Chapter 22. Why Theory Must Simplify
I. Simplification Is the Starting Point of Judgment
When judging a job, a person cannot unfold their entire life, their whole family, all their local relations, all their historical indices, and all their forms of eligibility within state systems at once. Someone looking for work carries more than a question of wages: there are also age, medical history, elders at home, children's schooling, rent, debt, household registration, local relations, past employment, whether their body can cope with staying up at night, and whether they have a fallback. A school receiving a child likewise cannot fully understand every aspect of the family's circumstances. School admission is not simply a question of student registration; behind it lie family migration, the parents' work, residence permits, social insurance, rental housing, household registration, school districts, school capacity, and local institutions.
Still less can a theory unfold every differentiating unit, every weight, every node, and every historical index in a single sentence. This is not because these differences do not exist. They are all there, but judgment must begin. Once judgment begins, it cannot unfold all of reality at once; it must undertake limited distinction extraction. Simplification does not abolish complexity. It is a form of compression necessary for action.
Judgments in everyday life, institutional judgments, and theoretical judgments must all proceed from limited positions. Limited extraction in everyday life lets people address the most pressing responsibilities first. If everyday judgment had to shoulder every relation at once, it could not form an immediate response. If institutional judgment rejected form fields and ordering, it too could not process materials. Rules, form fields, materials, and processing deadlines bring particular differences into judgment to be identified first, while other differences remain temporarily at the periphery. Without concepts and images, theoretical judgment could not even compare different consequences. All three need a limited beginning, but such a beginning authorizes no judgment to replace the rest of reality. All three kinds of judgment may omit important relations.
Volume VI addresses erroneous explanation, but it cannot begin by opposing theory. If simplification is treated as an error from the outset, every explanation will be struck down, leaving only the lament that “society is very complex.” That is not theoretical clarity but the abandonment of judgment. Conversely, if simplification is treated as all-powerful, and the set of differences one has extracted is taken to be the whole of reality, explanation will move toward illusory construction. The first step in this volume is therefore to establish simplification's legitimate place rather than criticize it: theory must simplify reality, or it cannot judge. Without simplification there is no judgment; without compression there is no transmission; without limited distinction extraction there is no action.
Examining simplification and rejecting theory are two different questions. Examining how institutions extract materials does not amount to denying that institutions need simplification. To examine simplification is to ask how these selections and omissions came about, whether the extracted differences still correspond to current consequences, when the simplification ceases to work, and whether a return remains possible. The test of a theory is not whether it needs compression. Examination must instead address its particular ways of extracting: whether it knows how it compresses, what it extracts, what it downweights, and whether it still retains a channel for unfolding complex reality again. Rejecting theory cancels these questions as well, pushing every judgment back to its starting point.
II. Limited Distinction Extraction Is Not a Copy of Reality
Judgment does not copy reality; it undertakes limited distinction extraction within a concrete relational field. Take the same family of a worker from elsewhere. The differences extracted in judgment vary completely with the events the family faces. When wages go unpaid, those concerned and those intervening do not begin by unfolding the family's entire migration history, marital relations, and local reputation. They first extract wages, working hours, the labor contract, the timing of wage arrears, debt, rent, and work risks.
When the occasion is a child's school admission, judgment does not begin by asking what historical indices of kinship the family has back home. It first extracts household registration, residence permits, social insurance, student registration, length of residence, the school district, and available school places. If the question becomes whether the family can gain local recognition, the focus of differentiation shifts to accent, ties with people from the same place of origin, length of residence, exchanges with neighbors, participation in weddings and funerals, and local rites and customs. In relationships and marriage, family reputation, medical history, kinship responsibilities, local historical indices, and past disputes may suddenly gain weight. None of these extractions captures the family's entire reality. Yet any concrete judgment, if it is to be formed, must first extract some differences from complex reality.
