Differentiating Exposures

After a disaster, one of the most difficult problems is understanding cause and effect—explaining which health issues occurred because of the disaster and which might have occurred otherwise. This is particularly true of an industrial disaster, and this particularity has been discussed for decades, most notably in the writing of Ulrich Beck. Beck (Reference Beck1992) argues that the uncertainties engendered by manmade hazards will increasingly overwhelm civic responses, destabilize private insurance systems, and undermine official scientific authorities. In his analysis, technological harms become globalized harms, in both their source and their consequence. As Beck has emphasized, these new, globalized, uncertain harms—harms with no clear single cause and thus no sense of who owes or is owed redress—meant that nation-states no longer had a monopoly over the assessment and quantification of injury. This can be contrasted with an analysis through the lens of biopower, wherein the quantification of state technical experts and agencies remains deterministic of who is made to live or left to die (Foucault Reference Foucault1990, 138), regardless of the accuracy of their biometric data.

Yet the history of Bhopal has conformed well to neither Beck's predictions nor Foucault's systems. While the disaster was indeed global in nature, it was successfully made local and governmental in terms of both science and relief efforts, with the liable transnational company disappearing behind the corporate veil. And though knowledge has indeed been power, it has been deployed mainly through its strategic absence; bio-obscuration rather than biopower. The case of Bhopal shows that major state investment and commitment do remain central to scientific studies after a disaster, and that the state will powerfully influence the aftermath through resources, rules, or indeed negligence. Long-term epidemiological studies in particular require an appropriate scale, resources, time, and a favorable political climate. The state is necessarily an important actor, if not always a trusted authority.Footnote ² Although Bhopal is an oft-cited illustration of the problem of transnational corporate accountability, that connection has been unfortunately marginal to the experience of survivors: the central role has been held by the Indian state since the 1989 civil settlement, and speaks to the way in which it became simultaneously a hyper-local, irreducibly “Asian” catastrophe (Oakes, this issue). For the vast majority of survivors, meanwhile, it has never for one moment left Bhopal.

One Night in Bhopal

The Bhopal Gas Disaster was a massive leak of chemicals, mostly toxic methyl isocyanate (MIC), from the American-owned Union Carbide Corporation (UCC) pesticide factory in central India during the night of December 2–3, 1984. It was not without some precedent in Asia and beyond: survivors of Hiroshima (1945) and Minamata (1932–68) were still coping with the aftermath of these tragedies in Japan; the chemical leak in Seveso, Italy (1979) had given a tiny suggestion of what Bhopal might be like; and Chernobyl (1986) would quickly follow it in the annals of catastrophe. But in its time, Bhopal was utterly unique. MIC was a manmade chemical, a precursor product used in the production of carbamate pesticide. The company knew very little about the consequences of human exposure, the medical community knew even less, and many of those living next to the factory were completely unaware of the potential toxic hazard.

On the night of December 2–3, the gas, leaking from a ruptured tank on the factory grounds, moved across the city in a slowly dissipating cloud that mimicked the effects of a deadly gas chamber. It left behind a dramatic swath of death and human misery—whole families staggering, vomiting, and dying in the streets among a rush of others blinded by the gas in the dark; animals crumbling or falling where they stood; birds falling out of the trees; plants turning brown; milk turning green.

Any description of the event made clear the lethality of the chemical. Still, doctors, scientists, and survivors continued to argue for decades over MIC, and over whether the gas had broken down into other chemicals—most controversially hydrogen cyanide. But despite some advanced toxicological science, the basic statistics of the disaster, such as the number of fatalities, were and would remain debated. Estimates of deaths in the immediate aftermath of the leak ranged from below 3,000 (by the government's accounting) to over 10,000 dead (by independent observers; Amnesty International 2004). This imprecision was at least partially because the primary population of exposed people was from the old city of Bhopal, and the worst affected were those living in poorly documented slum areas near the factory. Although it was not true that the slum area was entirely undocumented—as land permits had recently been issued—the badly affected parts of the city did have high levels of recent migration and a railway station where many homeless and indigent slept.

