Title: Fetishism as social creativity
Subtitle: or, Fetishes are gods in the process of construction
Author: David Graeber
Date: 2005
Source: Retrieved on 28th November 2021 from davidgraeber.org
Notes: Published in Anthropological Theory Vol 5(4): 407–438. doi: 10.1177/1463499605059230.


Originally, the term ‘fetishes’ was used by European merchants to refer to objects employed in West Africa to make and enforce agreements, often between people with almost nothing in common. They thus provide an interesting window on the problem of social creativity – especially since in classic Marxist terms they were surprisingly little fetishized. Starting with an appreciation and critique of William Pietz’s classic work on the subject, and reconsidering classic cases of Tiv spheres of exchange and BaKongo sculpture, this article aims to reimagine African fetishes, and fetishes in general, as ways of creating new social relations.


BaKongo • fetishes • fetishism • social contract theory • social creativity • Tiv

In this article, I would like to make a contribution to theories of social creativity. By social creativity, I mean the creation of new social forms and institutional arrangements. Creativity of this sort has been the topic of some discussion in social theory of late, although up to now anthropology has not played much of a role in it. Here I would like to bring anthropology into an area that has traditionally been seen as its home turf: by looking at the literature on ‘fetishism’ in Africa.

Now one could argue that creativity of this sort has always been one of the great issues of social theory, but it seems to me that the current interest can be traced to two impulses. Or perhaps more precisely, the desire to work one’s way out of two ongoing dilemmas that have haunted social theory for some time. One, mapped out most clearly, perhaps, by Alain Caillé (2000), French sociologist and animateur of the MAUSS group, is the tendency for theory to endlessly bounce back and forth between what he calls ‘holistic’ and ‘individualistic’ models. If one does not wish to look at human beings simply as elements in some larger structure (a ‘society’, a ‘culture’, call it what you will), doomed to endlessly act out or reproduce it, and if one does not want to fall back on the economistic ‘rational-choice’ option, which starts from a collection of individuals seeking personal satisfaction of some sort and treats larger institutions as mere sideeffects of their choices, then this seems precisely the point at which to begin formulating an alternative. Human beings do create new social and cultural forms all the time, but they rarely do so just in order to further their own personal aims; in fact, often their personal aims come to be formed through the very institutions they create. Caillé proposes that the best way to develop an alternative to the currently dominant utilitarian, ‘rational-choice’ models is by setting out, not from market relations, but instead from Marcel Mauss’ famous exposition of the gift, which is all about the creation of new social relations. He is not the only one working in this direction. Hans Joas (1993, 1996, 2000) has been trying to do something quite similar, setting out not from Mauss but from the tradition of American pragmatism. I have myself been trying to do something along these lines in my book Towards an Anthropological Theory of Value, where, inspired in part by ideas developed by my old professors Terry Turner (e.g. 1979, 1984) and Nancy Munn (e.g. 1977, 1986), I attempted to broaden the Marxian notion of production to include the fashioning of persons and social relations.

The other impulse is more explicitly political, and has to do with the concept of revolution. Here the problematic stems broadly from within Marxism. Marx, perhaps more than any other classic social theorist, saw creativity and imagination as the essence of what it means to be human; but as Hans Joas among others have remarked, when he got down to cases he tended to write as if all forms of creative action really boiled down to two: the production of material objects, and social revolution. For Joas, this makes Marx’s approach so limited he prefers to discard it entirely; I prefer to keep what I take to be his most profound insights and apply them to other forms of creativity as well; but what’s at issue here is the relation between the two. Because there is a curious disparity. Marx assumes that both the human capacity for creativity and human critical faculties are ultimately rooted in the same source, which one might call our capacity for reflexive imagination. Hence his famous example of the architect who, unlike the bee, raises her building in her own imagination before it is raised in reality. If we can imagine (as yet non-existent) alternatives, we can see the existing world as inadequate; we can also cause those things to exist. This is the ambiguity, though: while our ability to revolutionize emerges from this very critical faculty, the revolutionary, according to Marx, must never proceed in the same manner as the architect. It was not the task of the revolutionary to come up with blueprints for a future society and then try to bring them into being, or, indeed, to try to imagine details of the future society at all. That would be utopianism, which for Marx is a foolish bourgeois mistake. So the two forms of creativity – the creation of houses, or other material objects, and the creation of new social institutions (which is, after all, what revolution actually consists of) – should not work in at all the same way.

I have written a little about this paradox before.[1] What I want to emphasize here is how it has contributed to a fundamental problem in revolutionary theory: what precisely is the role of creativity, collective or individual, of the imagination, in radical social change? Unless one wishes to adopt completely absurd formulations (the revolution will come about because of the inexorable logic of history; human agency will have nothing to do with it; afterwards, however, history will end and we will enter a world of freedom in which human agency will be utterly untrammeled ...) this has to be the key question, but it’s not at all clear what the answer is supposed to be. The revolutionary theorist who grappled with the problem most explicitly was Cornelius Castoriadis, whose Socialisme ou Barbarie group was probably the single most important theoretical influence on the student insurrectionaries of May 1968. Castoriadis was the effective founder of the Autonomist tradition, which has come to be probably the dominant strain of Continental Marxism,[2] and eventually took Marx’s starting point – his faith in the critical role of the creative imagination and, hence, our capacity to revolutionize – so seriously that he ended up abandoning most other tenets of Marxism entirely. For Castoriadis, the great question became the emergence of the new.[3] After all, most of the really brilliant moments of human history involve the creation of something unprecedented, something that had never existed before, whether Athenian democracy or Renaissance painting, and this is precisely what we are used to thinking of as ‘revolutionary’ about them. History, then, was a matter of the constant pressure of the imaginary against its social containment and institutionalization. It is in the latter process, he argued, that alienation enters. Where Marx saw our dilemma in the fact that we create our physical worlds, but are unaware of, and hence not in control of, the process by which we do so (this is why our own deeds seem to come back at us as alien powers), for Castoriadis, the problem was that ‘all societies are instituted by themselves’ but are blind to their own creativity. Whereas a truly ‘democratic society is a society which is instituted by itself, but in an explicit way’ (in Ciaramelli, 1998: 134). By the end, Castoriadis abandoned even the term ‘socialism’, substituting ‘autonomy’, defining autonomous institutions as those whose members have themselves, consciously, created the rules by which they operate, and are willing to continually re-examine them (Castoriadis, 1991).[4]

This does seem a unique point of tension within radical thought. It is probably no coincidence that Roy Bhaskar, founder of the Critical Realism school (1979, 1993, 1994, 2001), found this exactly the point where he had to break with the western philosophical tradition entirely. After arguing for the necessity of a dialectical approach to social problems, he found himself asking: when contradictory elements are subsumed in a higher level of integration which is more than the sum of their parts, when apparently intractable problems are resolved by some brilliant new synthesis which takes things to a whole new level, where does that newness actually come from? If the whole is more than the sum of its parts, what is the source of that ‘more’, that transcendent element? In his case he ended up turning to Indian and Chinese philosophical traditions and arguing that the main reason why existing Marxism has produced such disappointing results has been its refusal to take on such issues, owing to its hostility to anything resembling ‘spiritual’ questions.

What is important for present purposes is merely to underline that all these authors are, in one way or another, dealing with the same problem. If one does not wish to see human beings simply as side-effects of some larger structure or system, or as atoms pursuing some inscrutable bliss, but as beings capable of creating their own meaningful worlds, then their ability to create new institutions or social relations does seem just the place to look. Radical thinkers are just dealing with the same issues from a more pragmatic perspective, since as revolutionaries, what they are interested in is precisely the creation of new social institutions and new forms of social relation. As I say, it is obvious that people do, in fact, create new institutions and new relations all the time. Yet how they do so remains notoriously difficult to theorize.

Can anthropology be of any assistance here? It is not obvious that it could. Anthropologists have not exactly been grappling with this kind of grand theoretical issue of late, and have never had much to say about revolution. One could of course argue that maybe this is all for the best, that human creativity cannot be, and should not be, subjected to anyone’s theoretical model. But a case could equally well be made that if these are questions worth asking, then anthropology is the only discipline really positioned to answer them – since, after all, the overwhelming majority of actual, historical, social creativity has, for better or worse, been relegated to our academic domain. Most of the classic issues even of early anthropology – potlatches, Ghost Dances, magic, totemic ritual and the like – are precisely about the creation of new social relations and new social forms.

Alain Caillé would certainly agree with this assessment: that is why he chose Marcel Mauss’ essays on the gift as his starting point. Mauss himself saw his work on gifts as part of a much larger project, an investigation into the origins of the notion of the contract and of contractual obligation. (That is why the question that really fascinated him was why it was that someone who receives a gift feels the obligation to return one.) This has proved a highly fruitful approach but in this article I would like to suggest another one, that I hope will be equally productive, which opens up a slightly different set of questions. This is to begin with the problem of fetishism.


‘Fetishism’ is of course a much-debated term. It was originally coined to describe what were considered weird, primitive, and rather scandalous customs, and as a result most of the founders of modern anthropology – Marcel Mauss prominent among them – felt the term was so loaded it were better abandoned entirely. It no doubt would have been, had it not been for the fact that it had been so prominently employed – as a somewhat ironic technical term to describe certain western habits – by both Karl Marx and Sigmund Freud. In recent years, the word has undergone something of a revival, mainly because of the work of a scholar named William Pietz, who wrote a series of essays called ‘The Problem of the Fetish’ (1985, 1987, 1988), tracing the history of the term, its emergence in intercultural enclaves along the West African coast from the 16th through 18th centuries CE. Pietz is that most unusual of things, an independent scholar who has had an enormous influence on the academy. His essays ended up inspiring a small literature of their own during the 1990s, including one large and well-received interdisciplinary volume in the US (Apter and Pietz, 1993), two different collections in the Netherlands (Etnofoor, 1990, Spyer, 1998), and any number of essays. The overriding theme in all this literature is materiality: how material objects are transformed by becoming objects of desire or value, a value that often seems somehow displaced, inordinate, or inappropriate. My own interest here is slightly different. What is especially interesting to me is Pietz’s argument that the idea of the ‘fetish’ was the product neither of African nor of European traditions, but of a confrontation between the two: the product of men and women with very different understandings of the world and what one had a right to wish from it trying to come to terms with one another. The fetish was, according to Pietz, born in a field of endless improvisation, that is, of near pure social creativity.

In what follows, I will first consider Pietz’s story of the origin of the fetish, then try to supplement his account (drawn almost exclusively from western sources) with some that might give insight into what the African characters in the story might have thought was going on, and then, return to our initial problem – and to see how all this relates to ‘fetishism’ in the more familiar Marxian sense. To summarize a long and complex argument, basically what I will argue is this: we are used to seeing fetishism as an illusion. We create things, and then, because we don’t understand how we did it, we end up treating our own creations as if they had power over us. We fall down and worship that which we ourselves have made. By this logic, however, the objects European visitors to Africa first labeled ‘fetishes’ were, at least from the African perspective, remarkably little fetishized. They were in fact seen quite explicitly as having been created by human beings; people would ‘make’ a fetish as the means of creating new social responsibilities, of making contracts and agreements, or forming new associations. It was only the Europeans’ obsession with issues of value and materiality and their almost complete lack of interest in social relations as things valuable in themselves that made it possible for them to miss this. This is not to say they were completely unfetishized. But this is precisely what is most interesting about them.


