Medical anthropology has produced some of anthropology's richest ethnography but surprisingly little cumulative theory. This article argues that the problem lies less in particular theoretical schools than in a recurrent classificatory operation inherited from Enlightenment thought and hardened by the nineteenth-century division between the natural and the human sciences. Living Value Theory calls this operation a SWIPE: a split world with incommensurably paired epistemologies, in which two domains are not merely distinguished but reified as independently constituted and assigned separate, mutually untranslatable ways of knowing. Medical anthropology's foundational disease/illness distinction reproduced the larger nature/culture partition by granting biomedicine pathology and anthropology experience and meaning. Subsequent approaches repeatedly discovered phenomena excluded by earlier partitions, generating concepts such as embodiment, structural violence, local biology, biosociality, syndemics, materiality, and multispecies relations. Yet the underlying operation usually survived, now accompanied by claims that its two poles were relational, porous, or entangled. The result has been a theoretical history in which empirical discoveries repeatedly undermine the ontology through which they are interpreted, while the ontology itself survives each undermining.
Health makes this failure especially consequential because health and sickness are intrinsically multimediated phenomena in which nonsymbolic processes remain causally decisive regardless of articulation. The article proposes replacing binary repair work with an account of living as recursive mediation across multisensorial embodiment, being-with, multimaterial forming, multiversal dwelling, and multisymbolism. Medical anthropology can then retain virtually all of its ethnographic discoveries while relinquishing much of the theoretical machinery built around them.
Introduction: The Strange Conceptual History of an Empirically Brilliant Field
Medical anthropology has produced outstanding ethnographies of suffering, caregiving, psychiatry, pharmaceuticals, epidemics, reproduction, inequality, technology, disability, and healing. Arthur Kleinman showed why technically correct treatment can fail when clinicians ignore what illness means in a patient's life (Kleinman 1980). Paul Farmer showed that tuberculosis and AIDS follow the contours of poverty and racism as reliably as they follow the biology of infection (Farmer 2004). Margaret Lock showed that menopause is not the same biological event in Kyoto and Massachusetts (Lock 1993). Annemarie Mol showed that atherosclerosis is not one object seen from several angles but several practices that only sometimes cohere into something singular enough to be called a disease (Mol 2002). Elizabeth Roberts showed that a lead-glazed cooking pot in Mexico City is simultaneously a hazard, a flavor, a grandmother, and a nation (Roberts 2025). These are not modest achievements. They are among the most sophisticated ethnographic accomplishments in the discipline.
Yet ask what medical anthropology's accumulated theory now tells us, in general, about the nature of health and sickness, and the answer is strangely disappointing. There is no cumulative theoretical architecture comparable to the empirical richness that stands beside it. Instead there is a succession of concepts and turns: disease and illness, explanatory models, embodiment, structural violence, local biologies, biosociality, syndemics, pharmaceuticalization, biological and therapeutic citizenship, materiality, assemblage, multispecies relations, planetary health, datafication. Each generated productive scholarship. Yet the relationship between them remains largely additive. A graduate seminar can walk through all of them in a semester, in roughly chronological order, and finish with a longer vocabulary rather than a sharper account of what health actually is.
The usual history of the field is too flattering to this outcome. It runs roughly as follows: biomedicine was reductionist; interpretive anthropology added meaning; critical anthropology added power; biocultural anthropology reintegrated bodies; the material turn added things; multispecies approaches added nonhumans; contemporary anthropology, chastened and sophisticated, has finally learned to appreciate complexity, relationality, and entanglement. On this account the field's theoretical history is a story of steady enrichment, each generation folding in something the last generation missed.
This article argues that this history mistakes repeated conceptual repair for cumulative theoretical progress. What actually happened, again and again, took a more specific shape. A generation of scholars drew a distinction that split reality into two domains and assigned each domain its own way of knowing. Ethnographic fieldwork then discovered, reliably, that living processes violated the distinction. A new concept was invented to describe the relation between the two now-troublesome terms: embodiment to reconnect body and meaning, structural violence to reconnect political economy and physiology, local biology to reconnect universal biology and cultural variation, materiality to reconnect symbol and thing. The reconnecting concept was received as a theoretical advance. In every case, the deeper classificatory operation that had made the reconnection necessary was left standing, available to generate the next round of trouble. Call this the binary machine: SWIPE, empirical failure of the SWIPE, repair concept, apparent theoretical advance, survival of the original partition. This article traces that cycle through medical anthropology's history and proposes a way to stop it.
One qualification belongs here from the outset, because it is easy to caricature this argument as a demand that anthropologists stop drawing distinctions. That is not the claim, and the difference matters enough to state formally before proceeding. Living Value Theory distinguishes ordinary binaries and gradients from a more consequential kind of division it terms a SWIPE, or split world with incommensurably paired epistemologies. A SWIPE has three features together, not any one alone: it treats two terms as reified, independently constituted domains rather than as analytically differentiated aspects of one process; it assigns each domain a separate and mutually untranslatable epistemology; and it treats the resulting partition as a starting point for inquiry rather than as a finding inquiry might overturn. Hot and cold sit on a gradient and cause no particular epistemic damage. Nature and culture, as reconstructed below, do all three things at once. The criticism developed here is therefore not that biomedicine and anthropology draw too many distinctions. It is that one founding distinction did all three things, was not required by the object it was applied to, and was applied to a body that had never respected it.
What would happen if medical anthropology retained its ethnographic discoveries but discarded the classificatory architecture that made those discoveries repeatedly appear as theoretical surprises? That question drives the rest of this article.
I. Before Medical Anthropology: How a Modern Epistemic Settlement Divided the Object
To see why medical anthropology's theoretical history took the shape it did, it helps to notice that Kleinman is not the origin of the problem. The distinction that organized medical anthropology's disciplinary settlement in 1978–1980 descends from a much older settlement, one that crystallized across Enlightenment philosophy and hardened into institutional form during the nineteenth-century separation of the Naturwissenschaften from the Geisteswissenschaften.
The canonical version of this settlement splits nature from culture. But it was never merely a split of substances into two kinds of stuff. From the outset it was simultaneously an epistemological division: nature was assigned explanation, causality, and law-like regularity, to be known through causal explanation, or Erklären; culture was assigned meaning, interpretation, and historical specificity, to be known through interpretive understanding, or Verstehen. This opposition became institutionalized in the Methodenstreit of the late nineteenth century and runs through every subsequent argument about whether the study of human life can or should be scientific (Weinberg 2021 gives the clearest recent statement of how deeply this partition still organizes the ontology of disease itself, a point Section II returns to directly).
