CROSS‑CULTURAL PERCEPTIONS OF DEVIANCE:

THE CASE OF BHOPAL

MICHAEL J. LYNCH

MAHESH K. NALLA

KEITH W. MILLER

Journal of Research Crime and Delinquency, Volume 26 Number 1, February 1989 Pp. 7-29 Sage Publications

Incorporating theory from several disciplines and data concerning media coverage of the Bhopal disaster, this article examines perceptions of the Union Carbide incident in Bhopal, India, from a cross‑cultural perspective. More specifically, we attempt to employ a content analysis of three Indian periodicals and corresponding accounts in a matched sample of periodicals from the United States to test certain assumptions that impinge upon consensus‑conflict issues. Several factors that structure the creation of definitions (i.e., ideological, political, economic, proximity of the event, seriousness, and cost‑benefit analysis) are discussed. Findings from this study indicate little cross‑cultural consensus concerning the definition of the Bhopal incident in India and the United States. Consensus was found, however, on other levels. U. S. periodicals, for instance, all seemed to portray this incident in a similar manner, while there .vas some variation among Indian periodicals. Furthermore, there was cross-cultural definitional agreement between American and Indian business periodicals.

Within criminology, issues pertaining to understanding and evaluating the definitions of behaviors and events as crimes have typically been addressed vis-à-vis the use of seriousness scales (see Rossi et al., 1974; Cullen et al., 1985, 1982; McCleary et al., 1981; Newman, 1976). It is commonly thought that this type of research empirically demonstrates the degree to which people agree (exhibit normative consensus) or disagree (exhibit normative conflict) concerning the definition of particular events and/or behaviors as crimes. It has been pointed out, however, that the seriousness scale approach contains numerous methodological biases. (These biases have been reviewed in detail elsewhere. See, for example, Meithe, 1982; Cullen et al., 1984; Newman, 1976; Rossi and Henry, 1980.) Briefly, the literature that provides a critique of the seriousness scale approach suggests that empirical findings supporting either consensus or conflict are the result of methodological artifacts. From this suggestion, we have concluded that alternative methodological strategies that investigate the consensus/conflict debate more fully are in order.

Barring methodological considerations of seriousness scale approaches, there are also theoretical reasons for expanding the investigation of consensus and conflict through the use of alternative research strategies. The most noteworthy of these theoretical concerns is found in the work of Emile Durkheim‑one of the founding fathers of the consensus position. Speaking to the issue of measuring the extent of consensus or conflict in society (what Durkheim called the extent of social solidarity), Durkheim warns, "social solidarity is a complex moral phenomenon that, taken by itself, does not lend itself to exact observation nor indeed to measurement. . ."(1968, p. 64, our emphasis). Continuing this discussion, Durkheim notes that as a measure of social solidarity "Law . . . reflects only part of social life and furnishes us with incomplete data for the solution of the problem" (1968, p. 65). These cautions, which Durkheim devised nearly 100 years ago, have had little impact on criminological and sociological investigations of social solidarity, although the remainder of his approach has been widely embraced. What do Durkheim's cautionary words mean for the study of consensus/ conflict?

In effect, Durkheim's warnings also cause us to question the methodology employed in studies that examine social solidarity by measuring agreement or disagreement with legal statutes (i.e., those employed in most seriousness scales). Such studies miss a broad spectrum of behavior and public opinion in which social solidarity or the lack of social solidarity may be evident. Social solidarity is, in other words, a multidimensional problem that has only been treated unidimensionally. Furthermore, seriousness scale approaches tend to approach the study of definitions of deviance in a static manner. They address the nature of definitions, but not the emergence of new definitions or the changing nature of definitions of crime and deviance. In order to do so, one has to observe the process of defining in action.

The above observations also give rise to a related methodological critique. In applying the same method of analysis repeatedly, social scientists are able to assess the reliability of measurement instruments. The validity of these instruments, however, remains in doubt. Over time, for example, it is possible that the same methodology yields the same results. However, by focusing on a specific set of behaviors, events, opinions, and so on, over a repeated number of applications, such a strategy precludes the possibility that unmeasured values are changing or that dissensus exists in areas not included within the framework of the measurement instrument.

With these criticisms in mind, it is evident that new approaches to examining the extent of consensus and conflict are in order. These approaches need not be new as much as they must be novel (little used) in a particular area. One such alternative methodology can be found in content analysis (Farrell and Swigert, 1976; Evans and Lundman, 1983).

