Relational Degradation - The Reconfiguration of the Relational Subject - Discussion I

ENGLISH

Relational Degradation

The Reconfiguration of the Relational Subject

Discussion I

Wanhong Huang · huangwanhong@serendip.ngo

---

Abstract

Problem. A class of relational deformations resists the vocabulary in which relational harm is ordinarily described. Some have no identifiable perpetrator, some proceed without malice, and some are assembled entirely from judgments no party was wrong to make. The vocabulary of act, intention, and fault has nothing to attach to in such cases, and this is argued to be a consequence of a premise it carries without stating: that the subject is prior to the relation.

Approach. The paper proceeds deductively from the ontology of generative relational being, on which the subject is not an entity standing in relations but is what is generated within them. A consequence follows by definition rather than by inference: any degradation of a relation is a change in the conditions under which subjects are generated. Seven cases are examined, not as evidence for the thesis but as exhibitions of the paths by which the alteration transmits.

Findings. Two corrections cut against the natural formulation. Degradation does not diminish the subject’s dynamics but reconfigures them, and the reconfigured dynamics are frequently the more intense. And the reconstitution of subjects cannot be the wrong, since every relation reconstitutes its participants. A criterion is therefore required, and it cannot be framed in terms of generativity without collapsing into the utilitarian norm the framework exists to displace. The criterion the ontology entails is revisability: degradation is present when the reconfiguration forecloses the possibility that the conditions of generation be further revised through common experience.

Contribution. Revisability is proposed as the sixth principle of the ethics of relational reproduction, and as a meta-principle rather than a further virtue: it governs not relations but the five principles themselves, since any principle that has lost its revisability may degenerate into an instrument of the party it was framed to constrain. A second result concerns distributive justice, shown to be limited in principle, and not merely in execution, with respect to abandonment.

Keywords: relational degradation; relational subject; generative relational being; recognition; revisability; symbolic power; distributive justice.

A note to the reader. This is a discussion paper. It is circulated in order to be argued with, and objection, counter-example, correction, and disagreement of every kind are actively sought rather than merely tolerated.

The invitation is not a formality. The paper proposes, as its principal contribution, that no norm may be held in a form that forecloses its own revision through what is subsequently found, and that the framework advancing this proposal is bound by it. A claim of that kind is answerable to what is brought against it, and it is worth no more than the objections it has survived. The conditions under which the argument would be false are stated explicitly in §22, and the questions it cannot answer are recorded in §21. Both are set out so that they may be used.

Correspondence of any kind is welcome, and disagreement is welcome most of all.

huangwanhong@serendip.ngo

AI usage statement. The ideas, claims, and theoretical framework of this paper originate with the author. AI tools were used as an aid in drafting, in the location and verification of sources, and in revision. The origin of each idea has been verified by the author, who takes full responsibility for the content.

§1 Introduction

Relations deteriorate, and the ways in which they do so have been studied in literatures that do not readily speak to one another. Moral philosophy examines the wrongs that parties do to one another. Recognition theory examines the injury sustained when a subject is denied the acknowledgment it is owed. Psychoanalysis examines what a subject does with the deprivations it undergoes. Political sociology examines the arrangements within which such deprivations become routine, and international relations examines the conditions under which parties who intend no harm arrive at catastrophe. Each has produced results, and none of them is contested here. What is contested is that the phenomena they treat are separate.

This paper puts three questions. The first is whether the forms in which relations deteriorate, betrayal, war, abandonment, the withdrawal of response, public categorization, administrative indifference, and the loss of trust, are seven kinds of harm that happen to resemble one another, or one alteration transmitted along seven different paths. The second is what, if they are one alteration, is altered, since the vocabulary of act and fault locates the alteration in the subject and cannot then explain the cases in which no one acted. And the third is what distinguishes a degradation from a mere transformation, since relations transform their participants continually and it is not obvious, and must not be assumed, that transformation is a harm.

The questions are worth putting because the vocabulary in which relational harm is ordinarily described fails before a class of cases that is neither small nor peripheral. It fails where the injury sustained from an intimate differs in kind, and not merely in degree, from the injury a stranger could inflict, since it can record only that the wrongdoer was close and must treat a difference of nature as an aggravating circumstance. It fails where catastrophe arrives without malice, as in the situation the realist literature has described, since a vocabulary that must find a fault will find one, and where none exists it will invent one. And it fails most completely where a community ceases to sustain one of its members and no member decided it: there was no occasion on which the question was put, no moment that could be named, and no participant whose conduct, taken singly, was wrong. There was a logic, running to its limit, and a vocabulary constructed to attach a predicate to an agent has nothing to attach.

It is argued here that these failures share a source, and that the source can be stated. The vocabulary carries, without stating, a premise about what a subject is: that the subject exists prior to the relation and is affected by it. On that premise, whatever befalls a subject must have been done to him, and what is done must have been done by someone. The cases in which no one did anything are then not merely difficult but unintelligible, and the vocabulary’s response to unintelligibility is to supply an agent it does not have.

The paper proceeds by denying the premise. On the ontology of generative relational being, developed by the author elsewhere, the subject is not an entity that stands in relations. It is what is generated within them, and its properties, its capacities, and its continuity are sustained by the relation’s ongoing generation and are not otherwise sustained. The consequence is immediate and it is definitional rather than inferential: if the subject is what is generated in the relation, then the conditions of the relation are the conditions of the subject’s generation, and to alter the one is to alter the other. No further mechanism is needed to carry the alteration across, because there is no gap across which it would have to be carried.

The cost of this approach should be admitted at once. The argument is conditional upon its premise in a manner that admits of no repair from within. A reader who holds that the subject is prior to the relation will find that nothing here obliges him: the cases will appear to him as a collection of harms, variously administered, and the criterion as one moral intuition among the available ones. He will not be mistaken to think so. He will be disagreeing with the premise, and the premise is where the disagreement belongs. This paper does not argue for the ontology, and the cases are not offered as evidence for it.

The questions have a place in the framework’s development, and it should be stated plainly since it bears on what the paper is for. The framework has advanced an account of what relations generate, of the wealth that a shared and unrepeatable history bears, and of the conditions under which the circulation of that wealth is just. It has said comparatively little about what happens when generation goes wrong, and the little it has said has been organized by the vocabulary it was elsewhere at pains to displace, treating degradation as a wrong done by someone to someone. The present paper is an attempt to bring the account of degradation under the same ontology as the account of generation, and its principal result is a normative one that the framework did not previously possess.

They have a place in a longer movement as well. The site to which value has been anchored has shifted before. It was anchored, under conditions in which the terms of life appeared as given from without, in a source beyond the human. It was anchored thereafter in the sign, which could carry worth so long as the sign remained scarce. As the reproduction of signs has cheapened, that anchoring has failed, and value has migrated toward what the reproduction of signs cannot cheapen, which is relation. But a value that has migrated to relation is a value exposed to whatever can deform relation, and the deformations are not new. What is new is that they now reach the site at which value is held. A theory of relational value that has no theory of relational degradation is a theory of an asset without a theory of its destruction, and the omission is not one that a further account of the asset will repair.

The paper’s departure from the nearest existing frameworks may be stated in a sentence each, and is set out at length in §3. Recognition theory holds that recognition is owed to what a subject already is; the present account holds that recognition is among the conditions under which a subject comes to be, from which it follows that the subject who has achieved nothing is not thereby without a claim, and that a dense and particularizing attention may be a form of recognition and may nonetheless foreclose the one who receives it. Psychoanalysis locates the source of the subject’s transformation within the subject; the present account locates it in the structure of relational generation, and asks not what the subject did with what befell it but what the relation was doing such that this befell it. Structural realism takes the anarchic condition as given and derives the dilemma; the present account takes that condition as produced and asks after its production, and disputes nothing the realist says about the world once the production is complete. And generative justice concerns the path by which value returns to those who generate it, which presupposes that something has been generated and taken; the case of abandonment lies outside it, since what the abandoned lacks is not the return of what he made but a relation in which making would be taken up.

Two corrections govern everything that follows, and both cut against the natural way of speaking. The first is that degradation does not diminish the dynamics of the subject generated within a relation. It reconfigures them, and the reconfigured dynamics are frequently the more intense, so that the observation of vigorous generation is no evidence whatever that a relation is sound. The second is that the reconstitution of subjects cannot itself be the wrong, since every relation reconstitutes its participants and education is not thereby a species of violence. A criterion is therefore required, and the requirement is not a formality. One of the seven cases, administration by category, alters the conditions of generation with the greatest deliberateness of any considered here, sets aside the particularity of every subject it touches, and is nevertheless the condition of a form of justice that no one proposes to abandon. An account that could not distinguish it from the others would have produced a complaint and not a criterion.

The criterion offered is revisability, and it is shown to be entailed by the premise rather than selected by the author. A relation that can no longer be revised through common experience does not cease to generate subjects; it generates one subject, repeatedly, and that is what it is for a condition of generation to have been closed. From this a normative claim follows which is the paper’s principal contribution. Revisability is proposed as the sixth principle of the framework’s ethics of relational reproduction, and it is proposed as a meta-principle and not as a sixth virtue standing alongside the five: it governs not relations but the five principles themselves, since a principle whose interpretation can no longer be revised is interpreted by the party it was framed to constrain, and a constraint interpreted by its object is not a constraint. The framework that states the principle is bound by it.

The remainder of the paper is arranged as follows. §2 sets out the three failures of the vocabulary of act and fault and diagnoses their source. §3 locates the argument with respect to the literatures nearest to it. §4 states the premise, derives the consequence, and supplies the definition of degradation that the framework requires and the ordinary use of the word does not. §5 makes the two corrections and introduces the three-layer structure used throughout. §7 sets out a taxonomy of the seven cases along three independent axes, the distinction each removes, the form of attribution each admits, and the availability of repair, so that each case may be read as occupying a determinate place rather than as one more instance in a list. §8 to §14 examine the paths of transmission, and §15 compares them. §16 eliminates the candidate criteria, including the one the framework would most naturally propose and which it must refuse. §17 states the criterion and exhibits its entailment. §18 states the constraint separating the description of a mechanism from the justification of what it produces. §19 and §20 recapitulate the five principles, show what they cannot settle, and propose the sixth. §21 records what the paper has not done, §22 states the conditions under which it would be false, and §23 explains why it is offered as a discussion.

§2 The Vocabulary of Act and Fault Presupposes a Subject Prior to Relation

The vocabulary in which relational harm is ordinarily described is a vocabulary of act and fault. There is an agent who acted, an intention with which he acted, a fault in the acting, and a liability that follows from the fault. The vocabulary is old, it is serviceable, and it runs so smoothly across the ordinary cases that what it cannot say is seldom noticed.

Three things it cannot say. They were named in the introduction and are set out here, since the diagnosis that follows depends upon their being seen precisely.

The first concerns the specific quality of the injury sustained in betrayal. That the betrayed party is harmed, the vocabulary records. That he is harmed in a manner unlike the harm a stranger could inflict, it does not, and the difference is not one of magnitude. A stranger may do more damage and be forgiven the more easily. It has the resources to say that a wrong was done and that the wrongdoer was close, but it has no resources to say why the closeness alters the kind of the injury rather than its degree, and it therefore treats as an intensifier what is in fact a difference of nature.

The second concerns the possibility of catastrophe without malice. In the situation the literature of international relations has described as the security dilemma, each party arms in response to the arming of the other, each acts on a correct assessment of its position, and the outcome is a war that neither sought and both were rational to have prepared for. The vocabulary of act and fault can describe this only by locating a fault, and if no fault can be located it must either invent one or fall silent. It commonly invents one. To call the outcome an aggression is to have converted a structural predicament into a moral one, which is to say that it is to have misdescribed it.

The third is the most difficult and it will occupy this paper at length. A person who has become unable to contribute may find that a community ceases to sustain him, and it is characteristic of such cases that no one decided it. There was no meeting at which the question was raised. There is no moment that could be identified as the moment of abandonment, and there is no participant whose conduct, taken by itself, was wrong. What there is, is a logic, running to its limit. The vocabulary of act and fault is constructed to attach a predicate to an agent, and where no agent is available it has nothing to attach.

These are not cases at the margin of an otherwise sound account. They are among the central forms in which relational life deteriorates, and the vocabulary of act and fault is silent before them or, worse, speaks and misdescribes.

It is tempting to treat the difficulty as a matter of scope. The vocabulary was made for the acts of persons, one might say, and these are not acts of persons but conditions of systems; two vocabularies are needed, and each may be sound within its province. The response is inadequate, and its inadequacy is instructive. The three cases are not remote from the province of moral judgment. They are the occasions on which moral judgment is most urgently demanded and most conspicuously fails, and a vocabulary that must cede precisely those occasions to another discipline has not been given a boundary. It has been given a diagnosis.

The diagnosis this paper offers concerns a premise that the vocabulary of act and fault carries without stating, and it is a premise about what a subject is. On that vocabulary the subject is prior to the relation. He exists, he possesses properties and intentions, and he then enters into dealings with others; the relation is a thing that happens between subjects who were already there. On that premise the three cases are indeed anomalous, since if the subject is already constituted, then whatever befalls him must befall him from outside, and what befalls him from outside must have been done by someone.

The premise is not obviously true, and this paper proceeds by denying it.

§3 Recognition Belongs Among the Conditions of the Subject’s Generation

The argument stands close to four literatures and takes something from each. Because the proximity is close enough to invite assimilation, the departures are stated before the argument proceeds.

3.1 Honneth grounds recognition in what the subject already is

The account is nearest to the theory of recognition, and the difference is not one of emphasis.

Recognition, as Honneth develops it, is owed to what the subject already is: to his personhood, which love acknowledges; to his standing as a bearer of rights, which law acknowledges; to his achievements and his contribution, which esteem acknowledges. Misrecognition is accordingly a failure to acknowledge something that is there to be acknowledged, and the injury is the injury of having one’s existing worth denied. The account is retrospective in its structure, and the retrospection is not incidental to it: what grounds the claim to recognition is a property the subject possesses in advance of the recognizing.

The present account cannot take this form, and the reason is the premise. If the subject is generated in the relation, then there is no antecedent personhood standing behind the relation and waiting to be acknowledged. Recognition is not the acknowledgment of a subject who was already there. It is among the conditions under which a subject comes to be there at all.

Recognition, in this framework, is not retrospective but constitutive. It does not acknowledge what the subject is. It is among the conditions under which the subject is generated, and this is why its withdrawal does not merely wound a subject but alters what a subject can be.

Two consequences follow, and both mark a departure.

The first concerns the case that recognition theory handles least well. If recognition is owed on the basis of what a subject has achieved or can contribute, then the subject who has achieved nothing and can contribute nothing has no ground on which to claim it, and the theory must either supply a supplementary ground, an appeal to bare personhood, which is a different principle, or leave him unprotected. On the present account the difficulty does not arise in that form, because recognition is not owed on the basis of anything. It is a condition of generation, and a condition is not a reward.

The second concerns Honneth’s assumption that recognition is a good of which more is better. §12 exhibits a form of attention that is intense, sustained, and elaborately particularizing, and that forecloses the subject as thoroughly as any neglect. The categorized subject is not ignored. He is watched with a fixed attention, and every observation confirms the category. That such attention is a species of recognition, and that it is nonetheless degrading, is not a possibility for which a theory that treats recognition as a scalar good has provided.

3.2 Psychoanalysis locates the transformation within the subject

The account uses the vocabulary of the symbolic and it does not adopt the explanatory structure that vocabulary ordinarily carries.

Psychoanalysis locates the source of the subject’s transformation within the subject: in the operation of the symbolic order upon it, in what it does with its constitutive lack, in the fantasies by which it makes that lack bearable. The present account locates the source in the structure of relational generation. The distinction is not a preference among vocabularies. It is a disagreement about where an explanation should begin, and it has a consequence that will be visible in the cases: where a psychoanalytic account asks what the subject does with what befalls it, this account asks what the relation was doing such that this befell it.

What is retained is the observation that the subject does not confirm itself. Some external confirmation is required, and its withdrawal is not a discomfort the subject might tolerate but a condition of its continuing to be a subject. §11 depends upon this, and §21 records that the paper has not established it.

What is refused is the assimilation of the whole account to that observation. That the subject requires an external confirmation does not settle what may supply it, and the framework’s interest lies precisely in the possibility that relational generation might supply what has elsewhere been supplied by a fixed and unquestionable authority. Whether it can is not established here. It is the question the framework exists to pursue.

3.3 Realism takes as given the condition this account takes as produced

The realist account of the security dilemma is not disputed. It is, within its domain, correct, and §9 says so without qualification.

The disagreement concerns where the analysis begins. The realist takes the anarchic condition as given, and derives from it that a rational party must read another’s capacity as a threat. The derivation is sound. What the framework does not permit him to ask is how the conditions of generation came to be such that capacity is legible as threat rather than as capacity, and this is not a question he has declined to answer but one his premises do not raise. The present account takes the anarchic condition as produced, and asks after its production.

The difference may be put sharply. Realism describes with precision the dynamics of a world in which a certain recoding has already occurred. It does not describe the recoding, and it does not possess the means to.

3.4 Generative justice presupposes that something has been generated and taken

The framework’s account of value, and of the circulation that returns value to those who generate it, is indebted to the theory of generative justice, and the debt is acknowledged in the framework’s prior work.

One departure is registered here because it bears on the criterion. Generative justice is an account of value’s return to its producer, and it is therefore an account concerning producers. Its central injustice is extraction: value is generated and does not come back. This is a real injustice and the account of it is not disputed.

But the case of §10 lies outside it. The abandoned subject is not one from whom value has been extracted. He is one whose generation has no recipient, and there is nothing to be returned to him because nothing has been taken. A framework organized around the return of value has no purchase here, and the difficulty is structural rather than an oversight: an account of just circulation presupposes that something circulates.

Generative justice concerns the path by which value returns to those who generate it. The case of abandonment concerns admission to the circuit at all, and it is not an instance of extraction. What the abandoned lacks is not the return of what he made but a relation in which making would be taken up.

The sixth principle proposed in §20 is offered as an addition on this point. It does not replace the criterion of just circulation and is not in competition with it. It states a condition prior to circulation, which is that the relation must remain able to revise the terms on which it admits and generates its participants, and it is a condition that a circuit may fail while circulating justly among those it has already admitted.

3.5 The account is silent on motive and derives no political conclusion

Two disclaimers, and they are made because the vocabulary invites the misreadings they exclude.

The account is not a theory of the individual’s psychology and makes no claims about motive. That the participants in a degraded configuration act without malice is asserted repeatedly, and it is asserted because it is true and because the explanatory burden of the account does not require it to be false. No case in the cases is explained by the wickedness of anyone.

And the account does not derive political conclusions from its dynamical claims. That a mechanism can be shown to sustain the conditions of a system’s continued operation is a description, and §18 states why it must never be permitted to become a justification. The vocabulary of generativity is unusually exposed to the collapse of that distinction, and the constraint is stated formally for that reason.

§4 The Subject Is Generated in the Relation, and the Consequence Is Definitional

The premise is stated in its strong form, because the weak form will not do the work and it is important to see why.

The weak form would say that the subject is influenced by his relations. This is a causal claim, it is uncontroversial, and it is compatible with everything the vocabulary of act and fault assumes. A subject who is influenced by his relations is a subject who was there to be influenced, and the influence is a thing that happens to him. On the weak premise the three omissions of §2 remain omissions, since the question of who did the influencing is still available and still unanswerable.

The strong form says something else.

The relational subject. The subject is not an entity that stands in relations. It is what is generated in them. Its properties, its capacities, and its continuity are not possessions carried into the relation but are sustained by the relation’s ongoing generation, and they are not otherwise sustained.

This is the premise of generative relational being, and it is not adopted here for the convenience of the argument. It is adopted because it is the framework’s ontological commitment and because the argument of this paper is an unfolding of that commitment rather than an addition to it. What must be shown is that the commitment has the consequences claimed. That it is true is not shown, and a reader who declines it will find nothing in what follows that compels him.

The distinction between the two forms may be put as a distinction between two questions. On the weak premise, one asks what a relation does to a subject. On the strong premise, that question is malformed, and the question one asks instead is what a subject is generated as.

4.1 The consequence follows by definition

Grant the premise and the principal claim of this paper follows immediately.

If the subject is what is generated in the relation, then the conditions of the relation are the conditions of the subject’s generation. A change in the former is a change in the latter, and it is not a change that must be transmitted by some further mechanism whose operation could be doubted. There is no gap across which a transmission would have to occur. To alter the conditions under which a subject is generated is to alter what the subject is generated as, and the two descriptions pick out the same event.

In generative relational being the relational subject is not an entity outside the relation but is continuously formed within its generation. Any degradation of a relation is therefore not merely a change in the relation. It is a change in the conditions of generation, and thereby a reconfiguration of what a subject can be generated as and of how the subject’s dynamics can unfold.

The cases examined below are accordingly not attacks upon subjects. They are alterations in the conditions of relational generation, and the subject changes in consequence. This is the respect in which the account differs from the accounts with which it will be most readily confused. Psychoanalysis locates the source of the subject’s transformation within the subject. Psychology locates it in cognition and affect. The present account locates it in the structure of relational generation, and it does so not as a matter of emphasis but as a matter of what the subject is taken to be.

4.2 Degradation is a change in the capacity to generate subjects

The title of this paper contains a word that, left undefined, would render it empty. Degradation, in ordinary use, means that a relation has become worse, and a paper that established that some relations become worse would have established nothing. The premise supplies a definition, and it is a definition with content.

Relational degradation. Not that a relation has diminished, and not that a relation has become worse. A change in the relation’s capacity to generate relational subjects, such that the space within which a subject may be generated is constrained, redirected, or closed.

Two things follow from this definition and both are needed.

The first is that degradation is not a moral term. It names a structural alteration, and whether the alteration is to be condemned is a further question that the cases must answer and that the definition itself leaves open. This is a feature and not a defect. A theory that built the condemnation into the description would have no way to accommodate the case of institutional indifference, which alters the conditions of generation with the greatest deliberateness and is nevertheless the condition of a form of justice we have no wish to abandon.

The second is that degradation is important for a reason that is not about relations. It is important because the relation is the site at which subjects are generated, and it is only in virtue of that fact that a change in relations is a change of any consequence at all. Ordinary relational theory can say that relations matter to people. The present account says something stronger and stranger, which is that people are what relations are doing.

