I. The Strange Absence of Rivalry
Social theory is usually taught as a contest. Students meet Marx against Weber, structure against agency, the individual against society, materialism against idealism, Marxism against Foucauldian analysis, Bourdieu against rational choice. Introductory textbooks arrange the field as a sequence of perspectives, each presented as an alternative to the ones before it, and examination questions ask candidates to assess which perspective best explains some feature of social life. The image that results is of an arena: a bounded space in which rival accounts of one object, usually called society, confront one another, and in which the task of the theorist is either to pick a side or to engineer a reconciliation.
The image is so familiar that one question is almost never asked of it. What exactly are these theories disagreeing about?
The question sounds naive, and it is meant to. Answering it requires a definition of rivalry more stringent than the one the textbooks rely on. Two theories are explanatory rivals when they attempt to explain substantially the same specified phenomenon and offer incompatible accounts of why or how it occurs. The elementary structure is this:
X happens because of A.
X happens because of B.
Where A and B cannot both be true of X, the theories are rivals. Everything depends on the third term. Disagreement between A and B is insufficient, because A and B may simply be answers to different questions. For rivalry in the strict sense, X must stay substantially the same across both statements.
Now perform a thought experiment. Take any passage of canonical social theory and strike out the most capacious nouns: society, modernity, capitalism, power, culture. Then ask what concrete phenomenon the passage is explaining. Ask what Marx is explaining when he analyses the working day, what Weber is explaining when he analyses the calling, what Durkheim is explaining when he analyses the totemic rite, what Goffman is explaining when he analyses the passage of hotel staff between kitchen and dining room, what Foucault is explaining when he analyses the examination, what Bourdieu is explaining when he analyses the preference for one kind of music over another.
The startling result is that X keeps changing. The length of a working day, the anxious discipline of a seventeenth-century Puritan merchant, the effervescence of an assembled clan, the management of front and back regions, the ranking of pupils, and the distribution of musical preferences across occupational groups are six different things. The theorists who explained them were, for the most part, explaining different things.
This yields the first thesis of the article. Much of what has been represented as theoretical rivalry in social theory is explanatory non-overlap. The canonical theories direct attention toward different phenomena far more often than they supply incompatible explanations of the same phenomenon.
The second thesis concerns the exceptions. There are places where social theorists do converge on a common problem and argue about it for generations: structure and agency, individual and society, micro and macro. These debates are real. Yet the objects they concern are of a peculiar kind, and I shall argue that the problems themselves depend on ontological abstractions that deserve more scrutiny than they have received.
The article therefore asks two questions. Were the social theories ever really rivals? And where they were rivals, were they always arguing about something that needed explaining? The answers lead to a third question, taken up in the later sections: what would a framework have to look like in order to compete over explanation instead of over attention, and what would count as evidence that findings in the human sciences could accumulate within it? Living Value Theory (LVT) enters the argument first as the instrument that makes the problem visible and only afterwards as a contrasting possibility.
II. How Shared Nouns Create Imaginary Shared Objects
The illusion of rivalry has a mechanism, and the mechanism is lexical. Canonical social theory runs on a small stock of extraordinarily capacious nouns: society, capitalism, modernity, religion, individual, structure, agency, culture, economy, state, class, power. Each gathers an enormous number of heterogeneous processes under one heading. That is what such words are for. Nobody could write a sentence about the last three centuries of European economic life without some term that sweeps together factories, contracts, banks, ledgers, enclosures, shipping routes, wage packets and stock exchanges.
Capitalism is the clearest example. To say that Marx and Weber both produced “theories of capitalism” creates an immediate impression of explanatory competition: one object, two accounts. But “capitalism” is far too large to function as a specified explanandum. It has to be decomposed before anything can be said to explain it.
Decomposed, the Marxian explananda include the exchange of commodities at prices that gravitate around some regular magnitude, the origin of a surplus where equivalents are exchanged for equivalents, the existence of a class of people with nothing to sell but their capacity to work, the lengthening and intensification of the working day, the compulsion on each capital to accumulate, the recurrence of crises, and the historical transformation of one set of relations of production into another. The Weberian explananda include the methodical, calculating conduct of economic life, the affinity between particular religious orientations and a particular ethic of work, the conditions under which commands are obeyed as legitimate, the spread of bureaucratic administration, the formal rationalisation of law, the role of capital accounting in the rational enterprise, and the historical distinctiveness of Occidental modes of organisation.
The two lists intersect. Both men wrote about the factory, about classes, about the long history of European economies. But intersection is not identity, and these remain lists of different questions. A reader who came to them without the word “capitalism” would not obviously file them under a single heading.
I shall call the mechanism nominal aggregation. It proceeds in steps. Heterogeneous phenomena are grouped under an abstract noun. Several theorists use the noun. Subsequent readers infer from the common noun a common explanandum. The theorists are then arranged as holders of rival positions on that explanandum. Finally the rivalry is taught, and the teaching confirms the existence of the object. Schematically:
heterogeneous phenomena → abstract noun → shared usage → inferred common explanandum → apparent rivalry
The formula for the whole sequence is simple: shared nouns create imaginary shared objects.
None of this amounts to a complaint against abstraction, which is indispensable. The difficulty arises at one specific point: when an abstract classificatory object is treated as though it identified a unitary phenomenon requiring a unitary explanation. Whitehead (1925) called the general form of this error the fallacy of misplaced concreteness, and Ryle (1949) gave it its most memorable illustration in the visitor who is shown the colleges, the libraries and the laboratories and then asks where the university is. The reader who asks which theory best explains capitalism, having been shown surplus value, the calling, bureaucracy and capital accounting, is in a similar position.