The first scientists charged with accounting for the harm caused by the exposure were doctors and pathologists who knew little of what they were dealing with beyond the evidence carried in the bodies accumulating around the hospital. Still many had grave concerns from the beginning that, in addition to the survivors’ immediate symptoms of exposure (which included a range of ills, mostly affecting the respiratory, ocular, gastrointestinal, and neuropsychiatric systems; Sriramachari Reference Sriramachari2006, 3), they might develop other diseases, such as cancer or birth defects, in the long term, either as a result of genetic mutation or through the persistence of chemicals lodged in the body after the exposure. A doctor with the ICMR commented to the press in 1985 that the studies on gas exposure should “be continued over a long term, as in the Hiroshima study” (Morehouse, Subramaniam, and Citizens’ Commission on Bhopal Reference Morehouse and Subramaniam1986, 31), which had been ongoing for nearly forty years.Footnote ³ But there was little centralization of demographic information, and almost no preexisting medical records exposed. As many sick lay starving and dying in their shanties, there was no clear leadership and no one seemed to know quite where to start.

Tracking a Disaster

In the affected communities, the situation was truly dire. The cloud of gas that escaped from the UCC factory that night in 1984 spread over an area of forty square kilometers that had a resident population of nearly half a million people. Many were unable to leave home or go to work. There was no food to cook, and there was no one to cook it. And the government response was not swift. Although official and volunteer relief teams distributed analgesics and painkillers in the streets and the state government eventually began to distribute milk and food rations, the medical crisis developed quickly into a humanitarian crisis. Families who had been dependent on the income from day labor ran out of necessities almost immediately. Yet most of these groups had access to only these emergency measures for a long time; the financial compensation took years to be distributed.

In the meantime, the nongovernmental sector jumped into action. Volunteer scientists arrived to attempt a quick assessment of the disaster's consequences. They conducted three studies within the first week and seven more within the first year (Narayan Reference Narayan1990), though all were crippled by lack of funding and their relatively tiny cohorts: the cohorts ranged from 33 to 1,109 individuals to represent a diverse affected population estimated at 200,000–500,000. So although all these studies found significant pathology and morbidity among the survivors (and collectively, these ad hoc studies established an important, if difficult to instrumentalize, baseline for injury), statistical significance was difficult to establish, not least because each researcher had to produce his or her own definition of what a “gas-exposed person” was. As yet, there was no official definition. The state, in the form of its premier research body, the ICMR, was the only group with the resources to establish a major cohort and the authority to define what constituted “gas exposure.” However, the early studies clearly highlighted the need for more research, illustrating the inadequacy of the ICMR's initial efforts despite an early influx of funding.

One study author complained bitterly about the situation in a 1985 op-ed. He accused the ICMR not only of failing to do important research, but also of failing to gather and share basic demographic toxicological and population data, actually impeding the research efforts. In his view, government research in Bhopal was a hospital-based boondoggle; overnight, a “research empire” had come into existence, an “empire with no clothes … built on the corpses and comatose bodies of the Bhopal victims” (Banerji Reference Banerji1985, 1330).

In Bhopal, most of the over-extended staff at the Gandhi Medical School, many of whom became the de facto core of the ICMR center in Bhopal, were in fact working hard. Many were obsessed with understanding the toxicology of the event and the link between exposure and symptoms. This was crucial from a legal standpoint, but it was far from straightforward. The responsible corporation was withholding information, samples of the gas were nearly impossible to come by, and clinical data was enormously difficult to parse. In hospital wards, MIC exposure could mimic the symptoms of tuberculosis and other endemic illnesses, undermining some patients’ claims to special treatment as disaster victims; the limited toxicological data available to these clinicians did not necessarily support the victims’ claims of exposure, especially because these doctors knew nothing of the mechanisms of comorbidity or the effects of multiple exposures in this case. There were also debates over unique markers, such as high levels of urinary thiocyanate; these markers indicated cyanide poisoning, which many argued ought to have been treated with the known antidote sodium thiosulphate, but neither testing for thiocyanate nor treatment with thiosulphate was ever scaled up beyond a few hundred patients.Footnote ⁴

Many activists and survivors, however, saw this focus on the mechanism that caused the illness (rather than on the epidemic of symptoms) as reductionist. They complained that the lab scientists and hospital-based doctors were avoiding the human chaos going on in communities and streets in favor of a more sanitized medico-legal question—the question of causality. This dynamic unfortunately pitted doctors against a desperate population who knew only that they were suffering and neglected. The medical professionals, under pressure and untrained for this kind of disaster, began to resent this the swell of survivors presenting with difficult-to-define complaints; they began to use words like “imposter” and “psychosomatic.” Only epidemiology could rescue the gas-affected from the margins. And eventually, it seemed that it would. Six months later, in the summer of 1985, the ICMR's prospective epidemiological study finally got underway, promising to discriminate the effects of general suffering from the suffering caused by the gas to the general population.

