In early 2009, while walking to an interview on the Zege peninsula, Abebe said to me: ‘Life in Zege has got worse. People are afraid to eat together.’ The reason, he told me, was widely thought to be competition in the market, which was driving traders to seek the services of sorcerers (Amharic: t'enqway) to boost their business by occult means. The problem was that while the sorcerers’ medicine was effective, it carried a side effect: it turned the client into a buda. This meant that they would make others sick, usually inadvertently, usually by looking at them or their food. Everyone I spoke to agreed, both that there was a problem, and that the root cause was merchants competing to get ahead. Buda are an everyday hazard of the environment in Zege – probably they are the most consistent source of anxiety in daily life, because they are ever present, unpredictable, and potentially fatal. But there was a prevailing sense that this was an unusually intense moment of buda activity, approaching crisis. The reasons people offered for this were: first, the world food price crisis of 2008 that had doubled the cost of some staples and had led to an increased sense of pressure and competition on all sides; and, second, the increased volatility of foreign commodities in the local market, which were making it harder to predict whose business would succeed, and at whose expense.
This scenario, as it was presented to me, resonates with a substantial literature on witchcraft, which has seen African discourses of sorcery and spirit attack as a kind of localizing discourse for dealing with the uncertainties, exploitations and opportunities of incorporation into the capitalist world system (see, for example, Austen Reference Austen, Comaroff and Comaroff1993; Masquelier Reference Masquelier, Comaroff and Comaroff1993; Geschiere Reference Geschiere, Roitman and Geschiere.1997: 10; Comaroff and Comaroff Reference Comaroff and Comaroff1999). A young man in Zege, Zebirhan, gave me an explanation of buda that could have been a boilerplate from Africanist literature: ‘Buda is when a person gets rich for no obvious reason, without any special ability.’ We will see that the buda phenomenon does indeed reveal deep concerns about the morality of exchange. But there is also reason to be suspicious of attempts to reduce buda or witchcraft discourse to a sort of cultural morality play about global capital (Moore and Sanders Reference Moore and Sanders2001: 13; West Reference West2005).
For one thing, the local notion of traders becoming buda by means of sorcery turns out to be quite unusual, and possibly unique, in the literature on buda in Ethiopia. Spirits called buda are widely known across Highland Orthodox Ethiopia and beyond, but in the majority of cases these are thought to be hereditary, essential to the person rather than acquired, and associated with certain marginal groups: archetypally, weavers, potters, or the Beta Israel or ‘Falasha’ (Rodinson Reference Rodinson1967; Pankhurst Reference Pankhurst1992; Reminick Reference Reminick1974; Salamon Reference Salamon1999; Finneran Reference Finneran2003; Lyons Reference Lyons2014). This kind of classification recalls a different literature on the power of African smiths and artisans (see, for example, McNaughton Reference McNaughton1993; Finneran Reference Finneran2003; Lyons Reference Lyons2014; Rasmussen Reference Rasmussen2013). Indeed, the hereditary form of buda is widely present in Zege, too: certain families are said to be buda and are unable to marry with non-buda.
These broad resonances with literature on malign spirit power from other parts of Africa create a strange problem for the analysis of buda: there are almost too many similarities. The concept suddenly seems overdetermined, so loaded with possible comparisons that it becomes difficult to locate with precision what is actually happening. As we will see, buda in Zege refers sometimes to traders seeking to maximize their income by magical means, and sometimes to a state of pollution that causes the descendants of potters, weavers or slaves to harm their neighbours, deliberately or inadvertently. The common factor is a spirit that, via association with certain people, causes invisible but grave harm to those nearby. But beyond that, the etiologies of harm are varied, complex and ambiguous.
Behind the variation in buda, it is possible to deduce a common logic of violating the hospitality norms of the dominant landed culture. Buda sometimes looks like a local cipher for capitalism, but this is because of a deeper historical opposition between the hierarchical hospitality mores of the landed nobility and the creative and dangerous potential of craft and exchange. In Zege, the particular forms of buda are shaped by distinctive local histories of slavery, the coffee forest, and the Orthodox Church.
While the logic of buda is consistent, its means of recognition contains a constitutive ambiguity: it is unclear whether buda is caused by actions (for example, of greedy traders) or essences (for example, slave descent). In either case, it is also unclear whether buda attacks are initiated on purpose, or inadvertently. There is therefore a slippage in local moral discourse: do buda cause harm because people deliberately do bad things, or because certain kinds of people are just inherently impure and dangerous? Is buda a kind of person or a spirit that inhabits people? Do my neighbours harm me because they are doing bad things, or – and here is the key question in Zege – are they basically different kinds of people from me (Lyons and Freeman Reference Lyons and Freeman2009; Lyons Reference Lyons2014)?
There follows from this a blurring of responsibility – buda are animal spirits, but their actions are traceable to humans, and some humans are irrevocably associated with the buda. Thus, the question of who is really human is mixed in practice with the question of what counts as immoral action, and social marginalization proceeds along both lines. This ambiguity is characteristic of the Amhara Christian class system in general, in which Christian universalism sits uneasily with steep hierarchy (Levine Reference Levine1965).
The question of what buda represents is not separable from the question of how buda is known – how the concept becomes associated with events and experiences, and how it is verified, authenticated, or cast into doubt. Buda is not reducible simply to a category of interpretation: because it describes reflexive subjects, the buda concept exerts a formative agency of its own upon the people it describes (Hacking Reference Hacking2000: 33–4; Mazzarella Reference Mazzarella2004; Reference Mazzarella and Dube2009: 303). The emergence of buda in actual episodes of fear, anxiety and sickness, in houses and at dinner tables, actively shapes the way in which buda becomes intelligible. Profound uncertainties exist concerning the nature and behaviour of buda, as well as the individuals responsible for any particular attack, and the nature of that responsibility. It is only when people must decide how to act, or how to respond to a situation, that these uncertainties become fixed and consequential.
