Dying for the Dead: Sati in Universal Context

By: JÖRG FISCH (Universität Zürich)

On 4 September 1987, in the Indian village of Deorala (Rajasthan), eighteen-year-old Roop Kanwar died with the body of her husband in the flames of his funeral pyre. The news spread around the world, and everyone was or pretended to be shocked.1Sati is the Indian custom of burning widows, or of widows burning themselves (the expression, originally denoting a faithful wife, says in its modern use nothing about the agent and thus nothing about the voluntary or involuntary character of the act) on the funeral pyres of their husbands’ bodies. It is probably the best-known among those customs the Europeans found overseas that they thought shocking to the utmost and therefore suppressed or tried to suppress in due course. The British prohibited sati in their own possessions in 1829 and had it prohibited by the Indian princes by 1862, but it lingered on. India tightened the law immediately after the death of Roop Kanwar. Nevertheless, there have been occasional new instances since then,2 and there probably still are a certain number of undetected cases every year. Accordingly, there has been a great deal of research,3 a huge number of accounts by eyewitnesses and others, and a considerable production of fiction and even poetry on sati, from Goethe and Southey to Jules Verne and M. M. Kaye, to mention but a few names.41
      In scholarly works it is frequently said that sati was not an isolated custom but that similar rituals were to be found all over the world in the course of history. Yet this assertion is never supported by references to serious research or to primary sources but only to similarly vague passages in the literature, or it remains completely uncorroborated. It is the purpose of this article to check these claims. To what extent can or could sati be found outside India? What similar customs have there been in the course of history? This necessitates a definition of the subject in broader terms. Then a short comparative survey of the incidence of the defined custom is given. On this basis, more systematic questions can be asked and, to some extent, answered: What were the origins and the functions of such customs? Were they sanctioned by religion? What were the social structures underlying them, and did they in turn influence those structures? The main thesis is that such customs link in a unique manner religious beliefs in a hereafter with power struggles in a society, either between classes or other social groups or between the sexes. Each aspect is irreducible to the other. Both idealized otherworldly and crudely materialist and political aspects are indissolubly bound to each other. Any attempt to suppress such customs was bound to have religious as well as social and political consequences—and this remains true even in a postcolonial world.2
      A cautionary word about the sources is necessary. A modern state bureaucracy usually does not administer customs of the kind investigated here, but it prohibits and suppresses them. Therefore, the bulk of the sources are not administrative but ethnographical and often of questionable authenticity.5 Nevertheless, they frequently give a fairly detailed and reliable picture. There is only one important exception. While all other colonial powers and the British themselves in all other places suppressed such customs as soon as they had the power to do so, the British in India at first tried to administratively regulate sati by distinguishing between legal and illegal cases according to Hindu law. For obvious reasons (mainly because British officials were not prepared to accept the notion of a widow lawfully burning herself to death in public in what was defined as a voluntary act), this attempt misfired, and eventually even the British in India tried to suppress the custom. But the aborted attempt at regulation produced an enormous account of sources. There is nothing comparable for other parts of the world. Moreover, of all the areas in which such customs could be found, India was the most easily accessible for Europeans who, up to the late eighteenth century, had no power to interfere. Therefore, we have more eyewitness accounts for India than for any other part of the world because such accounts are usually written only by outsiders. There seems to be, for example, not a single precolonial description of a widow burning in India by an Indian or in Bali by a Balinese.3
      There is also archaeological evidence. But its interpretation poses extraordinary problems that cannot be dealt with here. Suffice it to state that only in a very few cases it is possible to prove with reasonable certainty from archaeological evidence that the death of a person has occurred in connection with the custom investigated here.4
  
Definitions: Following into Death 
Sati in the manner in which it is known in India is not a universal custom. But, as will be shown, a slightly different form occurred on Bali, while more diverging patterns can be located in China and on many islands in the South Pacific. The scope of investigation should thus be broadened by looking at similar but not identical customs. This leads to the question of the basis of comparison. I propose to compare not the outward appearances but the basic content, the meaning and the purpose of a custom. In the case of sati this means that the manner of dying is irrelevant. Many different customs are connected with the (mainly involuntary) burning of human beings, such as burning of heretics, witches,6 prisoners of war, or, a “burning” problem in contemporary India, dowry murders, when wives are burned to death because their families do not fulfill the dowry expectations or demands of the husband’s family.7 If, instead, the content and the intention of sati are considered, this leads to the following definition: on the occasion of the death of a person, one or several persons follow that person into death, voluntarily or involuntarily, in a public ritual. The essential point is not the manner of dying (or killing), but the occasion of the death of the follower and, even more, its public character. The latter means that state and society tolerate and accept—and in many cases even enjoin or at least encourage—the custom. No state or political community can pretend that it does not, at least implicitly, accept an act regularly watched by a peaceful multitude with nobody interfering to prevent it, which here would mean to save the victim from herself or from those who are going to kill her. The public performance also requires a ritualistic character, as it has to be arranged for a certain number of spectators or at least witnesses. If state and society disapprove of the custom, it will be considered murder, manslaughter, abetment to suicide, or suicide, and those involved will be prosecuted.5
      The case is more difficult when the attitudes of state and society do not coincide. If the state approves of the custom while society rejects it, it probably will not happen so long as the state does not try to enforce it. If, on the other hand, society, or important sectors of it, encourage or even demand the custom while the state opposes or prohibits it, there will be a conflict; the state will be resisted. This has been the case in India ever since the British prohibition in 1829.6
      The requirement of publicity excludes many cases that in the situation of those immediately involved often do not differ significantly from public ceremonies, especially suicides at the death of a beloved person, particularly well known in literature, from Romeo and Juliet to Tristan and Isolde. The social meaning of such occurrences is quite different from public ceremonies. Moreover, they are not and cannot be ritualized because they lack continuity, for which assistance of the public or at least of representatives of the public is required.7
      There is no technical term in English for the phenomenon defined here. In German I have called it Totenfolge (following the dead). In English I shall usually speak of “following into death” and occasionally of “following in death.” Sometimes, other paraphrases are used.8
  
The Universal Occurrence of a Custom 
As a result of the scarcity of sources—which, moreover, are unevenly distributed—it is impossible to write a history of following into death. Only scattered indications can be given. Usually we do not know where and how such a custom developed and whether it was introduced in a society independently or as a consequence either of contacts with other societies or of more diffuse cultural influences. The foundation for comparisons is thus very weak. Apart from India, which is a special case owing to the quality and quantity of its sources, following into death has been established with reasonable certainty for various periods of the history of many areas of the world, notably for ancient China since the fourth millennium B.C.E. up to the twentieth century, for ancient Egypt from the third millennium B.C.E. to the sixth century C.E., and for the ancient Orient especially in the third millennium C.E. For Japan we have sources referring to cases between the third and the eighteenth centuries C.E. For most other areas we are restricted to European sources, referring notably to Africa, with a clear emphasis on West Africa, Southeast and Central Asia, the Pacific Islands, Eastern Europe, and the Americas (usually in a rather summary manner). It is impossible, and probably will always remain impossible, to present a map indicating the occurrence of following into death in space and time. We are restricted to knowing that in a particular area such customs were found or at least reported. This is the reason why the following remarks are sketchy rather than systematic and why the conclusions have to remain tentative. Any other presentation would be presumptuous.9
      Despite these precautionary remarks, it can be said that following into death is both a universal and a very exceptional custom. It can be traced in most parts of the world, but it is always quite rare. The closest similarity to the Indian custom of widow burning8 was to be found in Bali and, up to some time between the fifteenth and the seventeenth centuries, in Java. On both islands there was pervasive Hindu influence. Although there are no sources proving the Indian origin of the custom in Southeast Asia, there is a strong presumption in favor of it. Widow burning disappeared in the course of the spread of Islam. It seems not to have survived on Java after ca. 1700, when there were hardly any Hindus left, while Bali remained Hindu. The difference was that in Bali and Java sati was restricted, if not de jure then at least de facto, to the wives and concubines of rulers and members of the aristocracy, while in India in principle all Hindus, including casteless women, were allowed sati (it occurred even among Muslims converted from Hindusim). On Java and Bali, the women were either burned on the funeral pyres of their husbands or stabbed to death.9 This seems to have reflected an earlier stage of the custom, which only in India was developed further.10
      There is more evidence for a spread from India. The earliest written source for sati is a Greek account of the burning of a widow with the body of an Indian military leader in the army of one of Alexander the Great’s successors.10 The killing of widows is also reported from many Islands in the South Pacific, for example, Fiji, the Solomon Islands, and the New Hebrides (now Vanuatu). Here, strangling was the usual method. As a rule, it was done by relatives.1111