§5 Degradation Reconfigures the Subject’s Dynamics

The consequence drawn in §4 is easily stated and it is easily stated wrongly. Two corrections are required before it can be put to work, and both cut against the natural way of speaking.

5.1 The reconfigured dynamics are frequently the more intense

The natural formulation would be that the degradation of a relation degrades the dynamics of the subject generated within it. The formulation is wrong, and its wrongness is not a matter of wording.

Consider what is observed in the relations that have most conspicuously deteriorated. A community organized around a shared enmity does not exhibit a diminished capacity for action. It commonly exhibits an intensified one: its members act with greater decision, identify with greater conviction, and coordinate with greater efficiency than they did before. A group that has fixed upon one of its members as an object of contempt does not become listless. It becomes inventive. It generates fresh descriptions, discovers new occasions, and elaborates the categorization with a fertility that no equivalent effort would have produced in its absence.

The dynamics have not weakened. They have been reorganized, and the reorganized dynamics may be the more intense.

Relational degradation does not diminish the dynamics of the subject. It reconfigures them. Generation continues, and may continue with increased vigor; what changes is the space within which it can occur.

This correction is not a concession. It is what saves the thesis from vacuity. To say that degradation degrades is to say nothing; to say that degradation reconfigures is to say something that could be false, and the modes of reconfiguration can then be distinguished. Four may be named at the outset, and the cases below will exhibit them: generation may be constrained, so that fewer paths remain open; distorted, so that the paths that remain lead elsewhere than they did; closed, so that generation along a given path ceases altogether; or redirected, so that the subject is generated toward a configuration it would not otherwise have approached.

The correction has a further consequence, and it will be needed below. If degradation does not consist in the diminution of generation, then the presence of vigorous generation is no evidence that a relation is sound. A relation may generate copiously and generate only one thing. The question is never whether generation is occurring, and any diagnostic that asks only that question will report health in precisely the cases where health is most absent.

5.2 The reconstitution of subjects is universal and is not itself the wrong

The second correction concerns what the account must not claim.

If the relation is the site at which subjects are generated, then every relation constitutes its participants, and every alteration of a relation reconstitutes them. Education reconstitutes the student, and it is intended to. The student who emerges from a long apprenticeship is not the student who entered it, and the difference is the whole point of the exercise. A friendship of many years has constituted both parties in ways that neither would have chosen in advance and neither would now surrender.

It follows that the reconstitution of a subject cannot be the mark of degradation, since reconstitution is what relations do.

That a relation reconstitutes its subjects is not a defect but a definition. The criterion of degradation must therefore be sought elsewhere than in the fact of reconstitution, and any account that condemns reconstitution as such has condemned relation as such.

This is a serious constraint and it is the reason the criterion is necessary. It rules out, at a stroke, the most natural formulation of the normative claim. One cannot say that degradation occurs when a relation determines what its participants become, since that is the ordinary operation of every relation that is worth having. Something must be said about how the determination is effected, and the saying of it is the work of the criterion.

5.3 Degradation originates at a layer that intervention does not reach

It is convenient to distinguish three layers, and the distinction will be used throughout.

At the first layer stands the mechanism of relational generation: the conditions under which relations produce what they produce. At the second stand the relational subjects continuously formed within that generation. At the third stand the dynamics of those subjects among themselves, by which new relations are produced and the system continues to generate.

Degradation occurs at the first layer. It is transmitted to the second, and it alters the third.

Degradation originates at the layer of relational generation. Intervention is ordinarily directed at the second layer, by protecting the subject who has been injured, or at the third, by sanctioning the subject who has injured him. This is why intervention so often fails. The pathology is at a layer that neither intervention reaches.

The observation is not offered as a counsel of despair, and it does not entail that protection and sanction are worthless. It entails something narrower and more useful, which is that protection and sanction address consequences and that a relation whose generative conditions have been altered will continue to produce the consequences after each of them has been addressed. The role vacated by a sanctioned participant remains available, and it will be occupied.

§6 The Cases Exhibit the Paths of Transmission

The seven cases that follow are not evidence for the thesis established above. The thesis follows from the premise, and no case could strengthen it. Their function is different, and it is the function that a unifying account most requires.

A thesis which held that all relational degradation is a reconfiguration of relational subjects would be empty if it could not say how the reconfigurations differ. A unification that abolishes the differences among the things it unifies explains nothing, and the charge is easily earned. The cases are therefore presented as paths of transmission, and the question asked of each is a single question, put in the same words each time.

By what path does an alteration at the layer of relational generation transmit into a reconfiguration of the subject generated there?

The answer differs in each case, and the differences are the substance of this part. Where a case has been misdescribed in the author’s own earlier treatment, the misdescription is stated and corrected, since a path that has been misidentified is worse than one that has not been examined.

One further constraint is observed. The mechanisms at issue are visible in ordinary relations, and they are described here in ordinary relations. An argument that borrows its force from the emotional weight of an extreme example is not carrying its own weight, and the extreme cases are omitted for that reason rather than from delicacy.

§7 The Cases Are Distinguished Along Three Independent Axes

The seven cases are set out before they are examined, so that each may be read as occupying a determinate place in a structure rather than as one more instance in a list. Three axes distinguish them, and the axes are independent: a case’s position on one does not fix its position on the others.

7.1 The distinction that is collapsed

Each case removes a discrimination the relation was formerly able to make, and the discriminations removed are not the same.

CASE	THE DISTINCTION REMOVED	WHAT IS LOST
Betrayal	this relation vs. any relation of its type	specificity
War	a subject vs. an object of strategy	interiority
Abandonment	this person vs. any person of whom nothing is expected	singularity
Withheld response	present vs. absent	standing to be addressed
Public categorization	this person vs. one who may be so treated	the entitlement not to be categorized
Institutional indifference	this case vs. any case of its type	particularity
Loss of trust	this other vs. any other who might fail me	the standing to be entrusted

The distinctions are mutually irreducible, and the irreducibility is easily missed. A subject may be reduced to an object of strategy and remain an irreplaceable object of strategy; a relation may be reduced to its terms while both parties continue to acknowledge one another as subjects. That the seven are one alteration does not entail that they are one loss, and an account which allowed them to merge would have abolished the differences it was constructed to explain.

7.2 The form of attribution

The second axis determines whether the vocabulary of act and fault has anything to attach to, and it is the axis on which the difficulty of §2 becomes tractable.

CASE	STRUCTURE OF THE ACT	IS THERE A PERPETRATOR?
Betrayal	unilateral, intended, retrospective	yes
War	bilateral, may be unintended, self-fulfilling	possibly none
Abandonment	systemic	ordinarily none; a logic
Withheld response	coordinated, each act permissible	none, and none acted
Public categorization	three-party, intended, public	yes, and he knows it
Institutional indifference	institutional, intended, principled	yes, and he acts rightly
Loss of trust	incremental, each step warranted	none, and none erred

The column on the right exhibits a spectrum that moral philosophy is not equipped to traverse. It runs from a perpetrator who intended what he did, through configurations in which no one acted, to one in which the administrator acted deliberately and acted rightly, and terminates in a degradation assembled from judgments no party was wrong to make. A vocabulary that attaches a predicate to an agent can operate at one end of this spectrum and is silent at the other, and the silence is not a gap in its application but a consequence of its premise.

7.3 The availability of repair

The third axis anticipates the criterion and is the reason the cases are arranged as they are.

CASE	WHAT STANDS IN THE WAY OF REPAIR
Betrayal	a new specificity may be built; the old is beyond recovery, since the claim was about the past
War	both parties must cease objectifying at once, and each party’s ceasing exposes it
Abandonment	transfer cannot reach it; what is required is a different condition of generation
Withheld response	the faculty that would perform the repair has been enlisted against it
Public categorization	the audience’s discriminations must change, and the subject cannot address the audience
Institutional indifference	nothing does; appeal, hearing, and legislation are available
Loss of trust	there is no fault to acknowledge and no act to undo

Two rows are marked. The third carries the paper’s sharpest practical consequence, and §10 argues it: distribution operates within the conditions of generation, and abandonment alters those conditions, so that the limitation of distributive justice here is one of principle and not of execution. The sixth carries the paper’s criterion, and §17 and §20 argue it: the case in which the conditions of generation are altered most deliberately is the case in which repair remains available, and what distinguishes it is neither the gentleness of the alteration nor the intentions of those who administer it.

7.4 The order of composition

The cases are not independent, and two orderings among them will be used.

Objectification makes abandonment available. Once the other has been recoded as an object of strategy, an object of no strategic value is a discardable object, and the recoding is a precondition of the discarding. The converse does not hold. To render a subject indistinguishable from those of whom nothing is expected is not to render him an adversary, and the asymmetry supports a proposition that is uncomfortable and, so far as the author can determine, true: the enemy is more fully recognized than the abandoned. War accords the other the standing of an agent. Abandonment accords him no standing at all.

And the legitimate collapse licenses the illegitimate one. Where an institution has established that categorical treatment is proper, the categorizer of §12 need not manufacture an injustice. He need only invoke a category the institution has already authorized, and §12 shows that this is what he does. That the legitimacy of the one furnishes the cover of the other is the direct cost of the concession made in §13, and it is not a cost the account may decline to pay.

§8 Betrayal Rewrites the Relation’s Past

8.1 The alteration and its path

What the betrayer does is not, in the first instance, to end a relation. It is to redescribe it, and the redescription is retrospective. The claim implicit in the act is not that the relation will henceforth be of a different kind. It is that the relation was always of that kind, and that the other party had misunderstood it.

The subject generated within that relation had been generated on terms which the redescription cancels. He discovers not that a support has been withdrawn but that the support he took himself to have had was never of the character he supposed. The generative mechanism he relied upon is not destroyed. It is disclosed never to have existed in the form he relied upon it in.

He is generated thereafter as a subject who was mistaken about the conditions of his own generation. This is the specific quality of injury that §2 found the vocabulary of act and fault unable to name, and it explains why the closeness of the betrayer alters the kind of the harm rather than its degree. A stranger cannot inflict it, because a stranger was not among the conditions of one’s generation and cannot retrospectively cancel what he was never party to.

8.2 Forgiveness cannot retract a retrospective claim

Forgiveness can establish a new relation on new terms, and frequently does. What it cannot do is retract the retrospective claim, since that claim is not an act to be undone but an assertion about what was already the case. A new relation may be generated. The old conditions of generation cannot be restored, because they are now held never to have obtained.

Betrayal is irreversible not because trust is laborious to rebuild but because its operation is retrospective. What it alters is not the future of the relation but the account of its past, and no subsequent act has authority over the past.

8.3 The injury lies in the account of the past

Betrayal has been treated principally as a violation of trust, and the philosophical literature on trust supplies the resources for that treatment. Baier’s account of trust as the acceptance of vulnerability to another’s goodwill, and the accounts of betrayal that follow from it, explain why betrayal is a wrong of a distinctive severity: it exploits an exposure that the betrayed party undertook on the betrayer’s account [2]. Margalit’s distinction between thick and thin relations explains why the wrong is graver where the relation was thicker [13].

What these accounts share is that they locate the injury in the future. Trust was extended, the extension is now revealed to have been unwarranted, and what is lost is the capacity to extend it again. The account is correct and it is incomplete, and the incompleteness is exactly where the present framework contributes.

What the betrayed party reports is not, characteristically, that he can no longer trust. It is that he does not know what the relation was. The injury is to his account of the past, and the loss of future trust is a consequence of that injury rather than its content. On the vocabulary of trust this is difficult to say, since trust is a forward-looking attitude and a forward-looking attitude cannot be damaged retrospectively. On the present account it is straightforward: the relation was among the conditions of his generation, the redescription cancels those conditions, and he is generated thereafter as one who was mistaken about what generated him. The distinctive severity that Baier explains by the exploitation of vulnerability is explained here by something prior, which is that the conditions of one’s own constitution have been retrospectively withdrawn.

§9 War Recodes the Other Before It Destroys Him

9.1 The recoding precedes the violence

The relation is recoded as adversarial, and the recoding precedes the violence. This is the order that matters and it is the order the ordinary account inverts.

The sequence runs from subject, to enemy, to target, to casualty. It does not run from subject to casualty. The evidence is available in the preparations that precede any organized hostility: the construction of the adversary as a category, the elaboration of the narratives that make membership in that category legible, and the production of the figures in relation to whom one’s own participation becomes intelligible. Were the destruction of persons the object, none of this would be required. That it is invariably undertaken indicates that what is being built is not a capacity to destroy but a system for the production of subjects, without which the destruction could not be sustained by those who are to carry it out.

The subject is thereafter generated only as an enemy, or only as one who confronts an enemy. Here the first correction of §5 shows its necessity. The dynamics of the subject so generated are not weakened; they are heightened. Resolution, solidarity, and coordinated capacity all increase, and it is the increase, and not any lassitude, that makes the configuration self-sustaining.

9.2 Realism describes the world the recoding has already produced

The realist account of the security dilemma is not contested here, and this paper does not undertake to refute it. The realist describes with accuracy the dynamics of a world in which the recoding has already occurred, and within that world his conclusions follow. What he does not ask, and what his framework gives him no means of asking, is what altered the conditions of generation such that the other’s capacity is legible as threat rather than as capacity. He takes the anarchic condition as given [21]. The present account takes it as produced, and the difference is not a disagreement about the dynamics but a disagreement about where the analysis begins.

The constructivist literature has pressed this objection before, and what the present account adds to it should be said. Wendt’s argument that anarchy is what states make of it, and the wider claim that the identities and interests of states are socially produced rather than given, already deny that the anarchic condition is a datum [19]. The present account does not dispute this and is indebted to it.

Where it departs is in what the production is taken to produce. For the constructivist, what is constructed is an identity or an interest, which is to say a property of an agent who is there to bear it and who might in principle bear another. The agent survives the construction, and this is what makes the constructivist proposal a hopeful one, since identities may be reconstructed by altering the intersubjective understandings that sustain them. On the present account the agent does not survive the construction, because the agent is what the construction generates. What is at stake in the recoding is not which identity a party bears but what kind of subject is generated at all.

The difference is visible in what each account predicts about reversal. The constructivist expects that a reconstruction of shared understandings will restore the parties to a condition in which cooperation is available. The present account expects that it will not, and §17 gives the reason: the recoding has closed the channel through which understandings would be revised, and it is that closure, rather than the content of the understandings, that must first be undone.

§10 Abandonment Leaves the Subject’s Generation Without a Recipient

10.1 The alteration and its path

The relational system ceases to include the subject in its future. Nothing is done to him. What ceases is that he is no longer among the parties with respect to whom the system’s forward generation is organized.

His generation loses its recipient. He continues to generate, he acts, he speaks, he attempts, but the generation is not taken up, and the relational uptake that would have made it generative of anything further does not occur.

10.2 The abandoned subject loses his singularity

It is natural to describe his position by saying that his value has fallen to zero, and the description is wrong. Zero is a position. To be assigned it is to remain locatable within the order, to be a determinate element of it, and to be someone concerning whom the order can still say something, namely that his contribution is nil. That is not the situation of the abandoned. The situation of the abandoned is that he is no longer generated as a particular at all.

What abandonment removes is not the subject’s value but his singularity. He does not become worth nothing. He becomes indistinguishable from any other of whom nothing is expected.

10.3 A transfer restores a magnitude, and singularity is none

A transfer restores a magnitude, and what has been lost is not a magnitude. Resources supplied to a subject who is no longer generated as a particular are received by the category and not by him. He remains a member of the class of those from whom nothing is expected, distinguished now only by being a funded member of it.

The repair required is not a redistribution but a different condition of generation, and distribution operates within the conditions of generation and cannot alter them. This is a limitation of principle and not of execution, and it is the sharpest consequence the present account carries for the theory of justice.

10.4 Capability, misrecognition, and exclusion each stop short of singularity

The condition has been approached from three directions and each supplies something the others do not.

The capability approach holds that what is owed is not a quantity of resources but the real freedom to do and to be, and Sen’s insistence that equal resources may yield unequal capabilities already denies that a transfer settles the matter [16]. It is nearer to the present account than any other and it stops at a definite point. A capability is a capacity of a person and is assessed by asking what that person can do. The abandoned subject may retain every capability and be abandoned nonetheless, because what he has lost is not a capacity of his own but a relation in which his exercise of it would be taken up. What is missing is not on his side of the relation.

Fraser’s distinction between maldistribution and misrecognition marks the same territory and divides it differently [5]. Misrecognition, on her account, consists in institutionalized patterns of cultural value that deny a person the standing of a full partner in interaction, and the present case would be classed under it. What the classification does not supply is why the two remedies do not substitute for one another, and the present account gives the reason: misrecognition of this kind is not a further harm standing alongside maldistribution but an alteration of the conditions under which the subject is generated, and a transfer, which operates upon a subject already generated, cannot reach it.

The literature on social exclusion supplies the empirical texture and characteristically declines the theoretical question, treating exclusion as a compound of deprivations rather than as a condition with a structure of its own [17]. The structure is what the present account offers here. What the excluded subject loses is neither resources nor capabilities nor esteem, but singularity, the property of being generated as this one rather than as any of those; and the reason no transfer restores it is that a transfer restores a magnitude, and singularity is not a magnitude.

§11 Withheld Response Compels the Subject to Name Himself

11.1 Coordinated silence satisfies every condition of a symbol

Response is withdrawn. The subject speaks and the exchange continues past him; he is present and is not addressed. No act is performed against him and no name is applied to him, and it is characteristic of this case that each of its constituent non-acts is, taken singly, unimpeachable. No one is obliged to reply.

It is tempting to conclude that this case, alone among the seven, operates without symbolic mediation. The conclusion should be resisted. The withdrawal is coordinated. It is not that each participant independently happens not to respond; the non-response is consistent, is directed, and is anticipated by all parties including the one who suffers it. Coordination of that kind requires a shared rule, and a shared rule is a symbol whatever its mode of expression.

What is distinctive is the form the symbol takes. It is not a predicate. It does not assert what the subject is, and it therefore offers nothing that could be contested. It asserts a negated membership, and a negated membership requires no positive content. It need not hold that he is deficient, only that he is not among us.

The absence of a name is not the absence of a symbol. Silence, where it is coordinated, directed, and anticipated, satisfies every condition a symbol is required to satisfy. What it withholds is not its meaning but its predicate.

11.2 The subject supplies the account, and the faculty supplying it is thereby disabled

Here the transmission runs through a gap rather than through an imposition, and this is the crucial structural difference from the case of §12. Where a categorization supplies the subject with an account of himself, the withdrawal of response supplies him with none, and leaves him with a question that will not be left alone. The subject requires an account, and where none is furnished he will furnish one. It is he who names himself.

The interpretation he generates is generated by the very faculty upon which any revision of it would have to depend. This is the structural peculiarity of the case and it may be stated without metaphor. The subject’s interpretive capacity is enlisted in the production of an account whose effect is to disable that capacity. He is not persuaded of anything. He has concluded it, and a person does not doubt his own conclusion in the manner in which he might doubt an assertion made to him.

11.3 External kindness is absorbed, and the configuration is self-sustaining

The first is that external kindness is characteristically ineffective. A reassurance is received as consolation, an attention as pity, and a response, where one occurs, as the exception the rule anticipated. Each counter-instance is processed by an interpretation that has already specified how counter-instances are to be read.

The second is that this configuration requires no maintenance. A categorization must be continually re-performed, and §12 will make something of that requirement. The withdrawal of response, once established, is completed by the subject himself, and the community that initiated it need do nothing further.

11.4 Internalization is this case completed

It has been usual to treat the internalization of a degraded self-understanding as a distinct phenomenon, remarkable for the coincidence of the one who degrades with the one degraded. It is not a distinct phenomenon. It is the completed form of the case described here, and the coincidence is not a curiosity but the mechanism.

11.5 The subject supplies the category that symbolic violence would have imposed

The psychological literature on ostracism has established the phenomenon and much of its structure. Williams’s programme has shown that exclusion is registered rapidly, that it is registered even where the excluding party is known to be arbitrary or is not a party the subject values, and that it threatens the fundamental needs of belonging, self-esteem, control, and meaningful existence [20]. The findings are not disputed and the present account depends upon them.

The theoretical accounts of internalization supply the other half. Bourdieu’s symbolic violence names the mechanism by which the dominated come to apprehend their position through categories that are the categories of the dominant, and Fanon’s account of internalization exhibits its operation upon a subject who has no external standing from which to contest them [3, 4].

What is not supplied by either literature, and what the present account offers as its contribution here, is why the internalization is not merely a frequent sequel to exclusion but its completion, and why the interpretation the subject generates is uniquely resistant to correction from outside. The reason is structural rather than psychological. In the case of a categorization, an interpretation is imposed, and an imposed interpretation retains a mark of its origin: the subject may still hold that it came from elsewhere and is therefore contestable in principle, however unavailing the contest. Here nothing is imposed. A gap is left, the subject is required to close it, and the interpretation that closes it is his own conclusion. A subject does not doubt his own conclusion in the manner in which he might doubt an assertion made to him, and every subsequent kindness is received by an interpretation that has already specified how kindness is to be read. The literature on symbolic violence explains how the categories of the dominant become the categories of the dominated. It does not explain the case in which no category was supplied and the dominated supplied his own.

§12 Public Categorization Alters What the Audience Can Distinguish

12.1 The three-party structure is constitutive

The rule by which the community distinguishes among its members is rewritten, and the rewriting is public. This is the constitutive feature of the case and it is the one most often mistaken for an incidental one.

The operation is not performed upon the subject. It is performed before an audience, and the alteration it effects occurs in the audience’s capacity to distinguish, not in the categorizer’s. The categorizer need not believe what he asserts, and frequently does not. What is required is that others come to find the assertion available. This is why the operation is almost never conducted in private: conducted in private it would accomplish nothing, since the party whose discriminations must change would not be present.

The operation is a demonstration, and its site is the audience. What is altered is not what the categorized subject is, but what the community is able to say about him, and thereafter about anyone.

12.2 Bystander silence completes the operation, and resistance confirms the category

The silence of the bystanders is central rather than peripheral. It is not a failure to intervene, and to describe it so is to mislocate it. It is the evidence that the audience has accepted the categorization, and that acceptance is what the demonstration was for.

