SOCIAL SCIENCE RESEARCH COUNCIL

VOLUME 28 . NUMBER 2 . JUNE 1974 605 THIRD AVENUE· NEW YORK, N.Y. 10016

THE ROLE OF THE SOCIAL SCIENCE RESEARCH COUNCIL IN THE ADVANCE OF MATHEMATICS IN THE SOCIAL SCIENCES

THE CURRENT LEVEL of mathematical training for social scientists in this country was not quickly achieved, nor did it grow by itself through natural evolution; instead, it has come about through a long, fairly deliberate pro­cess that has depended upon the ideas and contributions of a great many people and organizations. To bring mathematical education and use to its present state re­quired the efforts of several successive academic gener­ations of mathematicians and social scientists, as will be apparent from the names that arise in the course of this description.

One organization that worked most systematically at developing a proper base for mathematics in social science was the Social Science Research Council. The Council was established in 1923 to provide a means for the several social science disciplines to improve research.

• The author is Professor of Mathematical Statistics at Harvard Uni­versity. He has been long and closely associated with the Social Science Research Council. serving as a member of its board of directon for twelve years. 1953-58 and 1965-70; as chairman of the board. 1965-69; as a member of the Committee on Problems and Policy. 1955-61 and 1965-69. and of the Executive Committee. 1969-70. His participation in Council activities dates from 1948. when he served fint as consultant in the work of the committee appointed to analyze the experience of the Research Branch. Information and Education Divi· sion. ASF. and then as chief of staff of the Committee on Analysis of Pre·election Polls and Forecasts. and senior author of its report. The Pre-election Polls of 1948 (Council Bulletin 60. 1949). Mr. Mosteller played a major role in all the programs of the Council on which he reports here. He was chairman of the 1951 Interunivenity Summer Research Seminar on Mathematical Models for Behavior Theory. which resulted in publication of Stochastic Models for Learning, by Robert R. Bush and Frederick Mosteller Uohn Wiley Be Sons. 1955). And he was an active member of the following committees: Mathematical Training of Social Scientists. 1952-58; Research Training. 1956-59; Scaling Theory and Methods. 1950-57.

This article was written at the invitation of the President of the Council as part of the commemoration of its 50th anniversary year.

by Frederick Mosteller-

One way to improve research is to provide specialized training for graduate students and postgraduates. As an early example, Samuel A. Stouffer was awarded a Coun­cil fellowship in 1931 to study statistics with Karl Pear­son and R. A. Fisher in England. More generally, fellowship programs, research institutes, conferences, research seminars, and research planning and appraisal projects offer training.

A mechanism that the Council has found most suc­cessful in organizing such efforts through the years is the committee. A committee may be appointed to assess the status of research on a topic or a field, or to develop new methods, or to train a new group of workers, or to make plans for new research. The committee may endure for one or two years or for many. Typically a group of social scientists, often an interdisciplinary group, proposes to the Council that it encourage a new development. The proposal is reviewed, and the Council takes steps that seem appropriate. Sometimes no further steps are taken because the problem may not be ready, or the Council may not be a good vehicle for the work. But sometimes the Council can act.

The interest of the Council in mathematics is that it is a special area where training may be useful or even essential to the prospective social science research worker. The Council has long carried on programs that support promising individuals, adding a needed special research skill through further study. What was different about mathematics was that it was viewed as being needed across a broader front of the social sciences, not on quite such an individual basis. The feeling in the 1950's was that the whole of social science was lagging in mathematical training. At that time, except possibly for economists, people entering social science were in effect

17

being self-selected for a low level of mathematical train­-- ing. The Council undertook to do something about this:

Many members of the board of directors and the staff encouraged and participated in the effort. To the extent that its success can be attributed to one man, that man was William G. Madow.

William G. Madow dreamed of a community of social scientists whose mathematical strength was equal to its mathematical problems. These dreams slowly turned into reality because he was willing to devote so much energy to the leadership and execution of the required program, because the Council gained the planning ser­vices of many unpaid experts over long periods and could facilitate through its staff 1 sustained efforts by its committees, and because foundations vigorously ,fi­nanced the many facets of the growing enterprise. The Council, though it speaks, as do foundations, of starting enterprises and then letting them find their own way at maturity, has sometimes helped a project for a long time. For -the mathematical training of social scientists, a 12-year program led to a natural end point through the successful creation of a new organization. Founda­tions have played and still playa vital role in mathemati­cal research in the social sciences. The preparation of materials, the summer training institutes, and the re­search projects could not have been carried out under the Council's auspices without their support.

Although elementary mathematical training has now again become largely the responsibility of the colleges and universities, we have today a much different cur­riculum in mathematics for social scientists than ex­isted in 1951, and we have materials much more appropriate for the effort. Without becoming a mathe­matician, a social scientist today can get a reasonable mathematical education to equip him for his research.

The growth was not a simple product of World War II or a creation stimulated by Sputnik. Instead, it was a continuation of previous efforts.

The American Mathematical Monthly in December 1932 published the first glimmerings of concern for

1 The staff of the SSRC traditionally maintains a low profile in com· mittee work. Nevertheless, the mathematical community has been most fortunate that Elbridge Sibley acted as staff officer for so many years for Council committees with quantitative and mathematical interests. His own studies of demography kept alive his interest in mathematical research, and his New England disposition made light of the cantanker· ousness of mathematicians. And as a veteran committee man he often found no problems where committee members scared up many. Both Pendleton Herring, then President of SSRC, and Paul Webbink, then Vice·President, maintained a most constructive attitude toward this de· velopment. Earlier Donald R. Young, a quantitative sociologist, had been President, and a number of chairmen of the board have been mathematically oriented, for example, Wesley C. Mitchell, Frederick Mosteller, Herbert A. Simon, Conrad Taeuber, S. S. Wilks, and E. B. Wilson.

18

mathematical trammg on the part of the Coundl-a Council subcommittee's report entitled "Collegiate Mathematics Needed in the Social Sciences." By 1954, Elbridge Sibley wrote that although the minimum stan­dards stated there were accepted, at least in theory, the methods used in the social sciences had advanced so much that the mathematics recommended in 1932 was no longer adequate for a reading knowledge of social science literature.

Let us move away from training to consider research for a moment. The Council's Subcommittee on Predic­tion of Social Adjustment, chaired by Stouffer, directed the preparation of The Prediction of Personal Adjust­ment (Council Bulletin 48, 1941), by Paul Horst et al,2 The list of prepublication readers alone represents the best of quantitative social science at the time.

In addition to other fine things in this volume, Louis Guttman laid the mathematical foundations for his "cumulative" method of scaling, now usually called Guttman scaling. It was a time when the case study method was under heavy attack. Stouffer contributed a short treatment of the unique case in terms that a statis· tician today would appreciate and one that might en­courage a modem Bayesian to write considerably more. For its period, the Guttman statement was mathemati· cally sophisticated for the social sciences, dealing as it did in 1941 with matrices, iterations, expansion of matri­ces in series, and so on. This book well illustrates my point that the later rise of mathematical developments in the social sciences was not merely a World War II phenomenon, ,but rather an extension of ideas and hopes that were being worked on earlier. Measurement and Prediction, Volume IV of "Studies in Social Psychology in World War II," also produced under the direction of a Council committee,S applied the Guttman methods and discussed scaling extensively.

The interest in scaling persisted, and a Council com· mittee on scaling theory and technique under the chair­manship of Harold Gulliksen 4 arranged for Warren S. Torgerson to prepare his book Theory of Scaling, pub. lished in 1958, a comprehensive mathematical treat­ment of this subject.

2 The other members of the subcommittee were E. W. Burgess, Leon­ard S. Cottrell, Jr., E. Lowell Kelley, and M. W. Richardson. Its work was supported by Council conference and planning funds provided by the Rockefeller Foundation.

8 Appointed to supervise analysis of the materials collected and methods used in the statistical studies of the social psychology of American soldiers conducted by the Army's Research Branch during World War II. The committee's work was supported by funds from the Carnegie Corporation of New York.

4 Other members of the committee were Paul Horst, J. E. Karlin, Paul F. Lazarsfeld, Henry Margenau, Frederick Mosteller, and John Volkmann. Its work was supported by the development funds granted to the Council by the Ford Foundation.

VOLUME 28, NUMBER 2

Another illustration of the Council's early encourage­ment of the use of mathematics in research in the social sciences was the summer seminar in mathematical models for learning held in 1951 at Tufts College.1I This two-month effort produced several papers and brought the Harvard and Indiana University groups in mathe­matical learning theory into close touch. The decision by Bush and Mosteller to write Stochastic Models for Learning was made late in the seminar.

Mention should be made of one non-SSRC conference that was extremely important for the ultimate develop­ment of mathematical applications in social science research: the eight-week seminar on the Design of Ex­periments in Decision Processes held at Santa Monica in the summer of 1952. This meeting helped create a community of scholars interested in mathematical meth­ods of doing research in the social sciences. The mem­bership included many who contributed directly to later SSRC efforts, for example, Bush, Robert L. Davis, Wil­liam K. Davis, Jacob Marschak, Howard Raiffa, and Robert M. Thrall.

Training and research in mathematical economics and econometrics does not playas large a role in this report as one might expect considering both their im­portance and their advanced state. The reason is that mathematical economics was a well-developed subject even by the beginning of the twentieth century, and econometrics fairly well developed by the 1930's and certainly the 1940's. A tradition had already been estab­lished for economists to study mathematics and, for the most part, they were doing this in their own universities in either departments of mathematics or of economics. Much research training went on in research projects within the universities. Even though some of the Coun­cil committees were directly concerned with mathe­matics in the field of economics, the training aspect was not prominent. Special conferences were sponsored by the training committees, and they will be mentioned.

For some time the obstacles besetting the student of social science who wanted training in mathematics had been under study both inside and outside the Council. In 1951 the Council asked eight specialists to prepare memoranda on the mathematical background desirable for Ph.D. candidates in social anthropology, social psy­chology, and sociology.

At about the same time, outside the Council, a Com­mittee on the Mathematical Training of Social Scien­tists, chaired by Madow and consisting of representa-

G Under the Council's program of Interuniversity Summer Research Seminars supported by the John and Mary R. Markle Foundation. The participants in the seminar were Frederick Mosteller (chairman), Cletus J. Burke, Robert R. Bush, William K. Estes, George A. Miller, David Zeaman; assistants: William McGill, Katherine Safford Harris.

JUNE 1974

tives of 12 scientific societies,6 also discussed what steps should be taken by social scientists. This committee raised with the Council the possibility of getting sup­port for a project on mathematical training, with the specific purpose of producing a source book on the mathematical methods used in the social sciences. The result was a two-month project under Madow's direc­tion, the Interuniversity Summer Research Seminar on Source Materials for the Mathematical Training of So­cial Scientists, held at Dartmouth College in the sum­mer of 1952.7 The technique was to prepare sets of prob­lems on single topics, beginning at an elementary level and continuing to a point where the results were of direct interest to social scientists. In addition, the par­ticipants in the seminar explained typical mathematical methods applicable to large numbers of social science problems. Finally, they prepared translations of social science ideas into mathematical terms together with appropriate mathematical references.

