Chapter 04 Society Is Not a Combination of a Few Grand Concepts: From a Platter of Concepts to Differentiating Units
I. How a Platter of Concepts Occludes Consequences
A single governing axis cannot be presupposed for society. The clues supplied by class theory come from the real differences it extracts, but that does not make the class image a safe, ordinary, limited model; the concepts of the state, the market, the nation, culture, and the family may all provide important clues in particular questions. When any one concept is placed in advance at the highest level of explanation, other differences are easily translated into its appendages, and complex relations are flattened into a single plane.
After rejecting a single axis, people often retreat to “the joint action of multiple factors”: society has the state, the market, culture, the nation, and the family in addition to class. It is not necessarily wrong to acknowledge that several factors may operate together, but the formulation easily stops at a parallel list of concepts; the concepts have multiplied, while the formation of consequences still has not been explained more precisely. “They are all relevant” at most points to several directions. It does not say who extracts the concrete differences, how they acquire greater operative weight, which qualification or restriction they connect to, or which differences, though present, never enter the result at hand.
A child’s school admission may involve institutional thresholds, income, the ability to rent housing, the ability to take time off, information channels, and the ability to pay at the same time. Wedding and funeral cash gifts may likewise involve amounts, reciprocal exchanges, local evaluation, and mutual aid; a workplace injury may involve labor relations, evidence, procedure, and waiting. If these questions are merely distributed into drawers labeled “the state,” “the market,” “culture,” and “the family,” analysis has not yet begun to track how the consequence is formed. Listing concepts completes only a classification; the consequence still has to be traced through concrete relations. A single axis sends all material toward one answer; a platter of concepts leaves several answers, but still sorts the material into pre-given names first. The former compresses differences into one line, the latter packages differences into several grand terms. Both can miss the steps that actually change the consequence: whether material can be obtained, whether a qualification can be recognized, whether waiting can be endured, whether a relation can continue, and who ultimately bears responsibility.
A platter of concepts appears prudent because it acknowledges that a matter has more than one side. But acknowledging several sides does not yet explain how those sides meet. If income, household registration, local reputation, care responsibility, and institutional procedure are merely named one after another, they remain like unrelated blocks. What must actually be made clear is which condition enables another condition to take effect, and which restriction prevents an existing resource from becoming a choice. This parallel listing also easily produces a false sense of balance. Once every grand term has been mentioned, people assume that the material has been treated fairly; in fact, some relations remain in the background, and some differences are named without anyone asking how they operate. Fairness in parallel listing cannot replace fairness toward the chain of consequences.
Listing concepts side by side has another problem: it easily mixes things from different levels. Household registration is a form of institutional eligibility; whether a landlord cooperates is a concrete relation; taking time off is a time condition; a shortage of places is a distributional restriction. They may all appear in a school-admission problem, but they do not acquire the same position merely because they are all called “state factors.” If analysis does not separate the levels, it cannot know whether to ask about the institutional text, the practice at the service window, support provided by private actors, or family time. Grand terms also often describe an ongoing process as a static background. When people say that an institution is present, they still have to ask how it enters the result through registration, verification, interpretation, supplementary documents, and deadlines; when they say that the family is present, they still have to ask how care, economic dependence, and emotional demands change the choices a person can make. Only by returning the background to the process can concepts stop occluding where the consequence actually passes through. A concept can temporarily focus analysis when it is used as an entry point; when it is used to close the question, it prevents further inquiry. The former says, “Let us first see how the administrative procedure intervenes”; the latter says, “This is a state problem,” and stops there. The difference is not whether a concept is used, but whether the concept leaves an exit back to the material.
II. Grand Concepts as High-Level Compression Labels and Indexes
“The state,” “the market,” “culture,” “the family,” and “the class image” are high-level compressed names for many concrete differences and relations. Grand concepts are not useless. They can provide direction and an entry point, allowing an outline to become visible first. Saying “the state” can gather together and direct attention to law, administration, household registration, registration systems, public finance, schools, borders, service windows, and enforcement; saying “the market” can gather together and direct attention to prices, wages, credit, rent, transactions, various opportunities including employment opportunities, risk, and ability to pay; saying “culture” and “the family” can likewise bring rites and customs, social face, habits, care, dependence, debt, inheritance, and emotional relations into view. Such an outline keeps discussion from losing all direction.
