant, grew up and went to school in Florida, finished high school when I was fourteen in my senior year, fifteen about a month before graduation. That would be a Depression year in Florida, '28--the Depression hit Florida earlier--and I worked four years, wholesale drug house, office boy and then bookkeeper, and saved up enough money to go to college, University of Florida, but only enough for three years, so I had to go through in three years, in mathematics. Then I spent two years as a teaching assistant getting a master's, went to the University of Wisconsin for a doctor's, and taught at Florida.

In the war years, I was a physicist for the Navy, by virtue of the Civil Service Commission, my only claim to being a physicist, at Taylor Model Basin, Bureau of Ships, and then taught at the University of Florida, assistant professor of mathematics, got a job at RAND and moved in '49 which would be, I guess, the second year in which RAND was a private corporation as opposed to a Douglas project, which it began in 1 46. I got invited to RAND because a couple of friends whom I'd known at Florida were here, Ed Quade, Hal Germond. Quade was living in the Palisades.

Collins: Let me just quickly ask, what were your areas of interest in mathematics in graduate school? Specht-2

Specht: Applied math, which is somewhere in between engineering and mathematics.

Collins: Did this involve you in taking engineering type courses?

Specht: No. The closest I've come to engineering, I taught engineering at UCLA briefly, linear systems, which is more nearly mathematics than it is engineering, for that matter.

Collins: To move on to the next period when you were with the Navy, what was the nature of your activity as a physicist? Were you doing applied mathematics problems?

Specht: Yes. The Model Basin, as you may know, which is out past the end of the streetcar line 20 which runs along the river to Cabin John and back, was created because the Navy was paying large amounts of money for designing ships that would go faster than the specifications in the contract. It was to their interest to be better predictors, and they built a towing basin in which they would tow models and measure drag and get the wave lines on the side, and the rails had to be curved to fit the curvature of the earth, because over a mile or whatever the length was if it's just a few inches, the model has been lifted a few inches out of water. So for that reason you had to have that precision which you would not otherwise have expected. But then at the time of the war they were doing all kinds of work for the Navy, physicists, and mathematicians and Jesse Ormondroyd of the University of Michigan, a first-rate mechanical engineer, headed the engineering group, a good man, and I've forgotten whatever I did useful. Maybe nothing.

Collins: I just wanted to get the flavor of the activity. In the immediate postwar period, what were you thinking about in terms of what you might do with your career? You said you went back to the university.

Specht: I probably didn't know. Well, I went from there to the University of Wisconsin to teach. I probably never was well enough organized to plan ahead, think ahead and so on. As today, you know: one year at a time.

Collins: But academic life in some sense was attractive to you.

Specht: When I went to Wisconsin as a graduate student, I wanted to combine math and physics, and they suggested I forget about math and enter physics. They didn't mean that as a compliment, you understand. And so I took the hint and just studied mathematics. I never did learn any physics, but it has always been attractive to me.

Collins: Do you recall what your dissertation topic was on at Wisconsin? Specht-3

Specht: Something in elasticity, strength of metal beams, which was Sokolnikov•s specialty. When I was teaching there I had only one doctoral student who did his dissertation under me and he did better on it than I could have, so, good man.

Collins: So how did you first become aware of RAND and what attracted you to the possibility?

Specht: I think I'd never heard of it until I had letters from Quade and Germond who were already here, whom I'd known at Florida, and they described it in glowing terms, and I came out as a visitor to interview or be interviewed.

Collins: By Quade and Germond?

Specht: No, it would be by John Williams, who was one of the world's unique people.

Collins: Do you recall anything about that initial exposure to RAND?

Specht: Not really. No. This was before the current building, of course. They were down on 4th and Broadway in Santa Monica. The people, of those, those I do recall, and there were some weird and wonderful characters, but now I'm not sure they were there the day I came in but later. Women from the very beginning, there were very few but very gifted women. If you look at the RAND "Index of Selected Publications, 1946-1962 11 (the unclassified publications) you will find a number of women represented--Selma Arrow, Elsa Bernaut, Bernice Brown, Janet Chapman, Leola Cutler, Alice Hsieh, Hilde Kallmann, Bella Kotkin, and Ruth Wagner, for example.

Harriet Kagiwada came to RAND in 1961, with a Master's degree in optics to work as a programmer for Dick Bellman. She went on to do important research in automatic control, radiative transfer, mathematical biology, and numerical methods. As a result of this work at RAND she was given a PhD degree in astrophysics by Kyoto University. And Harriet didn't speak a word of Japanese. consultant Margaret Meade wrote one of the early RAND books, Soviet Attitudes Toward Authority, An Interdisciplinary Approach to Problems of Soviet Character (McGraw-Hill, 1951). Nancy Nimitz, the admiral's daughter, and who now lives on a hilltop in Topanga, was writing on Soviet and agriculture, and, more recently, Soviet research and development.

When came in as president of RAND in 1967 the number of women increased markedly, but mostly at the--not quite scutwork level, but as research assistants and the like. When Don Rice became president in 1972 the number of senior women increased markedly. An economist friend of mine said I was wrong in saying that they were senior women because he said that most Specht-4

of them were at the level of beginning PhD's, but I argued since only forty percent or there-abouts of the professional staff have PhD's, that's senior enough, you know, not to quibble. But I've left behind your question about the very early years.

Collins: Well, let's look at what some of your initial responsibilities and activities were when you first joined up with RAND. Specht: I think, but I'm not positive, that John Williams essentially turned his people loose and had them go hunt, particularly in those days, other people's projects, and find any on which they could be useful. It was John who was responsible for the existence of the economics division and the social science division, although Olaf Helmer told me he, Olaf, invented the idea, which I believe. But John carried it into being, and it was John who selected Charlie Hitch as the first head of the economics division. John told me--it was a John Williams story, whether at first hand telling me or second hand--that he selected Charlie Hitch to be the head of the economics department because when he was in Arizona Charlie was the only economist he'd known.

I told that story to Harold Lasswell, and Lasswell--you know, fine lawyer, philosopher--Lasswell said that's a John Williams story, not to be believed. But Charlie left RAND to go to the Department of Defense, essentially as comptroller, whatever his title may have been, and then president of the University of California and then president of the Resources for the Future. But like John, Charlie sent his people into all corners of RAND to stick their noses into everybody's business without waiting to be asked.

The social scientists under Hans Speier, the bulk of them were situated in Washington where I suspect they felt that the power and influence were where they could, you know, be of the most use, and I had the prejudiced view that their view of themselves was, the guru on the mountaintop who waits to be asked important questions and then gives profound answers, while writing his book. But as I say, Charlie sent his people into all corners of RAND, as did John, so that if you were to wander through RAND and find a project that was--in which you could be useful, and which you'd be invited to join, which interested you. So John didn't specify, you work on this, you work on that, and so on.

Collins: Going, back to your comment about the difference between John Williams' approach in the mathematics department and Charlie Hitch's approach in handling economics and Hans Speier on the other hand, was that partly a function of geography, or just disciplinary training or inclination?

