Socio-technical Design: Strategies in Multidisciplinary Research

By the same author

Autonomous Group Functioning Behavioural Worlds

Socio-technical Design STRATEGIES IN MULTIDISCIPLINARY RESEARCH

P. G. HERBST

TAVISTOCK PUBLICATIONS

First published in 1974 by Tavistock Publications Limited 11 New Fetter Lane, London EC4 Printed in Great Britain in 10/12 pt Times New Roman by Willmer Brothers Limited, Birkenhead

© P.

G. Herbst, 1974

ISBN 0 422 73980 4

Distributed in the USA by HARPER & ROW PUBLISHERS, INC. BARNES & NOBLE IMPORT DIVISION

To John McNally and his team of Durham miners, the Israeli kibbutzniks, and the workers and managers of the Norwegian firms who took part in the Industrial Democracy Project, who each in their own ways explored the future

CONTENTS page

ix

Preface

INTRODUCTION

1 The Development of Socio-technical Research

3

PART I DESIGN OF SOCIO·TECHNICAL SYSTEMS

2 Approaches to Socio-technical Design 3 Socio-technical and Psychodynamic Variables in Shio

13

Organization Design 4 Socio-technical Design of Ship Organization 5 Organizational Learning and Organizational Change on Merchant Ships: Matrix Organization 6 Emerging Characteristics of Socio-technical Organizations: A Summary

28 45

54 61

PART 11 APPROACHES TOWARDS THE INTEGRATION OF THE PHYSICAL AND THE BEHAVIOURAL SCIENCES

7 The Operational-unit Paradigm 8 The Multiple-perspective Paradigm 9 The Psycho-physical Transformation Paradigm vii

65 82 95

CONTENTS PART Ill CHARACTERISTICS OF TASKS AND ORGANIZATIONAL STRUCTURE

10 Production Tasks and Work Organization 11 Research Tasks and Research Organization 12 The Structure of Science and Developmental Trends

113 157 170

13 Maps of Knowledge and the Design of Educational Organizations

181

EPILOGUE

14 The Evolution of World Models 15 The Product of Work is People

201 212

Bibliography

219

Index

235

viii

PREFACE This book, a companion volume to Autonomous Group Functioning and Behavioural Worlds, brings together papers on socio-technical theory, method, and design that have previously circulated mostly in m.imeographed form. The papers were written over a period of nineteen years while I was working at the Tavistock Institute of Human Relations in London, the Institute for Industrial Social Research in Trondheim, and the Work Research Institutes in Oslo. These three institutes have over a number of years been engaged in joint research programmes. The introductory chapter presents a brief overview of the history of socio-technical research and outlines possible directions of future development. In Part I, the types of problem encountered in socio-technical design are discussed, first in a historical context, and then by means of a case study of a project concerned with evolving a new form of ship organization. Part 11 considers several approaches to the fundamental problem of socio-technical theory. An integrated analysis of technology and social organization is possible only in so far as links can be established to bridge the gap between the physical and the behavioural sciences. Each of the three conceptual approaches chosen for more detailed study can be translated into operational terms and can therefore be tested and applied to the construction of methods for organizational analysis. Part Ill is concerned with the construction of methods for studying the relationship between task structure and work organization, and with their application to organizational problems encountered in industry, research, education, and science policy. The epilogue explores further one of the implications of the ship organization study: the interdependence of accepted principles of organization and the basic assumptions of our partly implicit world model. Fundamental changes in socio-technical organization either presuppose or create a concurrent change in the predominant world ix

PREFACE

model, a change both in man•s view of himself and in his relationship to the environment. The conclusion arrived at in the final chapter is that the ultimate product of work is we ourselves as human beings. The book is a result of collaborative research over a long period of time, and I am particularly indebted to the following: to Eric Trist and Fred Emery, successive chairmen of the Human Resources Centre of the Tavistock Institute in London, and Einar Thorsrud, director of the Institute for Industrial Social Research in Trondheim and later of the Work Research Institutes in Oslo, who made up the core group of the socio-technical research programme; to my colleagues in the early days on the Durham coalmining project, Gerth Higgin, Hugh Murray, and Alec Pollock; to those on the Norwegian Industrial Democracy Project, Per Engelstad, Jon Gulowsen, Knut Lange, and Julius Marek; to Jacques Roggema on the shipping project; to Louis Davis on the exploratory sociotechnical design project; and last, but by no means least, to my colleagues and friends in the European group, Hans yan Beinum, Herman Hutte, Jaap Koekebakker, Mauk Mulder, and Peter SchOnbach. From 1960 to 1967 the work reported was supported by a grant from the Norwegian Council for Science and the Humanities. To Toril Hungnes and Bergljot Brun my thanks for their kindness and their help with the secretarial work. For permission to reproduce material that has already appeared in print, thanks are due to the following: the Editor of Tidsskrift for Samfunnsforskning and Universitetsforlaget, Oslo, in respect of Chapters 1, 3, 4, and S; the Editor of the European Journal of Social Psychology and Mouton & Co., The Hague, in respect of Chapter 3; the Editor of Human Relations in respect of Chapter 8; the Editor of Nordisk Forum and Munskgaard International Publishers Ltd, Copenhagen, in respect of Chapter 13. Acknowledgement is due also to the Tavistock Institute of Human Relations for permission to include the material in Chapters 2 and 10. P. G. Herbst

Oslo, 1973

Introduction

CHAPTER 1

The Development of Socio-technical Research1

---·--Most of the basic concepts employed in the field of socio-technical studies can be traced back to a paper by Trist & Bamforth {1951) on the social and psychological consequences of the longwall method of coal-getting. The starting-point of the early coal studies was provided by psychiatric investigations by Morris (1947) and Halliday (1948), which gave evidence of an epidemic incidence of psychosomatic disorders among miners working under mechanized conditions. As a consequence of mechanization introduced into coalmining to increase productivity, the small self-regulating work teams, in which each man carried out the total task of mining, were broken up. The new system required for its operation forty to fifty men, each working on a single task. The resulting work organization was one in which a number of teams worked independently, on different piece rates, but, by the nature of the task, they were dependent on one another to get their work done. Each group of workers, optimizing conditions for itself, created and passed on bad conditions to the work groups responsible for subsequent tasks. Instead of enabling the workers to cooperate with one another, the new system created insoluble conditions for interpersonal and intergroup conflict, resulting in psychological defence mechanisms in the form of reactive competitive individualism, mutual scapegoating, and a high level of absenteeism, all of which contributed to a low level of performance. The analytical model applied is shown below. The technological system determines the characteristics of the social system through 1 Based on a research policy document written in 1966 for the Institute for Industrial Social Research in Trondheim, and used later also at the Work Research Institutes in Oslo. First published in Norwegian in 1969 in Tidsskrift for Samfunnsforskning, Vol. 10, No. 3-4, pp. 225-35. Revised version in P. G. Herbst (ed.), Demokratiseringsprosessen i arbeidslivet, 1970.

3

INTRODUCTION

the allocation of work roles and the technologically given dependence relations between tasks. Performance is a function of the joint operation of the social and technical systems. Dysfunctional consequences Psychosomatic . Social structure disorders; Technologlcal ____,._ 0 f the work ----..interpersonal and system~ /tem intergroup conflict Product quality and quantity

of the social system are not easily modified in so far as the social structure is based on the requirements of the technological system. 1 The principle that began to emerge at this stage was that, if the technological system is optimized at the expense of the social system, the results achieved will be sub-optimal. The same would be true if the social system were optimized at the expense of the technological system. The aim to be achieved would need to be the joint optimization of the technical and the social systems. At the same time, where conflicts are built into a work organization there is relatively little that can be achieved by means of a Human Relations approach to conflict resolution. One promising lead at this stage was that, at a higher level of mechanization, in the form of the Bolsover technique, the tasks could again be integrated, thus providing the conditions for the operation of relatively small autonomous teams (Wilson, Trist & Bamforth, 1951). The idea that a given work organization is a necessary consequence of a particular technological system was not discarded until the second series of coalmining studies started in 1954. This second series was concerned with carrying out more systematic and, where possible, quantitative investigations. In the course of fieldwork in the Durham area, a number of composite autonomous work teams were discovered, which had been organized by the men themselves. Comparative studies of conventional work organizations consistently showed the superiority of the composite autonomous work organizations both in terms of productivity and in terms of social-psychological criteria. The most interesting findings related to 1 Research along similar lines began independently in a number of countries: see Westerlund (1952), Walker & Guest (1952), Touraine (1955), Morse & Reimer (1956); and, later, Blauner (1964) and Jordan (1968).

4

THE DEVELOPMENT OF SOCIO·TECHNICAL RESEARCH

Iongwall faces where a total group of more than forty men working on a three-shift cycle had organized themselves as an autonomous group. Cohesion was in this case maintained by a work rotation scheme which had evolved both within and between shifts. It became clear at this stage that the same technological system can provide a choice of social systems, at least within a range of feasible alternatives (Trist et al., 1963). The only experiment undertaken at this time was triggered off unintentionally when the trade union secretary at one of the pits where fieldwork bad been done for about two years decided, with the cooperation of the pit manager, to try out the ideas by himself by organizing a small autonomous group for an experimental three•month period. It was possible in this case to obtain a complete day-by-day record of the work behaviour of the team and to construct the first of several quantitative case-study techniques which were later developed and which make it possible to carry out systematic theory-testing at the level of single cases (Herbst, 1962, 1970). At about the same time Rice (1958) completed the first experimental study of composite work organization in an Indian weaving shed, which indicated that the ideas that had been developing were applicable in a factory setting and also in a completely different culture. 1 The second phase of socio-technical studies came to an end in 1959. Until this time the chief interest had been in trying to find a solution to the problems of the mining industry. There appeared to be little chance of implementation at the time apart from anything that might result from the publication of findings. During the next years, work continued on theoretical problems. Emery (1959), in a review of the field, stressed that it was essential to look at socio-technical organizations as open systems, and, in a later paper, Emery & Trist (1965) showed how different types of organization could be looked at in terms of the need to adapt to different types of environment. At the same time, further attention was given to the development of methods for the study of sociotechoical systems, and to the quantification of socio-techoical 1 A convergent approach in the field of job design was independently developed: see Davis & Canter (1955, 1956), Davis, Canter & Hoffman (1955), and Davis (1957). For an application to job-training, see King (1964).

s

INTRODUCTION

l

~rinciples; and the implications of s~cio-t~h~ical concepts fo~ ~

mtegration of behavioural and phys1cal pnnc1ples were exaillln~ The third stage of development became possible with the start in 1962 of the Industrial Democracy Project in Norway. For the first time, conditions became available for socio-technical experiments in a number of industries. The aim was to utilize autonomous and composite types of work organization as a basis for extending the participation of workers in decision-making. During the same period, socio-technical experiments were carried out in Holland in the postal and telegraph services, and in Ireland in the public bus service {Thorsrud & Emery, 1966, 1969; Van Beinum, 1965; Emery & Thorsrud, 1969). A review of the experience gained so far suggests that there are at least three possible ways of inducing changes in a work organization, depending on whether primary changes are made in the social, economic, or technological system. In most of the studies that have been done up to the present time, the approach has been to carry out a socio-technical study of a particular technological system, and subsequently to design a more appropriate correlated social system. Finally, either at the time of implementation or as a result of it, necessary changes take place in the form and distribution of payment. Up to this point the technological system was taken as a given, and effort was concentrated on redesigning and implementing the socio-economic system in ways consistent with the demands and properties of the technological process structure. Two difficulties are encountered here. The first is that present-day technological systems have been designed specifically to give a maximum breakdown of jobs into simple repetitive work components which require of the workers a minimum of initiative and training. The production engineer has in fact functioned as a social engineer but with only one type of organizational structure in mind. The second difficulty is that to implement changes in an existing work organization requires a great deal of effort and working through. In 1966 a work group came together to examine the possibility of designing the total socio-technical unit. Specifically, this implies that, for a given product and raw materials, the total set of feasible production processes is examined. It should then be possible to derive the social-organizational requirements for each of these, and finally to arrive at a choice based on the joint optimization of the 6

THE DEVELOPMENT OF SOCIO·TECHNICAL RESEARCH

technical and social systems (Davis & Engelstad, 1966; Emery, 1966; Herbst, Chapter 2 of this volume). Fundamental changes are now occurring with the transition to automated process industries. The total number of operators required to run a factory may be as few as five to ten men. They have to function as a team, with adequate understanding of the whole process. The team organization has to be quite flexible in order to be able to cope with any problems that may emerge, and crucial decisions have to be made with the shortest possible time-lag. Further, the team, especially during the stage of running in a new factory, needs to have a capacity for continuous learning both at the technical and at the social-organizational level (Thorsrud & Emery, 1969). The present indications are that self-regulating work organizations will be specifically appropriate in these conditions, and that they can be matched better to the requirements of the new technologies than to the technologies based on job breakdown which emerged at the time of the first industrial revolution. There exists a wide field of choice with regard to the ways in which automation and computer technologies can be designed and utilized. They can be utilized to increase the range and possibilities of human choice, or in such a way that decision-making is abdicated to a computer program. The psychological conditions that lead to the decision to employ computers as instruments of mystification and irresponsibility present a problem with which we shall need to be concerned. A possibly more critical emerging problem is a consequence of the increasing rate of technological change. Once some degree of selforganization of work groups has been established it is possible to go further and build in a capacity for organizational adaptation and learning; that is, work teams can redesign their own organization and members may be enabled to carry out research to improve both their technical and their organizational skills with the help of outside specialists utilized as resource persons. However, work on the organizational side of the problem by itself will not be sufficient. In some industries the rate of technological change is approaching the point where, before a new form of work organization can be established (together with new training and recruitment schemes, new career structures and payment systems), further technological change 7 B

INTRODUCTION

has already disrupted the operational conditions for the maintenance of the new social organization. A major difficulty here is that we still tend to look at technological development as a process over which we have no control. It is possible that at the time of the industrial revolution there existed only a single techno-economically feasible production system to which society had to adapt. This is no longer the case today. The development of technology over the past decades now makes it possible for a choice to be made between a wide range of techno-economically feasible alternatives. The necessary condition for determini~g the direction of social change in industry and society as a whole is that we utilize the option of technological choice. The critical step in this case is to achieve joint policy decisions with respect to technological and organizational development over the same time-period, so that the type of technology chosen is consistent with and supports the direction established for social and organizational change. In practice, this means that the initial task of the design engineer will be to present the widest range of techno-economically feasible alternatives. Of these, a smaller number will be selected for further study and tested for consistency with social and organizational requirements before a commitment to a final technological design is made. To the extent that workers and staff are able to participate in the process of choosing, the conditions will be provided for achieving an initial joint commitment to the direction established for organizational change. At the same time, it will become possible for those who operate the new technology to test the extent to which the principles of choice turn out to be valid in practice, and to contribute at a later stage to the process of redesign. The same type of principle could be applied to architectural planning and town-planning. Possibly the best way in which social scientists could help would be to assist with setting up the conditions for systematic follow-up studies which would enable the designer to test the extent to which the assumptions and theories utilized in the design were sufficient and valid, thus creating the conditions for continuous learning. The long-term trend would in this case be away from exclusive reliance on centralized research institutions and in the direction of building research capacities into existing professional and work organizations, thereby providing an essential condition for an organizational learning process to be self-maintaining.

8

THE DEVELOPMENT OF SOCIO·TECHNICAL RESEARCH

Further theoretical and methodological work will need to be concerned with the extension of the principle of joint optimization and with the extension of the analytical technique to larger system units.

1. Joint Optimization of Industry and Society The early studies were mainly concentrated on the internal sociotechnical structure of organizations. At the next stage, criteria were developed for discovering optimal organizational structures relative to different environments, where the environment was taken as a given. However, industrial organizations do not only adjust to their environment; the new socio-technical systems that are developing within industry also induce or require changes in the environment of which they form a part. 1 To the extent that basic changes are introduced in the form of new types of work role and interpersonal relations within the work setting, these can be expected to spread out into society either directly or in the form of a model. If, over the next decades, technological and industrial development accelerates but there is inadequate knowledge of and regard to its societal consequences, then the outcome may well be a weakening of the social structure on which industrial development depends. We need therefore, over the next years, studies of the effects on society of changes within industry as well as studies of the societal conditions for industrial development; these studies need to be coordinated so that the conditions for the joint optimization of social and industrial development can be investigated.

2. The Social Ecology of Industry Industry forms part of an ecology of social organizations. Industrial development led to a rising standard of education, the formation of new social classes, and changes in family organization. Each of these changes, again, affected the further development of industry. 1 An example is found in the Norwegian Industrial Democracy Project which required the setting-up of a joint national steering committee of the Norwegian Trades Union Congress and the Confederation of Employers (Emery & Thorsrud,

1969).

9

INTRODUCTION

Moreover, all these changes are cross-linked. Changes in social classes and in family organization determine the educational interests and career choices of the younger generation, which then determine the educational and professional skills available to industry. At the same time, the steady expansion of university and postgraduate education has led to a steady growth of research organizations which play a central role in further societal and industrial development, although their place in the societal structure and their means of subsistence have not yet been clearly established. Just as in the ecology of nature so in the social ecology, each part is intimately dependent on the other parts; changes introduced within any one organizational sector or a lack of adjustive change can have an almost immediate influence, directly or indirectly, on other sections of the ecology. Organizational changes introduced to optimize any one sector of the ecology have at present unknown consequences. Their effects on other sectors may be direct or indirect, supportive or reactive; they may become damped or be intensified to an explosive degree. 3. The Problem of Optimizing Social Ecologies If we consider the social ecology as it exists today in the Western world, we find that equal weight is not given to the different organizational sectors. Instead, all sectors of the social ecology are to a greater or lesser extent subordinated to the need to maintain industrial development. This was essential during the early period of industrialization. It is maintained today by competitive pressures and military demands, and through the application of science without the ability to understand and control the consequences. Once a greater understanding of social ecologies can be achieved, it will become possible to work, in the first place, towards stabilizing the growth of the industrial and applied research sector and, at the next step, towards finding optimal directions of growth for social ecologies as a whole and for smaller societal units, at the national and possibly the international level.

10

PART I

Design of Socio-technical Systems

CHAPTER 2

Approaches to Socio-technical Design1

---·--COMPLETE SPECIFICATION DESIGN

Before the industrial revolution, a craftsman and his task corresponded to what will be referred to as a viable socio-technical unit. The craftsman was able, within his range of competence, to decide what type of goods to produce and their quality. He could choose his raw m~terials, his tools, and his work techniques. Furthermore, he had to manage his relationships with his environment, in the way of marketing his goods and looking after his customers. He could, and did, engage in research on new tools, new work techniques, and new products. In fact, the development of modem science owes at least as much to the ingenuity of the instrument-makers as it does to the theoretical scientists. Many scientists, from Newton to the Cavendish, combined both tasks, and the modem scientist or the team of researchers responsible for a total project are among today's surviving craftsmen. The displacement of the industrial craftsman at the time of the industrial revolution was made possible by the emergence of a new concept of production design. The craftsman looked at his task as an art which he learnt from a master. These partly empirical, partly systematic skills were not supposed to be passed on to any laymen, nor were non-members of the craft guild permitted to engage in the craft. 2 The monopoly 1 Originally prepared as a consultancy report (1966) for the Human Resources Centre of the Tavistock Institute of Human Relations, London. 2 The guild type of organization still survives in medical and other professional associations, except that university education has for the most part taken the place of apprenticeship training. Also, the industrial revolution still continues. The attempt to replace medical diagnosis based to some extent on intuitive skills by diagnosis based on computer programs is an example.

13

DESIGN OF SOCIO-TECHNICAL SYSTEMS

power of the guilds was broken by what was essentially a scientific analysis of tasks. It was found that most of the known industrial processes could be analysed into a sequence of simple operations. Once that is done, the task is no longer an art, but becomes a predictable determinate mechanism. It does not matter at this stage whether the operations are performed by a machine or a human being. In fact the most complex mass-produced human artefacts even today, whether TV sets or computers, are created by human machines consisting of a chain of hundreds of girls using their hands or only the simplest tools. When the complex production task was analysed into a sequence of elementary operations, at least some of these were simple enough to be taken over by machines. To begin with, what was looked for was the possibility of substituting mechanical power for human muscles. The amount of power that can be generated by a human is limited, irregular, and subject to fatigue. The steam engine supplied more power at a constant level, and wheel transmission provided regular repetitive operations. The worker becomes at this stage restricted not only to the production task but to a small set of operations within it or even to just one quasi-mechanical operation. However, since human beings, unlike machines, are capable of an extremely wide range of behaviour and variability, coercion has to be introduced to elicit from the operator the required set of responses and to prevent him from engaging in any other type of behaviour. The work autonomy of the craftsman depended primarily on his intrinsic task satisfaction, and the excellence of his product assured him both of his customers and of his community status. For the factory worker there is little if anything left to which he can meaningfully relate or from which he can derive self-respect. Motivation now has to be predominantly extrinsic. Reward can be used to motivate him to carry out a prescribed set of operations, and coercion has to be maintained to ensure that he does not engage in other forms of behaviour or carry out the prescribed operations in some other way. The changes that have occurred since the industrial revolution, apart from increased technical sophistication and increased process specification, are mainly concerned with the balance between coercion and extrinsic reward, and with the way in which coercion or manipulation is applied. 14

APPROACHES TO SOCIO-TECHNICAL DESIGN

Stage J: Foreman Control Prior to the development of trade unions, autocratic control was vested in the foreman. Wages were close to the survival level. They could be reduced for poor workmanship and men could be hired or fired on the spot. The response of the worker to such a system was apathy or resentment, or behaviour that could be interpreted as irresponsible. It was therefore necessary to maintain or increase coercive control.

Stage 2: Work-method Control During the next phase, the rise of the trade unions together with the increasing accumulation of industrial capital led to higher wages for the workers and gave them greater power to influence their work situation. This influence was, however, scarcely ever applied to the work situation as such, except in so far as there was government enforcement of safety and minimal employment regulations. At the same time, hours of work were reduced and pension and insurance schemes were introduced. During this period the coercive power of the foreman decreased and this trend was supported in part by a gradual acceptance of democratic ideology. Control now relied somewhat less on direct coercion and more on extrinsic reward, on the building up of loyalty to the firm, and on training. These gains were, however, paid for by the workers in terms of an even greater restriction of their freedom in the work situation. Time-and-motion study techniques developed early in the century were used to prescribe in minute detail the operations to be performed. The worker now operated like a machine. The only incidental advantage was that the operation sequence often had, like a machine, an intrinsic rhythm, which can provide some degree of pleasure. Whether it does so, however, depends on whether the imposed rhythm matches that which can be established by the operator. We have now come a very long way from the original work situation of the craftsman. First, the worker's relations with the environment have been severed. Next, instead of being related to a total task he is related to some small set of operations which is generally, but not necessarily, defined for him by a machine. Finally, 15

DESIGN OF SOCIO•TECHNICAL SYSTEMS

his work method is no longer left free but is prescribed by time-andmotion study. The production system at this stage makes use of only one ability of workers, namely their ability to simulate machine-functioning. The weakness of this type of organization is not simply that it does not make adequate use of human resources but also that, to the extent that workers do make use of their intelligence and skills, they will be in conflict, either unintentionally or intentionally, with the requirements of the production process. Typically, social groups develop where the production design makes no provision for them, ingenuity is employed to beat piece rates or make them ineffective, and games are invented to counteract the monotony of work. While the initial design was to suppress all internal variance, variances are now generated by the work force and additional work has to be carried out to offset this. The result at this stage is either chronic workermanagement conflict or a complex system of collusion. Thus the production system is now not only less efficient than was provided for by the original design, but far more costly in terms of non-productive costs in the form of management and supervisory work, inspection and security costs, grievance handling, and piecerate bargaining. Even in mechanized non-automated industries, labour costs may be as low as 5 per cent since only production labour is estimated. The real cost disappears under overheads. A realistic costing will need to include the costs of management, supervision, and clerical and professional work required to maintain and adjust the operation of the production unit. Stage 3: Machine Control of Workers The introduction of continuous process techniques led to a radically new development. On a production line, well constructed in terms of technical criteria, such as continuous press lines in car factories, the work is reduced to a simple repetitive operation such as to get hold of material and place it in a machine - get hold, place; get hold, place. The work pace is set by the machine. The operation is practically foolproof. There is no foreman to complain to, and no relationship to fellow-workers can be set up. No creativity can be employed or counteraction taken, and workers typically settle down to apathetic compliance with the work pace imposed by the machine. 16

APPROACHES TO SOCIO·TECHNICAL DESIGN

The worker now has no freedom left except to go on with the work or to leave. This may be a highly efficient system in terms of machine and manpower utilization. The internal system can produce little variance apart from possible walk-outs and strikes, but these, when they do happen, can be very costly. The operator has functionally becOme a machine component who can be controlled with almost the same freedom with which a machine can be controlled. It should be noted, however, that none of the production organizations considered so far depends critically on the use of machines. The continous process TV and computer assembly plants referred to earlier consist entirely of human components; moreover, they do not require direct foreman intervention. Stage 4: Process Automation Once the task of the worker can be reduced to a single simple repetitive operation, then in many cases it requires very little to substitute transfer and positioning mechanisms and to arrive at a fully automated process line. From the point of view of production design, process automation is essentially the logical end of the first industrial revolution rather than the beginning of a new one. From the point of view of the human operator, however, it implies fundamental changes. The lack of clarity about what the human requirements are is due to some extent to the fact that many of the existing, often only partly automated, plants are transitional stages. The socio-technical requirements that emerge are towards total unit management. There are, in present-day automated processes, basically three types of personnel left in the production unit itself. These are: (a) Process workers who remain in the production process itself on operations that have not yet been automated out. These are found chiefly in relation to process input and output. (b) Monitors who act as a direct or secondary check on potential process breakdown. Their role is essentially that of a signal device. (c) Process control staff whose task is effectively that of managing a total production unit. In addition, staff are needed to carry out (d) preventive maintenance work (e) repair work.

17

DESIGN OF SOCIO•TECHNICAL SYSTEMS

At low levels of automation, monitoring and process work may be combined. At higher levels of process automation, monitoring and process control are generally combined, with the latter as the dominant component. The task of the process control staff is no longer comparable to that of the previous factory workers. The process worker at Stage 3 had a task that was almost foolproof; he could scarcely make an error that could have any effect on the process. An error made by a member of the process control staff can, on the other hand, lead to serious equipment and production losses. His task, at least at the present level of automation, can make demands on vigilance, on skill in solving technical problems, on self-initiated task-congruent behaviour, and on social and communication skills. What disappears in these conditions is the traditional fractionated and hierarchical organization. There is no time for information to filter through different levels, and to become distorted or possibly get lost on the way. Also, all relevant information has to be combined and evaluated if effective action is to be taken. This cannot be achieved by traditional types of work organization. There are at the same time conditions where automation principles do not apply: for example, the product has to be tailor-made; the material worked on or the material work situation is idiosyncratic; the market is turbulent, requiring a high degree of internal flexibility; or the task itself is a creative and problem-solving one. Such conditions do not mean that computers or automated equipment cannot be employed, but rather that the total task cannot be programmed in detail, or that the creation of a program is the actual task. Here again, traditional principles of work organization are not adequate. Summary

The first industrial revolution was based on the application of a principle of production design. It was found that there were a number of tasks which could be analysed into a sequence of elementary operations. Tasks of this type are detenninate and can be represented by a functional equation, as follows. Given an input /, there exists a sequence of operations 'IT, such that if this sequence of operations is applied to the input a predictable output P results: 7r(J) ~P. 18

APPROACHES TO SOCIO·TECHNICAL DESIGN

A determinate task of this type can in principle always be carried out by a machine. What happened in practice was that each operational unit was analysed into a further sequence of even simpler elements, each of which could be performed by a machine. These were now linked up, and at that stage the need for direct productive labour input disappeared. Automated production units have since then become more sophisticated and complex through the introduction of feedback devices, which are relatively new, and programmed tapes, which had already been employed in the weaving industry. The basic design principle has, however, remained the same- namely, the successive decomposition of the process into the simplest possible elementary components and detailed specification of the component sequence. If process automation is, in terms of production design, the end of the first industrial revolution, then the second industrial revolution which is likely to occur over the next decades can be expected to be based on the formulation of new design principles. CRITICAL SPECIFICATION DESIGN

There are a number of recent developments in molecular engineering, in biosimulation, and in the study of socio-technical systems that point to the emergence of new design principles. Early engineering techniques were based on the method of building up increasingly complex machine structures which produced simple components, which were then assembled to produce a final product. In molecular engineering the structure of material itself, either as it exists or as it can be produced to specified criteria, is used to effect required transformations. Further, since many materials have metastable states their structural form can change in response to signals in the form of heat, pressure, or light, so that the same material can operate as a different machine, depending on environmental conditions (von Hippel, 1965). This points to the emergence of new forms of production engineering which are no longer based on the principle of successive decomposition, linkage of components, and hierarchical control structure. From the point of view of production design, the key development lies in the study and design of autonomous systems. Here we find two 19

DESIGN OF SOCIO•TECHNICAL SYSTEMS

lines of development, one from biophysics and the other in the sociotechnical study of work organizations. The last few decades have seen the emergence of cybernetics (Wiener, 1961), showing that self-adjustment requires the existence of cyclic feedback processes; of communication theory (Shannon & Weaver, 1949), which provides measures of structure and error variance of discrete processes; and of open-system theory (von Bertalanffy, 1950), which demonstrates that open systems can maintain steady-state functioning without the use of a separate control mechanism. An interesting point of departure is the non-specification technique described by Beurle (1962) in a paper on the properties of random nets. These are abstractly a set of elements with random connections. Beurle argues that the nervous system may initially be somewhat like this but that later, in response to transactions with an already structured environment, an internal adjusted structure which corresponds to a biased network structure should gradually emerge. Now, clearly, a random network can learn practically any desired response, but something has to be added to get to a workable model. 1 Without going into the detail of recent, more sophisticated, work, what we find here is a new approach to the problem of design which is no longer concerned with complete detailed specification but with minimal critical specification. The main reason for this approach was a concern with systems that can learn and that can adjust themselves to environmental changes. Adjustment, learning, and creative and intelligent behaviour require minimally: -internal variability to create alternative response patterns -the testing of alternative response patterns and evaluation of the outcome -selection of the most appropriate response. This is one of the lines of development that led to the study of autonomous systems. What was made clear at this stage was that variability, and thus making errors, was not a bad thing and that, on the contrary, systems must have sufficient potential and mobilizable 1 The simplest model of this type is found in stochastic learning theory which is based on a single type of response element that can change its probability of response. What is lacking here is the possibility of structural growth.

20

APPROACHES TO SOCIO-TECHNICAL DESIGN

internal variability and mechanisms for the self-correction of error in order to be able to adjust to a variable environment. The third and to some extent parallel line of development is more direCtlY relevant to the design of production systems. Before the second world war, the problem of optimizing the functioning of industrial and work organizations was looked at either from a tecbno-economic point of view or in terms of improving the social organization and human relations. What was left out of account was that the social organization is not independent of the technical production system. It is possible, as has been done in the past, to look for an optimal technical solution. However, if the correlated social system required is inferior, then the total production system may be far short of the optimal. A series of studies of socio-technical systems undertaken in the coalmining industry showed that this was indeed the case, and demonstrated further that, with the possible exception of Stage 3 systems, a given technological system can be operated by several different types of work organization. The variables that may remain to some extent free are (a) the pattern of task allocation, (b) the allocation of task responsibility, and (c) the method of payment. What emerged at this stage was the concept of autonomous work groups that would overcome the dysfunctional properties of fractionated work organizations (Trist & Bamforth, 1951; Wilson, 1951; Rice, 1958; Emery & Trist, 1960; Herbst, 1962; Trist et al., 1963); and this work converged with new principles that were being developed in the field of job design (Davis, 1962, 1966). The principle of critical specification design can be stated as that of identifying the minimal set of conditions required to create viable self-maintaining and self-adjusting production units. An optimal solution is obtained if the unit requires no external supervision and control of its internal functioning, and no internal staff concerned with supervision, control, or work coordination. The management function should primarily be supportive, and concerned with mediating the relationship of the unit to its environment. There is consistent evidence that work systems of this type are superior in terms of relevant social and psychological criteria. Chronic conflict between men, and between men and management, disappears. For individual members, the task provides the opportunity for learning and for participating in technical and organiza21

DESIGN OF SOCIO-TECHNICAL SYSTEMS

tional problem-solving. The group as a whole can learn on the basis of its experience, and becomes able to utilize experts as consultants. Conditions are created for the development of mutual trust and respect and thus also of self-respect. Just as internal conflicts and warring factions export their conflict into their environment, equally, cooperative relationships within a group provide the conditions for the growth of cooperative relationships with the environment. At the same time, a considerable reduction of unproductive overhead and management costs can be achieved. It cannot be expected that social organizations of this type will be ideal for and attractive to everyone. Since human beings differ in their emotional and social maturity, and human needs change in the course of development and growth, no single type of social system can be optimal. If in the case of autonomous groups there is less of a problem, this is because such groups cannot be imposed but depend for their development and maintenance on the consent of all involved. At the same time, a significant point about self-maintaining systems is that they do not possess a single rigid structure but have the characteristics of a matrix organization which can adapt its internal structure to meet internal and external task demands. There are two problems that will need to be considered in more detail:

1. What are the critical conditions for the operation of selfmaintaining socio-technical units? What we are looking for here is a minimal set of necessary and sufficient conditions. It should be noted that, in so far as we are dealing with a set of interdependent variables, there will be more than one possible set of critical conditions. We shall in that case need to find a set of variables that can be included in the set of design criteria. 2. Given this, we need a new design technique based not on the iteration of techno-economic variables only but on the joint iteration of techno-economic and social-psychological variables. Supporting conditions for a viable self-maintaining production unit are the following: (a) A clearly definable total task with an as far as possible easily measurable outcome state, which may be in the form of the quantity and quality of a product, and also an easily measurable 22

I

APPROACHES TO SOCIO•TECHNICAL DESIGN

set of relevant input states. These provide the necessary information both for evaluation of the system's performance and for J.D,aintenance and adjustment of the internal process. (b) A single social system is responsible for the total production unit. The unit should include as far as possible all the equipment and skills required for process control and technical maintenance. (c) Given that the functional elements of the production process are interdependent with respect to the achievement of the outcome state, the social organization should be such that individual members do not establish primary commitment to any part function - that is, do not lay claim to ownership of or preferential access to any task or equipment- but are jointly committed to optimizing the functioning of the unit with the outcome state as the primary focal goal. (d) In traditional types of work organization, doing and deciding tend to be split and decision-making functions are allocated to higher levels of the hierarchy. Self-maintenance requires that relevant decision-making functions are brought down to the lowest possible level and reintegrated into the operational work organization. This becomes of particular importance where the decisionmaking content of component tasks has become depleted by means of computer-programming and automation. (e) Responsible autonomy cannot generally be established and maintained unless the available tasks require personal responsibility based on some degree of competence, judgement, and skill. Similarly, unless the total task allocated to a production unit requires the development and use of personal competence, acceptance of joint responsibility for the organization and functioning of the unit may not be achievable. Emery & Thorsrud (1969) have formulated relevant criteria for job design in the form of hypotheses about the way in which tasks may be more effectively put together to make jobs and at the same time to satisfy general psychological requirements. Any new design technique will need to incorporate the basic set of techno-economic variables. However, instead of providing a detailed specification of all variables, the critical specification technique requires the identification of a minimal set of variables that have to be specified, and the identification of other variables that have to be left 23

c

DESIGN OF SOCIO·TECHNICAL SYSTEMS

free. The free variables are those that are required if the system is to achieve self-maintaining properties. The initial set of variables that require specification and are thus turned into fixed structural parameters may later on be even further reducible, since, given system properties, the specification of a given set of characteristics may lead to the emergence of steady-state properties of other system characteristics. The existing production design technique is based on the successive decomposition of the total production process into part-product processes; these are then decomposed into operational units; and these are finally decomposed into elementary man-machine operations. At each level there is an iteration cycle which provides the specification for the next lower unit until, ideally, every movement of operators and machine operation is rigidly specified: Level1 Level2 Level 3 Level 4 Level 5 Level 6

V V V V V ?

Product design Selection of total process Part design Selection of operational unit Design of operational unit Action design

What is produced in the end is not a functioning unit. In order to· coordinate the often thousands of split-off process elements, to counter the variances that arise in each processing and transport segment, and also the variance produced by the social-organizational links created by workers, and finally to adjust the system to variances in input and changing product specifications, a superstructure of work is required in the form of supervision, inspection, control, planning, scheduling, and personnel work. This additional system again requires coordination and produces variances for which a next higher level has to be provided, and so on. In so far as the 24

F' .,

APPROACHES TO SOCIO-TECHNICAL DESIGN

111 finally produces more variance than it can control at any level,

~ports

the unmanageable surplus variance to the environment berever it can be absorbed, compensated for, or simply got rid of. ;,th the rapidly growing rate of environmental pollution and the increasing incidence of chronic mental health disorders show that the problems exported can no longer be absorbed by the environment or effectively compensated for by the social and professional organizations created for this purpose. The creation of viable systems at the production-process level is aimed at: 1

-avoiding the production of variance due to incompatible technicalprocess requirements and social-organizational requirements -providing the conditions, where possible, for variance to be controlled within the unit itself. An alternative design procedure will therefore need to include at each iteration level the corresponding set of social-system variables. This will require, at each level, methods for studying the socialorganizational implications of technical decisions. The critical level for viable system construction is Level4, which is concerned with the selection and linkage of operational units. While at this point the design problem is made more complex, this should turn out to be more than compensated for by cutting off fixed specification at this level, since Levels 5 and 6 contain variables that should almost all remain free in order to provide the necessary conditions for the production unit to operate as a self-maintaining socio-technical unit. The design process would in this case take the form shown below: TECHNO-ECONOMIC VARIABLES

level 1

level 2

level 3

level 4

~ ~

[L> L>

SOCIAL-ORGANIZATIONAL VARIABLES

Product design

Selection of total process Part design

Structure of depa rtmenta I unit

Operational unit type and interlinkage

Structure of autonomous work unit

25

DESIGN OF SOCIO-TECHNICAL SYSTEMS

The design technique will require: (a) definition of the relevant social-organizational variables {b) socio-technical methods for inferring the organizational implications of a given technical-process structure (c) construction of a feasible joint iteration procedure. EVOLUTIONARY SYSTEM DESIGN

Nature does not create in the way in which factories do. A seed does not contain a complete specification of the organism, and the information given by the genes does not provide for a one-step implementation. Yet in spite of the fact that the information given by the gene structure is quite limited, the growth process proceeds with self-maintenance properties at each stage until a viable organism is completed which structurally reproduces the original one with a very high degree of reliability. Let us simply note at this stage that: 1. Reliable production does not require a complete specification of either the production process or the final product. 2. The creation of a complex structure does not require the initial production of elements which are later connected to produce the final product. The production process is not one step from specification of structure to structural implementation, but always goes through successive growth stages. 3. It is not simply the final product that is a viable system, but a viable system exists at every stage of growth. 4. A biological organism is not created but creates itself, given an initial structure and a correlated succession of suitable environments which maintain and feed the growth process. It appears likely that production processes incorporating these

principles will eventually be developed. However that may be, an understanding of growth principles is necessary for an understanding of the conditions for psychological and organizational development. A technological system can be designed and implemented by construction. A social organization cannot be created in the same way. The conditions for both psychological and organizational growth are more similar to biological

26

'f

APPROACHES TO SOCIO-TECHNICAL DESIGN

srowth processes, as against the mechanical construction type. This 111eans

that:

1. If we want to implement viable autonomous social systems, the design will not consist of a specification of the final system (although the characteristics of this system, which are aimed at, will have to be defined and accepted); rather, what has to be specified and implemented is the conditions that make it possible for a system of this type to develop. 2. The social system aimed at can rarely be implemented in one step but will need to go through successive stages of growth. The technical design should in this case be such that a viable sociotechnical system exists at each stage. Thus if a system is designed for composite group operation there should also be provision for the possibility of more fractionated operation during the initial learning stage, and also for the possibility of regression to a more primitive organizational form.

27

CHAPTER 3

Socio-technical and Psychodynamic Variables in Ship Organization Design1

---·--The problems encountered in designing ship organizations differ in a number of respects from those met with in developing new forms of organization in factories. In designing a factory organization we can generally start off with the specification of an established or a new technology, and generate possible types of work organization in terms of the requirement of achieving joint optimization of the total socio-technical system. In ship design, on the other hand, the critical decisions that have immediate implications for the social and work organization on board are concerned with the choice that exists with respect to the allocation of tasks requiring human intervention which can be located either on board or ashore. These in turn create alternatives in terms of manning by a continuous crew or a temporary crew, or by means of shore-based personnel. Since there exists in this case a wide range of possible technological alternatives, we can, instead of taking a specific technological system and working out the requirements for a supporting social system, consider the possibility of working the other way round. That is, we can attempt to specify initially the essential requirements for a social organization on board, and then work backwards to discover the critical supporting technological conditions that would need to be satisfied with respect to ship design. 1 This chapter, and Chapters 4 and S, were first published in Norwegian as a single article in 1969 in Tidsskrift for Samfunnsforskning, Vol. 10, No. 3-4, pp. 371-400. A revised version of the material was published in P. G. Herbst (ed.), Demokratiseringsprosessen i arbeidslivet, 1970. A first English version of Chapter 3 appeared in 1971 in the European Journal of Social Psychology, Vol. 1, No. 1, pp. 47-58.

28

VARIABLES IN SHIP ORGANIZATION DESIGN THB CONVENTIONAL SEQUENCE OF SOCIO·TECHNICAL DESIGN

The basic design variable is the allocation of tasks on board and ashore. The basic tasks include: _navigation and engine control _ engine and instrument maintenance -ship maintenance _ ship-shore communication -loading and unloading -catering. In principle, each of these task sectors can be wholly or partly shorebased. If tasks are split up so that one part is carried out on board and the other part ashore, then the significant decision variable is the location of task components that involve: (a) decisions requiring a high level of skill and judgement (b) work and decisions at technician level (c) unskilled and semi-skilled labour. Decisions made at this point are crucial since they have direct implications for: {i) the extent to which the total task allocated to the ship provides conditions for autonomy and self-regulation; (ii) the communication requirements between ship and shore {this is not a purely technical problem since a great deal of relevant information on the ship cannot easily be recorded, transferred, and adequately responded to ashore); (iii) the possible work-role and social structure, and, given this, (iv) the possible career structure; (v) educational and training requirements. The possible manning requirements are: (a) continuous crew on board (b) supporting transient crew (c) land-based manning. The unit for socio-technical analysis will need to be the total set of tasks required for effective ship operation, wherever they happen to

29

FIGURE 3-1 TECHNOLOGICAL DESIGN VARIABLE

SOCIO-TECHNICAL DESIGN SEQUENCE

MANNING REQUIREMENTS

WORK ORGANIZATION DESIGN

SUBSYSTEM DESIGN

Shore -------------.. Communication and ..-control and coordination of / ship-shore activities authority

.

