Given the way in which we have approached this paper, a word of introduction is necessary in order to provide the reader with some indication of our intentions. This special issue was set up to provide the opportunity for theorists from different perspectives to take an imaginative look at the relationship between complexity theory and narrative approaches to the study of organizations. Therefore, we should preface our paper with a brief biographical comment. Heather Höpfl is a psychologist and Professor of Management at the University of Essex, UK. Sumohon Matitlal is a Graduate Teaching Assistant and doctoral student in the Department of Accounting, Finance and Management, also at Essex. Heather Höpfl has for many years taught courses and given workshops on disasters and worked for several years as an adviser to British Airways Safety Systems on the development of the British Airways Safety Information System (BASIS). Sumohon Matital trained as an accountant and teaches on courses in both Accounting and in Management. Neither of us are complexity theorists, therefore, this paper will disappoint those readers who are looking for analysis from a complexity theory standpoint. However, where we think this paper can make a contribution to this special issue is in looking at the failure of a complex system. The paper considers how the antenarrative approach proposed by David Boje (2001) can bring to the surface issues which are normally subsumed within the rhetoric of justification and explanation which follow failure. In this sense, the intention is to provide a somewhat naïve and uncritical account of the emergence of the Bhopal Disaster and then to indicate how a narrative approach would provide a means of getting behind the post-disaster rhetorical stance and give clues as to the complex nature of the disaster in emergence. In the first draft of this paper, this account was provided without explanation by the authors in order to provide a disjuncture in the text. This was a deliberate narrative strategy undertaken in order to establish a particular account of the way in which the Bhopal Disaster unfolded and then to break with this account and to offer insights into how such a complex systems failure might be reviewed using a narrative approach. One might consider Italo Calvino’s (1981) book, If on a winter’s night a traveller as an example of a novel which seeks to impose this type of disjuncture in the expectation of continuity within a narrative structure. Here, our intention was to offer an account of what happened in Bhopal and to present some insights into the history of the disaster since complex failure usually has a long history. However, our intention went further than this. We wanted to introduce some sense of the speculative interpretation of the event, to insinuate conspiracy theory into the account, to offer descriptive passages and unchallenged assertions, to include journalistic comment alongside empirical research. The purpose here was to offer, within the narrative scope of the paper itself, some sense of the apparent simplicity of post disaster accounts and rhetoric and, optimistically, to leave the reader questioning the validity of many of the assertions which form part of this narrative. Behind this apparently simple and straightforward story of the Bhopal Disaster many questions are unanswered, many are reduced to simple causal relationships and many aspects of the disaster are not addressed. This paper seeks to provide insights into several distinct areas. First, it attempts to restore a temporal dimension to the story by pointing to the long incubation period that precedes disasters; secondly, it seeks to look at motives and motivations in the construction of the narrative of Bhopal; and, finally, and perhaps most importantly, it seeks to restore the body to the text by reminding the reader that the victims of Bhopal were flesh and blood, that this story has real casualties and fatalities, which recourse to analytical description and capture must not remove from the story. The production of narratives which describe complex failures cannot deal purely and simply in abstraction. This paper argues for an approach which embraces physicality and gives attention to the human tragedy of Bhopal. In this way our own biographies become entangled in the story. The paper is our attempt to demonstrate this entanglement as we attempt to offer our own take on the Bhopal story. Consequently, what follows is an uncritical account of the Bhopal Disaster, a discussion of the characteristics of disasters and a discussion of how the antenarrative approach might provide valuable ways of understanding and disentangling the complex narrative of Bhopal.

The Bhopal disaster outlined

“The story of Bhopal is one of unrelieved human misery. It is also a story of greed, opportunism and of sloth and rank incompetence at official level. Rarely has injustice been so visible, or so cynically unredressed” (Brown, 2002).


The Union Carbide Company was formed in 1898 by a group of businessmen from the Peoples’ Gas and Light Company of Chicago. Originally, the company provided lighting and heating services in the city using acetylene gas. In 1917, four companies, viz, Linde Air Products, National Carbon Company, Prest-O-Lite, and the Union Carbide Company merged to form the Union Carbide & Carbon Corporation, which was incorporated in New York. The name was changed to the Union Carbide Corporation (UCC) in May 1957 (Draffan, 2000). By 1960, the UCC had expanded into chemicals production, mining and metals, consumer products, electronics, hydrocarbons, ferroalloys and was a globally renowned company.