In this study, we employ content analysis of a matched set of Indian and U.S. periodicals to assess interpretations of the Union Carbide gas leak that occurred in Bhopal, India, December 3rd and 4th, 1984. Drawing on consensus/ conflict literature as well as literature from other areas (i.e., the political science of the environment; environmental toxicology) we investigate perceptions of the Bhopal disaster from a multidisciplinary perspective. Our concern is whether cross‑cultural and intracultural perceptions of the Bhopal disaster exhibit similar (consensus) or dissimilar (dissensus) interpretations of this event. We employ qualitative methodology in order to capture the changing nature of definitions of deviance/ crime, and in order to express the richness of the data.

Consensus and Conflict:

The Sociological Approach

One method for examining perceptions of events cross‑culturally draws heavily on "consensus‑conflict" literature (Nettler, 1978, pp. 216-219). Briefly, the consensus model relies on structural‑functional thought advanced by Durkheim and the social contract theory of Hobbes. At the heart of this theory is the assumption that society's components must be "stable" and highly integrated if it is to function effectively (Vago, 1981, pp. 12‑13; Durkheim, 1968). In this perspective, "value and normative agreement facilitates social integration while value conflict impedes social integration and renders society incapable of effective functioning (Michalowski, 1977, p. 23). Furthering this assumption, consensus theorists find that law reflects the collective will of society, with all members of society agreeing on basic definitions of right and wrong and the contents of the criminal law (Durkheim, 1968, pp. 65‑66). In this respect, the law is "merely a written statement of this collective agreement" (Michalowski, 1977, p. 23) that helps integrate and maintain social equilibrium. As a result, law or definitions of appropriate behavior helps society function effectively by establishing "moral boundaries," alleviating conflict and coordinating social integration or social solidarity.

Standing in direct contrast to the assumptions of the consensus model are conflict models. Conflict models begin with the assumption that there are many diverse social groups existing within any particular culture or society. These different social groups generate different definitions of right and wrong because each has its own set of norms and values. Certain groups, for whatever reason, have more power than others. Through competition or domination, a culture's powerful groups come to have its interests or values represented in law (Vago, 1981, p. 14; Vold and Bernard, 1985). As a result, law is designed to advance the interests of the powerful and is subject to controversy, conflict, and change (Turk, 1969; Michalowski, 1977, p. 26; Vago, 1981, p. 14; Chambliss and Seidman, 1982; Lynch and Groves, 1986). There is not, in other words, one set definition of criminal behavior that everyone can agree upon. And, unlike the consensus theorists, the conflict theorists views conflict as being beneficial to society: It helps society recognize "problems," and promotes social change and growth.

In theory, neither approach suggests that legal definitions encompass all normative value judgments made within a particular society. Within criminology, however, these approaches have been defined rather narrowly and have been applied almost wholly to criminal behavior (Rossi et al., 1974; McCleary et al., 1983; Newman, 1976; Wolfgang et al., 1985). We will attempt to expand upon this literature by examining definitions of an event that has no preestablished definition. In theory, both approaches should apply equally well to a situation in which events have not already acquired labels. From a consensus position that emphasizes shared norms and values, events that threaten these norms and values should be defined similarly within and across cultures. From a conflict perspective, the reverse should be true: Definitions of Bhopal will vary cross‑culturally and intraculturally as well, depending on the relationship of particular groups (those doing the defining) to the power structure of society.

TESTING FOR CONSENSUS

AND CONFLICT

In general, studies that attempt to assess the applicability of consensus and conflict theories concentrate on attitudinal surveys presented in the form of seriousness scales. However, there are several methodological problems one encounters when attempting to do research of this nature (see Miethe, 1982; Cullen et al., 1985). The major drawbacks of the seriousness scale method (for our purpose) relate to data collection, which is a time‑consuming, expensive process. Furthermore, the seriousness scale approach is not well suited for assessing the immediate impact of specific events, nor is it useful for assessing the dynamic nature of definitional construction.

In light of these difficulties and trends present in previous research (see Evans and Lundman, 1983), we choose to concentrate on media coverage of the Bhopal accident to assess questions relating to cross‑cultural perceptions of this incident. We suggest that issues relating to consensus and conflict over values can best be assessed by methods of analysis that are not limited by the scope of legal definitions of behavior. Media coverage of events provides us with one such alternative data source.