Specht: Well, disciplinary training to the extent that the economists studied enough mathematics and statistics and were able at least to talk to engineers and physicists and the like. Specht-5

You realize all of my comments are prejudiced, of course. The social scientists, political scientists they were rea~ly ~ut they were called social scientists; in universities econorn1cs 1s one of the social sciences, but the social science division of political scholars does not normally have that much training to be able to be at horne with physicists and engineers as the economists did. Now, interdisciplinary work is often a swindle, in the sense that, you know, the engineer retires to his tent to write his chapter, the economist retires to his tent to write his chapter; but I was talking a few years ago with an economist who had worked on problems in health policy with consultant physician Bill Schwartz of the Tufts Medical School. The economist told me that on the final report every sentence was argued out, discussed by all of the authors together. This means that the economist has to learn enough medical science and health policy to talk to MDs and the MD has to learn some economics. It's a painful but useful experience for everybody.

Collins: Was there a mechanism like that at RAND on the interdisciplinary projects here in the first decade or fifteen years of activity?

Specht: Well, I don't know that what the economist was describing was really anything formalized, but merely a practice. That's the first statement. Second, the very first RAND report is "Preliminary design of an experimental world-circling space ship" (SM-11827, May 2, 1946)--the study. This was written when RAND was part of Douglas. This study is impressive by the large number of experts involved and by the fact that it was not only an engineering design study--should you have a one­ stage rocket, a two-stage, etc.--but it speaks about what you could do with it in science. It speaks about what would be the impact upon this country if the Russians got there first, as they did. And incidentally, Joe Krieger, of the physics division, predicted the launching of Sputnik and was off by only seventeen days. He made the prediction in two senses: first, he had enough knowledge of the Russian program to know they were at the stage at which they were capable of launching an earth satellite; and second, he picked September 17, 1957, the 100th anniversary of the birth of K.E. Tsiolkovsky, the founder of the science of astronautics. (See RM-1922, "A Casebook on Soviet Astronautics, part II," F.J. Krieger, June 21, 1957, page 10. Sputnik was launched October 4, 1957.)

You were asking about interdisciplinary work. The physicists had special clearance problems because they worked on Atomic Energy Commission problems, beginning in 1950. So they didn't tend to mingle as much. In fact, there was a locked door and additional security and so on. The majority of the social scientists were in Washington. But the rest of the projects looked for people from all disciplines who could be useful. Specht-6

For example, Ed Quade and I, in the Mathematics Division, worked with Engineering Division members Gene Root and Dick Schamberg on what was called the multiple-strike study--a comparison of airplane systems for strategic bombing.

Collins: You mentioned that Williams gave you the directive, either explicitly or implicitly, to go out and hunt up things to do.

Specht: Be useful. Sure.

Collins: What do you recall finding of interest that you wanted to become involved in?

Specht: Well, there were many things of interest, you know, that I found interesting, but unlikely that I had any talents or information or skills to be useful on throughout RAND. For example, in 1955 I had the privilege and pleasure of working on tactical air problems with Ed Barlow and Norm Peterson from the Engineering Division, Bernard Brodie from Social Science, Sam Eastman from Logistics, from Physics, and Andy Marshall from Economics. But then at some point I became John Williams' gofer, administrative assistant, second in command, whatever, and so that I had a fair amount of nonscientific activities, sometimes interesting. RAND at some point had a connection with the graduate school at the University of California at Irvine, and I was the gofer for that.

The Philippine Government sent three of its people to work and learn at RAND: Ramon Katigbak, Lieutenant Colonel Gregorio Vigilar, and Ronnie Zamora. They spent several months over here, learning by doing--joining the RAND research program, finding some corner of it that was relevant to their interests, and working on it and on their problems. I don't remember now what they did, but I was their den mother.

Collins: The purpose of which was essentially to introduce them to approaches to analyzing policy issues?

Specht: Sure. So that there were a good many administrative nonscientific research jobs like that working for John, and by now I'm compressing time so I would have to look at the index to have any idea of what happened this year or this year or this year.

Collins: Did your assuming this position come about because of the increased size of the staff, increased managerial demands upon John, or was it something that worked out because you two had an easy working relationship?

Specht: Probably just the chemistry of John, I don't know. I assume. We may never know without being able to ask John. But he was an incredible man. At one time he had a souped-up Jaquar Specht-7

and thought he could lift the record for whatever class that car would come in, but somebody lifted it higher than he could go.

The Saturday Evening Post had a series of articles, with authors including C.P. Snow; these were later published in two hardback volumes, Adventures of the Mind (Alfred A. Knopf, 1961). I think the best essay by far is John Williams' "The Small World" (Saturday Evening Post, August 6, 1960; the second volume of Adventures of the Mind; RAND Paper P-1772-RC, May 28, 1959). John had written an earlier version of this in response to an invitation from the National Council of Teachers of Mathematics ("Some Attributes of the Changing Society," RAND Paper P-1579-RC, December 15, 1958).

The Post took "The Small World" and edited it. All of John's short powerful sentences became long, dull and pedestrian. I volunteered for the job of persuading the Post that they were wrong, and I went through analyzing sentence by sentence, and I had to be so clear that anybody it would say, "That's the only way it can be." They went back to the original in almost all cases. In a few cases they'd improved on John. When Sports Illustrated was about to begin publishing, they mocked up a dummy issue to show the potential advertisers, and for one of the articles in it they ran the paper of John's on television wrestling. Real lovely essay.

John was a disciple of Warren Weaver's. Warren Weaver was a guide to him, an I'm sure it was John who was responsible for RAND getting into , that I mentioned to you earlier, in the two fields of machine translation and in , which was an attempt to make the simulate in some sense a human, whether satisfactorily or not. (See, for example, "Studies in Machine Translation--Survey and Critique" by H. P. Edmundson, K.E. Harper, and D.G. Hays, February 25, 1958, RAND RM-2063. Also "The Logic Theory Machine: a Complex Information Processing System" by A. Newell and H.A. Simon, July 12, 1956, RAND P-868.)

Some years before this work in computer applications, it was John Williams' people who in 1949 began to build RAND's own computer. They were inspired by the von Neumann-Goldstine machine at Princeton, and made their own modifications. I think it was John Williams who was responsible for the acronym and the name of this computer: JOHNNIAC, for John von Newmann's original high speed numerical integrator and computer. The JOHNNIAC was one of the first "modern" . Its capabilities were much greater than those of the punched-card machines then available on the commercial market. When magnetic cores replaced vacuum tubes as memory elements JOHNNIAC became, in 1954, the first operational computer with core memory. (See the discussion and pictures on page 10 of the anniversary volume "40th Year, the RAND Corporation.") It was Cliff Shaw who wrote the first friendly computer language, JOSS--for JOHNNIAC open Shop system. Specht-a

Open Shop in the union sense means that anyone can work, union member or not. Open shop in the RAND computer sense meant that anybody could use the computer, expert or not.

Collins: Let's examine a little bit some of these prominent consultants who worked with the mathematics department. I'm very much interested in Warren Weaver's role in working with RAND, and whether you had any contact with Weaver.

Specht: I don't believe Warren Weaver was a RAND consultant--he was John Williams' friend and inspiration. But, speaking of consultants, not just to Mathematics but to all of RAND, if we skim the Author Index of the "Index of Selected Publications of the RAND Corporation, 1946-1962 11 we see a number of names that my unreliable memory says were consultants, including Louis Alvarez, , Ed Beckenbach, Murray Gell-Mann, Abe Kaplan, Steve Kleene, Tjalling Koopmans, , Oskar Morgenstern, , Jimmy Savage, Martin Shubik, Herb Simon, Edwin Teller, and John von Neumann.