The UCC decided to move into India in the early 1930s. In 1934, Union Carbide India Limited (UCIL) was established with the Union Carbide Corporation holding 50.9% of the share capital and the rest being held by Indian financial institutions and private investors in India (The Bhopal Plant, 2005). UCIL was a diversified manufacturing company, employing approximately 9,000 people and operating 14 plants in five divisions; its shares were publicly traded on the Calcutta Stock Exchange (Chronology, 2005). “It manufactured chemical products, plastic goods, photographic plates, films, industrial electrodes, polyester resin, laminated glass and machine tools. The company also had its own fleet of seven trawlers on the Bengal coast, specializing in deep water shrimping” (Lapierre & Moro, 2003: 66-67).

The Bhopal plant

In 1969, the UCC set up a small plant in the city of Bhopal, the capital of Madhya Pradesh to manufacture pesticides. The specific site within the city was zoned for light industrial and commercial use, not for hazardous industry (Broughton, 2005: 1). The procedure involved buying stable ingredients locally, mixing them and packaging them for sale to the government for use in an anti-malaria campaign (Shrivastava, 1996: 16). According to Cassels (1993: 39) the decision to manufacture the pesticides in India as opposed to relying on imports was based on India’s goal of conserving foreign exchange and its focus on industrialization. Bhopal was chosen as the site for the plant because of its location in the centre of India, good railway connections, a large lake that provided a water supply and a good source of local labor.

In 1974, UCIL received a government license to manufacture 50,000 tons of methyl isocyanate (MIC) based pesticides each year. To cut production costs, UCIL decided to manufacture key ingredients in its own plant instead of buying them from suppliers (Shrivastava, 1996: 19). In 1981, an MIC facility was established in the existing pesticides plant in Bhopal. Located to the north of the city, the plant was barely two kilometers from the railway station and very close to a residential neighborhood.

However by 1984, the Bhopal plant was working at one quarter of its production capacity due to decreased demand for pesticides. “Widespread crop failures and famine led to increased indebtedness…” (Broughton, 2005: 2) among Indian farmers. They had become too poor to invest in pesticides. According to Shrivastava (1987: 184) due to decreased profitability, in July 1984, local managers were directed to close the plant and prepare it for sale. But no ready buyer was found. Consequently UCIL made plans to dismantle key production units of the facility for shipment to another developing country. In the meantime, the facility continued to operate with sub-standard safety equipment and procedures (Broughton, 2005: 2). Shrivastava (1987: 184) points out that although the local government was aware of the safety loopholes, it chose not to place industrial and pollution control burdens on the struggling industry fearing the economic effects of the loss of such a large employer.

The disaster

On December 3, 1984 a highly toxic cloud of methyl isocyanate[1] spread invisibly over the city of Bhopal. Of the 800, 000 people living in Bhopal (Ungarala, 1998) at that time, 3,800 died immediately[2] and more than 300,000 were injured (Ungarala, 1998). The poisonous vapor had leaked from the Union Carbide pesticide plant.

According to Cohen (as cited in Ungarala, 1998) a series of studies made five years later showed that many of the survivors were still suffering from one or several of the following ailments: partial or complete blindness, gastrointestinal disorders, impaired immune systems, post traumatic stress disorders, and menstrual problems in women. A rise in spontaneous abortions, stillbirths, and offspring with genetic defects was also noted. This incident, the Bhopal Gas Tragedy, was one of the worst commercial industrial disasters in history. Over the last 20 years, exposure to the toxins have killed another 15,000 and brought about chronic or debilitating illnesses for thousands of others for which there is no cure (Clouds of Injustice: Bhopal disaster 20 Years on).

According to the Union Carbide Web Pages on Bhopal, the Bhopal plant produced pesticides for use in India “to help the country’s agricultural sector increase its productivity and contribute more significantly meeting the food needs of one of the world’s most heavily populated regions. The plant was owned and operated by Union Carbide India Limited (UCIL), an Indian subsidiary of the Union Carbide Corporation in the United States.