THE USE OF MEDIA ACCOUNTS IN

CROSS‑CULTURAL RESEARCH ON DEVIANCE

Research that employs mass media materials assume that media accounts are somewhat indicative of people's attitudes toward particular events (Graber, 1980; Lynch, 1987). In echoing this assumption we have disregarded a perplexing and perennial problem in the study of media reporting: "Are media reports reflections of cultural interpretations of a specific event?" (Stated more generally, does the mass media reflect public opinion?) Or, "does the mass media cause or influence public Opinion about events in a specific direction?"

For our purposes we may disregard the causal sequencing problem noted above for two reasons. First, it is not our intention to predict or measure the effect of media accounts on public perceptions. In other words, the cause‑effect sequence does not alter the outcomes we will be assessing. Second, our intention is to examine perceptions of Bhopal as they are reflected in readily accessible, open forums. We assume that media coverage given this incident is indicative of prevailing attitudes within a particular culture. We realize, however, that this assumption needs qualification.

For example, periodicals are subscribed to by subgroups within a population, and for that reason a periodical is likely to reflect the attitudes and values of a particular group. On one hand, consensus theory suggests that there should be no variation in normative evaluations of events across periodicals drawn from one culture (since all subgroups have the same values). On the other hand, the assumptions of conflict theory suggests that normative evaluations of events would vary by periodical, especially if periodicals represent particular subgroups of the population who have varying degrees of power.

In order to assess whether or not subgroup variation exists, the periodicals chosen for study represent three different communities or outlooks: (1) business/ conservative, (2) liberal, and (3) middle‑of‑the-road. Again, if the consensus position is correct, we should see no difference between media presentations of Bhopal intraculturally‑that is, across the groups of periodicals surveyed within one culture. If, however, the conflict perspective is more accurate, media interpretations of Bhopal should vary intraculturally or across the periodicals surveyed.

The issue of normative agreement/ disagreement may also be assessed cross‑culturally using the same methodology. Cross‑cultural normative variations may be assessed on two levels: within and across (aggregate) periodicals. If we find agreement cross‑culturally on both levels, then the consensus position is the appropriate theoretical explanation. Again, if we find disagreement cross‑culturally on both levels, then the theoretical explanation offered by conflict theorists is more appropriate.

Overall, we expect media interpretations of the Bhopal incident to vary quite dramatically between the United States and India. In the following section we describe some of the variables (i.e., proximity, ideology, culture, technology) that may affect media interpretations or the definition of a situation as criminal or noncriminal. Some of these variables affect the extent to which norms and values are shared within a particular society, while other variables act independently of normative and value agreements.

VAIRIABLES AFFECTING

DEFINITIONS OF THE SITUATION

Proximity and Harm

The use of media accounts will allow us to assess how, the proximity of an event to a particular culture affects the definition of that event. Proximity and harm are two variables that act independently of normative evaluations. Considering proximity alone, we expect that Indian periodicals will be more likely than U.S. periodicals to depict the Bhopal incident as a "crime," "horrible event," and the like. On the other and, we expect that U.S. periodicals will be more likely than Indian periodicals to define Bhopal as an "accident" or "industrial mishap." This, of course, is due to the proximity of the event to India itself and the direct effect (harm or cost) this incident had on the people of India.

However, we cannot overlook or minimize the fact that definitions of rime vary from one country to the next for cultural, political, ideological, or economic reasons (discussed below). All three of these variables have a direct influence on the structure of normative agreement. Therefore, proximity as well as other factors affect definitions lent the Bhopal incident by media accounts in the countries under study. Additionally, we expect perceptions of the incident to vary by the type of periodical surveyed. For example, business periodicals, even within India, are expected to view this incident differently than popular or liberal periodicals. We continue the discussion of these variables below.

THE ENVIRONMENT AS AN ISSUE

Environmental pollution and environmental hazards have only recently become important issues in many nations (Encole, 1975, pp. 11‑12). Some countries have always been concerned with protecting the environment from technological disasters, while other countries hove less concern with these problems. For example, "Swedes and Germans . . . possess collective norms that promote issueness for environmental affairs [while] Americans' traditional outlooks obscure such promotion. As a result, the achievement of issue status for the environment in Sweden and Germany may ‘cost’ less culturally than it will in the United States" (Encole, 1975, p. 13). As Encole's statement suggests, the degree to which environmental pollution/hazards are recognized as issues varies from one country to the next. In some countries these events may be defined criminally, while in others they do not even receive issue status.