Society is not a plane that can be seen all at once, nor is an individual's social position determined by one label. The same person may be a worker, a person from elsewhere, a father, a patient, or a debtor; someone from a particular village or within a particular national community; someone outside a particular school system, a descendant within a particular family historical index, or someone with or without documents in a particular state record. These identities are not decorations. They all gain weight in different relational fields: wages in labor disputes; household registration when a child enters school; accent in local recognition; medical history in employment and in relationships and marriage; kinship responsibilities in borrowing money, elder care, and household division; national memory in boundary conflicts, commemoration, and shared responsibility; and historical indices in new distributions, disputes, and recognition.
The presentation of differences is also constrained by the positions from which materials can be obtained. Some people can present formal proof; others can preserve only oral accounts or memories within relations. Materials not yet presented have not entered the current image; this does not mean they have been judged nonexistent. If an explanation were required to unfold all differences at once, it could not begin. If an institution were required to understand a person's entire life at once, it could not operate. If a family were required to reinvestigate its entire history whenever borrowing money, considering relationships and marriage, or arranging elder care, everyday family judgment would also come to a halt. The abundance of reality is invoked to explain why judgment must be limited, not to cancel judgment. Limitation is not an error; it is a condition for judgment to occur.
Limited distinction extraction does not require everyday judgment to become a complete theory. In making a provisional judgment, people can refrain from rewriting unprocessed differences as an absence of differences, while retaining a position from which to look back and re-examine them later. Materials appear at different times, people's attention is occupied by current responsibilities, and procedures arrange different materials at different positions. Three states must be distinguished here. Not yet extracted means that a difference has not entered the image immediately before us. Temporarily downweighted means that the difference has been seen but has not acquired relatively high weight at this moment. Denying its existence cancels its standing to re-enter the field of view. Only by distinguishing these three can we avoid mistaking the limits of our own view for the boundaries of reality without abandoning judgment.
Processing capacity also limits how many relational chains can be maintained at once. When materials exceed present capacity, some differences recede temporarily, but their positions of retreat can still be recorded. When consequences show that the original ordering is no longer appropriate, judgment can rearrange priorities. This is fundamentally different from declaring that the differences which receded never existed. Distinction extraction does not copy the whole of reality, because every image leaves some parts unprocessed. Nor is it arbitrary selection: the extracted differences must still be capable of being brought back to the question, relational positions, and actual consequences for examination.
The questioner does not stand outside relations and cut them up at will. The question faced, the materials available, the position occupied, and the consequences to be borne constrain the direction of extraction. People within families can form judgments through care and responsibility; people within local relations can form judgments through exchanges and recognition. Everyday life need not wait for formal procedures before distinction extraction begins. Non-arbitrariness comes from examinable connections between extraction and the question, materials, and consequences, not from approval by a single procedure. Everyday judgments within families, localities, or communities can also form examinable connections. They do not have to be converted into institutional materials before they can be non-arbitrary.
Limited extraction within institutions can likewise be compared, because different ways of handling things leave different consequences. Comparing those consequences does not require an institution to shoulder the whole of life at once. It examines whether a particular extraction has hidden the positions bearing the results. Extracted materials are temporarily presented in a working image along a particular direction, enabling the questioner to judge the next action. The scope within which that image is effective must be bounded by the question and consequences faced. How materials are seen and connected affects the direction of the working image. This does not require every judgment first to gain recognition through the same kind of procedure; it requires those judging to explain their position of extraction and scope of application.
Users of theory, too, can undertake only limited distinction extraction. A working image can be formed along axes of labor, income, property, debt, housing, risk, and relations of domination. Working images addressing other questions also form along axes of documents, archives, borders, certification, and implementation, or of prices, rents, wages, consumption, and competition. None of these images is a copy of reality. Users of theory must still explain which differentiating units they extracted, which differences they downweighted, and precisely which concrete questions the image applies to.
