Abstract
Accurate information can be difficult to access in authoritarian regimes. As more and more countries fall under the sway of illiberal leaders, ethnographers may find themselves unable to easily corroborate knowledge that comes from native informants. This article proffers several strategies for researchers to utilize when analyzing unverifiable ethnographic data. Following from sociological and psychological studies of rumor, the article discusses the specific role that rumor plays under conditions of authoritarian governance. Without legal access to information and robust civil liberties and rule of law, it may be risky for researchers and informants alike to dig too deeply into secret information. Assuming that corroboration is impossible, how might an ethnographer analyze the rumors without believing or disputing their factual authenticity? Several possible approaches are offered.
Subfields: Rumor, Russia, authoritarianism
In 2013, I was living in Irkutsk, Russia, conducting participant observation with environmental activist groups. One morning, I arrived at the office of one of the environmental nonprofits in my study to see that their storefront had been defiled with black spray paint. The graffiti slogan—“Foreign Agent ♥ USA”—had been splashed across the front wall. The defacement suggested to bystanders that the environmentalists inside were spies, working to subvert Russia’s national interest on behalf of a Western power. The local environmentalists were convinced that whoever had vandalized the storefront had done so at the behest of the Russian government. The vandal was never caught. There were no confessions or revelations, only a murky belief that the activists held, built upon their collected stories, suspicions, and circumstance.
As a social scientist and an ethnographer, I am called upon to make sense of this happening. The event forms a data point in my fieldwork: evidence for my eventual claims. Data of this nature can shape subsequent ethnographic analysis; unfortunately, though, the contours in this occasion are not clear-cut. There is the empirical fact of a defiled storefront. But this material object cannot speak for itself. Its existence is a fact, but it falls short of being a “social fact” (Durkheim 2014). It is my job to uncover the theoretical import of this event, but claims remain speculative without understanding who committed the act and why. Social theory is built upon the analysis of evidence. But in this particular circumstance, what is the evidence upon which I might build social theory? There is spray paint on a wall, but the rest is conjecture.
I cannot say that the vandalism was, in fact, a state-sponsored action designed to suppress and silence a critical voice. I do not have access to the halls of power where such a strategy might have been discussed and authorized. Neither can I say without a doubt that the government played no part in this harassment. If these activists are indeed the subject of state repression, then it is inadequate to consider their suspicions as nothing more than a cultural artifact—a narrative that helps them to make sense of the graffiti in a way that accords with their identity and their worldview. In this particular instance, it does matter whether or not their belief is true. There are vital social repercussions and major theoretical implications that hinge upon the reality of this piece of information. Would the Russian state orchestrate an extralegal (even illegal) attack on a fairly small and relatively harmless nonprofit, or are activists inventing villains and chasing phantoms? The answer to this question is one of pressing sociological significance, but it is also one that cannot be answered definitively with accessible resources. It is one of many such occurrences that plague the ethnographer of authoritarian regimes.
Ethnographers and those who employ qualitative social scientific methods rely upon the knowledge, information, and interpretation offered to them by native informants. These methods acknowledge the existence of grounded, place-based understandings that can only be grasped through active observation and direct communication. Oftentimes, the analytic goal of ethnography is to grasp the underlying narrative and interpretative frames that a particular cultural group uses in order to make meaning of life’s events.
Sometimes, though, ethnographers want to know whether what their informant has shared is, in fact, true. In such circumstances, qualitative methodology endorses triangulation (Flick, 2004) whereby information learned from informants may be cross-checked through regular channels, such as the news media or official statistics. However, these methods imply that official channels are themselves reliable and trustworthy, which is an assumption that cannot be made in authoritarian regimes. And while ethnographers have developed some tactics to dig out veracity from deceptive and deviant informants (e.g., Ben-Yehuda 1990, Douglas, 1976, Marx, 1984, Shulman, 1994), these methods may not translate to contexts where rule of law is weak and where digging too deeply may jeopardize the research endeavor. There are places and occasions when qualitative methods will produce information that simply cannot be substantiated by the scholar. The content is possibly accurate, but there is no way to ascertain the veracity of the report. In these instances, the qualitative researcher must make a choice on how to approach the problematic data, generally without recourse to any systematic or formalized methodology to guide the way.
The goal of this article is to sketch a framework for analyzing rumor in qualitative data specifically as it arises in authoritarian regimes. Knowledge is power, as the old adage goes, and authoritarian regimes maintain control, in part, through their opacity. In such contexts, rumors proliferate precisely because citizen access to information is limited. Authoritarianism is on the rise in countries around the globe (Diamond, 2015). As social scientists plunge into worlds where accurate information is difficult to come by, where rulers maintain their authority by means of deniability and doubt, we must be careful in how we assess information that cannot be corroborated, but must not be ignored.
No matter the subject of study, there are bound to be rumors in authoritarian contexts. Not all of these rumors can be verified; and, given the research question, it may not be expedient to interpret or explain them all. But does that make these rumors useless as data? I would argue to the contrary. Rumors can be usefully analyzed as data, but it does require the researcher to tread carefully and deliberately. There is always the temptation to believe what our informants tell us—especially those with whom we have developed a strong rapport. Neither does one wish to give a pass to corrupt and criminal conduct simply because it cannot be “proven” to have taken place. But if we assume that a rumor is true, we miss out on the opportunity that rumor provides by its very ambiguity—real data that, if analyzed well, can yield far more information than it purports to at face value.
