Common sense commodifies the secret, alienating the value of its content from its social context. But a secret perfectly kept dies in its circle of initiates. Few secrets, however, are dead on arrival, since their seduction lies precisely in their revelation. Most things said to be hidden are in fact nurtured through the processes of calculated concealment, allusion, and revelation, the secrets propagating themselves through circles of conspiracy, rumor, and gossip. As Tim Jenkins observed, “What is concealed, and the reasons for its concealment, serve to distract attention from the dynamic of the secret: what at first sight appears to be static and indeed dead, possessed by and known to only a few, kept in some dark place, in fact has a life and movement of its own; the secret propagates itself through a structure of secret and betrayal” (Reference Jenkins1999: 225–26).

The studied movement of disguise and disclosure of secrets is intrinsic to power (Canetti Reference Canetti and Stewart1962) and, as Simmel (Reference Simmel1906) and Goffman (Reference Goffman1959) taught, is a primary process of social life. Secrets incite suspicion and intimacy, jealousy, schism, and trust. Their power lies not in their content, which often proves a disappointment upon revelation, but in the rhetoric of their mystification. Mystified knowledge—the secret—can be owned, flaunted before or withheld from others, exchanged for allegiance, or sold (through treason or blackmail). Gossip, then, is not merely witness to the weakness of chinwags who cannot keep a secret. And to dismiss the stuff of such conversations as the flotsam of boredom and fantasy is to ignore an entire domain of social production.

This essay is about the life of one secret in the circles of its relentless revelation. More specifically, it is about the rhetoric of secrecy surrounding a caste of professional thieves called Kanjars in a rural part of the western Indian province of Rajasthan. Much of my material consists of what I initially dismissed as “mere gossip,” but which eventually came to the fore of my understanding not only of the social location of the Kanjar community but also of some basic aspects of local society. In popular rhetoric, official narratives, and their own presentation of selves Kanjars are a secret people possessed of secret knowledge, a secret tongue, and magical powers. Secrecy surrounding Kanjars is a matter known and circulated in the open. Unlike “public secrets” that are “generally known but cannot be articulated” and that are protected by the silence of their keepers (Taussig Reference Taussig1999: 5; also Herzfeld Reference Herzfeld1987: 95–122; Reference Herzfeld2005), Kanjar secrets are enthusiastically articulated by bazaar gossips, policemen, and Kanjars alike. Much like “the active milling, polishing, and promotion of the reputation of secrets” described by Paul Johnson in a Brazilian context (2002: 3, his italics; following Simmel Reference Simmel1906: 486), the discourse of secrecy that surrounds Kanjars generates mystified knowledge that acts as a screen for the absent, the well-known, and the banal. Whereas the public secrecy described by Taussig protects content that ought to remain hidden, mystification that surrounds Kanjars creates secret content where there is none. And it is not what is concealed, but the fact that something is hidden that is significant. The process of talking about Kanjar secrets creates a hidden and, as I shall show, essential social domain. Secret spaces, relationships, activities, and people like Kanjars play a crucial part in the maintenance of polite and proper life on display, making the generation of secrets a central aspect of public life. While rhetorically excluded from proper society, in practice Kanjars occupy some of its innermost domains.

Secrets Revealed

A few years ago I was in Jaipur, Rajasthan's capital, drinking midday whiskeys with some aristocratic Rajput families, people I like to visit periodically to keep myself abreast of current gossip. Our conversation drifted leisurely from polo ground drama to the new zinc factory inauguration party, the matrimonial matches to be made with the Jodhpur Rathores (a high-ranking Rajput clan), and the recent kills by man-eating panthers on the border with the neighboring province of Madhya Pradesh. I brought up Kanjars, an “ex-criminal” community connected in the minds of many with illicit liquor production, prostitution, and petty theft.Footnote ¹ At the time, I had come to frequent some Kanjars in a local slum. I mentioned to my Rajput acquaintances, rather cautiously, that I had heard that certain families in their midst maintained mutually beneficial relationships with “criminal castes,” and that I wondered if the present company could help me to learn more about Kanjars and other such groups. Expecting tacit suspicion at best, I held my breath, prepared to blame the midday heat and the spirits for the implied accusation of criminal involvement.