III. Relational Fields and Weights Bound the Direction of Simplification
Simplification does not arbitrarily reduce reality, nor does it mean theory sitting on high and deciding in advance what will always matter and what will never matter. A relational field first presents a demand for judgment. A relational field describes a concrete combination of questions, positions, materials, and consequences; it must not be written as an independent personality issuing demands. The field itself neither chooses nor acts for anyone. Different explanations extract different differences from the same family of a worker from elsewhere because the questions themselves differ, not because interpreters can choose materials at will. The relational field first presents a demand for judgment; only then does explanation begin distinction extraction.
These acts of extraction must be traced back to concrete positions: the people concerned present their circumstances, reviewers identify materials according to the procedures in which they operate, and relatives or people within the relations respond through existing forms of recognition. In a question of unpaid wages, the relational field calls for labor, wages, working hours, contracts, and risks to be extracted; the parties to the dispute and those handling it bring earnings and risks into judgment. In school-admission procedures, what must be extracted is household registration, residence permits, social insurance, student registration, and the school district. Both families and schools or reviewers identify eligibility materials.
Local recognition calls for the extraction of accent, exchanges with neighbors, ties with people from the same place of origin, and local rites and customs. Relationships and marriage call for family reputation, medical history, kinship responsibilities, and local historical indices; those concerned, together with relatives or people within the relations, reassess exchanges and responsibilities. Historical migration calls for ancestral place of origin, migration routes, old family accounts, burial grounds, and local memory.
Without asking about the relational field, theory will extract the differences most familiar to it everywhere: economic oppression whenever it sees a family, forms of eligibility within state systems whenever it sees a school, class displacement whenever it sees a nation, and narratives of victimization whenever it sees history. Such explanation may look clear, but it is already being used in the wrong relational field. When this happens, specific users have moved a familiar image to an ill-matched position; no abstract environment has made a wrong choice on its own. Different relational fields extract different differences from the same object, and the same set of differences carries different weights in different relational fields. Different fields give simplification different directions. Without attention to the relational field, simplification turns from a tool into occlusion. Misplacing materials need not change the materials themselves, but it changes the consequences to which they are connected. Household registration can acquire weight in school-admission eligibility, but when moved into the handling of unpaid wages it cannot replace working hours and earnings. Family memory can affect evaluations in relationships and marriage, but it cannot automatically acquire the same position when moved to other questions.
The existence of differences, their extraction, and their acquisition of weight are continuous but distinct levels. Existence means that differences are already present in the world. Extraction means that concrete judgment brings differences into the foreground. Weight assignment concerns the intensity with which differences enter into consequences within current relations. If these three levels are collapsed, seeing materials will be mistaken for their always being important. Separating them makes it possible to explain why the same difference carries different weight in different relational fields. A difference may exist for a long time without necessarily being presented in every judgment. Wages can exist in life and can be brought into a dispute by those concerned, yet even after being presented, a difference may still not have entered the current relational chain. The existence of materials does not guarantee that they can change eligibility, responsibility, or evaluation. Limits arising from procedural positions are not abstract defects. They explain who can present materials, who can respond to them, and which differences will change form in transmission. These differences of position can offer clues for later revision. The visibility of materials also changes over time. Differences not presented earlier may become entry points for renewed judgment after results have formed. Temporary downweighting is therefore not permanent exclusion.
Extracting differences is only the first step. Explanation must also judge their relative weight and assign weight to them within concrete relational fields. There is no table of weights fixed in advance, nor any single threshold through which differences acquire weight. Weight must always be observed within concrete relational fields. Wages, household registration, accent, medical history, kinship responsibilities, local relations, national memory, and historical indices may all coexist in the same family of a worker from elsewhere, but they will not be equally important in every relational field.
