Anyone who thinks that epidemics of sorcery or spirit possession in conditions of rapid social upheaval are already well-understood phenomena should read this impressive and eye-opening book. On one level it’s a familiar story – the massive dislocations of the transition from socialism accompanied by widespread ‘occult’ violence – but Pedersen shows in richly observed and brilliantly argued ethnographic detail that the specific forms in which shamanic spirits have made themselves manifest in northern Mongolia in the last few decades have been quite distinctive, and that those forms are not a mere reflection of supposedly more material realities, but are themselves distinctively shamanic. The impression that notions such as ‘occult economy’ explain these phenomena is superficial and misleading. The real story of the revival of shamanism in northern Mongolia is much more interesting and thought provoking.

When the military and financial support of the USSR was suddenly removed in the early 1990s, the institutions of the Mongolian socialist state and economy promptly went into terminal collapse. This was wholly unexpected locally, because viewed from inside, the fragility and dependence of the socialist system had been quite invisible: it seemed all-powerful, all pervasive, and as if it had crushed its enemies beyond all hope or even memory. The suppression of religious practice in the Mongolian People’s Republic had been among the most brutal and uncompromising of any socialist regime; all the more dramatic because the order socialism replaced had been so pervasively structured by Buddhist and shamanic institutions.

This seemed nowhere more true than in the remote and isolated northern valley of Shishged. Its plain had been dominated by a massive Buddhist monastic estate, controlled ultimately by the highest-ranking reincarnate lama in Mongolia and theocratic emperor of the short-lived independent state, but its extensive forested hills (taiga) had been at best lightly governed by Buddhist institutions, and long regarded as some of the ‘most shamanic’ places in Mongolia. The imposition of Buddhist civilization and the domestication of the landscape had never been completed, and the forests were largely left to shamans and hunters. The Darhad people of the region had a reputation for being ‘wilder, poorer, cruder, funnier, and (above all) more shamanic’ than anyone elsewhere in Mongolia. But by the height of the socialist order shamanism, along with Buddhism, appeared to have been wiped out here even more completely than in the rest of Mongolia. Not only were there no shamans and no practice of shamanic possession; gone also was all everyday experience of the presence of shamanic spirits.

But the economic trials and hardships people in this region had to endure in the 1990s were exacerbated by an epidemic of young men sent into terrifying homicidal rages by shamanic spirits. The spirits, long forgotten, had returned with a vengeance, but there were no institutions and specialists to control them. The possessed young men could not become shamans proper. Indeed there has been no ‘revival’ of shamanism in the most obvious sense: still virtually no one whom anyone recognizes as a ‘genuine’ shaman. Unable to learn how to subdue the spirits and so choose when to become possessed and when not, these young men remained permanently stuck as what Pedersen calls ‘not quite shamans’.

Pedersen argues nevertheless that shamanism has flourished in northern Mongolia since the demise of socialism, not in spite of but because there are virtually no ‘genuine’ shamans. To such an extent has it flourished, in fact, that the state that has emerged from the rubble is in certain respects itself ‘shamanic’ in form. The upheavals of the last decades have been such that now socialism tends to be remembered as stasis, as ‘a deep sense of sameness, predictability, and inevitability’. Pedersen does not of course mistake this memory for a description of what things were actually like. His point indeed is that shamanism was one of the many things that became unimaginable within the ‘hegemony of representation’ (Alexei Yurchak’s expression) that characterized late socialism. But the apparently monolithic socialist state, ever-present and unchanging, yet driving society inexorably forward towards ‘modernity’, was replaced in very short order by a state with an ephemeral spirit-like presence. The way persons might constitute themselves as significant political agents – might come to be seen like a state – has changed. Political agency has come to look like what shamanic power had always looked like: ‘visible in fleeting, conflictual encounters arising from confusing, unpredictable, and unintended events’ (p. 60). Pedersen illustrates the new shamanic form of state power with a series of brilliant descriptions: a shaman’s display of election posters, the conduct of strongmen political bosses, and the robust electoral support for anti-communist political candidates in the region. The last of these phenomena points, Pedersen suggests, to an affinity between shamanism and both democracy and the emerging capitalist market (Buddhism, by contrast, locally has an affinity with socialism (p. 76) – although this observation isn’t developed). Shamanism’s multifarious form mirrors these political phenomena, encompassing contradictions in a way that modernist socio-political projects with a single logic never could do.