Tompocacyocow/

Allocatoo" of lO'k•Y.eont;""ou' ,,... aboard and ashore

REQUIREMENTS

Car~r and

' ' ' " r ' x e recruitment

., Oepa
structure ~

I

Wock roles

~

Shift structure

Social aod

psychological opti~ization requirements

VARIABLES IN SHIP ORGANIZATION DESIGN

be located. It would appear to be feasible to look at the design for J]18DDing chiefly from the point of view of optimizing the social system and then look at the supporting conditions required in terms of tasks or task components which should be allocated to a continuous crew on board. Given the manning requirements on board, the next decision variable is the departmental structure established, which further restricts the possible work organization and career structure. The final decision variable is the shift structure and shift-allocation pattern. The socio-technical design problem can thus be broken up into a sequence of decisions (Figure 3-1). If we look at what can be done in terms of immediate organizational changes that are required on board in consequence of the changes in technology that have been introduced over time, the decreased size of crew, and the increasing difficulties of recruitment, then it is clear that the change process will have to go in the reverse direction of the design sequence. Thus, in experimental programmes concerned with the integration of deck and machine crew, changes in shift structure have been used to facilitate changes in interdepartmental relationships {A. Trist, 1968). Changes of this type cannot, however, go beyond a certain point in so far as major decisions are already built into the ship design and into the existing work-role, career, and status structure. As long as technological change was relatively slow, it was possible to find ways of adjusting the social organization to a given technological system. The main contribution of socio-technical analysis at this stage lay in showing that, even within the restrictions imposed by a given technological design, a choice of alternative types of work organization existed. It was therefore possible to work towards joint optimization of the techno-economic and the social systems. But this type of static socio-technological analysis is no longer adequate to cope with the current problems of the shipping industry. The present rate of technological change is such that, before a new form of organization on board can be established (together with new training and recruitment schemes, new career structures and pay systems), further technological changes will already have disrupted the conditions for the maintenance of the new social organization. This appears to be a major contributing cause of the emergence of 31

DESIGN OF SOCIO-TECHNICAL SYSTEMS

turbulent and potentially uncontrollable environments (Emery & Trist, 1965). It is no longer sufficient, then, to utilize the possibility of organizational choice unless the possibility of technological choice is utilized at the same time. Changes in technology have to be directively correlated with changes in social organization over the same period. Policy decisions with respect to changes in social organization over initially the next three to five years have in this case to be coordinated with the choice of new ship designs, so that the type of technology and design chosen for new ships that will come into operation is as far as possible consistent with and supports the direction of social, educational, and organizational development. This possibility did not exist earlier when as a rule only a single techno-economically feasible solution to a problem was available. At present, the limitations lie not so much in the possibility of generating alternative types of technological design as in the possibility of being able to specify, within the limits of techno-economic feasibility, the essential social and psychological conditions that have to be satisfied by the technological design we wish to implement. As a first step it is necessary to consider the characteristics of the existing culture and organization on board merchant vessels. Whatever new organization develops has to grow out of the existing one. Field studies were carried out on a number of Norwegian ships: a car bulk carrier on the Europe-East Coast of America Line, a cargo vessel on the Europ~West Coast of Africa route, and a factory fishing vessel in the North Sea (Thorsrud, Gulowsen & Kolltveit, 1967; Roggema, 1968). 1 While the purpose of the initial field studies was to collect data on existing technology and organization, a later field study was concerned with identifying potential directions of organizational change and development. Just as autonomous work groups were originally discovered in the course of fieldwork in coalmines in northern England, where in a number of places they had been designed and implemented by the workers themselves (quite independently of the theoretical socio-technical analysis which had indicated some years previously that this type of organization would 1 So far, only relatively few social scientific studies of seafaring have been carried out. Of particular relevance are Aubert & Arner (1959), Arner & Hersson (1964), and Barth (1966). Novels on life at sea should not be overlooked as a source of material and insights.

32

f.-

VARIABLES IN SHIP ORGANIZATION DESIGN

be optimal in the light of existing task requirements), so it appeared "ble that, at least on some ships, crew members might already ::becOme sufficiently concerned to explore for themselves potential cfireCtions of organizational development. pSYCHODYNAMIC AND SOCIAL-SYSTEM CHARACTERISTICS OF THE EXISTING ORGANIZATION ON MERCHANT SHIPS

Excessive Fragmentation Over a period of time the size of crew has decreased and it is likely to decrease further. Given the conventional departmental and role divisions, the hierarchical structure, and the shift structure, then: - a large number of crew members become isolated - the possibilities of collegial interaction in both work and non-work activities are minimized - even if the total territory of the ship is large, it becomes split up into private, work, and non-work territories, thus minimizing the effective living-space for nearly all crew members (Roos, 1968).

Hierarchical Structure The existing basic values and traditions of ship culture are intimately related to the existing status hierarchy. The present structure emerged under conditions that are not dissimilar to those that are found in factories ashore, with, however, the following differences: (a) all crew members may have to meet physical survival crises (b) the accepted way to the top is from the bottom of the hierarchy (c) the basic assumptions are those of a military organization. The last point is of some importance since perhaps a key issue is whether to retain the military organization model or to look for a different type of organizational model altogether. On the other hand, if it is judged that a military organization model should be retained, then it will be relevant to consider in some detail innovations and new forms of naval and military organization that have developed during the past generation.

33

DESIGN OF SOCIO·TECHNICAL SYSTEMS

If the need to maintain unskilled and low-skilled crew members on board as part of a continuous crew disappears, then it would become possible to restructure both the content and the responsibility of officers' roles. In this case possible alternatives to the hierarchical status structure could be considered.

Exchangeable Component Structure This model is one that makes it possible for any man in any position to be replaced, ideally without altering the effectiveness of the total organization. The basic assumption is that all the relevant work and interpersonal relationship requirements can be built into each role. Over and above this, specific psychological attributes are built into each role. Thus men at the bottom of the hierarchy tend to be given the attribute of being irresponsible and incompetent, and the captain's role has almost godlike superhuman attributes. Unless crew members at the bottom of the hierarchy are perceived to lack, or actually lack, competence and willingness to accept responsibility, then the justification for the existing authority structure largely disappears. However, if we simply designed a new organization in which the bottom level was made up of, say, junior officers, then almost inevitably attributions of irresponsibility and incompetence would be transferred to the junior officer group. Consideration of the existing organizational structure shows that there are a number of reasons why the development of personal, collegial, and friendship relations is difficult, and why such relations are, on the whole, both exceptional and unstable: 1. The existing work organization design does not require the development of personal relationships and these, where they do occur, are more likely to introduce a variance into the organizational system, which has to be dealt with, than to contribute to its effectiveness. 2. The high level of labour turnover unpredictably disrupts relationships that are formed, although some pair relationships may survive. 3. A two-class structure develops, with proletarian and gentlemantype values respectively. This allows greater freedom of interaction within each group. There are, however, problems in that potential 34

VARIABLES IN SHIP ORGANIZATION DESIGN

membership of each group is reduced by decreasing crew size, and that intermediate and specialist crew members cannot easily be integrated in either group. 4. Given the conditions required for maintaining the conventional status structure, it is difficult to switch over to a different type of social system during the leisure-time period. There is considerable evidence that the development of collegial and friendship relations is perceived to be inconsistent with the maintenance of the conventional authority and status structure. Since this is regarded as a central problem, it needs to be examined in more detail.

s.

Distance-regulating Mechanisms There are a number of social-psychological processes that contribute to distance maintenance on board. The most frequent reason given for distance maintenance is that familiar relations lead, if not to contempt, then at least to loss of respect. It is, of course, characteristic of an authoritarian structure that respect is the attribute of a role and not of the person who occupies the role. At the same time it is possible that: (a) The work role of some officers does not provide them with a feeling of competence. This may be the case for deck officers in so far as they can no longer utilize their navigational skills. (b) Some officers may be given responsibility for operations for which, owing to technological and administrative changes, they have insufficient training. (c) Officers are not able to demonstrate the competence that they do possess. At the same time: (i) it is precisely the distance-maintaining mechanism that makes it difficult for officers to demonstrate competence and at the same time makes it possible for the superior to protect himself from a judgement of incompetence by subordinates; (ii) the higher the status position the more the role content is looked at as a kind of mystique by subordinates, in which case distance maintenance is consistent with the needs of both superiors and subordinates;

35

DESIGN OF SOCIO·TECHNICAL SYSTEMS

(iii) the tendency over time to transfer high-level decisions to head office, coupled with decreased crew size, has reduced the effective authority of officers both upwards and downwards; the actual competence requirements for high-level roles will in this case be reduced. (d) As the size of the crew decreases, each crew member experiences himself as being individually more visible to other crew members. He can now less easily distance himself by disappearing as a member of a group, and he is also more likely to be physically isolated. As a result, the feeling of loneliness increases. The interview material (Roggema, 1968) suggests that officers seek to cope with the problem of loneliness by keeping themselves busy all day long if possible, and it would appear that the desire of crew members to work as much overtime as possible is not simply economically motivated. Thus, again, the possibility of social interaction during non-work time is decreased. We may note here that closer relationships between crew members are not necessarily inconsistent with the maintenance of a status hierarchy. However, this presupposes: - a joint value orientation of men and officers, rather than the existing separate men and officer cultures together with specialists who belong nowhere - commitment to a clearly defined joint goal or mission to which each crew member can effectively contribute - respect relationships based on demonstrable and clearly perceived competence. These conditions are to quite a large extent consistent with the operation of fishing vessels, but they are not easily applicable to merchant vessels. So far, we have looked at the distance-maintaining mechanisms that derive from the formal organizational structure. A somewhat deeper level of this issue is indicated by comments such as: 'I cannot say why, but it is better for the atmosphere to keep a little distance' (officer). 'It's a stress to be in such a small community. One has to defend one's private life' (officer). 36

VARIABLES IN SHIP ORGANIZATION DESIGN

•It's a psychic stress. You work together, you eat together, you see the same faces all the time. Even in leisure-time. You even see the same faces when you watch a film' (officerV A number of interrelated aspects are touched upon here. To use Lewinian terminology, the life-space on board ship is not internally segmented. This means that tension, when it arises, floods the total life-space; and there is nowhere to get away from it, since there is no outside region. The central problem then becomes that of affect control in conditions in which the obvious mechanisms for affect control that exist on land are not available. On land it is possible to play different roles in different social contexts, so that whenever tension builds up in one social context, one can transfer to the next. There is always a temporary shutting-off mechanism available. On board, the role-shifting mechanism cannot operate in so far as there exists a single social context, and within this each crew member is fixed in a single role of captain, bosun, or cook. On land, it is possible to release positive or negative affect in so far as it is generally possible to break off a relationship or escape from a social context. On board ship, positive and negative affect expressions spread and reverberate within the closed unsegmented social space. The spreading effect is unpredictable in so far as the crew is made up of persons who do not know one another. Moreover, in view of the independent tendency of psychological tensions to build up over time, the triggering-off of affect expressions can relatively easily lead to explosive and destructive reactions. The problem here is twofold: (a) At the interpersonal level the crew has the characteristics of an aggregate rather than of a structured community. (b) Interpersonal love-hate relationships which do not form part of the formally sanctioned structure cannot be closed off or played out within a private space by the participants, and there is no immediate escape possible. The participants cannot easily separate themselves again, nor can they jointly remove themselves from the response of the total social system. The necessity for interpersonal affect control under these conditions appears to be a major cause of the distance-maintaining mechanisms 1

The interview material in this chapter comes from Roggema (1968).

37

DESIGN OF SOCIO-TECHNICAL SYSTEMS

which, while independent of the status hierarchy, are at the same t · consistent with the maintenance of it. The psychological consequences are: (a) Interpersonal affect control and distance maintenance ten to lead to an impersonalization of interpersonal relationships.j The experience of others as just faces, to the extent that this occursJ also implies a corresponding relationship to oneself. (b) Similarly, the experience of a non-segmented social spa implies a corresponding absence of segmentation and bounded regions within oneself. Under these conditions the individual cannot respond to problem• within one of several roles and within one of several social contexts~ rather, in these conditions the way he meets and responds to pro~ lems affects his total personality. There are three possible ways of compensating for the reduction of personal affect relationships: (i) At a superficial level, the telling and retelling of yarns permits the vicarious enjoyment or reliving of memories concerned with affect-laden interpersonal events in the non-private sphere of the individual's life, allowing them to become the public property ot the community. By this means a crew member can also define for himself and others a personal identity apart from his role identity. (ii) Telling yarns and stories about oneself can also be a way of protecting one's personal life-region. (iii) Retreat into the personal life-region is a third type of compensatory behaviour. Concern for one's family ashore and protection of the personal life-region appear in the interview material as two major themes that are interrelated. To a man on land his home is his private territory. He can retreat into it and shut the door. It contains what he treasures. He allows access to it only to his friends. It has a semi-sacred character. It is a focal and central region of his life-space. The personal private region has similar properties. To allow it to become public property would be to lose it; to protect it means to have a private space into which one can retire, but primarily to 38

,

,

VARIABLES IN SHIP ORGANIZATION DESIGN

ilqJerience in fantasy or relive in memory positive affect relationships, bich may later be reactivated in reality. VI Lons separation from home may give grounds for quite realistic and concerns about those whom one loves, in which case the availability of at least one person to whom one can talk freely will be of considerable help. More problematical is the tendency for the content of the personal private region to become idealized over time. While this makes it possible to enhance the positive affect that can be experienced in fantasy, it will at the same time almost inevitably also produce internal doubts and uncertainties, and increase the unreality of the content of the private region. Under these conditions one may have an almost compulsive and indiscriminate need for communication to provide reality support for an idealized fantasy that is already threatened by one's own uncertainty and doubts. The risks in this case are that one's private life becomes public property and that the response received is likely to increase one's doubt and uncertainty, with the result that the idealized fantasy now becomes the reverse. Whereas before it provided a surplus of enjoyment and happiness, now it appears full of fear and negative feelings, and a depressive phase sets in. An alternative risk-lowering strategy is to reduce communication with others, thus increasing the need for distance maintenance. The major significance of the private personal region for most crew members is that it provides a bridge back to shore. To the extent that this region becomes idealized, what typically occurs is that, on return to land or home, discordance between fantasy and reality leads to increasing discontent and to a reverse idealization of life on board ship, and thus to an oscillation pattern which may also be repeated during shorter shore visits. The sailor is thus caught in a condition of double ambivalence. When at home he has a fantasy of life at sea which may not match reality, and when at sea he has a fantasy of being at home or in port. Eventually he may give up the one or the other. We know very little about the consequences if the personal region that binds a crew member to shore is given up. A transitional pattern in this direction is expressed in the following response to shore visits:

1

rears

•The first things you see are the docks, ugliness and dirt, filthy factories, terrible. It's the same everywhere. One often feels inclined 39 D

DESIGN OF SOCIO-TECHNICAL SYSTEMS

i

)

to stay on board. You have to get through a sort of barrier befoq you see something of the normal world' (sailor). In a novel on life at sea Geelmuyden says on this point:

'If a seaman has a horror of ports and longs to be at sea, thCII he is a proper seaman, then his chance to get away, if he woul~ like to, is much less. He has thrown away his last bridge an~ become part of the ship.' We need to refer briefly to another method of regulating inter. personal relationships. At lower-status levels we find some referen~ to the concept of 'shipmate', and at higher levels to the concept of 'style'. Both form part of the implicit culture of the shipboar4 community. It would be worth while to find out to what extenl there is consensus regarding these role definitions and to wha1 extent they are still significant and actualized. We note here oni, that both refer to impersonal role requirements with respect tc proper ways to behave, in the one case for persons with peer-grou11 status and in the other for those with officer status, and that theiJ function is to extend the formalization of interpersonal relationshipa consistent with the requirements of the existing role and statUI organization. For the purpose of considering possible changes iJJ the formal organization, an understanding of the present core values of shipboard culture will be needed. So far in our discussion of distance-regulating mechanisms we have examined, first, those deriving from the authority structure and, second, those relating to the problem of interpersonal affect relationships. The latter ar.e linked to: (1) a mono-role system (2) non-segmentation of the socio-physical space on board (3) the aggregate properties of the crew.

Both (1) and (3) are directly modifiable by means of socio-technical design; (2) is only indirectly modifiable since it is based on a situationally induced non-segmentation of the self. A still deeper level of the problem of distance maintenance finds very clear expression in the following interview extract: 'I am afraid to look at the captain as a person. Seeing the captain in this way would mean an attack on his authority' (officer). 40

.,

VARIABLES IN SHIP ORGANIZATION DESIGN

~ose

contact between captain and officers would according to another comment show 'that he is only a human being'. If in at least ne aspect of his role the captain is not a human being, what is he? ;eferences to him by the men as 'God' and 'the father' clearly indicate what his perceived role is, but then, why is it of such iJDportance, especially to the officers, to maintain this extra-human role attribution? After all, not even a present-day king or president re<J.uires these attributes. We noted earlier that the shipboard community has the characteristic of non-segmented social space in the sense that it provides a single social context that one cannot leave and within which interpersonal tension cannot normally be contained locally, but will easily spread and build up. Now if the captain becomes personally involved in this, then he is not in a much better position to achieve control than is any other crew member. So, in order to be able to control, he has to be located in a sense outside the human community and not be part of it, except that on board ship it is not physically possible to create an outside region. The characteristics attributed to God are that he is omnipotent and all-seeing. What is relevant is that the attribution of omnipotence is based on his not (normally) entering into direct human relationships and not exerting direct control, for in this case his power would be finite and limited and thus not much greater than that of an ordinary human being; and the attribution of being all-seeing is based on his not being seen. The captain is the person in whom all power and authority on the ship ultimately reside. By minimizing his personal interaction he becomes a focus of concern, and creates the condition whereby an image of himself becomes internalized in the crew members. In so far as such an image has been internalized it acquires characteristics of internal control and at least in this sense the captain becomes all-seeing. At the same time, direct communication, where this is required, will proceed not on a person-to-person basis but via the internalized image, and in this case may not require more than a hint or a gesture. It will be of considerable importance to crew members that the ultimate source of authority is completely trustworthy, benevolent, just, and, if need be, willing to offer himself for their safety. Again, since the captain is the source of all authority, and officers 41

FIGURE 3-2

ORGANIZATIONAL CONDITIONS THAT CONTRIBUTE TO DISTANCE MAINTENANCE

Hierarchica I status structure

Status-maintenance requirements

gra~~ti~ groups-.......__

Small crew size

Small peer

High fragmen

Departmental structure

No self-selection

~"~

...._ .... Distance --...........maintenance

~

Exchangeable - - Lack of component structure mutual trust

Protection of private personal region

t

Mono-role structure •

l

Single social context

Non-segmentation C?f the accessible life-space

/

Tension spread within the total I ife-space ~

Interpersonal affect control Non-segmentation ____. Tension spread within/ of the socio-physica I the tota I socia I system space

*

Extrusion of the captain from the social system

VARIABLES IN SHIP ORGANIZATION DESIGN

)lave their power only by delegation, it becomes of importance to

officers that, if not the actual, then at least the perceived power of the captain is as great as possible, for as his power decreases, so also does their share of it. In this case distance maintenance between captain and officers is required also to maximize perceived power. Now, to the extent that the captain is perceived to have godlike attributes, then other crew members acquire correlated roles, and so at this level all crew members can become actual or potential participants in a cosmic drama. According to the ancient and medieval myths, the cosmic drama scarcely ever involves the human coDllD.unity, but is played out in the upper regions. Therefore, if it is actualized anywhere in some form, it will be in the officers' mess - that is, among those who to some extent share the captain's authority and power - and not in the men's mess - that is, among those who are content to remain ordinary human beings. At the same time, to the extent that the captain is perceived to be omnipotent and to carry the responsibility for ship and crew, then the men at the bottom of the hierarchy will perceive themselves to be non-responsible and to have little actual or potential competence. This is quite consistent with the fact that officers scarcely ever refer to the need for the men to be efficient or competent over and above the level of following instructions, but consider the men's primary role to be that of having respect for officers, not for themselves as persons, but as representatives of the captain. Every social system implies a world view. This may be less immediately realized in the case of factory organizations, but becomes significant in the case of social systems in which people both work and spend their lives. The principles on which a social system is based do not become invalid if they are technically inefficient. If techno-economic criteria become a major aim to which people are prepared to subordinate themselves, then this is itself part of a world view. However, if the social system becomes inconsistent with task requirements and also the world view on which it is based becomes eroded, that is, people are no longer interested in, or able to play, the implicit roles, then the social system no longer possesses survival possibilities. The conventional organization combines the hierarchical status structure with an exchangeable component system, which can operate like a machine. If the former becomes simply weakened, then the 43

DESIGN OF SOCIO-TECBNICAL SYSTEMS

latter becomes dominant. This is quite consistent with the trend towards ship rationalization. The effect of this, which is already becoming clear, is the opposite of what is found in factories on shore. In factories the prevalent result is a reduced involvement in work. On board ship the typical result is a compulsive overinvolvement in work, which is frequently referred to: 'You are like a machine that belongs to the ship; you like the work, but it's work and nothing else. You cannot get rid of it' (sailor). 'You don't have a problem as long as you work, you don't notice a thing' (officer). 'It's all right as long as you work' (officer). In such cases, leisure-time has a negative connotation, and even shore visits can appear empty. Here we encounter the problem of alienation in its fundamental form. Both the self and the environment appear to be empty, and there seems to be nothing with which to fill the void, except work. To cope with the psychological problem at this level demands a good deal of maturity. Normally, one should avoid creating conditions where this type of problem is situationally induced and experienced as deprivation. The design for a new type of socio-technical system will need to start at this level.

CHAPTER 4

Socio-technical Design of Ship Organization

---·--oRGANIZATIONAL REQUIREMENTS

Every socio-technical system has to satisfy several requirements. The number of requirements for an effective ship organization will be larger than that for a factory organization, since we have to create a social system in which people have both to work and to live. The problem is not to create a new social organization for a new technology, but an organization that can cope with a steadily changing technology. At the same time, conditions have to be provided for the development of a micro-community which can function in a situation of isolation. The minimal conditions for effective task performance have to exist from the start. The requirements have to be: - feasible individually - mutually consistent, and if possible mutually supportive. The set of requirements would appear to include the following. The organization should: (1) be adaptable to technological change (2) facilitate the effective use of leisure-time (3) provide conditions for both autonomous and group-based activities (4) be consistent with an exchangeable component structure; that is, it should not be too difficult to replace leavers and to integrate new crew members (5) be either an overlapping role structure or a multiple role structure (6) if possible, link mutual respect relations to perceived and demonstrable competence 45

DESIGN OF SOCIO·TECHNICAL SYSTEMS

l

(7) if possible, be consistent with and provide conditions for thol

development of collegial and friendship relationships (8) minimize the build-up of psychological tensions (9) provide effective control over interpersonal tensions (10) provide greater stability of crew membership. Work roles should: (1) provide a basis for technical or professional competence (2) facilitate both transition to and recruitment from shore with a minimum of retraining (3) be consistent with career advancement requirements.

The task and task elements allocated to the ship should: (1) be consistent with the requirements of the social system (2) consist as far as possible of complete task regions (3) provide to some extent conditions for operating towards a joint aim for the total crew.

There are indications that at least some of the requirements form a cluster. Few data are available on the interrelations of relevant variables. The interview material provides only some general pointers. Tension Symptoms on Board

The analysis in Chapter 3 showed that a fundamental requirement of any new type of ship organization is that it must not be less effective than the existing type in coping with personal and interpersonal tensions. Results obtained from alcohol studies are revealing in this connection. At the time of the ship studies, 38 per cent of patients in institutions for alcoholics in Norway were ex-sailors. Of the Norwegian population as a whole, it is estimated that approximately 20 per cent are ex-sailors. Since of these 20 per cent quite a high proportion will have spent only a short time at sea, the actual degree of risk for the sea-faring population is likely to be even greater. Within the population of sailors the risk of alcoholism tends to be limited to a well-defined group. Brun-Gulbrandsen & Irgens-Jensen (I 964) carried out a study of 3,440 young sailors, aged twenty to twenty-one, in which they employed a composite 'handicap index'. 46

SOCIO-TECHNICAL DESIGN OF SHIP ORGANIZATION

The index was based on interview and objective-type data, and included the following variables: _ abnormal home environment _ low education _ high neurotic score. It was found that sailors who come from a normal home environment, and have a high level of education and a low neurotic score, have practically no alcohol problem; and that there is a steady and pronounced increase in alcohol abuse with ascending position on the handicap index. At the same time, alcohol abuse is found to increase markedly with increasing length of time at sea. Further, there is a higher rise over time, in both incidence and degree of alcohol abuse, among sailors of a lower educational level (and this group will also be likely to contain those who have no further interest in education). Thus there are no indications here that long periods at sea have a therapeutic effect; quite the contrary. A survey by Arner (1961) of relevant Norwegian studies shows that the same basic trends found with regard to alcohol abuse are found also with regard to criminal offences, suicide, accidents, and mental illness specifically of a psychotic type (the findings do not apply in the case of manic-depressive illness). In each case where data are available, incidence for sailors is above the average for the Norwegian population as a whole. It is of special interest to note that the corresponding rates for men on fishing vessels are markedly below those for the general population. INTERDEPENDENCE STRUCTURE OF ORGANIZATIONAL REQUIREMENTS

As a starting-point for the design of a socio-technical system for ships we shall take it as a given that: Over the next ten years there will be relatively frequent changes in ship design, ship operation, and fleet operation in the direction of greater automation and more efficient data-handling and datatransmission techniques, which will provide the basis for an integrated technical-system design.

47

DESIGN OF SOCIO·TBCHNICAL SYSTEMS

Whatever organization is established must, then, be consistent with this condition. Technical redesigns and changes take relatively little time, whereas the development of an appropriate new work and social organization takes a relatively long time. Since, therefore, by the time a new organization is developed it is almost certain that fundamentally new technical changes will have occurred, it is clear that the organization established must, from the start, be such that it is adaptable to, and able to cope with, relatively frequent changes in technology at a possibly quite sophisticated level of automation. In the light of the consequences of the above for a social and work organization, what we are looking for is: (a) the derivation of a set of organizational conditions that have to be satisfied (b) an examination of the consistency of the conditions, which will indicate whether the required organizational properties will be mutually supportive (c) a general specification of the resultant type of organizational structure, which could later be specified in more detail with the introduction of additional requirements. Given a steadily changing technology in the direction of greater automation, change-adjusting properties need to be built into the organization. The size of crew will decrease, and this in itself will Inake it difficult to Inaintain the traditional hierarchical status structure in which authority and work execution are separated. Instead, work execution and authority will coincide to a greater extent for the crew members on duty at a given time. In order to permit adjustment to technological changes, only a minimum of structure should be built into the organization, because the organizational structure will be constantly disrupted. Instead, adjustment capacities will have to exist to a greater extent in the crew members. This implies that crew members will require a fairly high level of education and an orientation towards further education. This would suggest developments in the direction of a small crew all of whose members have a professional or technical work role- that is, basically an all-officer crew. The expected consequences of these suggested lines of development are: 48

f'

SOCIO·TECHNICAL DESIGN OF SHIP ORGANIZATION

(i) a reduction of status-induced fragmentation and isolation (ii) a greater educational and cultural homogeneity of the crew.

Both should facilitate a capacity for more effective individual and joint utilization of leisure-time, and should contribute to improved !Dilnagement of personal tension. At the same time, the data on alcohol abuse indicate that, at least under conditions on board, a higher educational level, and possibly also commitment to further education in either a teaching or a learning role, will contribute to better management of personal tension. Whatever can be done in the way of improving the management of personal tension would lead to a reduction of labour turnover and therefore to a greater stability of the crew, and thus conditions would exist that would encourage the self-selection of group members. At the same time, given crew members all of whom have professional or technical competence, then, provided the technological structure is such that their actual and potential skills can be effectively utilized, and provided further that the correlated conditions for the improved management of personal tension are in operation, the conditions exist for an organization based on the autonomous responsible behaviour of individual crew members. In this case the external hierarchical control structure would no longer need to be the dominant authority and control system. It could possibly be allowed to fade into the background and be activated only under emergency conditions. However, the type of organization required to cope with emergency conditions has still to be considered. There is one possibly relevant finding from the coalmining studies (Trist et al., 1963; Herbst, 1962). The autonomous groups studied were not created by individuals who came together, but in most cases by pairs who came together to form a larger unit. A detailed analysis of the group process (Herbst, 1962) showed that subsequently, depending on task requirements, pair units joined together, split, or exchanged members. Consequently, under conditions of stress, the group did not regress to an aggregate, but regressed only partially, the complex structure temporarily breaking down into smaller subgroup units. Trist and his colleagues, in studying a twenty-four-hour cycle unit, found that the development of subgroups that rotated members across shifts played a significant role in avoiding 49

DESIGN OF SOCIO•TECHNICAL SYSTEMS

shift fragmentation, which would have destroyed the self-regulating properties of the system. Figure 4-1 shows the argument so far in diagrammatic form. It will be seen that the organizational requirements listed earlier are not independent, but are all related to one another; they appear to a considerable degree to be mutually supportive, which is essential for a feasible organizational system. The implication obtained at this stage from the total set of requirements is that an appropriate type of organization for ships would be similar to some extent to that found in some arctic camps. The basic crew would consist of officers. The range of task competence would need to include mechanical and electronic engineering, navigation, communication, computer-programming, and data analysis. Even on a fully automated vessel the crew should not be allowed to decrease in size below a-minimum of eight to ten persons, so that at least a pair would be available for each shift. Each officer, in addition to having specialist competence within his task region, would need to acquire some degree of skill and competence in one or more other task regions. This would mean that an officer would be in a command or an assistant position, depending on the nature of the task. Also, if some non-technical tasks were allowed to rotate, it would be possible to increase further the number of role relationships of crew members. The relationship between officers should in this case be basically of a collegial type. They should be able to engage in joint problem-solving where necessary, with the captain playing a senior role. The primary task role of the captain would need to lie in the field of ship-shore relationships, that is, in the management of boundary relationships between the ship and the various parts of the larger system of which it constitutes a part. H a sufficient degree of crew stability could be achieved, then, within defined safety limits it would be possible to allow the crew to evolve their own work organization and to decide on shift structure, allocation, rotation patterns, etc. This would allow each ship to develop to some extent an idiosyncratic culture, and would give crew members a better chance of finding the type of ship that suited them best. Another possibility would be to allow a more idiosyncratic culture to grow around the utilization of leisure-time. This would be particularly aided by the self-selection of crew membership. From the point of view of technological design, the implication is

so

FIGURE 4-1

TENTATIVE INTERRELATIONSHIP OF DESIGN CONDITIONS AND REQUIRI!MI!NTS

Organization adaptable~ High educational ~Professional o r - - - - - - - - - - - - - - - - - . . .

tot.d.7Jc~ '"""' . ~ techolcalwoM< role•

l

utJIJZJng leisure-time

Increasing automation

Small crew

G~~~er capacity for

Re!at•.vely little

Improved manag~ment

l

b"'Jt-oo~ pe""""l

!cess turno£. greater stability of crew

1!

Capacity for autonomous responsible behaviour

Possible self~ selection of

' ""'["......

ten"""

/

Organization based on responsible autonomy of crew members who are capable, where appropriate,ofteam action

Reduct:n of ' Educational and status-induced.__ fragmentation cultural homogeneity of crew and isolation

DESIGN OF SOCIO-TECHNICAL SYSTEMS

that relevant decision-making and problem-solving tasks are located on board. This means that a computer installation would be put aboard, and a possible link provided to a central computer installation at headquarters. A major aim of automation and ship design should in this case be to eliminate task components requiring no or low-level skills on board, or as far as possible to locate such components ashore. The total design is, then, one in which the permanent crew has the position of an elite, and the length of time spent aboard is related to the level of skill. Trainees would initially spend only shorter periods aboard as part of their training programme. They would thus have a transitional period before they committed themselves to joining a more permanent crew, and this would make it easier for them to reintegrate in the community ashore if they wished to. The ship design should be based on the principle of creating clearly demarcated areas as follows: -work area - non-work leisure territory -private territory. The leisure territory might include part of the deck area and should be designed to permit a wide range of possible uses. The territorial design should provide at least the possibility of creating distinct behaviour settings. In terms of recruitment and of maintaining relatively highly skilled personnel aboard, the conditions that have to be satisfied are that: (1) the technological design permits the location on board of a sufficient range of challenging tasks involving high-level skills (2) from a technological point of view the ship constitutes a relatively autonomous unit (3) the permanent crew aboard acquires the characteristics of an elite. These conditions are interdependent in so far as it will be difficult to recruit and maintain relatively highly skilled men on board if the ship is depleted of high-level tasks, if total task control and task management are not possible, if task roles do not provide sufficient responsible autonomy, and if the perceived status of crew

52

SOCIO·TECHNICAL DESIGN OF SHIP ORGANIZATION

JDCtnbers is low. It should not be overlooked that hiring and induction procedures, and shore-leave facilities, will need to be consistent with the status level required of personnel. To refer again to Figure 4-1, it will be noted that the organization considered so far pivots and is critically dependent on the extent to which more effective management of personal tension is achieved. If there is failure in this respect, the system might in fact not be viable. On the basis of the analysis in the previous chapter, which showed that a mono-role structure together with a single social context and a non-segmented socio-physical space creates not only a considerable amount of tension but also considerable problems of tension management, the additional suggestions have been concerned with the possible development of a multi-role structure and of distinct and separate behaviour settings, and with the provision of sufficient freedom to permit tension adjustment by means of changes in allocation, work rotation, and organization structure. Finally, it is necessary to consider a design which is practically the mirror-image of the one considered so far. Given the already existing capacity for controlling vehicles in space, it will become feasible to build ships that can be steered from land-based installations. This would imply that professional skill would be shore-based. A transient repair and overhaul crew would need to join the ship, and at most a routine watchkeeping and maintenance crew would remain on board. What would be required in addition is a set of strategically placed air stations, both on land and at sea, from which a professional stand-by crew could be flown out in a short time in case of emergency. Such a service would probably need to be organized on a national or international basis. The social-psychological problems for such a crew would be considerable, given the lack of significant content in their tasks and the long periods of isolation. From the point of view of long-term planning, the best policy would appear to be to provide all crew members with educational opportunities and with facilities for employing professional and technical skills, so that, if technological development should go in this direction, they could be readily transferred to shore establishments or to specialized emergency crews. Whatever way is chosen, a clarification of future career policy is essential to solve recruitment and organization problems over the coming years. 53

CHAPTER 5

Organizational Learning and Organizational Change on M erchant Ships Matrix Organization

---·--The preceding analysis of the existing social and work organization on merchant ships has shown that emergent technological requirements and social organizational requirements point in the direction of a future organization on board that will consist of a small team of officers, all of whom will have a technical or professional task role. Each officer will need to be able to perform several tasks. This is not simply in order to provide greater work variety, but because, with increasing automation, the performance of some tasks will be only occasionally required. At the same time, more composite and overlapping task roles will provide the conditions for smaller teams to cooperate on specific tasks. This will imply a development away from a hierarchical status structure to a matrix organization, within which different team subgroupings can come into operation to carry out specific tasks or planning roles. A matrix organization is one that does not have any single division of functions, such as deck and machine, but permits the formation of these and other subgroups according to the nature of the task to be performed. The matrix organization does not contain built-in status differences; it is based on the assumption that each officer has a specialist role together with a range of task competence which partly overlaps the competence of other officers. Any officer may thus, depending on the nature of the task to be done, take on a leadership role or act as a member of a specific task group. To the extent that crew size decreases and the need for technical competence increases, it will become important that the organization remains viable even if one or more officers are absent owing to recruitment difficulties or sickness. A matrix organization based on 54

LEARNING AND CHANGE ON MERCHANT SHIPS

composite roles will make possible a flexible redeployment of officers. A.t the same time, a decrease of built-in occupational and status differentials will make it easier for the organization to adjust to technical changes. It appears unlikely that a direct change from the existing organization to the new type of organization outlined can take place. What is required is not simply a new technical learning programme for the existing staff of officers and for younger crew members, but a process of learning how to operate a new type of organization on board. The learning programme must therefore be based on a total crew and may require a planning horizon of five to eight years. The main immediate problem is in this case to find possible transitional types of organization aimed at modifying the existing FIGURE 5-l

TRANSITIONAL TYPE OF MATRIX ORGANIZATION

A

Captain B Navigation Cargo-handling Deck maintenance Oerical work E

Mechanical engineering Mechanical maintenance Engine stores Clerical work

c

D

Navigation Radio communication Oerical work

Navigation Computer use Data-handling Instrument maintenance Clerical work

F Mechanical engineering Electrical maintenance

G Mechanical engineering Electronic maintenance

Navigational group Cargo group Administrative group Communications Data-handling and Records Engine operational group Mechanical maintenance group Electrical maintenance group Electronic maintenance group

ss B

BCD BDE ABDE CD

BCDE EFG EFGB FGD GDC

DESIGN OF SOCIO·TECHNICAL SYSTEMS

FIGURE 5-2

TRANSITIONAL TYPE OF MATRIX ORGANIZATION

A Captain

c

B

Cargo Navigation Engineering

E Mechanical engineering Electrical engineering Navigation

D

Navigation Electronic maintenance Engineering F

Radio communication Navigation Cargo Electronic maintenance Engineering G

Electrical engineering Mechanical engineering Navigation

Electronic engineering Mechanical engineering Navigation

organization in the direction of a matrix-type organization. The transitional organization should be consistent with the operational requirements of both existing vessels and automated vessels - automation will permit ships to be operated from a single control centre on board. The organizational diagrams presented here (Figures 5-1 and 5-2) are intended not as actual proposals for a new type of organization but as illustrations of the various possible ways in which further training of the existing staff of officers could provide a bridge towards the development of a matrix organization. The illustrative examples are based on a core of six officers. At this stage, all officers would require technical-school or technical-college training. A second type of matrix organization, suggested by David Moreby, 1 is one in which all officers can carry out bridge and engine control, watchkeeping, and also maintenance (see Figure 5-2). A major difference between the traditional organization and the matrix organization in its transitional form is that the traditional single division of functions between deck and machine becomes now only one of a number of possible subgroupings. Moreover, at this 1Personal

communication.

56

LEARNING AND CHANGE ON MERCHANT SHIPS

tage the deck and machine subgroups overlap, and at the next stage

~f technological development the distinction in its present form may

no longer be relevant. Instead of the present subordinate crew there would need to be a number of officer cadets. The initial aim oftraining would be to make every cadet capable of taking over at least one of the basic officer roles. The training scheme for cadets would need to be coordinated with a training scheme for officers that would both increase their level of competence and give them the opportunity to act as teachers. At the next stage the officer trainee scheme would have the function of providing a second generation of officers who could operate a more complete matrix organization. Officer cadets would need at this stage to be trained in at least two adjoining roles, such as CD, FG, DF, leading to an organization with a smaller number of more composite roles based on combinations of: X Administration Maintenance Cargo-handling

y Navigation Computer use Data-handling Communications

z Machine Electronic Electrical

The organization design outlined is essentially of the developmental type (discussed in Chapter 2). The initial problem is that of creating the conditions for a self-sustaining and continuous learning process both at the individual and at the organizational level. Sustaining a developmental process requires, however, that the necessary external supporting conditions are continuously evaluated, and appropriately modified. To give an analogy: if a child's development is to be maintained, both the teaching he receives (content and method) and the relationships he has with his teachers and other relevant persons in his environment have to change over time in an appropriate way. The immediate problem is how to initiate and provide the conditions required for the development on board of an organization that has the capacity for continuous learning. Preliminary findings from a field study (carried out by the author and Jacques Roggema) indicate a possible starting-point. The findings relate to a number of partly unanticipated consequences of a recently

57

DESIGN OF SOCIO·TECHNICAL SYSTEMS

introduced trainee scheme. During the first year of this scheme, elementary training is provided both on deck and in the engine room. During the second year (or the first year in the case of technical-school graduates), further training is provided for qualification either as able-bodied seaman or as mechanic. The study revealed that officers who had previously felt that the men showed no interest and had no sense of responsibility were impressed by the boys' interest and eagerness to learn, and considered that they had achieved remarkable results in a few months' time. One of the engine officers, reviewing the experience, began to see that the earlier difficulties were not all attributable to the men: 'I now start treating people in another way. For example, one of our greasers made a bad start; so we more or less cold-shouldered him. I did too. But I decided to give him more and more difficult work. I gave him jobs which were just a bit above his level. He changed completely, he has become interested, he does a fine job now. It all depends on leadership, but we never had the opportunity to learn anything about it.' 1 The training scheme in this case led to a relationship between officers and men that was different in quality; it had the characteristics of a personal teacher-pupil relationship in which the officers involved re-experienced their own tasks as something of value that could be passed on to their juniors, while at the same time the men had greater autonomy and felt increased satisfaction with their tasks. Furthermore, the officers were able to learn a better leadership role. It should be noted that teaching, although not in its present form, was in earlier days an accepted part of the officer role, so that officers who were willing to act as teachers in a sense rediscovered a valuable component of their traditional role. Like many teachers before them, they found that they had to do quite a bit of studying and thinking themselves in order to keep a jump ahead of their 1 The interview material was obtained by Roggema, and the final recommenda· tions are based on joint discussions in the course of the field study referred to above. Roggema's more recent field studies concerning long-tenn change projects in ship organization are reported in Roggema (1971) and Roggema & Thorsrud (1974). An independent study carried out by the Westfal-Larsen Shipping Co. (1972) indicates that better results are obtained by having a fairly large group of trainees on board one ship rather than by having smaller groups on a number of different ships.