Weight assignment has different points of emphasis in concrete situations. When wages go unpaid, wages, working hours, labor risks, debt, and rent weigh more. When a child enters school, household registration, residence permits, social insurance, student registration, and the school district weigh more. In local recognition, accent, ties with people from the same place of origin, length of residence, and exchanges with neighbors weigh more. In relationships and marriage, family reputation, medical history, kinship responsibilities, and local historical indices may gain weight. In household division and elder care, historical indices, care records, houses, cash-gift ledgers, and things elders have said may gain weight. In boundary conflicts, commemoration, and shared responsibility, national memory may gain weight.
This weight assignment is not arranged at will; it arises from the demands of the relational field. Wages gain weight in labor disputes because the current question calls for judgments about labor earnings, the bearing of risk, and fallbacks in life. This does not mean wages always override everything. Household registration gains weight in school-admission eligibility because current institutional nodes are extracting household registration, student registration, and residence eligibility, not because household registration is always more real than wages. Family historical indices gain weight in relationships and marriage because new relations are entering an old kinship network, triggering past differences, not because historical indices always govern the family. The relative weight of wages, household registration, or family memory depends on how the current question connects them to fallbacks in life, access through eligibility, or evaluations of responsibility. A material's gaining weight in another relational field does not mean it can govern every question forever.
Dangerous explanation often begins not with extracting the wrong differences, but with mistaking weight assignment within one relational field for eternal weight. Real differentiating units such as labor, debt, housing, and risk, together with their extraction, weight assignment, connection, and compression, may form powerful explanations within a limited scope on relevant questions. But if interpreters using class theory turn such explanations within a limited scope directly into the highest weight in every relational field, they will occlude nation, locality, family, institutional eligibility, and historical indices. A state image may be powerful on questions of documents, archives, borders, and implementation. Yet if it is turned directly into the supreme explanation of all society, it will occlude family, markets, localities, communities, and historical memory. Healthy explanation must recognize that weights form within concrete relational fields; they are not an eternal underlying reality.
IV. Explanation Must Connect Differences into Relational Images
Listing wages, household registration, accent, medical history, kinship responsibilities, local relations, national memory, and historical indices is not explanation. It merely heaps society back into a pile of materials. A list can enumerate differences such as wages, debt, housing, and family fallbacks; multifactor analysis likewise only places differences side by side. Social Differentiation Network Theory asks how these differences connect to form positions and consequences. Explanation means finding how differences connect, rather than listing more factors. A relational image must return differences to nodes, transitions between nodes, and relational chains, tracing how they jointly change positions.
When attention is limited, distinction extraction creates an order of priority, but that order is not an eternal ranking of value. A difference not given priority now may still become decisive in another question. Listing wages, debt, housing, and family fallbacks together means only that the materials are visible at the same time. Explanation also requires tracing how wages are determined at employment nodes, how debt changes the risk that can be borne, how housing constrains action, and how family fallbacks ease or worsen consequences. The role of counterexamples is limited to requiring a fresh look at whether a connection holds or a node has been omitted. They are not to be made into a new principle governing every explanation.
Positions of transition between nodes may also change the form of materials. A material may first be evaluated within a family and then enter institutional procedures, or it may first be fixed in an archive and then return to the action space of those concerned. Once old family accounts enter a record, some relational evaluations become fixed while other background recedes. When institutional results return to a family, they may become new grounds for responsibility. Different sequences of nodes and different ways of recasting may produce different consequences. If explanation misses these transitions between nodes, it can hardly show how differences actually enter into consequences for action.
Only when wages, debt, housing, fallbacks, and risk are connected do they form an image for judging a labor situation. Someone with low wages but no debt, no rent to pay, and family fallback support is not in the same situation as someone on equally low wages who carries debt, rents, supports a sick person, and has no fallback. Only when household registration, residence permits, social insurance, student registration, and school districts are connected do they form an image for judging school-admission eligibility. A child does not abstractly “want to go to school.” The child must enter a set of institutional nodes: where they live, what their parents have provided, whether their documents are continuous, how school districts are drawn, and how places are allocated.
