Chapter 9. Why Differences Carry Different Weights: How Weight Forms within Relations
I. Why the Same Difference Carries Different Weights in Different Relations
A difference's existence and its coming to matter in the relation at hand are two different things. The former says only that a certain difference can be identified; the latter concerns whether it is extracted, assigned operative weight, and changes people's circumstances. Nonlocal status exists throughout. When someone buys vegetables at a market, that status may relate only to accent, familiarity, and whether one is a regular customer. What matters more to the stallholder at that moment may be whether one pays, bargains, or comes often. When a child enters school, however, nonlocal status may connect with residence permits, rental contracts, continuity of social insurance, school places, document deadlines, the landlord's cooperation, and the parents' ability to take leave. It no longer manifests only as an accent, but enters the handling of eligibility, deadlines, and supplementary documents. In medical insurance reimbursement, rental registration, or eligibility review, this difference enters relations again through different materials and thresholds. The difference has not become something else. What has changed is how it is extracted and the consequences to which it connects.
The same holds for poverty. In everyday food and clothing, it may appear as economizing; at relatives' weddings and funerals, it connects with cash gifts and the parents' standing; when paying hospital bills, it bears on whether treatment can continue; in a bank loan, it connects with credit and guarantees. When a child moves on to further schooling, poverty also connects with training, accommodation, transport, and waiting for opportunities. These different consequences cannot all be attributed to a fixed power belonging to poverty itself. Its operative weight in each instance must be judged within concrete relations. Accents have no fixed operative weight either. Among people from the same hometown, an accent may bring closeness; in a job interview, it may bring distance; in negotiations over renting, it may change the landlord's expectations; in a local conflict, it may become a marker of exclusion. The same difference can become lighter or heavier. Weight concerns precisely these changes within relations, rather than a ranking of differences that holds forever.
When analyzing a particular identity or status, we must also ask whether it is adopted in this interaction, to what extent, and how it changes the handling together with other available materials. Otherwise, differences between situations can easily be misrepresented as sudden increases or decreases in the difference itself.
In judging these levels, we cannot leap straight from "already seen" to "certain to produce consequences." Someone who notices a difference may treat it only as background. Someone who extracts it may still keep it from becoming decisive because materials are insufficient, its relevance is not recognized, or other conditions impose limits. Only by explaining the grounds at each step can we know why operative weight rises here but may not rise elsewhere. Within the same situation, someone may see a difference yet choose not to adopt it, or adopt it while giving it very little operative weight. Analysis must specify where this turn occurs; otherwise, "differences have an influence" is an empty generalization. The operative weight of the same difference is therefore not determined in advance outside the situation. The scope within which it can be adopted becomes apparent only in concrete judgment.
Judgments of weight must attend to the capacity to bear consequences. The same monetary loss has different consequences for different families because their income buffers, debt deadlines, care responsibilities, and alternative channels differ. This has nothing to do with people's worth. These conditions must be explained through concrete materials and relations; they cannot be deduced directly from the names of identities or statuses.
Visibility and efficacy must also be distinguished. Readily available numbers, official documents, and public records often stand out more, yet may not change consequences more than care, pain, local recognition, or community responsibilities that are difficult to fit into a format. The visibility of materials is only one condition of extraction. Their actual operative weight still depends on how they are adopted and connected to action space.
II. Weight Is Not a Ranking of People
Weight does not mean that a difference is inherently higher. Nor is it a ranking of people: it does not mean that one person is higher or lower, much less refer to moral worth. Accent, household registration, the set of differentiating units extracted by class theory, gender, and medical records do not carry great weight forever. Whether they enter into consequences depends on whether current relations extract them, give them operative weight, and connect them to concrete handling. Weight is the degree to which a differentiating unit, within concrete relations, is extracted, is assigned operative weight, and affects judgment, channels, responsibility, recognition, risk, time, and choice. This definition does not treat weight as a fixed numerical value carried by a difference itself. It may take the form of thresholds and eligibility, or of responsibility, risk, shame, the capacity to wait, the ability to exit, and the range of choices. The same difference can connect to different consequences in different relations.
A difference not extracted within current relations may remain mere background. Even if it has been seen, it may not actually change circumstances if it has not been given operative weight. Weight therefore does not mean being seen. Being seen, being recognized, being given operative weight, and connecting to consequences are different relations that cannot be collapsed into one. Nor does weight mean a "root cause." It indicates only the strength with which a difference enters consequences in this relation. To declare abstractly, apart from concrete relations, that it is always the weightiest already exceeds the bounds of judging weight.