If the reader will allow me a highly simplified version of Pietz’s complex and layered argument: the notion of the fetish was not a traditional European concept. Medieval Europeans tended to interpret alien religions through very different rubrics: idolatry, apostasy, atheism. Instead the idea seems to have arisen, in the minds of early Italian, Portuguese, and Dutch merchants, sailors, and maritime adventurers doing business in West Africa starting in the 15th century, primarily from a confrontation with the threat of relativism. These foreign merchants were operating in an environment which could hardly fail to cast doubt about their existing assumptions about the nature of the world and of society: first and foremost, with the relativity of economic value, but also of the logic of government, the dynamics of sexual attraction, and any number of other things. By describing Africans as ‘fetishists’, they were trying to avoid some of the most disturbing implications of their own experience.

The first Portuguese merchants who set up ‘castles’ on inlets and river islands along the West African coast were brought there by one thing: the belief that this part of the world was where most – if not all – of the world’s gold originally came from. In the 16th and 17th centuries gold was the main product being extracted from the region (it was only somewhat later that attention shifted primarily to slaves). These were extremely practically minded individuals, entering into a very complex world full of an apparently endless variety of unfamiliar languages, religions, and forms of social organization, none of which, however, they had any particular interest in understanding as phenomena in their own right. They were simply after the gold. The very experience of moving between so many cultures, Pietz suggests, encouraged a kind of bare-bones materialism; in their writings, he notes, early merchant explorers tended to describe a world in which they perceived only three categories of significant object: tools, potential dangers, and potential commodities (1985: 8). And for obvious reasons they also tended to assess the value of just about everything by the price they thought it could fetch in European markets.

The problem was that in order to conduct their trade, they had to constantly confront the fact that the Africans they met had very different standards of value. Not entirely different. ‘Gold is much prized by them’, wrote an early Venetian merchant named Cadamosto, ‘in my opinion, more than by us, for they regard it as very precious: nevertheless they traded it cheaply, taking in exchange articles of very little value in our eyes’. To some extent this led to the familiar rhetoric of beads and trinkets. Merchants were always going on about how Africans were willing to accept all manner of junk – ‘trifles’, ‘trash’, ‘toys’ – for gold and other valuable commodities. But at the same time, Africans were clearly not willing to accept just anything, and one could never tell in advance what sort of junk a given group would fancy. Anyone who has pored over ‘travelers’ accounts’ from this period will likely have noticed how much time and energy merchants had to put into figuring out which particular variety of worthless beads, what color or type of worthless trinkets would be accepted at any given port of call.

Situations like this can very easily lead one to think. To reflect on the arbitrariness of value. After all, it is important to bear in mind that these early merchant adventurers were not only seeking gold, they were doing it at very considerable risk to their own lives. Coastal ‘castles’ were malarial pest-holes: a European who spent a year in one had about a 50:50 chance of coming back alive. It would be very easy, in such circumstances, to begin to ask oneself: why are so many of us willing to risk death for the sake of a soft yellow metal, one which isn’t even useful for anything except to look pretty? In what way is this really different than a desire for beads and trinkets?[5] It was not as if people of the time were incapable of such reflections: the absurdity of such overweening desire for gold became a stock theme for popular satirists, particularly in the age of the conquistadors. The merchants in West Africa, however, instead seem to have come to the brink of such a conclusion and then recoiled. Instead of acknowledging the arbitrariness underlying all systems of value, their conclusion was that it was the Africans who were arbitrary. African societies were utterly without order, their philosophies utterly unsystematic, their tastes utterly whimsical and capricious:

the most numerous Sect [in Guinea] are the Pagans, who trouble themselves about no Religion at all; yet every one of them have some Trifle or other, to which they pay a particular Respect, or Kind of Adoration, believing it can defend them from all Dangers: Some have a Lion’s Tail, some a Bird’s Feather, some a Pebble, a Bit of Rag, a Dog’s Leg; or, in short, any thing they fancy: And this they call their Fetish, which Word not only signifies the Thing worshipped, but sometimes a Spell, Charm, or Inchantment. (William Smith, 1744, in Pietz, 1987: 41)

So Africans were evidently like small children, always picking up little objects because they look odd or gross or brightly colored, and then becoming attached to them, treating them as if they had personalities, adoring them, giving them names. The same thing that inspired them to value random objects in the marketplaces caused them to make random objects into gods.

The commonest explanation of the origin of fetishes begins something like this. An African intends to set out on some project, to go off trading for example. He heads out in the morning and the first thing he sees that strikes him as in any way unusual or extraordinary, or just that randomly strikes his fancy, he adopts as a charm that will enable him to carry out his plan. Pietz calls it the ‘chance conjuncture of a momentary desire or purpose and some random object brought to the desirer’s attention’; Le Maire put it more simply: they ‘worship the first thing they meet in the Morning’. Bosman writes of one of his informants:

He obliged me with the following Answer, that the Number of their Gods was endless and innumerable: For (said he) any of us being resolved to undertake any thing of Importance, we first of all search out a God to prosper our designed Undertaking; and going out of Doors with this design, take the first Creature that presents itself to our Eyes, whether Dog, Cat, or the most contemptible Animal in the World, for our God; or perhaps instead of that any Inanimate that falls in our way, whether a Stone, a piece of Wood, or any Thing else of the same Nature. (in Pietz, 1987: 43)

It was not the ‘Otherness’ of the West Africans that ultimately drove Europeans to such extreme caricatures, then, but rather, the threat of similarity – which required the most radical rejection. So too with aesthetics, particularly the aesthetics of sexual attraction. European sources wrote of the odd practices of the women they encountered in coastal towns, who ‘fetishized themselves’ by making up their faces with different kinds of colored clays, or wore ‘fetish gold’ in their hair, intricately worked ornaments, frogs and birds along with glass beads and similar adornment. The descriptions here are not normally morally condemnatory, but they usually adopt a kind of sneering tone, one of contempt for what seems to pass as beauty in these parts, what Africans found alluring or attractive. But again they obviously protest too much. If European sojourners were entirely immune to the charms of women with earth on their faces and frogs in their hair, they would not have fathered hundreds of children with them; indeed, there is no particular reason to assume that the numbers of such children would have been substantially higher had the women in question behaved like proper European ladies and put grease on their lips and gold rings in their ears instead.

The same dynamic recurs when Europeans talked about African modes of government. First, observers would insist that the basis of African social life was essentially chaotic, that it was utterly lacking in systematic public order; they would usually end up by admitting that laws were, in fact, quite systematically obeyed. According to some, almost miraculously so. The attitude is summed up by a later British administrator, Brodie Cruickshank, Governor General of the Gold Coast in the 19th century:

The local govt of the Gold Coast must have the candor to acknowledge its obligations to Fetish, as a police agent. Without this powerful ally, it would have been found impossible to maintain that order, which characterized the country during the last twenty years, with the physical force of the govt. The extraordinary security afforded to property in the most remote districts, the great safety with which packages of gold of great value are transmitted by single messengers for hundreds of miles, and the facility with which lost or stolen property is generally recovered, have excited the astonishment of Europeans newly arrived in the country. (Cruickshank, 1853, in Pietz, 1985: 25)

The reason, they concluded, boiled down to the most primitive of instincts: fear of death, or the terrible punishments fetishes were thought to bring down on those who violated their (somewhat arbitrary) principles.

Again, the problem was not that the picture was so alien, but that it was so familiar. That government was an institution primarily concerned with threatening potential miscreants with violence was a longstanding assumption in western political theory; that it existed primarily to protect property was a theme in the process of emerging at this very time. True, the fetish was said to operate by invisible, supernatural means, and hence to fall under the sphere of religion and not government. But these observers were also, overwhelmingly, Christians, and Christians of that time insisted that their religion was morally superior to all others, and particularly to African religions, on the very grounds that their God threatened wrong-doers with the systematic application of torture for all eternity, and other gods did not. The parallels were in fact striking, although this was an area in which Europeans found it particularly difficult to be relativistic. It was their assumption of the absolute truth of the Christian faith which probably made any broader move to a relativistic attitude impossible.

On the other hand, this was an area where common understandings made a great deal of practical difference, because especially before Europeans came as conquerors, oaths sworn on fetishes and contracts made by ‘making’ or ‘drinking’ fetishes were the very medium of trust between Europeans and Africans engaged in trade. If it were not for their common participation in such rituals – often newfangled ones improvised for the occasion combining Bibles and beads and bits of wood all at the same time – the trade itself would have been impossible. And of course this is what especially interests us here.


Now, as the reader might have noticed, Pietz is almost exclusively concerned with how things seemed to the Europeans who came to Africa. There is almost no speculation about what any of the Africans with whom they traded might have thought was going on.[6] Of course, in the absence of documentary evidence, certain knowledge is impossible; but there is a pretty voluminous literature on more recent examples of the sorts of object these Europeans labeled ‘fetishes’, as well as on African cosmological systems more generally, so one can make some pretty good guesses as to what the Africans who owned and used such objects thought they were about. Doing so does not, in fact, invalidate any of Pietz’s larger points. Actually, it suggests that the ‘threat of recognition’, if I may call it that, runs far deeper than one might otherwise suspect.

Allow me to begin here with some probably unwarranted generalizations about the relation between European and African cosmologies. My interest in Pietz, and in fetishism more generally, originally arose as part of a comparative study of beads and other ‘currencies of trade’ (Graeber, 1995, 2001), which included cases ranging from Trobriand kula shells or Iroquois wampum to Kwakiutl coppers. For someone brought up in a religious environment largely shaped by Christianity, moving from the cosmological systems of Oceania or North America to Africa is moving from very alien to far more familiar territory. It is not just that throughout Africa one can find mythological topoi(the Garden of Eden, the Tower of Babel) familiar from the Old Testament, that just do not seem to be present in other traditions. There is a sense that African theologians seem to be asking mostly the same existential questions.[7] Max Weber made a famous argument that every religion has to come up with some answer to the question of ‘theodicy’, or the justice of God. How is it that if God is both good and all-powerful, that human beings must suffer? Now, it’s pretty obvious that, as a generalization, this is simply untrue. The question probably would not have even made sense to a Maori theologian, let alone, say, an Aztec poet or Trobriand chief. While every tradition does seem to see the human condition as inherently problematic in some way, in most, the reasons for human suffering is just not the issue. The problem lies elsewhere. Mythic speculation in Africa, on the other hand, focuses on the question endlessly (e.g. Abrahamsson, 1952) – even if many African theologians came up with what were from the Christian perspective very disturbing answers (i.e. who says God is good?)[8]

I said ‘unwarranted generalizations’ because, as any number of authors have reminded us, terms like ‘Africa’, ‘Europe’, ‘the West’, are fuzzy at best and probably meaningless. I cannot claim to know why European and African theologians seem to have been asking the same existential questions – perhaps it is because Europe and Africa were, for so much of their history, peripheral zones under the influence of the great urban civilizations of the Middle East. What I want to stress, though, is that here, 17th or 18th century European seafarers found themselves in much more familiar territory than they did when they ventured to places like China or Brazil. It was the underlying affinity, I suspect, which accounted for the common European reaction of shocked revulsion and dismay on being exposed to so many aspects of African ritual: a desperate denial of recognition. Because in many ways, African cosmological ideas seemed to take the same questions and come up with precisely the conclusions Europeans were most anxious to avoid: such as, perhaps we suffer because God is not good, or is beyond good and evil and does not care; perhaps the state is a violent and exploitative institution and there is nothing that can be done about it; and so on.