Why Dartmouth? A lucky accidentl Originally it was chosen because the summer climate in Hanover is attractive, and there were good library facilities. In later years it turned out that under the chairmanship of John G. Kemeny, the Department of Mathematics at Dart­mouth had as encouragingly forward a view toward so­cial science, new courses in mathematics, mathematical models, and computing as was to be found in the nation.

The next spring, the Council appointed its own Com­mittee on the Mathematical Training of Social Scien­tists.s With preliminary materials apparently available from the 1952 summer session, the new committee planned a 1953 summer training institute for social scientists to be held at Dartmouth. It discussed the ques­tions whether to include statistics in the teaching, and whether the mathematics should be primarily abstract and theorem-proving or directly applied to social sci­ence. As it turned out, statistics was not included, and three of the courses were taught on an abstract basis. As a rule in this committee, amiable agreement seldom sur­vived the first statement of fundamental principles.

8 This committee emerged from the 1949 meetings of several mathe­matical societies at Boulder, Colorado, where in a joint session members of the Institute of Mathematical Statistics, the Mathematical Associa· tion of America, and the Econometric Society recommended the forma­tion of such a committee.

7 The participants, in addition to the director, were Oswald H. Brownlee, David A. Grant. George A. Miller, E. William Noland, Howard Raiffa, and Robert Solow. The project was supported from the development fund granted to the Council by the Ford Foundation.

S The members and their years of service were William G. Madow (chairman). 1952-58; Jacob Marschak, 1952-54; George A. Miller, 1952-54; Frederick Mosteller. 1952-58; Robert M. Thrall, 1952-58; E. P. Hutchinson, 195!l-58; Paul E. Meehl, 195!l-54; W. K. Estes, 1955-58; John G. Kemeny. 1955-58; Howard Raiifa, 1955-58; Robert Solow. 1955-58; Robert R. Bush, 1956-58; staff: Elbridge Sibley, 1952-58. Fundl for the committee's program were provided by the Ford Foundation.

19

Nevertheless, consensus on what to do came fairly easily once we abandoned discussion of principles.

The 1953 institute was directed by Madow, who was also a member of the teaching faculty.9 The four courses offered were algebra, including matrix. theory and quad­ratic forms; calculus; other topics in analysis; mathe­matical models in the social sciences. The lecture program was very heavy, the problem material was not extensive, and since much of the effort was devoted to theorems and proofs, relatively little time was given to setting up problems and solving them and to practice in basic manipulative skills. Applications for membership in the institute had been gratifyingly large in number, 234 applications for what turned out to be 41 appoint­ments. The group was extremely strong as the list of members shows.10 Of those attending, about one third already had the doctorate, the rest were in graduate training. About half were psychologists, one fifth econo­mists, one fifth sociologists, and one tenth from other disciplines.

After the first institute the committee and the faculty of the institute took a very hard look at what had been done and reported in detail to the Council. On the plus side they noted that all the students stayed all summer. On the basis of a written examination, the faculty con­cluded that probably three quarters of the students gained a good bit from the courses. Two thirds of the students could see what mathematical devices were needed for different problems. They could gather mathe­matical information by taking enough time, and they found it useful. Half seemed now to be prepared for courses beyond a first year of calculus. On the negative side, even though there had been considerable prepar­ation of materials, it was not enough, and the teachers were hard pressed to write up their notes and distribute them. Students and faculty agreed that teaching the first three courses at an abstract level without social science content reduced the effectiveness of the seminar.

At the September 1953 meeting of the Council's board of directors, Bush suggested that to achieve a two-way understanding, the Council should consider the social science training of mathematicians, a suggestion that S. S. Wilks and Leonard Cottrell found immediately attractive. Ways and means of improving communica­tion with departments of mathematics about mathemati­cal training for social scientists occupied much of the Council's meeting time. Herbert Shepard, a student participant in the 1953 summer institute, spoke on its

II The other teachers were Robert R. Busb, Howard Railfa. and Robert M. Thrall: assistants: R. L. Davis and R. E. Priest. Visiting lecturers included Paul F. Lazarsfeld and Frederick Mosteller.

10 Social Science Research Council Anntlal Report, 1952-195J, pages 62-63.

20

strengths and weaknesses, as viewed Iby the students, and emphasized a need for homogeneity of readiness on their part.

The committee decided to postpone sponsorship of further summer institutes until 1955, by which time adequate teaching materials could be prepared, and to administer a questionnaire to the students of the 1953 institute to find out more about its results.

At the March 1955 board meeting, the Council learned the results of the follow-up questionnaire. The students ordered the goals of studying mathematics as follows: (1) to facilitate reading social science literature that uses mathematics; (2) to formulate and analyze social science problems with the help of mathematics, where appropriate, and to recognize where it is not; (3) to understand the nature of mathematical concepts and reasoning; (4) to prepare for the study of mathemati­cal statistics; (5) to communicate professionally with mathematicians and statisticians. Over three quarters of the students returned the questionnaires. All had stayed in their social science fields, quelling an earlier fear that some might leave. All but 2 said that they had made progress in reading the mathematical literature in their own fields, and 13 said they had made progress in formulating and solving their own mathematical prob­lems in the social sciences. This was particularly gratify­ing because the institute had not worked especially hard on this aspect, emphasizing instead the theoretical under­pinnings of the mathematics. Results of their institute work were being used by 8 in their teaching; 18 had taken further work in mathematics of a formal sort; and 6 more had participated in informal kinds of further study-all in all, a much stronger record than we could have reasonably hoped to find.

The main recommendations students made for further summer institutes, which the committee implemented in the 1955 institutes, were: (1) more homogeneous groups-one for students with calculus, one for those without; (2) smaller groups-about 15 members; (3) more calculus and more probability; (4) more material directly from the social sciences; (5) all materials to be available the first day of the session.

For discussion with mathematical organizations, the committee prepared a policy statement and outlined a mathematics curriculum particularly designed to meet the needs of social scientists.

Meanwhile the committee began taking steps to im­prove the availability of appropriate materials. It spon­sored preparation by Samuel Goldberg of a monograph on difference equations for use by social scientists. Bush, Robert P. Abelson, and Ray Hyman prepared a manu­script giving mathematical problems and examples for psychologists; this was to be an adjunct to mathematics

VOLUME 28, NUMBER 2

texts such as algebra and calculus books that did not include such materials. Harold Kuhn was asked to pre­pare a manuscript, including exercises, on applications of the theory of games and linear programming in economics. Gerard Debreu developed materials for the study of economics from an algebraic and topologi­cal view. Funds were made available to the Mathemati­cal Association of America to assist a group working at the University of Kansas in the summer of 1954, under the direction of W. L. Duren, Jr., on the preparation of experimental text materials for a general freshman course in mathematics that would be more suitable for social science students. This group (originally called the Committee on the Undergraduate Program, later the Committee on the Undergraduate Program in Math­ematics) maintained a very close relationship with the Council's committees on mathematical training through­out the years.

Nearly all these efforts to improve materials came to fruition over a period of years,l1 and much fresh ma­terial was available when the 1955 summer seminars were held.

The relationship begun with the Mathematical As­sociation of America was especially valuable because it provided a natural outlet for the committee's recom­mendations on the mathematical curriculum for social science students. Members of the committee spoke at a great variety of meetings of mathematicians and applied mathematicians dealing with training. Mathematicians became aware of mathematical problems in social sci­ence. Courses especially for social science students were beginning to be given in a few mathematics departments and some subject-matter departments. Dartmouth and the University of Michigan were among the leaders, and Harvard now had Bush teaching such a course.

The 1955 eight-week summer training institutes were held in four parts, two at Stanford University and two at the University of Michigan, with two sets of programs at each place, one for students who had studied at least a year of calculus and one for non-calculus students.12

The total number of students accepted was 65. The

11 The following publications resulted: Robert R. Bush, Robert P. Abelson, and Ray Hyman, Mathematics for Psychologists: Examples and Problems, Social Science Research Council, 1956; Gerard Debreu, Theory of Values: An Axiomatic Analysis of Economic Equilibrium, John Wiley Be Sons, 1959; Samuel Goldberg, Introduction to Difference Equations: With Illustrative Examples from Economics, Psychology, and Sociology, John Wiley &: Sons, 1958; Universal Mathematics, Part I, Functions and Limits: A Book of Experimental Text Materials, pre­liminary edition, University of Kansas Book Store, 1954; Universal Mathematics, Part II: Structure in Sets, preliminary edition prepared by W. L. Duren, Jr. and D. R. Morrison, Tulane University, 1955.

12 The faculty at Michigan included Robert L. Davis, Gerard Debreu, Harold Kuhn, and Robert Thrall; that at Stanford, Robert R. Bush, William G. Madow, Howard Raiffa, and Patrick Suppes.

JUNE 1974

programs weut off without a hitch, or at least as nearly so as such things ever do.

Although the committee recognized the need for more such training sessions, it did not believe that it should sponsor more of them because the necessary materials had now begun to become available, and the ability to conduct such programs had been shown, as mathe­maticians say, by an existence proof. It seemed reason­able to hope that educational organizations would offer further programs on their own initiative.

Consequently the committee, with the approval of the Council's Committee on Problems and Policy, turned to the second stage of its program-providing training in the applications of mathematics to social science prob­lems for both social scientists and teachers of under­graduate mathematics.

In 1957 two summer institutes were held at Stanford. The one for social scientists had five workshop seminars with about seven students each, all engaged directly in research. The workshop on learning was directed by Bush; that on activity analysis, by Robert Dorfman; international trade, Lionel W. McKenzie; language and communication, George A. Miller; decision processes and measurement theory, Patrick Suppes. This institute was attended by alumni of the 1953 and 1955 institutes, and others with similar preparation. One product of the learning workshop was Studies in Mathematical Learn­ing Theory, edited by Bush and William K. Estes.ls

Thrall directed the second institute, essentially the one proposed to the Council by Bush in 1953, intended to acquaint college teachers of mathematics with appli­cations of their courses to social science problems in the way they were familiar with applications in the fields of engineering and physical science.H The institute was cosponsored by the Mathematical Association of Ameri­ca, whose representative in the planning was Albert Tucker. Most of the mathematicians attending had 5-15 years of teaching experience. In addition to about 12 hours of lectures there were one or two workshops each week. The group also reviewed the committee's state­ment of recommended policies published in Items, June 1955.

These two institutes completed a five-year program of the committee. As its chairman, Madow summed up progress as follows (Items, December 1957):

Let us compare the present situation with that in which the committee was established. The differences are perhaps most strik­ing in the following respects:

18 Stanford Mathematical Studies in the Social Sciences, lil, Stanford University Press, 1959.

U The other faculty members were William K. Estes, Tjalling Koop­mans, and R. Duncan Luce.

21

There is greater acceptance among social scientists of the desir­ability. if not the need. of the study of mathematics.

There is greater understanding among mathematicians that the applications of mathematics in the social sciences are serious mathematics.

There is fairly general acceptance of the need for the revision of the undergraduate mathematics curriculum to make it more satisfactory for the social scientist. Texts to facilitate this develop­ment are appearing. and the number of courses is growing.