A story, apropos of nothing. Once John Williams and Fred Mosteller, as a diversion, studied the problem of the thick penny--how thick does a penny have to be before it's equally likely that it will land heads, tails, or stand on edge? They made a model and put it on the computer and ran off numbers, and I'll invent a number because I don't remember the ratio of thickness to diameter. Let's say .357. And then because you never know whether a model or any scientific analysis really fits the real world, they went down into the machine shop, which was part of John's domain, and took a cylinder and cut off, made pennies of various thicknesses and spun them. Von Neumann was on one of his periodic visits to RAND, and as I said, to the physics department. John was walking him out to his taxi and John was telling him about this problem and about their program, and their computer run, and von Neumann's interest apparently wandered off somewhere else. And as he got in the taxicab he said, ".357." "You son of a bitch!" said John.

Collins: Did you have any involvement with Von Neumann's activities?

Specht: Never laid eyes on von Neumann.

Collins: But your sense was that his Consultancy was with the physics department, not the mathematics department.

Specht: I think so. It would have been sensible. After all, he made a major contribution to modern physics in his book on the foundations of quantum mechanics--a book written when he was very young. And this reminds us of one of the most incredible episodes in the history of human thought: much of modern physics was created by men in their twenties--Bohr, Dirac, Heisenberg. Specht-9

Einstein was in his 20's when he published the special theory of relativity. But I never laid eyes on von Newmann.

Collins: At least in some accounts of RAND's history, one of his fundamental areas of contribution was to game theory work here.

Specht: The RAND Index, 1946-1962 lists two papers on game theory by von Neumann. I think his influence on RAND work in this area came rather from the monumental work, Theory of Games and Economic Behavior by John von Neumann and Oskar Morgenstern, published in 1944, just two years before RAND began at Douglas. I never worked in game theory, but there was a great deal of activity in it at RAND, because John thought that RAND could contribute significantly and usefully. This tells you something about Frank Collbohm, the president of RAND. Frank put up with many things that I suspect he did not think too highly of. And John Williams was able to devote a significant amount of RAND effort to important work in games theory and without being sure what would eventually come out of it. Olaf Helmer was one of the in-house leaders in this work. Olaf has two Ph.D.s--one in mathematics from the University of Berlin and one in logic from the University of London. He's living in Carmel now. You'd have to ask somebody who knows more than I do what impact this work had on work in other fields. I do not know.

Collins: Another way in which outside expertise was brought into RAND in this period was the occasional conduct of so-called summer studies. Were you involved with any of that kind of activity? Do you know what I'm referring to?

Specht: If you run through the pages of Brownlee Haydon's The RANDom News, the in-house news journal, you will find a number of courses, conferences, symposia that provided RAND with opportunities to teach and to learn from visitors (and not merely in the summertime). For example, Ed Quade's course on systems analysis was first given to (and by) RAND people: it was given to Air Force people for the first time in 1955. In 1956 members of the State Department took part in a political gaming exercise that dealt with the problems of Western Europe. Also in 1956 RAND offered a conference and course on computing techniques. In 1957 more than forty Air Force logistics specialists took part in a logistics conference. And, again in 1957, the Electronic Warfare Committee held its quarterly meeting at RAND: participants included experts from Air Force, Army, Navy, RCAF, RAF, and industry. In 1959 the RAND Symposium on Protective Construction had six hundred participants from u.s., Canada and other countries, representing the professions of architecture, physics, geology, and others. Also in 1959 the program of the symposium on mathematical programming included more than forty experts from business and academic institutions. And so on. Specht-10

TAPE 1, SIDE 2

Specht: About the history of RAND. I mentioned earlier The RANDom News, which appeared monthly (or less frequently) beginning, in its unclassified form, in 1955. The first formal history of RAND was published in 1963 and celebrated The First Fifteen Years. I saw a preliminary version of this small book. I started to rewrite it; I got John Williams interested, and the two of us re-did the history. The next formal history (by RAND) appeared ten years later; RAND: 25th Anniversary Volume. This includes the kind of material found in earlier annual reports-­ descriptions of research programs, administration, financing, and publications. But it begins with a sixty-page Sampler--two-page excerpts from RAND books, reports, and papers, from "The World­ Circling Spaceship" of 1946, to "Trains Faster Than Airplanes" of 1972. The Sampler is followed by a set of essays on topics of RAND research: education, nuclear deterrence, Soviet strategic policy, and others. Now descriptions of programs are normally written by committees, edited endlessly, and hard to read and stay awake. But in the Sampler I used a prose style that has the unique advantage that it cannot be edited; the original language of the book or report or paper, reduced to a two-page "sample" by the strategic use of ellipses. The RAND annual reports, beginning with the 1974-75 volume, use this Sampler structure for the initial fifty or sixty pages, though, of course, using contemporary publications. A very attractive history-plus­ description-of-today•s-RAND has recently been published: 40th Year, The RAND Corporation. (See pages 104-106 for a bibliography of ninety-three articles about RAND, published from 1951 to 1988. See Also Bruce L.R. Smith's book, The RAND Corporation, Case study of a Nonprofit Advisory Corporation, Press, 1966.) Collins: Right. Specht: In spite of the histories I have just mentioned, RAND, in some ways is deficient in its corporate memory. I give you one example. Some years ago newspapers were reporting concerns that visitors to nuclear tests--at the Nevada Proving Ground or in the Pacific--may have been exposed to nuclear radiation that increased their chances of coming down with cancer. Gus Shubert, very properly, wanted to know who from RAND had gone to nuclear tests. For the years after Arnold Mengel started publishing RAND items, this periodical of RAND activities would report who was traveling where and would answer Gus• question. For the earlier years, however, information was skimpy. The first obvious place to look is the file of visit security clearances. We didn't find any surviving files. The next obvious place to look is the file of travel orders. Same result--no surviving files. The only Specht-11

source of information we found was the chronological file of carbon copies of outgoing correspondence. Not completely satisfactory because we would find a letter from RAND saying, in effect, "We'd like to send the following people. If you have insufficient space, please deal from the top of the deck down as far as you can go." Not totally satisfactory because we found no file of responses to such letters. As I said, RAND's corporate memory is not in the best shape.

To change the subject from history to building design and its effect on interdisciplinary research, we can see over on that wall a diagram of the RAND building with its eight patios. It was John Williams who in 1950 wrote a document for the RAND staff pointing out that if you had, say, a corridor with offices opening into it on each side--a straight-line corridor one thousand feet long--and if you formed that corridor into a square, enclosing a single patio, then the maximum distance between people is halved. If you shape the corridor into a grid around more patios, then you cut the distance between people even more (although the distance decreases more slowly as the number of patios increases). This means that people are more apt to run into others, including those from other disciplines. And there is another interdisciplinary effect of this checkerboard structure--there are so many different ways to get from one place in the building to another that, if you're not totally a creature of habit, you run into different people as you follow this route or that route. We moved into this new building in 1953, and John took half of the mathematic group and scattered it across all the rest of the building--for mutual education. (For excerpts from John Williams' document and pictures of the building in process and completed, see pages 16, 17 of 40th Year, The RAND Corporation.

