India has strong claims to being the world’s largest democracy. It has genuinely free elections and these have twice turned the once-dominant Congress party out of office. Its press is exceptionally outspoken and its judiciary, especially in the higher courts, is aggressively independent. Moreover, a variety of private and government agencies promote social welfare not only through traditional relief activities but by bringing lawsuits on behalf of the poor.

Nevertheless, if one believes that a democracy should include a system of checks and balances, formal or informal, that would prevent the government from acting lawlessly, India fails. This failure has become apparent in the government’s response to the separatist agitation in the Punjab and Kashmir. By official count, 3,212 civilians—apart from armed militants and the troops fighting them—were killed in 1990 during the government’s campaign to crush these movements. Many believe the totals are much higher and that at least in Kashmir most of the victims died at the hands of the armed forces and the police. In Punjab, the militant Sikhs and government forces opposing them were probably equally responsible for the killing that took place. But the government’s behavior in Kashmir and Punjab is not so much a departure from a general rule of law as a highly dramatic demonstration of how the government fails to exercise restraint in enforcing the law. In a recent investigation of the Indian criminal justice system on behalf of Asia Watch we found that abuses by officials are flagrant and commonplace.

In the major cities anyone unlucky enough to be arrested and jailed faces a far greater likelihood of torture and physical abuse than in many other countries that lack India’s democratic institutions. We expected to find that detainees and prisoners would be badly treated to some degree, if only because life is harsh for many law-abiding Indians. What surprised us, however, was the extent of the brutality and corruption to which most prisoners were subjected.

In India an official commitment to egalitarianism is taken for granted, and the intense and often violent debate over civil rights mainly concerns the extent to which special benefits should be given to untouchables and other “backward castes.” But the criminal justice system itself is explicitly biased in favor of the more prosperous classes. Prisons, for example, give special privileges to virtually all upper-and middle-class inmates, even if they have committed extremely violent crimes. In some respects, moreover, prisoners who are guilty of violent political offenses are treated as an elite group; they have a more varied and ample diet than other prisoners and they can read books and articles that are denied to the others.

What accounts for such extensive police lawlessness in a working democracy? According to Indian law, every person who is arrested and held in custody must be brought before a magistrate within twenty-four hours of arrest. If the magistrate does not issue an order authorizing a prisoner to be held for up to fourteen days, he or she must be set free, and under no circumstances can the police hold a suspect without trial for more than ninety days.

In practice, however, except for the fortunate few who can afford to hire a lawyer and put up bail, prisoners often are kept in jail without trial for weeks or months. In some cases, the police evade the twenty-four-hour rule by not immediately recording an arrest. In many more cases, judges sign orders that prisoners continue to be held in custody without even requiring that they appear in court and without setting any time limits, even after weeks and months have passed.

People arrested are in the greatest physical danger when they are held in local police jails, where they are usually kept in crowded cells with only the most rudimentary sanitation facilities. Torture is commonplace. Experienced lawyers told us that more than half the people arrested in Bombay and an even higher proportion of those arrested in New Delhi were subjected to the “third degree”—a term commonly used in India to describe heavy beatings and other practices that can be described as torture.

One of India’s civil liberties groups, the People’s Union for Democratic Rights (PUDR) in New Delhi, recently published a report on people who died while in police custody: almost all of them had came to its attention through newspaper accounts. (India has a number of local civil liberties organizations but their budgets are tiny, their staffs are all volunteers, and their impact on criminal justice issues is limited.) During the 1980s, five to six people a year, most of them under thirty, were reported to have died in police custody in Delhi. “Most of these people,” the PUDR said,

died due to severe beating and prolonged torture. Practically every person taken to a police station in connection with some or the other offense in our country is subjected to severe beating and torture…. Sticks, boots and belts and wooden rollers are the most common instruments of beating. Sexual abuse, designed not only to hurt but also to humiliate is part of the torture. Naked or seminaked men are a common sight in police lock-ups. It is this process of torture, regular and systematic, whose end product is sometimes death, as in the case of these unfortunate 48 people in Delhi [who died in police lock-ups from 1980 to 1989].