58

LEARNING AND CHANGE ON MERCHANT SHIPS

pupils. Moreover, unlike classroom-teaching, which rarely presents the need for reality-testing, in the teaching situation on board the trainees could and did ask for immediate checking of what they had learnt against actual operations on board and were thus in a situation where they could test the knowledge of their teachers. Officers who had acted as teachers, and also some who had not, became more aware that a good deal of the training they had received at maritime schools was obsolescent, that their own knowledge of the functioning of equipment and instruments, while sufficient for daily operations, was incomplete, and that they had no understanding of the new types of ship technology that were being introduced. Many of these officers expressed a wish to participate themselves in more specialized and advanced training courses or felt the need for more basic technical knowledge. The findings at this stage indicate the following:

1. A training scheme, in which officers become involved in teaching, may provide the initial condition required for a change in the basic culture of ship organization. Such a scheme is found to decrease the distance between officers and men in a task-related context, and in a form that is experienced as appropriate by both officers and men. 2. A good way to extend one's knowledge of a subject is to teach it. The involvement of officers in teaching, and especially in the design of courses and the adaptation of teaching material, creates the conditions that can lead them to become interested in evaluating and extending their own knowledge. It is essential in this case that those officers who have become involved in teaching, and at the next stage also other officers on board, are given the opportunity for further appropriate training. At the same time, other ordinary crew members should be encouraged, and given the opportunity, to participate in parts of the total trainee programme. 3. The emphasis should not be on the elementary training of young sailors, but on the provision of conditions for the development of a teaching-learning culture based on the total crew on board. It will be clear that the trainee scheme in its present form, while it is valuable as far as it goes, is neither designed, nor sufficient, to create fundamental changes in organizational structure. 59

DESIGN OF SOCIO-TECHNICAL SYSTEMS

One of its functions at this stage might be to provide a means of selecting ship crews that are able and willing to participate in an experimental total-crew training scheme. An experimental scheme of this type would be specifically appropriate for ships that are likely to be replaced by more highly automated vessels. It would in this case be possible either to transfer a total crew or to provide a pool of potential recruits to take over new ships. This could be expected to reduce both the time and the difficulties involved in bringing new vessels into efficient operation. A possibly equally important consideration is that there is at present a widespread concern, especially among officers, about future career prospects. There is likely to be an increasing tendency for officers with good technical training to seek jobs ashore, leaving on board those who have nowhere else to go either because they lack technical training or because they are too old to be able to start a new career. The active involvement of officers and men both in training and in the development of new forms of organization required on automated ships could contribute considerably to the reduction, and possibly the reversal, of this trend.

60

CHAPTER 6

Emerging Characteristics of Socio-technical Organizations A Summary

---·--DIRECTIONS OF CHANGE IN ORGANIZATIONAL DESIGN

1. From adapting organizations to technological requirements to adapting technology to human and organizational requirements. 2. From designing organizations with a single, prescribed, rigid structure to designing organizations: (a) capable of learning and of relatively continuous change in organizational structure (b) with a capacity for operating different types of organizational structure, depending on task requirements. 3. From organizational designs based on one man-one job to designs based on enlarged and partly overlapping jobs. 4. From designs whereby responsibility for organizational units is allocated to one person or group (thus generating a hierarchical structure) to designs whereby responsibility is allocated to nonhierarchical self-organizing groups.

5. From organizations using electronic data-processing (EDP) techniques as work directors to organizations utilizing computer equipment as consultants to provide relevant information for selforganization and to assist problem-solving by operating groups. 6. Specific forms of organizational design have to be evaluated in terms of: (a) technological requirements (b) environmental requirements 61

DESIGN OF SOCIO-TECHNICAL SYSTEMS

(c) (d) (e) (f)

work-role and career implications educational requirements psychodynamic and mental-health implications social-ecological implications.

Each of these leads to different types of implication for organizational design, which optimally should be consistent with one another. At present a static organizational design is not possible, since each of the above categories is subject to change. What needs to be aimed at is, therefore, a directive correlation of all the characteristics technological, social, environmental, educational - that affect organizational design requirements. 7. The emerging science of socio-technical organization will need to be non-disciplinary. It cannot be contained within any one discipline, nor can existing problems be 'solved' independently by existing disciplines. What one is looking for is criteria for policy decisions with respect to optimal directions of change. DEVELOPMENT OF NEW STRATEGIES OF CHANGE

I. From policy-making based on the formulation of pre-specified plans to policy-making as a learning process within organizations.

2. From the separation of research and doing to the building of research capacities into organizations; it is no longer feasible for the research role to remain a prerogative of a privileged profession. 3. From the use of professionals as experts and implicit policymakers to the utilization of professionals in collaborative research projects. 4. A transition in educational techniques from predominantly programmed teaching to predominantly project and researchoriented teaching. SOCIAL MONITORING

The development of capacities for the rapid recognition and evaluation of emerging and evolving social, psychological, and technological trends.

62

PART 11

Approaches towards the Integration of the Physical and the Behavioural Sciences

CHAPTER 7

The Operational-unit Paradigm

---·--It will be shown in this chapter that there exists a sequence of isomorphisms, which extends all the way from ordinary everyday activities to both physical and behavioural laws. The analysis is based on the concept of an operational unit, which is defined in terms of: (i) an initial state of a system (ii) an operation or set of operations performed on the initial state (iii) the final state arrived at. In so far as operational units can be used to represent both physical and behavioural events, they can be utilized to interrelate the behavioural and physio-technical characteristics of task performance and work organization. At the same time, the operational-unit concept provides a possible basis for the integration of the physical and the behavioural sciences.

TASK SPECIFICATION

A task may be defined as an operation or set of operations performed on an initial state of a system which leads to a specified final state of the system. A task can be represented as an operational unit of the form

w(SJ ~so where '11' is an operation or set of operations on the initial state S; which leads to the final state S0 • An activity is defined as the performance of a task. In terms of the definition, the following types of task can be distinguished.

65

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

A(l) Transformation Tasks Both the initial state and the final state are given, and an operation exists which leads from one to the other. Thus, in the task of constructing a machine from a set of components, the initial state is the set of components, the operation is that of assembly, and the final state is the specified machine. Washing the dishes, closing a door, going to the office are all of this type. The initial and final states are in this case physically definable. Tasks of this type are of special interest since they can serve as a bridge between behavioural and physical theory. Furthermore, the operational-unit representation is applicable also to the transformation of behavioural states such as learning a poem or getting a job, provided that in each case the final states are definable.

A(2) Control Tasks Here, the final state S0 is given and, whatever the initial state, an operation exists to achieve a specific state S 0 or to maintain successive states within a specified range of outcome states. Task performance is not basically distinct from task control. Control tasks can be looked at as a special case of transformation tasks. Examples are machine repair, where, whatever breakdown has occurred, operations can be performed to get the machine going again, and machine maintenance, where any potential breakdown can be countered before it occurs. Other examples are keeping a class in order or looking after a prisoner - this last task is analogous to maintenance if the guard has to counter all possible escape attempts, and analogous to repair if he has to be able to recapture the prisoner after every escape. Both transformation and control tasks constitute in principle completely specifiable operational units. In the following, one of the three elements is unspecified, which generates a problem. B(l) Transformation Problems Given a state s, and a specified final state S0 , an operation or set of

66

THE OPERATIONAL-UNIT PARADIGM

operations has to be found for transforming Si into S0 • A typical example is the discovery of a mathematical proof.

B(2) Exploratory and Prediction Problems The initial state Si is given and an operation.,., but the outcome is unknown. If the operation is performed to discover the outcome, we speak of an exploratory task. If we are required to state the outcome before the operation is applied, we speak of a prediction task. The operation need not always involve doing something but may take the form of waiting for a specified period of time to elapse. For instance, in the case of weather prediction we are given a set of measures and are asked to specify the subsequent state a day or two ahead. Similarly in the prediction of planetary motion.

B(3) Inferential Problems Here we are given the final state and either (a) information about the preceding state is available and we are asked to infer the intervening operations, or (b) information about the preceding operation is available and we are asked to infer the previous state. Historical research and 'whodunits' present problems of this type. We note that a problem requires the discovery, and where possible the testing, of a complete task specification. Task specifications may take the form of a recipe in a cookery book or a set of instructions for assembling a machine. The task specification is invariant with respect to the person who carries out the instructions; that is, whoever carries out the instructions, the same results should occur. A task specification thus corresponds to a natural law in its most restricted and specific form. If, within a given field of phenomena, it is possible to arrive at a task specification, then it is often possible to arrive at an operational principle from which a set of different task specifications can be derived as a special case. We shall next consider the form taken by operational principles, which may also be referred to as models.

67

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES TYPES OF OPERATIONAL MODEL

C{l) The Input-Output Model

This model takes two forms depending on whether we look at the situation as one where an operation is performed on the system or as one where an operation is performed by the system. In the one case causality is taken to be located in the environment and in the other case it is taken to be located in the system. Starting with the operational unit 1r(Sj) ~ (S0 )

we can take the operation to be an input (I) which is applied to the system (S) leading to an output (P): ~(P).

I(S)

Here the input is applied to the system in order to achieve a specific output. In the case of the black-box model, which is equivalent to a transformation problem, only the input and output are given and the structure of the system is to be inferred. Alternatively, we can consider the system as operating on an input and producing an output: S(I) ~(P). From a purely formal point of view these two models are identical. They are, however, opposite interpretations of the same situation. Instead of locating casuality either in the environment or in the system we can interpret the situation as one where both the system and the environment perform an operation, 1rs by the system and 1rE by the environment, which leads to a final outcome state 'ITs.

'IrE

~ (So)•

In this case the initial state, which elicits the correlated operations, has to be given:

68

THE OPERATIONAL-UNIT PARADIGM

This is a type of Sommerhoff model (1950) obtained by joining the two uni-causal models together. 1 We can, if we wish, consider .,.E and .,.s as a set of operations performed by the larger unit which contains both the system and the relevant environment, in which case the representation can be collapsed to the form w(S,) -+ (S0 ) which is our original operational-unit form. This shows that: (1) If we interconnect a set of operational units, then the resulting system is again an operational unit.

C(2) The Medium-Thing Model The distinction between medium and thing has been discussed by Heider (see Heider, 1959). Whenever we observe a thing, then, in so far as the observer and the thing are distinct, an intervening medium is required. Thus, in hearing, the air acts as medium; in seeing, light-waves act as a medium; and in science, concepts and symbols act as the medium. Medium and thing characteristics are not inherent characteristics. We can make sound- and light-waves the object of study, in which case we need a new intervening medium. The wellknown anecdote of the Inan who pointed at the moon but found that everyone looked at his finger is meant to convey that, if we look at the concept, we cannot see the phenomenon to which it is intended to point. While the roles of medium and thing may change, some objects are nonetheless more suitable to act as mediums and others have more thing characteristics, at least with respect to a given operational context. An object functions as an ideal medium with respect to a given operation if, for the operational transformation w(S) -+P

for each distinct value .,.to .,.2 , ••• .,.,., there is obtained a corresponding set of distinct values Pto P 2 , ••• P,. In the communication model 1 Sommerhoff's model includes time as an explicit variable, which is not the case for the present operational-unit model.

69

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

is the sent message, the medium is the communication channel, and P is the received message. An object will have the characteristics of an ideal thing with respect to a given operation if, for each operation 7Tto • • • TT11, the same initial state remains. The possibility of defining thing-Jike objects in terms of invariants of a transformation was pointed out by Max Born (1964). 1 It will be noted that these properties depend on the type of operation applied, so that an object that stands out as thing in one context will not do so in another context.

1r

Looking back at the operational models discussed we find that each model can give an internally consistent and empirically testable view of the world. However, it is possible to construct many different types of operational model, and these may contradict one another in terms of their interpretation of what is observed since they look at phenomena from different points of view. An operationally formulated model is therefore not yet a general theory. The formulation of a general theory requires the use of a form of representation that is in itself non-operational but permits different possible operational interpretations. It is possible to distinguish between four successive levels in the formulation of scientific principles and laws: (a) Task specification principles, which we have discussed, are given in the form of instruction and skills that may be handed on from one generation to another. The men who found ways of constructing pyramids and cathedrals, of refining metals or devising boomerangs, all formulated and tested laws of nature at this level. Task specifications are typically incomplete. They may include elements or operations that are redundant although they are assumed to be essential; or they may omit significant elements and operations either because these are taken for granted or because they are difficult to describe. 1 Born, in his Nobel Prize lecture, stated this as follows: 'Well-developed concepts are available which appear in mathematics under the name of invariants in transformations. Every object that we perceive appears in innumerable aspects. The concept of the object is the invariant of all these aspects.' In a letter to Einstein, Born notes that the concept of 'observational invariants' is 'a descendant of Wertheimer's Gestalt in a new form' (see Born, 1971). This is possibly one of the few examples where psychological theory has bad some influence on the development of the theory of physics.

70

THE OPERATIONAL-UNIT PARADIGM

At the next level of generality we find: · (b) Operational principles, which make it possible to derive a whole range of possible task specifications, so that instead of knowing how to do one thing we now have a principle that makes it possible to formulate and test the conditions required to achieve a whole range of different goals. At the same time, the formulation of operational principles makes task specifications more determinate by eliminating redundant elements and operations and including missing elements and operations. The weakness of operational models lies in their basis in a manipulative relationship to the environment. The theories formulated at this stage take a quasi-causal form in which one element is taken to be the cause that operates on other elements. We find in this case theories that regard persons as causal agents who operate on their environment to produce results. Other theories take causality to be inherent in external stimuli which operate on persons and produce response outcomes. The different operational models that can be formulated are found to conflict with or oppose one another. The next level of generality is arrived at by noting that the initial distinction between elements and operations is an arbitrary one which is meaningful only in goal-oriented behaviour where, given an initial state of the world, we look for a way to achieve a specific outcome state. {c) Functional laws are arrived at if the distinction between operations and states of a system is dropped. The more general law now simply states a relationship of mutual dependence between states of a system. Since the basic operational principles are formulated in terms of three elements, the corresponding functional law will consist of three variables. A functional law is somewhat like a street map. A task specification and operational principles correspond to particular paths that may be traced out from one location to another. Just as the different paths that may be chosen will have the appearance of contradicting or opposing one another, so different operational models that initially appear to conflict with one another are now found to be different possible operational interpretations of a functional network structure. 71 p

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

At the next level of generality are: (d) Generalized laws, which are based on postulates that generate different types of functional law. In physics we assume that it is possible to produce, so to say, a single map of functional laws for the total physical universe. In the behavioural sciences, on the other hand, we find that different persons and also groups constitute different universes each of which can operate in terms of its own laws and, moreover, that these laws are subject to evolution and can change over time. Behavioural science therefore has to be based on the formulation of a set of postulates which generate the functional laws that can operate within each of the possible behavioural universes that can evolve (Herbst, 1970). FUNCTIONAL LAWS

As long as we investigate the world in terms of what we do or someone else does and the results that are obtained, it is natural to express these findings in terms of quasi-causal operational units. A transition to a general law implies in a certain sense the abandonment of an egocentric way of looking at the world. We find, first of all, phenomena such as planetary motions where the original idea of a person or an agent operating on the planet to produce its motion no longer appears meaningful. We note, further, that the distinction with which we started, between operations and states of the world, is an arbitrary one. We can look at all three terms of an operational unit simply as states of a system. If, moreover, we can define and measure each of these states in such a way that an equality sign can be substituted for the transformation sign, then the operational unit with which we started is transformed into a state equation of the form /(S1 , S2 , S 3 ) = 0. This now states simply that, for a set of variables S 1 , S2 , S 3 , there are only certain sets of values that are consistent with one another. While within the operational frame of reference the world consists of persons or agents that cause things to happen, the formulation of a state law transcends, albeit in a limited way, the egocentric point of view. The person who perceives himself to be acting on his environment now disappears within a network of states of the world

·n

THE OPERATIONAL·UNIT PARADIGM

where no variable is intrinsically internal or external, intrinsically an operation or a state of a system, but each of these interpretations is possible; and this is essentially the advantage of expressing relationships in non-causal functional terms. Each of the possible operational units is based on one point of view. A law that attempts to provide a universal statement has to embrace all the possible individual points of view. pHYSICAL LAWS

Let us consider, as an example, the gas law

f(p,v,T) = 0 which in this most general form states that if we obtain simultaneous measures of the pressure, volume, and temperature of a gas then only certain triads of values of these variables will be empirically observable. In the context of prediction we can put the relationship in the form

p = / 1(v,T),

v = / 2 {p,T), or T =

f 3(v,p)

so that, having obtained measures of any two variables, we can predict the simultaneous value of the third variable. The law is clearly reducible to an operational form. We can consider temperature as an operation performed on a given volume of the gas to obtain a certain pressure

T(v) -+p or we can consider pressure as an operation performed on the gas at a given temperature to obtain a certain volume p(T)-+ v.

However, as Wold (1966) has pointed out, not every causal interpretation is realizable. Thus it is not possible to change the temperature of the gas by modifying pressure or volume. It appears, then, that the distinction between variables and parameters of a physical system is not sufficient. A distinction needs to be made between state variables, which can be affected by a change of other variables that form part of the system, and parametric variables, such as temperature in the present case, which cannot be

73

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES!

the:

changed in this way within the normal range of application of law. Similarly, Ohm's law can be put in the operational-unit fonn voltage (resistance) ~current, and Newton's first law as force (mass)

~acceleration,

where resistance and mass are parametric variables. However, Einstein's reformulation of the law showed that predictable changes of mass can be produced as a result of changing acceleration. Any variable that is not parametric can be chosen as an operation or as a state of the system. However, some variables are more easily perceived as taking on characteristics of an operation and others are more easily seen as attributes of a system or object. For instance, in some relationships, time is an explicit variable. We can consider time as an operation that affects the properties of a system. It is more difficult to consider operations performed on time, which is not generally considered as an object, although there appears to be no objection to this in principle. Physical laws are thus expressible as or reducible to operational units, and this is in fact the way in which they present themselves whenever experimental manipulation techniques are employed. In each case: (2a) The formulation of an operational unit leads to a functional law which consists of three variables. (2b) To every functional relationship in terms of three variables there correspond six theoretically possible operational-unit interpretations. It is possible, then, to construct several networks of operational principles which are internally consistent and empirically valid. The interpretation of what is observed will be different in each case. Each law permits a number of different operational-unit interpretations, and each operational principle, however correct in itself, is only a partial and incomplete formulation of a general law. Of the different possible operational interpretations of a law, only some would normally be considered. For instance, the normal interpreta-

74

THE OPERATIONAL-UNIT PARADIGM

tion of the classical law was that if a given velocity (v) is applied to a constant mass (m) then a certain quantity of energy (E) results. It did not occur to anyone before Einstein that the principle could be interpreted as implying the possibility of transforming mass into energy in the form

E=mc 2 where c is the velocity of light. BEHAVIOURAL LAWS

Physical laws and physical objects are definable in terms of invariant properties found in a set of observations. Different operational principles correspond to different intentions, whether we want to predict, desire to get from one state to another, or wish to infer the past. However, the observations, the desire to achieve a certain state, and the directed effort to achieve it, are all at the same time the basic elements that define the psychological and behavioural characteristics of a person. The operational unit is the basic component of a physical system and also the basic component of behaviour systems. This is the case in so far as the operational unit corresponds to an aim-directed activity of a person, and these activities constitute the elements of behaviour systems. At least in classical physics an attempt was made to formulate principles of nature in a way that would eliminate the person who acts, perceives, and experiences. In relativity and quantum theory we find the first, although quite limited, step taken in the direction of recognizing the essential role of the actor and observer. In behaviour theory we start with the same type of operational unit. However, now the frame of reference in the analysis of data is the person who carries out different activities. Here we do not necessarily have to be concerned with eliminating those aspects of the world that we conventionally recognize to be physical. Since a behaviour system is composed of an interconnected set of activities, and activities can be represented as operational units, the system as a whole has the characteristics of an operational unit, and it follows that:

15

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

(3) The functioning of a behaviour system in so far as it can be

represented as an operational unit is expressible in terms offunctional relationships which consist of three variables. Different operational-unit interpretations that can be made of functional relationships lead to different types of operational model, each of which corresponds to an existing behaviour theory. The possible operational-unit models can be classified into basically two types: (i) transactional models, where the system is looked at in terms of its internal structure and its transactional relationship with the environment; (ii) component-structure models, where the output characteristics are looked at in terms of the state of the component elements of the system and the structural relationship between them.

The Transactional Model There are three ways in which the transactional representation can be formulated, all of which are found in the literature. We can consider the input (J) as an operation performed on the system (S) which leads to a given output:

I(S)-+ (P). This is the basic model of the behaviouristic approach. We can equally well look at the system as operating on and transforming an input, which results in an output:

S(I) -+(P). There is, however, a fundamental difference in interpretation. In the first case causality is seen to lie in the environment, which operates on a person, resulting in a given response. This is the type of model that an experimenter may adopt in explaining the behaviour of a 'subject'. He perceives the situation to be one where he manipulates the behaviour of the subject. In the second case causality is seen to reside in the person who acts on his environment in order to manipulate and transform it. This may be the experimenter's model of himself and also the subject's model of himself. 76

THE OPERATIONAL-UNIT PARADIGM

Different operational units thus correspond to different ways of looking at a situation. Both models can be looked at as possible interpretations of a functional law which states that the output (P) is a function of the state of the system (S) and the state of the environment (E):

P =f(S,E). This is the type of functional relationship proposed by Lewin. It corresponds to Sommerhoff's model if we interpretS and E as correlated operations of the person and the environment, which now lead to a result that is due neither to the person alone nor to the environment alone. Unlike the operational model, the corresponding functional relationship overcomes the need to adopt a uni-causal model. In applying this equation we now come to a fundamental problem of psychology. Within the life-space of the observer there exist objects and persons. Each person within the observer's life-space has, however, a life-space of his own which includes the observer. We have, in other words, the choice of studying the behaviour of a person in terms of the extrinsic frame of reference of an observer or in terms of the intrinsic frame of reference of the person. There is a somewhat analogous problem in physical theory. The paths of the planets were traditionally studied with respect to the frame of reference of the observer, and they turned out to be incomprehensible. The paths turned out to be quite simple, however, when they were plotted around the sun which actually determines their motion. This is essentially the change that Lewin showed had to be adopted in behaviour theory. The Component-Structure Model

The output (P) of a system can be looked at as a function of the components (C) of the system and of the structural connection (G) between them: P =f(C,G). In a behaviour system neither the components nor the structure is static. The components normally constitute ongoing activities and the structure has the nature of a constantly shifting pattern of coordination and control. A possible operational-unit interpretation

77

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

is in this case that the integration process operates on the components of the system, resulting in an output:

G(C)-+P. The component-structure model is isomorphic with the inputoutput model. The difference is one of interpretation of the variables in the general functional equationf(x,y,z) = 0. Let us consider the condition where the contributions to the output made by the components (C) and by the integration process (G) are independent, so that

P = k1/(C)+k,J(G). Suppose that k 2 = 0, then the output is determined only by the components, which is the elementarist behaviour model. If, on the other hand, k 1 = 0, then the output is completely determined by the structure of the system, which gives the Gestalt theory model. Both models will hold only within a restricted range of extreme conditions. In most cases normally encountered, both components and structure determine the output. An example is the functional relationship found in work-group functioning between the output rate of the group (P), the work rate of group members ( W), and the level of integration (G) measured in terms of the interaction process of group members. Provided (a) all members engage in work that contributes to the output, and (b) the task has the characteristics of a sequential path, then the relationship takes the form

P = k 1 W+k 2 G, that is, the output rate is a function of the work rate of group members and of how well they coordinate their activities (Herbst, 1970, Ch. 6). Another example is found in the study of Heiderian configurations consisting of an evaluative relationship between two persons and an object. The affect in the form of like-dislike associated with a configuration (L8 ) is in this case a function of the affect associated with the system components (Le) and the configurative characteristics of the system (LG) (Herbst, 1970, Ch. 12). When the interaction

78

THE OPERATIONAL-UNIT PARADIGM

effects are disregarded, the relationship is found to have the form Ls = aLe+ fJLG. In both cases the value of the parameters determines the extent to which the components and the structure respectively contribute to the output state. However, in the latter case an operational-unit interpretation may not be feasible since the relationship refers to only one type of variable. An analysis of a physical system can frequently be carried out by means of a single law. This is not normally possible in the study of behaviour systems. Here we need to extend the representation to permit the analysis of a network of variables. Functional Networks

Given a set of n variables which define the state of a system, then a set of (n-2) functional relationships of the form f(x,y,z) = 0 will generate the total network of relationships between the n variables: {4) Everynetworkoffunctionalrelationshipsoftheformf(x,y,z) = 0 is convertible into a cyclic network.

This is evidently the case, since in a cyclic network every variable has two neighbours. Methods for the study of networks of this type are discussed in the two companion volumes. Of special interest are the functional networks that are composed of linear or monotonic relationships. An example found in the autonomous group study {Herbst, 1962) is the relationship between stress {S), work rate of group members (W), level of integration (G), and output rate (P), where stress is defined as interference with or blocking of processes required for output production. Together with the relationship already considered we need a second relationship which can be formulated in terms of the effect that stress has on the work rate and the level of integration. The set of equations has the form W-W0 ..... {1) P-P0 = +b(G-Go) a

W-W G-G0

S+c S+d

0 -----= e--

79

..... (2)

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

where P0 , W0 are minimum boundary values which give the operative zero point on each scale, with P0 , W0 , G0 , a, b, e, c, d> 0 and d> c. The set of functional relationships can be put in the form

which shows that: (i) both the work rate and the level of integration increase the output rate and (ii) stress increases the work rate and/or decreases the level of integration.

This four-variable network will be seen to be structurally related to the Sommerhoff-type model shown below, keeping in mind that we can look at variables either as operations or as states of the system. We can, if we wish, look at stress as an operation performed on the system which is composed of a set of correlated states (S1, Sz) that together define the output state:

or we can, collapsing the representation, look at stress as an input (/)which is operated on by the system (S), resulting in an output: S(l) -+P. 80

THE OPERATIONAL-UNIT PARADIGM

This shows again that if we interconnect a set of operational units,

then the resulting system is also an operational unit. To summarize the discussion so far: in developing a behavioural science, we have a choice of different operational-unit models each of which can generate an internally self-consistent operational theory. Each of these is, however, only a partial theory in so far as it is limited by the specific choice made in the representation of the operational unit. Each of these represents a different point of view of the system, and each will be derivable from a functional relationship theory. To give an analogy, a functional relationship theory is like a map that gives the total network of paths. Different operational theories correspond to different paths that can be traversed, which are the result of arbitrarily choosing an initial point and a final point and finding out how one can get from one to another. The formulation of operational units can thus be the first step in the formulation of a more general theory. However, it is frequently advantageous to bypass the operational stage altogether and to go straight on to the testing of functional networks by means of multivariate measurement techniques.

81

CHAPTER 8

The Multiple-perspective Paradigm1

---·--The questions to be considered in this chapter are:

1. What kinds of relationship can we establish between the psychological, sociological, technological, and economic aspects of human behaviour? 2. What kinds of theoretical principle can be formulated linking different disciplines? 3. What are the basic methodological requirements for multidisciplinary research? The first question concerns the content, range, and general conceptual structure of behavioural science. Depending on how we approach this question we arrive at preliminary definitions of what is a behavioural unit, what are its components, and what kinds of relationship need to be examined both between internal components and between units. This will involve formulating the types of variable needed to describe and measure the structure and functioning of behavioural units, and here we have a critical junction point which leads on the one hand to theory formulation and on the other hand to method formulation (Figure 8-1). Theory and measurement are linked in at least two ways. If we proceed in the way outlined, they will be linked by virtue of their derivation from a common conceptual base. The transition to theory formulation requires the translation of the conceptual structure into a network of functional relationships between variables. The transition to method formulation requires the translation of the conceptual structure into operational definitions and the specification 1 This chapter was first published in 1965 under the title 'Problems of Theory and Method in the Integration of the Behavioural Sciences' in Human Relations, Vol. 18, No. 4, pp. 351-9.

82

THE MULTIPLE-PERSPECTIVE PARADIGM

of measurement conditions. Their derivation from a common conceptual base is essential to achieve structural isomorphism between theoretical concepts and operationally defined variables. All this belongs to the preparatory stage. The essential stage of scientific development is not reached until a continuous process of interchange is set up between empirical and theoretical work. On the one side, the theoretical formulation defines, if only in a rudimentary way at the beginning, the measurement conditions and the form of data analysis needed for theory-testing. On the other side, the data obtained will lead to clearer articulation and, where necessary, to changes in the theoretical structure. The initial conceptual structure plays in this case the role of a scaffolding that guides the joint development of method and theory and may later become transformed as a result of changes in the theoretical structure in so iar as this needs a new interpretation. FIGURE 8-1

PROCESS STRUCTURE REQUIRED FOR

SCIENTIFIC DEVELOPMENT

Conceptual structure

I

I

I

I

I

I

I

I

I

I

I

I \\

Method formulation

\

\

\

\

\

\

\

\

\

\

Theory formulation

This is the type of process structure that is best represented by the development of the physical sciences. Here the initial starting-point may be seen as the adoption of a sceptical attitude to the traditional view that the laws of nature are logically derivable from a priori or self-evident principles. In a sense the previously widely accepted a priori principles were not so much dismissed as demoted to a tentative conceptual structure which could be subjected to testing and modified in the light of empirical research. Practically every physical law was formulated initially on the basis of empirical research, either by data-fitting, as in the case of Kepler's laws of

83

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

planetary motion, or by noting invariant characteristics in the relationship between variables, as in the case of Boyle's and Charles's thermodynamic laws. However, the possibility of finding these simple laws was dependent on knowing where to look: the recognition, for instance, of what constitutes the simplest possible physical units, which are in this case closed systems with stable structural properties such as homogeneity or equilibrium, or units that tend towards simple structural states definable in terms of a small set of variables. The development of general theories at the next stage took the form of formulating a minimal set of relatively simple relationships that could be linked together to form a network of functional relationships from which other more complex principles or special cases could be derived. The major revolutionary theoretical development turns out to have been, however, in every case, the axiomatic formulation of a principle of measurement: in the case of Newton, the axiom stating the possibility of absolute measurement of physical quantities; in the case of Einstein, the derivation of the theoretical consequences that follow if the axiom is rejected; and again, in the case of quantum theory, the rejection, as its point of departure, of the axiom that the process of observation has no effect on the object observed. It appears almost certain that the development of a behaviour theory will intimately depend on the axiomatic clarification of the conditions of behavioural measurement. If something is to be learnt from the development of the physical sciences, it is, I believe, more in the necessary strategical conditions for scientific development than in the use of special physical techniques, concepts, or axioms. The type of process structure needed for scientific development, shown in Figure 8-1, may give some insight into conditions that will block or inhibit progress. This will happen if effort is exclusively centred on one sector to the neglect of others, or if any one of the process links does not function effectively. In sociology the tendency in the past has been for development to become arrested at the stage of formulating conceptual structures. Conceptual structure by itself is not yet a theory. It becomes fruitful in so far as it guides the development of method and theory formulation. However, if this transition is not made, then its function will be limited to that of an ideology which will influence one's way of thinking about and evaluating a problem without being testable.

84

THE MULTIPLE-PERSPECTIVE PARADIGM

A quite different development is found in the history of economics. Here the attempt was made to go directly to the stage of theory formulation. Guided by the view that the basic laws of physics can in a certain sense be looked at as tautologies, it seemed to be possible to formulate a priori equations of a kind that, combined with self-evident assumptions about human behaviour, would lead to a general theory. The main block in development arose through the relative independence of empirical and theoretical work. The theory influenced the types of method employed but there was relatively little feedback possible by which empirical data could modify the theory. The history of psychology shows a relatively sophisticated development. Almost from the start an intensive effort was made to clarify the conditions of psychological measurement and to set up a balanced interchange between empirical and theoretical work. Nonetheless, no theory was able to go into a sustained growth phase. At least one of the difficulties appears to have been a lack of congruence between theory and method, since these were not derived from the same conceptual structure and consequently the theoretical implications of methods such as statistical models were not necessarily congruent with the theoretical model to be tested. In spite of the considerable amount of empirical data collected over the past century, little of it is of a nature or in a form usable for theory construction. The possibility of formulating and testing behavioural principles depends in the first place on an appropriate definitional choice of a behavioural unit. We thus come back to our initial question. Is it in fact feasible to construct purely psychological, sociological, and economic theories, or is the situation, as Ackoff (1968) puts it, that 'nature may turn out not to be organized into disciplines in quite the same way as universities are' ? Whatever kind of choice we make between research strategies, and we are certain to make at least an implicit choice in any kind of research we are engaged in, it is worth while examining possible alternatives. The first conceptual model I should like to consider can be stated thus: (1) Disciplines such as psychology, sociology, economics, physics, etc. each have their own independent subject-matter and each is in a position to develop independent theories.

85

INTEGRATION OF PHYSICAL AND BEHAVIOURAL

SCIENCE~

From a theoretical point of view this means that the universe can be1 segregated into psychological, sociological, and economic phenomena: so that we arrive at psychological, sociological, and economic systems each of which exists as an independent phenomenal unit. This viewpoint has been challenged repeatedly, especially by research workers in the field who have found that, in dealing with practical problems, one cannot tease out separate psychological, economic, or technological bits, for the problem nearly always involves working with a total integrated organizational unit. For instance, in studying the functioning of a work group we cannot isolate the observed structure of interaction relationships from the technological structure, which determines the nature of the task to· be done and the types of relationship required for task performance. Nor can we disregard the economic aspects of behaviour either from the point of view of individual group members or in relation to the conditions for group survival. There is a need for the development of interstitial disciplines, among which social psychology, socio-technical analysis, and socio-economic theory are of special interest, since they demonstrate different ways in which bridging or unifying disciplines can be attempted, and they provide natural growing-points for the development of a unified approach. FIGURE 8-2

LINKING DISCIPLINES BETWEEN THE BEHAVIOURAL AND THE PHYSICAL SCIENCES

socio-economics

psycho-physics

The integration of all the disciplines shown in Figure 8-2 in order to arrive at a theory of behaviour may seem an impossible task. Attempts have been made to use computers, since these can deal with any number of variables and any desired degree of complexity. However, the problem encountered is one that first of all requires conceptual and methodological analysis to reduce it to manageable

86

THE MULTIPLE-PERSPECTIVE PARADIGM

proportions. To begin with, I think, we need to recognize that, given the initial assumption that the universe can be split up into psychological, physical, economic, etc. phenomena, then there is no way in which the various disciplines can be put together again. In this connection we need to consider a second model - a radical attempt to solve the problem of integration which suggests that all phenomena can be reduced to one type. Such monistic solutions, which seek to reduce apparent diversity to a single principle, may appear in the following forms: (2a) A./I events are material phenomena, or (2b) A./I events are mental phenomena.

If everything can be reduced to one thing, then it may not matter whether we call the one thing to which everything is reduced material or mental. In practice, however, adoption of the first view leads to depriving psychological phenomena of their reality status, and adoption of the second, with equal conviction, tends to deprive physical phenomena of their reality status. The problem of linking disciplines takes a different form in each conceptual model and is limited by the basic assumptions made. The segmented model encountered in, for instance, body-mind theories is limited to the postulation of isomorphic structure. There is no way of accounting for possible structural identities, nor can any kind of link be provided between psychological phenomena on the one side and material phenomena on the other. Within the monistic model in its physical form, the discovery of isomorphic structures, such as the structural identities between equations which describe mechanical and electrical phenomena, was of considerable importance. However, by themselves these provided no basis for integrating the physical sciences, a step that was made possible only by the formulation of transformation rules based on the development of a general concept of energy and the demonstration that physical energy could be reversibly transformed into electricity or heat. This is the basic model of what has in recent years become known as general systems theory, although in its programmatic form it is restricted to the formulation of isomorphic physicalist concept structures. The aim here is to find a way of integrating all the disciplines within a physicalist model. Exploration of the total range

87 0

INTEGRATION OF PHYSICAL AND BEHAVIOURAL

SCIENCE~

and the limits of the physicalist model is, I believe, an important task. Nevertheless, by adhering to a monistic physicalist model we cannot go beyond the formulation of physical laws and we thus bypass the problems encountered in integrating the various boo havioural disciplines and the behavioural and physical sciences. The fallacy of physical monism does not lie in the assumption that all events can be analysed as material phenomena. This is an axiom we may retain. The fallacy lies in the assumption that the only relationships that exist between events are those that are expressible in terms of physical dimensions. The problem of integrating different sciences arises at the point where we recognize that the physicalist model is only one of several analytical schemes. Scientific disciplines are not different from one another if they can all be subsumed under the physicalist model. They are different only in so far as they employ different analytical schemes. In order to integrate different disciplines we need to look for relationships that exist between different analytical schemes. With this problem, the conceptual models considered so far cannot help us. As a third alternative I should like to consider a multipleperspective-type model and discuss some of its methodological and theoretical implications. It can be stated as follows: (3) Every event can be analysed with respect to its role within a network of physical relationships or with respect to its role within a network of psychological relationships, or with respect to its role within a network of sociological, economic, etc. relationships. This implies in a more general sense that no event can be said to possess intrinsic or exclusive properties of being physical or psychological or economic, etc. It can have no properties apart from the relationships it has to other events. If we examine the event with respect to a specified type of relationship, then we define its properties relative to this set of relationships. If we examine the same event with respect to another type of relationship, we define a further set of different properties with respect to a second type of relationship structure. In the segmented model the problem of integration implies that, on the one hand, we have a set of physical phenomena and, on the other, a set of psychological phenomena, and a way has to be found of linking the two. In the present model, however, it implies that in

88

THE MULTIPLE-PERSPECTIVE PARADIGM

fact we are dealing with a single set of events whose properties are defined in the first place within a network of physical relationships and in the second place within a network of psychological relationships. Since in each case we are dealing with the same set of events, analysed with respect to a different set of relationships, the question then becomes one of determining in what way the two modes of analysing events are related to one another. Before considering this problem further at the theoretical level, I should like to examine its methodological implications by showing how it can be applied at the level of data analysis. The example I shall discuss is taken from the field of family studies (Herbst, 1952, 1954a). Suppose that, for each member of the family, we record what family activities he engages in. The results arrived at may be represented in diagrammatic form. In Figure 8-3 the husband's field FIGURE 8-3

OVERLAPPING ACTIVITY DOMAINS OF FAMILY MEMBERS

Husband

Wife

Child

contains all the activities in which he participates. The region of overlap with the wife's field contains all the activities in which he engages together with his wife. There is a region comprising his joint activities with the child which excludes the wife, and there is a region for joint activities involving all the members of the family. Data so obtained can be analysed in four distinct ways: (i) At the psychological level the set of activities is analysed to obtain measures for each individual person. (ii) At the social-psychological level the set of activities is analysed to obtain measures of the interaction relationships between family members and to map out the group structure in terms of activity relationships.

89

INTEGRATION OF PHYSICAL AND BEHAVIOURAL

SCIENCE~

(iii) At the sociological level the data are used to obtain measureJ, for the family as a group, such as the degree of joint action or thei minimum participation level required for survival of the family aa. a group. (iv) At the task-structure level the relationship between activities can be analysed irrespective of who does them. The data can in this case be used to study the differentiation of the set of activities into task regions and the topological relationship between task regions which determines the possible paths of locomotion of individual members within the task structure of the family. This empirical application of the basic conceptual model illustrates in a particularly simple way that the different disciplines do not involve different types of data but represent different ways or analysing the same data. The basic element in each case is an activity. At the psychological level activities are analysed with respect to their relationships within the activity structure of each individual and are thus used to obtain measures for each person treated as a behaviour system. At the sociological level the same data are used to obtain measures for the family group as a behaviour system. A further set of measures can be obtained by analysing the same data in a different way again to determine the characteristics of the task structure. The problem of integrating different disciplines will now be seen to be quite different from that implied by the segmented conceptual model. The task with which we are faced is not that of establishing links between intrinsically different types of phenomena but that of devising means for coordinating different analytical schemes that refer to the same basic data, where each analytical scheme represents a structural analysis of the data with respect to a different network of relationships. There are at least three possible approaches to the coordination of different disciplines within the multiple•perspective model, which have been applied in studies of child behaviour and group functioning. (a) Isomorphic Principles

One approach is to establish isomorphic structural principles between 90

THE MULTIPLE-PERSPECTIVE PARADIGM

different analytical schemes. This approach is also implicit in previous conceptual models, but the present principle provides more guidance as to how it can be achieved. As a first step we note that, when activities are used as basic elements, both individuals and groups can be represented as behaviour systems. We can then formulate a set of variables that define the structure and functioning of behaviour systems, look for appropriate measures for each variable at the level of individual behaviour and at the level of group functioning, and test whether identical characteristics emerge in the networks of functional relationships obtained. An example is given in Herbst (1957a, 1961, 1970). The functioning of a behaviour system can be represented in part by a network of theoretically definable variables which include the rate of functioning of activity elements, the level of coordination between activity elements, output rate, task involvement, and stress. The types of measure employed in the study of individual behaviour and in the study of group functioning may be different. In some cases a similar type of relationship can be shown to hold, but this is not necessarily the case, since behaviour principles do not as a rule have a unique universally valid form. However, the available empirical data support the hypothesis that the construction rules for the possible range of behaviour principles at the levels of individual behaviour and of group functioning are identical. An alternative procedure is to start off at the methodological level. In this case we begin by constructing a corresponding set of measures for different sets of phenomena in so far as these can be represented in terms of the same type of conceptual framework, and then go on at the next stage to the formulation and testing of a common set of theoretical principles. An example presented in Herbst (1962) shows that the concept of an activity domain can be applied to the measurement of the behaviour characteristics of a person, a task, and a group, using the same set of data on work-group functioning. In the case of a person, the activity domain is defined in terms of the number of activities he engages in; the corresponding person domain of an activity is the number of persons who engage in the activity. At the level of group functioning, a corresponding activity domain can he defined in terms of the number of activities the group engages in, and a person domain in terms of the number of group members. 91

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

Since for every domain of this type corresponding sets of measures can be constructed - such as number of elements in the domain, rate of change of elements, and participation characteristics of elements - it becomes possible to test whether a similar set of relationships exists for person, group, and task characteristics. In actual research practice these are not mutually exclusive procedures, but, as in bridge construction, building on both sides has to be simultaneously and mutually coordinated. (b) Transdisciplinary Principles The relationships found between variables within one analytical scheme cannot be independent of those arrived at by employing an alternative analytical scheme. For instance, the way in which a group functions will be subject to limits and constraints arising from the structural characteristics of the task. In the special case where the task structure is held constant, the relationships between the variables of group functioning must involve physical-technical characteristics of the task as parameters. In the case of the work group discussed in Herbst (1970, Ch. 6), where all members participate in the group task, and the task is reducible to a sequence of operations, the output rate (P) can be expressed as a function of the work rate ( W) of team members and the level of group integration (G) by the equation P-P0 =

(W~Wo) + b(G-G0 ).

The parameter a specifies the extent to which work rate under given task conditions is converted into an output. The parameter b can be interpreted as the extent to which task components are well coordinated, since an increase in the value of b means that less coordination skill by group members is required (work rate being kept constant) for the maintenance of a given output level. The principle in this case relates a set of variables of group functioning, while the parameters refer to physio-technical characteristics of the task. This means that also within the present conceptual framework it is possible to formulate purely sociological, economic, or psychological principles. However, this is possible only in so far as variations with respect to other alternative analytical schemes are

92

THE MULTIPLE-PERSPECTIVE PARADIGM

kept constant, so that the principles obtained will necessarily contain parameters that derive from other disciplines.

