Almost every society has festivals that have retained a ritualistic character over the centuries. Of particular interest to the modern inquirer are observances involving the deliberate violation of established laws; for example, celebrations in which sexual promiscuity is not only tolerated but prescribed or in which incest becomes the required practice.

Such violations must be viewed in their broadest context: that of the overall elimination of differences. Family and social hierarchies are temporarily suppressed or inverted; children no longer respect their parents, servants their masters, vassals their lords. This motif is reflected in the esthetics of the holiday—the display of clashing colors, the parading of transvestite figures, the slapstick antics of piebald “fools.” For the duration of the festival unnatural acts and outrageous behavior are permitted, even encouraged.

As one might expect, this destruction of differences is often accompanied by violence and strife. Subordinates hurl insults at their superiors; various social factions exchange gibes and abuse. Disputes rage in the midst of disorder. In many instances the motif of rivalry makes its appearance in the guise of a contest, game, or sporting event that has assumed a quasi-ritualistic cast. Work is suspended, and the celebrants give themselves over to drunken revelry and the consumption of all the food amassed over the course of many months.

I have no doubt that these festivities commemorate a sacrificial crisis. It may seem strange to memorialize a traumatic event in such an uproarious manner, but the explanation lies ready to hand. The specifically “festive” aspects of the celebration, those that most effectively capture the attention, that dominate the spectacle, and that are the only ones to survive the evolutionary transformations of the festival—these aspects have nothing to do with the festival’s underlying cause. The fundamental purpose of the festival is to set the stage for a sacrificial act that marks at once the climax and the termination of the festivities. Roger Caillois has pointed out the sacrificial origin of festivals. If the crisis brought on by the loss of distinctions and the subsequent advent of reciprocal violence can be celebrated in such a jubilant fashion, it is because these holocausts are seen in retrospect as the initial stages of a cathartic process. The beneficial character of the generative unanimity tends to be projected onto the past, affecting the initial impression of the crisis and making it seem other than it was. The violent dismissal of distinctions now acquires a favorable connotation, which will eventually manifest itself as a festive display.

I have already advanced a number of interpretations that may prove relevant to the subject of festivals. For example, ritual incest ultimately acquires a beneficial aspect that seems to be almost wholly independent of its sacrificial quality. In certain societies the aristocrats, and even the artisans, have recourse to ritual incest, more or less furtively, to bring them good luck before a difficult or hazardous undertaking. The rites performed during the enthronement of certain African monarchs or in the course of renewal ceremonies often resemble festival practices. Conversely, in festivals in which the monarch plays no direct role we encounter a substitute king—sometimes a “king of fools”—who is himself nothing more than a sacrificial victim endowed with sacral privileges; at the conclusion of the festivities, he or his representative will be sacrificed. The king’s sovereignty—real or imagined, permanent or temporary— seems to derive from an original, generative act of violence inflicted on a surrogate victim.

The function of the festival is no different from the function of other sacrificial rites. As Emile Durkheim perceived, the festival revitalizes the cultural order by reenacting its conception, reproducing an experience that is viewed as the source of health and abundance; reenacting, in fact, the moment when the fear of falling into interminable violence is most intense and the community is therefore most closely drawn together.

Primitive peoples regard their cultural tradition as a fragile and precious inheritance to be carefully nurtured and protected from any change, for change could only serve to damage it, perhaps mortally. The skepticism and resentment we moderns feel toward taboos of any kind, which feelings we tend to assume are shared by primitive peoples, play no part in their festivities. The often-cited syndrome of “release of tensions” or that much-ridden hobby-horse of psychosociologists, the “necessary outlet,” has relevance only to a single aspect of the ritual process, and an exclusive emphasis on these syndromes distorts the original spirit of the ritual.

Festivals are based on the assumption that there is a direct link between the sacrificial crisis and its resolution. The crisis is inseparable from its happy ending and becomes itself a cause for jubilation. But this interpretation is not the only one possible. As we have already noticed in the case of royal incest, religious thinking on the relationship between the crisis and its conclusion can result in two divergent viewpoints. Either it is the continuity between crisis and conclusion that strikes the imagination, or it is the rupture; in each instance, the resulting interpretation must be both partially true and partially false. Yet religious thinking tends to adopt one or another of the two solutions and cling to it for dear life—even if, at the outset, the choice could easily have gone the other way.