Only when language, marriage, place names, shared memory, and shared consequences are connected do they form an image for judging national networks. A national community is not an isolated label; it consists of paths of recognition, boundaries, and responsibility transmitted over time through partially overlapping networks. The same holds for historical images. A place name, a grave, an old account, or an archival document does not act in isolation. They must be triggered, interpreted, and contested within a new relational field before they re-enter judgment. Explanation does not place differences side by side; it explains how those differences connect to form consequences. This is also why this book opposes multifactor collage. Society is not a few factors added together. Differences are extracted, assigned weight, and connected within relational fields, forming consequences for action.
V. Compression Makes Transmission of Materials Possible
A relation of distinction extraction exists between the questioner and the materials. A question does not arise from materials naturally arranging themselves, yet neither can the questioner invent it at will. Subsequent actual consequences can be used to examine whether this distinction extraction holds up. In this process, different carriers undertake compression in different directions.
As entry points for identifying eligibility, documents point to eligibility that can be identified. Records preserve the traces of procedures and the results of handling that were once written into them. Unlike these two, family accounts preserve relational evaluations that can be retold and carried forward. Categories facilitate comparison across positions, while concepts gather dispersed differences into forms that can be discussed, laying out common threads for comparison.
Every kind of carrier downweights some of what was present at the scene, but such downweighting does not necessarily mean error. Concrete judgment requires entry points appropriate to the current question. One must neither demand that a single carrier encompass everything carried by other carriers, nor neglect to examine whether the downweighted contents affect the current question and whether other entry points remain through which they can be unfolded again.
After limited distinction extraction, weight assignment within a relational field, and the connecting and combining of differences, explanation must still compress. Without compression, judgment cannot be transmitted. Compression lets complex reality leave the scene of its occurrence and enter forms that can be transmitted, compared, and judged. Without compression, the past cannot be transmitted, institutions cannot operate, and theory cannot be expressed. Family historical indices must be compressed into a few sentences to be passed to the next generation: “The eldest took more that year”; “Her family contributed money back then”; “That branch of the family did not look after the elders.” These statements are not the whole past, but they allow past differences to move across generations.
Local history must be compressed into place names, inscriptions, reputation, and stories to circulate. State records must be compressed into archives, documents, and categories to be mutually recognized. Schools must compress complex families into student registration, household registration, social insurance, and eligibility materials. Theory must compress complex reality into concepts, sentences, images for judgment, and accounts before it can be discussed.
Concepts also perform compression in transmission. They make differences comparable and capable of being passed on, but do not preserve together with them the positions where things happened, the relational chains, and what remains unsaid. Using a concept still requires returning to what it extracted.
Yet compression necessarily sacrifices detail. When a family historical index is compressed into one sentence, care, illness and debt, work away from home, differences of sex, and elders' favoritism may be downweighted. When local history is compressed into a place name, mixed settlement, exchange, intermarriage, and conflict during migration may be downweighted. When state records are compressed into a category, people's concrete circumstances may be downweighted. When theory compresses reality into an image for judgment, differences that have not entered the image also recede. Compression itself is therefore not the error. The error comes when compression is forgotten after it has occurred. Information loss does not automatically prove a judgment wrong. What matters is to examine whether this loss obscures precisely the crucial differences in the current consequences, and whether other materials still preserve an entry point for return.
VI. Limited Explanations Can Still Be Compared
Not exhausting reality does not mean that all explanations fail equally. This must be made clear in advance, or we will later encounter an illicit substitution: since every explanation is limited, every explanation fails; since every explanation fails, effective and ineffective explanations are indistinguishable. This is the error of Model C: it substitutes “all fail” for “all are limited.” Here we establish the distinction without developing Model C, which will receive dedicated treatment later. Acknowledging limitation serves to enable more precise comparison, not to cancel comparison. Limited, erroneous, and ineffective are therefore not synonyms. Still less can limitation entail that all accounts have the same value. Ordinary limited explanations can retain explanatory power provided they state their scope, accept counterexamples, and permit comparison.