Instead of suspicion, my question was met with enthusiasm, and brought forth a host of stories about the bizarre beliefs and habits of Kanjars. My Rajput informants, eager to see me fill my notebooks with Kanjar “ethnography,” overwhelmed me with “secrets” of Kanjar life. With a raised brow and in a lowered voice one of my informants said that Kanjars have many secret practices, the details of which he proceeded to describe; from the training of wall-climbing lizards (used in house burglary) to bizarre wedding arrangements, blood-thirsty goddess worship involving human sacrifice, and ritual rooftop defecation, my company claimed to know it all. In the heat of the moment someone even suggested that there was little need for me to spend time among Kanjars because I could learn most of their secrets from the present party and other similarly “knowledgeable people,” of whom, as I came to realize, there was never a shortage.

Common account describes Kanjars as an external people—socially, morally, and categorically. Their indecency is not simply common knowledge; it is proverbial. Across northern India, for instance, one often hears parents say to ill-behaved children: Kanjarõ kī taraf mat karo! (“Don't act the Kanjar way!” or “Don't do it like a Kanjar!”), and “Kanjar” used as a term of abuse akin to the English “bastard” or “whore.” People describe Kanjars as gandhā (literally, “dirty”), a term that refers to things vulgar, uncivil, and generally unsuitable for decent folk; the fact of a young “English girl” living in their midst provoked much disbelief and consternation in the area.Footnote ² The exclusion of Kanjars from mainstream society is reified in the observable sphere of everyday life. Kanjars almost invariably live in settlements (bastīs)—never villages (gāoṅs) proper—that are separate from ordinary multi-caste villages or in slums on the outskirts of cities and towns, where they are engaged (both in repute and in deed) in all sorts of immoral, illicit, and illegal activities: distilling, drinking, and selling country liquor (dāṛū); stealing, selling, and eating goats and sheep; and prostitution.

The concept of a secret community—which links notions of power, secrecy, magic, and vice—has a long history in South Asian narratives, which warrants a brief note here. Ancient Indian literature (ranging from legal, statecraft, and liturgical texts to stories and epics) describes communities of peripatetic and jungle-dwelling peoples not only as socially, ritually, and morally set apart, or hidden, from ordinary society, but also as necessarily involved in thievery, understood as the art of disguises and the business of hidden possessors of secret knowledge and power.Footnote ³ Respect and fear afforded to thieves in ancient writing has to do with their hidden nature, which has been cause for much fascinated elaboration in ancient narratives on the subcontinent. Much like my Rajput acquaintances in Jaipur, authors of texts about the secret thieving communities describe their hidden worlds in substantial detail, imagining them as hereditary trade guilds that are “banded, cartelized and organized groups, that live together” (Bloomfield Reference Bloomfield1926: 205). In various texts thieving communities are said to possess highly formalized sets of professional skills, wear distinctive clothing, and work under the auspices of their own deities.Footnote ⁴ In short, the thieving communities of Brahmanical narrative have been imagined as castes like any other. Thieving, nevertheless, is a special business unlike pottery or the weaving of baskets. Contra Chris Bayly (Reference Bayly1996) and Sandria Freitag (Reference Freitag, Anderson and Guha1998), who have argued that before the arrival of the British on the subcontinent theft was a legitimate part of South Asian politics, neither in ancient narratives nor in pre-colonial politics has thieving been treated as a plainly legitimate trade. In Brahmanical narratives, the hidden lairs of robbers are dens not only of magic and power, but also of various bloodthirsty and barbarous vices. Although the colonial rubric of Criminal Caste shifted emphasis to the moral depravity and social danger of such groups, the idea of a hidden community of professional, hereditary, and deviant thieves, contrary to what the bulk of modern historiography would have us believe, predates British presence on the subcontinent.Footnote ⁵

Describing Thuggee in the first half of the nineteenth century in terms reminiscent of older Brahmanical writing, William Sleeman (the British pioneer of the “discovery” and “eradication” of Thuggee) defined this “cult of murderous stranglers” as a secret society. His voluminous catalogue of the Thugs’ “underworld” included descriptions of their secret argot, ritual practices, omens, modi operandi, and bloodthirsty deities (1836; Reference Sleeman1839).Footnote ⁶ As Radhika Singha observed, it is through claims to “mystery unveiled and mastered that a group of officers of the Political Department had lobbied for operations against this [Thug] ‘murderous fraternity’” (1993: 83). By the late 1840s, Sleeman shifted his attention from Thugs to the newly discovered “fraternities of hereditary robbers” (Reference Sleeman1849), communities that by 1871 came to be defined as Criminal Tribes in colonial legislation. The Criminal Tribes, with Kanjars figuring prominently among them, continued to be defined as hereditary professional guilds engaged in a perverse and secretive trade. Although their criminalization in colonial law and the attendant policing and penal measures removed communities like the Kanjar still further from mainstream society,Footnote ⁷ the terms in which Criminal Tribes were identified were not solely a figment of colonial imagination.