However these Web Pages refrain from using the term subsidiary. Instead, Union Carbide use the phrase “an Indian company in which Union Carbide Corporation held just over half the stock” and thereby impose some distance between themselves and the fatal plant. Union Carbide was, however, not devoid of accidents in their history. Before Bhopal, the Union Carbide Corporation caused the largest industrial disaster in the US. “In the construction of the Hawk’s Nest Tunnel in West Virginia in 1934 nearly 2,000 company workers, most of them black, died of Silicosis – an occupational disease caused by hazardous working conditions” (10 Amazing Facts about Bhopal). Ironically, this was the same year that UCIL was established “when Union Carbide Corporation (UCC) became one of the first U.S. companies to invest in India” (Bhopal Information Centre, 2001)

The Union Carbide Corporation also had a plant at Cimanggis in Indonesia, where in 1978 the company doctor had described “402 employees (more than half the work force of 750), as suffering from kidney diseases attributable to workplace contamination. The doctor was asked by the company not to tell the workers that there was mercury in their drinking water or else the workers would become anxious” (10 Amazing Facts about Bhopal). Once again, ironically, the Bhopal Plant was designed and constructed in the late 1970s and the land was provided by the Indian State Government of Madhya Pradesh.

A number of questions might be asked about the role of the Indian Government in relation to the Union Carbide plant. At first glance, the ready permission to develop and site a dangerous plant appears to demonstrate a considerable neglect: inward investment being privileged above standards and safety procedures. However, it is also likely that the State Government of Madhya Pradesh were concerned about providing employment, gaining US investment and taking for granted that US standards of safety would operate. Less easy to answer is why they permitted a plant producing highly toxic gas and fatal to human beings to be sited so close to a densely populated city.

The disaster unfolds

According to the Union Carbide Corporation a large volume of water was introduced into the tank storing the methyl isocyanate gas which triggered a reaction that resulted in the gas release. A committee of experts appointed by the Indian Government reached a similar conclusion. The official response of the UCC on the tragedy also states that “independent investigation by engineering and consulting firm Arthur D. Little, Inc., concluded that the gas leak could only have been caused by deliberate sabotage; someone intentionally connected a water hose to the gas storage tank, causing a massive chemical reaction” (see Browning, 1993).

Other versions of the cause of the mishap (The Real Story: Midnight 2-3 December 1984) suggest that during routine maintenance operations in the methyl isocyanate (MC) plant, around 9:30 pm, a large quantity of water entered storage tank no. 610. This triggered off a runaway reaction resulting in a tremendous increase of temperature and pressure in the tank and 40 tons of MIC along with Hydrogen Cyanide and other reaction products burst past the ruptured disc and into the night air of Bhopal at around 12:30 am. Both versions of the precipitating event which caused the disaster contend that safety systems were grossly under-designed and inoperative.

According to the chemical industry archives this claim is justified. “Documents recently uncovered in litigation [Bano, et al. v. Union Carbide Corp & Warren Anderson, 99cv11329 SDNY, filed 11/15/99] demonstrate that Union Carbide cut corners and employed untested technologies when building the Bhopal plant.” These documents reveal that the methyl isocyanate process developed by the UCC had had only a “limited trial run.” However the UCIL found the level of business risk acceptable and the UCC concurred. The company forged ahead with the unproven design and the untested technology knowing that it posed a “danger of polluting subsurface water supplies in the Bhopal area” according to a Carbide document marked “Destroy after use” (The Inside Story: 18th Anniversary of the Bhopal Disaster).

What is even more striking is the fact that senior factory officials knew of the lethal build-up in the tank at least one hour before the leakage, yet the siren to warn the neighbourhood communities was not sounded until more than one hour after the leak started (The Real Story: Midnight 2-3 December 1984).

The aftermath

The immediate human impact of the disaster in terms of figures has been considered earlier in this section. Survivors of the accident report that all that they could see around them were bodies. “People were dying all around” says Mohammad Owais, a volunteer at Hamidia Hospital, Bhopal, India (Clouds of Injustice: Bhopal disaster, 20 years on). According to a survivor who was running to the hospital in a desperate effort to escape the fumes and save her two kids:

“Lots of shoes and shawls and other clothing were strewn about. White clouds had enveloped everything … lots and lots of people were running, screaming for help, vomiting, falling down unconscious. My children were vomiting. I was not able to control my bowels. Faeces were running down my legs. We had just one thought in mind and that was to reach the hospital. At Bhopal Talkies crossing we all had fallen on the ground. I was two months pregnant at the time. I had a miscarriage right in the middle of the street, my body was covered with blood. There was blood all over. I was unable to control my bowels and the faeces ran down my legs, mixing with the blood…” (International Campaign for justice in Bhopal: That Night).

