Culture and Comparison

Malinowski’s “native’s point of view” has been systematically misunderstood for over a century.

How Anthropology Made Cultures Comparable, and at What Cost

This essay argues that anthropology’s comparative method carries a hidden bias not in ranking cultures but in selecting the dimensions on which they are compared. Tracing the word culture from Cato’s cultivation of a field through Cicero, Herder, Tylor, and Boas, it shows how an activity became a bounded possession, enabling both systematic comparison and colonial hierarchy. Relativism blocked ranking but left selection unexamined: why kinship, law, religion, and exchange rather than whatever matters most to the people studied? Rereading Malinowski’s famous section on the native’s point of view as a methodological instruction rather than a representational ideal, the essay proposes an alternative: enter fieldwork without a fixed comparative grid, attend to what people themselves treat as consequential, and build comparison afterward. Examples from Martin’s egg and sperm and fieldwork in Kolkata show how inherited categories distort inquiry. Culture should be recovered as cultivation, ongoing, relational activity, and comparison grounded in shared human stakes rather than inherited categories.

I. ‘Let Us Compare These Two Cultures’

“Anthropologists compare different cultures to each other.” The instruction sounds so straightforward that no one thinks to question it, and that is exactly the problem. What is being compared? Where do the boundaries of a culture come from? Why should culture be the kind of thing that comes in bounded units at all, English culture, Scottish culture, student culture, Bengali culture, such that one unit can be set beside another and measured against it? And, hardest of all, in what respect are the two units supposed to be compared, since nothing about the raw fact of two populations existing tells anyone which of their differences matter and which of their similarities can safely be ignored?

This is not a pedantic quibble. It is close to the single most important methodological question a comparative discipline can ask, because a comparison is never simply given by its objects. Consider Edinburgh and Kolkata, two cities separated by several thousand miles and, on paper, almost everything else. Kinship arrangements differ. Architecture differs. Legal systems differ. The way people address each other, argue with each other and mark status differs. A cross-cultural comparison built on these differences writes itself. But an enormous amount is also shared. People in both cities sleep. They eat. They grow older. They quarrel with their families. They raise children, worry about money, and fall ill. None of this shared material typically enters a comparison at all, and the reason is not that it is irrelevant. The reason is that it is not strange enough to be noticed. The moment something is universally shared, it tends to drop out of anthropological attention, not because it stops mattering but because there is, on the face of it, nothing to explain: everybody sleeps, so why would an anthropologist bother with sleep?

This already smuggles in a serious bias before any fieldwork has begun. If a discipline only compares what strikes it as strikingly different, it has quietly decided, in advance of any inquiry, that difference is where the interesting material lives and that similarity is background noise. The choice of what counts as comparable is made before the comparison starts, and it is worth asking where that choice actually came from. Anthropology did not stumble upon a natural world already divided into discrete, comparable cultures and then work out how to line them up against each other. The very idea of a culture, as a free-standing, bounded thing capable of being compared to another such thing, has its own history, and that history is considerably shorter and stranger than the confidence with which the word gets used today.

The rest of this essay follows that history from its beginning, because the deep history of the word culture explains a great deal about what is wrong, and what can be salvaged, in the ordinary anthropological instinct to compare.

It is worth being clear from the outset that the target here is not comparison itself. Comparison is one of the oldest and most productive tools any discipline has, and nothing that follows argues for abandoning it. The target is a specific, usually invisible step that happens before any actual comparison begins: the selection of a respect, a dimension along which two or more cases are to be set beside each other. Two societies, or two people, or two texts, resemble and differ from each other in an unlimited number of ways, and a bare instruction to compare them says nothing at all about which of those ways should organise the inquiry. Whenever that selection is made silently, by habit or by inherited convention rather than by an argument that could in principle be examined and rejected, the comparison that follows inherits a bias it can never see from the inside, because the very terms in which the two cases are being read off against each other were fixed before the reading began.

II. Cato’s Field

The first sustained use of the Latin word cultura comes from Roman agricultural writing, and Cato the Elder is usually credited with the earliest full use of the term. The setting is worth dwelling on, because it is not what a modern reader expects. Cultura, for Cato, is agricultura: doing something with a field. In one striking passage he describes an annual rite of purification performed on the farm. A pig, a lamb and a calf are led around the boundary of the land. Incantations are spoken. Sacrifices are offered to the gods. The farmer prays for a good harvest, for the health of his household, for wellbeing all around. This is what cultura names: not a possession but a practice, tending, caring for, looking after a piece of land through a combination of labour and address to divine powers who might help or harm it.