The significance of this settlement is institutional rather than merely philosophical, and its consequences for the human sciences were not a matter of inheriting one half of a completed picture. Living beings coordinate their lives through several irreducible and simultaneous mediations, not through nature and culture as separate compartments. The settlement cut those mediations apart and distributed the pieces among incompatible disciplines: embodiment to biology and physiology; the non-human environmental conditions of dwelling to natural history and later to geography; material arrangements, tools, and infrastructures to economics and engineering, with only their symbolic aspects left to social inquiry; and the relational fabric of life narrowed to human-human interaction, formalized through norms and institutions, stripped of the human-animal and human-more-than-human relations ethnography would spend the following century rediscovering as constitutive.
Anthropology took shape inside this settlement and partly defined itself by accepting jurisdiction over its cultural half. This produced real intellectual possibilities and an equally real constraint: culture became the discipline's trump card, and meaning became a privileged concern, sometimes privileged well past the point where it could do justice to what anthropologists were observing in the field. Every major tradition of social thought that grew up around the settlement inherited a fragment and tried, in its own way, to repair the wound rather than to notice that the wound had been avoidable; that broader argument belongs to the study of social theory generally and need not be rehearsed here in detail (Ecks 2026c develops it at length). What matters for the present argument is narrower and more specific: medical anthropology inherited both the fragment and the disciplinary permission structure that made the fragment hard to see as a fragment.
This arrangement damaged the human sciences broadly, but medical anthropology suffered from it with unusual severity, because its objects refuse to stay within symbolic or social jurisdiction. Illness immediately implicates bodies, sensations, microbes, food, temperature, sleep, kin, clinicians, architecture, work, medicines, transport, money, diagnosis, memory, and fear, all at once, and none of these can be bracketed away as merely contextual without distorting what is actually happening to a sick person. Political and religious anthropology can, for considerable analytical stretches, proceed through symbolization and institutional classification without excessive damage to the object; a claim Section V returns to and qualifies more carefully. Health resists this far more completely. A person with a fever is not experiencing a text.
Medical anthropology, in other words, did not inherit the wrong theory. It inherited an object that had already been divided before it began studying it, together with a set of disciplinary permissions that made the division unusually difficult to notice, because noticing it meant trespassing across boundaries the wider apparatus of modern knowledge had been built to enforce.
II. 1978–1980: The Consolidation of the Disease/Illness Settlement
If Section I explains where the wound came from, this section examines the moment medical anthropology gave that wound its most durable and clinically successful medical form. In 1978, Kleinman, Eisenberg, and Good distinguished disease, the practitioner's account of pathological process, from illness, the patient's experience and interpretation of suffering (Kleinman, Eisenberg, and Good 1978). Kleinman's 1980 book Patients and Healers in the Context of Culture extended this into the explanatory-model framework, showing that patients and healers routinely understand the same episode through different accounts of cause, seriousness, treatment, and prognosis (Kleinman 1980).
The distinction was not new in 1978, and recognizing this strengthens rather than weakens the argument that follows. Andrew Twaddle had already proposed a tripartite separation of disease, illness, and sickness a decade earlier (Twaddle 1968), and Horacio Fabrega had spent the preceding years arguing that disease concepts carry deep logical and social assumptions that medicine and social science handle in incompatible ways (Fabrega 1972, 1974). Darin Weinberg has since argued, from medical sociology, that the disease/illness distinction as such rests on an a priori nature/culture partition, assigning objective disease to nature and subjective illness to the domain of meaning (Weinberg 2021). Kleinman's achievement was therefore not the invention of a SWIPE that had not existed before. It was the consolidation of an already-circulating social-scientific SWIPE into a form so clinically compelling, so pedagogically portable, and so institutionally useful that it became medical anthropology's founding settlement rather than one contribution among several. That is a stronger claim than novelty, not a weaker one: it means the operation this article is describing was not a local mistake made once, in one field, by one author, but a general pattern that a series of scholars, working independently across sociology, anthropology, and psychiatry, converged on and then, in Kleinman's case, gave a form durable enough to organize an entire subdiscipline.
The credit due to this settlement should be stated before the criticism, because the criticism means something only against it. The disease/illness distinction legitimized patient experience as a serious object of inquiry at a moment when biomedicine had few resources for taking it seriously. It explained why technically competent treatment routinely fails when clinicians do not understand what sickness means in a patient's life. It brought causation, expectation, meaning, and treatment-seeking directly into the clinical encounter. And it gave anthropology a durable and genuinely useful role inside medical institutions, one that continues to improve care.
The price of this achievement has been underexamined for as long as the achievement has been celebrated. The settlement organized itself as a division of labor: medicine retained disease, anthropology received illness. Mapped onto the settlement described in Section I, this division reproduces biology against culture, pathology against meaning, the objective against the subjective, explanation against interpretation. It satisfies all three criteria set out above for a SWIPE rather than an ordinary binary: it reified two domains, assigned each an incompatible epistemology, and installed the division as medical anthropology's starting point rather than as a hypothesis ethnography might overturn.
An observation about the field's own history exposes something the standard account has largely missed. DSM-III, formalized in 1980, and the explanatory-model tradition consolidated in the same period are usually narrated as opposites, one biomedical and reductive, the other anthropological and humanizing. They are better understood as structural twins. Pre-1980 American psychiatry was not simply an ontology of psychoanalytic compromise formations; it was heterogeneous, encompassing biological, psychodynamic, and social currents in uneasy coexistence, and the reliability crisis that motivated DSM-III was real and well documented (Decker 2013 gives the fullest account). What DSM-III did was reorganize this heterogeneity around operationalized, checklist-based criteria designed to be applied consistently regardless of theoretical orientation, treating an inherently self-revising phenomenon as though it were fixed. Kleinman's tradition looks like a critique of this move. It functions, structurally, as its complement. Where DSM-III renders the patient a bearer of symptoms to be classified, the explanatory-model tradition renders the patient a bearer of cultural meanings to be elicited and interpreted. Both operations treat the patient as something to be read rather than as someone actively shaping, and being shaped by, the clinical encounter, and both locate the interpretive and transformative work primarily on the professional's side of the exchange. This structural isomorphism between symptom-reading and meaning-reading has gone largely unremarked because the two traditions have spent decades presenting themselves to each other as opponents (Ecks 2026g).
The stronger version of this criticism does not require claiming that Kleinman's framework makes the patient a wholly stable input incapable of negotiation; his own 1978 model explicitly described explanatory-model elicitation as a transaction between clinician and patient, and negotiation was never absent from it in principle. The sharper problem lies elsewhere. Even a negotiated explanatory-model framework represents the products of an ongoing recursive process, a patient's shifting, strategic, self-revising engagement with illness and with the clinician, as though they were a bounded object: an explanatory model, elicited once or periodically, rather than a continuously unfolding activity. The patient may change from visit to visit and often does. The analytic vocabulary available to describe this change continually restabilizes what they say into "their explanatory model," a noun phrase that performs closure on a process that has not, in fact, closed. Calling clinical interaction translation compounds the problem, because it assumes two pre-existing, internally coherent frameworks waiting to be bridged, when what is more often occurring is a vertical compression of richly layered, temporally extended articulation into an institutionally actionable category, a compression that is frequently necessary and often well performed, but that the translation metaphor does not describe because it is not translation.