(c) Transformational Principles Going one step further let us assume now that simultaneous variations exist relative to different analytical schemes. Some of the parameters will now be turned into variables and the principles we arrive at will link variables deriving from different disciplines. An example is the relationship found, under similar conditions of group functioning as before, between the number of employees (N) of a retail shop and the sales turnover (S) obtained. This has the form

N-No = a 1 (S-S0 ) and relates a sociological to an economic variable. The formulation of interdisciplinary relationships need not always require the construction of new principles but may involve little more than a redefinition of terms, which implies at the operational level a transformation of measurement scales. Output rate in the first case and sales turnover in the second are in fact different measurement scales applied to the same referent. In the first case output rate is defined and measured as a component within a network of variables of group functioning. In the second it is measured in terms of the market value and thus related to a network of economic variables, in which case the variable of group functioning now becomes transformed into an economic variable. This method is discussed in more detail in Chapter 9. SUMMARY

The three methods for integrating different disciplines considered so far may be swnmarized as follows. The method of isomorphism consists in finding a set of variables (At> A 2 , • • •) within one discipline that have an isomorphic structure with a set of variables (B 1 , B 2 , •• •) from another discipline. Isomorphism by itself does not constitute an integration of two disciplines. The second method is based on the fact that functional relationships between variables (A 1 , A 2 , •••) within one discipline may contain 93

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCBii

parameters (b~o b2 , •••) from another discipline, resulting in • functional relationship that integrates two disciplines, in the form f(A.l> A2 ,

••• b~o

bz)

~

0.

The third method is based on the possibility of transforming a parameter into a variable b 1 -+ B~o or finding a transformation rule by which a variable defined within one analytical scheme is trans. formed into a corresponding variable within another analytical scheme, say A. 2 -+ B~o so that in either case we arrive at a functional relationship of the form f(A~o B~o

... b2 , b3 )

~

0

where variables within one discipline (A. 1 , A. 2 , •••) are now linked to variables within another discipline (B1 , B2 , •••). Finally, I should like to suggest a fourth possible method which exists theoretically, but for which, as far as I know, no empirical demonstration has as yet been found. This potential method ia based on the possibility of stating the dual of a function arrived at by transforming all variables into parameters and all parameters: into variables, so that /(At, Az, · · · b~o hz)-+ f(a~oa z, · · · Bt, Bz).

We have already considered a behaviour principle that contains physio-technical characteristics as parameters. This suggests that there may be physical principles that in some form contain behavioural characteristics in their parameters. Suppose, then, that in a behaviour principle the physical parameters are transformed into variables, and the behaviour variables are transformed into parameters or constants. If this could be done, then it would be possible, starting with a behaviour principle, to turn it inside out, so to say, by stating its dual and thus arrive at a physical principle, or, starting with a physical principle, by stating its dual arrive at a behaviour principle. In view of the fact, however, that behaviour principles can take a range of different functional forms, a simple transformation of this type may be difficult to achieve.

CHAPTER 9

The Psycho-physical Transformation Paradigm

---·--Human behaviour that involves a transformation of the physical environment cannot be described and understood as a purely psychological phenomenon, but is given as a psycho-physical event. A theory of behaviour cannot therefore be formulated in terms of purely psychological concepts, but has to provide explicitly for an integration of psychological and techno-physical variables. If we observe any kind of human activity, such as lifting an object, writing a letter, driving a car, etc., it is not possible to split the observations obtained into separate and distinct physical and psychological phenomena. Consequently we cannot utilize a conceptual representation based on the assumption that there are distinct physical and psychological systems standing in a relationship of interaction with one another. Instead, the type of model required is one that will make it possible to specify the relationship between two different, but overlapping, conceptual representations of the same total process. Given that a behavioural act can be represented in terms of a set of physical variables (and parameters) p, and in terms of a set of psychological variables (and parameters) t/J, then the sets of physical and psychological variables will overlap, so that some corresponding variables will appear in both representations, as shown below:

P1 Pz P3 P4 (physical variables)

I I

+s +6 t/11 t/12

(psychological variables).

In the physical representation there will be a set of variables {p 3 , p 4 ) that do not have corresponding variables in the psychological model, but that are required in order to formulate a CQ.mplete physical

95

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

model expressible in a set of equations, say, ft(pt,PloPJ)=O} p h. ys1cal modl e. /2(p2, Pl• P4) = 0 Similarly, there will be a set of psychological variables (t/1 5, ,P6) such as, for example, task involvement, interest, anxiety about the outcome, performance satisfaction, etc., none of which has a corresponding variable in the physical model, but which are required in order to formulate a complete behavioural model, say,

gt(tfot, t/12, tfos) = 0 }

psychological model. gl(t/12, 1/Js, t/16) = 0 Our main interest is in the set of variables and parameters that appear, in different forms, in both the physical and the psychological representations of events. Fechner's contribution lies in pointing out that it is here that a bridge can be built between physical and psychological variables, and in demonstrating that there exists a functional relationship

"'=

w{p) which transforms a physical variable (p) into the psychological variable (tfo). The parameters of the functional relationship may vary from person to person. Given that a physical measurement scale can be transformed into each of the possible psychological measurement scales in terms of which different individuals operate, then the same principle also makes it possible to transform the psychological measurement scale in terms of which one person operates into the psychological measurement scale in terms of which another person operates. There are two forms of the psycho-physical principle that have been widely used in empirical data analysis. These are the FechnerWeber principle, which has the form of a logarithmic transformation,

t/1 = a 1 log (p-b1)+c1, and Brentano's (1874) and Stevens's (1957) principle, which has the form of a power transformation,

t/1 = a2(p-b2Y2 • Luce (1959) shows that the possible forms of the psycho-physical principle are theoretically derivable if the scale properties of both

96

THE PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

the physical and the behavioural measurement scales are given, since these are sufficient to determine the functional form of the transformation principle. The practical difficulty is that the scale properties of behavioural measurement scales are not given a priori. It is true that the researcher can impose different scale properties on the behavioural measurement instruments he employs. However, the measurement scales used will have no validity unless they match those on which individual behaviour is based. If the scale characteristics of behavioural measurement scales are, to begin with, treated as unknown, Luce's approach can still be applied in a different form, as shown below: Scale properties Scale properties of the physical of the behavioural measurement scale measurement scale

Luce's approach Present approach

Specified Specified

Specified Derivable

Psychophysical principle

Derivable Specified

If in this case the conditions that have to be satisfied by the psycho-physical transformation can be independently specified, then the scale properties of behavioural measurement scales become derivable. This problem is evidently insoluble within the limits of traditional psycho-physical theory in so far as no independent criteria for a transformation principle can be formulated. However, a possible solution exists if we consider a behaviour variable not as an isolated phenomenon, but as part of and defined by a principle linking a set of behaviour variables, and then ask what conditions have to be satisfied by transformations applied to a functional relationship between behaviour variables. The minimal condition that can be made is that the transformation should not change the functional form of the relationship. The most general form of the algebraic transformation that satisfies this condition is known to be: 1 1 This transformation was first formulated without proof by Waring, and later developed further in the geometries of Plucker and Mobius (cf. Coolidge, 1963, p. 269). In the companion volume, Behavioural Worlds (Herbst, 1970), a transformation of this type is applied to the derivation of the possible forms of behaviour principles, given that the dimension of behaviour variables can be

specified.

97

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCES

X=

Ax'+By'+C Y = ax' +by' +c

<XX'+Py'+y ax'+by'+c

.... (la)

Applied to a single variable, the psycho-physical principle is in this case found to have the form of a projective transfonnation: .... (lb) If this transformation is applied to physical variables measured in terms of the usual Cartesian metric scale, then behavioural measure· ment scales will satisfy the conditions of a projective scale. (It should be noted that in the simpler but non-homogeneous form of equation (lb) the principle will not ensure invariance of the functional form of an equation unless a restriction is introduced based on the dimension of the variables to which the transfonnation is applied.) We shall now consider a number of examples illustrating the application of the basic psycho-physical model, and then examine further the implications of the projective form of the psycho-physical principle. For the purpose of illustration we choose the simplest type of task, where the output rate P is a direct function of the work rate W employed. If we assume that both variables are measurable in terms of physical variables, then the physical relationship is given by the equation

w

P=-

a

•... (2a)

where the parameter a represents the extent to which the work rate is transformed into an output rate. The greater a is, the greater the work effort that has to be employed to produce the same output. Thus if the task involves turning a wheel, a may correspond to the frictional pressure that has to be overcome. If the parameter a is unity, then P= W. Then work rate and output rate become identical. Since both variables are in this case measurable by means of the same measurement scale, they have the same (physical) dimension. Let us denote

98

THE PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

the dimension of these variables as [Q], then equation (2) takes the form

1

[Q] =- [Q].

a

This implies that the parameter [Q]

a= [Q] = [1]. The parameter a is in this case said to be dimensionless. The dimensional equation, which after substitution becomes [Q] = [Q], is said to be dimensionally homogeneous, since both sides have the same dimension. In order to construct the corresponding behavioural equation we need to transform the physical measurement scale into all the possible individual (psychological) measurement scales in terms of which different individuals may operate. Thus the same (physical) work rate may correspond to a low work effort for one person and a high or impossible work effort for another; and a given increase in the work rate may be experienced as a small increase by one and a large increase by another; and similarly with regard to the output rate. If we denote the psycho-physical transformations of the (physical) work-rate and output-rate scales by 17( W) and 17(P), then the relationship between the physical and psychological models is P ,

W

I

I

1l{P), 17( W)

(physical variables) (psychological variables).

If we apply the psycho-physical transformation given by equation (la) to the physical relationship, then we obtain the corresponding behavioural relationship cxP+fJW+8

= a(AP+BW+C)

which can be put in the form 8(P-P0 )

=

ae(W- W0 ).

. ... (2b)

P0 and W0 are the zero points on the individual behavioural measure-

99

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCE.

ment scale, and these generally correspond to minimal boundat) values below which the behaviour system will cease to operate: E gives the rate at which an increase in the physical work rate ia experienced and responded to as an increased work effort; and a gives the rate at which an increase in the physical output rate ia experienced and responded to as an achievement. If we denote the dimension of the psychological work-effort and output scales by [P], then, since the zero points of each scale have the same dimension and since, further, the rate parameters E and 3 are dimensionless, the dimensional equation has the form [P] = [P]. If we now apply the same transformation again, we can transform the behaviour principle in the form in which it operates within the behavioural universe of one individual into the form it may take in the behavioural universe of another individual. However often we apply the tranformation, neither the linear form of the functional relationship nor the dimensional homogeneity of the equation will be changed. In other words, the transformation 'tT{p) satisfies the condition 1T"(p) = 'tT{p). It will be noted that the behavioural equation (2b) still contains the physical parameter a. However, when behavioural equations are fitted to data, such parameters cannot generally be independently estimated, since they combine multiplicatively with scale transformation parameters. We can now consider the application of the transformation to an example encountered in a study of autonomous group functioning (Herbst, 1962, 1970). The task of the group is in this case determinate, i.e. the output rate is predictable and depends both on the work effort of individual team members and on how well they coordinate their activities. Performance satisfaction (F) is a function of the output rate achieved (P) and the work effort deployed (W). We then have P ,

W

I

I

(physical variables)

F, n{P) , n{W) (psychological variables).

100

THE PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

Performance satisfaction under the above conditions can be shown to be a function of the output rate achieved and the extent to which a high output rate is achieved with a relatively low work effort. Performance satisfaction can then be put in the form of a psychophysical transformation of the physical output rate and the work rate

where a is again the physical parameter that corresponds to the extent to which the work rate is translated into an output rate. The psycho-physical transformation of the right-hand side of the expression is obtained by substituting the results from equation (2b), which are: P-+8(P-P0 )

W-+ €(W- W0 ). The behavioural equation can then be put in the form

F = g(P-Po)

[(P-Po)-~(W- Wo)]

where f can be identified as a performance expectation parameter. If/is low, then for a given work effort a relatively low performance result will be sufficient to achieve performance satisfaction. If f is high, then for the same work effort a high performance result has to be achieved for performance satisfaction to be experienced. This equation was found to fit the data obtained, by means of a continuous work record over a three-month period, in a longitudinal case study of an autonomous work group (Herbst, 1962, 1970). The relationship was found to hold for all except the initial one and a half weeks during which the group established its work organization. The basic data in this study were a set of mostly physicalist data on group functioning, and the projective transformation was applied to formulating the network of relationships between behaviour variables. The parameters g, J, a, are slope coefficients of the linear components of the equation and thus dimensionless. The dimensional equation in this case is [P2]

=

[P2].

Repeated application of the measurement-scale transformation again 101

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCE&

leaves the functional form and dimensional homogeneity of the equation unchanged. Performance satisfaction is found to have the dimension [P2 ]. This is one way in which the dimension of behaviour variables can be determined and empirically tested. Behaviour variables such as work effort, (internal) coordination, output rate on determinate tasks in so far as this is independent of external events, and also performance satisfaction, are all examples of internal variables which are definable in terms of the internal state of a behaviour system. Measurement-scale transformations of internal variables in the cases discussed take a linear form. Behaviour variables of a second type are those that are transactional, that is, they can be defined in terms of the state of the behaviour system and a simultaneous correlated state of the environment. An example is stress experienced with respect to a given situation. What types of condition produce stress depends on what the individual concerned seeks to achieve. We restrict ourselves to conditions where aim achievement is linked to physically measurable variables. Let us assume that a person is engaged on a task where the performance level achieved depends on the time available. Provided in this case that performance achievement also constitutes an operative aim for the individual, then the degree of stress will be linked to the amount of working-time lost, which may be due to breakdown of required technical equipment. Stress will in this case be linked to a physical variable with a time dimension. However, for different types of task and different criteria of aim achievement, stress may be linked to other physical variables that have a different dimension. We now require that stress should have the same dimension, whatever the specific dimension of the physical stressinducing event. This is po~sible only if the scale transformation has a non-dimensional form, since in this case the specific properties of the physical measurement scales are cancelled out. The dimensional expression for stress will in this case take the form

[S)

rsr

The simplest algebraic form that corresponds to this dimensional expression is aS+b cS+d 102

THE PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

which is again the projective form of scale transformation. This is possibly a simpler and more direct way of deriving the projective form of the psycho-physical principle. The linear transformation previously discussed is a special case of the projective transformation obtained by setting c = 0. In order to apply this form of scale transformation to each variable separately, we need to know whether to apply the transformation in its projective or linear form. It is shown in Herbst (1970) that the available data for group functioning and individual behaviour, obtained with techniques for studying single cases, are consistent with the following postulate. Given a simple behaviour system that allows measurement of any given behaviour variable, either internal or transactional, by means of a single measurement scale, then: (la) All internal variables are subject to a linear scale-transformation. (lb) All transactional variables are subject to a projective scaletransformation. The principle has not so far been systematically tested directly by means of the type of data found in psycho-physical research. One problem is that data are frequently presented as averages for a population. Another more fundamental problem lies in the difficulty of determining directly the characteristics of the specific measurement scale in terms of which a given in
P-+P 2 then by repeated application of the transformation we obtain

P, P 2 ,

r, ....

Thus the functional form of equations will change every time the measurement-scale transformation is applied. The reason for requiring that the identity transformation should

103 H

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCBI

TABLE 9-1

EVALUATION OF DIFFERENT FORMS OF THE PSYCHO-PHYSICAL PRINCIPLE

Type of psycho-physical principle Logarithmic transformatum

Power transformation

Projective transformation

Repeated application of the transformation maintains unchanged the functional form of behaviour principles and their dimensional homogeneity

No

No

Yes

Accounts for the identity transformation as a special case

No

Yes

Yes

Allows all-or-none response as a special case

No

No

Yes

Maximum number of parameters required for fitting data

3

3

3

be possible as a special case is as follows. If, in a traditional form of psycho-physical experiment, physical distance obtained by means of a measurement stick is compared with estimated distance obtained without the use of a measurement stick, then it should at least be possible for the two measures obtained to be identical. This condition is satisfied if the transformation takes a linear form as a special case. It is therefore satisfied by the power transformation, but not by the logarithmic transformation. With regard to the third criterion, psycho-physical data generate as a rule a continuous curve. However, it is possible in the study of single cases to obtain a qualitative all-or-none response instead of a response that varies quantitatively. For instance, in the example discussed, where experienced stress depends on time lost owing to task interruption, it is possible for the response to be related only to the incidence of task interruption as such and to be independent of the amount of time lost. In a study of pupil-task relations at school (Herbst, 1970) a case is found where it is not the intensity of anxiety

104

THE PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

that is related to work effort, but simply whether anxiety is present or absent. Anxiety is in this case elicited as an all-or-none response. Neither the logarithmic nor the power transformation can account for cases of this type. All these three criteria are satisfied by the projective transformation. The projective transformation in the form of equation (la) can be applied repeatedly without altering the functional form or the dimensional homogeneity of an equation. If we write the projective transformation p(X) in the form in which it can be applied to a single variable, (X) = aX+b P cX+d then, with c formation

= b = 0 and a = d = 1, we obtain the identity transp(X) =X.

If a= c = 1, b = 0, and d quantity, we obtain

= E,

where

E

is an infinitely small

X p(X) = X+E.

The expression has a zero value when X is zero and is unity for all positive values of X, and thus operates in this form as an all-or-none function. There is no difference between the three principles with regard to the fourth criterion, namely the number of parameters required for fitting data. As far as reliability of parametric estimation is concerned, the power function does appear to have certain advantages. In order to test the empirical validity of the projective principle, particular attention has to be paid to qualitative characteristics of the function that distinguish it from both the logarithmic and the power function. One difference is that the projective function can account for all-ornone responses. The other basic difference is that, while both the logarithmic and power functions increase without limit, the projective function can reach a limit beyond which no further effective discrimination occurs. 1 No further discrimination will occur if the 1 Junge (1966) concludes that the power function provides a good approximation to data within the low and medium intensity ranges, but that a levelling

105

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENC

function takes the form of an all-or-none response. At the time, the validity of any psycho-physical transformation is restri to conditions where parametric steady-state conditions are tained. There are a number of cases where quantitative increases · intensity, such as increases in sound, brightness, pain, etc., beyond a certain point, induce a change in parametric values that the behaviour system alters its mode of functioning or b down. In connection with this review of some of the differences betw the three forms of the psycho-physical principle, it should be no that the logarithmic and power principles cannot be tested in same way as the projective principle. As Luce (1959) has sho the functional form of the psycho-physical principle depends on scale properties imposed on psychological measurement scales. Thi · however, does not solve the problem of finding the scale properti of the measurement scales on which individual behaviour is based~ The conclusion arrived at on this point is that the measureme~~ scales in terms of which individuals operate may be based not on th~ geometry of a Euclidean metric space but on the geometry of ~ non-metric projective space. A non-metric projective scale doest! however, contain the Euclidean metric scales, both ratio and interval~ as special cases. 1 i The most interesting implication of the projective theory o~·' psycho-physical transformations is that the formulation and testinofbehaviour theories need not depend on the possibility of developina! equal-interval or ratio scales for behaviour variables, although it: will be of considerable advantage if at least some behaviour vari.i: abies can be linked with physicalist variables, measurable by means; of a ratio scale. ' ; Projective geometry is based on the definition of a set of point~; and a set of (straight) lines, together with a relationship of incidenooJ tendency towards a limit is observed for high intensities. However, his data;i from a study by de Valois et al. (1959), are based on physiological responses light. ; 1 A projective scale that leaves interval distances undefined can be trans-: formed into a Euclidean metric scale by applying a logarithmic transformation to the cross-ratio of points on the scale. This might make it possible to re-; formulate Fechner's logarithmic principle. The power principle, however, in so far as it assumes that human behaviour is based on the properties of a Euclidean metric space, appears to have only restricted validity.

to:

106

THB PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

between the points and lines. The set of points may be finite. The possibility of defining a straight line implies methodologically the possibility of constructing a uni-dimensional scale. The possibility of defining discrete points, incident with a line, implies methodologically the possibility of defining minimally discrete quantal intensity levels for psychological variables. The traditional approach to projective geometry has been to start with the definition of a Euclidean line, which is submitted to a projective transformation leading to the construction of a projective metric space (see Figure 9-1). There is an interesting analogy here with the traditional strategy of psycho-physics where, starting with physical variables measurable in terms of a Euclidean metric, we look for a transformation into the metric that operates in a psychological space. FIGURE 9-1

PROJECTIVE TRANSFORMATION OF A BUCLIDEAN METRIC

Euclidean

_..--metric

Proj~ctive

mefr1c

pl3) The construction of a coordinate projective geometry is based on the fact that, although distance between points and lines cannot be defined, any three points on a line can arbitrarily be defined as the zero point X0 , unit point X 1, and the point of infinity X 00 • In the construction of a psychological measurement scale, any three scale points may be arbitrarily denoted as 0, 1, oo. However, with regard to any fourth point, the requirement for a projective scale is that, for any given four points Xto X 2 , X 3, X4 , the cross-ratio

{Xt- Xz)(X3- X4) (X1-X4)(X3-Xz) remains invariant.

107

INTEGRATION OF PHYSICAL AND BEHAVIOURAL SCIENCE,

If a projective psychological scale is derived by a transformati~ of a physicalist interval scale, then the cross-ratio will retain thei value -t for any set of equal interval points. Thus in Figure 9-IJ where the values given are 0, 1, 2, 3, the cross-ratio is

{0-1)(2-3) = (0-3)(2-1)

-t

and this value will be maintained for any projective transformation of these scale points. There are at least four different types of problem with which psycho-physical theory can be concerned. Type 1 : The relationship between two forms of measurement of a physical variable. For example, the relationship between measures of distance obtained by means of a meter stick and by means of direct judgement. In both cases the information obtained is based on the sense of sight, but in the one case it is mediated by the use of an instrument and in the other case no instrument is employed. Another example would be measurements of weight by means of a balance and by means of lifting. Here, different senses are used. However, in both examples we are concerned with different forms of measurement of a physical variable, and one cannot necessarily say that a form of measurement that utilizes an instrument constitutes a physical phenomenon while one that does not utilize an instrument constitutes a psychological phenomenon. What appears to be meant by subjective measurement in this case is the use of instrumentally unmediated sensory experience. Type 2: The relationship between a physical variable and a physiological variable. For example, the response of specific types of visual cells to frequency or intensity of light. The relationship in this case is between two physically measurable variables. Type 3: The relationship between a physical variable and a qualitatively different experienced sensation. For example, the relationship between the physical characteristics of light-waves and perceived colour, or the relationship between the physical characteristics of soundwaves and experienced sound. Unlike Type 1 relationships, direct experience does not have the same characteristics as the physical 108

THE PSYCHO-PHYSICAL TRANSFORMATION PARADIGM

phenomenon that is mediated by an instrument, and the relationship

is between qualitatively different variables. Type 4: The relationship between a physical variable and a linked psychological variable. For example, the relationship between

experienced achievement level and variations in productivity rate, or the relationship between stress response and working-time lost owing to technical breakdowns. The relationship here is between physical variables on the one side and qualitatively different experiential or behavioural-response variables on the other. Unlike each of the previous types, Type 4 relationships are not necessary, but constitute linkages that come into operation within specific intentional structures. Thus a particular productivity rate will be experienced as an achievement only in so far as the production rate of a particular object is that which an individual seeks to achieve. Similarly, task interruption will be experienced and responded to as stress only in so far as it interrupts something the individual seeks to achieve. (In the case of the projective transformation of the stress variable (p. 102), the linkage between the physical event and the psychological response will cease if the parameters a and b are both zero.) It cannot be assumed that the specific characteristics of the transformation wiii be the same for all four types of relationship. We note particularly that Types 1-3 constitute necessary linkages, which means that the transformation will contain specific parametric values. This is not the case for Type 4 relationships. Also, Types 1-2 can be looked at as relationships between physicalist variables. The theory of projective transformations has been developed and applied to the study of Type 4 relationships (Herbst, 1967, 1970, 1972). The basic principles of the approach should be applicable to the other three types as well. Whether, and to what extent, the projective transformation is applicable across all four levels is a problem for further empirical research. Recent work by Atkin (1972) indicates that the theory and laws of physics can be based on a projective geometry and do not require a Euclidean metric. If this is the case, then there is a possibility for a basic and convergent reformulation of both the phy11ical and the behavioural sciences.

109

PART Ill

Characteristics of Tasks and Organizational Structure

CHAPTER 10

Production Tasks and Work Organization1

---·--INTRODUCTION This chapter comprises a series of papers outlining an approach to the conceptual representation and measurement of the relation between social and technological organization. The viewpoint adopted is that the problem requires the use of concepts that permit the simultaneous representation of both social and technological facts. Unless there is a common language for expressing behavioural and technical processes their interrelationship cannot be clearly traced out. Although for certain purposes it may be helpful to look at the social and technological structures as separate components of a production system, from the present point of view it will be more useful to consider them as two alternative frames of reference in the study of the complex interrelated processes within a production system. In practice there is no social or behavioural event that has no technical, material, or ecological components, and there is no technical process or ecological boundary that has no behavioural and social components. The concept of a pure social system without technological components or of a technological system without behavioural and social components is, therefore, somewhat of a fiction. The basic concept employed is that of an activity which may be analysed both with respect to its behavioural and with respect to 1 A collection of working papers written during 1954-56 at the Tavistock Institute of Human Relations in London. I am grateful to Dr F. E. Emery for help in making a selection from the original papers, which was later circulated under the title 'Task Structure and Work Relations' (TIHR Doe. 528, 1959, mimeo). Since these papers represent one of the first attempts made to develop techniques for studying socio-technical systems, they have been left largely in their original form apart from the substitution of the term 'variance', which came into use later, for the term 'disturbance'.

113

TASKS AND ORGANIZATIONAL STRUCTURB

its technological components. In other words, activities are treated] as the point of bifurcation between behavioural events on the one' hand and physical or technological events on the other. An activity may be altered by changing either its behavioural or its material and technological components, and it is by no means self-evident that these can be separated as neatly as it is generally assumed. The behavioural world and the physical world are not in practice two distinct and separate worlds, but two alternative conceptual analytical schemes with respect to which an event can be evaluated. For instance, a technical breakdown is a technological event that has certain effects on the technological process. It is also a behavioural event that, in the form of stress or in the form of support of aims, impinges on a person or group, interfering with the work carried out (which is the technological process looked at from the socialbehavioural point of view) and resulting in a response such as counteraction, to reduce the effect of the breakdown, or disregard of the breakdown (which again corresponds to the observed effect of the event on the technological process). To take another example, a behavioural record of what a physicist does in his laboratory can be used to analyse the behavioural structure and role performance of the scientist; or the same data, looked at in terms of the relationships between the operations carried out, can be used to provide information on the nature of the physical world. An event can be analysed with respect to its role within the technological frame of reference and with respect to its role within the social-behavioural frame of reference; and, by implication, every technical event has a corresponding behavioural representation, and every behavioural event has a corresponding representation within the technological frame of reference. An activity with respect to a production system can be defined as an operation carried out by a person, by means of a tool or a machine on material, which results in some observable material change. If activities are looked at from the point of view of who carries them out, and with whom, we arrive at a representation of the roles, the relationships between roles, and the types of relationship between persons which define the social system. If operations, independent of who carries them out, are looked at in terms of their sequential and mutual dependence relationships, we arrive at a representation of the production task structure. If activities are analysed with 114

PRODUCTION TASKS AND WORK ORGANIZATION

respect to the sequence and interrelation of material changes, we arrive at a representation of the material production process. A record of events within a production system may thus be utilized to determine the characteristics of both the social-behavioural and the technological structure, depending on how the data are analysed. Measurement does not imply that the raw data need to be numerical. Most of the quantitative measurement techniques discussed are based on qualitative data. It will generally be sufficient if records are made in terms of all-or-none characteristics, i.e. whether or not two activities are sequentially dependent, whether or not two persons work together, or whether their tasks are similar or different. The main requirements of an objective approach are the complete coverage of certain agreed dimensions of the work process, and the use of comparable methods of data collection. Section 1, 'Outline for a Dimensional Analysis of Production Systems', sets out the levels and components of socio-technical systems. Operations defined with respect to individuals, machines, and materials are used as a basis for the construction of higher-level units going up to, respectively, the social system, the plant layout, and the manufacturing system. The interrelations of the levels are indicated, together with the type of recording technique generally employed at each level. Section 2, 'Work Relationship Structure', attempts to provide a methodological approach to the study of work relations, based on the conceptual approach to the study of work organizations developed by Trist & Bamforth (1951). In Section 2(a) the task structure is looked at in terms of the dependence relationships and potential variance channels between tasks that form part of the production process. In Section 2(b) a method of measuring the work relationship structure is developed, which takes into account the following variables: (i) the relationship between persons in terms of the dependence relationship between the tasks to which they are allocated; (ii) the activity relationship between the persons who carry out the various tasks; (iii) the extent to which the content of the tasks in which they are engaged is similar or different; (iv) the extent to which independent or mutual goals operate. Section 2(c) considers a method of measuring the activity relationship structure between persons where the pattern of activity relationships is not fixed but changes in response to changing task conditions. 115

TASKS AND ORGANIZATIONAL STRUCTURE

Section 3, "The Variance-control Process', discusses a methodological approach based on the principle that a good way of finding out how an organization works is to study how it reacts to conditions of stress. A method is presented of recording the control sequence structure both with respect to the successive phases of the control process and with respect to the successive organizational units involved. The possibility is examined of measuring the operative aims of a person or group in terms of his or its observed reaction to variance in the work process. The final section, 'Work-domain Structure', looks at an organization in terms of its differentiation and segmentation into work regions, and at the type of information that can be obtained from study of the flow of personnel between these regions. Section 4(a) indicates that measures of the role rigidity and the degree of self-maintenance of a group can be derived from its level of participation in the differentiated work regions of its work domain. The degree of role differentiation within a group can be measured in terms of the pattern of task allocation. Section 4(b) examines the effects of the differentiation of the work system with regard to the development of role segregation and the emergence of a status hierarchy. Section 4(c) shows that the status hierarchy may be determined in terms of the relocation flow structure. The skill range of workers can be used as a measure of the potential flexibility of a group, and a method is set out for measuring the amount of change in group composition. 1. OUTLINE FOR A DIMENSIONAL ANALYSIS OF PRODUCTION SYSTEMS The smallest unit of a socio-technical system is taken to be an operation which produces an observable modification of some aspect of the system. In order to determine the characteristics of a sociotechnical system in general, and the characteristics of a production unit of the size of a factory or plant in particular, the following basic three elements are required: et an operation performed by an individual {J an operation performed by a machine y an operation performed on material.

The choice of component elements will be regarded as adequate to 116

PRODUCTION TASKS AND WORK ORGANIZATION

the extent that they permit the derivation of all relevant aspects of a production system with the exception, at this stage, of economic factors. In order to form a succession of higher-level units, the following operations on the basic elements and units formed by these elements will be employed: (i) the formation of sets of elements (ii) the formation of temporal sequences of elements (iii) the formation of interdependence relationships between

elements.

Element Sequences Q(ac) Sequences of operations performed by individuals: these are represented by means of operation charts. Q(ft) Sequence of machine operations: machines may be capable of one or more than one operation; in the latter case their sequence is either fixed or modifiable; in the case of a fixed sequence we may speak of intra-machine production flow. Q(r) Sequences of operations on material: these are represented by means of process charts. 1

Sequence Interaction Q(acp) Joint representation ofthe sequence of operations performed by the machine and those performed by the operator set out along a time dimension, by means of man-machine charts.

First-order Units (a)

A completed sequence of operations by a person defines an activity. Examples of activities are: assembling an apparatus, filling in forms, and sweeping the floor, together with a null activity representing idle time.

1 The use of operation, process, and machine charts is discussed in Barnes (1944, 1949), Maynard & Stegemerten (1939), and Rodgers (1950).

117

TASKS AND ORGANIZATIONAL STRUCTURE

(b)

The set of operations performed by a machine with a given set-up. Example: trims metal sheets of given dimension and cuts two slots, plus null operation when idle.

(c)

A completed sequence of operations on a set of materials leading to an intermediate product defines a production or service process. Examples of production processes are: welding, spraying, and assembling. Service processes are generally classified under: inspection, transport, store, and temporary delay. Clerical activities are generally classified under: preparing documents, transferring information from one document to another, file, forward, and temporary delay (Barish, 1951).

First-order Interaction (ab) Allocation of men to machines: the aim may be to minimize idle machine and man time, and to satisfy the requirements of priority schedules and process sequences (Moore, 1951; O'Donnell, 1952).

(be) Allocation ofprocesses to machines determines a given machine set-up, and is set out in the machine schedule. (ac)

Allocation of men to processes may be determined either by initial job selection or by subsequent training.

Second-order Units (A)

Sets of activities which are grouped formally or by expectation define jobs. These may be classified under: production, service (inspection, stock, etc.), clerical, supervisory, and managerial. The system of activities carried out by the person within the larger production system in which he is located defines the individual behaviourally with respect to the production system.

(B)

The total set of operations that a machine can perform with resetting and retooling defines the type of machine. A general classification of machines may be made into: forming, material removal, and joining, each of which permits sub-classification (Connelly, 1943). (This should be expanded to include machines used for transport, inspection, and clerical work.)

118

pRODUCTION TASKS AND WORK ORGANIZATION

(C) The sequence of processes performed on a set of raw materials leading to a given end-product defines a manufacturing process. These are represented by means ofjlow process charts (Maynard & Stegemerten, 1939).

FIGURE 10-1

DIMENSIONAL ANALYSIS OF PRODUCTION SYSTEMS USING THE METHOD OF ELEMENT INTEGRATION

ELEMENTS

ELEMENT SEQUENCES

Operations by

individuals

t.

Operations by

Operations on

machtnes

materials

t.

•

0 perat1on

0 perat1on

Sequences of machine

sequences

~}?,l;

sequences shown by

process charts

iernltions

l

"\an-machtne charts

FIRST-ORDER UNITS

Machine A specific individuals functioning production \ ~ ~~~~h given ·or service \ ..x:t-up Xplcess Allocation of men to

Allocation of men to

machines

SECOND-ORDER UNITS

SYSTEM STRUCTURE

l

Activities of

processes

Allocation of processes

to machines

1

Jobstarried out by persons

Maclne type

!

Mantactunng process for a

specific product shown by flow

process chart

Social Interaction and organization

j ""'-

Plant layout

Mantfacturtng system

_~ / Routtng and materialshandling system

/

The factory as a socio-technical system

119

TASKS AND ORGANIZATIONAL STRUCTURE

System Structure I

S(A)

The total set of activities of factory members, together witb their interaction and interdependence relationships, definel the social organization of the factory. S(B) The total set of available machines, together with their relative location, represents the plant layout of the factory. S(C) The total set of ongoing manufacturing processes, and their interdependence, defines the manufacturing system of thci factory. 1 System Interaction S(BC)

The manufacturing process, together with the plant layou~ determines the possible routing and materials-handling system. ;

S(ABC)

Generally, the manufacturing process determines the planC layout, and both together determine the materials-handlina system. The plant layout and the handling system, togeth~ with sub-unit interactions, in turn become important determinants of the social organization. We arrive at thill level at a possible representation of the factory as a whole~ as a socio-technical system (see Figure 10-1). ·

While this approach has some theoretical interest, and shows the possibility of interrelating a number of traditional measurement techniques, it does not appear to be entirely satisfactory as an ana.;. lytical scheme for socio-technical systems since, in its present form, it does not provide an adequate representation of the social-psychological characteristics of work organizations.

2. WORK RELATIONSHIP STRUCTURE (a) TASK DEPENDENCE STRUCTURE The types of relationship between tasks with which we are concerned at this stage are relationships of interdependence. These are of two kinds: 120

PRODUCTION TASKS AND WORK ORGANIZATION

(i) sequential dependence relationships (ii) variance-transmission relationships.

Work regions are said to be sequentia/ly dependent if Task A has to be completed before Task B can be begun. A variance-transmission

channel is said to exist if a breakdown or faulty work on Task A interferes with the carrying out of Task B. A sequential dependence link may be regarded as a potential variance-transmission channel at least where a breakdown within the antecedent work region is concerned. A variance link may exist also, however, in the absence of sequential dependence. For instance, two activities may depend on the same limited supply of material, in which case progress on one may interfere with progress on the other. We may distinguish four general types of sequential dependence pattern: divergent, chain, convergent, and cyclic (see Figure 10-2). FIGURE 10-2

TYPES OF TASK DEPENDENCE PATTERN

a• /·~ ----...y

~-~ .........Y-~

v·'.a... 6•--..•a

divergent

chain

convergent

""·5

~·/

a·,_;-·~ V cyclic

In the divergent pattern, a: is a key task since it has to be completed before any of the following tasks can be started. Its direct centrality can be measured by the number of tasks that are directly dependent on it. Its centrality within the system as a whole will be given by the total number of tasks that are sequentially dependent on it. In the chain pattern the centrality of a task decreases as one moves from the head to the tail end of the chain. Similarly, in the convergent pattern, the degree of direct dependence ofTask a: is determined by the number of directly antecedent tasks. An example of a cyclic pattern is the relationship between shifts in longwall mining. Each shift has the same centrality so that the system as a whole is equally vulnerable to breakdown in any one shift. A representation of sequential dependence structure and variance channels will be referred to as a task dependence structure. Figure 10-3 shows an approximate representation of the task dependence structure in conventional longwall mining. For a complete repre-

121

TASKS AND ORGANIZATIONAL STRUCTURE

sentation it would be necessary to check, for each pair of work regions, the existence of sequential dependence and variancetransmission relationships. 1 From the task dependence structure it is possible to locate work regions that are most likely to disrupt the system, and also work regions, such as filling, that are most vulnerable to disruption. Then it is possible to go one step further and to derive quantitative expressions of disruption potential and vulnerability (both of individual work regions and of the system as a whole) and of the flexibility of the system in terms of available alternative routes in case of breakdowns. FIGURE 10--3

TASK DEPENDENCE STRUCTURE

R.

I

I

~·~'Pi

l 1 ·~ I I

Bore

y

• ---?')t;\--:

Cut

:

Guml

II

•

) •

Break belt

1Build belt

SHIFT 1

I

~

---£>

SHIFT 2

~

•Fill

I

II I

SHIFT 3

Sequential dependence Variance channel

Disruption potential is defined as the extent to which a reduction in the level of functioning within a work region will affect other· work regions. This would tend to be high in a divergent pattern. Vulnerability is defined as the extent to which the level of functioning within a work region is affected by changes in the level of functioning· in other work regions. This would tend to be high in a convergent pattern. The additional data that would be required are the frequency 1 An analysis of the task dependence structure found in a composite cuttins Iongwall team is reported in Trist et al. (1963).

122

PRODUCTION TASKS AND WORK ORGANIZATION

of breakdowns within every region that affect other regions during a standard period of time, and a record of which other regions are affected. The same type of analysis may be made of the relationships of activities within each work region. A matrix representation of task relationships is useful at the recording stage, since it provides an immediate check on whether all pairs of activities have been taken into account. The matrix can at a later stage be used for quantitative analysis. In the following matrix of a task dependence structure the symbols used are: d = sequential dependence; s = variance channel: To Tasks

~

From

a;

fJ

y

lff

The matrix shows that Tasks {J and y are sequentially dependent on Task a:, and also that there is a variance channel from a: to {J.

TABLE 10-1

VARIANCE RECORD

Average amount of Variance Frequency disruption level

Direction of spread

Type of variance

a;-7{J

St

/1

tl

/tX lt

Sz

/z

tz

fzX tz

A more detailed analysis of variances will require a classification of types of variance, a record of the frequency of each type during a specified period, and some measure of the amount of disruption caused. A possible measure of disruption is that of time lost that would otherwise be spent on the assigned task. A measure of the overall effect of a variance can be obtained by multiplying frequency and average disruption caused, giving a measure of variance level for each channel. Table 10-1 shows how the data can be recorded. 123

TASKS AND ORGANIZATIONAL STRUCTURE

On the basis of these data it is possible to determine the extent which interference at each point is absorbed or transmitted. Data o this type can be used to study the process of variance control and·. the potential for self-maintenance. 1 (b)

MEASUREMENT OF THE WORK RELATIONSHIP STRUCTURBj

Given a specific task dependence structure which defines the inter- j dependence relations of activities, the choice that is made in the' allocation of persons to activities will to an important extent • determine the resulting social structure. · If, for instance, there are two activities that are sequentially re-; lated and between which a variance transmission may occur along< the direction of work flow, then, if one person is allocated to each! activity, a structure is set up in which B will be dependent on A and: vulnerable to faulty work on his part (Figure J0-4a). An alternative: relationship pattern is separate allocation to each task under con- • ditions where A helps, or avoids passing on variance to, B. FIGURE 10-4a SEQUENTIAL DEPENDENCE, SEPARATE ALLOCATION

A

8

____Ta_s_f_a_·___,) [,__ta_s_·k~P~~_:--_,.> Another pattern might be joint allocation of both activities to one or more persons. The cyclic pattern is of particular interest. Suppose that two independent teams are allocated to work regions that form part of a cycle such that the sequential dependence relationship includei variance transmission (Figure J0-4b). The resulting system may develop positive feedback, making it intrinsically unstable: for if a disruption occurs in work region ex, this is transmitted to {3, then 1 The variance matrix technique has been developed further by Marek, Lango & Engelstad (1964) and Engelstad (1969, 1970).

124

PRODUCTION TASKS AND WORK ORGANIZATION

passed back to

oc where disruption is now increased. Disruptions that occur will in this case be steadily built up until they get out of band and the cycle breaks down. The inter-shift relations in the conventional longwall method of coalmining studied by Trist &

FIGURE 10-4b

CYCLIC DEPENDENCE, SEPARATE ALLOCATION

Bamforth (1951) provide a relevant example. A contrasting example is provided by composite teams which under similar conditions have a mutual goal with respect to the maintenance of the work cycle. In this case interference is reduced or absorbed by performance control carried out by each team within the cycle. The negative feedback that operates in this instance provides the condition for system stability. 1 Once the task constraints are given and the task allocation is made, then how the organization operates will depend on the types of aim pursued by individual workers and work teams. If tasks are independent, then goal independence will cause no local interference. If tasks are interdependent and a mutual goal exists, then there is a co~dition where mutual help will be given. If tasks are interdependent and goals are independent, difficulties are certain to result, such as in the case of two cart-horses that operate under the task constraint of interdependence but are independent with respect to the direction in which they want to go. The measurement of the work relationship pattern requires the joint specification of the following variables for every pair of persons in the study of a group, and for every pair of groups in the study of an institutional structure: 1

See, particularly, the literature on autonomous group functioning, Herbst

(1962), Trist et al. (1963).

125

TASKS AND ORGANIZATIONAL STRUCTURE

(1) activity relationship-activities are carried out together or separately (2) role differentiation -tasks carried out are identical or different (3) task dependence -tasks may be dependent, interdependent, or independent (4) goal dependence -goals are shared, independent, or unreciprocated supporting. Let us start with a specification of the activity relationship. For every pair of persons there are the following possibilities: a they work together ii they work separately.

