Sacrifice and the problem of beginning: meditations from Sakalava mythopraxis


Department of Anthropology, London School of Economics and Political Science, Houghton Street, London WC2A 2AE, UK. (January-June); Department of Anthropology, University of Toronto, 100 St George Street, Toronto ON M5S 3G3, Canada (July-December).


This article addresses the general problem of beginning in human thought and action. It argues for complementing the emphasis on transition in the analysis of ritual with attention to beginning and for supplementing the relative passivity of liminality with the resoluteness of initiating action, while also attending to both the transitive and intransitive aspects of beginning itself. Drawing from representations of the foundation of a Sakalava monarchy in Madagascar, the article presents sacrifice as an exemplary form of beginning. Describing sacrifice in this manner obviates the need for any theory of sacrifice while offering new insights on the gift, ethical personhood, and the temporality of tradition.

Imagine you are standing at the podium about to deliver a public lecture. Your voice cuts into the silence and you begin. No moment is so sheer, so existentially chilling. Having begun, the lecture will run its course; practice follows its routines, speakers their texts. This is not a metaphor of cast dice. Inadvertent circumstance may disrupt the speaking; practice reshapes itself to meet contingency; writing itself takes turns the author does not expect at the outset. But no event is so autonomous, so simply itself as beginning.

Of course, a paradox about beginnings is that when we stop in the middle of things to think about them, they are always already in the past. So, whether by means of myth or by memory, we are often trying to understand a present open to the future that is itself already located imaginatively in the past.

‘The problem of beginning’ wrote Edward Said, ‘is the beginning of the problem’ (1975: 42).1 Nowhere in theory is this problem fully resolved. If beginning was unrecognizable to structuralism, it has fared little better in practice theory's revisions of that doctrine. Beginning is an embarrassment for social theory – and yet it is a regular part of our experience as we gird our loins, face the day, start our computers, embark on a journey, a book, or a love affair; or as we collectively set our sights, step forward to battle or toward a new world – or imagine our predecessors having done so.

Beginning is a problem not only for theory, but for practice itself. This essay addresses the problem of beginning in Sakalava practice and the enactment of beginning in Sakalava historical consciousness, how descendants of the polity of Boina in northwest Madagascar have imagined the beginning of their monarchy (events which took place around 1700 AD) and how that beginning forms a kind of prototype for other beginnings, how beginning again is always a beginning from the beginning. Following the Sakalava lead, my article is also about sacrifice. That is to say, it is about sacrifice as a kind of beginning – at once a means, an act, and a sign of beginning – and possibly about beginning as a kind of sacrifice.

Sakalava beginnings

Scene one: the curation of the knife

The village of Bemilolo sits in a pastoral setting of oxbow lakes along the Betsiboka River, a place of abundant rice fields, water fowl, and fish, about a half day's travel by bush taxi from the port of Mahajanga. Among the houses is a fenced-off enclosure containing a small building. This belongs to Ndramandikavavy, the founding queen of Boina. Her name means ‘Noble-lady-who-surpasses-all-women’. The shrine contains her material artefacts, notably a knife (viarara). The shrine is cared for by local inhabitants who claim descent from the Queen. They form a clan, the Tsiarana (‘Peerless’), whose duty it is to attend to their ancestress as well as to matters of cattle sacrifice in royal ritual. On the occasion of the annual ceremony honouring the Queen many other Tsiarana gather in the village, along with members of the royal family (mpanjaka), who also descend from her (through a different offspring than the Tsiarana). Ndramandikavavy herself speaks from one of her spirit mediums seated on a raised platform outside the enclosure. The knife is deployed whenever there is a burial service at the royal cemetery nearby (at Betsioko). Cattle sacrificed at a royal burial must have their throats cut with Ndramandikavavy's knife.

Sakalava are concerned with the enactment of fanompoa, a word that means service to royalty and, specifically, care of royal ancestors – named personages in the royal genealogy. While care is ongoing – especially taking care not to offend them – specific ceremonies are directed to the curation of the ancestors and their property – their bones and relics, tombs, houses, and artefacts (Lambek 2002). Some of the ceremonies, like the Queen's, are annual events. The largest one, the Great Service (Fanompoa Be), held in a temple on the outskirts of Mahajanga, is a new year's ritual. And yet, as different shrines across the countryside hold their own annual services, each serves in part as a necessary beginning, a prelude, for successive ones at other locations. Thus the Great Service cannot itself begin until other services have taken place. It is as if the problem of beginning is itself continuously displaced, pushed back into a kind of infinite regress. But such deferral cannot be infinite; if there is to be a conceptualization of beginning, there must be a first beginning – or a first step in the act of beginning. How that is indicated – the problem of beginning – is the problematic of this article.

The Queen's knife also figures prominently in the Great Service, although it is not literally the same object as the one in Bemilolo but a different materialization of the same conception. It is carried at the head of the annual procession that marks the climax of the new year festival, during which the four male ancestors whose relics reside at Mahajanga are paraded around their shrine immediately after their annual bath. Each year the Queen's knife is purified along with the relics of her menfolk; each year Sakalava begin their historical journey.

Why is the knife of sacrifice such a critical element? And why does it belong to the Queen? The Queen herself is markedly absent from the main shrine, which is commonly known as the Shrine of the Four Men (Doany Efadahy). These include the Queen's husband and their joint son.2 The Queen can never appear on the shrine precincts and yet the annual service cannot take place without her permission. More than that, the Great Service begins with a sacrifice of cattle in her name.

I do not want to treat either myth or sacrifice as discrete things that can be isolated and inspected in and of themselves. Rather, I look at sacrifice as it appears in Sakalava narrative about the past, in characters instantiated in the present, and as enacted in performance. Our idea of myth – sacred narrative abstracted from its context – is itself secular (cf. Asad 2003) and therefore cannot give us much sense of what it might be like in a more holistic universe. For Sakalava, myth is living or lived, quite literally. Ancestral personages are invoked in prayer, accounted for in the observance of taboos (and manifest in symptoms of their violation), and encountered in embodied performances of spirit mediums. Narrative has no intrinsic priority. Conversely, ‘sacrifice’ is not only an act carried out, but an idea or image explored in narrative and other kinds of performance. My narrative, now well begun, is therefore about neither sacrifice nor myth per se, but rather about a set of Sakalava images and practices from which both sacrifice and myth could be precipitated as analytic constructs – by an earlier generation of anthropologists, perhaps. It is better to say succinctly that my article is about beginnings.

In The weight of the past: living with history in Mahajanga, Madagascar (Lambek 2002), I portray the texture of historical experience among present-day Sakalava – how the past structures and sanctifies the reception of new events in the present (cf. Sahlins 1985) as well as how the past is continuously realized – made real – by the work of the present. But it could be said that I took beginnings for granted, colluding in the common anthropological wisdom that the explanation of beginnings lies beyond our reach of competence. Yet if we cannot say why things begin, perhaps we can still say how they begin, or what beginnings are, what it means to begin. This is not an offer to replace one myth with another, as Freud (1958 [1913]) does in Totem and taboo, for example. The question is less about origins, or the problem of discovering what happened, than it is understanding beginning as a problem.