This meditation on the utility of rumor-as-data is borne of my own research in Russia—a country with a long history of state repression and illiberal governance. While I entered the field with the purpose of studying the impact of globalization on local environmental activism, I was startled by the sheer preponderance of unverifiable information in my data. I had not entered the field to study rumor, and yet it appeared in my field notes unbidden—and copiously at that. Far from being the result of isolated individuals holding to crazy conspiracy theories, rumor formed a constant undercurrent in my data—arising from many informants, in divergent settings, and touching on numerous topics. But despite this diversity, the circumstance in question usually involved some form of corruption by powerful people who were in a position to prevent disclosure of the truth. My dubious data were stamped with authoritarian rule.
One or two rumors that show up in interviews may be safely ignored in analysis, but given the frequency of rumor in my field site, it seemed folly to throw it all out as hearsay. But it also seemed irresponsible to give more credence to these stories than they deserved. I began to reconsider what might be analytically useful about all of this unsubstantiated information. I had not come to Russia to hunt down evidence of government wrongdoing. Neither was I inclined toward a deep dive into the semiotics of any particular rumor I had heard. I sought for other options—for strategies that might help me navigate the space between sureness and doubt, so that I could use uncertain information to make real knowledge claims.
What I offer below is an exploration of the opportunities to utilize unverifiable ethnographic data in circumstances where uncertainty is a given. I do not offer a novel means to uncover the truth, but rather a road map to grappling with uncertainty. In so doing, I hope to offer a useful methodological meditation to researchers working in authoritarian regimes, where scarcity of information yields a surfeit of rumor. Under such conditions, what is an ethnographer to do?
“Reality” and qualitative epistemology
Ethnography has always produced multiple levels of truth. These levels are sometimes parsed into the emic and etic approaches to knowledge and understanding; emic refers to within-group truths, which are only accountable to standards held by the group itself, while an etic approach takes an outsider’s perspective and holds external standards for knowledge and truth. Historically, scholars have favored the etic. While categorizing the belief systems of various tribes around the world, anthropologists did not take descriptions of witches or spirits at face value—these were real to the believer (emic truth), and as such would be real in their consequences; but the reality of witches and spirits could be dismissed by Western observers relying upon scientific methods (etic truth).
This traditional relationship between the emic and etic created a representational crisis of conscience for ethnography, as the question of voice, identity, and power came increasingly to the fore (e.g., Clifford and Marcus, 1986, Said, 1978). Much could be said about the raging debates and attempted correctives that sought to assuage these problems of power and representation. However, the practice of ethnography—including the study of the Other—continues, and it does so at least in part because there is still valuable knowledge to be gained through the practice. The ethical standards have been heightened, but the value of scholarly distance remains. Ethnographers are encouraged to guard against “going native”; instead, an outsider analysis brings additional validity by virtue of its “critical distance” (Delamont, 2004). Perhaps in sociology more so than anthropology, the requirements of nomothetic theory encourages scholars to elevate certain types of knowledge—the rational–scientific, measurable, testable, replicable, and verifiable—above all else. Sociology, as an empirical social science, will settle for little less.
Perhaps because ethnographers began their enterprise as Western colonizers infiltrating “primitive” tribal societies, a previous generation of scholars felt no compunction for assuming that their research subjects were empirically wrong when thinking about their own beliefs. There was not a great concern as to whether the stories people told were accurate as such; stories were only important insofar as they accurately described a group’s belief system. The outcome of interest was a belief system, not a correct depiction of reality. For that reason, there is no formalized methodology for handling unverifiable information that crops up in qualitative data (as far as I am aware). It is all “cultural,” and therefore “real in its consequences” (Merton, 1968). Its ontological reality is hermeneutically unimportant.
But ethnography is not devoid of factual, empirical, substantiated information. Scholars seek to learn as much about their field sites before entry, relying upon a knowledge base that is previously established using traditionally-accepted sources. The cultural content of the study is situated within a larger, known empirical environment. In many cases, the researcher can access certain facts about the field site without recourse to asking informants themselves. The data-driven contour of the general circumstance in question is known and available to the ethnographer. In these situations, the researcher may ask informants for information, but the purpose is only to learn the local understanding or interpretation, not to gain hard evidence, which has been previously established.
Alternately, informants may offer the scholar some piece of information, which the researcher then seeks to verify. In this case, the ethnographer learns the fact from an informant, but it can be cross-referenced later against a more robust evidentiary base. In other words, the empirical truth of an informant’s claims may be taken as “fact” after the information has been corroborated in a manner deemed more reliable than mere hearsay. Ethnographers often try to triangulate particular claims made by informants (Flick, 2004). Options for verification include official statistics, newspaper reporting, other media, multiple methods or similarly conventional sources. However, there may be occasions when an ethnographer encounters information in interviews or observed conversations that cannot be readily substantiated or verified. Without any systematic methodological guidance, individual ethnographers are left to decide on their own whether and how to analyze this problematic data.