The mystification of thieving communities took a new lease on life after India's independence. Retaining their criminal identity in official practice, if no longer in statute, Kanjars and related groups continue to be understood as a people hidden from moral society and the gaze of the law.Footnote ⁸ In 1998, for instance, the chief minister of the Indian province of Madhya Pradesh lamented that the state's educational programs had little effect on “the criminal instincts” of Pardhis (another “ex-criminal tribe”) because, as he explained, their instincts were nourished by the “hidden nature of their society, which is resilient to the ideas of modern education” (Telegraph 1998; my emphasis). A recent newspaper article about Kanjar peacock poachers described them as descendants of “famed highway plunderers … said to be habitual criminals and always carry country-made pistols and crude bombs with them.” The alleged inability of the police to track down the poachers is written off with the claim that Kanjars “disappear on the spot into their secret lairs” (Srivastava Reference Srivastava2005). While local monographs on “ex-criminal castes” call for their “upliftment” (Bhargava Reference Bhargava1949; Garg Reference Garg1965; Shah Reference Shah1967; Gandhi Reference Gandhi2008), administrative descriptions of Kanjars propagate old notions of their secret nature.

Similar allegations fill the records of local police stations. In areas that house Kanjar settlements, “Village Crime Note Books” (or VCNBs), which are meant to be chronicles of crime in each village within the jurisdiction of a police station, emphasize the inherent secrecy of the Kanjar community. A characteristic account opens one such document:

Another document, the “Compendium Concerning Kanjar Gangs,” an ethnography of sorts of local Kanjars compiled by police officers and kept on file in the district police headquarters,Footnote ⁹ opens with the following statement: “The Kanjar caste is a criminal caste. From ancient times these people have roamed about committing group crime [including] theft, roadside burglary, looting, and dacoity [gang robbery]. They are a caste that is addicted to crime. They are very difficult to find because they can run very fast and when they commit a robbery, they disappear into the jungle or cross over the [state] border” (Compendium n.d.).Footnote ¹⁰

Kanjar secrecy comes almost as an afterthought and an apology in both documents, each of which makes a point of noting the inability of the police to round up Kanjars. As we shall see, there are reasons for such disclaimers, and the mystifying rhetoric they invoke masks a certain set of relationships meant to be hidden from the official gaze. The same terms and images—lizards, addiction to crime, magical healing techniques, secret tongues, and supernatural endowment—are as much part of the official story as of rumor in the bazaar. VCNBs are curious places where popular images fill official documents and are cast as expertise and consequently as official knowledge. A section of regular entries in one such notebook reads as follows:

The primary sources for such chronicles are the stories villagers tell about their Kanjar neighbors. The constables inevitably “hear that so and so visited the town,” “the village people believe that so and so associates with the Kanjars,” and so forth. The information is always gathered from “reliable sources” and the entries invariably end up “complete and correct.” This particular VCNB, which documents thirty-four years of patrolling one village, records only a single instance of direct communication between police staff and local Kanjars: a particularly earnest officer noted that he conveyed to the residents of one Kanjar settlement the evils of meat eating, drinking, and theft. Otherwise, the record is filled with reports of Kanjar disappearing acts. Whilst one may expect official accounts to be free of rumor and fantasy, in fact police and bazaar knowledge share a great deal. Contrary to the global rhetoric of modern statehood, state authorities are not separate from “the people”; police officers are people too, and police accounts of Kanjars are as full of folkloric stereotypes as is drunken gossip. Popular narratives not only inform, but also legitimize police records. While statements regarding Kanjar magical disappearances may raise an anthropologist's smile, for villagers they are matters of obvious fact. When I relayed the content of some police descriptions of Kanjar to local farmers, they nodded with approval: “Of course,” said one, “everyone knows that Kanjars can disappear—they have magic [jādū]—nobody knows how they speak and how they steal and where they go and whence they come.” The police, however, make it their business to know more about Kanjars than do ordinary tattlers. But such knowledge is not meant for the official chronicles.