During the days that followed “people had to dig three tiered graves since there was no option but to pile up one body on top of another” (Bhopal Disaster still unresolved 20 years on: Greenpeace). The hundreds of thousands, who survived, however, face a fate worse than death. They suffer from acute breathlessness, brain damage, menstrual chaos, loss of immunity; “some journalists have dubbed it chemical AIDS” (The Real Story: Midnight 2-3 December 1984).

The official response from the UCC states that a team of medical experts were immediately deployed to Bhopal with medical supplies and equipment. A disaster program was started to study effects of over exposure to MIC. The UCC established a fund for victims of the tragedy – the (UCC) Employees’ Bhopal Relief Fund. The official response also highlights the facts that in the days, months and years following the disaster, Union Carbide provided approximately $2 million in aid to the Indian Prime Minister’s Relief Fund, provided a $2.2 million grant to Arizona State University to establish a vocational-technical centre in Bhopal, provided an additional $5 million to the Indian Red Cross, offered an initial $10 million to build a hospital in Bhopal (Bhopal Information Centre, 2001).

Five years after the disaster, in May, 1989, Union Carbide and Union Carbide India Limited (UCIL) entered into a $470 million legal settlement with the Government of India. The settlement was affirmed by the Supreme Court of India, which described it as “just, equitable and reasonable,” and settled all claims arising out of the incident. Ten days after the decision, Union Carbide and UCIL made full payment of the $470 million to the Indian government (Bhopal Information Centre, 2001). The official response emphasizes that “the settlement award was much larger than any previous damage award in India” (Bhopal Information Centre, 2001).

The responsibility for disbursing the funds lay with the Indian Government which delegated the same to the Madhya Pradesh State Government. Figures from the Bhopal Gas Tragedy Relief Department (internet) show that a compensation of only $230 per deceased person was awarded to his/her family.

Other sources, for example, The Bhopal Express, suggest that the first suit filed against UCC claimed damages up to $15 billion. Later the Indian Government unexpectedly assumed the sole power to represent all the victims, filed a suit for upwards of $3 billion. Four years after filing the suit without informing the victims, the government settled for a sum of $470 million, nearly one-seventh of the original claim. It continues to ignore the repeated directives of the Bhopal court to extradite Warren Anderson and other senior officials of Union Carbide who face charges of manslaughter and other criminal offences (The Real Story: Midnight 2-3 December 1984). Once again the political dimension of Bhopal comes into question. Why did the government agree to such a lower figure? Why does it hesitate to take action against these officials? These questions are a source of speculation but remain unanswered.

Even 20 years later, according to Amnesty International (Clouds of Injustice: Bhopal Disaster 20 years on) the survivors still await just compensation, adequate medical assistance and treatment, and comprehensive economic and social rehabilitation. The State Government fails to provide basic medical care and health services. It has not set up any effective mechanism to ensure that compensation reaches the survivors before it is too late. Claims are still being processed by the state government officials. Next to nothing has been done to rehabilitate the survivors with hygienic living conditions and appropriate livelihood options (The Real Story: Midnight 2-3 December 1984).

The plant site has still not been cleaned up. As a result, toxic wastes continue to pollute the environment and contaminate water that surrounding communities rely on. New environmental studies indicate that tons of toxic materials dumped at the old plant have now seeped into the groundwater, affecting a new generation of Bhopal citizens (Baldauf, 2004). Women and children are beaten up by Bhopal police for demanding that local politicians should obey a ruling of India’s Supreme Court and provide them with safe drinking water (International Campaign for Justice in Bhopal). According to one of the survivors:

“We have to travel at least two kilometers to get clean water… My health is so bad that it prevents me from carrying the water I need from there” (Clouds of Injustice: Bhopal Disaster 20 years on).

Dow Chemical, which bought Union Carbide in 2001, refuses to take responsibility for the company they bought, living up to its history of valuing profit more than human life (Greenpeace International).