Unverifiable information in ethnographic data
Unverifiable data take many forms, and these are often context dependent. Before undertaking to address or analyze such data, it is good to know what type of unverifiable story is being told. To better understand this uncertain information, there is a sizable body of research on social story-telling to which ethnographers may turn for assistance. Folklorists, social psychologists, sociologists, and other social scientists have made the conditions and content of social story-telling their primary subject, including stories that hold only a tenuous relationship to the truth. These studies have yielded important findings that ethnographers can utilize to better grapple with such data when they arise. Following these studies, I differentiate four different types of unverifiable tales that occur in ethnographic data: evasion, gossip, urban legend, and rumor. While the definition of these terms varies within the literature, there are some general rubrics to help understand their differences. The choices available to an ethnographer for analysis are contingent on the phenomenon at hand. I will briefly address each in turn. Evasion. Sometimes stories are unverifiable because they are highly personal, but potentially embarrassing for the informant. These tales confront the ethnographer who suspects an unreliable narrator. As Douglas (1976) points out, the relationship between the researcher and the researched can be riddled with conflict. People may don masks or “fronts” to protect their secrets and secure an appropriate “presentation of self” (Goffman, 1959). Such informants misrepresent themselves and misinform the researcher by their evasions, silences, lies, and half-truths. Arguably the strongest asset that ethnography offers is the ability to enter a field site for an extended time, familiarize oneself with insider knowledge, and built rapport with multiple informants. With time and trust, evasion becomes less likely as the research relationship becomes less adversarial. Additionally, Douglas (1976) offers several other means to tackle evasion, including the ethically dubious decision to dissemble and deceive informants about the nature of the research relationship. Gossip. Another type of unverifiable information that is also highly personal is gossip. But unlike evasion, which implies engaging in deceit about one’s own behavior, gossip provides potentially inaccurate stories about other people’s behavior (Eder and Enke, 1991). Gossip comprises unverified information about another individual’s action, character or behavior (DiFonzo and Bordia 2007). Gossip has elements that are individual (the subject) and social (the network that shares the gossip). Gossip builds in-groups and out-groups, and forms outlines of social solidarity and social conflict. Although there are elements of power in gossip, this power–struggle is interpersonal and the stakes are access to reputational status and network linkages (Kurland and Pelled, 2000). A plethora of gossip in ethnographic data would suggest that the field site is fraught with social striving. It would encourage the ethnographer to map status hierarchies and trace shifting alliances among various individuals in a singular social group. Urban legend. Sometimes an ethnographer encounters a story whose truth cannot be verified, but is told by multiple informants—particularly if they want to convey some warning about how to behave. While its repetition by multiple informants may encourage the researcher to believe its accuracy, it is also possible that the story informants are sharing is in fact an urban legend. Folklorists define an urban legend as a frequently repeated story that contains a moral message (Barnes, 1996; DiFonzo and Bordia, 2007). Urban legends “reflect many of the hopes, fears and anxieties of our time” (Brunvand, 1981: 2). Importantly for the ethnographer to note, an urban legend is often recounted to make sense of a shifting cultural context. If multiple informants are sharing a variation of the same unverified story, particularly if the story offers a transcendent moral, then the savvy ethnographer might examine the tale in light of any rapid cultural changes that are upsetting collective norms. It may be that the oft-repeated but unverifiable story is performing the cultural work of a modern-day legend. Rumor. A rumor is a piece of information shared person to person, but whose veracity has not been ascertained. Unlike evasion, a rumor is not about one’s own personal experience. Unlike gossip, it not about a particular individual’s character. Nor does a rumor impart a moral lesson as does an urban legend. Instead, a rumor is told to inform, because the rumor is thought to be potentially useful to the person with whom it is shared. Neither the speaker nor the receiver can be sure of the rumor’s truth (at least in the moment), but its potential importance justifies the choice to share it. Rumors are not necessarily false—they are simply less secure in their accuracy than the socially accepted standard that is set for establishing the truth (Shibutani, 1966). Rumors exist precisely because situations are ambiguous and potentially threatening (DiFonzo and Bordia, 2007). Rumors thus constitute unverified information told about an uncertain situation that is considered socially important (Peterson and Gist, 1951).
The above definition points to general social conditions as a variable that may be more or less likely to produce rumors. There are social conditions that erode institutional legitimacy and thus cause some groups to doubt the socially accepted standards for verification. One such example is social inequality. Racial and gendered caste systems promote rumored knowledge, as some demographics find themselves consistently excluded from institutional decision-making and are left relatively powerless within them (Fine and Turner, 2001, Parsons et al., 1999). In these cases, rumors may persist in a particular field site despite the fact that the ethnographer may be able to disprove them conventionally. In other contexts, ethnographers may be no abler than informants to ascertain the truth of the rumor.
Rumor is a social phenomenon, and it is inevitable that rumors will appear in ethnographic data from time to time. More often than not, there is no need for ethnographers to approach rumors in their data systematically. A rumor may crop up here or there, but usually rumor is sufficiently infrequent that it is not problematic for the overall project. More often than not, a rumor that cannot be corroborated by the researcher is simply dropped from the analysis. The exception, of course, is when ethnographers choose to make rumor the center of their project (e.g., Bass, 2008, Yezer, 2008). Unless rumor is the topic of study, though, it is difficult for researchers to utilize rumors to make knowledge claims.
However, given that there are social contexts that produce a substantial amount of rumor, certain field sites can be expected to produce far more data of dubious reliability than may be safely discarded. Even though a researcher did not intend to study rumor directly, the data may be saturated in rumor simply by virtue of the study’s location. In those cases, it is worthwhile to consider how one might successfully turn rumor into data, as well as the various implications for doing so. Ethnographers in rumor-dense field sites would benefit from an analytical roadmap to begin addressing the unverifiable stories that inevitably arise. While there are a number of social conditions that promote the spread of rumor, there is arguably no social context in greater need of workable strategies for rumor data than that of modern authoritarianism.
Rumor in authoritarian regimes
After three waves of democratic expansion around the world, authoritarianism is again on the rise (Diamond, 2015). Yet because of liberal democracy’s global hegemony, modern authoritarian regimes still frequently pay homage to democratic institutions or norms in order to legitimate their rule. Called either “competitive authoritarianism” (Levitsky and Way, 2002) or “electoral authoritarianism” (Schedler, 2009), these regimes are largely defined by what they are not: liberal democracies. A general definition of liberal democratic governance includes: free, fair, competitive elections; universal suffrage; political rights and civil liberties; and independent, civilian-led branches of government. Many modern authoritarian regimes offer some institutional similarities to democracy, but violate these democratic standards in a variety of ways sufficient to undermine their liberal democratic intent.