Collins: Did you have much contact with Ed Quade and Paxson as they were beginning to mature their ideas about systems analysis?

Specht: As I mentioned earlier, Ed Quade and I worked with engineers Gene Root and Dick Schamberg on a systems study comparing airplane systems for strategic bombing. This--the multiple strike study--was a sequel to Ed Paxson's bombing study. I'm not sure what the title of the report actually is.

Collins: That's it.

Specht: I'd known Quade at the University of Florida. I'd not known Paxson before I came here. But they were both impressive characters in many ways.

Collins: What is this notion of developing a formalized understanding of an approach to dealing with systems? Did you get into the technicalities of that issue with them? Did you participate in working through that problem? Specht-12

Specht: When we were talking earlier about courses, conferences and symposia, I mentioned Ed Quade's course on systems analysis. It was given first to RAND people. Then in 1955, and again in 1959 RAND offered an intensive five-day course, "An Appreciation of Analysis for Military Decisions" to armed forces officers and civilians. This material is published as a RAND book, Analysis for Military Decisions, edited by E.S. Quade (Rand McNally, 1964). As the Preface says, "Designed primarily for decision makers and not for analysts, this course did not attempt to teach but to point out the weaknesses and possible abuses, as well as the effectiveness of an analytic approach to long-range military planning." Because you spoke of Ed Paxson, I mention that one section of the book is Ed's "An Introduction to the Lunar Base Problem." I wrote the chapter on "The Why and How of Model Building."

Collins: Yes. Let me put the question a slightly different way. For example, when you were working on the multiple strike study, were you consciously examining the methodology as you went along, this notion of a systems study? Or did it seem more natural than that?

Specht: Probably not in any sense of formally thinking about whether in what sense could the methodology be advanced, or in what sense would it apply. I was more I think living in an atmosphere in which the idea of--well, for example, in systems studies, no systems study, no scientific study of any decision problem is ever divorced from value judgements. But in many studies throughout the world, the value judgments are never apparent. They also may never even have realized--you know, they're sort of build in, embedded in one's choice of criteria, for example. You're going to be efficient--what are you going to be efficient about? What are the criteria? Well, the choice of criteria is a matter of values, not a matter of science. But the business of being explicit in setting down, as far as one could, you never know to what extent you've divorced yourself from prejudice, for example, but nevertheless, to try to be explicit about what are the criteria, what are the bases for judgment-­ that was a kind of, in the atmosphere, you know, sort of in the wind, sort of, that's what people were doing I think, as opposed to considering formally, now, how can I apply this systems analysis? I think RAND probably developed systems analysis without realizing until after the fact when as I say they created these courses in which the index again must have.

Collins: I've seen some of the literature. To what degree do you recall being aware during these early studies that you were involved in, your sensitivity to these value issues? Is that something that was recognized then or did it come a little bit later?

Specht: Well, I'm not sure that anybody thought about this matter of separating values from science, let's say. But just Specht-13

the matter of explicitly setting down, these are the criteria by which we'll judge which is the best system, that was just kind of the way one did things, and I don't think anybody thought about any philosophical issues about values and judgments as opposed to science and analysis.

Collins: That kind of consideration or discussion did come into play later when RAND began to assess the ways in which systems analysis could be genuinely--

Specht: Well, I suppose the first time that you tried to sit down and say: this is systems analysis, were these groups of lectures, seminars and discussions. I suppose the moment that you start doing that you find yourself facing these questions. Quade writing and editing Analysis for Military Decisions (and the lectures that preceded the book) must have faced these issues.

Collins: Right. What became the balance in your own work between the, for lack of a better word, managerial activity you were doing as John Williams' second or assistant, and research? Did you try to keep some kind of balance between those two roles, if you will?

Specht: I did try to remain active on RAND research projects-­ tactical air studies, gaming exercises, etc. But as I may have mentioned to you, I took John's place for a year as acting head of the department or division.

Collins: That would have been approximately when, do you recall?

Specht: John resigned, briefly, in 1957. At this time the Mathematics Division included a small group of people who had been involved in the work of the System Development Laboratory, work that led to the creation of soc, System Development Corporation. Before getting to John Williams' resignation, I need to say something about the history of this System Development Laboratory--then Division, then Corporation.

A group of psychologists in the Mathematics Division, headed by John Kennedy, wanted to study how people and machines work together under stress--how the people set up their own informal patterns of information and direction. The psychologists and a group of computer experts simulated an Air Defense Direction Center where information on an enemy raid comes in from radars, has to be analyzed and evaluated, and then orders are sent to a fighter group. Initially they used college students as guinea pig crews. The Air Force was impressed by the productivity of these student crews during simulated air raids and saw that this simulation would be a useful training tool. The work grew in size--first into the System Development Division of RAND, formed in 1954, and then into the System Development Corporation (SDC), spun off in 1957. John Kennedy did not join the new corporation Specht-14

but left RAND in August 1957 to become Dorman T. Warren Professor of Psychology at Princeton, and later Chairman of Princeton's Psychology Department.

As I said, in 1957 the Mathematics Division, headed by John Williams, included a small group of people, the Psychology Research Laboratory, who had been involved in this system simulation work but who did not choose to go with soc. The president of RAND had a financial problem which he solved by cutting the budgets of all the departments by a fixed percentage; in addition, he cut to zero the budget for the Psychology Research Laboratory; finally, he subtracted from the mathematics budget a number of dollars without specifying any name, but to the nearest dollar it was the salary of one of our more distinguished researchers. John Williams resigned as head of the Mathematics Division. The president said he allowed as how he assumed I would take the job, and I allowed as how it wasn't clear that RAND was such an interesting place to work. I set out to think of a price to charge the president for taking the job. All I wanted was dollars to save jobs. I'm slow of thought, and it was the next day when it occurred to me that it would be much more fun to persuade John Williams to set his own price for staying. John told the president, "You've heard Specht's price, now hear mine"--which was, in addition to the dollars, two other things: first, any time John had something to communicate to the Board of Trustees, the president would send it on to the Board; and, second--! think a political stroke of genius--the next time a similar budgetary problem arose, the president would solve it by cutting all department budgets by a percentage somewhat greater than necessary, and then use this "surplus" to restore funds to departments where the cuts threatened projects that the president valued. As I say, I think president Frank Collbohm's positive prejudices were extraordinarily good. So the Management Committee had, not a mortgage burning, but a resignation burning. John then spent a year on leave from the division head's responsibilities, and I filled in for him during that time.

A footnote to my sketchy history of the System Development Corporation. soc was originally created nonprofit. Some time later it went profit, and the assets were turned over to the System Development Foundation. Sometime later--after I left RAND in 1979, but was a consultant--the foundation was about to distribute money. RAND wrote a pitch as to why some of that money should go to RAND. Someone in RAND had the weird idea that I should read this draft pitch and comment on it. Now this pitch was written in the old oaken bucket syndrome--give RAND money to restore RAND to ancient glory. I wrote that, speaking as a foundation executive, I would give at most seventeen bucks on the basis of this pitch--who cares about ancient glory? The question is what are you going to do interesting next year? I suggested which way they bend their pitch. I haven't the foggiest idea what happened to the pitch, but in fiscal year 1982 the RAND Endowment Fund (set up in 1974) received a ten million dollar Specht-15

unrestricted grant from the System Development Foundation (see the RAND Annual Report, 1982-83, pages 168-169). How much more RAND would have gotten if I had stayed out, we'll never know.

