Introduction: What Social Differentiation Network Theory Aims to Explain
I. Judgment Must Begin with What Is Happening
This book is not about which grand term best sums up society, but about where people should actually begin their judgment when faced with a concrete matter. Someone has not received the wages they are owed; a child cannot enroll in school because the required documents are incomplete; a family suddenly can no longer hold on after a patient is hospitalized; a local relationship changes course because someone gives testimony; a debt approaches its due date; a movement draws people who previously had nothing to do with one another into the same public event. Many things do not initially present themselves as complete images of “class,” “state,” “market,” “nation,” or “family.” In concrete events, people often first encounter materials, money, illness, relationships, responsibilities, records, deadlines, and risks; grand concepts can help people organize these consequences, but they cannot replace tracing the event itself.
This book therefore does not begin with a ready-made name for society. First, look at how a particular matter lands in people's lives: which differences have already entered into consequences, who or what procedure is handling them, how a determination changes eligibility, risk, evaluation, responsibility, and ways out, and how the current handling leaves records that enter later life. Starting from concrete matters does not shrink society into everyday trivia. Only in this way can we know what abstract explanations actually capture and what they leave out.
Starting with grand terms is convenient: it can gather disorderly phenomena into a clear image and quickly bring certain burdensome relationships into view. The trouble begins when a concept is treated as society itself: events then fall into the compartments it has already prepared. A person's wages, labor relations, debt deadlines, housing conditions, household registration documents, physical condition, responsibilities to relatives, local reputation, language, credit history, and institutional status will not carry equal operative weight in the same event. Some differences weigh heavily at this moment, some sink into the background for the time being, and others have not even gained recognition yet. Neat names may be easy to understand, but they can wipe away the event's own texture.
II. Stepping Back from Grand Concepts to Concrete Differences
What this book calls a differentiating unit is a unit of difference in the world that can be distinguished, represented, subjected to weight assignment, and brought into comparison, and that may enter into consequences within a concrete relational field. It is not society's smallest atom, nor does it break a person down into an endless table of attributes. Its significance lies in reminding interpreters: do not rush to name an event; first identify which differences have already been extracted, which remain in the background, which are subject to occlusion, and which are changing a person's position, eligibility, risk, evaluation, responsibility, action, or consequences.
When wages are not paid on time, we cannot leave it at “low income”; the matter also connects to contracts, working-time records, contracting relationships, debt deadlines, rent, whether fellow workers can testify, and whether the case will be accepted for arbitration. Whether a child can enroll in school is likewise not just an education problem; whether actual residence can be recognized through documents, whether the landlord has registered, whether the enrollment deadline is approaching, and whether the family has the time and cash flow to navigate the procedures can all change the outcome. These observations are not meant to establish a universal checklist. Each matter engages different differences; what truly needs to be traced is how they enter into consequences in the event at hand.
The “relational field” here is neither a background nor a scene casually sketched by an interpreter. It refers to the concrete state of relations in which an event is already situated and which is causing certain differences to be extracted and subjected to weight assignment. Unpaid wages, school enrollment, seeking medical care, dividing a family household, migration, registration, giving testimony, transactions, and public conflicts bring different materials, relationships, and records to the fore. Discussing whether a difference matters apart from its relational field makes it easy to mistake abstract prominence for weight in reality.
A relational field will not have answers ready for the interpreter. It only lets certain questions surface first: which document is being required, which relationship can provide testimony, which record is being recognized, which deadline is approaching, who consequently has more to shoulder, and who loses choices. Interpreters cannot take in every possible difference at once; they can only follow the consequences that have already become apparent to determine which parts currently need to be traced. This limited scope must be adjusted as materials, nodes, and consequences change; it must be neither so narrow that it misses relationships already at work nor so broad that it crams everything into a single analysis.
Weight addresses precisely this problem. It is not a fixed ranking, much less a way of declaring a root cause for society in advance. A difference's weight consists in how strongly it can enter into real consequences within a concrete relational field. Household registration may not become an issue when buying groceries day to day, yet may suddenly gain weight in school enrollment, medical care, housing, and eligibility determinations; a past negative review ordinarily has no conspicuous effect, but once a procedure extracts it again, it may change subsequent opportunities. Weight requires the interpreter to explain: why this difference gains weight at this moment, what steps it has passed through, and where it will transmit the consequences.
Weight also requires comparison. Why do two apparently similar events produce different outcomes? When a family is just as short of money, why can one document be deemed valid while another is sent back; when two people are both injured, why can the fact of one person's labor enter into a determination, while the other's pain can go no further than a private account? Comparison does not automatically identify a single cause, but it can force explanation away from empty generalizations and back to which differences actually changed the consequences. A counterexample is not a nuisance; it makes the original judgment aware of its own boundaries and gives differences that have been downplayed a chance to reenter the analysis.