Each stage of the argument is necessary if we are to appreciate how active a concern buda is for people in Zege: an ever-present, unpredictable environmental threat; a source of constant intrigue and speculation; a field of legal and metaphysical debate; and a mechanism of severe social exclusion.
Let me start with an example that brings out some key themes. One night, the baby daughter of the couple I was staying with, Haregwa and Thomas, would not stop crying. This was unusual for her, and they immediately suspected buda. Dreadfully worried, they made the difficult uphill trek for an hour through the forest to the home of the traditional doctor and buda expert, Beza. Beza provided a snuff made from local plants, and the baby was soon better. To my mild surprise, given what I had read in the literature, no accusations followed, and no obvious attempt to find the party responsible for the attack. While buda is an animal spirit, their attacks are seldom thought to be autonomous; rather, they are usually associated with intentional or unintentional human involvement. During my time living with Haregwa and Thomas, Haregwa would occasionally have us breathe incense out of a gourd, as a prophylaxis against further buda intrusion.
I recalled the event with Haregwa four years later. She remembered it vividly. This time, though, she revealed that they had indeed had strong suspicions about who caused the attack – she named a fairly successful local hotelier whom we both knew quite well. I asked why she had not accused him or tried to have him divested of the buda spirit. She laughed and said, ‘Yasasärn näbbär’ (‘He would have us arrested’). As she explained, the government did not permit buda accusations on the grounds that this was superstition: ‘Mängest säw säw aybällam yillal’ (‘The government says people don't eat people’). So nothing had come of the episode. I would later hear whispers that a young man I had known had publicly called this hotelier a buda, and the subsequent conflict had been serious enough that the accuser had to leave town; he had joined the army and gone to the south, although I suspect he would have had other reasons for leaving. The complicating factor was that this accuser was a slave descendant and susceptible to buda accusations in his own right.
Following up my interview with Haregwa, I asked a friend (from the Zege forest, of high birth) about the legality of accusation. He confirmed that buda accusation was forbidden because it was ‘not scientific’. He told me that buda had multiplied in recent years, partly because of coffee traders travelling to more dangerous parts of the countryside, but that it was increasingly difficult to ‘remove them from the society’. Somewhat alarmed, having heard enough tales of the violent ostracism of buda suspects, I asked what he meant, and he said that it was difficult to get suspected buda out of community associations, church groups and other positions of local influence. It remained characteristically ambiguous whether he meant that the bad spirits had to be removed, or the people possessed by them.
There are a few points to note about this episode. One is that buda becomes relevant in, and is knowable through, conditions of fear and sickness. My friends saw their child become ill, and their attempts to figure out the cause were carried out in a state of high anxiety. There is a strongly affective element to the experiences that are marked and verified as buda attacks. Other examples include seeing people collapse on the street or have fits of shouting and growling. All of this stands as fairly compelling evidence that one should be concerned about buda. Much of this will be experienced in early childhood and is likely to influence subsequent experiences and attitudes, and contribute to the experience of buda as an immanent danger. Only later comes the extended work of verification and trying to establish responsibility.
Another point to be aware of is that I had always thought this couple to be quite good friends with the trader they were blaming. I knew him well and had always found him friendly, as most people did. He ran a small hotel near the marketplace and did good business serving beer and food. But I also heard occasional rumours around town that he was indeed a buda. Some said that he was descended from buda and had married a weaver's daughter; others that he was using buda tricks to help his business. It was quite possible, in the end, for him to be both a friend and a dangerous accomplice of spirits. It is also significant that the boy who ended up openly accusing him was himself of slave descent – a suspect turned accuser, playing on the duality of buda as an essential type, and as indicating selfish and antisocial accumulation.
Finally, there is general local agreement that buda exist, but considerable debate around attributing responsibility. For the government, on the other hand, which still works from a broadly Marxist materialist definition of reality, this was pure superstition and a prime example of harmful traditional practice (cf. Marsland Reference Marsland, Kelly, Harper and Khanna2015). I suspect that this is part of the reason why people in Zege describe an ongoing crisis. Common agreement that buda are active finds no means of public outlet, and so the discourse circulates in the form of gossip and innuendo. Given that buda accusations have historically involved violent ostracism, it is hard to be too critical of the government's position. But it is important to recognize the existence of a genuine ontological disagreement. In practice, to figure out what buda denotes, we have to ask how it is identified – how people go about applying the concept to experience, and how new experiences are in turn mediated by the knowledge of the concept. This attention to the work of understanding starts to show how it is that buda can become such a capacious entity, capable of describing a range of different types of things, and yet understood as integral to the historical experience of a place such as Zege.
Finally, the various actors here are drawing on quite deep-rooted categories of historical experience, especially the histories of slavery and land ownership in Zege. It is to this historical-memorial dimension that I turn next.
The context of the crisis
The Zege peninsula (current population approximately 10,000) juts into Lake Tana from west to east, some 15 kilometres by boat from Bahir Dar, the regional capital of Amhara Province. The area has been part of canonical Ethiopian Orthodox Christian territory since the thirteenth or fourteenth century, and Zege, like the islands of Tana, has been known for its monasteries since roughly that time (Rahel Reference Rahel and Babiker2002; Bosc-Tiessé Reference Bosc-Tiessé2008; Tihut Reference Tihut2009; Boylston Reference Boylston2012; Binayew Reference Binayew2014). According to local church accounts, Zege has been home to a coffee forest for just as long, the result of a pact between the missionary saint Betre Maryam and God: so long as the people of Zege preserved the monasteries and the forest, no natural harm would come to the people or the land. Outside historians, and some in Zege, would place the arrival of coffee in the eighteenth or nineteenth century, but it does seem that the peninsula has long been supported by trade from the forest rather than plough farming (Cerulli Reference Cerulli1946; Pankhurst Reference Pankhurst1968; Abdussamad Reference Abdussamad1997).