Authoritarianism is not identical wherever it is found, either through time or across space; each regime has its own unique idiosyncrasies and institutional configurations. But whether these regimes represent the near-totalitarian social control of the mid-20th century (e.g., Nazi Germany, the Khmer Rouge) or the electoral authoritarianism of the early 21st century (e.g., Iran, Venezuela, Russia), they share a commitment to a centralized state power that is maintained at the expense of individual rights and liberties. Chief among these curbed rights, for our present purposes, is the right to know.
In liberal democratic ideal, there is a reciprocal relationship between citizen knowledge and citizen power. When the broad public becomes the institutional gatekeeper to state power, then political leaders will enact laws that the broad public desires—such as law that secures public access to knowledge. And when the public is well-informed, it will also better fulfill its role as gatekeeper to the sovereign state. So important is an informed citizenry to a functional democracy that states will constitutionally enshrine the freedom of speech and press.
The relationship between knowledge and power is no less reciprocal under authoritarian rule. The citizenry is neither expected nor encouraged to govern itself. Instead, authoritarian leaders view the public with suspicion, as a potential powder keg that might, at any moment, revolt. To stave off the threat of dissent, information—particularly regarding the workings of government—is deliberately obscured. Transparency is not a hallmark of dictatorship. To achieve its compliance, the public is left in perpetual uncertainty, navigating daily lives amidst institutions that perhaps appear predictable right up to the moment that they are not.
The problem of knowledge is thus tied to the problem of power. The more centralized the latter, the more limited the former; in such conditions, rumor is bound to flourish (Kapferer, 1990). As was discussed in the previous section, rumor thrives under conditions of uncertainty, threat, and social distrust. In the societal proclivity toward rumor, authoritarian regimes take pride of place. Indeed, history shows that rumors proliferate under conditions of authoritarian governance. The Soviet-bloc countries were notorious for the vibrancy of their rumor mills (Bauer and Gleicher, 1953), and conspiracy remains a potent part of Russian society today in Putin’s neo-authoritarian regime (Rogers, 2019). Rumor in China is embraced as a form of social resistance against state censorship (Gao, 2012; Liu, 2013), and it fuels an alternative political imaginary among dissidents in Iran (Rahimi, 2013). In these societies where the state controls the officially sanctioned “public sphere,” rumor becomes an informal—and unreliable—medium for an excluded citizenry to convey potentially useful information.
But that is not the only reason that rumors spread under conditions of authoritarianism. Rumor can also be an effective means for social control by the authoritarian state itself, helping the regime project greater authority than it actually commands (Wedeen, 1999). An authoritarian state may instigate rumors to secretly spread its own message, to anonymously malign its opponents or to sow a distracting confusion and discord in the public sphere (Crozier, 1996). Rumors may encourage or discourage certain behaviors in the broader public that operate to benefit the regime (Saavedra and Rinke, 2012). Rumors also assist authoritarianism indirectly because their very indeterminacy fosters an aura of uncertainty, which in turn dampens the public imaginary for collective action (Brown, 2016). To the extent that rumors heighten fear and anxiety, they keep the population cowed and tractable (Perice, 1997). But it is worth noting that the price of a frightened and mistrustful population is the love and patriotism that often accompany consent of the governed in stable democracies. Rumors represent one more manifestation of the precarious position of modern authoritarianism (Schedler, 2013; Wedeen 1999). Whether true or untrue, rumors generally damage the reputation of an authoritarian regime (Huang, 2017). Deploying rumor as a tactic of power is essentially a gamble by the ruling elites that popular enmity is less threatening popular sovereignty. They forsake public affection to achieve social control.
In these rumor-saturated authoritarian environments, an ethnographer may find herself grasping for solid empirical ground on which to stand while evaluating the meaning of informants’ claims. Sometimes, the truth can be ascertained using conventional methods; but other times, accurate information is hidden, even from the elite position of the scientific observer. In some respects, the problem of “hidden and dirty data” (Marx, 1984) is not unique to authoritarianism. But authoritarian regimes have added burdens in the quest for certitude. First of all, authoritarian states do not offer the legal right to access state records. Laws such as the Freedom of Information Act (FOIA) in the United States or Access to Information Act (ATIA) in Canada, while not without their own difficulties, are superior to regimes that offer no such public access (Cordis and Warren, 2014, Larsen and Walby, 2012). Separate, independent branches of government offer the opportunity for subpoena power that can disclose information among diffuse power holders (Lee, 1978). Whistleblower protection laws allow for greater cooperation by individuals who have important information on wrongdoing (Miceli et al., 1999). These institutional mechanisms work in favor of the intrepid social researcher in liberal democracies.
But still more importantly, it is the rule of law and its guarantee of individual rights that render data collection more possible in democratic regimes. It is with the assumption of an impartial judicial system backed by individual rights that previous researchers might confidently suggest strategies that could be precarious in authoritarian regimes, such as digging through trash (Shulman, 1994) or deception through impersonation (Douglas, 1976, Marx, 1984). The fact is, tactics for uncovering dirty data in authoritarian regimes could be dangerous, not only for the researcher, but also for various individuals who might otherwise be willing to assist. When one’s fate is decided by the unpredictable whims of an opaque authority, it is difficult to determine when pulling on a particular thread would lead to personal jeopardy.
In authoritarian regimes, the ethnographer has no better access than the informants do to the governing circle or the halls of power where the truth resides. When ethnographers are studying groups that rely upon rumor, the data likewise become plagued by a preponderance of questionable knowledge claims. Such circumstances demand a more systematic and rigorous approach to rumor. Such analytic tools are particularly necessary for scholars whose research topic is not directly political. It is one thing if the purpose of a study is to uncover acts of state repression or violence. In those cases, ethnographers can devote care, time, and attention to delving into the veracity of hidden government actions (e.g., Ben-Yehuda, 1990). But there are other scholars in authoritarian conditions whose interests may be more prosaic. Perhaps the research question involves gender in higher education, parenting strategies across social class, or, as in my own research, transnational trends in local environmentalism. In these circumstances, rumors may proliferate, and their content may have implications for the analysis, but rumor itself—as fact or fiction—is not the central purpose of the study. In such circumstances, what should an ethnographer do?