      While so far the decisive criterion was the distinction between the sexes, that women (usually wives) follow men but men do not follow women, there was a very different form of the custom to be found in many places, in its most elaborate form in some parts of Africa, especially in the west of the continent. Here, rulers, and in many states and chiefdoms, members of the ruling group in general, were granted (or managed to usurp) the privilege of being accompanied into death by some of the people among whom they were living. The followers were in the first place members of the court, from ministers to servants and slaves and to wives and concubines. Sometimes there were additional, anonymous categories: prisoners of war, slaves, criminals,12 and even uninvolved bystanders rounded up in the streets. The followers were both men and women, usually with a majority of men, while those followed were mostly men, but they could be women as well. According to John Roscoe, at the funeral of a ruler of Baganda, the following men were put to death: the chief cook, the chief brewer, the chief over the herdsmen, and he who had the care of the king’s well; also the following women: the cook and they who had charge of the beer, the bedchamber, the water, the king’s clothing, the milk pot and the milk, together with “hundreds more.” They were “sent to attend the king, who was supposed to need them in the other world.”1312
      Such ceremonies were not limited to the occasion of the death or rather the funeral of the person to be followed. Especially for rulers there were, in some states, regular, usually yearly, remembrance ceremonies with more victims.14 Moreover, in some places messengers were from time to time dispatched. Having been beheaded, they had to transmit news to the previous ruler.15 The sources—which are almost exclusively European—frequently speak of thousands of victims or followers. Such reports are certainly exaggerated. But there can be little doubt that in some states, especially in West Africa, the numbers were considerable, with dozens and sometimes perhaps even hundreds of victims in one ceremony.1613
      This kind of custom, which is not built upon the differentiation between the sexes but on the social and political status of those involved, can also be found elsewhere, although it is usually less well documented. It seems to have played an important part in the kingdoms and empires of Central and South America, where great lords sometimes had to offer some of their servants and followers to accompany the ruler into the hereafter.17 Basically similar customs, but on a smaller scale, corresponding to the smaller political units, are reported for many places of the Americas, when chiefs and other important persons sometimes were followed by women and slaves.14
      There are also some scattered notices about such customs in Central Asia. Moreover, archaeological evidence points in the same direction in several places all over the world. Here it is important to remark that archaeology can prove such a custom with a reasonable probability only in spectacular cases. If, for example, only two spouses are buried together, we can never know with certainty how and when exactly they died. If, however, a tomb groups a great number of bodies around an obviously dominating figure and simultaneous burial is certain, then there is some likelihood of the minor figures having followed the major one into death. Such evidence is available for some periods of Egyptian history and of the ancient Near East, especially in Sumer.18 Similar discoveries have been made for the Scyths in several areas of Eastern Europe and Central Asia.15
      Special cases are presented by China and Japan. In both countries, there is written as well as archaeological evidence of members of the court following deceased rulers into their tombs, sometimes in large numbers. In China, archaeological evidence reaches as far back as the fourth millennium B.C.E., while the earliest case on record dates from 678 B.C.E.19 In both China and Japan there was early opposition to the custom, especially by Confucianism. This eventually led to suppression. But the custom was revived in both countries in a different style. In China, women committed suicide after the death of their husbands, sometimes many years afterward. The state was opposed to the custom in principle, while in practice it countenanced and often even honored such acts of connubial virtue right into the twentieth century20 by placing memorial tablets in temples and even erecting honorary gates.21 In Japan, the custom revived much later, especially in the Tokugawa period, since the sixteenth century, in the shape of samurai committing suicide on the occasion of the death of their master, a daimyo. Here, the state was able to suppress the custom almost completely by the late eighteenth century, despite its being deeply rooted in and supported and glorified by society.16
      Another form of the custom was probably to be found in several continents, but there is reliable evidence only for Southeast Asia.22 Here, on many islands and in some mainland areas headhunting had the function of following into death: the persons killed had, in the view of their killers, to become servants in the hereafter either of one of the killers or of a person on the occasion of whose death the headhunting had been organized. While headhunting in itself originally has no connection with following into death, it was later interpreted and justified in that sense. In some societies, headhunting could be replaced by the killing of slaves. This meant the transition to the indiscriminate killing of followers or of ordinary people on the occasion of the death of a ruler or another important person.17
      There is only one area for which it can be confirmed with reasonable certainty that there was no following into death over an extended period. While there are some indications that among Eastern European peoples such customs could be found up to the late Middle Ages, there is neither written nor archaeological evidence for other parts of Europe, despite the enormous number of sources we have since the time of ancient Greece. There is as yet no satisfactory explanation for this phenomenon. What so far seems certain is that Christianity is not the main reason, as following into death cannot be proved for pre- Christian times, and Christianity, although it never countenanced the custom, did not reject or combat it right up to the late Middle Ages.18
  
Statistics 
The information given so far has deliberately been kept very vague with respect to quantitative questions. There is and there never will be any way of knowing the frequency of such customs in particular places and eras. The reason is that modern bureaucracies, which would be expected to leave reliable information, do not register such customs but suppress them. It can, however, be asserted with a high degree of probability that there were relatively few cases. If only the rulers were followed into death, then the ceremonies occurred but once every so many years. If the killing or the suicide of widows had been prevalent at one time in a certain society, this society would soon have suffered a decline of its population or even died out, and this would in all likelihood have been noted in some sources. There are complaints by missionaries in the South Pacific during the nineteenth century that the population suffered a decline because of the frequent strangling of widows.23 But the assertions—taken up by anthropologists in the early twentieth century24—are not corroborated by any numbers and seem to be rather exaggerated. The missionaries were, of course, strongly biased against the custom and considered it a crime.19
      The only exception merely confirms the rule. The British hesitated to suppress sati in India out of fear of violent Indian reactions, and instead tried to administer it bureaucratically. The attempt failed and led, in 1829, to the unconditional prohibition of the rite.25 But in the meantime, it had produced fairly reliable statistics at least in one part of India, the Bengal Presidency, for the years 1815–1828. The result was that on average there were 581 cases a year out of a population of about fifty million. This meant that perhaps one widow in 430 died on the funeral pyre of her husband. In 1818, the year with the highest number recorded, on the same basis of calculation, one widow in 298 died, while in single districts with particularly high incidence it could be between one in forty and one in seventy. The highest relative number recorded was in 1819 for the suburbs of Calcutta. Here, in a population of 360,360 there were fifty-one cases of sati, or about one widow out of thirty-six.26 But these were clearly exceptional cases. For the whole of British India at this time, a reasonable estimate would suggest that about one in one thousand widows became a sati.20
      It is impossible to say whether widow burning was increasing or decreasing between the British conquest of Bengal and the abolition of the custom in 1829—there is simply not enough data. It is even less possible to indicate trends for other periods.2721
      Sati was to be found in all castes and in all social strata, even among outcastes and among recently converted Muslims, although it was comparatively more frequent among the upper castes and wealthy families. The custom was most recurrent among widows between the ages of twenty-one and sixty, but it could be found even among young girls and very old women. Some had many children, while others were childless; on the average, a sati had 2.3 surviving children.22
      The importance of following into death does not lie in the quantitative field. It probably has never had any significant impact on a considerable population. Much less has it had—objectively or intentionally—the function of population control. There is no reason to believe that it was more frequent in any place outside India than in India itself, and it probably was much less habitual in most other areas of the world. An indication for this is the fact that the suppression of the custom was nowhere so difficult as in India. The circumstances suggest one possible exception. So long as the custom of following the dead occurs within a society, it is in that society’s interest to not let the custom grow out of hand, to limit the frequency of the ceremonies and the number of followers in each case. Otherwise there might be serious social unrest, which in Africa indeed occasionally took place. It is different if the victims are taken from other, hostile societies, as is the case with headhunting. Theoretically it is imaginable that two or more societies might virtually root each other out by headhunting—but there is no evidence that this ever really happened.23
      The significance of following into death is thus not in the numbers, but in the individual cases and their symbolical meaning.24
  
Preconditions: Belief in a Hereafter and Social Inequality 