I distinguish between beginning and origin. We often think of myth as stories of creation; myths of origin are about the emergence of something from nothing, of order from chaos, of absence replaced by presence. In contrast, beginnings occur in the midst of life, rarely as a very first beginning, but each time a new beginning. Beginnings emerge against what precedes them.

Beginnings are located in time and in society, whereas origins may be situated in a pre-temporal or pre-historical horizon. Evans-Pritchard, for example, distinguishes Nuer accounts of past events which ‘have therefore a position in structure’ from ‘the horizon of pure myth ... One mythological event did not precede another, for myths explain customs of general social significance rather than the interrelations of particular segments and are, therefore, not structurally stratified’ (1940: 108). Beginnings are acts and therefore require specific actors or agents. Origins are events and imply extra-human forces. Beginnings entail human intention. Beginnings also imply a narrative that leads forward from their present. An origin is the culmination of its narrative.

The acts of Sakalava narrative belong to a human age. Indeed, in northwest Madagascar there are no prominent first acts of creation and it is said that people came to the island rather than originating there. Key acts are ones of planting, burying, and setting down roots rather than of coming up autochthonously through the soil.

Edward Said, to whom I am much indebted, argues that ‘beginning is ... an activity which ultimately implies return and repetition rather than simple linear accomplishment ... beginning and beginning-again are historical whereas origins are divine’ (1975: xiii). He says further that

the designation of a beginning generally involves also the designation of a consequent intention ... when we point to the beginning of a novel, for example, we mean that from that beginning in principle follows this novel ... The beginning, then, is the first step in the intentional production of meaning (Said 1975: 5, his emphasis).

Thus, beginnings are active, humanly made, not naturally or divinely given.3 Along the path through the forest, they are the first step into a clearing, the first stroke of the axe to mark the trail.

Scene two: from the Schwartzwald to Bourg la Reine and thence to Ithaca: the rites, wrongs, and rings of passage4

The ring of the axe is no idle metaphor. Chopping, cutting, spilling, killing – in a word, sacrifice – is one of the ways that we mark time, one of the clearest acts of beginning. In beginnings, the flux of homogenous time is overcome, the stream is parted into a before, a now, and a henceforward; into a past, present, and future. Beginnings signal the self-conscious insertion or intervention of the human act and the recognition of time's place in human experience. Simultaneously, they make spatial discriminations as beginning is materialized – in the chopping of a tree or the spilling of blood – and localized, marking a place in the extension of space, a point, a path, and a direction.

In The rites of passage, cutting is described as a symbolic act of separation. Arnold Van Gennep (1960 [1908]) understands ritual as both temporally constituted and temporally constitutive, enacted and enacting, moving its subjects along trajectories of successive social statuses. He pictures the life cycle as the successive habitation of a series of rooms. The passage between them is enacted through ritual, as though one paused at each threshold to rearrange one's habit and demeanour – and took a deep breath before stepping through.

Victor Turner developed the model in a series of brilliant analyses (e.g. 1967; 1969), but with hindsight it may be fair to say that this elaboration became a kind of appropriation. Where Van Gennep distinguishes three phases of ritually produced movement – separation, transition, and incorporation – Turner famously focuses on the middle, liminal phase. Turner's exploration of transition is certainly true to an aspect of ritual and temporality, but as a result we have come to think of the liminal phase as primary, and the other phases as mere brackets or containers of the precious liminal cargo. The power of this model of transition entails two emphases that I wish to counter. First, the action of separation, the bracing of the shoulders and decisive first step, is displaced or over-weighted by the passion characteristic of liminal beings and their rebirth. Second, the intransitive quality of the act is overshadowed by the transitive.

Consider the term ‘initiation’. Whereas initiation as the prototypic rite of passage has come to connote some kind of drawn-out process, the verb ‘to initiate’ more fundamentally means simply ‘to begin’ or ‘set going’.5 Think also of the meanings of ‘initial’ and ‘initiative’.

Furthermore, in the sense of ‘beginning’, the verb ‘to initiate’ does not have the strong transitive sense of initiating someone, but only a weaker transitive sense of initiating a course of action. The verb ‘to begin’ can also be used in the intransitive sense of simply beginning. Here the end or object is intrinsic to the means and thus the action is understood as valuable in itself rather than instrumental.

In other words, the model of ritual as transition (and essentially transitive) is ready for another look. Viewed as wholes, rites de passage might sometimes be better characterized by a shift of emphasis as rites of beginning. Indeed, the first phase, the new departure, is critical insofar as it anticipates the point of the entire ritual sequence. As a codification and realization of intention and commitment, it subsumes what follows. Enactment of the remaining phases of the ritual is but fulfilment of what is established in setting forth. Initiation means literally ‘beginning’, and then ‘admission’. Transformation’ and ‘induction’ are secondary meanings, reinforced by the tripartite Van Gennep-Turner model.6 Turner's elaboration of the middle phase obviates understandings of sheer beginning or final end.7

Liminality is a spatial metaphor of waiting and passivity; I would supplement it with a temporal phenomenology of action. Rituals are frequently culminations of what is past and anticipations of what will follow. Often they are both simultaneously. Sacrifice is both a passionate culmination for the victim and a significant initiative by the person who offers it. It draws a line in blood between ‘before’ and ‘after’. Once you have killed something there is literally ‘no going back’ for either victim or killer. Sacrifice is thus a materialization of intention and a consummation of resolution.

How much more powerful all this is when victim and sacrificer are one and the same person.

Scene three: the shrine of the four men

The Great Service sets the Sakalava community on a fresh start every year. But critical to the success of the ceremony of new beginning is its own beginning. In Sakalava fashion, the key acts happen backstage. Each year, ancestral approval for the Service must be sought and confirmed; rising in the bodies of spirit mediums, royal ancestors ‘open the gate’ or ‘clear the path’ to move ahead. Thus intentionality is marked from the outset – both the intention of the living producers of the ritual, who state their aims very clearly in addressing the ancestors, and the intention of the ancestors, who explicitly acquiesce and join in the wish to see the ceremony successfully completed (vita tsara).

A week before the public beginning of the new year's festival, and consequent to her acquiescence, descendants of the Queen gather at the shrine to make a bovine sacrifice on her behalf.8 They signify their commitment by offering the beast – and the animal itself, like all Sakalava sacrificial beasts, must indicate acquiescence by not bellowing or struggling when it is thrown.

The skin of the dead beast is then stretched to dry and used to refurbish the shrine's two sacred drums (mañandria be, ‘great ennoblers’). The vitality inherent in the life of the beast and the intentionality of the offering are transposed into the vibrancy of the drums (cf. Ruel 1990). There is a literature associating percussion with transition (Knauft 1979; Needham 1967), but I emphasize here its significance as beginning and enactment. Drumming is both energetic and unambiguous. The drummer beats the drum; each movement of the arm is specific and deliberate. The energy of the arm is transferred to and enlarged by the vibrations of the drumskin. The rhythm of the drum continues to emit the intention of the sacrifice. The passivity of the beast silently submitting to the knife is transcended by dynamism, noise, and vitality. The drums are a medium for the active force of the victim.