Strategies for analysis of rumor in authoritarian regimes
In the remainder of this essay, I proffer four strategies for dealing with ethnographic data gathered in authoritarian conditions particularly as it pertains to rumor. Under such conditions, uncertainty must generally be accepted as a given. The researcher can never establish the “facts of the matter.” But that does not mean that the rumors are useless. Below are four general approaches to this data to help guide subsequent analysis. They are not intended to be definitive, but rather to help begin the process of reconciling the knowable and the unknown under difficult circumstances. Neither does use of one strategy preclude the others. To the contrary, these strategies target different elements of the data and thus complement one another in order to build a fuller understanding. I will describe each approach in turn below.
The evidential approach: Engage in/with investigative journalism
As ethnographers, our pride in craftsmanship comes from our analysis: we take as data a public life that anyone might also observe, but with our analysis and theoretical insight, we can explicate for others the meaning and social importance of the everyday. Cultural knowledge is generally available for the intrepid ethnographer. As Geertz (1973) encouraged, our role is to understand the difference between a wink, and a wink, and a wink. The problem is not to find out whether a wink actually took place. But sometimes the claim to study “meaning” can become an excuse not to be burdened with the truth.
The task, then, is to start thinking like an investigative journalist (Scheper-Hughes, 2004), or perhaps a private detective (Shulman, 1994). Detectives must have sufficient evidence to overcome reasonable doubt in a court of law, and newspapers will not run a story if they cannot find two independent sources to corroborate the information. Ethnographers are generally let off the hook on this front because they can claim it is only an informant’s “understanding” that is sought, not verifiable truth. However, in practice, ethnographers take much of what they are told for granted. Perhaps because our informants are kind enough to allow us to observe them, we reciprocate with a “hermeneutic generosity” that trusts, rather than suspects (Scheper-Hughes, 2004). But by adopting the journalist’s, or the detective’s, mantra of verification, we can potentially do a better job of elucidating the institutional context in which this thing called culture plays itself out, which would only improve our analysis overall.
Very often, the rumors that appear in qualitative data in authoritarian regimes simply cannot be corroborated. But sometimes some aspects of the unverifiable story can be. Perhaps there is some paper trail that can be found or a second source who might corroborate some piece of the larger picture. Maybe a secretary or a janitor can recall something about a particular circumstance, whether a certain person visited a certain place, if more powerful individuals are less than forthcoming. It an authoritarian regime, “proving” the veracity of a rumor may not be a reasonable goal; but lowering the bar from “proof” to “support” can at least render the rumor more robust than mere hearsay. In my own work, I heard a rumor of government repression whereby an agent had impersonated a tax authority to force the suspension of activists’ bank account. In this case, I was close enough to the story’s origin that I was able access the letter that was sent to the bank to at least verify that it was fraudulent. I cannot confirm that it was sent by a state agent; but it was sent deliberately, under false pretenses, by someone who had knowledge of government documents. Little corroborations, such as securing a related document, can help minimize the suspicion that scientists are trained to employ when evaluating anecdotal claims.
Clearly, an ethnographer should exercise some discretion when deciding to pull on a thread in an authoritarian tapestry. Depending upon the claims made, or the personages involved, it might place the overarching project in jeopardy simply to follow-up on a rumor. These are not places bound by the rule of law; neither are they known for robust protections of individual rights. Moreover, foreign scholars working in such countries do so with permission from the regime itself. That does not mean surrendering verification, but it does mean that a researcher might want to think creatively about how not to draw attention to oneself in the process of verification. Keeping low on the information totem pole—by befriending reception staff or service workers—may allow for some corroboration of a rumor, while still keeping a low profile from those who may be watching.
Another problem with verification, though, is that it takes time. As ethnographers, we treasure our limited time in the field; time spent trying to verify a rumor is time spent away from gathering more data. There is a tension between achieving breadth and depth on any subject, so corroborating rumor can penalize the greater project if the researcher fails to cover fully the field of study. Moreover, because most of the intellectual heavy lifting is done in analysis and writing up, the researcher seldom knows in the moment exactly which of many rumors are the ones that will become critical for the analysis. Especially in a social context with a large amount of rumor, then how do we decide which rumors to try to verify? Adding to the complication, for all of our training in theory, method, and analysis, we do not often have training in investigative journalism sufficient to the task at hand—how might we even go about trying to confirm what powerful individuals do not wish to be known?
One means to overcome some of these difficulties is to form relationships with local journalists. Simply having the ear of a helpful local independent journalist can provide a sounding board for suggested tactics for verifying rumors and can be enormously helpful. Local journalists have the skills and contacts to help verify the information that we, as ethnographers, have painstakingly uncovered using the methods that we normally employ for cultural analysis. Sometimes the relationship can even become a true collaboration, where a journalist will also utilize the interviews and information gathered cooperatively for their own professional publications (e.g., Scheper-Hughes, 2004). If such a working relationship develops, it might even be possible to ask a collaborator to look into verifying particular rumors after the researcher has left the field—which is especially important when the relationship between a rumor and the larger project only becomes apparent later in the analysis.