I shall return to this theme in a moment.

Throughout much of Africa, ceremonial life is dominated by what anthropologists have labeled ‘rituals of affliction’. Those powers considered worthy of recognition are almost invariably those capable of causing human misery, and one comes into contact with them when they attack one in some way. A typical chain of events (I’ll use a Malagasy example out of familiarity) might run like this: one offends a power without knowing it, say by bringing pork into a spot inhabited by a Vazimba spirit; the offended spirit causes one to become ill, or to experience nightmares; one goes to a local curer who identifies the spirit and tells one how to propitiate it; doing so, however, causes one to become part of a congregation of former victims all of which now have a special relationship with the spirit, which can help one or even direct its powers against one’s enemies. Suffering leads to knowledge, knowledge leads to power. This is an extremely common pattern. Victor Turner for instance estimates that among the Ndembu of Zambia, there are essentially only two types of ritual: rituals of affliction, and ‘life-crisis rituals’ such as initiations and funerary rites. And he adds that even the latter always ‘stressed the theme of suffering as a means of entry into a superior ritual and social status’ (1967: 15–16); normally, because initiation rituals passed through physical ordeals (suffering) to the attainment of some kind of ritual knowledge. Most of the African objects labeled ‘fetishes’ were enmeshed in precisely this ritual logic.

Let me take two representative examples. The first is the Tiv of central Nigeria, c.1900–1950. They are a good case to start with because they are well documented and lived not too far from the region dealt with in Pietz’s texts. The second is the BaKongo of the central African coast, who have a much longer history of entanglement with European trade. The Tiv are a classic example of a ‘segmentary’ society; before they were conquered by the British, they recognized no centralized authority of any sort, beyond the confines of a typical extended family compound (L. Bohannan, 1952; P. Bohannan, 1959; Bohannan and Bohannan, 1953). Larger society was instead organized on a genealogical basis, through an elaborate system of patrilineal lineages, which, however, had no permanent officials or ritual officers. Where the ritual life of most segmentary societies in the region centered on an elaborate cult of ancestors or of earth shrines, the Tiv lacked these too. Instead, their ritual life revolves largely around warding off witchcraft, and the control of objects called akombo, or ‘fetishes’.

The names of most akombowere also those of disease. In a certain sense, the akombo quite simply were those diseases,[9] though they were also embodied in material ‘emblems’. These emblems might be almost anything: a pot of ashes, a whisk broom, a piece of elephant bone. These existed in certain places, and were owned by ‘keepers’, and they were always surrounded by a host of rules and regulations indicating what could and could not be done in the vicinity. One came into relation with an akombowhen one broke one of those rules – this is called ‘piercing’ it – and became sick as a result. The only way to set things straight was to approach its keeper in order to ‘repair’ the akombo or ‘set it right’. After victims have so freed themselves from the effects of the fetish, they might also decide to take possession of it, which involves a further ritual of ‘agreement’ and sacrifice in order to give the new owner the power to operate (‘repair’) it, so as to help others so afflicted, and also to gain access to whatever other powers the akombo might have (Bohannan and Bohannan, 1969). All this is very much on the model of a typical ‘cult of affliction’.

What I have said so far applies to minor, or ordinary, akombo. There were also major akombo, which had broader powers. Probably the most important of these were those that protected markets. According to Tiv informants of the colonial period, what really distinguished these great akombofrom the ordinary variety was, first of all, that they could protect a whole territory from harm; second, that they could be passed on from father to son; third, that they ‘either contain a part of a human body as a portion of their emblems, or they must be repaired by a human sacrifice ... or both’ (Bohannan and Bohannan, 1969 IV: 437).

To understand this, one has to understand something, I think, about traditional Tiv conceptions of social power – at least as they existed in the early 20th century. The Tiv combined very hierarchical domestic arrangements – with household compounds constructed around some important older man, almost invariably with numerous wives, surrounded by a host of frustrated unmarried adult sons – and a fiercely egalitarian ethos which allowed next to nothing in the way of political office outside the compound. Certain older men manage to gain a larger influence in communal affairs, but such accomplishments are viewed with extreme ambivalence. Social power, the ability to impose one’s will on others, is referred to as tsav; it is seen in quite material terms as a fatty yellow substance that grows on human hearts. Some people have tsav naturally. They are what we would refer to as ‘natural leadership types’. It can also be created, or increased, by eating human flesh. This is ‘witchcraft’, the definition of evil:

Tiv believe that persons with tsav form an organization called the mbatsav. This group is said to have a division of labor and a loose organization. The mbatsav are said to meet at night, usually for nefarious purposes; they rob graves in order to eat corpses;

they bewitch people in order to put corpses into graves which they can rob. There is thought to be a network of ‘flesh debts’ which become established when someone tricks you into eating human flesh and then claims a return in kind; the only thing you can do is to kill your children and your close kinsmen – people over whom you have some sort of power – and finally, because no one can ever win against the organization, you must give yourself to them as a victim because you have no kinsmen left to give. (P. Bohannan, 1958: 4–5)

As Paul Bohannan succinctly puts it: ‘men attain power by consuming the substance of others’. While one can never be certain that any particular elder is also an evil cannibalistic witch, the classes overlap, and it would seem that in recorded times at least, every generation or so, a witch-finding movement would sweep through the country unmasking the most prominent figures of local authority (Akiga, 1939; P. Bohannan, 1958).[10] This is not quite a system in which political power is seen as intrinsically evil, but it

is very close. It only stands to reason, then, that akombothat have power over communities should have a similar predilection to absorb human flesh. The information we have about most of these ‘great akombo’ is somewhat limited, because most were destroyed during a witch-finding movement in the 1920s, but the one sort that did tend to survive were the akomboof markets. Fortunately, these are the most relevant to the issues under consideration here.

Tiv markets are dominated by women, who are also the main producers. Over the last few centuries markets have also been the principal context in which most Tiv come into contact with those with whom they can trace no close genealogical ties, and therefore, towards whom they have no necessary moral obligations. In markets, then, the destructive powers of akombocould be used to keep the peace. Every significant market had its own fetish (Bohannan and Bohannan, 1968: 149, 158–62), which Tiv of the colonial period, interestingly, often compared to an authorization certificate from the colonial regime. Essentially, they embodied peace agreements between a series of lineages who shared the same market, by which their members undertook to deal fairly with one another, and to abstain from theft, brawling, and profiteering. The agreement was sealed with a sacrifice – nowadays said to be a human sacrifice, though the Bohannans suspect most often it was really just a dog – whose blood was poured over the akombo’s emblem. This is the sacrifice by day; in addition, the (male) elders, in their capacity as mbatsav, kill others of their own lineages ‘by night’ – that is, by witchcraft (Bohannan and Bohannan, 1968: 159–60). Henceforth all those who violated the agreement would be struck down by the akombo’s power. And in fact, the existence of such agreements made it possible for marketplaces to become meeting places for the regulation of local affairs, judgments, and the taking of oaths.

This gives some idea, I think, of the logic by which ‘fetishes’ also came to mediate trade agreements with European merchants in the 16th and 17th centuries. The similarity with European theories of the social contract, which were developing at precisely this time, need hardly be remarked. I will return to these parallels in a moment. The Tiv themselves had little to do with Europeans before the British conquest; they came into relation with the trade largely as victims, being raided for slaves by more powerful neighbors. As a result their recorded history is very shallow. The BaKongo, famous for their minkisior ‘fetishes’, many considered brilliant works of art, on the other hand have one of the longest recorded histories in Africa. In 1483 the Kongo kingdom entered into an alliance with Portugal; the royal family converted to Catholicism. At the time its capital, São Salvador, was the largest city south of the Sahara. Within a century the kingdom was torn apart by the pressures of the slave trade, and in 1678 the capital was destroyed; the kingdom broke down into a series of smaller successor states, most of whom officially recognized the authority of a nominal Kongo monarch stripped of almost all real power: a classic hollow center (Thornton, 1987). Later centuries witnessed even greater fragmentation, the centers of most of the successor states hollowed out in similar fashion, leaving a highly decentralized social field in which former chiefly titles increasingly became prizes that could be bought and sold by successful merchants and slave-traders. Certainly this was the case by the 19th century, during which power gradually shifted to commercial towns along the coast. This is also the period from which we have most of our information on minkisi, as recalled in documents recorded by Christian converts, in the KiKongo language, at the very beginning of the colonial age. In a lot of ways the BaKongo might seem as different from Tiv as can be: matrilineal where the Tiv were patrilineal, hierarchical where the Tiv were egalitarian, with a cosmology centering on the ancestral dead which was totally alien to Tiv conceptions. But the basic assumptions about the nature of power in both cases are remarkably similar. First of all, we find the same logic of affliction: here too, one comes into contact with powers largely by offending them; once that power has caused one to suffer, then one has the opportunity to master it and, to an extent, to acquire it for oneself.[11] This was the normal way in which one came into relation with a nkisi: one first appealed to its keeper to cure one of an ailment; as such, one became a member of what might be broadly called its congregation; later, perhaps, if one was willing to undergo the expensive initiationprocess, one could eventually become a keeper oneself.

BaKongo and Tiv theories of the relation of political power and witchcraft were also remarkably similar. The power of chiefs was assumed to be rooted in a physical substance in the body – in this case, called kindoki. This was also the power of witches. The main difference was that Kongo witches operated on a level that was somewhat more abstract than Tiv witches; while they too became entangled in ‘flesh debts’, they were mainly represented as consuming the spiritual substance of their victims, through invisible means, sucking up their souls rather than literally dining on them. Also, while at first witches feed on their own relatives, those who have sucked up, and thus gained the power of, a large number of souls can eventually become powerful enough to attack almost anyone. It is the responsibility of chiefs to thwart their evil plans, using their own ndoki. However, as Wyatt MacGaffey emphasizes (1977, 1986, 1988), the difference between a chief and a witch is merely one of motive: witches are simply those who use their nocturnal powers for their own selfish purposes, greed or envy rather than the good of the community. And since the latter is a notoriously slippery concept, while no one without kindokiis of any real public account, no one with it is entirely above suspicion.

There are two key differences, though, with Tiv akombo, and these appear to be linked. One is that Kongo minkisitend to become personified. They have not only names and histories, but minds and intentions of their own. This is because their powers are really those of ancestral ghosts: most nkisistatuettes, in fact, contained in their chests both a series of medicinal ingredients which gave them their specific capacities for action (cf. Graeber, 1995) and grave dirt, which effected their connection with the dead. The second difference is that they tend to act largely when someone intentionally provokes them. While the Tiv might say that one who unintentionally offends an akombo‘pierces’ it, with minkisithis was no mere metaphor. Those operating nkisiwould often quite literally drive nails into the object to provoke it into action. This was not, I should stress, at all like driving pins into a voodoo doll, since the idea was to provoke the nkisito anger – though Wyatt MacGaffey (1986) stresses that, in a larger sense, the figures represented both the aggressor and the victim simultaneously, the assumption that the infliction of suffering creates a kind of unity between the two.