Why are there such customs? Why do people want to be accompanied into death, why do others want third persons to accompany dead individuals, and why do people themselves sometimes even want to accompany the dead? In this section I shall not search for actual causes but only for conditions that make the custom possible and plausible, without really explaining it. It is about necessary but not sufficient conditions; the actual explanation has to be sought mainly on the level of the individual cases. Otherwise there is a risk of overdetermining a custom with structural causes. Sati, for example, has often been explained by referring to the undeniably bad situation of women and especially of widows in India. But the situation of women was not better in many other parts of the world that did not have sati, and even in India most widows, among whom there were certainly many whose situation could hardly be worse, nevertheless did not become satis.25
      The results both of empirical research and systematic reflection suggest two necessary conditions for following into death. First, most sources either emphasize or take for granted a specific belief in a future life. Life in the hereafter is thought to be a continuation or a duplication of this life, with the individual having the same position there as in this world. But life in a hereafter can become a continuation of this life only to the extent to which the whole environment of the deceased person, including human beings, is transferred into the other world. If somebody wants to find his full surroundings there, he has to take them with him, objects, animals, and humans.26
      If the belief in a future life along these lines fades away, following into death becomes, in the eyes of a society, either pointless (if it is done voluntarily and suicide is tolerated), or criminal (if it is done involuntarily or if suicide is not allowed). In contrast or opposition to the view described, there may be either a complete lack of belief in a future life, or the belief that the position of the individual in the hereafter is determined by a final judgment in accordance with his behavior in this world. In such a case it makes little sense to take followers into the hereafter, as they may end up in a different place there and the individual will not be awarded followers according to his position in this world but according to his merits in the final judgment—if there are followers at all.27
      There is, however, a very important exception to this plausible rule. In India, there seems to have been at first a belief in a future life replicating life in this world. But it was gradually transformed into the view of a future retributive justice, in connection with the theories of karma.28 Once the future status of a person depended on his behavior, and if moreover the hereafter was replaced, under normal conditions, by a cycle of rebirths in this world, following into death, whether as sati or otherwise, would no longer have made sense. But sati did not disappear. Instead its justification was changed in accordance with the new beliefs. It was declared to be such a meritorious act that it could undo the consequences of the new belief. The widow who became a sati could, by her act, liberate her husband (it was implicitly presupposed that she did want to do this in any case), even if he had committed the worst sins, and thus be reunited with him. In this way the custom of sati became resistant against the abandonment of traditional beliefs in a hereafter that were unconnected with moral views, and it became more deeply rooted in society than any similar custom in the world. This is probably the main reason why its suppression has proved so difficult.28
      This condition of a specific belief in an afterlife need not be shared by every individual in the respective society, least of all by those who are compelled to follow a dead person. But it must be the prevalent or rather the dominant or hegemonic belief in that society. Otherwise society will no longer support such rituals, which means that they no longer can be performed in public, being considered crimes. But then, even if people are still killed secretly, this is no longer following into death in the sense defined. In an ideally working system both followers and those to be followed share the belief described. The more the followers themselves are convinced that it is their duty to follow, so as to preserve their society, the less the system has to rely on compulsion in order to enforce and stabilize the custom.29
      Second, the belief in, or rather the idea of, a hereafter replicating this life is self-destructive. If everybody attempts to take his full surroundings with him, mankind will soon die out. The idea can at best be realized selectively by limiting the number of followers. As it is a matter of life and death, it is not to be supposed that this selection will be left to chance. There must be a decision about who is to be followed and who is to follow or at least about who is, in principle, entitled to followers and who has the basic duty to follow, even though in practice he or she may not actually be required to follow somebody. The foundation for such momentous decisions will be social inequality. Only if some people are thought and declared to have more rights than others and if such a view is enforced, not necessarily by open violence but perhaps by a certain ideology, is a system of following in death compatible with the survival and the functioning of a society. Otherwise this society runs the risk of ending up in anarchy, in a bellum omnium contra omnes. There are different types of inequality that can serve as a foundation for a system of following into death. Before they are dealt with, however, it is necessary to remember the fact that even the combination of an appropriate belief in a hereafter with outspoken social inequality does not necessarily lead to following into death. There have been many societies fulfilling both conditions that nevertheless have not known such customs.30
  
Institutional and Individual Following into Death 
Different forms of social inequality result in different forms of following the dead. It is useful to distinguish between two basic types. They are ideal types, not descriptions of reality. Real occurrences of such customs usually will combine elements of both types. These have already taken shape in the description of the various customs all over the world. The most typical representations are sati on the one hand and the ruler who is followed by his court and a number of commoners on the other.31
      The first type of social inequality is based on originally natural distinctions such as race, kinship, age, or, in this case, most frequently gender. One group has to provide the followers, while the other is entitled to be followed. This type, which I call “individual following into death,” usually is built upon the distinction between the sexes: women have to follow men; husbands are followed by their wives or widows. The theoretically conceivable case of a reversal with men following women seems never to have been realized so far. The only partial exception on record is from Asante in West Africa, in the nineteenth century. If a man who did not belong to the aristocracy married a sister of the ruler, he was obliged to accompany her into death, whereas she had no corresponding duty.29 In this case, social standing, the fact of belonging to the ruling group, obviously was more important than the fact of being a woman. The sources do not say whether and to what extent the rule was kept.32
      The second type of inequality is built upon social stratification, on classes, castes, estates, or any other differentiation between privileged and discriminated, not naturally given strata of society. Here, inequality has not to be added to a given natural distinction, but it is part of the distinction itself. Basically, the lower strata are obliged to follow the upper strata, the commoners have to follow the aristocracy, the courtiers have to follow the ruler, and so forth. I call this type “institutional following into death.” Within its framework, various forms are possible. The distinction between followers and those to be followed need not necessarily encompass the whole society. Often, such a custom is a phenomenon remaining within the upper strata, for example, when only the ruler, or the members of the ruling family, have the right to be followed, while the aristocracy or those with access to the court have to provide the followers. In this case, the lower classes are excluded altogether, except perhaps as spectators. This does not mean a more powerful position in society of the excluded groups (although it means, of course, an improved condition for the individuals, insofar as they are not liable to this form of loss of life), but rather it is an exclusion from the struggle for power. There can even be iteration: whereas the aristocracy is liable to provide followers for the ruler, they themselves have or at least claim the right to be followed by their own retainers. This seems to have occurred frequently in Africa.33
      Logically, the two types mutually exclude each other. In practice, however, most of the empirical cases contain elements of both in varying degrees. An existing custom is thus liable to constant change, to movements from a more institutional to a more individual character and vice versa.34
      Where the individual aspects of the custom prevail, there is frequently a tendency to influence it in the direction of becoming at least a relative prerogative of the wealthy and powerful. This does not necessarily mean that the upper classes intend to monopolize the custom for themselves—it can also be the result of circumstances. Public ceremonies, which are indispensable, are expensive, and they are often not paid for by the state or by society at large. Thus it is easier for wealthy families to perform the ceremonies, and if nevertheless the poor do perform them too, then the rich can render their own ceremonies still more sumptuous and imposing. Moreover, there is some likelihood that the upper classes will be more concerned about the hereafter and especially about their own position there, as they have more to lose than the poor.35
      The sources confirm these reflections. In India, about 50 percent of the satis belonged to the upper castes (varna) of the Brahmans and the Kshatriya, while these represented only about 10 percent of the population at large.30 Their ceremonies were of course more sumptuous and spectacular than those of the lower castes and of the outcastes, attracting greater crowds. There were even attempts to exclude lower castes and outcastes altogether, but they were never successful. Thus, within a basically individual custom there were many institutional elements. Among the rulers this tendency went one step further. In many of the Hindu principalities, even in those in dependent alliances with and thus subordinate to the British, it was an unwritten rule that at the funeral of a ruler there had to be at least one and, if possible, a whole group of satis, thus outshining even the richest and most influential subordinates.31 Although there are no statistics, the sources leave no doubt that in other areas of the world similar predispositions were to be found. Thus in the South Pacific islands, chiefs were followed more frequently and by more women than commoners were. Such tendencies could even be intensified when it became the purpose of the rulers or of the upper classes to exclude the lower classes, by monopolizing the right to be followed. This seems to have happened (without there being a strict law) at some time before the seventeenth century in Bali. The sources mention the killing of widows only at funerals for rulers and members of the aristocracy, while in earlier times the custom probably had been practiced among other castes or classes as well. It is obvious that such restrictions are not the result of a preoccupation with saving lives, to avoid a custom that might threaten the survival of a society, but rather with monopolizing power.36
