in search of a new paradigm

Tom Dwyer

IFCH/UNICAMP

Paper Presented to the XVth World Congress on Occupational Safety and Health, S‹o Paulo, Brazil. 12-16 April, 1999.

(This paper serves as an introduction to Tom DwyerÕs presentation. It is a modified and revised version of a paper presented to the Third European Conference on Monitoring the Working Environment, given on October 10-11 1996, European Foundation, Dublin, Irish Republic.)

The Great Western Paradigm

The 'Great Western Paradigm,' founded by Descartes, would come to be associated with far more than just science, it would become associated with what is called 'modernity' and this would spread its tentacles, carrying the spirit of reason into all nooks and crannies of Western life. It met with extraordinary success as its rationalising power helped build the wealthiest nations ever seen and, through science and education, the highest average levels education and of knowledge distribution known in human history. This was achieved through a general process that sought to crush traditions and all other imagined obstacles to the spread of reason. The idea of reason built up came to be exported, by imposition or by imitation, as though it were universally applicable.

One reason for the spread of the paradigm was the rising capacity of scientific activity to explain, and in its applied forms, to transform the world. As science grew its diverse languages became increasingly esoteric and distant from citizens. Science contributed to concentrating powers in the hands of people who had no role in producing it, the heads of large corporations and of nation states. A darker side of the modernisation process became visible, the building and concentration in human hands of destructive powers hitherto unimagined in the history of the World and the creation of ecological imbalance. From the late 1960s there was increasing questioning in the West of science and of the society it had played such an important role in producing. The very idea of universal principles of reason would come be contested, and this would have profound effects in many parts of the World.

The Industrial Safety and Health Paradigm

The idea of work safety and health predominating in pre-industrial Europe was based around the general notion of causality operating, summed up in the words of Ricoeur: "if you suffer, if you are sick, if you fail, if you die, it is because you have sinned". With the rise of industrial society causes of misfortune came to be blamed on human agents, and in the case of accidents on the owners of industrial plants.

Notions of responsibility, ideas of cause and prevention emerged that were closely linked to the development of science, technique and a legal system which responded to the problems of an industrial age. These measures together would come to form the presently dominant paradigm on safety and health at work which is, as to be expected, closely linked to the Great Western Paradigm. The world was separated into the world of things and the world of thinkers, into objects and subjects. In industrial safety this eventually translated into the concepts of 'unsafe conditions' and 'unsafe acts'. With regards health the objects were microbes, dusts and other noxious agents to be combatted through vaccines, sanitary measures and personal hygiene. It took over a century, from Sir Humphrey Davy's invention of the safety lamp until the publication of Heinrich's Industrial Accident Prevention, for the presently dominant paradigm to crystallise in the area of work safety. The scientific disciplines of industrial engineering and medicine, through armies of professional practitioners, would come to deal with the conditions of work. Industrial psychology was brought in to help deal with the acts, these were to be controlled through education, incentives and the resort to discipline. At a more general level, legislation, at first implemented in the name of social peace, public health and humanitarian values, built up impersonal systems of rules and regulations to provide administrative controls in order to resolve the social problems and political conflicts that accompanied the development of industrial society. It sought to simultaneously regulate conditions and to punish behaviour. In Europe the building blocks of the modern welfare state were constituted through worker compensation and those of a controlling and interventionist state through safety legislation and inspectorates. The system of thought behind these institutions constitutes the industrial safety and health paradigm.

It is a paradigm different from the other dominant one which is the industrial health paradigm the history of which is now extrem4ely well known at a world level.

The Structure of Scientific Revolutions

Thomas Kuhn is widely attributed with giving the word paradigm its current signification. Science is seen as a self destructive activity, it advances by undermining its own bases. The central hypotheses is that science progresses when an existing theoretical system of knowledge becomes undermined by inquiry, the anomalies produced are typically left to a side and accumulate as research proceeds under the old paradigm. However, a paradigm change occurs when a new theoretical body of knowledge is produced and accepted, one that is capable of integrating both the validated knowledge accumulated under the old paradigm and its anomalies.

The Collapse of the Industrial Health and Safety Paradigm

The current situation is one of a contest of the paradigm built up with industrial society, this has occurred for two specific reasons:

1. a growing series of research results contradicted the dominant paradigm (this is the sense given by Kuhn to the basis of a scientific revolution) and, in a more practical sense, the consequences of the paradigm's limits were measured by rising accident rates in a number of countries observed in the 1970s and increasing preocupations about worker health. Further, major industrial disasters from Flixborough to Three Mile Island and Chernobyl were widely interpreted as evidence of the inadequacy of the paradigm.

2. the nature of economic activity suffered considerable mutation in the more advanced Western countries. Shrinking size, the 'putting out' of productive activities, the flattening of hierarchies and an increasing resort to consultancies were among the changes transforming the face of large productive organisations. Investment in knowledge replaced investment in the division of labour and machinery as the central growth producing process. This led some to hypothesise that a new form of productive society was being built. To mark its difference from the one that had gone before it some call it 'post-industrial'. New sources of power and new cultural patterns emerged, definitions of productive activity changed, certain social conflicts consequently lost their force or became increasingly institutionalised (ex. between capital and labour in the industrialised countries) and concommitently social conflicts emerged around new themes. For the purposes of this paper one significant new site of conflict can be identified, the production and control of 'risk' became a central theme of political and technical debates and social movements in many complex societies.