What gets lost when the patient is treated, even provisionally, as a stable source of an elicitable model is not a minor methodological nicety. Patients anticipate their doctors, change their stories between visits, withhold information strategically, experiment against clinical advice, comply in ways calibrated to what compliance will earn them, and reorganize their relationships in response to a diagnosis. The framework's own vocabulary works against capturing this: calling something an explanatory model treats it as a bounded object rather than an unfolding process, and calling something context already establishes a core and a surrounding, a distinction the multimediated character of illness does not support.
The unbundling this article proposes should not be mistaken for a new two-column mapping in which disease sits on one recursive level and illness on another; that would simply install a second SWIPE where the first one stood. A tumor, an infection, or a kidney stone may be a largely nonrecursive biological process, unaffected by how it is named. The institutional label "stage two carcinoma" is an L4 stabilization. The ache that made someone seek a scan may register at L2, and the story they tell a friend about it at L3. All four of these can belong to a single case, in almost any combination, and both disease and illness, as ordinary words, span several mediations and several recursivity levels at once rather than mapping cleanly onto either one. Disease is not simply L4 and illness simply L1 through L3. What the disease/illness distinction actually does is bundle several different processes together, biological process, felt disturbance, articulation, and institutional classification, and then distribute the bundle across two disciplines as though it were a single joint with a single correct place to cut. The more precise description replaces the bundle with an account of how a particular disturbance moves through mediations and levels in a specific case, which is an empirical question, not a definitional one.
Kleinman did not invent the SWIPE this section has described. He gave an already-circulating social-scientific partition one of its most durable and clinically productive twentieth-century forms. Because that form worked so well for a specific institutional purpose, the following four decades of medical anthropology were substantially organized around escaping a settlement the field's founding synthesis had consolidated.
III. The Repair Industry: Fifty Years of Rediscovering What the Original Cut Excluded
The history that follows 1980 can be organized as a series of repair concepts: theoretical innovations that restored to view something the disease/illness settlement had placed out of bounds. The argument here is not that these concepts were mistaken or that the scholarship built on them should be discarded; almost none of it should be. The argument is that their empirical findings survive intact while their status as theoretical breakthroughs looks considerably less secure once the underlying pattern is visible, and that not every case fits the pattern equally well, which is itself informative.
Embodiment is the clearest case. The claim, associated with Csordas and the broader phenomenological turn in medical anthropology, was that bodily practice, sensation, and skill are not mere vehicles for cultural meaning but sites of meaning-making in their own right (Csordas 1990). The ethnography this generated, on pain, ritual, and bodily comportment, remains valuable. But the breakthrough character of the claim depended entirely on the prior settlement having ceded the body to biomedicine and made symbolic interpretation anthropology's privileged jurisdiction. Embodiment is, in substantial part, the return of something that should never have been allowed to disappear from anthropological view.
Structural violence deserves more careful handling, because Farmer's contribution is not reducible to a binary. His demonstration that poverty, racism, and political arrangements shape the distribution of tuberculosis, HIV, and maternal mortality is an empirical achievement that survives completely (Farmer 2004). What repays scrutiny is the architecture the concept repairs rather than its substance: a prior separation between structure and embodied life. The metaphor that structural violence "gets under the skin" presupposes a question that should not need asking: where was structure imagined to be before it got there? Political economy is lived, directly and without residue, through food, housing, labor, transport, and care. When Pfeiffer and Chapman traced how structural adjustment policies imposed by international financial institutions eventually appeared as empty shelves and absent clinicians in particular clinics, nothing crossed from a structural realm into a biological one (Pfeiffer and Chapman 2010). A single multimediated process unfolded, with no seam at which one register handed off to another. The concept performed indispensable work in making this visible against a discipline still organized around the disease/illness settlement, but the getting-under-the-skin metaphor only needed inventing because the skin had first been imagined as a boundary.
Local biology is perhaps the cleanest case, because it names its own repair operation almost explicitly. Lock's comparative research on menopause in Japan and North America showed that biological process, reported symptom, medical category, diet, and life history cannot be cleanly separated (Lock 1993), and Lock and Kaufert extended this through the concept of local biologies more directly (Lock and Kaufert 2001), later broadened through epigenetics and microbiome research (Niewöhner and Lock 2018). The SWIPE this concept repairs is universal biology set against local culture, biology as constant substrate, culture as variable overlay. Once stated plainly, the theoretical astonishment of "local biology" as a discovery is harder to sustain, because the implicit alternative it argues against, a body developing outside any developmental environment, was never coherent. Lindenbaum's account of kuru made the point vividly decades earlier: mortuary practice and prion transmission were not two separate explanations of the same epidemic (Lindenbaum 1979). The epidemic existed through their conjunction.
The hyphenated vocabulary clustering around this problem, biosocial, biocultural, sociosomatic, psychosocial, deserves a brief note rather than a lengthy case study. Every hyphen commemorates the wound it is meant to heal. Each term promises integration while grammatically reproducing the terms whose prior separation supposedly requires it. This is not always a useless move; Jenkins and Cofresi's sociosomatic analysis of depression and political violence, and Dressler and Bindon's demonstration that cultural consonance carries measurable cardiovascular consequences, are genuine contributions (Jenkins and Cofresi 1998; Dressler and Bindon 2000). But the hyphen is conceptually revealing precisely because it shows a discipline reaching for integration using a grammar built for separation.
Syndemic theory is a harder case, and it deserves treatment as one rather than as a predetermined casualty of the same pattern. Singer's demonstration that diseases cluster and interact under conditions of social disadvantage identifies a specific mechanism: mutually intensifying pathological processes patterned by structural inequality, such that HIV, substance use, and depression can worsen one another's course in ways a single-disease model will miss (Singer 2009; Singer et al. 2017). This is not merely a relabeling of an old binary. It names something a multimediation account would also predict, cross-mediational effects between co-occurring conditions, but names it with enough specificity to guide clinical and epidemiological attention in ways the general framework alone does not. Syndemic theory should therefore be read as surviving in a narrower and more precise form than it is usually claimed for: it is a genuine descriptive and mechanistic concept for a real class of interactions, not a general repair of the biological/social divide, and not a claim that disease entities were ever properly isolable to begin with. Multimorbidity, the now-common situation of a patient managing several chronic conditions simultaneously while coping with employment and financial strain, is what should be expected once one stops assuming that living processes decompose into single-disease units; work on how treating one condition can worsen another through fragmented, multiply medicated care shows this concretely (Ecks 2021). But syndemics earns a place in the vocabulary anyway, because naming a specific interactive mechanism has real analytical and clinical value that a purely general multimediation claim does not supply on its own.