Cultural, Ideological,

and Technological Influences

On a theoretical level, "national issue status" appears to be linked to (1) cultural values and (2) ideological differences that exist from one country to the next. For example, industrial dumping of waste materials

or real estate land exploitation are regarded as community concerns in socialist Sweden. However, these same events are viewed as the "natural exercise of entrepreneurial initiative" in capitalist America (Encole, 1975, p. 14). But the relationship between issue status and cultural values or ideology is not clear cut. Treating environmental threats from industrialization or technology as issues of concern might also be connected to the types of economic production carried out in a society. Viewed in this light, it becomes clear that environmental hazards go hand in hand with technological advancement (Hohenemser, Kasperson, and Kates, 1980, p. 1). We might expect, then, that the more a country relies on technology in its production process, the less likely people within that country are to view technology as hazard creating, and the less likely countries of this type are to label technological accidents or hazards as crimes.

However, the relationship between the mode of material production (which involves technological advancements) and the labeling of environmental hazards is quite complex. For example, in the U.S.S.R. environmental hazards such as pollution are dismissed as "capitalist maladies," while in China‑a country that shares the same general ideological outlook and economic system as the U.S.S.R.‑environmental hazards or pollution are seen as the price of technological advancement (Encole, 1975, pp. 17‑18), meaning that pollution, environmental hazards, or environmental disasters may arise even in socialist countries (Encole, 1975, p. 18). Here, we modify our hypothesis to suggest that the more a country relies on technology in the production process and the more profit oriented a country is, the less likely it might be to define potentially hazardous corporate practices as criminal events.

Durkheim's (1968) analysis of society presented in the Division of Labor in Society, leads to an alternative hypothesis. In Division of Labor, Durkheim argued that societies with simple divisions of labor or societies in which technological innovation was low (mechanical societies) should exhibit a greater degree of social solidarity than societies with more complex divisions of labor and a heavier reliance on technological innovation (organic societies). Conversely, organic societies in which work becomes a fragmented social process should exhibit greater social divisiveness than organic societies. From this perspective, industrialized societies should exhibit less consensus concerning the definition of events than less industrialized nations. If this perspective is correct, then in our study, India, a more mechanic society than the United States, should exhibit a higher level of normative agreement than the United States.

The Costs and Benefits of Technology

The brief review given above suggests that cultures, for whatever reason(s), do not define environmental hazards in a similar manner. For some (i.e., United States, China), environmental hazards are the "price" of industrial development; for others (i.e., Sweden, Germany), the environment is seen as worthy of protection or defined so that technological advances and environmental concerns can coexist. Of course, the coexistence of society, technology, environment, and economic activity might require that profit is sacrificed to protect the environment. In this context, cost‑benefit analysis stemming from profit considerations has created a field of specialized risk management concerned with identifying, predicting, and preventing technological and toxic disasters (Hohenemser, Kasperson, and Kates, 1980). Hohenemser, Kasperson, and Kates's (1980) review of this area suggests that industrial toxicologists and government agencies within the United States have become more concerned with blocking or limiting outcomes than blocking initiating events, modifying technology, or modifying wants. In other words, the inevitability of technological advancement and technological disaster has been accepted by many U.S. regulatory agencies that "police" and regulate industries. The goal of these agencies is to contain rather than prevent disasters.

The Concrete Events: The Catalyst Toward Action

We have noted above that three things appear to lend an overall shape to public awareness and its ensuing definitions of corporations' treatment of the environment or general corporate behavior: (1) culture,

(2) ideology, and (3) technological advancement or methods of production. Of course, this list is not exclusive. In addition, the event or catalyst may also influence the definition of the event. Here, we can rely on two traditional measures to access how events will be interpreted: the degree of harm inflicted by the event and the proximity of the event to a Particular culture. We hypothesize that the more harm inflicted by a technological hazard, the more likely it is that it will be perceived criminally, with the exception that (A) the more a country relies on technology for its livelihood, the less likely it will be to regard these events as criminal; and (B) exported hazards will be treated more criminally by the cultures that have the hazard imposed upon them than by the culture(s) imposing the hazard (i.e., the proximity of the hazard to the culture coupled with the perceived source of the hazard).

Due to limitations in our data, we are unable to assess this aspect of our theoretical approach. We have mentioned this possibility so that future research can be conducted to address this issue.