It can be assumed almost a priori that certain societies will opt for the second solution, the one that emphasizes the rupture between the crisis and the founding violence. In such cases another kind of festival will arise, one that, in comparison with the festival we have been describing, might perhaps be called an antifestival. The rites of sacrificial expulsion are not preceded by a period of frenzied anarchy, but by an extreme austerity and an increased rigor in the observance of all interdicts. Extraordinary pre- cautions are taken to prevent the community from falling prey once again to reciprocal violence.

In fact, both solutions can be observed. In some societies we encounter ritual occasions that resemble festivals—there is the same periodicity, the same interruption of normal activities, the same rites of sacrificial expulsion—and yet are so very different from festivals that they constitute a vexing problem for ethnologists, one similar to the enigma of royal incest, which is accepted in some societies and forbidden in others. Far from being temporarily suspended, in the antifestival all cultural prohibitions are strongly reinforced.

The rites of the Swazi Incwala correspond closely to the definition of the antifestival. Throughout the period of observance all sexual activity, including the most legitimate, is forbidden. Sleeping late in the morning is regarded as a crime, and physical contact between individuals is to be avoided, even physical contact with one’s own body (the celebrants are not supposed to wash or to scratch themselves). A threat of imminent pollution—that is, of violence—hangs over the entire community. All singing and loud noise are prohibited. Children are scolded if they grow noisy at play.

In The Golden Bough Frazer offers a fine example of an antifestival. For several weeks in the year the Cape tribe on the Gold Coast permit no sound of tom-toms or musket fire. Public conversations are forbidden. If a dispute arises and voices are raised, the contesting parties are summoned before the chieftain, who deals out stiff fines to everyone involved. To avoid arguments over strayed livestock, all lost animals become the property of their finders, and the original owners are obliged to relinquish all claim.

It seems clear that such measures are intended to prevent the outbreak of widespread violence. Frazer offers no explanation for them, but his anthropological intuition (far superior to his theorizing) prompted him to classify these practices under the rubric of festivals. The logic of the antifestival is as strict as that of the festival. The goal is to reproduce the beneficial effects of violent unanimity while abbreviating as much as possible the terrible preliminaries—which, in the case of the antifestival, are perceived in a negative light. The longer the interval between any two purifying rituals, the greater the danger of a violent explosion. Impurities accumulate; and in the period immediately preceding the celebration of the rite, a period saturated with the memory of the sacrificial crisis, every- one moves with extreme caution. It is as if the community had suddenly become an arsenal piled high with gunpowder. The Saturnalia has been transformed into its opposite, the feast changed to a fast; but the purpose of the ritual remains the same.

In addition to the festival and the antifestival, one finds, as might have been expected, “mixed” ceremonies resulting from a more complex, more nuanced concept of the relationship between the crisis and the restoration of order, a concept that takes into account both the continuity and the discontinuity between them. In some instances at least these variations can be seen as a late development, resulting from the sheer remoteness of the original violence; that is, from the cumulative effect of mythological elaboration.

All too often we go astray when examining the nature of the festival and allow our attention to be diverted to secondary aspects. Under these circumstances the events hidden behind the rite become increasingly inaccessible, and the rite’s unity of purpose splinters into many incompatible segments. At the very point when the religious aspects of the rite have begun to reflect an ignorance equal to our own, the rite suddenly appears to have a timely and original function, whereas this function is in fact belated and derivative. The asceticism and mortifications of the antifestivals seems very far removed from the kind of activity we associate with a festival. We fail to grasp that they share a common origin and that in those communities where the ritual has retained its greatest vitality they often achieve a “dialectical” equilibrium. The more the rites diverge from their true function, the more differentiated they appear and the more interesting they become to scholars, who can sort them out in different categories. Modern scholarship, notably since Frazer, is no longer unaware that certain festivals entailed human sacrifice. Nonetheless, we are still far from suspecting that the distinctive traits of this practice and their innumerable variations can be traced back directly or indirectly to a generative act of collective violence, a liberating gesture of mob anger. The origin of the festival can still be discerned, even in those instances where sacrificial immolation has been eliminated from the proceedings. The dis- appearance of the sacrificial event may lead to new rites whose sacrificial character is easily identifiable—rites of exorcism, for example. These rites occupy the same place as the vanished sacrifice, and even when they are not directly linked to sacrifice they serve the same function in the ceremonies. In short, they can be said to be a replacement for the sacrifice. What is the correct procedure for ridding a person or place of devils and evil spirits? Often it is a matter of shouting, clanging weapons or cooking vessels, and beating the air with a stick. Nothing seems more natural than to take a broom to the devil—if, that is, one is stupid enough to believe in his existence. The modern intellectual, the “liberated” Frazerian, there- fore concludes that primitives liken the spirit of evil to some great beast that takes to its heels when frightened. The rationalistic mind does not bother with customs that seem not only puerile in conception but lacking all reason.