Even chiefly office could be drawn into the same logic. In much of Central Africa, leopards were symbols of royal power. So too here. One 19th-century notebook (no. 45, MacGaffey, 1986: 159) describes how, should someone kill a leopard, a man wishing to be invested in an important chiefly title might rush to the scene to ‘desecrate its tail’ by stepping on it. This was a period in which such titles could be acquired fairly easily by men who had gained fortunes in trade; after desecrating the object, the man could proceed to acquire the title through what is a kind of ‘purchase’ which might typically involve, for example, the payment of 10 lives ‘by day’ (slaves delivered to the current holder), and 10 ‘by night’ (members of the chief ’s own kin group killed by witchcraft; cf. Vansina, 1973).

The following gives something of the flavor of their power (note that a ngangais curer and keeper of minkisi; bangangais the plural):

Lunkanka is a nkisi in a statue and it is extremely fierce and strong. It came from Mongo, where many of our forebears used to go to compose it, but now its banganga [keepers] have all died out. When it had a nganga it was very strong, and so it destroyed whole villages. Its strength lay in seizing [its victims], crushing their chests, making them bleed from the nose and excrete pus; driving knives into their chests, twisting necks, breaking arms and legs, knotting their intestines, giving them nightmares, discovering witches in the village, stifling a man’s breathing and so on. When it was known that Lunkanka was exceedingly powerful, a great many people trusted it for healing, placing oaths and cursing witches and magicians, and so on. (in MacGaffey, 1991: 127)

The text goes on to explain that if two men make an agreement – say, one agreed to be the other’s client, or pawn, and thus be bound to his village – they might both drive nails into Lunkanka to seal the agreement; the nkisiwould then act as its power of enforcement. According to Wyatt MacGaffey (1987), in the 19th century every aspect of BaKongo economic life, from the policing of marketplaces to the protection of property rights to the enforcement of contracts, was carried out through the medium of nkisi, and the nkisiso employed were, in every case, forms of crystallized violence and affliction.

The underlying logic seems to have a remarkable similarity to social contract theories being created in Europe around the same time: MacGaffey has even found KiKongo texts which celebrate the existence of nkisias a way of preventing a war of all against all.[12] Once again, there is a striking parallelism in underlying assumptions: in this case, the same background of competitive market exchange, the same assumption that (at least outside of kin relations) social peace is therefore a matter of agreements, particular agreements to respect one another’s property, that must be enforced by an overarching power of violence. The main difference seems to lie in the assumed reasons why such violence is necessary. The Judaeo-Christian tradition goes back at least to Augustine (himself an African), having been based, as authors like Sahlins have much emphasized (2000), on an assumption that human desires are in their essence insatiable. Since we can never have enough pleasure, power, or especially material wealth, and since resources are inherently limited, we are all necessarily in a state of competition with one another. The state, according to Augustine, embodies reason, which is divine. It is also a providential institution which by threatening punishment turns our own base egoism – especially our fear of pain – against us to maintain order. Hobbes (1651) merely secularized the picture, eliminating the part about the endless desires being a punishment for original sin, but keeping the basic structure; then Adam Smith, Enlightenment optimist that he was, brought divine providence back in to argue that God had actually arranged things so that even our competitive desires will ultimately work for the benefit of all. In every case, though, the western tradition seems to combine two features: the assumption that humans are corrupted by limitless desires, and an insistent effort to imagine some form of power or authority (Reason, God, the State ...) which is not corrupted by desire, and hence inherently benevolent. God must be just (despite all appearances to the contrary); a rational man can rise above bodily passions; it should at least be possible to have rulers who are not interested in their own aggrandizement but only about the public welfare. The result was that the effects of power tend to be endlessly euphemized or explained away. African cosmological systems seemed to lack both features: probably, because they were less inclined to see human motivation as, say, a desire for wealth, or pleasures that could be abstracted from, or imagined independently from, the social relations in which they were realized. They tended to assume that what people desired was power itself.[13] It was impossible thus to imagine a form of political power which was not – at least partly – constituted by the very form of evil which the western tradition saw as the means to transcend.[14] Perhaps for this reason, what Europeans nervously euphemized was exactly what Africans seemed to selfconsciously exaggerate. One might consider here the difference between the famous ‘divine’ kingships of much of Africa, whose subjects insisted that any ruler who became weak or frail would be promptly killed, but in which, in actual fact, this seems to have happened only rarely, with an institution like Augustine’s Roman Empire, which claimed to be the embodiment of rational law and guardian of public order but whose actual rulers murdered one another with such savage consistency that it is almost impossible to come up with an example of an emperor who died a natural death. Similarly, in 17thand 18th-century Europe, African states developed a reputation for being extraordinarily bloodthirsty, since their representatives and subjects never saw any point in disguising the essentially murderous nature of state power. This despite the fact that the actual scale of killing even by the Ganda or Zulu states was negligible in comparison with the devastation wreaked in wars within Europe at the same time – not even to mention with what Europeans were prepared to do to anybody else.


Another way to understand the difference is to look at the contrasting ways in which power was seen to take on material substance or tangible form. For Pietz’s merchants, of course, the emphasis was on material valuables, beautiful or fascinating objects – or sometimes artificially beautified people – and their powers to enchant or attract. The value of an object was its power. In the African cases we have looked at, at least, power is imagined above all as a material substance inside the body: tsav, ndoki. This is entirely in keeping with the distinctions sketched out earlier, but it also has an interesting corollary, which is, in a sense, to systematically subvert that principle of representation which is the very logical basis of any system of legitimate authority. Here I can only refer to an argument I have made at greater length (Graeber, 1997): that any system in which one member of a group can claim to represent the group as a whole necessarily entails setting that member off in a way resembling the Durkheimian notion of the sacred, as set apart from the stuffs and substances of the material world, even, to a certain degree, abstracted from it. Much of the etiquette surrounding figures of authority always tends to center on denial of ways in which the body is continuous with the world; the tacit image is always that of an autonomous being who needs nothing. The ideal of the rational, disinterested state seems to be just one particular local variation of this very common theme, inherent, I have argued, to any real notion of hierarchy.

It is not that the logic of hierarchy is not present – one might well argue it always is, in some form or another – but rather that things seem to work in such a way as to constantly subvert it. It seems to me one can’t really understand even the famous Tiv system of spheres of exchange without taking this into account. The system, as mapped out by Paul Bohannan in an essay in 1955 (see also Bohannan, 1959), is really quite simple. Everything considered worth exchanging, all things of value, fell into one of three categories; things of each category could, ordinarily, be exchanged only for each other. The resulting spheres of exchange formed a hierarchy. At the bottom were everyday goods like food or tools or cooking oil, which could be contributed to kin or friends or sold in local markets. Next up were prestige goods such as brass rods, slaves, a certain white cloth, and magical services such as those provided by owners of akombo. The highest consisted in nothing but rights in women, since all marriages, before the colonial period, were considered exchanges of one woman for another – or more exactly, of their reproductive powers – and there was a complicated system of ‘wards’ whereby male heads of household could acquire rights in women seen as owed them in one way or another and marry them off in exchange for new wives, even if they did not have an unmarried sister or daughter of their own. On the other hand, division between spheres was never absolute. It waspossible to convert food into valuables, if one found someone sufficiently desperate for food, or, under other circumstances, valuables into additional wives. To do so took a ‘strong heart’, which according to Bohannan was inherently admirable (‘morally positive’), though one has to imagine somewhat ambivalently so, since having a strong heart meant, precisely, that one had that yellow substance on one’s heart which also made one a witch.[15]

Obviously, the system is all about male control of women. The sort of goods that are largely produced and marketed by women are relegated to the most humble category; those controlled by men rank higher; the highest sphere consists solely of men’s rights in the women themselves. At the same time, one could say as one moves up the spheres, men are increasingly gaining control of the capacity to create social form (households, descent, genealogy ...); converting upwards from food and tools that can merely keep people alive, to objects with the capacity to assemble clientages, and then finally, to the power to create descent itself. Since, after all, when one assembles wives and wards one is not, technically speaking, trafficking in women so much as in their reproductive capacities. All of this one does by manipulating debt, in its various manifestations, placing others in a position of obligation. This in turn makes it easier to understand what’s really going on with stories about witchcraft and the flesh debt, what I would propose should really be considered the fourth sphere, since it marks the ultimate fate of those with ‘strong hearts’. This is where the whole system collapses on itself, the direction is utterly reversed: since those who are most successful in manipulating networks of debt to gain such powers over creation are discovered, here, to be in a position of limitless debt themselves, and hence forced to consume the very human substance the system is ostensibly concerned to produce. In striking contrast with the western version, the insatiable desire for consumption, when it does appear, is not a desire for wealth but for the direct consumption of human beings, indistinguishable from the political power which, in the European version, is usually imagined as the only thing capable of controlling it.

Now, all this might seem appropriate to an egalitarian society like the Tiv, which one would expect to be somewhat ambivalent about the nature of social power and authority. The surprising thing, then, is how much of this is reproduced, almost exactly unchanged, in the BaKongo material, where the political situation was so different. Granted it was not entirely different – this was an area where centralized authority had been being effectively broken down for generations (Ekholm Friedman, 1991); but the parallels are striking, even down to the small details like the payments ‘by day’ and ‘by night’. The few salient differences do seem to reflect a greater acceptance of social hierarchy among the BaKongo (at least in principle). There is more of an overt willingness to see kindoki as capable of serving the common good, and, significantly, I think, also a tendency to treat the whole matter of witchcraft more abstractly. While there is some occasional talk of feasting on disinterred bodies, the usual imagery is of a kind of disembodied vampiric power feeding off the soul-stuff of its victims – which, if nothing else, shows a reluctance to challenge the fundamental logic of representation through abstraction on which any system of legitimate rule must, it would seem, eventually rest. Ultimately, though, these are minor differences.


The first Portuguese and Dutch sources, as I mentioned, seem entirely oblivious to all this. Caught up as they are with their own newfound materialism, questions of economic value – and in particular, value in exchange – were the only ones that really concerned them. The result is that, oddly enough, at the moment when Hobbes was writing his famous theory of the social contract (1651), he seems to have been entirely unaware that, in Africa, social contracts not so different from the sort he imagined were still being made on a regular basis.

This brings one back to the questions with which we began: about the nature of social creativity. The main way of talking about such matters in the western intellectual tradition, for the last several centuries, has been precisely through the idiom of contracts, social or otherwise. As I mentioned at the start of the article, Marcel Mauss claimed that his essay on the gift (1925), in fact, was really part of a much larger project on the origins of the notion of the contract and of the notion of contractual obligation. His conclusion – a rather striking one – was that the most elementary form of social contract was, in fact, communism: an open-ended agreement between two groups, or even two individuals, to provide for the other; within which, even access to one another’s possessions followed the principle of ‘from each according to their abilities, to each according to their needs’. Originally, he argued, there were two possibilities: total war, or ‘total reciprocity’. The latter informed everything from moiety structures (where those on one side of a village can only marry the daughters of those on the other, or only eat food grown on the other, or only the others can bury their dead) to relations of individualistic communism such as applied between close friends, or in-laws or, in our own society, husband and wife. This later gets refracted into various more specific forms of gift relation, and then of course eventually you get the market, but ‘total reciprocity’ remains the kind of baseline of sociality, even to the present day. This is why, Mauss suggests, wage labor contracts seem so unsatisfying to those on the receiving end; there is still that underlying assumption that voluntary agreements (such as, say, marriage) should involve an open-ended commitment to respond to one another’s needs.