The problem is that an entry point cannot replace an analysis of consequences, and compression also occludes internal differences. Saying “the state” does not explain whether household registration, school places, registration with the subdistrict administration, administrative procedure, legal text, an explanation at the service window, or fiscal allocation is creating the restriction; household registration and school places are not the same kind of threshold. Saying “the market” likewise does not say which of housing prices, wages, credit, rent, employment opportunities, opportunity costs, or risk pricing has connected to the person’s choice; housing prices and wages do not change choice in the same way. Care, debt, control, and inheritance do not all lead to the same consequence either. A concept points toward a field of relations, but that does not mean every relation within it has the same operative weight. When differences of different kinds, operative weights, and directions are gathered under one name, the name itself can no longer substitute for detailed judgment. A grand term may appear to have explained the problem when it has merely wrapped up the material that still requires further questioning.
Compression can also misrepresent differences that constrain one another as pointing in the same direction. A rise in wages may not immediately relieve housing pressure; acquiring institutional eligibility may not make up for a shortfall in time for care; local reputation may bring mutual aid while simultaneously making refusal more difficult. If only one general name is retained, these tensions are smoothed away, and analysis is left arguing only over which concept is more important. Because concepts are convenient for argument, people also often mistakenly assume that concepts already correspond one-to-one to consequences. In fact, relations pointing in opposite directions can exist under the same name: the family can provide care, but it can also increase responsibility; local relations can provide mutual aid, but they can also increase pressure. Unless compression is unfolded again, these directional differences are easily erased by a single overall judgment.
The conceptual outline, the concrete institution, the points within relations where things are taken up and carried forward, and the actual consequence must be kept separate. An institutional name can indicate which kind of threshold to investigate, but it cannot replace questions about documents, deadlines, records, service windows, and modes of enforcement; the name of the family can indicate care and dependence, but it cannot replace questions about time, debt, health, and available exits. If a name is treated as the answer as soon as it appears, the points that can be examined and changed disappear instead. A compressed result cannot itself extract material, determine weight, arrange classifications, or make a ruling. Those acts are always performed by interpreters, organizations, or institutional procedures: someone chooses which records count, a procedure stipulates which documents are valid, and a form of naming puts relations that could previously be discussed separately into the same category. Asking about these processes preserves the concept while preventing it from occluding the actual extraction of material and the judgment applied to it.
Grand concepts do not need to be abolished; they can remain indexes for entering a question. They indicate which material is worth seeking, but cannot decide in advance which material is primary or secondary, still less act in place of concrete interpersonal, institutional, and resource relations. The point of taking them apart is to make names accept the constraints imposed by the material again; their public use remains. When people use “the market” or “the state,” they should be able to say which relations it points to here, which parts have not yet been proved, and which counterexamples might force the judgment to contract. This restriction does not deprive a concept of force. It prevents the concept from pretending to be a conclusion that has already been completed.
The convenience of compression also changes the way questions are asked. When people first ask, “Is this really a state problem or a market problem?”, attention easily concentrates on choosing a name for the event. If they first ask which difference changed a qualification, a cost, how things are taken up and carried forward, or an exit, the name can remain until later as a temporary tool for organizing the material. This latter question does not deny concepts, but it does not let a concept distribute the conclusion among the materials in advance. A concept can still be used as an index, but each use must return to one simple question: what material does it point to here, and what material does it occlude? If that cannot be answered, the concept may only be giving an existing judgment a different name. The same question applies to the state, the market, the family, culture, and the nation, as well as to the class image.
Reopening grand concepts does not mean looking for some social atom or final material. The analysis must pursue which concrete differences and relations are brought into comparison in this matter, how they are recognized, downplayed, or amplified, and how they enter the consequence along institutional, interpersonal, and resource conditions. Society is a differentiation network, but this judgment does not provide another complete total image. How material is extracted and how it enters a ruling must still be traced to concrete people and procedures.
III. Unfolding the School-Admission Eligibility Chain and Institutional Responsibility
In some regions and particular admission years, household registration or residence conditions, school places, application rules, and local provisions issued by education authorities may affect admission. Income, the ability to rent housing, the ability to take time off, information channels, and the ability to pay additional expenses are composite situational factors a family may face while going through the procedure, not uniform formal thresholds proved by the policy sources cited. When a child enters school, whether a residence permit is valid, whether social insurance has been continuous, and whether a rental contract is compliant can all change the result. Whether a landlord is willing to cooperate with filing, whether the document deadline can be met, whether the parents can take time off to go to the service window, whether anyone reminds them of the application procedure, whether school places are scarce, and whether the community is willing to issue a certificate all operate at different stages. Each can turn the facts of everyday residence and work into institutional eligibility, or make that conversion fail. The policy examples here are used only within the jurisdictions, periods, procedures, and material requirements specified by the admission policies discussed in Chapter 20.