Two things about this scene deserve attention. The first is that cultura is not a bounded, comparable entity in the way the modern noun culture will later become. It is a verb rendered as a noun: the cultivating of a field is cultura of that field, always attached to its object, never a free-standing thing that a Roman farmer simply possessed. The second is that this earliest use of the word makes no distinction whatsoever between nature and culture in the sense that later becomes second nature to anthropology. Agriculture is precisely nature being cultivated, worked, tended. There is no sense here that the field belongs to one domain and the farmer’s practice to another; the rite, the harvest, the divine favour and the tending of the soil belong to a single, continuous activity.

The word travels from the field to the person not long afterward. Cicero writes that cultura autem animi philosophia est, the cultivation of the soul is philosophy, extending the agricultural logic to the mind without breaking the grammatical pattern that had governed it from the start: a person cultivates their own mind exactly as a farmer cultivates a field, through sustained attention, labour and care. Anyone working through a difficult argument, or trying to change how they think about a question, is doing a recognisable version of the same thing Cato’s farmer was doing with his field: tending something in the hope of a better outcome than it would produce untended.

This is, on the argument developed here, the good use of the word culture: culture as an ongoing activity, as something one does rather than something one has. What happens over the following centuries is a slow and largely unremarked transformation of an activity into a possession, and it is worth being precise about how that transformation occurred, because the consequences for anthropological comparison are considerable.

III. From Cultivation to Possession

Somewhere across the eighteenth century, the word culture loses its object. Where cultura had always required a complement, of a field, of a mind, of a soul, culture becomes able to stand entirely on its own. A person or a society no longer cultivates something in particular; they simply have culture, in the way one has property, and once the word can be used this way it starts to behave like a scalar quantity, something a person or a nation can possess in a greater or a lesser degree.

The consequence follows almost automatically. Once culture is a quantity rather than an activity, ranking becomes available in a way it had not been before. Eighteenth and nineteenth century writers increasingly speak of cultured peoples and their opposite, and early anthropology inherits this vocabulary wholesale, dividing humanity into the cultured and the primitive as though this were simply an observation about how much of a substance called culture different populations happened to possess. This is the point at which the deep grammar of the word starts to do real political work. A concept that once named a relationship between a person and something they tended now names a possession that can be measured, compared, and used to justify a hierarchy among human populations, with the observer’s own society reliably placed at the top of the scale it has just invented.

It is worth pausing on how much has changed in this shift, because it is easy to read past it. Cultura could never have organised a ranking of peoples, because it never named the kind of free-standing thing a ranking needs. One cannot meaningfully ask whether one farmer has more cultura than another; one can only ask whether a field is being tended well or badly, according to some further standard of what a well-tended field looks like. Once culture becomes a possession, ranking is no longer merely available: it becomes close to unavoidable, since a scalar quantity practically demands the question of who has more and who has less. The invention of the free-standing noun and the emergence of a colonial hierarchy of peoples are not two separate historical events that happened to occur around the same time. The second follows from the grammar of the first.

It is against this background, and specifically against this newly invented ranking, that the next figure in this history makes his intervention, and it is worth taking real care over what he actually did and did not achieve, because the received version of the story gets it importantly wrong.

IV. Herder’s Strange Move

Johann Gottfried Herder, writing in the second half of the eighteenth century, is arguably the single most consequential figure for how social anthropology still defines culture today, more so, in some ways, than any twentieth-century theorist. What Herder does is pluralise the noun. Instead of one culture, measured out along a single scale from primitive to civilised, there are cultures, plural, each with its own internal logic, its own coherence, its own reasons for being exactly as it is. Crucially, Herder insists there is no meaningful, non-arbitrary way to take one people’s culture and declare it superior to another’s. Edinburgh culture cannot be measured against Bengali culture on some shared yardstick, because there is no neutral yardstick standing outside both of them.