In this case as in many others a complacent intellect and the seeming “naturalness” of the circumstances can serve to conceal their most interesting aspects. In principle, the act of exorcism is an act of violence perpetrated against the devil or his associates. In some festivals this terminal violence is preceded by mock combat between the exorcists themselves. We recognize here a pattern repeated in many sacrificial rites: the actual immolation is preceded by ritual disputes between the sacrificers, in which the violence is to some extent simulated. This phenomenon of mock combat must stem in all instances from the same general source.

In an example adduced by Frazer, the young men of the village go from house to house, pausing at each to perform the rites of exorcism. The tour begins with a quarrel about which house to visit first. (As a good positivist, Frazer takes care to include even those details that do not fit neatly into his own scheme; for this reason alone he deserves our gratitude and respect.) The preliminary quarrel is a reminder of the sacrificial crisis; the sacrifice or exorcism that follows emulates the unanimous violence which is, in effect, promptly grafted onto the reciprocal violence and distinguished from it only by its miraculous results.

At the conclusion of the quarrel unanimity is achieved, and the moment has come for the surrogate victim, for the performance of the sacrifice. The object of the quarrel is ostensibly the sacrifice itself; that is, the selection of the victim. In the course of the quarrel each disputant strives to put in a final word, reducing his antagonist to silence; each wants to get in the decisive blow, the one that permits no response and that will therefore serve as a model for the rite itself.

Greek legends often contain vague reference to a sacrifice—a human sacrifice—offered by a community, city, or army to some god. The persons involved agree on the need for such a sacrifice but disagree on the choice of victim. To understand the situation the investigator must reverse the order of events. First comes the violence, spontaneous and senseless; then comes the sacrificial explanation, genuinely sacrificial in that it conceals the senseless and basically intolerable aspect of the violence. The sacrificial explanation is rooted in an act of terminal violence, violence that can only be labeled sacrificial retrospectively, because it brought all hostilities to an end. These stories may represent the minimal form of mythological fabrication. A collective murder that brings about the restoration of order imposes a kind of ritualistic framework on the savage fury of the group, all of whose members are out for one another’s blood. Murder becomes sacrifice; the angry free-for-all that preceded it is transformed into a ritual dispute over the choice of the most suitable victim, one that satisfies the piety of the faithful or has been selected by the god. In effect, the real question behind these preliminaries is, Who will kill whom? The dispute concerning the first dwelling to be exorcised screens the same conflictual process, leading to the violent resolution of the crisis. Exorcism represents the last chain in a series of reprisals.

Having succumbed to reciprocal violence, the celebrants as a group vent their fury on the empty air. We see here manifested a truth common to all rites, but never more clearly displayed than in this type of exorcism. Ritual violence awakens no hostility, confronts no antagonist; as long as their blows are directed as a group against an insubstantial presence, which for excellent reasons shows no tendency to retaliate in kind, the exorcists are not likely to resume their quarrel. And here the rite reveals its origin and function. The unanimity attained through the intervention of the surrogate victim must not be lost. The community stands united before the onslaught of “evil spirits” and remains faithful to its vow to reject mutual hostility. The rite reaffirms and reinforces this resolution. And religious thought returns again and again to that supreme wonder, that last word of violence, which is all the more precious for being pronounced so late in the day. Sacrifice is the boon worthy above all others of being preserved, celebrated and memorialized, reiterated and reenacted in a thousand different forms, for it alone can prevent transcendental violence from turning back into reciprocal violence, the violence that really hurts, setting man against man and threatening the total destruction of the community.