Pedersen draws a comparison with hunters, who need to be able to switch at will between human subjectivity and that of their intended prey, but unlike accomplished hunters, not-quite shamans are not in control of which subjectivity to adopt at any given time, an incapacity that would be fatal for a hunter. Because they might be possessed at any moment, they are permanently in some respects half spirit and half human, and although they may live in villages among neighbours and kin, they become socially marginalized. A whole generation of men is stuck in ‘transition’, not quite having become shamans, a shamanic equivalent of Yurchak’s ‘last Soviet generation’. Pedersen describes how the male antecedents of one not-quite shaman have each in their turn occupied equivalently marginal social positions, replicating in each generation an analogous affliction, but the precise form of their marginalization was substantively as different as the socio-political regimes to which they were successively marginal.

Most literature on religious revival in post-socialist contexts describes more or less voluntary affiliations or conversions, but in northern Mongolia shamanism is a predicament people find themselves in: an involuntary calling. Not-quite shamans are ambivalent, trickster figures, who constantly break proprieties and conventions and speak about their spirits in riddles and paradox because their own experience of spirits takes such forms. They tell wild stories that deliberately confound any attempt on the part of a listener to sort out truth from falsehood, because, Pedersen suggests, this is the only way accurately to talk about shamanic experience. It’s for this reason that one of the distinctive art forms of the Shishged area was a kind of satirical and mocking joking song. These were appropriated in Soviet times to state-socialist ‘nationalities’ policy and folklorized versions of them, reified as ‘Darhad culture’, were performed in state celebrations. But in the context of comprehensive state suppression, these were the only contexts Pedersen has found where any vestige of shamanism could be made manifest – in a revealingly parodic way. So there are tales now of one lone surviving secret shaman, employed to act the part of a shaman in ‘house of culture’ performances. But actually he was not acting. He really did become possessed while on stage. ‘Indeed the only way he could respond to his shamanic calling was by pretending to pretend that he was a shaman’ (p. 209) – which was I suppose another shamanic way of being not quite a shaman.

Pedersen notes that the situation in northern Mongolia now has clear parallels not only with other post-socialist contexts but also with the widespread revival of ‘witchcraft’ in Africa, and he concedes how neatly the most influential accounts of the latter phenomenon might seem to apply. But very carefully and courteously, Pedersen argues against reading these shamanic phenomena as the expression on a ‘symbolic’ level of the ‘real’ structural disorders of that academic occult phenomenon, ‘neo-liberalism’, which haunts so many recent anthropological accounts and is invoked to explain just about everything about contemporary realities that observers regard with distaste. He notes the capacious ambiguity of these accounts – sometimes ‘occult’ phenomena are a symbolic comment on real-life changes, sometimes they are interpreted psychologically as an attempt to cope and ‘give meaning’ to them. Either way, these ‘symbolic-functionalist analytics’ fail to take them seriously. Since the analysts already know all about the structures and forces that make up ‘reality’, the ethnographic data have no chance to teach them anything they don’t already know. ‘Occult’ phenomena are reduced to fulfilling one or more of a few universal ‘functions’, already stipulated in the theory: ‘whether rubberized as “witchcraft”, “sorcery”, or “shamanism” [they] are collective representations – or to use a more fashionable term, social imaginaries – within a neo-Marxist (and indeed neo-Durkheimian) model of social production’ (pp. 33-4).

Pedersen gives us a penetrating description of this kind of bad social science, in which explanatory concepts are ‘fixed and therefore “dead” containers of sociocultural content and their purported political-economic context’ (p. 36). And he inveighs eloquently against restrictive ideas of social form as, ‘mental and ideological schemata through which some fixed structure of order or symbolic meaning is imposed on an inherently disorderly world of social and material practices’.