The results are set out in the form of a matrix. In the case below, the group activity structure shows that A and C work together and B works on his own: Persons C B Persons A B

rei d

A method of determining the activity relationship structure where the pattern of subgroupings changes in the course of a work cycle in response to changing tasks demands is discussed in Section 2(c). With respect to role differentiation, the following possibilities exist:

i

tasks are identical tasks are different i* tasks are different and are exchanged or rotated. i

The results are again set out in the form of a matrix. In the case below, the role differentiation structure shows that A and C exchange tasks and B works on a task that is different from theirs: Persons C B Persons A B

r,·-· i

126

PRODUCTION TASKS AND WORK ORGANIZATION

A finer distinction that can be made is given by a measure of the similarity of the sets of tasks between any pair of persons. A possible measure of similarity in the composition of two element domains is given in Section 4(c). A possible measure of the degree of similarity in the tasks carried out by the members of the group as a whole is given in Section 4(a). Combining the two variables discussed so far gives the following possible patterns: ai ai

ai tii ai*

ai*

work together on the same task work together on different tasks work on their own on the same task work on their own on different tasks work together and exchange tasks work on their own and exchange tasks.

The next variable is that of task dependence, where the following possibilities exist: d* dAB

d

tasks are interdependent or mutually dependent B is sequentially or one-way dependent on A tasks are independent.

The following matrix of a task dependence structure shows the condition of cyclic dependence from A to B to C and back to A: Persons

C

B

A

dcA d.u

B

dsc

Persons A relationship of interdependence may be set up either owing to the nature of the task on which two or more persons work together, or owing to a common limited supply of services, materials, or tools in the case of persons working on their own. The last variable is that of goal dependence, where the possibilities are as follows: g

g K..tB

there is a mutual goal goals are independent A supports B or avoids behaviour that would act as a variance for B. 127

TASKS AND ORGANIZATIONAL STRUCTURE

The presence or absence of mutual goals is determined empirically in terms of the presence or absence of mutually supportive or helping behaviour, or behaviour directed towards the reduction of variance transmission. When the task- and goal-dependence variables are combined, the following are among the types of pattern obtained:

d*g

tasks are interdependent, there is a common goal tasks are interdependent, each person has his own goal dg tasks are independent, there is a common goal tasks are independent and goals are independent dg d.tBKAB B is task dependent on A., and A. controls variance transmission B is task dependent on A., they pursue separate goals.

d*g

When all four variables are combined, the work relationship structure is obtained. The following matrix would be obtained in a coal-hewing team where each man worked on his own stint. The men are not dependent on one another with respect to the task carried out. A. and B help one another out if either runs into difficulties or gets behind. C provides help for D which is not reciprocated: Persons A

Persons B C

D C B iiidg iiidg iiidg iiidg iiidg iiidgcD

The same method can be applied to the measurement of both intra-team and inter-team structure. The matrix below would be obtained in the case of a long-wall cycle with three shifts, where each team had a different task and where Teams A. and B controlled incoming lag in order to reduce, where possible, the lag transmitted to the subsequent team: Teams

C A

iiidcAi

B

iiidsc8sc

Teams 128

B iiid,..gAB

PRODUCTION TASKS AND WORK ORGANIZATION

It appears very likely that the type of structure that exists within a team will affect the types of link that the team can establish with other teams. This is an important organizational problem, which Bott (1957) began to investigate in the fields of family structure and kinship network. A method of the type proposed, where both intraand inter-team structure are measured in terms of the same set of variables, would make it possible to trace out the relationship both empirically and conceptually and may provide a guide for organizational planning. There are two ways in which the relationship matrix can be analysed. The first is by constructing indices for each variable or pair of variables in turn, which can be used for testing quantitative formulations with respect to the relationships between indices of group functioning. Possible measures will be: the degree of joint or separate activities; the degree of role segregation; the degree of task interdependence; and the frequency of dependence links associated with goal independence as a measure of conflict potential. Some of these are discussed in greater detail in other sections of this chapter. The second method of analysis is concerned with determining the overall work relationship structure. There is no difficulty in the case where patterns are homogeneous and fixed, such as in the examples used. However, if there is a large amount of variation in the form of pattern differentiation within or between teams it will be necessary to classify structures in terms of the predominant type or types of pattern observed. The type of structure found can then be used as a basis for predicting: the extent to which a team or organization can cope with stress; attitudes to the work situation; and the direction of hostility expression. It would also be possible to apply Heider's condition of balance and imbalance in system structure (Heider, 1946). If, for instance, in a cyclic system some teams minimize variance transmitted and others do not, a state of imbalance will exist. The Heider principle can in this case be generalized to predict either that pressures will be exerted to induce other teams to employ variance control so that all share a mutual goal, or that all teams will become goal independent, resulting in disruption of the cyclic dependence network. With the present set of variables it is possible to construct fifty-four basic types of work relationship structure. These may be further

129

TASKS AND ORGANIZATIONAL STRUCTURE

differentiated in terms of structural type. For instance, sequential dependence structures may be differentiated into sequential chain and sequential cyclic structures. Table 10-2 lists only a few of the TABLE 10--2

Structure

TYPES OF WORK RELATIONSHIP STRUCTURE

Nomenclature

Example

iiidg

Homofunctional isolate aggregate

iiidg

Homofunctional connected aggregate

iiid*i

Homofunctional isolate interdependence

aidi

Heterofunctional isolate dependence

didg

Heterofunctional connected dependence

iii*d"'g

Isolate composite

ai*d"'g

Joint composite

130

Men work independently on the same task; tasks are independent. Same as above; tasks are independent but men give mutual help to one another. Men work on the same task each towards his own goal, but they are jointly dependent (for instance, on a common supply of tools and materials). An example on a larger scale is a number of factories producing similar goods and competing for supplies and markets. Men work on their own on different, sequentially dependent, tasks without taking each other's needs into account. Same as above, but each man minimizes and controls lag or variance transmitted. Men work on their own towards a mutual goal (i.e. a common pay note); the tasks are interdependent and are rotated. Same as above, but men work together; each man participates in various tasks.

PRODUCTION TASKS AND WORK ORGANIZATION

possible types, which have been chosen as examples. The nomenclature is at this stage tentative. The examples given are with reference to team structure.

(c) MEASUREMENT OF THE GROUP ACTIVITY STRUCTURE

In response to the temporal sequence of tasks and the types of breakdown that may occur in the course of a work period, the pattern of relationships between persons will show an ongoing change. The degree of change in the activity relationship between persons may vary all the way from the case where men are on fixed positions on a flow line day after day to the case where there are composite teams and the pattern of relationships changes at a relatively fast rate. The fact that the pattern of relationships shows a continual change does not necessarily mean that the group lacks a structure. The following is concerned with a method of analysing the interaction process in order to determine the structure of a group in terms of the activity relationship between individuals. The interaction process may be analysed at two levels, either with respect to the changing size of subgroups, or by specifying who belongs to each subgroup, which makes it possible to analyse the changing pattern of subgroup membership. An activity relationship between persons during a given moment of time has the following properties: (1) It is reflexive, since a person may engage in an activity with, or by, himself. (2) It is symmetric, since if A is engaged in an activity with B, then B is engaged in that activity with A. (3) It is transitive, since if A and B are engaged in a joint activity, and B and C are engaged in a joint activity during the same moment of time, then A and Care engaged in a joint activity. Relationships between persons that do not have these properties will not be regarded as activity relationships. A relationship that is reflexive, symmetric, and transitive is said to be an equivalence relation. A fundamental property of an equivalence relation is that it divides a set of elements into mutually exclusive or non-overlapping subsets. We can, therefore, represent 131

TASKS AND ORGANIZATIONAL STRUCTURE

the momentary state of a group in terms of the activity relationships between members by means of a partition written in the form (4, 3, 1, I) showing that, in a group of nine persons during a given period of observation, there is a subgroup of four who work together, a subgroup of three who work together, and two persons who work on their own. Table 10-3 shows the number of ways in which a group consisting of four individuals can be subdivided, ranging from the case where at a given moment no subdivisions exist to the case where there are as many subgroups as there are individuals in the group. These two extreme cases are represented by the partitions (4) and (1,1,1,1), respectively; the latter partition may more conveniently be written (1 4 ). There are, in all, five different partitions, or five different ways in which a group of four can be subdivided. For a group of N persons, the number of possible subdivisions can be determined from tables of partitions (Whitworth, 1901). In order to determine the most likely way in which a group will subdivide, we need to know the number of ways in which each partition may be realized. Let us consider as an example the partition (3,1), where three persons are in one subgroup and one person is by himself, and let A,B,C,D represent the four persons who are members of the group. The following are the possible ways in which we can allocate these individuals to subgroups of three and one respectively: (ABC)(D), (BCD)(A), (CDA)(B), and (DAB)(C). We shall refer to each possible allocation as an interaction pattern, since it shows with whom each individual is acting at any given moment. All the interaction patterns corresponding to the partitions of a group of four persons are set out in Table 10-3. The characteristics of a group in terms of activity relationships will be determined by the observed frequency, or the probability of occurrence, of each of the possible activity patterns over a period of time. A group in which all activity patterns are equally likely to occur represents a perfectly fluid group. It will be shown later that this definition implies that, if a group is in a state of fluidity, then the probability of engaging in a joint activity is the same for all pairs of persons. Consequently, no permanent subgroups or cliques will exist in a fluid group. 132

PRODUCTION TASKS AND WORK ORGANIZATION

The probability of finding the group in any given state of subdivision will be proportional to the sum of probabilities associated with the interaction patterns that correspond to that subdivision. Table 10-3 shows the probability of occurrence of each of the possible partitions of the four-person group in a state of fluidity. The probability values are obtained by dividing the number of interaction patterns corresponding to each partition by the total number of interaction patterns of the group. The most likely state of the fluid four-person group is found to be a subdivision of the group into a subgroup of two, and two single individuals. The least likely states are that of no subdivision and that of complete dissociation. In order to determine the number of interaction patterns and probabilities of subdivision for groups of larger size, general formulae are needed. These will be found in the appendix to this section (p. 137 below). TABLE 10--3

PARTITIONS AND CORRESPONDING INTERACTION PATTERNS FOR A GROUP OF FOUR PERSONS

Partition Interaction patterns

Probability of the partition in the fluid group

(4)

(3,1)

(2,2)

(2,1,1)

(ABCD) (ABC)(D) (AB)(CD) (AB)(C)(D) (BCD)(A) (AC)(BD) (AC)(B)(D) (CDA)(B) (AD)(BC) (AD)(B)(C) (BC)(A)(D) (DAB)(C) (BD)(A)(C) (CD)(A)(B)

1 15

-

3

4 -15

-15

6

-15

(1,1,1,1) (A)(B)(C)(D)

1 15

The possible subdivisions or partitions of a group differ from one another in the extent to which at any given moment the members of the group are dissociated from one another. If we have a measure that orders partitions in terms of increasing dissociation, and has a unique value for each partition, then it is possible to investigate the dynamics of changes in the momentary state of the group in terms of changes in the degree of its dissociation. 133

TASKS AND ORGANIZATIONAL STRUCTURE

Let a group of N individuals at a given moment be divided into subgroups n 1 , n2 , n 3 , ••• , so that

If / 1 represents the fraction of the group in subgroup n 1 , that is,

then

A possible measure of the degree of dissociation (D), that will have the desired properties, is given by the expression D = -

Lfi log/1• i

If there are no subdivisions at all, then n 1 = N, and in this case / 1 = 1. But, when / 1 = 1, the function / 1 log / 1 = 0, and therefore the degree of dissociation takes its minimal value, zero. In the case of 1 maximum dissociation, there will be N fractions of value N' so that the degree of dissociation will be 1 1 D = -N Nlog N =log N.

The maximum degree of dissociation consequently increases as a logarithmic function of the number of individuals in the group. Since a logarithmic function increases at a decreasing rate, the addition of a single individual will have more effect on the degree of dissociation in a small group than in a larger group. Matrix Representation of the Group Activity Structure

A necessary condition for a group to be structured in terms of activity relationships is the existence of a quasi-stationary distribution of 134

PRODUCTION TASKS AND WORK ORGANIZATION

interaction patterns over a period of time. If, under the influence of forces internal or external to the system, a change in the distribution of activity patterns occurs, then the group is said to restructure. A convenient way of analysing structures is by means of matrix representation. An interaction pattern can be represented by a matrix if, for pairs ofindividuals who have a joint activity relationship, we make an entry of 1 in the appropriate cell of the matrix and, for pairs of individuals who do not have a joint activity relationship, we make an entry of 0. The matrix for the pattern (ABC)(D) is shown in Table J0-4a. An entry of 1 is made in cells AB, BC, and AC, and also in cell DD on the diagonal for individual D, who is engaged in an activity by himself. Since the complete matrix is symmetrical, a triangular matrix may be employed. The set of sequential interaction patterns is then summed into a single matrix of the type presented in Table J0-4b. TABLE 10-4a

TABLE 10-4b

MATRIX OF THB

MATRIX OF THB FLUID FOUR•

INTERACTION PATTERN (ABC}(D)

A B

D 0 0

D

0 1

c

c 1 1 0

PERSON GROUP

B A 1 0

D

c

B A

5 5

5 5

c

5 5 5

D

5

A B

0

5

5

Suppose that each of the fifteen possible activity patterns of the four-person group appears just once during a given period of observation. If each of these patterns, listed in Table 10-3, is entered into a matrix, then we obtain the matrix of a fluid group shown in Table J0-4b. It will be seen that in the fluid group each person has the same frequency of joint activities with every other member of the group. 1 The characteristics of the fluid and unstructured group can be used as a base-line for the determination of the structure of a group 1 The probability of isolation in terms of acceptance-rejection relationships in the group has been investigated by Katz (1952).

135 K

TASKS AND ORGANIZATIONAL STRUCTURE

from any set or sample of activity patterns. The activity patterns are first entered into a matrix. We then proceed by the following three steps: (1) For a person to form part of an activity structure, he must have an activity relationship with at least one other member of the group. We shall say that an activity relationship exists between two persons if the frequency of their joint activities is not less than it would be in the corresponding fluid group. We eliminate, therefore, as isolates, all persons whose frequency of joint activities with every other member of the group is less than the frequency of joint activities in the fluid group for the same number of interaction patterns. (2) The hypothesis is tested that the group is structured. This is done by setting up the null hypothesis that the frequencies of the various types of interaction. pattern for all individuals who are not isolates can be accounted for if we assume that all interaction patterns have an equal probability. (3) If the hypothesis of fluidity can be rejected, then a matrix is constructed in which a cell entry of 1 is made for each pair of persons whose frequency of joint activities is not less than that for persons in the fluid group. The resulting matrix represents the activity structure of the group.

Activity structures are conveniently represented by a type of topological diagram known as a linear graph. 1 A linear graph is constructed by representing each person as a point element and connecting, by a line, those pairs of points that represent persons who have a joint activity relationship. The appropriateness of representation by a linear graph lies in the fact that in this type of structure the relationship between elements is symmetric but intransitive. This is in fact the basic difference between the momentary interaction pattern and the activity structure. In the latter, the activity relationship 1 A similar type of diagram has been used by Bavelas (1948) for the representation of available communication channels between the members of a group. A purely statistical approach to the measurement of group structure may not be satisfactory, especially if this does not take into account the nature of the tasks and the dependence relationship between them. The subsequent development and empirical application of the measures of group process and structure are reported in Herbst (1962).

136

PRODUCTION TASKS AND WORK ORGANIZATION

is no longer transitive. It is, for instance, possible for a chain structure to exist between persons such that there is a relationship between A and B, and between B and C, but no relationship exists between A and C. This is because the activity structure defines the characteristics of a group over a period of time, so that, in this instance, the joint activities B engages in with A may be at different times, and with respect to different tasks, from those he engages in with C. APPENDIX TO SECTION

2

The total number of activity patterns for a group of N persons (AN) will be given by the number of ways in which N different persons may be allocated to 1, 2, 3, ... , N different subgroups, which can be shown (see Whitworth, 1901) to be AN = Nl (coefficient of x! in the expansion of e<e><- 1>)

where 5

e<e><-1)

5

13

= 1+x+x2+6x3+8x4+14xs+ .••

For a group of four persons we have

5

AN= 4!8 = 15,

which is the number of interaction patterns we found by listing the number of possible allocations of individuals shown in Table 10-3. The number of interaction patterns increases very rapidly as the size of the group increases. A general formula for the number of interaction patterns corresponding to any given partition can be worked out in the following way. The number of ways in which N persons can be allocated to subgroups of size n~> n2 , n3 , ••• will be Nl Thus for the partition (3,1) we have

4! 4. . 3 ! 1 ! = mteract10n patterns. 137

TASKS AND ORGANIZATIONAL STRUCTURE

In the case where there are subgroups of equal size, however, such as in the partition (2,2), we will not be able to distinguish between the patterns (AB)(CD) and (CD)(AB), which are obtained by exchanging all the members of one subgroup with the members of the other subgroup. We shall, therefore, need to divide further by the factorial of the number of subgroups of equal size. The number of interaction patterns that corresponds to a partitionn 1a,n/,n 3 ", ••• of a group of N persons, where oc, {3, ~·is the number of subgroups of size n 1 , n2 , n3 , will thus be given by the expression N! [(n 1 !t(n 2 !)P(n 3 !)" ...][oc!f3!y! ...

r

In the case of a fluid group, the probability of each of the possible partitions of the group is given by the number of interaction patterns that corresponds to that partition. Since the above expression takes a minimum value of unity either if there are no subdividions (N 1) or if the group is completely subdivided into isolate individuals (1 N), these states will be the least probable. Their probability decreases rapidly as the size of the group increases, since this leads to a rapid increase in the number of alternative activity patterns. 3. THE VARIANCE-CONTROL PROCESS A work-system variance may be defined as a deviation from a steady state of the production and social process or a change in the rate of progress towards an organizational goal if the deviation is in the direction of reduced productive efficiency. Whether a system variance exists is thus judged with reference to an organizational aim which may or may not be identical with the aims of individual members of the work organization. The efficiency of a variance-control cycle will depend on: (i) the speed with which variances are detected and if necessary anticipated (ii) the speed with which appropriate variance-control processes are set in motion and completed. Variance detection is clearly a necessary condition for control. Formally, variance detection may be the responsibility of a foreman working together with a specialized inspection staff, but these people 138

PRODUCTION TASKS AND WORK ORGANIZATION

will be able to cover only part of possible system variance and to do so intermittently. In practice, the efficiency of variance control will depend on the reaction of every member of the organization to events that lie within his perception radius. This again will be a function of his cognitive structure of the work situation and, more specifically, of the extent to which each person has internalized organizational values. Members' reactions to variances that appear in their work situation thus serve a double function of evaluating the efficiency of the control process and providing a method of obtaining objective measures of the personal relevance of, and attitudes to, the task. (It may be noted that the latter is also the 'common-sense' method used in practice. The identification of the individual with organizational aims is assessed not, except provisionally, by his expressed attitudes and interests, but by his actual performance during an organizational emergency.) The reaction of individuals to the incidence of variance as a function of personal relevance and attitude is shown in Table 10-5. It will TABLE 10-5

REACTIONS TO VARIANCE

Personal relevance of the task

Attitude to the task

Variance perception

Behaviour reaction

Experiential reaction

high high low

positive negative neutral

immediate immediate delayed

immediate delayed delayed

concern relief indifference

be noted that level of personal involvement and attitude to task requirements can be derived from the reaction to variance. Quite generally, it appears that, provided the individual is capable of variance discrimination, perception latency will be a function of the personal relevance of the existing state or goal, and reaction speed will be a function of the corresponding attitude. Control Sequence Structure The control cycle has a sequential phase structure. We may have an occasional short-circuit if parts of the cycle become routinized, or 139

TASKS AND ORGANIZATIONAL STRUCTURE

we may have back-tracking if the sequence is stopped at any point· or re-examination is required. The sequence will, however, remain constant since the antecedent phase is always a necessary condition for the phase that follows. The phase sequence involved in the control process is set out in Table 10-6. The final step is that of inspection, to ascertain whether the variance has in fact been removed. If not, the process retracks to find out where a mistake was made. Part of the frustration experienced, if the control sequence does not lead to the desired goal, is that it now becomes necessary to deal with an additional variance located in the control cycle, which has to be removed before the original variance can be dealt with. TABLE 10--6

Phase

PHASE SEQUENCE OF THE CONTROL PROCESS

Description

Perceiving that something is wrong Finding out what is wrong Evolving possible methods for getting rid of the trouble Deciding on method to be used and the allocation of men and resources 5. Performance Implementing the decision 6. Inspection Checking whether the variance has been removed

1. 2. 3. 4.

Detection Diagnosis Planning Decision

Each of the various phases of the control process may occur in a different part of the organization, in which case the findings obtained at each point have to be communicated at least to that part of the organization in which the following phase is to proceed. The control phases may be sub-differentiated into smaller units. Thus different aspects of the decision process may occur in different parts of the organization, one person deciding on the method to be used while another decides on the personnel to be allocated. The level of differentiation of control phases will to some extent depend on the level of differentiation employed in describing the organization. For our present purpose it would appear sufficient to distinguish whether the process occurs at the level of an individual worker, or the work team, or the group leader, i.e. the foreman, or whether it occurs within the organizational environment of the group. If we want to

140

PRODUCTION TASKS AND WORK ORGANIZATION

study a larger system, then the external environment may be further differentiated into service and governing units, etc. The control sequence structure may be represented in the form shown in Figure 10-5. The control diagram shows that in this instance the variance was detected by one of the workers; he passed the information on to the other team members, who communicated it to the leader, who passed the information on to an external agency which was called in to discover the cause. The agency's diagnosis was then communicated to the leader, who considered possible ways of dealing with the problem, decided on a method, instructed one of the workers under him to carry it out, and subsequently checked that the variance had been removed. FIGURE 10-5 worker Detection Diagnosis Planning Decision Performance Inspection

team

CONTROL DIAGRAM

leader

external

X

* * * ~

*

(Type: leader

control,external consultation)

The advantage of a diagrammatic representation is not only that it demonstrates the total process with greater clarity but also that it enables the theoretically possible types of process to be determined and may suggest, in specific cases, better methods of control. Additional information desirable for the purpose of evaluation would be the amount of time taken at each point. For purposes of recording and analysis, the data may be presented in tabular form. The following code is used: P T L E

individual worker work team leader external

TL means that team and leader act together. For the purpose of summing observations, T,L represents that sometimes the team and

141

TASKS AND ORGANIZATIONAL STRUCTURE

sometimes the leader carries out the activity. Similarly, P,T means the person affected by the variance or other team members. Table 10-7 shows some of the possible types of control pattern that may be obtained by summing the control diagram horizontally. The type of control pattern obtained is likely to differ for different types of variance. It will be seen that the analysis of control patterns makes TABLE 10-7 Effect on

TYPES OF VARIANCE-CONTROL PATTERN

Detee- Diaglion nosis

Planning

De- Per/or- Inspeccision mance tion

p

p

p

p

p

p

p

p

p

p

P,T

P,T

P,T

P,T

P,T

P,T

p

p

L

L

L

p

L

P,T,L

T,L

T,L

T,L

P,T

T,L

E

E

E

E

P,E

E

P,T p

p

Type

Stimulus response Autonomous control Team control Leader control Group decision External control

it possible to represent phenomena ranging all the way from a stimulus response to a group decision procedure. The data required are either a study of a representative sample of variances or a continuous record and follow-through of all variances occurring during a limited period of time. The data can be analysed in a number of ways: (1) We can determine the type of control process activated by a particular type of process variance. (2) We can examine the manner in which specific individuals or groups are involved in the control process. (3) We can obtain an overall measure of the types of control process used in different types of work system.

142

PRODUCTION TASKS AND WORK ORGANIZATION

If the control diagram is summed vertically, the data make it possible to determine the types of role played by different units of the organization and to obtain measures of the relative degree of control exerted by workers, the team leader, and the external system. 4. WORK-DOMAIN STRUCTURE (a) BOUNDARIES OF WORK REGIONS AND TASK ALLOCATION The system of allocating activities to persons is as a rule quite complex and cannot be adequately represented by formal job descriptions. On the one hand, a worker may carry out a large (or small) number of tasks apart from those he is supposed to be doing. On the other hand, every task in the total set of tasks in which he may participate at one time or another does not necessarily have a job title attached to it. We begin this section by considering a possible way of representing the activity domains of work teams and individuals, and then (Section 4(b)) we look at the developmental stages in the evolution of work systems. This leads on to a discussion (Section 4(c)) of methods that may be used to study the effects of segmentation and flow within the system on its flexibility and status structure. The Work Domain of a Group The total set of work activities carried out by a team will be referred to as its work domain. Within the work domain of a group we may distinguish: (1) A core region, which defines the formal task of the group. In the case of a coal-filling team the core region may include filling and setting props.

(2) A maintenance and service region, which includes all activities required to enable work in the core region to proceed. These would include, in the case of a filling team, belt-mending, supplying props to the face, etc. It should be noted that some service activity may be carried out by persons outside the group, either on their own or in conjunction with the team, for instance in the case of a serious mechanical breakdown of equipment. The extent to which

143

TASKS AND ORGANIZATIONAL STRUCTURE

the service region is included in the work domain of the group can be used as a measure of the degree of self-maintenance of the group. (3) An extra task region, which includes activities that are formally allocated to other teams but may also be carried out by the work group if it is necessary for them to be done before work in the core region can proceed. Thus the filling team may do belt-building if this has not been completed on the previous shift. The extent to which a group does not engage in extra task activities may be used as a measure of its role rigidity. The situation in which unskilled workers in factories are frequently found can be described as follows. The team has an assigned job; the setting up of the job, and service and maintenance, are carried out by specialized staff. No member of the team would normally be asked to carry out tasks that are not part of his job. In this instance, the high role rigidity and low degree of self-maintenance are largely externally imposed. The relationship between work regions may be represented by means of a path-field employing a method for analysing participation data developed in a study of task-allocation structures in the family (Herbst, 1954a). The illustrative diagram in Figure 10-6 shows that FIGURE 10-6

core region

PATH-FIELD REPRESENTATION OF A WORK DOMAIN

-

maintenance and service region

extra task region

The arrows show the point of entry and the direction of movement when the level of participation is increased.

participation cannot normally fall below the level of the core region. If additional activities are engaged in, then these will be in the maintenance and service region; and if activities are carried out beyond that point, then these will be extra task activities. In the case of a role-segmented work team, the point at which the 144

PRODUCTION TASKS AND WORK ORGANIZATION

boundary of the work domain of the group is located is an important determinant of the efficiency of the work system as a whole. For example, if the filling team finds on arrival at the face that belt-building is not completed and it treats this activity as lying outside its work domain, then the team will not be in a position to start work and the shift cycle breaks down. Work organizations differ in the sharpness of their internal boundary and in their rigidity. A sharp boundary may be said to exist if, for each activity, it is known whether it belongs inside the work domain or not. Alternatively, there may be a boundary zone which includes activities with respect to which there is no general agreement as to whose job they are, or with respect to which disagreement exists between the work team and management. The set of activities that the group accepts as belonging inside its work domain may be referred to as its region of task acceptance. The incidence of activities that lie in the work domain but outside the region of acceptance may be used as a measure of the tension between the group and its environment. Task Allocation within Work Teams

Having briefly considered the work domain of the group we may next consider how the various group activities are distributed among team members. Allocation may be completely undifferentiated, where each member of the team carries out all the activities in the work domain of the group. At the other extreme, the work domain may be completely differentiated so that different subgroups carry out different tasks which do not overlap one another. The set of activities carried out by any individual member of the team will be referred to as his task domain. It appears desirable to have a measure of the degree of role differentiation. In the matrices below (Table 10-8) all the activities that are carried out by the team are shown along the horizontal axis, and the members of the team are shown along the vertical axis. An entry of 1 is made under each activity for every team member who carries out that activity either regularly or sometimes. No entry is made for a member who never does it. Matrix (a) shows a case of complete differentiation: every team member carries out a different activity, so that there is no overlap

145

TASKS AND ORGANIZATIONAL STRUCTURE

TABLE 10-8

ROLE DIFFERENTIATION MATRICES

, Persons A~ B 1 c 1 «

,

Team activities

Team activities

p

«

A PersonsB

c

(a)

p

1 1 1 1 1 1 1 1 1 (b)

between their task domains. In matrix (b), every team member carries out all activities. The degree of overlap between the task domains of any pair of team members is determined by the number of activities carried out by A that are also carried out by B. Table 10-9 presents a hypothetical example for three persons and three activities. It indicates that A and B have two activities in common, that B and C have two activities in common, and that A and C have three activities in common. TABLE 10-9

NUMBER OF ACTIVITIES COMMON TO EACH PAIR OF A THREE•PERSON TEAM

Persons

C

B

A

3

2

B

2

Persons

For N persons the number of possible overlaps between their task domains is N(N-l) 2 For n activities, the maximum number of activities that can be carried out by both A and B is n, so that the maximum sum of cell entries in Table 10-9 is nN(N-1)

2 146

PRODUCTION TASKS AND WORK ORGANIZATION

If~

denotes the sum of cell entries in Table 10-9, then degree of role differentiation

=1

2~

nN(N _ 1) .

For the example given: number of activities (n) number of persons (N) sum of cell entries (l:)

=

3

=3 =7

degree of role differentiation = 1- 3:;:2 = 1 - 0· 78 = 0·22. The measure has a value of 0 if everyone carries out all tasks, and a value of 1 if everyone carries out a different activity. It may be noted that complete differentiation is not possible if the number of persons exceeds the number of activities. If the pattern of allocation changes over time, as in the case of composite teams, then the data will need to be recorded on the basis of successive work-cycle periods. (b)

STAGES IN THE GROWTH OF WORK SYSTEMS

It has been shown so far that a work team may engage in a number of activities apart from its core activities. Correspondingly, the task domain of an individual member will contain, apart from some or all of the core activities, some or all of the service and extra task activities engaged in by the team. In addition, shifting of personnel may occur between teams to make up for absences or to reinforce certain teams. In practice, then, each face-worker, for example, may move and be shifted between a number of task domains. The total set of tasks in which a person participates over a period of time defines his job domain. Under ideal conditions the formal job boundaries coincide with the boundaries of the actual job domain. We look now at various stages in the process of segmentation of the work system into job domains, and then go on to consider the problem of flow within, and shifts between, job domains. The illustrative examples are taken from coalmining. The simplest type of work system is one that is completely undifferentiated (Figure 10-7). Each worker carries out all the activities

147

TASKS AND ORGANIZATIONAL STRUCTURE

involved in coal-getting by himself. Historically, this appears to be the earliest pattern. FIGURE 10-7

UNDIFFERENTIATED SYSTEM

Any work system can maintain itself only in so far as it provides for the training of new entrants. In the early type of work system, a new entrant was attached to a collier as a trainee. The advent of the trainee led to a differentiation and partial segmentation of the work domain into two distinct job regions (Figure 10-8). The collier transferred service and supporting activities, such as the supply of tubs and materials, to his trainee while retaining activities included in the core region of his work domain. FIGURE 10-8

TRANSITIONAL SEGMENTATION IN ORDER TO PROVIDE AN ENTRY ROUTE

Job domain of trainee

Job domain of collier

In the course of training, the job domain of the trainee gradually expanded until it overlapped all the activities included in the collier's job domain, after which training was completed and the new collier was allocated to a place of his own. The introduction of mechanization and longwall mining resulted in a disruption of the traditional work system. The immediate result 148

PRODUCTION TASKS AND WORK ORGANIZATION

was a differentiation of the total work domain into its task elements. Figure 10-9 shows a work domain that is differentiated but not yet segmented. Each worker may carry out any one of the tasks included in the work system. However, at one time he may carry out Task A while others carry out Tasks B, C, and D, and at other times he may move on to any of the other tasks. The unsegmented system requires that all face-workers are trained in all tasks. FIGURE 10-9

DIFFERENTIATED SYSTEM

In practice it is found that an increase in the degree of differentiation results in an increasing degree of segmentation. Little is at present known about the dynamics of the segmentation process. The following are some of the factors that appear to be involved: (1) A worker kept on one task for any length of time tends to develop special skills at this task and to some extent to lose skills that he may have acquired previously on other tasks. {2) At the same time, a process of perceptual inertia operates. The task that a worker carries out tends to become one of his attributes. Thus the person who is engaged on filling becomes identified as a ffiler and is no longer perceived as a person who can carry out other types of work. (3) One of the reasons for introducing a segmented system is that it reduces the amount of training that needs to be given to workers who will each be engaged in perhaps one task only. Segmentation may thus be imposed by differential training. (4) A developing tendency towards the segmentation of the work domain leads to the ordering of segmented parts with respect to relative status. This in turn increases the tendency towards a 149

TASKS AND ORGANIZATIONAL STRUCTURE

sharper definition of job boundaries, which further increases the level of segmentation. The tendency towards a sharper definition of boundaries will be expected to be strongest for those engaged in desirable and emerging high-status jobs. Figure 10-10 illustrates two systems, with medium and maximum segmentation respectively. In the former the work domain is split into three job domains, two of which contain two or more tasks. Within these two job domains, workers are able to shift between tasks. In the latter system the work domain is split into as many job domains as there are tasks, and thus the temporary shifting of workers between tasks is not permitted. FIGURE 10-10

SEGMENTED SYSTEMS

Medium segmentation

-..0-@ -@ Maximum segmentation

Some Effects of Segmentation of the Work Domain Segmented work systems have a number of properties not possessed by undifferentiated systems: (1) They automatically decrease the size of job domains of workers and impose a greater role rigidity on them.

(2) They make possible the emergence of more complex social structures. (3) More specifically, they make possible the development of the tendency to attribute status to individuals by virtue of their occupation of specific job domains.

ISO

PRODUCTION TASKS AND WORK ORGANIZATION

(4) Since the greater the degree of segmentation the greater also the need for coordination of activities, the greater the amount of system control that has to be provided. (5) In the undifferentiated system an entrant needs to engage from the start in all group activities, whereas the segmented system provides the possibility of an entry route consisting of transitional activities. It was noted earlier that an undifferentiated system may undergo temporary segmentation in order to provide an entry route. (6) The degree of segmentation affects the flexibility of the system with respect to the effects of absences, turnover, and emergency conditions. In the differentiated, unsegmented system, individuals can easily be moved from one task to another where necessary. The degree of relocation possible decreases with increase in the degree of segmentation.

(c) FLOW STRUCTURE OF WORK DOMAINS To begin with, it will be necessary to distinguish between shifts of workers within their job domain, and shifts of workers from one job domain to another. While by definition a worker may be shifted freely between tasks within his job domain, so that all shifts within a job domain are reversible, a shift from one job domain to another involves abandoning all tasks in the previous work domain for tasks in the new domain in so far as the two domains are completely segmented. Further, shifts between job domains are predominantly one-directional. To put it in another way, one may say that the tasks within a job domain are compatible with one another with respect to participation in them, whereas the tasks contained in different job domains outside their region of overlap are incompatible in this respect. To give an example: the work carried out by a puller and the work carried out by a pay clerk constitute segmented job domains, since carrying out any job within the puller's work domain, such as belt-shifting, is incompatible with carrying out any task included in the work domain of the pay clerk, such as making out pay slips. A shift from one job domain to another will be referred to as a transfer.

Two types of shift within a job domain may be distinguished. Workers may be temporarily shifted from one team to another to

151 L

TASKS AND ORGANIZATIONAL STRUCTURE

,

make up for absences, or to help out if a team falls behind scheduiJ and in these cases they return to their old team as soon as their hel~ is no longer required. Such a shift will be referred to as a relocation.' A more permanent shift from one team to another will be referred' to as a reassignment. Looked at from the point of view of the work system as a whole, the relocation system corresponds to some extent to a theoretical' flow model discussed in Herbst {1954b). Each team needs to keep up a certain strength so that if workers are missing from one team or additional individuals are required in another team, a flow of persons between teams results. Coalmining studies show that the direction of relocation is by no means haphazard. A team generally has a body of spare men who are on datal work. In some pits the number of spares, that is workers who are qualified for face-work but for whom no place can be found, is always sufficiently large to make up for wastage through turnover or absence. Under these conditions no shifting of workers between teams will be necessary. In other pits, on the other hand, the number of spare men available is rarely sufficient, so that relocation of workers between teams must be arranged. It may ~ stated as a general principle that workers will be shifted from tasks that are perceived as less essential to tasks that are perceived as more essential. The direction of flow, therefore, provides a measure of the perceived relative centrality of tasks within the work system, and, by implication, of the status hierarchy of roles associated with tasks. Whether a worker is shifted from one task to another wiJl depend on: - the relative importance of the task within the work system - the extent to which various work teams are ahead of or behind schedule relative to the tasks carried out by other teams.

Figure 10-11 depicts a relocation structure that was found for men working on a conventional longwall face. It will be seen that the structure is completely hierarchical. Without exception, relocations made are in one direction only. Rippers are used to reinforce stonemen, and stone-men are transferred to pulling, but no relocations are made in the reverse direction. This is an interesting example of a case where, in order to cope with technical breakdowns, maintenance 152

PRODUCTION TASKS AND WORK ORGANIZATION

men were used to reinforce men on production work; as a result, little maintenance work got done, and there was a state of permanent 'dysfunction'. FIGURE 10-11

DIRECTION AND NUMBER OF RELOCATIONS

ON CONVENTIONAL LONGWALL FACES IN A SEAM OVER A SINGLE PAY PERIOD

Rippers

Hewers

;/

,jjflec:

Stone-men

~10 Pullers

Fillers

Cutters

On the basis of the data presented in Figure 10-11 it can be seen that there are three separate hierarchies with a partially common base (Table 10-10). Pullers have a higher status than stone-men, and below the latter are the hewers and the rippers. Cutters are above scuffiers who are above hewers. In the third hierarchy, fillers are above hewers. TABLE 10-10

Status Low High

ROLE HIERARCHY

I

li

Ill

Hewers, Rippers Stone-men Pullers

Hewers Scuffiers Cutters

Hewers Fillers

The extent to which relocation is possible defines the flexibility of the social system. The degree of flexibility will be a function of the degree of segmentation of the work domain. A measure of flexibility should have a maximum value of 1 when the work domain is unsegmented and every worker can be shifted at any time to any of the tasks included in it, and should have a minimum value of 0 153

TASKS AND ORGANIZATIONAL STRUCTURE

when the work domain is completely segmented and no relocation is possible. If the tasks for which a worker is qualified do in fact set the limit to the tasks that he may carry out, then the former can be used as a measure of potential flexibility. Table 10-11 shows two extreme cases of potential flexibility. In the first case (a), each worker is qualified for only one task; in the second case (b), each worker is qualified for all tasks.

TABLE 10-11 POTENTIAL FLEXIBILITY MATRICES, SHOWING TASKS THAT EACH MAN IS QUALIFIED OR ABLE TO CARRY OUT

Tasks

F. IX

Persons B

A c D

Tasks

fJ

)'

A

Persons B

1

c

1

1

D

IX

fJ

,

1 1 1 1

1 1 1 1

1 1 1 1

(a)

(b)

For N persons and n tasks, the sum of cell entries is N in the first case and nN in the second case. If~ is the sum of cell entries in any given case, then a measure that varies between 0 and 1 is . I flex1"b"l" ~- N ). degree of potentia 11ty = N(n1

For case (b) above: number of tasks (n) number of persons (N) number of cell entries (~)

=

3

= 4

12 12-4 degree of potential flexibility = 4(3 _ 1) = 1. =

Actual flexibility is given by the number of tasks to which fillers, pullers, etc. are in fact relocated.

154

PRODUCTION TASKS AND WORK ORGANIZATION

The Effects of Internal Flow on Team Cohesion While a high degree of flexibility may have beneficial effects from the point of view of the efficient functioning of the technical system, it may, if carried beyond a certain point, have detrimental effects on the social structure in so far as it involves a serious disruption of existing teams. The higher the degree of flow, the greater will be the rate at which the composition of teams will change. If that rate is relatively low, then the team may well be able to deal with it. If it rises beyond a certain point, the team will not be able to maintain a stable social structure and interpersonal relations will tend to become depersonalized. A group changes its composition both by virtue of the people who may enter it and by virtue of those who may leave it. A measure of the rate at which a group changes its composition therefore needs to take both entrants and leavers into account. What is required is a measure that will make it possible to state to what extent a group remains the same from one period of time to another. A change in group composition will occur: (i) if there is an exchange: one man leaves the group and is replaced by another (ii) if the group increases in size (iii) if the group decreases in size. A measure of the change in composition has to combine the effects of these three types of change, and to vary between 0 and 100 per cent. A measure that may be used for this purpose is shown in Table 10-12. First, the number of men who have left and the number of men who have joined are recorded. The difference between these numbers gives the amount of change in group size. Then, take the number who have left or the number who have joined, whichever is the larger, and subtract the amount of change in group size. This gives the amount of change due to exchange of group members. The amount of change in group composition is then given by: number of men exchanged+ change in group size size of group (max.) That is, the amount of change due to the replacement of men who have left is added to the amount of increase or decrease in group 155

TASKS AND ORGANIZATIONAL STRUCTURE

size, and the total is divided by the size of the group at either the earlier or the later period, whichever is the larger, in order to avoid values exceeding 100 per cent. TABLE 10-12

RECORD FOR CALCULATION OF THE AMOUNT OF CHANGE IN GROUP COMPOSITION

Size Period of no. group

-1

9

2

12

Amount of change No. who No. who due to: have Change have joined left Exchange in size

5

2

2

3

Change in group composition

2 + 3 = 41·7% 12

The measure equates the effects of the three possible types of change that may occur as follows. A value of 50 per cent change in group composition is obtained: (i) if half of the group members are exchanged (ii) if the size of the group is halved (iii) if the size of the group is doubled.

156

CHAPTER 11

Research Tasks and Research Organization'

---·--An application of socio-technical analysis to questions of research organization and policy will involve consideration of the following: 1. The characteristics of research tasks. 2. The implications of these characteristics for appropriate and feasible forms of research organization. 3. The requirements of research organizations in terms of their linkage to specific client systems. 4. The properties and optimization requirements of the ecological network of research organizations and client systems. Since socio-technical analysis was originally developed for the study of production systems, an extension of the framework is needed if it is to be applied to research organizations. 1. CHARACTERISTICS OF RESEARCH TASKS In the first place, there is a basic difference between research tasks and production tasks. Production processes have to a significant degree the characteristics of a determinate sequence of operations which, with a relatively high level of reliability, leads from a given input to a predictable output. Research tasks are generally not of this type, but are indeterminate - with respect to the required initial state conditions, or the method to be followed, or the final result that will be obtained. Research may be defined as work concerned with the conversion of an indeterminate task into one that has a lower degree of 1 Written in 1968 at the Technion, Haifa, and the University of Jerusalem (Department of Sociology), in the course of an exploratory study of science policy carried out with the support of the Norwegian Council for Science and the Humanities.