The provisional character of a working image is not weakness. It allows the image to enter concrete comparison with other materials and consequences. The more clearly the image specifies its scope, the easier it is to know when it needs reopening. Ordinary limited explanations can be compared by what they extract. What they extract, what they downweight, and whether this is proportionate to the weights of the current question are all differences that can be examined concretely. Accuracy in identifying materials is one measure of comparison here. Accuracy does not mean piling up quantities of material within an image, but whether the extracted materials can explain the consequences now occurring. Whether nodes and relational chains are seen affects an explanation's accuracy. Merely listing materials without tracing their transitions between nodes, leaving crucial nodes and transitions out of the image, may yield an image whose wording is clear but which cannot explain consequences.
Fit with the relational field is another important measure of comparison. Fit is not an abstract verdict of good or bad; it concerns whether an image is placed in an appropriate relational field. If the same image explains consequences in one question but occludes relations when moved to another, the results in both fields are not thereby equally effective. Connected with this is the capacity to revise in response to counterexamples. Such capacity does not mean changing one's account at will. It concerns whether interpreters acknowledge difficulties when contrary materials appear, whether they can adjust the scope, and whether they specify which weights, nodes, or connections need re-examination. An explanation can be compared with others as long as it states its scope and makes clear what it extracts, what it downweights, the relational fields in which it is effective, and whether it can face counterexamples and permit revision. A limited explanation is not an ineffective explanation.
Under these explicit constraints, ordinary limited explanations taking different approaches can indeed be compared with one another. For example, interpreters using a state image can trace documents, archives, borders, certification, and implementation. Interpreters using a market image can trace prices, rents, wages, consumption, and competition. Interpreters using a historical image can trace how past differences are preserved, triggered, and assigned weight anew. These ordinary limited explanations can be compared using the measures just stated, rather than relying on closed systems that cannot be compared.
Class theory does not belong among these ordinary limited models. Real differentiating units such as labor, income, property, debt, housing, risk, and capacity for domination, together with the processes of extracting, assigning weight to, connecting, and compressing them, can form partial explanatory power. An image can also present certain real circumstances, but this presentation of circumstances is not attributable to any autonomous explanatory power of the image itself. The class image is only a low-dimensional output of this process. It cannot on this account be placed alongside Model A and Model B as a limited explanation.
VII. The Same Object Requires Multiple Limited Images
The same person or family exists simultaneously within relations of labor, institutions, kinship, locality, and history. Different questions give different differences within these relations greater weight. Faced with the same family of a worker from elsewhere, someone handling unpaid wages can first trace wages, working hours, debt, housing, and risk. Someone handling the children's school admission must trace household registration, residence permits, social insurance, student registration, and administrative eligibility. When the question concerns local recognition, relationships and marriage, or historical memory, accent, exchanges, kinship responsibilities, migration routes, and historical indices come to the foreground. Each image has its own scope of materials and consequences. One cannot require any image to shoulder all relations at once, still less let any of them decide the answers for every relational field in advance.
Simplification has force because it brings dispersed differences into a single image through which judgment becomes possible. Someone who finds life hard may be unable to say exactly what weighs them down: low wages, high rent, pressing debt, dangerous work, no fallback, dependents at home. When these differences exist separately, they are simply many pressures. Class theory and the interpreters who use it place labor, income, property, debt, housing, risk, and capacity for domination within the same image, enabling people to see a burdensome situation concretely.
Likewise, a family meeting obstacles everywhere in the city may be unable to say where the problem lies: materials are required for the child's school admission, renting cannot prove stability, social insurance is insufficiently continuous, household registration is elsewhere, and the residence permit falls a few months short. Interpreters using an image of state incorporation of differences place documents, archives, borders, certification, mutual recognition, and implementation within the same image. People can then see why they cannot readily pass through the procedures of many institutions.