This move belongs to the wider current of European romanticism, which prized the deep, irreducible individuality of every living thing and extended that same principle from persons to whole peoples. Every nation, on this view, has something unique and untranslatable about it, in the way every individual does, and this idea fed directly and rapidly into the political movements of the following century. If a people is genuinely, deeply unique, the argument runs, then it deserves its own political form, its own boundaries, its own state. The romantic theory of cultural uniqueness and the rise of nineteenth-century nationalism are not separate stories either.

Herder’s second legacy is relativism, and it is worth stating the position in its strongest form before asking what is wrong with it. Relativism says that a practice should be judged by the standards internal to the people who hold it, not by the standards of an outside observer who happens to find it strange or objectionable. The position is put under real strain by cases where a practice one community endorses looks, to an outside observer, like a serious harm. Female genital cutting is the example that recurs in these discussions precisely because it forces the question in its sharpest form: if a community holds that the practice is necessary, valued and meaningful within their own moral world, does an outside observer have any standing to object? The strong version of relativism answers that the observer does not: that is their world, this is ours, and passing judgement across that boundary is itself a form of arrogance. Whatever the appeal of that answer in principle, most people find it very hard to hold in practice once a specific, serious harm is actually in front of them, and this is not a failure of nerve so much as a sign that relativism, taken to its logical conclusion, asks something few people are willing to give.

The interest of this essay is not to relitigate that debate, which anthropology has argued over for a century without resolving it, but to notice something relativism leaves completely untouched. Herder’s contribution is to block one kind of question, whose culture is better, while leaving a different and, on the argument here, more important question entirely unexamined: out of everything two populations could conceivably be compared on, why this respect and not another. That second question survives every version of the relativist settlement, and it becomes visible only once the discipline had built the machinery for comparison in earnest, which is where Tylor, and soon after him an entire generation of colonial fieldworkers, come in.

V. Tylor’s Complex Whole

Edward Burnett Tylor’s 1871 definition of culture is usually credited as the founding definition for the discipline, and the credit is deserved, though not always for the reason usually given. Tylor writes that culture, taken in its widest sense, is that complex whole which includes knowledge, belief, art, morals, law, custom, and any other capabilities and habits acquired by a person as a member of society. The phrase that deserves the most attention is complex whole. Tylor is not merely describing culture; he is doing something with the definition that has enormous downstream consequences. He is turning culture into a bounded container with clearly identifiable internal compartments: here is knowledge, here is law, here is custom, each a sub-region within the larger, bounded whole.

Once culture has this shape, a further method becomes available that was not available before. One can compare two cultures as wholes, or, more usefully, one can decompose each whole into its named compartments and compare compartment against compartment: the law of one society against the law of another, the kinship system of one against the kinship system of another. This decomposition is precisely what makes systematic, institutionalised comparison possible, and it is formalised, twenty years after Tylor, in a manual called Notes and Queries on Anthropology, issued by what would become the Royal Anthropological Institute in 1892 and circulated to missionaries, colonial administrators and travellers headed to unfamiliar parts of the world.

Before this manual existed, an observer’s report on an unfamiliar society was whatever struck that particular observer as interesting or strange, an entirely unsystematic and person-dependent record. Notes and Queries changes this by telling the traveller in advance what to look for: always record kinship, always record political organisation, always record systems of exchange, and so on through a long list of headings. The stated purpose, and a genuine achievement, was to remove the idiosyncrasy of the individual observer and put ethnographic reporting on a more systematic, comparable footing.

The achievement, however, came with a cost that went almost entirely unexamined at the time. The list of headings was decided in advance, by the discipline, in London, before any particular fieldworker ever set foot among any particular people. Whatever the traveller was told to look for was, by construction, being imposed on every society visited, whether or not those categories carried anything like the same weight for the people actually living there. The bias that Notes and Queries proudly claimed to have removed, the personal idiosyncrasy of the individual observer, was real enough. But a second and much less visible bias took its place: the assumption, built into the manual itself, that kinship, law, religion and exchange were the universally relevant categories worth asking about everywhere, an assumption nobody had actually tested and that the manual’s own existence made it very easy never to notice needed testing.

It is worth noticing how durable this arrangement turned out to be. Long after Notes and Queries itself stopped being handed to travellers, its underlying logic survived in the standard chapter headings of introductory textbooks, in the standard sections of an ethnographic monograph, and in the standard fields of a university curriculum: kinship gets a chapter, religion gets a chapter, economy gets a chapter, politics gets a chapter. None of this is illegitimate in itself, and a great deal of excellent scholarship has been produced within exactly this division of labour. The point is only that the division was inherited rather than derived anew from what any particular people studied turned out to care about, and an inherited division has a way of feeling like a discovery about the world rather than a decision made, for entirely understandable administrative reasons, by a committee working from London in 1892.