There is fairly general agreement on the type of mathematics curriculum that social scientists should study.

Many universities and colleges have mathematical and statisti­cal faculty members who have worked sufficiently on applications of mathematics in the social sciences to be able to make such ap­plications. advise others about them. and teach mathematics to

social scientists. Many universities and colleges have social science faculty mem­

bers who have used mathematics sufficiently in their own research to be able to guide. advise. and teach others to apply mathematics to social science problems.

There is more general understanding of the role of mathematics in social science research: that it is best used when casually used. that it is not a substitute for social science thinking. that it is not a temporary development but a continuing development.

The number of social scientists who use mathematics or can read mathematical social science literature has greatly increased.

Madow then went on to describe possible new direc­tions of work, emphasizing that previous work had been oriented primarily toward training, and that next steps should deal with research. In recognition of the new mission and of the extensive service of the members, the old committee was disbanded and a new one appointed in 1958, the Committee on Mathematics in Social Sci­ence Research, with Madow as chairman.16

By 1959 a report 16 prepared for the Committee on the Undergraduate Program in Mathematics showed that 15 percent of students in four-year colleges had available a course in finite mathematics. This approach was new but was encouraged by both its sponsor and the Council's committee.

Under the chairmanship of George A. Miller, the committee held a three-day conference, at the Center for Advanced Study in the Behavioral Sciences early in 1959, on the use of mathematics in undergraduate courses in psychology. The 21 participants were psy­chologists who had either taught introductory psycholo­gy in mathematical terms or planned to do so.

The committee also sponsored two exploratory proj­ects initiated in the seminar of 1959. James G. March undertook to identify areas of political science where

1G The other members during 195~O were Carl F. Christ. Sanford M. Dornbusch. John G. Kemeny. James G. March. Philip J. McCarthy. George A. Miller. and Anatol Rapoport.

18 F. Mosteller. K. Choi. and J. Sedransk. A Catalog Suroey of College Mathematics Courses. CUPM Report Number 4. Mathematical Associ­ation of America. December 1961.

22

mathematical models would be most profitably devel­oped. And Bernard P. Cohen, Joseph Berger, and J. Laurie Snell met at Dartmouth to consider the appli­cability of mathematical models to research on the behavior of small groups.

In 1960 the committee was reconstituted with Suppes as chairman,17 He suggested to the Council four lines of activity: a postdoctoral research fellowship program for the various social sciences, further training institutes like those held in 1957, establishing a sort of "Woods Hole" laboratory for mathematical work in the social sciences, and a new effort in the field of statistics. He reported the formation of the new Panel on Biological, Management, and Social Sciences of the Committee on the Undergraduate Program in Mathematics. This Panel, chaired by Kemeny, had a fair number of mem­bers from groups previously named. In 1964 it produced Tentative Recommendations for the Undergraduate Mathematics Program of Students in the Biological, Management, and Social Sciences. This set of recom­mendations assumed that students from these areas would begin with calculus, taking-in all-a basic four­semester course mathematics sequence plus a two-year sequence in probability and statistics. The main areas of study were analytic geometry and calculus, including both difference and differential equations, linear alge­bra, and many-variable calculus. It was expected that in addition to elementary probability and statistics there would be complex stochastic models. High-speed com­puters were expected to play important roles in all areas. (Most of the people involved in the various committees discussed here used high-speed computers in their work, but computing did not become a major branch of mathe­matics for these committees.)

The "Woods Hole" idea struck a responsive chord, and it was to be the subject of many further discussions not only by the SSRC committee but also by others working in the area. The awkward point was that the idea of selecting anyone location, except possibly for climate and intellectual atmosphere, did not seem very apt. Neither Woods Hole, Massachusetts nor any other location was envisaged as a site. Everyone found it neces­sary to explain this frequently; it was a phase we appar­ently had to go through.

The committee meanwhile developed a two-year pro­gram of (a) summer research institutes for advanced graduate students and recent recipients of the Ph.D. in social sciences who wanted to apply mathematical models in their research, and (b) senior conferences of

11 The other members during 1960-64 were David Blackwell. James S. Coleman. Clyde H. Coombs. Robert Dorfman. and Howard Raiffa; W. K. Estes served during 1960-61. and R. Duncan Luce. 1962-64; staff: Francis H. Palmer. 1960-62. and Elbridge Sibley. 1963-64.

VOLUME 28. NUMBER 2

established social scientists whose research was mathe­matically oriented.1s

In the summer of 1962 four research institutes were held. Their subjects, chairmen, and locations were as follows: two-person interactions, Cletus Burke, at Stan­ford University; models of social decision-making mechanisms and their implications for political science and welfare economics, John C. Harsanyi, at Princeton University; bargaining, negotiation, and conflict, Har­old W. Kuhn, at Princeton University; psychology of choice and decision, Frank Restle, at Stanford Univer­sity. These sessions had a total of 41 participants plus 4 assistants.

Concurrently a sequence of conferences for senior re­search workers was held at Stanford. The topics treated were learning theory and measurement and choice theory; 13 psychological research workers attended.

During the summer of 1963 two more summer re­search institutes were held: on measurement and data analysis, under the chairmanship of Lincoln Moses, at Stanford University; and on mathematical models of social structure, James M. Beshers, chairman, at the University of Wisconsin. A total of 26 scholars attended these sessions.

During the same summer two senior conferences were held, both at Stanford, one on psychophysics led by R. Duncan Luce and one on learning theory led by Estes and Suppes. Luce also prepared a most informative paper, "The Mathematics Used in Mathematical Psy­chology," 19 in which he thoughtfully discusses the role of models, fundamental measurement, probability mod­els, learning models, preference models, psychophysical models, latency models, psychometrics, and nonnumeri­cal models. Then in the light of these illustrations he explains aspects the student needs of fundamentals, analysis, probability theory, algebra, topology, and geometry.

In the summer of 1964 the committee sponsored one research training institute and two senior conferences. At Stanford, Anatol Rapoport and Julian H. Blau con­ducted a six-week institute on mathematics for political scientists and sociologists. There were 25 participants. Also at Stanford, Bush and Luce held a two-month senior conference on mathematical learning theory, with 6 additional participants. At the University of Rochester Lionel McKenzie led a six-week senior conference on mathematical models of economic growth, with a daily attendance of 10-13 participants.

The main topic of discussion by the committee dur­ing 1964 was the possibility of establishing an Institute

18 Funds for this program were provided by the National Science Foundation.

19 American Mathematical Monthly, April 1964. pages 364-378.

JUNE 1974

of Mathematical Social Science, essentially the "Woods Hole" idea, but adapted to mathematical work. The discussions were held jointly with the Council's Com­mittee on Simulation of Cognitive Processes.

One must not get the impression from this version of history that thought, planning, and development are limited to the specific committee assigned a task. For example, a strong push in the "Woods Hole" direction was provided in a report from the Behavioral Sciences Subpanel of the Life Sciences Panel of the President's Science Advisory Committee, entitled "Strengthening the Behavioral Sciences," April 20, 1962. It includes the following recommendation (page 14):

There is a special need for summer institutes. or other short­term instructional arrangements. to bring research workers and selected teachers up to date in new techniques and experimental procedures. Experience suggests that such arrangements would be more effective if they were set up on a relatively long-term basis and in a suitable research environment. There should be a small core staff to plan during the entire year for the instruction pro­gram as well as to work on research. As one specific step in this direction. such a special instructional program should be centered upon the application of mathematics and computers to the be­havioral sciences.

The "experience" mentioned here is, of course, that of the Council's committees.

At this time one natural group to consider a response to the Life Science Panel's recommendation was the Division of Social Sciences of the National Science Foundation, then headed by Henry W. Riecken, who had been a student member of the first SSRC summer train­ing institute in mathematics (1953). And so the people involved in the future development were beginning to

include scholars who had received some idea of the program by actual participation.

After the usual lengthy discussions, a plan was sub­mitted to the Divisions of Social Sciences and of Scien­tific Personnel and Education of the National Science Foundation with a request for $500,000 to set up an organization under the wing of, but not necessarily geographically at, the Center for Advanced Study in the Behavioral Sciences. By now, of course, extensive ne­gotiations were involved. Among those who worked es­pecially hard in preparing the proposal to the Foun­dation were Suppes, Bush, Estes, and Luce. They were in Stanford for SSRC conferences. By accident I visited Stanford just at this time and joined some of the work­ing sessions because Suppes could not bear to see idle hands. The name of the newly created organization turned out to be the Mathematical Social Science Board.

During the period from 1964 through 1972, the Math­ematical Social Science Board held 85 events, including

23

some trammg institutes, many conferences, research seminars, workshops, and so on. Some 1,900 people attended these scientific events (the total number of participants enrolled was 1,957, but some attended more than one event). A number of the events were in fields more newly concerned with applications of mathe­matics: anthropology, 9 events; history, 17; linguistics, 15; and political science, 7. In psychology, economics, and sociology many events continue to be held. In 1973 the support provided by the National Science Foun­dation was at the level of $220,000 per year; there were 30 further events, including 7 workshops on computers and information processing psychology.

In addition to continuing to facilitate research by means of conferences and workshops, the Board is cur­rently attempting to gain support to resume training (not now permitted by the NSF), not in elementary mathematics, but in more specialized techniques and models. These involve both advanced mathematics and particular techniques and models that have grown up in one area of social science but appear to be transfer­able to another. One example is the use of econometric models being made in history. The proposed training would probably be provided by means of small summer institutes and year-long pre- and postdoctoral fellow­ships. In addition, the Board is looking into the need for national facilities, such as a computer network suitable for simulation research to be available at many colleges and universities.

Since the establishment of the Mathematical Social Science Board, the Council has continued to foster new developments in this field. One example of later efforts is the publication in 1973 of Structural Equations Mod­els in the Social Sciences, edited by Arthur S. Gold­berger and Otis Dudley Duncan. The production of this volume had its origin in a discussion at the Council's September 1968 meeting at Skytop, which led to a con­ference on causality. Out of that meeting came a research conference directed by Goldberger at the University of

Wisconsin; the volume is the product of the confer­ence.20 Since simultaneous equations models have be­come extremely important in several social sciences, this exposition is most welcome.

As a result of the same 1968 discussion, the Council also sponsored preparation and publication of a forth­coming volume on social experimentation, its problems and methods. This covers planning and evaluation studies that use the experimental method (random as­signment of subjects to treatment) in substantive areas such as income maintenance, education, and health. It is designed for use by managers and staff of governmental intervention programs and by social scientists who must evaluate such programs. The manuscript has been pre­pared by the Committee on Experimentation as a Method for Planning and Evaluating Social Interven­tion, chaired by Riecken.21

Here we have followed the development of mathemati­cal work in the social sciences in this country from very tentative training efforts to systematic large-scale pro­grams. In 1953, in addressing the Council, I explained that the results would look meager for a long time, but that "the carefully thought out program of the Coun­cil's Committee on the Mathematical Training of Social Scientists may move us slowly to a place we will be proud of in 20 years." I think that has happened.22

20 The conference was supported by the Russell Sage Foundation; the volume was published by Seminar Press.

21 The other members are Donald T. Campbell. Nathan Caplan. Thomas K. Glennan. Jr .• John W. Pratt, Albert Rees. and Walter Williams; staff: Robert F. Boruch. The volume. Sodal Experimenta. tion, edited by Riecken and Boruch. will be published by Academic Press in 1974. The project is supported by the National Science Foundation.