Four distinct responses to these twin movements can be observed: the application and extension of the principles of the industrial safety and health paradigm, the submission of actions carried out by professionals and the state to cost-benefit analysis, the depthening of hyper-rationalism in the form of systemic safety and, in the health area, what Diana Chapman Walsh calls 'workplace wellness programs,' and, least well defined of all, the developments of approaches based in the social sciences. The responses mix theoretical reflection, strategies for professional, employer and state action, and research. They constitute concrete signs of both the shoring up of and the disintegration of the old and yet are unable to provide clear signposts on what new approaches might include and how to get to them.

Towards a Cartography

The industrial safety and health paradigm operates in three distinct spheres of social reality: as a macro level frame of reference and site of conflict for the whole society, institutions of various kinds seek to build up knowledge and subsequently apply it aggregating power, economic and organisational resources and at a micro level the knowledge and conception of the world built up within the paradigm are applied so as to reshape workplaces in the name of safety and health. As has been stated diverse disciplines have traditionally studied health and safety: psychology, engineering, medicine and, more recently, ergonomics.

Traditionally sociology has avoided the area. In recent times, however, because of the limits reached by the industrial paradigm some have sought sociological insights. The negation of the notion, explicit in the dominant paradigm, that socially constructed subjects are an important element in the industrial safety and health equation, has spurred demand for sociological input. Such input is seen to fill a chasm in the paradigm - the ignorance of social relations as structuring social activity. How do such insights change our knowledge? At a macro level historical sociologies of science and of the State are equipped to provide explanations as to the origins and working of the industrial paradigm. In this manner variations from one country to another are explained in terms of the relative capacities of political and social actors to mobilise resources and skills in different directions and with different strengths. It is at the institutional level that concrete practices by scientists, professionals and administrators are produced, the sociology of the professions is but one sub-discipline which seeks to analyse and explain interventions made by social actors in their distinct areas of expertise. Finally, at the micro level, a sociology of work concentrates on the social processes which lead to those actions which result in the production of accidents and illness in workplaces.

The industrial paradigm provided a coherent view of the interventions necessary at the institutional and micro levels, its questioning poses a series of challenges, and leads to calls to rebuild 'the system'. In each of the disciplines which study the question researchers and theoreticians can be found seeking to provide responses to the new context without abandoning the knowledge acquired in the old. The question becomes "where to start?" A sociologist embarking on the adventure of writing about a new paradigm is obliged to navigate in the turbulent and waters produced these by disciplines and by the action of those who represent them. It seems prudent to start with those disciplines which are closest to a sociological approach. Uta Gerdhardt provides a warning that must be heeded in this enterprise: "sociology is an analytical science while medecine (and the same may be said for other disciplines in the field of industrial safety and health) is a practical endeavour. The latter may avail itself of various insights of the former, if they are carefully researched. But never should sociology which, within the confines of medicine, is at best one of its theoretical disciplines attempt to overcome or replace medicine. The danger hidden in such an undertaking is that sociology itself forgoes its roots and abandons its tenable aims. Its roots are in the theoretical reflection of societal goings-on. Its tenable aims are the systematic description and methodical explanation of social life and social structures."

Macro level

The overall risk of accident or illness that a category of workers has is the result of complex set of processes in the cultural, political and organisational spheres in a given society. Rosner and Markowitz's book about the history of silicosis in the United States this century detailed the mechanisms by which a known disease inducing agent was kept out of the public debate and not officially recognised as a cause of industrial illness. This occurred at the same time as different scientific research policies and political relations kept silicosis in the public arena and guaranteed its recognition in Germany and England. In a way similar to silicosis transformations in work processes may lead to new risks, these may creep up undetected and remain unrecognised, as did repetitive strain injuries (RSI), for a relatively long period because the available diagnostic and monitoring systems ignored them. However, the mere fact that a problem has been detected does not mean that attempts will be made to resolve it! For this to occur science and technique need to mobilised to devise prevention strategies, and such mobilisations are typically political in nature. Subsequently strategies are implemented to control exposure to those agents defined as damaging. Specific mobilisations of science and technique are to be analysed at the institutional level and the actual implementation of preventive measures is to be analysed at the micro level.

Not only do different countries have different industrial and labour market structures but they may exhibit quite varied approaches to similar industrial risks, furthermore the empirical results of similar preventive efforts may be quite different.

Most monitoring and control processes used today are based in the industrial health and safety paradigm. This is, as previously stated, a paradigm based in a worldview grounded in a particular type of reason and even though it is under increasing attack its advocates seek to impose this vision in producing knowledge and interventions.

The rise of new technological threats has accompanied a loss of public confidence in traditional ways of managing risk. In particular the idea, propogated within the industrial paradigm, that there exists only one kind of rationality, has come under attack. The management of peoples' relationships to risks in the economically developed societies is, as shown by social movements from ecologists to Aids activists, relevant to more than just scientists, professionals and bureaucrats, workers' representatives and employers. The triadic relationship between employers, workers and state and professional control agencies built up around the industrial safety and health paradigm in the First World is unbalanced by the introduction of a social new actor which, for want of a better conceptualisation at this stage, can be called 'the public'.

The journal Risk Analysis clearly reflects the importance of 'the public' as a force to be dealt with. Over the years it has included a large number of articles which place emphasis on public participation in decisions affecting the installation and management of plants which may result in the exposure of citizens to hazards. Such participation is seen as a strategy to educate the public and to guarantee its confidence and control. Articles also discuss the need for "scientific assessment of risk and a responsive consideration to the public's perception of risk". Decisions which the industrial safety and health paradigm had isolated as being merely technical, administrative and economic are thus thrown into an arena where they are treated politically, however, the aim of professionals and industry is that this treatment occur under conditions controlled and built up with reference to the dominant industrial paradigm- reinforced through reference to the esoteric equations of systemic safety or cost-benefit analysis.