VI. Boas and the Line He Did Not Cross

Franz Boas belongs to the next generation of this story, and his importance to the discipline is difficult to overstate. Boas emigrated from Germany to the United States late in the nineteenth century and effectively founded the American tradition of cultural anthropology from his position at Columbia University, where he trained nearly everyone who would go on to shape the discipline for the following generation, Ruth Benedict among the best known of his students. Boas took Herder’s romantic insight, that cultures are plural and that no one culture is straightforwardly superior to another, and turned it into the working doctrine of an entire national tradition of anthropology.

Boas’s own words state his position without hedging: civilisation is not something absolute, but is relative, and our ideas and conceptions are true only so far as our civilisation goes. This was, at the time Boas was writing, a direct assault on the dominant anthropological orthodoxy, which held to a confident evolutionary ranking of human populations with white Europeans placed, unsurprisingly, at the summit and everyone else arranged beneath them according to how closely their institutions were judged to resemble European ones. Boas’s insistence that this ranking was ethnocentric, unscientific and morally indefensible is among the genuine, lasting achievements of the discipline, and it has not been seriously questioned since.

But notice exactly what this achievement does and does not accomplish. It dismantles ranking: the claim that one people’s institutions can be placed above another’s on some universal scale of civilisational advancement. It does nothing at all to touch a different and prior question, namely why the discipline should be comparing kinship, religion, law and exchange across societies in the first place, rather than some other set of concerns entirely. Relativism answers the question by whose standards should a practice be judged. It has no answer at all to the question why compare on this particular dimension rather than any of the countless others available. These are genuinely different questions, and it is entirely possible to be a scrupulous, committed relativist, refusing to rank any society above another, while still inheriting, without ever examining it, an entire grid of supposedly universal comparative categories that nobody in the society being studied necessarily thinks of as the natural joints of their own social life.

This is the gap that survives Boas’s revolution fully intact, and it is worth naming clearly because the history just told makes it easy to see why it survived. Ranking was the visible, politically explosive problem, the one that provoked outrage and demanded a response. Selection, the prior question of which dimensions of a life are worth comparing at all, produced no such outrage, because it does not look like a moral failing in the way ranking does. It looks like ordinary method. And ordinary method, unlike ranking, felt as though it did not need defending, so nobody, for the better part of a century, thought seriously about defending it, or attacking it either.

Boas’s institutional legacy compounded this. Because he trained an entire generation of American anthropologists directly, and because his students went on to train the generation after them, the doctrine of cultural relativism became something closer to a professional reflex than an argued position: the kind of thing a student absorbed early in training and rarely needed to defend again, because everyone in the room already agreed with it. That kind of unanimity is a genuine achievement when the position in question, the rejection of racial and civilisational ranking, deserves to be unanimous. But unanimity of this kind has a cost of its own. A position nobody in the room disputes tends to stop being examined, and the question relativism never answered, why compare on these dimensions rather than others, had the bad luck of resembling the question relativism did answer closely enough that generations of students absorbed the second as though it had settled the first.

VII. When Similarity Catches the Attention: Martin’s Egg and Sperm

A discipline trained to notice difference develops a particular attentional profile. What catches an anthropologist’s eye is reliably the strange, the symbolically loaded, the politically contested, the story that can be told and retold. Difference of this kind is, in a sense, the discipline’s native currency. Similarity, by contrast, tends to sink into an unremarked background: if something is shared by everyone everywhere, there seems to be nothing left for an anthropologist to say about it.

There is, however, a well-known case that inverts this pattern entirely, and it repays close attention precisely because the inversion is what makes the argument work. Emily Martin’s essay on the egg and the sperm looks at how standard medical and biology textbooks describe human conception, and it draws its evidence not from fieldwork or interviews but from published scientific writing, a reminder that ethnographic method is broader than long-term participant observation and can be turned on texts as readily as on villages. What Martin notices is a strikingly consistent gendered narrative running through this supposedly neutral scientific literature: the egg is described as large, passive and essentially inert, waiting to be acted upon, while the sperm is described as small, active, competitive, aggressively pursuing its target. The biology does not require this narrative in the least; nothing about the relative size or motility of the two cells demands that one be cast as a hero and the other as a prize. Martin’s real discovery is not a difference between two cultures but a similarity: a resemblance between the supposedly objective language of science and the entirely ordinary, unremarkable sexism of popular language about men and women. The scientific text turns out to be far more soaked in ordinary cultural assumption than its own claim to neutrality would suggest.