The simplification achieved by words like “capitalism” or “society” can therefore be deceptive in a precise sense. A word of this kind compresses complexity. Compression is useful, and it is a different achievement from discovering an underlying process. A noun that abbreviates a thousand phenomena has not thereby shown that the thousand phenomena share a cause. When two theorists are then said to offer competing theories of what the noun names, the appearance of a shared object is produced by the abbreviation itself.
The problem is compounded by the fact that the canonical theorists used these nouns themselves, often in their titles. That makes it natural to read their work as offering theories “of” the noun. But a title is a classification, and the explanatory work is done sentence by sentence, at a level of specificity where the grand noun has usually dropped out of sight. The test proposed here, which I shall call the explanandum test, simply insists on reading at that level. For any explanatory proposition, state what is being explained without using the aggregating noun. If two theorists are rivals, their statements will match. More often than the history of the field would lead anyone to expect, they do not.
III. Marx and Weber: The Paradigm Rivalry Reconsidered
The hardest case should be taken first. Marx and Weber are presented in virtually every curriculum as the paradigm rivals of classical social theory, and with better warrant than most pairings, since Weber wrote with Marx and the Marxists of his day constantly in view. If the explanandum test finds little direct rivalry here, the broader history deserves reconsideration.
There is real disagreement, and it should be acknowledged at the outset. Weber rejected strong versions of historical materialism. He denied that religious ideas could be treated as reflections of economic situations, and he denied that any single class of causes could claim general priority in historical explanation. His account of stratification separates class, status and party where the Marxian tradition tends to derive the latter two from the first, and it defines class by market situation where Marx defines it by relation to the means of production (Weber 1978). On these points the two bodies of work do address substantially the same X and say incompatible things about it. Whether a given status order can persist against the grain of the class structure is a question to which a Marxian and a Weberian may give different answers, and evidence can bear on which is right.
Now apply the test more widely. The question to ask is what specific phenomena Marx’s conceptual machinery was built to explain, which is a narrower question than what Marx thought about capitalism. The machinery of Capital, with its value, abstract labour, surplus value, organic composition of capital and reserve army, was built to explain how a surplus arises systematically from exchanges that appear equal, why the working day and the labour process take the forms they do, and why accumulation proceeds through expansion and crisis (Marx 1976). The machinery of Economy and Society, with its types of social action, its types of legitimate domination and its contrast of formal and substantive rationality, was built to explain why people orient their conduct to an order they regard as valid, how administrative staffs are organised and obeyed, and how spheres of life come to be governed by calculable rules. These are different machines for different jobs. Very little in Weber’s sociology of domination contradicts the theory of surplus value, because very little in it is about the origin of surplus. Very little in Capital contradicts the typology of legitimate domination, because Marx was not asking why officials are obeyed.
The Protestant Ethic is the instructive text, because it is the one most often cited as Weber’s reply to Marx. The textbook contrast runs: Marx explains capitalism materially, Weber explains it culturally. Weber explicitly declined that reading. In the closing pages he states that it was no part of his aim to substitute a one-sided spiritualistic causal interpretation of culture and history for an equally one-sided materialistic one (Weber 2001). What he investigated was narrower and stranger: how a particular doctrine of predestination, through the psychological pressures it placed on believers, fostered a methodical conduct of life in a worldly calling, and how that conduct had an affinity with a specific ethos of acquisition. The X is a form of conduct and its religious sources. Marx’s X, in the chapters on primitive accumulation that are sometimes set against it, is the forcible separation of producers from their means of subsistence. One could hold both accounts to be entirely correct without contradiction. They explain different features of an early modern world that both men called by one name.
To describe these relations more exactly, three categories are needed. The first is explanatory contradiction: the same X, with incompatible explanations. The second is differential explanation: the same broad empirical domain, but a different X within it, as when two theorists both write about the nineteenth-century factory and one explains the length of the shift while the other explains the authority of the foreman. The third is explanatory non-overlap: substantially different empirical objects, as when one theorist explains crises of overproduction and another explains the routinisation of charisma.
Conventional histories of social theory dramatically overestimate the first category because they collapse the second and third into it. Differential explanation is especially easily mistaken for contradiction, since the shared domain supplies shared vocabulary, and shared vocabulary, by the mechanism described above, supplies the appearance of a shared object. The literature on Marx and Weber is full of this. A century of commentary has asked whether Weber refuted Marx, completed Marx, or conducted a lifelong dialogue with Marx’s ghost. All three formulations presuppose that the two oeuvres are answers to the same questions. On inspection, a modest zone of contradiction, concentrated in the theory of stratification and in general claims about historical causation, is surrounded by very large territories in which each simply did not address what the other explained.
The significance of the case lies in its difficulty. Marx and Weber belonged to adjacent generations of one intellectual world, wrote about the same centuries and the same continent, and are joined by an explicit polemical link. If even here the proportion of direct explanatory competition is small, there is little reason to expect it to be larger between Durkheim and Goffman, or between Freud and Luhmann, where no such link exists.
IV. Social Theory as an Archipelago of Attention
Move outward from the paradigm case and the pattern holds. It would be pointless to offer miniature summaries of each canonical thinker. A more revealing exercise is to ask what each made newly visible.