Other Indian civil liberties groups have reported that in 1988, twenty people died in police custody in Andhra Pradesh; fifteen in Bihar; eight in Kerala; twenty-two in Uttar Pradesh; and nineteen in West Bengal. The claims by officials that some of the prisoners had been drunk or ill, or had committed suicide, can’t be supported by medical evidence since it is almost unheard-of for medical representatives of the family or of independent agencies to be allowed to participate in an autopsy. To complicate matters further, the Hindu practice of cremation often makes independent autopsies impossible when doubts are raised about the causes of death. Under these circumstances, the evidence cannot be conclusive, but it appears that each year throughout India at least several hundred people die after being tortured. A deputy inspector-general of police, Shailendra Misra, concluded:


The brutal behavior of our police is established beyond doubt…[in the] police commissions, surveys of public opinion and reports on specific instances of brutality.1

For women the greatest danger of police detention is rape. The National Expert Committee on Women Prisoners, whose chairman was the retired Supreme Court judge Krishna Iyer, has reported that only six institutions throughout the country are exclusively for women, and their inmates are serving long-term sentences; fewer than 20 percent of women in pretrial detention are in a separate prison. Everywhere else women prisoners are in mixed prisons, mainly in separate wings or cells but as the Iyer commission found, always under the authority of male guards.

How many of the guards, we asked, sexually exploit the women, most of whom are migrants from rural areas and lack the protection that family or community connections can provide? The answers are particularly difficult to obtain, for in India, to an even greater extent than in Western countries, victims risk punishment or ostracism if it becomes known that they have been raped. If they are married, they are likely to be abandoned by their husbands and families, and if they are single, their chances of marriage are virtually lost. Thus rape is seldom reported, and when it happens to women in police custody, the silence is all the greater. Even if the victim were able to accuse her assailants, it is almost certain that she would suffer more than they would. The only clues to what happens come from incidents uncovered more or less by chance. The Delhi police themselves recently acknowledged that between 1988 and 1990 fourteen rapes took place of women in custody, and twenty-four police officers were named as responsible. Suspensions and dismissals have followed, but, so far as we could find, none of the twenty-four officers has as yet been convicted of a crime.

No doubt some police mistreat the people in their custody in order to force them to confess and plead guilty, thereby making their work easier and improving their reputation for efficiency. But these motives should not be given much weight. According to Indian law, confessions made before a judge are inadmissible at trial; in that respect, Indian defendants are better protected than they would be in the United States, where the Miranda rules only entitle suspects to be warned about their rights and not to be questioned without a lawyer present if they so choose. The Indian rule is less protective than it seems, because evidence obtained as a result of a confession to the police, such as stolen property, can be used in court. Still, if the crime does not involve property, the value of a confession diminishes.

Moreover, if the police simply wanted a conviction they would have little reason to beat up petty criminals, since in many cases a simple statement by the arresting officer before a magistrate is enough to obtain a guilty verdict. People familiar with police lockups told us, for example, that typical victims of police abuse were poor rural migrants who took trains to the city without paying the fare. In most cases, the person arrested has no ticket and there is no need to coerce a confession. The police are also frequently called to deal with cases of assault involving people who know each other; here too, beatings to obtain a confession are unnecessary—the victim and neighbors would testify to what happened—but they occur anyway.


Police torture in India has a long history. In 1854, when the British governor of Madras established a commission “for the Investigation of Alleged Cases of Torture,” it received almost two thousand complaints. The commission concluded that torture was a standard practice and gave gruesome details:

Among the principal tortures in vogue in police cases we find the following: twisting a rope tightly around the entire arm or leg so as to impede circulation; lifting up by the moustache; suspending by arms while tied behind the back; searing with hot irons; placing scratching insects, such as the carpenter beetle, on the navel, scrotum and other sensitive parts; dipping in wells and rivers, till the party is half suffocated; squeezing the testicles; beating with sticks; prevention of sleep; nipping the flesh with pincers; putting pepper or red chilies in the eyes or introducing them into the private parts of men or women; these cruelties occasionally persevered until death sooner or later ensues.2

In the 136 years since this report was issued many other commissions—some national, others sponsored by state governments—have made similar findings. The report of Inspector Misra, who served as director of a national police commission between 1977 and 1980, concluded that “the current methodology of third degree does not show any remarkable refinement over the methods described by the Torture Commission of 1854.3

Police brutality continues to flourish, we found, largely because of widespread corruption. Policemen have low status in India and are poorly paid; as with many other Indian officials they feel driven to supplement their incomes. The detainees themselves, or their families, are threatened with torture if they do not bribe the police—a threat that can work only if those who do not pay, or cannot pay, are in fact tortured. As Misra and the Police Commission observed:

With a policeman the reputation for expert brutality fetches more money than actual brutality, but of course the reputation has to have a solid foundation. Once a policeman acquires a reputation for expert third degree he makes enormous quantities of money in daily crime work by simply with-holding his customary brutality.4

In such a system people with money or connections can usually find a way to avoid being tortured, while virtually the only people subjected to severe police brutality are from the lowest classes and castes. Since they have little political influence, the issue of police brutality is mostly ignored in Indian political life.