Public participation in such decision-making processes may not follow the designs of those who seek to control it. Should it come to be widely accepted as a principle for the managing of the relationship between the public and hazards in the context of a democratic society it can be hypothesised that this would open up a space for the contest of and possible social control over the complex built up by scientists and the decision-makers they serve. The institutionalisation of such a space would make a specific contribution to the dilution of the enormous power this complex has built up. In another context Goulet has observed that "it is preferable to have large numbers of nonexpert people gain some understanding of what consequences technical choices have on their society than to have a tight knit team of experts committed to make certain technologies work."

At a macro level the appearance of the new social actor is forcing institutional moves that place scientific and technical decisons in a new perspective. Citizens, previously ignored in the construction of technological choices, except through the building of technical and administrative responses through political processes, now become integrated into these processes. Changing cultural values come to be reflected in new institutional practices and in redefinitions of basic concepts of democracy.

The Institiutonal Level

Many institutions have been built up to develop and administer the industrial safety and health paradigm. Bodies of knowledge are developed which may be theoretical or practical, and members of the institutions seek to apply them in concrete work situations. They claim legitimacy on the basis of their competence in the sphere where information is produced, distributed and applied and, in the case of the State, legitimacy is claimed on the bases of competence to exercise roles accorded by law.

Institutions mediate between the micro and the macro levels, members of institutions, be they safety engineers, industrial doctors, ergonomists, government inspectors or safety and health researchers act both within institutional systems and intervene in workplaces in ways that can affect peoples' day-to-day lives. What is relevant or irrelevant in constructing or responding to hazards is, in theory at least, determined solely by the member's disciplinary or administrative training. In times of changing values and technologies the relevance of earlier training may be put into question, members may no longer be able to perceive problems and suggest strategies to combat them.

Like all other types of workers, professionals and public servants (who I shall call administrators) work within systems of social relations. What differentiates professionals from other workers are their lengthy training periods, their specialist skills and the reference they make to their own autonomy. Administrators differentiate themselves by their claim to serve the State through the application of what Weber characterised as legal-rational authority. However, let us imagine for a moment that institutional members trained within the industrial safety and health paradigm have perfect knowledge of all relevant aspects of the task as defined by their discipline or institution. They are frequently found to claim that their autonomy insulates them from working within systems of social relations that are capable of distorting their judgement. The sociology of the professions has spent a great deal of effort in analysing professional action and claims of autonomy. One of of the sociology of professions discoveries that is of relevance to this paper is that employment status affects professional judgement. Eliot Freidson discusses the limits of professional action in large firms and develops certain generalisations.He indicates that doctors working in these firms normally have less autonomy to make their diagnoses in a proper manner than do liberal professionals. In addition those who work in large organisations appear to have more political autonomy than those employed in smaller organisations. The literature on industrial safety and health professionals demonstrates that the time allowed for medical diagnoses may be eroded in fee for service payment systems. Cost-benefit calculi or pressure from superior actors influence decisions thus, for example, inducing engineers to alter their professional judgements that certain types of equipment are unsafe, or that work should not be executed in specific situations.

The industrial safety and health paradigm assumes that actions are carried out by people who are guided only by considerations based in a pure technical rationality. This assumption is clearly inadequate since it ignores the types of action that have just been described. Any new paradigm must include specific reference to the social relations within which professionals work and the constraints upon their action which these lead to: the roles played by power and economic relations in shaping actions and that of limited and deficient qualifications which combine to limit the validity of professional claims to a monopoly of competence within their defined areas of action.

reflections from the institutional sphere

Traditional work safety and health management has sought to control the hazards to which people are exposed through standardising protective procedures. Members of institutions contribute to investment decisions which always result in the presence of certainÔresidualÕ hazards in a given workplace. Peoples' relationships with these hazards are then managed. The industrial paradigm is partial and omitted a whole series of factors well known factors that are relevant to life at work, workplaces can be poorly organised and people are unaware of certain hazards. People can be motivated to expose themselves to hazards for personal gain and, in situations where safety and health are traded off against profits, people may be pressured to work with known hazards. Beyond this new hazards are constantly produced.

Researchers and theoreticians, it should be of no surprise, introduce new reflections into the literature and oblige their disciplines to confront new themes. I will now draw a few insights from this literature that I believe are of importance for the development of a new paradigm.

As early as 1972 psychologists Hale and Hales' pioneering review of the relevant European literature led them to observe the insufficiencies of accident research conducted within the dominant paradigm. For them research should generate new prevention techniques, in order to do this they concluded that "radically new theories are needed." Only later on could a loose form of consensus be seen to emerge as to the directions in which theoretical research should go.

It would be startling were the search for a new overarching body of knowledge in the area of health and safety at work dissociated from wider movements in the economy and society. In 1989 the Head of the European UnionÕs FAST programme stated: "What the economy and the firm need is more and better human and organisational innovation. This calls for a qualitative jump in the understanding of how humans interact with technology and, above all, in the design of increasingly complex techno-organisational systems. Most of the current inefficiencies, failures and accidents in advanced automated production systems are due to system design approaches which do not sufficiently take into account interactions between human, organisational and technical elements." Such words give high-level institutionally-based backing to the views of numerous researchers.