An eligibility chain can be seen here: residence must leave records that can be recognized; the records must be assembled within the deadline; and assembling the documents requires someone who can arrange time, obtain information, and complete the necessary trips. When a landlord does not cooperate with filing, families with similar incomes and similar labor positions may still lose the same possibility of admission; when social insurance is interrupted for one month, a set of documents that appeared complete may change eligibility. These consequences are not directly deduced from abstract names. They take shape where concrete differences meet procedures. Eligibility chains also interconnect. A residence permit, social insurance, a rental contract, and filing are not isolated items; the absence of one document may mean that even with another document available, the procedure cannot be completed. The ability to take time off and information channels are not purely private conditions either: they determine whether a family can turn existing facts into documents that a service window can recognize before the deadline. This is why tracing relations matters more than listing concepts.
The same rule does not necessarily produce the same result for every family. When school places are scarce, whether the documents are complete, the speed of supplying additional documents, residential stability, and the availability of a community certificate may all change where a family ends up. This does not shift institutional responsibility onto the individual; it refuses to let an abstract institutional name occlude how an institution connects with living conditions through concrete procedures. Time is not an incidental background either. Document deadlines, time off, and trips to the service window bring work arrangements into the conditions of admission; information channels affect whether a family knows when to apply, to whom, and with what documents. The ability to pay extra expenses may not determine the result by itself, but when expenses for housing, transportation, or replacement documents, or losses incurred by taking time off, arise, it changes whether the family can continue to endure the process. Tracing these relations does not deny institutional critique; it brings the institutional threshold down to a place that can be explained.
The role of information channels is not merely “knowing more.” It concerns whether a family knows the deadline, what kind of filing is needed, which certificate can be supplemented, and when it must go to the service window. The ability to take time off determines whether that information can become actual action. When work arrangements conflict with document deadlines, the same procedural requirement produces different time costs for different families. The availability of a community certificate also shows that a procedure consists of more than formal rules. How rules are explained, how documents are verified, and whose cooperation is needed can make the same fact of residence appear with different degrees of visibility. Registration with the subdistrict administration, a residence permit, rental filing, and school places each have their own requirements. A family must make the necessary connections among them before it can turn its living arrangements into eligibility for admission.
IV. Wedding and Funeral Cash Gifts, Marriage Pressure, and Workplace Injury: Tracing Relations and Consequences
Assigning wedding and funeral cash gifts to “the market plus culture” can do no more than mark the two ends of the problem. Amounts, income, return gifts, and burdens affect a family; wedding and funeral exchanges, social respectability, the order between elders and juniors, and local evaluation are also involved. Yet the same five hundred yuan is not the same thing in different relations. It may be ordinary reciprocity, or it may be understood as estrangement, compensation, or pressure. The difference lies not only in the amount, but also in who attended, who was absent, how past exchanges left records, and how local reputation was preserved. Attendance records and return gifts in gift-giving are not fixed moral scales. They become conditions of pressure or mutual aid because local relations continuously preserve, interpret, and respond to past exchanges. Whether the parents still live locally, whether the person has not returned home for years, whether help is expected later, and whether the person will still need mutual aid in the future can all change how the same payment is understood. A name cannot make the final judgment for these relations.
The pressure of income and return gifts cannot be erased, and local reputation and future mutual aid cannot be reduced to an equivalent measure and then treated as nonexistent. One person loses room to maneuver because of an expense, while another receives support through past exchanges; the same relation can bring burden and help at the same time. Handing all of this over to the names of the market or culture only mixes different consequences into one conclusion.