In contemporary Sakalava sacrifice, ‘value’ moves from the labour of raising the animal through its market price; but also from the intention of the donor through the acquiescence of the beast; and from the throb of life through the spilled red blood to the pulsing of the drums made from the victim's skin; then from the drums that excite and incite to the acts of those who respond to its rhythm, in anticipation, in attention, in dance, and in labour and loyalty, in virtuous acts of commitment to the monarchy, obeisance to the ancestors, and acknowledgement of hierarchy, hence in granting power to Ndramandikavavy. This is a version of the process of the ‘dissolution of death ... into authority’ analysed by Maurice Bloch (1989) for the neighbouring Merina that combines mystification with morality.

The initial sacrifice is described as an offering on behalf of the Queen, but it is also a repetition of her original sacrifice. Those who perform the offering stand in for their ancestor who made the first sacrifice. Just as Ndramandikavavy's original act began the polity, so does it begin the new year, which is also a celebration and re-enactment of the polity and includes offerings to the original monarchs by their subjects. Because many of the offerings end up in the hands of the living monarch and because acts of work, offering, and obeisance form the substance of the polity, the celebration is a literal as well as symbolic enactment. But my point here is simply that both the historical polity and the annual enactment begin with the Queen's sacrifice. Indeed, just as the liturgical cycle enacts the beginning of the polity, so the beginning of the polity is equally the initiation of the liturgical order.

However, the Queen's original offering was not composed of cattle. This is a point to which we shall return.

Sakalava emphasize and elaborate the beginning of their Service with various deliberate steps before the public display, putting the weight, one could say, more on the intention than on the accomplishment (which is a kind of anti-climax, or at least it was to me). The sacrifice of the Queen's beast forms a preface to the ensuing festival. It is relatively private (concerning only her own direct descendants, mediums, and shrine workers), whereas the festival is public, incorporating thousands of people through numerous forms of attachment. The sacrifice is also quite discreet; most people do not know the connection between the Queen and the drums. But it is absolutely critical. The festival cannot take place without the Queen's acquiescence, as provided in her speech through a medium and performatively accomplished in the sacrifice. Moreover, the drums are one of the most important heirlooms and signs of royal legitimacy; they must be renewed each year, their skins must be those of the Queen's cattle, and their rhythm must punctuate the ensuing events.

Not only the rehearsal of a genealogical charter, these acts are also conditions that enable the felicitous performance and the performative effects of the subsequent main ritual events. But they are also performative events in their own right. The Great Service is front-loaded with a series of such events that establish the intention and willingness to proceed. It is as though Sakalava are preoccupied with the sort of infinite regress of intentionality – the ghost in the machine – described by Gilbert Ryle (1949), albeit in the Sakalava case this is located in the external, inter-personal world of the polity, rather than the internal, private one of the mind. Ryle, of course, saw regress as the product of a category mistake. But it may well be that this category mistake is not so easy for humans to avoid, that mind and body and the mind/body problem are part of the human condition, that the Durkheimian problem of social solidarity and moral commitment is formally similar, and that the problem of beginning is one of its manifestations. The series of ritual acts productive and demonstrative of intention and commitment, and each a necessary condition for the succeeding one, are Sakalava expressions of the problem.

Scene four: Morningside Heights (Manhattan)

Some years before he became famous as the author of Orientalism, Edward Said wrote Beginnings, subtitled Intention and Method, from which I have already quoted. Here Said is interested in creative production. Every work of art has its beginning, its conception, and the way that conception is activated or materialized in a first sentence, note, or brushstroke.

Said does not explicitly consider beginnings characteristic of social practice, and hence he ignores the field of ritual. And yet his account is extremely relevant for social practice. ‘What sort of action,’ he asks,

transpires at the beginning? How can we, while necessarily submitting to the incessant flux of experience, insert (as we do) our reflections on beginning(s) into that flux? Is the beginning simply an artifice, a disguise that defies the perpetual trap of forced continuity? Or does it admit of a meaning and a possibility that are genuinely capable of realization? (1975: 43).

Said speaks of the necessity ‘to acknowledge the mind as providing self-concerned glosses on itself over time, the mind as comprising its own philosophical anthropology’ (1975: 43); surely these glosses are found in ritual no less than in literature. Indeed, Said argues that beginnings are necessary. ‘Without at least a sense of a beginning, nothing can really be done, much less ended ... A beginning gives us the chance to do work that compensates us for the tumbling disorder of brute reality that will not settle down’ (1975: 49-50).9

Said makes a careful distinction between transitive and intransitive beginning. These are ‘really two sides of the same coin. One, which I call temporal and transitive, foresees a continuity that flows from it. This kind of beginning is suited for work, for polemic, for discovery.’ However,

In attempting to push oneself further and further back to what is only a beginning, a point that is stripped of every use but its categorization in the mind as beginning, one is caught in a tautological circuit of beginnings about to begin. This is the other kind of beginning, the one I called intransitive and conceptual. It is very much a creature of the mind, very much a bristling paradox, yet also very much a figure of thought that draws special attention to itself (1975: 76-7).

In sum,

The point of departure ... has two aspects that animate one another. One leads to the project being realized: this is the transitive aspect of the beginning – that is, beginning with (or for) an anticipated end, or at least expected continuity. The other aspect retains for the beginning its identity as radical starting point: the intransitive and conceptual aspect, that which has no object but its own constant clarification ... These two sides of the starting point entail two styles of thought, and of imagination, one projective and descriptive, the other tautological and endlessly self-mimetic. The transitive mode is always hungering ... for an object it can never fully catch up with in either space or time. The intransitive ... can never have enough of itself (1975: 72-3).

Scene five: the countryside west of Mahajanga, circa 1700

The Sakalava have reached the country of Boina where they would like to settle and rule. An oracle advises the King, ‘In order to succeed, you must sacrifice what is dearest to you (raha tsifoinao).’ The King offers the best cattle from his herds, but they are not acceptable. He offers gold, and then fine Chinese porcelain, but they too are insufficient. Finally his wife announces, ‘I am what is dearest to you.’ And so she sacrifices herself.

This is an extreme distillation of more elaborated and divergent narratives of the founding of the Sakalava polity10 but sufficient for my purpose here.

We can see now that if Ndramandikavavy is absent from the main shrine it is precisely because its very existence is predicated on her death. Moreover, it is not merely the sacrificers who take the Queen's place in the annual ceremony, but the cattle themselves. They are killed not only on the Queen's behalf, or in her honour, but with her knife and in her stead. The throat that this knife first cut was the Queen's. Each sacrifice is a repetition and recognition of her original act; each victim is identified with her. It is her skin and her life that animate and empower the drums. It is her intentionality that produces and permeates the ensuing ceremony and that infuses the resolution of the celebrants as loyal subjects.

As a character in the Sakalava poiesis of history (Lambek 2002), Ndramandikavavy symbolizes determination. It is her determination that is harnessed to ensure and energize the enactment of the Great Service. It is her determination that is diffused to her descendants and her subjects – the general population (vahoaka) who have produced and renewed the polity through their own acts of sacrifice, service, and celebration over the course of three hundred years. Her resolve is evoked in her every appearance and lends itself to every act carried out with her knife and in her name.

‘Determination’ is an interesting word, seemingly pulled between free agency and inevitable coercion. Ndramandikavavy's sacrifice is deeply intentional, yet to call it choice trivializes it, as though Abraham or Isaac had a choice. To draw upon Said again, it is

a necessary certainty, a genetic optimism, that continuity is possible as intended by the act of beginning ... Consciousness of a starting point, from the vantage point of the continuity that succeeds it, is seen to be consciousness of a direction in which it is humanly possible to move (as well as a trust in continuity)’ (1975: 47-8, original emphasis).