Sociologist Charles Perrow analysed human factors engineering (ergonomics) and demonstrated that its assumptions precluded considerations relating to the role of a series of important social elements of the workplace. He urged a strengthening of the discipline through reference to such elements. Leading ergonomists, perceiving the limits of the industrial paradigm in various spheres have, for quite some time, been urging a reformulation of key principles of the discipline. Such a search is complex because the grounds of the discipline are shifting especially due to the emergence of new problems. De Montmollin provides an overview of the empirical content of the discipline: "Today the former supremacy of the physical, and above all the physiological sciences is questioned by the rise in cognitive sciences and in particular cognitive psychology which, thanks to the widespread introduction of computers, allows the human factors ergonomist to focus his attention less on lifting and more on the interfaces between operator and computer.... Here, 'workload' obviously implies mental workload, the object of much experimental research." Beyond this change in empirical focus the bases for analyses are shifting. Thus, 'macro-ergonomics', 'participative ergonomics' and 'activity oriented ergonomics' each seek to provide new foundations for the discipline. Although strictly minority movements the participative approaches are important for the rupture they make, which also follows in the tradition of French activity oriented ergonomics, by including reference made in constructing their analyses to a previously omitted factor, the workers' voices, as an expression of perceptions of lived out reality.

In the medical area Armstrong observed that a barrier was held up by British doctors who objectified conceptions of diagnosis, an approach which led to only visible ailments being defined as capable of treatment. In an article published in the early 1980s he saw that medical diagnosis had reached a point where "the conditions of possibility" for "an extended patient's view, whatever its empirical support, have begun to occur over the last few decades and this of itself (besides its ramifications into patient representation, community politics, patient's rights, and so on) signifies a change in the status of patienthood." In other words, an industrial doctor working within the emergent approach referred to by Armstrong captures worker's voices and actively employs them in constructing diagnoses. Despite widespread support for the use of the patient's view Armstrong doubts that general clinical practice in England had been modified in function of this newer vision. Dodier, on the other hand gets down to specifics and conducted a study providing close observational detail on how some French industrial doctors use the patient's view in constructing their diagnoses. A systematic attempt to incorporate such a perspective into industrial medicine was made with the Italian workers' models. Developed from the early sixties these now constitute a distant landmark. In more recent times Behavioural Medicine, supported by a considerable body of psychological theory, has arrived on the scene in the United States.

In engineering de Greene has seen a need for a "paradigm shift" in order to confront the challenges of complex technologies. His prescription states, among other things, that "Equipment and jobs should be designed simultaneously, and to the extent possible the users should also be the designers." In "Beyond Mechanization" Hirschhorn develops compatible arguments in raising doubts as to whether a phenomenon emergent in recent times, highly complex industries, can function safely without taking into account the workers, their perceptions, actions and views. Reflecting on Three Mile Island and other nuclear reactor incidents he argued, "we must design jobs in such a way that workers can effectively control the controls, modifying them and regulating them to prevent failures and errors unanticipated by the engineers." In the systems he proposes workers must be able to form an integrated vision of what goes on in their area, and have the capacity to take corrective action. Workers need to be able "to understand the consequences of their control decisions." For this to happen a whole set of new considerations need to be integrated into the design of workplaces: "[N]ew technologies... demand that we develop a culture of learning, an appreciation of emergent phenomena, an understanding of tacit knowledge, a feeling for interpersonal processes, and an appreciation of our organizational design choices."

Whilst de Greene urges a "paradigm shift", and Hirschhorn was quite precise in specifying in 1984 that the form of integration between the technological and social system is of major importance, it is Sheridan who, writing in 1989, provides a blunt view of the state of the art: "The author has been working for many years to get engineers to see the more predictable aspects of human behavior as legitimately within engineering design. It is a bigger chore to get engineers to see technology as an infrastructure of artifacts placed between people, a medium through which people communicate with themselves and each other, which both enhances and constrains how they interact. Eventually, however, it is essential that the systems design engineer appreciate the responsibility for shaping human behavior. In this sense the design of a technological system with which people interact is design of human behavior, or to put it bluntly, design of people." Sheridan's analysis still gives a primordial role to the professional, however it is a role structured in a new way, as with Hirschhorn it is seen that professionals must alter their manner of conceiving the conditions and constraints within which peoples' relationships to their work are managed. Such conditions and constraints are carriers of workplace hazards. However, he also observes great difficulties in converting fellow engineers, prisoners of the industrial paradigm, to a viewpoint that incorporates human behaviour.

Thus far in this paper we have seen consistent reference being made to the knowledge, motivations and actions of those who work with hazards. What the authors examined propose constitutes a clear break with the industrial paradigm. In sociological terms the acquisition of an understanding of the role played by the objects of disciplinary action is viewed as inseparable from the knowledge and motivation that subjects have in relation to these objects. This principle - common to the views highlighted here - can be wagered on as a necessary but not a sufficient viewpoint from which to build a new paradigm.

Any new paradigm must necessarily incorporate both elements included in and omitted from the former paradigm, be clearly structured, obey the principles of formal logic and have clear criteria for explanation. These are prerequisites for scientific progress. Let us now consider some of the proposals made by representatives for their disciplines before moving on to outline a sociological theory.

Proposals

British ergonomist John Wilson asks whether access to workers' views, as expressed in participative strategies, can be seen as a "framework and a foundation for ergonomics". He builds up his case through reference to a series of classical themes found in the literature on participation, including considerations as to the roles of knowledge and power in the process. He gives strong emphasis to the importance of the French approach to ergonomics which seeks to systematically capture worker knowledge and orientations when designing interventions. Wilson then goes on to formulate a prescription for implementation, he states that the participative "process must be voluntary, collaborative, subjective as well as objective, relaxed, non-directive (but must have direction and purpose), iterative, dynamic, flexible, allow compromises, and must aid creativity. A tall order! However, that is not all; in the wider literature we also see that the process requires adequate time, and that the participants must have access to information and external expertise and also have sufficient knowledge and experience.... To achieve all this the ergonomist must play the roles of facilitator, group consultant, technical advisor, and problem solving assistant, as well as (subtle) educator, advocate, and supporter." Participative ergonomics, although appearing in close harmony with the general lines of reflection set out by members of other disciplines, has yet to build a clear view of its contribution to new and coherent theoretical reflection within the discipline.