157

TASKS AND ORGANIZATIONAL STRUCTURE

indeterminacy. The reduction, or, where feasible, the elimination, of indeterminacy leads to the possibility of being able to state that under given initial conditions a given set of operations will lead to a specified result. Once this is achieved, a research task is, at least in principle, converted into a production-type task. At this level there is no intrinsic difference between theoretical and applied research. Whether a given research task is classified as theoretical or applied depends more on the immediate context in which it is carried out than on its intrinsic characteristics. To the extent that differences exist in the nature of research tasks and production tasks, the implication is that the types of work organization appropriate to production will not be appropriate to research. Moreover, in so far as we are likely to find different types of research task, there will be a different range of work organizations appropriate to each of them.

2.

RESEARCH TASKS AND RESEARCH ORGANIZATION

A major problem encountered in research organizations derives from the uncertainty characteristics of research tasks. This problem arises for individual researchers and for research management, as well as for institutions, such as client systems and foundations, that are linked to research organizations. A competent researcher could possibly be defined as one who is capable of coping productively with uncertainty; who is capable of admitting to himself and others that he does not from the start know all the answers. This capacity, in some form, may need to be possessed by all institutions that are linked to research organizations. Many of the potential causes of sub-optimal functioning of research organizations can be traced to defences against anxiety generated by the intrinsic uncertainty characteristics of research. Defence may take the form, at the personal level, of excessive concentration on safe routine-type projects; at the management level, of control by means of close supervision of personnel; and in clients and foundations, of demands for complete specification of the work to be done, which means that the research task is contracted as a production task. At the other extreme, the research worker may look for or be put into monastic seclusion to pursue his work. However, there are conditions where each of these organizational

158

RESEARCH TASKS AND RESEARCH ORGANIZATION

trends may be functional. Institutes that specialize in the application of routine methods, paid for by clients, will tend to an organizational structure appropriate to production organization, whereas high-risk, innovative research, on a low budget, will be more appropriately done by lone researchers. In a more systematic analysis each of these variables would need to be taken into consideration. As a general principle, research tasks that can be defined as activities concerned with the reduction of uncertainty require work organizations that possess a corresponding level of requisite variety. This implies that the research team must possess an adequate range of competence and skill, and also have sufficient autonomy and internal flexibility to be able to adopt organizational patterns and work procedures appropriate to emerging task requirements. The results obtained from earlier studies of the conditions required for autonomous group functioning should therefore be of some relevance. 3. RESEARCH-CLIENT RELATIONSHIPS In studying the relationship between research and production organizations we cannot assume from the outset that the present almost complete organizational separation of operating and research functions is the only possible model. On the one hand, the increasing complexity and increasing automation of production processes have resulted in an upgrading of task requirements for operators. The tasks of operators no longer lend themselves to a techno-economic design in which the required activities of each operator can be specified in complete detail. To the extent that task requirements become more indeterminate and subject to change, then production tasks increasingly acquire the intrinsic characteristics of research tasks. The most successful of the firms that have participated in the Norwegian Industrial Democracy Project turn out to be those in which, in the course of establishing autonomous work groups, workers have to an increasing extent taken up research functions, thereby creating the necessary conditions for a continuous process of technical and organizational change. The tasks of operators can be extended to include both technical and organizational research, covering evaluation of existing conditions, formulation of possible directions of change, implementation of joint decisions, and evaluation of results.

159

TASKS AND ORGANIZATIONAL STRUCTURE

At the same time, technological development, especially the introduction of computers and the industrialization of research, leads to a relative downgrading of research tasks which increasingly acquire the characteristics of production tasks. What is beginning to emerge, then, is a society in which: (i) There will be relatively little difference in the task characteristics of production and research organizations. These will differ more in their predominant orientation than in intrinsic task characteristics. (ii) The existing rigid walls between production and research organizations will disappear. Increasingly their tasks overlap, and this will make possible an increasing mobility of personnel in both directions. (iii) There will be a fundamental change in the role requirements for researchers and specialists in terms of their relations with clients. Whereas, at present, researchers and specialists tend to have an authoritarian and prescriptive role, they will need to accept the role of consultants to operating teams in a joint search for solutions to technical and social problems. (iv) Temporarily, at least, specialists and researchers will feel their status and special prerogatives threatened. Neither universities nor research organizations may be able to continue their traditional role of providing almost the only accepted protected retreat from social involvement. While the need to conserve open spaces in nature is being recognized, the need to conserve open and free spaces in society may not be understood and recognized until it is too late. The major determinants of the characteristics of a research organization are likely to be the nature of its source of funds, the kinds of client it has, and the type of relationship it has with funding and client organizations if these are distinct. A systematic typology of the transactional structure of research organizations and an analysis of the dynamic properties of possible configurational structures would appear to be of particular importance. Both the analysis of research tasks and the analysis of the transactional relationship structure of research organizations require the development of new techniques. The study of the internal structure

160

RESEARCH TASKS AND RESEARCH ORGANIZATION

of research organizations presents fewer problems since here we can apply already existing techniques for studying the structure of work organizations. 4.

THE ECOLOGY OF RESEARCH ORGANIZATIONS

An essential problem of science policy is concerned with the optimization of the total ecology of research organizations, client systems, and funding sources within, minimally, a national region. However, in order to tackle problems at this level, some understanding of each of the subordinate levels is required. As far as research strategy is concerned, the best point of entry might be at the third level, since the development of techniques for studying the transactional structure of research organizations would help to identify the variables required for mapping out ecological network structures. Furthermore, by starting at this level it would be possible to test the adequacy of possible conceptual frameworks by means of relatively easily accessible data. A CONCEPTUAL FRAMEWORK FOR THE STUDY OF RESEARCH TASKS In the remainder of this chapter we shall consider, first, the intrinsic characteristics of research tasks and, second, their implications for research organization. DEFINITIONS

We shall employ as a basic concept an operational unit defined in terms of: (i) an initial state, S 1 (ii) an operation or sequence of operations performed on the initial state, 71' (iii) a final outcome state achieved, S(J. A sequence of operations that leads from a given initial state to a final state is referred to as a route. A route generally refers to a method or technique. An operational unit provides a definition of a task.

161

TASKS AND ORGANIZATIONAL STRUCTURE

A task is said to be determinate if the application of a specific set of operations to a given input state always results in a predictable output state. The principle may be determinate or stochastic. This may be put in the form 7T

(SJ--+ S0 •

A task is said to be indeterminate if one or more elements are unknown.

Assumption 1: Research tasks involve the reduction of the indeterminacy level of a task. We shall show in the following that, when the operational-unit concept is used, there are at least three dimensions in terms of which a research task can be analysed. The three dimensions are referred to as indeterminacy level, innovation level, and generality level. INDETERMINACY LEVEL

Given that a task can be defined in terms of three elements, an initial state, an operation performed on the initial state, and an outcome state, then we can define a first-order research task as one in which any two elements of the operational unit are given and the third element has to be found. A second-order research task is one in which only one element (generally an initial state or an outcome state) is given and the remaining two elements of the operational unit have to be found. A zero-order research task can in this case be defined as one in which all three elements are given and therefore it does not constitute a research task. Our initial assumption may thus be rephrased to state that the ideal aim of research is to achieve a zero-order research task. INNOVATION LEVEL

In terms of decreasing level of innovation a research project may be: (a) innovative, if no previous solution to the problem exists and it is not soluble by means of known routines or techniques; (b) alternative innovative, if an alternative solution to a problem is found, the requirement generally being that the new method or 162

TABLE 11-1

RESEARCH TASKS DEFINED IN TERMS OF INDETERMINACY LEVEL AND INNOVATION LEVEL

Indeterminacy level Initial state

Route

Outcome state

Secondorder task

given unknown unknown

unknown given unknown

unknown unknown given

Firstorder task

given given unknown

unknown given given

given unknown given

given

given

given

Zero-order task -~---

-

-----~

-

-~

--~

Innovation level

Innovative

Factfinding

Alternative intwvative

Marginal

Rep/icative

X

X

X

X

X

X X

X

X

X

X

X

X

X

X --~

-

---

- -

X -

Note: x signifies combinations that are not possible.

---

X - -

----

~-

TASKS AND ORGANIZATIONAL STRUCTURE

solution is more economical than the previously known solution or gives superior results; (c) fact-finding, using established methods or techniques to obtain new findings; (d) marginal, improving or optimizing existing methods or techniques to make them more economical or more effective; (e) replicative, checking on or repeating previous findings. The above order does not necessarily imply a descending scale in terms of the value of the project.

Theorem I: No first-order research task can be innovative unless it involves finding a route from a given initial state to a final state. Theorem 2: All second-order research tasks are innovative or fact-finding. In terms of the two variables considered so far, the possible types of research task can be set out as shown in Table 11-1. Those combinations that are not possible are indicated in the table by x . It will be noted that the replicative task is a zero-order one, since the initial state, route, and final state are all given. What is tested in this case is that the operational principle does in fact hold. For an analysis of the intrinsic characteristics of research tasks, a third variable is still needed.

GBNBRALITY LBVEL

(la) Working Principles Many of the most significant discoveries began as simple working principles. For instance, certain herbs were found to have curative effects; some substances when mixed were found to be explosive; some metals when allowed to rotate freely were found to come to rest in a constant direction; some chemical reactions were found to proceed only if another substance was added as a catalyst. The research technique employed is basically one of trial and error, and systematic replication of findings is required in each case since no other method of providing evidence is available at this stage. What is discovered is a route that leads from an initial state to a final state. What is not known is which of the initial state conditions

164

RSEARCH TASKS AND RESEARCH ORGANIZATION

and which of the operations implemented are essential and which are irrelevant.

(lb) Operational Principles At the next stage the essential conditions under which a principle operates are determined. In the case, say, of herbs that have been found to be curative, the active ingredient is identified. A new research task may in this case be to find a chemical synthesis of the active agent. What is at this stage still unknown is why the principle operates; that is, the complete sequence of operations that, given a specified initial state and a set of implemented operations, leads to the final state. Thus, while the principle of catalytic action has been known and applied for a long time, it is only relatively recently that some progress has been made in discovering how different catalysts operate.

(le) Scientific Operational Principles A scientific operational principle requires the specification of all the essential initial conditions and the complete sequence of operations (including those that have to be implemented) that lead predictably to a given outcome state. Stochastic principles are a special case that come into the same category. In this case the structural details and phases of the stochastic process have to be completely specified. All operational principles have the nature of a model with quasi-causal characteristics.

(2) Functional Laws If we use the analogy of a street map, then all the operational principles considered so far provide us with information as to what route to follow in order to get from one particular location to another one. A functional law is a principle that enables us, using the same analogy, to derive all the possible routes that can be taken from any one location to any other and provides us, so to say, with the total street map. 165

TASKS AND ORGANIZATIONAL STRUCTURE

Thus the gas law, relating pressure (p), volume (v), and temperature (T),

pv = R T tells us, for any initial state of the volume of a gas, what will be the effect of the operation of increasing the pressure at any given temperature. It also tells us, for any initial state of the pressure, what will be the effect of the operation of increasing the temperature for any given volume of the gas. Quite generally, functional laws should be reducible to or transformable into a set of operational principles that can be empirically tested. 1 Finally, the aim is to arrive at a postulate system. (3) J>ostulate ~)lster,n

A postulate system makes it possible to derive the total network of functional laws within a given discipline from a minimal set of postulates. It will be seen, from Table 11-2, that the analytical scheme is applicable to research tasks ranging all the way from abstract theoretical work to production engineering. The table shows that there are basically three types of secondorder task, depending on the type of question asked, as follows:

T)lpe of task

Question asked

Convergent How can we achieve it? Divergent What can we do with it? Method-oriented How can this method or technique be utilized? Each type of problem can arise in basic and applied research, and it will be noted that, as far as the intrinsic characteristics of research tasks are concerned, no necessary relation exists between the nature of the task and whether it is in the field of basic research, applied research, or production engineering. With regard to the implications of the intrinsic characteristics of 1 However, given a functional law, all theoretically possible operational sequences are not necessarily realizable. In the present case, no change of temperature can be achieved by alteration of the pressure and volume (Wold,

1966).

166

TABLE 11-2 Second-order tasks Given Unknown

Outcome Initial state state, Route

I

TYPES OF SECOND-ORDER RESEARCH TASK

Chemical research

Production engineering

or theoretical derivation of an empirical principle

To discover a method of synthesizing a chemical substance

To discover a method, and the components required, for the manufacture of a specific mechanism

Theory formulation

I To discover proof of a theorem,

Initial state

Route, Outcome state

Given a set of conceptual elements and definitions, to discover the theorems that can be derived from them

Given a class of chemical substances, to discover which new compounds can be synthesized from them

To develop production techniques for utilizing specific raw materials

Route

Initial state, Outcome state

Given a new theory, model, or measurement technique (i.e. information theory, opensystem theory, factor analysis), to investigate its utilization within a given field of study

Given a new analytic technique or a new synthesizing technique, to discover its possible utilization

To discover alternative utilization of a production-process technique

TASKS AND ORGANIZATIONAL STRUCTURE

research tasks for research organization, the analytical scheme suggests the following: 1. To the extent that the task has a determinate structure, the organizational requirements will approximate to those found to be applicable to production processes, and the research organization could well be located within or attached to the client system (e.g. industrial laboratories in the case of industrial research). 2. Different types of organizational requirement appear to be indicated for convergent, divergent, and method-oriented research. In the field of applied research, convergent tasks are undertaken in mission-oriented and action-research institutes. Divergent tasks are carried out in institutes concerned with discovering alternative uses and exploitation possibilities for available resources and products. Method-oriented research institutes appear to be particularly prevalent in the social sciences (possibly because they provide practitioners with a basis of demonstrable competence). Methodoriented research is found, for instance, in institutes for survey research, Gallup polls, personnel-testing, factor analysis, scaling techniques, etc. There is possibly a tendency also for institutes, whatever their starting-point, to develop in the direction of method-oriented research if they begin to accumulate costly equipment, and personnel with particular skills. The utilization of existing equipment (such as an expensive computer) and available skills can in this case become a primary concern. Convergent mission-oriented research requires that members of the institute share a common set of values with respect to the aims to be achieved. Divergent research requires relative independence and autonomy to generate and explore alternatives, and a primary orientation towards economic feasibility and market requirements. In method-oriented research the method itself tends to become an object of value, and research tends to be either its routine application provided as a service to clients, or marginal improvements in the technique itself. There is, finally, an important class of first-order problems that require the discovery of an effective route from an initial state to an outcome state. An example is desalination research. Here research

168

RESEARCH TASKS AND RESEARCH ORGANIZATION

and development work is being carried out, employing completely different processes based on physical, biological, and chemical principles. The organizational requirement will be similar to that for divergent research, namely a number of relatively independent competing units with a provision for sharing information obtained. Not all research projects can be looked at as problem-solving tasks of the kinds considered so far. In the development of a new field of study, or the development of a theory, none of the elements - initial state, route, or outcome state - may initially be given. Within the present analytical framework this would correspond to a third-order task that would always be innovative or alternative innovative. Research of this type, which cannot be pre-specified in detail, generally requires individual research workers with sufficient freedom to formulate their own research projects. The aim in this kind of research is to reduce the problem to an identifiable second- or first-order research task. Once this is achieved, the research task may become sufficiently specifiable to be taken over by appropriate research teams.

169

CHAPTER 12

The Structure of Science and Developmental Trends 1

---·--Developments in any one scientific field generally have consequences for other fields as well. In this chapter an attempt is made to show how a mapping-out of the structural relationship between different sciences makes it possible to discern possible developmental trends in the structure of science as a whole. This approach also has policy implications in the way of identifying possible support requirements of different applied fields. THE STRUCTURE OF SCIENCE IN TERMS OF LINKING DISCIPLINES

The extent to which scientific fields are linked to one another can be mapped out, to begin with, in terms of connecting disciplines. We shall take as basic subjects physics, chemistry, biology, psychology, sociology (including anthropology, political science, and international relations), and economics. The major linking disciplines at present are between physics and chemistry (physical chemistry), chemistry and biology (biochemistry), psychology and sociology (social psychology), and sociology and economics (socio-economics). Weaker linking disciplines exist, such as biophysics and psychochemistry (effects of drugs on behaviour, and brain chemistry). There is no direct linking discipline at present between physics and any of the behavioural sciences. If these relationships are mapped out, it will be seen (Figure 12-1) that there exists an ordinal distance scale going from physics on the one hand to economics on the other. Strong linking disciplines 1

Written as a working paper for a study group on science policy of the

Scandinavian Summer University, 1966.

170

THE STRUCTURE OF SCIENCE AND DEVELOPMENTAL TRENDS

exist between adjacent fields; weaker links exist between fields that are two positions apart; and no linking disciplines exist between fields separated by a greater distance. The appearance of two distinct clusters- with physics, chemistry, biology on the one hand, and psychology, sociology, economics on the other - seems to be due primarily to the weakly developed link between biology and psychology. Mathematics is, surprisingly, linked most strongly to the two most distant fields, physics and economics, so that the total structure takes on a cyclic form. We can use the existing structure to make some predictions about developmental trends, at least in the immediate future. FIGURE 12-1

THE STRUCTURE OF SCIENCE IN TERMS OF EXISTING LINKING DISCIPLINES

/

~~------~<-------,, .., ........... ......

.,. ~,---------, .........

1

Phys~i~s-Chemi~ry Bio\~gy----Psy~~6vlogy~~~;iology-Ecoilomics \\

I \

'

',

/

I \

\

I

/ ,"'

_,"'

"'

/

Mathematics

- - strong linking disciplines - - - - weak linking disciplines

First, it will be noted that it is between the adjacent fields of biology and psychology that a second-order linkage is only beginning to be developed. A few significant pioneering studies have been carried out, such as studies of the ecology of animal populations, tracing out the relation between the psycho-biological survival mechanism at the level of individual animals and the control mechanisms found within the interaction relationships among animals. Since, over the next decades, serious problems will emerge owing to the accelerating increase in human population density and the resulting threat to the balance of the total human, animal, and plant ecology, this is a critical area of study. Work that has been done on animal populations indicates that basic biological and psychological changes occur when population density increases beyond a critical level, which can take the form of either excessive apathy or excessive aggression. Both are control mechanisms that reduce population pressure.

171

TASKS AND ORGANIZATIONAL STRUCTURE

Second, the development of a stronger linkage between psychology and biology, which appear at present as structurally the most central disciplines, is likely to have fundamental consequences. The work of ethologists, which is still largely disregarded, may in this case play a crucial role in the development of science as a whole. At the same time, progress made in ethology is immediately relevant to progress that can be made in the study of larger ecological systems. Given this direction of development, which is indicated by urgent practical needs and consistent with the developmental direction within science itself, then the relationships between what will become central disciplines at the next stage will be as shown in Figure 12-2. FIGURE 12-2

EMERGING CENTRAL STRUCTURE WITHIN ECOLOGY-RELATED DISCIPLINES

biochemistry

So far, the discussion has been restricted to relatively short-term developmental changes. If we look again at Figure 12-1, we find that the hierarchical order of disciplines is the same as that discussed by Comte at the end of the last century. He pointed out that, beginning with physics, each discipline provides the basis for the succeeding discipline, in this case chemistry; chemistry provides the basis for biology; and so on. At the same time, as we go up this hierarchy, the size of the unit of study tends to increase. 1 This hierarchy has thus remained fundamentally unchanged as a basic structure, and the very successful progress that has been made over the past hundred years has been to fill in the details. It is nonetheless possible that a different order might ultimately develop which might change the character of science itself. At present, the so-called natural sciences and the behavioural sciences are still largely unconnected. Suppose we want to integrate them. Is the 1 1bis point has recently been stressed by general system theorists. However, the relationship between sciences cannot be accounted for solely in terms of size of unit.

172

THE STRUCTURE OF SCIENCE AND DEVELOPMENTAL TRENDS

development of a closer link between biology and psychology the only way to proceed? In terms of the presumptive hierarchical order of sciences it would appear so. However, on theoretical grounds another way of linking the natural and the behavioural sciences looks more promising. In the natural sciences, physics supplies the basic principles which become more complex as we move on to chemistry and biology. Similarly, in the behavioural sciences, psychology provides the basic principles which then become applied to small groups and to larger organizations. In this case the development of a direct linkage between physical and psychological theory construction may turn out to be relatively more straightforward. To the extent that progress in this direction can be made, the consequences in terms of practical implications may not be appreciable; however, our view of the nature and implications of science would be basically altered. DEPENDENCE RELATIONSHIPS BETWEEN SCIENCES

We shall next examine the possibility of using a method that has been applied in recent years to the study of complex industrial technologies. Given a set of elements or processes, we determine for every pair of elements whether the effective performance of one is dependent on the effective performance of the other. We obtain in this case a binary matrix which can be represented by a graph. Since there are different types of dependence relationship, it is important to make explicit the criteria used. We shall ask in the following: 'To what extent is progress with respect to theory and method in discipline A dependent on progress made in discipline B ?' With respect to relationships between applied fields, the criterion used is: 'To what extent is the level of progress made in one applied field dependent on the level of progress made in another field?' Since there is at least in some cases some degree of uncertainty in judgements of this type it would be best to obtain expert judgements, preferably from those engaged in multidisciplinary research. The results could then be combined and possibly converted into a probability scale. What one is looking for in this case is not so much consensus on each specific judgement as one or possibly more than one explicit model of the total structure of the scientific field. 1 1

A method of mapping of this type was introduced by Bemal (1939) but

173

TASKS AND ORGANIZATIONAL STRUCTURE

Figure 12-3 shows an illustrative diagram. It includes the applied sciences of medicine and engineering, together with three applied fields, industrial organization, agriculture and food technology, and education and health. FIGURE 12-3

DEVELOPMENTAL DEPENDENCE LINKS BETWEEN SCIENCES AND FIELDS OF APPLICATION

Education & Health

__., strong links __ ,...relatively weak. links

The central star within this structure is, not surprisingly, mathematics. The behavioural and the natural sciences again form two distinct clusters, each of which appears to have a different structure. The natural sciences have a hierarchical structure where each discipline provides the basis for the next and each of these links is appears to have received little attention. Since the structure of science changes over time, it would be useful if maps of the field could be constructed on the basis of expert evaluation of each discipline at regularlintervals. The need for systematic studies of this type has been emphasized by Price (1962).

174

THE STRUCTURE OF SCIENCE AND DEVELOPMENTAL TRENDS

highly developed. In recent years a major new development has been in the field of mathematical biophysics, where some progress is being made towards the application of mathematics to the study of networks of biological processes. While the original intention was to extend the methods of mathematical physics to the study of biological phenomena, the present trend is towards the formulation of a specific mathematical biology in terms of organizational processes and structure, so that a basis is now emerging for the future development of a new discipline embracing mathematics, biology, sociology, and psychology. Relations between the behavioural sciences are of a somewhat different form. At the present stage sociology occupies structurally the central position. The weak link between economics and psychology reflects a basic weakness in the present situation. On the one hand, the axioms of economics are based on common-sense psychological assumptions which are unrelated to psychological and empirical research. On the other hand, psychological theorists are scarcely aware that their models embody economic behaviour principles. One of the major foci of integration in recent years has been in the field of mathematical model-building. RELATIONS BETWEEN SCIENCES AND APPLIED FIELDS

There are essentially two points I wish to make here. The first is that competition between different sciences is at present primarily due to the competitive practical demands that are made on science. The second is that the need for the integration of different sciences arises primarily in the applied fields. So far, a great deal of attention has been given to translating basic research into applied research. This is no longer a one-way street and there is an increasing need now to feed back theoretical and methodological developments from the applied to the basic sciences. This has always been obvious in physics. Since the seventeenth century, progress made in physics has led to progress in engineering, and progress made in engineering has provided the basis for further progress in physics. The existence of this feedback cycle with very little lag has been the major reason for the effective and sustained development of physics. It is in this case almost as true to look at physics as a basic field relative to engineering as it is to look at

175

TASKS AND ORGANIZATIONAL STRUCTURE

engineering as a basic field relative to physics. To some extent biology and medicine constitute a similar couple. Here, however, medicine provides a perhaps too restricted applied basis for the development of fundamental biology. The significant point is that both physics and biology have intermediate, highly developed, applied fields. One of the weaknesses of the behavioural sciences is that a distinctive applied behavioural science has not yet emerged. Thus the situation is often rather like taking students trained in physics and asking them to solve civil engineering problems. At the present stage a sharp differentiation between fundamental and applied behavioural science would not be functional since little exists in the way of firmly established principles and methods. What is needed, however, is a better feedback between applied and fundamental work. Every applied field receives direct contributions from a number of sciences. In the case of industrial organization these are, in order of seniority, engineering, economics, and sociology. The problem at this stage is that the engineer is trained to look for the best engineering solution, which may not be the best from the viewpoint of economics. The economist is trained to look for an optimal economic solution, which will not take into account social and psychological factors and is likely to produce sub-optimal results even in terms of purely economic criteria. Since there is scarcely any practical problem of industrial operations that is a purely engineering, economic, or social organizational one, we find at this point a major source of the need for the integration of different disciplines. This is at present only at the beginning stage. The new discipline that is emerging, based on the integration of technology, sociology, and economics, is shown in Figure 12-4. The structure of science that is emerging is then likely to be somewhat similar to the successive stages in an industrial project, which are research, development, implementation. We have: (1) fundamental sciences (2) applied sciences (3) implementational sciences. What is interesting is that it is in the implementational sciences that the most challenging theoretical and methodological problems are at present found. Whatever progress is made here will have immediate implications for the fundamental sciences, so that a cyclic process is

176

THE STRUCTURE OF SCIENCE AND DEVELOPMENTAL TRENDS

being set up which is likely to lead to an even more rapid scientific development in the future. Facilities are needed, therefore, for the relatively easy transition of scientists between the fundamental and the implementational disciplines. FIGURE 12-4

EMERGING STRUCTURE INTERRELATING TECHNOLOGY, SOCIOLOGY, AND ECONOMICS

Sociology

socio-economics

---...,If---

Technology __\-+-.

'"\TT'

jlndustrial organization

Economics

j

Up to the present time, most applied research has been concentrated in the industrial sector, and within that sector an increasingly large portion in most developed countries has gone into armaments research. It is in that field that the most rapid developments have taken place and the major capital resources are available. The agricultural base of industrial society, which during this time has benefited from industrial development, could be taken almost for granted. We are now moving into a period where this is no longer so. Given the present rate of population increase, the problem will be no longer that of increasing the standard of living but that of maintaining sheer survival. The sciences that contribute directly to agriculture and food technology are biology, medicine, chemistry, engineering, economics, and sociology. We have noted that there are interdisciplinary trends within the fundamental sciences; however, this is still at a highly theoretical level. Within the implementational sector the degree of integration is still quite weak. This means that decisions are likely to be implemented that may optimize with respect to one criterion but

177

TASKS AND ORGANIZATIONAL STRUCTURE

will have unexpected consequences with respect to other criteria. Some of these consequences may mean only that the outcome is sub-optimal; others, brought about on a large scale, may be catastrophic. For instance, the intense agriculture of virgin lands has led to wide-scale erosion in Australia, Russia, and America. Pesticides and drugs have as yet incompletely understood effects on biological development, leading in some cases to the survival of the fittest and most resistant pests, and their consequences for the natural ecology as a whole are unknown. Some sectors of research, such as the production of artificial foods, can perhaps be allocated to the industrial sector. At the same time, there is need for a rapid build-up of agricultural, ocean research, and applied biology institutes to about the level of technological universities, and these have in turn to be linked to implementing organizations. We have so far considered only the relationships between some TABLE 12-1 DEVELOPMENTAL EFFECTS OF SCIENCES ON EACH 12-3

OTHER AND ON FIELDS OF APPLICATION, BASED ON FIGURE

Fundamental sciences

Fields of application

Applied sciences

;...

otl~

~J

~

(,)

";ii~ .l:! 'I:!

t(,<:l

"'!~

t t

Biology Psychology Sociology Economics

1 t 1 1 1 1 0 X 1 1 f t 1 0 0 X 1 f t 1 0 0 0 X 1 ! t 0000x11 0 0 0 0 1 X 1 0 0 0 0 t 1 X

Medicine Engineering

0000000

01!tt1t

Mathematics Physics

Chemistry

X

178

t

t

1

1

1 1

0 0

t

t t t t t

0 0

1 1

X

0

t

1

1

1

X

1

1

1

t

1

t 1

1

t

1 1

t t

t t

1 1

t

THE STRUCTURE OF SCIENCE AND DEVELOPMENTAL TRENDS

applied fields and the sciences that are directly linked to them. In Table 12-1, the distances between different sciences and applied fields are given, on the basis of Figure 12-3. Sciences that are immediately linked are assigned the value 1 ; those that are connected via one science are assigned the influence potential t; and, quite generally, sciences connected via N sciences are assigned the influence potential 1/N +I. It will be seen that the relationship between the fundamental sciences appears at present to be an almost perfectly hierarchical one. The weighting used is quite crude but is probably sufficient to get some idea of the priority that needs to be given to each science for the purpose of supporting the development of different applied fields (Table 12-2). TABLE 12-2 ORDER OF PRIORITY OF SCIENCES RELATIVE TO DIFFERENT FIELDS OF APPLICATION, DERIVED FROM TABLE 12-1

Industrial organization

Agriculture & Food technology

Engineering Economics Sociology

Biology Chemistry Medicine Engineering Sociology Economics

Physics Mathematics Chemistry Psychology Medicine Biology

Education & Health Psychology Sociology Medicine Engineering

Physics Psychology

Biology Chemistry Economics Mathematics

Mathematics

Physics

The problem of science policy needs to be separated into two distinct but related levels. At one level, decisions are concerned with the judged optimal priorities of the fundamental sciences, given their intrinsic state of development and their judged contributions to and effects on other sciences. Such decisions have to take account of the existing state of both national and international development. The results arrived at, at this level, have then to be further modified ~

179

TASKS AND ORGANIZATIONAL STRUCTURE

in terms of the priorities given to different fields of practical application. It will be noted that if, for instance, the agriculture and food technology sector is provided with major scientific resources, then this will to some extent interfere with the scientific backing essential for the industrial sector. Equally, if excessive weight is given to the sciences that are relevant to fields of application, then this may upset the optimal balance required for the development of the fundamental sciences.

180

CHAPTER 13

Maps of Knowledge and the Design of Educational Organizations1

Educational organizations are facing growing demands for change both in highly industrialized and in developing countries. The problems of such organizations are likely to become increasingly serious and chronic in the years to come. It cannot be said that there has been a lack of change- whether in architectural design, curricula design, subject content (e.g. updating), or teaching techniques- and these types of change may be necessary or desirable. However, piecemeal and isolated changes, especially when they are imposed from without, may add to the internal turbulence and to the uncertainties of both teachers and students. It is no longer a question of reforming an organizational structure in order to bring it up to date, nor is it simply a question of creating a new kind of organization. Rapid internal changes both in the field of knowledge and in techniques, and external changes in industrial and social organization, all point to the need to build a capacity for change into existing educational organizations. In order to do this, we need to be able to identify the essential characteristics of educational systems, which determine how the total organization operates. The characteristics of an educational organization are determined by the model used for structuring the educational task. If, for instance, the production-process model is applied to the structuring of the content and process of education, then the educational organization acquires the organizational characteristics of the traditional factory, and also their social-psychological consequences. Now that the 1 An invited lecture given at a conference on 'Teaching and Research' held at the University of Uppsala in 1970. First published in Norwegian in 1971 in Nordisk Forum, Vol. 6, No. 3, pp. 171-88.

181

TASKS AND ORGANIZATIONAL STRUCTURE

simple production-process model is no longer adequate to define the task and work characteristics of the new emerging types of industrial organization, it is likely to be easier, and in fact it becomes necessary, to reconsider a number of the basic assumptions that have been built into existing educational organizations. The development of alternative forms of educational organization depends on possible ways of redefining the educational task. Study of the relationship between the structure of educational tasks and the characteristics of educational organizations may be referred to as socio-didactic analysis and constitutes an extension of socio-technical theory. 1 TYPES OF TASK

Let us begin by making a distinction between two fundamentally different types of task. There are tasks that are determinate in the sense that every element is specifiable and the outcome is predictable. We can represent a task of this type as an operational unit where a specifiable operation '1J' applied to a given initial state Si leads to a predictable outcome state S0 : 'TT(Si) ~So. This is essentially a production-type task. 2 Beyond this we have a range of tasks that are indeterminate. For instance, we may be given an initial state and a required outcome state and the problem is to find a set of operations that will make it possible to get from the initial state to the outcome state:

?(SJ ~so. If in the educational setting we provide the student with the necessary instruction and training to achieve this, then we degrade what is essentially a research-type task into a production-type task. At the next stage we have tasks such as, given a new material, what can we do with it or how can we use it: ?(Si)~?

The concept of socio-didactic systems was proposed by Roggema (1969). The relationship may be deterministic or stochastic. Blichfeldt (1973) has suggested the use in this context of the terms 'closed' and 'open' tasks. 1

2

182

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

Finally, beyond this, there is the type of task we have increasingly to cope with, where no element can be specified at the outset:

?(?)-? For example, if we wish to create a democratic school or university organization, we cannot initially define in detail what this type of organization will look like and how it will function. We are given neither the relevant facts that define the existing situation, nor the necessary steps for implementation. It is this type of task that Thorsrud (1972) refers to when describing policy-making as a learning process. Those who are involved in a process of change are required to be able to identify relevant characteristics of the existing situation and to discover potential directions of change towards the objective. The exploratory implementation of one or more alternative first steps of a change process is then evaluated to define new alternative options which only now become visible. It is in the process of change that the objective aimed at becomes more clearly structured and definable and the characteristics of what was the initial situation become more clearly defined. As a general principle, success at any task level reduces the problem to the next task level below. Whenever a problem is found to be reducible to a determinate task, it can be programmed on a machine, and there is no need to bother about this any more except in so far as an understanding of various types of determinate task is a necessary step in learning to deal with higher-level tasks which require capacities that machines lack. Ideally, learning experience should be provided at each task level. However, the problem here is that each type of task requires a different type of organizational structure; and, in so far we impose or need to maintain a particular type of organizational structure, then this structures the task in a way consistent with the organizational requirements. TASK STRUCTURE AND EDUCATIONAL ORGANIZATION The relationship between task structure and work organization in the educational context can be set out as follows: 183 N

TASKS AND ORGANIZATIONAL STRUCTURE

1. The way in which the subject taught is structured as a task, and the criteria established for performance evaluation, determine the type of relationship that is possible between teacher and student and the task-mediated relationship between students. 2. The type of relationship that is possible between subjects is determined by the task characteristics of each subject. 3. The relationship between different subjects determines the type of relationship that is possible between teachers. 4. The task-mediated relationship between teachers and the characteristics of the student-staff relationship determine the type of control and authority structure that is possible within the total system. 1. Task Structure and Teacher-Student Relationships Roggema (1969), in a study carried out in Dutch teacher training colleges, shows that students are able to distinguish, with a high degree of consistency, between 'schoolish' and 'non-schoolish' teachers. Schoolish teachers are found to utilize the productionprocess model in structuring their task, while non-schoolish teachers utilize a research-type model. · The schoolish teacher splits his subject into small isolated bits, which have to be worked on and learnt one at a time. Students are required to follow rigid instructions. The performance of students, both in terms of following instructions and in terms of the results obtained, is judged simply as right or wrong. The teacher claims complete autonomy for himself as an expert, while allowing little or no autonomy to his pupils. Subjects that especially lend themselves to being taught in this manne1 are mathematics and foreign languages. A machine teaching programme is an extreme example of this type of teaching technique. Non-schoolish teachers give their students autonomy to investigate, discuss, and find out for themselves. The teacher defines his role as a resource person for the activities of his students. The teaching is problem-oriented. Where drill is needed, the purpose and meaning are explained. Judgement of performance is not simply in terms of right or wrong, but in terms of the development of increased ability, competence, and independence. Teachers of this type are more

184

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

often found in subjects such as geography and the social sciences. Subjects such as mathematics and languages, which have intrinsically a highly determinate structure, lend themselves more easily to being taught by means of a production-process technique. Subjects such as the social sciences are more in the nature of an exploratory research field and are not so easily accommodated to this method of teaching unless the teacher is prepared to be very dogmatic. An important finding in Roggema's study is that any subject can be taught in either way. 1 The structure of an educational task is thus not given, but the way in which the task is structured determines both the type of teacher-student relationship and the appropriate criteria for performance evaluation. Perhaps the most fundamental difference between the two teaching techniques lies in the evaluation criteria employed. The schoolish teacher is concerned with the ability of his students to follow instructions and to perform precisely a predetermined programme. The student has the choice of being either passive or rebellious. An attempt by the student to display originality is seen both as an attack on the teacher's expertise and as non-compliance with the task to be learnt. The conditions that create boredom with and alienation from the task, especially in schools, but increasingly also in universities, are thus very similar to those found in industry. This is not an accident. The schools as we know them today have very successfully produced the type of person who was needed by the society that came into being after the industrial revolution - a society that, in order to function, had to use human beings who were wiiiing to act as substitutes for machines and who were trained to subordinate themselves to the requirements of authoritarian hierarchical work organizations. 2. The Student in a Segmented Universe of Knowledge The way in which a task is structured determines the way in which it can be related to other tasks. A task that is completely specifiable has the characteristics of a closed system. To the extent that each subject can be structured and taught by means of a production-process model, then each subject can be taught independently of other 1 See also Wertheimer (1959), on alternative techniques for teaching mathematics.

185

TASKS AND ORGANIZATIONAL STRUCTURE

subjects. The universe of knowledge thus becomes split into independent segments. As a result, pupils at school are made to go in the course of a day through an incoherent sequence of unrelated subjects, establishing little more than a superficial relationship with any teacher. At university level most subjects are split even further into independent segments. For example, psychology is split into clinical, social, experimental, and developmental psychology, child psychology, gerontology, work psychology, organizational psychology, etc. The need for specialization has to be recognized; however, to the extent that each segment of knowledge becomes a self-contained unit, it becomes unrelated and therefore irrelevant to the world outside itself. The main difficulties encountered by those who leave schools and universities equipped with over-specialized knowledge relate to the increasingly rapid obsolescence of what they have learnt and to the fact that task requirements in industry and the types of problem with which professionals increasingly have to deal can no longer be handled within the confines of any one specialized field of compe-

tence. In contrast to programmable production-process tasks, research and problem-oriented tasks do not split up the field of knowledge and they require a type of organization based on cooperation rather than on individual competition. The types of problem encountered on the shop floor, in a hospital, or in a family require an understanding of social-psychological, economic, technological, and political aspects and their interrelationships. It appears likely that for some time to come a leading role in research will be taken by non-specialized, independent institutions, which are client- and problem-oriented and only loosely linked to the universities. Such institutions can, however, develop effectively only in so far as correlated changes take place in educational organizations. The initiative in developing new forms of educational organization should if possible come from the universities, since these to a large extent determine the structure and content of the subjects taught at the lower levels of the educational system, both indirectly in the form of a model and directly in terms of university-entry requirements. 186

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

3. The Relationship between Subjects and the Relationship between Teachers A schoolteacher is primarily responsible for the teaching of a specific subject. In so far as each subject is an independent and self-contained unit, then no relationship between teachers is required. In so far as teachers are unable to recognize their mutual relevance, they tend either to have an individualistic and competitive relationship with one another or to withdraw reactively into their private subjects. This indicates that a democratic type of educational organization cannot be achieved by concentrating primarily on the individual teacher-pupil relationship, since the characteristics of the existing organizational structure are largely determined by the relationship between teachers. One of the advantages of the present system is that each teacher can function as an exchangeable component, that is, he can be expected to do the same type of job whichever school be comes to, and he can carry out his task without building up personal relationships with either his pupils or other teachers, to the extent that these relationships are built into his formal role. The socio-technical organization of schools is thus similar to that found in the traditional type of factory. Here, however, we need to distinguish between two types of organizational system. We have, on the one hand, bureaucratic organizations, which are designed to be maximally impervious to organizational change, and which can be relied on to go on operating in the same way indefinitely and with a high degree of reliability. There is no point in complaining that school organizations, which are of this type, are difficult to change. This is not what they were designed for. What they do allow for is a reprogramming of the content of individual subjects at a very slow rate, since this does not affect the organizational structure. Factories, on the other hand, are normally not able to operate in this way unless they have an assured control of the market. Industrial organizations today generally have some degree of slack both in terms of staff positions and in terms of managerial staff who are responsible for initiating and working through change processes. By providing slack at all levels of the hierarchy so that members at all levels can participate both in defining the direction of organizational change and in working through the process of change, it 187

TASKS AND ORGANIZATIONAL STRUCTURE

becomes possible to create organizations with the capacity for continuous organizational learning. 1 It seems almost paradoxical to state that what existing educational organizations lack is the capacity for learning. However, at present few if any of the teaching and administrative staff have time for more than just keeping the existing system running, and under these conditions it does not help to set up a separate organization and give it the responsibility for introducing change. One of the first steps that need to be taken would appear to be to build in some degree of slack at all levels of the organization. A decentralization of the school system will require a redefinition of the role of the schoolteacher, which may need to become more like that of the university teacher in that classroom teaching will constitute only part of the total task. POSSIBLE STRATEGIES FOR CHANGING EDUCATIONAL ORGANIZATIONS

The relationships between task and organizational characteristics discussed so far are shown in Figure 13-1. FIGURE 13-1

TENTATIVE RELATIONSHIPS BETWEEN THE ELEMENTS OF SOCIO-DIDACTIC SYSTEMS

Student--student relationship

/ Task structure/ of subjects

tt

Relationship between subjects

.

Teache!-student relationship

t

Teac_her--:teacher relat1onsh1p

The interdependence of the different aspects of educational organizations has a number of implications: (a) If, as is at present the case in Norway, regulations and reforms concerning each aspect are imposed independently by the higher 1 1bis type of strategy, which was utilized in the Norwegian Industrial Democracy Project, is discussed by Thorsrud & Emery (1969) and Thorsrud (1972).

188

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

levels of the school hierarchy, then teachers find themselves submitted to inconsistent demands and are forced into a defensive position. (b) Since the total system within which teachers work does not have the characteristics of a democratic organization, the creation of representative pupil-teacher committees at the school level, while it may satisfy formal requirements for democratization, cannot within the given structure achieve this aim in practice, because the teaching staff have little freedom to explore, implement, and test organizational alternatives.