In fact, Ndramandikavavy extracted a promise from her husband as a condition of her sacrifice, namely that only their joint descendants would rule; his continuity would be equally hers. In its transitive aspect the beginning thus conforms to Malinowski's notion of a charter in that it establishes legitimacy to rule and both specifies and restricts succession to office. Yet it differs from many genealogical charter myths in validating an ancestral couple. The Queen effectively cuts off all the King's collaterals and other descendants. The King and Queen's respective children by other partners form two separate clans, closely associated with royalty but distinct from it. Descendants from the Queen alone are Tsiarana – the Peerless – who maintain and wield the sacrificial knife.11

In comparison with the Judaeo-Christian-Islamic tradition, and especially the story of Abraham and Isaac, the following points are salient: (a) the victim is a woman; (b) the victim is said to ‘kill herself’ (mamon teña); (c) the victim is not an offspring but a spouse and parent; (d) the sacrifice is fully carried out, there is a human death; and (e) there is no direct recipient, no divine being to whom or on whose demand or behalf the sacrifice is given. There is much to say about gender and about mothers and sons, but having begun an essay on beginnings, I must stay on track.12 Germane to the present analysis is the determined yet voluntary quality of the sacrifice, the completion and finality of the act, and the absence of a divine arbiter or recipient. Thus, whereas biblical sacrifice seals God's promise to look after His people, Sakalava sacrifice seals the people's promise to care for Ndramandikavavy, her son, and his descendants.13

Sacrifice as exemplary beginning

There are many theories of sacrifice. I do not intend to review them or to adjudicate among them. Nor do I propose to add another theory; I merely provide a redescription. I want to show that sacrifice is an exemplary mode of beginning and hence that beginning affords one way to interpret sacrifice.14 This is certainly not the most intuitive interpretation, since – for the person, animal, or object destroyed – sacrifice would appear to form most saliently an ending. In a sense, each new beginning signals simultaneously the end of what it departs from and perhaps the death of alternatives not taken. But I suggest more strongly that endings form and substantiate beginnings. I do not suggest that this is the only way to understand sacrifice, but I admit that one way to read my argument is that no theory of sacrifice – no explanation that draws upon forces external to the act itself – is necessary.15

Sacrifice as a literal act

By ‘literal’, I mean here univocal, exact, unmediated, and unambiguous, hence where certainty can be distinguished from uncertainty. The body is often the ground for the literal.16 The body is literally grounded in burial, and authors such as Astuti (1995), Bloch (1971), and Middleton (1999) have noted how in Madagascar this – finally – establishes exclusive social identity. Sacrifice is one of the most literal acts possible, both as an action and in terms of its end product, that is, a dead body.17

To be sure, the sacrifice of an ox for a human, or a cucumber for an ox, can be understood metaphorically. But these are merely replacements of one object of sacrifice for another, not metaphors for other, more consequential kinds of acts; at bottom, they are still sacrifice. For the ox, his life is no metaphor. Cut flowers die; liquid cannot be put back into its container; like Humpty Dumpty, a sliced cucumber cannot be put back together again. There may be some ambiguity as to whether a given act of killing is sacrificial or what that might mean (Ruel 1990), but killing in and of itself is always a literal act. For the victim there is no return, nor can the killer withdraw her act.

In this sense, killing is like ritual. One of the things about rituals is that you cannot take them back in protest that you did not mean it. Once done they cannot be undone, at least not without going to the trouble of holding another ritual of a different kind. Sacrifice ensures and epitomizes this quality of ritual action.

Cutting throats and spilling blood are literal acts. The object can be metaphorically displaced – from royal self to human subject, from animal to vegetable – but the act of sacrifice is no metaphor.

The gift and its limits

It is often argued, following Mauss, that there can be no ‘pure’ gift, since the obligation to give is always met with the obligations to receive and to return a gift. The recipient who does not offer a return is beholden to the donor and the donor achieves a higher moral status in that respect. Hence no gift can be said to be characterized by pure disinterest.

What of sacrifice? In Ndramandikavavy's case, while her sacrifice was noble and courageous, no Sakalava would suggest it was undertaken without self-interest. The Queen struck a bargain; she asked for something in return. However, what her sacrifice shares with the idea of the pure gift is the impossibility of a full return. Here lies the source of her power.

Why can there be no equivalent return? To answer that, l must address the question of who her recipients are and how I know her gift was not disinterested.

There is no third party, no deity to receive or be honoured by Ndramandikavavy's sacrifice. The instigator is a diviner and he is responding to the workings of a neutral, amoral destiny. The immediate beneficiary is posited as the King, her husband, but her death is simultaneously his loss. Indeed it is phrased in those very words as that with which he could not bear to part, raha tsifoinao. Yet Ndramandikavavy's quick response actually turns things in a more interesting way; in sacrificing herself, she seizes his role as the offerer of sacrifice and achieves the perdurance of her own name and rule. What the King loses is not just his wife but his agency (what is dearest to him indeed).

By establishing his exclusive right to succession, the Queen makes their son the direct beneficiary. It is only towards him that her sacrifice might be said to be disinterested and yet it is precisely because of her identification with her son that it is most self-interested. It is his – and therefore her – descendants who succeed him. His paternal half-siblings – the other children of her husband – are specifically excluded and disinherited.

More generally, the beneficiaries are all subsequent members of the kingdom of Boina. This has a sharp ideological angle: the Queen's subjects are expected to honour her for effecting the conditions under which her descendants rule over them, as well as identifying with her as an ideal subject.

When Ndramandikavavy rises in the bodies of her mediums it is evident that she went willingly, but not selflessly. She retains great attachment to the things of this world, to her son in particular. And she is especially rude to her husband. He shows his wife great affection and puts his arm around her but she soon gets impatient. After a few minutes, she says brusquely, ‘What are you still doing around here? No one needs you any longer!’ And so the conqueror and first king of Boina slips meekly away from his shrill wife and out of the body of his medium. He is, in her eyes, irrelevant. He is eternally beholden. He cannot return her gift.

If the Queen gave up her life on behalf of her husband, it is clear that he in turn has given over his authority to her. In response to supplicants, the King speaks in platitudes and is generally amenable to whatever is proposed; the Queen is forthright, assertive, and tough, expecting people to uphold the standards of the polity for which she gave her life. She inspires respect and fear not only in her husband but also in her human subjects. She is known to be harsh (mashiaka) and her devotees are afraid of her wrath.

Ndramandikavavy's sacrifice is a gift that must be received yet cannot be returned in full and hence cancelled. The debt to her is in a sense what keeps the polity running. Generation after generation must bow to her authority, respond to the interpellation of the drum; each royal succession must ensure that the promise to her remains fulfilled. Acknowledgement is mandated, but no return could ever be sufficient to cancel the debt.

The Queen's gift cannot be cancelled not only because she courageously gave her own life – and no larger gift could be imagined – or because the existence and persistence of the political community are the outcomes of her act; not only for the obvious reason that she is no longer fully present to benefit from a return gift; but also because what she offered was manifestly a first gift. Any and all returns – even were they to be equivalent – are quite literally secondary. Only one act in a sequence can be first.