The utility of this practice depends enormously upon the regime in question. Before befriending members of the press, an ethnographer should take care to evaluate the unique position of journalism in the country at hand. Authoritarian regimes differ considerably on the freedom of their media, and members of the profession may have unique viewpoints or values depending upon their social position (Hanitzsch et al., 2011). Before enlisting the help of journalists, ethnographers should first interrogate what journalism means in that particular country and whether journalists would be generally trusted, reliable, and beneficial partners. Neither are all media created equally within a country (e.g., Heinrich and Pleines, 2018). While national outlets may be close to the regime, it may be that local reporters have greater independence. In Russia, I have—often coincidentally—befriended journalists in the sites where I have worked, and these relationships have proved helpful to me. But it is worth noting that this approach would not necessarily be beneficial in all authoritarian states, nor in all levels or types of media within the same regime. Therefore, understanding the position of journalism within each regime is critical.
Additionally, collaboration with journalists can have major ethnical implications. Reporters working in authoritarian regimes are modern heroes, risking their own well-being and that of their loved ones to shine the light of truth in dark corners of society. They need no encouragement from scholars to undertake this work—it is a calling that they come to of their own accord. But that does not absolve us of all responsibility if we ask them to join us in particular projects that involve potentially unflattering exposés in contexts where the protection of human rights and the rule of law are spotty at best. Russia, for example, has a dismal track record, with 58 journalists killed since 1992, and a high number of suspicious suicides of journalists and editors uncovering corruption. Scholars can retreat from the consequences of “muckraking” when they leave the field, which means that they must be especially careful when and whether to enlist the assistance of others who must remain there.
There is also another ethical hurdle when enlisting the assistance of journalists. The confidentiality of informants remains sacrosanct, and adding a local journalist to the project should be treated with at least the same care as adding any other member to a research team. In authoritarian regimes, knowledge shared with a local journalist opens another avenue for an oppressive state to target an informant. The means by which scholars hide the identity of informants in scholarly publications (adopting pseudonyms and altering identifying characteristics) is often insufficient to eliminate suspicion for people who know a particular field site intimately. Informants should be made aware of any aspects of their story that might be shared with a local journalist and agree to sharing this information. The researcher should also exercise discretion. Occasionally, informants in authoritarian regimes are less concerned with potential repression than we, as researchers, are willing to accept on their behalf. Even with permission, we should be attentive to the potential outcomes of sharing data and err on the side of caution.
It would be delightful to uncover the truth behind every rumor in an authoritarian regime. And as the investigative approach above implies, even some corroboration may be better than none. However, in authoritarian states, uncertainty of rumored information must generally be accepted as a given. Sometimes, no matter how much we would like to know the truth, we simply cannot. The next three strategies for analyzing rumor data assume that verification is impossible and asks: how else might these data be used?
The cultural approach: An ethnographic gold standard
The most solid strategy for tackling a rumor in ethnographic data is also the one that is most common and comfortable for the ethnographer: cultural analysis of the rumor’s content. What is this rumor about and what does it convey regarding meaning and subjective understanding? This kind of analysis is ethnography’s bread and butter, and I do not need to belabor the methods of cultural analysis in the pages of this journal. Suffice it to say that the content of rumor can provide insight into general social understandings because, even though the rumor may not be true, it has sufficient plausibility to be deemed important information to pass on (Fine, 2007). A false tale may succeed in spreading as rumor because of its fidelity to a larger social narrative that is commonly accepted; the story makes “social sense” (Goffman, 1966; White, 1980). Uncovering social sense often motivates the ethnographic endeavor.
The content of any rumor is necessarily specific to the field site and to the topic of study, and there is no singular rubric for interpreting its content. As Kapferer (1990: 145) states, “Interpretation of a rumor must incorporate the place, country, or culture in which it circulates as well as the rumor’s particular public.” But as the field of cultural studies has long maintained, stories offer symbolic representations, often with a dichotomous moral valence (Levi-Strauss, 1966). Analyzing a rumor might profitably begin by asking: Who are the people involved, and what social groups might they represent? What is the assumed motivation for the actors in this rumor? Are they good or evil? Polluted or pure? These dichotomies can be utilized to uncover some general cultural understandings.
But in the authoritarian context, it is particularly important to attend to the subtleties of power that rumor often conveys. We are aided in this effort because rumors often have within them embedded assumptions about social action. The analyst might ask: Who has agency in this rumor, and who does not? What is at stake for the different people involved? The same question asked of the people represented in the rumor can also be asked of the individual to shares it: In telling this rumor, what is the speaker signaling about his/her values? Why is this information deemed important to share? Does the rumor suggest a course of action, and what are the power dynamics implicit in that suggestion?
When taking a cultural approach to rumor content, ethnographers in authoritarian regimes may have an added advantage that other research settings do not. The sheer preponderance of rumor in these settings allows the researcher to test the emergent symbolic system across multiple rumors, rather than relying overly much on any single rumor. While certainly there is profitable cultural knowledge to be gained by a deep interrogation of one prevalent rumor, the rumor-dense environment of authoritarian governance allows for new and expanded types of analysis. Do the general cultural codes contained in different rumors match one another, or is there a divergence? For example, is the villain always the state, or do some rumors differentiate between low- and high-level bureaucrats? If there is a divergence between rumors in terms of their underlying cultural schema, is it systematic in some way? What might explain that, if it occurs?
In addition to the content of a rumor and the cultural understandings it offers, the analyst ought not to neglect the social context in which rumors are shared. This information might also yield important insights. Because rumors are shared, they can help trace the contours of important social boundaries within the field. The ethnographer can ask: Who is sharing this rumor with whom, and in what context? Who is deemed trustworthy, and who cannot be trusted? Who comprises the in-group and who are the outsiders? Also informative might be in investigation into the geography of rumor in the field site. In addition to the social networks of sharing, the analyst might also attend to the occasion and setting wherein rumor-telling takes place. When do informants engage in rumor-spreading practices and, importantly, when do they refrain from doing so? Are there social spaces where rumor-spreading is less welcome? Why is rumor more likely in this environment and less likely in that one? What kind of work might rumor-telling be doing in these particular social environments? In addition to elucidating social boundaries, thinking about the “space” and the “place” of rumor might have analytic utility for moving beyond the often dichotomous symbolism of the rumor’s content.