The material turn in medical anthropology, exemplified by work on the social lives of pharmaceuticals, requires explaining why a turn toward materiality was necessary at all (Whyte, van der Geest, and Hardon 2002; Hardon and Sanabria 2017). It was necessary because meaning had been made analytically primary by the settlement described in Section II, leaving drugs, devices, and infrastructures analytically thin by comparison. Pharmaceutical anthropology is an especially clear illustration because medicines are simultaneously chemical, commercial, relational, symbolic, bodily, and institutional at once and without priority among these registers, their worth negotiated across all of these registers together rather than fixed by pharmacology alone (Ecks 2014, 2022). The empirical richness of this scholarship survives its reframing entirely intact; what looks less secure is the idea that a turn was theoretically required to notice what should never have needed separate noticing.
The feminist and decolonial critique that runs alongside this material turn deserves the same treatment. Chaparro-Buitrago's account of forced sterilization in Peru shows that a rights framework organized around reproductive choice and infertility can register one harm while missing what women themselves described as a loss of fuerza, bodily strength connected to work, family, and everyday capacity, a loss that a narrower legal or biomedical vocabulary has no way to name (Chaparro-Buitrago 2022). The methodological lesson generalizes beyond this case: an adequate account should not assume in advance which mediation an intervention has damaged, but discover what has actually become difficult, fragile, or unavailable in a person's life, which is precisely the kind of question a mediational analysis is built to ask rather than to answer for it.
Multispecies anthropology completes the sequence. The turn toward ethnographies of animals, plants, fungi, and microbes alongside humans (Kirksey and Helmreich 2010) has been especially consequential in the microbial register: research on the microbiome and on antibiotic-driven losses of microbial diversity has destabilized the bounded individual organism that biomedicine, and much of anthropology alongside it, had long assumed as a starting point (Blaser 2021; Paxson and Helmreich 2014). But microbes did not become constitutive of human health only once anthropologists began attending to them. They needed rediscovering as part of human health only because human sociality had been analytically purified from the wider living world by the same settlement that separated nature from culture in the first place.
Read together, these cases permit a rewriting of medical anthropology's history that differs from the standard account offered in the Introduction. The field's theoretical trajectory was not a sequence of expanding theories, each building cumulatively on the last. It was a sequence of returns, ethnography rediscovering mediations an earlier, avoidable cut had placed out of view, with theory naming the rediscovery as a breakthrough more often than it recognized the rediscovery as evidence against the cut itself. Syndemics is the clearest reminder that the pattern is not universal, and that reminder matters, because a diagnostic test that fails every case it is applied to is not a diagnostic test. It is a verdict reached in advance.
IV. The Great Non-Solution: Complexity, Relationality, and Entanglement
A contemporary medical anthropologist reading the preceding sections has an obvious objection ready: nobody in the field still believes in rigid binaries. Biology and culture are entangled. Body and environment mutually constitute one another. Material and symbolic dimensions are inseparable. Local and global are co-produced. This vocabulary is now pervasive enough in the field that it can seem to have already accomplished what this article is calling for.
It has not, for a reason that follows directly from the three-part definition of a SWIPE given in the Introduction. Entanglement is a relation between terms, and a relation between terms presupposes that the terms have already been individuated enough, and reified enough, to require the further claim that they are entangled. Saying that biology and culture are entangled does not undo the operation that produced biology and culture as separable, reified candidates for entanglement in the first place. It restates the operation and then apologizes for it in the same breath.
The obvious rejoinder deserves a direct answer rather than a dismissal, because it identifies the real difficulty this article's own framework has to clear. Living Value Theory distinguishes five mediations, embodiment, being-with, dwelling, materiality, and symbolism. Why is this not simply a five-way version of the same operation, five domains instead of two, entangled instead of split? The answer is the criterion set out in the Introduction, and it needs to be stated as a formal test rather than a rhetorical gesture. A SWIPE requires all three of reification into independently constituted domains, the assignment of incompatible epistemologies to each domain, and treatment of the resulting partition as a starting point rather than a hypothesis. The five mediations fail the first criterion by design: they are explicitly analytic differentiations within one process, not separately existing substances that happen to interact, and no claim is made that embodiment could exist, even in principle, without being-with, dwelling, materiality, and symbolism simultaneously in play. They fail the second criterion as well: there is no proposal that embodiment requires a different epistemology than dwelling, no Erklären for the body and Verstehen for the symbol. And they fail the third, because the mediational grid specifies what to examine in a given case without predetermining which mediation will turn out to matter most, or whether the five-way distinction itself will prove to be the most useful cut for a given problem. Differentiation without reification, without epistemological incommensurability, and without a priori jurisdiction is not a SWIPE. This is not a technicality. It is the difference between analysis and amputation.
Complexity language performs a related but distinguishable substitution. Roberts's Mexico City ethnography of lead-glazed cookware shows people negotiating taste, memory, kinship, material practice, urban infrastructure, and household care at once, in relation to objects that are simultaneously dangerous and irreplaceably valuable (Roberts 2025). It would be natural to describe this as complex, meaning that many factors interact. Multimediation asserts something stronger than multifactorialism. Taste is not a variable of the same kind as blood lead level. Kinship is not a variable of the same kind as cookware composition. Memory is not a variable of the same kind as dose. These can be related, and the relation matters enormously for how people live, but they cannot be rendered commensurable without loss, because they belong to different mediations with different temporalities and different recursive structures. What Living Value Theory calls the enclosed exposure model of toxicology fails precisely by treating multimediated life as though it were merely multifactorial: it first separates body, object, chemical, environment, and meaning into isolable variables, and only afterward attempts to reconnect them through risk analysis (Ecks 2026b). It begins after the wrong cut has already been made. Saying that many factors interact does not, by itself, register that some of the interacting elements are not factors in the same ontological register at all.
The same substitution, formalized statistically, appears in the health-behavior literature on self-efficacy, collective efficacy, and social capital. Bandura's self-efficacy construct located the determinant of health behavior inside the individual, a cognitive property to be measured and boosted (Bandura 1977). Collective efficacy research and social capital theory arose to correct its individualism by relocating the causal variable to the group (Ecks 2026d). All three constructs share a common architecture: each locates the determinant of health behavior in a container, individual or collective, and treats that container as the proper unit of intervention. More sophisticated versions have since been developed in direct response to exactly this criticism, multilevel models that treat self-efficacy as a level-one variable and neighborhood deprivation as a level-two modifier. But methodological sophistication of this kind formalizes the individual/society binary as a statistical architecture rather than dissolving it, operationalizing it as levels in a hierarchical model. This is a more sophisticated form of the original error, and the analogy to medical anthropology's own theoretical history could hardly be closer: a discipline that keeps discovering a binary's inadequacy and responds by building elaborate scaffolding around it rather than asking whether the binary was the right place to start.