Insofar as her act is original, Ndramandikavavy points to a problematic of the gift literature left largely unaddressed by Mauss and his successors. These are the questions: What constitutes a first gift? Which donor stands out as the original source of the hau to which the gift must return? Moreover, if every gift engenders a return, how do we know that any given gift is not always already a return? How is a gift marked as first? Where and how does the chain of reciprocity begin? How can one break it and start anew?18 We are back to the question of beginnings.

My point is that sacrifice is one way to mark a gift as ‘first’. Indeed, this may be implicit in our understanding of the distinction between sacrifice and gift.

Blood sacrifice or self-sacrifice makes an especially good beginning, a first gift, because it is a gift that does not – indeed, cannot – return to its origins. It is the gift that cannot be returned, or cancelled, or withdrawn. The only thing to do with it is to honour it.

Words and deeds

Ndramandikavavy's sacrifice is not a gift to a transcendental being, but dictated by fate. It is in this respect unmotivated and unreceived. Thus I do not wish to place undue emphasis on sociality, communion, commensality, appeasement, or supplication. A sacrifice is purely what it is and not something else.19

Goethe's Faust rewrites the New Testament to begin, ‘First was the act!’20 Sacrifice (cutting, killing) is manifestly neither word, nor object, but act. What distinguishes sacrifice from simple giving or killing is that it is explicitly performative, bringing into being a new conventional or moral state.

Performative acts – in Austin's sense – are frequently beginnings and they are spoken in the first person. They can be understood as public expressions and objectifications of resolution. Austin provides examples like: ‘I... take this woman to be my lawful wedded wife ...; I name this ship the Queen Elizabeth ...; I give and bequeath my watch ...; I bet you sixpence ...’ (1962: 5). Each of these casts intentionality forward – as vow, promise, contract – most generally, as commitment. Henceforward, the state of affairs, the name, the relationship, the expectation, shall be understood in the terms established in the performative act. Thus, the state of marriage is brought into being in the enactment of the wedding service. As Rappaport notes (1999: 133), acts subsequently committed by persons subject to the conditions initiated by means of the performative act are to be judged with respect to those conditions.

In Muslim societies a simple and ubiquitous act of beginning is found in the utterance of the ‘bismillah’. Trobriand spells are also beginnings and performative acts in this sense. As Malinowski portrays it, Trobrianders utter a spell at each new phase of the kula, from felling the tree to make a canoe (1922: 126-7) to meeting the kula partners (1992: 334-49). The utterance of each spell is the beginning of each new phase. The spell can be understood to mark a reorientation in the stream of practice and to begin what lies ahead. The kula expedition as an ethical enterprise is thereby constituted as a series of marked intentions, each initiated by a performative utterance that constitutes the quality of the time/space and action of what is to follow (cf. Munn 1986; Tambiah 1973).

Austin's insight is that such utterances are simultaneously acts. Rappaport describes ritual as comprised of both acts and utterances, but it could be said that he also privileges utterances. Thus he singles out ultimate sacred postulates, phrases like the bismillah which are unquestionable and invariant, and which sanctify the acts and more specific utterances they accompany, as when a political pronouncement includes the phrase ‘In the name of God ...’. Yet if utterances are acts, certain highly conventionalized acts may be considered simultaneously utterances (Lambek 1992).21 We can speak of sacred acts no less than sacred postulates, and sacrifice is surely among them.

Of course, sacrifice is an act generally composed of both word and deed, but it is one in which the deed speaks more powerfully than the word.

Just as Rappaport describes a hierarchy of postulates with respect to their sanctity (1999: 314), so, I suggest tentatively, we may think of acts.22 This hierarchy is one of unquestionableness, invariance, precedence, and inclusiveness; one could add irreversibility. As a pure beginning, the first sacrifice sanctifies and establishes the condition for the acts that follow. Indeed, it sanctifies the entire subsequent history of the kingdom, for so long as it is respected. It also establishes the grounds of value or meaning, providing the standard against which other acts can be understood. Subsequent sacrifices of lesser value – ranging from human subjects, through cattle, down to a few coins or drops of rum – can only approximate, rehearse, and evoke the ultimate sacrifice. But like it, they are performative acts that sanctify the conventional and moral states they initiate.

My point is not that in the beginning lies the deed, or lies the deed alone, but rather that in the performative deed lies a beginning.


Heidegger has written that ‘to the anticipation which goes with resoluteness, there belongs a present in accordance with which a resolution discloses the situation’ (quoted by G. Steiner 1978: 110). The question is, how is resolve disclosed or realized – made real – both for actors and for their public? The irreversibility of blood sacrifice is certainly one way.

A blood sacrifice is an execution. According to the dictionary, to ‘execute’ means:

1. Carry (plan, command, law, will, judicial sentence) into effect; perform (action, operation, etc.); make (legal instrument) valid by signing, sealing, etc.; discharge (office, function). 2. Carry out design for (product of art or skill); perform (musical composition). 3. Inflict capital punishment on (Oxford Illustrated Dictionary 1975).

To execute a living being in sacrifice is simultaneously to execute a plan or course of action.23 In both Malagasy (where the relevant verb is manapaka) and English we see how central killing is to the idea of carrying out a definitive and irreversible social act.24

Outright killing is not the only form of executive blood-letting. Scarification is one example; defloration another. In Mayotte a woman's relinquishment of her virginity is marked. Done in the wrong context, it is simply an ending and thus destructive. Done in the right context, it is a beginning, the first moment of adult sexuality and the consummation – as we say – of marriage. The act of submitting to defloration is sacrifice in this sense, a formal, decisive, and irreversible execution and a gift that can never be fully returned. It is no coincidence that it is likened to the act of male circumcision (Lambek 1983).25

Action and passion

Rituals are decisive deeds, yet they are acts in which people submit to something larger than themselves. In this respect, as Heidegger would put it, ‘man is not the opener of truth ... but the “opening for it” ’ (G. Steiner 1978: 115). Rituals thereby combine action and passion in a heightened manner. These two dimensions correspond respectively to Said's transitive and intransitive modes of beginning.

Whenever I mention Heidegger I am skating on very thin ice, and you may already have concluded that I have fallen in. But I cannot resist allusion to Heidegger's concept of Gelassenheit, releasement or letting go, which is not simple ‘inaction or passivity’. As Dallmayr notes,

an engagement in being ... challenges or decenters the customary focus of action theory on desire, will, or deliberate intentionality ... instead, the accent is shifted to ontological participation in which the actor is released at least partially from the dictates of an instrumental pursuit of objectives. This shift ... does not ... remove moral-political responsibility (1993: 58).

Sacrifice, especially self-sacrifice, exhibits this inextricable connection of action and passion or ‘resoluteness of released engagement’ (Dallmayr 1993: 58-9). It is not dying in a simple, passive sense, or killing in a simple, active sense; not accepting the inevitability of death but, as Lienhardt (1961: 296) emphasizes for Dinka, converting it into life. Sacrifice is a negation of the act of giving birth and yet strangely like it. It is ‘giving death’, submitting to it, yet bearing it forth; like birth, an act at once transitive and intransitive. And it can provide a sharper sense of beginning even than birth.26 A birth can be undone – through death. But a death cannot be undone.