      Institutional following into death shows an even more outspoken tendency to constant rivalries and struggles for power because the assignment of the roles of followers and of those to be followed is much more open than in the context of the individual type, as it does not rest on a natural distinction. What are the social limits of the right to be followed and of the duty to follow? Does only the ruler have this right or is it extended to the whole of the nobility, or does some other social group, class, or estate have it? Struggles of this kind can best be seen in Africa, where rulers sometimes tried to monopolize the right to be followed, but often did not have the power to enforce it. Within the group of those who are entitled to be followed, additional restrictions based on other criteria, such as age and especially sex, can be introduced. Basically, the institutional type indicates that women have to be followed in the same cases, meaning not less frequently and in the same manner as men. There is, however, only one such instance on record: among the Natchez Indians in the lower Mississippi Valley.32 Here we have excellent sources for the seventeenth and eighteenth centuries. An almost “perfect” system of the institutional type seems to have prevailed. The right to be followed depended exclusively on social stratification, and within this framework, there was no distinction between men and women. The central feature was the possibility for commoners of both sexes to marry into the nobility, at the price of having to follow one’s spouse into death in the case of his or her dying first, while the spouse of noble origin, who could marry only commoners, had no corresponding duty. In other institutional cases, women usually were not categorically excluded from being followed. But at least in practice and often also in theory they “enjoyed” their privilege much less frequently than men. This did not necessarily mean that they had to become followers more frequently than men. If the followers were mainly members of the court, they tended to be men. In addition, however, there were very frequently spouses as followers— and this right or duty was usually reserved for women. This meant that male rulers in most cases were followed not only by courtiers, but also by some of their wives and concubines, while women rulers or other important female members of the court were usually followed by only male and female courtiers and attendants, and only in very rare cases by their spouses.3337
      The custom is a burden for any society and especially for the groups that have to bear its costs in human lives. But it also conveys power and prestige to some individuals or groups within the respective society. Therefore, there are frequent attempts by lower classes to acquire the right to be followed. If this right extends to the lowest classes, the character of the custom necessarily becomes individual, as there is no social stratum below those to be followed. Instead, the differentiation between bearers of rights and bearers of duties develops between the sexes. This is most likely to occur when the upper classes become either weak or no longer are interested in the custom, without wanting to prohibit it. This may be due to a change of beliefs in the hereafter or to a change to rulers, often of a new dynasty, who never have had such a belief. In India, both the Muslim and the European conquerors abhorred sati and had no interest in maintaining it as a privilege for themselves. For them it was a question of tolerating sati or prohibiting it altogether, not of maintaining and defending it as a prerogative of their own. So long as they did not dare to abolish it altogether, they had no interest in privileging one group. It was somewhat similar in China and Japan. Here, Confucian criticism led to the abolition of the custom in favor of the rulers. But in time, similar customs were introduced among the lower classes in China, with widows killing themselves at the death of their husbands, and in Japan, with samurai killing themselves at the death of their daimyo. The latter remained a custom with institutional character, although its adherents were not placed at the top of the social ladder, while the former was of a mainly individual type.38
      In India, there was yet another social factor that weakened the institutional character of sati. Originally, it seems to have been mainly institutional, being performed at the death of rulers and important warriors. It conveyed enormous social prestige. The priestly caste, the Brahmans, opposed it. But they were unable to suppress it. This was a disadvantage in a struggle for social prestige between the two castes. In order to deprive the warrior caste of their de facto monopoly, the Brahmans started to regulate sati with religious prescriptions, thereby gradually lowering the requirements and thus opening the custom both to themselves and to the lower castes and even the outcastes. While the question of the right to be accompanied became a bone of contention between the two upper castes, the lower strata took advantage of this competition by appropriating to themselves a right that the warriors were no longer able and the Brahmans were unwilling to deny them. The losers were the women.39
      The various social configurations and gradual shifts outlined above show that the relation between institutional and individual elements often is subject to changes as a result of power struggles and changing beliefs. The same holds for the functions of the custom, which by now have become clear. Basically, they correspond to the two necessary causes.40
      In the first place, following into death is both an expression and a result of power struggles in a society. In theory, it can bear on most members of a society even with respect to life and death, but in practice it always has an exceptional character. Its function usually is much more symbolic than real, if considered from the point of view of the whole society, as only a small number of individuals are directly involved. But for them, who thus themselves become the symbols, it is terribly real. This may be one of the reasons for the comparative attractiveness of such customs in so many societies. Following into death symbolizes and expresses in a fatal reality social inequality to the society at large. Others might become victims themselves. The public rituals in which the victims are killed or kill themselves weld together the potential (future) followers and those with the right to be followed. Thus, while it clearly draws a line between privileged and discriminated groups, it also joins them in one apparently indivisible society. This aspect is even more emphasized when the ceremony is declared to have beneficial consequences for the whole society. In Africa, it was frequently believed that if one dispatched followers to the previous ruler he would procure prosperity for the entire people. The often spectacular public ceremonies give to the victim a special status, higher than any other status she could attain in her society. This holds in particular for satis in India. The widow becomes, as a sati, ipso facto a goddess, who is deeply venerated and whose place in public memory is high above the place of most members even of the privileged groups. But the price for attaining this position is death. Only in death can the social distinctions and discriminations be surmounted. While the unity of society is emphasized, the distinction between the social groups is confirmed with blood.41
      This leads to the second function and at the same time shows that both functions can be neither separated from nor reduced to each other. The victim, or the follower (even the difficulty or rather the impossibility of finding a generally satisfying terminology shows that the different aspects belong together), is a concrete individual and member of society, and at the same time she is looked upon by society first as a prospective and then as a real dweller in the hereafter. A belief in the hereafter is inseparable from the custom. Although this is not its sole function, it certainly operates to provide followers to dead persons, people who care for them in the next life, whether in official or in private functions. The conviction—which perhaps is not shared by everybody but at least deeply imbues society at large—that followers are needed on its part intensifies social struggles, as it is impossible to send people indiscriminately into the hereafter. Sometimes, the custom may be a consequence primarily of the need for followers; sometimes class or gender struggle may lead to the introduction or at least to the intensification of following into death.42
  
Free Will or Compulsion? 
Among outsiders, the question most often and most controversially discussed in the context of following into death is whether the death of the followers occurs voluntarily or involuntarily, by compulsion. The respective convictions can become almost a matter of faith. The answer certainly depends to a great extent on whether emphasis is put mainly on social struggles between classes and genders, so that the followers become victims of this struggle, or on the belief in a life in the hereafter and the desire of the followers to join the person they follow. As has been shown, neither of these aspects can be isolated.43
      To begin with, a closer look at the manner of killing is indispensable, although it cannot provide a full answer. There is both murder or manslaughter (or, speaking more neutrally, killing by others) and suicide involved. The first usually leads to the presumption of coercion and violence. But there are cases in which it occurred at the demand of the person killed, and this kind of killing is in many societies not considered involuntary. The strangling on the South Sea islands, for example, was often done after the repeated demand of the widow. On the other hand, suicide has the presumption of being voluntary. But there can be demands for suicide attended with the threat of murder that leave little choice to the victims; in such a case their death certainly cannot be considered voluntary.44
      There can be no doubt that following into death was in many cases nothing but murder. Extreme cases are on record in Africa: sometimes at the death of a ruler people were more or less indiscriminately apprehended in the streets and either killed on the spot or, usually, slain later in a ceremony. They were never asked whether they wanted to accompany the dead.34 Chinese historiography reports how sometimes great numbers of ordinary people were lured to the mausoleum of an emperor only to be locked up and thus buried alive.35 In India too, shocking scenes are on record. On 20 November 1820, the following happened in the district of Gorakhpur:
One Seetloo, a Brahmin, died when absent from his family. A fortnight afterwards, his widow, Hoomuleea, a girl of about fourteen years of age, proceeded to burn herself, the pile being prepared by her nearest relations then at the village she resided in…. Whether the sacrifice was originally a voluntary one, has not been ascertained; it must be presumed it was so.