The blessed status of the peninsula comes from the monasteries and the forest together; living from the products of the forest places the monks at a remove from the labour of farming and the curse of Adam, and draws on widely understood tropes of sacred forests set apart from agricultural land (Tsehai Reference Tsehai, Sheridan and Nyamweru2007). But, unlike purely mercantile work, the produce of the peninsula still proceeded from land ownership, and therefore fit squarely in the imperial hierarchy of top-down land grants (Hoben Reference Hoben1973; Crummey Reference Crummey2000). The monastic community quickly drew other Christian settlers in – according to one local scholar, the original attractor was the healing ability of the founding saint. Then, from the seventeenth century, larger secular landowners began to arrive, receiving land grants from Emperor Iyasu I (who reigned 1682–1706) (Binayew Reference Binayew2014: 9). The landholders appear to have quickly developed into a quasi-clerical class, holding high lay offices in the local church administration and controlling much of the produce of church land, and this alliance of Church and nobility controlled land in Zege with relatively little state interference until the twentieth century (Binayew Reference Binayew2014: xii, 13). As in the rest of the country, the Church lost its landholdings after the 1974 revolution, but it retained significant local influence, and the descendants of the former nobility, while stripped of much of their former wealth and power, retain a high level of local status.
The coffee trade became a serious business in the latter part of the nineteenth century, as British and Italian interests in the region created new markets, and the newly founded Addis Ababa rapidly became a trade hub connecting the hinterlands to these wider networks (Merid Reference Merid1988; Abdussamad Reference Abdussamad1997; Fernyhough Reference Fernyhough2010). Local scholars describe this as a time of great wealth for the peninsula, but also as a high point of slavery. Slaves were captured from Benishangul and Gumuz regions and sold by Muslim and Agew traders on the Afaf market, after which they would usually end up working the church coffee lands (Triulzi Reference Triulzi1981; Abdussamad Reference Abdussamad1997; Boylston Reference Boylston2012). The continued presence of slave descendants in Zege is something of a public secret and plays a major part in the formation of the current buda crises. Slave descendants cannot marry into noble families, placing them in the same category as weavers, tanners and other marginalized local artisan groups – the same people most likely to be associated with a hereditary form of buda. As we will see, this dynamic of partly acknowledged marginalization, along with the landed classes’ fears of spiritual revenge from the margins, forms a significant part of the historical background of buda in Zege.
Since roughly the same period (starting between 1880 and 1920 or so), the market town of Afaf (population approximately 3,000) has acted as a gateway between the peninsula and the mainland. Plots were originally controlled by local landholders and taxed by the state; they are now rented out by the local qäbällé authority. This was the base of my fieldwork in 2008–09 and the focal point of buda incidents. The market draws together people from Zege and plough farmers from a broad spread of open countryside on the mainland, as well as being a base for traders in goods brought in via Bahir Dar: bottled beer and soft drinks, plastic jerrycans, torches, textiles, plastic shoes, soap, cooking oil, condoms, mobile phone cards, and various other commodities. Around the marketplace are a number of hotel bars that sell food and drink, play music and show football games, as well as general stores, a few veterinary pharmacies, some smaller tea- or beer-houses, and some residences. Much of this trade has picked up since the town gained regular electricity in around 2005, allowing for a greatly extended nightlife in the bars and teahouses.
Afaf town is therefore the trade centre that connects the Zege peninsula to the outside. The peninsula itself is often cast as a religious haven, but, nonetheless, it has an active political and economic life – in fact, one of the upshots of having the forest in the first place was that the inhabitants would not be self-sufficient but would be connected by trade relations to the farmlands beyond. The monasteries were never meant to be isolated – they served the populace, but their work had to take place at a remove from the toil and sweat of plough agriculture.
I have heard gossip about at least three of Afaf's more successful businessmen being buda. Much of this gossip, as far as I can gather without being rude, derives from their non-noble backgrounds. People are becoming (comparatively) rich who would not have done under the old system, at least as people remember it. The key turning point here is the post-1974 land reforms carried out by the Derg government, which stripped the Church and the nobility of most of their lands and redistributed them at about 0.75 hectares per person (Rahel Reference Rahel and Babiker2002). The old system was upheld by feasting – to be a legitimate wealthy person you had to feed your neighbours – whereas now to be a legitimate landholder you just have to be registered. At the same time, nobles lost control of taxation rights on plots in the Afaf market, with the taxes now going to the state.
Everyone I spoke to at the time described Easter 2009 as a crisis point in buda activity, partly due to the rise in food prices. There had been at least one town meeting in which the police had been urged to arrest the sorcerers responsible. But, in fact, the government forbids public buda accusations, and so no actual cases were brought. This clash between the local perception of an intense buda problem with a rationalist government's efforts to stamp out accusations is highly significant and will be discussed below – it may contribute to a continuing sense of unresolved crisis. In the return visits I made, every year between 2009 and 2014, I found that ambient buda concerns had sometimes receded and sometimes re-emerged, as if the 2009 crisis was just a particularly intense moment in an ongoing process of waxing and waning uncertainty.
The concept of buda is known throughout highland Ethiopia, and similar dynamics exist well beyond the traditional Orthodox heartlands (Pankhurst Reference Pankhurst, Freeman and Pankhurst2003). Almost all of the work that exists on buda describes it as a hereditary phenomenon associated with more or less endogamous groups of blacksmiths, potters or other kinds of craftspeople (Finneran Reference Finneran2003). Special attention has been devoted to the Beta Israel, still known in Ethiopia as ‘Falasha’, who have suffered particularly visibly from buda accusations (see especially Quirin Reference Quirin1992; Kaplan Reference Kaplan1995; Salamon Reference Salamon1999; Seeman Reference Seeman2009). Some authors even describe buda as an ethnicity distinct from Amhara (Reminick Reference Reminick1974; Pankhurst Reference Pankhurst1992).