METHODOLOGY

Selection of Periodicals

For purposes of analysis, we have selected three Indian periodicals. The choice of these periodicals was not random, and was influenced by the following factors: (1) the periodicals were written in English, (2) availability of domestic (Indian) editions as opposed to export editions, and (3) general orientation (conservative, liberal, and so on) of the periodical. Thus, given these conditions and limitations, it was impossible to choose a sample of Indian periodicals randomly. We choose to include "leading" (respected) periodicals written in English that represented conservative/business, middle‑of‑the‑road, and liberal sectors of the population. Periodicals were chosen on the basis of their reputation rather than readership, since readership data are largely unavailable for most Indian periodicals. Once we had selected the Indian periodicals we were to study, we matched them to a set of comparable American periodicals.

The Indian periodicals chosen were Sunday, India Today, and Business India. Sunday represents a liberal perspective, India Today a middle‑of‑the‑road stance, and Business India conservative/business attitudes. Both Sunday and India Today are general information periodicals while Business India caters to a limited business readership.

The American periodicals chosen for study were Newsweek, Time, U.S. News & World Report (U.S. News for short), Fortune, Business Week, and The Progressive. Each parallels an Indian periodical in the sample. Newsweek, Time, and U.S. News are middle‑of‑the‑road, general information magazines, while Fortune and Business Week are geared toward business‑oriented populations. The Progressive assess liberal views of the incident from America.

All articles relating to this incident appearing the year following the disaster (December 1984 to December 1985) in any of the above periodicals were included in the analysis.

BASIS OF COMPARISON

The contents of articles relating to the Bhopal incident covered in Indian and American periodicals were examined using two major categories of analysis: (a) visual contents (photographs and cartoons), and (b) lingual contents.

Visual contents. Photographs accompanying stories have considerable impact on readers. The types of photographs chosen for publication reflect editors' decisions to convey a particular message to its readership. These messages are, we suggest, in some way related to the periodicals overall integration into cultural life, and are not merely "editorial decisions." In this respect, the publication of any picture represents, in our opinion, clear political values held by those who control the various periodicals under study. Thus the values imparted in the pictures presented reflect three levels of values: (1) those of the editorial staff, (2) those of the readership, and (3) those related to the periodical position in A) the market economy of the world system (see Wallerstein, 1974; Bollen, 1983) and (B) the market economy of the nation/state.

For the purpose of this article we have categorized photographic contents into 5 types: those depicting (1) death, (2) injury, (3) relatives of victims, (4) Union Carbide plant, and (5) miscellaneous. Pictures were measured in square inches and the percentage of space devoted to pictures in relation to the text was also calculated (a picture to text ratio. This ratio standardizes square inches devoted to pictures in each periodical as a percentage of the total square inches of the article, thus allowing us to compare these figures across periodicals.) (See Table 1.)

Lingual contents. The total space devoted to describing the incident " as measured in square inches. With the exception of Newsweek, Indian periodicals in the sample contained an average of four times the space devoted to the Bhopal incident than American periodicals. In terms of square inches, Newsweek's coverage of the event ranked second only to India Today's coverage.

However, the quantitative aspects of these articles (the space devoted to text describing the incident) may not mean much since it is the quality of the reporting that effects the readership's opinion. In order to assess qualitative aspects of media coverage, we have paid particular attention to the general theme of articles, vocabulary (specific words used to

TABLE 1: Visual Presentation Summary

Death Injured Relatives & Victims Equipment & Machinery

Sq. In. PIT Tot. N Sq. In. PIT Tot. N Sq. In. PIT Tot. N Sq. In. PIT Tot. N

Sunday 48 5 12 102 11 5 0 0 0 12 1 1

India Today 301 25 8 153 13 5 81 7 3 86 7 3

Business India 67 10 4 7 1 1 0 0 0 64 10 3

U.S. News 0 0 0 0 0 0 0 0 0 6 4 1

Time 18 9 1 0 0 0 0 0 0 14 7 1

Newsweek 283 23 5 56 5 5 0 0 0 83 7 3

Fortune 72 26 1 0 0 0 0 0 0 0 0 0

Business Week 0 0 0 0 0 0 0 0 0 31 3 3

describe the incident), and analogies by counting the number of times specific words or phrases that describe the event (i.e., crime, disaster) were used. (See Table 2.)

FINDINGS

Our findings are presented below according to categories discussed above. Where possible, we standardize our findings by turning our observations into rates to make figures comparable across periodicals.