As Rappaport argues, in lending themselves to a ritual, performers inevitably accept its terms and effects (1999: 117 ff.). There is no more complete mode of participation in this sense than allowing oneself to be sacrificed. Self-sacrifice is the limit act of committing oneself to the liturgical order and hence of confirming, establishing, and realizing it through one's very being. This is equally a form of identification with it. Others then realize their own commitment through their attachment to, or identification with, the sacrificial victim or sacrificer. Instances of exemplary acceptance of this kind include Jesus on the cross, Dinka spear masters buried alive (Lienhardt 1961), and dying Hindu pilgrims who prepare their bodies as offerings to the gods on the model of Vishnu's act of creation (Parry 1994: 30-2, 188-90). As the sacrificer is identified with the sacrifice, so is the ending coextensive with the beginning and the transitive with the intransitive.

The sacrificer conjoins in her act the indexical and canonical aspects of ritual and does so more completely than can be imagined through any other means. We see in this fusion a deeper significance for sacrifice than that presupposed by communication or communion with a deity. In submitting to the liturgical order, it becomes a part of her being and she too becomes an intrinsic part of that order (Rappaport 1999: 119). Actors and action can then only be described, interpreted, and evaluated with respect to prior commitments – with respect to what has been begun.

Ndramandikavavy simultaneously acts and submits. Such action by means of submission, or subjection through action, is the hegemonic form of belonging in the tributary Sakalava polity; in this sense Ndramandikavavy is an exemplary subject and, like all Sakalava rulers, expected to be simultaneously more powerful than the average citizen but also more bound by taboo and obligation to the ancestral order, more intimately part of it. Equally, we can see in sacrifice an epitomization of the Durkheimian understanding of society in which each member must give up some individuality in order to be transformed, at a higher level, into a moral person.

Social being more generally entails a dialectic of action and passion, of taking in, doing, and becoming what is expected of us (or what is unexpected) and in taking a purchase on what we are becoming and have become and making it ours. The idea of beginning captures this double edge of social subject and agent. In beginning a new project – even as one is the agent of its production or enactment – one throws oneself into it, one subsumes oneself in the practice that it entails and one is thereby partially consumed by it. And so, if beginning is signified or realized in sacrifice, that is in part because beginning is, in a way, itself an act of sacrifice.27

Time and intention

‘We do not live “in time,” as if the latter were some independent, abstract flow external to our being. We “live time”; the two terms are inseparable.’ So says George Steiner in his explication of Heidegger's Sein und Zeit (1978: 78). If so, what other way to punctuate time than by means of death? Sacrifice executes a beginning and in producing a present realizes both the past that it displaces and rehearses and the future that it anticipates. Sacrifice founds beginnings; it is foundational.28

My argument here is less functionalist than poietic or rhetorical. Functionalism assumes the prior existence of the whole, whereas I am describing how certain kinds of wholes are brought into being. Having learned about rites of passage in school, Simon Lambek (my son) asked insightfully how sacrifice could be a ritual if it had no closure. This is exactly the crux of the problem and the basis of my critique.

Both Van Gennep and Turner understand rites of passage as transitions. I have been interested more in sheer beginnings, in beginning as an existential problem and also in how people conceptualize, announce, and enact beginning – how they in fact begin. In Said's terms, this is to shift the emphasis from transitive to intransitive acts. Understood in its intransitive aspect, beginning, in the form of sacrifice, is neither a gift, nor an instrumental act, nor a referential statement; it is simply itself. Here we approach Wittgenstein's critique of Frazer: it does not make sense to see ritual as an expression of something other than itself; it simply is what it is, a necessary part of the life of a ceremonial animal (Wittgenstein 1979 [1931]; cf. James 2003).

Many rituals have both transitive and intransitive aspects. For example, if a wedding ritual entails marrying in the transitive sense, something one does to or with a partner or to one's growing children as subjects of a ritual performed on their behalf, there are nevertheless moments in marriage festivities when the intransitive mode prevails, where the act of marrying is experienced or understood as just that, ‘getting’ married. Here is a sense of ‘the blossom breaking from the bud’ (G. Steiner 1978: 137), sheer potentiality without any object. This constitutes a ripeness, a sense of capacity rather than utility (cf. Macpherson 1973: 4-5). Such a state of being (Dasein) shares features with Turner's compelling depiction of ‘liminality’, but, understood as intransitivity, it is a different way of conceptualizing the ritual act.

Writing as a professor of literature, Said is naturally most interested in beginning as poiesis, the ‘bringing into being’ (Agamben 1999: 69) of a new cultural artefact, a work of literature or art. I have emphasized less poiesis than practice, seeing beginning as the intentional engagement in social endeavours, commitments, or games. This is partly amatter of anticipation of outcome, partly of submission to the rules, craft, or ethics that the practice entails. Aristotle distinguishes poiesis and practice, making and doing, but in beginning – in authoring a book, performing a ritual, or founding a polity, the distinction may be negligible.29

Following the Sakalava cue, I have argued that the resolution entailed in beginning is well signified in the act of sacrifice. Problems of uncertainty and alternative come into focus most clearly at beginnings. Sacrifice can – and with a stroke – end them. Sacrifice combines the ethical consequentiality of judgement with the decisiveness of choice. It thereby forms not merely a kind of ‘natural symbol’ (Douglas 1970) of beginnings but a natural means to execute them.

I argue both that the story of Ndramandikavavy addresses the problem of regress in Sakalava historicity, providing – retroactively – a founding moment and a coalescence of intention, and that the continuation of sacrifice – henceforward figured in her name, by means of her knife – continues to make this available in the ongoing life of the polity.

Rites de passage belong to long cycles of generational repetition, and that is an important part of the experience for participants. But Ndramandikavavy's act is portrayed retroactively as original, a ‘first’ act of beginning, an initiation in the literal sense of the term, a particular and deliberate intervention in history. The sacrifices that punctuate ordinary life realize, focus, and display resolution and add weight, direction, and significance to specific acts and utterances and to the projects and states of affairs they initiate, introduce, and inform. Their force is greater insofar as they draw upon the paradigmatic act of intervention, firstness, and resolve, the perduring sacred prototype or ultimate sacred act of Sakalava beginning.

Ethics and ending

Without beginnings it is hard to see how action could be conceived in ethical terms. Beginnings cast intention forward; in drawing from the past, they render valuable what follows from them and, as performative acts, they set the terms of reference against which actors, events, and practices are to be judged. Who we are as moral persons is in large part a consequence of the beginnings we have enacted or participated in. Some of these happen before we reach adult consciousness, as we are named, baptised, circumcised, and so forth. Later in life we take up charges with fuller consciousness and resolve and we indicate this to ourselves and to others, just as we realize it for ourselves – sometimes retroactively – in beginning. It is through such explicit engagements and undertakings that our specific ethical lives and characters are formed; they provide the basis for both the cultivated dispositions and the ends and means that constitute us as ethical persons. As a powerful and definitive mode of beginning, sacrifice exemplifies these ethical functions.