Treating weight as a label that can be preserved apart from relations readily occludes the conditions of extraction, the grounds of operative weight, and the direction of consequences. If any of these links cannot be explained, the judgment of weight must remain provisional. Such a judgment must therefore explain its grounds, scope, and provisional character: it must specify the handling in question, the materials supporting it, and the kind of consequences affected. Merely saying that a difference is "weighty" without explaining these conditions slips back into an abstract label that conceals rather than explains the actual handling. Whenever the scope, materials, or points of reference adopted by participants change, the original judgment must be checked again. This also requires separating judgments of weight from language that evaluates people. To say that a difference has greater operative weight in a particular handling only explains the consequences of its adoption; it cannot be extended into a judgment of the person's worth, ability, or deserved position.
III. How Differences Gain Weight within Relations
A difference gains weight not because it grows larger on its own, but because specific participants begin to require, adopt, or compare it within relations. We can approach this through several connected points of observation that must not be treated as a fixed process, such as extraction, the assignment of operative weight, comparison, and connection to consequences. The differences participants extract vary with the handling. When buying vegetables at a market, they may extract whether one pays, is a regular customer, or bargains. In admission review, they may extract household registration, residence permits, rental contracts, continuity of social insurance, school places, and document deadlines. When hospital bills are paid, those involved may extract medical insurance status, medical records, balances, reimbursement rates, and the capacity to advance cash. At relatives' weddings and funerals, closeness and distance, records of exchanges, the parents' standing, past dealings, and attendance may also enter judgment.
Extraction, however, does not mean that operative weight has already been assigned. Consequences cannot be inferred directly from the mere existence of a differentiating unit, and handling does not always begin with complete materials and proceed in sequence. Some differences are merely seen, some become decisive, some are seen but not recognized, and some are recognized yet lack sufficient operative weight. Existing medical records may carry insufficient operative weight if they do not meet reimbursement or compensation requirements. An accent can bring closeness at a gathering of people from the same hometown, yet create distance in an interview. In some settings among relatives, poverty may bring shame rather than sympathy. Even where a workplace injury has occurred, the operative weight bearing on responsibility may be insufficient without a contract and evidence. Someone may first face time pressure and only then seek medical records, contracts, or available witnesses; alternatively, the materials may already be complete, yet temporarily unable to change the result because they have not been adopted. If analysis mistakes an order of observation for a procedure that reality must necessarily follow, it will miss the concrete nodes at which extracted materials are not adopted, fail to take effect, or are interrupted.
Comparison brings extracted differences into the same handling, giving them concrete positions. Whose materials are more complete, whose social insurance is more continuous, who owns property and who does not, who can wait and who cannot, and who has a local kinship network and who does not may all become objects of comparison. Who gives more or less in cash gifts, who seems more like an insider or an outsider, who has savings and who can only borrow, and who can provide proof and who cannot also change how the handling at hand can proceed; they do not rank people in an eternal hierarchy. Different points of reference may also produce different results in such comparison. The same materials may not be viewed in the same way in different questions. Both the points of reference and the consequences must be recorded to distinguish insufficient materials from a change in their operative weight caused by how they are adopted.
Differences must ultimately connect to consequences: residence permits to school-admission eligibility; medical record documents to reimbursement, leave, and compensation; accents to trust and distance; and debt deadlines to time pressure. Family responsibilities connect to migration and choices of work; savings to the capacity to wait; local kinship networks to available exits, information, and care; and national memory may connect to trust, fear, and community boundaries. The more direct the consequences, the more readily weight rises. But this is only an empirical tendency, neither a fixed formula nor something that can determine weight on its own. We must still ask what other differences are extracted, downweighted, or assigned weight anew at the same time. Someone may first feel the pressure of consequences and only then look back to identify which differences have been treated as grounds. These points of observation are connected but cannot replace one another. Every link must remain open to examination against materials and consequences, allowing analysis to verify each part rather than supply a prefabricated answer to the circumstances.
Weight may also change repeatedly over time. A record may carry great weight in one application, then temporarily recede into the background once correction or supplementation is complete. Local recognition may provide support in a crisis, only to become a constraint later through conflict. Analysts need to record the conditions under which weight increases, recedes, and increases again, so that a ranking on one occasion is not presented as a permanent hierarchy.