Alain Caillé (2000) sums up the difference between the first sort of contract, and gift relations in general, and the more familiar contract as between ‘conditional unconditionality’ and ‘unconditional conditionality’. The first is an unlimited commitment, but either party is free to break it off at any time; the second specifies precisely what is owed by each party, no more and no less – but within that, each party is absolutely bound. My own work on trade currencies, and in particular what happened to beads or shell currency once they left the circuits of the trade (Graeber, 2001), revealed some striking patterns. Everything seemed to turn on the presence or absence of an internal market. In North America, belts of wampum, originally acquired in the fur trade, were never used as money by indigenous people when dealing with each other (in fact there were no market relations between indigenous people of any kind at all); instead they became a key element in the construction of social peace. The Iroquois Confederation, for example, saw themselves as emerging from a kind of Hobbesian period of war of all against all, but it was caused not by competition over wealth and power but by the power of grief and mourning, which twisted humans into monstrous creatures craving vengeance and destruction. Wampum, in comparison, was never seen as causing anyone to hurt anybody else. Wampum was crystallized peace, a substance of light and beauty with the power to heal and open those wounded and cramped by rage; gifts of wampum cleared the way to open-ended relations of mutual responsibility of just the sort Mauss seemed to have in mind (1947). In Madagascar, in contrast, where buying and selling was everywhere, trade beads and, later, ornaments made of melted silver coins, became elements in charms (ody, sampy, and so on) that operated very much like West African fetishes: they might not have embodied diseases, quite, but they were capable of being highly punitive in their effects. If anything, in Madagascar the Hobbesian logic becomes much more explicit, because this was also the way one created sovereign power and the state.

Here again I can only summarize a much more elaborate argument (Graeber, 1995, 2001), but the gist goes something like this. Silver coins, which came into Madagascar largely through the slave trade, and which were melted down to create ornaments and broken up to create smaller denominations of currency which people actually used in daily life, were also used, in Imerina, to create the power of kings. Every major event at which the ruler appeared was marked by ‘giving hasina’, the presentation of unbroken silver coins by representatives of the people to the king – unbroken to represent the unity of the kingdom created by this act of recognition. The ultimate message was that by doing so, the people created royal power, in exactly the way that one created a charm or fetish. Even more critically, in the Merina kingdom, every time two people came to any sort of business agreement, or for that matter every time members of a community came to agreement on the disposal of property or the maintenance of irrigation works, they invariably sealed the contract by ‘giving hasina’ to the king (Graeber, 1995: 96–109), recreating that power of violence which bound them to their contractual obligations.[16] It is not that contracts of the more open-ended, Maussian variety, did not exist in Madagascar or for that matter in West Africa. Most often, they are referred to in the literature under the rubric of ‘rituals of blood brotherhood’. In Malagasy these are called fatidra. In 19th-century texts gathered by missionaries – the Tantara ny Andriana (Callet, 1908: 851) or Fomba Gasy (Cousins, 1968: 93–4; see also Ellis, 1838, vol. 1: 187–90; Sibree, 1875, 1897) – they are indeed treated as the most basic, even primordial, form of contract (most business partners, for instance, seem to have been bound together in this way). The two parties would each put a little of their blood together in a piece of liver, eat the liver, and then would swear always to be responsive to one another’s needs, never refuse help in a crisis, never refuse food when the other is hungry, and so on. However, the actual body of the oath takes the form of imprecations, invoking an invisible spirit created by the ritual and calling on it to wreak every sort of disaster and havoc upon them should they ever fail to live up to these obligations. The same is true of the creation of communal ties: people insisted (in fact, they still insist) that even before there were kings, those creating new communities would begin by ‘giving hasina’ to some stone or tree or other object which would then have the power to enforce their communal obligations, to punish or at least expel those who did not respect the social contract.

When Mauss described ‘total reciprocity’ he was thinking of the sort of agreements that would be made in the complete absence of market institutions: here, we are dealing with societies deeply entangled in market relations; in fact, often, relations between people had little else in common. It is hard to escape the conclusion that the generic power of money – as the one thing already binding the parties together – itself became the model for that invisible power which was, as it were, turned back against itself to maintain commitments even when it might have been in one party’s short-term financial interest not to. Hence, even the ‘individualistic communism’ of blood brotherhood ends up subsumed under that same logic.

The comparison of North America and Madagascar is telling, I think, because in both cases stuff which is an embodiment of pure value, and which is seen as coming from very far away, becomes the basic medium for the creation of new social ties – for social creativity. The Iroquois of the Six Nations used wampum to create peace, but in fact what we call society was, for them, peace: the ‘League of the Iroquois’ was called ‘The Great Peace’, and the presentation of wampum became the medium for creating all sorts of contracts, mutual agreements and new institutional forms (see Graeber, 2001: 125–6, 132–4). In the Malagasy – and also African – cases we are looking at the media for the creation of agreements, communities, even kingdoms.

That this should so often involve manipulation of objects of alien, and apparently universal, value should perhaps come as no surprise. No doubt we are just dealing with the familiar structural principle that a social field, or logical domain, cannot be constituted except in relation to something which is not part of it, something transcendent or anyway alien. A constitution cannot be created by constitutional means; beings capable of establishing a system of justice cannot themselves be bound by that system of justice; always one needs something else. This much is straightforward enough. But it’s also important to stress that these objects were, ultimately, only the medium. Hence what they are is ultimately somewhat arbitrary: one can use valuable objects from faraway lands, or one can, in fact, use pretty much any random object one lays one’s hands on, ‘a Lion’s Tail ... a Bird’s Feather ... a Pebble, a Bit of Rag’. In this, Pietz’s sources had a point, because this is exactly the moment where the arbitrariness of value comes fully into focus. Because really, creativity is not an aspect of the objects at all, it’s a dimension of action. In this sense the new does in fact emerge from the old, and the numinous, alien nature of the object is really the degree to which it reflects on that aspect of our own actions that is, in a sense, alien to ourselves.


Here of course is where we start, finally, moving in the direction of the Marxian notion of the fetish: objects which seem to take on human qualities which are, ultimately, really derived from the actors themselves.

Not that we are speaking of pure mystification here. As I have tried to demonstrate in my analysis of the Merina Royal Bath ceremony (2001: 232–9), and hasina ritual in general, people were not entirely unaware that it was the ritual that made the king, that what constructed royal power was not the coin, but the action of giving it. This was tacit in the ritual itself, and stated explicitly just off-stage. Similarly, Malagasy charms involved the giving of an oath or pledge by those protected by them, or over whom they had power; without that, it was simply a powerless object. On the other hand, once given, the object was treated as having a power of its own. Something similar seems to have been widely recognized by West African ‘fetishists’. In fact, if one looks over the literature surveyed by Pietz, one sees the exact same emphasis on action: here, taking a collective oath could be called ‘making’ or ‘drinking’ or ‘eating’ fetish, phrases which appear to be direct translations from African languages. A fetish is something one makes, or does:

Obligatory Swearing they also call, making of Fetiche’s; If any Obligation is to be confirmed, their Phrase is, let us as a farther Confirmation make Fetiche’s. When they drink the Oath-Draught, ‘tis usually accompanied with an Imprecation, that the Fetiche may kill them if they do not perform the Contents of their Obligation. (Bosman, 1967 [1705]: 149)

The basic sequence here – people create (‘make’) something; then they act as if that thing has power over them – is of course just the sort of thing Marx was thinking of when he spoke of ‘fetishism’. There are two curious things here. One is that those involved seemed not entirely unaware that this was happening: both that these objects were constructed and, at the same time, that they came to have some kind of power over those who constructed them. This is very important I think and I will try to consider the full implications in a moment. The other curious thing is that Pietz does not even consider any of this. In fact, even when he turns to look at Marx’s own work (Pietz, 1993), Pietz considers every definition of fetishism, every aspect, other than the simplest and most common one: that ‘fetishism’ occurs when human beings end up bowing down before and worshipping that which they have themselves created.[17]

Now, this is a peculiar oversight.

The reason seems to lie in the structure of Pietz’s argument: that ‘the fetish’ is a concept that emerged within a peculiar intercultural space in which neither existing European, nor existing African categories really applied. He calls it a ‘space of cultural revolution’,[18] in which the ‘conceptualities, habits and life forms, and value systems’ of a number of radically different social systems (feudal Christianity, proto-capitalist mercantilism, African lineage systems and so on) were suddenly juxtaposed and forced to come to terms with one another. It was therefore a space of continual innovation and cultural creativity, as each side found their existing practices and categories inadequate in dealing with the others, that a kind of pidgin culture emerged, particularly among figures like the tangomaos, ‘Portuguese speaking adventurers and traders who made their home on the Guinea mainland, in defiance of the orders of the crown, and who married there and established mulatto families’ (Donelha in Pietz, 1987: 39).

In this situation, Pietz argues, the standard Christian rubrics for dealing with alien religious practices just didn’t seem to work. The most common of these had been ‘idolatry’. Pagans worshipped idols. Idols were material images, made by human beings, that represented invisible powers – conceived as a god, a spirit, though the Christian knew them to be really demons – with whom the worshipper came into relation by some kind of verbal compact. Here was the key difference with fetishism. Fetishes – at least in the descriptions of the first Portuguese and Dutch traders – did not represent anything; they were material objects seen as having power in and of themselves; imaginary products, in effect, of the merchants’ own materialistic cosmology. As Wyatt MacGaffey noted early on (1994), this materialistic emphasis was precisely what was missing from the way Africans talked about these things (making one wonder how much one is really talking about a ‘pidgin culture’ at all). Some of the items labeled ‘fetishes’ took the form of images, many did not; but verbal compacts and invisible spirits were almost invariably involved. The foreign missionaries who were the first to establish themselves in Imerina, for instance, did not hesitate to label their Merina equivalents ‘idols’ instead of ‘fetishes’, even though sampyonly rarely took representational form. The difference between Malagasy ‘idols’ and West African ‘fetishes’ seems to be quite simply that the former were first named by missionaries and the latter mainly by merchants, men really only concerned with exchange and questions of material value. Questions of production, creation, let alone the production or creation of social relations, were simply of little interest to Pietz’s sources. As a result, what is to me, at least, the most fascinating aspect of the whole complex of ideas drops away: that is, the notion of ‘making fetish’

that by a form of collective investment one can, in effect, create a new god on the spot

even though this seems to be what really startled European newcomers to Africa, and ultimately caused them to launch into peculiar fantasies about people who worship the first thing they see in the morning. It was the improvisational quality of the ritual surrounding fetishes which made it appear to them that in many African societies, at least, it was particularly in the domain of religion – what should have been the domain of eternal verities – that everything was up for grabs, precisely because this was also the main locus for social creativity. In this sense, as we will see, the issue is not so much that these were objects that existed in a ‘space of revolution’, but rather that they were themselves revolutionary objects.


So what is a fetish, then?

A fetish is a god under process of construction.

At least, if ‘fetish’ can still be used as a technical term in this context – and of course there is no consensus on this point – this is what I would suggest.