22 I wish to express appreciation to Michael L. Brown. William G. Cochran, Preston S. Cutler. John P. Gilbert. David C. Hoaglin, Nan Hughes. Eleanor C. Isbell. John G. Kemeny. William H. Kruskal. R. Duncan Luce. William G. Madow. George A. Miller. Franco Modigliani, Henry W. Riecken. Elbridge Sibley. Robert Thrall, and Donald R. Young for advice and new material in the preparation of this history. Preparation was facilitated by Grant 32327Xl of the National Science Foundation.

THE COMMITTEE ON ECONOMIC GROWTH, 1949-68

UNDER THE effective and imaginative leadership of its chairman, Simon Kuznets, the Committee on Economic Growth functioned for nearly twenty years and con­tributed notably to both empirical and theoretical in­quiry into economic growth and its determinants. The

• The author is James B. Duke Professor of Economics at Duke University. As a member of the board of directors of the Social Science Research Council, 1945-50. of the Committee on Problems and Policy for an even longer period. 1949-61. and as its chairman. 1958-61. he

24

by Joseph J. Spengler ""

committee was established 'by the Social Science Re­search Council in response to a memorandum submitted to it by Kuznets in 1948. In this memorandum he urged the need of a committee to explore how the study of economic growth might best be carried on, "to establish

was actively associated with the development of Council programs in many fields during those years. A member of the Committee on Eco­nomic Growth throughout its existence. 1949-68. he participated as organizer. author. or discussant of papers at many of its conferences

VOLUME 28. NUMBER 2

how -fruitful empirical study of economic growth can best be planned; and, in areas in which the groundwork is not ready for empirical studies, to stimulate thinking and discussion leading toward formulation of the neces­sary framework." Kuznets pointed out that such a com­mittee would have to include members from a variety of disciplines other than economics, since economic devel­opment was affected by many factors such as science and technology, natural resources, the efficiency of the state and other organizations and social mechanisms, and the whole pattern of culture. Three years after the termi­nation of the committee, the Alfred Nobel Memorial Prize in Economic Science was awarded to Kuznets for his illuminating inquiries "into the economic and social structure and process of development."

The original membership of the committee included, besides the chairman, Edgar M. Hoover, Wilbert E. Moore, J. J. Spengler, J. M. Clark, and Lauriston Sharp. In 1950 Shepard B. Clough and Morris E. Opler joined the committee, in effect succeeding Clark and Sharp. Richard Hartshorne, added in 1953, served through 1964. Melville Herskovits replaced Opler in 1955 and served until his death in 1963. Bert F. Hoselitz joined in 1954, succeeding Clough, and served through 1968. Neil J. Smelser was a member during 1960-65; Richard A. Easterlin, 1964-68; and Moses Abramovitz, 1965-68. The late Paul Webbink, Vice-President of the Social Science Research Council, was its liaison officer with the committee. Webbink contributed remarkably for two decades to the removal of all the barriers, financial and otherwise, in the way of the committee's work, and pro­vided wise counsel as well-contributions paralleling those he made to research and its organization in many other areas of concern to the Council.

BEGINNINGS

The emergence of the study of economic growth into the forefront of economic inquiry found economists and other social scientists unprepared, since, as Kuznets ob­served, "in economics the problems of the growth of

and at its 1956 Interuniversity Sumlner Research Seminar on Theories of Economic Growth, and he contributed as author or editor to the important publications that resulted. He was a member of the Com­mittee on Grants-in-Aid, 1947-49, and of the Committee on Auxiliary Research Awards, 1961-62; a consultant to the Committee on Histori­ography in the preparation of its report The Social Sciences in Historical Study (Council Bulletin 64, 1954); a participant in conferences and author of papers published under the auspices of the Committee on Comparative Politics; and a participant in, and contributor to the publication that resulted from, the 1959 conference on the history of quantification in the sciences held by the Joint Committee on the History of Science (NRC - SSRC).

The present article was written at the invitation of the President of the Council as part of the commemoration of its 50th anniversary year.

JUNE 1974

nations have been lying dormant practically since the middle of the nineteenth century." 1 Even so, at the time the committee was established, many forces-recent and current-were converging to intensify interest in economic development, theretofore the concern prin­cipally of economic historians.

Both practical and theoretical concerns were arousing interest in economic development, among them recently expressed fears of economic stagnation and even of de­population, increasing emphasis on "economic plan­ning," and differing inferences from J. M. Keynes' General Theory. On the theoretical level R. F. Harrod and Evsey Domar were drawing attention to the need to dynamize Keynes' treatment of investment and thereby introduce a growth dimension. Earlier, in the later 1920's and the 1930's, inquiries carried out under the auspices of the National Bureau of Economic Re­search-by Kuznets, Arthur F. Burns, Frederick C. Mills, Wesley C. Mitchell, and others-had directed attention to the importance of growth phenomena. Recrudescence of interest, especially on the part of demographers, in interpretations of population growth and diffusion processes stimulated interest in turn in the nature of the growth processes that might underlie S-shaped aggregate growth curves. It was becoming evi­dent, also, that conventional inputs did not account for all growth; thus, Tinbergen noted the significance of technical progress (Weltwirtschaftliches Archiv, May 1942), and T. W. Schultz was searching for the causes of unexplained increases in agricultural output. In 1940 Colin Clark published his estimates of the extent of world poverty (Conditions of Economic Progress) and called attention to the major sources of economic growth, some of them unconventional. Cobb-Douglas production functions were at hand for use by those searching for unexplained residual output and its sources.

At this time there was also concern about the possible inadequacy of natural resources. The success of the Marshall Plan was raising unfounded hopes respecting the applicability of such plans in economically under­developed countries. The establishment of the United Nations and related international agencies, together with plans for inquiries into economic conditions and their improvability in many parts of the world, was pointing to the need for hard data and understanding of growth processes. There had long been concern in ad­vanced countries over the weakness of forces of economic convergence and the consequent laggardness of some regions (e.g., the "South" in the United States and Italy). Accordingly, interest in the stimulation of the

1 Items, June 1959, page 13.

25

development of "backward" regions and in the role played therein by cities had reinforced other sources of interest in growth and development in general.

Commenting in 1959 on how the economic growth of nations had "become a topic of leading concern to economists and other social scientists during the last 15 to 25 years," Kuznets found reasons for the shift of this topic into central focus to lie "in the emergence of widely perceived problems, in this instance, the danger of secular stagnation of the 'mature' capitalist econo­mies, the challenge of forced industrialization under authoritarian auspices behind the iron curtain, the risks of failure in the struggles for development on the part of the 'underdeveloped' areas of the world." 2

The favorableness of the climate of opinion to exten­sion and intensification of the study of growth, as well as the catholicity of the kinds of inquiry indicated, made it relatively easy to place inquiry on an interdisciplinary base. It must also have made it easier both to enlist the interest of economists and other social scientists and to win foundation support.

While the Committee on Economic Growth could accomplish much more than a mere catalytic agent, its own capacities, even if reinforced by those of knowl­edgeable and interested individuals, were limited. But it could pursue at least five courses of action: (a) initiate inquiries into strategic or critical areas, always with the aid of competent scholars; (b) work with organized bodies whose primary concerns might overlap in some degree those of the committee; (c) carry to completion inquiries into areas by bodies which for one reason or another had not pursued them as far as was warranted by the state of the data and the importance of the prob­lems under consideration; (d) facilitate inquiry by both individual scholars and ad hoc or organized groups into relevant areas; and (e) foster the accumulation of needed statistical and related information, past and current, by encouraging the initiation of such accumulation, where nonexistent, and the expansion of efforts at accumula­tion already under way.

The procedure adopted by the committee was as follows: Assessment of past action and evaluation of prospective alternatives would be made at meetings for this specific purpose. When a course of action was de­cided upon, a subcommittee would prepare a program which would in tum be examined and presumably ap­proved with modification by the whole committee. Steps would then be taken to identify and select participants and deal with logistical problems. Upon completion of an activity (e.g., a conference), the worthwhileness and feasibility of publication of the findings would be con­sidered. Not all the projects proposed could be brought

21bid.

26

to fruition, because of lack of data or other conditions. For example, the role played by large corporations in the economies of small countries, today a subject of inquiry under United Nations auspices, did not prove as easy of examination as anticipated. Similarly, study of cultural conditions bearing on growth in nonmodern countries did not prove sufficiently feasible.

A great deal of groundwork remained to be done after 1948. There had been considerable inquiry into growth, of course, even though much theoretical and empirical work remained to be undertaken before many effective studies could be launched. This had been indicated by Abramovitz 8 and by participants in a Universities­National Bureau Committee for Economic Research conference held in 1948. Kuznets noted, in his foreword to the volume that resulted from that conference, "the relative scarcity of sustained empirical work, and the absence of an agreed upon body of theoretical hypothe­ses concerning factors determining economic growth." 4

Since a comparative quantitative approach was essen­tial to the making of international comparisons and the derivation of explanatory information, it was essential that competent scholarly personnel be enlisted in West European countries, Japan, and Australia to undertake studies in these countries for comparison with one another and with those made in the United States and Canada. Wider historical perspective was thus gained,1I not only in the wealth and income growth studies out­side the United States, for the promotion of which Kuz­nets assumed responsibility, but also in the committee's final series of studies of postwar economic growth, di­rected by Kuznets and Abramovitz.

CONFERENCES AND SEMINARS

A major instrument of the committee was the inter­disciplinary conference focused on a set of the factors which affected or might affect economic growth. Inter­university summer seminars were also considered, but only one of these was held, under the leadership of Bert F. Hoselitz, at Dartmouth College, during part of the summer of 1956.6

Sixteen major conferences, eight of them jointly co­sponsored with other organizations, were held during the years 1951-64. Their visible products were of six kinds: (I) 11 books embodying the papers presented, usually with revisions growing out of conference dis-

8 "Economics of Growth" in B. F. Haley, ed., if Survey of Contempo­rary Economics, Richard D. Irwin, 1952.

4 Problems in the Study of Economic Growth, National Bureau of Economic Research, 1949.

B See Kuznets's accounts in Items, December 1955 and June 1959. o It resulted in Theories of Economic Growth, Free Press, 1960, edited

by Hoselitz.

VOLUME 28, NUMBER 2

cussion; (2) conference papers published in appropriate journals; (3) the setting in motion of streams of inquiry that eventually found expression in the research publi .. cations of participants: (4) opportunities to explore the multifacetedness of growth and developmental processes and become aware of the diverse contributions that could be made by various disciplines; (5) as a sequel to these opportunities, identification of areas of "moderni­zation" (e.g., politics and government) in need of fuller study by other social scientists than economists; (6) stress on quantification insofar as feasible and pursuable with­out undue simplistic abstraction.

There follows a list of these conferences, their sub­jects, locations, dates, and cosponsorship (if any), and indication of the publications resulting. It is evident that in some degree earlier conferences discovered gaps in knowledge that later conferences were designed to fill.