Marx made relations of production, exploitation and accumulation recursively salient: after Marx, a wage is something one can ask questions about. Weber made meaningful conduct, legitimacy, bureaucracy, rationalisation and the configurations linking religious and economic life salient. Durkheim opened forms of solidarity, ritual assembly and collective classification to inquiry (Durkheim 1995). Simmel opened the forms of sociation: the dyad and the triad, the stranger, the secret. Freud opened the afterlife of early attachments in adult conduct. Goffman opened the minute organisation of encounters and the presentation of self (Goffman 1959). Foucault opened historically specific arrangements of knowledge, classification, discipline and government (Foucault 1977). Bourdieu opened sedimented dispositions, fields of positioning, distinction and reproduction (Bourdieu 1984). Luhmann opened the self-reference of communication. Latour opened the work that things do in holding associations together.
The vocabulary used to teach these thinkers is revealing. Lecturers and textbooks say that a theorist “focuses on,” “emphasises,” “draws attention to,” “foregrounds,” or “is concerned with” some range of phenomena. These are verbs of attention. They are rarely verbs of contradiction. One seldom reads that Goffman “refutes” Durkheim or that Bourdieu “falsifies” Simmel, and when one does read that Foucault “challenges” Marx, the challenge on inspection usually consists in directing attention to something Marx did not analyse, such as the capillary techniques by which bodies are trained, and insisting on its importance.
This suggests a redescription. Many canonical social theories function as organised systems of recursive relevance. They teach researchers what to notice, what to make askable, and what to regard as requiring explanation. A student who has absorbed Goffman can no longer sit in a restaurant without seeing the swing door between kitchen and dining room as a boundary between regions. A student who has absorbed Bourdieu can no longer hear a remark about wine without hearing a position being taken. The theory has reorganised what is salient, and the reorganisation recurs: it shapes the next observation, and the one after.
This is an achievement and should be stated as one. Making previously unnoticed phenomena visible can transform human understanding, and some of the greatest works of the social sciences are great for exactly this reason. Merton’s (1968) case for theories of the middle range, and Abend’s (2008) demonstration that sociologists use the word “theory” in seven distinct senses, both register that a great deal of what is called theory is something other than a set of explanatory propositions. What has been less often noticed is the consequence for rivalry. Attention and explanation are different intellectual operations, and they stand in different logical relations to their neighbours. Two explanations of one thing can contradict one another. Two directions of attention cannot. To look at the factory floor and to look at the confessional is to look at two things, and no quantity of looking at one will refute what is seen in the other.
An image may help. Social theory may have mistaken a library for a debate. Its canonical works occupy neighbouring shelves because disciplinary history has classified them together, and a reading list reproduces the shelf. Bibliographical co-location is then taken for ontological competition, as though books that stand side by side must be talking to one another about one thing. Some of them are. Most are doing something else: each is conducting its own inquiry into its own range of phenomena, and the conversation between them is largely the work of later commentators.
If that is right, the usual cartographic metaphor needs replacing. Instead of densely overlapping rival maps of a single territory called society, the field resembles an archipelago: a scatter of partially connected explanatory projects, some joined by causeways of real disagreement, most separated by open water, and all of them drawn onto one chart by the conventions of the discipline that teaches them.
V. Why Theoretical Pluralism Has Been So Easy
The redescription explains a feature of social science that is otherwise puzzling: the ease of eclecticism.
Consider a sentence of a kind found in countless dissertations. Marx illuminates the economic dimension of the case, Foucault its disciplinary dimension, Bourdieu the dynamics of distinction and reproduction, and Goffman the level of interaction. Such sentences are offered, and accepted, as theoretical synthesis. Examiners rarely object that the four frameworks are incompatible, and they are right not to, because in the use being made of them they are compatible. But nothing has necessarily been synthesised at the explanatory level. The analyst has changed explananda four times. In a study of a hospital, say, the wage relation, the classification of patients, the bearing of the consultants and the conduct of the ward round are four objects, and four lenses have been applied to them in turn.
Compare what happens where theories are rivals in the strict sense. A geologist cannot write that plate tectonics illuminates one dimension of mountain formation while a contracting Earth illuminates another. A physician cannot combine germ theory and miasma theory as complementary perspectives on cholera. Incompatible explanations of one phenomenon cannot be accumulated; one has to choose, or show that the phenomenon was two phenomena. Systems of attention carry no such restriction. They can be added together indefinitely, because looking here does not preclude looking there. The tolerance of social theory for pluralism, often praised as intellectual generosity and sometimes lamented as a lack of rigour, is better understood as a structural consequence of non-overlap.
The Frankfurt School is an illuminating case. The combination of Marx, Weber and Freud that runs from the early Institute through Dialectic of Enlightenment (Horkheimer and Adorno 2002) is usually narrated as a heroic synthesis of traditions that pulled in different directions. It was certainly an intellectual achievement. But the combination was possible in large part because each source supplied analytical machinery for a substantially different range of phenomena: Marx for commodity production and class relations, Weber for rationalisation and administration, Freud for the formation of drives and the appeal of authority. The three were fitted together much as one fits together instruments that measure different quantities. Had Freud offered a rival account of surplus value, or Marx a rival account of the Oedipal conflict, no synthesis would have been available, and one of the two would have had to give way.
The apparent success of synthesis is therefore itself evidence. It is usually read as showing that rival perspectives can be reconciled by sufficiently subtle minds. It may show something simpler: that the perspectives were seldom in conflict. The paradox can be stated directly. Theories that contradicted one another should be difficult to combine. The extraordinary combinability of social theories indicates how rarely they were rivals in the first place.