Holism fares no better, and it is worth addressing directly because it is sometimes proposed, in clinical settings especially, as the antidote to the reductionism this article is diagnosing. The biopsychosocial model, the most institutionally successful holistic framework in contemporary medicine, retains the biological, the psychological, and the social as separable domains and asks clinicians to consider all three (Engel 1977). This is an improvement over considering only the first, but the underlying logic remains additive, bio plus psycho plus social, three boxes checked rather than one, with the boxes themselves left conceptually untouched.
The Hardest Cases
Fairness to the field requires asking which approaches came closest to escaping the pattern altogether, because a critique that treats every predecessor as equally trapped is not credible, and because the near-misses show most clearly what a genuine escape would require. Mol's account of atherosclerosis as multiple, enacted differently by different clinical practices rather than viewed from different angles, comes unusually close (Mol 2002). Mol does not claim that pathology and interpretation are separate domains subsequently connected by practice; she claims that practices bring the disease into being differently, which dissolves the objective/subjective SWIPE more thoroughly than most alternatives on offer. What remains missing is an account of why some enactments cohere into a livable, actionable disease and others do not, a question about recursive stabilization that ontological multiplicity alone does not answer.
Scheper-Hughes and Lock's proposal of "the mindful body" as an object simultaneously individual, social, and political comes close for a different reason: it explicitly refuses to let the body be assigned to biology alone (Scheper-Hughes and Lock 1987). But the three bodies remain three bodies, analytically stacked rather than treated as aspects of one recursive process, which reproduces an additive structure not unlike the biopsychosocial model it otherwise improves on. Haraway's cyborg and companion-species work refuses the human/nonhuman and organism/machine SWIPEs with genuine force, but its primary object is the boundary itself rather than the recursive structure of coordination on either side of it, which makes it a powerful critique with less to say about how health specifically is sustained or lost (Haraway 1991). Latour and actor-network theory dissolve the human/nonhuman divide by symmetrical treatment of actants, but the resulting flatness has difficulty explaining why some disturbances remain unfelt for years while others are registered and remediated within seconds, a distinction this article's recursivity levels are built specifically to capture (Latour 2005).
What these cases share is instructive. Each dissolves one SWIPE with real success and leaves at least one other operation, reification without incommensurable epistemology, additive stacking, or the absence of a graded account of salience, intact. None supplies a general architecture that would let a researcher move to the next case and know in advance what to look for. That is the gap the remainder of this article tries to close.
The conclusion of this section should not be mistaken for a claim that medical anthropology's contemporary vocabulary is unsophisticated. It is, if anything, too sophisticated for its own good. Entanglement, complexity, multilevel modeling, and biopsychosocial integration can make a bad cut harder to see, because each new layer of qualification looks, from the inside, like progress toward exactly what this article is calling for.
V. Why Medical Anthropology Suffered Most: Health Is Multimediated All the Way Down
The preceding sections have argued that medical anthropology inherited an unnecessary cut and then built increasingly elaborate machinery to repair it without abandoning it. This section argues that health, more than most anthropological objects, exposes this problem starkly, and that the field possessed, throughout its history, ethnographic evidence sufficient to have seen the problem clearly at almost any point along the way.
The claim rests on five mediations Living Value Theory treats as irreducible and simultaneous rather than as separate domains that subsequently interact: multisensorial embodiment, being-with, multimaterial forming, multiversal dwelling, and multisymbolism (Ecks 2026e). These are analytically distinguishable but practically inseparable aspects of a single, ongoing process of living coordination, and the distinction from a SWIPE, spelled out formally in Section IV, is what allows them to be named without being reified.
An ordinary illness episode makes this vivid faster than an abstract argument can. Consider knee pain. There is tissue damage, a biomedical fact available to imaging and diagnosis. But there is also, without any temporal gap in which the biological fact exists before social consequences are added, a change in how stairs are navigated, whether a partner now helps with tasks that were previously independent, whether work can still be performed, whether analgesics are affordable, and how long a waiting list will make all of this last. There is no sequence here in which a purely biological disease precedes a set of separable social consequences. The episode unfolds multimediationally from its first moment, and any account that isolates the tissue damage as the real object and everything else as downstream consequence has already performed the cut this article has been describing.
Depression exposes the specific bias the disease/illness settlement built into the field: a bias toward treating suffering as anthropologically real only once it achieves narrative articulation. Sleep changes, appetite shifts, bodily heaviness, and withdrawal frequently develop before a person names a condition, and families often adapt their behavior without discussing why. A framework organized around the disease/illness distinction, or its later descendants, structurally privileges the moment of naming, because naming is where meaning becomes available for anthropological analysis. The pre-articulate disturbance is not an incomplete version of the real phenomenon waiting to become narratable. It is a large part of the phenomenon, and a theory that becomes competent to describe suffering only once it has been spoken has already missed a great deal of what suffering is.
Toxicity supplies the sharpest illustration in the health corpus, because it removes any temptation to treat multimediation as a special feature of mental health alone. In Roberts's ethnography, a lead-glazed cooking pot is simultaneously a chemical exposure, a taste, a source of warmth, a marker of kinship, a repository of memory, and a thread connecting national history to a single kitchen (Roberts 2025). Health here "was never properly enclosed to begin with." The pot cannot be understood by measuring lead leaching and then, as a separate and subsequent step, asking what it means to the family that uses it.
A comparative point sharpens the argument and forecloses an obvious objection. Are religion and politics not equally multimediated, such that health has no special claim on this argument? The comparison is instructive precisely because it reveals a difference of degree rather than of kind, and the difference should be stated carefully rather than as a categorical distinction between domains. Political and religious anthropology can, for considerable analytical stretches, proceed through symbolization and institutional classification without the costs of doing so becoming immediately visible, because a delayed or contested symbolic reading of a ritual or a policy does not, on its own, kill anyone by next week. Health does not offer this grace period, because nonsymbolic processes, infection, toxicity, metabolism, pain, and sleep, remain causally consequential regardless of whether or how they are articulated, and the consequences of misreading them arrive on a timescale that leaves little room for a slow theoretical correction. Dehydration happens before discourse. Toxic exposure does not wait for symbolic recognition before acting on a kidney. Microbes were shaping human metabolism and immunity for the entire span of human existence before anthropology treated them as constitutive rather than external. Health is not uniquely multimediated. It is unusually unforgiving of the delay between a SWIPE's empirical failure and its theoretical correction.