Fetishes exist precisely at the point where conventional distinctions between ‘magic’ and ‘religion’ become meaningless, where charms become deities. Frazer of course argued that magic is a technique, a way of humans trying to shape the world to their will – if only by mistaken techniques – while religion was instead a matter of submitting to an external authority.[19] For Durkheim, magic was ritual pursued for purely individual ends; it becomes religion when it acquires a church, a congregation, because religion is about society. Fetishism then is the point where each slips into the other: where objects we have created or appropriated for our own purposes suddenly come to be seen as powers imposed on us, precisely at the moment where they come to embody some newly created social bond.

This may sound rather abstract but if one looks carefully at the ethnographic evidence, this is exactly what happens. Ordinary life in rural Madagascar is still full of different sorts of ‘medicine’ (fanafody), a term which covers everything from herbal infusions to charms with the power to bring bolts of lightning down on an enemy’s head. Most people know how to make or work one or two sorts, or at the very least, are willing to allow others to speculate that they might. The simplest charms are improvised for a specific occasion, others are more permanent: very important, older charms which affect whole communities – charms which guard the crops against hail, or protect villages from thieves – which have names and histories and keepers, or even have to be renewed (like kings) by periodic sacrifice. In earlier centuries, certain of these went on to take on a more general role as protectors of communities, and these came to be known as ‘sampy’. They were ultimately collections of bits of rare wood, beads and silver ornaments, kept hidden under cloth or in boxes, usually with little houses of their own; sometimes they spoke through their keepers; they had names and stories, wills and desires, they received homage, gave blessings, imposed taboos. They were, in other words, very much like gods. Especially so when they came to be adopted into the royal pantheon: at any given time, the king would adopt 12 or so to be the guardians of the kingdom, and these would be borne before the royal army during campaigns; they were present at important rituals; their ritual days were national holidays; their keepers a de facto priesthood. These were also the ‘idols’ – with names like Kelimalaza, Manjakatsiroa, Ravololona – that so offended the English missionaries in the 19th century. Yet this was also a very unstable pantheon. If these were gods – and in fact they were called ‘gods’ (Andriamanitra, the same word used for the Creator, or later the Christian God) – their hold on godhood seemed remarkably tenuous. New ones would appear; older ones might slip into obscurity, or else be exposed as frauds or witchcraft and purged from the pantheon. There literally was no clear line between ordinary ‘magic’ and deities, but for that reason, the deities were a constant process of construction. They were not seen as representing timeless essences,[20] but powers that had proved, at least for the moment, effective and benevolent.

West African ‘fetishes’ were not exactly the same as Merina sampy – they tended to be more destructive in their powers, more caught up in the logic of affliction; there were other subtle differences – but there too we find the same continuum between casual charms and quasi-deities, the same sense of objects created through human actions, property that could be possessed, inherited, even bought and sold; tools, but at the same time objects of obeisance and adoration, capable of acting with potentially devastating autonomy.

So what does all this strange theology have to do with social creativity per se? Here I think we can finally return to Marx.

For Marx, the ‘fetishism of commodities’ was one particular instance of a much more general phenomenon of ‘alienation’. Collectively, human beings create their worlds, but owing to the extraordinary complexity of how all this creative activity is coordinated socially, no one can really keep track of the process, let alone take control of it. As a result, we are constantly confronting our own actions and creations as if they were alien powers. Fetishism is simply when this happens to material objects. Like African fetishists, the argument goes, we end up making things and then treating them like gods.

The actual argument in Capital(1967 [1867]: chp. 2) is of course much more complicated. In it, Marx is mainly making a point about value.[21] For Marx, value always comes from labor, or, to be more precise, value is the symbolic form through which our labors become meaningful to us by becoming part of some larger social system. Yet in capitalism, consumers tend to see the value of commodities as somehow inhering in the objects themselves, rather than in the human efforts required to put those qualities in them. We are surrounded by objects designed and produced for our pleasure or convenience. They embody the intentions of people who anticipated our needs and desires and sank their energies into creating objects that would satisfy them; but owing to the workings of the market system, we normally do not have the slightest idea who any of those people are or how they went about it. Therefore, all those intentions seem as if they are properties of the object itself; objects therefore seem to be things we can enter into personal relations with; we become indignant, hit them or kick them when they don’t work, and so on. Actually, capitalism seems rife with such subject/object reversals: capital grows, money is always fleeing one market and seeking out another, pork bellies doing this, the bond market doing that ... In every case, what is happening is that we are operating in a system so complicated we could not possibly see all of it, so we mistake our own particular perspective on the whole, that little window we have on it, with the nature of the totality itself. Because from the point of view of the consumer, products might as well have simply jumped out into the market with a personal commitment to play their DVDs or vacuum their apartments; from the perspective of the businessman, money might as well be fleeing some markets, and so on.[22]

Now, all this jumbling of agency might seem innocent enough; particularly since if really challenged on the matter, few would defend the premise that commodities really have minds of their own, or that money really flees markets all of its own accord. For Marx this becomes dangerous for two reasons. First, because it obscures the process of how value is produced, which is of course very convenient to those who might wish to extract value. Money represents the value of labor, but wage laborers work to get money; it thus becomes a representation that brings into being what it represents; it is therefore easy to see it as the source of that value, or asvalue (since again, from the laborer’s perspective, it might as well be). In the same way tokens of honor (rather than honorable actions) can come to seem the source of prestige; tokens of grace (rather than acts of devotion) the source of divine favor; tokens of conviviality become the source of fun; and so on. Second, all of this makes it much easier to treat the ‘laws of the market’, or tendencies of whatever system it may be, as natural, immutable, and therefore completely outside any possibility of human intervention. This is of course exactly what happens in the case of capitalism, even – perhaps especially – when one steps out of one’s immediate situated perspective and tries to talk about the system as a whole. Not only are the laws of the market taken to be immutable, the creation of material objects is assumed to be the whole point, the commodities themselves the only human value, so that in contemporary Africa, for example, one can witness the bizarre spectacle of government officials and their World Bank advisors declaring that the fact that in some areas half of the population is dying of AIDS is a real problem because it is going to have devastating effects on ‘the economy’ – apparently oblivious to the fact that until fairly recently ‘the economy’ was universally assumed to be the way we distribute material goods so as to keep people alive.

The emphasis on value theory makes it easier to understand the strange disparity – with which I began – between Marx’s view of material production and the way he talks about what I have been calling social creativity, or revolution. In producing a house or a chair, one first imagines something and then tries to bring it into being. In fomenting revolution, one must never do this.[23] The main reason for the disparity seems to be that, as Hans Joas points out, Marx does seem to reduce human creativity to two modalities: production (which happens all the time), or revolution (which happens only occasionally). Not in principle: in The German Ideology(1970 [1846]), for example, Marx states very clearly that the production of material goods was always, at the same time, the production of people and social relations, and all this was a creative process and therefore in constant transformation. But Joas is right to say that in Marx’s concrete analyses of events of his own day, all of this does rather tend to fade away. Social creativity tends to get reduced to political action, even, to dramatic, revolutionary change.

One reason is that in carrying out this kind of value analysis, one has to assume that the social system surrounding production is pretty much stable. Let me illustrate. To say that in fetishizing commodities, or money, one is confusing one’s partial perspective on a system with the nature of the system as a whole, does at the very least imply that a whole system exists and that it is possible to know something about it. In the case of a market system this is a perfectly reasonable claim: all economic study is premised on the assumption that there are things called ‘markets’ and that it is possible to understand something about how they work. Presumably the knowledge required is not comprehensive: one need not know exactly who designed and produced the pack of cigarettes or palm pilot in one’s pocket in order to avoid fetishizing it.[24] One simply needs to know how these things generally tend to work, the logic of the system, how human energies are mobilized, organized, and end up embodied in objects. But this, in turn, implies the system tends to work roughly the same way over time. What if it doesn’t? What if it is in a process of transformation? What if, to take an extreme example, the system in question does not yet exist, because you are, in fact, trying to bring it into being through that very act of fetishism?

In the case of many of these African fetishes, this was exactly what was happening. Merchants who ‘drank’ or ‘made fetish’ together might not have been creating a vast market system, but the point was usually to create a small one: stipulating terms and rates of exchange, rules of credit and regimes of property that could then be the basis of ongoing transactions. Even when fetishes were not explicitly about establishing contracts of one sort or another, they were almost invariably the basis for creating something new: congregations, new social relations, new communities. Hence any ‘totality’ involved was, at least at first, virtual, imaginary, and prospective. What is more – and this is the really crucial point – it was an imaginary totality that could only come into real existence if everyone acted as if the fetish object actually did have subjective qualities. In the case of contracts, this means: act as if it really will punish them for breaking the rules.

These were, in other words, revolutionary moments. They involved the creation of something new. They might not have been moments of total transformation, but realistically, it is not as if any transformation is ever really total. Every act of social creativity is to some degree revolutionary, unprecedented: from establishing a friendship to nationalizing a banking system. None are completely so. These things are always a matter of degree.[25]

Yet this is precisely where wefind the logic of fetishism cropping up – even the origin of the word ‘fetish’ – and it doesn’t seem to be misrepresenting anything. Of course it would also be going too far to say that the fetishistic view is simply true: Lunkanka cannot really tie anyone’s intestines into knots; Ravololona cannot really prevent hail from falling on anyone’s crops. As I have remarked elsewhere (Graeber, 2001), ultimately we are probably just dealing here with the paradox of power, power being something which exists only if other people think it does; a paradox that I have also argued lies also at the core of magic, which always seems to be surrounded by an aura of fraud, showmanship, and chicanery. But one could argue it is not just the paradox of power. It is also the paradox of creativity. This has always been one of the ironies of Marxism. Marx ultimately wanted to liberate human beings from everything that held back or denied them control of their creative capacities, by which he meantfirst and foremost, all forms of alienation. But what exactly would a free, non-alienated producer look like? It is never clear in Marx’s own work. Not exactly like an independent craftsperson, presumably, since the latter are usually caught in the shackles of tradition. Probably more like an artist, or a musician, or a poet, or even an author (like Marx himself). But when artists, musicians, poets, or authors describe their own experience of creativity, they almost invariably begin evoking just the sort of subject/object reversals which Marx saw as typical of fetishism: almost never do they see themselves as anything like an architect rationally calculating dimensions and imposing their will on the world. Instead one almost invariably hears how they feel they are vehicles for some kind of inspiration coming from outside, how they lose themselves, fragment themselves, leave portions of themselves in their products. All the more so with social creativity: it seems no coincidence that Mauss’ work on the ‘origins of the idea of the contract’ in The Gift(1965) led him to meditate endlessly on exactly these kind of subject/object reversals, with gifts and givers becoming hopelessly entangled. Put this way, it might seem to lead to a genuine dilemma. Is non-fetishized consciousness possible? If so, would we even want it?

In fact, the dilemma is illusory. If fetishism is, at root, our tendency to see our own actions and creations as having power over us, how can we treat it as an intellectual mistake? Our actions and creations do have power over us. This is simply true. Even for a painter, every stroke one makes is a commitment of a sort. It affects what she can do afterwards. In fact this becomes all the more true the less one becomes caught in the shackles of tradition. Even in the freest of societies we would presumably feel bound by our commitments to others. Even under Castoriadis’ ideal of autonomy, where no one would have to operate within institutions whose rules they had not themselves, collectively, created, we are still creating rules and then allowing them to have power over us. If discussion of such matters tends towards metaphoric inversions, it is because it involves a juxtaposition of something that (on some level) everyone understands – that we tend to become the slaves of our own creations – and something no one really understands, how exactly it is we are able to create new things to begin with.