In the 1950s RAND was organized into divisions, some of which were broken into departments. For example, the Engineering Division comprised Departments of Operations, of Aeronautics, and Electronics; the Economics Division was made up of Departments of Economics, Cost Analysis, and Logistics. In 1960 the division structure was abolished, the departments became autonomous, and the former division heads now formed the Research Council. Jim Digby and I were on the staff of the Council, were the Research Council gofers. Speaking of gofers, once we had a collection of visiting Russians, Academy of Science types, hosted by Nancy Nimitz, and I was among the RANDites who were there, probably because I'm a friend of Nancy's. We from RAND each had to explain how we fitted into the world, what we did and why we were there and so on, and I made the dumb mistake of explaining with a joke, which ain't translatable: I said I was the president's gofer. I saw them looking puzzled, and their translator moved his hands saying something about a little animal. So as I say, I was "the president's gopher" for some years.

Collins: You mean Frank's? Specht: No, this would be Harry Rowen.

Collins: Okay. I'd be curious to have you elaborate on what you thought were Frank Collbohm's positive prejudices, and then his negative ones.

Specht: On the positive side, John Williams got Frank Collbohm's support for an extraordinarily broad program of research in the Mathematics Division--work, for example, in linear programming, , network theory, the mathematical theory of control processes, game theory, and in biological and medical studies. And this work was supported even though some of it did not have as much of a direct application to Air Force problems as some may have expected.

On the negative side, Herman left, of course.

Collins: Herman Kahn, you're referring to.

Specht: Yes. And I've always assumed, without really knowing, that Frank felt that--Herman had carried on the civil defense study, a very massive study, and Herman was a missionary and was preaching this gospel and reporting his study around the country. I've always assumed that Frank felt he'd sort of lost control, but I don't know. We can't ask Herman now. LIFE Magazine once had a picture of Herman in profile on a double page, and he's on the left side of the page, but his great belly slurps over into the right hand side of the page which makes it hard to remember Specht-16

that Herman weighed 140 pounds when he was in Burma. That's all I know of Herman's Burmese connection, if any.

Collins: What are the two morals about Frank's approach to the organization that you'd extract from those two examples?

Specht: My impression was that Frank did not value as highly as he should have John Kennedy and some of Kennedy's associates whose work helped lead to the eventual development of the System Development Corporation. My own interpretation, which may have nothing to do with reality, was that Frank's positive prejudices, I thought were first rate. The things that he favored and supported or the things that he tolerated because he appreciated the people, as he must have appreciated John, he brought him in.

Collins: John Williams.

Specht: John Williams. Frank's negative prejudices in the case of John Kennedy and some of his people is a better example. I wouldn't say that he had been prejudiced against Herman Kahn until possibly, I'm conjecturing here, he felt he had lost control of Herman and events, that Herman was preaching his gospel of civil defense around the country. But in John Kennedy's case, as I told you, his budgetary solution eliminated the one group of survivors from Kennedy's Psychology Research Lab. In addition, he cut out the salary of one senior RAND character without naming any names. He just subtracted the number of dollars which, to the closest dollar, was Olaf Helmer's salary. So as I say, I though his negative prejudices tended to be mad. His positive ones were extraordinarily good.

Collins: I guess in both instances he felt he was doing what was best for the corporation.

Specht: Oh, I'm sure. But for the most part he operated in a hands-off way, you know, for most of the time, it seems to me leaving things to Charlie Hitch or John Williams or the people that he'd chosen and trusted.

Collins: I guess it kind of raises a possible interesting question. Did perhaps Frank or others in the organization have a certain view of what the "RAND type of person" was? I mean people like Williams and Hitch fit in, but people like Kennedy and Helmer didn't fit in as well?

Specht: He must have. I haven't the foggiest idea of what aroused him in Olaf's case. It was Olaf, as I mentioned, who told me that he had invented the idea that RAND should have an economics and a social science division, and it was John Williams who made it happen.

Collins: You mentioned Herman Kahn's role in giving briefings on his particular project, and briefings were clearly an interesting Specht-17

element of RAND's presentation of its research product to its clients, primarily the Air Force. Did you ever get involved in this briefing activity substantially? Was that something that came out of some of the studies that you were involved in?

Specht: Not quite a briefing. I gave a lecture at Paris, at the NATO meeting, AGARD, Advisory Group in Aeronautical Research and Development. The lecture was on some aspects of systems analysis, as was my part in Ed Quade's course in systems analysis. I took part in briefings on the multiple strike study, a successor to Ed Paxson's first bombing study. And some others. But I was not an important briefer in any sense.

Collins: You mentioned the presence of the Psychology Research Lab in the department. I guess that offers the question, what was the composition of the department in this early period? There were psychologists, there were mathematicians--any others?

Specht: The division, those were the days before the Research Council and the abolishment of divisions--the division in mathematics' case was initially made up of two departments, mathematical analysis and numerical analysis (later called Computer Sciences). But the individuals had varied backgrounds. While they were principally mathematicians and engineers, I mention that John Williams did his undergraduate work in astronomy; one of Olaf Helmer's two Ph.D's was in Logic; Paul Armer, the head of the Numerical Analysis Department from 1952 on had a degree from UCLA in meteorology. Later, Stan Azen was an accomplished musician and composer; Kathleen Archibald was a sociologist.

The psychology research group, John Kennedy and his psychologists, first entered RAND and were part of the social science division. Then, for whatever reasons, they decided that they would feel more comfortable in mathematics. (This would be about 1952 or 1953.) Maybe we were wilder and more loosely structured. Who knows? But John invited them in to the division as a third entity, the Psychology Research Laboratory, so there was Mathematical Analysis, Numerical Analysis and the Psychology Research Laboratory. The last group was certainly interested in artificial intelligence and in the machine translation that I mentioned to you earlier on. But their major interest and involvement was with this simulation of groups of people working under stress, and they were raising the question, how do people set up their own informal lines of communication and command, as opposed to whatever formal structure normally exists. People working together find themselves inventing their own channels of command and communication, and that was what they wanted to simulate and study, and it was for that that they began this Air Defense Direction Center Simulation which led to System Development. Specht-18

Collins: Right. Since this group was a part of the mathematics division--

Specht: Until it grew large and then it became its own division, System Development Division in 1954.

Collins: Did you and John have any role in effecting that transition from something that came under mathematics to an independent entity?

Specht: I don't remember any of that. Although I think it came automatically and on its own, in the same way that its separation from a division of RAND to a separate corporation because it became so large, and secondly, by the time they were this well along they understood exactly what their mission was, what their goal was, and that made them unfit to be part of the mathematics division. You know, they saw they were driving straight ahead toward what they wanted to achieve, and quite properly they knew what they wanted to achieve. Then they were lager in size, so that that made them, I think, kind of automatically, first, no longer part of the mathematics division, and a little later, as they grew and grew, no longer a part of the RAND Corporation, but Systems Development Corporation.

Collins: Right. Another area of activity within the mathematics department, I don't know where it was located, was investigations into dynamic and linear programming.

Specht: Dick Bellman began writing RAND publications on dynamic programming and on linear programming in 1951-52. They were both extraordinarily prolific. If you look at the RAND Index of Selected Publications, 1946-1962, you will see that the subject index takes almost ten pages to list the more than four hundred publications--not to mention later publications from 1963 on.