Medical anthropologists sat with sick people, in clinics, households, psychiatric wards, and kitchens, for the better part of five decades. Their ethnography continually showed bodies that exceeded biology, relations that exceeded "the social," things that exceeded material context, and symbols that acted back upon what they classified rather than merely representing it. Ethnography kept discovering multimediation. Theory kept restoring SWIPEs. The field possessed, in its own primary evidence, more than enough material to have abandoned the disease/illness settlement and its descendants decades before it eventually got around to complicating them, which means the failure was never a failure of data. It was a failure of the categories through which the data were read.
VI. Starting Before the Cut: Living Value Theory as Theoretical Compression
Living Value Theory is introduced only now, deliberately, so that it does not read as one more school announcing its arrival alongside embodiment, syndemics, and assemblage theory on the same terms. Its claim is not that it sees more than its predecessors. Its claim is that it starts somewhere different.
LVT does not claim that biology and culture are deeply intertwined; it denies that biology/culture is the appropriate starting point for an account of health. It does not propose a more sophisticated integration of disease and illness; it refuses to begin by classifying sickness into the two categories at all. The core proposition: life is recursive mediation in a mesocosm, the lived field in which reality becomes available to a living being through embodiment, being-with, material forming, dwelling, and symbolism at once. The living process comes first. The five mediations are analytical differentiations made within that process, not separate domains the process is built out of.
A second architecture does comparable work for how disturbance becomes visible, replacing the objective/subjective, pathology/meaning binary examined in Section II with something more precise. LVT distinguishes five levels of recursivity (Ecks 2026f; Ecks 2026h). L1 is absorbed coordination, the state Gadamer described as health's characteristic self-concealment, in which embodied and social life proceeds without becoming an explicit problem to itself (Gadamer 1996). L2 is felt misalignment, a pre-symbolic sense that something is off, ordinarily remediated spontaneously and without generating a symbolic articulation, a clinical encounter, or a data trace. L3 is articulation, the point at which a disturbance is named and made available to another person without yet being fully stabilized into a fixed category. L4 is institutional stabilization, the level of diagnoses and protocols, which enables communication and coordinated action but which also, characteristically, suppresses the texture of the L3 articulation it was built from. L5 is meta-repositioning, reflection on the entire system of levels, asking what a given L4 category makes visible, what it hides, and what it does to the living process it claims only to describe.
Health should not be equated with L1 coordination, and doing so would recreate, at one remove, the very static binary this article has spent five sections dismantling. Someone managing diabetes through deliberate L3 and L4 routines, tracking glucose, adjusting insulin, negotiating a diagnosis with a clinician, can be living remarkably well, and a framework that defined health as seamless, unarticulated coordination alone would have no way to say so. The more defensible formulation is this: health is the continuing capacity for sufficiently fluid recursive coordination across mediations, a capacity that can be exercised through absorbed L1 functioning, through spontaneous L2 remediation, or through deliberate, well-fitted movement into L3 articulation and L4 classification and back again. In human phenomenology, one of its more characteristic manifestations is that successful coordination tends to recede from attention, which is what gives Gadamer's observation its force. But recursive fluidity, not L1 self-concealment, is the operative concept, and self-concealment is one mode this fluidity can take rather than its definition.
Illness, correspondingly, frequently produces a recursive opening: a disturbance that forces formerly background processes into explicit attention, breathing that becomes noticeable, a familiar walk that becomes exhausting, a medicine that stops working. This should not be treated as a definition of illness as such, because it leaves out cases that matter, subclinical disease that generates no felt disturbance at all, slowly developing incapacity that a person adapts around without ever registering it as a problem, pathology discovered incidentally on a scan for something else. What recursive opening names precisely is what happens when a disturbance does cross into felt or articulated registers, which is a common but not universal path illness takes, and the qualification matters because a theory that quietly redefines illness as whatever produces a recursive opening would simply relocate the disease/illness SWIPE inside its own vocabulary.
The value of this architecture lies in compression rather than proliferation, and this is the standard against which it should be judged. A five-mediation, five-level framework should account, without inventing a dedicated subtheory for each, for illness experience, embodiment, care, structural inequality, pharmaceutical action, toxicity, multimorbidity, clinical recognition, diagnosis, materiality, and digital health. If it instead generates a recursive pharmaceuticals, a recursive care, and a recursive psychiatry, each requiring its own literature, it will have failed by its own standard, reproducing the proliferation this article has spent five sections diagnosing in medical anthropology's existing vocabulary.
VII. What Survives When the Theories Are Binned?
The preceding argument risks sounding more destructive than it is. The test proposed here is not whether a concept survives contact with its opposite theoretical school. It is whether a concept survives the removal of the SWIPE it was invented to repair, and a test that fails every candidate is not a test; it is a verdict dressed up as one, which is why Section III's treatment of syndemics matters as much as its treatment of local biology.
Run the experiment on the field's classic ethnography directly. Remove the proprietary theoretical framing from each and ask what disappears. Very little does. Lock's account of local biological variation in menopause survives; nothing about the finding depends on the "local biology" label rather than on the ethnographic evidence itself (Lock 1993; Lock and Kaufert 2001). Farmer's demonstration that inequality is patterned and that patterning is bodily survives (Farmer 2004). Kleinman's clinical observation that eliciting a patient's own account of illness improves care survives, even once the explanatory-model architecture built around it is set aside (Kleinman 1980). Kitanaka's account of the institutional conditions under which depression became a recognized public category in Japan survives (Kitanaka 2012). Mol's demonstration that atherosclerosis is enacted differently across clinical practices survives (Mol 2002). Roberts's account of lead pots in Mexico City survives with unusual force, because it was never organized around a repair concept to begin with (Roberts 2025).
A parallel distinction between empirical finding and surrounding theoretical infrastructure appears in Living Value Theory's critique of digital-phenotyping research in psychiatry, where behavioral correlations with later diagnoses are real while the infrastructure built to explain them, the assumption that enough nonrecursive behavioral traces will disclose a hidden psychiatric object, is not (Ecks 2026a). Generalizing this distinction produces a reusable test. For any concept in the field's inherited toolkit, ask what SWIPE it presupposes, what phenomenon it was invented to make visible, whether that phenomenon remains intelligible once the concept's proprietary vocabulary is removed, whether the concept does more than reconnect two terms generated by an earlier, avoidable cut, and whether it identifies a mechanism a multimediation account could not otherwise distinguish.