      The preparatory rites completed, Hoomuleea ascended the pile, which was fired by her uncle, the prisoner Sheolol. The agony was soon beyond endurance, and she leaped from the flame, but seized by Sheolol Bhichhook, and others, she was taken up by the hands and feet, and again thrown upon it; much burnt, and her clothes quite consumed, she again sprang from the pile, and running to a well hard by, laid herself down in the watercourse, weeping bitterly. Sheolol now took a sheet, offered for the occasion by Roosa, and spreading it on the ground, desired her to seat herself upon it. “No” she said, “she would not do this, he would again carry her to the fire, and she could not submit to this; she would quit the family, and live by beggary; any thing, if they would have mercy upon her.” Sheolol upon this, swore by Ganges, that if she would seat herself on the cloth, he would carry her to her home. She did so; they bound her up in it, sent for a bamboo, which was passed through the loops formed by tying it together, and carrying it thus to the pile, now fiercely burning, threw it bodily into the flames. The cloth was immediately consumed, and the wretched victim once more made an effort to save herself, when at the instigation of the rest, the moosulman Buraichee, approached near enough to reach her with his sword, and cutting her through the head, she fell back, and was released from further trial by death.36There is one approach that categorically excludes any free will. This is the theory of structural violence.37 Even in the case of apparently voluntary suicide, a society that admits to and indeed fosters such a custom is imbued with violence to such an extent that in reality the follower is the victim of those structures, of structural violence. While this view in itself is quite logical, it is unable to categorize the real, manifold, and important differences between the various manners of death mentioned in the sources. There can be little doubt that not only within the legal and moral framework of the respective societies but also according to modern standards there were sometimes genuinely voluntary acts by persons who fully knew what they did. There are numerous examples on record. The European observers of sati in the nineteenth century not only unanimously abhorred the custom, but the prevailing intellectual atmosphere in Europe was such that it was next to impossible to maintain the opposite view—an adherent of sati would almost have had the position that an adherent of torture would have today. Thus European eyewitnesses constantly were on the lookout for any kind of compulsion. Therefore, if they affirm free will, it can be taken for granted. In November 1829, a few weeks before the abolition of sati, Wiliam Henry Sleeman, a British official in India, had to deal with a widow of about sixty-five years who wanted to burn herself with the body of her husband. Sleeman refused permission and had the body of the husband burnt at a safe distance from the widow. She, however, insisted. She began fasting. Sleeman, fearing she would starve herself to death, threatened to confiscate the family’s rent-free lands while promising her a good provision if she would live. It was all to no avail, and after five days Sleeman gave in. “The old lady … seemed extremely pleased and thankful.” When the fire was lit, she “walked up deliberately and steadily to the brink, stepped into the centre of the flames, sat down, and leaning back in the midst as if reposing upon a coach, was consumed without uttering a shriek or betraying a sign of agony.”38 On Fiji, in the 1840s, the British Wesleyan missionary Thomas Williams saved a widow from being strangled by taking her into a Christian village. But she escaped and asked the relatives of her deceased husband to strangle her, which eventually they did.39
45
      Of course it is possible to explain these and similar cases by referring to the extraordinary situation in which the victims were. It is not unlikely that a month or even a few days later they would have decided to stay alive. Nevertheless, there is no doubt that these cases are worlds apart from the indiscriminate slaying of victims who perhaps do not even know that somebody has died, or from the murder of Hoomuleea. In addition, there is another structural consideration. The essential feature of the customs described is their public, ritual character. In the long run it is difficult to maintain such proceedings with completely unwilling victims who try to resist as much as possible. It is not impossible, as the case of Hoomuleea (and other, similar Indian cases) or African examples, which rather resemble public executions, show.40 But the very brutality of such scenes had the effect that they could not become the rule: the public in all probability would not have tolerated them as regular features of the custom. In India, the decisive question was whether, immediately after her husband’s death, the widow consented to die with him. There was probably often enormous psychological pressure on her; but usually no outright violence was applied. Thus, if she remained steadfast, she had a fair chance of surviving, although the outlook for her remaining days (25 percent of the sati were under thirty, 43.1 percent under forty, and 62.9 percent under fifty years old 41) was utterly bleak. Once she had consented, however, violence would probably be used if she tried to retract afterward.46
      This leads to the conclusion that within the framework investigated here, the cases range from ordinary and sometimes even outra geously cruel murder to almost stubborn suicide. In each case there was a particular mixture of consent and compulsion, of free will and violence. By what factors was this mixture determined?47
      The most important element was certainly the personal relationship between follower and followed. The closer it was, the more likely the former would be prepared and even demand to follow the latter. It is probably no coincidence that the cases reported by Sleeman and Williams were of widows who had been married for a long time (whereas Hoomuleea was only fourteen). The fact that there are no examples of husbands (some of whom probably did not love their wives less than some wives loved their husbands) following their wives into death does not disprove the argument. The framework was given by society, and it could not be transcended by the individual. It meant that women had to follow men, and men did not follow women. It was maintained and enforced by society, which had to countenance the act by making it a public and legal ceremony. Nobody could commit sati against the will of the society surrounding him; he only could commit suicide, which meant that he had to kill himself in secret.48
      The second factor was belief in a hereafter. If someone was convinced that following a dead person would enhance her status in the next world or reunite her with a beloved person, this increased the likelihood of her killing herself or of letting others kill her. The sources never allow us to prove such a belief beyond any doubt; nevertheless there is much evidence that it played an important part.49
      There were, in addition, this-worldly incentives for the victims, although their importance should not be exaggerated: the followers could acquire an exalted position for the remaining (usually very short) period of their life and great fame after their death. Thus, in India a sati became ipso facto a goddess who frequently was venerated in a temple built exclusively for her.4250
      The motivation a strong belief in a hereafter could exercise on a follower should not be mistaken for another incentive that was directed not at the victim but at her relations: frequently, the fact of having had a follower among its members enhanced the social and material status of a family. On Bali, for example, the widow’s family usually acquired privileges and material possessions.43 In India, there were similar inducements, although they did not come from the rulers. In some parts of the country, especially in Bengal, the D�yabh�ga law of inheritance left the widow’s property to the family of the husband, who thus had a direct interest in her death. At least in one case, however, such material advantages in the first place benefited the (prospective) victims. Among the Natchez, the commoner who married a member of the nobility gained an important enhancement of status—for the price of having to die if his or her noble spouse died first.4451
      While the interest of the family in the death of the victim suggests compulsion, there can be no doubt that in most cases it is not a sufficient reason to explain the death of the follower. Each case has to be accounted for separately. Psychological factors have to be taken into account as well. The prospective victims had often been prepared and educated during their whole lives for their supposed duty. When the decisive moment came, it was difficult to avoid the seemingly logical conclusion. But both in China and in India, widows sometimes killed themselves only many years afterward, having had ample time to reflect. In such cases another standard explanation comes to mind. The life of a surviving widow was so dreadful that she often preferred death to a continuous suffering. But this leads to the question of why so few widows actually did kill themselves.52
      Considering these various elements, there seems to be no doubt that both compulsion—structural and individual—and free will have to be taken into account in explaining the phenomenon. Both belong inextricably together, in the same way as material and spiritual aspects, this-worldliness and otherworldliness.53
  
The Origins 
There is no way of telling how following into death historically developed. Whenever a custom of this description appears in the written sources, it is already fully developed, and it is impossible to determine the length of time it has been in use by then. This holds even more for archaeological sources. While they reach much further back, they can give conclusive evidence only in very rare cases, and they never can tell the story of how the custom was introduced in a particular society.54