Pictures

Death and injury. Of the magazines that had pictures depicting death and injury, India Today contained a 38% death/injury picture to text ratio followed by Newsweek (28%), Fortune (26%), Sunday (17%), Business India (11 %), and Time (9%). U. S. News, Business Week, and The Progressive (all American periodicals) did not contain any pictures depicting death or injury. In addition, Business Week (U.S.) and U.S. News devoted all or most of its picture space to miscellaneous materials that consisted of photographs of Union Carbide executives and the physical plant.

In terms of pictures alone, Fortune devoted 87% of its visual content to death and injury, followed by India Today (73%), Sunday (68%), Newsweek (59%), Time (50%), Business India (44%), and U.S. News, Business Week, and The Progressive (0%).

Although space devoted to pictures of the dead and injured show a significant trend (Indian periodical and general information periodicals from the United States tended to present a higher ratio of pictures concerning death and injury), we also need to account for the number of pictures devoted specifically to death or injury in each periodical. For example, Sunday had 12 pictures depicting the dead and 5 depicting the injured (17 pictures) representing 17% of its picture to text ratio. On the other hand, Fortune had 1 picture of death and Newsweek 5 death and 5 injury‑related pictures representing 26% and 28% of their picture to text ratios, respectively. India Today had a greater picture/text ratio than Sunday, but fewer pictures depicting death. (See Figure 1.)

If we compare Indian and American magazines it becomes apparent that American magazines have attached lesser significance to the effects of Bhopal in terms of its devastating effects on the population exposed to the gas leak (fewer pictures of death and injury, as well as smaller

TABLE 2: Use of Select Words

Magazines

Sunday India Today Business India Newsweek Time Fortune Business Week

Accident/Industrial accident 1 12 5 9 6 6 4

World's worst industrial accident 4 2 1 0 0 0 0

Crime/Criminal 7 0 0 0 0 0 0

Cover‑up 4 0 0 0 0 0 0

Culpability 2 1 1 0 0 0 0

Death 44 21 19 18 1 12 4

Disaster 7 6 2 22 4 3 4

Murder/Manslaughter 4 0 1 0 0 0 0

Negligence 11 2 5 4 0 0 0

Responsibility 9 1 1 0 1 0 0

Tragedy 11 2 1 9 3 2 3

Victims 17 12 5 9 3 2 7

amount of total space devoted to death and injury). However, even among Indian magazines, some have given more importance than others to death and injury. For example, Sunday had 12 pictures depicting death (5% picture/ text ratio), India Today 8 pictures (25% ratio), and Business India only 4 (10% ratio).

It may be premature to draw any conclusions from this information since linguistic content also affects the message these periodicals tried to convey. However, two trends are clear. First, Indian periodicals stress the harmful effects of the Bhopal gas leak more fully than the American periodicals surveyed. Second, within each country, business periodicals are more likely than general information periodicals to overlook the death and tragedy caused by the Bhopal gas leak. These variations in reporting can be seen as existing on two levels: intraculturally and cross‑culturally. These conclusions are, however, tentative, and need to be assessed with larger samples. An examination of text‑themes would also help untangle this problem and reveal if there is any concordance between visual and textual presentation.

Text

In this section we divided the analyses into three areas: (a) Title of Text, (b) Major Themes in the Text (including analogies and opinions), and (c) Specific Words and Phrases.

TITLE OF TEXT

Coverage of Bhopal's Union Carbide incident in Indian and U.S. periodicals revealed several interesting findings. To begin with, Indian headlines read: "The Cover‑up: Why Union Carbide Should Be Thrown Out of India"; "The Crime Continues"; "Bhopal: The Dangers Ahead"; and so on (Sunday, April 3, 1985, p. 7 ff.); "City of Death" (India Today Dec. 31, 1984); and "The Catastrophe at Bhopal" (Business India, Dec. 17‑30, 1984). On the other hand, U.S. magazines cover story headings read "Can it Happen Here"; "Tragedy in India: The Poison‑Gas Disaster" (Newsweek, Dec. 17, 1984); "Frightening Findings at Bhopal" (Time, Feb. 18, 1985); "Inside Story of Union Carbide's Nightmare" (U.S. News & World Report, Jan. 21, 1985); "Union Carbide: Coping with Catastrophe" (Fortune, Jan. 7, 1985); "Union Carbide Fights for its Life" (Business Week, see also Progressive, Feb. 1985, p. 13) and "Valley of the Shadow of Death" (Progressive, March, 1985).

