We all know the Celestial Emporium of Benevolent Knowledge, Borges’ fictional Chinese encyclopedia, not least because Foucault put it in his Order of Things. The Celestial Emporium provides a taxonomy of “animals” including: those that belong to the Emperor, embalmed ones, trained ones, suckling pigs, mermaids, fabulous ones, stray dogs, those that are included in the present classification, those that tremble as if they were mad, innumerable ones, those drawn with a very fine camelhair brush, others, those that have just broken a flower vase, and those that from a long way off look like flies.
Let me be pedantic and explain the joke. The joke is not that fabulous animals and mermaids don’t exist. This is not about propositional truth and Popperian falsification. The joke is that the taxonomy developed in the encyclopedia has no obvious logic and no sense of whether it is sufficiently complete or not. The taxonomy lacks what Foucault called a common locus: a logic in which all these things could be brought together.
And if we’ve been teaching Foucault in social anthropology courses, chances are that we talked about Borges’ thought experiment and laughed at the joke. Because it was so obviously fantastical. But maybe the joke is on us, and we don’t even get it. This article argues that social anthropology has been moving into Borgesian encyclopedia territory over the past few decades. We have a lot of -logy about anthropos, but we have no sense anymore of how this knowledge of humans might cohere.
I take the Routledge Encyclopedia of Social and Cultural Anthropology, edited by my former colleagues Alan Barnard and Jonathan Spencer (Second Edition, 2010), as my example. I focus on this particular encyclopedia out of sheer convenience: I have been a contributor to it, and I have been using it in my teaching, so I feel more familiar with it than with other anthropology encyclopedias.
The Routledge Encyclopedia contains a large number of individual entries. Most of them are, on their own, excellent entries written by excellent scholars. I am not making any argument here about the quality of individual entries. What interests me here is: what is the organizing logic that brought all these entries together like this? What makes these entries instances of the same systematic inquiry? A critical reader might already say: “you are asking the wrong question! It’s not the job of an encyclopedia to provide a complete theory of what it means to be human. We are not doing grand theory anymore!”
Such rejection of grand theory is already suggested by the introductory explanations. Barnard and Spencer state the aim of the encyclopedia quite modestly: to offer “an accessible and provocative guide to the many things that anthropologists have had to say,” a map that will help readers “find their way around” a pluralistic and diverse set of perspectives. The editors point out explicitly that they did not want “to police what counts as anthropologically correct knowledge about the world.” And aiming for exhaustive, authoritative coverage was a mirage: “We recognize how dangerous it would be to claim that this book is complete.” That modesty seems appropriate, especially within the discipline's increasing suspicions of grand theory. It is also, as I will try to show, a symptom of a much deeper problem: that social and cultural anthropology has lost a clear sense of what an encyclopedia of social and cultural anthropology should actually be about. “What the world does not need,” they write in the introduction, “is an encyclopedia which promises the last word and the complete truth on all that anthropologists know.” I am not saying that an encyclopedia of anthropology should promise the last word and complete truth, but I am saying that an encyclopedia of anthropology should have an idea about what anthropology, as a discipline, is about.
The argument proceeds through four moves. (1) I will briefly introduce Living Value Theory as the conceptual framework that I am working with. I re-specify social anthropology's object in terms of five irreducible mediations and five levels of recursivity. Then (2) I use LVT to read the encyclopedia's organisational logic, identifying a systematic pattern of mediational imbalance. The third part examines three specific entries to show how that imbalance operates in individual entries. The fourth part situates the problem historically, arguing that the discipline has drifted toward much higher levels of symbolic recursion in ways that have progressively thinned its mediational range and mesocosmographic grounding. A brief comparative observation about archaeology, drawn from my years as an Area Editor for the International Encyclopedia of the Social and Behavioral Sciences, sharpens the diagnosis.