Once committed and remembered, sacrifice can be intrinsic to who one is, as a person or a society. Who one is is thereby understood existentially rather than essentially, through the sheer initiation of being. And yet, despite the emphasis on intention and resolution, this is hardly the continuous ‘freedom’ that Sartre advocates. That is because each act of sacrifice is simultaneously a passion; each turns us irrevocably in a certain direction, locates us on a certain path; and each invites identification and repetition. Each undertaking we initiate must be made with respect to commitments always already engaged, subject positions already assumed, each new beginning with respect to beginnings already begun.

I am describing regulation or channelling in the flow of social and existential value, the mastering of creative and ethical potential – that is, setting its direction and establishing and accepting its consequences. Of course, this does not happen equally effectively everywhere and not without considerable debate. We might contrast an ostensibly relatively orderly tradition, like that of Sakalava, with more chaotic scenes in which society appears to have lost its way. These may be characterized by epidemics of witchcraft which, by inverting or perverting the logic of sacrifice, both demonstrate and effect the collapse of ethical order (De Boeck & Plissart 2004). The rhythm of regular beginning is displaced by rupture and increasingly frantic attempts at establishing new beginnings. The place of ethics and the formation of ethical persons will be distinctive in these circumstances.

I have been using beginnings as a means to read across the tradition/modernity divide, but I should add that Said himself worked with a distinction between tradition and modernity that has quite different implications. Indeed, he sees beginnings as distinctive of modernity. He argues that contemporary writers and critics can no longer easily imagine themselves within a tradition, ‘a place in a continuity that formerly stretched forward and backward in time’ (1975: 9). After Nietzsche, a text has become ‘an invitation to unforeseen estrangements from the habitual’ (1975: 9). In a masculine idiom quite foreign to Sakalava, Said distinguishes father-son succession, ‘dynastic, bound to sources and origins, mimetic’, from that of brotherly displacements, ‘complementarity and adjacency; instead of a source we have the intentional beginning’ (1975: 66). He contrasts these in Viconian language as ‘sacred’ and ‘gentile’, respectively (1975: 13).

This distinction doubtless owes its formulation to the structuralist contrast between the syntagmatic and paradigmatic no less than to Vico. I find it over-drawn and indeed something of a rehearsal of modernity's own ideology; an ideology – dare I say it – of ‘traditionalism’ comparable to that of ‘orientalism’. I have demonstrated the relevance of Said's analysis of beginnings precisely for the succession characteristic of tradition, which is not nearly so uninformed by intention, reflection, uncertainty, contestation, or women's agency as Said would appear to imply.

Tradition is neither timeless nor mindless: that is to say, it is not without beginnings. Beginnings are not routine; at least, in Sakalava ‘service’ every effort is made not to take them for granted. Each beginning actively solicits intention and explicitly demands it. Such intentionality, as Rappaport underscores, is transformed through ritual action into commitment. A difference from the situation of the modern writer described by Said is that commitment is not simply to the work at hand, hence fragmentary and fragmenting, but to a wider order within which specific works and beginnings are construed and authorized.

If beginnings per se do not serve to distinguish between tradition and modernity, perhaps the gentiles can no longer encode them as sacrifice. Marshall Berman (1988) finds modernity's myth of beginning in Goethe's Faust. Faust's pact with Mephisto is a precise inversion of sacrifice: you reap the material benefits and postpone the ethical reckoning. It is life lived on credit, or capital.30 Yet even Faust must sign his contract with a drop of blood. ‘Blood's a very special ink, you know,’ says Mephisto.31

If endings are beginnings, so too, I have indicated, are beginnings endings of a sort. And so I end this somewhat recursive discussion with a final word from Said: ‘Beginning is a consciously intentional ... activity ... whose circumstances include a sense of loss’ (1975: 372).


  • Versions of this article have been delivered as an inaugural public lecture at the London School of Economics and Political Science, 14 February 2006 and the presidential address to the Society for the Anthropology of Religion, 8 April 2005; also to the Departments of Anthropology at Berkeley and Oxford. From many helpful comments and much hospitality at each occasion, special thanks go to Maurice Bloch, Stephan Feuchtwang, Bill Hanks, Cori Hayden, Wendy James, Simon Lambek, Harriet Lyons, Johnny Parry, Laura Rival, Anne de Sales, and Charles Stafford. Thanks also to Joshua Barker, Aram Yengoyan, David Graeber, and an anonymous reviewer, and to Glenn Bowman. Fieldwork in Madagascar, conducted in six trips of four to six weeks each between 1993 and 2001, has been generously supported by the Social Science and Humanities Research Council of Canada. Further acknowledgements can be found in Lambek (2002).

  • 1

    Here is the larger context of the phrase:
    ‘Beginning’ alternates in the mind's discriminations between thought that is beginning and thought about beginning – that is, between the status of subject and object. Paraphrasing both Hegel and Vico, we can say that formally the problem of beginning is the beginning of the problem. A beginning is a moment when the mind can start to allude to itself and to its products as a formal doctrine (Said 1975: 42).

  • 2

    Strictly, it is Doany Ndramisara Efadahy, Ndramisara's shrine of the four men; Ndramisara was diviner and architect of the polity (Lambek 2002: 93-4, 97) and sometimes identified with the Queen's brother.

  • 3

    ‘As consistently as possible I use beginning as having the more active meaning, and origin the more passive one: thus “X is the origin of Y,” while “The beginning A leads to B”’ (Said 1975: 6).

  • 4 

    References here are to dwelling places associated, respectively, with Heidegger, Van Gennep, and Turner (Ithaca, NY as well as the prototypic end of the voyage). This section owes a good deal to a conversation with Joshua Barker.

  • 5

    ‘Initiate v.t. 1. Begin, set going, originate. 2. Admit (person), esp. with introductory rites or forms (into society, office, secret, in mysteries, science, etc.)’ (Oxford Illustrated Dictionary 1975).

  • 6

    And further reinforced by the ubiquitous practices of fraternities and sororities on American campuses that served as undergraduate essay topics.

  • 7

    The relevance of both modes is amply discussed by Rappaport, to whose analysis I am much indebted. Following Rappaport (1999: 95), any ritual qua event takes the preceding (analogic) flux of uncertainty and indecision and, in a clear (digital) act, definitively transcends it.

  • 8

    I observed this event in July 1994. That year the Tsiarana pooled their resources to purchase the animal. There are no rules regarding the sex or age of the bovine offering. Compare the killing of a bull to initiate Merina rituals of blessing, an act Bloch‘hesitates’ to call a sacrifice (1986: 59, 98).

  • 9

    For example, while weddings are commonly seen, in accordance with alliance or descent theory, as acts of exchange or affinity or the transmission of rights in offspring, property, and services, a characteristic of the weddings I have observed or participated in, both in North America and among Malagasy speakers in Mayotte, is the sort of optimistic refocusing ascribed by Said to beginnings. Indeed, I would say that ‘beginning’ forms both a central function and a self-interpretation or theme of weddings (Lambek 1983; 2004).

  • 10

    The king is Ndramandisoarivo. The identity of the oracle varies from version to version, as does the method by which the Queen dies.

  • 11

    The clan descended from the King alone holds authority over royal burial grounds.

  • 12

    For exposition of the gender of sacrifice, see Lambek (in press a and in press b).

  • 13

    The Queen's act draws on symbolism that stretches back earlier, to the kingdom of Menabe, as does the royal genealogy.