Buda is described as a spirit, or a person possessed by a spirit, that harms people by ‘eating’ them, and that may transform into an animal, particularly a hyena, at night (Reminick Reference Reminick1974; Tubiana Reference Tubiana, Lewis, Al-Safi and Hurreiz1991; Pankhurst Reference Pankhurst1992). Those accused of being buda may be forced into humiliating confessions and purifications and are sometimes in real danger of being killed (Rodinson Reference Rodinson1967; Salamon Reference Salamon1999). In all cases, non-buda will refuse to eat, marry or have sexual relations with buda. Of particular interest for this essay, some sources note that ‘buda’ are not just craftspeople but specifically those who sell their products in the market (Rodinson Reference Rodinson1967: 58; Lyons Reference Lyons2014). Related to this, those accused of being buda are almost always described as landless people, or at least those who cannot adequately support themselves from the land (see especially Reminick Reference Reminick1974; Pankhurst Reference Pankhurst, Freeman and Pankhurst2003; Lyons Reference Lyons2014). There is broad recognition of a pervasive class dimension to buda ideology, in which the powerful use buda accusations to restrain and stigmatize other groups (Reminick Reference Reminick1974; Salamon Reference Salamon1999; Finneran Reference Finneran2003; cf. Galt Reference Galt1982 for a similar argument about ‘evil eye’ in Italy).
Academic explanations of the phenomenon fall roughly into two camps, not necessarily mutually exclusive. On the one hand, there are those who emphasize the material and symbolic properties of crafts, especially metalworking, which appear to a settled farming population as occult, even outright evil, since they are opposed to the principles of the growth and fertility of crops (Finneran Reference Finneran2003; Lyons and Freeman Reference Lyons and Freeman2009; Lyons Reference Lyons2014). The other approach focuses on the historical experience of dispossession as generative of stigma, with craftworking being the only profession available to those without access to land (Quirin Reference Quirin1992; Kaplan Reference Kaplan1995; Freeman Reference Freeman, Freeman and Pankhurst2003). Quirin and Kaplan both demonstrate that the ‘Falasha’ did not become a stigmatized group, much less a quasi-caste, until they had been stripped of their rights to rist land in the fourteenth and fifteenth centuries – and, even then, it took hundreds of years for a fully stigmatized identity to coalesce. These efforts to show the historical emergence of dispossession and buda accusation are persuasive, but they do not explain why the accusations take such strikingly similar form in quite culturally distinct parts of Ethiopia – even in a work such as Freeman and Pankhurst's Peripheral People (Reference Freeman and Pankhurst2003), which considers a wide range of cases. Some part of the symbolic argument must therefore be necessary in order to reveal the shared underlying logic of buda and related phenomena.
Part of what makes buda interesting is that it is capable of incorporating a great diversity of ideas, histories and practices under a monolithic rubric that is recognized throughout the country. After all, ‘buda’ designates an invisible, inscrutable being. It is powerful precisely because its referent is not well fixed, and its motivations uncertain. Even more than this, and critical to understanding the flexibility of the phenomenon, buda is always something that violates fundamental, normative social distinctions: male and female (Lyons Reference Lyons2014), human and animal (Salamon Reference Salamon1999; Baynes-Rock Reference Baynes-Rock2015), or master and slave (Reminick Reference Reminick1974). This boundary violation may be buda’s definitive aspect – a sort of anti-definition that would go some way to explain the variance in the literature (cf. Hannig Reference Hannig2014).
As we are beginning to see, the connections between buda, market exchange and slavery appear to be much more apparent in the environs of Zege than elsewhere, where narrower definitions around craftspeople are more usual. I would suggest that this is due to the particularities of Zege's history – the monasteries and the coffee forest, which, as we have seen, have enabled a narrative of Zege as a sacred space while it simultaneously became a focal point of the slave trade. However, it is worth noting that the former province of Gojjam, in which Zege lies, is notorious across Ethiopia as a hotbed of buda. This is something that locals acknowledge, sometimes identifying the more rural areas as the real danger spots. Ethiopians from other areas do not find it surprising that spirits there would be more multifarious and complex, or that the Gojjamé would know more about them than others do.
The locally specific dynamic of buda acquired by visiting sorcerers helps to reveal a common structural and symbolic logic, by giving us a situation in which it makes intuitive sense to think of buda in a non-hereditary way because it still shares some core aspect that is recognizable as ‘buda-ness’. That commonality, I argue, is dependent on wealth or value creation that does not proceed from the land. This is not just a response to capitalism, but indicative of older tensions between landowning, hospitality and market exchange in Zege. Buda stands for anti-sociality, for the inhuman – but it is rooted within humans and in society, and the divisions that it denotes and blurs are historical in nature.
Buda in historical discourse
I had been in the field for a long time before I worked up the courage to ask direct questions about slavery. When I finally did ask Abebe if there had been slaves in Zege, and what had happened to their descendants, he began reeling off a list of names – many of whom were friends of us both – who were yäbariya zär, seeds of the slaves. Seamlessly, he also began listing people who were yäbuda zär, seeds of the buda. While buda and bariya were not the same thing to him, he made it clear that they belonged in the same category: they were people whom nobody would marry, and many would not eat with them either, especially in Abebe's father's generation. I was surprised because I had attended weddings with many of these people and had never noticed any divisions. Abebe said that, indeed, there were ways to be subtle about which table you ended up on, so that the rejection would not be explicit. I asked him if there had been slaves in his family and he responded, mock-indignantly, ‘Näs'uh nän’ (‘We are clean’). I conducted a number of further interviews on the topic that elicited the same picture.Footnote 1
He had learned all of this – which families are ‘pure’ and which not – from his father talking in the home. Establishing purity is not simple because of the Amhara cognatic descent system, which means that there are no clear lineages or descent groups of any kind. Instead, there are clusters of competing land claims and patriarchal households in which slaves, servants, women and children have historically been subordinate in just about the same way (Hoben Reference Hoben, Tuden and Plotnicov1970; Reminick Reference Reminick1974). People do keep careful patrilineal genealogies, though. Many men, especially local scholars and those from important families, can trace the names of their fathers back eight generations or more. These genealogies have traditionally served to prevent incest according to the seventh generation rule (Hoben Reference Hoben1973) but also to exclude impurities of slave or buda from the line. But the genealogies are not explicit; they are not displayed in corporate social groups, but are kept in the home and in people's memories. They play a key role in structuring relations of inclusion and exclusion, but, as with the food at the wedding, they do so largely below the surface of social interaction. This helps to explain why descendants of the ‘pure’ are often on openly friendly terms with slave or buda descendants, even when profound rifts exist between them.
