Reference:Evolution of socio-technical systems/Developments at the level of the primary work system

In reviewing the developments which have taken place since the socio-technical concept was established, it is appropriate to begin with the primary work system since this has been made the organizational building block.

The advent of the Norwegian Industrial Democracy project faced the research team with the task of intervening in the design of work systems. The situation was different from that of the fifties when the key innovations had appeared spontaneously. In the development of socio-technical studies, concepts and methods have evolved in relation to the demands of the field situation.

The principles of work design

A set of principles was needed to improve work design so that the ideal of joint-optimization could be approached. Basic to this was some knowledge of the psychological requirements individuals have of their work beyond what is usually included in an employment contract. Herzberg et al. (1959) seemed to be right in separating the extrinsic from the intrinsic dimensions of job satisfaction, whatever the statistical arguments about 'dissatisfiers' and 'motivators'. What the trade unions had fought for had to remain or be won where it did not exist - adequate and fair pay, job security, benefits, safety, health, due process. These constituted the conditions of employment. What had also to be considered was how far jobs gave opportunities for an additional set of requirements that could only arise from the character of the jobs themselves and of the work organization in which they were embedded.

Drawing on Lewin's (1935) Berlin experiments on person-task relations (as well as on his and Bion's later work with groups), Emery (1964, 1976) identified six intrinsic characteristics (listed on the right of Table 2-1, which compares them with the extrinsic characteristics). The intrinsic characteristics may be spelled out as follows:

The need:

1. For the content of a job to be reasonably demanding in terms other than sheer endurance and to provide some variety (not necessarily novelty).
2. To be able to learn on the job and go on learning. Again it is a question of neither too much nor too little.
3. For an area of decision-making that the individual can call his own.
4. For a certain degree of social support and recognition in the work place for the value of what he does.
5. To be able to relate what he does and what he produces to his social life, for it to have meaning and to afford dignity.
6. To feel that the job leads to some sort of desirable future (not necessarily promotion).

These intrinsic requirements are not confined to anyone level of employment. It is not possible to meet them in the same way in all work settings or for all kinds of people. They cannot always be judged from conscious expression. When there is no expectation that any of the jobs open to him will offer much chance of learning, a person will soon learn to 'forget' such a requirement.

The requirements are too general to serve as principles for work redesign. For this purpose they need to be linked to the objective characteristics of industrial jobs (Davis. 1957). Table 2- 2 (c. f. Emery, 1978a) summarizes the linkage.

The redesigning of work leads beyond individual jobs to the organization of groups of workers and, beyond that, the organization of support services (such as maintenance). The wider implications affect organization design more generally. A congruent set of principles of work design and of 'core dimensions' of jobs has been identified by other writers (Hackman and Lawler, 1971; Hackman, Oldham et aI., 1975; Herrick and Maccoby, 1975; Walton, 1975a; Hackman and Suttle, 1977). This degree of agreement is exceptional in so new a field and has placed work redesign on a firmer foundation than is. commonly realized.

Individual differences In motivation

A wide range of individual differences has been found in work force motivation. All workers do not want 'enriched' jobs or to take more responsibility. This was investigated by Hackman and Lawler (1971), whose findings indicate that employees with some desire for 'higher order' need satisfaction perform better and feel more positi ve when their jobs rate high on the 'four core dimensions', which are similar to those identified by Emery. Those whose motivational pattern is not particularly oriented to higher need satisfaction do not show so strong a pattern of association, though most of the correlations are in the same direction.

Much of the literature on job satisfaction has attached too much importance to responses given at only one point in time - especially to questionnaires. People change over time and learn through experience. Many workers do not know what their real feelings will turn out to be until they have had actual experience of redesigned jobs. They need to know also how far such jobs are likely to be an enduring feature of their work-lives or how much they represent merely a temporary change.

It makes a difference whether one is considering solely individual attitudes or also social change involving norms and values. Employees who have convincing evidence that their organization has committed itself -long term - to joint optimization are more likely to commit themselves than those who do not. This is especially so if they sense that the norms and values of the wider society are also changing in the direction of the new paradigm.

A work group offers scope for a range of abilities and preferences. There is more room for individual differences in work groups than in standardized individual work stations.

Motivation in the work place has been reconceptualized by Susman (1976) in terms of a theory of directed action, which draws on the object relations tradition in psychoanalysis (Klein, 1932, 1959) and on the work, deriving from Lewin and gestalt psychology, of Chein (1954, 1972). Directed action is transactional. It is connected with the completion of an object relationship. It is a molar concept coproduced by the characteristics of the object and the meaning which the subject imparts to the situation. The individual and the work place become directively correlated. This view enables Susman to spell out the conditions under which directed action can be incorporated into work design so that the individual experiences self-enhancement. These conditions are consistent with the principles summarized earlier and provide the enabling context in which commitment can develop.