III. How Differences Enter into Social Consequences
A differentiating unit does not enter into consequences on its own. It needs to be extracted, undergo weight assignment, enter concrete nodes and relational chains, and, after processing, leave traces that can be extracted again. Extraction is not arbitrary selection. A difference may be pushed into current consequences by an event, procedure, relationship, or record; it may also enter a judgment because the interpreter frames a question and chooses an entry point for analysis. In both cases, the basis for extraction must be explained and subjected to checks against consequences, comparisons, and counterexamples. Weight assignment, too, is not an abstract statement that something is important; it is when a difference begins to change eligibility, responsibility, risk, evaluation, action space, or available fallback options.
Nodes are locations, mechanisms, or carriers where differences are stored, given definite form, formatted, identified, extracted, certified, passed on, executed, or subjected to write-back. Documentation requirements, files, contracts, schools, hospitals, judgments of responsibility within families, local testimony, administrative procedures, and other places where processing actually takes place can all become nodes. A node is not an abstract dot, nor need it be an individual with a will; its role becomes apparent through what, concretely, it stores, rejects, recognizes, or passes on.
Some consequences become more severe precisely because differences have not been successfully extracted. If the fact of actually living somewhere cannot be converted into recognized material, that does not mean the person did not live there; if an injured body lacks records that can enter a determination, that does not mean the pain does not exist. The problem is that certain nodes handle only particular forms of difference, so facts of lived experience are rewritten, downplayed, or set aside as they enter a procedure. This gap cannot be fully accounted for by simply saying, “That is how the system works.” We must also examine who preserves the materials, how certification is obtained, who bears the consequences after a refusal, and whether people can reconnect to the processing chain through other relations.
Relational chains, in turn, reveal how differences undergo transmission, recasting, branching, amplification, attenuation, or recirculation between nodes. A case of unpaid wages may move from an employment relation into debt deadlines, then into rent and household expenses; if a piece of material cannot be accepted as valid, the consequences may extend from a single service counter to school enrollment, residence, and the next eligibility determination. A relational chain is not simple linear causation. It can break off or take a detour, and it may reconnect at another node. Merely saying that “some structure caused everything” often amounts to skipping precisely these intermediate processes.
The historical index and write-back keep explanation from stopping at what is immediately before us. Past differences do not govern the present on their own; they must re-enter the present through family memory, archives, institutional records, local reputation, community commemorations, or other nodes. Nor is write-back ordinary archiving. When the result of a determination, recording, handling, or action is preserved and can later alter eligibility, risk, evaluation, and action space anew, that is write-back. It shows that people do not bear only the consequences of the present in the present; what is written down today may also become a difference extracted anew tomorrow.
These concepts are not a set of procedures to be imposed on everything. Their functions differ: the relational field grounds judgment in the state of relations currently unfolding; weight keeps judgment from placing all differences on the same plane; nodes reveal how differences are handled; relational chains allow people to trace how consequences flow; the historical index and write-back reveal how time enters positions and choices. Together, they serve not to explain society in its entirety, but to enable a limited judgment to make its own grounds clear.
If an explanation retains only the endpoint, it will quickly erase the intermediate processes: unpaid wages become proof of some overall condition, blocked school enrollment becomes an example of some label, and local testimony becomes an incidental phenomenon. Tracing the process does not negate broader pressures; it allows people to know where those pressures take effect and where they are amplified, blocked, or redirected. This is also why this book does not understand comparison as selecting the one and only answer. Comparison means subjecting different explanations to the test of the same event, to see which consequences each can make clear and where each begins to leave parts it cannot handle.
IV. The Differentiation Network Is Not a New Total Image
In concrete relational fields, differentiating units undergo extraction, weight assignment, connection, transmission, and write-back, forming a constantly changing differentiation-network order. This is the book’s core judgment. It does not mean that a new entity called a “network” lies hidden behind society, nor does it seek to replace the original big picture with an even bigger one. What the differentiation network points to is the order through which differences enter into consequences by way of nodes and relational chains; this order overlaps, branches, flows back, and fractures, and also reconnects in new events.
This book therefore cannot be reduced to an ordinary list of factors. Merely listing income, household registration, education, family, locality, ethnicity, the state, the market, and platforms one by one does not explain how they act. Conversely, presenting the differentiation network as the sole truth would likewise commit the error of the total image. This book puts forward a core theoretical judgment about how society takes shape in concrete relational fields and enters into consequences; it neither compresses everything back onto a single axis nor excludes other explanations that can explain events within their effective scope.
The state, the market, the nation, the family, locality, schools, platforms, and community can all be used, but each must be brought back to its own traceable mechanisms. The state involves identity documents, files, registration, certification, mutual recognition, enforcement, and write-back; the market involves prices, contracts, credit, deadlines, liquidity, debt, and risk shifting; the family preserves care, property, responsibility, memory, silence, favoritism, harm, and long-term evaluation; local familiarity, reputation, testimony, and spatial nodes can constitute real recognition networks; the national community, too, cannot be reduced to a myth of bloodline, a registered name, subjective imagination, or class occlusion. It may form through long-term interaction, language, rites and customs, shared memory, shared harm and protection, demands of responsibility, experiences of boundaries, and intergenerational write-back. Breaking these concepts back down into mechanisms does not abolish them; it requires them to submit to comparison and calibration.