“Culture plus the family” gathers several sources of marriage pressure, but it cannot show how that pressure changes with age, health, care, and work conditions. Age milestones, the parents’ health, anxiety about old age, employment stability, the cost of having children, economic dependence, and relatives’ evaluations can make the pressure change direction at different moments. Age, employment stability, the cost of having children, and parental health in marriage pressure do not simply add up. When the parents’ health begins to deteriorate, marriage pressure is no longer merely pressure at the level of ideas. Changes in parental health may bring care and anxiety about old age closer together; care, old-age support, family continuity, and the taking up of economic responsibilities may enter the conversation together. This does not mean that all families will be like this. It means that the sources of pressure cannot be closed in advance by the two names “the family” or “culture.” Economic dependence may also make refusal difficult; when work is unstable, marriage arrangements are not necessarily only a private choice. Calling all these pressures culture or family merely removes from view the conditions that can be questioned separately.
In the case of workplace injury, the summary “class plus the state” is still too coarse. Contracts, medical records, surveillance footage, testimony from coworkers, training records, the employer’s attitude, the assessment procedure, advance payment of medical expenses, family savings, and waiting time may all change whether a remedy can be carried through. Labor relations and institutional procedures are present, but they do not automatically produce evidence or make it possible for a family to endure the wait. Whether a contract exists, whether medical records are complete, whether footage is preserved, and whether coworkers are willing to testify each connects to different conditions of recognition. In workplace-injury remedies, training records and the employer’s attitude may affect whether responsibility is recognized, while advance payment of medical expenses and family savings affect whether the wait can be endured. When the assessment procedure is prolonged, the relations among evidence preservation, coworker testimony, and family resources become clearer. These differences do not deny the importance of labor relations; they show how labor relations produce asymmetrical consequences within concrete procedures.
All three sets of circumstances point to concrete relations and their consequences. Differences are not a flat list of factors. What matters for analysis is when they acquire operative weight, whether they can be recognized, and whether they connect to a consequence. Consequences are often not decided once and for all: exchanges continue after a gift is paid, marriage-pressure conversations change with health, care, and work, and a workplace-injury remedy passes through evidence collection, assessment, and waiting. Naming an event with a grand concept at only one point in time easily misses how relations reallocate pressure and the support that sustains people through the subsequent process.
Concrete tracing therefore also prevents the sources of pressure from being reduced to mere individual choice. A gift is not only a matter of whether one is willing to give; marriage pressure is not only a matter of whether one is willing to marry; and a workplace injury is not only a matter of whether one is willing to pursue a remedy. Income, records, health, evidence, waiting, and conditions for receiving support constrain how action can take place. Making these constraints clear prevents abstract concepts or personal character from replacing actual relations.
This turn also requires preserving what is unknown. When a concept first acquires the status of an answer, material not yet seen is often treated merely as an exception; returning to concrete differences requires acknowledging that some relations have not yet been clarified and that an existing judgment may change with new material. Primary and secondary status cannot be declared before the material arrives. Retreating from grand concepts to concrete differences can still produce judgments. Analysis can still indicate which relations are provisionally closer to the consequence, but it must explain where that judgment comes from and what material would make it fail. A name provides a starting point; concrete tracing determines whether a judgment can stand.
This turn also places critique at a more concrete point. If one merely says that a grand concept oppresses people, it is often difficult to tell where it can be changed; if one can trace which document, which deadline, which way of taking things up and carrying them forward, or which relation changed the consequence, it becomes possible to distinguish which links require further questioning. This does not reduce responsibility to technical detail. It prevents critique from stopping at a name that cannot be tested. Analysis therefore no longer settles for asking “what factors are there?” but asks “how do differences acquire consequences within relations?” The first question easily produces a long list; the second requires tracing the intersection of material, procedure, time, and the taking up and carrying forward of what is involved. It does not guarantee a simple answer at once, but it prevents explanation from stopping at a parallel list of concepts.
V. The Turn to Differentiating Units: Procedural Recognition, Actual Operation, and Judgments of Operative Weight
To say only that society “has class, the state, the market, culture, and the family” appears comprehensive, but it soon comes to a halt. This way of listing existing names does not explain how a consequence reaches a particular person, or why similar situations produce different results. If analysis keeps moving among several grand terms, it easily treats known concepts as explanations that have already been completed. School admission, gifts, marriage pressure, and workplace injury show that certain concrete differences acquire relative weight within relations and connect to qualifications, positions, actions, and consequences. What this chapter marks with “differentiating unit” is precisely the analytical turn from grand concepts back to these concrete differences and their relational processes.