Other proposals are less applied in their focus. De Greene, in a recent essay, saw the need for a "new cognitive/emotional paradigm in order to comprehend and to fit the emerging systems dynamics of irreversible evolution, transformation and structural change." However, he writes of the difficulties of imagining the philosophies and designs that might best fit the overall environment of the future. "With imagination should come insight also into needed new kinds of systems indicators." In Sheridan's words: "Essentially what is called for is design of the 'social system', including organization policy, and motivational elements, jointly with the design of the technical system." De Montmollin, reflecting on the future of ergonomics, goes on to propose a programme that could "link up the current plans of anthropological ergonomics... and in this way come closer to the ideal objective, still in a stage of wishful thinking at present, of an autonomous science of work."

These proposals have been built up, as one might expect with pioneering efforts in diverse disciplinary fields, in ways that are unintegrated. However, at least four elements can be retained as building blocks of further investigations:

1. the importance for initiatives taken within the natural sciences (eg. the design of equipment and other elements of the technical system) to be interpreted from the viewpoint of the social sciences (eg. how the equipment is processed within social and cultural settings)

2. the importance of arriving at an understanding of workers' knowledge of the workplace

3. the importance of understanding worker motivations to act.

4. the difficulties in building new understandings and, from de Greene, the importance of creativity in the process.

Preamble My principal wager is that the atoms that will make up the basis of a new theoretical approach are to be found in the analysis and interpretation of the actions of individuals, as carried out in systematic fashion, in the process of executing work on a day-to-day basis. Work occurs within the context of constraints built up by prior and concurrent actions at the macro and institutional levels which articulate into the workplace. Thus, legislative, investment, cultural, biological, political, technical and economic factors acquire relevance in workplaces. The theory built up is independent of the size of the workplace or of the type of contract within which the worker is employed. In the process of producing goods and services workers enter into relationship with the above listed and other factors and in the process of their organised interaction with these accidents and ill-health are produced.

Workplace accidents and illnesses are produced systematically. Key theoretical categories of classical sociology of work can be used to conceptualise the workplace. A key concept used to understand the execution of work is that of the social relation, social relations are the modes in which peoples' relationships to their work are managed. That social relations of work produce industrial accidents and illnesses is the key proposition underlying the theory to be outlined. It is important to state here that accidents and illness are analytically separate phenomena, however they have sufficient similarities to permit them to be treated in the same breath in this paper, I shall briefly return to this point in the next section.

Sociological Theory - Introduction

A sociological theory of accidents and illness is composed of two analytically distinct moments. In the first the workplace is seen as made up of a series of 'givens' - machines, labor force, materials, trade union systems, legislative prescriptions, product markets - that are formed by the working through of social relations external to the workplace, social relations discussed previously at the macro and instituional levels. Such 'givens' are changed either through the alteration of the social relations that produce them or through policies which determine that different sets of 'givens' be assembled within a workplace. In a second moment these 'givens' are assembled in the workplace. Through their insertion into systems of social relations of work, and through the transformation of these, goods and services and industrial accidents and ill health are produced. It is this second moment that I will discuss from now on.

Social relations of work can be conceptualised as existing at three levels: rewards, command and organisation. In addition there exists a level of reality, the individual member level, at which behaviour occurs which is not explicable in terms of a theory social relations. The use of the term 'level' does not imply the existence of a hierarchy, rather the term refers to a conceptual distinction between analytically distinct systems of social relations. Although the levels are distinct from an analytical viewpoint they are interrelated. A change in one may alter the way the other operates and this may affect productivity, quality, industrial action or exposure to hazards. Questions are constantly raised in the minds of individual workers about the knowledge they have of the tasks to be performed, the orientations they have towards working with known hazards and their power to resist these. To restate what has been already laid out, peoples' relationships to the hazards of their tasks are managed through social relations and through such management accidents, ill health and well-being are produced.

Accidents and ill health can be put into a wider context if one thinks about the current emphasis to be found on 'benchmarks' within managerial theory and practice. The execution of work has, as the multiple 'benchmarks' invented to measure performance demonstrate, many faces: production, productivity, quality, scrap, waste, breakdowns etc. Managerial efforts seek to align as many of these as possible in the same positive direction, however, this does not mena to say that all obey the same logic, frequently the emphasis on achieving one benchmark leads to slippage in another. In an analgous fashion two 'benchmarks' of central relevance in this paper are health and safety. Health and safety, like any other benchmark, can be decomposed into various further sub-parts. Shannon, for example, has suggested that small accidents have different causes to big ones. Some chemical agents used in workplaces cause instant ill health the impact of which has the sudden nature that characterises accidents whereas others have long latency periods. Other measures of ill health such as medically defined stress as demonstrated by Karasek and Theorell or as in much researched and albeit ill defined (from a traditional medical viewpoint) repetitive strain injuries and certain psycopathological reactions as revealed by Dejours and Roberts have been each demonstrated to be intimately linked to the way in which peoples' relationships to certain task structures and their contents are managed.

Organisational level

A worker arrives in a workplace with adequate or inadequate knowledge to successfully complete the task, 'successful completion' may, of course, be defined in many ways. Task related knowledge is a concept too complex to investigate here, it has been studied principally by psychology and, in recent times, by ergonomics. However, task performance may be interrupted or affected by phenomena produced in areas outside the responsibility of the worker. Here we speak of disorganisation. Furthermore, where task structure is narrow and tasks repetitive the worker is faced with little variety, in certain cases these features produce the well documented negative effects of routine work.