IV. Why the Differentiating Units Extracted by Class Theory Often Carry Weight
The class image readily attracts interpreters because the set of differentiating units extracted by class theory often bears on whether someone can wait, exit, or proceed by trial and error. What must be acknowledged is not class as what society fundamentally is, but the real weight of these differentiating units in many relational fields. Class theory extracts a set of differentiating units whose weights are amplified together within economic relations and relations of labor, property, risk, and domination. It does not extract a single difference. Wealth, labor, resources, domination, risk transfer, the capacity to wait, the capacity for trial and error, and the ability to exit may all belong to this set.
In illness, having or lacking savings changes the path of treatment. For someone without savings, treatment quickly connects with borrowing money, stopping work, reimbursement, pressure from relatives, and debt deadlines. In unemployment, whether a family provides a fallback changes the capacity to wait. Some people can wait a month, three months, or half a year; others face pressure to pay rent and buy food after three days without income. In children's education, property, the capacity to obtain training, access to transport, and time available to spend with a child change the channels. In workplace injury and debt, this set of differentiating units also changes bargaining power, the capacity to wait, and the ability to bear failure. Thus, in circumstances involving wages, unemployment, workplace injury, debt, medical care, housing, and buffers against risk, this set of differentiating units extracted by class theory may indeed press to the fore. In opposing the class image, we cannot write off the differences extracted behind it as irrelevant.
But this acknowledgment cannot be inverted into an assertion of a class entity. What is real and weighty consists of concrete differentiating units and their relations. What is called class is an image formed when class theory subjects a set of these differences to low-dimensional compression, projection, and naming.
This capacity for distinction extraction within a limited scope cannot change the basic structure of class theory. Class theory is precisely Model C: total-image subsumption. It is neither Model A nor Model B, nor an ordinary limited model. On the basis of a set of real differentiating units that may carry high weight, it compresses them into a total explanatory image and thereby crowds out or subsumes other explanatory spaces. This is precisely why the class image must be dismantled back into concrete differentiating units, relations, and consequences, and why the image cannot be treated as society itself.
Differences extracted by class theory and compressed into the class image often carry real weight, but this cannot replace checking concrete relations. Unless we go on to ask who adopts property, labor, debt, or risk in what kind of handling, it is easy to mistake the limited explanatory capacity formed through real differentiating units and their extraction, weight assignment, connection, and compression for the action of an entity itself.
Three levels of distinction must be maintained here. Concrete differentiating units may carry great weight in certain relational fields; this explains only the material source of class theory's explanatory appeal. When class theory extracts these differences, assigns them weight, connects them, and subjects them to low-dimensional compression, projection, and naming as a class image, it has already organized them into a Model C structure in which the total class image prearranges the explanatory positions of other differences. The class image is only the low-dimensional output of this process; there is no antecedent, safe class image of an ordinarily limited kind. These three levels cannot replace one another. The distinction also prevents another misunderstanding: the weightiness of the differences marked by the image does not mean that class theory has extracted every weighty difference. All relations placed outside the image that nevertheless affect recognition, responsibility, or available exits must re-enter examination. The image's appeal comes from a set of recurrent pressures. The problem is that, from the very formation of the class image, class theory organizes this set of pressures into a Model C structure that prearranges the explanatory positions of other differences.
V. Why No Set of Differences Can Be Presupposed to Carry the Greatest Weight Forever
The set of differentiating units extracted by class theory can carry great weight, but cannot be declared in advance to carry the greatest weight on every occasion. The same low income does not mean the same circumstances. Someone with a local kinship network may be able to borrow money, obtain information, receive care, or have someone introduce them to a job; someone without one may be more isolated. These relations change how the same income conditions enter consequences. Two people whose low-dimensional positions as marked by the class image are similar may nevertheless differ: one has a local fallback and the other does not; one knows local rules and the other does not; one has help caring for older people and children and the other does not. Their structures of weight are not the same.