If so, the real question is how one gets from this perfectly innocuous level to the kind of complete insanity where the best reason one can come up with to regret the death of millions is because of its effects on the economy. The key factor would appear to be, not whether one sees things as a bit topsy-turvy from one’s immediate perspective – something like this seems inevitable, both in the realization of value, which always seems to operate through concrete symbolic forms, and especially in moments of transformation or creativity – but rather, whether one has the capacity to at least occasionally step into some overarching perspective from which the machinery is visible, and one can see that all these apparently fixed objects are really part of an ongoing process of construction. Or at the very least, whether one is not trapped in an overarching perspective which insists they are not. The danger comes when fetishism gives way to theology, the absolute assurance that the gods are real.

Consider again the confrontation between Pietz’s European merchant adventurers in the 16th and 17th centuries, and their West African counterparts – many being merchants themselves. I have already argued that while both arrived with a number of broadly shared cosmological assumptions – for instance, that we live in a fallen world, that the human condition is fundamentally one of suffering – there were also a number of profound differences which the Europeans found deeply disturbing. (Whether their African partners were equally disturbed by the encounter we are not in a position to know.) To reduce the matter to something of a caricature: the European merchants were, as Pietz stressed, budding materialists. They were Christians, but for the most part their interest in theological questions seems to have been negligible; the main effect of their Christian faith was to guarantee the absolute assurance that, whatever spiritual ideas Africans had, insofar as they were not Christian, they had to be profoundly mistaken. This in turn had an effect when they confronted what they really cared about: matters of trade, material wealth, and economic value. Confronted with abundant evidence of the arbitrariness of value, they instead fell back on the position that Africans themselves were arbitrary: they were fetishists, willing to ascribe divine status to a completely random collection of material objects.

In the European accounts, social relations tend to disappear. They were simply of no interest. For them there was therefore virtually nothing in between God and the world of material objects. But the Europeans could at least compliment themselves that, unlike Africans, they managed to keep the two apart. Of course they were wrong; the whole thing was largely a projection; they were in fact already well on the way to the kind of fetishism described by Marx where social relations, for the very reason that they are made to disappear, end up getting projected onto objects. All this was in dramatic contradistinction with the Africans, for whom social relations were everything. As Jane Guyer (1993, Guyer and Belinga, 1995) has pointed out, conventional economic categories are hard to apply in such contexts, because people (rights in women’s fertility, authority over children, the loyalty of followers, disciples, recognition of titles, or status, or accomplishment) were the ultimate form of wealth. Material objects were interesting mainly insofar as they became entangled in social relations, or enabled one to create new ones. Since wealth and power could not, ultimately, be distinguished, there was no way to idealize government (which disturbed Europeans); it also made for an enchanted world – one in which, for that very reason, the mechanics of enchantment were never very far from the surface (which disturbed them even more). It was as if everything existed in that middle zone which the Europeans were trying to evacuate; everything was social, nothing was fixed, therefore everything was both material and spiritual simultaneously.

This was the zone in which we encounter the ‘fetish’. Now, it is probably true that most gods have always been in the process of construction. They exist at some point along the passage from an imaginary level of pure magic – where all powers are human powers, where all the tricks and mirrors are visible – to pure theology, with an absolute commitment to the principle that the constructive apparatus does not exist. But objects like akombo, minkisi, sampy– or, for that matter, the improvised ‘fetishes’ made of Bibles and bits of wood through which half-Portuguese tangomaos negotiated business deals – seem to have existed at a midpoint almost exactly in between. They were both human creations and alien powers, at the same time. In Marxist terms, they were fetishes from one perspective and, from another, not fetishized at all. Both perspectives were simultaneously available. But they were also mutually dependent. The remarkable thing is how much, even when the actors seem perfectly aware that they were constructing an illusion, they also seemed aware that the illusion was still required. It rather reminds one of the practice of shadow puppetry in Southeast Asia: the whole point is to create an illusion, the puppets themselves are supposed to be invisible, mere shadows on the screen, but if you observe actual performances, you usually find the audience is ranged around in a big circle so that many of them can only see the puppets and cannot actually see the illusion at all. There doesn’t seem to be a feeling that they are missing out on much. Nonetheless, it would not be a performance if the illusion did not take place.

This is what one might expect in a world of almost constant social creativity; in which few arrangements were fixed and permanent, and, even more, where there was little feeling that they really should be fixed and permanent; in which, in short, people were indeed in a constant process of imagining new social arrangements and then trying to bring them into being. Gods could be created, and discarded or fade away, because social arrangements themselves were never assumed to be immutable.

What does this teach us about the grand theoretical issues raised at the beginning? If nothing else, that if one takes seriously the idea of social creativity, one will probably have to abandon some of the dreams of certainty that have so enchanted the partisans both of holistic and individualistic models. No doubt processes of social creativity are, to some degree, unchartable. This is probably all for the best. Making it the centerpiece of a social theory regardless seems like it would be an increasingly important gesture, at a time when the heirs of Pietz’s merchants have managed to impose their strange, materialist theology on not just Africans but almost everyone, to the extent that human life itself can be seen as having no value except as a means to produce fetishized commodities.


David Graeber would like to thank the following people: Alain Caillé, Lauren Leve, Stuart Rockefeller, Jennifer Jackson, Michael Duncan, Massimo de Angelis, Hylton White, Ilana Gershon, Sylvia Federici, George Caffentzis, Marshall Sahlins, Rupert Stasch, Nhu Le, Anne Rawls, Yvonne Liu, Stephen Shukhaitis, Andrej Grubacic, Terence Turner, Michael Taussig, Mike Menser, Heather Gautney.


Abrahamsson, Hans (1952) The Origin of Death: Studies in African Mythology. Studia ethnographica Upsaliensia; III. Uppsala: Almquist.

Akiga (1939) Akiga’s Story: The Tiv Tribe as Seen by one of its Members(translated and annotated by Rupert East). London/New York: Oxford University Press (published for the International African Institute).

Apter, Emily and William Pietz, eds (1993) Fetishism as Cultural Discourse. Ithaca, NY: Cornell University Press.

Baudrillard, Jean (1972) Pour un critique de l’économie du Signe. Paris: Gallimard. Bhaskar, Roy (1989 [1979]) The Possibility of Naturalism. Hemel Hempstead: Harvester Wheatshaft.

Bhaskar, Roy (1993) Dialectic: The Pulse of Freedom. London: Verso. Bhaskar, Roy (1994) Plato Etc. London: Verso.

Bhaskar, Roy (2001) Reflections on Meta-Reality: Transcendence, Emancipation, and Everyday Life. New Delhi and Thousand Oaks, CA: Sage Publications.

Bohannan, Laura (1952) ‘A Genealogical Charter’, Africa22: 301–15.

Bohannan, Paul (1955) ‘Some Principles of Exchange and Investment among the Tiv’, American Anthropologist57: 60–7.

Bohannan, Paul (1957) Justice and Judgement among the Tiv. London: Oxford University Press.

Bohannan, Paul (1958) ‘Extra-Processual Events in Tiv Political Institutions’, American Anthropologist60: 1–12.

Bohannan, Paul (1959) ‘The Impact of Money on an African Subsistence Economy’, Journal of Economic History19: 491–503.

Bohannan, Paul and Laura Bohannan (1953) The Tiv of Central Nigeria. London: International African Institute.

Bohannan, Paul and Laura Bohannan (1968) Tiv Economy. Evanston, IL: Northwestern University Press.

Bohannan, Paul and Laura Bohannan (1969) A Source Notebook on Tiv Religion, 5 volumes. New Haven, CT: Human Relations Area Files.

Bosman, Willem (1967 [1705]) A New and Accurate Description of the Coast of Guinea, Divided into the Gold, the Slave, and the Ivory Coasts. New York: Barnes & Noble.

Caillé, Alain (2000) Anthropologie du Don: le Tiers Paradigme. Paris: Descelée de Brouwer.

Callet, R.P. (1908) Tantara ny Andriana eto Madagascar. Tananarive: Académie Malgache.

Castoriadis, Cornelius (1987) The Imaginary Institution of Society(trans. Kathleen Blamey). Cambridge: Polity Press.

Castoriadis, Cornelius (1991) Philosophy, Politics, Autonomy: Essays in Political Philosophy(ed. David Ames Curtis). New York: Oxford University Press.

Ciaramelli, Fabio (1998) ‘The Circle of the Origin’, in Lenore Langsdorf and Stephen L. Watson with Karen A. Smith (eds) Reinterpreting the Political: Continental Philosophy and Political Theory. Albany: State University of New York Press.

Cousins, William E. (1968)Fomba Gasy(ed. H. Randzavola). Tananarive: Imarivolanitra. Ekholm Friedman, Kajsa (1991) Catastrophe and Creation: The Transformation of an African Culture. Philadelphia, PA: Harwood Academic Publishers.

Ellis, William (1838) History of Madagascar, 2 vols. London: Fisher & Son.

Etnofoor(1990) Special issue on ‘Fetishism’, Etnofoor3(1).

Graeber, David (1995) ‘Catastrophe: Magic and History in Rural Madagascar’, endlessly unpublished manuscript.

Graeber, David (1997) ‘Manners, Deference and Private Property: The Generalization of Avoidance in Early Modern Europe’, Comparative Studies in Society and History 39(4): 694–728.

Graeber, David (2001) Toward an Anthropological Theory of Value: The False Coin of Our Own Dreams. New York: Palgrave.

Graeber, David (forthcoming) ‘Turning Modes of Production Inside Out: Or, Why Capitalism is a Transformation of Slavery (short version)’, Critique of Anthropology scheduled 2006.

Guyer, Jane I. (1993) ‘Wealth in People and Self-Realization in Equatorial Africa’, Man 28(2): 243–65.

Guyer, Jane I. and Samuel M. Eno Belinga (1995) ‘Wealth in People as Wealth in Knowledge: Accumulation and Composition in Equatorial Africa’, Journal of African History36: 91–120.

Hobbes, Thomas (1651) Leviathan. London.

Jameson, Fredric (1981) The Political Unconscious: Narrative as a Socially Symbolic Act. Ithaca, NY: Cornell University Press.

Joas, Hans (1993) ‘Institutionalization as a Creative Process: The Sociological Importance of Cornelius Castoriadis’ Political Philosophy’, in Hans Joas Pragmatism and Social Theory, pp. 154–71. Chicago, IL: University of Chicago Press.

Joas, Hans (1996) The Creativity of Action. Chicago, IL : University of Chicago Press. Joas, Hans (2000) The Genesis of Values. Cambridge: Polity Press.

MacGaffey, Wyatt (1977) ‘Fetishism Revisited: Kongo nkisi in Sociological Perspective’, Africa47(2): 172–184.

MacGaffey, Wyatt (1986) Religion and Society in Central Africa. Chicago, IL: University of Chicago Press.

MacGaffey, Wyatt (1987) ‘Lulendo: The Recovery of a Kongo Nkisi’,Ethnos52: 339–49. MacGaffey, Wyatt (1988) ‘Complexity, Astonishment and Power: The Visual

Vocabulary of Kongo Mikisi’, Journal of Southern African Studies14(2): 188–203. MacGaffey, Wyatt (1991) Art and Healing of the Bakongo Commented by Themselves: Minkisi from the Laman Collection. Stockholm: Folkens Museum – Etnografiska.