If you happen to pass by the derelict Union Carbide pesticide factory in Bhopal, nothing would strike you as out of the ordinary. It looks like any other abandoned factory. Rust has set in and creepers are growing over the machinery. The only sign of life is a bored watchman who is not sure what exactly he is supposed to watch.

Twenty years on, most aspects of the Bhopal disaster are far from resolved.

The construction of the narrative of Bhopal

The story of Bhopal as presented above is discontinuous and in many ways unsatisfactory. It has been assembled. It is not coherent. The plot-line is approximately chronological but there are digressions and sidesteps which make the trajectory of the plot erratic. Levels of analysis shift. The story offers spurious coincidences with all the enthusiasm of the best conspiracy theories. Much of the detail supplied as ‘evidence’ is dubious and perhaps spurious. The reader has to make judgements. Some of the accounts are from the internet and are sensationalist, others are from serious pieces of scholarly work and based on empirical evidence. References are incomplete or missing. The reader is not supplied with enough information to be able to decide which is which. Some accounts are formal others informal. Some stories are personal and tragic, others cold and impersonal. Some accounts are written to serve a distinct political agenda or to support corporate impression management. They reflect different hypothetical worlds and standpoints. References are used to indicate sources and to lend authority to the force of various arguments. Again, little distinction is made in the text between the relative significance of some accounts vis á vis others. Some accounts emerge from the poetry of the tragic, others from the rhetorical demands of justification. The analysis reflects this shifting as it does between vernacular accounts of the experience to sophisticated legalise about responsibility and response. The story is replete with different motivations, judgements and justifications. It is a narrative in formation. A positivistic researcher might struggle with the piece as poor science. A more interpretative researcher might comment on its incoherence, the inadequacy of its description: that it is not the full story. This paper seeks to take forward some of these issues in order to evaluate whether or not narrative approaches have any contribution to make to the study of complexity and, in this case, in the context of catastrophic failure.

Science and simplification

Jack Cohen and Ian Stewart (1994) “have argued that the role of science is to seek simplicity in a complex world.” They argue that they are concerned with what they describe as “the remarkable regularities and patterns” that are to be found in nature and go on to argue that science is concerned with identifying the “laws of nature” that govern such patterns. However, and what is interesting here, they invite us to be cautious in adopting such a “comfortable” view of the world and rather to look to complexity theories and patterns of emergence in order to challenge the “explanation paradigm” of the sciences (Cohen & Stewart, 1994: 648). They suggest that there is a lack of a “coherent ‘formal’ theory of complicity” and say that “without it, many of the most important questions for science cannot even be framed coherently” (1994: 648). Rather then, they direct our attention to complexity theory in order to give emphasis to the idea “that highly complex interactions taking place in large populations of systems can conspire to create large-scale but simple patterns” (1994: 649). In a recent private conversation with Jack Cohen, he argued that these patterns can emerge in entirely predictable ways so that it is the case that events take place, as he put it, “because the story calls for them.” This shift from the paradigm of simple explanation to complexity involves a shift towards the narration of patterns of emergence. At this stage it is useful to consider the contribution of Boje’s (2001) work on ante-narrative since this provides a way of linking Cohen’s meta-rules and myths with the idea that of the process of narration.