The record contained in police stations is what officers call a “formality” (using the English term), necessary as evidence of their work. “Real information” (pakkī sūchnā), as one sub-inspector explained, does not make it to paper. “True Kanjar knowledge” (sachchī Kanjar jānkaṛī) is almost never written down because what is written is no longer hidden. Instead, records of contact with and information about Kanjars are passed down orally from one officer to the next within police stations and district police circles. In stations with Kanjars or similar “ex-criminal castes” under their jurisdictions, a constable and an assistant sub-inspector are typically designated for communicating with and gathering information about Kanjars. These officers often remain in their posts for many years (well beyond the maximum two-year term prescribed by the Rajasthan Police Act [2008: 14]), during which they cultivate close relationships with local Kanjars. Some bonds between officers in the lowest ranks may reach across generations, since positions in the police are often inherited. These officers are usually assigned to the circle of regular inspection in Kanjar villages, and can be often found in civilian clothes, chatting with Kanjars in villages or the bazaar.

Some of these men become what are known in the police as “Kanjar experts.” These are district-level officers, no lower than sub-inspector in rank, responsible for collecting intelligence from and about Kanjar gangs. It is under their supervision that the Kanjar “ethnographies” kept in the district headquarters are compiled. Most Kanjar experts have extensive experience and knowledge about the life and work of local Kanjar communities and most have long-standing relationships of patronage with Kanjar villages that they “adopt” (infin. god lenā) and with influential Kanjars in them.Footnote ¹¹ The practice of “adopting” thief-villages is a common institution of patron-client bonds between “criminal” communities (whether thieving, liquor distilling, or prostituting) across Rajasthan and northern India at large.Footnote ¹² Relations of patronage are typically initiated by the senior house officers (SHOs) of local police stations, who establish connections with village leaders, usually heads of thieving parties, who become both informers (mukhbīrs or mukhbars in Kanjari) and mediators between the police and their community. In exchange for information and a share of spoils, SHOs direct their staff to turn a blind eye to the activities of informers and their parties, stop or minimize the filing of false cases, “write off” arrest warrants at little or no cost to their informers, and make the filing of court cases redeemable with moderate payments.Footnote ¹³ The best Kanjar clients are, of course, those most heavily involved in thieving, ideally gang leaders.Footnote ¹⁴

Claims to Kanjar magic and secrecy written into police ethnographies leave the reader mystified, as they are meant to; within the police, information about Kanjars is framed as privileged knowledge, or “expertise.” Transmitted orally from one expert to the next and masked on paper by allusions to magic and secrecy, what the police do know about Kanjars is preserved within and through the rhetoric of secrecy. The relations and activities screened by this rhetoric are of interest, and I will return to them, but first I shall turn to what Kanjars themselves make of such claims.

Nurturing Secrets

When I first arrived at my field research site, my Kanjar informants insisted that nobody outside of their community knew their language.Footnote ¹⁵ Not only was it unknown to outsiders, they said, but it was intrinsically unknowable to them. As Kanjars (much like their neighbors) explained it, this language (āpas kī bolī, literally “insider language”) did not lend itself to learning as such. Rather, it propagated itself as instinctual knowledge, acquired only through being born and raised within the Kanjar community. As I slowly picked up the Kanjari dialect (which turned out to be a slightly modified form of the regional Mewari language), consternation spread: either I too had magical powers, much like the Kanjars, or I had been sent in by sarkār, the government, itself subject to much mystification. As I picked up more words, Kanjars themselves began to insist that there was yet another level of secrecy to their language, a secret tongue (pārsī) beneath their everyday speech. This tongue turned out to be a professional argot consisting of no more than a few dozen words. As my friends in the settlement taught me more of this language, others insisted on the existence of two, four, or even a dozen secret tongues aside from that which I was learning, so that no matter how much I tried I would never have the “total knowledge” (ṭoṭal jānkaṛī) of their secret language.