Cover story headlines indicate two major themes. The first refers to the citizens of Bhopal as victims of the incident while the second theme infers that Union Carbide was the victim of the disaster. Having identified these themes it becomes clear that Indian magazines portrayed the incident in more serious terms than American magazines. Not unexpectedly, we also note that Indian periodicals were more likely to focus on the effects the disaster had on the people of India, while American periodicals were more highly concerned with the effects of the disaster on the reputation of Union Carbide. In a similar vein, if we categorize magazines according to readership, we see that business-oriented magazines, regardless of the country of origin, clearly depicted Union Carbide as a victim. For example, Business India discusses (though not at length) Union Carbide as a victim of the disaster while other Indian periodicals do not address this issue. Similarly, Business India's headlines do not depict the incident in as serious a light as other Indian periodicals. Clearly, Business India's portrayal of the incident is influenced by its relationship to the business world.

Among the Indian magazines surveyed, Sunday was the only periodical consistently to treat the incident as a "crime." India Today, for example, refrained from value judgments except to call the result a "City of Death," and discussing the event as a "disaster," "tragedy," or "accident." Surprisingly, the only other periodical to label this event criminally was Business India, which, in one instance, equated Bhopal with murder. American periodicals were content to label the incident "an industrial accident," and in no case did an American periodical discuss this event as a form of criminal behavior.

MAJOR THEMES OF TEXT

While titles indicate the general theme of an article, they are of little value in extracting the opinion hidden in the text of apparently value-neutral approaches (for example, "Can It Happen Here," Newsweek, Dec. 17, 1984). All magazines reported the accident and extent of the damage caused by the gas leak in terms of death, injury, financial implications, and after effects on the survivors. However, the importance given to each of these aspects varied from one magazine to the next. Among Indian periodicals Sunday was most critical of the incident. It highlighted its narration with statements such as "[Union Carbide] . . . is trying to wriggle out of its responsibility for this unthinkable crime" that is "much more horrible than murder" (April 7‑13, 1985, p. 14). The magazine also blamed employees of Union Carbide America for negligence and the Indian government for not keeping a strict check on safety measures at the plant.

On the other hand, India Today (Dec. 31, 1984) maintained a strict value‑ free narration of the incident, though it covered all aspects of the incident in detail. Its conclusion, however, was interesting:

For people so recently ravaged by a silent invisible killer from the sky. . . it wasn't the immediate suffering that alone has left an indelible black mark in Bhopal's collective conscience. It is the unanswerable questions that will continue to haunt its population for generations to come, as survivors tell their children, why Bhopal? In part the answer has to do with the ethic of industrialization and the demands it makes on a still feudal rural society. But, the greater part of the question cannot be answered, and will not vanish easily from the conscience of those guilty of negligence .... The guilty will undoubtedly be taken to task, but there can be no retribution just enough, nor any turning back of the clock [p. 24].

Though India Today does not specifically call the incident a "crime" it does raise three major issues: "guilt," "retribution," or "just" punishment (which imply a crime has been committed) and the "ethic of industrialization." We will speak to this latter issue later in this text.

It is also interesting to consider the final position taken by the conservative coverage Business India gave this incident. In concluding their position on the Bhopal disaster, Business India indicates a decisive shift to a more liberal position: A position that condemned rather than defended business' interests. It concluded that "the catastrophe at Bhopal underscores the fact that certain industries, whatever their benefits, threaten life in a way no society can accept" (Dec. 17‑30, 1984, P. 74).

Of all magazines surveyed, Sunday (April 1985, p. 23), the most liberal Indian periodical surveyed, painted the most vivid picture of the event in an article called, "The Crime Continues":

Bhopal has become a laboratory of human guinea pigs . . . . Deadly methyl isocynate . . . [gas left] 2,500 dead, 50,000 maimed and 1,500 severely affected, [all] . . . hapless victims of medical experiments and selfish political design . . . . On the average, two MIC‑affected residents of Bhopal die every day‑either adults who were exposed to the gas, or deformed newly‑borns or children born prematurely. Even fetuses [sic] have been affected; today doctors are advising every woman who was pregnant at the time of the tragedy to undergo an abortion. Clearly, the catastrophic gas leak has converted Bhopal into a mini‑Hiroshima, and the brutal effects will be felt by the generations to come [emphasis added].