Posession of knowledge about the hazards of one's job constitutes a first mechanism with which to avoid defined negative effects of underqualification. However, knowledge of one's job is not sufficient to avoid the other two social relations identified at the organisational level. Typically disorganisation can be combatted via the adoption of mechanisms that seal off discrete parts of the workplace from others, other methods include tight coordination, feedback and routine maintenance. Routine work can be combatted through a variety of techniques including job enlargement, job rotation, rest pauses and, where certain types of postures prove unhealthy, by exercise. All such mechanisms can be developed via two distinct strategies, via management or worker efforts.

Technical choices, and this is a constant theme in recent engineering and ergonomics literature, need to studied in order to examine how they articulate into the day to day lives of people in the workplace and thereby produce consequences. Hirschhorn details how certain tasks have been designed to be so complex that workers are simply unable to develop the level of qualifications necessary to run them and, as a consequence, accidents, physical and mental disorders are produced. He suggests that design projets explicitly consider worker capacities to learn and to run complex systems, in other words, that systems be designed in such a way as to systematically erradicate disorganisation and underqualification.

Professionals have an important role in diagnosing ailments. A keyboard worker may feel shoulder pain, an industrial doctor with a traditional clinical orientation will arrive at a different understanding to a doctor whose investigation begins by considering the worker's perceptions of her own condition. In the traditional diagnostic process the perceptions and knowledge workers have of their workworld are not seen as legitimate unless validated by the existence of visible symptoms. Diagnoses are made in a manner that excludes detailed investigations of the complaints as to the states of pain and the subjective meanings which workers attach to their cause, the investigation of the latter would be of use in revealing the mechanisms through which reported disorders are produced. The lack of a detailed examination of the nature of pain frequently results in workers' experiences being disqualified as, for example, they are told that their complaints are 'nothing to worry about'. In such a manner the doctor-patient relationship becomes a factor in both the production of a lack of knowledge and of ill health. Lack of knowledge maintains the status quo, new knowledge produces demands for change.

"Right to know" legislation is an institutional response to the existence of disorganisation and underqualification, it seeks to foster a perfect state of knowledge with regards the hazards of the job. Where workers labour with complete knowledge of all relevant hazards produced outside of their own bodies neither underqualification nor disorganisation can be diagnosed as producing their accidents and ill health. It thus becomes necessary to investigate the roles played by social relations at other levels in order to understand their actions.

At the command level the relationship between workers and the hazards of their work are managed through the use of power, this may be covert and accepted, as in the case of voluntary servitude, or overt and resisted, as in the case of authoritarianism. The operation of this level to the employer's benefit is guaranteed only through the absence of a resort to collective action on the part of the workers, resort that is capable of opposing the employer's use of power, the basis of this guarantee is the absence of integrated workgroups. Thus the third social relation at this level, workgroup disintegration, is, as its name implies, a mode of managing people's relationships to their jobs whereby interdependent workers involved in the same task lack the capacity to communicate effectively.

Workers are aware, for example, that certain chemicals have dangerous long-term consequences but define working with them as a 'part of the job', as normal. While such definitions may result, as Dejours has found, in workers developing psychopathological conditions, it can be said that the relationship between people and the hazards of their jobs under such conditions is managed by voluntary servitude. In other cases the hazards may be more immediate and the motivation to work with them come from an entirely different source - fear of punishment. Here the person's relationship to the hazards of her work is being managed by authoritarianism. Finally, ineffective integration between workers engaged in interdependent tasks that involve hazards has been demonstrated to lead to accidents. Thus, work designers must ask if jobs require interdependent work and, should they do so, if adequate provisions have been made for communication among workmates. Such an interrogation can only be replied to adequately when reference is made to real and not hypothetical practices in workplaces.

'Consciousness raising' may attempt to convince workers that hazards should not be viewed as normal conditions of work. The strengthening of union and collective power serves as a basis through which authoritarianism can be combatted. Finally, efforts may be made to improve communications within workgroups. As occurs at the organisational level, such changes can be either employer or employee produced.

In recent times, and in a clear break with the industrial safety and health paradigm, legislators in some countries have sponsored 'right to refuse work' legislation thereby attempting to give workers adequate support to erradicate those cases of accidents and illness produced by authoritarianism. A Canadian study related that unionised workers used "right to refuse dangerous work" laws more than twice as frequently as their non-unionised colleagues. In other words, the non-unionised personnel work with and do not protest about dangers that their unionised colleagues refuse to tolerate.

Rewards level

It is common in management literature to see reference to two completely different motivational techniques, the stick and the carrot. When one does not result in desired behaviour it is suggested recourse can be made to the other in order to guarantee work. When employees are aware of the hazards of their work, incentives of diverse kinds may be used to induce performance.

Work is produced at the rewards level through the workers accepting offers of recompense, be these monetary or symbolic, in return for greater effort, such effort can be obtained through intensification of work (as in piece payments) or through its extension (as in overtime work).

The rewards level is seen as producing hazardous work through three social relations: financial incentives, extended work and symbolic recompense. No single one of these constitutes by itself a social relation capable of producing ill health and accidents and the effects of each may be differentiated by worksite or by sector. Social relations must be adequately conceptualised in order that hypotheses be formulated relating to their roles in the production of accidents and illness.