One of the foremost gang and community leaders in the village, who became my best informant and friend, nevertheless persisted in teaching me the tongues, which constitute a kind of pig Latin, formed by standard substitutions of phonemes and additions of prefixes. Our lessons were a transgression, since while others are to know about Kanjar secrets, they are not to be privy to their content. (It was such deviousness that made for his thrill in teaching me the tongue.)Footnote ¹⁶ Each time I came up with a “secret” word, my Kanjar neighbors and hosts assumed that I had not “learned” it as one does English in school, but had absorbed it on the basis of a natural, albeit a strange, predisposition. As it grew increasingly apparent that I was beginning to grasp the content of most “secret” conversations, local Kanjars reached a general consensus: I must have been a Kanjar in a previous life.Footnote ¹⁷

Another sphere of secret knowledge is comprised of the “eighty-four wisdoms” (chaurāsī buddhiyāṅ), known only to Kanjars.Footnote ¹⁸ At the end of my stay, my friend (the one who had taught me the secret tongues) and I drove my motorcycle beyond the boundary of the settlement and settled under a banyan tree for my long-promised lesson about them. The eighty-four wisdoms he revealed turned out to be a varied collection of thieving modi operandi, ancestral practices, and regulations regarding matters like bride price and incest. As many of the wisdoms replicated the speculations that I so often heard in the bazaar, I had to hide my disappointment. The long-awaited revelations of ancient mores and secret practices, from the use of wall-climbing lizards to human sacrifice, reiterated what I had thought to be tall tales on the lips of others. When I asked my friend why I had not observed any of this in their village, he explained that these were very old practices and that, although humans were no longer sacrificed and lizards no longer reared, it was important that all Kanjars knew these “secret signs of [their community's] distinction” (gupt pahchān). The divulgence of these wisdoms, I was warned, would bring about the “ruin” (barbādī) of their community. Much like their neighbors, my Kanjar hosts also claimed that each member of their family had a treasure (arj) hidden in the jungle. I never learned the location of any such treasure, and neither, I believe, did my Kanjar informants. Such revelations echoed an observation made by Paul Dresch, anthropologist of Arabia: “The experience [of fieldwork] was not well described as simply ‘dialogue.’ Nor was it ‘data collection.’ It was more like a trade in secrets, though in a curious way. The more thriller-like the material, the less difficult to gather, to the point where one [is] constantly being told things unfit for print: the more mundane and empirical the published facts, by contrast, the more collecting them resembled clandestine intrigue” (Reference Dresch, Dresch, James and Parkin2000: 109–10).

While anthropologists are spoon-fed sensational material, the plain stuff of ethnographic interest may long remain obscure to, and deliberately obscured from, the ethnographer's sight. While many secrets are imposed on the researcher, matters that ought to be more mundane are often learned obliquely in the long and laborious course of fieldwork. The basic data of ethnographic writing are others’ secrets—family relations, household economics, or religious affinities—which are rarely mentioned in everyday conversation and are often revealed only uncomfortably. The very process of asking questions about such matters can raise alarm, making otherwise innocuous information suddenly unfit for discussion, leaving the interview not only a less than desirable but a potentially detrimental ethnographic tool. The “secrets” you ask for are then replaced with the secrets informants are happy to tell. Do you want secrets? We will supply you with some! Such readily divulged secrets, although often un-publishable, carry ethnographic weight of their own. Not only do they forge a sense of intimacy among their keepers, as Dresch suggests, but they also define a sphere of their keeping as special. By relaying them, informants make anthropologists feel that they are “in” of exceptional value and interest, often without actually letting them inside.

While little of the contents to which I thus became privy came as a revelation, the presentation of certain knowledge as hidden proved instructive. As it turned out, secrecy was just as central to Kanjars’ presentation of selves as it was to their mystification by others. The trick here lay in revealing that secrets exist while concealing what they were (cf. Johnson Reference Johnson2002: 3). Unease surrounding the revelation of wisdoms and tongues had nothing to do with a fear of policemen learning things that they either already knew (such as the modi operandi) or did not care to know about (such as bride wealth prescriptions). Instead, revelations of secrets, by making obvious things that are meant to be hidden, threatened the mystery and the invisibility that are so central to the role of Kanjars in local life. As one of my Kanjar informants succinctly, if enigmatically, put it, “Our secrets [gupts] are our watering well [kūṅā],” invoking a local metaphor for a source of livelihood.

Secrets Concealed

Contrary to the claims of popular discourse, Kanjars are centrally involved in the proceedings of local life: they are connected to various “patrons” (jajmāns), employers, and “friends” (dosts) for whom they perform crucial services. Such relationships can constitute one-off employment or outlast several generations. Village communities hire Kanjars as watchmen and a number of Rajput landlords (zamīndārs) in the area continue to employ them as watchmen, thieves, and intelligence agents.Footnote ¹⁹ One Rajput of my acquaintance, for instance, told me that his Kanjar “servant” (kamīn) had been in his family's service for three generations, helping them to “push down others who stuck their heads up too high.”