Boundaries between national communities, memories of war, and religious relations may also outweigh differences in income, employment, and property. Two poor people do not necessarily trust one another simply because they are poor. Language, religion, historical injuries, community memory, and boundaries may change alliances, fear, recognition, and exclusion. Interpreters adopting class theory can extract some resource differences, but this explanation cannot automatically generate trust and community. Locality, language, clan, ethnicity, religion, institutional status, gender, age, medical records, and responsibilities to relatives may all gain weight within certain relations. Listing these differences is not meant to add them to another total table, but to reject the advance designation of a single center of high weight. Judgments of weight must return to concrete relations. They neither deny that differences extracted by class theory may be weighty nor allow that set of differences to subsume other differentiating units that have already changed consequences.
This limitation still allows valid comparison. Every judgment must establish which differences actually change the consequences at hand more strongly, while acknowledging that such comparison is constrained by materials, time, the ways matters are taken up and sustained, and changes in relations. Refusing to presuppose the highest weight does not mean assigning equal weight to all differences. When materials are sufficient, we must still judge which differences more directly change channels, responsibility, waiting, or available exits in the immediate question. But this greater operative weight holds only for this relation and this handling; it cannot be passed off as a permanent overall ranking. Economic, labor, and property differences may be the first to change whether people can wait and exit in some circumstances. That judgment, however, must be supported by the consequences at hand, rather than declared before any comparison. This is precisely why opposition to assigning equal weight to all differences and opposition to a total image can both hold.
Macro-level judgments across events cannot rely solely on high frequency. A difference may recur because the same institution repeatedly extracts it, or simply because the materials are easy to record. We must also compare whether it actually changes eligibility, risk, or action space, and which counterexamples show no such effect. Frequency is a clue, not sufficient proof of weight.
VI. What Changes When Differences Gain Weight Is Action Space
When a difference gains weight, this is more than another item in an explanation: people's paths may narrow or widen. Once a difference acquires higher weight within concrete relations and connects to consequences, people's channels, responsibility, recognition, shame, waiting, available exits, and choices may all change. When a residence permit gains weight, it may change the channel to school admission; when medical records gain weight, they may change reimbursement, leave, and compensation; when an accent gains weight, it may change trust, distance, and the degree of scrutiny one faces. When a debt deadline gains weight, it changes the ability to wait, resign, rest, or refuse poor-quality work. When family responsibilities gain weight, they change migration, shifts, leave, commuting, and physical wear. When a local kinship network gains weight, it changes information, borrowing, care, and available exits; when national memory gains weight, it may change recognition, vigilance, alliances, and exclusion. These consequences need not occur simultaneously, nor can they be written as a fixed chain.
A person may not immediately be ordered to do anything, yet action space has already changed: some paths open and others close; some choices become costly, some refusals dangerous, and some waiting impossible. Weight therefore cannot serve merely as decoration in an explanation. Analysts must also explain how a difference concretely changes the costs of choice, the bearing of responsibility, and the width of the paths that can be taken.
Analyzing a difference's weight is not meant to fix a person in a position. It is meant to explain what certain choices cost at the time, which support remains available, and which support is already insufficient. Only by making these conditions clear can we understand why the same refusal, waiting, or exit poses different degrees of difficulty for different people.
Channels, waiting, proof, refusal, exit, taking matters up and sustaining them, and available exits are not the same thing. Someone may still have a channel through which to make a demand, yet lack the conditions needed to wait for a reply. Someone may be able to offer proof but have no fallback for bearing failure. Someone may refuse one arrangement but be unable to bear the cost of repeated refusals. Only by separating these can we see how action space concretely contracts or expands. Action space also depends on whether different conditions can be obtained at the same time. The ability to provide proof does not necessarily mean the ability to wait; the ability to wait does not necessarily mean the ability to exit; having a fallback does not necessarily mean that matters will be taken up and sustained for one. No single one of these conditions can therefore substitute for a judgment of the circumstances as a whole. Changes in channels and available exits also affect, in turn, whether people can wait, supply further proof, and bear refusal, so we cannot look only at a single result.
VII. Why Judging Relative Weight Cannot Stop at a List
A list makes differences visible; their relative weight brings circumstances into view. In school admission, one can list household registration, social insurance, rental contracts, residence permits, school places, parents' leave, the landlord's cooperation, and information channels. Yet without knowing which of these outweighs other differences under current rules and at the current temporal node, analysis still stops at the list. The same holds for paying hospital bills: medical records, medical insurance, balances, reimbursement rates, patient accompaniment and care, and transport costs are all present, yet what pushes a family to the edge of borrowing may be its capacity to advance cash at that moment. At relatives' weddings and funerals, records of exchanges, closeness and distance, the parents' standing, past dealings, and attendance enter judgment again. The source of shame still depends on whose evaluation, which past relation, or which absence has acquired greater operative weight.
