/^%^CH&s^^ •iJ^'vM... ALFRED P. SLOAN SCHOOL OF MANAGEMEN THE KNOWLEDGE BASE FOR EDUCATION OF RESEARCH MANAGERS Donald May, 19b6 G. Marquis 1^200-66 MASSACHUSETTS INSTITUTE OF TECHNOLOGY 50 MEMORIAL DRIVE CAMBRIDGE, MASSACHUSETTS 02139 •> i* -sPi' <• • & THE KNOWLEDGE BASE FOR EDUCATION OF RESEARCH MANAGERS Donald May, 19fc>t3 G. Marquis #200-bb Paper presented at the 19th National Conference on the Administration of Research, Newport Beach, California, October b-8, 19o;5, and published in the Proceedings by the Denver Research Institute, University of Denver, April, 19bt>. iw.-iu._. ,'..rrj .. . , p ^ ^"fr(- iUG'7l9g? Third Session '5 o- THE KNOWLEDGE BASE FOR EDUCATION OF RESEARCH MANAGERS Donald G. Marquis Professor of Industrial Manui^ement Alfred P. Sloan School of Manaf>emenl Massachmetls Dr. Williams has been looking like to turn interested in research we cause in my now our attention are all in management, I In the latter part of the nineteenth century, are management is at this moment cal difference the history in in financial want to tell you why I believe this. think I During we would and distribution, with the advances in communicaand transport making mass markets possible. At the present time the critical difference between an effective and a less effective organization is in the way they manage research, innovation, and the utili- At various organization has depended upon competence trade and company was how well made they the difference new technologies. Research management managed was manage- all the functional activities. Compare, uncertainty. now trial you can do public and he held the Chairmanship their eral Electric he is such fields his Ph.D. certainty new product development. and engineers Scientists from other employees in their expectations, values, their attitudes, and their motivations. many is so great that it results as a basis for is hard to use planning in the future. This analysis leads to the proposition that a good is one who can deal capably with and therefore he should be problems, these special Ford research manager and Space Administration. a consultant for the Genother corporations. feedback loop knowledge of at paid more than other managers. It is also of interest to note that in these times of rapid change, the top Hi": management tasks in an organization sound more and more like those of a and publication have been social psychology, and as learning, human behavior degree of in the Company and contributions in research the and University of Michi- private organizations, including the time, the partments of psychology: Yale University (1942-45) has served as considtani to the National Aeronautics is The third source is the diflSculty in measuring results when each research task is unique and never repeated. Even if you could measure results, the delay two distinguished de- Foundation, the National Science Foundation, and At the present in a professional person. differ gan (1945-57). Dr. Marquis obtained He as Before join- ing the Institute in 1959, of Yale in 1932. first example, The second source of difficulty is that you are managing a new kind of employee who views himself Massachusetts Institute of Technology. for is the at There are three sources with which you can plan and schedule production or inventory or sales or cash flow compared with what professor of Indus- Management not only the critical dif- The of this special difficulty. Donald G. Marquis is ference between a good organization and an average one, but research is the most difficult to manage of industrial revolution commerce. After the the difference that zation of between a successful and critical difference unsucessful dif- in Prior to the eighteenth cent- ferent critical activities. in management. ing times in history the difTerence between an effective ury the was part of the twentieth century tion versities. I first agree that the critical difference has been in market- critical for the overall effec- performance of industry, government, and uni- tive when the great business empires were constructed, the criti- of course, be- first Technology ment of production. would We the business, but second, at least judgment, because research at the past. to the future. Institute of ager. in organizations. 77 are coming to man- research Third Session 3 a- Top tention to their reply but then asked a second ques- executives have to deal with the uncertainties "How do you know?" He found he could put of long-range planning, foreign competition, govern- tion: ment policy, and so forth. They have to deal more and more with professional personnel in all parts of their organizations. They have to deal with the social, economic, and political environment of their the answers into four categories. and lowest in terms of accuracy of perwould be "no information at all." Next would be "information comes to me without my seeking it, and I interpret it to suit my wishes." Third, "information comes to me, and I pay careful attention to it." Fourth and most accurate, "I go out actively and deliberately seek information directFirst, ception, organization with the inevitably long delays in the feedback loop that brings them knowledge of I firmly believe that improved, it will if results. management can be research be the best training groimd for top executives. The ly." American Scientific what we mean, then, by accurate percepsample of only 20 companies there was a clearly significant relation between the score on accuracy of perception of customers and employees, documented the recently men increasing proportion of technically trained This tion. in top management, and I know one science laboratory manager in a very large company who says that one of his missions is to and the hard facts of rate of increase in profitability and growth of the company. Othf studies conducted by Edward B. Roberts have examined the formation of new technical enterprises by individuals who left government-funded develop alumni of his laboratory become presidents of all