Thieving groups have played an important part of local affairs, and continue to do so, through their involvement in rural systems of watch and ward, intelligence supply, and conflict management on political levels ranging from princely disputes to disagreements between farmers. Others have noted the policing and spying functions of such groups (e.g., Bayly Reference Bayly1996: 163ff, 372ff),Footnote ²⁰ but little has been said about their role in disputes. A clandestine nocturnal raid (gemī in Kanjari) of a house is the standard method of “pushing down” (infin. dabānā) others. While such raids can bring material gain for participants through thievery, their main aim is not resource extraction, and the commissioning party often gains little or no material profit. Instead, the raid is employed to intimidate opponents by penetrating the rivals’ homes so as to “expose weakness” (infin. kamjorī dekhānā) and leave them in a subordinate position relative to the supplicant. If successful, raiding contests are resolved with the reestablishment of rapport between the opponents and the restitution of stolen property.

Almost a year ago, a Rajput of my acquaintance awoke to find the contents of his living room (including a coffee table, chairs, curtains, stationery) neatly arranged in his garden.Footnote ²¹ The burglars took nothing, but had left a clear message; the terms of sharing a field, on which he had been unable to agree with his cousin, were soon negotiated and courteous rapport between the two restored.Footnote ²² Such raiding tactics are not restricted to conflict negotiation within the Rajput elite, but are a common means of communication in local society.Footnote ²³ In the decades since independence, however, many bonds between Kanjars and their Rajput sponsors have been severed along with other types of local patronage (jajmānī) ties. Today only two Rajput families in Fararpur continue to employ their family Kanjar clients.Footnote ²⁴

Kanjars also find employment as “watchmen” (chaukīdārs) in villages, orchards, fields, and family homes. For them, watchman-ship (chaukīdārī) is one of the most reliable sources of income and almost two-thirds of adult men in the settlement of my research are currently so employed. In fact, it is the only legal source of income in the community aside from small-scale cash crop agriculture. Yet this legal and fully legitimate business depends on illegal activity. Watchmen's duties lie both in policing their assigned domains (including preventing theft, tracing culprits, and retrieving stolen property) and in negotiating conflicts within the communities of their watch. Kanjars are liable, in effect, for protecting villages from themselves, their responsibilities being restricted to the control of their own gangs, families, and villages. More often than not, however, watchmen are engaged in negotiations of disputes within their patron-communities. When family conflicts arise, it is the watchman who is called upon to facilitate communication, and, if that fails, to “put pressure” on the unruly party through the thieving tactics already described.

Secret Agents

Involvement of Kanjars in local politics and social control is not confined to the “shadow economy” of informal relations, but is also central to the formal arena of local state governance. I have already noted that many Kanjars work for the police, providing them with crucial intelligence, means of control, and a source of income.Footnote ²⁵ In return for their services, a blind eye is turned to their activities, a practice that both Kanjars and police officers term “adoption.”Footnote ²⁶ Such arrangements explain why most active Kanjar gangs remain protected on the territories of local police stations, and why “history sheeters” (recidivists with a criminal “history sheet” on file with the police) remain so elusive in police conversations and official records. It is also why, these days, Kanjars rarely raid areas outside the jurisdiction of their patron-stations, and target neighboring villages with increasingly frequency. Such local concentration of their activity has provoked a rise of violence both within and against their community, something I describe in greater detail elsewhere.Footnote ²⁷ While standing beyond the pale of formal legitimacy, such relationships with the police constitute the habitual and often socially sanctioned realities of local political life. As raiders, watchmen, informers, and mediators, Kanjars are employed when communication between different parties falters.