Listing can tell us only which differentiating units are present. Judging their relative weight requires asking within what relations they are extracted, compared, and assigned weight, and what consequences they thereby connect to. Such judgment does not cancel comparison; it demands more concrete comparison. The multiplicity of differences cannot justify writing them all as equally important, nor can complexity be an excuse for abandoning judgment. Lists and comparison do not exclude one another: a list offers points of entry that can be checked, while comparison must explain when those points of entry actually lead to consequences. Without materials, we should not pretend that comparison is complete; once materials appear, we cannot invoke an earlier provisional judgment to refuse revision.
Valid comparison also requires knowing what is being compared: different materials within the same handling, different ways of taking matters up and sustaining them in similar circumstances, or the effect of the same difference at different temporal nodes. However abundant the materials, unclear objects of comparison cannot yield a reliable conclusion. Retaining a provisional judgment when materials are insufficient is not an abandonment of analysis; it avoids writing the unknown as a fixed position. Comparison yields provisional conclusions that can be questioned, rather than turning materials into unalterable verdicts. If consequences differ from what was expected, we must also return to examine what went wrong in extraction, operative weight, or comparison. The revisability of judgment is not weakness. It is a necessary condition for keeping abstract names from closing comparison in advance.
When several differences gain weight at the same time, weight analysis need not rank one difference above all the rest. It can explain how multiple paths run in parallel, who bears which part of the consequences, and which relations counteract or amplify one another. This differentiation into levels preserves the real structure of an event better than converting all differences into a single total score.
A judgment of weight must leave an argument sufficient for others to understand its grounds. Where materials come from, who extracts them, at which node they act, how far the consequences extend, and what contrary materials still await handling together explain why a difference gains weight. These need not be converted into scores or fitted into a fixed format. What matters is that readers can follow the argument to identify omissions and distinguish operative weight supported by materials from an interpreter's intuitive ranking.
VIII. How Weight Rearranges Other Differences
Once a difference gains weight, the operative weight of other differences may also be rearranged. Without local household registration, if local admission policy lists residence, social insurance, or rental materials among application conditions, the weight of some of these materials may rise; whether the landlord cooperates in producing materials is a living condition faced by the family as it goes through the procedure. Without savings, debt deadlines, dates when wages are credited, and rent dates may also gain weight. Formal institutional requirements and informal conditions for carrying things forward must be understood separately. Greater weight in care responsibilities may amplify the weight of shifts, commuting, and leave rules; incomplete medical records may amplify the weight of the burden of proof, witnesses, surveillance footage, and the capacity to wait; greater weight in local kinship networks may also amplify the weight of local reputation, past relations, forms of address, and cash-gift exchanges. Differences do not propel other differences on their own. Concrete handling and arrangements for living make them alter one another's operative weight within the same relation. Which connections actually occur must still be checked against materials, timing, and concrete relations.
When specific participants increase the operative weight of certain differences in explanation or handling, other differences may be downweighted accordingly. When the class image takes the center in an explanation, ethnicity, locality, religion, the private sphere, language, and rites and customs may be downweighted. When a cultural explanation is central, housing prices, working hours, debt, and institutional eligibility may be downweighted; when a family explanation is central, market pressures, labor institutions, and social security may be downweighted as well. When institutional materials are given very high operative weight, care responsibilities, information asymmetry, bodily conditions, and difficulties in taking leave may lose their position as grounds. Being downweighted does not mean nonexistence; they may continue to produce consequences.
Weight cannot be understood as isolated operative weight. When specific participants increase the operative weight of certain differences within the same handling, whether other differences can be seen, recognized, compared, and connected to consequences may change accordingly. It is when participants adopt certain materials, bring them into comparison, or act on them in handling that they change the opportunities for other differences to enter judgment. This change may come from different points of reference, supplementary materials, the emergence of consequences, or changes in relations of taking up and sustaining. Once conditions of entry have changed, we must still ask who adopted what materials, to what extent, and whether the consequences really changed as a result.