the divisions of his company. If we are going to talk about research managers we should acknowledge the fact that they are to many not a single species, but that there are varieties of research managers. as we two entirely One large people is new a small in which the the technical entrepreneur enterprise which may who starts eventually grow into a large organized business. A study carried out last MIT year at by Harry new tablished bridge region. of 20 small technical He found who had esenterprises in the Cam- individuals that the rate of profitability of these enterprises to certain was growth clearly a there motivation for achievement cess of the which the of enterprise He this. During a long were accurate perception of talk with the founder-presi- company he would "What do the customers think of your product''" Hi- paid no at- ask, from the larger MIT Lincoln organization father's with Laboratories, perhaps in the 2000 last 15 50 new companies by individuals who are technical entrepreneurs, who want to take the technology that they have learned in the parent laboratory and use it as a basis for forming a new business. Of these 50, there are 45 still viable and successful. The aggregate volume of business of the spin-off companies from these two laboratories is rapidly approaching the point where it will exceed the volume of the two parent companies, and 40 percent of their devised a very ingenious method of getting dent of each a which the his years the customers and accurate perception of employees. at is (in in probability of suc- employees, there have been formed as — type some other), then the new company is less. Similarly rale measured on McClelland's test; they had a lower desire for power and authority they didn't try to hold the strings too tightly; and most important of all, they had an accurate perception of critical aspects of their environment. The two most important aspects for this two or three years a gap of business or characteristics of the founder. higher is founder of the new business works in related The men whose companies showed a higher of increase in profitability had, compared to others, is — 1965 issue of the Harvard Business Review, examined characteristics of One is how directly the technology of the parent company is taken into the new company the more directly, the more successful the new enterprise. If Schrage, which will be published in the Nov. -Dec. the is from government-supported R and D to commercial research. For example, the MIT Instrumentation Laboratory, which got started shortly after World War II, has graduated, or spun off, or lost, 27 employees who have started their own companies. Of these, 26 are successful and still in business, although they differ in their rate of growth and profitability. The differences depend upon some interesting factors. requiring organization. a large laboratory is other is This transferred manager of a organized laboratory, and anything over 50 find research managers. The situations different interest to us as a own companies. channel by which technology laboratories to start their The job of research manager differs by industry, was made clear yesterday. Moreover there are certainly is In a business already We 78 is commercial. would go on and talk about other types of Third Session i/» ' research managers. Certainly who we ought traditions of research who search are from can also distinguish research managers by team leaders or managers several at first- manager a research much in who ought production. many there are Recognizing, then, kinds of research manag. can we say about the education, training or seems to me we can know whether with sufficient need to know there are certainty to an average research manager. This of title my report refers —unlike If We bilities and of say a little —who we all word is it rules of thumb — is some Investigation is memory to carry in It is about it Now, is worth the let's in Mes- we have our actions. The fourth I want to managers is a quite try. formation on research managers we have. As a mat- know is only one large-scale systematic one reported by Clarence Randall in ter of fact, there study. This is based upon very careful assessments of re& Hamilton. It sur- whom 300 were R and D in companies of managers, by systematic procedures all sorts On young, perhaps grown up very —because turn to what kind of systematic in- veyed 1427 executives, of because traditions always grow quickly myths, I would like to suggest that in no field of management than research management is it more necessary to do systematic research. In no field is it more difficult, but, for reasons that I hope to have made clear The second and perhaps more common knowledge is tradition. The management of years. But the traditions have fables, ago; besides, times have changed and this Tradition. ly, much different situation. spoke of yesterday. 20 for absence of learn- difficult for 1956, is of results show up the results of an action five years search managers by Booz, Allen organized research and development in systematic investigation of the results research. source of been not expect Results. of Mesthene to the not basis for manage about approval the sort of things that Dr. the a posieriori a priorisiics that Dr. directly — D may we would of research management. operate how and just don't exist. In the put in a plug, and who give speeches as if they went up on top of the mountain and received the R source of knowledge, and the one for which some research managers of course none of them are here, but This for already, Revelation. There are other fields in the case, is Systematic each. some this to believe that there to. me budgeting for has been of a spec- thene was talking about yesterday you are teaching polymer chemistry or magnetohyrodynamics, there is an established body of knowledge, and you can test whether a student has mastered that body of knowledge. Is there anything comparable to that in research management? The body of knowledge in research