I. Five Mediations, Five Recursivity Levels
Living Value Theory proposes that human life is constituted through five irreducible mediations that are always co-present in any instance of coordination. Embodiment refers to the lived, pre-symbolic condition of bodies in continuous engagement with the world: fatigue, affect, perception, vulnerability, metabolic demand, and the felt sense of misalignment when things go wrong. Being-with refers to inter-recursive coordination among humans: the relational, hierarchical, obligating, and excluding processes through which people act with, against, and for one another. Dwelling refers to the non-human environmental conditions within which coordination occurs: terrain, climate, hydrology, soil, pathogen ecologies, seasonality, and the affordances and resistances of particular places. Multimateriality refers to coordination through things, tools, infrastructures, substances, built forms, and media: the material channels through which coordination extends, stabilises, and transforms across time and space. Multisymbolisation refers to coordination through classification, narrative, law, ritual, language, and other symbolic systems: the processes through which meaning is generated, shared, contested, and compressed into forms that enable rapid collective action.
These five are irreducible in the strong sense: none can be fully explained through any of the others. Embodiment cannot be reduced to symbolic representation, nor materiality to social relations, nor dwelling to cultural construction. They are always co-present: any instance of human coordination involves all five, even when analysis foregrounds only one. The methodological implication is that an adequate account of any anthropological object should show how that object operates across the full mediational field. The question is not which mediation a given phenomenon belongs to, but which mediations are foregrounded in an account and which are suppressed, and what the consequences of that suppression are.
LVT also proposes five levels of recursivity that describe how symbolic self-consciousness relates to the coordination it partially articulates. L1 refers to seamless, unreflective coordination: the ordinary flow of embodied practice in which tools are used, paths are walked, conversations proceed without requiring articulation of their own conditions. L2 refers to felt misalignment: the moment where something goes wrong, where the taken-for-granted breaks down and attention is drawn to the conditions of coordination. L3 refers to articulation: the naming, describing, and explaining of what is happening, including ethnographic description and informant explanation. L4 refers to category stabilisation: the production of concepts, schemas, and frameworks that enable rapid coordination across cases, including theoretical concepts in anthropology. L5 refers to meta-reflection: systematic inquiry into the conditions and limits of knowledge production itself, including reflexive anthropology and epistemological critique.
These levels are not a hierarchy of intellectual value. L3 description is not inferior to L5 critique; L1 coordination is not less significant than L4 theorisation. What matters for the argument that follows is how the distribution of work across these levels has shifted within the discipline, and what that shift has done to anthropology's mediational range.
II. The Mediational Map of the Encyclopedia
The Barnard and Spencer encyclopedia organises its 275 main entries in alphabetical order. The entries cover “ethnographic surveys, history of the discipline, subdisciplines and neighbouring fields, anthropological concepts and methods, and anthropological objects.” Each category contains well-selected and highly competent entries. What is missing is any principle of coordination among them. The analytical table of contents places environment, ecological anthropology, settlement patterns, and nomadism alongside postmodernism, reflexivity, discourse, and essentialism with no indication that these entries operate at different generative levels, address different mediations, or have different relationships to the conditions of possibility for human coordination. The list looks ordered. It does not specify what makes it a list of things belonging together.
Reading the encyclopedia through LVT's framework reveals a systematic pattern of mediational imbalance. Being-with is the dominant and largely invisible mediation. Kinship, descent, alliance, marriage, class, ethnicity, gender, nationalism, community, patronage, the state, colonialism, migration, violence, war, civil society, and forms of political life together constitute the largest single cluster in the volume. This dominance is so complete that it has become invisible: being-with is treated not as one mediation among five but as the substance of anthropology itself. The field has come to identify its own object with inter-recursive human coordination, leaving the conditions under which that coordination is possible as background or supplement.
Multisymbolisation is the second dominant mediation and the most explicitly theorised. Ritual, myth, religion, cosmology, classification, language, discourse, ideology, text, structuralism, symbolic anthropology, reflexivity, and postmodernism: these entries not only foreground symbolic mediation but frequently treat it as sufficient for analytic purposes. The study of symbolic systems has been one of anthropology's genuine intellectual achievements, producing analyses of remarkable internal coherence. Its cost has been a tendency toward what might be called symbolic overreach: the progressive absorption of other mediations into symbolic terms, so that embodiment becomes 'the symbolism of the body,' materiality becomes 'material culture as meaningful artefact,' and dwelling becomes 'the cultural construction of place.' In each case, a real mediation is acknowledged and then reframed as a variant of symbolic mediation.