Who Counts?

The Bhopal Gas Disaster posed an enormous methodological challenge to the Indian scientists: indeed it would have been an enormous challenge anywhere. The scale of the event was unprecedented in India, in Asia, and in the entire world. Still, the eventual practicalities were straightforward. The ICMR's fieldworkers walked through some of the gas-affected neighborhoods in Bhopal. In their door-to-door canvas, they collected base data from each family (names, ages, occupations, etc.), and then screwed a metal plate with a cohort number to the outside of each home. In doing so, they produced the first official definition of gas exposure: you were gas-exposed if you had a metal plates affixed to your door (though, importantly, not all gas-exposed areas were covered by the study) (Sriramachari, Reference Sriramacharin.d., 19–22).

The study designers also assigned each neighborhood an exposure level, which set the ICMR's study apart from most other research and relief efforts. The ICMR's early clinical studies (as well as the other independent epidemiological studies) had shown that a range of factors—geographical location at the time of the leak, respiration rate, housing quality, age, and preexisting conditions, among others—were correlated with exposure, which was strongly correlated with injury. So the ICMR's epidemiology team, led by Dr. M. P. Dwivedi, created a tier of exposure levels. Their system established a tripartite division between severe, moderate, and mild exposure, based on the mean mortality rate in each neighborhood in the first few days after the disaster. (They also had a “no exposure” tier for the controls.) It worked as follows: the wards (electoral districts) that had an average death rate of 22 per 1,000 residents (range 20–23.8) in the first two days after the accident were marked “severe,” “moderately exposed” wards had 1.33 deaths per 1,000 residents, and wards with “mild” exposure had less than one death per 1,000 residents in that time period (0.20/1000) (Sriramachari Reference Sriramachari2006, 20). The study registered 95,000 people, 16,000 of whom were controls. It was not comprehensive, but it was a vast undertaking.

The ICMR tiers were based on averages of averages, calculated from tenuous accounts of (hospitalized) deaths alone. They did not account for the mortality that continued for years afterwards, but, importantly, attempted to mimic quantitatively the path of the gas cloud itself through the city. The ICMR's division assumed that exposure itself was both suffering and harm, and that the criteria for being exposed was to have been in a certain place at a certain time.

Amazingly, this version of exposure was and would remain a radical one, completely different from the definitions later used to determine compensation level or hospital access. Despite the many criticisms of the ICMR, therefore, it is important to note that its definition of gas exposure was the only one to acknowledge exposure itself as a kind of harm. Tragically, however, the most significant impact that the ICMR study would end up having on the lives of most survivors would come from its missing findings, not its insightful results. The first (and only) decade of results would not be published for over twenty years. This was rendered doubly tragic because the government hospitals that were beginning to provide treatment to survivors made no effort to track their patients’ progress or assess the severity of their exposure, nor were they mandated to do so. Doctors treated each visit as a unique event, giving treatment based on presented symptoms, and did not retain copies of patient records in the hospital, nor, often, give them to the patients. Monitoring, which could have been a secondary function of the ICMR study, also failed. In a few cases, data collected as part of the epidemiological study was used in claims courts to help determine whether a given person deserved financial compensation for their injury (Singh Reference Singh2006). The establishment of the cohort was therefore a major, if flawed, achievement. By 1986, the study began to accrue the promised data, the data that could have made the corporeal mystery of the gas both visible and traceable. But it was not to be.

Secrets

The initial findings of the epidemiological study were striking, but they were not accessible to the public. The secrecy was theoretically a way of protecting the legitimacy of both the gas victims and the state from a hostile legal adversary. UCC's lawyers were threatening to challenge each and every individual claim of illness in the court system, potentially dragging out the litigation (and the crisis on the ground) for decades. Officials were concerned that UCC could use the study's findings against the survivors. Still, the logic appeared more political than legal. If the findings were strong, why hide them? And if they were weak, that too would come out in time. Meanwhile, enforcement was aggressive; students were arrested and accused of being traitors for attending ICMR meetings (Rajagopal Reference Rajagopal1987), and every internal report of the ICMR was stamped “Confidential … do no publish without express permission.” Many in civil society were angered by this. As one doctor wrote, “the lack of scientific information has led to an underestimation of the damage caused by Union Carbide Corporation. It has also impeded the emergence of a proper line of medical treatment for the gas victims” (Lochan Reference Lochan1991, 1322).