Resistance by the categorized subject is characteristically unavailing. The contest is not between him and the categorizer. It is conducted in a forum he cannot address, and any conduct of his is available for redescription as confirming the category. Conduct well and he is the exception. Conduct badly and he is the instance. Protest and he is what such persons do.

The categorizer commonly bears the subject no particular animus. The subject is not the object of the exercise. He is the medium through which the demonstration is conducted, and a medium is not hated.

12.3 Where rules are forming the operation instructs; where they are fixed it invokes

Where the community’s rules of discrimination are still forming, the operation is pedagogical. It exhibits, to participants who have not yet settled the matter, what the available categories are, and it is the most vivid instruction they will receive.

Where the rules are institutionally fixed, the operation cannot proceed by instruction and does not attempt to. It proceeds by invoking a category that already carries authorization, with the effect that resistance to the categorization becomes indistinguishable from resistance to the institution that authorized it.

In the first setting the operation creates a category. In the second it places the subject into a category the institution has already licensed. The second is the harder to resist, since to dispute it is to dispute the institution.

12.4 Labelling theory misses the audience and misreads the silence

Labelling theory has held the essential insight for six decades. Becker’s formulation, that deviance is not a quality of the act but a consequence of the application of rules and sanctions by others, and Lemert’s distinction between primary and secondary deviation, already locate the operation in the community rather than in the one who is labelled, and already recognize that the label produces the conduct it purports to describe [12, 11]. Goffman’s account of stigma supplies the phenomenology of the labelled subject’s position with a precision this paper does not attempt to improve upon [7].

Two things are missing from these accounts and they are what the present framework adds.

The first is that the operation is a demonstration addressed to a third party, and that the third party is where the alteration occurs. Labelling theory locates the label in the community and does not ask why the labelling is characteristically performed in public, nor what the audience is for. It treats the audience as the aggregate of those who come to hold the label. The present account holds that the audience is not a recipient of the label but the site of its installation, and that the operation is therefore not a communication about the labelled subject but an alteration in what the community is able to distinguish, after which it is able to distinguish so about anyone, and not only about him.

The second concerns what a bystander’s silence is. On labelling theory it is a failure to intervene, which is to say a non-event, and the literature on the bystander effect treats it accordingly, as an absence explained by diffusion of responsibility [14]. On the present account it is not an absence at all. It is the completion of the operation, since the demonstration is conducted precisely in order to establish that the categorization may be made without objection, and a silence is the evidence that it may. To call it a failure to intervene is to have described the constitutive element of the mechanism as though it were a missing one.

12.5 Attribution fails, and responsibility does not

The categorizer, examined closely, is not the author of the configuration either. He must sustain it continually, since a demonstration that ceases to be performed ceases to persuade, and he cannot easily withdraw, since the position he occupies is the position that would otherwise be available for him to occupy. He is a participant in an arrangement no participant authored.

The consequence for intervention is the one §5 predicted. To remove the participant is to leave the role, and the role will be filled. But the consequence must not be permitted to become an exculpation. That responsibility cannot be located exhaustively in individual psychology does not entail that there is no responsibility. It entails that responsibility has been sought at the wrong layer.

§13 Institutional Indifference Forecloses Particularity and Keeps Revision Open

This case is different in kind from the six others, and the difference is the pivot of the paper.

13.1 Setting aside particularity is the institution’s purpose

Relations are categorized. The subject is processed as an instance of a type, and his particularity is set aside deliberately and as a matter of principle.

This must be stated without hedging. The moral force of legal equality consists precisely in the setting aside of particularity. To treat like cases alike is to refuse to generate the subject as a singular, and the refusal is not an unfortunate side-effect of an otherwise admirable arrangement. It is the arrangement. An administration that attended to the particularity of every case would not be a better administration. It would be a system of favor.

If the reconfiguration of subjects were itself the wrong, then legal equality would be a wrong, and universalism would be a wrong, and procedural justice would be a wrong. This is absurd, and its absurdity is the reason a criterion is required.

13.2 Sacrifice of the part for the whole is the form every imposition takes

It is tempting to justify the case by observing that the local generation of subjects is sacrificed so that the general generation may be sustained. The observation is available, it is not obviously false, and it must nevertheless be refused.

The form of the sentence, that the part is given up so that the whole may continue, is the form in which every systemic imposition has justified itself, and an account that accepts it as an explanation has forfeited its standing to reject it as a justification. §18 states the constraint and the reason for it.

The legitimacy of this case does not rest upon its service to the whole. It rests upon something else, and identifying what that something is is the task of §17 and §20.

13.3 Weber’s description is accepted and the critical inference declined

Weber’s account of bureaucratic rationality states the case with a clarity that has not been improved upon: administration sine ira et studio, without hatred and without affection, is not a deformation of administration but its perfected form, and the elimination from official business of love, hatred, and all purely personal feeling is precisely what makes calculable administration possible [18].

The critical literature has drawn the darker inference. Bauman’s argument that the bureaucratic mode of organization is not merely compatible with atrocity but supplies its conditions, and Arendt’s account of the functionary who does his work without thinking, both hold that the setting aside of the particular is what permits the unthinkable to be administered [9, 1].

The present account accepts Weber’s description and declines the critical inference, and the ground on which it declines it is the paper’s contribution here. The critical accounts treat categorical administration as degrading because it sets aside the particular, from which it follows that the arrangement is dangerous in proportion to its success. If that were right, the remedy would be less administration, and the conclusion is one that few of the critics have been willing to draw and that none has been able to make coherent.

The present account holds that the setting aside of the particular is not what makes an arrangement degrading, and it must hold this, since to hold otherwise is to condemn legal equality. What distinguishes the administrative case is that its determinations remain revisable: there is an appeal, a hearing, a proceeding in which the categorization may be contested, and a legislature by which the categories themselves may be altered. The arrangement that reconstitutes the subject furnishes the means by which he may be reconstituted again.

This yields a prediction that the critical accounts do not, and it is a prediction on which the present account may be tested. It is not the categorical form that renders an administration dangerous. It is the closure of the channels of revision, and an administration that categorizes with the greatest severity while keeping those channels open is, on this account, less dangerous than one that categorizes gently and has closed them.

§14 The Loss of Trust Cancels the Common Future While Cooperation Persists

14.1 Trust is the capacity to generate a common future

The common future is cancelled. The parties continue to transact, to cooperate, and to communicate, and an observer would find the relation intact by every behavioral measure. What has ceased is that they generate anything forward together.

It has been said, and it was correct as far as it went, that trust is not prediction. What was not said is what trust positively is, and the omission left the account thin. Trust is the capacity to generate a future in common. Its loss is therefore not a downward revision of one’s estimates concerning another’s conduct, since such a revision is compatible with a perfectly functional relation, but the withdrawal of the relation from the business of jointly producing what neither party could have specified in advance.

What is lost when trust is lost is the capacity to generate a common future. Cooperation, exchange, and communication may all continue unimpaired, which is why the loss is invisible from outside and is frequently invisible to the parties themselves.

The subject is left without a counterpart for joint generation. He may still act, and his acts may still be reciprocated, but they are reciprocated within a settled frame rather than taken up into anything new. His generation contracts to the repetition of what is already specified.

14.2 The degradation is assembled entirely from correct judgments

This is the feature that distinguishes it from the six others, and it is the feature that will trouble the reader. There need be no occasion on which anything occurred. Each successive withholding is a reasonable revision in light of the evidence, no step is mistaken, and no party has erred. The degradation is composed entirely of correct judgments.

A degradation may be assembled from steps each of which is rational and none of which is wrong. If this is so, then rationality is, under certain relational conditions, among the mechanisms by which subjects are reconfigured, and this is a consequence the account must accept rather than evade.

14.3 Trust as estimate registers no loss here; trust as joint generation registers it

Two literatures treat trust and they treat it as two different things.

The decision-theoretic tradition treats trust as an estimate. To trust is to act on a judgment concerning another’s likely conduct, and the rationality of trusting is assessed by the accuracy of the judgment and the payoffs that attend it [8]. On this account, the withdrawal of trust in the face of adverse evidence is not a degradation at all. It is learning, and to describe it as a loss is to have misdescribed a correct inference as an injury.

The sociological tradition treats trust as a mechanism for the reduction of complexity. Luhmann’s account holds that trust permits a party to act as though a range of possibilities were foreclosed, and thereby makes possible undertakings that could not survive the entertainment of every contingency [10]. This is nearer to the mark and it remains an account of what trust does for a party rather than of what it does between them.

The present account holds that trust is the capacity to generate a common future, and the difference from both is not terminological. On the decision-theoretic account there is nothing to explain in the configuration described above, since each withholding was warranted and the outcome is what warranted withholdings produce. On the present account there is a great deal to explain, and the explanation is that a relation may satisfy every party’s estimates and cease to generate anything the estimates did not already contain. What is lost is not a belief about the other. It is the standing to undertake, with him, what neither could have specified in advance.

This is the paper’s most uncomfortable contribution and it is stated as such. If trust is an estimate, then rational updating cannot degrade a relation. If trust is the capacity for joint generation, then a sequence of rational updatings can extinguish it, no step being mistaken and no party at fault. The account is committed to the second, and it accepts the consequence, which is that rationality is, under certain conditions, among the mechanisms by which relations are reconfigured. It does not follow that the parties acted wrongly. It follows that the vocabulary of act and fault has, once again, nothing to attach to.

§15 The Paths Diverge at the Channel of Revision

The taxonomy of §7 distinguished the cases by what each removes, by the form of attribution each admits, and by what stands in the way of repair. Having examined them, a fourth column may now be added, and it is the column on which the remainder of the paper turns.

CASE	ALTERATION AT LAYER ONE	PATH TO THE SUBJECT
Betrayal	the relation’s past is rewritten	he was mistaken about the conditions of his own generation
War	the relation is recoded as adversarial	he is generated only as enemy, with heightened dynamics
Abandonment	the system ceases to include him in its future	his generation has no recipient; singularity is lost
Withdrawal of response	response is withheld, in coordination	he must supply his own account, and the faculty supplying it is thereby disabled
Public categorization	the rule of discrimination is publicly rewritten	he is generated only as the category, and resistance confirms it
Institutional indifference	relations are categorized by principle	his particularity is set aside; the channel of revision remains
Loss of trust	the common future is cancelled	he has no counterpart for joint generation

One entry in the third column is unlike the others, and it has been marked. Six of the seven paths close the possibility that the subject may be generated otherwise. The seventh does not, and it is the case in which the alteration at the first layer is the most deliberate of all.

Whatever distinguishes it from the six is not the gentleness of the alteration, and is not the good intentions of those who administer it. The section that follows argues that what distinguishes it is the only thing that could.

§16 The Candidate Criteria Fail, Including the Framework’s Own

The cases have established the difficulty in its sharpest form. Six paths close the possibility of the subject’s being generated otherwise; a seventh does not, and the seventh is the one in which the alteration of generative conditions is most deliberate and most complete. A criterion is therefore required, and before it is offered the candidates that will not serve should be eliminated, since each is more attractive than the criterion that survives.

16.1 Reconstitution is universal

This was settled in §5 and is repeated because it is the candidate to which the argument most naturally reverts. Every relation reconstitutes its participants. If reconstitution were the wrong, then relation would be the wrong, and the account would have condemned the condition of its own subject matter.

16.2 Generation may increase under degradation

This was settled in §5 also, and it is worth pressing. A degraded configuration may generate with greater energy than the configuration it displaced, and commonly does. The observation that a relation is generating vigorously is therefore not evidence that it is sound, and a criterion framed in terms of the quantity of generation would return a verdict of health in the cases where health is most absent.

16.3 Foreclosure of future generation reinstates the utilitarian norm

This candidate is the most tempting and it must be refused with the most care, because it appears to be the framework’s own.

The proposal would be that a symbol which fixes a subject in a position, closing the future he might have had, is a degrading symbol, while one which leaves that future open is not. The proposal is circular, and worse, it reintroduces the very norm the framework exists to displace.

Suppose the wrong done to the abandoned consists in the closure of his future generation. Then a subject who genuinely will not generate again, and such subjects exist, and their existence is not a marginal complication, is not wronged by the closure, since there was nothing there to close. The criterion protects the productive and abandons precisely the party for whose sake the account was constructed.

To replace usefulness with generativity is still to legislate a utilitarian norm. It preserves the structure in which a subject deserves recognition because he will bring something about, and merely renames what he will bring about. “He deserves recognition because he can still generate” and “he deserves recognition because he can still contribute” are the same sentence.

The consequence is one the framework must accept and must not soften. Generativity is the phenomenon that this account describes. It cannot also be the reason for which recognition is owed, and an account that made it so would have dismantled itself in the act of grounding itself.

§17 A Relation That Cannot Be Revised Generates One Subject Repeatedly

What distinguishes institutional indifference from the six others is not that it reconstitutes less, nor that it generates more, nor that it is more gently administered. It is that the subject it reconstitutes may still be reconstituted again, and by common experience.

The criterion. Degradation is present when the reconfiguration of a subject’s conditions of generation forecloses the possibility that those conditions may be further revised through common experience.

The criterion was not selected. It is entailed by the premise, and the entailment should be exhibited rather than asserted.

If the subject is what is generated in the relation, then his continued generation requires that the relation remain capable of taking up what is new. A relation that can no longer be revised by common experience has not ceased to generate, this was the burden of the first correction, but it can generate only what it has already specified. It produces, and it produces the same subject, indefinitely.

A relation that can no longer be revised through common experience does not stop generating subjects. It generates one subject, repeatedly. And that is what it is for a condition of generation to have been closed.

17.1 Administration keeps the channel open; the six others close it

Administration by category reconstitutes the subject and sets aside his particularity, and it does both by design. It nevertheless leaves the channel of revision open, and the openness is not incidental to it but is among its constitutive commitments: there is an appeal, there is a hearing, there is a proceeding in which the categorization may be contested, and there is a legislature by which the categories themselves may be altered. The subject may be reconstituted again, and the means by which he may be are furnished by the arrangement that reconstituted him.

Education reconstitutes the student more thoroughly than any of the seven cases reconstitutes anyone, and no one proposes that it is thereby a violence. The student may dispute the teacher, may exceed him, and may become something the teacher did not envisage. The reconstitution is complete and the revision remains available.

The six others close the channel, and they close it by different routes. In betrayal the revision would have to reach a past that is no longer available for revision. In war the subject may be revised only within the category of the adversary. In abandonment there is no counterpart to take up a revision. Where response is withheld, the faculty that would perform the revision has been enlisted in producing what would have to be revised. Where categorization is public, any attempt at revision is received as confirmation. Where trust has been lost, the joint generation in which a revision would consist has itself been withdrawn.

Institutional indifference passes and public categorization fails, and not because the first is gentle and the second cruel. The first keeps open the channel through which its own determinations may be revised. The second closes it, and closes it as the condition of its operation.

17.2 A harsh relation may pass and a benevolent one may fail

Three consequences follow, and none is comfortable.

The first is that a relation may be harsh, categorical, and impersonal, and may not be degrading. This will strike some readers as a defect. It is the price of a criterion that can distinguish a court from a mob, and any criterion that could not make that distinction would be worthless whatever else it could do.

The second is that a relation may be warm, attentive, and well-intentioned, and may be degrading, if the attention it supplies is of a kind that forecloses the subject’s revision of what he is taken to be. The account is committed to this and the commitment should be stated. Benevolence is not a defense.

The third is the one raised by the case of §14. If a degradation may be assembled from steps each of which is individually rational, then the criterion condemns a configuration that no participant erred in producing. The account accepts this. It holds that responsibility is not thereby dissolved but is relocated to the layer at which the conditions of generation are determined, and it acknowledges that the machinery for assigning responsibility at that layer is not well developed, in this paper or elsewhere.

§18 The Wall Between Dynamical Description and Normative Justification

One constraint governs the use of everything above, and it is stated separately because the vocabulary of this paper is unusually exposed to its violation.

The account describes mechanisms. It observes that certain reconfigurations of subjects can be found, across many systems and at many scales, to sustain the conditions under which those systems continue to operate. The observation is available and it is worth pursuing.

What may not be done is to convert the observation into a reason.

That a reconfiguration sustains the conditions of a system’s continued generation is a description. It is not a justification, and it must never be permitted to become one. The sentence “the part was given up so that the whole might continue” is the form in which every systemic imposition has ever defended itself, and an account that accepts it as an explanation has surrendered its standing to refuse it as a defense.

The exposure is specific to a framework of this kind and it should be understood precisely. A vocabulary organized around generativity can state, in its own terms and with perfect fidelity, that the destruction of a subject’s generative capacity enlarged the generative capacity of the system containing him. In a coupled system this may be true. It is a proposition about dynamics and it may be established or refuted by evidence.

It supplies no reason for anything.

The account is therefore obliged to be able to say, and does say, that a reconfiguration which forecloses a subject’s revision is a degradation even where it can be shown to have enlarged the generative capacity of the whole. This is sayable only because recognition has not been grounded in generativity, and it is the second time in this paper that the refusal of §16 has done necessary work. A framework that grounded recognition in generativity would find the sentence unsayable and would discover, too late, that it had constructed an apparatus for the justification of exactly what it was built to name.

§19 The Five Principles Are Satisfiable Within a Closed Relation

The criterion of §17 does not stand alone. It joins a set of principles the author has developed elsewhere, under the heading of an ethics of relational reproduction, and the relation between the criterion and those principles requires stating, since the criterion is here proposed as an addition to them and an addition must be shown to be needed.

The five principles state the conditions under which a relation may reproduce itself justly.

The ethics of relational reproduction: five principles. A just relational reproduction satisfies five necessary conditions. Generativity: it does not foreclose the participants’ capacity to keep generating. Co-reproduction: the relation is reproduced through joint participation and not by the decision of one party alone. Non-exploitation: it does not depend on the instrumentalization or sacrifice of any participant. Restraint and non-entrenchment of power: since the generation of value carries power, it generates at the same time the counter-power that restrains power and prevents its entrenchment. Subject preservation: it does not purchase unity by effacing the other as a subject; coincidence is not the aim, and the preservation of difference and of the subject is what makes continued generation possible.

The five are not in question here, and the argument of this paper has been conducted within them. What is in question is whether they suffice, and the case of §13 shows that they do not.

19.1 Administration by category satisfies all five and forecloses every subject it touches

Administration by category may be examined against each of the five, and it satisfies each.

It does not foreclose the participants’ capacity to keep generating. It sustains that capacity at the scale of the arrangement as a whole, and it is ordinarily instituted for that purpose. It is reproduced through joint participation, since the categories are legislated and the legislation is answerable. It does not depend upon the sacrifice of any participant, and it is characteristically designed to ensure that no participant is sacrificed to the convenience of any other. It restrains power and prevents its entrenchment, and this is not merely compatible with it but is very nearly its definition. And it does not efface the other as a subject; it declines to attend to his particularity, which is a different thing, and it declines equally for all.

Administration by category satisfies the five principles and collapses the singularity of every subject it touches. Public categorization violates the third and the fifth, and collapses the singularity of one. The two are formally alike in what they do to the subject, and the five principles separate them by pointing to violations in the one case that are absent in the other.

This is not nothing, and it should not be dismissed. But it does not reach the difficulty, and the reason is worth stating exactly.

The five principles are conditions upon a relation as it presently stands. Not one of them asks whether the relation remains open to being altered. A relation may satisfy all five and be closed, and a closed relation that satisfies all five will continue to satisfy them, since it is the relation itself that supplies the terms in which the satisfaction is assessed.

The point may be put as a difficulty about self-assessment. Each of the five is a predicate applied to a relation, and the application requires a description of the relation under which the predicate can be evaluated. Where the relation has closed the channel through which its own descriptions might be revised, it furnishes the descriptions itself, and it furnishes them in terms that its own operation has fixed. A configuration that has foreclosed the revision of its categories will report that it does not exploit, since exploitation is defined in terms it has fixed; that it preserves subjects, since what a subject is has been fixed likewise; and that it restrains power, since the power it recognizes is the power its own descriptions admit.

The five principles are not thereby refuted. They are shown to be insufficiently protected, and what protects them is what must now be added.

§20 Revisability Governs the Five Principles Themselves

Revisability. A relation must preserve the possibility that the conditions of its own generation be revised through common experience. This is the sixth principle of the ethics of relational reproduction, and it is proposed here as this paper’s addition to that ethics.

The ground of the principle was given in §17 and is restated in a sentence. A relation that can no longer be revised through common experience does not cease to generate subjects. It generates one subject, repeatedly, and that is what it is for a condition of generation to have been closed.

20.1 The sixth principle governs the five

Its relation to the five is not that of an item added to a list, and it is not a further excellence a relation might exhibit alongside the others. It does not govern relations. It governs the five principles themselves, and it is for this reason that it belongs to a different order.

The argument is short and it is general. Every principle requires interpretation, and the interpretation is performed by those to whom the principle applies. Where the channel through which that interpretation might be revised has been closed, the principle does not thereby cease to be satisfied. It is satisfied, and it is satisfied in terms the closure has fixed. What was a constraint upon a relation becomes an instrument of the relation, and the conversion requires no bad faith and may be accomplished by parties acting in perfect sincerity.

Any principle that has lost its revisability may degenerate into an instrument of oppression. Not because the principle was mistaken, but because the party who interprets it is the party it was meant to constrain, and a constraint interpreted by its object is not a constraint.

Each of the five, held without the sixth, degenerates in a manner that may be stated exactly, and the degeneration is not a failure to satisfy the principle but a satisfaction of it under a description the closed relation supplies.

Generativity becomes I define what counts as generation. This paper has insisted throughout that a degraded configuration generates copiously, and the insistence bears its fruit here: a relation that produces a single subject with unfailing fertility satisfies the first principle by every measure the first principle supplies. The principle does not distinguish fecundity from repetition, and it cannot, since the distinction is unavailable at the moment of assessment and becomes available only across the revision of what is being assessed.

Co-reproduction becomes jointly, but only in the manner already fixed. That the parties act together does not entail that anything they do together may alter what they are doing. A relation may be reproduced by every participant and changed by none of them.

Non-exploitation becomes I define what counts as exploitation. The configuration of §14 exhibits the difficulty in its purest form: no participant instrumentalizes any other, no step is mistaken, and the degradation is assembled entirely from correct judgments. There is no exploiter to be found and the principle has nothing to attach to. What is required is not the search for a violation but the observation that the relation can no longer be revised by what the parties experience within it.