For Boje (2001) stories and narratives are not the same. “A story is an ante state of affairs existing previously to narrative … narrative adds plot and coherence to the story line” (Boje, 2001:1). Thus stories for Boje, are antenarratives – “fragmented, nonlinear, incoherent, collective, unplotted and pre-narrative speculation” (Boje 2001:1). When a story, i.e., an antenarrative, becomes a narrative, the narrator imposes “counterfeit coherence and order on otherwise fragmented and multi-layered experiences of desire.” The account of Bhopal at the beginning of this paper resembles a narrative in terms of its through-line of argument but, in fact, it is a series of disjunct accounts and fragments which have been given short-lived coherence by a tenuous chronology. The antenarrative analysis is particularly relevant in relation to the analysis of the post disaster period, because of the taken for granted assumptions which are laid bear by the disaster. In this case, the accounts of survivors and the contrasting rhetorical stance of Union Carbide show themselves to be different strands of story. “Antenarrative is constituted out of the flow of lived experience … it directs our analytic attention to the flow of storytelling as a sense making to lived experience” (Boje, 2001:3). This is particularly pertinent to a post disaster situation in which the predictability and expected patterns of continuity no longer hold. The justification and rhetorical position of management is as much a dimension of a story emerging from lived experience and it is for the direct victims of the tragedy. It is also pertinent to the way in which Boje’s notion of ‘counterfeit coherence’ becomes evident in the competing interpretations which are offered of the event in the immediate aftermath and, indeed in the case of Bhopal, have continued to be offered over the two decades which have passed since the disaster. The paper seeks to focus on the “multi-stranded stories that lack collective consensus … where characters collide rather than interact” (Boje, 2001: 5). This is the period in which strands of stories, fragments, tragic accounts, personal misery are thrown into flux and competition with the formal, rhetorical and political narratives of the organization, international opinion, and the response of the Indian government. By disentangling the formal from the informal, the personal from the impersonal and the embodied from the rhetorical, antenarrative provides an opportunity to perform the unravelling which complexity theory calls for in order to gain a better understanding of the relationship between systems and their components and the human tragedies which simple explanatory accounts tend to preclude. Clearly then, antenarrative, by its very nature, opens up a political agenda, in this case, not least the catastrophic failure of the Bhopal disaster itself. This is without doubt one of the major disasters of the twentieth century and the circumstances which led to the failure deserve some comment. However, the analysis here is primarily concerned with the desire to find a coherent narrative for the disaster. That is to say, the intention is to provide some insight into the post disaster stories and their own complex strands as they become occluded by the coherent privileged narratives of organizations and governments. Certainly there is scope here for an examination of the complexity approach adopted by Cohen and Stewart in relation to the approach adopted by Boje for examining the multi-stranded patterns which emerge in stories of organization failure and disasters.

Towards an interpretative environment

First, however, it is instructive to consider how the narration of catastrophic failure has changed over the past twenty years. The issue of risk has gained widespread significance in the past twenty years. The world has witnessed a series of major disasters affecting such diverse technologies as nuclear installations, chemical plants, oil tankers and ferries, railway networks, oil platforms, commercial and military aircraft and an escalating degree of actions by terrorists. Such disasters are no longer remote tragedies but events which global telecommunications thrust into the immediate experiences of a global audience (Boje, et al., 2004). Tragedy has become both a source of fear and a consumable spectacle for a world-wide audience. Consequently, there is a desire for a greater appreciation of complexity and for ‘solutions’ to complex problems. The challenge is not so much to address the issues which can be specified, albeit imprecisely, but rather to give attention to the unknown and unknowable aspects of situations. Of course, this is a formable task since mere awareness of complexity or the nature of what is hidden does not, in itself, change complex situations or make them safer. Studies of the relationship between knowledge and its tacit counterpart have thrown some light on experiential dimensions of systems. However, what seems to be repeatedly ignored in such analyses is the competing status of explanations, accounts and experiences. This is where the strands of story which come together in both the events preceding a disaster and those which follow can be used to distort the singular spectacle of disaster as presented and consumed.

Common characteristics

First, it is useful consider how the understanding of disasters has changed in recent years and to point out that despite the obvious differences in the industries involved and their technologies, it has become apparent from the analysis of disasters that, at a contextual level, there are many common characteristics (Reason, 1990). As a result, recent attention has been given to the socio-technical aspects of safety systems, to the complexity of the contributory causes in accident analysis, to the multiplicity of ways in which systems can fail, to the predominance of human factor contributions to failure, to perceptual and information difficulties and, not least, to the appreciation of the historical dimension, the fact that disasters often have a long incubation period. This widening of the boundary around safety issues has resulted in a move away from what Toft has described as a “propensity to look for simple causal solutions … shaped by the technical concerns of the engineering community” (Toft, 1992) towards a commitment to the recognition of the social and organizational context of incidents and accidents. This, in itself, has been a significant move in terms of the analysis of disasters and catastrophic systems failures. However, most research has been conducted from an organizational perspective. Human factors have become more prominent and there has been much attention to the interpretative environment of systems and systems failures. Some organizations have adopted approaches derived from studies of the Learning Organization, for example, British Airways Safety Services, others have been persuaded that a change in organizational culture would bring potential systems failures to the surface more readily. Such approaches are laudable and useful. However, in Boje’s terms they are concerned with coherent explanations, with producing narratives with structure and plot-line, and in Cohen and Stewart’s terms with simple explanations, with science and all its attendant problems of reductionism. In other words, organizations seek to either apply science to identify causal relationships, albeit now complex causal relationships, or else they are concerned with providing a good story: frequently the rhetoric of justification. However, it is often the very desire for resolution and explanation which precludes the possibility of a deeper understanding of what lies behind the potential for catastrophic failure.