Multimateriality is present in the volume but fragmented. Technology, material culture, money, markets, food, house, property, land tenure, pharmaceuticals, science and technology studies: each appears as a distinct specialist entry with no analytic thread connecting them as instances of the same general mediation. The encyclopedia contains the materials for a theory of multimateriality; it does not contain the theory. Things appear as topics rather than as a general mode through which coordination is conducted, sustained, and transformed.
Embodiment is partially present but consistently rerouted into specialist domains. Body, affect, emotion, medical anthropology, sexuality, reproduction, psychological anthropology: embodiment is acknowledged as a domain of inquiry but is notably absent from entries where it is in fact continuously operative, including kinship, labour, political life, and ritual. The entry on body, for example, treats embodiment primarily as a site of symbolic inscription and cultural construction, a treatment that is not wrong but that leaves the pre-symbolic, L1-L2 dimensions of embodied coordination largely unaddressed.
Dwelling is the most systematically erased mediation in the volume. The entries on environment, ecological anthropology, landscape, place, settlement patterns, nomadism, transhumance, and fishing together constitute a small cluster at the margins of the analytical table of contents. More revealingly, the ecological anthropology entry itself notes that 'during the first half of the twentieth century, social and cultural anthropology, whether in the British versions of Malinowski and Radcliffe-Brown or the American version of Boas, examined relationships within the social and cultural realm, with little direct attention to relations with the environment.' This observation is offered as a historical remark about a corrected deficit. From an LVT perspective it reads as an inadvertent acknowledgement of structural bias: dwelling has been systematically treated as background rather than as one of the constitutive conditions of all human coordination.
This distribution is not random, and its explanation does not require attributing carelessness to individual scholars. It tracks a deeper relationship between mediations and their availability to symbolic articulation. Being-with and multisymbolisation are the mediations that most readily generate narratable content: interactions, meanings, debates, genealogies, and competing interpretations. Dwelling operates on temporalities that resist symbolic narrativisation: climate cycles, soil formation, hydrological regimes, and pathogen ecologies do not unfold in the timeframes of participant observation or produce the kind of quotable material that anthropological prose handles well. Embodiment, at its most generative level, is precisely the pre-symbolic: the seamless, taken-for-granted flow of coordination that becomes visible only when it breaks down. So the encyclopedic map does not simply fail to represent certain mediations. It has been organised around what its dominant epistemological form can most readily capture.
III. Demediation in Practice: Three Entries
Close reading of specific entries confirms and complicates the pattern identified in the encyclopedia's overall organisation. The following three entries are chosen not as extreme examples but as characteristic ones, representative of broader tendencies.
The entry on resistance, authored by Spencer, synthesises the major theoretical influences on ethnographies of resistance in the 1980s and 1990s: Gramsci's hegemony, Foucault's power-knowledge, Bourdieu's practice theory, Williams and Thompson's cultural Marxism, and the Subaltern Studies historians. The theoretical combination is described as 'a heady one,' and the entry captures accurately both the intellectual excitement and the interpretive risks of the resistance literature. The entry's own conclusion identifies the central problem: the concept of resistance could lead to 'a kind of neo-functionalism in which all social and cultural phenomena might be reduced to the role they play in maintaining or subverting power relations.' This is a perceptive observation about symbolic overreach within the resistance literature. What the entry does not supply is an account of resistance grounded in the mediations that make it possible and costly: the bodily risk and endurance that resistance requires, the material resources and infrastructural conditions it depends on, the terrain within which it is conducted, and the environmental constraints that shape its timing and form. The correction offered to symbolic overreach is further theoretical refinement rather than mediational rebalancing.
The entry on food, authored by Pottier, is the closest approximation to a fully multi-mediated account available in the volume. It covers ecological conditions of production, agrarian change and labour relations, gender and household food provisioning, market structures and price volatility, famine coping strategies, and nutritional embodiment. Dwelling is present through climate, soil, and environmental availability. Multimateriality is strong in the treatment of production technologies, land rights, and market infrastructure. Being-with is sustained throughout the analysis of gender, class, and household negotiation. Embodiment appears through hunger, malnutrition, and the physical costs of labour. This mediational range is genuine and unusual within the volume. Yet even this entry leads with a symbolic frame: food is described on its opening line as 'the most powerful instrument for expressing and shaping interactions between humans' and 'a repository of condensed social meanings.' The multi-mediated richness accumulates through specific empirical sections on production and distribution; the theoretical frame still privileges symbolic mediation as the primary register of analytic attention. The food entry is the strongest case of mediational thickness available in the encyclopedia, and even it cannot fully escape the gravitational pull of the symbolic opening.