Ultimately, the civil suit against UCC was settled in 1989 for a one-time payment of $470 million, which represented “full and final” settlement of all civil claims arising from the catastrophe, including the preemptive settlement of compensation claims for any potential long-term consequences. Because there was no trial, none of the ICMR's epidemiological findings beyond gross estimates of the total affected population were ever presented in court. Yet the ICMR's policies regarding the dissemination of the epidemiological data were never revised to make the studies publicly available after the settlement. Instead, the studies were quietly wrapped up until the last one closed in 1994 and the ICMR terminated its center in Bhopal, publically saying that there was no more work to be done there. But what was known about its research did not support this. Reports smuggled out of the ICMR showed that, over the first five years, the health status of survivors declined precipitously: the general morbidity rate went from 15.26 percent in 1987 to 30.52 percent by 1990. In addition, 48.13 percent of the severely exposed subjects suffered from mental health disturbances (Lochan Reference Lochan1991, 1322).

The ICMR is a directorate in New Delhi with a network of affiliated researchers, many of whom were passionate about the work. But the institution has remained secretive about several of its fateful decisions. A review of the ICMR's internal newsletters over the period from 1984 to 1994 reveals no references to even the occurrence of the Bhopal disaster, much less to any research about it. This striking omission seems to suggest that decisions about the ICMR's work and its status as a state secret came from high up in government, and they came very early on.

After the settlement, the survivors’ health care became the responsibility of the state. This would have been a logical moment to lift the embargo on publication of ICMR research and to integrate it with a program of medical monitoring in the hospitals. In 1991, there were still sixteen active studies, including the epidemiology study; by 1994, instead of continuing the research “as in the Hiroshima studies,” all of the ICMR studies were shut down, less than a decade after the disaster. In 2006, a summary of the clinical studies was finally released, followed in 2010 by summaries of the toxicological, pathological, and epidemiological studies. The 12–16-year-old material showed, among other things, that when the ICMR ended its work in Bhopal the relative mortality rates in gas-affected areas were still high; the gas-affected areas were still accruing eleven “excess” deaths per day, and the study documented 9,667 “excess” deaths between 1985 and 1993 (BGIA 2010, table 4). But the documentation itself was sketchy. One researcher I spoke to described trying to decipher degraded and destroyed research pages, long after his collaborator had passed on.

After the ICMR left Bhopal, responsibility for basic maintenance of the disaster cohort was passed to a smaller, Bhopal-based group called the Centre for Rehabilitation Studies (CRS), which employed many of the same fieldworkers who had begun the work in 1985. However, although CRS continued to do maintenance on the cohort, data collection slowed, attrition from the cohort accelerated, and the findings were still essentially kept secret. The Centre was underfunded, uninspired, unable to publish, unsupported by senior researchers, and visibly demoralized. “I feel like I've been resting my brain for fifteen years” one fieldworker told me. A senior official volunteered woefully that “we could have done better.”

Seeing the Exposed

In parallel to the ICMR studies, another bio-legal process had begun in Bhopal in 1986. This was the aforementioned process of “categorization,” or “registration and evaluation,” of survivors’ claims. This person-by-person evaluation was meant to test the validity of each claimant's sickness, not their exposure—their legitimacy under biopower, not their presence during the catastrophe. First, it purported to establish whether or not an individual was “really” gas-exposed and therefore entitled to use the gas-relief hospital infrastructure. Second, it assigned a category to each survivor's exposure levels, establishing a hierarchy of exposure with six levels, ranging from “temporary” to “utmost severe.” The category in which each survivor was placed determined the amount of settlement compensation that the individual would be allowed to claim.

Unlike the epidemiological study, the exposure categorization process did not use the geographical location of the survivor's home as an index of exposure; it instead based its judgments on limited physical examinations, conducted five to ten years after the event, and on medical records presented by the individual. Physical exams were only conducted in about half of cases (and the criteria used for them was deeply problematic; see Das Reference Das1995). The medical records tended to illustrate less about the level of injury than about the level of medical care received. For example, most of the free treatment given by aid agencies and government hospitals in the early aftermath—the emergency aid received by the most severely injured survivors—was not accompanied by any documentation; survivors in more distant, wealthier areas, particularly those who had visited private doctors, were more likely to have documentation.

Ultimately, many people from severely affected areas were unable to prove that their injuries were new, much less related to the gas, and 92 percent of claimants were placed in the lowest category of exposure (“temporary injuries”), which assumed that they had zero ongoing morbidity after the crisis of the event itself. This entitled them to a lifetime compensation of only 25,000 rupees (about $795 in the 1990s). Although a robust and publicly available epidemiological study would have made it obvious that the categorization process was systematically underestimating injury levels, the ICMR's ongoing policy of secrecy and the closure of its studies made this comparison impossible.