management accumulated over the past 19 years (which Dr. Williams is working to summarize) is pretty thin. I tend to think of this body of knowledge as derived from four sources; revelation, tradition, experience, and Let it by experience we see the growth of ing If systematic investigation. resources, projects learning on the basis of experience, because the possi- what the announced is the in research manage- three, five, ten, or fifteen years. about foundation is very hard to operate on the basis kind, and- ial the difference between a superior and sound management knowledge is it of selecting only one or a few laboratories, we need any principles known be worth teaching. management his of experience because your experience has that de\'"' talk improve you don't know for a long Whether you that is a managerial action in that before to going to challenge that statement. laboratory, or termination, courses or curricula or on-the-job training, to The reason for your ment of research managers? It am planning the allocation are as what "s, I contend that one of the hardest things ment. about research as he does about finance or marketing or am I time the results of what you have done. to be knows how his experience I the laboratory director, the sense that he of re- experience; and is world to learn from experience the vice president for research and development, and the president of the organization, management, then, of research. managers, the functional levels, from the principally sure that everyone here believes that he has learned keting. line supervisors, the project management, of management of production facilities. Experience. The third source of knowledge search and development into manufacture and mar- their level of responsibility; the forms other are concerned with carrying the products of re- We The management have come from sence of other more soundly based knowledge. to distinguish managing a science-or discipline-oriented activity from those who are managing an applied research or advanced technology activity, those who are managing a development activity, and those those throughout the country. the basis of the assessments and talking with colleagues and superiors, the managers were classi- rapid- fied as those the ab- 79 who were clearly promotable, those who '5 o- ^ « Third Session I/I '5 IT o would never be promoted, and a middle group of marginally promotable. About one-third were clearly good research manager. promotable. though The We are depending here upon a measure of performance based on evaluation by people who observed their work and who perhaps shared the myths and fables and rules of thumb of the person being It turns out that eight characteristics dis- the promotable from the not promotable all fields. These are performance in his evaluated. tinguish manager present in position, drive, intellectual as sounds to me like the Boy Scout oath honest, obedient, courteous, etc. ful, any idea how tics to to move from this the selection of research I list some of social to the acceptability. (Well, if but there are some promising ap- years ago, those of you in The current report of the MIT Research Program on the Management of Science and Technol- more ogy, copies of which are available here, represents Another is you have no other creativeness, whatever that even good research people by the use of bioThe work that I described at that meeting was largely a promise, a direction, and a hope, and you were kind enough to say, "Go ahead, and when you have some results, let's see them." the is necessary, graphical information. fields. One is to analyze systematically tentially faith- turns out little is who were at Esles which Dr. Donald Pelz reported some very interesting results on factors in research effectiveness. Dr. Calvin Taylor, on the same program, talked about ways of selecting po- research managers than for managers in critical for other these characteristics are a It sounds, Two wouldn't have managers or it Park heard a session of characteris- design of a training program for them. that — think I proaches. ness. It that difficult, outcomes of research management by some obmeasure of results. TTiis is not nearly as easy the leader- ability, seems jective administration, initiative, motivation, creative- ship, alternative it first time we have felt confident enough of our is. results to issue a report. now good one; white Anglo-Saxand Protestant). Less important than in other fields, according to this very thorough systematic criterion that's always a to try to I will not take your time survey the results of the 38 published papers and 70 theses which are summarized therein. on, I hope, instead, that this organization years, invite reports of the and drive. This is what you can get with a very thorough study of people's opinions about who is a jective research studies tions for 80 management will, over the more substantial and obwhose results have implica- study, are initiative, administrative ability, leadership, policies and procedures, Third Session c o DISCUSSION OF THIRD PANEL SESSION 3 u I/I FRENCH, Panel Chairman While you are thinking of something to question these gentlemen about I would like to ask some ques- of communication tions that occurred during the speeches. government, and J. D. First of ment. Yet, is believe more I Is that very manage- few professional mean government. political me to that a great deal of the in- down being fed stems from any relationship between what you said and the need to have more people in Congress, and perhaps in the administrative branch of the government, who have had experience is wonder 1 to us all there if is in professional research. gather you MARQUIS, D. G. structured and predictable methods are going to emerge successfully? me In your research: there any pattern emerging, since 1 seems fluence that Nevertheless, about this divergence of considera- view it government. in that interval of time). management of the development of top in occurred to people, certainly