Among U.S. magazines, Newsweek maintained a middle‑of‑the‑road approach throughout its discussion of the disaster with statements such as: "residents of Bhopal have become victims of an age when the rush to produce dangerous chemicals has often overstepped concern for human safety." This approach is similar to the concern showed by India Today. Both seem to imply that more consideration needs to be given to the cost and benefits of expanding industrial capacity and dangerous industrial endeavors. Also echoing a cost‑benefit concern was Fortune. For its part, however, Fortune seemed to suggest that the costs were those associated with the ability of Union Carbide to expand its business rather than the human costs of disasters such as the Bhopal gas leak. Along these lines, Fortune raised the question of "how Chairman Anderson and his team . . . will influence managers elsewhere who are imagining the unimaginable happening to them," while U.S. News reported that "with world attention focused on the tragedy in Bhopal, a multinational corporation works to regain its bearings‑and its reputation" (Jan. 21, 1985). Again, these two statements show more concern for the effects of the incident on Union Carbide than the people of India. However, the common theme linking all periodical coverage of the Bhopal event relates to the "ethic of industrialization." To some extent, all periodicals viewed industrialization and technology as hazard‑creating processes. We will address this issue more fully in our discussion of the Bhopal incident.

Specific Words and Phrases

Indian magazines tended to use more words describing the death and injury that resulted from this incident (for example, dead, killed, corpses) than U.S. periodicals. Sunday recorded 44 such usages related to death alone. Most prevalent uses among other periodicals were: India Today (N = 21) Business India (N = 19), Newsweek (N = 18), Fortune (N = 12), and Time (N = 1). In comparison to Sunday, India Today referred to instances of death about half (47%) as often, while Time (2%) made little reference to this issue (see Figure 2). In addition, Sunday used certain words that other magazines have not used at all, such as "crime" (N = 6), "cover‑up" (N = 4), "culpability" (N = 2), and "murder" (N = 2). Sunday also made reference to this incident with phrases such as "human guinea pigs" and "mini‑Hiroshima." This signifies the very strong position Sunday has taken toward describing the incident. On the other hand, U.S. magazines depicted a different picture of Bhopal with preferred terms such as "disaster" (Newsweek, N= 22) and "industrial accident" (Newsweek, N = 5; Time, N = 6).

DISCCSSION

What do these findings mean? There is a clear pattern in reporting this incident by Indian and American magazines. Indian magazines portrayed the incident much more seriously than U.S. magazines and discussed loss of life and suffering of the people of Bhopal in greater detail. American magazines were more inclined to make reference to the incident as a disaster and label Union Carbide as the victim. Similarly, Indian magazines were more likely than American periodicals to label the incident a crime, or to call the act "irresponsible and negligent." In contrast, American magazines were content to call the incident an industrial accident.

It would appear that the conflict perspective is a more suitable explanation of normative evaluations of this event than the consensus position, both cross‑culturally and intraculturally. Since our data are qualitative as opposed to quantitative, the strength of this conclusion is in question. However, the consensus explanation failed in two regards. First, there are strong normative disagreements across cultures and within the sample of Indian periodicals. Second, as noted earlier, Durkheim's discussion of social solidarity hypothesizes a greater degree of normative agreement in mechanic as opposed to organic societies. Our findings reveal exactly the opposite conclusions: stronger normative agreement in the organic (American) society. It is possible, however, that Indian society is in the midst of social change, thus raising the possibility that India can be characterized by anomic social relationships that would lead to greater normative dissensus.

It should also be noted that the conflict position did not hold in every case. Two exceptions were noted. First, there is cross‑cultural agreement concerning the interpretations of Bhopal presented in business or conservative periodicals. Such agreement, though, does not necessarily violate the assumptions of a conflict perspective. Radical conflict theorists, especially those who posit a direct relationship between economic production and value formation, would not be surprised to find the business sectors in these two countries expressing similar interpretations of this event.

Second, and more detrimental to the claims of a conflict theory, normative evaluations of the Bhopal disaster did not vary intraculturally within the United States. The appearance of consensus within American society could be explained in a variety of ways (for instance, one could discuss the type of normative hegemony that pervades U.S. culture, suggest that consensus is an artifact of false‑consciousness, or that periodicals do not represent opinions of the entire range of the American public). However, we see no need to provide this type of postfactum explanation that can be equated with a theoretical apology.

At this point, we would like to turn our attention to some factors (discussed earlier) that might contribute to explaining the definitional differences we noted cross‑culturally.