People may or may not decide to work harder in order to earn incentives offered in return for effort. When workers expose themselves to greater hazards in order to earn financial incentives it can be hypothesised that the probability rises, other things being equal, of their being subject to illness and/or accidents. One particularly clear example is provided by the offering of 'danger' and 'dirt money' payments. Another is that when paid on a piece rate bases those working at routine jobs associated with the production of RSI may substantially increase work rhythms thereby aggravating their ailment.

Work can be extended in return for payment. When this occurs we are talking of a social relation that is different from financial incentives which aim to intensify work performance and, in the case under examination, promote a more intense relationship between workers and the hazards of their jobs. Subject to extended work people labour for long hours or outreach their physical capacities. The results are well documented, physical exhaustion, accidents because of fatigue or over-exposure to noxious substances.

When workers' relationships to work are managed by symbolic rewards compensation for effort is not material, rather, esteem, status, and prestige are among the incentives used to produce orientations favourable to increased work either through its intensification or its extension. The use of symbolic rewards is extensively discussed in management literature. In this way, for example, work in a hazardous environment may be carried out solely by those who, for psychological or social reasons, value their contact with such an environment.

This cursory examination of the rewards level allows us to perceive that peoples' relationships to the hazards of their jobs may be managed entirely at the rewards level. However, as at the other levels observed, both workers and managers (who act as the agents of employers) can attempt to limit the effects of this level on the production of illness and accidents.

This level refers to that part of the worker which is not organised, commanded or rewarded. It is made up of the autonomy that remains to workers after the operation of the systems of social relations which seek to structure their actions.

At this level accidents and illness are not produced in a manner that is capable of being analysed sociologically. Rather, they are produced when workers act with the context of the hazards of their jobs in manners that are autonomous from the systems of social relations. Staff selection programs, incentives, rules, and work routinisation constitute some of the managerial practices designed to limit the space for behaviour at this level.

This theory has been elaborated so as to permit science, technique and administrative measures, whatever the cultural and political contexts in which they are elaborated and applied, to be examined through concentrating on their roles in the production of various phenomena in workplaces, and among them safety and health. This type of conceptualisation thus permits the theory to be applied in many different coutries and for research to be carried out within it. This has been shown in the posters of Ester Galli and Myrian Matsuo displayed in this conference and less directly in the read papers produced at the Oswaldo Cruz Foundation in Rio and elsewhere. It is indeed correct to say that as we get closer to a genuine sociological understanding in Brazil the more the establishment refuses dialogue. The case of the recently scuttled Brazil-Sweden cooperative reasearch and action programme is absolutely scandalous and I am of the opinion that someone in Fundacentro should loose their job because of this. It is a part of a much larger and systematic struggle within this organisation which congregates at the same time some of the best and some of the worst civil servants in the country, a good medium to high level of research skills, poor direction accentuated by the corporatist spirit within the house and a refusal to open the activities to scientific evaluation.

Knowledge acquired in the natural sciences is integrated, through technologies and diagnoses, into the social world, a theory such as that outlined in this paper permits the researcher to clarify the terms upon which such integration takes place and to trace the consequences of different strategies of integration. Limitations inherent to the process of production of knowledge in the natural sciences and in the solutions applied by engineering, medicine and ergonomics are analysed by some through explicit reference to social processes. Through paying close attention to these two these two types of social analysis the grounds for developing a new approach to safety and health at work can be located. The sociological theory outlined permits that the interventions of professionals and administrators be analysed in terms of their effects on the ways in which people manage their relationships to the hazards of their work. As such it should be capable of stimulating reflexive thinking on the part of these actors, inviting them to evaluate the underlying knowledge about, motivations for, and consequences of their professional and administrative actions. Such consequences are never given from the outset but, rather, depend on how actions articulate within systems of social relations in the workplace. The integration of a diversity of themes from competence to corruption, from pro-active approaches to prevention to safety management-by-propaganda, becomes possible because the theory, as elaborated at the micro level, places worker knowledge and motivations for actions at its centre and because it is firmly anchored in a disciplinary and theoretical tradition which has made a long contribution to developing understandings of work even though it has little tradition in the area of safety and health.

Implications of the Sociological Theory for Safety and Health Research

First, it is important to recapitulate a few principles. In the same workplace different groups of workers can have their relationships to their work managed on a routine basis by different sets of social relations.

The key theoretical proposition being advanced in this part of the paper is that the production of ill health and accidents in a given workplace is inextricably related to the social relations used to manage the relationships between workers and the various hazards of their jobs. Changes in social relations are linked to changes in the production of such effects.

Acting within such a theoretical perspective a researcher must seek to establish the ways in which people come to work with hazards. Where such hazards are known to workers the researcher must seek to identify the motivations that lead people to work with them. In other words, investigation is to be carried out into the understandings that people have of their workplaces and the sense which they give to their actions. In field studies a series of questions arise, among them: are certain known hazards treated as a normal part of the job (voluntary servitude), do workers execute tasks involving hazards because they fear punishment should they refuse (authoritarianism), or do they work with such hazards because they see themselves as being rewarded for such action (financial incentives or symbolic rewards)?

From one workgroup to another, from one workplace to another, responses to such questions will differ. A research sociologist will seek to build up valid explanations of differing patterns of accident and illness causation. To be considered valid, explanations must exhibit causal and meaning adequacy, in other words, statistical (causal) relationships must be established between the existence of patterns of social relations and exposure to hazards, the explanation established must be meaningful to the workers involved. For example, should a statistical relationship be established between financial incentives and accidents such an explanation can be considered valid only if workers are shown to be motivated to work more dangerously in order to earn the incentives offered. Within this type of sociological analysis preventive strategies are adopted subsequent to consideration of the relationship established between different social relations and the production of accidents and illness at work.