The turn does not produce another, longer table of factors. If residence permits, contracts, medical records, gift records, and parental health are merely placed side by side, we still have not answered when they are extracted, by what procedure they are recognized, or how they connect to qualifications, waiting, responsibility, or choice. The work of analysis is not to collect as many names as possible, but to find the differences and relations that genuinely change the consequence in a concrete question. Grand concepts can remain entry points: the state points to institutional thresholds, the market to payment and credit, the family to care and dependence, and the nation and culture to memory, rites and customs, and shared responsibility. After the entry point, however, analysis cannot stop at the name. It must return to the concrete material and ask which difference has entered the comparison and where it was taken up and carried forward in a way that changed the result.
When judging operative weight, one cannot equate “recognized by a procedure” with “actually operative”; they are not the same thing. Some conditions are downweighted because they leave no record, yet may still affect whether a person can obtain material, endure waiting, or maintain a relation; some conditions are formally listed but may not change the result in a particular situation. Tracing this gap keeps analysis from merely repeating existing classifications. Judgments of operative weight must also distinguish the direct from the indirect. A condition may not immediately determine a qualification, yet may affect whether the material needed to determine that qualification can be obtained; a relation may not directly produce a remedy, yet may determine whether support is available during the wait. These indirect relations cannot be treated as irrelevant simply because they are absent from the surface rule, nor can they therefore be exaggerated into the sole cause.
Being able to distinguish differences does not mean that all differences are equally important in the current question. If analysis lists every detail as equally decisive, the result is only a fragmented list; if it returns to grand concepts, it flattens the differences again. Between these two errors, one must judge which differences have acquired greater operative weight in the relation at hand. A rental contract may not stand out in ordinary life, but may carry greater weight at school admission because of filing and qualification; records of wedding and funeral exchanges may not ordinarily determine life, but may connect to local reputation and mutual aid in weddings, funerals, and borrowing money; surveillance may ordinarily be only equipment, but after a workplace injury it may connect to evidence and the assessment procedure; parental health may ordinarily look like family background, but may redirect pressure when marriage is urged and old-age care is arranged. These comparisons do not establish a fixed ranking in advance. They require judgment to change with the concrete question. What analysis seeks is not an eternal ranking table, but a checkable judgment about how the current consequence was formed.
Operative weight is not an eternal property carried by a difference, nor is it decided once by a conceptual name. How material operates in a concrete situation depends on the handling by people, organizations, and institutional procedures: someone extracts material in the situation, stipulates which records are valid, and connects certain conditions to a qualification or responsibility. When a procedure, relation, or event changes, a former background may move forward, and material once considered weighty may cease to work; when new material appears, the difference formerly given greater weight may also need to be checked again. A difference with greater operative weight need not remain weightier forever. This correctability is the restriction required to avoid both reducing analysis to a list and organizing it around a governing axis.
The key, then, is not to claim that everything matters, but to make clear what has affected the consequence in this question, what remains only background, and what, though not recognized by the current procedure, may reappear in another relation. This is only a transitional discussion; it does not write relative weights in advance as a complete rule for later chapters. Such questioning also preserves the place of counterexamples: people with similar incomes may end up with different consequences because of housing, medical history, care responsibility, or support available through local relations. This is not a denial of labor and wages. It refuses to let a total image predetermine that every difference can only be read as the same position. Concrete differences can have very great force, but a total image is not thereby exempt from being taken apart.
VI. Dismantling the Class Image and Real Differences
The differences that class theory captures are often persuasive precisely because wages, labor, housing, debt, and risk really do compress people’s choices. Taking the image apart does not erase these pressures, nor does it dilute them with culture or the family. It requires each pressure to be returned to a relation that can be checked: who pays, who waits, who has a buffer, who can exit, and which form of institutional eligibility makes the pressure heavier or changes its direction. Taking the class image apart does not yield a list of differences that can be completed once and for all. Different questions will extract different material, and wages, housing, debt, and responsibility for workplace injury need not always appear in the same combination. Analysis therefore cannot first judge someone through a low-dimensional position; it must check whether the extracted differences are actually connected to the consequence at hand. This is especially relevant to judgments of responsibility. If one summarizes everything through the class image, different restrictions and burdens are easily written in advance as the same situation; returning to concrete relations of wages, housing, debt, procedure, and care makes it possible to see who carries the pressure, where it is shifted, and where it might be changed. This does not reduce the weight of labor and property relations. It refuses to let the image make the final ruling for them.