MacGaffey, Wyatt (1994) ‘African Objects and the Idea of the Fetish’, Res25: 123–31. Makar, Tesemchi (1994 [1936]) The History of Political Change among the Tiv in the 19th and 20th Centuries(foreword by A.P. Anyebe). Enugu, Nigeria: Fourth Dimension.

Marx, Karl (1970 [1846]) The German Ideology. New York: International Publishers. Marx, Karl (1973 [1857–8]) The Grundrisse. New York: Harper and Row.

Marx, Karl (1965 [1858]) Pre-Capitalist Economic Formations(trans. Jack Cohen). New York: International Publishers.

Marx, Karl (1970 [1859]) Contribution to the Critique of Political Economy. New York: International Publishers.

Marx, Karl (1967 [1867]) Capital, 3 volumes. New York: New World Paperbacks.

Mauss, Marcel (1925) ‘Essai sur le don. Forme et raison de l’échange dans les sociétés archaïques’, Annee sociologique, series 2, 1: 30–186.

Mauss, Marcel (1947) Manuel d’ethnographie. Paris: Payot.

Mauss, Marcel (1965) The Gift: Forms and Functions of Exchange in Archaic Societies (trans. I. Cunnison). New York: Norton.

Mauss, Marcel (1968–69) Oeuvres, 3 volumes. Paris: Editions de Minuit.

Munn, Nancy (1977) ‘The Spatiotemporal Transformations of Gawan canoes’, Journal de la Société des Océanistes33(54–55): 39–53.

Munn, Nancy (1986) The Fame of Gawa: A Symbolic Study of Value Transformation in a Massim (Papua New Guinea) Society. Cambridge: Cambridge University Press.

Negri, Antonio (1999) Insurgencies: Constituent Power and the Modern State(trans.

Maurizia Boscagli). Minneapolis: University of Minnesota Press.

Negri, Antonio and Michael Hardt (2000) Empire. Cambridge, MA: Harvard University Press.

Pietz, William (1985) ‘The Problem of the Fetish I’, RES: Journal of Anthropology and Aesthetics9: 5–17.

Pietz, William (1987) ‘The Problem of the Fetish II: The Origin of the Fetish’, RES: Journal of Anthropology and Aesthetics13: 23–45.

Pietz, William (1988) ‘The Problem of the Fetish IIIa: Bosman’s Guinea and the Enlightenment Theory of Fetishism’, RES: Journal of Anthropology and Aesthetics16: 105–23.

Pietz, William (1993) ‘Fetishism and Materialism: The Limits of Theory in Marx’, in Emily Apter and William Pietz (eds) Fetishism as Cultural Discourse. Ithaca, NY: Cornell University Press.

Pietz, William (1995a) ‘The Spirit of Civilization: Blood Sacrifice and Monetary Debt’, RES: Journal of Anthropology and Aesthetics28: 23–38.

Pietz, William (1995b) ‘Death and the Deodand: Accursed Objects and the Money Value of Human Life’, in Judith Farquar, Tomoko Masuzawa, and Carol Mavor (eds) (Un)Fixing Representation. Minneapolis: University of Minnesota Press.

Sahlins, Marshall (1972) Stone Age Economics. Chicago, IL: Aldine.

Sahlins, Marshall (2000) ‘The Sadness of Sweetness: The Native Anthropology of Western Cosmology’, in Marshall Sahlins Culture and Practice: Selected Essays. New York: Zone Books.

Sibree, James (1875) Madagascar: The Great African Island. London: Trübner & Co. Sibree, James (1897) ‘The Malagasy Custom of “Brotherhood by Blood”’, Antananarivo Annual and Malagasy Magazine21: 1–6.

Spyer, Patricia, ed. (1998) Border Fetishisms: Material Objects in Unstable Spaces. New York: Routledge.

Thornton, John (1987) ‘The Kingdom of Kongo, ca. 1390–1678: The Development of an African Social Formation’, Cahiers d’Études Africaines22: 325–42.

Tseayo, Justin Iyorbee (1975) Conflict and Incorporation in Nigera: The Integration of the Tiv. Zaria, Nigeria: Gaskiya.

Turner, Terence (1979) ‘Anthropology and the Politics of Indigenous Peoples’ Struggles’, Cambridge Anthropology5: 1–43.

Turner, Terence (1984) ‘Value, Production and Exploitation in Non-Capitalist Societies’, unpublished essay, based on a paper presented in the AAA 82nd Annual Meeting, Denver, Colorado. Ostensibly to appear in Critique of Pure Culture (forthcoming). New York: Berg Press.

Turner, Terence (1987) ‘The Kayapo of Southeastern Para’, unpublished monograph prepared for CEDI, Povos Indigenas do Brasil, Vol. VIII, Sul do Para, Part II.

Turner, Victor (1967) The Forest of Symbols: Aspects of Ndembu Ritual. Ithaca, NY: Cornell University Press.

Vansina, Jan (1973) The Tio Kingdom of the Middle Congo 1880–1892. Oxford: Oxford University Press.

[1] In the last chapter of Towards an Anthropological Theory of Value(Graeber, 2001), subtitled ‘The Problem of the Fetish, IIIb’. What follows in this article was, in large part, originally written for that chapter but ended up having to be cut for reasons of space. I was tempted to call it ‘The Problem of the Fetish, IIIc’ but it seemed unlikely many readers would get the joke.

[2] Especially in Italy. The most familiar representative for most readers in the Anglophone world is Toni Negri, but most of the ideas presented in Empire(Negri and Hardt, 2000) are the products of a long tradition involving many other writers and activists.

[3] For Castoriadis, history is no longer a matter of the development or play of productive or class forces but the work of ‘the imaginary, which is creation ex nihilo’, such that change is ‘the positing of a new type of behavior ... the institution of a new social rule ... the invention of a new object or a new form’ that is ‘an emergence or a production which cannot be deduced on the basis of a previous situation’ (Castoriadis, 1987: 3, 44).

[4] The tie to the Autonomist school can be seen by looking at the early work of Toni Negri, on constituent power (1999). Essentially he is trying to work out exactly the same problems: what is that popular power of creativity that emerges during moments of revolution and how would it be possible to institutionalize it?

[5] Actually there is no particular reason why gold should be a better medium of exchange than beads. Economists of course might make the argument that the supply of gold in the world is inherently limited, while glass beads can be manufactured in endless number; however, there is no way that European merchants of that day could have had the slightest idea how much of the earth’s crust was composed of gold; they saw it as precious because they got it with difficulty from very far away, just as Africans did beads.

[6] At least, there is none in his first three, best-known articles (Pietz, 1985, 1987, 1988). He does address West African ideas in two later articles concerned with debt and human sacrifice (1995a, 1995b); these essays, however, are concerned with a later historical period, and somewhat different sorts of questions.

[7] This would be one reason why Africans have been, from such an early period, comparatively receptive to religions like Christianity and Islam.

[8] Most African cosmologies posit the creator as in one way or another beyond good and evil, as, for instance, an otiose creator who has abandoned the world, or a force of violence beyond all moral accounting whose very arbitrariness demonstrates his local priority to, and hence ability to constitute, any system of human justice.

[9] More precisely, symptoms.

[10] Bohannan interprets these movements as regular features of Tiv social structure. More recently Nigerian sociologists (Tseayo, 1975) and historians (Makar, 1994 [1936]) have placed them in the colonial context, as a result of British efforts to force a highly egalitarian group into the framework of a state based on indirect rule. In fact, there is no real way to know whether such movements did occur earlier, but it seems reasonable to assume some such mechanism existed, at least, for as long as Tiv egalitarianism itself did.

[11] MacGaffey (1986) suggests the archetypal BaKongo ritual cycle leads from affliction to sacrifice to retreat to receiving gifts to new status.

[12] Personal communication, March 2000. Just as in Hobbes, by creating some overarching power of violence, which can ensure people fulfill their contractual obligations and respect one another’s property rights. Which if we look again at Pietz’s material, becomes especially ironic. Here we have European merchant adventurers swearing oaths and making agreements with Africans over objects they called ‘fetishes’, at exactly the same time as authors such as Hobbes were inventing social contract theory back home, but it was apparently the Africans who were seeing the act as creating a sort of social contract; the Europeans seem to have had other fish to fry. All this obviously raises the question of whether there is any reason to believe that Hobbes, among others, was aware of what was going on in Africa at the time; in Hobbes’ case at least, though, I have managed to find no concrete evidence. While Hobbes grew up in a merchant household, in his entire corpus his only mentions of Africa, as far as I am aware, are via Classical references.

[13] Clearly, what I am suggesting here could be considered a variant of the famous ‘wealth in people’ argument (see for instance Guyer, 1993, Guyer and Belinga, 1995).

[14] Obviously, this is a bit of a simplification.

[15] Bohannan and Bohannan (1968: 233): having a ‘strong heart’ means you have ‘both courage and attractiveness’.

[16] This is by no means unique to Madagascar. In the BaKongo case, too, royal power was seen as created through the same means as fetishes.

[17] In fact, the word ‘fetish’ derives from a Portuguese term meaning ‘something made’, or even ‘artificial’; this is why the term was also used for cosmetics – ‘make-up’ – (Baudrillard, 1972: 91).

[18] The phrase is adopted from Fredric Jameson. Jameson’s notion of ‘cultural revolution’ (1981: 95–7), in turn, goes back to a certain strain of Althusserian Marxism: the idea is that as one ruling class is in the gradual process of replacing another, the conflict between them can become a crisis of meaning, as radically different ‘conceptualities, habits and life forms, and value systems’ exist alongside one another. The Enlightenment, for example, could be seen as one dramatic moment in a long cultural revolution in which those of the old feudal aristocracy were ‘systematically dismantled’ and replaced with those of an emerging bourgeoisie. In the case of the West African coast one is of course speaking not of one class replacing another but a confrontation of different cultural worlds.

[19] This is why, as I have suggested (2001: 239–47), Marxists have such a difficulty figuring out what to think about magic.

[20] That is to say there was nothing like the fixed, mythological pantheon one finds among the Greeks, or Babylonians, or Yoruba, where objects of cult could be identified with some enduring figure like Zeus, Athena, Shango, or Marduk.

[21] Value being, as I have previously defined it, the way our creative actions take on meaning for us, by being placed in some larger, social, framework, by being embodied in some social ‘form’ like money or commodities (Graeber, 2001).

[22] As Terry Turner and others have argued at some length (see Graeber, 2001: 64–6), all this is pretty much exactly what Piaget was talking about when he described childish ‘egocentrism’: the inability to understand that one’s own perspective on a situation is not identical to reality itself, but just one of an endless variety of possible perspectives, which in childhood too leads to treating objects as if they had subjective qualities.

[23] Even this is somewhat deceptive language because it implies the production of people and social relations is not itself ‘material’. In fact I have argued elsewhere (forthcoming) that the very distinction between ‘material infrastructure’ and ‘ideological superstructure’ is itself a form of idealism.

[24] In point of fact, if one does, this can lead to fetishism of a different sort, as in the sort one sees in heirloom valuables in many gift systems which are seen as embodying or including the personalities of certain former owners.

[25] From a Marxian perspective it might be rather disturbing to see business deals as a prototype for revolutionary activity; but one must bear in mind it comes with the argument that the prototypical form of contract, even between business partners, is communism.