The structural approach: Deploying the counterfactual
Rumors proliferate because (1) information is scarce and/or (2) the sources of information are deemed untrustworthy. Rumor indicates a breakdown of institutional trust because the regular channels by which established truths are created or disseminated have lost their legitimacy. Such a loss may be due to ignorance or experience. Either the population sharing rumors is distant enough from the institution in question that its inner workings are opaque (ignorance), or previous encounters with the institution have left a legacy of distrust (experience). However, it is not simply ignorance and experience that breeds distrust—a social group must also perceive itself to be powerless to alter or avoid the institution in question. In authoritarian regimes, all three conditions are often present in a perfect storm for the production of rumor. While cultural analysis is always important for ethnography, a glut of rumor in the data should point the researcher toward a structural analysis. It is not simply subjective understandings, but social structures, that abundant rumor makes evident.
Rather than considering rumors to be true or untrue, a structural approach would examine the institutional configurations that prevent these rumors from becoming facts or falsehoods. Ethnographers who cannot verify a rumor can play with the counterfactual instead: What would it take to create a situation where this information could be verified? What factors are inhibiting the production of verifiable knowledge in this instance? In those cases where rumors differ from official accounts of the same phenomenon, then the issue is not lack of information, but rather lack of trust in the source. When the problem is one of trust, then the counterfactual would ask: What mechanisms are missing from the social structure that would help to establish confidence? How would the institution have to change in order to incorporate the excluded groups who are presently sharing rumors instead?
The point of asking counterfactual questions is not simply to acknowledge that institutions are opaque and untrustworthy; this condition often accompanies authoritarianism and is itself unsurprising. But in the process of answering these counterfactual questions, the researcher may be prompted to view the field site from a different perspective. The counterfactuals yield a diversity of possible “other” conditions. They move the analysis beyond a broad brushstroke of “authoritarian” versus “liberal democratic,” and toward more specific gradations in the field site at hand. Now the analyst has specific alternatives that can become points of attack when thinking about the implications of the existing institutional structure.
Of course, no structural analysis can do without the added question: Who benefits from the current state of affairs? The ruling regime is not necessarily the only party that has a vested interest in institutional obscurity and bendable, if not breakable, legal regimes. While teasing out the counterfactuals, the ethnographer also may be prompted to look at moments in history that can elucidate the present situation. Is there evidence in the past of widespread trust of this institution—goodwill that has since been lost? Have there been attempts in the past to secure access for different, excluded social groups? What was the result of these attempts?
Perhaps the most important opportunity that comes from the counterfactual analysis of rumor is how it can be brought to bear on non-rumor data. The insight gained into the social structure by means of rumor can be turned into new hypotheses for interrogating field notes. As I have mentioned before, not every ethnographer in an authoritarian country intends to study rumor, but the data end up offering rumors, regardless. Having established a clearer understanding of the social structure by examining these rumors that crop up, the analyst may then ask: What else besides rumor does this institutional context engender? How might another issue (say, the research topic) be impacted by this same structure? If the public is in a state of uncertainty and distrust for specific, identifiable reasons, then rumor is probably not the only consequence that will emerge from the field notes. Thus rumor—even when not the topic of study—can work to enhance insights made using other ethnographic data. Posing counterfactual questions about rumors can begin a process whereby the social structure may be analyzed by means of cultural data. In the case of authoritarian regimes, such an analysis is crucial.
The social–psychological approach: Selective use of auto-ethnography
Because rumor is information of dubious veracity, its consequences in terms of human behavior are variable. Much depends upon the social–psychological reaction to rumored information. Some of this we can access through interviews and observations. But there is still another important avenue to understanding the social–psychological reaction to rumor: ourselves. When conducting fieldwork in a rumor-heavy context, it can be analytically beneficial to reflect on whether and when you—the researcher—being to deploy information learned from rumor in your own life. For lack of a better term, I am suggesting a kind of auto-ethnography vis-à-vis rumor knowledge. Auto-ethnography is the process of gathering field notes about one’s own experiences and internal state for the purpose of later conducting ethnographic analysis on oneself (Chang, 2016; Ellis, 2004). As a method, auto-ethnography is not without its problems, and has a sufficient number of critics (e.g., Delamont, 2007; Freeman, 2011; Jackson and Mazzei, 2008). To illustrate the usefulness of auto-ethnography for analyzing rumor in authoritarian regimes, I offer the following example from my own research.
One of the major events that took place during my fieldwork in Irkutsk, Russia, was the implementation of a law designed to limit NGO activity within the Russian Federation. Colloquially called the “Foreign Agent Law,” this piece of legislation had the pernicious effect of targeting, harassing, and, eventually, shuttering several nonprofit organizations that were particularly critical of the regime. On the pretext of limiting foreign intervention in Russian politics, the nation’s prosecutors began a purge of particular social organizations—some of which were participating in my study.
To learn more about the law and how it was being implemented in Irkutsk, I paid a surprise visit to the city prosecutor’s office. Having previously been denied interviews by state bureaucracies, I hoped that I might gain an audience if I just “dropped in.” From this visit, I learned that the regional prosecutor had provided the city prosecutor with a list of organizations to target. With that information in hand, I planned to pay a similar visit to the regional prosecutor, to follow-up on the origin of this list. But a couple days later, when I was walking down ulista Karla Marksa, one of the main streets in Irkutsk, I slowly became aware of a man about 30 feet behind me who seemed to be tracking my movements. It was an unnerving sensation. I made some sudden changes to my route to see whether he continued to follow my path, and he did. The logical part of my brain tried to convince me that it could be a coincidence, but an animal fear had gripped me and I was tense with adrenaline. I ducked suddenly onto a departing bus and left the fellow behind.