Collins: And they were both in mathematics?

Specht: They were both in mathematics.

Collins: Where did they fall in that framework you were just describing?

Specht: Well, not in the numerical analysis but in the mathematical analysis. That was itself a collection of all kinds of characters from people trained in mathematics to--well, sociologist Kathleen Archibald had been Miss Canada. John hired her because he was impressed by her, as he should be. She told John when she came in that Miss Canada was not selected on grounds of beauty, but she thought that she would mention that less and less as time went on. But John looked for people who were interesting, no matter whether they were in statistics or mathematical analysis or economics or sociology or what. Specht-19

TAPE 2, SIDE 1

Specht: I told you, I think, in January, the story of Sam Genensky whom John hired, and who changed the world slightly--he created the Center for the Partially Sighted. Now, Sam was trained in mathematics, and I think probably applied mathematics at . But later on when Sam spent more and more of his time on the partially sighted, you know, that was fine with John, and, as I say, sam has changed the world somewhat. Sam came into RAND with no vision in one eye and 8/500 in the other, but not handicapped. His friends at RAND built for him a system which you may now have seen, it's in libraries around the world, even private individuals have them, a closed-circuit television system. Imagine a television camera with a zoom lens looking downward at a platform that slides on roller bearings easily to the right or left, on which you put the book you're reading or the pad of paper you're writing on, and so that it slides easily to the right or left to go across a line, and so the zoom lens of the television camera is looking at it, and then there's a twelve inch black and white TV set on which with the zoom lens you can make the characters as big as you need them to be. And as I say you slide the platform so that you go across a line. You would not at that time, maybe still today, I don't know, get your eye up close to a twenty-three inch color set, the high voltage electrons hitting the face of the tube lead to soft x-rays coming off the face. Whether they've filtered those out by now, I do not know, but we measured, we tested the twelve inch low voltage black and white set and no measurable radiation. And Sam, being essentially the Rabbi of RAND, doesn't rest till he finds somebody to manufacture it. The first manufacturer was Visual Tek, now called VTEK. It's in Santa Monica. Now there are at least two, maybe more, manufacturers. Sam gets no money from it. RAND gets no money from it.

I think I told you this story in January, that I heard from Sam only in the past year, that when John Williams was in Brown University trying to persuade Sam to come to RAND, Sam had just finished his doctoral degree at Brown, John must have brought with him personnel department forms to fill out and one of the questions was, Do you have any disabilities, and Sam asked John how he should answer that, and John paid no attention to him and went on talking. sam asked him again, John paid no attention, and finally Sam cried, "John, John! How do I answer this?" Said John, "Walk over to the window." Sam walks over to the window, stands in front of it, looks out for a while. John says, "Now come back." John went on talking about something else. "But John," says Sam, "How do I answer this about any physical disabilities?" John says, "Well, when you stood in front of the window I didn't see any light coming through. You don't have any physical disabilities." Specht-20

So Sam while at RAND created the Center for the Partially Sighted that was part of santa Monica Hospital. Sam's work at RAND was supported in part by RAND's own funds--it was partly RAND-sponsored research. But John was happy to see somebody do useful and interesting work, no matter in what way. As I say, Sam has changed the world slightly. The Center has since left Santa Monica Hospital, has its own quarters in Santa Monica, 720 Wilshire. Sam turns out to have a great talent for raising money. He has to. And one of their fund raising devices that gets maybe $85,000 or so a year, they have a big bash in Beverly Hills and give an award to the man of Vision of the Year or the Woman of Vision of the Year, and it's been, I think, Carol Burnett, Lena Horne, Sammy Davis, Jr., Sid Caesar and like that. And they concentrate on showing people, well, first that they're not--they never use the term "partially blind" but rather "partially sighted," you see. That tells you something about Sam.

Collins: As John's second, did you ever participate in the so­ called Management Committee meetings, the meetings between the department heads and Goldy and Frank?

Specht: In the year when John first of all resigned, and then canceled his resignation but then essentially went on leave for a year while at RAND, I had his job so I sat in on the meetings. But I don't remember anything interesting.

Collins: In terms of how this group worked to set policies and discuss projects that were under way--

Specht: It may be that it's a trick of memory and dismissing anything unpleasant. Who knows? That's intended as a joke. But Goldy, I thought was an extraordinarily fine man and administrator, Dick Goldstein.

Collins: How did you like the experience of essentially assuming the role of department or division head and the kinds of responsibilities that that entailed?

Specht: Well, I find something to be amused by and enchanted by in any place, so this wasn't all that different, you know. Nothing particular either glorious about assuming the job or painful in leaving it. I still had the same people to move around with.

Collins: But I assume, as acting head, you had to be more concerned about what people were doing, what projects were under way, personnel decisions.

Specht: Yes. I don't remember any that I made or whether they were good or bad. Specht-21

Collins: One thing that you've alluded to a couple of times that's intrigued me was the sense, at least in the early period, that many of the studies were done in an intellectual framework that could be characterized as still fighting World War II.

Specht: I wouldn't characterize all of the early studies that way because there were many that were engineering studies on the design of , for example. Will Kellogg and Stan Greenfield got a plaque later from the meteorologists for inventing the meteorological satellite. I think what they did was to show what you could see from three hundred miles up and how you could relate that and tie that in with meteorological measurements on the earth's surface. But in the war games, I spent a summer at the Army War College on a war game, on an exercise. But the military exercises were, you know, map studies over the map of Europe, and they weren't dropping atomic weapons, they were manning bombers on bombing missions and fighters against bombing missions, you know, and like that. I think that's far closer to World War II than to World War III. Unless we're lucky in World War III, in which case that showed great foresight.

Collins: Right. One of the things that RAND is obviously quite noted for is its involvement in strategic studies and especially as you move into the missile area.

Specht: Herman's study of civil defense was very much World War III. Or his image of how one guards against the nuclear weapons, you know.

Collins: I guess what I'm trying to get is a kind of a question of ambience as the character of the studies changed more from a World War II mentality more to one that adapted itself to the new requirements of the missile age.

Specht: Well, about World War II, as I say, I was thinking only of some of the war games. The first systems study was the Paxson bomber study, and it was really an aircraft design study: what do you do in terms of speed, range and altitude. The answer to that is of interest today as well as it was forty years ago. But in doing this study, as part of the study they carried out these simulated exercises of let's say bombing missions. And for the military context in which they operated those that were not an atomic weapon, that looked more like World War II. But that was only a tiny piece of the whole schmear.

Collins: This is perhaps too broad and too slippery a question to answer well, but what I was trying to get at is the sense of how the notion of gaming, the Kriegspiel, these kinds of things that were an intimate part of the RAND culture during this period of the fifties, changed over time or the degree to which you were aware of them and your reaction to them. Specht-22

Specht: Well, first of all, the gaming from Kriegspiel to the war game exercise I suppose influenced research in the sense of people thinking about trying to simulate the environment in which these things might operate. Whether that simulation was like the psychologists' later Air Defense simulation, or whether that simulation was a map exercise on the map, or that simulation was a computer exercise in which you sort of formalized some assumptions about attrition and fighters attacking bombers or whatever--so that in all those cases, you had the sense, I suppose, of trying to simulate, imitate, adapt some aspect of an environment, of the actual operating environment or whatever.