(1) Quantitative Description of Technological Change, Prince­ton University, April 6-8, 1951, cosponsored by the Coun­cil's Committee on Social Implications of Atomic Energy and Technological Change. Papers were published by Joseph L. Fisher, W. R. Maclaurin, and Jacob Schmookler. Two of Schmookler's later studies were continuations of his papers presented at the conference,7 as was the later conference on inventive activity in 1960 (see below).

(2) Economic Growth in Selected Countries-Brazil, India, and Japan, New York City, April 25-27, 1952. The papers were edited by Simon Kuznets, W. E. Moore, and J. J. Spengler and published as Economic Growth: Brazil, India, Japan, Duke University Press, 1955.

(8) Strategic Factors in Periods of Rapid Economic Growth, New York City, April 9-10, 1954. Papers were published by Henry G. Aubrey, Shepard B. Clough, and Penelope Hartland.

(4) The Role of Cities in Economic Development and Cultural Change, University of Chicago, May 24-26, 1954, cospon­sored by the University's Research Center in Economic Development and Cultural Change. The papers were published in Economic Development and Cultural Change, 1954-55, under the editorship of Hoselitz.

(5) Investment Criteria and Economic Growth, Massachusetts Institute of Technology Center for International Studies (cosponsor), October 15-17, 1954. The collected papers were made available in photo-offset, by the Center.

(6) Entrepreneurship and Economic Growth, Harvard Uni­versity Research Center in Entrepreneurial History (co­sponsor), November 12-18, 1954. Papers were published by John Pelzel, Fritz Redlich, and Charles Wilson.

(7) The State and Economic Growth, New York City, October 11-12, 1956. The papers were edited by Hugh G. J. Aitken and published by the Council as The State and Economic Growth,1959.

(S) Commitment of the Industrial Labor Force in Newly De­veloping Areas, Chicago, March 2S-30, 1955. The papers

T Invention and Economic Growth, Harvard University Press, 1966, and Patents, Invention, and Economic Change, edited by Zvi Griliches and Leonid Hurwicz, Harvard University Press, 1972.

JUNE 1974

were edited by W. E. Moore and A. S. Feldman and pub­lished by the Council as Labor Commitment and Social Change in Developing Areas, 1960.

(9) Natural Resources and Economic Growth, University of Michigan, April 7-9, 1960, cosponsored by Resources for the Future, Inc. The papers were edited by Spengler and published as Natural Resources and Economic Growth, Resources for the Future, 1961.

(10) Economic and So~ial Factors Determining the Rate and Direction of Inventive Activity, University of Minnesota, May 12-14, 1960, cosponsored by the Universities-National Bureau Committee for Economic Research. The papers were published as The Rate and Direction of Inventive Activity, Princeton University Press, 1962.

(11) Relations between Agriculture and Economic Growth, Stanford University, November 11-12, 1960. Papers were published by William H. Nicholls, Philip M. Raup, and Boris C. Swerling.

(12) Economic Trends in the Soviet Union, Princeton, May 6-S, 1961. The proceedings were edited by Abram Berg­son and Simon Kuznets and published as Economic Trends in the Soviet Union, Harvard University Press, 1968.

(18) Economies of Sub-Saharan Africa, Northwestern University, November 16-18, 1961. The papers were edited by Melville J. Herskovits and Mitchell Harwitz and pub­lished as Economic Transition in Africa, Northwestern University Press, 1964.

(14) The Role of Education in the Early Stages of Economic Development, University of Chicago Comparative Educa­tion Center (cosponsor), April 4-5, 1968. The papers were edited by C. A. Anderson and Mary Jean Bowman and published as Education and Economic Development, AI­dine Publishing Company, 1965.

(15) Demographic and Economic Trends in the Developing Countries, New York City, October 10-12, 1968, cospon­sored by the Population Council. This conference was a sequel to an earlier one sponsored by the Univesrities­National Bureau Committee for Economic Research. Papers were published by Paul Demeny, M. A. EI-Badry, Carmen A. Mir6, and George J. Stolnitz.

(16) Social Structure, Social Mobility, and Economic Develop. ment, San Francisco, January 80 - February 1, 1964. The papers were edited by Neil J. Smelser and S. M. Lipset and published as Social Structure and Mobility in Eco­nomic Development, Aldine Publishing Company, 1966.

FOREIGN PUBLICATIONS

The results of the research of foreign scholars, whose participation in the committee's earlier series of studies was enlisted by Kuznets as its chairman, have been of especial importance for subsequent research in its field. The published reports are far too numerous to list here, including over 50 articles, monographs, and chapters of books, by some 25 scholars from a dozen countries, in at least 7 languages. (Lists of all publications that have resulted from the committee's activities are available from the Council upon request.)

With acceleration of economic growth after World War II, the committee in 1963, in response to interest

27

expressed by the Ford Foundation,8 undertook sponsor­ship of parallel studies of the sources of this postwar economic growth in the United States, Japan, and four European countries-France, Germany, Italy, the United Kingdom-to which Sweden was later added. A meeting of the European and other scholars engaged in this undertaking was held in London in January 1964, the better to coordinate the studies and overcome com­mon deficiencies. Not all of these studies have been completed, but some 20 published reports have appeared and more are in preparation.

KUZNETS' PUBLICATIONS

Since Kuznets was the leading spirit in the commit­tee's work and in its efforts to launch studies in various countries, many of his growth-oriented studies completed during the lifetime of the committee are integrally re­lated to its work. Other than his several papers included in the conference volumes already cited and a series on the "Quantitative Aspects of the Economic Growth of Nations," which appeared as supplements to issues of Economic Development and Cultural Change, 1957-67, Kuznets' major publications relating to economic growth included the following: Six Lectures on Economic Growth, Free Press, 1959; Postwar Economic Growth: Four Lectures, Belknap Press of Harvard University Press, 1964; Economic Growth and Structure: Selected Essays, W. W. Norton & Co., 1965; Modern Economic Growth: Rate, Structure, and Spread, Yale University Press, 1966; Economic Growth of Nations, Belknap Press of Harvard University Press, 1971; Population,

S As reported by Paul Webbink, Items, December 1971, page 42, "The committee's initial modest outlays were defrayed from Council funds. Subsequently the Rockefeller Foundation enabled Simon Kuznets to spend half of his time during five years on research on economic growth, a contribution later taken over by the Ford Foundation, which also provided generous funds for most of the committee's foreign studies and its conferences."

Capital, and Growth: Selected Essays, W. W. Norton & Co., 1973.

CONCLUSION

It is not possible within the compass of a short report to enumerate or assess the contributions of the com­mittee's efforts to the development and improvement of the study of economic growth. The publications resulting from the activities of the committee have in­creased manyfold our knowledge of the forces shaping growth and development in the West and in Japan since and even before the early nineteenth century. They pre­sent much of this knowledge in quantitative terms, thus enabling students to sift out the separate contributions of various kinds of inputs while appreciating the long­time dimensions of some of the forces at work. The studies make for caution against accepting at face value some of the commonly employed indicators of develop­ment and welfare and against too exclusive resort to simplistic models whence recalcitrant elements have been abstracted. From the standpoint of the Council, especially important is the evidence of the complexity of growth processes, of the inadequacy of any particular discipline to cope satisfactorily with this complexity, and of the need for solid interdisciplinary approaches to many of the problems faced, some old and some precipi­tates of the growth process itself.

By way of closing this report and giving assurances to scholars, employed and unemployed, that much remains to be done, I can do no better than refer the reader to Simon Kuznets' Nobel Memorial Lecture entitled "Mod­em Economic Growth: Findings and Reflections" 9 and addressed to (among other aspects) the social implica­tions of modem economic growth, the significance of what is known for less developed countries, and the problems emerging in both more and less advanced countries.

o Reprinted, American Economic Review, June 1973, pages 247-258.

JAPANESE INDUSTRIALIZATION AND ITS SOCIAL CONSEQUENCES: A REPORT ON THE CONFERENCE HELD ON AUGUST 20-24, 1973

JAPANESE economic development has been a source of fascination for foreigners and Japanese alike, not only in its purely economic context and in a broader social,

• The author is Professor of Economics at Yale University. At the invitation of the Joint Committee on Japanese Studies, American Council of Learned Societies and Social Science Research Council, he

28

by Hugh Patrick •

political, and cultural context but also for purposes of international comparison. Much research by economists has focused on establishing the general contours of

served as chairman of the committee named by it to plan the confer­ence on which he reports here, and he is editing the conference papers Cor publication by the University of California Press.

VOLUME 28, NUMBER 2

japan's process of economic growth in quite aggregative terms. Even discussions of industrialization-particu­larly those in Western languages-have tended to be aggregative in nature. At the same time other social scientists have examined other features of the process of Japanese change, usually taking as given the concurrent process of economic development.

When in 1969 the Joint Committee on Japanese Studies decided to sponsor a series of conferences on various aspects of contemporary Japanese development, industrialization and its social consequences for modem Japan was one of the five topics selected. The writer was invited to organize an international conference on this topic, with the assistance of a committee which included John W. Bennett, Solomon Levine, Kazushi Ohkawa, Henry Rosovsky, Koji Taira, Tsunehiko Watanabe, Kozo Yamamura, and Yasukichi Yasuba. The conference of 28 economists, sociologists, and anthropologists from Japan, the United States, England, and Israel was held on August 20-24, 1973 at the University of Washington Continuing Education Center.1

The theme of Japanese industrialization and its social consequences is broad in scope and in time. In planning the conference, we attempted to cover the entire period from the Meiji era (1868-1912) to the present. Inevi­tably, we had to limit the topics to be considered. The authors of papers were asked to examine certain indus­tries and selected aspects of industrialization itself, and to focus mainly on how selected social factors have been changing in Japan in response to the process of indus­trialization. We excluded from formal consideration the other side of the coin: the social causes of industrializa­tion, as distinct from consequences. All recognized that much of the social change that has occurred in Japan over the past hundred years is intimately connected with the process of industrialization in very complex, interactive chains of causality.

1 The American participants in the conference were John W. Den­nett, Washington University; Martin Dronfenbrenner, Duke University; Robert E. Cole, University of Michigan; Susan B. Hanley, University of Washington: Solomon Levine, University of Wisconsin: Larry Meissner, Yale University; James Nakamura, Columbia University: Hugh Patrick, chairman of the conference: William V. Rapp, Morgan Guarantee Trust Company: Henry Rosovsky, Harvard University: Gary Saxon­house, University of Michigan: David L. Sills, Social Science Research Council; Koji Taira, University of illinois at Urbana-Champaign: John Wisnom, University of Washington: KOlO Yamamura, University of Washington.

The Japanese participants were Masayoshi Chubachi, Keio Univer­sity; Hiroshi Hazama, Tokyo Kyoiku University: Ryoshin Minami, Hitotsubashi University; Chie Nakane, Tokyo University: Hiroshi Oh­buchi, Chuo University: Kazushi Ohkawa, Hitotsubashi University (emeritus): Akira Ono, Seikei University: Michio Sumiya, Tokyo Uni­versity; Ken'ichi Tominaga, Tokyo University: Tsunehiko Watanabe, Osaka University: Yasukichi Yasuba, Kyoto University.