(c) If the major aim of the centralized educational authority system remains primarily that of retaining an identical structure for all schools, then this does not permit a process of evolutionary change. Evolutionary change requires a greater degree of autonomy and self-regulation for individual schools within which representative committees can have a meaningful role. Centralized departments would then acquire more of a consultative role. (d) The fact that the problems at present faced by both schools and universities manifest themselves primarily at the interface of student-teacher relationships does not mean that this is the strategic point within the total system where change needs to be implemented. The teacher-student relationship is embedded in and dependent on the existing task structure of subjects, together with the relevant performance evaluation criteria and the relationship between teachers. The type of relationship that is found between teachers is in turn dependent on the relationship between subjects. The dependence structure shown in Figure 13-1 suggests that the central component of the total structure is the interrelationship of subjects. Different modifications of this will enable different types of organizational structure to emerge. 1 1 F. E. Emery has suggested that, in Figure 13-1, the arrows between teacherteacher relationship, teacher-student relationship, and student-student relationship should be reversed. It would appear that the direction may be different depending on whether we look at the organizational design and development sequence, or the dynamic relationship within a functioning organization. In the latter case the dependence relationship is likely to operate in both directions.

189

TASKS AND ORGANIZATIONAL STRUCTURE MAPS OF KNOWLEDGE AND EDUCATIONAL ORGANIZATION

We shall in this section consider four types of relationship between subjects, each of which generates a different type of educational organization. The Segmented Model

It is this type of model that underlies the existing school organization and also the departmental and faculty structure of the traditional university. The school curriculum consists of a set of independent and self-contained subjects such as I History 11 Religion 11 English 11 Chemistry 11 Mathematics I

each of which can be taught without any reference to the others. At the university level we have

I

Natural sciences 11 Philosophy 11 Social sciences 11 Languages I

which are further segmented into many different disciplines each of which is defined and organized in a way that maximizes its selfcontained character. This map of knowledge has a number of implications. Since the field of knowledge is split into a number of independent parts, an understanding of the whole becomes an inconceivable and superhuman task. Sound knowledge can be achieved only by intensive concentration on some small segment of the field. Cross-disciplines can and do emerge, but each of these becomes again a new subdiscipline or floats temporarily in an academic void. In terms of application the assumption is that every social or technical problem can be subsumed under some particular specialty. The task of the client is simply that of finding the right specialist for his particular problem. Sometimes the problem has to be redefined to fit the discipline, or a new discipline may emerge to solve a new problem. The significant characteristic of the segmented model is that it provides no basis for intrinsic organization in so far as those who work in different disciplines cannot see their relevance to one 190

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

another. The relationship between subjects is in this case primarily competitive. The Hierarchical Model

The hierarchical model is a result of the assumption that everything is ultimately reducible to one particular discipline. The most influential model of this type derives from Comte. It was thought that the principles of sociology would be reducible to simpler principles of psychology, these to principles of biology, these in turn to principles of chemistry and physics, and these ultimately to mathematical principles: Mathematics

t

Physics

t

Chemistry

t

Biology

t

Psychology

t

Sociology This model has not been found suitable in terms of an organizational structure. It leaves out too many, admittedly mostly 'unscientific', subjects which do not fit into the hierarchy. However, the model has provided the traditional status ratings of the various sciences. Nobel prizes are awarded for physics and for chemistry, with mathematical physics having the highest status. Biology has been accepted into the fold, but psychology and sociology still wander in the academic wilderness however much they may claim for themselves scientific status. The Centralistic Model

This is the only model that up to the present has provided an accepted structure for integrating, at some periods, very nearly all subjects (see Figure 13-2). 191

TASKS AND ORGANIZATIONAL STRUCTURE

The medieval synthesis survived for many centuries. Within this synthesis theology allied with philosophy constituted the central subject, and provided the ideological structure for the whole. Science, history, art, geography, and political theory had to be formulated and could develop only in conformity with the basic tenets of the catholic teaching. Since each subject conformed to the same structural basis, each could be understood and seen in relation to the others. FIGURE 13-2

CENTRALISTIC MODELS

Philosophy History

..----

tt

Theology

+

~

Natural philosophy

Political theory

Research Education ..----

t

Economic ~ development

Natural sciences

•

Social sciences

Philosophy

.t •

History

..----

Political ideology

-

Natural sciences

t

Social sciences

Every centralized system survives only so long as it can suppress or devaluate all other viewpoints that would structure the world of knowledge in a different way. Throughout the Middle Ages there existed a large group of subjects with low status outside the academic establishments, such as alchemy and astrology, metallurgy, and the 192

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

various mechanical crafts such as instrument-making, clock-making, lens-grinding, etc. The link-up between the mechanical crafts and natural philosophy, which already potentially existed in ancient Greece, led to the break-up of the medieval synthesis. The Renaissance, with the emergence of new ways of looking at the world, had a truly liberating effect, but it did not produce a supporting system of social organization. Throughout the nineteenth century we find one attempt after the other, especially in German philosophy, to re-create a new centralistic system until, at the present time, we have two competing centralistic models. We have one system in which the natural sciences, education, and the social sciences become structured to supporting and accelerating the utilization of physical and human resources in the service of industrial production. In the competing centralistic system, political ideology linked to philosophy occupies the central position and structures the content of the social sciences, history, and the natural sciences. Since each subject is made to conform to the same structural basis, it is possible again to see each in relation to the others. Neither the conservative nor the politically radical position goes beyond the conceptual structure and the associated type of human and social organization inherited from the Middle Ages. They are different versions of what Marcuse (1964) calls the 'one-dimensional society' in which one way of structuring the world suppresses all other world views. If there is at present at least an indication that a possibility exists of going beyond the medieval type of world model then this is chiefly due to the emergence of spontaneous social mutations in the form of non-hierarchical social organizations. The Multiple-perspective Model A transcendence of the centralistic model becomes possible as soon as it is realized that each subject is capable of structuring and illuminating the whole world of knowledge. Each subject displays a different structure of the whole. There is thus nothing wrong with any of the centralistic models except in so far as it is claimed that one viewpoint is right and others are wrong. In the multipleperspective model the world no longer has a one-dimensional structure. Each subject allows some aspect of the whole to become 193

TASKS AND ORGANIZATIONAL STRUCTURE

visible so that it stands out as the figure while other aspects become part of the background and thus invisible. When another viewpoint is adopted, aspects of the whole that were previously invisible now become visible as part of a figure, while other aspects, previously visible, become part of the background. From the point of view of classical physics, the world is shown as a set of tiny solid atoms moving at various speeds in conformity with mechanical laws. One aspect of the world becomes visible and structured, while others become invisible. From the viewpoint of the artist or the psychologist, other aspects of the whole become elements and structures. However, each subject can also structure any other subject. Mathematics can be applied to the study of literature and linguistic structure; moreover, the linguist or the social scientist can look at mathematics as a form of language. Mathematics can be used to analyse the structural characteristics of a painting; but again, few would have devoted their life to mathematics if they had not recognized and experienced the aesthetic pleasure of an elegant proof or the intrinsic beauty of mathematical structures. The multiple-perspective model is shown in Figure 13-3. It will be seen that each of the previous models is a special case of the multiple-perspective model. FIGURE 13-3

MULTIPLE-PERSPECTIVE MODEL

hierarchical model segmented model ab denotes that subject a is applied to subject b; ba denotes that subject b is applied to subject a.

Each horizontal row gives a centralistic model. In the present case, subject c is applied to, and structures, every other subject. The central diagonal gives the segmented model in which each subject exists as a self-contained unit, unrelated to other subjects. 194

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

The remaining diagonals give different hierarchical models. In the present case, a provides the foundation for b, b for c, c for d, and so on. The multiple-perspective model is consistent with a matrix-type organization, which allows different subgroupings and structures to emerge depending on the nature of the task. Let us first consider how the model can be applied to a research organization (Figure 13-4). The larger technical research organizations are typically composed of a set of independent institutes, each FIGURE 13-4

MATRIX OP RESEARCH ORGANIZATIONS

of which is concerned with a specialist field of research, and they have, in addition, a central adininistrative and to some extent policy-making unit. While each unit may have to maintain its specialized function, it is possible to allow project teams to emerge, composed of members of several institutes depending on the nature of the problem. These teams can utilize the more specialist-oriented staff members as resource persons. In this case policy formulation can be reintegrated into the system and the condition for a process of organizational development exists. The transition to a matrix organization means essentially a transition from a rigid static system to one in which learning is possible both at the individual and at the organizational level. An application of this type of organization to tertiary education would not necessarily imply a fundamental change in the existing system of university education since training and research in specialties will need to be maintained. Instead, the possibility exists of creating 195

TASKS AND ORGANIZATIONAL STRUCTURE

a post-university centre to provide a polytecbnical type of education. If we assume that, as at present, students acquire, over an initial two- or three-year period, a basic knowledge of one or more subjects, then some could continue at university with more intensive specialization in one subject, while others could go over to a 'polycentre' organized on a different basis. The polycentre would consist of self-recruiting work groups. These could be composed, as in the case of the Scandinavian Summer University, of students and staff members interested in working on a specific problem, or they could be groups of students utilizing full-time or associated staff members as resource persons and consultants. Unlike the present Summer University, work groups could go beyond the exploratory stage to intensive work in the chosen problem area. Given that each student would be a member of several work groups, formal examinations would not be needed. As in any research institute, a group that had been working together for a sufficiently long period would be able to evaluate the competence of its members on the basis of their contribution to the work of the group, and independent evaluations from several work groups could be obtained for each student. Beyond this, independent assessments could be made of the work of each group and of the papers and research reports of individual students. The total organization is shown in Figure 13-5. FIGURE 13-5 AN EDUCATIONAL MATRIX MODEL Specialization

Basic subjects

Secondary university

Primary university

The polycentre would need to remain in close association with the university, so that students could study additional subjects there, at either a basic or a more advanced seminar level, in accordance with their needs, and without having to take examinations unless they wished to do so. The transition to a multiple-perspective model presents difficulties 196

THE DESIGN OF EDUCATIONAL ORGANIZATIONS

that are probably not simply related to intellectual capacity. Intensive specialization, in so far as it imposes a single structure on one's way of comprehending the world, may make it gradually increasingly difficult for a person to switch over to a different structure, logic, and mode of experience. The model requires that the world can be seen not just as a single Gestalt but as several Gestalts, depending on which viewpoint is taken, together with their interrelationships. Every conceptual model, as soon as it is formulated, becomes a closed system and also carries within itself the possibility of selftranscendence. If each discipline can structure the whole world and make it comprehensible, then no one viewpoint can claim to provide us with an understanding of the 'real' world; nor could we achieve such understanding by seeing the world simultaneously from all possible viewpoints, even if that were feasible. A true perception is possible only if the world is seen from no viewpoint whatsoever. The problem lies in the inherent limitations of the processes of perception and conceptualization. A structure or figure becomes visible in so far as we suppress and make invisible the background from which it stands out. Thus perception is not possible without distortion of what is seen, and conceptualization is possible only by virtue of what is left out. To go beyond this we would need to see figure and ground simultaneously. However, if figure and ground become one, then there is no figure to be seen, no object that can be got hold of, and no basis for conceptualization. Conceptually, the step beyond takes us over the edge of the knowable, which is at the same time both possible and inconceivable.

197

Epilogue

CHAPTER 14

The Evolution of World Models

---·--I shall use the term world model to refer to that aspect of the world that is judged by man at any given time to be relevant to human action and survival. A world model is valid in so far as it correctly describes the predominant characteristics of a given present. At the same time, each world model represents and forms part of what at any given time is perceived to be the 'real' world. We can at present distinguish between at least three successive world models. Their basic characteristics are most easily approached in terms of man's changing relationship to his environment. Each world model is different from the next. There is no necessary implication that one is intrinsically better than or superior to another. Each may be optimal or more suitable with respect to a given purpose. WORLD MODEL

1

If we use as an analogy the conditions for hitting a target, then in World Model 1 the targets are fixed. In this world what appears to be eternally valid is facts. Facts about nature and society can be handed down from generation to generation, and so can requisite skills. The world is closed. There is nothing new to be discovered. Technology and social structure remain unchanged, or the change is so slow that it is not perceptible. The relation of man to nature and man to man is not problematical. This world is known in its static and stagnant form in primitive societies. It is also known in an advanced and dynamic form. It is dynamic not with respect to technology or social structure, but with respect to individual development. This world differs in its geography from ours. It is different both in space and in time. The golden age is not long past. Heaven and hell both interpenetrate 201

EPILOGUE

and coexist with the present. What stands out as unknown is the individual fate of each human being in this world and in the afterdeath world. The ultimate state of perfection has already been achieved by some in the here-and-now and can be achieved by others. The ultimate state of perfection is not something new. It is the original state which living beings have lost and which they can regain, each for himself. Whereas in the stagnant world of this type the life and future of the individual are determined by fate, in the dynamic world of this type the locus of change and development is the individual himself. The environment has no inherent power. It can assist or obstruct him. The ultimate source of responsibility is seen to lie within each individual. While it is true that no new knowledge and understanding need to be achieved, it is still up to the individual to acquire and utilize what is available. At the same time it is realized that there is an infinite amount of knowledge that it would be possible to seek but that would be of no human use or significance. WORLD MODEL

2

From the point of view of World Model 1, World Model 2 will seem like a further fall from grace. From the point of view of World Model2 the present is seen to be inherently superior to the past, and the future inherently better than the present. However, evolutionary theory is part of the assumptions of World Model 2, and so is the concept of the struggle for survival, and both are projected into the past. In terms of our initial analogy, in World Model 2 the targets aimed at are not fixed but moving. Technology and with it social structure are no longer static but seen to be subject to change. Against this background of change, human nature appears static. Appearing static, human nature becomes opaque and unknowable and its potentialities seem no longer relevant or significant. The world is no longer closed but open. The knowledge to be gained about the environment appears unlimited. The locus of change lies in the environment. What appears problematical and significant is the properties of the environment. In contrast to the nature of human beings, the environment and specifically the physical world appear translucent and knowable. 202

THE EVOLUTION OF WORLD MODELS

The person has become split from the changing environment. He stands apart from and against it. He is no longer part of the world. He stands outside it. It is this that creates the basic contradiction in this world model. In the active mode, standing godlike outside the world, man controls, masters, and subjugates his environment. In the passive mode it is the environment that shapes, determines, and governs his behaviour. In World Model 2, the predominant relationship is not man to himself, but man to external objects. It is because of the fundamentally different characteristics of man and the physical environment that the primary split between self and environment arises. Since the physical environment is infinite in extent, the world is seen as an open system. The relationship of man to man is mediated by the physical environment. It is by means of a changing technology that social structure becomes modified and the requirement for human adjustment arises. Man here becomes the medium and facilitator of technological development. The locus of change and power lies in the physical environment. The pathologies of this model emerge when man begins to treat man as part of the physical environment. In the active mode he perceives and masters others as objects. In the passive mode he experiences himself as object, as a cog in the machinery. Since the locus of change and power is seen to lie in the environment, the environment appears potentially threatening and overwhelming. Life is seen as a struggle for survival. The geography of the world has altered. Personal existence is bounded by birth and death. It is seen as arising out of nothing at birth and returning to nothing at death. There is no personal project beyond this. The aim of personal salvation becomes meaningless. Man spends his existence seeing himself fighting against his environment for survival, convinced that there is no survival. The self appears ephemeral and unreal. What appears objective and real is the external physical universe which exists from the beginning to the end of time. The environment is seen to be in a state of change. In this world what appears eternally valid is not facts but the laws of change. Reality becomes split off and moves behind phenomena. Reality can no longer be known directly, as it can in World Model!, in which this is the essential possibility open to man

203

EPILOGUE

and the necessary basis for liberation. In World Model 2 it can be apprehended and understood only indirectly and conceptually. The function of insight and understanding is no longer a liberating one but the pragmatic one of meeting the challenge of the environment. In its passive form the world is seen to be ruled by impersonal and eternal laws of change, to which human beings have to adapt: laws of matter and mechanics, laws of historical determinism, technological development over which man has no control, economic laws, and universal and deterministic laws of human behaviour. In the active form of this world model, it is knowledge and the application of the laws of nature that make it possible to modify the environment. Human nature does not seem to be essentially changeable. The environment, technology, and social structure are seen to be modifiable and changeable. The project has become the creation of an environment which will satisfy the needs of, and provide the summum bonum for, all human beings. The aim is not individual but universal salvation; not here and now, but in the future. WORLD MODEL

3

If World Model 2 is now becoming superseded it is not because of its internal contradictions but because of a fundamental change in the relationship of man to his environment. The clearest indication of the emergence of a new world model is found in Emery & Trist's (1965) identification of a new type of environment which they refer to as turbulent. A turbulent environment is one in which directive correlations established with the environment, on which the survival of an organization depends, unexpectedly break down. Actions that are initiated become attenuated or build up uncontrollably, and goal-directed strategies can lead to the opposite of the intended result. There are indications that the turbulence observed is not a necessary characteristic of the environment but a transitional phenomenon, which will continue to occur as long as we respond to World Model 3 as if we are still living in World Model 2. At the same time, it is by understanding the conditions that lead to turbulence that we have the best chance of understanding the emerging characteristics of the present. This is no longer a matter of theoretical

204

THE EVOLUTION OF WORLD MODELS

concern but increasingly a matter of human survival. Or, to put this a little differently, it is precisely the theoretical understanding of the present that is becoming an urgent concern, for the essential feature of World Model 3 is that the principles in terms of which our world operates are no longer constant but are becoming subject to quite rapid change. In World Model 2 the predominant characteristic of the environment was that it comprised an aggregate or cluster of elements. This was the model in terms of which classical science built its theories of universal and immutable deterministic laws and modem science built its stochastic principles. These types of law no longer correspond to the predominant features of the emerging present. A basic characteristic of systems that consist of an aggregate of elements is, as Heider pointed out, that they can function as a medium (see Heider, 1959). In World Model2, the physical environment, and specifically technology, functioned as a medium for the relationship of man to man. As long as the environment with which man was concerned had the properties of an aggregate or cluster of elements it could act as a medium. It could be modified and changed without altering its essential nature. Now that the environment has become reactive and turbulent, it has lost its characteristic of a medium. It is at this stage that man, facing his environment, discovers himself. The turbulent environment is man himself. The struggle for survival turns out to have a suicidal aspect. In the emerging World Model 3, the predominant relationship is man to man. The world is again closed. It is the dense interdependent ecology of life on this globe. Humanity, in the words of Teilhard de Chardin (1961), has folded in upon itself. There is no outside to man's world. The existential loneliness of humanity is not overcome by space flights. Man takes his environment along wherever he goes. The system has taken the environment into itself. Each part has the rest of the system as its environment and each part is the environment for others. What has receded as a dominant characteristic is the independent physicalist-type environment, which could be conceptualized as an aggregate or cluster of elements and which provided the basis for the immutable laws of classical science. It is not that these laws have ceased to operate in so far as they are valid, but that they no longer correspond to the conceptual model in terms of which the 205

EPILOGUE

predominant characteristics of the present can be understood. This is because the behaviour of man, the relationship of man to man, and the social ecologies that have come into being do not conform to the universal and immutable principles of classical science. The principles in terms of which interpersonal structures and social ecologies operate are subject to change. This change occurs as the predominant conceptual and organizational model in terms of which the world is experienced and responded to is transformed. The basic characteristic of World Model 3 is that the principles in terms of which the present operates become subject to recurrent change. In World Model 2 the leading part in developmental change was played by technological development. Technological change was responded to as an environmental phenomenon over which man had no control and to which he had to adapt. The transition to the new world model is a result of technological change in the form of the increased scale of technological and biological intervention, and the increased speed of electronic communication and transport, together with increasing population density. However, in World Model 3, technological change no longer plays the leading part. Human and societal adaptation to independent and non-directed technological change has ceased to be possible. The essentially new factor is the potential for technological choice. Provided that we know what we want, we can to an increasing extent select and produce the technology required. The critical condition for gaining control over social change in industry and society as a whole becomes, then, that we utilize the option of technological choice. To the extent that the choice of future technologies is directively correlated with social and educational changes over the same period, stability and directionality of societal development can be achieved. This is a necessary but not sufficient condition. Man has become able to create his future society, not as he has done so far, blindly and unknowingly, but, within limits, consciously. What he is faced with is the problem of deciding what kind offuture he wants. The condition for survival in World Model 3 is not competition in the present but cooperation with respect to the future to be created. Only in this way can the relevant here-and-now decisions concerning technological, social, and educational changes be directively correlated and implemented.

206

THE EVOLUTION OF WORLD MODELS

In World Model 2, social organization was based largely on the socio-technical machine principles that were developed at the time of the industrial revolution. 1 Each human component within an organization was given a single narrowly prescribed function together with a set of prescribed relationships to other human components of the organization. Both work functions and work relationships had to be rigidly adhered to and unchanging. The human components that required a minimum of training could be easily replaced and could do their job without being concerned with or able to affect the aims of the enterprise. A hierarchical structure developed to control and coordinate the split-off parts. Within this world model, managers saw themselves as standing outside the organizational machine that they controlled. The basic principle adhered to was the separation of actions and decisions. With the development of new types of technology designed as integrated systems that can be operated by small staff teains, and particularly in view of the increasingly rapid rate of change in technological design, it has now become necessary to build learning capacities into the organization of industrial work teams. This can be achieved only by creating relatively autonomous matrix organizations in which neither task roles nor work relationships are fixed. As this happens, work organizations become capable of carrying out research both to find ways of improving the production process and to discover and test possible strategies for coping with changing task requirements. At the same time, research organizations are beginning to expand into experimental production projects, and universities are becoming involved in societal projects that can no longer be tackled by single specialist skills. As a result, organizations are beginning to establish new links with one another which in the course of societal development and adjustment will change over time. The traditional machine model now becomes irrelevant. What emerges at this stage is an ecological matrix in which component organizations have no single specialized function, no 1 The preceding organizational model for industrial activities was that of autonomous professions or quasi-professions. Production units consisted of a master with apprentices, or production was based on the family unit. Some industries, such as textile manufacture, were based on extended kinship structures. lnd~tries such as mining which required unskilled labour and capital investment created serious problems of organization.

207

EPILOGUE

single rigid structure, and no single centralized organizer or leading part. In World Model 3, societal systems consist of organizations that possess an almost total range of capacities, but each has a predominant role which at any given time is determined by the links established with other organizations. Thus not only individual organizations but also society as a whole develops in the direction of a matrix organization. At this stage, societies can develop learning capacities and increase their capacity for determining the direction of their development and growth. The direction of development is towards a society in which there will be relatively little difference in the educational level and status of those who work in industrial, educational, research, and service organizations. People will differ more as regards their focus of orientation than as regards the nature of their work. The leading elements in the transitional stage of development are the rapid increase and diffusion of complex technologies which can be operated by a small number of persons, and the rapid increase and diffusion of higher education. In terms of their operational requirements these will up to a point be mutually supportive. As development continues, the traditional hierarchical type of organization based on the separation of doing, planning, and deciding will become increasingly inappropriate and will be replaced by primary work groups in which these functions are integrated. The members of these groups will to an increasing degree be able to participate in policy decisions and be capable of using specialists as consultants. Whereas, in World Model 2, the basic principles in terms of which organizations and society were built remained stable, 1 in World Model 3 the principles in terms of which the world operates are subject to evolutionary change. At this stage practically all socialscience research acquires the characteristics of applied research, with time-limited validity. Increasingly, the tasks of the social sciences are: rapid diagnosis of the present, the recognition of emergent trends, and the discovery of developmental options and strategies of future development. At the same time, a clearer distinction emerges between applied 1 This appears to be true of both capitalist and communist societies. The latter, following Lenin's interpretation of Marx's theory, turn out to be essentially an attempt to apply the traditional factory model to the whole of society.

208

THE EVOLUTION OF WORLD MODELS

and theoretical behavioural science. Theoretical science will be concerned with the behavioural and social worlds that can exist, and the transitional paths between them. This defines the realm of the possible. It is pure theory in the sense that it has no implications for action, but provides the structure within which principles of applied science can be formulated. Applied science can tell us only where we are at any given time, and in what direction and by what means we can go from there, but it again provides no criteria for choice. It becomes evident that the key problems of this world model lie no longer in how to achieve, but in what to achieve. In World Model 3, values and ethics become the central concern. It is at the point where human behaviour is seen to be subject to causal developmental principles (and this has implications that go beyond World Model3) that the foundation is supplied for a science of human behaviour which is essentially a science of ethics. Before, means and methods seemed relatively immaterial and what was significant was the goal to be achieved. What is seen now is that it is the means and methods that we apply that create the future. It is in the choice of methods and means that values become operative, and it is by understanding what the effects on the future will be of actions performed today that rational action becomes possible. However, the unification of ethics and values with an understanding of the consequences of actions performed in the here-and-now is ultimately possible only at the level of individual behaviour and is not yet part of the reality of World Model 3. The individual at this stage is still a means, a contributor to the creation of a joint future, and concern is with problems of interpersonal ethics and societal values. The individual cannot as yet perceive a reality for himself outside his social role and responsibility except, as in all ages, at the inevitable moment of death, which remains for him the great mystery. There are several reasons why problems of social ethics and values become crucial at this stage: 1. Within the internal environment of social ecologies, actions tend to elicit immediate results, and therefore learning becomes not only possible but essential. 2. As soon as social, technological, and educational change processes become directively correlated, the rate at which a new social organization can be implemented can be considerably

209

EPILOGUE

speeded up. It is therefore essential that the effects of actions on the relatively immediate future are understood, in order that control can be maintained over the development process. 3. The maintenance of directive correlations in respect of the relatively autonomous units of a matrix organization presupposes a shared set of values and commitment to the creation of a joint social future. The shared set of values has to be sufficiently explicit to provide all participant members with criteria for identifying the optimal actions required of them as individuals. It is generally difficult to understand the role and significance of new developments. It is only in retrospect that one can see that the emergence of World Model 3 did not start off as a theory but in the evolution of kibbutz organizations in Israel about fifty years ago. That an evolutionary process had been started became clearer later, when a wide range of communal matrix-type organizations embodying different value orientations emerged, and in the more recent phase these began to evolve further in the direction of testing out the problem of creating larger regional units without endangering the autonomy of member units. What was demonstrated at the same time was the possibility of selecting technology and adapting it to the requirements of the type of social organization that members wished to achieve and maintain. This means that one could study archaeological reconstructions of water conservation systems employed 2000 years ago with the same interest and attention to possible relevance as the most recent technological developments. Just as in Israel it was initially assumed that the new type of organization was relevant only for agricultural settlements, so the work groups based on an autonomous matrix organization that were discovered in British coalmines during the early stages of the development of socio-technical theory (Trist et al., 1963; Herbst, 1962) were judged initially to be relevant only to the special conditions ofthe mining industry. The significance of the more recent Industrial Democracy Project in Norway (Thorsrud & Emery, 1969) lies in showing not only that matrix-type organizations are viable, but also that they are particularly appropriate to the requirements of modern technology, provided, however, as far as future development is concerned, that technological and social change become directively correlated.

210

THE EVOLUTION OF WOR.LD MODELS

Significant social changes, when they do occur, generally happen spontaneously and almost as a matter of course, with those who are involved in them not necessarily being aware of more than their local implications. The emergence of World Model 3 coincides with a significant development within the framework of World Model2 brought about by the transition to technologies based on the transformation of information. To its adherents it seems that the basic aim- complete, rational, machine control which requires no human participation or intervention- is now achievable. In this final phase of World Model2 the predominant relationship is not man to object but man to data. Given that computers can unravel the structures inherent in data assemblies, the world appears in principle fully predictable and controllable to those who, as before, perceive themselves as standing godlike outside the system. However, it is precisely when the ego is on the point of becoming exclusively dominant that the unreality of its content may be perceived. The world now consists of data, but nothing real besides them to which they refer. Neither mind nor matter is any longer part of the reality of this world. Materialism has been turned inside out, and it is in grasping that there is nothing substantial to be grasped that the final phase of World Model 2 points to the future emergence of World Model 4. During the transitional phase of development the possibility exists of a fixation, or a regression to World Model 2. It is possible that godlike mysterious power will be projected upon computer programs, to which effective decision-making authority will be transferred. In this way the emerging computer and media technology may be used to re-create a substitute external environment to which society has to adapt, and this may for a time permit the survival of centralized hierarchical organizations.

211

CHAPTER 15

The Product of Work is People 1

---·--The assertion that the product of work is people is evidently either false or a gross oversimplification;for one cannot entirely disregard that, as a result of work, we also produce cars, colour TV, and baby napkins, not to speak ofpapers for conferences, all of which may have some value as well.

The significance of what we do is not always evident while we are doing it. As children, most of us were absorbed in various games either by ourselves or with others. At the time we were intensely involved in what we were doing, which we took to be an aim in itself. It is only in retrospect that we are able to recognize that we were then developing qualities that were needed to help us to make the passage into adulthood. The same can be said of the tasks we engage in during the adult phase of our life. At that time our concern is almost entirely with our relations to the outside world, and we become for a time almost oblivious of ourselves. What absorbs us then is our career, our family, and whatever we do in the way of changing and modifying the nature of our environment. The nature of the outer world is seen as the source of the various forms of sensuous enjoyment, of our happiness and our distress, and our intellectual, social, and physical skills are developed in relation to our environment and become centred on understanding it and finding ways of controlling it. At this stage our success or failure appears to be almost objectively measurable by ourselves and also by others in terms of how much we have achieved. 1 Paper contributed to a conference on 'The Quality of Working Life', held at Harriman, N.Y., in 1972.

212

THE PRODUCT OF WORK IS PEOPLE

It is as we begin to enter the post-adult period that something like a figure-ground reversal may take place in our values and perceptions. That which appeared to be the product begins to be seen almost as a byproduct. We may then become aware that it is not, after all, how much or how little we achieved that matters, but how we went about achieving it. That which we have achieved is always in the past, and as soon as it is achieved it is no longer truly related to us. However, that which we have become-our qualities as human beings and our potential for future development -is always in the here and now. After everything else is gone, that which we have become remains, whether we recognize this or not, as the product. The products of our adult life are rather like a shell that protects the growth of the fruit that is inside. As soon as one realizes that the shell that one has produced is not that which is of essential value, one may discover and make use of what one finds in oneself. However, one can use one's work and one's inner strength to produce no more than an outer shell of success, possessions and pretensions, and then towards the end of one's career one may quite literally suddenly experience oneself as an empty, burnt-out, and rigid shell, and find that one has gained little of value. The period of adulthood is a phase in our life when we may become almost completely absorbed by our concern with our environment, and to this extent deeply self-alienated. We are then not yet able to develop our wisdom, so we do not yet really know what is of value and what is not, what is success and what is failure. During the early part of one's adult life, choosing what to set out to achieve appears to be the critical decision. One may wish to accumulate a sizeable amount of money or just enough to buy a motor cycle; or one may wish to win a Nobel prize or at least to become a university professor. Alternatively, one may feel that if one's aim is to help others or to create a better kind of society, then this is a superior choice; but it is possible to be just as selfish and obtuse in the way we seek to help others as in any other kind of pursuit. The significant choices that we make are not simply choices of goals. They are more often less obvious choices of the means we adopt, at any point in our career, to achieve our aims. It is choices of means that at any time express our human qualities and determine what qualities we create and develop in ourselves. We make such choices sometimes consciously, sometimes intuitively, and 213

EPILOGUE

often without realizing that we have made them. And frequently we do not at the time fully understand the implications of our choice. Thus those working as scientists may in their work be able to develop qualities such as non-attachment to theories and beliefs; the capacity to accept that one does not know, or what Nicholas of Cusa called learned ignorance (see Heron, 1954); the capacity for sustained concentration on a problem; and the capacity to refrain from biasing one's facts in favour of a view that one likes or against a view that one dislikes. Qualities of this kind, whatever kind of work one does, become essential during the post-adult period when, having developed oneself in relation to the environment, one turns back to gain, to the extent to which one is able, understanding and control of oneself. However, it is also qualities of this kind that, in the course of one's career, enable one to work more effectively. Whatever success one then achieves becomes a source of satisfaction in so far as it confirms that one has developed qualities that one did not have before. However, if this is not the case, then whatever success is achieved will not give one the experience that one has achieved something of value. The relevance of human qualities to the society in which we live is more easily seen by going back in time. Up to the end of the Middle Ages, published scientific data were not only often of doubtful accuracy but also occasionally altered to maintain consistency with accepted theory. This at the time made it practically impossible for Kepler to develop and test his theory of planetary orbits until he was able to work with Tycho Brache, whose observational accuracy and personal integrity he could trust. Moreover, at the time even distinguished scientists could not be trusted not to steal discoveries from their colleagues, so that it was not unusual for findings to be written in cryptograms or secret languages. The same conditions that inhibited the development of science as a social enterprise also inhibited the national and international development of trade and industry. If Europe produced something distinctive at the time it was not simply the development of industrial organizations and science, but the human qualities and values of the Protestant ethic which, from the time of the Reformation, were built into the institutions that we have inherited, and which both in the Western world and in the Eastern European countries continue to make these developments possible. 214

THE PRODUCT OF WORK IS PEOPLE

Institutions and organizational forms that survive over many centuries can become objects of veneration because they enshrine people's values, although what these values are may no longer be known. Moreover, institutional ideals can remain visible and more or less unchanged even when the value base has shifted or reversed itself. So people may retain their criteria as to what is a good civil servant or a good worker although they may not be clear why some qualities are good and others are not good, or they may change their view of what they are good or not good for. Thus while the Protestant ethic justified itself through achievements in industrial, economic, and scientific development, there is scarcely anybody today who knows what this was. At the same time, there was already, from the beginning, the expectation that what were initially the elite values of the emerging middle class would, as the general standard of living and education increased, eventually apply to the working class as well. Even at the end of the nineteenth century, however, it was not yet predictable that this would occur. As the value basis was reversed, industrial, economic, social, and scientific progress became aims in themselves, or it was taken for granted that they would more or less automatically ensure human contentment and happiness. For some time the ideal society seemed just round the corner. There is no necessary gulf between the older and the younger generations as far as these expectations are concerned, but more a feeling of mutual frustration due to an out-ofphase development. The rising standard of living has led to an increased level of satisfaction and also to increasing discontent and human unhappiness. However, while the increased level of satisfaction can be cancelled out by increased expectations, discontent and unhappiness remain. Furthermore, to the extent that experienced satisfaction is cancelled out it can no longer be used to compensate for the experience of distress. · For the older generation, the progress achieved since the end of the second world war has gone so much beyond what was thought possible at the time that distress tends to be well compensated for by the experience of satisfaction with what has been achieved. The group that is at present most exposed to distress is the younger middle-class generation for which no achievable satisfaction provides sufficient compensation. The gain achieved by adding further cars

215 p

EPILOGUE

or swimming pools to the family property is little more than marginal at best, and there is a feeling of surfeit. Consequently, while the one side is predominantly concerned with increasing the level of satisfaction the other side is concerned with decreasing distress. It is because these are two quite different things that the two generations have come to be out of phase. 1 In a beautiful paper, Schmidt (1970) suggests that both sides may be right in their views about the future, and in this he may be right as well. However, it is also possible that in a significant way both sides may be wrong, and they may be wrong for the same reason. There is no question that economic, industrial, scientific, and technological development are eminently worth-while goals. They are aims that can be achieved by societal change and they can and do increase human satisfaction and comfort. The first walk on the moon created a considerable amount of enthusiasm, the motor car may be a convenient form of transportation, and the modern supermarket may be a better place for shopping than the old corner grocery store (or it may not be). But whether so or not, none of these achievements has any enduring effect on people's experience of distress or happiness. If something of enduring value has resulted, then this lies in the exploration and cultivation of previously unknown modes of human potentiality and in the conditions provided for the development of human qualities, intellectual, social, and moral - in effect, in the now almost forgotten initial purposes of the original reformers. If we speak today of the developed part of the world, what we find in the Western world is a high level of development in terms of what provides satisfaction but not in terms of understanding how to overcome distress. Satisfaction derives from whatever can be obtained from the environment, and our society is geared to achieve this. However, contentment, peace of mind, happiness, and wisdom- these are not obtainable from the environment in the same way, but are found in oneself, and to achieve them involves a different form of work. They are not obtainable from the environment, but they determine how we are able to relate ourselves to our environment. 1 It has been argued that there is always a generation gulf. This view, however, overlooks the fact that at the time of the German National Socialist revolution the two generations were in phase in their search for a way of overcoming distress; again, there was no basic conflict in Western Europe immediately before the first world war.

216

THE PRODUCT OF WORK IS PEOPLE

Whenever we encounter what appears to be an eitherfor situation, we need to investigate whether we have not constructed for ourselves a false dilemma. The production of food and shelter and other basic necessities of life is essential for human existence. However, if this becomes an all-absorbing and single-minded social pursuit to which people sacrifice, or are forced to sacrifice, their life, then they run the risk of ending up by losing their capacity to determine the direction of their own life and can no longer find what is worth while. On the other hand, societies that become totally committed to aims beyond the satisfaction of material needs may not survive long enough to enable people to achieve anything of value. A dilemma emerges at the point when it is assumed that a choice has to be made.

217

BIBLIOGRAPHY

Towards an Idealised University. Management Science 15, 121-33. ALLESCH, G. J. VON (1931). Zur Nichteuklidischen Struktur des phaenomenalen Raumes. Jena: Fischer. AMBER, G. H. & AMBER, P. S. (1962). Anatomy of Automation. Englewood Cliffs, N.J.: Prentice-Hall. ANGY AL, A. (1941). Foundations for a Science of Personality. Cambridge, Mass.: Harvard University Press. ANGYAL, A. (1965). Neurosis and Treatment. New York: Wiley. ARNER, o. (1961). Skipet og Sj0mannen. Oslo: Institute for Social Research. (Mimeo.) ARNER, o. &HERSSON, L. (1964). Sjemannsyrkets Saerpreg. Tidsskrift for den Norske Legeforening 84, 1223-8. (English summary.) ASHBY, w. R. (1952). Design for a Brain. New York: Wiley. ASHBY, w. R. (1962). Principles of the Self-organizing System. In H. V. Foerster & G. W. Zopf (eds.), Principles ofSelf-organization. Oxford: Pergamon. ATKIN, R. H. (1972). From Cohomology in Physics to q-Connecitivity in Social Science. International Journal of Man-Machine Studies 4, 139-67. AUBERT, v. (1965). The Hidden Society. Totowa, N.J.: Bedminster Press. AUBERT, V. & ARNER, 0. (1959). On the Social Structure of the Ship. Acta Sociologica 3, 200-19. A UBER T, V. & ARNER, 0. (1962). The Ship as a Social System. Oslo: Institute for Social Research. (Mimeo.) ACKOFF, R. L. (1968).

BALDAMUS, w. (1961). Efficiency and Effort. London: Tavistock. BARISH, N. N. (1951). Systems Analysis for Effective Administration.

New York: Funk & Wagnalls. Work Methods Manual. London: Chapman & Hall. BARNES, R. M. (1949). Motion and Time Study. New York: Wiley; London: Chapman & Hall, 1951. BARNES, R. M. (1944).

219

BIBLIOGRAPHY

The Role of the Entrepreneur in Social Change in Northern Norway. Bergen and Oslo: Norwegian Universities Press. BARTH, F. (1966). Models of Social Organization. Occasional Paper No. 23. London: Royal Anthropological Institute. BA VELAS, A. (1948). A Mathematical Model for Group Structures. Applied Anthropology t, 16-30. BELL, D. A. (1960). The End of Ideology. New York: Free Press. BENNIS, w. G. {1966). Changing Organizations. New York: McGrawHill. BERNAL, 1. {1939). The Social Functions of Science. London: Routledge. BERTALANFFY, L. VON (1950). The Theory of Open Systems in Physics and Biology. Science 3, 25-9. BERTALANFFY, L. VON (1951). Theoretical Models in Biology and Psychology. Journal of Personality 20, 24-38. BERTALANFFY, L. VON (1968). General System Theory. New York: Braziller. BEURLE, R. L. (1962). Functional Organization in Random Networks. In H. V. Foerster & G. W. Zopf (eds.), Principles of Selforganization. Oxford: Pergamon. BION, w. R. (1961). Experiences in Groups, and Other Papers. London: Tavistock; New York: Basic Books. BLAU, P. (1955). The Dynamics of Bureaucracy: A Study of Interpersonal Relations in Two Government Agencies. Chicago: University of Chicago Press. BLA UNER, R. (1964). Alienation and Freedom. Chicago: University of Chicago Press. BLICHFELDT, 1. F. {1973). Bakerst i K/assen. Oslo: Nordli. BOGUSLAW, w. {1965). The New Utopians. Englewood Cliffs, N.J.: Prentice-Hall. BORN, M. (1964). A Statistical Interpretation of Quantum Mechanics. In Nobel Lectures in Physics. Amsterdam: Elsevier. BORN, M. (1971). The Born-Einstein Letters. London: Macmillan. BOTT, E. (1957). Family and Social Network. London: Tavistock. Second edition, 1971. BOULDING, K. E. {1964). A Pure Theory of Conflict Applied to Organizations. In G. Fisk (ed.), The Frontiers of Management Psychology. New York: Harper. BOX, G. E. P. (1957). Evolutionary Operation. Applied Statistics 6 (1). BRAY, J. R. (1958). Notes toward an Ecology Theory. Ecology 9, 770-6. BR ENT ANO, F. (1874). Psychologie vom empirischen Standpunkte. Leipzig: Dunker und Humbolt. BAR TH, F. (1963).

220

BIBLIOGRAPHY BRIDGEMAN, P. w. (1931). Dimensional Analysis. London: Oxford University Press; New Haven, Conn.: Yale University Press. BRUN-GULBRANDSEN, S. & IRGENS-JENSEN, 0. (1964). Alkoholmisbruk blant Unge Norske Sjemenn. Oslo: Universitetsforlaget. BUCK Low, M. (1966). A New Role for the Work group. Administrative Science Quarterly 11, 59-78. BURNS, T. & STALKER, G. M. (1961). The Management of Innovation. London: Tavistock. Second edition, 1966. CHARDIN, P. TEILHARD DE (1961). The Phenomenon of Man. London: Collins. CHEIN, I. (1954). The Environment as a Determinant of Behavior. Journal of Social Psychology 39, 115-27. CHEIN, I. (1972). The Science of Behavior and the Image of Man. New York: Basic Books; London: Tavistock. CHURCHMAN, C. W. & EMERY, F. E. (1966). On Various Approaches to the Study of Organizations. In J. R. Lawrence (ed.), Operational Research and the Social Sciences. London: Tavistock. COCH, L. & FRENCH, J. R. P. (1948). Overcoming Resistance to Change. Human Relations 1, 512-32. COLEMAN, J. S., KATZ, E. & MENZEL, H. (1966). Medical Innovation: A Diffusion Study. Indianapolis: Bobbs-Merrill. CONNELL Y, J. R. (1943). Technique of Production Processes. New York: McGraw-Hill. COOLIDGE, J. L. (1963). A History of Geometrical Methods. New York: Dover Publications. COOPER, R. (1972). Man, Task and Technology. Human Relations 25, 131-57. CRAIK, H. K. (1970). Environmental Psychology. In T. M. Newcomb (ed.), New Directions in Psychology, No. 4. New York: Holt, Rinehart & Winston. CURLE, A. (1947). Transitional Communities and Social Reconnection: A Follow-up Study of the Civil Resettlement of British Prisoners of War, Part I. Human Relations 1, 42-68. CURLE, A. (1949). Incentives to Work: An Anthropological Approach. Human Relations 2, 41-7. DA VIS, L. E. (1957). Towards a Theory of Job Design. Journal of Industrial Engineering, Sept., 19-23. DAVIS, L. E. (1962). The Effects of Automation on Job Design. Industrial Relations 2 (1), 53-72. DAVIS, L. E. (1966). The Design of Jobs. Industrial Relations 6, 21-45. 221

BIBLIOGRAPHY

& CANTER, R. R. (1955). Job Design. Journal ofIndustrial Engineering, Jan., 1-10. DA VIS, L. E. & CANTER, R. R. (1956). Job Design Research. Journal of Industrial Engineering, Nov., 11-18. DA VIS, L. E., CANTER, R. R. & HOFFMAN, J. (1955). Current Job Design Criteria. Journal of Industrial Engineering, March, 5-10. DA VIS, L. E. & ENGELSTAD, P. H. (1966). Unit Operations in Sociotechnical Systems: Analysis and Design. Doe. 894. London: Tavistock Institute of Human Relations. DA VIS, L. E. & TA YLOR, J. C. (eds.) (1972). Design ofJobs. Harmondsworth: Penguin. DENBIGH, K. G. (1951). The Thermodynamics of the Steady State. London: Methuen; New York: Wiley. DICKSON, D. E. {1936). The Morbid Miner. Edinburgh Medical Journal, p. 696. DURKHEIM, E. (1966). The Division of Labor in Society. New York: Free Press. DA VIS, L. E.