  • 14

    A precursor of this view is found in Hubert and Mauss's classic Sacrifice (1964 [1888]). They highlight the presence and destruction of a victim as well as the Vedic model of the creation of the universe as a sacrifice. Indeed, Malamoud observes both that Vedic creation presupposes a sacrificer, oblation, and model (hence that it is a beginning rather than an origin) (1990: 188) and that the rites themselves elaborate the process of beginning and begin with an act of violence, a ‘first slice’ (une entame) (1990: 190). I am very grateful to Anne de Sales for providing the relevant text. Parry (1994) offers an exemplary ethnographic exposition of Hindu death as sacrifice and creation.

  • 15

    This is akin perhaps to Lévi-Strauss's argument with respect to totemism (1963 [1962]) or Franz Steiner's with respect to taboo (1956).

  • 16

    This is the case despite the fact that the body itself is often grasped by means of metaphor.

  • 17

    However, Sakalava leave the death of royalty ambiguous and unspoken. Rotting flesh is separated from enduring bones and the royal ancestors perdure as tromba (possessing spirits).

  • 18

    See Schwimmer (1973) on the ‘starting mechanisms’ of exchange and on exchange as sacrifice; Graeber (2001: 176-88) on the hau and the meaning of Maori initial gifts; and Caillé (2000) for a lively discussion of sacrifice as utilitarian exchange. Like Caillé, my discussion of sacrifice does not include the sense of abnegation.

  • 19

    For this reason also, I do not emphasize the feast. In this I follow Sakalava practice – the meat of animals sacrificed at the shrine is divided up and, for the most part, taken to individual households to be cooked and eaten separately. In contrast to commensality, the act of killing is a salient aspect of animal sacrifice virtually everywhere. As an indirect piece of evidence I note that depiction of animal slaughter is practically de rigueur in ethnographic film. Of course, consuming the sacrificial object could be seen as the final, incorporative phase of the rite; and, like cutting and spilling, eating and digestion are irreversible acts and processes that cannot be taken back.

  • 20

    Faust, part I, line 1264 (Goethe 1992 [1797]: 39).

  • 21

    Acts of lime-cutting in Sri Lanka (Obeyesekere 1981) or sharing kola nuts in Nigeria are examples. Note they both entail cutting.

  • 22

    Indeed, utterances are acts, a point made not only by Austin but also, from an alternative philosophical tradition, by Arendt (1998 [1958]). Systematic analysis would show similarities and differences between sacred postulations, in Rappaport's sense and other kinds of acts. In any case, alongside Rappaport's ultimate sacred postulates (1999: 263 ff.) – USPs – one could add ultimate sacred acts – USAs – of which sacrifice may be a prototypic example.

  • 23

    The executive significance of sacrifice can also be gleaned from its refusal. Evans-Pritchard describes the Nuer leopard skin chief threatening a curse against those reluctant to accept mediation.
    He would take an ox and rub its back with ashes and begin to address it, saying that if the injured party were to insist on revenge many of them would be killed in the endeavour ... he would then raise his spear to slaughter the animal, but this would be as far as people would care to let him go. Having asserted their pride of kin, one of the dead man's family would seize his upraised arm to prevent him from stabbing the ox, saying: ‘No! Do not kill your ox. It is finished. We will accept compensation’ (1940: 175).

  • 24

    This might be linked to Heidegger's formulation of the violence of being in his Introduction to metaphysics and described by Nelson (n.d.) as ‘an originary strife ... that is intended to contest the reification of identity’.

  • 25

    Ideally, defloration is followed by a liminal period and final incorporation in Van Gennep's sense. Nevertheless, it is evident in the manner in which the defloration is immediately celebrated (with congratulations, ululations and sexy dancing) that a new beginning for the bride has been achieved. As for circumcision, insofar as cutting the foreskin is a repetition of Abraham's sacrifice of Isaac, it is itself a sacrifice. Circumcision is a mark not only of difference or singularity but also of intentionality, a performative act establishing the kind of person one (or one's child) is intended to be and in respect to which one accepts subsequent conduct to be evaluated. Among Dogon, sacrifice introduces differentiation, ‘the birth of a discontinuous world’ (De Heusch 1985: 134); circumcision as sacrifice is linked to menstruation, periodicity, and alternation, but compare Bloch (1986) on non-Muslim Malagasy circumcision.

  • 26

    In a companion essay (Lambek in press a) I interpret giving birth as a form of sacrifice. Conversely, Ndramandikavavy's sacrifice is a form of birth – a claim to being mother of her son and an act that gives birth to the polity of Boina. On Sakalava birth see Feeley-Harnik (2000); on action as natality see Arendt (1998 [1958]: 9).

  • 27

    I wish to emphasize that I am not attempting to romanticize sacrifice or intending to allude to contemporary acts of violence. The suicide bomber is not beginning anything nor transforming death into life. While the act is surely decisive and irreversible, it is neither part of a liturgical order nor explicitly performative and its aim is purely instrumental. On the dangers of applying the religious language of martyrdom to politics, see Hasan-Rokem (2003).

  • 28

    Sakalava make blood sacrifices beneath the corner posts of new temples, at doorways, and along paths that people must step over.

  • 29

    Compare Arendt (1998 [1958]), for whom the difference between making (‘work’) and doing (‘action’) is highly consequential. Arendt emphasizes the radical unpredictability and irreversibility of action in contrast to the routines of practice. A full theory of action would bring together the philosophical literature on intentionality and ethics (e.g. Cavell 1976) with sociological discussions of ‘social creativity’ and social action (Graeber 2005; Joas 1996 [1992]).

  • 30

    Goethe also works with a contrast between inward activism and outward worldliness (Berman 1988: 46-7), a theme of German culture that entailed additional contrasts between Christians and Jews as well as between Kantian idealist and Aristotelian practical ethics. Berman argues (1988: 46-7) that Faust's inversion of ‘In the Beginning was the Word’ was understood as replacing Christ with the Old Testament God who makes the world. It was part of Weber's brilliance to complexify and transform these oppositions.

  • 31

    Faust, part I, line 1759 (Goethe 1992 [1797]: 54).

Le sacrifice et le problème du commencement : méditation sur la mythopraxie des Sakalavas


L’auteur aborde le problème général du commencement dans la pensée et l’action humaines. Selon lui, l’accent mis sur la transition dans l'analyse du rituel doit être complété en accordant de l’attention au commencement, et la relative passivité de la liminalité doit se doubler du caractère résolu de l'action initiatrice, tout en s'intéressant aux aspects transitifs et intransitifs du commencement lui-même. À partir des représentations de la formation d’une monarchie Sakalava à Madagascar, l’article présente le sacrifice comme une forme exemplaire de commencement. Cette description du sacrifice rend caduque la nécessité d'une théorie du sacrifice, tout un offrant de nouveaux éclairages sur le don, la nature éthique de la personne et la temporalité de la tradition.

Michael Lambek currently alternates between the London School of Economics and Political Science and the University of Toronto, where he holds a Canada Research Chair in the Anthropology of Ethical Life. His books include Human spirits (Cambridge University Press, 1981); Knowledge and practice in Mayotte (University of Toronto Press, 1993); The weight of the past (Palgrave, 2002); and Illness and irony (Berghahn 2003, ed. with Paul Antze).