Applying this test across Section III's cases yields a consistent but not uniform pattern. Disease and illness: the phenomena survive almost entirely, redescribed as a disturbance traced across recursivity levels rather than sorted into two pre-given categories. Local biology: the empirical fact of developmental variation survives; the implied contrast with an unlocal biology does not. Structural violence: the patterned relationship between political economy and bodily deterioration survives completely; the getting-under-the-skin metaphor need not survive for the finding to remain intact. Embodiment: the ethnography of bodily practice survives entirely; its status as a theoretical breakthrough does not, once the prior exclusion it corrects is made explicit. Syndemics passes the test in its narrower form: it names a specific, real mechanism, mutually intensifying disease processes under patterned disadvantage, and that mechanism does not simply dissolve into the general claim that everything is multimediated, because the general claim does not by itself specify which conditions will worsen which others or why. Biosociality, the concept describing social formations organized around biological classifications, similarly survives as a useful description of a real and specific social phenomenon without needing to be elevated into a general claim about the nature of biology and sociality as such.
The conclusion here should not be mistaken for the claim that theory is useless, which would undermine the framework the article goes on to build in Sections VI and VIII. The conclusion is narrower: bad abstraction is expensive. It consumes graduate training and journal space that could be spent on the actual organization of living processes. Medical anthropology does not need fewer findings. It needs fewer middle-range classificatory entities standing between its findings and the living processes those findings describe, and it needs a test, applied honestly enough to let some candidates pass, for telling the difference.
VIII. How to Move On: A New Research Programme for Medical Anthropology
The preceding sections have been diagnostic. This one is meant to be usable, organized around a single instruction: stop organizing research around the domains the field inherited, and start organizing it around the question those domains were always a poor substitute for.
The place to begin is not biology against culture, individual against society, or disease against illness. It is a question with no smaller or more convenient version: what has to hold together here for living to continue? Answering it in practice means analyzing mediational configuration case by case rather than assuming in advance which mediation will matter most. What is happening in embodiment, in the rhythms and capacities sustaining or disrupting coordination? What relations are sustaining or disrupting it, and what do they obligate? What materials, tools, and infrastructures are enabling or constraining what is possible? What dwelling conditions, climate, terrain, built environment, is the coordination happening within? What symbolic forms are being produced, and how are they feeding back into the process they describe? And how are all of these changing over time, since a configuration that holds today may not hold next month.
Tracing recursive movement supplies the second half of the method. What remained unnoticed at L1? What first became salient at L2, and through what channel? How was it articulated at L3, to whom, in what words? How was it stabilized at L4, and did that stabilization improve coordination or overwrite something the L3 articulation had captured more accurately? Can an L5 repositioning reopen a classification doing more harm than the disturbance it was meant to manage? This gives medical anthropology a more exact intervention science than a familiar but blunt question, whether an intervention was culturally appropriate, by asking instead what mediations an intervention altered and what that alteration did to a person's recursive capacity to continue living.
Clinical anthropology benefits from this immediately. The clinician should not be imagined as a broker moving between two pre-existing explanatory frameworks. Clinical encounters are interrecursive: patients alter clinicians over the course of a relationship, diagnoses alter patients, and treatments alter illness trajectories in ways that feed back into how the treatment is subsequently understood. A clinical anthropology built on this premise studies the recursivity shift itself as its central object, rather than treating diagnosis as an endpoint to be evaluated for cultural sensitivity afterward.
Public health requires a comparable but more demanding shift. Adding social determinants to a biomedical model is not the move proposed here, because it leaves housing, food, and infrastructure in the position of external determinants acting on an otherwise complete organism, the same transactive model Sections III and IV both dismantle. The alternative studies mesocosmic conditions as partly constitutive of what health capacities are possible, not as a supplementary layer added to a biology that would be complete without them.
Mental-health anthropology requires the largest reorientation, because it is the subfield in which the symbolic bias traced in Section II has been most durable. It should stop treating articulated narrative as the privileged point at which suffering becomes anthropologically real and study pre-articulate disturbance directly: sleep, withdrawal, pacing, fatigue, and the changes in being-with and dwelling that often precede, by weeks or months, the moment a condition is named. This requires the kind of sustained ethnographic presence the field has always been better positioned to provide than most disciplines sharing its object.
Digital-health research supplies a concrete demonstration of what this reorientation looks like. Digital technologies can force health into continuous symbolic legibility, converting unnoticed functioning into compulsory self-monitoring and manufacturing felt misalignment where none arose from the living process itself, as when a wearable's Readiness Score announces fatigue before its wearer has formed any impression of their own state (Ecks 2026a). This is not primarily a story about commodification, real as that harm is; it is a story about L4 representation substituting for, rather than remaining answerable to, L1 and L2 coordination. A research programme organized around mediational configuration and recursive movement can specify exactly when a given architecture crosses this line and when it does not, a distinction blanket technological optimism and blanket suspicion are equally unable to draw.
Conclusion: Medical Anthropology after Medical-Anthropological Theory
Medical anthropology does not suffer from a lack of theory. It may suffer from having had too much of the wrong kind, applied with genuine sophistication to a problem sophistication alone could not solve, because the problem was never a shortage of nuance. It was a starting point. The field's theoretical history repeatedly begins after an unnecessary division has already been made, then spends its considerable ingenuity finding better ways to repair the division rather than asking whether it needed making. Disease and illness was the paradigmatic case, consolidating an already-circulating Enlightenment settlement into its most durable medical-anthropological form. Subsequent theories, embodiment, structural violence, local biology, materiality, multispecies relations, discovered the empirical inadequacy of whatever cut had preceded them, and mostly responded by introducing a new category of connection between the terms the cut had produced rather than asking why the terms had been produced as separable at all. Syndemics is the reminder that this was not universally true, and that reminder is part of what makes the diagnosis credible rather than merely polemical.
The irony running through this history is that ethnography was better than theory throughout. Ethnography kept encountering bodies that exceeded biology, relations that exceeded "the social," things that exceeded material context, symbols that acted back upon what they classified, and living processes that refused every jurisdictional boundary theory tried to draw around them. The evidence for this article's argument was never missing from medical anthropology's own archive. It has been sitting in the field's most celebrated ethnographies for decades, described with extraordinary care, and read back through categories that were never adequate to what the description showed.
The alternative proposed here is not another turn. It is not another hybrid concept promising to reconnect what an earlier concept separated, not a more elaborate biopsychosocial model with a fourth box added to the three it already checks, not another declaration that everything is relational or entangled, since entanglement talk, as Section IV argued, presupposes the very separation it claims to overcome. It is a change in where the account begins. Before disease and illness, there is living. Before biology and culture, there is multimediated coordination that has never actually decomposed into the two. Before individual and society, there are selfrecursive and interrecursive living beings coordinating within a mesocosm neither term adequately describes. Before objective pathology and subjective experience, there are disturbances that may or may not become felt, articulated, classified, and institutionalized, each step carrying consequences a two-term framework has no vocabulary for weighing.