The unknown stories

The ways in which the unknown and undesirable might be construed and conceptualized is important in order to explore the dynamic tensions between those things which a scientific approach to complex systems might include, and those aspects of the system which are destined always to remain outside the scope of precise data capture, but which feature significantly in the interpretive domain of the broader system. This appreciation of the dialectics of complex systems, between the rational and irrational aspects of systems, between those categories which can be used to capture and aggregate data and those which remain elusive, is important from the point of view of making transparent those aspects of systems which are irreducible and, therefore, potentially the most threatening. In this respect, the power to construct, define and manipulate information into categories and classes of answer deserves attention. It is important in relation to the construction of the notion of ‘risk’ to consider the power both to define risk and to issue authoritative statements about the nature of risk. In other words, decisions about risk in context should be seen in terms of the socio-political environment in which they occur. It could be argued that this is clear in the accounts of deplorable safety standards at the Union Carbide plant at Bhopal that very different standards of safety were being applied than would have appertained to a comparable plant in the USA. Consequently, the power to define what are acceptable levels of risk, and to do this in relation to specific locations and groups, demonstrates the extent to which the power to define and judge risks is political.

The ability to classify information as carrying this or that risk is often associated with the rhetorical position of the organization which is making such statements. Hence, the capacity to construct statements regarding risk rests on a power relationship within organizations and between organizations and the outside world. In this way, one definition of a situation may carry more significance than another. This may be the result of a hierarchical difference or a difference in apparent expertise. Workers at the Bhopal plant were not told about the lethal toxicity of the gases they were working with. Poor maintenance and unacceptable practices were not dealt with by the company. Even after the disaster medical workers were unable to advise the public because they themselves were not aware of the seriousness of the situation they were dealing with. The point being, that employees, townspeople and medical workers were all excluded from the knowledge of the dangers presented by the plant. Therefore, where one group has the power to define, another group is disempowered: one story prevails and transforms itself into narrative and other stories remain personal, fragmentary and lost.

The power to offer explanations

The construction of definitions and social meanings rests on the use of language and ritualistic behavior, on territorial and spatial division, on hierarchical arrangements, on theatricality and on the exclusion of certain people from the place where definitions are created. The construction of risk is such that categories of acts and actors are created and then responded to as if such constructions were objective realities. In the case of Bhopal, Union Carbide placed considerable emphasis on the fact that the plant was designed and constructed using Indian consultants and workers and was operated by an Indian plant management team. By implication, they suggest that Indian design and construction is inferior to American design, that Indian plant operation is not as thorough or sophisticated as American management. By this rhetorical manoeuvre Union Carbide insinuates two distinct meanings. First, that the American parent company had very little to do with the tragedy and, secondly, that Indian management and consultants are inferior vis á vis their American counterparts: that things would have been different if the plant was under American control. In political parlance, this type of rhetorical shift is nowadays generally referred to as spin. What is being spun here is one of the yarns of antenarrative into a strong narrative thread. It is a yarn which is destined to compete with those small fibres of stories offered by managers, workers, local people and medical workers whose stories tell of repeated complaints about neglect, poor procedures, inadequate maintenance, lack of information. Corporate definitions not only seek to compete with alternative stories but frequently, as in the case above, to discredit them.