There is a corollary for the sociology of the discipline. A field of explanatory rivals would be expected to show a history of elimination, with defeated theories leaving the curriculum as phlogiston left chemistry. Social theory shows a history of accretion. The canon lengthens, and almost nothing is removed from it. This too is usually explained by the special difficulty of testing claims about human affairs. The difficulty is real, but the accretive pattern is also exactly what one would predict if the canon were chiefly a growing inventory of things worth noticing.
VI. Where the Real Debates Occur: Structure and Agency
The argument now has to face its most obvious counterexample. There are sustained, direct debates in social theory, in which successive generations address what all parties agree is the same problem. The debate over structure and agency is the most conspicuous, with the debate over individual and society as its close relative and the debates over micro and macro, material and ideal, subject and object as its cousins.
These debates meet the criterion of overlap. Durkheim’s (1982) insistence that social facts are external to and constraining upon individuals, methodological individualism’s insistence that only individuals act, Parsons’s (1937) attempt at a voluntaristic theory of action, Bourdieu’s (1977) habitus as a mediating device, Archer’s (1995) morphogenetic critique of conflation, and Giddens’s theory of structuration are positions on one question, and several of them are incompatible. Giddens is especially important because he names the problem as foundational. He treats the dualism of structure and agency as the central problem of social theory and proposes to overcome it with the duality of structure, in which structure is both the medium and the outcome of the practices it recursively organises (Giddens 1979, 1984).
At first sight this falsifies the first thesis. Here is a common X, and here are incompatible accounts of it. I want to suggest that it reveals a second problem, and that the second problem is the more interesting of the two.
Ask what the X is. What exactly are “structure” and “agency”? Under the explanandum test, both dissolve at once into heterogeneous collections. “Structure” has been used for grammatical rules, distributions of property, kinship terminologies, legal codes, the layout of buildings, occupational hierarchies, patterns of expectation, and statistical regularities in life chances. “Agency” has been used for bodily movement, deliberate choice, improvisation, refusal, creativity, the capacity to have done otherwise, and the bare fact that persons are the ones who do things. Neither term picks out a specified phenomenon, since each is an aggregate formed in the same way as “capitalism” and “society.”
The debate thus has an unusual shape. It begins by stabilising two heterogeneous sets of processes as two distinct objects. Having produced them as distinct theoretical entities, it then confronts the considerable difficulty of explaining how the one relates to the other: whether structure determines agency, whether agency produces structure, whether each presupposes the other, and in what measure. The sophistication of the proposed solutions is real. Yet every solution accepts the separation as its starting point, including those that aim to transcend it. Structuration theory describes a duality where earlier theories saw a dualism, but a duality remains a relation between two terms, and the theory is still an answer to the question of how the two are joined.
The same shape recurs across the family of debates:
individual ↔ society
micro ↔ macro
material ↔ ideal
subject ↔ object
In each, the theoretical machinery first performs a separation and then devotes enormous effort to reconnecting what it separated. The effort is proportional to the cleanness of the cut. Once “the individual” has been constituted as an entity whose properties can be stated without reference to others, and “society” as an entity with properties of its own, the question of how such different things could affect one another becomes pressing and close to insoluble, and a century of ingenuity can be spent on bridges, mediations, dialectics and dualities.
From a process-ontological perspective there is a prior question, which the debate does not ask because its terms presuppose the answer. Should these entities have been granted foundational status at all? That is a question about what kind of things “individual,” “society,” “structure” and “agency” are, and it requires a distinction that the standard debate has not had available.
VII. Socially Real Without Being Ontologically Basic
The distinction is this: something can be socially real without being ontologically basic.
The vocabulary of Living Value Theory offers a way to say it exactly. LVT distinguishes levels of recursive engagement, among them coordination that proceeds unnoticed (L1), misalignment that is felt (L2), articulation in symbols (L3), and the institutional stabilisation of what has been articulated (L4). Binary classifications are characteristically L4 operations: a distinction that someone once drew becomes fixed in law, in administrative categories, in curricula, in forms and registers, and from then on is encountered as a feature of the world. In these terms, an L4 stabilisation can be socially real without being ontologically basic.
“Individual,” “society,” “economy,” “nation,” “organisation” and “class” are categories of this kind, and they have become extraordinarily consequential. Laws operate through them. Statistics measure them. Institutions reproduce them. People understand themselves through them. Governments intervene in their name. A census counts individuals; a constitution addresses citizens; a central bank acts upon the economy; an army mobilises the nation. None of this reality is being denied. A stabilisation that organises the conduct of millions has effects as hard as any.
But causal efficacy does not establish ontological primitiveness. Money is socially real without being a primitive constituent of living existence. Corporations are socially real, own property and can be sued, without being organisms. Nations can kill and mobilise millions without thereby becoming foundational units of what there is. In each of these cases it is uncontroversial that the entity is a stabilised outcome of processes that preceded it and that could be described without it. Nobody asks how money and living beings, considered as two independently constituted kinds, manage to interact.
Individual and society can be treated in the same way, and so can structure and agency. This is the point at which LVT dissolves the problem where other frameworks set out to solve it. It proposes no better bridge between two ontological poles, because it refuses the initial partition. On its account living beings are already embodied, already being-with others, already dwelling within conditions they did not make, and human life, in addition, forms materials and, in symbolising, recursively articulates itself and its worlds. There is no moment at which an individual exists first and then enters into relation with a society, and no moment at which a structure exists apart from the living coordination that carries it. The processes are continuous. “Structure” and “agency” subsequently emerge as L4 descriptions of aspects of those processes: the first collects what is sedimented and resistant in coordination, the second what is open and unfixed in it. As descriptions they can be useful. As primitives they generate a problem that the processes themselves do not contain.