Medical anthropology went wrong not because any particular generation of scholars reasoned poorly, but because it accepted, largely without noticing it had a choice, the partitions through which modern thought had already divided living processes, and then made the sensitive, ingenious, empirically grounded repair of those partitions its central theoretical vocation. Moving on does not require repairing the divisions more sensitively still. It requires beginning before them, with the living process itself, and letting whatever divisions turn out to matter emerge from what that process actually does, rather than installing them in advance and asking ethnography to reconnect what they should never have been allowed to separate.
Medical anthropology does not need another theory of how the pieces fit together. It needs to stop cutting living into pieces before it begins.
References
Bandura, A., 1977. Self-efficacy: toward a unifying theory of behavioral change. Psychological Review, 84(2), pp.191–215.
Blaser, M.J., 2021. Missing Microbes: How Killing Bacteria Creates Modern Plagues. Rev. ed. New York: Picador.
Chaparro-Buitrago, J., 2022. Debilitated lifeworlds: women's narratives of forced sterilization as delinking from reproductive rights. Medical Anthropology Quarterly, 36(3), pp.295–311.
Csordas, T.J., 1990. Embodiment as a paradigm for anthropology. Ethos, 18(1), pp.5–47.
Decker, H.S., 2013. The Making of DSM-III: A Diagnostic Manual's Conquest of American Psychiatry. Oxford: Oxford University Press.
Dressler, W.W. and Bindon, J.R., 2000. The health consequences of cultural consonance: cultural dimensions of lifestyle, social support, and arterial blood pressure in an African American community. American Anthropologist, 102(2), pp.244–260.
Ecks, S., 2014. Eating Drugs: Psychopharmaceutical Pluralism in India. New York: New York University Press.
Ecks, S., 2021. Depression, deprivation, and dysbiosis: polyiatrogenesis in multiple chronic illnesses. Culture, Medicine, and Psychiatry, 45(4), pp.507–524.
Ecks, S., 2022. Living Worth: Value and Values in Global Pharmaceutical Markets. Durham: Duke University Press.
Ecks, S., 2026a. Algorithmic affects: how digital health tech manufactures unease. Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026b. Beyond enclosure: transactive dualism, multimediation, and health in Mexico City. Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026c. The impossible object: beyond social theory. Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026d. Case studies in unearned confidence, part 2: the disastrous applications of Bandura's "self-efficacy" in health research. Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026e. Health across five dimensions. Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026f. Health value is what makes life better (even if you cannot measure it). Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026g. The recursive and the nonrecursive. Living Value Theory, livingvaluetheory.org.
Ecks, S., 2026h. Two thresholds: recursive descent and the biomedical reorganization of the mesocosm. Living Value Theory, livingvaluetheory.org.
Engel, G.L., 1977. The need for a new medical model: a challenge for biomedicine. Science, 196(4286), pp.129–136.
Fabrega, H., 1972. Concepts of disease: logical features and social implications. Perspectives in Biology and Medicine, 15(4), pp.583–616.
Fabrega, H., 1974. Disease and Social Behavior: An Interdisciplinary Perspective. Cambridge, MA: MIT Press.
Farmer, P., 2004. An anthropology of structural violence. Current Anthropology, 45(3), pp.305–325.
Gadamer, H.-G., 1996. The Enigma of Health: The Art of Healing in a Scientific Age. Stanford: Stanford University Press.
Hardon, A. and Sanabria, E., 2017. Fluid drugs: revisiting the anthropology of pharmaceuticals. Annual Review of Anthropology, 46, pp.117–132.
Haraway, D.J., 1991. Simians, Cyborgs, and Women: The Reinvention of Nature. New York: Routledge.
Jenkins, J.H. and Cofresi, N., 1998. The sociosomatic course of depression and trauma: a cultural analysis of suffering and resilience in the life of a Puerto Rican woman. Psychosomatic Medicine, 60(4), pp.439–447.
Kirksey, E. and Helmreich, S., 2010. The emergence of multispecies ethnography. Cultural Anthropology, 25(4), pp.545–576.
Kitanaka, J., 2012. Depression in Japan: Psychiatric Cures for a Society in Distress. Princeton: Princeton University Press.
Kleinman, A., Eisenberg, L. and Good, B., 1978. Culture, illness, and care: clinical lessons from anthropologic and cross-cultural research. Annals of Internal Medicine, 88(2), pp.251–258.
Kleinman, A., 1980. Patients and Healers in the Context of Culture. Berkeley: University of California Press.
Latour, B., 2005. Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford: Oxford University Press.
Lindenbaum, S., 1979. Kuru Sorcery: Disease and Danger in the New Guinea Highlands. Palo Alto: Mayfield.
Lock, M., 1993. Encounters with Aging: Mythologies of Menopause in Japan and North America. Berkeley: University of California Press.
Lock, M. and Kaufert, P., 2001. Menopause, local biologies, and cultures of aging. American Journal of Human Biology, 13(4), pp.494–504.
Mol, A., 2002. The Body Multiple: Ontology in Medical Practice. Durham: Duke University Press.
Niewöhner, J. and Lock, M., 2018. Situating local biologies: anthropological perspectives on environment/human entanglements. BioSocieties, 13(4), pp.681–697.
Paxson, H. and Helmreich, S., 2014. The perils and promises of microbial abundance: novel natures and model ecosystems, from artisanal cheese to alien seas. Social Studies of Science, 44(2), pp.165–193.
Pfeiffer, J. and Chapman, R., 2010. Anthropological perspectives on structural adjustment and public health. Annual Review of Anthropology, 39, pp.149–165.
Roberts, E.F., 2025. Grappling with exposure in Mexico City. Current Anthropology, 66(6), pp.891–915.
Scheper-Hughes, N. and Lock, M.M., 1987. The mindful body: a prolegomenon to future work in medical anthropology. Medical Anthropology Quarterly, 1(1), pp.6–41.
Singer, M., 2009. Introduction to Syndemics: A Critical Systems Approach to Public and Community Health. San Francisco: Jossey-Bass.
Singer, M., Bulled, N., Ostrach, B. and Mendenhall, E., 2017. Syndemics and the biosocial conception of health. The Lancet, 389(10072), pp.941–950.
Twaddle, A.C., 1968. Influence and Illness: Definitions and Definers of Illness Behavior among Older Males in Providence, Rhode Island. PhD dissertation. Providence, RI: Brown University.
Weinberg, D., 2021. Diagnosis as topic and as resource: reflections on the epistemology and ontology of disease in medical sociology. Symbolic Interaction, 44(2), pp.367–391.
Whyte, S.R., van der Geest, S. and Hardon, A., 2002. Social Lives of Medicines. Cambridge: Cambridge University Press.