There is a further step available here that goes beyond what Martin’s own essay draws out explicitly, and it concerns exactly where in the process this gendering enters. Turning a biological process into a scientific description requires, at some point, translating it into a system of symbols: written language, certainly, but also photographs, diagrams, mathematical notation, chemical formulae, statistical tables. Modern science relies overwhelmingly on these last few, on mathematics, chemistry and quantitative graphing, and it does so for a specific reason: none of these notations carry an inherited gender bias the way ordinary language does. Nobody examining an equation, a graph or a chemical formula can locate within it a preference for one sex over the other; the notation simply will not support the kind of narrative Martin found running through the surrounding prose. Even the diagrams and micrographs in the textbooks Martin studied are, for the most part, free of the bias she is diagnosing. The bias enters specifically where the account has to leave notation behind and be retold in ordinary language, because ordinary language cannot be scrubbed clean of the history of the society that produced it. Ordinary words carry with them, whether a given speaker wants them to or not, the sediment of every inequality, gendered, political and otherwise, that has shaped the society in which those words have circulated.

A small, almost domestic example makes the same point without any of the charge that surrounds conception and gender. Nearly every human society treats the right hand as good, trustworthy and correct, and the left hand as suspect, unlucky, or worse, a moral asymmetry with no basis whatsoever in any real difference between the two hands. Nobody actually believes, on reflection, that a right hand is morally superior to a left one. The asymmetry is not a fact about hands. It is a fact about what happens the moment a society starts talking about hands, since ordinary language, once it exists, cannot help encoding the inherited values of the people who speak it. Martin’s textbook authors are not unusually sexist scientists smuggling their private prejudices into their research. They are doing something close to unavoidable: writing, in ordinary language, about a process that ordinary language has never had a neutral way of describing.

VIII. Two Ways of Doing Anthropology

All of this history points toward a genuine fork in how the discipline can proceed, and it is worth stating the fork plainly, because most anthropological training presents only one branch of it as though it were the whole of the discipline. One way of doing anthropology walks into a new field site already equipped with the categories the discipline has inherited, kinship, law, religion, exchange, political organisation, and sets out to describe and compare each of them systematically across societies. This is the Tylor and Notes and Queries inheritance, and it remains, in various updated forms, the default method taught to most students.

The other way of doing anthropology walks into the same field site with a much simpler and much harder question: what matters most to the people here? Not what the discipline has decided in advance is worth asking about, but what the people being studied themselves treat as consequential, worth arguing over, worth organising a life around. On this second approach, the anthropologist does not arrive with the comparative grid already fixed. The grid, if there is to be one, is built afterward, out of what turns out to matter to the people actually studied, rather than imposed on them beforehand out of what happens to matter to the discipline, or to the observer’s own society.

This is not a minor variation on the first method. It relocates the entire authority over what counts as relevant, from the anthropologist and the discipline behind her to the people being studied. And it is, on the argument of this essay, both the harder method to practise and the one that gives anthropology whatever genuine claim it has to being a distinctive way of knowing other people, rather than merely a more polite version of a colonial questionnaire.

IX. Malinowski’s Actual Point

The clearest statement of this second method was made, and largely misunderstood, at the very founding of British social anthropology. Bronislaw Malinowski’s Argonauts of the Western Pacific, still among the most cited books the discipline has ever produced, contains what is probably its single most quoted sentence: that the ethnographer’s goal is to grasp the native’s point of view, his relation to life, to realise his vision of his world (Malinowski 1922, Introduction). For most of the twentieth century this line was read as a statement about the problem of representation: how can an outsider write about another people fairly, without condescension, without imposing categories that distort what they are trying to describe. That is a real and serious problem, and anthropology has spent a great deal of energy on it.