Implications of the Sociological Theory for Safety Policy

This indeed is a wide topic. As we are in Brazil I shall refer specifically to this country, These suggestion are based on extensive research carried out within the type of framework elaorated in this paper, it is non-exhaustive and divided into four areas: Action (short term), research, evaluation and action of a medium and long term nature.

Short term

- erradicate slavery, certain forms of child labour, chronic authoritarianism and chronic overwork

- drastically reduce the number of workers excluded from the protection and benefits that accrues to workers in the formal economy

- implement in (non-pelego) unions centres of ÔEduca‹o sindical para a cidad‰niaÓ and with special attention to areas where workers are disorganised

- implementation of an Ôefficient CIPAÕ scheme based on the principles of sociological and related analyses which have identified Ôbest practicesÕ

- implementation of a monitoring and control system and the elaboration of viable evacuation proceedings in order to reduce the likelihood and the damage inflicted by major accidents

- Abolition of labour justice in the area and instant and bureaucratic settlement of worker claims (this may serve as a pilot project for the abolition or transformation of the labour courts)

- Begin the transformation of Fundacentro into a research and educational organisation of world-recognised excellence in the area

- Begin the transformation of Fundacentro into the major policy advice body in prevention, rehabilitaiton and compensation

- Launch two major series of integrated research projects, one of general research and one of evaluation studies. Each series of projects to be completed within 2 years and the results are to be used in constructing medium and long-term goals.

Research

- compliation of acceptable national statistics

- action research that seeks to spread the best of the spirit and practice of occupational rehabilitation as it currently exists and thereby extending its notorious sucesses to all other areas of the society where people are maimed

- a profound investigation of corruption in the area with interviews of all corrupt and gaoled officials etc. with a view to fighting it in the future

Evaluation

- analysis of the factory inspector system

- cost benefit analysis of the safety professionals

- cost-benefit analysis of the lack of regulation in the Brazilian economy (ie. of those workplaces outside of th eformal controls system)

- Employer education schemes

- Worker education schemes

- evaluation of all safety and training courses in the area

- Fundacentro its mission and its capacity to achieve it with the resources it has (note: such evaluations have been constant yet the institution has got far worse over the last few years)

Medium term

- improvement of the statistical apparatus

- Implementation of adequate and progressive reform projects based on evaluation studies

- develop a national educational and vigilance system (perhaps via internet)

- Development of TV programmes, reporting systems and propaganda which includes educational material with a sociological basis.

Long-term

- Based on the best available knowledge reform of the legal and institutional apparatus so as it reflects the culture associated with the emergence of a new paradigm. That this lead to the combat against work accidents to be seen as an element of the modernisation of the country, as installing greater participation and citizenship and producing a healthier work-life.

Conclusion 3

The prescriptions built up in this paper probably appear to be overly focused on the micro level and on workers and workplaces to be of use in developing monitoring programmes. At the present point in time case studies of workplaces constitute the only method that appears viable in the building up of such understandings. Thus, comparative studies can be conducted within the matrice of the sociological theory outlined on such important themes suggested in this paper, for example, the consequences of the insertion of design principles mooted by Sheridan or de Greene versus traditional principles, the effects of the articulation of the patient's view into a diagnostic processes compared with traditional views and comparisons of ergonomic interventions carried out under participative and traditional approaches.

Historically new theories have frequently required the development of instruments to permit their testing and the subsequent development of applications. Copernicus' theory of the earth's heliocentric orbit was only able to be confirmed from the moment that a product of the technical revolution, the telescope, was available for use. Present techniques used to monitor and predict industrial health and safety are predominantly products of the industrial paradigm. As such they appear like fishing nets ill-adapted to the demands of new seas. Theory, analysis, monitoring and eventually prevention will only be able to advance if researchers, to use Karl Weick's words, "drop their tools" - "organizational researchers" he suggests "are in a... threatened position to that of firefighters (caught in the middle of a brush fire) and face a similar imperative to drop their heavy tools or they will be overrun. To drop one's tools is simultaneously to accept mutation and to modernize remembered values or to believe the past as well as doubt it." The challenge is to build monitoring techniques that depart from a new and basic question- what social relations lead people to work systematically with hazards and what social processes sustain such activity?

A first practical consequence of this paper is contained in the suggestion that the categories of cause used herein be employed in the processes of investigation of accidents and industrially produced illness. Thus, the knowledge and motivations of human beings who suffer the consequences of the hazards would be placed at the centre of investigation and prevention efforts. In this way the proposed analysis would break radically with the industrial approach which, referring to 'unsafe acts' and 'lack of personal hygiene', excluded reference to workers, their knowledge and their motivations in explaining the production of their misfortune.

The social processes that sustain hazardous work are not, as was illustrated with reference to silicosis in the United States, only given at the micro level of the workplace. They are also produced at the institutional and macro levels. Sociology as an analytical discipline can elucidate the processes by which hazards come to be produced, come to be known about and come to be questioned. In so doing it renders society more transparent to its members. Once hazards are known about and understood widely they can become subject of research, policy development, public knowledge and contest.

The sociological approach outlined also seeks to make actors, and this is particularly relevant at the institutional level, aware of their responsibilities for their own actions. For this to occur, professionals and state administrators must necessarily develop reflexive thinking about their own actions, seeing these as produced by social relations. The understanding of the evolution and spread of such thinking and its capacity to transform practices is of particular importance.

Finally, at the macro level a new orientation has been observed, one that breaks with the monopoly accorded to scientists, whereby a conscious public is to become involved in decision making about themes involving risk. Here, indeed, is a phenomenon worthy of major attention. It too would appear to constitute a sign of the erosion of the industrial paradigm and to serve as a possible indicator of the new.