As I reflected on that bizarre afternoon, the “known unknowns” of life under an illiberal regime came unbidden to my mind. The coulds, the possiblys, the might-haves that had saturated so many of the conversations I have had with Russian activists and others suddenly seemed less academic and more existential. All the rumors of burglarized apartments, staged muggings, and framed killings began to define the parameters of a liminal situation, where a questionable event begat even more questionable futures. I had to decide what this incident meant to me, and my mental debate went something like this: If it were really the FSB then I shouldn’t have been able to spot them, right? If the government was spying on me then they wouldn’t have been so obvious about it. But some people say that the FSB will deliberately show that they are watching you, to mess with your mind. But so what if the FSB was following me? What is the worst that could happen? Anna Politkovskaya, Mikhail Beketov, Oleg Kashin, Alexander Litvinenko, Sergei Magnitsky [i.e. individuals rumored to have been assassinated by the Russian government.]…Except they were all uncovering massive corruption. I’m just writing a dissertation on environmental activism. I’m a nobody. If you are a nobody, then what does it matter if you are hit by a car or violently mugged? Alright, that is just crazy talk. Something could happen to your husband. He doesn’t speak Russian – he would be an easy target. Remember what happened to [another activist, whose family member who was framed for murder]? No! This is crazy. This is all crazy – I probably wasn’t even being followed! Your apartment could be rifled. That’s what happened to [other activist]. Your laptop and files could be ‘stolen,’ like [other activist]. They wouldn’t bother to do that…would they? You could lose your visa. Shoot. I really could lose my visa. I know of people who were deported for less. You could lose your visa. Yes, I could lose my visa.
To this day, I cannot even say that I am absolutely certain that I was being followed by anyone that day, let alone that my alleged stalker was indeed a government agent. What I can say for certain was that I allowed uncertainty to alter my behavior to the benefit of the authorities. My research is potentially less complete than it might otherwise have been, and, at some level, I have only myself to blame. Rumors of state repression created the possibility of repression in my mind beyond what my own experience would have anticipated. I was not directly harassed, threatened or warned. 1 I took the warning on myself. Even if I did not believe that something extremely bad would happen to me, the fact that I suspected that something extremely bad had happened to other people made it easier for me to believe that something more reasonable, but still bad—like losing my visa—could happen to me. Rumor made the possibility of extreme state repression feasible, and that made the threat of moderate repression (losing a visa) feel downright probable.
Although I did not make my subjective experience the center of my ethnographic fieldwork in Russia, I can say that paying attention to my own thoughts and behavior in relation to the rumors I was encountering eventually helped me to improve my analysis of other ethnographic data at my disposal. My insight into the social–psychological consequences of the rumors I had heard could be brought to bear on my observations of others, helping to explain their behaviors and beliefs. I did not publish any results as “auto-ethnography,” but the practice of attending to my inner experience with rumor helped me to construct better theory about the social meaning of the rumors I was hearing and their subsequent indirect effects in my informants’ lives. The subjective experience of living with uncertain knowledge taught me far more about how power operates in authoritarian regimes than my repeated readings of Hannah Arendt or Michel Foucault had ever provided.
Conclusion
Social science does no favors by assuming the “correctness” of information that cannot be substantiated. Unverifiable information remains unverifiable. Just because a government is authoritarian and the rumor comes from the subaltern does not make it true. But discarding the unverifiable is a false assurance for claiming epistemological accuracy. Every instance of lung cancer is not caused by smoking, every drought is not caused by climate change, and every murder of a government critic is not committed by an agent of the state. That does not change the fact that smoking causes cancer, climate change causes droughts, and authoritarian regimes kill their critics. Rumors may be right or wrong—acknowledging our inability to verify certain claims is honest and it is necessary. However, while ensuring that readers understand the limits of rumor as knowledge, it is also incumbent upon us to explain the sociology of rumor and what this unverifiable information can yield. If we have done our work as analysts by, for example, following the strategies above, this task will be made easier. If done well and done right, then the empirical uncertainty of the rumor need not weaken the data-driven analytic conclusion.
At the end of the day, rumors remain the content of people’s minds, shaping the structure of thought and patterning people’s behavior. Believing (or not believing) in rumored information is akin to believing (or not believing) in the “evil eye.” The standard dictum of the social scientist—that these become real in their consequences—still applies. The same framework that undergirds all qualitative methodology continues to guide the analysis of data that contains possibly true, but unverifiable, rumored information. However, it is the contention of this article that there are social contexts that give rise to higher levels of rumor, and that when ethnographic research yields an unexpectedly large amount of rumor, there should be additional effort made by the analyst to examine the data beyond that of narrative or belief system. Rumor—unverifiable though it may be—can be useful data, but it requires treading carefully.
Recognizing the risks involved, this article has suggested several strategies for confronting unverifiable information in ethnographic and qualitative data, particularly when these occur in great abundance, such as happens within authoritarian regimes. The strategies presented here are not intended to exhaust the methodological possibilities and potentialities. But hopefully they can provide some guidance to the ethnographer confronting a disconcerting amount of rumor in informant conversations. By incorporating findings from the study of rumor itself, ethnographers who are attempting to study other social phenomena (e.g., activism, gender relations, the workplace) in contexts rife with rumor may have an added avenue for analyzing this problematic data. As authoritarian governance appears to be experiencing a global resurgence, it is likely that ethnographers will find their data increasingly populated with unverifiable information. Hopefully, with these strategies for analysis, they can be better prepared to utilize it.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