Restraint of power is defeated by its own success, and the counter-power congeals into a power. A counter-power instituted to restrain power is itself a structure, and structures entrench. The account of asymmetry given here bears on the point directly: the asymmetry of a relation continuously produces power, and what prevents that power from settling is not a countervailing structure but the continuous absorption of the power into the relation’s ongoing generation. Absorption is work, and work stops. Where it stops, the power need not be seized. It is subsided into, and no one need have done anything for it to have been.

Subject preservation becomes only a subject answering to my description is worth preserving. To hold the other as a subject, and to hold him as the subject one has determined him to be, are not distinguishable at any single moment. They are distinguishable only across time, and only by whether he may become someone else.

The five principles ask whether a relation is presently just. The sixth asks whether the relation still permits that question to be raised again. Without the sixth, the five may be satisfied within a closed system, since it is the system that supplies the terms in which their satisfaction is assessed.

This is the general form of what the cases found in the particular case of the symbol. It was said there that what is dangerous in a symbol is not that it exists but that it can foreclose its own revision. The sixth principle is that observation raised to the level of the norm. It is not directed at symbols. It is directed at every norm whatever, and at every norm that this framework has itself proposed.

20.2 Power is absorbed into the relation’s ongoing generation

The fourth principle, as formerly stated, holds that the generation of value carries power and that it therefore generates at the same time the counter-power that restrains power. The account of this paper suggests an amendment, and the amendment is offered for consideration rather than asserted.

Power is not, on the account given here, restrained by a counter-power that stands over against it. It is absorbed, which is to say that it is taken up into the relation’s ongoing generation and becomes part of what the relation generates in common. The teacher’s power over the student is not held in check by an opposing power the student wields. It is absorbed by the joint work of learning, by the reciprocal correction that work permits, and by the growth in which the asymmetry is consumed. We do not notice the power because it is being used up.

Power need not be balanced. It needs to be absorbed, and absorption is continuous work rather than a standing arrangement. Oppression need not be manufactured; it need only cease to be absorbed.

The amendment matters because it locates the failure in the right place. A relation does not become oppressive at the moment someone seizes an advantage. It becomes oppressive at the moment the work of absorption stops, and the advantage that was always present ceases to be consumed. This is why good relations degrade without anyone doing wrong, and it is why the sixth principle is not an optional refinement of the fourth but the condition under which the fourth can be sustained at all. Absorption requires that the relation be able to take up what is new. A relation that cannot be revised cannot absorb.

20.3 The principle applies to the framework that states it

A meta-principle that exempted its author would not be a meta-principle. It would be a device for the protection of one norm against the scrutiny it demands of every other, and the exemption would be the first instance of the closure the principle was formulated to name.

The framework is therefore bound by it.

If generative relational being cannot accept the revision of its own theory through new common experience, then it has violated its own sixth principle. The principle is not directed at symbols, nor at the norms of others. It is directed at every norm whatever, and the first norm to which it applies is this one.

The self-application is not a rhetorical gesture and it carries a cost that should be stated. It means that the five principles are not a foundation, and that the sixth is not a completion. It means that a further principle may be required, that one of the six may be shown to be mistaken, and that the framework has no standing to resist such a showing on the ground that its principles are its own.

This is the position the framework in fact holds, and it holds it for a reason that is not modesty.

20.4 The norm is itself in historical motion

The ethics of relational reproduction was not derived from an axiom and is not presented as the discovery of a permanent truth. It was developed as an account of the conditions under which relations reproduce justly, under material conditions that the author has elsewhere argued are themselves in motion: the site at which value is anchored has shifted before, and appears to be shifting again, and the norms adequate to one configuration of that anchoring are not thereby adequate to another.

A norm produced within a historical movement remains answerable to that movement. It does not stand outside the process that generated it, surveying it; it is a part of the process, and it is revised as the process discloses what the norm could not have anticipated.

A norm generated within a historical movement must remain revisable by that movement. To hold otherwise would be to claim for the norm a position outside the process that produced it, which is precisely the position that the framework denies to the subject and cannot consistently claim for its own principles.

The addition made by this paper is offered in that spirit and is itself an instance of it. The five principles were adequate to the cases against which they were framed. They were shown, in §19, to be inadequate to a case that this paper’s analysis brought into view, and the inadequacy was not a defect in their formulation but a limit of the experience from which they were drawn. A sixth is accordingly proposed, and it is proposed under the same condition it imposes: that it be revisable by what is subsequently found.

That the framework has been enlarged in this manner is not a concession. It is the operation of the sixth principle upon the framework itself, and it is the only evidence available that the principle is more than a demand made of others.

The form of the present paper is governed by the same consideration, and the point is easily mistaken. To hold a claim revisably is not to hold it tentatively, and it is not satisfied by an author’s willingness to issue a second version. A revision made by an author is an exercise of the interpretive authority the principle is concerned with, and an account revisable only by the hand that wrote it has been opened to no one. What the principle asks is that the conditions of refutation be supplied to the reader, so that the account may be defeated by what he brings and not merely amended by what its author later concludes. The claims of this paper are advanced without hedging, and the means of defeating them are set out in §22. That is what it is for a norm to be held revisably, and it is the whole of what the form of this paper is meant to accomplish.

20.5 The principle is necessary and insufficient

It must be said plainly, since a principle that claimed more would be worse than one that claimed less.

The sixth principle is a necessary condition and not a sufficient one, and it does not constitute a practice. It states that the channel of revision must remain open. It does not state how the asymmetries a relation produces are to be absorbed, and an open channel through which nothing passes is an open channel in a relation that is degrading. It says nothing to a subject who has no reserves elsewhere, and for whom the availability of a channel he has not the standing to enter is no availability at all. And it cannot restart what has stopped: revisability is a property of a relation that is generating, not a therapy for one that has closed, and to name the property is not to have supplied the means of its restoration.

What the practice would consist in, by what arrangements the surplus a relation generates is returned to those who generate it, by what means a gaze that has become singular is made reciprocal, and what is to be done where the symbol has already congealed, is not settled here. §21 records these as questions the paper leaves open, and they are the questions on which the value of the whole account will finally turn.

§21 Open Questions the Argument Leaves

Four questions are raised by the argument and are not answered in it. They are stated here rather than left to be discovered, since an account that conceals the questions it cannot answer has not been given a boundary but a disguise.

21.1 The degrading symbol generates copiously

It has been said, at several points and without elaboration, that a degraded configuration generates copiously. The observation was used and was not examined, and it will not survive examination unchanged.

A symbol that fixes a subject in a category does not merely persist. It elaborates. It produces fresh descriptions, discovers new applications, and furnishes those who employ it with an inexhaustible interpretive work in which they are, in a perfectly ordinary sense, creative. The consequence is uncomfortable and it should be faced. One cannot oppose such a configuration by generating more, since it is already generating, and generating with an inventiveness that a countervailing effort will not easily match.

What distinguishes the generation of a symbol that closes from the generation of one that opens is therefore not the quantity of generation and cannot be. The author’s provisional view is that the distinction lies in where the generation unfolds, whether in the interpreter, who is thereby enlarged, or through him, in which case what he produces is the symbol’s confirmation and not his own. But this is a claim about the phenomenology of interpretation and it requires an argument this paper has not made.

21.2 The subject cannot confirm itself

The account has said that a symbol deprives the interpreter and has not said why the interpreter remains.

The answer, if there is one, is likely to be that the subject cannot confirm himself, and that a confirmation from outside is not a comfort he may forgo but a condition of his continuing to be a subject at all. If this is so, then leaving the configuration is not an option he declines but an option he does not have, and the choice presented to him is not between comfort and integrity but between his feeling and his existence, which is not a fair choice and is not usefully described as one.

This is a claim about the constitution of subjects and it is a large one. It is compatible with the premise of the argument above and it is not entailed by it, and the present paper is not the place to establish it.

21.3 A closed relation reports that matters are in order

The criterion of §17 is stated in terms of revisability, and revisability is a structural property that can be examined from outside. There is reason to doubt that the examination can always be conducted from outside.

A configuration that has closed the channel of revision will not report the closure. It will report that matters are in order, and it will do so sincerely, since the language in which the closure could be named is the language the closure has removed. If this is right, then the criterion, though correct, may not always be applicable by the means the paper has supplied, and something else may be required to detect what it correctly defines. What that something is, the author suspects, is not a further structural test but a form of perception, and the suspicion is recorded without being argued.

21.4 The conditions of generation have been treated as given

The account has treated the conditions of relational generation as given, and has asked what happens when they are altered. It has not asked how those conditions come to be what they are, nor by what dynamics they change of themselves, nor what determines, when a relational system can no longer absorb the asymmetries its own generation produces, whether it will reconstitute its arrangements, reconstitute its membership, or reconstitute its adversaries.

That question is the one the author takes to be the most important the framework faces, and it requires a formal apparatus the present paper does not possess. It is the subject of the metamodel developed separately, and it is the reason that work was undertaken.

§22 The Argument Stands or Falls With Its Premise

The argument is deductive and its structure should be evident. A premise concerning the constitution of subjects entails a consequence concerning what the degradation of a relation is. Two corrections keep the consequence from vacuity: that degradation reconfigures rather than diminishes, and that reconstitution as such is not the wrong. A criterion is then required, and the criterion is shown to be entailed by the premise rather than chosen by the author. The cases exhibit the paths by which the consequence transmits, and they do not establish it.

The paper is therefore conditional upon its premise in a way that admits of no repair from within. A reader who holds that the subject is prior to the relation, that it possesses its properties independently and enters into dealings that leave its constitution untouched, will find that nothing in the argument obliges him. The cases will appear to him as a collection of harms, differently administered, and the criterion will appear as one moral intuition among the available ones, with no particular claim upon him. He will not be wrong to think so. He will be disagreeing with the premise, and the premise is where the disagreement belongs.

What the paper claims to have shown, conditionally upon that premise, is five things.

That the omissions of the vocabulary of act and fault identified in §2 are not marginal but structural, and that they follow from its assumption of a subject prior to relation.

That relational degradation admits of a unified account which does not abolish the differences among its forms, the unity residing in what is altered, the conditions of a subject’s generation, and the differences residing in the paths by which the alteration transmits.

That the reconstitution of subjects cannot be the criterion of degradation, that generativity cannot be its criterion either without collapsing into the utilitarian norm the framework exists to displace, and that revisability through common experience is entailed by the premise as the criterion that remains.

That the five principles of the ethics of relational reproduction, which the criterion joins, are insufficiently protected without it, since a relation that has closed the channel of its own revision supplies the terms in which its satisfaction of those principles is assessed. Revisability is accordingly proposed as the sixth principle of that ethics, and it is proposed as a meta-principle and not as a further virtue standing alongside the five: it governs not relations but the five principles themselves, since any principle that has lost its revisability may degenerate into an instrument of the party it was framed to constrain. This is the paper’s addition to the framework’s normative apparatus, and the addition is offered under the condition it imposes, namely that the framework which states it is bound by it, and that a norm produced within a historical movement remains answerable to that movement.

And that distributive justice is limited in principle, and not merely in execution, with respect to the case of abandonment, since distribution operates within the conditions of generation and what abandonment alters is the conditions themselves.

The sharpest of these is the last, and it is the one on which the author would most welcome correction. If the repair of abandonment can be shown to be achievable by transfer, if the supply of resources can be shown to restore the singularity whose loss the account describes, then §10 fails, and with it the strongest practical consequence the paper carries.

Objection, counter-example, and correction are welcome at huangwanhong@serendip.ngo. The argument has been set out at this length in order to make it easy to disagree with, and the account is held open for the revision that the sixth principle requires of it.

§23 Revisability Requires That the Conditions of Refutation Be Supplied

This is a discussion paper, and one objection to the designation should be met at the outset, since it is the obvious one. A preprint may be revised. If the paper’s claims are provisional, a subsequent version will say so, and nothing is gained by announcing provisionality in a title.

The objection would be decisive if the designation meant that the author intends to revise. It does not mean that, and the intention to revise is not what the sixth principle of this paper requires. A revision issued by an author is an exercise of the author’s interpretive authority. It leaves the terms of assessment where they were, namely with him, and an account that can be corrected only by the party who wrote it has not thereby been opened to anything. What the principle requires is not that the author remain willing to change his mind but that the reader be given the means to defeat him.

The paper is written accordingly. It states, in §22, the five claims it makes and identifies which of them it is most exposed on. It names the single finding whose refutation would carry the strongest practical consequence of the whole account down with it. It records, in §21, the four questions it cannot answer and the reasons its instruments are inadequate to them. These are not the gestures of modesty. They are the conditions of refutation, and they are set out because an account that supplies no such conditions has not made itself revisable by declaring that it might one day be revised.

The argument is deductive in form. It begins from an ontological premise, draws a consequence that the premise entails, and then examines seven cases in which the consequence may be observed to transmit. The cases are not evidence for the thesis and are not offered as such. If the premise is rejected, no accumulation of cases will restore the argument; if the premise is granted, the principal claim follows without them.

The cases earn their place for a different reason. A thesis that a relation’s degradation is a reconfiguration of its subjects would be empty if the paths of that reconfiguration could not be distinguished from one another, since a unification that abolishes its own differences explains nothing. The function of the cases is therefore to keep the differences, and its adequacy is to be judged by whether the seven paths are genuinely distinct and whether each is recognizable in the terms the argument supplies.

The concluding sections carry the weight. It argues that the reconstitution of a subject is not the wrong, that a criterion is therefore required, and that the criterion is entailed by the ontology rather than selected by the author. The test of this part is a single case. Institutional indifference reconstitutes subjects by category, forecloses their singularity, and is nevertheless the condition of a form of justice we have no wish to abandon. An account that cannot distinguish it from the others has not produced a criterion but a complaint. Whether the distinction offered here holds is the question on which the paper turns.

Two matters are excluded by design. The paper does not argue at the level of extreme cases, since an argument that draws its force from the emotional weight of its examples is not carrying its own weight, and the mechanisms at issue are visible in ordinary relations without recourse to atrocity. And the paper does not derive any political conclusion from its dynamical claims. The distinction between the description of a mechanism and the justification of what the mechanism produces is maintained throughout, and it is maintained deliberately, since a vocabulary of generativity is unusually exposed to the collapse of that distinction. §18 states the constraint and the reason for it.

Correspondence, objection, and counter-example are welcome at huangwanhong@serendip.ngo.

---

End of Discussion I. The conditions under which subjects are generated have been treated here as given, and the question of how they themselves change has been deferred. It is the question that matters most, it is not answered here, and it is the subject to which a subsequent discussion would have to turn.

References

[1] Arendt, H. (1963). Eichmann in Jerusalem: A report on the banality of evil. Viking Press.

[2] Baier, A. (1986). Trust and antitrust. Ethics, 96(2), 231–260.

[3] Bourdieu, P. (1991). Language and symbolic power (G. Raymond & M. Adamson, Trans.). Harvard University Press.

[4] Fanon, F. (1967). Black skin, white masks (C. L. Markmann, Trans.). Grove Press. (Original work published 1952.)

[5] Fraser, N. (2003). Social justice in the age of identity politics. In N. Fraser & A. Honneth, Redistribution or recognition? A political-philosophical exchange (pp. 7–109). Verso.

[6] Eglash, R. (2016). An introduction to generative justice. Teknokultura, 13(2), 369–404.

[7] Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Prentice-Hall.

[8] Hardin, R. (2002). Trust and trustworthiness. Russell Sage Foundation.

[9] Bauman, Z. (1989). Modernity and the Holocaust. Cornell University Press.

[10] Luhmann, N. (1979). Trust and power. Wiley.

[11] Lemert, E. M. (1951). Social pathology. McGraw-Hill.

[12] Becker, H. S. (1963). Outsiders: Studies in the sociology of deviance. Free Press.

[13] Margalit, A. (2017). On betrayal. Harvard University Press.

[14] Darley, J. M., & Latané, B. (1968). Bystander intervention in emergencies: Diffusion of responsibility. Journal of Personality and Social Psychology, 8(4), 377–383.

[15] Honneth, A. (1995). The struggle for recognition: The moral grammar of social conflicts. Polity Press.

[16] Sen, A. (1999). Development as freedom. Oxford University Press.

[17] Silver, H. (1994). Social exclusion and social solidarity: Three paradigms. International Labour Review, 133(5–6), 531–578.

[18] Weber, M. (1978). Economy and society (G. Roth & C. Wittich, Eds.). University of California Press. (Original work published 1922.)

[19] Wendt, A. (1992). Anarchy is what states make of it: The social construction of power politics. International Organization, 46(2), 391–425.

[20] Williams, K. D. (2007). Ostracism. Annual Review of Psychology, 58, 425–452.

[21] Waltz, K. N. (1979). Theory of international politics. Addison-Wesley.

[22] Lacan, J. (1998). The seminar of Jacques Lacan, Book XI: The four fundamental concepts of psychoanalysis (A. Sheridan, Trans.). Norton. (Original work published 1973.)

[23] Marx, K. (1998). Theses on Feuerbach. In The German ideology. Prometheus Books. (Original work written 1845.)

[24] Ostrom, E. (1990). Governing the commons. Cambridge University Press.

[25] Huang, W. (2025). Responding to the crises of symbol and subject in modernity: Towards a generative nomos of relational presence. SSRN. https://doi.org/10.2139/ssrn.5752104

[26] Huang, W. (2026). Relational wealth: Definition and characteristics. Knowledge Commons. https://doi.org/10.17613/nq7js-vz283

[27] Huang, W. (2026). A theory of relational reproduction: Definition and ethical principles. Knowledge Commons. https://doi.org/10.17613/qncwr-zbt82

[28] Huang, W. (2026). Relational value under the symbolic crisis. Knowledge Commons. https://doi.org/10.17613/bm32z-aqt14

[29] Huang, W. (2026). A metamodel for ontologically heterogeneous social dynamics. Part I. Serendip Commons Society.

---

中文

关系性退化

关系性主体的重构

讨论之一

黄万宏 · huangwanhong@serendip.ngo

---

摘要

问题。 有一类关系性变形抵抗着人们通常用以描述关系性伤害的那套词汇。其中有些没有可指认的加害者，有些在毫无恶意的情况下发生，还有些完全由任何一方都没有作错的判断所拼装而成。在这类情形中，行为、意图与过错的词汇没有任何可供附着之处；本文论证，这是该词汇携带却未加陈述的一个前提所致，即：主体先于关系。

进路。 本文自生成性关系存在的本体论出发进行演绎推论。依该本体论，主体并不是一个站在关系之中的实体，而是在关系之内被生成的东西。一个后果由定义而非由推论得出：任何关系的退化都是主体被生成的条件的改变。本文考察七个案例，它们不是作为论题的证据，而是作为这一改变得以传递的诸路径的展示。

发现。 有两项修正与自然的表述方式相抵触。退化并不削弱主体的动力学，而是重构它，而被重构的动力学往往更为强烈。并且，主体的重构不可能就是那个错，因为每一种关系都在重构其参与者。因此需要一个判准，而这个判准不能以生成性来表述，否则将坍缩为该框架为之而存在、意在取代的功利主义规范。该本体论所蕴涵的判准是可修正性：当重构封闭了生成条件经由共同经验而被进一步修正的可能性时，退化便在场。

贡献。 可修正性被提出为关系性再生产伦理的第六条原则，并且是作为一条元原则而非又一项德性：它所治理的不是关系，而是那五条原则本身，因为任何丧失了可修正性的原则，都可能退化为它本欲约束的那一方的工具。第二项结论关乎分配正义：就遗弃而言，分配正义被表明是在原则上、而不仅仅是在执行上受限的。

关键词： 关系性退化；关系性主体；生成性关系存在；承认；可修正性；象征性权力；分配正义。

致读者的一则说明。 这是一篇讨论稿。它之所以被传阅，正是为了被辩驳；反对、反例、纠正以及任何形式的异议，都是被积极寻求的，而不只是被容忍的。

这一邀请并非形式上的客套。本文作为其主要贡献所提出的，正是：任何规范都不可以一种封闭了它自身经由后续发现而被修正之可能的形式被持有；并且，提出这一主张的框架本身也受其约束。这一类主张必须对针对它而提出的东西负责，而它的价值不会超过它所经受住的那些反对。使本文论证为假的条件，明确陈述于 §22；它无法回答的问题，记录于 §21。二者之所以被列出，正是为了被使用。

任何形式的来信都受欢迎，而其中最受欢迎的是异议。

huangwanhong@serendip.ngo

AI 使用声明。 本文的思想、主张与理论框架均出自作者。AI 工具被用作起草、文献的查找与核验以及修订过程中的辅助。每一个思想的来源均已由作者核实，作者对内容承担全部责任。

§1 引论

关系会恶化，而它们恶化的诸种方式，一直是在彼此难以对话的若干文献中被研究的。道德哲学考察各方彼此所作的错事。承认理论考察当一个主体被拒绝给予其应得的承认时所承受的伤害。精神分析考察一个主体如何处置它所经受的剥夺。政治社会学考察此类剥夺得以成为常规的那些安排，而国际关系考察那些无意加害的各方竟走向灾难的条件。它们各自都产出了成果，而其中没有一项在此受到争议。受到争议的是：它们所处理的现象是彼此分离的。

本文提出三个问题。第一个是：关系恶化所采取的诸种形式，即背叛、战争、遗弃、回应的撤回、公开归类、行政的漠然以及信任的丧失，究竟是七种碰巧彼此相似的伤害，还是沿着七条不同路径传递的同一个改变。第二个是：如果它们是同一个改变，那么被改变的是什么；因为行为与过错的词汇把这一改变定位在主体身上，于是便无法解释那些无人行动的案例。第三个是：什么把退化与单纯的转变区别开来；因为关系持续不断地转变着它的参与者，而”转变即是伤害”这一点并不显然，也绝不可被假定。

这些问题值得提出，因为人们通常用以描述关系性伤害的那套词汇，在一类既不小也不边缘的案例面前失效了。当来自亲密者的伤害在种类上、而不仅仅在程度上不同于一个陌生人所能施加的伤害时，它失效了；因为它只能记录下加害者是亲近的，于是不得不把一种性质上的差异当作一项加重情节来处理。当灾难在毫无恶意的情况下降临时，它失效了，正如现实主义文献所描述的处境；因为一套必须找到过错的词汇总会找到一个，而在过错并不存在之处，它便会发明一个。而它最彻底地失效，是在一个共同体不再维系它的某一成员、却没有任何成员作过这个决定之时：没有任何场合上这个问题被提出过，没有任何时刻可以被指认，也没有任何参与者的行为单独来看是错的。有的是一套逻辑，运行到它的极限；而一套被构造来把谓词附着于施动者的词汇，此时无物可附着。