Latent failures

In the early 1990s, a body of work emerged to encourage the investigation of hidden factors in catastrophic failure. Reason’s work in particular, following earlier work by Turner (1976), received considerable attention for its emphasis on what is hidden, on ‘latent failures’ (Reason, 1990: 28) which may only become evident when they occur with a “precipitating event” (Turner, 1976) which causes the system to fail. Moreover, Reason contended that “there is a growing awareness … that attempts to discover and remedy these latent failures will achieve greater safety benefits than will localized efforts to minimize active failures” (Reason, 1990: 476-7), for example, in the nuclear industry, failure to perform necessary maintenance activities, that is, latent failure has played a major role in incidents and accidents in nuclear installations (Rasmussen, 1980). This had clearly been an aspect of the Bhopal disaster. Part of the approach to the detection of latent failures requires an understanding of where the power to define the nature of the activity and its attendant risks actually lies. The need to attend to what is hidden is well understood by theorists working on the nature of risk. However, the extent to which it is possible to expose what is deferred by the privileging of specific interpretations is a matter of debate. It is here that Boje’s antenarrative makes a valuable contribution. Boje’s assertion that “A story is an ante state of affairs existing previously to narrative … (and that) narrative adds plot and coherence to the story line” (Boje 2001:1) provides a way of conceptualizing a situation in its raw state: of maintaining a plurality of perspectives on a situation before a prevailing narrative emerges. In disjunctive situations, from the conversational to the catastrophic, taken for granted situations are laid bare, assumptions do not hold, multiple possible interpretations co-exist, before narrative structures are imposed, coherent plot-lines offers, competing definitions excluded.

There is a need to see behind the rhetoric of complex systems and to expose the tacit assumptions, privileged constructions and authority to define, in order to permit alternative ways of understanding a problem, or the way in which a social situation is defined. The need to give greater attention to data which is not immediately apparent, and which does not yield easily to data capture, has a growing currency with researchers of complex systems. What is concealed behind the construction of appearances, of order and normality, may resist attempts to interpret, expose or deconstruct. The power of such factors, however, is considerable and continually threaten to subvert understandings which organizations may cherish about themselves as numerous studies of major disasters have shown.

In a similar way, Turner’s (1976) influential work on the causes of disasters argued that large-scale accidents have an “incubation period” in which there are a series of unnoticed events which are likely to run counter to established beliefs about the way that the system operates or that risks are defined. Turner encouraged safety researchers to concern themselves with “the cultural disruption which is produced when anticipated patterns of information fail to materialize” in order to develop an appreciation of the way in which individuals “gradually come to develop and rely on a mistaken view of the world” (1976: 193). “The problem of understanding the origins of disaster is the problem of understanding and accounting for harmful discharges of energy which occur in ways unanticipated by those pursuing orderly goals” (1976: 201).

The incubation period ends when some precipitating event draws attention to the discrepancy between the environment as it is believed to be and the environment as it actually is. This forces into the open the “hidden, ambiguous or anomalous events which have accumulated during the incubation period” (Turner, 1976: 201) producing a sudden shift in information levels. Consequently, Turner argues that sensitivity to information, including informal accounts and stories, is vital to the prevention of disasters. However, this is more difficult than it might at first seem. Some information is completely unknown, some may be known but not fully appreciated; some information may be available to some members of the organization but not to others sometimes because of political asymmetries, sometimes because of discontinuous communications; some information may be available but cannot be appreciated within current modes of understanding (1976: 195).

What Boje’s (2001) analysis contributes is a way of analyzing stories in order to look for oppositions. It invites researchers to look beyond surface appearances and to question dominant story lines, to be sensitive to absent voices, absent accounts. The antenarrative approach draws attention to issues of hierarchy and the assumptions that accompany organizational structures. Boje advocates that such structures should be reinterpreted in order to give voice to those accounts which are missing or ignored in relation to privileged accounts, for example, corporate positions versus those of employees. By taking this approach, Boje suggests that it will be easier to get access to alternative stories and accounts. Alternative plot lines should be considered so that, in the case of Union Carbide, for example, the righteous and justified position of the company, a heroic plot-line, might be considered from a different angle: as tragedy, for example. Finally, Boje advocates an approach which tests the limits of the rules to find out at what point they become absurd. In the case of Union Carbide this can be applied in reserve for example, to ask how far the company might have permitted lax safety procedures to extend. All these suggestions for antenarrative analysis indicate a way forward in terms of understanding complex systems failure. They provide a way to surface hidden meanings and assumptions; to change perspectives in order to see narratives in formation.

The story of Bhopal presented here is not intended to be a full or comprehensive account. Rather, it is provided in order to move closer to an appreciation of what antenarrative can reveal. To hold onto incoherence is the challenge here. It is necessary to hold onto the various strands of stories and not to be seduced too quickly by a desire for cosy and coherent narratives. Frequently, in disaster analysis, the co-existence of multiple and fragmented stories is paramount to gaining an insight into the complexity of the system which has failed.