The situation resembles other historical pseudo-problems. Nineteenth-century physics invested great ingenuity in determining how bodies move relative to the luminiferous ether, and the failure of experiments to detect that motion produced increasingly elaborate constructions. The question was never answered. Once the underlying physics was reformulated, the need to relate the two entities disappeared, because one of them was no longer required. Ryle’s (1949) treatment of the Cartesian problem of how mind moves body has the same form: the difficulty of connecting two substances was produced by having posited them as two substances.
The claim being made here is deliberately limited. It is less that LVT has finally solved structure and agency than that LVT provides reasons for asking why structure and agency ever appeared to require reconciliation. A problem that arises only once two L4 stabilisations have been mistaken for the basic furniture of the world is a problem about the mistake.
VIII. The Strange Inverse Pattern of Social-Theoretical Rivalry
The two halves of the argument can now be brought together, and together they describe a pattern that is odd enough to deserve notice.
At lower levels of abstraction, where explananda are concrete, the canonical theories are well grounded empirically and seldom rival one another, because they explain different phenomena. At very high levels of abstraction they do rival one another, directly and at length, but over objects that were partly generated by theoretical stabilisation. Schematically:
concrete phenomena → strong empirical grounding, weak theoretical overlap
highly abstract objects → strong apparent overlap, questionable ontological grounding
Rivalry and grounding vary inversely. Where the field knows most securely what it is talking about, its theories talk past one another; where its theories engage one another most intensely, it is least clear that there is something there to be talked about.
The interesting category is the one that falls between. It can be specified exactly: the same clearly specified phenomenon, with incompatible process explanations. This is what explanatory rivalry looks like in fields where theories are regularly eliminated. How much canonical social theory falls into it?
The answer is that nobody knows, because the question has not been put in a form that permits counting. Cases certainly exist. Durkheim and Tarde gave incompatible accounts of how conduct spreads through a population. Rational-choice theorists and Bourdieu give incompatible accounts of why working-class students leave education earlier than middle-class peers with similar marks. Marxian and Weberian theories of stratification, as noted, collide over the autonomy of status. It is notable that these are among the debates in which evidence has been brought to bear most productively, that they are conducted at the middle range, and that they attract less attention in theory curricula than the grand oppositions.
The matter should be treated as an empirical question, since a rhetorical accusation, in either direction, settles nothing. I therefore propose an instrument: an explanatory-overlap matrix. Take a set of canonical theorists. From each oeuvre, sample explanatory propositions, that is, statements of the form “X occurs because of A” or “X occurs through process A.” Code each proposition by its explanandum, stated without aggregating nouns. Arrange the theorists along both axes of a matrix and enter in each cell the proportion of sampled propositions in which the two theorists address substantially the same X. Then, for the propositions that do overlap, code separately whether the explanations offered are compatible, complementary or contradictory.
The exercise would be laborious and would involve judgement at the coding stage, as content analysis always does. Intercoder reliability could be measured, and the criteria for “substantially the same X” could be varied to test the robustness of the result. The prediction that follows from this article is specific. The matrix will be sparse. Most off-diagonal cells will be close to empty; the filled cells will cluster along a few known lines of dispute; and among the filled cells, the proportion coded as contradictory will be lower than the reputation of the pairing suggests. A dense matrix, with frequent overlap and frequent contradiction, would refute the first thesis of this article.
IX. Why LVT Keeps Getting into Fights
Living Value Theory now moves into the foreground, and it does so by way of a puzzle about its own conduct.
A curious feature of the LVT project has been its repeated collision with established explanations. In work on pharmaceuticals, mental health, economics, anthropology, value, sleep, narrative, ritual, law, culture, comparison, technology, education and, most recently, entrepreneurship, the experience has been the same. LVT can accept most of the empirical observations of the literature it meets and still finds itself rejecting the theoretical explanation attached to them. From outside this can look like belligerence: a new framework picking quarrels with every established one in turn. Given how little the established frameworks quarrel with each other, the contrast needs explaining.
The preceding sections explain it structurally. Conventional theories largely occupy different explanatory territories and can therefore coexist. A new theory of the conventional kind would be expected to behave likewise: it would identify a neglected range of phenomena, make them salient, and take its place on the shelf. Instead of claiming another territory, LVT moves underneath the territorial divisions and addresses the processes through which the phenomena of each territory occur. Since every territory already has a resident theory, and every resident theory comes with some account of process, however implicit, the collisions are unavoidable.
The contrast can be put in a sentence. Conventional social theories mostly compete for attention; LVT competes over explanation. The qualifier “mostly” matters, since the previous section listed exceptions, and nothing here depends on denying them.
A few rapid examples show the pattern. Bourdieu observes sedimented dispositions and patterned differences in taste. LVT has no need to deny them; the correlations in Distinction are what they are. It asks what processes produce the phenomena subsequently collected under “habitus,” and it holds that a term which names the sedimented outcome of coordination cannot also serve as the explanation of that coordination. Foucault identifies disciplinary classifications and techniques of government. LVT does not need to deny these either. It asks what living processes make classification, uptake, compliance, resistance and institutional stabilisation possible, and it declines to let “power” stand as the answer, since “power” is one of the aggregating nouns. Marx identifies commodities, wage labour, exploitation and accumulation. LVT can accept these as historical formations of great consequence while refusing to elevate them into ontological primitives from which the rest of human life is to be derived. Goffman’s observations of interaction can be extraordinarily precise, and they still leave open the process ontology of being-with through which any encounter occurs: what it is for one living being to discern another who is discerning it in turn.