But at times Pedersen seems to draw back from the implications of his own critique, and retreats to a position that implies a strange kind of ontological apartheid. Thus on occasion (e.g. p. 38) he says this unsatisfactory conception of form, and its associated reductive functional explanations, may have some merit for Africa. They are only definitely inadequate, ‘in Melanesia, the Amazon, and northern Asia’, and he limits the alternative view he proposes by designating it ‘non-Euro-American’ (and ‘non-Cartesian’ as if these were the same thing), even though his own exposition of that alternative view draws heavily on a galaxy of indubitably ‘Euro-American’ thinkers. The implication is that the very stuff of reality, and therefore what will work as an explanatory account, is different in these two realms. Among the oddities of this formulation is that it leaves ontologically ambiguous everywhere other than ‘Melanesia, the Amazon, and northern Mongolia’ on the one hand and ‘Euro-America’ on the other, and therefore a good deal of the world and much of human history, including of course the sources and most of the life story of the Buddhism that has been so pervasively influential in northern Mongolia, and, to his credit, is integral to Pedersen’s own account of the region.

Why should this generally bold and persuasive book have occasional resort to a less cogent and more timid position that doesn’t really suit its purposes? The reason seems to be that it has been intermittently possessed by a theory, and, like a not-quite shaman with his spirits, has not quite yet managed to subdue that theory.

The theory in question is variously labelled here perspectivist, post-plural, post-human, ontological (i.e. post-epistemological), and a few other designations, which generally most insistently announce its newness. Indeed, for a view that sets itself so resolutely against ‘Euro-America’, this seems to betray a strikingly Modernist faith in the avant-garde. The theory goes (or has gone) that the ‘radical alterity’ of certain societies (in ‘Melanesia, the Amazon, and northern Mongolia’) consists not in them having different ‘socially constructed’ viewpoints on the same (natural) world, but in them living in actually different worlds. The differences between them and Euro-America are not therefore epistemological (different ways of knowing the same reality) but ontological (fundamentally different realities). With this theory, because it asserts ‘the ontological auto-determination of the world’s peoples’ (the phrase is Eduardo Viveiros de Castro’s), the problems of ‘relativism’ are said to disappear; the picture we have to grapple with now is not multi-cultural but multi-natural, and the theoretical challenge for anthropology is to develop new concepts to enable us to understand these alternative ‘natural’ realities: new notions of truth, cause, relation, etc. Luckily, by an amazing historical coincidence, many of the necessary conceptual resources were prefabricated for us by Gilles Deleuze, even though the problem in this form had not yet been discovered when he wrote, and even though he lived in a Euro-American multi-cultural ontology. But that’s a whole other story.

The signs of the book’s struggle with this theory are scattered here and there, but we may take as an example a long footnote on page 35. ‘Ontology’ refers of course to the study of, or reflection on, the question of what there is – what are the fundamental entities or kinds of stuff that exist? ‘Ontologies’ therefore refers to ‘any theories or concepts of what exists’. So, Pedersen points out, in this sense the claim that different societies have different ontologies ‘far from referring to “what really really exists” and “the true nature of things” . . . denotes “the set of things whose existence is acknowledged by a particular theory, or of a metaphysical system [like shamanism] of having such-and-such an ontology’ (citing the philosopher Ted Honderich).

This reformulation certainly evades some of the obvious problems that confronted the bolder ‘ontological’ theory: what on earth happens at the boundaries between these different ontologies, and when things or people cross from one to another? What kind of meta-ontology does one have to postulate to make sense of the thought that the world could be made up of different stuff in different places? How has this ontological pluralism come about, and do different ‘peoples’ (a strikingly unproblematized term) make their worlds different (as the notion of auto-determination would seem to imply) or do their different ideas respond to and reflect different pre-existing realities? And so on. But the price for avoiding all this is that this use of ‘ontologies’ (to refer to views about what exists rather than putting forward a claim about what exists) doesn’t do the work the other sense of the word, in the original formulation, was there to do: it does nothing to establish that people in ‘Melanesia, the Amazon, and northern Mongolia’ live in different worlds, or enjoy ontological auto-determination; it delivers not new post-plural multi-naturalism, but merely the familiar old idea that different peoples have different theories about the world. So we have, by these means, not escaped the spectres of multi-culturalism after all.

And perhaps for this reason, some of the time, including in the second half of that same footnote on page 35, Pedersen seems to feel that the original formulation was better after all. So he insists, ‘such ontologies are not simply linguistic or mental (ideational) phenomena’, thus denying just what the retreat to ontologies as theories and ‘metaphysical systems’ had conceded.

But Pedersen then continues: ‘certain material things (shamanic costumes) and certain social forms (such as joking practices) may be said to constitute distinct concepts and theories – and therefore indeed ontologies – in their own right’. So after wavering, the footnote ends with the revised ‘different theories’ position after all: we are back to the familiar idea of diverse local theories, including as embodied in objects and practices.