Lloyd Shapley was the chief Kriegspieler, I think, and that was a diversion, a lunch time operation. And he was a player of "Go" which never had the great attraction of Kriegspiel. But I remember how ignorant I was, when I first saw , who led RAND's work in game theory, and he had a glass bottle filled with stones, and I asked him about it and he said they were his Go stones, and I was sufficiently ignorant that I translated it as "gallstones" and was marvelously impressed by it. So I suppose the impact of Kriegspiel was largely those same people who were interested in the mathematical theory of games, and in both cases you're considering the adoption of a strategy by one side or another side.

They along with the economists thought that they could use the theory of games to simulate conflict of all kinds, whether corporate or military, and where the war gamers, whether over the map or the people in the simulated Air Defense Direction Center, were trying to do a simulation of a different type, not formalized in the mathematical computer type programs. How effective either group was, I'm not sure, you know, but it was in both cases a magnificent ambition, to try to deal more realistically with conflict than had been the case before.

Collins: This was--clearly there seems to be a sense in which the support of research, i.e., game theory which at first doesn't seem to have much application comes to have some fruit.

Specht: Sure. As I say, I think it's a tribute to John and to Frank that this was able to go on for years and to be tolerated and money to be spent and so on. And then to the Air Force, which put up with this, and John's and Frank's ability to persuade the Air Force.

Collins: Right. Do you have any sense of how something broadly theoretical like game theory made the transition to more practical application here at RAND? How did that happen?

Specht: Well, first of all, we're assuming that really happened. I'm not enough of a scholar to know with any confidence the answer to that, as to what was the actual result, effects. I don't know the answer. You know there are books by Dresher and Specht-23

others, but they're not I think answering that question of the effect upon the systems studies. You'll have to ask somebody who knows more than I do to find out what effects they really had. I don't know.

Collins: Another broad kind of question, what do you think were the notable achievements or contributions of the mathematics division or department, whatever incarnation you want to talk about, to the RAND enterprise?

Specht: one was in selection of people. You know, if we looked at who, from here, and moving on outside, a selection of interesting people and important people, it seems to me, moved through, had some contact with the mathematics department. You know, whether it was System Development that came from the early experiments done when the group was a part of mathematics, or work on linear programming, not completely done by mathematics. Economists were in on this, too. Dynamic programming, game theory, simulation, again, the simulation by people as opposed to simulation by computer was broader than merely mathematics. Mathematics included the computers, and I think that, for example, JOSS was one of the first friendly languages. There have been many copies of JOSS. If Fortran had been designed by RAND types it would probably be a much more friendly language than I suspect it is. The development of the JOHNNIAC computer, which as I say was a modification of the Princeton machine, but the installation of individual stations throughout RAND that are wired in to the center--this was back before the existence of the personal computer--and now, well, I have been all day in areas I'm not qualified to say anything about, but that seem to me to be important.

Collins: Well, we want your sense from your point of view and your experiences, the significant areas in which mathematics contributed.

Specht: The development of systems analysis was in part the work of mathematics, but not in a major sense. The Economics Division (today "Economics and Statistics Department") has always been a landmark group within RAND. Olaf told me he invented the idea, but John carried it into fruition of the existence of political science, called social science, and the economics division with RAND, and it seems to me that was important beyond the walls of RAND, because otherwise there might have been a tendency for, you'll forgive the expression, "think tanks" to be merely homes for tired mathematicians and engineers, and this broadening was important. But that was not due to the mathematics division but to, you know, the guys, Olaf and John.

Collins: One area we haven't touched on is RAND relations with the Air Force, and the question of how a research organization like RAND maintains independence of activity in selecting Specht-24

projects and personnel when you have a sponsor like the Air Force.

Specht: Again, this would be something that Goldy, Dick Goldstein, and Frank Collbohm would be I think responsible for, and to their credit that the relations were as mutually supportive as I think they were, and something about which I know nothing. And speaking of relations with the Air Force, from the early years talented Air Force officers have been stationed at RAND, part of the research staff, and for mutual education. Today there are also Army and Navy officers at RAND.

Collins: Okay, I was just curious whether it impacted in any way in achieving this balance on the mathematics department as you recall.

Specht: Well,if it did the department was protected from it, so I do not know. There is a joke, as you may know, that LeMay said that RAND stood for: Research and No Development. But the Air Force relations are something of which I was ignorant.

Collins: One way in which this manifested itself was the mounting interest of the Air Force in having RAND assist with what they called short-term problems, things that needed to be done today or the next day, rather than things that looked to the long term.

Specht: Oh, sure. If I were in the Air Force, that's precisely what I would ask for.

Collins: Did that kind of request impinge on mathematics, as you recall?

Specht: Well, remember that from the first days of RAND John Williams had this mixture of some people sticking their noses into everybody's short-term projects, as did Charlie Hitch in economics, but then in addition John had the Lloyd Shapleys in game theory or the George Dantzigs in linear programming or the Dick Bellmans in all kinds of analysis, who were sort of guarded, protected and given their head to pursue their own passions.

Collins: So really it reflected and depended on personal skills and capabilities.

Specht: Well, and personal passions and prejudices and preferences.

Collins: One thing that RAND did along these lines, I'm not clear on the extent of the activity, was the so-called loan program, where RAND staff members would go and spend a period of time and sit in an Air Force office and work with them on problems. Do you recall people from mathematics getting involved in that kind of thing? Specht-25

Specht: From skimming the pages of RANDom News for 1955-1959 I see that: Arnold Mengel (Mathematics) was stationed in Wiesbaden for several months. Norton Kristy (System Development Division) was stationed at Air Defense Command Headquarters, primarily for liaison. John Springer (Cost Analysis) was on loan to HQ Strategic Air Command. Dick Schamberg (Engineering) was on loan as chief of the Air Proving Ground Command operations analysis office at Eglin Air Force Base. Fred Hoffman (Economics) was RAND's liaison representative in the Plans Executive Office, Directorate of Plans, HQ USAF. Collins: We talked a little bit about the mathematics division connection with universities, both through consultants and you mentioned this graduate program down at Irvine. Can you think of any other ways in which RAND cultivated contact with the universities' mathematics division?

Specht: One connection between RAND and scholars outside comes from the many RAND publications available from several hundred libraries here and abroad; also the many RAND books. There are other, more direct connections. (I do no limit this to the Mathematics Division.) Looking only at one year's issues of RANDom News (1955-56) I see that: Ted Harris (Mathematics) was Editor of The Annals of Mathematical Statistics. Alan Manne and (both of Economics) each spent a year's leave of absence to do research at The Cowles Foundation for Economic Research at Yale. Cal Gazley, Ed Williams, and consultant Julian Cole (all of Engineering) gave a course on hypersonic aerodynamics at UCLA. George Dantzig (Mathematics) was Vice President of the Institute of Management Sciences. Hans Speier (Social Science) was consultant to the Center for International Studies at MIT. Hans also had a year's fellowship at the Stanfr.r.d Center for Advanced Study in the Behavioral Sciences. Specht-26

Dick Bellman (Mathematics) lectured at Princeton and Columbia.

Bernard Brodie (Social Science) was a guest professor at the Graduate Institute of International studies in Geneva.