本文论证，这些失效共有一个源头，并且这个源头是可以被陈述的。该词汇携带却未加陈述地带有一个关于主体是什么的前提：主体先于关系而存在，并被关系所影响。依此前提，无论什么降临于一个主体，都必定是被施加于他的；而凡被施加的，必定是被某人施加的。于是那些无人作过任何事的案例，就不只是困难的，而是不可理解的；而该词汇面对不可理解之物的反应，是去供给一个它并不拥有的施动者。

本文的推进方式是否认这一前提。依作者在别处发展的生成性关系存在的本体论，主体并不是一个站在关系之中的实体。它是在关系之内被生成的东西；它的属性、它的能力以及它的连续性，都由关系持续进行的生成所维系，并且不由其他任何东西维系。其后果是直接的，并且是定义性的而非推论性的：如果主体就是在关系中被生成的东西，那么关系的条件就是主体被生成的条件，而改变其一即是改变其二。无需任何进一步的机制来把这一改变传送过去，因为并不存在任何需要被跨越的间隙。

这一进路的代价应当立刻承认。本文的论证以一种无法从内部修补的方式，条件性地依赖于它的前提。一位主张主体先于关系的读者将会发现，此处没有任何东西迫使他就范：诸案例在他看来将是一批以不同方式施行的伤害，而那个判准则是可选的诸道德直觉之一。他这样想并不算错。他是在与那个前提相分歧，而分歧本就属于前提那一层。本文并不为该本体论作论证，而诸案例也并非作为它的证据被给出。

这些问题在该框架的发展中占有一个位置，并且这一点应当明白说出，因为它关乎本文是为何而作。该框架已推进了一套关于关系生成什么的说明，关于一段共有而不可重复的历史所承载的财富的说明，以及关于该财富的流通在何种条件下为正义的说明。关于生成出错时会发生什么，它所说的相对很少；而它所说的那一点点，又是由它在别处竭力要取代的那套词汇所组织的，把退化当作某人对某人所作的一件错事。本文试图把关于退化的说明纳入与关于生成的说明相同的本体论之下，而它的主要结论是一个该框架先前并不拥有的规范性结论。

它们在一场更长的运动中同样占有一个位置。价值被锚定的场所此前曾经迁移过。在生活的条件显现为自外部给定的情形下，它曾被锚定于一个超乎人的源头。此后它被锚定于符号，而符号只要保持稀缺，就能承载价值。随着符号的复制变得低廉，那种锚定失效了，而价值向着符号复制所无法使之低廉的东西迁移，那就是关系。但一个已迁移至关系的价值，是一个暴露于任何能够使关系变形之物的价值；而那些变形并不新鲜。新鲜的是，它们如今抵达了价值被持有的那个场所。一套没有关系性退化理论的关系性价值理论，是一套关于某项资产却没有关于其毁灭的理论，而这一遗漏并不是关于该资产的进一步说明所能修补的。

本文与最切近的既有框架之间的分歧，可以各用一句话说出，并在 §3 中详加陈述。承认理论主张，承认之所应得，是就主体已然所是而言；本文的说明则主张，承认属于一个主体得以成为主体的诸条件之一，由此可得：一无所成的主体并不因此就没有诉求；并且，一种密集而个别化的注视也可能是一种承认形式，却仍然可能封闭那个接受它的人。精神分析把主体转变的源头定位于主体之内；本文的说明把它定位于关系性生成的结构之中，并且所追问的不是主体如何处置降临于它之事，而是关系当时在做什么，以致此事降临于它。结构现实主义把无政府状态当作给定的，并由此推出困境；本文的说明把该状态当作被生产出来的，并追问它的生产；一旦生产完成，对于现实主义者关于那个世界所说的一切，本文毫无异议。而生成正义关切的是价值返回其生成者的路径，这预设了某物已被生成并被拿走；遗弃的案例落在它之外，因为被遗弃者所缺的并不是他所造之物的归还，而是一种他的创造会被接纳的关系。

有两项修正统辖着以下的一切，并且两者都与自然的说话方式相抵触。第一项是：退化并不削弱在一段关系之内被生成的主体的动力学。它重构这些动力学，而被重构的动力学往往更为强烈，以致”观察到旺盛的生成”绝不构成”一段关系是健全的”的任何证据。第二项是：主体的重构本身不可能就是那个错，因为每一种关系都在重构其参与者，而教育并不因此就是一种暴力。因此需要一个判准，而这一需要并非形式上的。七个案例中的一个，即按范畴施行的行政，以此处所考察的任何案例中最为审慎的方式改变着生成的条件，搁置了它所触及的每一个主体的特殊性，却仍然是一种无人主张废弃的正义形式的条件。一套无法把它与其余六者区别开来的说明，所产出的将是一句抱怨，而不是一个判准。

所提出的判准是可修正性，并且它被表明是由前提所蕴涵的，而不是由作者所选取的。一段不再能经由共同经验而被修正的关系，并不停止生成主体；它反复地生成同一个主体，而这就是一个生成条件已被封闭的含义。由此推出一个规范性主张，它是本文的主要贡献。可修正性被提出为该框架的关系性再生产伦理的第六条原则，并且它是作为一条元原则、而不是作为与那五条并立的第六项德性被提出的：它所治理的不是关系，而是那五条原则本身；因为一条其解释不再能被修正的原则，将由它本欲约束的那一方来解释，而一项由其对象来解释的约束，不成其为约束。陈述该原则的框架，本身也受其约束。

本文其余部分的安排如下。§2 陈述行为与过错的词汇的三项失效，并诊断其源头。§3 就与本文最切近的诸文献定位本文的论证。§4 陈述前提，导出后果，并提供该框架所需要、而该词的日常用法并不提供的退化定义。§5 作出两项修正，并引入贯穿全文的三层结构。§7 沿三条彼此独立的轴线，即每一案例所移除的区分、每一案例所容许的归责形式，以及修复的可得性，列出七个案例的分类，以便每一案例都可以被读作占据一个确定的位置，而不是清单上又一个实例。§8 至 §14 考察诸传递路径，§15 加以比较。§16 排除诸候选判准，其中包括该框架最自然会提出、而它又必须拒绝的那一个。§17 陈述该判准并展示其蕴涵关系。§18 陈述那道把机制的描述与其所产之物的证成分隔开来的约束。§19 与 §20 重述五条原则，表明它们无法裁定什么，并提出第六条。§21 记录本文未曾做到的事，§22 陈述本文将为假的条件，§23 说明它为何作为一篇讨论稿被提出。

§2 行为与过错的词汇预设了一个先于关系的主体

人们通常用以描述关系性伤害的词汇，是一套行为与过错的词汇。有一个行动了的施动者，有一个他据以行动的意图，有行动中的一个过错，以及由该过错而来的一份责任。这套词汇是古老的，是好用的，并且在日常案例上运行得如此顺滑，以致它无法说出的东西鲜少被注意到。

它无法说出三样东西。它们已在引论中被点出，此处再加陈述，因为随后的诊断有赖于它们被精确地看见。

第一样关乎背叛中所承受的伤害的特定性质。被背叛的一方受到了伤害，这一点该词汇记录得下。而他所受的伤害在方式上不同于一个陌生人所能施加的伤害，这一点它记录不下；并且这一差异不是量级上的差异。一个陌生人可能造成更大的损害，却更容易被原谅。它有资源说出一件错事被做了、并且作错事者是亲近的，但它没有资源说出为何这份亲近所改变的是伤害的种类而不是它的程度；因此它把一种性质上的差异当作了一个加强词。

第二样关乎无恶意的灾难之可能。在国际关系文献所描述的安全困境的处境中，每一方都为回应对方的备战而备战，每一方都基于对自身处境的正确评估而行动，而结局是一场双方都未曾寻求、双方却都有理由为之作准备的战争。行为与过错的词汇要描述这一点，只能靠定位出一个过错；而如果没有过错可以被定位，它就必须要么发明一个，要么沉默。它通常发明一个。把这一结局称作一次侵略，就是把一个结构性的困局转换成了一个道德性的困局，也就是说，就是把它误描述了。

第三样最为困难，并且它将占据本文相当的篇幅。一个已变得无法作出贡献的人可能发现，一个共同体不再维系他；而此类案例的特征是：没有人作过这个决定。没有任何会议上这个问题被提出过。没有任何时刻可以被指认为遗弃的时刻，也没有任何参与者的行为单独来看是错的。有的是一套逻辑，运行到它的极限。行为与过错的词汇被构造来把谓词附着于施动者，而在没有施动者可供指认之处，它无物可附着。

这些并不是一套原本健全的说明的边缘案例。它们属于关系性生活得以恶化的诸种核心形式，而行为与过错的词汇在它们面前要么沉默，要么更糟，开口而误描。

人们很容易把这一困难当作一个适用范围的问题。有人会说，这套词汇是为个人的行为而制的，而这些不是个人的行为而是系统的状况；需要两套词汇，各自在其领域之内可以是健全的。这一回应是不充分的，而它的不充分是有教益的。这三个案例并不远离道德判断的领域。它们恰恰是道德判断被最迫切地要求、而又最显眼地失效的场合；而一套必须把恰恰这些场合让渡给另一门学科的词汇，所得到的不是一条边界。它所得到的是一份诊断。

本文提出的诊断，关乎行为与过错的词汇携带却未加陈述的一个前提，而它是一个关于主体是什么的前提。依该词汇，主体先于关系。他存在着，他拥有属性与意图，尔后他与他人发生往来；关系是发生在早已在场的诸主体之间的一件事。依该前提，这三个案例确实是异常的；因为如果主体已然被构成，那么无论什么降临于他，都必定自外部降临于他，而自外部降临于他之物，必定是被某人施加的。

该前提并不明显为真，而本文的推进方式是否认它。

§3 承认属于主体被生成的诸条件之列

本文的论证与四种文献相距甚近，并各自从中取用了一些东西。由于这种切近足以招致混同，故在论证推进之前先陈述其分歧所在。

3.1 霍耐特把承认奠基于主体已然所是

本文的说明与承认理论最为切近，而其差异并不是侧重上的差异。

依霍耐特所发展的承认理论，承认之所应得，是就主体已然所是而言：就他的人格而言，此为爱所确认；就他作为权利承担者的地位而言，此为法所确认；就他的成就与他的贡献而言，此为尊重所确认。相应地，误认便是未能确认某种本就在那里等待被确认的东西，而伤害就是自身既有价值被否认的伤害。该说明在结构上是回溯性的，而这种回溯性对它并非无关紧要：为承认之诉求奠基的，是主体在被承认之先就已拥有的一项属性。

本文的说明不能采取这一形式，其理由就在于那个前提。如果主体是在关系中被生成的，那么就不存在任何先行的人格，站在关系背后等待被确认。承认并不是对一个早已在场的主体的确认。它属于一个主体得以在那里存在的诸条件之列。

在本框架中，承认不是回溯性的而是构成性的。它并不确认主体是什么。它属于主体被生成的诸条件之列，而这正是为何它的撤回不只是伤害一个主体，而是改变了一个主体所能是的东西。

由此推出两项后果，二者都标示着一处分歧。

第一项关乎承认理论处理得最差的那个案例。如果承认之所应得，是基于一个主体已成就什么或能贡献什么，那么一无所成、亦无可贡献的主体，就没有任何据以主张它的根据；于是该理论要么必须补充一个根据，即诉诸赤裸的人格，而这是另一条原则，要么就任他失去保护。依本文的说明，这一困难不会以那种形式出现，因为承认并不是基于任何东西而应得的。它是一项生成的条件，而一项条件不是一份酬报。

第二项关乎霍耐特的这一假定：承认是一种多多益善的好东西。§12 展示了一种注视，它强烈、持续、且精细地个别化，却像任何忽视一样彻底地封闭了主体。被归类的主体并未被忽视。他被以一种固定的注视看守着，而每一次观察都在确证那个范畴。这样一种注视是承认的一个种类，而它却仍然是退化性的：对于一套把承认当作标量善物的理论来说，这不是它所设想过的可能性。

3.2 精神分析把转变定位于主体之内

本文的说明使用象征的词汇，而它并不采纳这套词汇通常所携带的解释结构。

精神分析把主体转变的源头定位于主体之内：定位于象征秩序对它的运作之中，定位于它如何处置其构成性的缺失之中，定位于它借以使那份缺失可忍受的诸幻想之中。本文的说明把该源头定位于关系性生成的结构之中。这一区分不是诸词汇之间的偏好。它是关于一项解释应当从何处开始的分歧，而它有一个将在诸案例中可见的后果：精神分析的说明追问主体如何处置降临于它之事，而本文的说明追问关系当时在做什么，以致此事降临于它。

被保留下来的，是这样一个观察：主体不能确认它自身。某种外部的确认是必需的，而它的撤回并不是一种主体或可忍受的不适，而是它继续作为一个主体存在的条件。§11 依赖于此，而 §21 记录了本文并未确立它。

被拒绝的，是把整个说明同化于那个观察。主体需要一种外部的确认，这一点并不裁定什么可以供给它；而本框架的旨趣恰恰在于这样一种可能：关系性生成或许能够供给那在别处一直由一个固定而不容置疑的权威所供给的东西。它是否能够，此处并未确立。这正是本框架为之而存在、要去追寻的问题。

3.3 现实主义把本文视为被生产之物的状况当作给定的

现实主义关于安全困境的说明未受争议。在其领域之内，它是正确的，而 §9 毫无保留地这样说。

分歧在于分析从何处开始。现实主义者把无政府状态当作给定的，并由此推出：一个理性的一方必须把另一方的能力读作威胁。这一推导是可靠的。本框架不容许他去追问的，是生成的条件如何变成了这样一种条件，以致能力可以被读作威胁而不是被读作能力；而这并不是一个他拒绝回答的问题，而是一个他的前提不会提出的问题。本文的说明把无政府状态当作被生产出来的，并追问它的生产。

这一差异可以说得更尖锐些。现实主义精确地描述了一个某种再编码已然发生的世界的动力学。它不描述那次再编码，而且它也不拥有描述它的手段。

3.4 生成正义预设了某物已被生成并被拿走

本框架关于价值、以及关于把价值返回其生成者的流通的说明，受惠于生成正义理论，而这份亏欠在本框架的先前工作中已被承认。

此处登记一处分歧，因为它关乎那个判准。生成正义是一套关于价值返回其生产者的说明，因而是一套关于生产者的说明。它的核心不义是榨取：价值被生成了，而它没有回来。这是一项真实的不义，而关于它的说明未受争议。

但 §10 的案例落在它之外。被遗弃的主体并不是一个被榨取了价值的人。他是一个其生成没有接受者的人；而没有什么可以被返还给他，因为没有什么被拿走。一个围绕价值之返还而组织起来的框架在此无从着力，而这一困难是结构性的、而不是一处疏忽：一套关于正义流通的说明，预设了有某物在流通。

生成正义关切的是价值返回其生成者的路径。遗弃的案例关切的是能否被接纳进这个回路，而它并不是榨取的一个实例。被遗弃者所缺的并不是他所造之物的归还，而是一种他的创造会被接纳的关系。

§20 所提出的第六条原则，正是就这一点作为一项补充而被给出的。它并不取代正义流通的判准，也不与之竞争。它陈述的是一项先于流通的条件，即：关系必须保持有能力去修正它据以接纳并生成其参与者的那些条件；而这是一项条件：一个回路可能在对它已接纳者之间正义地流通的同时仍然失效。

3.5 本文的说明对动机保持沉默，并不推出任何政治结论

两点免责声明，之所以作出，是因为本文的词汇容易招致它们所要排除的误读。

本文的说明不是一套关于个体心理的理论，并且不对动机作出任何主张。一个退化了的格局中的参与者们在毫无恶意的情况下行动，这一点被反复断言；之所以被断言，是因为它为真，也因为本说明的解释负担并不要求它为假。诸案例中没有一个案例是由任何人的邪恶来解释的。

并且，本文的说明不从其动力学主张中推出政治结论。一个机制可以被表明维系着一个系统持续运作的条件，这是一项描述；而 §18 陈述了为何它绝不可被容许成为一项证成。生成性的词汇格外容易使这一区分坍塌，约束正因此而被正式陈述。

§4 主体是在关系中被生成的，而其后果是定义性的

前提以其强形式被陈述，因为弱形式无法承担这项工作，而看清为何如此是重要的。

弱形式会说：主体被他的诸关系所影响。这是一个因果性主张，它无可争议，并且它与行为和过错的词汇所假定的一切都相容。一个被其诸关系所影响的主体，是一个本就在那里等着被影响的主体，而影响是发生在他身上的一件事。依弱前提，§2 的三项遗漏仍旧是遗漏，因为”是谁施加了影响”这个问题依然可提，也依然无法回答。

强形式说的是另一回事。

关系性主体。 主体不是一个站在关系之中的实体。它是在关系中被生成的东西。它的属性、它的能力以及它的连续性，都不是被带入关系的所有物，而是由关系持续进行的生成所维系，并且不由其他任何东西维系。

这是生成性关系存在的前提，而它在此并不是为了论证的方便才被采纳的。它之所以被采纳，是因为它是本框架的本体论承诺，也因为本文的论证是那项承诺的展开，而不是对它的添加。必须被表明的是：该承诺具有所主张的那些后果。它为真，本文并未表明；而一位拒绝它的读者，将在下文中找不到任何迫使他就范的东西。

两种形式之间的区分，可以表述为两个问题之间的区分。依弱前提，人们追问一段关系对一个主体做了什么。依强前提，那个问题是构造不良的；人们转而追问的是：一个主体被生成为什么。

4.1 后果由定义而来

承认该前提，本文的主要主张便立刻随之而来。

如果主体就是在关系中被生成的东西，那么关系的条件就是主体被生成的条件。前者的改变即是后者的改变，而这并不是一种必须由某个其运作可被怀疑的进一步机制来传送的改变。并不存在任何需要发生传送的间隙。改变一个主体被生成的条件，即是改变该主体被生成为什么；而这两种描述所指的是同一个事件。

在生成性关系存在中，关系性主体并不是关系之外的一个实体，而是在关系的生成之内被持续地形成。因此，任何关系的退化都不只是关系的一项改变。它是生成条件的一项改变，因而是对一个主体所能被生成为什么、以及主体的动力学能够如何展开的一次重构。

因此，以下所考察的诸案例并不是对诸主体的攻击。它们是关系性生成条件中的改变，而主体随之改变。这正是本文的说明区别于那些最容易与之混淆的说明之处。精神分析把主体转变的源头定位于主体之内。心理学把它定位于认知与情感之中。本文的说明把它定位于关系性生成的结构之中，而它这样做并不是侧重上的问题，而是关于主体被当作什么的问题。

4.2 退化是生成主体之能力的改变

本文的标题中含有一个词，若不加定义，便会使标题落空。退化，在日常用法中，意谓一段关系变糟了；而一篇确立了”有些关系会变糟”的论文，什么也没有确立。前提提供了一个定义，并且这是一个有内容的定义。

关系性退化。 不是说一段关系减损了，也不是说一段关系变糟了。而是关系生成关系性主体之能力的一项改变，以致一个主体可被生成于其中的空间被约束、被改道或被封闭。

由这一定义推出两件事，二者都是必需的。

第一件是：退化不是一个道德术语。它命名一项结构性的改变；而这项改变是否应受谴责，是一个进一步的问题，须由诸案例来回答，而定义本身对此保持开放。这是一项优点而不是一项缺陷。一套把谴责构筑进描述之中的理论，将无法容纳制度性漠然的案例；后者以最为审慎的方式改变着生成的条件，却仍然是一种我们无意废弃的正义形式的条件。

第二件是：退化之所以重要，其理由并不是关于关系的。它之所以重要，是因为关系是主体被生成的场所；也只有凭借这一事实，关系中的改变才成为任何有后果的改变。日常的关系理论能够说，关系对人们要紧。本文的说明所说的更强也更陌生，那就是：人们就是关系正在做的事。

§5 退化重构主体的动力学

§4 所导出的后果易于陈述，也易于被错误地陈述。在它可以被投入使用之前，需要两项修正，而两者都与自然的说话方式相抵触。

5.1 被重构的动力学往往更为强烈

自然的表述会说：一段关系的退化使在其中被生成的主体的动力学退化。这一表述是错的，而它的错误不是措辞问题。

考虑在那些最显眼地恶化了的关系中所观察到的东西。一个围绕共同敌意而组织起来的共同体，并不表现出行动能力的减损。它通常表现出一种被强化了的行动能力：它的成员们行动得更为果决，认同得更有信念，协作得比先前更有效率。一个把自己的某位成员固定为蔑视对象的群体，并不变得倦怠。它变得富有发明力。它生成新的描述，发现新的场合，并以一种若无此对象则不会由等量努力所产出的丰产性，去精心阐发那项归类。

动力学并未削弱。它们被重新组织了，而被重新组织的动力学可能更为强烈。

关系性退化并不削弱主体的动力学。它重构它们。生成仍在继续，并且可能以增强的活力继续；所改变的是它得以发生于其中的空间。

这项修正不是一个让步。它正是使论题免于空洞的东西。说退化使之退化，等于什么也没说；说退化进行重构，才是说了某种可能为假的东西，而重构的诸种模式于是可以被区分开来。开篇可先点出四种，而以下诸案例将展示它们：生成可以被约束，以致更少的路径保持敞开；被扭曲，以致仍然敞开的诸路径通向与先前不同之处；被封闭，以致沿某一给定路径的生成完全停止；或被改道，以致主体被朝着一个它本不会趋近的格局而生成。

这项修正还有一个进一步的后果，而它在下文中将被用到。如果退化并不在于生成的减少，那么旺盛的生成之在场，并不构成一段关系健全的证据。一段关系可以大量地生成，而只生成一样东西。问题从来不是生成是否正在发生；而任何只问这一个问题的诊断，都会恰恰在健全最为缺席的那些案例中报告健康。