very few people in research, go into all, Mr. Williams, you indicated earlier concepts of the almost chaotic method of managing research on the one hand, as opposed to a more ordered attempt to structure research. (Incidentally, I was struck by the fact that the chaotic group seemed to do rather well tion regarding the It Panelist Others yesterday spoke about the fact that governin the past has largely been in the hands of there any pattern ment emerging which indicates what the instructional problems these managers will face will be? good reasons, because of the upon the powers of government. Dr. Mesthene especially pointed out, and I don't need to elaborate on it, that the nature of government is lawyers, and for very legal constraints L. B. I WILLIAMS, Panelist am not quite sure. Dr. mean French, what you changing — that the objectives of government are shiftand we are swallowing them, and that in this new kind of government, especially on the executive side and perhaps later on the legislative side, 1 would agree with you that there is a place for people who have had experience and have learned something as re- by the word "instructional." ing CHAIRMAN FRENCH: Teaching or learning become an administrator of I,. B. WILLIAMS; Well. I to research. think it is very, very clear, managers. .scarch you cannot really expect a maximum of efficiency of your research of R and D organization to come from leaving it alone, leaving as I pointed out, that the situation in chaos. M. HARRIS, Ihe serious you read and reread these Proceedings, that certainly some kind of order and system in the planning and executing of research and R and D effort is recognized now as not only desirable but necessary. Peltz" studies seem to have brought that out (his studies being at least an academic and scholarly attempt), so that whether or not we are ready to start teaching these methods of two-way channels of influence 1 am not qualified to say, and I don't think that the Proceedings would in any way reflect this. see, as It occurred to ask Don I telling point when you suggested common management and that was management, and perhaps research management might be a direct a rather the this question. who spend ducing products such as their in the pharmaceutical industry, the Gillette Company), and dustries I call the con- still I own money on pro- chemical industry, the textile industry, and the find that people in these in- think that, for most types of work, the is more eff^ective. I am as to the source of the information rather curious that you have shown here, which distinctly showed that functional work is more productive. Is it more productive with a between the responsibilities of of research is stressed I have held discussions among what dustries (people me. Dr. there this sumer-oriented, generally non-government-funded in- Marquis, when you were speaking, that you made a great deal in Company have been pulling each other's importance of the project type of thing versus the functional, and I still stick to this. I would like to project system CHAIRMAN FRENCH: I some time on some things, but comment, because yesterday legs for here yesterday by Dr. Harris. More and more you can Gillette Don Marquis and The word "chaos" was used government-funded money, be- the large projects of that cause, line proach 91 if it so, that's a different story than what I of the way was discussing yesterday. to ap- Third Session c MARQUIS: You (A D. G. 3 u were, in 35 cases, for very large in million government-funded The reason industry. that Marquis are correct that the data — $4 —performed by questions that were asked about whether the projects started our studies with are Air Force projects, whether we are dealing with commercial development, and so on. The basic point that Don Marquis mentioned is the need to know much more about the evaluation of results. Although he and his group have moved very far beyond this kind of study that was mentioned earlier, the kind of study where you simply ask people for their opinions as to who is good and who is bad, I think that we are still a long way from the desired end product. That is the measurement of the actual government contract projects is that it is easier to get your hands around them. They have a beginning and an end, an initial estimate of cost, and so forth. We started this summer trying to do the same sort of thing in commercial laboratories, and wc welcome your offer of cooperation. We would like very much to study the eight laboratories in Gillette. GORDON, R. S. would I because think I Panelist to follow like up Dr. Harris' question results which, in the case of the Gillette evaluation of functional versus project organization. It my would be understanding, then, tegrative function — that is to say, the utilization of information for a mission this that the in- sir, — is really If this true, is is from the this not different kinds of criteria will The important beyond the concern of the people performing the contract. was done within the government. It even in it it can not ready for application so than first D. G. is we need that a You government operation. But it was the gov- N. One These were a mixture of Air Force, and other government agencies. usual morning cautionary note: to detract at all from Don Marquis' very excellent presentation, for which but great admiration, about how thing that I think a of these fables and myths, a just as this is This live, little some- is J. T. I which con- to take a lot think it is more of going to take time; it I man entirely. a little Norman, split them in a variety of ways, and the re- reported are true, regardless of the other is number interesting fact came out, since we JR., Thiokol Chemical Corporation look at these results from a somewhat different mind. My I What was raise the following question in the breadth of the technical con- programs? concern for that of late going the kinds of studies that GREY, tent in these think I asked previously about project versus