In local society, the unwillingness of people in a locality to accept marriages involving a particular family may likewise have more than one cause. Accent, old grievances, ancestral place of origin, migration, wedding and funeral exchanges, local reputation, and kinship designations are all at work. Only by connecting these differences can interpreters using an image of local networks explain why the obstacles facing this family cannot be reduced to “low income” or “a different culture.”
And when relations become tense again on a national community's commemorative day, the past has not returned automatically. Only when language, place names, shared harm, shared resistance, education, commemoration, external pressure, and shared responsibility are triggered together do they produce the consequence of hardened boundaries.
Images enable dispersed materials to be named, compared, and transmitted, and allow people at different positions to coordinate action around a consequence. Family accounts can arrange responsibilities; records and documents can support procedural handling; concepts can guide theoretical comparison. This power takes shape through these uses. This does not mean that an image has become an autonomous entity. In concrete operation, comparison makes dispersed circumstances visible on a common scale; transmission lets materials cross their original nodes; coordination enables people at different positions to connect their handling of matters; and action brings judgment back to responsibilities, qualifications, and choices. Each kind of power depends on concrete practice. None proves that the image has revealed an underlying reality that need not be unfolded further. If a successful image is applied repeatedly after leaving its original question, convenience in use may occlude the differences it has not processed.
Multiple limited images of the same object therefore cannot simply be added together, still less directly combined into a total image. Images capable of complementing one another usually each retain a clearly defined partial scope. Interpreters using a labor image can trace earnings and risk; those using an institutional image can trace eligibility and certification; those using a historical image can trace preservation and triggering. If one partial image addresses earnings and risk and another addresses access through eligibility, they can complement one another at their interface and check one another where they connect, rather than either image encompassing the whole.
Differences in weight assignment may also be affected by changes in relations. When new responsibilities, qualifications, or risks appear, materials that once seemed slight may acquire a new position, but this does not make them a fixed highest weight. Different explanations can extract different differences from the same object. The key to theoretical comparison is not which explanation can encompass everything, but which better fits the dominant weights in the current relational field. The error lies in passing off a set of differences as the whole, not in extracting that set. If users all claim that the images they employ grasp every consequence, their judgments may conflict. Conflicting images must return to the question and consequences. If one image elevates weights within a limited scope into the highest weights for every question, differences retained by another image will be downweighted. Merely placing their names side by side in writing cannot remove this conflict.
Even after a difference enters an image, we must ask along which relational chain it acts. If intermediate nodes remain unseen, the distance between differences and consequences will be improperly shortened. When judgments conflict, the solution is to compare materials, relational fields, and relational chains, not to invent another model encompassing every image. Complementarity requires users to acknowledge the partial scopes addressed by the different images. When conflict occurs, they must judge which image has been pushed beyond its scope of application. If labor differences really change school-admission or family consequences, interpreters can trace how they connect across nodes. Without that relational chain, however, two images cannot be combined directly merely because they address the same object.
VIII. Healthy Simplification Must Retain a Return Channel
Transitions between nodes turn a list into an image. Who preserves materials, who recasts them, and who accepts them affect whether they can enter subsequent judgment and how they return to the positions of those concerned. Transmission never moves content unchanged: every passing-on retains some differences and downweights others. Only with this understanding can users of carriers avoid mistaking form for complete reality. The danger of simplification arises not from failing to exhaust every detail, but from cutting off paths to what remains unprocessed. Forgetting boundaries lets a partial image expand outward; denying remaining differences rewrites the unprocessed as nonexistent; reification lets a tool of judgment masquerade as reality itself; using an image in the wrong relational field connects effective materials to erroneous consequences; and excluding counterexamples closes the entry point for renewed examination.