The entry on symbolic anthropology, authored by Spencer, is narratively fluent and competent as a summary of a particular disciplinary canon. It offers a clear account of the emergence of the Geertz–Turner–Schneider constellation, traces their intellectual influences, and summarises the internal critiques that followed. As a genealogy of a particular group of anthropologists who called themselves “symbolic” anthropologists, it is mostly accurate. However, as an account of symbols and symbolization in human coordination, it is thin.
The most basic omission is definitional. At no point does the entry ask what a symbol actually is. The term is introduced through the usage of the canonical authors and then taken as given. Schneider’s definition of a symbol as “something which stands for something else” is cited, but not interrogated. Turner’s analyses of ritual symbols are described, but the ontological status of those symbols is never examined. Geertz’s interpretive approach is summarised, but the conditions under which something counts as a symbol rather than as an action, an object, or a bodily event are left unspecified. The result is that “symbol” functions as a placeholder term, whose meaning is presumed to be self-evident because it circulates within a recognised tradition. This absence has further consequences. Because the entry does not specify what a symbol is, it also does not ask how symbolic mediation relates to the other mediations that are constitutive of any instance of coordination. Symbols appear as if they operate in a relatively autonomous domain, rather than as processes that are always materially instantiated, embodied in practice, situated within environments, and entangled in relations of being-with. The most immediate absence is multimateriality. There is no sustained attention to the fact that symbolic systems are inseparable from specific material forms and evolving inscription systems: oral transmission, tablets, paper, print, archives, recording devices, and digital media. The entry proceeds as if symbolic systems could be discussed without any reference to their multimaterial co-mediations.
That assimilation raises a further problem. The equation of symbols with language is simply assumed rather than demonstrated. Symbolic anthropology’s analogy between culture and language is reported, but not examined. There is no consideration of whether all symbolic systems are in fact language-like, or whether ritual performance, bodily gesture, spatial arrangement, and material form operate through different kinds of symbolic processes that cannot be reduced to linguistic structure. In the case of ritual, which figures prominently in Turner’s work, the question is particularly acute: what exactly are the symbols in a ritual performance? Are they discrete elements that can be decoded, as Turner’s analyses suggest? Are they patterns of action? Are they embodied states? The entry reports the debates but does not engage the underlying problem. What results is a serious reduction. The entry merely reproduces the self-descriptions of a group of anthropologists without ever asking even the most basic questions about symbols and how they relate to the other irreducible mediations.
The contrast between these three entries supports a broader diagnosis. The decisive variable in mediational thickness is not the quality of the analyst but the ontological pressure the topic exerts. Food resists demediation because survival makes the full mediational field visible: analysis that ignores climate, production, or hunger becomes visibly inadequate. Resistance invites symbolic compression because its object is already highly symbolised by participants and analysts alike. Symbolic anthropology is, in a sense, the limit case: an entry whose topic is symbolic mediation itself, where the analytic pull toward symbolic closure is at its strongest. The encyclopedia does not contain what might be called a fully multi-mediated entry in the strong LVT sense. What it contains are entries where demediation is harder to sustain, and entries where it can proceed almost without friction.
IV. The Recursivity Drift
The organisational pattern described above has a temporal dimension. The second edition of the Barnard-Spencer encyclopedia, published in 2010, is not simply a snapshot of a static imbalance. Compared with earlier reference works in the discipline, and with comparable volumes from the 1970s and 1980s, it shows a clear increase in entries organised at L4 and above. The new entries commissioned for the second edition include affect, cosmopolitanism, counterinsurgency, diaspora, engagement, ethics, finance, gay and lesbian anthropology, genocide, human rights, neoliberalism, NGOs, sovereignty, and well-being. These are not supplementary topics added to an unchanged field. They represent a reorientation of the discipline toward higher levels of symbolic recursion: toward concepts that operate on other concepts, that stabilise interpretive positions, or that reflect on the conditions of knowledge production itself.