Tuvia Blumenthal, Tel-Aviv University, and Ronald P. Dore, Uni­venity of Sussex, also participated in the conference.

JUNE 1974

Twelve papers provided the foci for the conference discussions. Participants were expected to have read the papers in advance, so it was not necessary for authors to present them. Instead, two or three discussants presented prepared comments on each paper; the author was given the opportunity to reply; and a general dis­cussion followed. The discussions were extraordinarily frank, direct, friendly, critical, and interdisciplinary_ It was not a conference in which most Japanese partici­pants took one position and most American participants took another. Americans criticized Americans and Jap­anese; Japanese criticized Japanese and Americans. I believe the fine rapport was achieved both because of the high level of professionalism of the participants and because most of them already knew each other.

The conference papers can be classified in several ways. Three dealt with important features of three in­dustries, each of which has been significant in Japan's industrialization process and each of which displays characteristics different from the others. Gary Saxon­house examined labor force recruitment and techno­logical diffusion in the cotton-spinning industry; Tuvia Blumenthal analyzed the growth of, technological in­duction in, and the role of government in the develop­ment of the shipbuilding industry; and Kozo Yama­mura examined the origins and growth of the large, general trading firms.

Three papers compared aspects of the economies of scale and production in large and small firms, a theme that came up in many contexts throughout the confer­ence. R yoshin Minami stressed the significance of elec­trification and particularly the development of small electric motors in enabling small firms to compete on relatively less disadvantageous terms with large firms. Yasukichi Yasuba tackled the problems of the emergence and widening of wage differentials by size of firm in a number of industries in both the prewar and postwar periods. William Rapp examined the evolving structure of export production and industrial development in terms of the changing shares of small and large firms in exports.

A major purpose of the conference was to break new ground in exploration of the social consequences of Japanese industrialization, which are many, varied, com­plex, and on the whole relatively unexplored, at least in publications in Western languages_ Perhaps one of the most fundamental changes in Japan is summed up as "demographic transition"; Japanese population growth initially accelerated with industrialization, and the pat­terns of fertility and mortality have changed dramati­cally. Hiroshi Ohbuchi considered this transition, with particular emphasis on the socioeconomic forces bring­ing it about. In another path-breaking effort, Akira Ono

29

and Tsunehiko Watanabe examined changes in income distribution, particularly between rural and urban areas, as a consequence of the process of industrialization. This, too, is an area in which data are poor and not much research has been done on either the historical or the postwar period. Masayoshi Chubachi and Koji Taira examined the concept and facts of poverty, par­ticularly urban poverty, over the course of Japanese industrialization-another important area of research.

The authors of two papers looked at the occupational structure and life styles of the labor force. Hiroshi Hazama summarized and generalized from his monu­mental research (published in Japanese) on the evolu­tion of life styles of industrial workers. Robert Cole and Ken'ichi Tominaga examined the changing occupa­tional structure of Japanese workers and the relative importance of the concept of "occupation" in Japan.

In the final paper, John Bennett and Solomon Levine attempted to expand upon the social consequences iden­tified in other papers and to provide their own over-all assessment. They concentrated particularly on the en­vironmental .impacts of Japanese industrialization in terms of both social costs and public response. Since their paper had to build upon the others prepared for the conference, they were able to present it only in tentative form, focusing mainly on the environmental issues that have developed in the postwar period. Revi­sion of their paper is under way on the basis of the other papers and the authors' own further research.

It is not possible here to summarize each of the con­ference papers; rather, some of the themes touched upon in a number of papers and in the discussions will be ex­amined and stressed.

The discussions fortunately did not bog down in dis­putes over methodology or terminology. Happily, the participants steered away from such vague and complex concepts as "modernization versus Westernization" and "modern versus traditional," although they did consider the concepts of "economic dualism" and "paternalism." While trying to isolate certain issues and utilize case studies fully, the participants struggled with the prob­lem of recognizing that everything relating to the con­ference topic depends on everything else. This was true not only in an input-output sense-that the use of elec­tric motors by small enterprises depended on both electrification and a motor-producing industry, and that shipbuilding and innovations in that industry depended on the availability and improved quality of steel, for instance-but also true of interdependence among a host of economic and social variables. Many of the con­sequences of industrialization have been unintended, or certainly not well understood when they first appeared. Who, for example, a hundred years ago would have an-

30

ticipated the effects of industrialization and urbaniza­tion on fertility and mortality rates?

With regard to methodology, it may be noted that although few papers were explicitly comparative, on the whole the approach was quite comparative. There were few assertions that the Japanese were either unique or were just like Westerners. It was pointed out that surplus labor and wage differentials by size of firm are characteristic of certain developing countries, as well as of Japan. On the other hand, the evolution of the gen­eral trading company to its prewar and contemporary roles is an institutional development not replicated else­where. The participants noted that Japanese firms were particularly skilled at absorbing foreign technology, al­though all were puzzled as to how and why.

In retrospect it seems that the conference was domi­nated by two interrelated themes: the conditions of people as industrial workers (in contradistinction, say, to consumers or farmers), and differences between large enterprises and small. This was mainly the consequence of the topics selected for papers, of the choice of authors, and particularly the choice of participants. It was clear throughout the conference that the most sig­nificant "interfacing" of knowledge, methodology, and interests among the participants from different social sciences had to do with workers-their life styles, occu­pations, mobility, distribution by sex, wage differentials and the causes thereof, etc. The discussions were on the whole fruitful but not always conclusive.

LARGE AND SMALL ENTERPRISES

The participants noted how certain constellations of features characterized particular phenomena when data were not adequate to determine either essential features or the relative importance of various features. For ex­ample, large firms were described as having new, usually imported technologies, skilled male labor, more capital per worker, higher output per worker, higher wages, and a special life style-in comparison with small enter­prises. Large firms were also described as either more or less paternalistic than small, according to the defini­tion used. Nakane stressed the involvement of close personal relationships in paternalism, with discretionary modes of behavior making it applicable mainly to small firms. Dore and Cole contrasted this kind of paternalism with the "managerial" or "institutional" paternalism of large firms in which benefits are determined by im­personal rules rather than by personal relationships. Yasuba suggested that large firms have to pay higher wages and fringe benefits to compensate for their lack of paternalism. Minami regarded this as one aspect of a fundamental behavioral difference between large and

VOLUME 28. NUMBER 2

small (family-owned) enterprises: large firms can be characterized as attempting predominantly to maximize economic goals (profits, growth), while owners of small firms, especially those using unpaid family workers, do not behave as economic maximizers but as the urban equivalents of agricultural households concerned with total income and average income-sharing rather than with marginalist calculations. Participants generally agreed with the propositions on large firms; but the char­acterization of small firms remained in dispute-a mani­festation of the "economic dualism" controversy between those who characterize Japan as having gone through a classical surplus-labor economy phase and those who reject that interpretation in favor of a neoclassicist his­torical model of abundant labor with low productivity and wages equal to its marginal product.

Minami's position reflects his synthesis of two quite different ways in which the participants, following the rather confused literature, used the concept of economic dualism. At one point in the conference they were asked to write down their definitions of dualism, and these were then circulated to clarify use of the concept. One use stressed the phenomenon of wage differentials by size of firm within the same industry. The other concept of dualism was the classical two-sector case, in which labor in the modern, manufacturing, large-enterprise sector is paid its marginal product because owners are profit maximizers, and (surplus) labor in the traditional, agricultural, smal1-scale sector receives more than its marginal profit because owners behave according to some sharing, average, or institutional (constant insti­tutional wage) principle different from profit maximiz­ing. This second concept was in the background of most of the discussion, but was explicitly incorporated into the papers by Minami and by Ono and Watanabe. The latter associated the postwar narrowing of income differ­entials with the ending of the surplus labor phase of Japan's development.

WAGE DIFFERENTIALS

The wage differential issue IS Important in under­standing not only the historical process of development in Japan, but the continuing process in many countries today, since it is a quite general phenomenon with not­able implications for policy in resource allocation. We all well understand that some wage differentials are in­evitable and desirable, for example, differentials arising from occupational differences in skill requirements and in attractiveness of work. These differentials are associ­ated with evolving demands for different types of labor and with evolving supplies of such labor based on educa­tion and on-the-job training. One might also expect

JUNE 1974

regional differences in wages because of local labor markets, moving costs, and differences in cost of living.

Dualism in wage differentials refers to contexts in which they remain-typically by size of firm or by sex­even after adjustments are made for differences in labor skills, abilities, regions, and types of work. Yasuba in his path-breaking study found that even after standardiza­tion for labor skills, wage differentials by size of firm do exist; they started prior to World War I (earlier than previously throught); and they widened within given industries and increased among industries by the 1930's.

The conference did not consider the explanations of wage differentials by sex, a topic on which little research on Japan has been done. Part of the problem is that men and women are usually in different occupational categories; relatively few occupations employ both male and female workers. The writer's own view is that sex discrimination occurs in Japan not so much in paying women lower wages than men for the same work as in preventing women from entering or being promoted into more highly skilled occupations, and in giving women lesser wage increments by seniority, lesser re­tirement benefits, and the like.

A further economic explanation of wage differentials lies in capital market imperfections, whereby differen­tials by size of firm in availability and cost of borrowed funds are greater than differences in degrees of risk of default and in transaction costs of loans. Several papers alluded to evidence that large firms not only have been able to borrow funds at substantially lower interest rates than small firms, but also have greater access to funds. This makes it profitable for large firms to use relatively more capital and less labor than small firms. Two theo­retical implications, supported by empirical evidence, are that the profit per unit of capital would be lower in large firms, and that output per worker would be greater in large firms. These capital market imperfections may result from a variety of possible causes: ignorance of actual risks, a high degree of risk aversion, govern­mental restrictions on interest rates and on operations of financial institutions, non-profit-maximizing behavior by large financial institutions, and/or the oligopolistic power of financial institutions. One analytical dilemma is that while capital market imperfections make it pos­sible for large firms to have more capital and output per worker, there is no explanation of why management actually pays workers more. Clearly more than ability to pay has to be involved.

A number of plausible causes of wage differentials by size of firm after adjustment for normal economic ex­planations were discussed. Cole made the important point that forces which may have caused wage differen­tials initially do not necessarily explain their continu-

31

ation, or their widening or narrowing. The participants also recognized the importance of distinguishing con­ceptually and empirically between issues concerning the existence of such dualism and issues concerning its degree, both in number of workers involved and in the size of wage differentials.

As causes of wage differentials, Dore and others em­phasized institutional features that distinguish large firms from small. Dore pointed to the practices by large firms of hiring workers directly from school, providing formal on-the-job training, and seniority wage increments; to changes in labor legislation; and in the postwar period to the rise of unions. Many economists, particularly Taira, have stressed that the development of permanent employment and seniority increments before World War I, which became more widespread among large firms in the 1920's and 1930's, were rational efforts by management to reduce costs of labor turnover and especially the loss of skilled workers. Thus, these insti­tutional patterns were deliberately created by profit­maximizing entrepreneurs; yet when firmly established, they have taken on an independent character. They are now institutions and have to be accepted as givens by large enterprises. And they have had considerable social consequences, for example, on workers' life styles.