Work analysis

Needed also was a method for analyzing work systems. This had not only to be academically defensible but communicable to workers, managers and staff specialists who could, after some practice, become able to- use it for themselves. The following nine-step model derives from the second field experiment of the Norwegian Industrial Democracy project at the Hunsfoss Paper and Pulp Mill which began in 1964 (Emery and Thorsrud, 1969, 1976) - where for the first time an 'action group' of workers, technicians and supervisors was created in order to diagnose the malfunctioning of the particular system they were concerned with. Emery was again the initiator. The condensed version, quoted below from Trist (1971), has been put in systems terms to make it as general as possible.

1. An initial scanning is made of all the main aspects - technical and social- of the selected target system - that is, department or plant to be studied.
2. The unit operations - that is, the transformations (changes of state) of the material or product that take place in the target system - are then identified, whether carried out by men or machines.
3. An attempt is made to discover the key variances and their interrelations. A variance is key if it significantly affects (1) either the quantity or quality of production, and (2) either the operating or social costs of production.
4. A table of variance control is then drawn up to ascertain how far the key variances are controlled by the social system - the workers, supervisors, and managers concerned. Investigation is made of what variances are imported or exported across the social-system boundary.
5. A separate inquiry is made into social-system members' perception of their roles and of role possibilities as well as constraining factors.
6. Attention then shifts to neighboring systems, beginning with the support or maintenance system.
7. Attention continues to the boundary-crossing systems on the input and output side - that is, supplier and user systems.
8. The target system and its immediate neighbors are then considered in the context of the general management system of the organization as regards the effects of policies or development plans of either a technical or social nature.
9. Recycling occurs at any stage, eventually culminating in design proposals for the target and/or neighboring systems.

This procedure was first used as a training method for departmental managers in the Shell Management Philosophy project in the U.K. (Hill, 1971). It has since been incorporated by Davis in the UCLA Short Course on QWL.

The model was originally tailored to the requirements of continuous process industries. A variant for office units was then introduced. While a number of alternatives are likely to be required for different technologies, the logic of relating any target system to the set of its surrounding systems would appear to be general.

Autonomous groups and primary work systems

No one group in any organization can be completely autonomous. It can only be conditionally or semiautonomous. There are, nevertheless, several dimensions and degrees of autonomy. The most systematic analysis of these has been made by Susman (1976), who, building on an earlier paper by Gulowsen (1972), approaches this question by distinguishing three classes of decision: those concerning task independence, those concerning self-governance, and those concerning types of self-regulatory activity. He separates boundarytransaction uncertainty from conversion uncertainty and proceeds to introduce categories of technically required cooperation, type of interdependence, and type of coordination. These concepts provide a framework for the analysis of autonomy more rigorous than that previously available.

A socio-technical theory of the efficacy of autonomous work groups is based on the cybernetic concept of self-regulation. The more the key variances can be controlled by the group, the better the results and the higher the member satisfaction. Over a large array of situations, the range of variances controllable by a group is greater than that controllable by individuals separately linked to an external supervisor. The difference in the underlying design principle is summarized in Emery's concept of 'participant design' (1974, 1976). The function of supervision is to manage the boundary conditions in the group's environment so that the group itself may be freed to manage its own activities. This is a very different concept from the bureaucratic theory of control.

Autonomous groups are learning systems. As their capabilities increase, they extend their decision space. In production units they tend to absorb certain maintenance and control functions. They become able to set their own machines. The problem-solving capability increases on day-to-day issues. ^[1] They negotiate for their special needs with their supply and user departments. As time goes on, more of their members acquire more of the relevant skills. Yet most such groups allow a considerable range of preferences as regards multi-skilling and job interchange. The less venturesome and more modestly endowed can find suitable niches. The overall gain in flexibility can become very considerable, and this can be used to enhance performance and also to accommodate personal needs as regards time off, shifts, vacations, etc.

Autonomous groups do not always succeed. A good deal has become known about the conditions affecting their success or failure. These will not be reviewed here, except to note that one of the most common reasons for failure is lack of support in the surrounding organizational milieu. A year or two ago, my research center at the University of Pennsylvania carried out a study of 'work teams' in a very large organization. Of the 90 that had existed at some time during a lO-year period, only two were extant when the study was made. In addition to the effects of the mid- '70s recession in disrupting work teams through layoffs, a principal reason for failure was lack of support in the wider organization. When the initiator departed, 'fade-out' occurred (MBSC, 1978), no matter how successful the project was economically.

Autonomous groups of the face-to-face kind are not the only type of nonhierarchical social formation that has appeared at the level of the primary work system. Herbst (1976) points to 'matrix' type groups in which there is limited overlap between member skills which are too complex for all to learn. In such groups, there may be considerable spatial and temporal scatter. Herbst also mentions clusters of network roles which 'boundary span' across primary work systems and also connect subgroups within such systems. Matrix groups and network clusters are becoming prominent as organizational interdependencies increase. This is apt to happen to a greater extent in advanced technologies and in organizations with large and varied clienteles.

Confusion has been created regarding the number of individuals suitable for inclusion in autonomous work groups. Small group theory supports an upper limit in the 8-12 range. The Tavistock mining studies, however, showed that 'composite groups' tended to be much larger. A number of other studies have reported groups beyond the limit of the face-to-face range. These all tend to be complex groups with several subsets.