      The development of the customs in question thus can never be the object of a historical narrative, but only of speculations, hypotheses, and guesswork. The necessary conditions in the shape of a certain belief in a hereafter and of social inequality can give no more than a framework, while the additional conditions of the individual case must be left open. However, at least a few additional remarks and hypothetical qualifications seem possible.55
      Following into death is not a custom that more or less automatically develops out of human nature. It presupposes a relatively high degree of social differentiation with the corresponding social inequality. Only where either some kind of classes, estates, castes, or other social strata or a clear-cut division of labor with matching differences of power between the sexes, age groups, or races exist can a custom of following into death develop. Therefore, the origins should probably not be sought too far back in history.56
      It seems appropriate to differentiate between the origins of institutional and of individual following into death. The roots of the latter might lie in a genuinely individualistic context: when persons, usually lovers or spouses, either die together, in a common suicide, or follow each other into death, regardless of whether the man or the woman dies first, killing themselves not in a ritualized, public ceremony, but in private. If such behavior becomes more frequent, society may become more concerned about it and try to control and to limit it by incorporating it into a public ceremony. The consequences are that society plays a much more important part in it and that it becomes a ritual in which social inequality, in this case the inequality between the sexes, is reflected. Society can prescribe who is allowed or required to participate in what capacity in the ceremony, with the consequence that usually men no longer are allowed to follow and women no longer are allowed to be followed. There is evidence from the South Pacific islands, from New Zealand, and from North America that there was indeed frequent private following into death in which either sex participated on both sides, but that there was a tendency to convert the private act into a public ceremony with fixed, one-sided rules for both sexes.57
      The institutional form presents, in the view of those who are followed, a dilemma. On the one hand, within the religious context of life in the hereafter, it is their main interest to transfer their accustomed environment into the other world, which means preference for those who are closest to them. On the other hand, the custom has to emphasize social inequality to demonstrate to the lower classes or groups their inferior position by obliging them to follow. But the more this aspect is stressed, the more likely it becomes that the person with the right to be followed is followed by his potential or actual enemies. This happens fairly often, for example, when innocent victims are indiscriminately slain in Africa, and even more in headhunting, when the survivors kill someone in order to serve a dead person in the hereafter. Such cases presuppose an additional mental operation: it must be ensured that in the hereafter enemies become perhaps not friends, but at least faithful servants so that the person to be followed can indeed have an interest in being followed by them. It is to be supposed that this construction will turn up only in a later phase of such a custom, that in the beginning preference will be given to close and familiar followers, but that then in some cases there will be a gradual extension to more and more distant and anonymous categories of victims, as a consequence of a supposed growing need to stress social inequality. This is done not only with a view to impressing and intimidating enemies. If too many friends and adherents are killed or induced to kill themselves, then the ruling group will be weakened and risk losing its paramount position; so the successors of the dead ruler will tend to restrict the number of “friendly” followers.58
      There is evidence that the problem of converting enemies into friends or loyal servants has been seen and taken seriously. Among the Lakhers, in Northeast India, there was a ceremony, after a head had been procured, which converted the slain enemy into a loyal follower of the person he had to accompany. Otherwise he would only have troubled him.4559
  
Opposition and Suppression 
The various customs of following into death are no longer practiced. They belong to history, with one exception: India. But even there sati has become extremely rare, and it is, at least on paper, very harshly suppressed. Legislation introduced in 1988 makes anyone connected with a sati ceremony liable to a mandatory life sentence.46 This leads to the search for the reasons of the discontinuance of such customs. In this case, the sources are not silent. The basic distinction is between opposition arising inside a society, for whatever reason, and opposition or rather suppression from outside. The former need not be more peaceful than the latter, but the latter is usually more efficient and successful.60
      Internal opposition can have many roots. The one that might seem most obvious, resistance of the victims who simply do not want to die, seems to have been extremely rare. There was of course frequent individual resistance, especially by those who tried to escape in the last moments. But attempts of this kind usually had little chance of success, and they had no consequences. A kind of organized resistance can be traced in the sources only in some parts of Africa, where occasionally potential victims fled and refused to return unless they were given guarantees against being used as followers. At times they even had some success.47 But this kind of resistance did not put into question the custom itself. If it became so vehement that it threatened the cohesion of the respective society, it could at best reduce the number of victims, while it remained a zero-sum game so long as other victims could be procured. There is no case on record in which a change in the distribution of power in society or more particularly a leveling out of social inequality led to the abolition or simply to the disappearance of the custom.61
      Opposition in connection with the belief in an afterlife was rather rare, and success was even rarer. The most important case is China, where there was, in the Confucian tradition, strong opposition from the fifth century B.C.E. Eventually it was fairly successful in that it managed to stamp out the institutional form almost completely. But the consequence was the development and spread of individual forms, mainly of widows killing themselves in public after the death of their husbands. The state did not attempt to suppress the custom with all its power, probably mainly because the officials themselves often admired the widows, who were held to be extremely virtuous.62
      The Chinese case allows two conclusions. First, there is little chance of thoroughly rooting out such customs without really rooting out corresponding beliefs in a hereafter. Two thousand years of opposition on the highest levels of state and society had not been able to do so in China. Second, it is easier to root out institutional than individual following into death, especially under premodern conditions. If the institutional form prevails, only a comparatively small group, the highest strata of society, has to be convinced. It will probably be more easily accessible for rational arguments, but more important is the fact that it will be much more dependent on the central authority and its favors than the mass of the people. Those who refuse to obey can be taken to account. If, however, the custom itself is accepted and the corresponding belief in a hereafter is shared by the whole society, it is almost impossible to root them out, as not only China but also India shows.63
      Japan is a similar case. There was also criticism along Confucian lines here, with the result of institutional following into death being suppressed. But the custom was revived in the warrior caste, to be suppressed again by the shoguns in the eighteenth century, mainly for political reasons. Thus the suppression was indeed successful, while the underlying beliefs lingered on, as was shown by the widespread admiration for General Nogi Maresuke’s suicide in 1912, at the death of the Meiji emperor.4864
      There may have been more internal opposition in other parts of the world—but it is not on record and it certainly was not successful. In India, there were always groups and sects not practicing sati and sometimes more or less opposing it. But those currents never became popular movements.65
      In most cases that can be reconstructed, the suppression and eventual disappearance of the custom was the result of external influence, pressure, or administrative action, mainly by the European colonial powers. Islam, Christianity, and Judaism all had an outspoken belief in a final judgment and consequently rejected all kinds of following into death. This did not mean, however, that such customs had to be rooted out even among nonbelievers; they were only incompatible with one’s own religion. Initially, neither Christianity nor Islam bothered much about what adherents of other religions did. This gradually changed, especially with Christianity after the beginning of European colonial conquests. The growing abhorrence of such customs was, however, not so much the result of growing religious zeal— rather, it abated—as the indirect consequence of the development of the modern state. The European state now took care of its subjects in a much further-reaching way than it had done previously. It held itself responsible for customs that so far had been considered exclusively matters of religion, belonging to society. Following into death was in this category, and it appeared more and more as cruel and inhumane, whatever its religious roots might be. It simply was no longer possible officially to tolerate such customs in the framework of a modern, rational bureaucratic (colonial) administration. This usually seemed self-evident, and the Europeans tended to suppress following into death whenever they gained political and administrative control, as a rule with quick success. After all, in many societies the customs were of restricted significance, and usually the opposition of those who had profited from them was rather muted. There was, however, an important exception that in the end only confirmed the rule. While the British suppressed following into death in all their possessions in a similar manner as other colonial powers did in theirs, they refused to do so in India, mainly for fear of resistance and unrest. Nevertheless, they could not just shut their eyes before occasional very shocking cases—after all, they had proclaimed themselves responsible for the well-being of their subjects. Therefore they tried to administer sati by distinguishing between legal and illegal cases. But this distinction ran counter to the moral feelings of the administrators who, with their European background, abhorred legal and illegal cases alike. As there was growing insubordination in the administration, followed by a public campaign both in Britain and in India, sati was prohibited altogether in 1829, with almost complete and immediate success.4966