In certain ways, the categorization process had to be individuated. The process of tort law was such that injury had to be established as the harm done; and a certain set of individuals had to be identified as those who had been subject to that harm. In addition, the structure of the settlement assumed that there would be a hierarchy of injury, but chose to medicalize that hierarchy rather than to look at it as a distribution of exposure—exposure being predictive of injury. The real question regarded what was seen as evidence. In this case, the framework of the categorization process rendered invisible those already on the margins and created perfect conditions for fraud, which was easily carried out by those with the means to generate the minimal medical paperwork necessary, but not by the impoverished or seriously ill survivors. Meanwhile, as knowledge of the widespread fraud became common among medical personnel, a dual erasure occurred: both of the authenticity of the exposure experience and of the authenticity of the exposed person.

Reimagining the Cohort

In 2010, the political winds began to change. Most attribute this to public outrage over the extremely low sentences given out to the Indian executives of Union Carbide in the criminal case finally concluded that year. But it also reflected the transformation in economic and political power that India had undergone since 1984. An empowered Group of Ministers convened in New Delhi to reconsider the state of affairs in Bhopal. They mandated, among other things, that the ICMR restart its research on the subject: reinvigorating and following up on the epidemiological cohort and initiating new studies on gas exposure. The ICMR soon sent staff and machinery to the old offices of CRS and began to try to revivify data collection. In 2013, the ICMR published a summary of the work done by CMS between 1996 and 2010 on the gas-affected cohort—work that CRS was the first to admit was feeble and problematic, but was now presented as authoritative.

But the project of reconstructing the period of the ICMR's dormancy as one in which meaningful research had continued was pushed forward anyway. Local activist groups condemned the report and called it “concocted and unscientific,” asserting that in an “overwhelming number of cases the research assistants have not visited the subjects let alone interview[ed] them and rarely has any Assistant Research Officer (Medical) visited any of the subjects let alone examine[d] them” (Bi Reference Bi2013, 1). Work had essentially been stalled in the absence of the ICMR's imprimatur, and everyone knew it.

At stake in the debate over the legitimacy of this report was not only the erasure of the decade during which the ICMR had been absent from Bhopal and the organization's attempt to create the illusion of a seamless research program, but also the medical history of the survivors themselves. Scientific rigor aside, the report suffered from a striking absence: that of the majority of the cohort. Between the enrollment in 1985 and the 2010 report, the fieldworkers had lost track of 89 percent of the study's original subjects—that is, they were unable to find them for follow-ups. By 2010, therefore, the study only reflected the experiences of those survivors who were still living in the same homes upon which the number plates had been affixed in 1985: the survivors who had enjoyed the highest levels of domestic stability since the disaster, and the men who were not required to move for marriage. The indigent population that the cohort had been meant to make visible had in fact disappeared from it with little fanfare, leaving a very different population behind. Yet survivors by this point were acutely aware of the danger of being lost—and actively fought against it. They knew that compensation payments, past and future, were contingent on the government being able to find them. They knew that access to many kinds of better and worse free medical care was dependent on their residence in Bhopal. Where would they go? Had anyone really looked for them?

Epidemiology and Invisibility

The lack of empirical, publicly available data on the long-term effects of the catastrophe has, over time, undermined the survivors’ claims to services. Thirty years after the disaster, the “gas relief” offered in hospitals remains generic, and medical providers who work in government hospitals tend to believe that there are no long-term effects resulting from the exposure. One medical researcher noted that the “off the cuff opinion” among doctors at the city's major medical teaching hospital was that, contrary to the (limited) epidemiological evidence, exposure was not a “persistent problem.” Yet without access to individual medical histories or ongoing studies, doctors feel they are doing the right thing by being dismissive of patients’ claims to special injury. They are blind to the progression of chronic conditions and see only the suffering caused by worry about the unknown effects of the gas on their bodies and their children. This dismissal by the medical profession has aggravated the traumatic uncertainty to which survivors are already subject. Although most are not highly scientifically literate, they nonetheless understand the possibility that their exposure will have long-term consequences. Unlike the Kazakhs described by Stawkowski (this issue), who have come to feel that they have “adapted” to radiation poisoning, gas-exposed Bhopalis instead tend to believe that every ill they experience is a potential consequence of their gas exposure—a situation aggravated by the ICMR's long-term secrecy. Meanwhile, activist groups, which have operated for decades with little information beyond the certain knowledge that the ICMR was suppressing medical research, have tended to emphasize and publicize any findings of harm and of dangers to health. Survivors’ post-exposure claims to both health and to illness were made almost equally suspect by the suppression of the ICMR data.