During our conversation about buda, Haregwa explained to me how the stigma or suspicion of buda can follow people from place to place. When she married Thomas, people had been suspicious because he came from elsewhere, and therefore his zär (seed or race) could not be known. But, she told me, if she had married a local weaver's son, it would have been even worse, and her family would have asked her why she was spoiling her seed. She also said she felt that this zär thinking was on the way out (eyyäqärrä näw), whereas her father's generation had been very strict.
An outcome of this system of reckoning descent is that, rather than a set of lineages or segmentary kin-based groups, there are just two categories: ‘clean’ and ‘unclean’. It is part of their semi-implicit nature that there is not an institutionalized term for either category. These categories are not pertinent to every social context. In relations of trade, of daily conversation or of friendship, they can be negligible; in sharing food and in marriage they can be absolute. These different levels of inclusion and exclusion, in which sociability in one context combines with exclusion in another, are defining mechanisms of Amhara life.
What makes buda sometimes seem like an ethnic or caste designation is that, if you divide the world up into pure and impure people, you create two de facto endogamous moieties. Those who are likely to be called buda, whether because of slave descent or being ‘Falasha’, or because of their profession, are likely to marry others in a similar position. Anecdotal evidence suggests that this is the case in Zege, with marriages between slave descendants and tanners’ descendants, for example, being commonplace. Through these means, a general opposition between pure people and impure, opposed to each other as honour is to shame (kebur/hefrät), is preserved and transmitted. The two groups, however, are not symmetrical. One part is acknowledged only in gossip and innuendo, and has no grounds for solidarity. Moreover, the division is not acknowledged in Orthodox Christianity, which is shared by most.
Almost everyone who gets called buda would contest the term. The label is shameful and humiliating and, crucially, unlike a group designation, it is isolating. There is no possibility of buda getting together and joining an ethnic or class movement or solidarity group, because no person could possibly want to be associated with the concept. It would be deeply offensive to interview anyone about being buda. This is why it is always incorrect to describe buda as an ethnicity. Buda is an anti-ethnicity; it dissolves solidarity. It is an appellation of mistrust and exclusion; it casts people as antisocial agents. Precisely because it blurs the line between individual actions of sorcery and polluting essence, buda can never become a unifying identity. By contrast, in many parts of Ethiopia, people who were called ‘Falasha’ have been able to organize themselves as a corporate group and to redefine themselves as Beta Israel, their own preferred designation, although many have converted to Orthodoxy in an attempt to escape stigmatization and have found themselves caught between the two groups (Seeman Reference Seeman2009).
In the Lake Tana region and around Zege, the most prominent marginal group is the Weyto, possibly the original inhabitants of the area, who are mostly canoe makers, and who endure comprehensive discrimination (Gamst Reference Gamst and Hess1979). The Weyto are now almost all Muslim and most claim to have always been so (based on my own interviews and also Cheesman Reference Cheesman1936), but they find comparable difficulties integrating into Muslim communities as in Christian, because of the prevalence of the stigma against them. I have spoken to some Weyto people in Bahir Dar who suggested that, if and when their material lives improved slightly, some out-marriages became possible. The few Weyto who remain in Afaf, by contrast, have no such prospects.
The idea that slave descent renders you impure and hence unfit for marriage is a common one, for example, in Madagascar, as described by Denis Regnier (Reference Regnier2012). It is easy enough to speculate on why this might be – in order to morally justify owning a person, you need to tell yourself a story in which they are different from you in some basic way. Regnier's argument revolves around the way in which people essentialize historically contingent differences. Mine, on the other hand, is that buda discourse in Zege blurs the lines between what is due to your essence and what to your actions.
Most of the gossip I have heard about people in the village who became buda after visiting sorcerers has revolved around some of the wealthier merchants – not coffee merchants, who have land, but shopkeepers, bar owners and other businessmen.Footnote 2 These are people who have become wealthy, but not from the land. This landlessness is the obvious thing that suspect merchants share with slaves, but also with the numerous other caste-like groups that experience forms of social exclusion in Ethiopia. And there is a general assumption that landlessness is accompanied by a desire for what others possess.
There is a curious, revealing reversal of the story of buda being slaves, which Reminick (Reference Reminick1974: 283) reports from the 1960s and which I have heard in the exact same form in Zege, in which a buda could enslave the soul of the recently deceased and turn it into a domestic slave. When any visitor came, the enslaved person would be bewitched to look like a household ornament. Reminick's interpretation of the story is persuasive: a landed man's status would be designated by the number of dependants in his household, including, at that time, the descendants of his or his family's slaves. It stands to reason that a buda should desire the same status, but having no land to support such dependants, he would of course achieve this by occult means. Buda, in this iteration, are a projection of an agrarian landed people's nightmare of the agency of the landless. Slave owners tend to become dependent on their slaves and to resent this fact. In the register of slavery and marginalization, buda articulates the anxieties of the dominant about the desires and the potency of those they dominate; it seems to speak to concerns inherent in the property system itself – or at least the property system of nineteenth- and twentieth-century Ethiopia, with its extractive aristocracy and long-distance trade markets.