5.2 主体的重构是普遍的，并且它本身不是那个错

第二项修正关乎本文的说明必须不去主张的东西。

如果关系是主体被生成的场所，那么每一种关系都构成它的参与者，而关系的每一次改变都重构他们。教育重构学生，而且这正是它的意图。从一段漫长学徒生涯中走出来的学生，不是当初走进去的那个学生，而这一差异正是整件事的全部意义所在。一段多年的友谊已以某种方式构成了双方，而这种方式是双方事先都不会选择、如今也都不会放弃的。

由此可得：一个主体的重构不可能是退化的标志，因为重构正是关系所做的事。

一段关系重构它的诸主体，这不是一项缺陷而是一项定义。因此，退化的判准必须在重构这一事实之外去寻求；而任何谴责重构本身的说明，都已谴责了关系本身。

这是一项严峻的约束，而它正是判准之所以必要的理由。它一举排除了那个规范性主张的最自然的表述。人们不能说：当一段关系决定了它的参与者成为什么时，退化便发生了；因为那正是每一段值得拥有的关系的日常运作。必须就这项决定是如何施行的说些什么，而说出它正是判准的工作。

5.3 退化起源于干预所不能抵达的那一层

区分三个层次是方便的，而这一区分将被贯穿使用。

第一层是关系性生成的机制：关系据以产出其所产之物的条件。第二层是在那种生成之内被持续形成的诸关系性主体。第三层是这些主体彼此之间的动力学，新的关系由之被产出，而系统持续地生成。

退化发生在第一层。它被传递到第二层，并改变第三层。

退化起源于关系性生成的那一层。干预通常被导向第二层，即保护受到伤害的主体；或被导向第三层，即制裁伤害了他的主体。这正是干预如此频繁失效的原因。病理处在两种干预都不能抵达的那一层。

这一观察并不是作为一则绝望的劝告而被提出的，它也不蕴涵保护与制裁毫无价值。它所蕴涵的是某种更狭窄也更有用的东西，即：保护与制裁处理的是后果；而一段其生成条件已被改变的关系，在这些后果逐一被处理之后，仍将继续产出这些后果。一个被制裁的参与者所空出的位置依然可供占据，而它将被占据。

§6 诸案例展示传递的路径

以下七个案例并不是上文所确立的论题的证据。该论题由前提推出，而没有任何案例能够加强它。它们的功能是另一回事，而这正是一套统一性的说明最需要的功能。

一个主张”一切关系性退化都是关系性主体的重构”的论题，如果它无法说出诸重构彼此有何不同，那么它就是空洞的。一次抹除了它所统一之物之间差异的统一，什么也没有解释，而这一指控很容易被坐实。因此，诸案例是作为传递的路径被呈现的，而向每一个案例所提出的是同一个问题，每次都以同样的措辞提出。

一次发生在关系性生成之层的改变，经由何种路径，传递为在那里被生成的主体的一次重构？

答案在每一个案例中都不同，而这些差异正是这一部分的实质。凡某一案例在作者本人早先的处理中曾被误描述之处，该误描述都被陈述并纠正；因为一条被误认的路径比一条未曾被考察的路径更糟。

还须遵守一项进一步的约束。所论的诸机制在日常关系中即可见，而它们在此也是在日常关系中被描述的。一个从极端例子的情感分量中借取其力量的论证，并不是在承担自身的分量；而极端案例之所以被略去，是由于这一理由，而不是出于顾忌。

§7 诸案例沿三条彼此独立的轴线被区分

七个案例在被考察之前先被列出，以便每一个都可以被读作占据一个结构中的确定位置，而不是清单上又一个实例。三条轴线区分它们，而这些轴线是彼此独立的：一个案例在其中一条上的位置并不固定它在其余各条上的位置。

7.1 被坍缩的那项区分

每一个案例都移除了关系此前能够作出的一项分辨，而被移除的诸项分辨并不相同。

案例	被移除的区分	所丧失者
背叛	这一段关系对比于任何同类型的关系	特定性
战争	一个主体对比于一个策略的对象	内在性
遗弃	这个人对比于任何一个不被期待任何东西的人	独一性
回应的撤回	在场对比于缺席	被称呼的资格
公开归类	这个人对比于一个可以被如此对待的人	不被归类的权利
制度性漠然	这一个案对比于任何同类型的个案	特殊性
信任的丧失	这一个他者对比于任何一个可能辜负我的他者	被托付的资格

诸项区分彼此不可化约，而这种不可化约性很容易被忽略。一个主体可以被化约为一个策略的对象，同时仍是一个不可替代的策略对象；一段关系可以被化约为它的条款，而双方仍继续彼此承认为主体。这七者是同一个改变，并不蕴涵它们是同一种丧失；而一套容许它们合并的说明，将已抹除了它本被构造来解释的那些差异。

7.2 归责的形式

第二条轴线决定行为与过错的词汇是否有任何可供附着之处，而正是在这条轴线上，§2 的困难变得可处理。

案例	行为的结构	是否存在加害者？
背叛	单方面的、有意的、回溯性的	有
战争	双方面的、可能是无意的、自我实现的	可能没有
遗弃	系统性的	通常没有；有的是一套逻辑
回应的撤回	协同的，每一个行为都是可允许的	没有，且无人行动
公开归类	三方的、有意的、公开的	有，且他知道
制度性漠然	制度性的、有意的、依原则的	有，且他行动得正当
信任的丧失	渐进的，每一步都有根据	没有，且无人有错

右栏展示了一道道德哲学无力穿越的光谱。它从一个意图了他所为之事的加害者开始，经过无人行动的诸种格局，抵达一个行政者审慎地行动并且行动得正当的情形，并终结于一种由任何一方都没有作错的判断所拼装而成的退化。一套把谓词附着于施动者的词汇，可以在这道光谱的一端运作，而在另一端沉默；而这份沉默不是它在应用上的一处空缺，而是它的前提的一个后果。

7.3 修复的可得性

第三条轴线预示了那个判准，而它正是诸案例如此排列的理由。

案例	什么阻碍着修复
背叛	一种新的特定性可以被建立；旧的则无从复得，因为那项主张是关于过去的
战争	双方必须同时停止对象化，而每一方的停止都使自己暴露
遗弃	转移支付无法抵达它；所需要的是一种不同的生成条件
回应的撤回	本会执行修复的那项官能，已被征用来对抗修复
公开归类	听众的分辨必须改变，而主体无法向听众陈情
制度性漠然	没有什么阻碍；申诉、听证与立法都是可得的
信任的丧失	没有过错可供承认，也没有行为可供撤销

有两行被标出。第三行承载着本文最尖锐的实践后果，而 §10 论证它：分配在生成的条件之内运作，而遗弃改变的正是那些条件，因此分配正义在此处的限度是原则上的限度，而不是执行上的限度。第六行承载着本文的判准，而 §17 与 §20 论证它：生成的条件被改变得最为审慎的那个案例，恰恰是修复仍然可得的案例；而把它区别开来的，既不是改变的温和，也不是施行者的意图。

7.4 复合的次序

诸案例并不彼此独立，而其中两种次序将被用到。

对象化使遗弃成为可能。一旦他者已被再编码为一个策略的对象，那么一个没有策略价值的对象便是一个可弃置的对象，而那次再编码正是弃置的前提条件。反之则不然。把一个主体弄成与那些不被期待任何东西的人无从分辨，并不就是把他弄成一个敌手；而这种不对称支持着一个令人不适、并且据作者所能判断为真的命题：敌人比被遗弃者被更充分地承认。战争给予他者以一个施动者的地位。遗弃则不给予他任何地位。

并且，合法的坍缩为不合法的坍缩提供了许可。凡一个制度已确立了按范畴对待是正当的地方，§12 中的归类者便无需制造一项不义。他只需援引一个该制度已然授权的范畴，而 §12 表明这正是他所做的事。其一的合法性为其二提供了掩护，这是 §13 中所作让步的直接代价，而这是本文的说明不可拒绝支付的代价。

§8 背叛改写关系的过去

8.1 改变及其路径

背叛者所做的，首先并不是结束一段关系。而是重新描述它，而这一重新描述是回溯性的。该行为中隐含的主张，并不是这段关系从此将属于另一种类型。而是这段关系一向属于那种类型，而另一方一直误解了它。

在那段关系之内被生成的主体，曾是依据那一重新描述所取消的条款而被生成的。他发现的不是一份支撑被撤走了，而是他自以为拥有的那份支撑，从来就不具有他所设想的性质。他所倚赖的生成机制并未被摧毁。它被揭示为：从未以他所倚赖的那种形式存在过。

此后他被生成为一个曾对自身生成条件有所误解的主体。这正是 §2 发现行为与过错的词汇无力命名的那种特定性质的伤害；而它解释了为何背叛者的亲近所改变的是伤害的种类而不是它的程度。一个陌生人无法施加这种伤害，因为一个陌生人不曾属于一个人被生成的诸条件之列，也就无法回溯性地取消他从未参与其中的东西。

8.2 宽恕无法撤回一项回溯性的主张

宽恕能够在新的条款上建立一段新的关系，而且常常如此。它所不能做的，是撤回那项回溯性的主张；因为那项主张不是一个有待撤销的行为，而是一个关于本已如此之事的断言。一段新的关系可以被生成。旧的生成条件无法被恢复，因为它们如今被认定为从未成立过。

背叛之所以不可逆，并不是因为信任重建起来费力，而是因为它的运作是回溯性的。它所改变的不是关系的未来，而是关于它的过去的记述；而没有任何后继的行为对过去拥有权威。

8.3 伤害在于关于过去的记述

背叛此前主要被当作对信任的违背来处理，而关于信任的哲学文献为那种处理提供了资源。贝尔关于信任是接受自身在他人善意面前之脆弱性的说明，以及由之而来的关于背叛的诸说明，解释了为何背叛是一种具有独特严重性的错事：它利用了被背叛的一方因信赖背叛者而承担的一份暴露 [2]。马格利特关于厚关系与薄关系的区分，解释了为何在关系更厚之处这项错事更为严重 [13]。

这些说明所共有的是：它们把伤害定位在未来。信任被给予了，而这一给予如今被揭示为没有根据，所丧失的是再度给予它的能力。该说明是正确的，并且是不完整的；而这份不完整之处，正是本框架所贡献之处。

被背叛的一方所陈述的，其特征并不是他不再能够信任。而是他不知道那段关系曾经是什么。伤害是对他关于过去的记述的伤害，而未来信任的丧失是那项伤害的一个后果，而不是它的内容。依信任的词汇，这一点难以说出；因为信任是一种朝向未来的态度，而一种朝向未来的态度无法被回溯性地损害。依本文的说明，这一点是直截了当的：那段关系曾属于他被生成的诸条件之列，而那一重新描述取消了那些条件，于是此后他被生成为一个曾对生成了他的东西有所误解的人。贝尔以脆弱性之被利用所解释的那种独特严重性，在此则由某种更为先行的东西所解释，那就是：一个人自身构成的条件已被回溯性地撤走。

§9 战争在毁灭他者之前先对他重新编码

9.1 再编码先于暴力

关系被再编码为敌对的，而这一再编码先于暴力。这是要紧的次序，而日常的说明恰恰把它颠倒了。

序列是从主体，到敌人，到目标，到伤亡。它不是从主体直接到伤亡。证据可见于任何有组织的敌对行动之前的种种准备之中：把敌手构造为一个范畴，精心阐发那些使该范畴的成员身份变得可辨识的叙事，以及生产出那些人物形象，一个人自身的参与正是相对于他们才变得可理解。倘若毁灭人们就是目的，这一切都不会是必需的。它总是被着手进行，这表明所建造的并不是一种毁灭的能力，而是一套生产主体的系统；没有这套系统，那些将要执行毁灭的人便无法维持这一毁灭。

此后，主体只被生成为一个敌人，或只被生成为一个直面敌人的人。此处 §5 的第一项修正显示出它的必要性。如此被生成的主体的动力学并未削弱；它们被增强了。决心、团结与协同的能力全都增长，而正是这一增长、而非任何倦怠，使得该格局自我维持。

9.2 现实主义所描述的是再编码已然造成的那个世界

现实主义关于安全困境的说明在此不受争议，而本文并不着手驳倒它。现实主义者准确地描述了一个再编码已然发生的世界的动力学，而在那个世界之内，他的结论是成立的。他所不追问的，以及他的框架不给他任何手段去追问的，是什么改变了生成的条件，以致他者的能力可以被读作威胁而不是被读作能力。他把无政府状态当作给定的 [21]。本文的说明把它当作被生产出来的，而这一差异不是关于动力学的分歧，而是关于分析从何处开始的分歧。

建构主义文献此前已提出过这一反对，而本文的说明在此之上添加了什么，应当说明。温特关于”无政府状态是国家造就的”的论证，以及更广泛的”国家的身份与利益是被社会地生产出来的、而不是给定的”这一主张，已然否认无政府状态是一项给定的事实 [19]。本文的说明不与此争辩，并且受惠于它。

它的分歧之处在于：这项生产被认为生产出了什么。对建构主义者而言，被建构的是一种身份或一种利益，也就是说，是一个本就在那里承载它、并且原则上可能承载另一种身份的施动者的属性。施动者在建构中存活下来；而这正是使建构主义的方案成为一个充满希望的方案的东西，因为身份可以通过改变维系它们的主体间理解而被重新建构。依本文的说明，施动者并不在建构中存活下来，因为施动者正是建构所生成的东西。在再编码中处于危急之中的，不是一方承载哪一种身份，而是究竟有什么样的主体被生成。

这一差异可见于每一种说明关于逆转所作的预测。建构主义者预期，共有理解的重新建构将使各方恢复到一种合作可得的状况。本文的说明预期它不会；而 §17 给出了理由：再编码已经封闭了诸理解本会经由之而被修正的那条通道；而必须首先被解除的，是那种封闭，而不是那些理解的内容。

§10 遗弃使主体的生成失去接受者

10.1 改变及其路径

关系性系统不再把该主体包含进它的未来。没有什么被施加于他。所停止的是：他不再属于系统的前向生成所围绕组织的那些各方之列。

他的生成失去了它的接受者。他继续生成，他行动，他言说，他尝试，但生成没有被接纳；而本会使它进一步生成任何东西的那种关系性接纳，并未发生。

10.2 被遗弃的主体丧失他的独一性

人们很自然地会这样描述他的处境：他的价值已跌至零。而这一描述是错的。零是一个位置。被指派到零，就是仍然可在该秩序之内被定位，就是它的一个确定的元素，就是一个该秩序仍能说出些什么的人，即：他的贡献为零。这不是被遗弃者的处境。被遗弃者的处境是：他根本不再被生成为一个特殊者。

遗弃所移除的不是主体的价值，而是他的独一性。他并未变得一文不值。他变得与任何一个不被期待任何东西的人无从分辨。

10.3 转移支付恢复的是一个量值，而独一性不是量值

转移支付恢复的是一个量值，而所丧失的并不是一个量值。供给一个不再被生成为特殊者的主体的资源，是由那个范畴、而不是由他所接收的。他仍然是”不被期待任何东西的那一类人”中的一员，如今只是因其为该类中一名获得资助的成员而有所区别。

所需要的修复不是一次再分配，而是一种不同的生成条件；而分配在生成的条件之内运作，无法改变它们。这是一项原则上的限度，而不是执行上的限度；而它是本文的说明为正义理论所带来的最尖锐的后果。

10.4 可行能力、误认与排斥各自都止步于独一性之前

这一处境已从三个方向被趋近，而每一个方向都提供了其余方向所不提供的东西。

可行能力进路主张，所应得的不是一定数量的资源，而是去做与去成为的真实自由；而森关于”同等资源可能产出不同等可行能力”的坚持，已然否认转移支付足以了结此事 [16]。它比任何其他进路都更接近本文的说明，而它止步于一个确定的点。一项可行能力是一个人的能力，并且是通过追问那个人能做什么来评估的。被遗弃的主体可能保有每一项可行能力，却仍然是被遗弃的；因为他所丧失的不是他自身的一项能力，而是一种他对该能力的行使会被接纳的关系。所缺失的东西不在关系的他这一侧。

弗雷泽关于错误分配与误认的区分标示出同一片领域，并以不同的方式划分它 [5]。依她的说明，误认在于那些制度化的文化价值模式，它们否认一个人作为互动中平等伙伴的地位；而本处的案例将被归入其下。这一归类所不提供的，是为何两种补救不能彼此替代；而本文的说明给出了理由：此类误认并不是与错误分配并立的又一项伤害，而是主体被生成的条件的一项改变；而转移支付作用于一个已被生成的主体，因而无法抵达它。

关于社会排斥的文献提供了经验上的质感，而其特征是回避理论问题，把排斥当作诸种剥夺的复合体，而不是当作一种自有其结构的状况 [17]。这一结构正是本文的说明在此所提供的。被排斥的主体所丧失的既不是资源，也不是可行能力，也不是尊重，而是独一性，即被生成为”这一个”而不是”那些当中的任何一个”的属性；而没有任何转移支付能够恢复它的理由是：转移支付恢复的是一个量值，而独一性不是一个量值。

§11 回应的撤回迫使主体为自己命名

11.1 协同的沉默满足符号的每一项条件

回应被撤回。主体说话，而交谈越过他继续进行；他在场，却不被称呼。没有任何行为被施加于他，也没有任何名称被加诸他；而这个案例的特征是：它的每一个构成性的”非行为”，单独来看，都是无可指摘的。没有人有义务作出回复。

人们很容易断定，这个案例在七者之中独独是无需象征中介而运作的。这一断定应当被抵制。撤回是协同的。并不是每一个参与者各自碰巧不作回应；这种不回应是一贯的、有指向的，并且被包括承受它的那个人在内的各方所预期。这一类协同要求一条共有的规则，而一条共有的规则就是一个符号，无论它以何种方式表达。

具有特色的是该符号所采取的形式。它不是一个谓词。它并不断言主体是什么，因此它不提供任何可供争辩的东西。它断言的是一种被否定的成员身份，而一种被否定的成员身份不需要任何肯定性的内容。它无需主张他有缺陷，只需主张他不在我们之中。

名称的缺席不是符号的缺席。沉默，凡在它是协同的、有指向的并且被预期的地方，就满足了一个符号所须满足的每一项条件。它所扣留的不是它的意义，而是它的谓词。

11.2 主体供给那份记述，而供给它的官能因此被废除

此处，传递是经由一处空缺、而不是经由一次强加而运行的；而这正是它与 §12 的案例之间关键的结构性差异。归类为主体供给一份关于他自身的记述，而回应的撤回则不为他供给任何记述，只给他留下一个不会自行了断的问题。主体需要一份记述；而在没有记述被提供之处，他将自行提供一份。是他为自己命名。

他所生成的那份解释，正是由那项官能所生成的，而对它的任何修正都必须倚赖这同一项官能。这是该案例的结构性怪异之处，而它可以不借助隐喻而被陈述。主体的解释能力被征用来生产一份记述，而这份记述的效果正是废除那项能力。他并未被说服接受任何东西。他是自己得出了它；而一个人不会以怀疑别人向他作出的断言的那种方式，去怀疑他自己的结论。

11.3 外来的善意被吸收，而该格局自我维持

第一，外来的善意其特征是无效的。一句宽慰被接收为怜恤，一份关注被接收为可怜，而一次回应，凡其发生之处，被接收为那条规则早已预期的例外。每一个反例都被一份早已规定了反例应当如何被读解的解释所处理。

第二，这一格局不需要维护。一次归类必须被持续地重新演示，而 §12 将就这一要求作出发挥。回应的撤回，一旦确立，便由主体自己来完成；而发起它的那个共同体无需再做任何事。

11.4 内化就是这一案例的完成形态

人们向来把一种退化了的自我理解的内化当作一个独立的现象来处理，并以”施加退化者与被退化者合而为一”这一点为其可注目之处。它不是一个独立的现象。它是此处所描述的案例的完成形态；而那种合一不是一件奇事，而是那个机制本身。

11.5 主体供给了象征性暴力本会强加的那个范畴

关于排挤的心理学文献已确立了该现象及其大部分结构。威廉斯的研究纲领已表明：排斥被迅速地登记；即便在实施排斥的一方被知为任意的、或并非主体所看重的一方时，它仍被登记；并且它威胁到归属、自尊、掌控与有意义的存在这些基本需要 [20]。这些发现不受争议，而本文的说明依赖于它们。

关于内化的理论说明提供了另一半。布迪厄的象征性暴力命名了这样一种机制：被支配者逐渐通过那些属于支配者的范畴来领会自身的位置；而法农关于内化的说明，展示了它在一个没有任何外部立足点可据以抗辩这些范畴的主体身上的运作 [3, 4]。

两种文献都不提供、而本文的说明在此作为其贡献所提供的，是：为何内化不只是排斥的一个常见后续，而是它的完成；以及为何主体所生成的那份解释，独独地抵抗着来自外部的纠正。理由是结构性的，而不是心理学的。在归类的案例中，一份解释被强加；而一份被强加的解释保留着它来源的印记：主体仍可主张它来自别处，因而在原则上是可争辩的，无论这争辩多么徒劳。此处则没有任何东西被强加。一处空缺被留下，主体被要求去填补它，而填补它的那份解释是他自己的结论。一个人不会以怀疑别人向他作出的断言的那种方式，去怀疑他自己的结论；而此后每一份善意，都被一份早已规定了善意应当如何被读解的解释所接收。关于象征性暴力的文献解释了支配者的范畴如何成为被支配者的范畴。它不解释这样一个案例：没有任何范畴被供给，而被支配者供给了他自己的范畴。

§12 公开归类改变听众所能分辨的东西

12.1 三方结构是构成性的

共同体据以在其成员之间作出分辨的那条规则被改写了，而这一改写是公开的。这是该案例的构成性特征，而它也是最常被误当作一项偶然特征的东西。

该操作并不是施行于主体之上的。它是在一群听众面前施行的，而它所造成的改变，发生在听众的分辨能力之中，而不在归类者的分辨能力之中。归类者无需相信他所断言的东西，而且通常并不相信。所要求的是：他人开始觉得那项断言是可用的。这正是为何该操作几乎从不私下进行：若私下进行，它将一无所成，因为那个其分辨必须改变的一方并不在场。