These forms of overstepping have concrete mechanisms in different images that cannot substitute for one another. One danger cannot stand in for the others. The risk of the class image lies in its brutal projection masquerading as underlying reality: differences of labor, income, property, housing, debt, and risk are projected into the sole underlying reality. National-community networks, however, are real historical relations. Their members, families, and those who transmit things across generations, record keepers, educators, organizers, or disseminators also simplify in practice, compressing intricate interactions into names, boundaries, designations, and responsibilities. This simplification is itself useful. Its distinctive danger lies in detaching long-term interaction, shared memory, boundaries, and responsibility from their historical course, then turning them into an essence, a matter of bloodline, and something eternal. This differs in origin from the danger of class theory's ontologizing projection.
Interpreters using state, market, or historical images likewise face different risks of overstepping. Those using a state image extract nodes such as documents, archives, borders, and implementation. Their danger lies in expanding institutional nodes such as certification and implementation into the whole of society and compressing them into a single personality. Users of a market image may expand exchange and prices into all of society, elevating them into the sole governing force. If users write historical materials as causes that automatically govern the present, they too overstep the concrete conditions of preservation, triggering, and renewed extraction. Identifying where these different images begin to overstep serves here to mark their respective limits, rather than to develop their subsequent mechanisms in advance.
Healthy simplification must therefore retain a return channel. A return channel is defined as what allows a compressed image to be reopened. Healthy explanation does not reject simplification; it knows where simplification occurs. It extracts some differentiating units from complex reality, assigns them weight within a concrete relational field, connects them into an image for judgment, and further compresses them into an account that can circulate and be transmitted. A return channel requires interpreters to specify three things together: which differences were extracted, which contents were downweighted, and which relational fields the image applies to. Only when the scope is specified does the position at which later materials can re-enter become clear. Differences that did not enter the image have not disappeared; they may gain weight again in other relational fields. As long as an explanation retains the capacity to reopen its image, it is a limited explanation. Once it cuts off the return channel and treats itself as reality itself, it begins moving toward illusory construction.
When an explanation begins to harden, first ask which differences it extracted and which contents it downweighted, then whether those differences have gained weight in the current relational field. In this examination, overlooked records and relational positions that never entered the procedure may both alter the weights originally assigned and the consequences derived from them. Re-examining records previously given little attention can fill gaps in the materials.
The next questions are which crucial nodes have gone unseen and which relational chains have been cut off. Severed paths of transmission may likewise alter the original weights and consequences. Identifying where materials were preserved and by whom they were recast allows a return to nodes; tracing the process through which consequences cross different positions allows a return to relational chains.
Whether historical indices have been misread, whether counterexamples have been treated as noise, and whether revision remains possible when materials cannot be accommodated are also crucial to unfolding the image again. Historical indices make it possible to examine how old differences are triggered in the present. Place names, old accounts, archives, and memories are not the past itself. Only after they have been extracted, connected, and assigned weight within new relational fields is it possible for them to change current judgment. The existence of counterexamples forces interpreters to confront materials their images cannot accommodate. Such unfolding again will alter the original bearing of responsibility, judgments of eligibility, and evaluations, and change the starting point of subsequent explanation. It need not overturn the entire original image. Once materials are added, the original bearing of responsibility, judgments of eligibility, and existing evaluations may be revised. The new explanation thus formed will in turn become an entry point for later people to continue examining.
An explanation that cannot be reopened will increasingly resemble a closed image. Only an explanation that can be reopened retains the capacity for calibration. The boundary of healthy simplification concerns whether omissions can still be opened up through materials, nodes, relational chains, historical indices, and counterexamples, rather than whether an image omits details. Once simplification no longer states its boundaries, no longer acknowledges remaining differences, and no longer retains a return channel, it enters dangerous compression. By elevating validity within a limited scope into validity for the whole and treating downweighted differences as nonexistent, compression exceeds its working scope, turns boundaries that should be clear into occlusion, and hardens an image into underlying reality. This is precisely the beginning of explanation's movement toward illusory construction.