In each case the same thing happens. LVT shares the explanandum and rejects the explanans. Taste differs by class: agreed. It differs because of habitus: contested. Bodies are trained in institutions: agreed. They are trained because power produces subjects: contested. By the definition given at the outset, that is rivalry in the strict sense: the same X, with an incompatible account of how it occurs.
This also explains the asymmetry of the encounters. The established theories do not, as a rule, feel obliged to reply, since from within a system of attention a challenge to its process assumptions can be received as a mere change of subject. From the side of LVT, the subject has not changed at all. That mismatch is uncomfortable, and it is informative. A framework that collides with everything is doing something different in kind from frameworks that collide with almost nothing, and the difference lies in the level at which it operates.
X. From Theory to Process Ontology
Why does LVT operate at that level? The answer is that it is increasingly difficult to understand as a social theory at all, because its basic concepts are not designed around a bounded social domain. Its foundational claim concerns living processes.
The architecture can be stated compactly. Life occurs through continuous mediation. For every living being there are three irreducible zoetic mediations: multisensorial embodiment, being-with, and multiversal dwelling, the last understood as the environmental and planetary conditions that no one made. Some forms of life generate two further, poietic mediations: multimateriality, the forming of materials, and multisymbolization. The five are always co-present, with only their salience shifting, and none can be reduced to another. Recursive engagement occurs at five differentiated levels, L1 to L5, of which the passage from coordination that works unnoticed, through felt misalignment and articulation, to institutional stabilisation has already been introduced. It occurs, further, in relation to three types of recursive counterparts: nonrecursive, selfrecursive and interrecursive.
What matters here is less the terminology than the status claimed for what the terms name. These are proposed as process-ontological conditions of living, conditions without which nothing that lives could continue to live, and they are therefore not variables peculiar to one empirical domain. Embodiment is no more the property of the sociology of the body than being-with is the property of the sociology of interaction. On this account both are dimensions of every phenomenon that any human science has ever studied, whether or not the science in question has noticed them.
Once this is seen, the enormous apparent range of LVT ceases to be mysterious. Consider a list of questions that the framework has been used to address. Why can a four-year-old enjoy Mozart? How does someone sleep well? What happens in Hindu image worship? How does Kula exchange work? Why does someone trust a pharmacist? How is clothing produced? How do legal processes reopen settled relationships? How does narrative buffer metabolic stakes? How should a startup be incubated?
By any ordinary standard this list is absurd. It runs across developmental psychology, sleep science, the anthropology of religion, economic anthropology, medical sociology, material history, jurisprudence, literary theory and business studies. A theory that claimed competence in all of them would normally and rightly be suspected of saying nothing in particular about any.
The suspicion rests on a premise, however: that the questions are radically heterogeneous. They appear so only if disciplinary classifications are taken for ontological divisions, as though the world came sorted in advance into the psychological, the religious, the economic and the legal. Those classifications are themselves L4 stabilisations with a recoverable history, and what they mark are the boundaries of faculties and journals. For LVT the nine questions concern different configurations of the same living processes. The child listening to Mozart and the trader waiting for a Kula partner are both embodied, both with others, both dwelling somewhere, both engaged with formed materials and with symbols, and both coordinating at several levels at once. What differs between them is the configuration, and the configuration is what the analysis sets out.
The comparison with canonical social theory can now be made precise. A system of attention has a range: the phenomena it makes salient. Its range is extended by pointing it at new phenomena, and its limits are reached where it has nothing to notice. A process ontology has no range in that sense. Either its conditions hold wherever there is life, in which case nothing living falls outside it, or they fail somewhere, in which case it is false. This is a much more exposed position than that of a perspective, since a perspective cannot be refuted by a case it does not cover, and it is why the question of evidence has to be faced directly.
XI. The Most Unusual Evidence: LVT Has Stopped Growing
The evidence I want to put forward is of an unusual kind. It concerns what has failed to happen.
Ordinarily, increasing theoretical scope produces conceptual proliferation. A framework developed for one domain is carried into another, meets phenomena it was not built for, and acquires new concepts, auxiliary hypotheses and qualifications to accommodate them. Over time the foundations swell. This is the normal life course of general theories in the social sciences, and it is one of the reasons such theories eventually become unwieldy: the machinery required to state the theory grows as fast as the territory the theory covers. The successive schemes of Parsons are the classic illustration.
The recent history of LVT appears to show the reverse. Since the latest substantial formulation of its core principles, roughly forty further articles have applied the framework to radically heterogeneous material. These applications have generated many derived insights without requiring significant alteration of the foundational architecture. The five mediations, the five levels and the three types of counterpart stand as they stood.
The derived insights are real discoveries, and some examples will indicate their character. Recursive opening names the way a settled coordination becomes askable again. Recursive discernment names discernment that alters the conditions of its own subsequent operation. Returnability names the discerned possibility of resuming coordination after departure, interruption or loss of access. Metabolic buffering names what narrative does when it retains consequential pasts so that they need not be lived through again at full cost. The passage from L3 to L4 was identified as the crux of ethnographic comparison, where a lived articulation becomes a portable class. The application to startup incubation yielded a method of discovery that precedes the market.
Each of these said something that had not been said before. Yet none is a new ontological primitive, since each is specified in terms of the existing architecture. Returnability is a configuration of discernment across mediations and levels, and adds no sixth mediation. Metabolic buffering is a function of multisymbolization in relation to embodied stakes, and adds no new level. The discoveries are theorems, so to speak, where they might have been expected to be axioms.