The class image cannot be supplemented to form a platter of “class plus the state, market, nation, culture, and family,” nor can it be kept at the center with all other relations treated as peripheral patches. Class theory does extract real differences that carry considerable weight in labor, wages, property, debt, housing, risk, resource buffers, employment, and domination. Yet the limited explanatory capacity formed by these differentiating units and by their processes of extraction, weight assignment, connection, and compression cannot change the structure by which class theory organizes concrete differences into a total class image from the outset. The problem is not that class theory was later overused or occasionally claimed a monopoly on explanation in advance. It has never been a safe, ordinary, limited explanation. When material such as labor, wages, and property is extracted, class theory simultaneously compresses, projects, and names it as a total image, making it easy for other differences to be arranged inside that image. The concrete differences extracted can return to a bounded analysis; the class image itself cannot be remade into a safe, ordinary, limited model.
Taking the image apart is not a denial of the weight of wage arrears, insecure housing, debt deadlines, unemployment risk, responsibility for workplace injury, or resource buffers. On the contrary, precisely because these consequences are real, we must ask separately when they become heavier and how they connect to household registration, medical history, care, institutional eligibility, support available through local relations, or the ability to endure waiting. In wage arrears, labor, wages, employment, and resource buffers may carry considerable weight; in waiting for a workplace-injury remedy, labor relations, medical records, procedure, and what the family bears may jointly change the result. Returning these concrete differences to their respective questions does not thereby produce a gentler class theory. The class image is attractive because the relations extracted behind it really do carry considerable weight. That attraction cannot give the image the status of what society fundamentally is, nor exempt interpreters from scrutiny in light of counterexamples, boundaries, and other relations.
VII. Society Is a Differentiation Network: Ontological Boundaries and Continuity with Volume I
The phrase “differentiation network” first points to a real, dynamic structure of social relations: society forms and exists as a differentiation network. At the same time, it preserves a processual requirement: do not treat a social consequence as directly caused by a single name; ask instead how differences enter a result through relations, procedures, the taking up and carrying forward of what is involved, and write-back. It does not provide an all-purpose guiding line that can replace other concepts, nor does it declare differences not yet extracted to be known answers in advance.
Differentiating units cannot be arbitrarily added together into social atoms. Society is a differentiation network: concrete differences are extracted within relations, assigned operative weight, connected, transmitted, and returned to consequences. The differentiation network is a real, dynamic, open, relational social structure situated in relational fields, and it replaces the class entity and the pre-existing class plane. But it is not a fixed, closed, pre-completed, personified ultimate entity that exhausts society; it is not the highest image, nor does it arrange society by itself. Extraction, naming, classification, recognition, and adjudication must still be traced to interpreters, organizations, and institutional procedures.
The name of a nation can be compressed, invoked, or used; criticizing the name of a nation and a plane of identity does not amount to denying a real national community. A real national community may form through long-term interaction, multiple layers of relaying, language, rites and customs, shared memory, shared harms, shared protection, demands of responsibility, experiences of boundaries, and intergenerational write-back. It cannot be reduced to a bloodline myth, a state registration, subjective imagining, a cultural label, a plane of identity, or class occlusion. Criticizing the direct adjudication of a problem through “the nation” as an unexpanded grand term is precisely a way of preventing the real relations of a community from being replaced by a convenient name.
Understanding society as a differentiation network does not promise a convenient total answer. It requires a fresh check in every concrete question: which differences were extracted, which relations gave them operative weight, which material was omitted, and how the result returned to people’s qualifications, responsibilities, and actions. This work is slower, but it prevents complex consequences from being handed over to a name too early. In this sense, rejecting a platter of concepts is not rejecting the use of concepts, nor rejecting the formation of provisional judgments. What it rejects is treating a name as a completed account of society before the relations have been unfolded. A concept can direct us toward material, but the relation between material and consequence must be explained separately.
Society is neither a presupposed axis nor a platter of a few grand concepts. Concepts can provide direction, but they cannot replace tracing how concrete differences enter qualifications, positions, actions, responsibilities, and consequences. Putting concepts side by side only expands the index; returning differences to relations is what begins to explain how consequences form. The differences extracted by class theory—labor, wages, property, debt, housing, and risk—must still be examined in concrete questions. What cannot be retained is the practice of collecting them into a total image of society. Retreating from a platter of concepts to concrete differences does not cancel judgment; it requires every judgment to state which relations support it, which material it has skipped, and under what conditions it would need to be revised.