But it is not, on a careful reading, what Malinowski’s sentence is actually about. The line continues, in terms that are quoted far less often than the famous half about grasping the native’s point of view: that the anthropologist must study man, and must study what concerns him most intimately, that is, the hold which life has on him (Malinowski 1922, Introduction). This is not a claim about how to represent someone accurately. It is a claim about where the anthropologist’s attention should be directed in the first place, and it says: not toward the categories the discipline already possesses, but toward whatever it is that most grips the life of the person being studied. Malinowski’s programme, read this way, is not don’t misrepresent people. It is: find out what is actually at stake for them, and build the inquiry around that, rather than around whatever the observer arrived already expecting to find.

This is a substantial claim about the discipline’s own history: that its founding text already contained, in more or less complete form, the alternative to the categorise-and-compare method that came to dominate the field’s actual practice for most of the following century. Malinowski said it first, and in some ways said it best, and generations of anthropologists trained in his own tradition proceeded largely as though he had said something much narrower, a point about fair representation, rather than the much more radical methodological instruction it actually is: let the people you are studying tell you what is at stake, and do not walk in having already decided.

The narrower reading is not hard to understand, and it is worth asking why it took hold so completely. Representation is a problem a discipline can teach directly: it can be turned into a set of writing conventions, a set of cautions about voice and authority, a checklist for what not to say and how not to say it, and generations of students have been trained accordingly, usually to real and lasting benefit. Finding out what is actually at stake for another person is not the kind of thing that can be taught as a checklist at all, since it cannot be specified in advance without contradicting itself; a method that tells a student in advance what she will find important has already failed at the one thing it claims to teach. A discipline under professional pressure to produce a teachable curriculum will, understandably enough, gravitate toward the version of Malinowski’s sentence that can be taught, and let the harder, less teachable version quietly recede into the background of the very quotation everyone kept citing.

X. Shared Stakes and a Stomach in Kolkata

If the comparative grid is not to be supplied in advance by the discipline, where does the ground for comparison come from at all? The answer offered here is that it comes from what is genuinely shared, prior to any particular society’s own elaboration of it. Every human population needs to sleep safely. Every human population needs to eat. Every human population raises children, and every human population must somehow manage its own vulnerability, its own risk of illness and harm. These shared stakes are what make it possible to understand another human being at all, across whatever surface differences separate two populations, and they are a far more solid ground for comparison than any inherited list of institutional categories, because they answer to something every human life actually requires rather than to a set of headings a nineteenth-century committee happened to settle on.

A personal example may make the method concrete. Fieldwork for a first book, Eating Drugs, was originally designed around a specific question: how do people in Kolkata understand and talk about environmental and air pollution (Ecks 2014). That was the proposal, and the first weeks of interviews in local clinics followed it faithfully, asking people directly how air pollution affected their health. The answers came back readily enough: air pollution blackens the lungs, everybody knows this. But they kept sliding, unprompted and repeatedly, toward a different organ entirely. It’s my stomach, people said. The soot and dust go into my stomach. That is what upsets my digestion. After roughly two weeks of hearing this same redirection from person after person, the honest response was not to keep steering the conversation back toward the original research question, but to notice that nobody being interviewed seemed to think pollution, as such, was the interesting or urgent topic, while almost everybody had a great deal to say, unprompted, about digestion and the stomach. The project changed accordingly, and the resulting book was about digestion and pharmaceutical use rather than about the environmental attitudes it had originally set out to document.

This is a small but exact illustration of the method Malinowski’s line actually recommends. The original research question was not wrong to ask; it simply was not what mattered most intimately to the people being asked, and the discipline’s own instinct, letting oneself be surprised and redirected by what turns out to be at stake for the people studied rather than defending a proposal written in advance, is close to unique among the social sciences. A structured quantitative survey, whatever its other virtues, does not permit this kind of redirection: a respondent presented with a fixed questionnaire about pollution has no space within the instrument itself to say that pollution is not actually what concerns them, and that something else, their stomach, their family, their debts, is where their attention and anxiety actually live. Long-term, open-ended fieldwork is one of the few methods in the social sciences built to let exactly that kind of redirection happen, and to treat it as data rather than as noise.