(1963). A Note on the Ceiling of Psychophysical Power Functions. Rep. Psycho/. Lab. No. 149, University of Stockholm. EMERY, F. E. (1959). Characteristics of Socio-technical Systems. Doe. 527. London: Tavistock Institute of Human Relations. EMERY, F. E. (1966). Report on a Theoretical Study of Unit Operations. Doe. 900. London: Tavistock Institute of Human Relations. EMERY, F. E. (1967). The Next Thirty Years: Concepts, Methods, and Anticipations. Human Relations 20, 199-237. EMERY, F. E. (ed.) (1969). Systems Thinking. Harmondsworth: Penguin. EMERY, F. E. & OESER, 0. A. (1958). Information, Decision and Action. Cambridge: Cambridge University Press. EMERY, F. E. & THORSRUD, E. (1964). A New Look at Industrial Democracy. XVth International Congress on Applied Psychology, Ljubljana. EMERY, F. E. & THORSRUD, E. (1969). Form and Content in Industrial Democracy. London: Tavistock. EMERY, F. E. & TRIST, E. L. (1960). Socio-technical Systems. In C. W. Churchman & M. Verhulst (eds.), Management Sciences: Models and Techniques. Oxford: Pergamon. EMERY, F. E. & TRIST, E. L. (1965). The Causal Texture of0rganizationa1 Environments. Human Relations 18, 21-32. ENGELSTAD, P. H. (1969). Socio-technical Approach to Problems of Process Control. Doe. 26. Oslo: Work Research Institutes. 222 EISLER, H.

BIBLIOGRAPHY ENGELSTAD, P. H. (1970). Tekno/ogi og Sosial Forandring pd Arbeidsplassen. Oslo: Tanum. FAULKNER, T. E. (1952). Projective Geometry. Edinburgh: Oliver & Boyd; New York: Interscience Publishers. FECHNER, G. T. (1860). E/emente der Psychophysik. Leipzig. FIEDLER, F. E. (1953). The Psychological Distance Dimension in Interpersonal Relations. Journal of Personality 22, 142-50. FLAMENT, c. (1963). Applications of Graph Theory to Group Structure. Englewood Cliffs, N.J.: Prentice-Hall. FOGEL, L. J. et al. (1966). Intelligent Decision Making through a Simulation of Evolution. Behavioral Science 11 (4). FORD, H. (1968). Mass Production. InC. R. Walker(ed.), Technology, Industry and Man. New York: McGraw-Hill. FOSTER, P. M. & GORMAN, L. (1965). A Pilot Study of Factors underlying Fleet Turnover in Irish Shipping Ltd. London: Tavistock Institute of Human Relations. FRANK, F. (1953). Untersuchungen iiber den Zusammenbruch von Feldmausplagen (Microtus arva/is). Zoo/. Jahrb. 82, 95-136. FRIEDMANN, G. (1955). Industrial Society. Glencoe, Ill.: Free Press. FRIEDMANN, G. (1961). The Anatomy of Work. New York: Free Press; London: Heinemann. GEELMUYDEN, H. (no date). Apent Hav. Oslo: Aschehoug. GLUCKMAN, M. (ed.) (1962). Essays on the Ritual of Social Relations. Manchester: Manchester University Press. GOFFMAN, E. (1956). The Presentation of Self in Everyday Life. Edinburgh: University of Edinburgh, SSRC Monog. 2; New York: Doubleday (Anchor), 1959. GOLDSMITH, M. & MACKA Y, A. (eds.) (1964). The Science of Science: Society in the Technological Age. London: Souvenir Press. GOMBERG, w. (1956). The Use of Psychology in Industry: A Trade Union Point of View. Management Science 2. GOULDNER, A. w. (1954). Patterns of Industrial Bureaucracy. Glencoe, TII.: Free Press; London: Routledge, 1955. GRoss, E. (1953). Some Functional Consequences of Primary Controls in Formal Work Organizations. American Sociological Review 18, 368-73. GROSS, E. (1956). Symbiosis and Consensus as Integrative Factors in Small Groups. American Sociological Review 21, 174-9. GUETZKOW, H. (1957). Differentiation of Roles in Task-oriented Groups. Carnegie Institute of Technology, Graduate School of Industrial Administration. 223

BIBLIOGRAPHY

& DILL, W. R. (1957). Factors in the Organizational Development of Task-oriented Groups. Sociometry 20, 175-204. GULOWSEN, J. (1971). Selvstyrte Arbeidsgrupper. Oslo: Tanum. GUSTAVSEN, B. (1969). Industristyret. Oslo: Tanum. GUETZKOW, H.

(1948). Psychosocial Medicine: A Study of the Sick Society. New York: Norton; London: Heinemann, 1949. HEIDER, F. (1946). Attitudes and Cognitive Organization. Journal of Psychology 21, 107-12. HElD ER, F. (1959, originally 1926). Thing and Medium. Psychological Issues Monograph 3(1}, 1-34. HERBST, P. G. (1952). The Measurement of Family Relationships. Human Relations 5, 3-35. HERBST, P. G. (1954a). Family Living: Regions and Pathways. In 0. A. Oeser & S. B. Hammond (eds.), Social Structure and Personality in a City. London: Routledge; New York: Macmillan. HERBST, P. G. (1954b). The Analysis of Social Flow Systems. Human Relations 1, 327-36. HERBST, P. G. (1957a). Situation Dynamics and the Theory of Behavior Systems. Behavioral Science 2, 13-28. HERBST, P. G. (1957b). Measurement of Behaviour Structures by means of Input-Output Data. Human Relations 10, 335-46. HERBST, P. G. (1961). A Theory of Simple Behaviour Systems, I & 11. Human Relations 14, 71-94 and 193-240. HERBST, P. G. (1962). Autonomous Group Functioning: An Exploration in Behaviour Theory and Measurement. London: Tavistock. HERBST, P. G. {1967). Postulates for a Generalized Behaviour Theory. Human Relations 20, 65-82. HERBST, P. G. (1970). Behavioural Worlds: The Study of Single Cases. London: Tavistock. HERBST, P. G. (1972). Behavioural Dimension Analysis. Scandinavian Journal of Psychology 13 {1), 1-4. HERON, G. {1954). Of Learned Ignorance. Translation of De Docta Ignorantia (1440) by Nicho1as of Cusa. London: Routledge. HIGGIN, G. & JESSOP, N. (1965). Communications in the Building Industry. London: Tavistock. HIPPEL, A. R. voN (ed.) (1965). The Molecular Designing of Materials and Devices. Cambridge, Mass.: MIT Press. HIRSCHMAN, A. 0. & LINDBLOM, C. E. (1962). Economic Development, Research and Development, Policy Making: Some Converging Views. Behavioral Science 7, 211-22. HOMANS, G. c. (1946). The Small Warship. American Sociological Review ll, 294-312. 224 HALLIDA Y, J. L.

BIBLIOGRAPHY

c. (1950). The Human Group. New York: Harcourt; London: Routledge, 1951. HORSFALL, A. B. & ARENSBERG, C. M. (1949). Teamwork and Productivity in a Shoe Factory. Human Organization 8, 13-25. HUNTER, G. H. & KITTRELL, J. R. (1966). Evolutionary Operations: A Review. Technometrics 8, 389-97. HUTTE, H. A. (1949). Experiences in Studying Social Psychological Structures in Industry. Human Relations 2(2). HOMANS, G.

c. (1960). Units, Dimension and Dimensionless Numbers. New York: McGraw-Hill Paperbacks.

IPSEN, D.

(1951). The Changing Culture of a Factory. London: Tavistock/Routledge. JAQUES, E. (1956). The Measurement of Responsibility. London: Tavistock. JASINSKI, F. J. (1956). Technological Delimitation of Reciprocal Relationships: A Study oflnteraction Patterns in Industry. Human Organization 15(2), 24-8. JAY, A. (1970). Management and Machiavelli. Harmondsworth: Penguin. JOHNSEN, E. (1968). Studies in Multiobjective Decision Models. Lund: Student Litteratur. JORDAN, N. (1963). Allocation of Functions between Man and Machines in Automated Systems. Journal of Applied Psychology 47(3), 161-5. JORDAN, N. (1968). Themes in Speculative Psychology. London: Tavistock. JUNGE, K. (1966). Some Problems of Measurement in Psychophysics. Oslo: Universitetsforlaget. JAQUES, E.

& KAHN, R. L. (1966). The Social Psychology of Organizations. New York: Wiley. KA TZ, L. (1952). The Distribution of the Number of Isolates in a Social Group. Annals of Mathematical Statistics 23, 271-6. KING, s. D. M. (1964). Training within the Organization. London: Tavistock. KATZ, D.

& LORSCH, J. W. (1967). Organization and Environment. Boston, Mass.: Division of Research, Graduate School of Business Administration, Harvard University. LEWIN, K. (1935). Dynamic Theory of Personality. New York: McGraw-Hill. 225 LAWRENCE, P. R.

BIBLIOGRAPHY LEWIN, K.

(1947). Frontiers in Group Dynamics. Human Relations

1, 2-38. LEWIN, K. (1948). Resolving Social Conflict. New York: Harper. LEWIN, K. (1951). FieldTheoryinSocial Science. New York: Harper;

London: Tavistock, 1952. (1939). Patterns of Aggressive Behavior in Experimentally Created 'Social Climates'. Journal of Social Psychology 10, 271-99. LIKERT, R. (1961). New Patterns of Management. New York: McGraw-Hill. LIPPITT, R. (1940). An Experimental Study of the effect of Authoritarian and Democratic Group Atmospheres. University of Iowa Studies in Child Welfare, 16. LOMBARD, G. F. F. (1955). Behavior ;n a Se/ling Group: A Case Study of Interpersonal Relations in a Department Store. Boston, Mass.: Graduate School of Business Administration, Harvard University. LUCE, R. D. (1959). On the Possible Psycho-physical Laws. Psychological Review 66, 81-95. LYSGAARD, s. (1967). Arbeiderkollektivet. Oslo: Universitetsforlaget. LEWIN, K., LIPPITT, R. & WHITE, R.

(1960). The Human Side of Enterprise. New York: McGraw-Hill. MANN, F. c. & HOFFMAN, L. R. (1960). Automation and the Worker. New York: Holt, Rinehart & Winston. MANNHEIM, B. F. & ROSENSTEIN, E. (1967). The Israeli Seamen: I, Social and Occupational Characteristics and Attitudes; II, Social, Occupational and Attitudinal Characteristics related to Occupational Behavior and Commitment. Haifa: Technion-Israel Institute of Technology. (Mimeo.) MARCH, J. o. (1955). Group Autonomy and Internal Group Control. Social Forces 33, 322-6. MARCUSE, H. (1955). Eros and Civilization. Boston: Beacon Press; London: Routledge, 1956. MARCUSE, H. (1964). One Dimensional Man. Boston: Beacon Press; London: Routledge. MAREK, J. (1962). Effects of Automation in an Actual Control Work Situation. Doe. 669. London: Tavistock Institute of Human Relations. MAREK, J. (1967). Technological Development, Organization and Interpersonal Relations. Acta Sociologica 10 (~). 224-57. 226 MCGREGOR, D.

BIBLIOGRAPHY

Wire Mill of the Christiania Spigerverk. Trondheim: Institute for Industrial Social Research. (Mimeo.) MAYNARD, H. B. & STEGEMBRTBN, G. J. (1939). Operation A1111lysis. New York: McGraw-Hill. MAYO, E. (1933). The Human Problems of an Industrial Civilization. New York: Macmillan. MAYO, E. (1945). The Social Problems of an Industrial Civilization. Cambridge, Mass.: Harvard University Press; London: Routledge. MAYO, E. & LOMBARD, G. F. (1944). Teamwork and Labor Turnover in the Aircraft Industry of Southern California. Boston. MENZIES, 1. E. P. (1960). A Case-study in the Functioning of Social Systems as a Defence against Anxiety. Human Relations 13, 95-121. MER TON, R. K. (1957). Social Theory and Social Structure. Revised and enlarged edition. Glencoe, Ill.: Free Press. MILES, M. B. (ed.) (1964). Innovation in Education. New York: Teachers College, Columbia University. MILLER, E. J. (1959). Technology, Territory, and Time: The Internal Differentiation of Complex Production Systems. Human Relations 12,243-72. MILLER, E. J. & RICE, A. K. (1967). Systems of Organization. London: Tavistock. MOORB, F. G. (1951). Production Control. New York: McGrawHill. MOREBY, D. (1969). The Costs, Causes and Effects of Labour Turnover in the British Merchant Navy. Plymouth. (Mimeo.) MO RENO, J. L. (1934). Who Shall Survive? Washington, D.C.: Nervous and Mental Disease Publ. Co. MORRIS, J. N. (1947). Coal Miners. Lancet, 2, 341. MORSE, N. c. (1953). Satisfaction in the White-collar Job. Ann Arbor: Survey Research Center, University of Michigan. MORSE, N. C. & REIMER, E. (1956). Experimental Change of a Major Organizational Variable. Jour1111l ofAbnormal and Social Psychology 51, 120--9. MULDER, M. (1958). Group Structure, Motivation and Group Performance. The Hague & Paris: Mouton. Second edition, 1963. MULDER, M. & WILKE, H. (1970). Participation and Power Equalization. Organizational Behaviour and Human Performance 5, 430--48. MUMFORD, L. (1967). The Myth of the Machine. London: Seeker & Warburg; New York: Harcourt. MUNDEL, M. E. (1950). Motion and Time Study: Principles and Practice. Englewood Cliffs, N.J.: Prentice-Hall. 227 MARBK, J., LANGB, K. & BNGBLSTAD, P. (1964).

BIBLIOGRAPHY NADEL, s. F. (1951).

The Foundations ofSocial Anthropology. London:

Cohen & West. o'DONNELL, P. R. (1952).

Production Control. Englewood Cliffs,

N.J.: Prentice-Hall. Development and Stability of Scientific and Engineering Interests and their translation into Career Decisions. Does. T 548, 567, 795. London: Tavistock Institute of Human Relations. PASK, G. (1962). The Simulation of Learning and Decision-making Behavior. In C. A. Muses (ed.), Aspects of the Theory of Artificial Intelligence. New York: Plenum Press. PEDERSEN-KRAG, G. (1951). A Psychoanalytic Approach to Mass Production. Psychoanalytic Quarterly 20, 434-51. PELZ, D. C. & ANDREWS, F. M. (1966). Scientists in Organizations. New York: Wiley. PEPPER, s. c. (1961). World Hypotheses. Berkeley: University of California Press. PRICE, D. 1. DE SOLLA (1962). Science since Babylon. New Haven, Conn.: Yale University Press. PUGH, P. s. (1966). Modern Organization Theory. Psychological Bulletin 66 (4), 235-51. PAASCHE, T. (1965, 1966).

Oscillations and Sociotherapy. Human Relations 9, 357-74. RICE, A. K. (1951). The Use of Unrecognized Cultural Mechanisms in an Expanding Machine-shop (The Glacier Project, Ill). Human Relations 4, 143-60. RICE, A. K. (1958). Productivity and Social Organization: The Ahmedabad Experiment. London: Tavistock. RICE, A. K. (1970). The Modern University. London: Tavistock. RICE, A. K. & TRIST, E. L. (1952). Institutional and Sub-institutional Determinants of Change in Labour Turnover (The Glacier Project, VIII). Human Relations 5, 347-72. ROBY, T. N. & LAN ZETT A, 1. T. (1958). Considerations in the Analysis of Group Tasks. Psychological Bulletin 55, 88-101. RODGERS, w. H. (1950). Process Charting Techniques in Methods Engineering. Motion Economy 1 (4), 12-18. ROETHLISBERGER, F. & DICKSON, W. 1. (1939). Management and the Worker. Cambridge, Mass.: Harvard University Press. 228 RAPOPORT, R. N. (1956).

BIBLIOGRAPHY

{1968). Survey of Interview Data collected on the Hoegh Aurore and Hoegh Merchant. Groningen: Institute for Social Psychology. (Mimeo.) ROGGEMA, J. (1969). De Schoolse School. Assen: Van Gorcum. ROGGEMA, J. (1971). Organizational Consequences of EO Operation. Doe. 20. Oslo: Work Research Institutes. ROGGEMA, J. & THORSRUD, E. (1974). Et Skip i Utvikling. Oslo: Tanum. ROOS, P. D. {1968). Jurisdiction: An Ecological Concept. Human Relations 21, 75-83. ROGGEMA, J.

w. H. {1970). Organizational Frontiers and Human Values. Belmont, Calif.: Wadsworth. SCHRODINGER, E. (1958). Mind and Matter. Cambridge: Cambridge University Press. SEASHORE, S. E. & BOWERS, D. G. {1963). Changing the Structure and Functioning of an Organization: Report of a Field Experiment. Monograph 33. Ann Arbor: Survey Research Center, University of Michigan. SELZNICK, P. (1948). Foundations of the Theory of Organization. American Sociological Review 13, 25-35. SELZNICK, P. (1957). Leadership in Administration. Evanston, Ill.: Row, Peterson. SHANNON, C. E. & WEAVER, W. (1949). A Mathematical Theory of Communication. Urbana, Ill.: University of Illinois Press. SIMON, H. A. (1954). Some Strategic Considerations in the Construction of Social Science Models. In P. F. Lazarsfeld (ed.), Mathematical Thinking in the Social Sciences. Glencoe, Ill.: Free Press. SINGER, c. et al. (1956). A History ofTechnology, Vol. 2. New York: Oxford University Press. SMITH, M. & CULPIN, M. (1924). General Psychological Problems Confronting an Investigator. Industrial Fatigue Research Board, 4th Annual Report. SOMMERHOFF, G. (1950). Analytical Biology. London: Oxford University Press. SPECTOR, A. J. (1956). Expectations, Fulfilment and Morale. Journal of Abnormal and Social Psychology 52, 51-6. STARBUCK, w. H. (ed.) (1971). Organizational Growth and Development. Harmondsworth: Penguin. STEVENS, s. s. (1946). On the Theory of Scales of Measurement. Science 103, 677-80. 229 SCHMIDT,

BIBLIOGRAPHY

s. s. (1957). On the Psychophysical Law. Psychological Review 64, 153-81. STINCHCOMBE, A. L. (1959). Bureaucratic and Craft Administration of Production. Administrative Science Quarterly 4, 168-87. STRAUSS, G. (1954). The Set-up Man: A Case Study of Organizational Change. Human Organization 13, 17-25. SUSMAN, G. I. (1972a). Automation, Alienation and Work Group Autonomy. Human Relations 25, 171-80. SUSMAN, G. I. (1972b). Process Design, Automation and Worker Alienation. Industrial Relations 11 (1), 34-45.

STEVENS,

w. (1911). The Principles of Scientific Management. New York: Harper. THOMPSON, J. B. & BATES, F. L. (1957). Technology: Organization and Administration. Administrative Science Quarterly 2, 325-42. THOMPSON, J. D. (ed.) (1966). Approaches to Organizational Design. Pittsburgh: University of Pittsburgh Press. THORSRUD, E. (1970). A Strategy for Research and Social Change in Industry: A Report on the Industrial Democracy Project in Norway. Social Science Information 9, 65-90. THORSRUD, E. (1972). Policy-making as a Learning Process. In A. B. Chems, R. Sinclair & W. I. Jenkins (eds.), Social Science and Government. London: Tavistock. THORSRUD, E. & EMERY, F. E. (1964). Industrie[t Demokrati. Oslo: Universitetsforlaget. Published in English under the title Form and Content in Industrial Democracy (see Emery & Thorsrud, 1969). THORSRUD, E. & EMERY, F. E. (1966). Industrial Conflict and 'Industrial Democracy'. In J. R. Lawrence (ed.), Operational Research and the Social Sciences. London: Tavistock. THORSRUD, E. &EMERY, F. E. (1969). Motenny Bedriftsorganisasjon. Oslo: Tanum. THORSRUD, E., GULOWSEN, J. & KOLLTVEIT, E. (1967). Skipet i Utvikling. Oslo: Work Research Institutes. (Mimeo.) TOURAINE, A. (1955). L'Evolution du travail ouvrier aux usines Renault. Paris: Centre National de la Recherche Scientifique. TRIST, A. (1968). Work and Life at Sea. HRC Doe. 120. London: Tavistock Institute of Human Relations. TRIST, E. L. (1953). Some Observations on the Machine Face as a Socio-technical System. Doe. 341. London: Tavistock Institute of Human Relations. TRIST, E. L. (1958). Human Relations in Industry. A paper delivered before a seminar on Workers' Participation in Management, held 230 TAYLOR, F.

BIBLIOGRAPHY

in Vienna under the auspices of the Congress for Cultural Freedom. The Professional Facilitation of Planned Change in Organizations. Proceedings of the XVIth International Congress on Applied Psychology. Holland: Swets & Zeitlinger. TRIST, E. L. (1970). The Organizing and Financing of Research. In R. Mahou (ed.), Main Trends of Research in the Social and Human Sciences. (Unesco.) London: HMSO. TRIST, E. L. & BAMFORTH, K. W. (1951). Some Social and Psychological Consequences of the Longwall Method of Coal-getting. Human Relations 4, 3-38. TRIST, E. L., HIGGIN, G. W., MURRA Y, H. & POLLOCK, A. B. (1963). Organizational Choice. London: Tavistock. TURNER, A. N. (1957). Foreman, Job, and Company. Human Relations 10, 99-112. TURNER, A. N. & LA WRBNCE, P. R. (1965). Industrial Jobs and the Worker. Boston, Mass.: Division of Research, Graduate School of Business Administration, Harvard University. TRIST, E. L. (1968).

UEXKULL, J. VON (1920). Umwelt und Innenwelt der Tiere.

Berlin:

Springer. & KRISZA T, G. (1934). Streifzuge durch die Umwelten von Tieren und Menschen. Berlin: Springer.

UEXKULL, J. VON

v ALOIS, R. L. DE, et al. (1959). Electrical Responses of the Primate Visual System, 11. Journal of Comparative Physiology & Psychology 1, 69-81. VAN BEINUM, H. 1. 1. (1963). Een Organisatie in Beweging. Leiden. VAN BBINUM, H. J. J. (1965). Factors underlying Attitudes towards the Trade Unions and their Relevance for a Programme of Change. Doe. 464. London: Tavistock Institute of Human Relations. VAN BEINUM, H. J. J. & DE BEL, P. (1968). Improving Attitudes to Work, especially by Participation. Proceedings of the XVIth International Congress on Applied Psychology. Holland: Swets & Zeitlinger. VICKERS, SIR G. (1965). The Art ofJudgment. London: Chapman & Hall; New York: Basic Books. VICKERS, SIR G. (1968). Value Systems and Social Process. London: Tavistock. & GUEST, R. H. (1952). The Man on the Assembly Line. Cambridge, Mass.: Harvard University Press. 231

WALKER, C. R.

Q

BIBLIOGRAPHY WALKER, C. R., GUEST, R. H. & TURNER, A. N. (1956).

The Foreman on the Assembly Line. Cambridge, Mass.: Harvard University Press. WATT,K.F.(1966).SystemsAnalysisinEcology.NewYork:Academic Press. WEINBERG, A. M. (1963, 1964). Criteria for Scientific Choice. Minerva 1(2) and 3(1). WEISS, R. S. & JACOBSON, E. (1955). A Method for the Analysis of the Structure of Complex Organizations. American Sociological Review 20, 661-8. WERTHEIMER, M. (1959). Productive Thinking. Enlarged edition, edited by Michael Wertheimer. New York: Harper; London: Tavistock, 1961. WESTERLUND, G. (1952). Group Leadership. Stockholm: Nordisk Rotogravyr. WESTFAL-LARSEN & eo. (1972). Heranger-prosjektet. (Mimeo.) WHITE, R. & LIPPITT, R. (1960). Leader Behavior and Member Reaction in Three 'Social Climates'. In D. Cartwright & A. Zander (eds.), Group Dynamics. Second edition. Evanston, Ill.: Row, Peterson; London: Tavistock. WHITEHEAD, A. N. (1958). The Aims of Education. London: Benn. WHITWORTH, A. (1901). Choice and Chance. Cambridge: Deighton Bell. WHYTE, w. F. (1955). Engineers and Workers: A Case Study. Human Organization 14(4), 3-12. WIENBR, N. (1961). Cybernetics. Second edition. Cambridge, Mass.: MIT Press. WILLIAMSON, D. T. N. (1972). The Anachronistic Factory. Proc. R. Soc.(London)B. 331, 139-60. WILSON, A. T. M. (1951). Some Aspects of Social Process: Lewin Memorial Award Lecture. Journal of Social Issues, Suppl. No. 5. WILSON, A. T. M. (1955). Some Contrasting Socio-technical Production Systems. The Manager, Dec. WILSON, A. T. M., TRIST, B. L. & BAMFORTH, K. W. (1951). The Bolsover System of Continuous Mining. Doe. 290. London: Tavistock Institute of Human Relations. WOLD, H. o. A. (1966). On the Definition and Meaning of Causal Concepts. In Model Building in the Human Sciences. Entretiens de Monaco en sciences humaines, 1964. Monaco: Editions 'Sciences Humaines'. wooDWARD, J. (1958). Management and Technology. London: HMSO. 232

BIBLIOGRAPHY WY ATT, S. & FRASER, J. A. (1929).

The Effects of Monotony in Work. Industrial Fatigue Research Board.

YOUNG, M. {1958).

The Meritocracy. London: Routledge.

Jr. {1964). The Sailor Aboard a Ship: A Study of Role Behavior in a Total Institution. Social Forces 43, 389-400.

ZURCHER, L. A.

233

INDEX

Ackoff, R. L., 85, 129, 219 activity behavioural and technological, 113-14 defined,65 in a production system, 114, 117 activity domain, 39, 91-2 activity relationships, 131,136 adulthood, 213 affect control, 37ft' alcoholism, among sailors, 46-7 alienation, 44 architectural planning, 8 arctic camps, 50 Arner, 0., 32n, 47, 219 Atkin, R. H., 109, 219 Aubert, V., 32n, 219 authoritarian structures, 35 automated process industries, 7 automation, effects of on workers, 16-18 autonomous work groups, 4, 5, 21, 22,32, IOD-1, 125n autonomous work organizations, 5, 6, 19ft', 45 Bamforth, K. W., 2, 3, 21, liS, 125, 231,232 Barish, N. N., liS, 219 Barnes, R. M., li7n, 219 Barth, F., 32n, 220 Bavelas, A., 136n, 220 behavioural events, 114 behavioural laws, 75-81 behavioural measurement scales, 97,98

behavioural science basis for, 72 conceptual structure of, 82 distinguished from theoretical science, 209 integration of with physical sciences, 86-7, 109 behavioural unit, 85 Behavioural Worlds, 97n behaviour systems, individuals and groups as, 91 behaviour variables, 97-103, 106 Bemal, J., 173n, 220 Bertalanffy, L. von, 20, 220 Beurle, R. L., 20, 220 biological growth systems, 26-7 biosimulation, 19 Blauner, R., 4n, 220 Blichfeldt, J. F., 182n, 220 body-mind theories, 87 Bolsover technique, 4 Born, M., 70, 220 Bott, E., 129, 220 boundary zones, 145 Boyle's law, 84 Brache, Tycho, 214 Brentano, F., 96, 220 Brentano's and Stevens's principle, 96 Brun-Gulbrandsen, S., 46, 221 Canter, R. R., 5n, 222 career policies, for ships' crews, 53,60 car factories, 16 change in environment, 9

235

INDBX

change-contd. in group composition, 156 model for, 9 organizational, effects of, 10 in organizational design, 61-2 processes, 183 in ship organization, 31 in social class and family organization, 9-10 in work 01ganizations, 6 in work relationships, 131 Chardin, P. Teilhard de, 205, 221 Charles's law, 84 choice, organizational and technological, 8, 32 coalmining studies, reference to, 3ff, 21, 49, 125, 147,210 coercion, 14ff collegial relations, 34-5, 51 collusion, 16 communication theory, 20 complete specification design, 13-19 component-structure model, 77-9 composite autonomous work organizations, 4-5 computers, 7, 52 Comte, A., 172 conceptual models, and research strategies, 85ff conceptual structure, 83, 84 conflict, 4, 21-2 Connelly, J. R., 118, 221 continuous learning, capacity for on ships, 57ff in work organizations, 7, 8, 61, 188 continuous process techniques, 16 control sequence structure, 139-43 control tasks, 66 convergent research, 168 Coolidge, J. L., 97n, 221 cooperative relationships, 22 craftsman, the 13 critical specification design, 19-26 cybernetics, 20 cyclic dependence, 124-5 cyclic networks, 79

data analysis, in multiple-perspective model, 89-90 data collection, 115 Davis, L. E., 5n, 7, 21, 221, 222 degree of dissociation, 134 Demokratiseringsprosessen i arbeidslivet, 3n, 28n desalination research, 168-9 design engineer, task of, 8 developmental processes, sustenance of, 57 differentiated work systems, 149 dimensional analysis of production systems, 115, 116-20 dimensionally homogeneous equation, 99 dimensionless parameter, 99 disruption potential, 122 distance-regulating mechanisms, 35-44 'disturbance', term, 113n divergent research, 168 ecology of research organizations, 161 ecology-related disciplines, 172 economics, scientific development in, 85 educational organizations characteristics of, 181-2 centralistic model, 191-3 hierarchical model, 191 matrix model, 196 multiple-perspective model, 193-7 segmented model, 190-1 strategies for change, 188-9 subject-teacher relationships, 187-8 task structure, 183-8 effects of on student, 185-6 and teacher-student relationships, 184-5 types of task, 182-3 educational requirements for ships' crews, 48-9 Einstein, A., 70n, 74, 75, 84

236

INDEX

element integration, in analysis of production systems, 119 (fig.) elements, in production process, 116-17 Emery, F. E., 5, 6, 7, 9n, 21, 23, 32, 113n, 188n, 189n, 204,210, 222,230 energy, concept of, 87 Engelstad, P. H., 7, 124n, 222, 223,227 environment changes in, 9 pollution of, 25 in world models, 203-4 environments, turbulent, 32, 204, 205 equivalence relations, 131 Euclidean metric scales, 106, 107 European Journal of Social Psychology, 28n evolutionary system design, 26-7 excessive fragmentation, 33 exchangeable component structure, 34-5 exploratory and prediction problems, 67 extrinsic reward, 14ff factory organizations, 28, 120 family studies, reference to, 89, 129, 144 fantasy, 39 Fechner, G. T., 96, 223 Fechner-Weber principle, 96 feedback devices, 19 flexibility, measurement of, 153-4 flow process charts, 119 fluid groups, 132, 133, 135 foreman control, 15 friendship relations, 34-5 functionallaws, 71, 72-3, 165-6 functional networks, 79-81, 82ff gas law, 73 Geelmuyden, H., 40, 223

generalized laws, 72 general systems theory, 87, 172n general theories, 70ff, 84 group activity structure matrix for, 126, 134-7 measurement of, 131-8 Guest, R. H., 4n, 231 guilds, 13, 14 Gulowsen, J., 32,224,230 Halliday, J. L., 3, 224 Heider, F., 69, 78, 129, 205, 224 Heider principle, 129 Herbst, P. G., 3n, 5, 7, 21, 28n, 49, 57,72,78, 79,89,91,92,97n, 100,101,103,104,109,125n, 136n, 144, 152, 210, 224 Heron, G., 214, 224 Hersson, L., 32n, 219 hierarchical structures, 33-4 Higgin, G. W., 5, 21, 49, 125n, 210, 231 Hippel, A. R. von, 19, 224 Hoffman, J., 5n, 222 Human Relations, 82n human relations approach, 4 identity transformation, 103-4, 105 ideology, 84 independent disciplines, 85-6 individual behaviour, measurement of, 106 individuals, defined behaviourally, 118 Industrial Democracy Project, see Norwegian Industrial Democracy Project industrial revolution, 7, 8,13, 18 inferential problems, 67 input-output model, 68-9 Institute for Industrial Social Research (Trondheim), 3n integration of disciplines, 93-4 interaction pattern, 132, 133, 135 internal variables, 102, 103

237

INDEX

interpersonal relations, 37ff interstitial disciplines, 86 lrgens-Jensen, 0., 46, 221 isomorphic principles, 90-2 iteration cycle, 24 job breakdown, 7 job design, 5n, 21 job domains, 147 job motivation, 14 jobs, defined, 118 joint optimization, 4, 6-7, 9 Jordan, N., 4, 225 Junge, K., 105-6n, 225 Katz, L., 135n, 225 Kepler's laws, 83, 214 kibbutz organizations, 210 King, S.D.M., 5n, 225 kinship network studies, 129 Kolltveit, E., 32, 230

measurement and theory, 82ff medium-thing model, 69-70, 205 mental health disorders, 25 method formulation, 82, 83, 84 method-oriented research, 168 Middle Ages, 214 mining, see coalmining studies Mobius, A. F., 97n molecular engineering, 19 monistic models, 87 mono-role system, 40 Moore, F. G., 118, 227 Moreby, D., 56, 227 Morris, J. N., 3, 227 Morse, N. C., 4n, 227 multidisciplinary research, 82 multiple-perspective model, 88-94 multi-role structure, 53 Murray, H., 5, 21, 49, 125n, 210, 231

labour costs, 16 Lange, K., 124n, 227 Lenin, V. 1., 208n Lewin, K., 37, 77, 225, 226 linear graph, 136 linear relationships in functional networks, 79, 80 linear transformation, 103 Luce, R. D., 96, 97, 106,226

networks, 88, 89 Newton, Sir Isaac, 13, 74, 84 Nicholas of Cusa, 214, 224 non-segmented social space, 37, 38, 40,41 non-specification technique, 20 Nordisk Forum, 181n Norwegian Council for Science and the Humanities, 157n Norwegian Industrial Democracy PToject,6,9n, 159, 188n,210

machine charts, 117 machines, classification of, 118 management function, 21 manufacturing process, defined, 119 Marcuse, H., 193,226 Marek, J., 124n, 226,227 Marx, K., 208n matrix organizations, 22, 54-60, 195, 196 Maynard, H. B., 117n, 119,227 measurement and data, 115

O'Donnell, P. R., 118,228 Ohm's law, 74 open-system theory, 20 operational interpretations, 74-6 operational models, types of, 68-72 operational principles, 71, 74, 165 operational units basic to behaviour systems, 75 defined,65, 161 interconnection of, 69, 81 physical laws as, 74 operation charts, 117

238

INDBX

operations in a socio-technical system, 116 organizational design, change in, 61-2 organizational learning and change, 7' 54--60, 188-9 participation data, 144 payment, method of, 6, 21 performance satisfaction, 101 personal life region, 38 personal private region, 38-9 phenomena, interpretation of, 87, 88,95 physicalist interval scale, 108 physicalist models, 87-8 physical laws, 73-5 physical monism, 88 physical sciences, process structures in, 83,84 Plucker, J., 97n Pollock, A.B., 5, 21, 49, 125n, 210,231 polycentres, 196 postulate system, 166 Price, D. J. de Solla, 174n, 228 problems, defined, 67 process automation, 17-18 process charts, 117 production process, defined, 118 production systems design of, 21, 24,25 dimensional analysis of, 115, 11~20 production task structure, 114 productivity, 4-5 progr.wrrunedtapes, 19 projective geometry, 1~7. 109 projective transformation, 100ff psychological and physical variables, 95-6 psychological measurement scale, 107 psychology, scientific development of, 85-6 psycho-physical events, 95

psycho-physical model applied,98-100 transformations, 100ff types of problem applicable to, 108-9 validity, 106 psycho-physical principle criteria for evaluating, 103ff forms of, ~7, 103 pupil-task relations, 104-5 quantum theory, 84 random nets, 20 recruitment of ships' crews, 52-3 Reimer, E., 4n, 227 relationships in a multipleperspective model, 88, 89 relocation systems, 152-3 research, defined,158, 168 research-dient relationships, 159-61 research organizations ecology of, 161 role of, 10 see also research tasks research strategies, conceptual models, 85ff research tasks characteristics of, 1S7-8 conceptual framework for definitions, 161-2 generality level, 164-9 indeterminacy level, 162 innovation level, 162-4 and research organization, 158-9,168-9 restructuring of groups, 135 Rice, A. K., 5, 21, 228 Rodgers, W. H., 117n, 228 Roggema, J., 32, 36, 37n, 57, 58n, 182n, 184,185,229 role differentiation matrices, 126, 145-6 role-shifting, 37 roles of ships' crews, 43, 51-2

239

INDEX

scale transformation, 96ff Scandinavian Summer University, 170n, 196 Schmidt, W. H., 216, 229 science policy, 179-80 sciences and applied fields, 175-9 dependence relationships between, 173-5 developmental effects of, 178 developmental trends, 171-2 and linking disciplines, 170-3 order of priority for, 179 scientific development, process structure for, 83, 84 scientific operational principles, 165 segmented models, 188 segmented work systems, 150-1 self-maintaining work organizations, 22, 23 self-regulating work organizations, 7 sequential dependence relationships, 121, 122 service process, 118 Shannon, C. E., 20, 229 shipboard culture, 40 ship organization conventional design, 29-33 design problems, 28 emerging characteristics, 61-2 interdependence structure, 47-53 psychodynamic and social system characteristics, 33-4 rationalization, effects of, 44 requirements, 45-7 studies of, 32-3, 57 tension, 46-7, 53 territories, 52 transitional matrix, 55, 56 variables in, 29, 31 ship's captain, 40-1,43, 51 ship's crew, 43, 51-3, 57--60 ship's officers, 54, 58-9 social ecology, 9-10 social monitoring, 62 social-psychological characteristics

of work organizations, 120 social systems, world view of, 43 society, human development of, 214-16 socio-didactic analysis, 182 socio-didactic systems, 188 sociology, scientific development of, 84,85-6 socio-technical experiments, 6 socio-technical research development of, 3-6 history of, 7-10 Sommerhoff, G., 69, 79, 229 Stegemerten, G. J., 117n, 119, 227 Stevens, S. S., 96, 230 stochastic learning theory, 20n stress, 79-80, 102, 104, 109 task,defined,65,161-2 task acceptance, 145 task allocation, 21, 145-7 task analysis, 14 task dependence structure, 120-4, 127 task domain, 145, 146 task relationships, matrix representation, 123 task specification, 65-7, 70 Tavistock Institute of Human Relations, 13n, 113n technical events, 114 Technion, Haifa, 157n technological change increasing rate of, 7-8 and ships, 31, 47-8 in world models, 206 technological choice, 8, 32, 206 tension, 46-7, 53 theory formulation, 82-5 Thorsrud, E., 6, 7, 9n, 23, 32, 58n, 183, 188n,210,230 Tidsskrift for Samfunnsforskning, 3n,28n time-and-motion study, 15 total unit management, 17 Touraine, A., 4n, 230

240

INDBX

town planning, 8 trade unions, IS training schemes, for ships' crews, 51-60 transactional model, 76-7 transactional relationships, 160 transactional variables, 102, 103 transdisciplinary principles, 92-3 transformation of behaviour variables, 97ff transformation principles, 93 transformation problems, 66-7 transformation tasks, 66 Trist, A., 31,230 Trist, E. L., 3, 4, S, 21, 32, 49, 115, 125,204,210,222,231,232 turbulent environments, 32, 204, 205

variance-transmission relationships, 121,122 vulnerability, defined, 122

uncertainty, in research, 158 undifferentiated work systems, 148 University of Jerusalem, 157n University of Uppsala, 181n Valois, R. L. de, 106n, 231 values and ethics, 209-10 Van Beinum, H. J. J., 6, 231 variability, and system functioning, 20-1 variables behaviour, 97ff in critical specification technique, 23-4,25 psychological and physical, 95-6,108-9 in ship organization design, 29, 31 state and parametric, 73-4 transactional, 102, 103 variance, reactions to, 139 •variance', term, 113n variance-control pattern, types of, 142-3 variance-control process, 116, 138-43 variance matrix technique, 123-4

Walker, C. R., 4n, 231 Waring, E., 97n Weaver, W., 20, 229 weaving industry, 19 Wertheimer, M., 185n, 232 Westfal-Larsen Shipping Company, 58n,232 Westerlund, G., 4n, 232 Whitworth, A.,137, 232 Wiener, N., 20, 232 Wilson, A. T. M., 4, 21,232 Wold, H. 0. A., 73, 166n, 232 work-domain structure, 116, 143-56 effects of internal ftow, tSS-6 effects of segmentation, 150-1 ftow structure, 151-6 path-field representation, 144 stages in growth of work systems, 145-7 task allocation, 145-7 worker-managementconftict, 16 workers foreman control of, IS machine control of, 16-17 participation in decision-making, 6,23 and process automation, 17-18 response motivation, 14ff and work-method control, 15-16 working principles, defined, 164 work organizations changes in, 6 composite, 5, 6 conftict in, 4 fractionated, 21 self-maintaining, 22, 23 self-regulating, 7 traditional, 18, 56 in world models, 206 work relationship pattern, measurement of, 125ff

241

INDEX

work relationship structure measurement of, 124-30 group activity structure, 126,

131-8 task dependence structure, 120-4, 127

types of, 129, 130 world model, defined, 164 world model1, 201-2 world model 2, 202-3 world model3, 204-11

242