At the same time, we were frequently told, middle-class people, who almost always are spared brutal punishment, tend to believe that it is the most effective method for deterring crime. A report by a West Bengal civil liberties group on people killed in custody found that “to everyone it appears quite natural that the police should beat up any arrested person.” As in other countries where police brutality takes place as a matter of course—Brazil, with its police death squads, is an extreme example—the wide acceptance of torture reflects a lack of faith in the legal criminal justice system. Illegal methods flourish partly because the legal ones seem inadequate.

In fact, India does not appear to suffer from soaring crime rates. No one places much confidence in the government’s statistics on crime, but academic experts who analyze the crime rates are convinced that the streets in Delhi or Calcutta are relatively safe, and most visitors would agree. In India, as elsewhere, probably the most reliable crime statistics are for murder, one of the few crimes that are universally reported. In 1986, the most recent year for which we could obtain data, there were 27,269 murders and another 4,195 criminal homicides, a combined total lower than that for the United States, although India’s population is three and a half times larger.

But whatever the rate of crime, and whatever the findings of commissions and civil liberties organizations, there are no administrative or legislative groups organized to check or discourage police brutality. In a society teeming with slums and desperate poverty, many people believe that the only sure way to maintain order is to have the police act first and ask questions later. That the police are notoriously corrupt itself contributes to the public’s willingness to tolerate police brutality. After all, some people say, if the culprits had not been beaten, it would have meant that the police had accepted their bribes.

Tradition and custom shape life in prison as well. More than forty years after the end of British rule, the prisons are still technically governed by the Prisons Act of 1894. Since prison officials are anxious to avoid being held accountable, however, the various state prison manuals based on the 1894 law are collectors’ items, not only in short supply but expensive as well. (We bought a copy of the Punjab Manual in a Delhi bookstore for twenty dollars, and were repeatedly told how lucky we were to find one.) A number of commissions have attempted to update and revise the code, but except in a few states, Indian state parliaments have not enacted new legislation. The mostly unread authoritative text remains the 1894 act.

Not only the regulations but the day-to-day reality of India’s prisons seem archaic. In fact, the use of prisons, which in the West grew swiftly with modernization, has not become widespread in India; imprisonment is probably not more extensive now than it was under British rule. The most thorough investigation of the prisons, the All India Committee on Jail Reform (whose chairman was the retired Supreme Court justice Anand Mulla) found that there were 1,220 prison facilities in the country as of December 31, 1980, of which 822 (67 percent) were police jails, while almost all of the others were state prisons; together they held some 160,000 inmates; more recent estimates put the figure at 184,000. Well over half of these inmates were awaiting trial. In view of the inaccuracy of the records in some Indian states, these figures may be too low; but if there were as many as 250,000 prisoners in a population of 850 million, the incarceration rate would rate be 29 per hundred thousand. In the United States, the incarceration rate per hundred thousand is 426, in the Netherlands, 40.

Such low rates would be cause for satisfaction in India if they did not also reflect the extent of police brutality. When the police torture unconvicted offenders and then let them go or even kill them, imprisonment becomes superfluous. If the number of inmates is low, it is partly because many people are not tried but are punished by brutal treatment.

Often we could find no clear reason why one person was beaten or tortured and sent home while another was beaten and kept in prison. Petty offenders make up about half of the prisoners serving time; the other half have committed felonies such as robbery and assault, or worse. About a third of the convicted inmates serve less than a year; 23 percent serve between one and ten, and the rest (44 percent) ten or longer.

The All India Committee observed that many of the inmates came from the “underprivileged sections of society,” noting that “persons who have means and influence generally manage to remain beyond the reach of the law even if they are involved in violation of the law” (p. 21). The degree to which low-caste Indians are imprisoned emerges from a study of the prisons of the state of Orissa. Situated on the northeast coast of India, Orissa’s population of more than twenty million has roughly the same proportion of urban and rural residents as in the entire country. The tribes and castes that are “scheduled,” i.e., have been declared legally to be victims of discrimination and in need of protection, accounted for almost two thirds of the prisoners—about four times their proportion in Indian society.5 By comparison, black men account for 43 percent of the inmates in prisons and jails in the United States, also about four times their proportion in the population.

But there the comparison ends. India still follows the standards of the 1894 act by categorizing prisoners according to their social class. (A similar system prevails in other countries of the Indian subcontinent, among them Pakistan, which share the British colonial legacy.) Inmates classified as A or B are persons who

by social status, education and habit of life have been accustomed to a superior mode of living. Habitual prisoners may be included in this class by order of the Inspector General of Prisons.