该操作是一次演示，而它的场所是听众。所改变的不是被归类的主体是什么，而是共同体能够就他说出什么，并且此后能够就任何人说出什么。

12.2 旁观者的沉默完成该操作，而抵抗则确证那个范畴

旁观者的沉默是核心的，而不是边缘的。它不是一次未能介入，而把它如此描述就是把它错置了。它是听众已接受那项归类的证据；而那份接受正是那场演示的目的所在。

被归类的主体的抵抗，其特征是徒劳的。这场较量并不在他与归类者之间。它是在一个他无法向之陈情的论坛中进行的；而他的任何举止都可以被重新描述为对该范畴的确证。举止良好，他就是那个例外。举止不良，他就是那个实例。抗议，他就正是这样的人会做的事。

归类者通常对该主体并无特别的敌意。主体不是这场操演的对象。他是这场演示据以进行的媒介，而人们不会仇恨一个媒介。

12.3 规则尚在形成之处，该操作施教；规则已然固定之处，它援引

凡共同体的分辨规则尚在形成之处，该操作是教学性的。它向那些尚未就此事定见的参与者展示可用的范畴有哪些，而这将是他们所受到的最生动的教导。

凡规则已被制度性地固定之处，该操作无法凭教导而推进，也不试图如此。它凭援引一个已然携带授权的范畴而推进；其效果是：对该归类的抵抗变得与对授权它的那个制度的抵抗无从分辨。

在第一种情境中，该操作创造一个范畴。在第二种情境中，它把主体置入一个制度已然许可的范畴。第二种更难以抵抗，因为对它提出异议就是对该制度提出异议。

12.4 标签理论错失了听众，并误读了沉默

标签理论持有那项本质洞见已有六十年。贝克尔的表述，即越轨不是行为的一项性质而是他人施加规则与制裁的一个后果，以及莱默特关于初级越轨与次级越轨的区分，已然把该操作定位在共同体之中而不在被贴标签者身上，并且已然认识到标签生产出它声称要描述的那种举止 [12, 11]。戈夫曼关于污名的说明，以本文无意加以改进的精确性，提供了被贴标签的主体的处境的现象学 [7]。

这些说明中缺失了两样东西，而它们正是本框架所添加的。

第一是：该操作是一次向第三方发出的演示，而第三方正是改变发生之处。标签理论把标签定位在共同体之中，却不追问为何贴标签其特征是公开进行的，也不追问听众是为了什么。它把听众当作那些逐渐持有该标签的人的总和。本文的说明主张：听众不是标签的接收者，而是标签被安装之处；因此，该操作不是一次关于被贴标签的主体的传达，而是共同体所能分辨之物的一次改变；此后，它便能够就任何人、而不只是就他，作出这样的分辨。

第二关乎旁观者的沉默是什么。依标签理论，它是一次未能介入，也就是说，是一个非事件；而关于旁观者效应的文献也相应地对待它，把它当作一种由责任扩散所解释的缺席 [14]。依本文的说明，它根本不是一种缺席。它是该操作的完成；因为那场演示之所以被进行，恰恰是为了确立”该归类可以在无人反对的情况下被作出”，而一份沉默正是它可以如此的证据。把它称作一次未能介入，就是把该机制的构成性要素描述得仿佛它是一个缺失的要素。

12.5 归责失效，而责任并未失效

归类者，若仔细考察，也不是该格局的作者。他必须持续地维系它，因为一场不再被演示的演示便不再具有说服力；而他也不能轻易退出，因为他所占据的那个位置，正是那个否则将空出来供他占据的位置。他是一个无人作其作者的安排中的一名参与者。

对干预的后果正是 §5 所预言的那一个。移除那名参与者，留下的是那个角色，而那个角色将被填补。但这一后果绝不可被容许成为一项开脱。责任无法被穷尽地定位于个体心理之中，这并不蕴涵不存在责任。它蕴涵的是：责任一直是在错误的层次上被寻找的。

§13 制度性漠然封闭特殊性而使修正保持敞开

这个案例在种类上不同于其余六个，而这一差异正是本文的枢轴。

13.1 搁置特殊性正是制度的目的

关系被范畴化。主体被当作一个类型的实例来处理，而他的特殊性被审慎地、并且作为一项原则而被搁置。

这一点必须毫不含糊地陈述。法律平等的道德力量恰恰在于特殊性之被搁置。同等对待同类个案，就是拒绝把主体生成为一个独一者；而这一拒绝并不是一项在其他方面令人钦佩的安排的一个不幸副作用。它就是那项安排本身。一个照顾到每一个案之特殊性的行政，不会是一个更好的行政。它将是一套恩惠的体系。

如果主体的重构本身就是那个错，那么法律平等就是一个错，普遍主义就是一个错，程序正义也就是一个错。这是荒谬的；而它的荒谬正是判准之所以必要的理由。

13.2 为整体牺牲部分是一切强加所采取的形式

人们很容易这样为该案例作辩护：局部的主体生成被牺牲，以便普遍的生成得以维系。这一观察是可得的，它并非明显为假，而它仍然必须被拒绝。

这个句子的形式，即部分被舍弃以便整体得以延续，正是每一项系统性的强加据以为自己辩护的形式；而一套把它当作解释来接受的说明，已然丧失了把它当作证成来拒绝的资格。§18 陈述了这项约束及其理由。

这个案例的正当性并不基于它对整体的效劳。它基于别的东西；而指认出那别的东西是什么，正是 §17 与 §20 的任务。

13.3 韦伯的描述被接受，而批判性的推论被拒绝

韦伯关于官僚制理性的说明，以一种未曾被改进的清晰度陈述了这一案例：不带憎恨与偏私的行政，即不带仇恨也不带情爱的行政，不是行政的一种变形，而是它的完成形态；而把爱、恨以及一切纯属个人的情感从公务中清除出去，恰恰是使可计算的行政成为可能的东西 [18]。

批判性的文献则推出了更为阴暗的结论。鲍曼关于”官僚制的组织方式不仅与暴行相容，而且为暴行提供了条件”的论证，以及阿伦特关于”不加思索地做着自己工作的公务员”的说明，两者都主张：特殊者之被搁置，正是使不可思议之事得以被行政地施行的东西 [9, 1]。

本文的说明接受韦伯的描述而拒绝那个批判性的推论；而它据以拒绝的理由，正是本文在此处的贡献。批判性的诸说明把按范畴施行的行政当作退化性的，其理由正是它搁置了特殊者；由此可得，该安排的危险程度与它的成功程度成正比。倘若那是对的，补救之道便是更少的行政；而这个结论，鲜有批判者愿意推出，也无人能够使之融贯。

本文的说明主张：搁置特殊者并不是使一项安排成为退化性的东西；而它必须如此主张，因为主张相反者就是谴责法律平等。把行政的案例区别开来的，是它的裁定保持可修正：有申诉，有听证，有可以对该归类提出异议的程序，还有可以据以改变范畴本身的立法机关。那个重构了主体的安排，同时提供了他得以被再度重构的手段。

这产出了一项批判性诸说明所不产出的预测，而这是一项本文的说明可据以被检验的预测。使一个行政成为危险的，不是它的范畴化形式。而是修正通道的封闭；而一个以最严厉的方式进行范畴化、同时使那些通道保持敞开的行政，依本文的说明，比一个温和地进行范畴化、却已封闭了那些通道的行政更少危险。

§14 信任的丧失取消共同的未来，而合作仍在持续

14.1 信任是生成一个共同未来的能力

共同的未来被取消了。各方继续交易、继续合作、继续交流；而一位观察者会发现，依每一项行为上的测度，这段关系都完好无损。所停止的是：他们不再共同地向前生成任何东西。

人们曾说过、并且就其所及而言是正确的：信任不是预测。未曾说出的是信任肯定地是什么，而这一遗漏使那份说明变得单薄。信任是生成一个共同未来的能力。因此，它的丧失并不是对自己关于他人举止之估计的一次向下修正，因为这样一次修正与一段完全正常运作的关系是相容的；而是关系从”共同产出任何一方都无法事先指明之物”这项事业中的撤出。

当信任丧失时，所丧失的是生成一个共同未来的能力。合作、交换与交流可能都毫无损伤地继续，而这正是为何这份丧失从外部看不见，并且常常连当事各方也看不见。

主体被留下，没有一个可供共同生成的对手方。他仍可行动，他的行动仍可能得到回应，但这些回应是在一个已然定型的框架之内作出的，而不是被接纳进任何新的东西之中。他的生成收缩为对已被指明之物的重复。

14.2 该退化完全由正确的判断拼装而成

这是把它与其余六者区别开来的特征，而它也是将令读者不安的特征。可以并不存在任何”发生了什么”的场合。每一次相继的扣留都是依据证据所作的合理修正，没有哪一步是错的，也没有哪一方有过错。该退化完全由正确的判断构成。

一种退化可以由一步步各自理性、无一为错的步骤拼装而成。如果确实如此，那么在某些关系性条件之下，理性本身就属于主体被重构的诸机制之列；而这是本文的说明必须接受、而不是回避的一个后果。

14.3 作为估计的信任在此登记不到丧失；作为共同生成的信任则登记得到

有两种文献处理信任，而它们把它当作两种不同的东西来处理。

决策论传统把信任当作一项估计。信任就是依据关于他人可能举止的一个判断而行动；而信任的理性程度，由该判断的准确性以及随之而来的收益来评估 [8]。依这一说明，面对不利证据而撤回信任根本不是一种退化。那是学习；而把它描述为一种丧失，就是把一次正确的推论误描述为一次伤害。

社会学传统把信任当作一种降低复杂性的机制。卢曼的说明主张，信任容许一方仿佛一系列可能性已被排除那样地行动，并因而使那些无法在穷举每一项偶然性的情况下存活的事业成为可能 [10]。这更接近要害；而它仍然是一套关于信任为一方做了什么、而不是关于它在双方之间做了什么的说明。

本文的说明主张：信任是生成一个共同未来的能力；而它与上述两者的差异并非术语上的差异。依决策论的说明，上文所描述的格局中没有什么需要解释，因为每一次扣留都是有根据的，而结局正是有根据的扣留所产出的东西。依本文的说明，则有大量东西需要解释；而那解释是：一段关系可以满足每一方的估计，而不再生成任何估计所未曾包含的东西。所丧失的不是关于他者的一项信念。而是与他一道去着手那些双方都无法事先指明之事的资格。

这是本文最令人不适的贡献，而它正是作为如此而被陈述的。如果信任是一项估计，那么理性的更新不可能使一段关系退化。如果信任是共同生成的能力，那么一系列理性的更新可以将它熄灭，而其中没有哪一步是错的，也没有哪一方有过错。本文的说明委身于第二者，并接受其后果，即：在某些条件之下，理性本身就属于关系被重构的诸机制之列。由此并不能推出各方行动得不当。由此推出的是：行为与过错的词汇再一次无物可附着。

§15 诸路径在修正的通道处分岔

§7 的分类依据每一案例所移除的东西、每一案例所容许的归责形式，以及什么阻碍着修复，来区分诸案例。既已考察它们，如今可以添上第四栏，而正是这一栏使本文其余部分得以展开。

案例	第一层上的改变	通向主体的路径
背叛	关系的过去被改写	他曾对自身生成的条件有所误解
战争	关系被再编码为敌对的	他只被生成为敌人，且动力学被增强
遗弃	系统不再把他包含进它的未来	他的生成没有接受者；独一性丧失
回应的撤回	回应被协同地扣留	他必须供给他自己的记述，而供给它的官能因此被废除
公开归类	分辨的规则被公开改写	他只被生成为那个范畴，而抵抗确证它
制度性漠然	关系依原则被范畴化	他的特殊性被搁置；修正的通道仍然存留
信任的丧失	共同的未来被取消	他没有可供共同生成的对手方

第三栏中有一项不同于其余各项，而它已被标出。七条路径中的六条封闭了主体可被以别样方式生成的可能性。第七条没有；而它恰恰是第一层上的改变最为审慎的那个案例。

无论是什么把它与那六者区别开来，那都不是改变的温和，也不是施行者的良善意图。下一节论证：把它区别开来的，是唯一可能做到这一点的东西。

§16 诸候选判准皆告失败，其中包括本框架自己的那一个

诸案例已以最尖锐的形式确立了这一困难。六条路径封闭了主体被以别样方式生成的可能性；第七条没有，而第七条恰恰是生成条件的改变最为审慎、最为彻底的那一个。因此需要一个判准；而在提出它之前，应当先排除那些不堪用的候选者，因为其中每一个都比幸存下来的那个判准更具吸引力。

16.1 重构是普遍的

这已在 §5 中被了结，此处重复，是因为它是论证最自然会退回去的那个候选者。每一种关系都重构它的参与者。如果重构就是那个错，那么关系就是那个错，而本文的说明便谴责了它自身论题的存在条件。

16.2 生成可能在退化之下增加

这同样已在 §5 中被了结，而它值得再推进一步。一个退化了的格局可能以比它所取代的格局更大的能量进行生成，而且通常如此。因此，”一段关系正在旺盛地生成”这一观察并不是它健全的证据；而一个以生成的数量来表述的判准，将恰恰在健全最为缺席的那些案例中给出健康的判决。

16.3 对未来生成的封闭重新确立了功利主义规范

这个候选者最具诱惑力，而它必须以最大的审慎被拒绝，因为它看上去正是本框架自己的那一个。

该提议会说：一个把主体固定在某个位置、封闭了他本可拥有的未来的符号，是一个退化性的符号；而一个使那个未来保持敞开的符号则不是。该提议是循环的；而更糟的是，它重新引入了本框架为之而存在、意在取代的那个规范。

假设对被遗弃者所作的错事在于对他未来生成的封闭。那么一个确实不会再生成的主体，而此类主体是存在的，并且他们的存在并不是一项边缘性的复杂情况，就并未因这一封闭而受到冤屈，因为原本就没有什么可供封闭。该判准保护了多产者，而恰恰遗弃了那个本文的说明为之而构造的一方。

以生成性取代有用性，仍然是在立一条功利主义的规范。它保留了这样一种结构：一个主体之所以配得承认，是 因为 他将带来某种东西；而它只是把他将带来的东西改了个名字。”他配得承认，因为他仍能生成”与”他配得承认，因为他仍能贡献”，是同一个句子。

这一后果是本框架必须接受、并且不可加以软化的。生成性是本文的说明所描述的那个现象。它不能同时是承认之所以应得的理由；而一套使之如此的说明，将在为自身奠基的行动中拆毁自身。

§17 一段不能被修正的关系反复地生成同一个主体

把制度性漠然与其余六者区别开来的，不是它重构得更少，不是它生成得更多，也不是它施行得更为温和。而是：它所重构的那个主体，仍然可以被再度重构，并且是经由共同经验而被再度重构。

该判准。 当一个主体的生成条件的重构，封闭了那些条件经由共同经验而被进一步修正的可能性时，退化便在场。

该判准并非被选取的。它是由前提所蕴涵的；而这一蕴涵关系应当被展示，而不是被断言。

如果主体就是在关系中被生成的东西，那么他的持续生成要求关系保持有能力去接纳新的东西。一段不再能被共同经验所修正的关系并未停止生成，这正是第一项修正的要旨，但它只能生成它已然指明的东西。它在产出，而它无限地产出同一个主体。

一段不再能经由共同经验而被修正的关系，并不停止生成主体。它反复地生成同一个主体。而这就是一个生成条件已被封闭的含义。

17.1 行政使通道保持敞开；其余六者封闭它

按范畴施行的行政重构主体并搁置他的特殊性，而它做这两件事都是有意为之的。它却仍然使修正的通道保持敞开；而这份敞开对它并非无关紧要，而是属于它的构成性承诺之列：有申诉，有听证，有可以对该归类提出异议的程序，还有可以据以改变范畴本身的立法机关。主体可以被再度重构，而他得以被再度重构的手段，正是由那个重构了他的安排所提供的。

教育对学生的重构，比七个案例中的任何一个对任何人的重构都更为彻底，而无人主张它因此就是一种暴力。学生可以与老师争辩，可以超越他，并且可以成为老师未曾设想的某种东西。重构是彻底的，而修正仍然可得。

其余六者封闭这条通道，并且它们以不同的途径封闭它。在背叛中，修正将必须抵达一个不再可供修正的过去。在战争中，主体只能在敌手这一范畴之内被修正。在遗弃中，没有对手方来接纳一项修正。凡回应被扣留之处，本会执行修正的那项官能，已被征用来生产那本该被修正的东西。凡归类是公开的地方，任何修正的尝试都被接收为确证。凡信任已丧失之处，修正本会在其中进行的那种共同生成，本身已被撤回。

制度性漠然通过而公开归类不通过，其理由并不是前者温和而后者残酷。前者使那条其自身裁定可经由之而被修正的通道保持敞开。后者封闭它，并且把封闭它当作其运作的条件。

17.2 一段严苛的关系可能通过，而一段仁慈的关系可能不通过

由此推出三项后果，无一令人舒适。

第一项是：一段关系可以是严苛的、范畴化的、非人格的，而可以不是退化性的。这在某些读者看来将是一项缺陷。这是一个能够把法庭与暴民区别开来的判准所须付出的代价；而任何无法作出那一区分的判准，无论它还能做到什么，都是毫无价值的。

第二项是：一段关系可以是温暖的、体贴的、出于善意的，而可以是退化性的，只要它所供给的那份注视，是一种封闭了主体对”他被当作什么”的修正的注视。本文的说明委身于此，而这份委身应当被陈述。仁慈不是一项抗辩。

第三项是由 §14 的案例所提出的那一项。如果一种退化可以由一步步各自理性的步骤拼装而成，那么该判准所谴责的，就是一个没有任何参与者在其生产中犯错的格局。本文的说明接受这一点。它主张责任并不因此被消解，而是被重新定位到生成条件被决定的那一层；并且它承认，用以在那一层上分派责任的机制并未得到良好的发展，无论是在本文之中还是在别处。

§18 动力学描述与规范性证成之间的那道墙

有一项约束统辖着以上一切的使用，而它之所以被单独陈述，是因为本文的词汇格外容易违反它。

本文的说明描述诸机制。它观察到，跨越许多系统、在许多尺度上，某些主体的重构可以被发现维系着那些系统持续运作的条件。这一观察是可得的，而它值得追究。

不可做的是：把这一观察转换成一项理由。

一次重构维系着一个系统持续生成的条件，这是一项描述。它不是一项证成，而且它绝不可被容许成为一项证成。”部分被舍弃以便整体得以延续”这个句子，正是每一项系统性的强加曾经据以为自己辩护的形式；而一套把它当作解释来接受的说明，已然交出了把它当作辩护来拒绝的资格。

这种暴露是这一类框架所特有的，而它应当被精确地理解。一套围绕生成性组织起来的词汇，能够以它自己的术语、并且以完美的忠实性陈述：一个主体的生成能力之被摧毁，扩大了包含他的那个系统的生成能力。在一个耦合系统中，这可能为真。这是一个关于动力学的命题，而它可以被证据确立或驳倒。

它不为任何东西提供理由。

因此，本文的说明有义务能够说出、并且确实说出：一次封闭了一个主体之修正的重构是一种退化，即便它可以被表明扩大了整体的生成能力。这一点之所以可说，仅仅是因为承认并未被奠基于生成性之上；而这已是本文中 §16 的那项拒绝第二次做了必要的工作。一套把承认奠基于生成性之上的框架，将会发现这个句子不可说，并且将会太迟地发现：它已经建造了一套装置，用以证成它本被建造来命名的恰恰那种东西。

§19 五条原则在一段封闭的关系之内是可被满足的

§17 的判准并不孤立。它加入作者在别处以关系性再生产伦理为题所发展的一组原则；而该判准与那些原则之间的关系需要被陈述，因为该判准在此是作为对它们的一项补充而被提出的，而一项补充必须被表明是必要的。

五条原则陈述一段关系得以正义地再生产其自身的诸条件。

关系性再生产的伦理：五条原则。 一次正义的关系性再生产满足五项必要条件。生成性：它不封闭参与者持续生成的能力。共同再生产：关系是经由共同参与而被再生产的，而不是由一方独自的决定所再生产。非剥削：它不依赖于任何参与者的工具化或牺牲。权力的约束与非固化：由于价值的生成携带权力，它同时生成那约束权力、防止权力固化的反权力。主体保全：它不以抹除他者之为主体来换取统一；重合不是目的，而差异与主体之被保全，正是使持续生成成为可能的东西。

五条原则在此不受质疑，而本文的论证一直是在它们之内进行的。受到质疑的是它们是否足够；而 §13 的案例表明它们并不足够。

19.1 按范畴施行的行政满足全部五条，并封闭它所触及的每一个主体

按范畴施行的行政可以逐一对照五条原则加以考察，而它满足每一条。

它不封闭参与者持续生成的能力。它在整个安排的尺度上维系着那种能力，而它通常正是为此目的而被设立的。它是经由共同参与而被再生产的，因为范畴是被立法的，而立法是须负责的。它不依赖于任何参与者的牺牲，而且它的特征恰恰是被设计来确保没有任何参与者为任何他人的便利而被牺牲。它约束权力并防止其固化，而这不仅与它相容，而且几乎就是它的定义。并且它不抹除他者之为主体；它拒绝去照顾他的特殊性，而这是另一回事，并且它对所有人一律地拒绝。

按范畴施行的行政满足五条原则，并坍缩它所触及的每一个主体的独一性。公开归类违反第三条与第五条，并坍缩一个人的独一性。二者在它们对主体所做的事情上形式相似，而五条原则借由指出一者中存在而另一者中不存在的违反，把它们分开。

这并非无足轻重，也不应被轻视。但它并未触及那个困难，而理由值得精确地陈述。

五条原则是施加于一段关系如其当下所是的条件。其中没有一条追问：这段关系是否仍然对被改变保持敞开。一段关系可以满足全部五条而仍是封闭的；而一段满足全部五条的封闭关系，将继续满足它们，因为正是这段关系本身提供了评估其满足与否所用的那些术语。

这一点可以被表述为一个关于自我评估的困难。五条中的每一条都是施加于一段关系的谓词，而这一施加要求一份关于该关系的描述，谓词得以据之被评估。凡关系已封闭了其自身描述本可经由之而被修正的那条通道之处，它便自行提供那些描述，并且它以其自身运作所固定的术语来提供它们。一个已封闭了其范畴之修正的格局，将会报告说它并不剥削，因为剥削是以它所固定的术语来定义的；说它保全主体，因为主体是什么同样已被固定；说它约束权力，因为它所承认的权力正是它自身的描述所容许的权力。