None of this was comfortable at the time it was happening. A funded research proposal describes a project to a funding body, and a funding body reasonably expects the project it paid for. Abandoning a proposal on the strength of what informants keep saying unprompted, rather than what they were asked, requires a kind of institutional and personal confidence that fieldwork training does not always supply, and it is worth being honest that the discipline’s own incentives sometimes pull in the opposite direction, toward finishing the project that was proposed rather than the project the people being studied kept insisting on. The redirection described here worked out, and produced a more interesting and more truthful book than the original proposal would have. It could equally have produced two years of fieldwork with no clear argument at all, and the difference between the two outcomes is not something a method can guarantee in advance. What the method can guarantee is only that the question actually being answered will be one the people studied recognise as their own, rather than one imported for their convenience by someone who had already decided what to ask before arriving.

XI. When Culture Really Is the Stake

None of this amounts to a claim that culture, or cultural identity, never matters to the people anthropology studies. Sometimes it plainly does, and sometimes it matters enormously. Nationalism is an obvious case: the conviction that one’s own people are fundamentally distinct from a neighbouring people, and that this distinctness justifies a particular political arrangement, is precisely a claim that culture is at stake, made with total sincerity by the people making it. Religious and communal identity supplies other, closer-to-home examples: in parts of the United Kingdom today, whether a family comes from a Protestant or a Catholic background can still carry real social weight, in exactly the way a cultural or communal difference is supposed to.

The point being made here is not that such differences are never real or never consequential. It is a point about method, and about the order in which things should be established. If cultural identity turns out to be a vital stake for the people an anthropologist is studying, that is a discovery to be made in the field, through attending to what the people themselves treat as mattering, not a premise the anthropologist is entitled to import before any fieldwork has begun. The error this essay has been tracing throughout is not the claim that culture sometimes matters. It is the assumption, built into the discipline from Tylor’s complex whole through Notes and Queries and beyond, that culture and its standard sub-categories must always matter, everywhere, simply because the discipline has a longstanding habit of assuming so. Let the answer be supplied by the people being studied, and treat it as a live question rather than a foregone conclusion.

XII. Abstraction Cannot Be Avoided, Only Done Honestly

None of this is an argument against abstraction as such, and it would be a mistake to read it that way. Even Malinowski, having spent years accumulating extraordinarily rich, situated detail about Trobriand life, still needed, at some point, to say that a given transaction was an instance of ceremonial exchange as opposed to ordinary economic exchange. Some abstraction of this kind is unavoidable the moment an ethnographer tries to say anything general at all, and there is nothing wrong with abstraction in itself. The point is not to avoid abstracting from the rich particularity of a life. It is to notice how large an abstraction is being made, and to make it honestly rather than by default.

Culture itself, on this reckoning, is one of the largest and least examined abstractions the discipline possesses. It compresses an enormous, heterogeneous range of activity, tending a field, cultivating a mind, raising a child, worshipping a god, organising a household, into a single bounded noun, and it does so so smoothly, and so early in most people’s training, that the compression itself becomes invisible. The corrective is not to stop using the word, which would be neither possible nor especially useful, but to remember what kind of word it originally was, and what work the noun form has been quietly doing ever since it broke free of its object in the eighteenth century.

XIII. Conclusion: The Instinct the Discipline Already Had

Cultures do not need to be treated as bounded, comparable wholes, because they were never bounded, integrated wholes to begin with. What actually exists, underneath the noun, is cultivation: an ongoing, relational activity, directed at a field, a mind, a household, a body, that only later, and only partially, hardened into the free-standing thing the discipline learned to compare. Recovering that older sense of the word does not require abandoning comparison. It requires comparing something else: not cultures as pre-given wholes, but the different ways in which different populations tend to the stakes they actually share, sleep, food, safety, the raising of children, the management of illness and risk.

The striking thing, on reflection, is how old this alternative actually is within the discipline’s own history. It is not a recent correction imported from outside anthropology. It is fully present, plainly stated, in the founding text of British social anthropology, in Malinowski’s own insistence that the task is to find out what holds a person’s life most intimately, not to arrive already knowing. Several generations of anthropologists trained in the very tradition Malinowski founded proceeded as though he had said something much narrower and much safer than he did, and only slowly, and only partially, has the discipline returned to what was there from the beginning. The founding instinct of the discipline was, in this sense, already the right one. What anthropology studies, when it is at its best, is not culture as a bounded possession to be inventoried and compared. It is whatever turns out to matter most to the people willing to let an outsider spend long enough among them to find out.