The criterion at work can be named: non-inflationary extensibility. A framework has it to the degree that it can expand its explanatory domain without proportionately expanding its foundational conceptual machinery. The corresponding measure, at least in principle, is a ratio: analytical scope relative to primitive conceptual complexity.
The criterion is needed because scope alone proves little. Large theories can achieve scope cheaply in two ways. One is semantic elasticity: a concept such as “power,” “exchange” or “system” is stretched until it covers everything, at which point it distinguishes nothing. The other is conceptual proliferation: a new term is coined for each new case, and the theory becomes a lexicon. Both produce frameworks of apparently universal reach. Neither produces cumulation, because in the first case the concepts no longer mean anything determinate and in the second there is no stable core for findings to accumulate around.
The striking possibility presented by the recent development of LVT is enormous scope with low conceptual strain. It has to be stated as a possibility. An absence of growth admits of a deflationary reading, namely that the concepts were loose enough from the start to absorb anything, and an author’s own impression that his framework has stabilised carries no evidential weight. The observation becomes evidence only if elasticity can be ruled out and if the stability can be checked by procedures that others could carry out. That requires tests.
XII. Three Severe Tests of a Cumulative Process Ontology
A claim of this kind should be turned into a research programme, since the alternative is self-congratulation. Three tests suggest themselves, and each is designed so that LVT could fail it.
The first is the conceptual-strain test. Select cases maximally distant from the domains in which LVT originated, which were medical anthropology and the study of pharmaceutical markets in South Asia. For each case, ask whether analysis requires modification of foundational concepts, the introduction of auxiliary concepts, metaphorical stretching, or special pleading. Each of these is a recordable event, and their frequency across cases is a measure of strain. The selection of cases matters greatly. A framework’s originator will tend, without intending it, to choose objects on which the framework performs well. For that reason the startup case carries particular weight. It arrived by accident, through a question put to the author by his son about how a new business should be developed, and it concerned a field in which the framework had no prior investment. Cases supplied by others, and above all by sceptics, would be more valuable still.
The second is the invariance test. Ask whether the same concept means the same thing across radically different applications. Does being-with retain the same definition and the same inferential role in analyses of childhood, of sleep, of Kula exchange, of pharmaceutical transactions and of entrepreneurship? Inferential role is the stricter requirement. It is easy to preserve a form of words while quietly changing what follows from it. If being-with licenses one set of inferences when applied to people who share a bed and an unrelated set when applied to people who share a company, then the term is two concepts under one label, and the apparent scope of the theory has been purchased through semantic elasticity. The test can be operationalised by extracting every use of a core term from the corpus, stating the inferences drawn from it in each context, and having readers who do not know the purpose of the exercise judge whether the uses are consistent.
The third is the retrospective reanalysis test. Return to the earliest LVT analyses and redo them with the current architecture. Three outcomes are possible. The first is substantive contradiction: the current framework yields conclusions incompatible with the earlier ones, which would show that the framework has changed its content while keeping its name. The second is additive enrichment: the current framework yields the earlier conclusions together with new ones that the earlier analysis could not have reached. The third is analytical stability: the later terminology makes the earlier analyses more explicit without changing their substantive conclusions. The third outcome would be the strongest evidence of stabilisation, and the second would be compatible with it. The first, if frequent, would be fatal to the claim.
This test has a feature that makes it unusually resistant to manipulation. The corpus is chronologically fixed. The early analyses were published at known dates, in a known vocabulary, before the later architecture existed, and they cannot be rewritten to agree with it. The historical development of LVT therefore becomes evidence about LVT. A framework’s past is ordinarily of biographical interest only. Here it functions as a set of commitments made in advance, against which the mature framework can be checked.
None of the three tests has yet been carried out with the independence that would make its result compelling, and I do not claim otherwise. What can be claimed is that the tests are well defined, that their outcomes are not foregone, and that they address the right question. For a would-be cumulative framework, impressiveness is beside the point; what counts is whether it has grown in scope while holding its primitives fixed and its meanings constant. That question has answers.
XIII. What Cumulation Would Mean in the Human Sciences
The larger issue behind these tests is an old one. The natural sciences and mathematics routinely generate enormous cumulative structures from relatively stable primitives. A result established in one decade becomes a tool in the next, and students are not required to read the founders in order to practise the science. Social theory has struggled to do the same. Its students read the founders because the founders have not been superseded, and its findings, however rich, do not readily stack.
The conventional explanation locates the difficulty in the object. Social reality, it is said, is too reflexive, too historically variable or too dependent on context for cumulation of the natural-scientific kind. Human beings react to what is said about them; societies change; meanings differ from place to place. There is truth in each of these observations, and they have been made with great force from Dilthey to the present.
The analysis in this article suggests an additional possibility, which the conventional explanation tends to obscure. Perhaps cumulation failed in part because social theory did not possess common explanatory objects or a common process ontology. If the canonical theories are largely systems of attention directed at different phenomena, then each has accumulated observations within its own domain, often impressively. Marxian political economy has cumulated; so has the Goffmanian study of interaction; so has Bourdieusian field analysis. What has not happened is cumulation across them, and on the present account it could not have happened, because their findings are stated in conceptual coordinates that do not map onto one another. A finding about habitus and a finding about disciplinary power cannot be added together, any more than a distance and a temperature can. The archipelago grows island by island.