This shift has produced genuine intellectual gains. Anthropology is considerably more self-aware about its own categories and their historical and political situatedness than it was in 1970. The critiques of evolutionism, structural functionalism, and naive cultural holism that accumulated through the 1980s and 1990s were substantively important, and the field is better for them. The attention to discourse, power, and representation has opened areas of inquiry that earlier frameworks could not reach.
The unintended structural consequence has been a progressive thinning of the mediational field. L4 and L5 work operates almost entirely through multisymbolisation: it produces and refines conceptual vocabularies, critiques inherited categories, and analyses the conditions under which knowledge is generated. As the discipline becomes more invested in work at these levels, the mediations that are not primarily symbolic become harder to sustain within the dominant analytic vocabulary. Dwelling, which operates on timescales and material processes that resist conceptual elaboration, becomes increasingly peripheral. Embodiment, at its most important pre-symbolic level, becomes legible only when it is already being articulated, which means it is already being partially misrepresented. Multimateriality persists as a set of distinct research areas but loses any unifying account of what makes them instances of the same general mediation.
The encyclopedic form accelerates this process. An encyclopedia accumulates entries without specifying their generative order. As L4 and L5 categories multiply, they acquire the same ontological flatness as L3 descriptions and the same apparent independence as distinct research domains. The result is a catalogue in which 'environment' and 'postmodernism' appear as neighbours with equal bibliographic weight and no mechanism to specify that one concerns the conditions of possibility for all human coordination and the other concerns a meta-reflective orientation toward symbolic accounts of coordination. The form enacts the very demediation it otherwise leaves unaddressed.
A related and more specific shift concerns what anthropology does with what might be called the native's point of view. The original commitment to understanding people's lives from within their own frames of reference contained two tasks that have since come apart. One was to represent accurately how people understand and describe their world: their categories, narratives, and interpretive frameworks. The other was to grasp what is at stake for people within those worlds: the embodied pressures, material constraints, environmental conditions, and relational obligations that give their actions their force. As the discipline has moved toward higher levels of symbolic recursion, these two tasks have diverged. The analysis of representation has become increasingly sophisticated while the analysis of what is at stake has become progressively thinner. The discipline has become, in a specific technical sense, more interested in how worlds are described than in the conditions within which people must coordinate their lives.
V. The Archaeology Contrast
During the author's years as an area editor for the International Encyclopedia of the Social and Behavioral Sciences, overseeing entries across anthropology, archaeology, health, and research methods, one pattern recurred with sufficient consistency to be striking. Archaeology entries tended to feel less theoretically elaborate than those in social anthropology. They were also more analytically secure, more resistant to interpretive drift, and more consistently anchored in the specific conditions of the phenomena they described. At the time this difference was easy to register but difficult to explain. From the perspective developed here, the structural reason becomes clear.
Archaeology works primarily with materials that are non-symbolic or only partially symbolic: soils, sediments, bones, settlement layers, artefacts, pollen sequences, isotope ratios, and environmental traces. These materials do not submit readily to purely discursive elaboration. Dwelling is structurally unavoidable: terrain, hydrology, climate, and ecological constraints are not background for archaeological analysis but primary evidence. Multimateriality is the discipline's basic evidential field, not a specialist subdomain. The consequence is that archaeology cannot drift as far into L4 and L5 symbolic closure as social anthropology can. Its object exerts ontological pressure that prevents certain kinds of mediational erasure.
The paradox is that what reads within anthropology as archaeological undertheorisation is often greater mediational completeness. Archaeologists do not typically forget dwelling because they cannot: their evidence is embedded in it. They do not typically reduce their objects to symbolic systems because so much of what they study was never symbolically encoded in ways that are recoverable. They cannot easily perform the move that symbolic anthropology perfected, of abstracting a cultural logic from practices and then treating that abstraction as the explanation of the practices.