functional orI would agree lot point of view, and my data might be harmful as the myths themselves, and I examination. indicated by the two questions which have been ganization. disagree with your be said to think was not stressed sufficiently. In the world in which most of us sists have nothing word ought these results can be used. I I I managers in this case: the ones who ran projects that were judged more successful had less experience than those who ran projects which were judged less successful. I wouldn't, on the basis of that, argue that experience is a bad thing, but I would argue that the myth that the man with a lot of experience is the best project manager deserves re- NASA KAPLAN: My Although are talking about project use in an operational system by a govern- Without wishing computer. of other factors of that sort. ment, and these, incidentally, were all development projects. That evaluation is given in terms of its potential for could sults that ernment's evaluation of the success of the develop- ment agency. Navy, Army, is more research. For this particular development projects, you could split them into large and small. The results are the same. and tween the industry's laboratory and the government's technical people. The follow-through, the implemena has in the past, and result rolls off the a bad thing, sub-.set of large is it MARQUIS: the integration involves a cooperative relationship be- tation, re- has to go forward general proposition that for a starving different, is on much more food Yes, of course that this research It as soon as the an academic laboratory, such as the brain MARQUIS: is search has to be encouraged. concern or surgeons team that our chairman runs? D. G. differ. point be viewed as something that inte- gration that occurs in laboratory mission or operational performance, say, within a private Company, would be what kind of product and how profitable is it. In the case of an Air Force project it would be how good is the weapons system, and so on. The does bear on Professor Marquis' it more knowledge of the parameters involved, of precisely the types of projects we talking about, with a lot is million to $6 my question arises from the fact we have had many discussions of the stim- ulation of the interdisciplinary project team. Don ever, in this case 92 it might appear that if How- the technical Third Session problem areas that have been reported over the past 18 years. I hope this is one of the outputs of your content of the project was rather narrow, the techni- were rather stimulated by association with cal people their technical peers. MARQUIS: When D. G. $30 million content were effort. which project, Don, I hope you look into how you transfer the knowledge you are generating and how you get the word across to the R and D manager. The other thing a a system, the technical is interdisciplinary, although in general these is all you are talking about aerospace, electronic and electric. These are exactly the ones for which project organization is GREY: T. I relates to his earlier If I Kaplan's example, and how comment. Let mc give you a were a lab manager and of projects under me Allen's data, which that will be used. J. with reference to it hypothetical case. recommended and, as a matter of fact, government agencies are now requiring it and insisting that the proposals outline the form of project management I had a large number along with the results of showed that the best project Tom teams consider two approaches at a time, would I use his data from the standpoint of looking at my many pro- address this question to Mr. Williams. and finding those project teams that are considThen I can do one of two things. The ones that are considering more than two approaches I can fire, or I can force them to consider only two at a time. jects NCAR Will your study of the Proceedings of the Norm is in- ering two approaches? clude a critical analysis to determine whether the apparent beneficial effects of the shift from chaotic to organized research may be a self-induced artifact of our concern for our raison d'etre as the research management establishment? WILLIAMS: I would like to respond to your comment. Certainly over the years we have recognized that several kinds of people have come into L. B. WILLIAMS: L. B. "artifact" a "straw tell Well, certainly, man." you mean by we can if don't think that I from the Proceedings whether such not, really. after I consideration the case or is don't think they will reveal that. are experiences, and, as all, first we heard of research administration from an academic viewpoint. had the people interested in These, yesterday, First of problems all, think I we the behavioral sciences and psychologists, who wanted no doubt research administration area, the sociologists happens, perhaps, and what some of the problems of motivation and control and some of the things I mentioned that Dr. is is still an art. What happening is the same to explore thing that happened, for instance, to the medical profession. I J. 1 think it don't know what relatively is MAHONEY, E. Peltz treated. the timing might be, but ple of science comparable. Administration The first comment is to I Les Williams, and hope is person, and he generally — Conference form here. I see in these gatherings have seen a couple over the years now a very unique bringing together of academic minds, industrial people, the guys who push the buttons on R and D management, and also the government people who, in some R cases, are sponsoring either contracts and work D in work, or they are actually doing under in things, I am all together. I effort, hope I R and D where he This proach, or a noble goal, I that interested is think. a noble type of ap- We are just so far dark ages on the academic point of view going to take quite a while before we learn in the it is to describe terms what is in mathematical or even heuristic the real situation in research and re- search management. So, I think that our Conferences have included in