Another variable important in the explanation of wage differentials, according to Yasuba, was the process of induction of foreign technology. Virtually all Japa­nese industries relied on foreign technology. Large firms, not small, typically imported foreign technology and adapted it to their specific organizations. Such technolo­gy required skilled labor, which in part was trained on the job. While some of the skills may have been general enough to be transferred if the worker moved to another firm, some were specific to the particular firm's technology; thus the worker was more productive in that firm than elsewhere. Moreover, the process of diffusion of technology to smaller firms, or indeed to other large firms, was relatively slow. A firm benefited from keeping skilled workers by paying higher wages, recouping the costs of training workers from that com­ponent of their higher productivity that was specific to the firm and hence did not have to be paid out in wages. This theory is supported by Yasuba's evidence that wage differentials were more predominant in industries undergoing rapid technological change. However, one would expect that once an innovation has been diffused to all firms, the productivity differentials and wage differentials would disappear. Thus, to explain the per­sistence of wage differentials, either firm-specific differ­ences in technology that benefit large firms must persist, or the Bow of technology importation and innovation by large firms must be continuous.

S2

This theory is also supported by Saxonhouse's nega­tive evidence on the cotton-spinning industry. Its tech­nology was indeed foreign; virtually all spindles were imported until 1925 from one British company, which maintained a staff of engineers in Japan. The technology did not change rapidly and it was relatively easy to learn. Moreover, Boren, the Cotton Spinners Trade Associ­ation (a cartel), diffused innovations rapidly among all cotton-spinning firms through technical publications, exchange of engineers, and the like. Thus, there were no firm-specific technologies that required firm-specific skills. A firm would not hesitate to hire a competent worker away from another firm. And what do we find? No system of permanent employment of production workers, low seniority increments (presumably reflecting the learning-by-doing that occurred), apparently rela­tively small wage differentials by size of firm, and con­tinuing high turnover of each cohort of entering workers even after factory living conditions improved over the early years of this century (about half of the new en­trants left the firm within six months). In other words, where there was no technological gap, wage differentials were not significant. This is all very neat-except for the fact that the cotton spinners were female, typically young and unmarried. One might argue that in Japan, even in the early stages of industrialization, women were not expected to work permanently to become highly skilled. They were expected to work for a few years and then quit to marry. So perhaps the firm-size wage differ­ential is a phenomenon that developed primarily for males; as permanent employment and seniority incre­ments become increasingly institutionalized in the post­war period, female workers in large firms also benefit, but much less than their male counterparts.

This example points to a major difficulty in the analy­sis of wage differentials and of many other phenomena: there are a number of plausible explanations. It is difficult to assess the relative importance of each because data-particularly for the prewar years-are inadequate, and insufficient research has been done. Moreover, one has to be cautious in specifying the structure of causal relationships; there may well be synergistic interactions among various causal variables.

The dichotomizing between large and small firms, and their workers, is overly simplistic in several respects. In size, enterprises of course range from miniscule to gigan­tic. However, many of the differential features we have noted do change smoothly as the size of firm changes. More important, such dichotomizing may seem to imply greater homogeneity in, say, large enterprises than in fact has existed. This was demonstrated quite clearly in the industry studies reported and in Hazama's paper on the evolving life style of industrial (blue collar)

VOLUME 28, NUMBER 2

workers, all of which referred mainly to relatively large firms.

LIFE STYLES OF WORKERS

Hazama delineated three modal types of workers, and their respective life styles, in the early phase of indus­trialization: relatively skilled male workers in machinery and shipbuilding; female workers in textiles; and male workers in mining. The first category of workers ema­nated in considerable part from artisan occupations, with their life styles based on skills, high mobility, high con­sumption, and short time horizons. Gradually they evolved into "large-enter prise-type workers," diligent, stable, more family-oriented, concerned with security, and with new patterns of consumption and use of leisure time. (They learned their life styles largely from govern­ment workers and white collar workers in large firms.) Today their lives are built around their company, their union, and their family. This type of worker has not only increased in absolute numbers and as a proportion of the labor force; more importantly, he appears as a model whose life style workers in small enterprises attempt to emulate. He is the pacesetter in Japanese life styles, patterns of consumption, tastes, and other forms of behavior and values. As Chubachi and Taira put it, "Japan is as solidly a middle-class society today as any country in the world."

One has been tempted in the past to dismiss female industrial workers as an unimportant category, only temporarily employed, with no distinctive life style, ana~

lytically uninteresting. An unexpected feature of the conference was the rehabilitation, as it were, of analy­sis of the role of women in Japan's industrialization. Cole and Tominaga emphasized that a high proportion of Japanese factory labor was female: over 50 percent until the early 1930's according to the census of manu­factures, somewhat less when workers in the very small firms excluded from the census are counted. Women have worked predominantly in textiles, and in the post­war period in electronics and other light manufacturing. Before World War II they were usually recruited from rural villages. They have lived in company dormitories, with life styles dominated by the firm.

The history of the evolution of the female factory workers' life style has been one of gradual improvements in company-provided living facilities, reduction in work hours, and lessening of company restrictions on personal freedom. In Hazama's words, "the life style in the textile industry has been criticized as approximating that of a desert which drove female factory workers to satisfying their hunger and their sexual desires by having trysts with men." Management voiced great fear for their fe-

JUNE 1974

male workers' morality, concerned too that they would become pregnant. Bad living conditions no doubt ac­counted for the high rates at which girls ran away from their factory jobs. Saxonhouse found that improvements in living conditions made by large firms in the 1920's and 1930's did nothing to slow down the runaway rate. Apparently these problems have been less severe in post­war Japan. Two points are worth mentioning: (1) Dif­ferences between large and small firms in industries which predominantly hire female workers are less than in male-oriented industries. (2) The expectations of young, unmarried females in large enterprises appear not to have changed substantially. Most plan to marry (now an industrial worker, rather than a farmer-a reflection of where the men are), to quit work when they marry, and to adopt the life style of their husband's colleagues. Recently the proportion of married female workers in industry has increased to over 50 percent; most are middle-aged workers who have returned to employment in smaller firms.

In the early phase of industrialization mine workers often were drifters, dropouts from society, hired via subcontracting arrangements with labor bosses, given to hedonism in consumption and in the use of leisure time. Eventually to some degree they evolved into, or more likely were replaced by, large-enterprise-type workers. As drifter types they fall out of Hazama's analysis. Pre­sumably their refuge was low-wage, small firms. At worst they became merged with the urban poor studied by Chubachi and Taira.

Chubachi and Taira note the special poverty prob­lems of minority groups in Japan-burakumin, Ko­reans, Ainu, Okinawans-but focus mainly on the mainstream urban poor, especially in ghettos. The urban poor have a distinctive life style, though not necessarily a "culture of poverty" which is inescapable. Theirs is a poverty of the working poor; in developing countries the poor cannot afford not to work. The occupational characteristics of the very poor in Japan evolved over time in response to the changing demands of industri­alization. Traditionally they were rickshaw pullers, day laborers, street vendors, scavengers. As some occupations died away, more became day laborers, especially in con­struction-related activities, and wage earners, probably in miniscule units of production (simple manufactur­ing). With the ending of labor abundance and with ever growing prosperity, the mainstream urban poor are dis­appearing in terms of both income and status.

THREE INDUSTRIES COMPARED

The heterogeneity of large enterprises and their work­ers is not only in life styles. Case studies of the cotton-

S!!

spinning industry, shipbuilding, and the general trad­ing companies reveal significant similarities as well as differences. In all three industries large firms are pre­dominant. The economic importance of these industries has been substantial over the course of Japanese devel­opment. In a sense, modern industrialization began and thrived until World War II with cotton spinning; now cotton has been largely replaced by synthetics. Ship­building, with an equally long history, has emerged as a major world competitor, producing half the world's merchant ships over the past decade. General trading companies not only have long been important, but their role is probably growing. All three industries have been importantly involved in foreign trade.

As for major differences, workers are mainly white collar in trading companies, male skilled (blue collar) in shipbuilding, and female semiskilled in cotton spin­ning. Life styles historically were different but, as we have seen, they have become increasingly homogeneous in many respects. All three industries relied initially on foreign sources of technology and know-how. This tech­nology was largely embodied in imported machinery in cotton spinning, less so in shipbuilding, and even less so in trading. Reliance on foreign technology decreased sharply in cotton spinning in the 1920's, while it has con­tinued to be important in shipbuilding, though perhaps domestic adaptation and improvement have had great­er impact. In trading companies, the high reliance on foreign methods had declined by 1900.

Most important, perhaps, have been industry-specific differences in the nature and degree of government sup­port. On the whole, the cotton-spinning industry de­veloped without special government assistance. In contrast, the shipbuilding industry has always relied heavily on government support-military orders before the war, direct subsidies, subsidies to Japanese shipping firms buying Japanese-built ships, subsidies in the form of low-interest export credits to foreign buyers. The trading companies are an intermediate case. They were subsidized heavily in the nineteenth century in order to reduce the share of foreign trading firms in Japan's trade. Since then they have perhaps been on their own. How long it took these industries to become competitive is of interest: within about 15 years cotton spinning firms were exporting substantially; it took shipbuilding some 70 years.

SOCIAL CONSEQUENCES

The social consequences of Japanese industrialization have been profound, but our understanding of them is tentative and far from conclusive. For example, a basic feature of the past century has been the tripling of

lI4

japan's population-essentially all of the increase has been in urban areas. Both accelerated population growth and increased urbanization have been conse­quences of industrialization; in turn they have had further significant, complex social consequences. No paper prepared for the conference dealt directly with urbanization, but it underlay Hazama's discussion of the evolution of workers' life styles, Cole and Tominaga's paper on changes in occupational structure, and of course Chubachi and Taira's treatment of the urban poor-to cite only a few examples.

Industrialization and urbanization were significant causes of the decline in fertility observed since 1925. Women married later. Gross reproduction rates were lower in industrial than in agricultural prefectures and declined more rapidly. These factors are probably also important in explaining the increases in fertility and in population growth rates during the Meiji era, though the data are too meager to test hypotheses. Ohbuchi made the conference participants aware of the current controversy over Meiji population statistics in apprais­ing the relative merits of several different projections. Ohbuchi also used cross-section regressions over time to test the effects of industrialization, urbanization, level of income, education, mortality, and housing on the decline in postwar fertility by age groups. He was able to explain about two thirds of the differential fertility by regions for the 1950's, and somewhat less for the 1960's.

Industrialization was the means of increasing family incomes and hence levels of living, providing the ma­terial resources for changes in life style. Increases in incomes and consumption were moderate, with some setbacks, until World War II. The war was dis­astrous; it took some 17 years to restore living levels. However, rises in average levels tell nothing about changes in the distribution of income, what is happening to certain groups absolutely and relatively. It is possible that all the benefits of industrialization accrue to only a few, but that does not appear to be true in an absolute sense even for prewar Japan. It appears that the material living conditions of virtually all Japanese improved be­tween the early Meiji years and the mid-1930's. How­ever, this is far from completely demonstrated; we need to know more about the rural poor and the poor among minorities. Yet the inequality of income distribution widened up to the mid-1930's, within both agriculture and other sectors, as higher-income families became rela­tively better off. After World War II the distribution of income became significantly more nearly equal, a consequence of zaibatsu dissolution and land reform in inflationary circumstances and, in the last 15 years, of the ending of abundant labor supplies with attendant

VOLUME 28, NUMBER 2

sharp rises in wages, especially for relatively unskilled workers. Data on income distribution are abysmally poor for the prewar period, and poor even for the post­war period. Probably equally important have been the major changes in the distribution of wealth. Owners of land and corporate stock have gained dramatically in comparison with holders of bonds, loans, and savings deposits.