Category C consists of “prisoners who are not classified in class A and B” (Punjab Manual, p. 196). People of higher classes and castes—those with property or from socially superior families or who have education—thus are set apart from the poor, the educated, the low caste. It is not what you have done (which in the United States determines whether you are sent to minimum, medium, or maximum security prison) or how you behave in prison, but who you are that counts.

The classifications make a huge difference to a prisoner’s life. A recent state investigation of the prisons of Tamil Nadu (the relatively progressive province that includes Madras) reported, matter-of-factly, that the daily allotment for feeding class A and B prisoners ranged from fourteen to seventeen rupees (around one dollar), while for C prisoners it was seven to eight rupees. The A and B prisoners were allowed to buy fruit and other food of “good nutritive value”; the others were not. The privileged prisoners were allowed to write and receive one letter a week, the others, one letter every two weeks. A and B prisoners could receive any newspapers, the others only those on a prescribed list. In still other jurisdictions, the A prisoners are exempted from menial labor and from such restraints as handcuffs and irons, all of which are standard for C-class prisoners.

The classification system increases the authority—and the corruption—of prison administrators. The official regulations provide that state officials classify the inmates, but in fact the prison administrators do so. The power to decide each prisoner’s grade and privileges increases their control over the prison population and their ability to exchange favors for money. Inmates with money are usually prepared to pay the going rate for A or B status, and although hard evidence of such payoffs is not easy to find, every student of Indian prisons we know of, whether inside or outside the system, agrees that such corruption is rampant. At the same time long-term convicts—that is, hardened criminals—are formally assigned most of the duties ordinarily performed by American prison guards. They are in a position to hand out or deny favors and to extract payment for them. Even if there are no payoffs, a prisoner’s wealth and status will usually decide his classification. Relatively well-off prisoners can also hire lawyers who will devote themselves to seeing to it that their clients are accorded a privileged status.

Almost every aspect of a prisoner’s life, we found, is determined by his ability to pay. In the Tihar jail in New Delhi, which holds between five thousand and eight thousand prisoners, more than 90 percent of prisoners are in class C. We were not allowed to enter the jail, despite repeated requests to do so, but reliable sources told us that most of the prisoners live in barracks, sleep on the floor, work in the prison factory for a paltry wage that is generally not paid, and are fed two small loaves of bread twice a day, along with some lentils and very small portions of vegetables. Inmates who have some money will buy—usually from guards, sometimes from inmates—a bed for the night or week, or a better work assignment. Since the diet is hardly adequate for survival, they buy food at the prison canteen or bribe the kitchen staff. If they have no money, they will try to win favors from better-off inmates by running errands for them, washing their clothes, or performing sexual services. The system, as a former prisoner put it, also “makes the doctor a king.” He can, for a price, recommend a special diet—milk, eggs, meat—or get a prisoner some time off by prescribing him a few days in the prison hospital. It is never much of a hospital—there is hardly any medical treatment in the prisons—but it is better than standard C conditions.

Conditions for most women prisoners are almost equally harsh. An investigating committee headed by Justice Krishna Iyer of the National Supreme Court found that the six special institutions for women were “better than other custodial centres although hardly adequate.” In the mixed institutions, “clothing, work, education, or even medical examination were generally not available,” and there were “no beds, bedsheets or pillows, just a cane mat.” “Human rights cannot survive in such jails,” the committee concluded, and this statement applies to virtually all the penal institutions in India.

Virtually the only groups that have protested the abuses and corruption of the criminal justice system are civil liberties organizations, but they have not had much effect. They have conducted investigations of the deaths of the people who were in custody, and the lawyers who volunteer their services are dedicated; but they have no national organization, no professional staff members, and depend entirely on volunteers who do civil rights work at home or at their regular business offices. As one student of the human rights movement in India recently observed, these organizations tend to become active when there is a crisis—for example a series of deaths in a local prison—and then “they fade away, even die out, only to resurface, often in a new form at the time of another crisis.”6 It is almost impossible for them to conduct the sustained campaigns that would be required to expose and stop such practices as police torture or to reform the prison system.

By contrast, virtually every country in Latin America and the Caribbean has one or more human rights organization with its own offices and professional staff. (The exceptions are Cuba, where the Castro regime has made this impossible, and some of the smaller Caribbean nations.) That India should lag behind these countries might seem surprising because efforts to defend civil liberties there go back to 1936, when Jawaharlal Nehru established the national Indian Civil Liberties Union, which was active until independence was declared in 1947. (Rabindranath Tagore was its honorary president and Krishna B. Menon was its first secretary.) After Nehru became prime minister he wrote a letter to his fellow members recommending that the ICLU be dissolved, apparently believing that his government could be counted upon to protect civil liberties. The ICLU soon ceased to exist.