Desire is central to this picture. Slaves are dangerous because it is assumed that they desire the possessions of their masters; the marketplace is dangerous as a whole because it is a field of competitive desire. At the same time, it was strong, long-distance markets for coffee that created the demand for slaves in Zege in the first place. Slavery brought in foreigners in large numbers and made them subordinate, in conditions of maximal desire and acquisitiveness in Zege society at large.
We have seen that the buda-through-sorcery model is much less common across Ethiopia than the hereditary model and may be unique to Zege. But when viewed in terms of landlessness, from the perspective of the landed, it is possible to deduce a common logic: those who do not possess land, and are therefore reliant on craft or trade, are likely to be suspected of buda. In Zege, this dynamic is particularly acute because of the reliance of the landed on coffee and their ensuing dependence on commodity trade. As the next section will make clear, trade and exchange are understood as the diametric opposites of the nobility's ideal form of sociality, which is hospitality.
Hospitality and exchange
Buda is a category of difference, signalling a hostile dimension in intimate relations (cf. Geschiere Reference Geschiere, Roitman and Geschiere.1997: 10). In Zege, slavery created strangers within: in the very household of the master. The systematic exclusion by landed Christians of all kinds of landless craftspeople, meanwhile, produced another kind of local strangeness – people who were economically and sometimes ritually integrated, but excluded from kinship and commensality. This alienation, however, coexists with spheres of inclusion: most importantly, in many cases these people are Christians. They therefore do the things that make an Ethiopian Orthodox Christian a moral person: in theory, they eat together, and they fast together. This contrasts with relations between Christians and Muslims, which are often friendly and cooperative but in which neither side ever eats the other's meat.
Here is a vital point: anxieties around buda crystallize around moments of sharing substance, or in which different people are joined together – either through marriage and descent, or through the sharing of food. This indicates one way in which the buda concept is more than a discursive element: it is experienced through material and embodied relations, and it is through these embodiments that buda experiences are transmitted down generations (Argenti Reference Argenti2007).
In the era of slavery, a Christian master had to baptize his slaves. For one thing, it is commanded in one of the key Christian law books, the Fetha Nägäst, which also forbids Christians from selling slaves to others (Pankhurst Reference Pankhurst1968: 88; Strauss and Paulos Reference Strauss and Tzadua2009). Christians would buy from Muslim traders and then the slaves would be set free on their owners’ death, or pass to their owners’ descendants and become dependent retainers (Reminick Reference Reminick1974; Abdussamad Reference Abdussamad1997). For another, slaves prepared food for the household, and so they had to be Christian. This immediately causes problems, because you need to bring the slave close enough into circles of commensality that they cook for you, but keep them distant enough that you can claim complete superiority. According to my research assistant Tefera, one solution was to tell slaves that they were accursed, and that any fine meat would give them leprosy, so as to stop them stealing food while they cooked it. As Tefera tells it, slaves would have to eat rats, thus marking them as substantially inferior and morally compromised.
Likewise, those who remain of the ‘Falasha’ people in Zege have not kept their distinct religion but converted to Christianity a generation ago. Nonetheless, they, too, retain the taint of buda, which means that their hospitality is always potentially poisonous, and their children are unmarriageable. Also, the merchants who are suspected of being buda are almost always Christians. This is why they are so troublesome – you are supposed to eat with them, but the spirit within them, because of its greed, threatens this commensality.
Abebe's comment that alerted me to the importance of buda – ‘people are afraid to eat together’ – summarized what for many people is the major trend in Zege in recent years. While conducting interviews in the forest with Tefera, one of our most frequent topics was the decline in hospitality because of economic hardship. The elders we spoke with framed Zege's history as one of collective feasting, in which all proper relationships were defined by acts of eating together, either in the great funerary täzkar feasts or in the clerical dining rooms (mäfraq) that organized church and lay hierarchies. The church dining room epitomizes proper hospitality, with its dignitaries sitting in prescribed order under the blessed authority of the Church. Now, as Tefera told me, people would rarely offer food to non-related visitors or ask it of their hosts. There are two overlapping reasons for this: one is consideration – not wanting to compel people to part with their very limited supplies. The other is fear of buda: because of an atmosphere of competition, hardship and envy, there is always a danger that people who serve you food may inadvertently harm you via the evil spirit. These fears are particularly pronounced when people go to areas slightly poorer or less central than their own. Fortunately, there are ways to offer hospitality without serving food: home-brewed beer and coffee are relatively safe and can always be served to guests.
Eating together is the main way in which Amhara act out who belongs and who is excluded, but buda are not only threatening because they disrupt hospitality; they actually eat you. This is the point where we are reminded that buda is both a kind of person and a kind of animal spirit. I remember a friend asking me in the course of discussion what the English word ‘cannibal’ meant. I said it was a person who ate people and he replied, ‘Oh, you mean buda.’ If buda creates ambiguity between action and essence, we might regard eating as the moment when these two dimensions come together. Eating is an action that makes you what you are – an action that constitutes essences (Strathern Reference Strathern2012).
The connection between eating and dominance has been a recurrent trope in postcolonial African studies (Mbembe Reference Mbembe1992; Argenti Reference Argenti2007; Bayart Reference Bayart2009). In Ethiopia, for example, the word mäblat – to eat – is used when one wins a game: the winner eats and the loser is eaten. Buda eating their neighbours, then, looks like a classic reversal. To some extent we can read it, again, as a projection of the dominant onto those they dominate: we are more successful than them materially, so they must take their revenge spiritually. This only goes so far, however, since the market traders who are called buda are hardly disadvantaged. More significant is the concern of a hierarchical relationship (beneficent feeding) turning into an exchange. You may eat the buda’s food, but rather than this acting to arrange an ongoing relationship of delayed obligation, as hospitality normally does, the buda exacts direct repayment in the form of your own vitality.