五条原则并不因此被驳倒。它们被表明是保护不足的；而保护它们的东西，正是如今必须被补上的东西。

§20 可修正性治理五条原则本身

可修正性。 一段关系必须保全其自身生成条件经由共同经验而被修正的可能性。这是关系性再生产伦理的第六条原则，而它在此作为本文对该伦理的补充而被提出。

该原则的根据已在 §17 中给出，此处以一句话重述。一段不再能经由共同经验而被修正的关系，并不停止生成主体。它反复地生成同一个主体，而这就是一个生成条件已被封闭的含义。

20.1 第六条原则治理那五条

它与五条的关系，不是一个项目被添加到一张清单上的关系；它也不是一段关系可能与其余诸项并立展现的又一项优异之处。它并不治理关系。它治理五条原则本身，而正因如此，它属于一个不同的秩序层级。

论证是简短的，并且是普遍的。每一条原则都需要解释，而解释是由那些该原则所适用的人来执行的。凡那份解释本可经由之而被修正的通道已被封闭之处，该原则并不因此就不再被满足。它被满足了，而且它是以那次封闭所固定的术语被满足的。原先是对一段关系的一项约束的东西，变成了那段关系的一件工具；而这一转换无需任何恶意，并且可以由完全真诚地行动着的各方来完成。

任何丧失了可修正性的原则，都可能退化为一件压迫的工具。这并不是因为该原则本身是错的，而是因为解释它的那一方，正是它本欲约束的那一方；而一项由其对象来解释的约束，不成其为约束。

五条中的每一条，若不与第六条同持，都会以一种可以被精确陈述的方式退化；而这种退化并不是未能满足该原则，而是在封闭的关系所提供的某种描述之下对它的满足。

生成性变成由我来界定什么算作生成。本文自始至终坚持一个退化了的格局大量地生成，而这份坚持在此结出果实：一段以不竭的丰产性产出单一主体的关系，依第一条原则所提供的每一项测度，都满足第一条原则。该原则并不区分丰产与重复，而且它无法区分，因为这一区分在评估的那一刻是不可得的，而只有跨越对”被评估之物”的修正才变得可得。

共同再生产变成共同地，但只以已然固定的方式。各方一同行动，这并不蕴涵他们一同所做的任何事情可以改变他们正在做的事情。一段关系可以由每一个参与者所再生产，而不被他们中的任何一个所改变。

非剥削变成由我来界定什么算作剥削。§14 的格局以最纯粹的形式展示了这一困难：没有参与者把任何他人工具化，没有哪一步是错的，而该退化完全由正确的判断拼装而成。找不到剥削者，而该原则无物可附着。所需要的不是去搜寻一项违反，而是去观察到：这段关系不再能被各方在其中所经验的东西所修正。

权力的约束被它自身的成功所击败，而反权力凝结成一种权力。一种为约束权力而设立的反权力，其自身就是一个结构，而结构会固化。此处所给出的关于不对称的说明直接关乎这一点：一段关系的不对称持续地产出权力；而阻止那份权力沉淀下来的，不是一个抗衡性的结构，而是权力被持续地吸收进关系持续进行的生成之中。吸收是工作，而工作会停止。凡它停止之处，权力无需被夺取。人们只是沉陷入它；而它之成为如此，无需任何人做过任何事。

主体保全变成只有一个符合我的描述的主体才值得被保全。把他者作为一个主体来持守，与把他作为一个人们已然认定他所是的那个主体来持守，在任何单一的时刻都无从区分。它们只能跨越时间被区分，并且只能凭借他是否可以成为另一个人而被区分。

五条原则追问一段关系当下是否正义。第六条追问这段关系是否仍然容许那个问题被再度提出。没有第六条，五条可以在一个封闭的系统之内被满足，因为正是那个系统提供了评估其满足与否所用的那些术语。

这正是诸案例在符号这一特殊案例中所发现的东西的普遍形式。那里曾说过，一个符号中危险的并不是它存在，而是它能够封闭它自身的修正。第六条原则正是那一观察被提升到规范的层级。它并不针对符号。它针对任何规范，也针对这个框架自己所提出的每一条规范。

20.2 权力被吸收进关系持续进行的生成之中

第四条原则，依其先前的陈述，主张价值的生成携带权力，因而同时生成那约束权力的反权力。本文的说明提示了一项修订，而这项修订是被提出以供考虑的，而不是被断言的。

依此处所给出的说明，权力并不是被一个与之对立而立的反权力所约束。它是被吸收的，也就是说，它被接纳进关系持续进行的生成之中，并成为关系所共同生成之物的一部分。老师对学生的权力，并不是被学生所握有的一种对立权力所制衡的。它是被学习的共同工作所吸收的，被那项工作所容许的相互纠正所吸收的，被那不对称于其中被消耗掉的成长所吸收的。我们之所以注意不到那份权力，是因为它正在被用掉。

权力无需被制衡。它需要被吸收，而吸收是持续的工作，而不是一项常设的安排。压迫无需被制造；它只需不再被吸收。

这项修订之所以要紧，是因为它把失效定位在了正确的地方。一段关系并不是在某人攫取了一项优势的那一刻变得具有压迫性的。它是在吸收的工作停止的那一刻变得具有压迫性的，而那份一直在场的优势不再被消耗掉。这正是为何良好的关系会在无人作错事的情况下退化；而这也正是为何第六条原则不是第四条的一项可选的精细化，而是第四条得以被维持的条件本身。吸收要求关系能够接纳新的东西。一段不能被修正的关系不能吸收。

20.3 该原则适用于陈述它的那个框架

一条豁免了它的作者的元原则不会是一条元原则。它将是一件装置，用以保护某一条规范免受它要求于其余每一条规范的那种审视；而这项豁免，将是该原则本被表述来命名的那种封闭的第一个实例。

因此，该框架受其约束。

如果生成性关系存在不能接受它自己的理论经由新的共同经验而被修正，那么它就违反了它自己的第六条原则。该原则并不针对符号，也不针对他人的规范。它针对任何规范；而它所适用的第一条规范，就是这一条。

这一自我适用不是一个修辞姿态，而且它带有一项应当被陈述的代价。它意味着五条原则不是一个基础，而第六条不是一次完成。它意味着可能还需要一条进一步的原则，六条中的某一条可能被表明是错的，而该框架没有资格以”这些原则是它自己的”为由去抵抗这样一种表明。

这正是该框架事实上所持的立场，而它持此立场的理由并不是谦逊。

20.4 规范本身处于历史的运动之中

关系性再生产的伦理并不是从某条公理推导出来的，也不是作为一项永恒真理的发现而被提出的。它是作为一套关于关系在何种条件下正义地再生产的说明而被发展出来的，而这是在作者曾在别处论证为其自身处于运动之中的物质条件之下：价值被锚定的场所此前曾经迁移，并且看来正在再度迁移；而适合于那种锚定的某一格局的规范，并不因此就适合于另一格局。

一条在历史运动之内被产出的规范，仍须对那场运动负责。它并不站在生成了它的那个过程之外去审视该过程；它是那个过程的一部分，并且随着该过程揭示出那条规范本无法预料的东西而被修正。

一条在历史运动之内被生成的规范，必须保持可被那场运动所修正。若主张相反，就是为该规范要求一个在产生它的过程之外的位置；而那恰恰是本框架否认给主体、因而也不能一致地为自己的诸原则所要求的位置。

本文所作的补充正是在这一精神下被提出的，并且其自身就是这一精神的一个实例。五条原则对于它们被表述以应对的那些案例是足够的。它们在 §19 中被表明，对于本文的分析所带入视野的一个案例是不足够的；而这份不足并不是它们表述上的一项缺陷，而是它们由之得出的那份经验的一个限度。因此第六条被提出，而它是在它自己所施加的同一项条件之下被提出的：即它须可被后续的发现所修正。

该框架以这种方式被扩充，这不是一项让步。它是第六条原则对该框架本身的运作，而它是唯一可得的证据，用以表明该原则不只是一项对他人提出的要求。

本文的形式受同一考虑所支配，而这一点很容易被误解。可修正地持有一项主张，并不是暂定地持有它；并且它不能由一位作者愿意发布第二个版本而得到满足。一次由作者作出的修订，正是该原则所关切的那种解释权威的一次行使；而一套只能由写下它的那只手来修正的说明，并未对任何人开放。该原则所要求的是：驳倒的条件被提供给读者，以便该说明可以被他所带来的东西所击败，而不只是被它的作者后来所得出的结论所修补。本文的诸项主张是不加含糊地提出的，而击败它们的手段陈列于 §22。这就是一条规范被可修正地持有的含义，而这就是本文的形式所欲成就的全部。

20.5 该原则是必要的而非充分的

这一点必须明白说出，因为一条主张得更多的原则会比一条主张得更少的原则更糟。

第六条原则是一项必要条件而不是一项充分条件，并且它并不构成一套实践。它陈述修正的通道必须保持敞开。它并不陈述一段关系所产出的诸不对称应当如何被吸收；而一条无物通过的敞开通道，是一段正在退化的关系之中的一条敞开通道。它对一个在别处毫无储备、并且对之而言一条他没有资格进入的通道之可得等于不可得的主体，无话可说。并且它无法重启已经停止的东西：可修正性是一段正在生成的关系的属性，而不是一段已经封闭的关系的疗法；而命名这项属性，并不就是提供了恢复它的手段。

那套实践将由什么构成，即：经由何种安排，一段关系所生成的剩余被返还给生成它的那些人；经由何种手段，一道已变得单向的目光被重新弄成相互的；以及在符号已然凝结之处应当做什么，此处并未裁定。§21 把这些记录为本文留待开放的问题，而它们正是整个说明的价值最终将系于其上的那些问题。

§21 论证所留下的开放问题

论证提出了四个问题，而在其中并未回答它们。它们在此被陈述，而不是留待被发现；因为一套隐藏其无法回答之问题的说明，所得到的不是一条边界，而是一层伪装。

21.1 退化性的符号大量地生成

在若干处曾说过、而未加详述的是：一个退化了的格局大量地生成。这一观察被使用了而未被考察，而它将无法原封不动地经受考察。

一个把主体固定在某个范畴之中的符号，并不只是持存。它加以阐发。它产出新的描述，发现新的应用，并为使用它的人提供一项无穷无尽的解释工作，而他们在其中，就一种完全日常的意义而言，是有创造性的。这一后果令人不适，而它应当被直面。人们不能靠生成得更多来对抗这样一个格局，因为它已经在生成，并且是以一种抗衡性的努力不易匹敌的发明力在生成。

因此，一个封闭的符号的生成与一个开启的符号的生成之间的区别，不是生成的数量，而且也不可能是。作者的暂定看法是：这一区别在于生成在何处展开，即：是在解释者之中展开，从而使他被扩大；还是穿过他而展开，在此情形下他所产出的是该符号的确证而不是他自己的东西。但这是一个关于解释之现象学的主张，而它要求一个本文尚未作出的论证。

21.2 主体不能确认它自身

本文的说明曾说一个符号剥夺了解释者，而未曾说过解释者为何仍然留下。

答案，若有的话，多半是：主体不能确认他自身；而一份来自外部的确认，并不是他可以放弃的一种安慰，而是他继续作为一个主体存在的条件。若果真如此，那么离开那个格局便不是一个他所拒绝的选项，而是一个他并不拥有的选项；而摆在他面前的选择，不是在安适与操守之间，而是在他的感受与他的存在之间，而这不是一个公平的选择，把它描述为一个选择也无益处。

这是一个关于主体之构成的主张，而它是一个大主张。它与上述论证的前提相容，而它并不由该前提所蕴涵；而本文不是确立它的地方。

21.3 一段封闭的关系会报告说一切正常

§17 的判准是以可修正性来表述的，而可修正性是一项可以从外部加以检查的结构性属性。有理由怀疑这项检查是否总能从外部进行。

一个已封闭了修正通道的格局不会报告这一封闭。它会报告一切正常，并且它会真诚地如此报告；因为那种本可用以命名该封闭的语言，正是该封闭所移除的语言。如果这是对的，那么该判准虽然正确，却未必总能凭本文所提供的手段而被应用；而或许需要别的什么东西，去侦测它所正确定义的东西。作者猜想，那别的什么东西不是一项进一步的结构性检验，而是一种知觉的形式；而这一猜想仅被记录，未被论证。

21.4 生成的条件一直被当作给定的

本文的说明一直把关系性生成的条件当作给定的，并追问当它们被改变时会发生什么。它未曾追问那些条件如何成为它们所是的样子，未曾追问它们凭何种动力学而自行改变，也未曾追问是什么决定了，当一个关系性系统不再能吸收其自身生成所产出的诸不对称时，它将重构它的诸安排，还是重构它的成员，还是重构它的敌手。

那个问题正是作者认为该框架所面临的最重要的问题，而它要求一套本文并不拥有的形式装置。它是那个另行发展的元模型的主题，而这正是那项工作被着手进行的理由。

§22 论证与其前提共存亡

本文的论证是演绎性的，而它的结构应当是显然的。一个关于主体之构成的前提，蕴涵一个关于”一段关系的退化是什么”的后果。两项修正使该后果免于空洞：退化是重构而不是削弱；以及重构本身不是那个错。于是需要一个判准，而该判准被表明是由前提所蕴涵的，而不是由作者所选取的。诸案例展示了该后果得以传递的诸路径，而它们并不确立它。

因此，本文以一种无法从内部修补的方式，条件性地依赖于它的前提。一位主张主体先于关系的读者，即主张主体独立地拥有其属性、并进入那些不触动其构成的往来之中的读者，将会发现论证中没有任何东西迫使他就范。诸案例在他看来将是一批以不同方式施行的伤害，而那个判准在他看来将是可选的诸道德直觉之一，对他并无特别的要求力。他这样想并不算错。他是在与那个前提相分歧，而分歧本就属于前提那一层。

本文所主张已经表明的，在那个前提之下，是五件事。

§2 所指认的行为与过错的词汇的那些遗漏，并不是边缘性的而是结构性的；并且它们出自它关于一个先于关系的主体的假定。

关系性退化容许一套统一的说明，而这套说明并不抹除其诸形式之间的差异；统一在于被改变的东西，即一个主体被生成的条件，而差异在于该改变得以传递的诸路径。

主体的重构不能成为退化的判准；生成性同样不能成为它的判准，否则将坍缩为该框架为之而存在、意在取代的功利主义规范；而经由共同经验的可修正性，作为余下的那个判准，是由前提所蕴涵的。

该判准所加入的关系性再生产伦理的五条原则，若无它则保护不足；因为一段已封闭其自身修正通道的关系，提供了评估其对那些原则之满足与否所用的那些术语。因此，可修正性被提出为该伦理的第六条原则，并且它是作为一条元原则、而不是作为与那五条并立的又一项德性被提出的：它所治理的不是关系，而是那五条原则本身；因为任何丧失了可修正性的原则，都可能退化为它本欲约束的那一方的工具。这是本文对该框架规范装置的补充；而这项补充是在它自己所施加的条件之下被提出的，即：陈述它的那个框架本身受其约束，并且一条在历史运动之内被产出的规范，仍须对那场运动负责。

以及：就遗弃的案例而言，分配正义在原则上、而不仅仅在执行上是受限的；因为分配在生成的条件之内运作，而遗弃所改变的正是那些条件本身。

其中最尖锐的是最后一项，而它也是作者最欢迎纠正的一项。如果遗弃的修复可以被表明是经由转移支付而可达成的，如果资源的供给可以被表明能够恢复本文的说明所描述的那种被丧失的独一性，那么 §10 就失败了，而本文所承载的最强的实践后果也随之失败。

反对、反例与纠正，欢迎寄至 huangwanhong@serendip.ngo。本文的论证被铺陈到这样的长度，正是为了使人易于与它相分歧；而本说明为第六条原则所要求于它的那种修正而保持敞开。

§23 可修正性要求驳倒的条件被提供出来

这是一篇讨论稿；而对这一称谓的一项反对应当在开篇就予以回应，因为它是显而易见的那一项。一份预印本是可以被修订的。如果本文的主张是暂定的，那么后续的版本会这样说；而在标题中宣告暂定性，并无所得。

倘若这一称谓意味着作者打算修订，那么该反对将是决定性的。它并不意味着那个；而修订的意图并不是本文第六条原则所要求的东西。一次由作者发布的修订，是作者解释权威的一次行使。它把评估的术语留在原处，即留在他手中；而一套只能由写下它的那一方来纠正的说明，并未因此就向任何东西开放。该原则所要求的，不是作者保持愿意改变主意，而是读者被给予击败他的手段。

本文正是据此而写的。它在 §22 中陈述了它所作的五项主张，并指认出它在其中哪一项上最为暴露。它点出了那唯一一项发现：对它的驳倒将把整个说明中最强的实践后果一并带倒。它在 §21 中记录了它无法回答的四个问题，以及它的工具对它们不敷使用的理由。这些不是谦逊的姿态。它们是驳倒的条件；而它们之所以被列出，是因为一套不提供此类条件的说明，并不能靠宣布它有朝一日可能被修订，就使自身成为可修正的。

本文的论证在形式上是演绎性的。它从一个本体论前提出发，导出该前提所蕴涵的一个后果，然后考察七个案例，在其中该后果可以被观察到得以传递。诸案例不是该论题的证据，也不是作为证据被给出的。如果前提被拒绝，那么再多的案例累积也无法挽救该论证；如果前提被承认，那么主要主张无需它们便可推出。

诸案例凭另一个理由赢得它们的位置。”一段关系的退化是它的诸主体的一次重构”这一论题，如果那次重构的诸路径无法彼此区分，那么它就是空洞的；因为一次抹除了它自身诸差异的统一，什么也没有解释。因此，诸案例的功能是保住那些差异；而它的适当与否，须由这七条路径是否真正彼此有别、以及每一条是否可在该论证所提供的术语中被辨认出来，来加以判断。

结论诸节承担着分量。它论证：一个主体的重构不是那个错；因此需要一个判准；而该判准是由本体论所蕴涵的，而不是由作者所选取的。这一部分的检验是一个单一的案例。制度性漠然按范畴重构主体，封闭他们的独一性，却仍然是一种我们无意废弃的正义形式的条件。一套无法把它与其余诸者区别开来的说明，所产出的不是一个判准而是一句抱怨。此处所提出的区分是否成立，正是本文成败所系的问题。

有两件事被有意排除。本文不在极端案例的层面上进行论证；因为一个从其例证的情感分量中汲取力量的论证，并不是在承担自身的分量；而所论的诸机制在日常关系中即可见，无需诉诸暴行。并且，本文不从其动力学主张中推出任何政治结论。机制的描述与该机制所产之物的证成之间的区分，被贯穿地维持着；而它是被审慎地维持的，因为一套生成性的词汇格外容易使那一区分坍塌。§18 陈述了该约束及其理由。

来信、反对与反例，欢迎寄至 huangwanhong@serendip.ngo。

---

讨论之一到此结束。主体被生成的条件在此一直被当作给定的，而它们自身如何改变这一问题已被推迟。它是最要紧的问题，它在此未被回答，而它正是后续的讨论所必须转向的主题。

参考文献

[1] Arendt, H. (1963). Eichmann in Jerusalem: A report on the banality of evil. Viking Press.

[2] Baier, A. (1986). Trust and antitrust. Ethics, 96(2), 231–260.

[3] Bourdieu, P. (1991). Language and symbolic power (G. Raymond & M. Adamson, Trans.). Harvard University Press.

[4] Fanon, F. (1967). Black skin, white masks (C. L. Markmann, Trans.). Grove Press. (Original work published 1952.)

[5] Fraser, N. (2003). Social justice in the age of identity politics. In N. Fraser & A. Honneth, Redistribution or recognition? A political-philosophical exchange (pp. 7–109). Verso.

[6] Eglash, R. (2016). An introduction to generative justice. Teknokultura, 13(2), 369–404.

[7] Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Prentice-Hall.

[8] Hardin, R. (2002). Trust and trustworthiness. Russell Sage Foundation.

[9] Bauman, Z. (1989). Modernity and the Holocaust. Cornell University Press.

[10] Luhmann, N. (1979). Trust and power. Wiley.

[11] Lemert, E. M. (1951). Social pathology. McGraw-Hill.

[12] Becker, H. S. (1963). Outsiders: Studies in the sociology of deviance. Free Press.

[13] Margalit, A. (2017). On betrayal. Harvard University Press.

[14] Darley, J. M., & Latané, B. (1968). Bystander intervention in emergencies: Diffusion of responsibility. Journal of Personality and Social Psychology, 8(4), 377–383.

[15] Honneth, A. (1995). The struggle for recognition: The moral grammar of social conflicts. Polity Press.

[16] Sen, A. (1999). Development as freedom. Oxford University Press.

[17] Silver, H. (1994). Social exclusion and social solidarity: Three paradigms. International Labour Review, 133(5–6), 531–578.

[18] Weber, M. (1978). Economy and society (G. Roth & C. Wittich, Eds.). University of California Press. (Original work published 1922.)

[19] Wendt, A. (1992). Anarchy is what states make of it: The social construction of power politics. International Organization, 46(2), 391–425.

[20] Williams, K. D. (2007). Ostracism. Annual Review of Psychology, 58, 425–452.

[21] Waltz, K. N. (1979). Theory of international politics. Addison-Wesley.

[22] Lacan, J. (1998). The seminar of Jacques Lacan, Book XI: The four fundamental concepts of psychoanalysis (A. Sheridan, Trans.). Norton. (Original work published 1973.)

[23] Marx, K. (1998). Theses on Feuerbach. In The German ideology. Prometheus Books. (Original work written 1845.)

[24] Ostrom, E. (1990). Governing the commons. Cambridge University Press.

[25] Huang, W. (2025). Responding to the crises of symbol and subject in modernity: Towards a generative nomos of relational presence. SSRN. https://doi.org/10.2139/ssrn.5752104

[26] Huang, W. (2026). Relational wealth: Definition and characteristics. Knowledge Commons. https://doi.org/10.17613/nq7js-vz283

[27] Huang, W. (2026). A theory of relational reproduction: Definition and ethical principles. Knowledge Commons. https://doi.org/10.17613/qncwr-zbt82

[28] Huang, W. (2026). Relational value under the symbolic crisis. Knowledge Commons. https://doi.org/10.17613/bm32z-aqt14

[29] Huang, W. (2026). A metamodel for ontologically heterogeneous social dynamics. Part I. Serendip Commons Society.