V. The Power and Boundaries of the Class Image
Labor, wages, employment, property, debt, housing, risk, relations of domination, and the width of available exits are all real differences, and they may carry great weight. Any judgment about society that ignores them will lose its sense of reality. Class theory’s appeal as an explanation within a limited scope lies precisely in its extraction of one set of high-weight differentiating units from among them. The question has never been whether these differences exist, but whether, after they undergo low-dimensional compression, projection, and naming, the class image is treated as a social entity, a natural community, or what society fundamentally is.
The class image is not an actor. Acts of compression, subsumption, demotion, or monopolizing explanation can be attributed only to class theory, interpreters, organizations, institutionalized interpretive procedures, or concrete handling mechanisms. When class theory elevates the results of its own distinction extraction into society itself and demotes family, locality, nation, the private sphere, community, the body, records, and other valid explanations to subsidiary material, what it produces is not an orientation that can be compared on an equal footing with other limited explanations. Class theory is precisely Model C: total-image subsumption, not Model A or Model B; the explanatory capacity within a limited scope generated by the real differentiating units mentioned above and by the processes of their extraction, weight assignment, connection, and compression cannot change this basic structure.
This judgment does not require readers to deny oppression, unpaid wages, workplace injuries, poverty, debt, or risk shifting. Precisely because these consequences are indeed so heavy, it is all the more necessary to ask how they pass through concrete nodes and relational chains and intertwine with family responsibilities, local recognition, institutional eligibility, national community, private life, and historical records. If all problems are first translated into the class image, many differences already acting in events will be downplayed, leaving only an image that looks clear but is in fact far too flat.
VI. How Limited Explanations Remain Open to Calibration
Being limited does not mean being invalid. Any explanation can extract only some differences, and every explanation leaves some parts unaddressed for the time being. Acknowledging this limit is healthy explanatory humility: it requires an explanation to state what it has seen, what it has downplayed, in which events it is valid, and which counterexamples would force it to revise itself. This humility preserves comparison; it does not abolish judgment.
Explanation can also run out of control at the other extreme. One form of this loss of control elevates a limited image produced through distinction extraction into society itself, leaving no place for other valid explanations; another illicitly substitutes the claim that society cannot be effectively explained for the fact that society cannot be exhausted and explanation is necessarily limited, thus placing models with explanatory power, limited but revisable models, erroneous models, and utterly groundless claims on the same level. The former is total-image subsumption; the latter is the erasure of distinctions between explanations. They move in opposite directions, yet both destroy the conditions for comparison, revision, and calibration.
Complexity itself is no reason to refuse judgment. Events may indeed involve multiple sets of differences, multiple nodes, multiple relational chains, and records left at different times; this calls for more detailed explanation, not for declaring everything equally important and equally impossible to compare. Only when “too complex” is used to cancel valid comparison, revision in light of counterexamples, and explanatory calibration does complexity-based conflation constitute Model C’s mechanism for erasing distinctions between explanations. Acknowledging omissions, boundaries, and counterexamples is not this mechanism; on the contrary, these are the conditions that allow limited judgments to remain revisable.
VII. The Book’s Course of Work
This book does not provide readers with an answer they can immediately apply to everything. It first addresses why grand concepts are attractive and why they so easily exceed their own valid scope; it then clarifies, step by step, how differences undergo extraction, weight assignment, preservation, transmission, comparison, and write-back, and examines their different modes of operation in communities, institutions, history, and various concrete domains. The discussions that follow do not ask readers to accept a closed system, but require every explanation to return to events: what it has extracted, what it has omitted, at which nodes it holds, along which relational chains it is transmitted, and in the face of which counterexamples it needs adjustment.
This course also defines the limits of the book’s use. Social Differentiation Network Theory cannot replace concrete historical research, institutional analysis, legal judgment, economic analysis, or people’s own experience; nor can it recast all this work as material subordinate to itself. What it requires is that, before making any broad judgment about society, we do not let concepts flatten events in advance. Judgments may be limited, but they must be traceable, comparable, and revisable.
The episodes involving school admission, workplace injuries, platform accounts, families, and local life used repeatedly in this book are composite scenarios that bring multiple verifiable mechanisms into a single analytical setting. Unless explicitly stated otherwise, they neither refer to a single real case nor imply that different regions, years, institutions, or platforms share the same lists of materials, procedures, and consequences. Specific institutional facts remain bounded by the relevant jurisdictions, periods, and sources.
Society is not constituted in advance by grand concepts such as class, state, market, nation, and family. In concrete relational fields, differentiating units undergo extraction, weight assignment, connection, transmission, and write-back, forming a constantly changing differentiation-network order. This statement is neither the endpoint of the book nor another total image; it is the direction this book must hold to as it begins to make judgments.