Compare this with Geschiere's (Reference Geschiere, Roitman and Geschiere.1997) analysis of the duality in witchcraft between ‘levelling’ and ‘accumulative’ mechanisms – in which witchcraft sometimes applies to the disadvantaged trying to gain a spiritual quid pro quo, and sometimes to the powerful using illegitimate means to get ahead. The buda example in Zege suggests, rather, a single common logic: in each of these cases, a relation of proper hospitality, in which hosts feed guests and both are honoured, is subverted and turned into a relation of cumulative exchange, in which each party is measuring and trying to increase their lot. The key point is that the moral baseline, from which buda and witchcraft deviate, is not one of generalized equality, but one of hospitable feeding, with its implications of morally sanctioned asymmetry.
At the same time, the trope of eating calls into question the dividing lines between humans and animals. Countless people have told me in interviews that buda are animal spirits, as distinct from zar, which are human-like. And yet, as this article demonstrates, people continually talk about buda as if it were an essential quality of certain kinds of people. In many parts of Ethiopia, the archetype of buda is that it transforms from a person into a hyena at night (Salamon Reference Salamon1999). Since the hyena eats carrion, the buda therefore steps out of the realm of proper personhood, and people afflicted by buda attacks are said to sometimes eat ash or faeces. Its predatory and gustatory habits are unpredictable and it lacks the repression or suppression of the appetites that make people moral. In addition, buda cannot speak – here again there is an explicit contrast with zar spirits, which can be coaxed to talk through a medium. A person afflicted by a buda attack may growl, grunt, scream or lose consciousness – the animality of the spirit is not just a point of discussion, but is displayed with great vigour and visceral force. Buda are both animal spirits that roam the village and the forest, and people who live among others. The conflation of one with the other is central to the logic of buda, and plays a key role in deciding what counts as a proper human. But this key question has a strongly affective dimension – it is evidenced through quite striking, distressing moments of self-loss on the part of the victim. These affective aspects – fits, growling and collapsing on the part of the victim, fear and anxiety for the witnesses – are an inseparable part of how buda becomes meaningful (attached to experiences and events), but also how the truth and existence of the phenomenon is established at all.
Truth, fear and responsibility
Investigating the social relations to which buda applies has brought us from historical discourse – thinking about slaves, land and craftspeople – to direct bodily experience. It is becoming clearer how historical relations of experience are lived out through bodily relations of eating together and of marriage. But a question remains as to how people are able to identify certain events as indicative of buda. Buda spirits cannot be seen, the local traditional doctor informs me, even with a microscope – this is the key way in which they differ from bacteria, which are visible but tiny. How, then, are they discernible at all?
The main answer is that buda are known by their effects. If you grow up in Zege, by a young age you will have heard your parents and other people discussing buda and speculating about whether a particular illness was buda-related. If your own mother was ever worried about buda attacking you, she would have held a gourd of incense and charcoal to your face, so the smoke would drive the spirit away. You might well have seen people collapse on the street, catatonic or in convulsions, as sometimes happens with buda attacks. As we have seen, if you did get sick you might have been rushed through the forest by concerned relatives to see the local doctor. It is only later that a child learns that her neighbours may be causing buda attacks in some way. She may have always worn a protective amulet containing a Ge'ez prayer scroll around her neck, which her mother placed on her because the family next door was known to be yäbuda zär, the race of buda. She will start to hear speculation and gossip, and to learn what is generally known about how buda behave and which people are responsible for them. Most people have become sick with a buda attack long before they really understand what that might mean. Experience of buda attacks is immediate and powerful, but the chains of responsibility that attach buda attacks to human perpetrators can be deduced only through retrospective inquiry.
There is a long tradition in anthropological thought that says we should not make assumptions, before the fact and on behalf of our informants, about what exists and what does not – one of the key touchpoints is Evans-Pritchard's famous witnessing of the mysterious light (Reference Evans-Pritchard1937: 34); another is Edith Turner's (Reference Turner1987) extended engagement with the numinous in the field. The latest round in this discussion has come out of the ‘ontological turn’ and debates about what it actually means to ‘take seriously’ the realities of others (Henare et al. Reference Henare, Holbraad and Wastell2007; Course Reference Course2010; Pederson Reference Pederson2012; Viveiros de Castro Reference Viveiros de Castro2015; Graeber Reference Graeber2015).
For my own part, I would feel disingenuous trying to claim to be something other than a sceptical and ritually tone-deaf methodological philistine. When it comes down to it, I do not think that buda spirits can harm me if I don't believe in them – although nor can I deny the power of the phenomenon or the capacity of buda discourse to act beyond the control of any of the social actors implicated within it. As I have tried to indicate, buda events are largely unpredictable, fraught with uncertainty, and by no means simply or consciously tools of the powerful.
The standard solution has been to say that, whatever the ultimate truth, spirits are real to those who live with them and therefore they have ‘social reality’. This is even more unsatisfactory, as it dodges the question of why spirits are real for ‘them’ and not for ‘us’. So, rather than try to arbitrate questions of existence, it seems better to ask how buda become true, persuasive or plausible (as any ‘social fact’ must go through some process of verification) – as people work to make experience and memory comprehensible in the light of previous experience and knowledge.