Peter Swerling (Engineering) was Research Assistant Professor for one semester at the University of Illinois' Control system Laboratory.

Joe Kershaw (Economics) spent a semester as Visiting Professor of Economics at Williams College.

Collins: Were there any kinds of analogous relationships with industry that the mathematics division participated in?

Specht: I don't know of any. But there was another way in which RAND may have had some effect on the world outside its halls-­ through the many RAND alumni, people who worked at RAND and then left to make contributions elsewhere. Since you mentioned industry, let's begin with some of the alumni whose subsequent histories included connections with business, industry, finance. (These notes only carry us up to 1978.)

Igor Ansoff (at RAND 1948-56) was later Vice President, Plans and Programs, Lockheed Electronics, still later, Dean, Graduate School of Management, .

Edward J. Barlow (1948-60), Vice President, Engineering, Aerospace Corporation; President, Instrument Group, Varian Associates. Robert w. Buchheim (1954-63; 1964-67), Chief Scientist, USAF; Executive Director, Research and Engineering, North American Aviation; President, CEO, Southwestern Research Corp., Deputy Assistant Director, u.s. Arms Control and Disarmament Agency.

Alain Enthoven (1956-61), Assistant Secretary of Defense; President, Litton Medical Products; Vice President, Litton Industries; Marriner s. Eccles Professor of Public and Private Management, Stanford Graduate School of Business.

W. Richard Goodwin (1955-57), President, CEO, Johns-Manville Products Corp.

Thomas V. Jones (1951-53), President, Chairman of the Board, Northrop Corp.

Jeffrey C. Kitchen (1956-61), Deputy Assistant Secretary of State; Vice President, Northrop Corp. Specht-27

Albert Madansky (1957-65), President, Data Plan; Professor of Business Administration and Director, Center for Management of Government and Non-profit Enterprises, Graduate School of Business, University of Chicago. James w. Petersen (1953-64, 1965-69), Chairman, Economics Dept., University of Vermont; Vice President, Southeast Banking Corp. Norman c. Peterson (1949-59), Director, Program Development, Hughes Aircraft; Vice President, Management Systems Inc.; Executive Vice President, RAPIFAX Corp.

Robert L. Slighton (1962-74), Vice President, Chase Manhattan Bank; Deputy Assistant Secretary of the Treasury.

John summerfield (1956-62), Vice President, Western Airlines; President, Pan American Airways.

Charles J. Zwick (1956-65), Director, Bureau of the Budget; President, CEO, Southeast Banking Corp.

Other interesting alumni include:

Thomas A. Brown (1961-74, 1977), Deputy Assistant Secretary of Defense.

Charles A. Cooper (1963-67, 1968-70), Minister for Economic Affairs, u.s. Embassy in Saigon; Deputy Advisor to the President on National Security Affairs; Assistant Secretary of the Treasury; Manager, Energy Policy, Exxon Corp.

R. Walter cunningham (1960-64), astronaut; Vice President, Century Development Corp. Joseph o. Fletcher (1964-71), Research Professor of Oceanography and Atmospheric Sciences, University of Washington; Head, Office of Polar Programs, National Science Foundation; Deputy Director, Environmental Research Laboratories, National Oceans and Atmospheric Administration.

W. Lawrence Gates (1966-76), Chairman, Department of Atmospheric Science, Oregon State University.

Henry Geller (1973-75), General Counsel, FCC; Director, Office of Telecommunications Policy.

Bernard Gifford (1972-73), Deputy Chancellor, New York Schools; Controller, Russell Sage Foundation. Specht-28

Victor Gilinsky (1961-71, 1973-75), Assistant Director, Office of Planning and Analysis, AEC: Commissioner, Nuclear Regulatory Commission. William Gorham (1953-62), Deputy Assistant Secretary of Defense: Assistant Secretary, Department of Health, Education, and Welfare: President, Urban Institute. Joseph Gross (1957-73), Professor and Head, Department of Chemical Engineering, University of Arizona. George R. Hall (1964-73), Economic Advisor, AEC: Deputy Assistant Secretary of Defense; Commissioner Federal Energy Regulatory Commission. Olaf Helmer (1946-68), President, Institute for the Future: Quinton Chair for Futures Research, Graduate School of Business Administration, usc. Charles J. Hitch (1948-61), Assistant Secretary of Defense: President, University of California: President, Resources for the Future. Fred c. Ikle (1955-63, 1967-73), Director, Arms Control and Disarmament Agency: Chairman, Conservation Management Corp; Vice President, Transat Energy, Inc. William A. Johnson (1964-72), President's Council of Economic Advisors: Special Assistant, Office of the Deputy Secretary of the Treasury: Assistant Administrator, Federal Energy Office. Herman Kahn (1947-61), Director and Trustee, . Abraham Kaplan (1947-52), Professor of Philosophy, UCLA: Professor, University of Michigan: Dean of the Faculty and Professor of Philosophy, University of Haifa. William w. Kellogg (1947-64), Associate Director, National Center for Atmospheric Research. Robert A. Levine (1957-65, 1969-75), Deputy Director, Congressional Budget Office. William Meckling (1955-62), Dean, Business Administration College, University of Rochester; President, Center for Naval Analyses: Council of Economic Advisors: Dean, Graduate School of Management, University of Rochester. Constantine Menges (1967-69), Deputy Assistant Secretary, Department of Health, Education and Welfare: Director, Bureau of International Affairs, Civil Aeronautics Board. Specht-29

Percy Pierre (1968-71), Dean of Engineering, Howard University; Assistant Secretary of the Army.

Henrys. Rowen (1950-61, 1967-72), Deputy Assistant Secretary of Defense; Assistant Director, Bureau of the Budget; Professor of Public Management, Graduate School of Business, .

James Schlesinger (1963-69), Chairman, Atomic Energy Commission; Director, Central Intelligence Agency; Secretary of Defense; Secretary of Energy.

John White (1968-77) Assistant Secretary of Defense.

Collins: Well, that I think covers the range of questions I wanted to touch on with you.

Specht: I've told you far more than I know.

Collins: What this is all getting at is just a sense of how RAND works as an institution and carries out its objectives. Do you have any general observations or thoughts about areas that one ought to be covering, or just reflections about the way in which RAND operated as an institution that might be helpful in getting a handle on its history.

Specht: No. It was led by people with, you know, a passion for a variety of studies, projects and goals, an interest in people who would similarly work well either within RAND, with other people or on their own. They were somehow able to judge, I think, very, very well the talents and the character of the individuals that they invited, because there was a continuing flow of interesting people and useful people and protective people, and people who were not hands-on guided day-to-day when they got here, but were given their head and encouraged to find spots in which they could find something that was interesting and useful. It seems to me that kind of philosophy went from Frank and John throughout RAND.

Collins: All right. Is that what attracted you? I guess we haven't really addressed that question when we first talked about your coming to RAND. Is that what attracted you to make the decision to transfer from academia to an organization like RAND?

Specht: I don't remember in what terms Quade and Germond described the place, or what it was that got me hooked, but they were both interesting types, and when I visited, John was an interesting type, fascinating type, as were a number of the others, so I think it was people.

Collins: Okay. That's all I have, unless you have any other reflections? Specht-30

Specht: Peace! Collins: All right. A good final note, thank you.