Another unit of analysis is required, namely the primary work system, which may include more than one face-to-face group along with others in matrix and network clusters. The primary work system is a functional system with a semi-independent operational identity, whether as a production or service unit. In the Saab 99 engine plant, all the assembly teams of three formed part of the same primary work system.

In a primary work system an individual is apt to have several group memberships. In the mining studies a miner considered himself as belonging basically to a 'seam society' in which he had established rights and privileges regarding employment and deployment. Within the seam he belonged to a face or cycle group; thereafter to a task group. This latter was more temporary than the former which, though focal, was less enduring than his seam belongingness.

These multiple but congruent memberships gave him considerable 'space of free movement' within a seam population of more than 100 people which still comprised a personal world. The seam group contained ancillary personnel, a district management system, and a complete territory, as well as all the equipment necessary for mining [c.f. Miller (1959) regarding the relations between territory, technology and time]. It constituted a polity, being a recognized worker constituency with representatives who conducted detailed negotiations for the seam within the overall union agreement for the colliery.

This was the type of group which Rice found in the automatic loomshed in Ahmedabad (though the overall number was smaller), which Melman found in the Standard Motor Company in Coventry (though the overall number was larger), and which Burden (1975) describes in the micro wax plant in Shell's Stanlow refinery in Cheshire. In my own work I have found such groups in a number of situations, including Alcan's Reduction Division at Arvida, where the quite large divisional workforce served as a reference community for the on-shift task groups of six. It functioned as their polity, taking all decisions concerning self-governance - including the decision to accept the proposal (which its representatives took part in shaping) to tryout autonomous work groups. The task groups made decisions concerning the self-regulation of their own activities.

Primary work systems of this scale and this complexity are the type of socio-technical unit which is emerging at the present time in a number of new plants in North America. Cummins Engines' diesel plant in Jamestown, N.Y. (Pava, 1979) and Shell's petro-chemical plants in Sarnia, Ontario ('Davis and Sullivan, 1980) are examples. In the latter the workers have the possibility of learning not only all the process jobs, but also a coordinating or service job in instrumentation, the lab, quality control, maintenance, etc. An advantage from their point of view is that they can maximize their time on days (over 50 percent) - a cherished improvement as regards the workhome interface.

Such arrangements increase the competencies that may be acquired, the number of deployment patterns available, and the career paths open. For the individual they create roles rather than mere jobs. For the organization they bring into being a variety-increasing system directively correlated with the complexities and interdependencies of the technology. Emery (1980) has described an elaborate system of this kind in a new metal mine in Australia where there is paramount need to share information and for all to have a cognitive map of the entire process. This cognitive map is the essential system control which has now become a new 'dissipative structure' (Prigogine, 1968) in the workers themselves. It represents a 'morphogenetic' change. The penalties of not evolving socia-technical systems of this kind have been suffered in more than one nuclear power plant.

Complex primary work systems will increase as computer-aided continuous process technology advances. Group, or cellular, technology offers a parallel opportunity in batch production (Williamson, 1972). These areas need an intensification of research efforts.

Multiple memberships in a larger group which comprises a social aggregate may be postulated as lessening the dangers of overinvestment which can easily occur if the individual is bound too exclusively in one face-to-face group. These dangers have been investigated by Miller and Rice (1967) in their work on the relations between task and 'sentient' group boundaries. The benefit in greater cohesion consequent on the coincidence of these boundaries tends to be offset by a greater propensity to self-sealing. The presence of a social aggregate introduces properties into a primary work system different from those of the constituent task groups (c. f. Churchman and Emery, 1966). It constitutes the ground on which they are the figures. To have some ground of its own on which to stand increases the power of a group in a wider organizational setting to relate to other groups having similar standing. This enables the members of a primary work system to become a polity. The ability of the system group to become a polity or worker constituency relates it directly to the trade-union organization. This may undo many fears which union people have concerning autonomous groups - that even though the consequence may be unintended, their effect will be to undermine the union by bringing into being a competing loyalty.

The development of self-standing primary work systems containing mixes of groups with commonly shared skills, matrices whose members have partly overlapping skills and networks of mainly specialist skills constitutes a new basis for the effectiveness of socio-technical organizations. They create organizational units of considerable 'robustness' which compose 'microsocieties' which have intragroup, intergroup and aggregate relations with a whole operational task. These microsocieties provide considerable space of free movement to the individual and are open to the interorganizational environment.

Self-standing primary work systems exemplify a holographic principle of organization in which the whole is represented in the part (Pribram, 1977). The forms through which holographic primary work systems may best become linked to the overall organization constitute an area requiring further research. The question is not so much that of small being beautiful as of finding ways of retaining small in large so that the advantages of both may be realized. Open systems planning as developed by Clarke, Krone and McWhinney (Jayaram, 1976; McWhinney, 1980) would seem to offer a promising new approach to this problem.

Notes