This involves paying due attention to the doubts and uncertainties that surround these modes of veridiction. Perhaps most important of all, it involves asking not just how people deduce what buda are and what they do, but how they decide whether these judgements matter or not – whether they are grounds for making accusations public or taking action against culprits. It is one thing to come up with ideas and explanations for what the world is like, as we all do. It is quite another when these notions become the basis for wanting to take legal action against people, commit violence against them or exclude them. There are moments when we are called to act on our convictions, and it is at these times that speculations become more concrete.Footnote 3
Three uncertainties emerge in buda discourse: how buda behave, who is responsible for them, and whether they exist at all. This last point is not really a question for most people in Zege, who have seen ample evidence in the form of attacks and sicknesses. But it does matter, because the government position is that buda is superstition, and public accusations are therefore not permitted. As we have seen, this question of legality is highly consequential – it probably prevents violent ostracisms, but it also produces a prevailing sense of paranoia, of a threat that is not being acknowledged. In fact, when a young man snaps and openly accuses our hotelier friend of being buda, it results in his leaving the community. In these flashpoints, what had been open questions of reality and responsibility become fixed and determinate. When you take action on a point of ontological uncertainty, certain things become irreversible.
Some questions of buda behaviour do touch on the possibility of doubting their existence. A point of discussion for many of my friends is why white people seem so unconcerned about buda. Some friends have told me not to worry, because buda clearly do not eat white folk; others have insisted I procure buda snuff to take home with me, in case I became sick on my return, and English doctors would not recognize the spirit. In that case, they might give me an injection, which would probably kill me. Some speculated that buda must be a local phenomenon and that Europe was just not their habitat. Again, what might fix some of these open questions would be a moment that called for action – perhaps if I were to get sick or come under suspicion myself. The general point is that it is important to know something about how buda behave, because only then can you protect yourself from their consequences by finding the correct prevention or treatment.
Finally, there is the question of who is responsible: who is known to be buda, or to associate with them, or who is responsible for a specific attack. This can be known by applying smoke or incense to the victim, or through genealogical work, or by consulting experts. In any case, it is vitally important to remember that buda, for the people of Zege as much as for an anthropologist, is a sociological phenomenon. We are in total agreement that the causes of buda that matter are to be found in intentional and unintentional human action.
In the past, in other parts of Ethiopia, dramatic practices of revelation existed in which buda spirits were compelled to speak the name of their masters through ordeals of smoke and prayer (Pankhurst Reference Pankhurst1992; Salamon Reference Salamon1999). Such revelatory practices are now strongly contraindicated. Responsibility is now ascertained largely through gossip and comes to light unexpectedly, in moments of conflict such as that in which my slave descendant friend made a buda accusation in a moment of anger, and ended up leaving the village. This was the result of years of talk and insinuation about both parties, but only the eruption of the conflict allowed something to become fixed and public.
Zegeña have recourse to Beza, the traditional doctor and unquestioned buda expert, for diagnosis and verification. But I do not actually know whether Beza is willing to diagnose the human party responsible for buda attacks. If he were, it is possible that I could get him into trouble by talking about it. Certainly, this is not part of his public work nor have I ever heard it insinuated that he would give the basis for accusations. Beza specializes in cures, and as such he is one of the most important people on the peninsula.
This harks back to Evans-Pritchard's (Reference Evans-Pritchard1937) original observations about witchcraft as, at heart, a form of social knowledge that sought to ascertain by what agency and motivation things happened. This social-epistemological aspect of witchcraft is what gets forgotten in studies that treat sorcery as simply a form of micro-politics (Douglas Reference Douglas and Douglas1970, cited in Geschiere Reference Geschiere, Roitman and Geschiere.1997: 218).
In this sense, witchcraft and spirit malady practices can be seen as ways of mapping obscured or repressed aspects of interpersonal and environmental relationships, and making them intelligible. In Zege, however, such mapping (which might in itself have been pathological) has been interrupted. This has not put an end to scapegoating practices, but has certainly destabilized them and contributed to the ambiguity of buda as a concept of personal morality, essentialist discrimination, and the social-historical causation of suffering.
What is being deduced in the aftermath of a buda attack is a relationship between culprit, spirit and victim. Such relationships are widely agreed to be conditioned by historic property and class relations, especially those surrounding slavery, trade and craft. We need to understand the interpretive discourse that surrounds buda as sociological work: people are trying to figure out what the world is like, but especially what kinds of relationships, interests and causal agencies are in play. This involves ontological speculation about what buda might be, but primarily it involves tracing responsibility. It is therefore an activity highly susceptible to legislation as to what can and cannot be made public.
Conclusion
To briefly recap the historical logic of buda as an idiom of human difference, a comparative account shows that the relationship of buda with greedy merchants, as seen in Zege, is part of a much wider politics of value. The landed nobility, which regards hospitality as the hallmark of moral sociality, is deeply mistrustful of other forms of value creation such as craft and trade, especially insofar as they involve exchange transactions. Commodities are here understood as opposed not to gifts, but to hospitality and hierarchy.
Market traders share with slaves, potters and tanners the fact that they do not make a living from the land. Because of this, they are thought of as avaricious – merchants because of the nature of their occupation, slaves because of their envy of their masters. This perceived lack of control over the appetites in turn opens up the whole symbolic repertoire of buda as indicating animality, the failure to fast, the lack of autonomy and honour, and the propensity to make spirit attacks due to either malice or a deeply shameful lack of control.
This symbolic logic helps to make sense of buda phenomena but is not sufficient. We have seen the ambiguity and uncertainty in which specific cases of buda emerge, and the legal and political difficulties in identifying relations between spirit attacks and individuals or groups of people. People in Zege are doing constant work to try to figure out whether and how buda are relevant to new and unexpected situations. They are tracing relations of association and responsibility, and constantly speculating on the deeply ambiguous relationship between personal intention, the moral qualities of others, their historic relations of desire and jealousy, and the harmful actions of spirit agents. Paying close attention to this epistemological and sociological work seems to be the best way to appreciate the full scope of the buda.
Acknowledgements
Research for this article was supported by a British Academy Postdoctoral Fellowship and an ESRC 1+3 Studentship. I am grateful to seminar respondents in Oxford (African studies) and Edinburgh (anthropology) for feedback on earlier versions.