      The success of the Europeans showed that it was easier to abolish such customs from outside than from inside, by a power that on the one hand had no vested interests in them and on the other hand was alien to all kinds of beliefs connected with them. This was the general framework for the success. Nevertheless, there were differences in its degree. Institutional forms were easier to suppress than individual ones. In Africa, the former were particularly important; following into death was often concentrated on the person of the ruler, and sometimes it was intimately connected with the cohesion of the state. Several European observers and witnesses were convinced that some African states would be threatened in their very existence if they had to abolish these customs. Richard Burton, the famous traveler, wrote: “It is evident, that to abolish human sacrifice here is to abolish Dahome.”50 The Europeans were corroborated in their opinions by African rulers,51 who no doubt were glad to have the opportunity to do so. But in all probability they believed no less in what they said than did their European visitors. Nevertheless, after the colonial conquest the Europeans easily suppressed the customs—but then the African states indeed had vanished.67
      Where following into death had mainly an institutional shape, it was more difficult to stamp it out completely. India is the foremost example, as it is still struggling against a custom that in the official view of the state and in legislation is considered simply a crime, while in the eyes of probably still a majority of the people it is a highly meritorious act.68
      In other parts of the world, and especially in Africa, there is an additional factor that contributed to the final disappearance, so that no revivals were reported even after decolonization, when European colonial pressure ceased. This factor is the spread of religions hostile to the custom—Christianity and Islam—in many areas where the custom could formerly be found.69
      Finally, there is a new and growing moral pressure, not dissimilar to the European moral pressure of the nineteenth century: of course, following the dead is utterly incompatible with what now are no longer standards of civilization but standards of human rights, so that a state that would openly tolerate or even propagate such customs would become an outcast. Nevertheless, the decisive factor remains religion, with its different forms of belief in a hereafter. So long as there are beliefs that are compatible with following into death, the possibility of its revival cannot be ruled out.70
  
Conclusion 
Sati, the burning of widows in India, has probably been the best-known (and the most despised or lamented) Indian custom in Europe since the ancient Greeks. Comparisons with other customs usually have been made on the phenomenological level, especially with the burning of witches and heretics. Here it is suggested to introduce comparisons on the functional level. The main question is thus not what happens, but what the function and purpose of the act are. Seen in this perspective, the central aspect becomes the connection between this world and the other world, between the living and the dead. On the basis of a belief in a hereafter that is an immediate continuation of this life, both worlds are connected by an act in which a dead person is followed, voluntarily or by force, in a public ceremony, by one or several living persons, thus emphasizing the continuity between the two worlds. The particulars of the ritual, whether it is killing or self-killing, and the manner of killing are, from a functional point of view, unimportant. Usually, this manner corresponds to the manner of disposing of the dead. Thus, for example, in Indian castes that bury the dead, sati is usually performed by burying alive. In other contexts, the method of killing is to preserve the body of the followers as intact as possible, so as to enable them to do their duties in the hereafter, which often leads to strangling.71
      Following into death thus defined can be shown to have occurred, in the course of history, in most areas of the world for a longer or a shorter time, with the notable exception of Western Europe, for which there is no satisfactory explanation so far, due to the lack of sources. Two further important questions cannot be answered either: we can only guess the context in which such customs had their origin, as we don’t know when they developed, and we do not know whether there was a kind of diffusion from one point of origin or whether the relevant customs developed independently from each other in different places. The only exceptions are Southeast Asia, where the occurrence of widow burning makes Indian influence very likely, and Japan, where there was probably Chinese influence.72
      Following into death is of special interest because it links two worlds and thus religious with sociopolitical aspects. It is a matter of life and death. As it is physically impossible to accompany every dead person, the question of who has the right to be followed and who has the duty to follow arises. Following into death thus not only becomes a mirror of social structure and political power; it also can influence them. There are two main functions in this framework. Especially in India, sati reflects, confirms, and strenghtens the subordination of men to women. In many other places it has the same effect for the superiority of the ruling groups. The ceremony of killing the followers shows the long-term results of social and political power struggles.73
      Following into death presupposes certain religious beliefs (as a necessary, not a sufficient, condition). Wherever beliefs in a final judg ment prevailed over beliefs in a transfer of this world into the other, following into death either vanished, if the beliefs of the people at large changed, or were suppressed, if foreign conquerors had sufficient strength to abolish them. This is what European colonialism did, mainly with humanitarian arguments, but basically because of its own religious background.74

Notes1 Cf. the press reports in Mulk Raj Anand, Sati: A Writeup of Raja Ram Mohan Roy about Burning of Widows Alive (Delhi, 1989), pp. 81–188, and the—somewhat one-sided— summary in Mark Tully, No Full Stops in India (London, 1991), chap. 7.2 See for example The Times of India, 8 July 2002, p. 9 on a case of 8 June 2002 in Madhya Pradesh. Also The Hindu, 8 August 2002, p. 5.3 For a survey and a detailed bibliography of the vast literature on sati see Jörg Fisch, Tödliche Rituale: Die indische Witwenverbrennung und andere Formen der Totenfolge (Frankfurt, 1998), pp. 498–500, 525–548. An early and still valuable monograph is Edward Thompson, Suttee: A Historical and Philosophical Enquiry into the Hindu Rite of Widow-Burning (London: George Allen & Unwin, 1928). Dealing with similar questions, but on the basis of modern research, is V. N. Datta, Sati: A Historical, Social and Philosophical Enquiry into the Hindu Rite of Widow Burning (New Delhi: Manohar, 1988). More thorough and concentrated on Bengal is Bhenoy Busan Roy, Socioeconomic Impact of Sati in Bengal and the Role of Raja Rammohun Roy (Calcutta: Pilgrims Publishing, 1987). An excellent article is Catherine Weinberger-Thomas, “Cendres d’immortalité: La crémation des veuves en Inde,” in Archives de Sciences Sociales des Religions 67 (1989): 9–51, and her book Cendres d’Immortalité: La crémation des veuves en Inde (Paris: Seuil, 1996) (English translation by Jeffrey G. Mehlman and David G. White, Ashes of Immortality: Widow Burning in India [Chicago: University of Chicago Press, 1999]). For an introduction to more specialized research, which frequently takes up postmodern approaches, cf. for example John Stratton Hawley, ed., Sati: The Blessing and the Curse: The Burning of Wives in India (Oxford: Oxford University Press, 1994); Lindsey Harlan and Paul B. Courtright, eds., From the Margins of Hindu Marriage: Essays on Gender, Religion, and Culture (Oxford: Oxford University Press, 1995); Lata Mani, Contentious Traditions: The Debate on Sati in Colonial India (Berkeley: University of California Press, 1998).4 Johann Wolfgang von Goethe, The God and the Bajadere (1797); Robert Southey, The Curse of Kehama (1810); Jules Verne, Around the World in Eighty Days (1873); M.M. Kaye, The Far Pavilions (1978). Cf. Rajeswari Sunder Rajan, Real and Imagined Women: Gender, Culture and Postcolonialism (London: Routledge, 1993), chap. 2, and Dorothy M. Figueira, “Die flambierte Frau: Sati in European Culture,” in Hawley, Sati: The Blessing and the Curse, pp. 55–72.5 See the bibliography in Fisch, Tödliche Rituale, pp. 496–498, 501–524.6 Pompa Banerjee, Burning Women: Widows, Witches, and Early Modern European Travelers in India (New York: Palgrave Macmillan, 2003).7 Cf. Werner Menski, ed., South Asians and the Dowry Problem (Stoke on Trent: Trentham Books, 1998).8 I use this formula because it is comparatively open as to agency. Strictly speaking, the verb is, of course, transitive and thus suggesting that the widow is burned, presumably by somebody else, and not reflexive in the sense that she is burning herself. But the formula is so short that the second meaning cannot be excluded. The question of voluntariness and compulsion should remain open here; it will be discussed later.9 For a survey, see, for example, Alfons van der Kraan, “Human Sacrifice in Bali: Sources, Notes and Commentary,” Indonesia 40 (1985): 89–121.10 Diodorus Siculus, Bibliotheca Historica, 19.32.3–19.34.6.11 There is, as for most other parts of the