LVT suggests another model, which can be laid out as a sequence:
stable process ontology → heterogeneous applications → derived concepts → transferable findings → further applications
The decisive feature is where the growth occurs. In this model it occurs on the output side. Applications multiply, derived concepts are formulated, and findings from one domain become available in another because they are stated in the same coordinates. A result about returnability reached in the analysis of children’s stories can be put to work in the analysis of sleep, and a result about the passage from L3 to L4 reached in the analysis of ethnographic comparison can be put to work in the analysis of law. The foundational machinery, meanwhile, stays as it was.
This is the difference between conceptual accumulation and conceptual proliferation. In proliferation, new concepts are added at the base and the theory becomes larger without becoming more powerful. In accumulation, new concepts are derived from an unchanged base and each derivation increases what can be derived next. The first is how lexicons grow, and the second is how a body of knowledge grows.
Whether the human sciences can sustain cumulation of the second kind remains open, and it would be rash to announce that the question has been settled by one framework’s recent history. What the argument of this article does establish is that the standard reason for pessimism is incomplete. The reflexivity and variability of human life may be real obstacles. They have never been tested against a common ontology, and the absence of one could not be counted among the causes of failure until the theories were seen to lack it.
XIV. Conclusion: What Were the Social Theorists Actually Theorising?
The article began with a naive question: what exactly are the canonical social theories disagreeing about? The traditional history of the field presents an arena of competing accounts of society. A closer examination suggests something stranger.
Many of the apparent rivals were explaining different things. Shared abstract nouns concealed the difference by supplying an imaginary common object, and the conventions of teaching turned the shelf on which the books stood into a debate among their authors. Where theorists did converge on a single problem and argue it out, the problem was often an opposition between two terms, structure and agency or individual and society, which their own abstractions had produced and institutional stabilisation had hardened. Explanatory rivalry in the strict sense, the same specified phenomenon with incompatible accounts of its process, exists in social theory but occupies a smaller part of it than the image of the arena implies, and how small a part is a question that can be settled by counting.
None of this diminishes Marx, Weber, Durkheim, Goffman, Foucault or Bourdieu. It allows their achievements to be described more accurately. They made particular aspects of human worlds visible and askable with extraordinary power, and anyone who has learned to see with them sees more. But systems of recursive attention are not necessarily rival ontologies, and they should not be asked to behave as if they were.
One consequence is worth stating in its most provocative form. Social theory has long lamented the absence of a successful synthesis of its competing theories. Perhaps no synthesis succeeded because the theories were never sufficiently commensurable to be synthesised. What was missing was something prior to synthesis: a common ontology within which their heterogeneous objects could first be located, so that it would become possible to say where two accounts concern the same process and where they do not.
That changes the question posed by Living Value Theory. The question is no longer whether LVT should join Marxism, Weberian sociology, Bourdieusian theory, Foucauldian analysis and the rest as one more perspective on social life, but whether these were ever the right comparison class. A process ontology of living that competes over explanation, extends without inflating, and can be tested for strain, invariance and retrospective stability is a different kind of undertaking from a system of attention, and it should be judged by different standards, including harsher ones.
Social theory may have spent more than a century searching for agreement among theories that were rarely disagreeing. The more fundamental task is to discover what kind of world made all their different objects possible in the first place.
Selected References
Abend, Gabriel. 2008. The Meaning of “Theory”. Sociological Theory 26(2): 173–199.
Archer, Margaret S. 1995. Realist Social Theory: The Morphogenetic Approach. Cambridge: Cambridge University Press.
Bourdieu, Pierre. 1977. Outline of a Theory of Practice. Translated by Richard Nice. Cambridge: Cambridge University Press.
Bourdieu, Pierre. 1984. Distinction: A Social Critique of the Judgement of Taste. Translated by Richard Nice. Cambridge, MA: Harvard University Press.
Durkheim, Émile. 1982 [1895]. The Rules of Sociological Method. Translated by W. D. Halls. London: Macmillan.
Durkheim, Émile. 1995 [1912]. The Elementary Forms of Religious Life. Translated by Karen E. Fields. New York: Free Press.
Foucault, Michel. 1977. Discipline and Punish: The Birth of the Prison. Translated by Alan Sheridan. New York: Pantheon.
Giddens, Anthony. 1979. Central Problems in Social Theory: Action, Structure and Contradiction in Social Analysis. London: Macmillan.
Giddens, Anthony. 1984. The Constitution of Society: Outline of the Theory of Structuration. Cambridge: Polity Press.
Goffman, Erving. 1959. The Presentation of Self in Everyday Life. New York: Doubleday.
Horkheimer, Max, and Theodor W. Adorno. 2002 [1947]. Dialectic of Enlightenment: Philosophical Fragments. Translated by Edmund Jephcott. Stanford: Stanford University Press.
Marx, Karl. 1976 [1867]. Capital: A Critique of Political Economy, Volume 1. Translated by Ben Fowkes. Harmondsworth: Penguin.
Merton, Robert K. 1968. Social Theory and Social Structure. Enlarged edition. New York: Free Press.
Parsons, Talcott. 1937. The Structure of Social Action. New York: McGraw-Hill.
Ryle, Gilbert. 1949. The Concept of Mind. London: Hutchinson.
Weber, Max. 1978. Economy and Society: An Outline of Interpretive Sociology. Edited by Guenther Roth and Claus Wittich. Berkeley: University of California Press.
Weber, Max. 2001 [1904–05]. The Protestant Ethic and the Spirit of Capitalism. Translated by Talcott Parsons. London: Routledge.
Whitehead, Alfred North. 1925. Science and the Modern World. New York: Macmillan.