Reconfiguration does not necessarily mean that certain differences are excluded from then on. It means only that, in the handling at hand, participants make less use of certain materials or have not yet connected them to consequences. A difference not adopted on one occasion cannot therefore be declared inoperative, nor can one configuration be written as a permanent structure. Once consequences become apparent or materials are supplemented, the original selections may still be rewritten. Thus, in saying that a difference is downweighted, we should also specify which materials were given priority, which kind of comparison occupied the position, and when the point of entry may need to be reopened. Only then does downweighting avoid rewriting complex relations as yet another abstract verb.
Different nodes may conflict in the weight they assign to the same difference. Medical staff focus on bodily conditions; workplace managers may focus on attendance at work; family members are concerned with income and care. No one side's judgment can automatically substitute for those at other positions. Analysts must trace what capacity for execution each judgment has, while also preserving the constraints that conflicting materials place on subsequent handling.
Weight is also affected by alternative channels. If a node rejects materials but the person concerned can supply further proof through another procedure, the original constraint may be lighter. If the same record or deadline cuts off all alternative paths, the weight increases. When judging a difference's operative weight, we must examine both the main channel and available exits.
Nor does high weight automatically mean a high degree of responsibility. Materials may be important to a result without those who produce them having authority to decide the consequences. Conversely, staff at the end of a process may carry out an action while remaining constrained by upstream configurations and authorization. Weight explains the strength of a difference's effect; responsibility must still be judged according to action, knowledge, authority, and the capacity to revise.
Weight in intergenerational relations cannot be measured by whether those concerned experienced something directly. Later generations may extract old differences anew through family memory, education, institutional records, and community practices. Only when these carriers give them consequences in the current event do past materials gain weight. Historical weight requires present relays; it is not the past governing the present on its own.
IX. The Boundaries of Judgments of Weight
Analyzing differences' weight reveals their relative weight, but cannot yield a permanent ranking detached from relations. To judge why a difference gains weight at this moment, we must still ask what relations it has entered, who extracted it, under what rules it was assigned weight, what consequences it connects to, and what other differences amplify its weight or downweight it. The same accent can bring closeness among people from the same hometown and distance in an unfamiliar institutional setting. The same poverty can take the form of economizing in daily life and direct pressure in hospitals and at weddings and funerals. These differences have not changed their essence on their own; what has changed is how they enter relations. Differentiating units such as labor, debt, risk, and domination, extracted by class theory and compressed into the class image, may carry great weight in workplace injury and debt, yet cannot automatically explain boundaries between national communities and religious memory. This does not deny their weight. It refuses to treat an image produced by distinction extraction within a limited scope as society itself.
Judgments of weight therefore cannot turn every increase in weight into irreversible fate, nor mistake a temporarily weighty difference for society's fixed substratum. Relations can change, materials can be supplemented, ways of taking matters up and sustaining them can emerge, and previously downweighted differences may re-enter judgment. Counterexamples, supplementary materials, and changes in relations do not interfere with judgments of weight; they are necessary conditions for correcting them. If new materials show that a previously overlooked difference actually changed consequences, or that a previously weighty difference did not act as expected, analysis should revise its judgment rather than defend the previous ranking. This caution neither refuses to explain consequences nor provides an excuse to delay judgment. It prevents the consequences that can currently be explained from being expanded into conclusions about all relations, so that every attribution of greater operative weight can state its grounds, scope, and conditions of refutation, and remain subject to counterevidence from consequences.
The calibration of weight must ultimately be grounded in the retractability of judgments and an honest account of what remains unknown. If materials previously judged central are shown by later comparison not to have entered actual handling, or if different results arose through another path, the operative weight attributed to them should be reduced. If the core path persistently fails, the conclusion should be withdrawn. Without this condition, weight becomes merely flexible language through which analysts preserve their original account. Likewise, if the materials establish only that a difference was raised but cannot establish how much operative weight it acquired, the conclusion should stop at "possibly relevant" or "still to be checked." Vague claims of high weight cannot conceal insufficient evidence. Reducing the strength of a judgment is itself part of keeping weight constrained by reality.
The responsibility of judging weight is not to assign people positions in advance. It is to explain which differences changed action space within which relations, which were downweighted, and which still need checking. Only by preserving these conditions, countervailing possibilities, and the responsibility of comparison can weight avoid becoming an abstract term of rank once again. It must guard against two errors at once: writing a temporarily weighty difference as the root of everything, and refusing to judge what is currently changing consequences simply because differences are numerous. Weight analysis can maintain an explanation open to examination and correction only between these two errors.