The archaeological record of agriculture, settlement, and urbanisation makes the asymmetry fully explicit. The emergence of agriculture in the Fertile Crescent, the Nile floodplain, the Indus basin, and the Yellow River valley was shaped primarily by very specific alignments of rainfall, river sedimentation, temperature stability, and plant domesticability. The rise of dense urban settlement depended on the ability to stabilise water, waste, and food flows under particular ecological constraints. The epidemiological transitions that accompanied domestication and density were shaped by proximity to animals, waste accumulation, and the specific pathogen ecologies of particular environments. None of these transformations are primarily stories about being-with or symbolic systems, though both were significantly reorganised by them. They are stories about dwelling conditions and the material-symbolic systems that emerged to coordinate life within and across those conditions. Archaeology cannot forget this. Social anthropology, which studies largely the downstream consequences of these long-term transformations, has organised itself as if the upstream conditions were not part of its object.
VI. The Four-Field Parallel and Its Failure
The five mediations proposed by LVT bear a striking, if deceptive, resemblance to the classic four-field structure of American anthropology. Biological anthropology aligns approximately with embodiment, though LVT's concept of embodiment extends well beyond physiology or evolutionary biology to include the lived, pre-symbolic field of bodily coordination. Social and cultural anthropology aligns with being-with, though it has consistently over-identified the discipline's object with this one mediation. Archaeology aligns with dwelling and multimateriality, though it has tended to collapse these into a single domain of material culture without distinguishing between non-human environmental conditions and human-built infrastructural forms. Linguistic anthropology aligns with multisymbolisation, though it captures only a portion of the wider field of symbolic coordination that includes ritual, law, classification, narrative, and institutional categories.
This parallel is not accidental. It suggests that the discipline has long carried an implicit structural awareness that human life unfolds across distinct but interrelated domains that cannot be reduced to one another. The four-field model was an attempt, however imperfect, to hold this plurality in view within a single disciplinary formation.
What the model never provided was an account of those fields as instances of a shared underlying object: as different foregrounded mediations of a single, always multi-mediated process of coordination. Without that account, the fields developed as parallel specialisms with their own methods, data, and theoretical traditions. The connections among them were acknowledged in principle but never theorised in a way that made them analytically necessary. The gradual erosion of the four-field model in recent decades, the retreat to more autonomous disciplinary formations, is not only an institutional development. It reflects the loss of whatever organisational intuition the model was obscurely tracking.
The contemporary encyclopedic form inherits this failure at scale. It presents plurality without coordination: a set of legitimate specialisms that can be listed and cross-referenced but whose mutual relationships have no specified basis. What appears as intellectual richness, the sheer range of anthropological work covered in 275 substantial entries, is in part the product of a fragmentation whose structural causes have not been diagnosed.
VII. A Note on Holism
The preceding argument will predictably be read as a call for holism: an attempt to grasp 'the whole' of a society or culture through a single integrated framework. The resemblance to holistic ambitions is not entirely accidental. Both the present argument and earlier holistic anthropology resist the reduction of human life to a single domain.
The difference lies in the underlying ontological commitments. Classical holism in anthropology typically posited a relatively stable social or cultural whole organised around a dominant symbolic or normative order. The analytic task was to reconstruct that order and show how it integrated the elements of the whole. Whether framed in functionalist, structuralist, or symbolic terms, the presumption was that coherence was given and that the analyst's task was to describe it. This presumption made holistic anthropology, paradoxically, a form of demediation: by attributing unity to the social or cultural whole through a dominant symbolic logic, it compressed the plurality of mediations into a single register. The appearance of completeness was purchased at the cost of the very plurality it claimed to describe.
The mesocosmic perspective that follows from LVT begins from irreducibility and instability rather than from integration and order. The five mediations are not components of a system but conditions of possibility for coordination: they are always co-present, always partially misaligned with one another, and never fully integrated into a coherent whole. Coordination is ongoing work, subject to friction, breakdown, and partial repair. What appears as social order at any moment is a temporary, incomplete, and contested outcome of multiple mediations being brought into provisional alignment. The analytic task is not to describe a given whole but to track the processes through which anything like a whole is temporarily achieved, and to attend to the mediations that are suppressed or erased in the course of that achievement.
The difference is not merely terminological. It changes what counts as a good anthropological account, what counts as significant evidence, and what counts as a satisfying explanation.