Observers of Indian political negotiations—whether in caste and village councils (panchāyats) or the National Parliament—have marked them as lacking dialogue. Communication in panchāyats has been described as a trade in complaints, accusations, apologies, and “conciliatory” speeches, directives that neither respond to what was said nor provoke a rejoinder; that is, they consist of a monologic exchange of declarations not meant as responses or contributions to a dialogue (Gupta Reference Gupta1997: 48ff). The same communicative asymmetry marks proceedings in the Indian National Parliament, where havoc often ensues as parliamentarians rise out of order to shout out speeches that have little to do with discussion agendas, the particular questions raised or the authority of the speaker (Spary Reference Spary2010).Footnote ²⁸ Such a monologic style of communication defines interaction not only in the parliament or the village council, but also in communicative contexts ranging from the household to the courtroom.Footnote ²⁹ As some of my informants explained, respect (izzat) due to superiors (baṛā) prevents rejoinders from inferiors (chhoṭā), for it causes “anxiety” (gabaṛ) within the community and brings shame (śarm) onto not only the offending party but also onto the community at large. Put simply, in the context of interaction between rank unequals, dialogue amounts to speech out of turn and, as such, to a violation of the relational order. In order to avoid such violence, people employ clandestine means of communication, such as raids, and hidden go-betweens like Kanjars.

The monologic communicative style bespeaks a key aspect of local social life. The order of ranked relations which structures local society—referred to as “hierarchy” by old convention—assumes an asymmetry of exchange, communication, and rapport. This means that the order of communicative precedence is set in accordance to the social structures of super- and subordination: fathers have communicative precedence over sons, husbands over wives, elders over youngsters, patrons over clients, and so on. When the order of rank is unclear (outside of the particular locality, among unranked cousins or “friends”), difficulties of communication arise.Footnote ³⁰ Conflicts of interest and the need to speak out of turn do transpire and they call for a medium of communication. Not everyone can take on this role. If dialogue is prevented by a set order of relatedness that defines relationships within and between communities, such as castes, villages, or households, then conflict management calls for mediation from outside the ranked structure order—it requires external agency.

Mystified and rhetorically hidden, Kanjars are external to polite society, and it is their position as outsiders that makes for their involvement in some of the innermost negotiations of local life. This is best illustrated with an example: a lengthy dispute between a Gujar (herder) man and the wife of his younger brother, which was mediated by the family's Kanjar watchman. The woman stole five goats from the elder brother and sold them in a village some twenty kilometers away, and her brother-in-law was at a loss as to how to resolve the matter. Over the following two months, the watchman made a series of visits to both parties, relaying the concerns, and finally threats, of one to the other. The conflict was ultimately resolved with the woman paying for the goats and the watchman receiving one-fifth of the sum. To my query as to why they could not just talk to each other, the Kanjar watchman, exasperated with yet another obtuse question, explained: “She is his younger brother's wife [chhoṭī bhābhī]! How could he speak to her?” The victim of theft could not himself discuss the matter with her; as in most jātis in Northern India, contact between Gujar men and their younger brothers’ wives is strictly constrained and the relationship is visibly marked with the veil and the physical distance which men keep from their younger brothers’ wives (Freed and Freed Reference Freed and Freed1964: 153; Mandelbaum Reference Mandelbaum1970, 1: 64–65). Verbal communication between them is almost entirely banned: inasmuch as a woman cannot directly address her husband's elder brother, she is also spared the obligation to respond to him. Any persistence on his part to provoke dialogue would be interpreted as a violation of her, his, and the family's honor. Besides, as my Kanjar host explained, as an elder (baṛā), it was shameful for him to openly “beg” (māṅgnā) for the return of the goats. The family needed an “outside man” (bahār kā ādmī) for “instructing” or “counseling” (samjhānā) the woman.Footnote ³¹ “I am,” he added, “just such a man!” It was on this prohibition that the younger brother and his wife attempted to capitalize: by deputing the wife to steal the goats, he deprived the plaintiff of any obvious response.

Quiet resolution of conflict and the maintenance of integrity of families, or other communities, relies on preserving “respect” or “honor” (izzat) in the eyes of the “public,” be this public one's caste, village community, or the crowds of the bazaar. While Kanjars are rhetorically removed from polite society, they have access to internal family dealings, and it is by employing them that families save their face. Because Kanjars are hidden from—and thus not part of—public life, families can make them privy to their internal dealings without fear of their exposure to people who matter, and on whose judgment family honor relies. It is the Kanjars’ discursive removal from society that gives them access to some of its most intimate social domains that ought to be kept apart from the public sphere. This presents us with a paradox: life on display relies on persons and relations that must be kept hidden. In other words, the hidden is an intrinsic part of the public sphere. The secrecy of Kanjar persons, which locates them within this invisible sphere, sustains the visible face of social life.