In organizing the conference we were well aware that there have been many social costs of Japanese industriali­zation, but we did not attempt to provide a compre­hensive accounting of those costs. Their estimation is empirically and conceptually difficult (as is true of social benefits, too). What is one group's benefit may be ano~her group's loss. Comparisons and weighting involve value judgments. For example, do we regard the increase in importance of nuclear families-a consequence of industrialization-as good or bad? Or, how do we com­pare increases in absolute incomes with worsening in relative income distribution? Chubachi and Taira, in

examining urban poverty, concluded that it was not caused by industrialization; indeed, the urban poor are eventually absorbed-but probably the last and the least to benefit. On the other hand, the absolute worsening of the environment-pollution of air, water, and land, crowding, and the like-is a social cost of growth, a cost which only recently, as income levels have risen and pollution explosively increased, has become widely perceived.

Good research begets frustration as well as temporary euphoria. It heightens awareness of unanswered ques­tions, and the need for further research. The conference demonstrated that Western knowledge of the Japanese economy and society has come a long way from that of prewar years or even a decade ago. We have a much stronger core of knowledge from which to operate. Yet, we must admit that Western social scientific knowledge and understanding of Japan are still rudimentary. It is hoped that the conference will stimulate fresh efforts in research on the issues discussed by its participants.

PERSONNEL FREDERICK BURKHARDT TO RETIRE

AS PRESIDENT OF ACLS

The retirement this summer of Frederick Burkhardt as President of the American Council of Learned Societies and hence as collaborator in many joint enterprises of the ACLS and SSRC is deeply regretted by officers and staff of the SSRC who have been privileged to be associated with him in these activities. The 17 years of his presidency have seen a remarkable growth and strengthening of the re­lations between the two Councils, centered mainly on the development of programs of research and training for research in most of the major areas of the world, and of scholarly exchanges with them. When he took office in 1957 there was one joint area committee of the Councils, that on Slavic Studies, appointed in 1948. The number has in­creased to 13, and a few more existed for various terms during the 17 years. Outstanding among these was the Joint Committee on the Foreign Area Fellowship Program, 1962-73, whose offerings are now integrated with the programs of the joint committees on the separate foreign areas. The intellectual and administrative tasks represented by this growth and change in foreign area programs have been numerous and demanding of time and talent. Without the wide knowledge of fields and personnel, vigorous attention, imagination, and wise counsel of Frederick Burkhardt these tasks might have been too formidable. His colleagues at the SSRC feel deeply indebted to him and will miss his skills, his wit, and his abundance of common sense.

Looking to the future, they extend a warm welcome to Robert M. Lumiansky, who will become President of the

JUNE 1974

ACLS on July I, 1974, and look forward to working directly with him. Mr. Lumiansky will succeed Mr. Burkhardt as Chairman of the International Research and Exchanges Board and as a member of the Conference Board of Associ­ated Research Councils.

GRANTS FOR SOUTH ASIAN STUDIES

The Joint Committee on South Asian Studies~ cospon­sored with the American Council of Learned Societies (which administers its program)-Charles J. Adams, Edwin D. Driver, Ainslie T. Embree, Rosane Rocher, Susanne H. Rudolph, John W. Thomas, and Helen E. Ullrich-at its meeting on February 23, 1974 awarded grants to the follow­ing 14 scholars:

Harry W. Blair, Assistant Professor of Political Science, Bucknell University, for research on the election process and socioeconomic determinantS in Bihar

Bernard S. Cohn, Professor of Anthropolo~ and History, University of Chicago, for research on Imperial rituals and ideologies in nineteenth-century South Asia

Jean Ellickson, Assistant Professor of Sociology and Anthro­pology, Western Illinois University, for research on the relationship between dietary changes and food beliefs in Bangladesh

Suzanne Hanchett, Assistant Professor of Anthropology, Queens College, City University of New York, for re­search on ritual and social structure in South India

Savak Katrak, Associate Professor of Foreign Area Studies, State University of New York, College at Oneonta, for research on the life and thought of Dadabhai Naoroji

David N. Lorenzen, Professor of Oriental Studies, College of Mexico, for research on North Indian devotional Hindu sects of the fifteenth and sixteenth centuries in their sociopolitical context

!5

Michelle B. McAlpin, Visiting Assistant Professor of Eco­nomics, Reed College, for research on farmers' responses to recurrent threat of famine: agriculture in Bombay Presidency, 1840-1939

Morris David Morris, Professor of Economics, University of Washington, for research on effects of urban development on rural economic growth in Gujarat, 1800-1914

Gananath Obeyesekere, Professor of Anthropolgy, Uni­versity of California, San Diego, for research on religion and social change in modern Sri Lanka

Theodore Riccardi, Jr., Assistant Professor of Middle East Languages and Cultures, Columbia University, for a study of traditional Nepalese historiography

Donald E. Smith, Professor of Political Science, University of Pennsylvania, for research on secularization and politi­cal change in Sri Lanka

Myron Weiner, Professor of Political Science, Massachusetts Institute of Technology, for research on early twentieth­century migrations in India and their effects on ethnicity

Theodore P. Wright, Jr., Professor of Political Science, Graduate School of Public Affairs, State University of New York at Albany, for research on the politics of the Muslim minority in India

Lawrence Ziring, Professor of Political Science, Western Michigan University, for research on the Pakistan Muslim League, 1947-58

GRANTS FOR SOVIET STUDIES

The Joint Committee on Soviet Studies, sponsored with the American Council of Learned Societies (which adminis­ters its program)-Herbert S. Levine (chairman), Mark G. Field, Peter H. Juviler, Walter M. Pintner, Irwin Weil, and Dean S. Worth-at its meeting on February 16, 1974 awarded grants for research relating to Revolutionary Rus­sia and the Soviet Union to the following 10 scholars:

Zvi Gitelman, Associate Professor of Political Science, Uni­versity of Michigan, for research on the political sociali­zation of Soviet and American immigrants in Israel

Antonina F. Gove, Assistant Professor of Germanic and Slavic Languages and Literatures, Vanderbilt University, for research on the role of woman in the poetics of Marina Cvetaeva

James C. McClelland, Assistant Professor of History, Uni· versity of California, Santa Barbara, for research on the role of higher education in the transformation of Russian society, 1917-41

William G. Rosenberg, Associate Professor of History, Uni­versity of Michigan, for research on railroad workers and Russian labor, 1917-21

Alvin Z. Rubinstein, Professor of Political Science, Univer-

sity of Pennsylvania, for research on the Soviet-Egyptian influence relationship since the June 1967 war

Donald V. Schwartz, Assistant Professor of Political Econo­my, University of Toronto, for research on recent adap­tations of systems theory to administrative theory in the Soviet Union

Robert M. Slusser, Professor of History, Michigan State University, for research on Soviet-American relations and the internal political struggle in the Soviet Union from November 1960 to June 1961

Peter H. Solomon, Jr., Assistant Professor of Political Econo­my and Sociology, University of Toronto, for research on specialists in Soviet policy-making: criminologists and criminal policy in the 1960's

Rex A. Wade, Professor of History, University of Hawaii at Manoa, for research on arming the Revolution: work­ers' militia and Red Guards in the Russian Revolution of 1917

Robert C. Williams, Associate Professor of History, Wash­ington University, for research on the genesis of early Soviet culture, 1905-30

PUBLICA rlONS The Allied Occupation of Japan, 1945-1952: An Annotated

Bibliography of Western-Language Materials, compiled and edited by Robert E. Ward and Frank Joseph Shul­man, with the assistance of Masashi Nishihara and Mary Tobin Espey, for the Joint Committee on Japanese Studies and the University of Michigan Center for Japa­nese Studies. Chicago: American Library Association, March 1974. 887 pages. $50.00.

Computer Simulation in Human Population Studies, edited by Bennett Dyke and Jean W. MacCluer. Papers prepared for a conference sponsored by the Committee on Biologi­cal Bases of Social Behavior, June 12-14, 1972. New York: Academic Press, February 1974. 539 pages. $16.00.

Modern Chinese Society: An Analytical Bibliography. 3 volumes. Prepared under the auspices of the former Subcommittee on Research on Chinese Society, Joint Committee on Contemporary China. Stanford: Stanford University Press, January 1974:

Vol. 1. Publications in Western Languages, 1644-1972, edited by G. William Skinner. 880 pages. $35.00.

Vol. 2. Publications in Chinese, 1644-1969, edited by G. William Skinner and Winston Hsieh. 877 pages. $38.00.

Vol. 3. Publications in Japanese, 1644-1971, edited by G. William Skinner and Shigeaki Tomita. 600 pages. $32.00.

SOCIAL SCIENCE RESEARCH COUNCIL

605 THIRD AVENUE, NEW YORK, N.Y. 10016

Incorporated in the State of Illinois, December 27, 1924, for the purpose of advancing research in the social science$

Directors, 1974: WILUAM J. BAUMOL, ALLAN G. BOGUE, LAWRENCE A. CREMIN, LEON EISENBERG, LEON D. EPSTEIN, SUSAN M. ERVIN-TRIPP, RICHARD

F. FENNO, JR., LEO A. GOODMAN, EDWARD E. JONES, LAWRENCE R. KLEIN, GARDNER LINDZEY, LEON LIPSON, CORA BAGLEY MARRETr, HERBERT

MCCLOSKY, SALLY FALK MooR'E, JAMES N. MORGAN, MURRAY G. MURPHEY, ALFONSO ORTIZ, JOHN W. PRATT, ALICE S. ROSSI, 'VILLIAM H. SEWELL,

ELEANOR BERNERT SHELDON, ELLIOTT P. SKINNER, M. BREWSTER SMITH, JANET T. SPENCE, EDWARD J. TAAFFE, KARL E. TAEUBER, JOHN M. THOMPSON,

ROBERT E. WARD, CHARLES V. WILLIE

Officers and Staff: ELEANOR BERNERT SHELDON, President; DAVID L. SILLS, Executive Associate; GORDON M. ADAMS, MICHAEL W. DONNELLY, JAMES

FENNESSEY, LoUIS W. GOODMAN, ELEANOR C. ISBELL, DAVID JENNESS, PATRICK G. MADDOX, ROWLAND L. MITCHELL, JR., ROBERT PARKE, MICHAEL

POTASHNIK, DOROTHY SODERLUND, DAVID A. STATT, ROXANN A. VAN DUSEN, NICHOLAS ZILL; NORMAN MANN, Business Manager; CATHERINE V.

RONNAN, Financial Secretary; NANCY CARMICHAEL, Librarian

86 8",