One other way in which this book is occasionally possessed by its theory may be seen in how Pedersen sometimes mischaracterizes some of his own best observations. He describes how he elicited many ‘consistently negative and even angry responses . . . when asking people in the Shishged to spell out the precise nature of the shamanic spirits’. Even specialists replied with teasing fantasies, mockery, and diverting comical stories: ‘Not only did people deny the presence of shamans . . . they kept making jokes about people who were clearly not shamans, endlessly pointing out to me this or that person (children included) while exclaiming, with suppressed giggles, “Watch out, Morten, he is a shaman”’ (p. 185). Pedersen once asked one of his key not-quite-shaman informants ‘Where does a spirit master go when it is not present in the game animal any more?’ The reply was dismissive: ‘He cut me off with “I have no idea,” clearly annoyed by my question. “I guess it just goes here and there”’ (p. 175). But our ethnographer won’t take ‘I don’t know’ for an answer, and brightly continues, ‘Answers like these suggest that spirit helpers are thought of as visible traces left by the invisible movements of spirit guardians, each new metamorphosis being like a shadow cast by the [spirit’s] travel across the duration of time’, which could be so, although it seems equally likely that the aggravated informant meant to suggest that he’d been asked a question to which there is no answer.

Pedersen does have an excellent idea here. He goes on to draw a contrastive parallel between these essentially motile spirits and Platonic Ideas. Instead of being unchanging entities of which people’s diverse fleeting impressions are imperfect representations, the unseen entities of shamanism are labile, as it were, ‘all the way up’. What you see is all you get. The confusing, fragmentary manifestations people encounter in a shamanic séance just is what there is. On this account, ‘genuine shamans’, those who are able to some degree to pin their spirits down and control them are, Pedersen argues, less shamanic than the not-quite shamans whose unpredictable behaviour more fully manifests the ‘fluid ontology of spirits’: ‘ontology’ here meaning merely ‘composition’.

All of this of course amply supports Pedersen’s argument that shamanism is most itself where it is fleeting, implicit, and incomplete.  But then, and at first sight unaccountably, he insists on calling this ‘a coherent cosmology imbued with a distinct logic’ (p. 180). Everything in his ethnography, his own brilliant commentary on it, and the explicit and forceful statements of his informants proclaims against there being any such thing. The reason Pedersen resorts to this idea of a ‘coherent cosmology’ emerges when he defends it not against the protestations of his informants, but against an imagined ‘social constructionist’ accusation that it is ‘antiquated’ to postulate a shared system of cultural meanings. He retorts that social constructionism is itself  ‘an obsolete paradigm’, since perspectivist multinaturalism has postulated multiple worlds. So the theory has briefly taken possession again: the ‘cosmology’ is a version of the distinct ‘ontology’ of ‘Melanesia, the Amazon, and northern Mongolia’. But although Pederson does make as good a case for it as I think may be possible, the elaborate super-ontology of perspectivist multinaturalism is neither required for his fascinating account of north-Mongolian shamanism, and nor does that account constitute evidence for it.

The central conviction of this book, I think, is that the anthropologist should take seriously the concepts and theories embodied in the ethnographic data, as concepts to think with and use in ethnographic analysis, rather than supposing that we already have all the concepts we need, because we already know all the kinds of stuff of which social reality might be composed – those dead ‘containers of sociocultural content and their purported political-economic context’. We should be open to the possibility that what we learn from our ethnography can tell us something we don’t already know about what kinds of things there are in the world. This is an important proposition for anthropology, but Pedersen seems once to have thought, and every so often to fear, that in order to sustain this conviction he needs a theory that makes ‘Melanesia, the Amazon, and northern Mongolia’ into a different ontological realm, where things can be true without having to be true for the rest of us. This seems to me both unsustainable – you have to oscillate between two different uses of ‘ontology’ to make it appear both true and interesting, but it can’t in logic be both at the same time – and also an unnecessary qualification to the central claim, which the substance of this book rather triumphantly vindicates. But this theory, although it seizes him from time to time has not taken full possession. Indeed, taken as a whole the book demonstrates brilliantly – in scrupulous and fascinating ethnography and inexhaustibly creative and insightful interpretation – that no post-anything overarching theory is needed to practise its precept.