The involvement of Kanjars in local relations does not always take the deferent form of “explaining” and much of it consists of penetration by raiding. The course of such raiding is meant to be hidden, from both state law and the public gaze. The clandestine ethic of thieving negotiations is maintained as much by perpetrators as by raid victims. Kanjars see their thieving activities as an art of disguise, an idea that, as I have already noted, has a long history. They pride themselves on entering, robbing, and leaving homes unobserved, and look down on those they call “new players,” who steal in broad daylight, leave evidence of their presence, or struggle instead of fleeing upon discovery. Victims of raids that they know were commissioned are no keener to have such violations exposed. In the course of my fieldwork, not a single such burglary was reported to the police or otherwise made public by victims and I learned about them only through my Kanjar contacts. If rumors about such actions do seep into the public sphere, they are assessed as matters of course. People describe raiding politics as the “internal matters” (andar ke māmle or andar kī bāt) of the parties involved, which are not meant for judgment. In other words, commissioned raids are commonly treated not as violent, detrimental, or criminal, but, on the contrary, as legitimate and indeed socially constitutive practice.Footnote ³²

The growing involvement of Kanjars in state policing creates yet another domain of relations meant to remain unseen. With the decline of Rajput patronage, new relationships of patronage have been forged between Kanjars and the police. That most Kanjars in the business of thieving are informers for the police, however, reveals more than the practical detail of such arrangements between Kanjars and their police patrons, bespeaking something much more general about the workings of modern governance and statehood at large. One sub-inspector described the space occupied by the “secret agents,” as officers jokingly call their Kanjar informers (using the English term):

The rhetoric of separation between the government (sarkār) and the criminals (aprādhī log) invoked by the officer creates a gap where crucial links ought to lie. The contrasts between formal and informal, official and unofficial, criminal and legal, or government and society that ground modern state ideology draw boundaries between persons and within spheres of relationships that are not, and cannot be, kept pure of each other (Abrams Reference Abrams1988; Mitchell Reference Mitchell1991; Fuller and Harriss Reference Fuller, Harriss, Fuller and Bénéï2001). The old Indian dictum that “it takes a thief to catch a thief” is contrary to the idea that those who steal and those who police are not only different people, but are in principle part of separate social and institutional spheres. As the sub-inspector points out, much policing in fact takes place through the agency of those formally designated as criminals, making the formal logic of modern statehood fundamentally incompatible with the observable realities of local practice. This disjuncture creates a sphere of social relations that must be kept secret, and this is where Kanjars step in. Relations within this sphere, which is habitually glossed over or written off as “criminal” or “corrupt,” must be kept opaque not only to maintain the reputable standing of state servants, but also to uphold the very idea of the state as distinct from society. The mystification of Kanjars, which propagates popular stereotypes in official paperwork—the claims to their invisibility or magical escapes—is not only a mask of this hidden sphere of relationships; it reinforces the very logic of modern statehood that makes this sphere “criminal” and, in principle, hidden. Much as behind-the-scenes negotiations and thieving contests serve to uphold the public honor of families and the respectability of Rajput families, it is the relationships that stand in the shade of officialdom that maintain the face of the state.

For Kanjars, this secret sphere is not only a space where the enterprising thrive, but also a domain of vulnerability that remains largely invisible to international watchdog organizations, Delhi officials, the press, and “the rule of law.” For the many Kanjars who do not have the good end of arrangements of patronage with the police, this carries tangible consequences. Those not under police tutelage are subjected to persistent predation, “erroneous” convictions, or prolonged incarceration without trial or evidence. Even for well-established informers, police patronage can be fickle, and they often find themselves as vulnerable as their unprotected caste-mates. As I write, my Kanjar host who taught me the secret tongues and the wisdoms, and who is one of the area's most successful informers, has been in prison for more than two months, put there by a new and “unfriendly” head of police station. When Kanjars disappear into jails for months at a time or are murdered by upper caste neighbors, nobody is surprised: they are, after all, master-illusionists, ever disappearing into “the jungle” or their secret lairs. This is how one jailer explained the absence of seven Kanjar inmates from his recordsFootnote ³⁴: “Nobody ever knows where Kanjars are—they are always coming and going [āte-jāte rahate]; sometimes they are here and other times they are not. How can I keep track of them?” While for some Kanjars the rhetoric of secrecy may be a “watering well,” for the community at large, it is a screen that conceals sites of vulnerability, violence, and abuse. The necessary outsider is also an expendable victim and even a required one.