Bio-Obfuscation

The history of the ICMR's epidemiological study after the Bhopal disaster illustrates the importance of comprehensive post-disaster epidemiological studies, as well as some of the consequences of their absence, or, perhaps worse, their hiddenness. More importantly, it exemplifies some of the consequences of the legalization and obfuscation of population data. In Bhopal, data collection was problematically partial, and access to that limited data was restricted. The consequences of the event were theoretically knowable but seemingly forever deferred: now, they are being reconstructed as scientifically whole but subject to so much fraud and attrition that they appear empty and insignificant. Although the original ICMR studies had many inadequacies and methodological problems, many of the researchers did remarkable and impassioned work in an unprecedented and chaotic situation; for them, as for the survivors, the truncation and suppression of their work for more than a decade has been another disaster.

Today, the ICMR's agenda in Bhopal has been reborn; once again, on the instructions of the Group of Ministers, the ICMR has founded a National Institute for Research in Environmental Health in Bhopal. The institute's stated mission is to carry out environmental health research in service of “becoming a Centre of excellence in capacity building for research and health interventions to meet challenges in environmental disasters in the country^” (NIREH, n.d.). Laudable and necessary though these efforts are, they are devastatingly tardy. The decades of lost opportunities and the attrition of the cohort have perhaps permanently obscured the fate of Bhopal, and survivors are left in a kind of biological limbo: neither healthy nor ill, neither visible nor invisible. Meanwhile, the ICMR has allowed the most valuable resource for understanding the exposure—the cohort itself—to slip through its fingers. Eighty-nine percent of the survivors are now lost, and any further reconstruction of the data will be only a shadow—perhaps not even representative—of the histories of the exposed.

As my field research in Bhopal in 2011 showed, many survivors, lacking reliable information, felt chronically pessimistic about their own and their family's future health and happiness. Most, even those who were part of the ICMR study, knew little about the work that was being done, and what they did know was not comforting. The status of the research was especially insulting to residents of Jai Prakash Nagar, the worst-affected neighborhood; these people were categorized as having suffered “severe” exposure by the ICMR, but 90 percent of them were classified as having sustained only “temporary injuries” during the parallel government categorization process. And over time, those who were lost to follow-up by the ICMR were bewildered by their failure to be counted even while those who remained in the cohort experienced mounting frustration at being studied year after year for no apparent reason. Fieldworkers themselves were frustrated and humiliated by this ongoing stasis: as one told me ruefully, “What can I do? They are always crying.”

The situation of the individual Bhopali survivor, trying to describe years of suffering in a few short moments for a suspicious doctor, evokes Foucault's description of plague victims: they are seen as “a form, at once real and imaginary, of disorder.” The Bhopali gas-affected, like plague victims, also had as their “medical and political correlative” the specter of chaos and obscurity (Foucault Reference Foucault1995, 198). And yet this disorder was a vivid catastrophe rendered willfully imperceptible.

In Bhopal, the epidemiological study of the survivor cohort was meant to make legible the harm from its “visible” signs: to inscribe meaning and legitimacy on the desperate and disorderly demographics of the survivors. Instead, disorder became a self-fulfilling prophecy—the harm could not be made visible because the cohort was allowed to appear chronically unmappable (although even in ideal circumstances, this kind of toxic injury might never be fully mappable—a point raised by Goldstein, this issue). The studies were supposed to unveil an abnormal injury in the population; instead, the disappearance of the studies and their subjects was instead normalized.

In the end, the loss of, and the loss to, science of the ICMR's studies in Bhopal has been massive, and the state has sanctioned this loss. It is a lesson that should be noted in the aftermath of Asian disasters to come. It is important to recall that although the history of industrial disasters is often presented as a race against technology, the example of Bhopal shows that disaster sufferers, in the long term, do not suffer only from the complexity of their symptoms. Failures of measurement, monitoring, caregiving, science, and clinical competency impede the ability to even ask complex questions, much less find complex solutions. Long-term disaster relief requires a consensus about suffering, not just about technology. What is unacceptable suffering? How do we find those who are experiencing it? And what are the long-term consequences of whole populations living out their lives in a mistrusted biological obscurity?