world, no comprehensive work on the South Pacific. A good contemporary account is Thomas Williams and James Calvert, Fiji and the Fijians, ed. George Stringer Rowe, 2 vols. (London: Alexander Heylin, Paternoster Row, 1858).12 There has been a controversy on the question whether the killings at the death of rulers were usually restricted to executions of criminals. But the sources do not allow this view. Cf. for example Clifford Williams, “Asante: Human Sacrifice or Capital Punishment? An Assessment of the Period 1807–1874,” International Journal of African Historical Studies 21 (1988): 433–441; Ivor Wilks, “Asante: Human Sacrifice or Capital Punishment? A Rejoinder,” International Journal of African Historical Studies 21 (1988): 443–452; Robin Law, “‘My Head Belongs to the King’: On the Political and Ritual Significance of Decapitation in Pre-Colonial Dahomey,” Journal of African History 30 (1989): 399–415.13 John Roscoe, The Baganda: An Account of Their Native Customs and Beliefs (London: Macmillan, 1911), pp. 106f.14 For a detailed description for Dahomey, cf. Richard F. Burton, A Mission to Gelele, King of Dahome, 2 vols. (London, 1864), chap. 14. A modern analysis is Thomas C. McCaskie, “Death and the Asantehene: A Historical Meditation,” Journal of African History 30 (1989): 417–444, esp. 431–437.15 Some detailed descriptions by a basically not unsympathetic observer can be found in J. A. Skertchly, Dahomey As It Is: Being a Narrative of Eight Months’ Residence in That Country With a Full Account of the Notorious Annual Customs, and the Social and Religious Institutions of the Ffons; Also an Appendix on Ashantee, and a Glossary of Dahoman Words and Titles (London, 1874), pp. 338–361.16 The most detailed and thorough calculations for Asante in 1867–1868 are to be found in McCaskie, “Death and the Asantehene,” pp. 431–437. McCaskie’s estimates are between 1,853 and 3,600 victims for this period.17 There is a summary of Spanish sources in Hubert Howe Bancroft, The Native Races of the Pacific States of North America, vol. 2 (New York, 1875), pp. 603–622. Cf. also Diego Durán, Historia de las Indias de Nueva España e islas de la tierra firme, vol. 1, 2nd ed., ed. Angel Ma. Garibay K. (Mexico, 1984), pp. 55f.18 The excavations are described in P. R. S. Moorey, Ur “of the Chaldees”: A Revised and Updated Edition of Sir Leonard Woolley’s Excavations at Ur (Ithaca, N.Y.: Cornell University Press, 1982), pp. 51–103.19 A good collection of Chinese material can be found in J. J. M. De Groot, The Religious System of China, Its Ancient Forms, Evolution, History and Present Aspect, Manners, Customs and Social Institutions Connected Therewith, vol. 2 (Leiden: E. J. Brill, 1894), pp. 721–827.20 For a case of 1935 see Florence Ayscough, Chinese Women Yesterday and Today (Boston: Houghton Mifflin Company, 1937; repr. New York: Da Capo Press, 1975), pp. 73f.21 Cf. Mark Elvin, “Female Virtue and the State in China,” Past and Present 104 (1984): 111–152.22 There is no monograph on these aspects of headhunting in Southeast Asia. For an older survey of some customs see Robert Heine-Geldern, “Kopfjagd und Menschenjagd in Assam und Birma und ihre Ausstrahlungen nach Vorderindien,” Mitteilungen der Anthropologischen Gesellschaft in Wien 37 (1917): 1–65. Cf. also Meinrad Schuster, “Kopfjagd in Indonesien” (unpublished PhD diss., Frankfurt, 1956).23 Williams and Calvert, Fiji and the Fijians, pp. 203f. Cf. also Charles Wilkes, Narrative of the United States Exploring Expedition, during the Years 1838, 1839, 1840, 1841, 1842, vol. 3, (Philadelphia: Lea and Blanchard, 1845), pp. 158f.24 Felix Speiser, Südsee/Urwald Kannibalen. Reisen in den Neuen Hebriden und Santa- Cruz-Inseln (Stuttgart, 1924), p. 94.25 Cf. Jörg Fisch, “Humanitarian Achievement or Administrative Necessity? Lord Wiliam Bentinck and the Abolition of Sati in 1829,” Journal of Asian History 34 (2000): 109–134.26 The numbers are from Fisch, Tödliche Rituale, pp. 239–241, and are computed from the yearly returns. They are partly printed in the British Parliamentary Papers (House of Commons) and partly in the India Office Records (Oriental and India Office Collections, British Library, London).27 It has frequently been observed that the number of satis reached an unprecedented peak under the British, as a kind of Indian self-assertion or even resistance. But this remains pure speculation. The best-known propounder of this thesis is Ashis Nandy in “Cultures of Politics and Politics of Cultures,” Journal of Commonwealth and Comparative Politics 22 (1984): 262–274, esp. 264; and in “Sati: A Nineteenth-Century Tale of Women, Violence and Protest,” in Rammohun Roy and the Process of Modernization in India, ed. V. C. Joshi (Delhi: Vikas, 1975), pp. 168–194, esp. 170.28 For an overview, see Pandurang Vaman Kane, History of Dharmasastra (Ancient and Mediaeval Civil and Religious Law of India), 2nd ed., vol. 2, pt. 1 (Poona, 1970); and Wendy Doniger O’Flaherty, ed., Karma and Rebirth in Classical Indian Traditions (Berkeley: University of California Press, 1980). Cf. Fisch, Tödliche Rituale, pp. 291–296.29 Thomas Edward Bowdich, Mission from Cape Coast Castle to Ashantee (London, 1819; repr. London, 1966), p. 291; A. B. Ellis, The Tshi-Speaking Peoples of the Gold Coast of West Africa: Their Religion, Manners, Customs, Laws, Languages, Etc. (repr. Oosterhout, NB: 1966 [1887]), p. 287.30 Cf. Fisch, Tödliche Rituale, pp. 260f. More detailed for Bengal proper is Roy, Socioeconomic Impact of Sati, pp. 36–48, 148–166.31 In the Bengal Presidency, which was under direct British rule and therefore had no princes, there were only a few cases with more than one victim: out of 8132 cases, 128 occurred with two victims, four with three victims, and three with four victims (Fisch, Tödliche Rituale, p. 261). In the princely states, as a contrast, rulers were often accompanied by several of their wives and concubines. Thus in Bundi (Rajasthan), in nine cases a total of 237 women were burned (Datta, Sati, p. 164). Cf. ibid., chap. 6, and R. K. Saxena, Social Reforms: Infanticide and Sati (New Delhi: Trimurti Publications, 1975), Appendix.32 For an introduction see John R. Swanton, Indian Tribes of the Lower Mississippi Valley and Adjacent Coast of the Gulf of Mexico (Washington, D.C.: Smithsonian Institution, 1911; repr. New York, 1970), pp. 2f, 45–257; Douglas R. White et al., “Natchez Class and Rank Reconsidered,” Ethnology 10 (1971): 369–388; and Jeffrey P. Brain, “The Natchez ‘Paradox,'” Ethnology 10 (1971): 215–222.33 For the only case on record, in Asante in the nineteenth century, see n. 29.34 For a description referring to Asante in 1867 see Friedrich August Ramseyer and Johannes Kühne, Vier Jahre in Asante. Tagebücher der Missionare Ramseyer und Kühne, 2nd ed., ed. H. Gundert (Basel, 1875), p. 285.35 Cf. De Groot, Religious System of China, vol. 2, pp. 400f., 726, 730; and Marcel Granet, Danses et légendes de la Chine ancienne, vol. 1 (Paris, 1926), pp. 221f.36 Letter of Judge R. M. Rattray to the Nizamat Adalat, 25 May 1821, in British Parliamentary Papers, House of Commons 1823, vol. 17, paper no. 466, pp. 67f. Also in British Parliamentary Papers, House of Commons 1830, vol. 28, paper no. 178, pp. 7f.37 See the seminal article of Johan Galtung, “Violence, Peace and Peace Research,” Journal of Peace Research 6 (1969): 167–191.38 William Henry Sleeman, Rambles and Recollections of an Indian Official. Revised annotated edition by Vincent A. Smith (1844; repr. Karachi, 1973), pp. 18–23; quotations pp. 22f.39 Williams and Calvert, Fiji and the Fijians, 1:202. Cf. Wilkes, Narrative, p. 96, and John Elphinstone Erskine, Journal of a Cruise among the Islands of the Western Pacific, Including the Feejees and Others Inhabited by the Polynesian Negro Races (London, 1853), p. 228.40 To mention but two of the numerous accounts: Skertchly, Dahomey As It Is, pp. 368, 416; and William Allen, Letter of 2 February 1842, in T. B. Freeman, Journal of Various Visits to the Kingdoms of Ashanti, Aku, and Dahomi in Western Africa, 3rd ed., ed. Harrison M. Wright (London, 1968 [1844]), pp. 193f.41 Fisch, Tödliche Rituale, p. 490, table 6.42 It is part of the Indian government’s renewed struggle against sati to prevent the building of temples. In Deorala, where in 1987 Roop Kanwar burned to death, the police are staving off any attempt to worship her with a presence around the clock.43 Cf. Van der Kraan, “Human Sacrifice in Bali,” pp. 120f.44 See n. 32.45 N. E. Parry, The Lakhers, ed. J. H. Hutton (Calcutta, 1976 [1932]), p. 213. Cf. Schuster, “Kopfjagd und Menschenjagd,” pp. 40–54.46 Commission of Sati Prevention Act, 1987 (Act No. 3 of 1988), Delhi, 1990.47 For a particularly successful resistance in Calabar, southern Nigeria, in the 1850s, see Basil Davidson, Black Mother. Africa: The Years of Trial (London: Victor Gollancz, 1961), pp. 196f.; Elisabeth Isichei, History of West Africa Since 1800 (London: Macmillan, 1977), pp. 11, 117; Robin Law, “Human Sacrifice in Pre-Colonial West Africa,” African Affairs 84 (1985): 53–87, esp. 80f.; and the contemporary account by Hope Masterton Waddell, Twenty-Nine Years in the West Indies and Central Africa: A Review of Missionary Work and Adventure 1829–1858, 2nd ed., ed. G. I. Jones (London: Frank Cass, 1970 [1863]) pp. 642–644. McCaskie, “Death and the Asantehene,” pp. 433f. describes resistance in Asante in 1867.48 Cf. Robert Jay Lifton et al., Six Lives, Six Deaths: Portraits from Modern Japan (New Haven, Conn.: Yale University Press, 1979), pp. 29–66.49 See Fisch, “Humanitarian Achievement.”50 Burton, A Mission to Gelele, 2:26. Cf. Law, “Human Sacrifice,” p. 86, and Catherine Coquery-Vidrovitch, “La fête des coutumes au Dahomey: Historique et essai d’interprétation,” Annales 19 (1964): 696–716.51 Some examples: Skertchly, Dahomey As It Is, pp. 180f, 235, 237, 462; Law, “Human Sacrifice,” p. 85; Waddell, Twenty-nine Years, p. 294; Winwood W. Reade, Savage Africa, Being the Narrative of a Tour in Equatorial, Southwestern and Northwestern Africa (New York: Harper and Brothers, 1864), p. 54.

Content in the History Cooperative database is intended for personal, noncommercial use only. You may not reproduce, publish, distribute, transmit, participate in the transfer or sale of, modify, create derivative works from, display, or in any way exploit the History Cooperative database in whole or in part without the written permission of the copyright holder.