A new civil liberties movement was organized in India in 1975 when Nehru’s daughter, Prime Minister Indira Gandhi, imposed a state of emergency and jailed several thousand political opponents. To resist her repressive policies, a long-time associate of Nehru and Gandhi, Jayaprakash Narayan, founded the People’s Union for Civil Liberties and Democratic Rights. The new group became less active in 1977 when Mrs. Gandhi was defeated at the polls, and many of its members went into the Janata government. By 1980, the movement split into two groups: the People’s Union for Civil Liberties, which concentrates on protecting existing constitutional rights, and the People’s Union for Democratic Rights, which seeks broader reforms of Indian society. The two groups have collaborated at times, particularly in investigating and denouncing the failure of the police to protect the Sikhs in 1984 when some three thousand of them were murdered in Delhi after Mrs. Gandhi’s assassination.

The Indian civil liberties movement has been mainly concerned with the need to defend peaceful dissenters from repression; but for most of the period since independence, peaceful dissenters have been able to make their views known, while much of the current opposition in Punjab and Kashmir has condoned violence, and such anarchist-minded groups as the Naxalites are committed to the violent overthrow of the government.

In India, moreover, the representatives of civil liberties groups that we met, unlike their counterparts elsewhere in the world, were invariably hostile to the idea of obtaining support from foreign donors such as the Ford Foundation, the Dutch and German Foundations, the EC, or church groups such as Oxfam. While the civil rights leaders badly need the funds such organizations could be willing to contribute, they say they fear that outsiders would impose their own concerns, and the Indian organizations would be seen as manipulated by foreign interests. The outright rejection, really disdain, for foreign assistance that we encountered in India has no counterpart in any other country we know; it seems to reflect a more general antagonism to anything that smacks of Western involvement in India’s social affairs. Indian leaders frequently recalled to us the case of the Asia Foundation, which provided financial support for many Indian institutions and claimed to be wholly independent of the US government but was revealed in the late 1960s to be a conduit for CIA funds. That some Indian organizations were shown to have been duped at the time makes other foreign agencies suspect now.

The prospects for social reform in India appear dim, and there is no reason for optimism about reforming the system of criminal justice. Since no powerful political constituencies are much concerned about police activities or prison corruption, and since the middle and upper classes have no personal stake in change and many people are persuaded that the current system works reasonably well, the Indian human rights movement seems the only force that could bring about reforms. When some of the movement’s leaders asked us, as outsiders, for our views on what might be done, we observed that because the prison system is relatively manageable in size, change could be accomplished without spending large amounts of money. Were the case otherwise, the prospect of reform would be unrelievedly bleak. Indeed, significant improvements could be accomplished through relatively inexpensive administrative reforms that are entirely consistent with the prevailing ideology (although not the practice) of the Indian government and the leading political parties.

Clearly the prisoner classification system should be eliminated, not only for reasons of fairness but to make the condition of the prisons a more urgent concern of people from the middle classes. So long as the “haves” can buy their way out, the “have nots” will be left to suffer the system’s worst abuses. It could be argued that abolishing the A and B classes will only serve to make the standards of class C life more pervasive; the leveling could be downward, not upward. Nevertheless, the risk should be taken, for there seems no other way to broaden awareness of police and prison practices.

Second, while corruption is never easy to weed out, it is so extensive among police and prison officials that even a limited effort to bring some honesty and accountability to the system would be telling. A few firings and prosecutions might work wonders.

Third, to curtail the abuses in police detention, the practice of turning over arrested people to police custody after they are arraigned before magistrates should be eliminated. The practice amounts to a judicial invitation to the police to use torture and, in the case of women, engage in rape. It would be difficult to enforce the rule that an arrested person be arraigned within twenty-four hours, since the police are the ones who count the time; but once, offenders are brought before a judge, the courts could release many more of them on bail or personal recognizance. If they must be held in custody to ensure they won’t run away, they should be placed in facilities under court supervision, not in police lock-ups.

Moreover, the prosecution of police who abuse detainees might deter others. It would cost little to require that independent doctors, not police doctors, conduct post-mortem examinations. And were families and reform organizations given access to these findings, still another deterrent to abuse would be provided.

Finally, ways should be found to strengthen the civil liberties and human rights organizations in India. In view of their fears of foreign funding and foreign interference, suggestions from outsiders will automatically be suspect; but if the Indian organizations themselves devise new strategies for promoting human rights, they will find their counterparts in the United States and Europe eager to help them.

This Issue

May 16, 1991