the past some of these academicians and results of scholarly efforts. is see a research this. In other words, I hope we use us say, as a research mechanism, possibly every year reporting on some of the general forum, research administration. how emphasis after this is together things and people into a working model for this back very happy to see Les' who More lately, of course, we have had the academiwho generally can be classified as the operations research people, who would like somehow to tie and of themselves. trying to pull this a person is not people. cians, I think, as Don Marquis brought out, we have over the years relied upon this Great Man coming forward and speaking on how he manages his project, and the peo- manager or business-administrator type of business this is evolution of the — in A little later, or maybe about the same time, we had another kind of scholarly and academic person coming into the picture. This is the economist or the National Aeronautics and Space with reference to what These are people interested and research. let I personally believe that tinue that. 93 What it would be good to con- the distribution between experience M M 3 u Third Session c 3 u and theory should be is not, of course, up to me, but up to the Conference management directly from year to year. am I when I say more of it. want I come the closest. to E. couldn't agree with I is first the most many of his reasons for to the extreme speculative learn my own man- in and 1 have belief on that times, my R end of the to the D and then If not, luding myself for 25 years me in believing I had done I would The question MARQUIS: From develop. that velopment contracts, what control system, as mercial sir, in . HERSHEY: My experience goes back to when way was no PERT, so I am disqualified to measure me tell you what I am talking about. This was a series of development projects in the 1930's, in which we had a problem of taking the discoveries of chemists and, as chemical engineers, converting them Now, we MARQUIS: successfully from a chemical reaction how to It is development That's not an easy question to hard to use statistical evidence of any about the unique events, the Lavoisiers, the Maxwells, and so Tom forth. You don't try to predict new book on Revolutions In Science. These revolutions occur when these great breakthroughs come, and in between we little ants work piling up pebbles and grains of sand to confirm, them. Read Pruden's to apply, to test aspects of a theory. clearly learned over a succession of half a dozen of these satisfy yourselves that the out? sort to talk Let into operating plans. concerns, to government, utility to industrial . there it. are supposed to de- answer. R. L. men environment that existed then, and may now be on D. G. . business gets how do you smoother, of such individuals was not actually the result of the PERT the value of a is compared with a simple much And, quite aside from the immediate com- how do you the your experience, management as this propose to insure that the essential nutrients are re- so, one necessarily learns from experience. This certainly is a myth. One does not learn from experience unless one has the ability and the desire to examine experience in order to learn from it. D. G. is, efficient, tained in the soil in which these remark on the myth like to borderline much more have been de- I question pertains to research as chaos. it, possible to is it My There are those who believe that the time is not when the dramatic advances in fundamental knowledge may still be made by individual intellectual giants. Such giants in the past, I think, have grown in the soil of what is known here today phases of the development process. Finally, don't entirely past spec- development side of purely industrial research, from experience. in certain I Gordon may have think, by the of the difficult experience would suggest to you move over at least in Dr. anyone here, not the contentious part at of application and development. 1 justify this discrimination on the ground that the title of this Conference did not contain "D," nor does the title of this session, and I can limit the question perhaps, and decrease the argument. I him more On the other hand, I was a little startled to hear him apply so broadly the statement that you cannot learn from experience. I think this is right if you go if somebody would that in the latter part. of the kind which will be called fundamental research, make a few remarks on Professor Mar- point. that hope it specifically. word "R and D." du Pont de Nemours & Co. I. myself have said so trum, but to Professor Marquis referred entirely to the single various functions a businessman has to face given the was done good and we should have is that the research function I part in the it quis' excellent talk. aging. first some remarks think HERSHEY, R. L. direct only expressing a personal opinion think this I end of the go more disagrees with the existing paradigm — New —-that knowledge is Kuhn's word but we explain it away. We try to fit it in with new ad hoc hypotheses; eventually it accumulates to the point where somebody with real insight can get us a new paradigm, and then we work for another 50 years building up both confirmatory and disconfirma- to the operating plant. D. G. MARQUIS: I couldn't agree with you more, and I don't want anything that I have said to carry the implication that don't have the highest admiration lor the accomplishment of the R and D people I over the last 20 or 30 done about ten times years. I only think it tory evidence. I could be don't think as effectively. we can J. E. MAHONEY: like to direct to This is a question which Professor Marquis. we need try to predict these great make the history of science. What breakthroughs that 1 held I it would this do, perhaps, is increase the effectiveness of intermediate level of work that lay the foundation for the genius. at the 94 is so essential to o r p h n .'• sn 1 Date Due Lib-26-67 MIT LIBRARIES 3 9080 01439 2028