VIII. Toward an Anthropology Re-anchored
The argument so far has been primarily diagnostic. The final section offers orientation rather than prescription, sketching what anthropological practice might look like if the mediational imbalance described here were taken seriously as an analytic problem rather than an incidental limitation.
The most consequential shift would be in where inquiry begins. Current anthropological practice tends to begin from already-named topics or domains: kinship, ritual, markets, violence, identity. These topics are stabilised at L4; they are concepts that organise fields of inquiry and make comparisons possible. Beginning from topics has real advantages, but it also builds in the mediational structure of the field's existing vocabulary. If dwelling and embodiment are systematically underrepresented in that vocabulary, beginning from topics will reproduce that underrepresentation even in research that aims to address it.
An anthropology re-anchored in mediational coordination would begin from L2: from felt misalignment, from the moments and sites where coordination strains or breaks down, where the taken-for-granted becomes visible as contingent. This is not a prescription for atheoretical empiricism. It is a claim about the generative site of inquiry: that the most productive anthropological problems arise from attending to what goes wrong in coordination, and that the mediations most relevant to any given problem should be determined by that problem's specific character rather than by the field's inherited vocabulary.
Every object would be treated as fully multi-mediated by default. This does not require that every study address all five mediations with equal intensity. It requires that mediational omissions be registered as such and their consequences considered, rather than assumed to be methodologically neutral simplifications. When dwelling drops out of an account of ritual, or embodiment drops out of an account of class, or multimateriality drops out of an account of religious practice, the question to ask is what work that omission is doing and what it is concealing about the coordination under examination.
The field's relation to archaeology, biological anthropology, and ecology would need to be reconceived along similar lines. The current model of interdisciplinary exchange, in which neighbouring fields are cited when their data are useful and otherwise held at disciplinary distance, perpetuates the very fragmentation this article diagnoses. An anthropology organised around mediational coordination would treat the conversation with archaeology not as generous border-crossing but as an analytic necessity: the mediations that archaeology foregrounds are the ones social anthropology most consistently suppresses, and any claim to comprehensive understanding of human coordination requires attending to both.
Finally, the encyclopedic form itself would need to change. A reference work organised around mediational coordination would not replace alphabetical lists with thematic sections; that would be a cosmetic revision. It would need to specify what makes its entries instances of the same inquiry, how different domains of study relate to a shared underlying process, and which entries describe conditions of possibility for coordination and which describe particular stabilised forms of it. Such a work would be considerably harder to write than the current encyclopedic tradition demands. It would also be considerably more useful as an account of what anthropology knows.
Conclusion: The Map That Cannot Navigate
The editors of the Barnard and Spencer encyclopedia describe their book as a map to help readers find their way around the anthropological landscape. The map metaphor repays examination. A map, in its most basic sense, presupposes bodies moving through terrain. It relates embodiment to dwelling: it is useful because it translates the co-constituted field of bodies-in-environments into a form that can be consulted when direct navigation fails. A map registers gradient, resistance, obstacle, and affordance. It is oriented toward the practical question of how to move through conditions that exceed immediate perception.
The encyclopedic map offered by the volume under discussion does the opposite. It maps terms to terms, debates to debates, concepts to concepts. It is organised around the two mediations most available to symbolic articulation, and it systematically marginalises the two most constitutive of the conditions within which that articulation occurs. It is, in the specific sense that matters here, a map that has quietly abandoned the mediations that make mapping meaningful.
Borges's Chinese encyclopedia is memorable because its categories are vivid and each is internally coherent, yet they lack a common locus. Contemporary anthropology's encyclopedic tradition is not absurdist: its categories are serious, its entries are often excellent, and its coverage is genuinely wide. The problem is structural rather than qualitative. The discipline has accumulated a great deal of knowledge about human coordination without specifying what principle connects that knowledge, what conditions of possibility it describes, and what mediations it has had to suppress to achieve its particular forms of clarity.
The recovery this article calls for is not the last word and complete truth of all human practices across all cultures and historical eras. What it calls for is a renewed recognition that to be human is not just to be with other humans in various constellations. Anthropology studies more than human-human interactions. To be human is to live, at once, through embodiment, being-with, dwelling, making, and symbolizing.