Constrained Generative Beings - On the Absence of Measured Generation - A Preliminary Discussion

Posted on 2026-07-15 In Philosophy , Theory of Justice Views: Disqus: Word count in article: 17k Reading time ≈ 1:04

ENGLISH

Constrained Generative Beings

On the Absence of Measured Generation

A Preliminary Discussion

Wanhong Huang · huangwanhong@serendip.ngo

Abstract

The generative relational framework holds that value is generated in relation and is not possessed by a subject in advance of relation. This paper subjects that commitment to its hardest case, namely the case in which no generation is measured at all. Theories of distributive justice concern the allocation of goods among participants, and theories of exploitation concern the appropriation of value from those who have produced it; neither addresses the subject who produces nothing. The paper’s principal result is that the absence of measured generation cannot be identified with the absence of generative being, and that a system which infers the second from the first has committed an error concerning the world before it has committed an error concerning the subject. The concept of the constrained generative being is introduced to designate a locus of generative possibility whose participation in generative relations is blocked, redirected, or rendered unrecognizable by relational conditions. The concept is distinguished from those of weakness and vulnerability, which locate a deficiency within the subject, and equally from the humane reply that the excluded subject possesses an unrecognized worth, which concedes that value is lodged within subjects and disputes only the accuracy of the inventory. Three topologies of constraint are distinguished. Under endogenous constraint the generative rule is altered; under exogenous constraint the rule is intact and no environment permits its execution; and under epistemic constraint, which the paper introduces, the rule is executed, the environment permits execution, generation occurs, and the system’s measure does not register it. The third case is not a weaker instance of the first two: under the first two the system observes correctly and infers wrongly, and under the third the system’s observation is false and cannot be corrected from within, since the error is located in the instrument rather than in the inference. From this the paper draws its normative consequence. Because constraint is located in the relational conditions of generation, the remedy is necessarily performed upon those conditions and never upon the subject, and the obligation to act follows from the location of the constraint rather than from any assessment of the subject’s worth.

Keywords: generative relational being; constrained generative being; constraint ontology; endogenous and exogenous constraint; epistemic constraint; abandonment; recognition; the relationality of the value metric; productivism; disability.

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 and revision. The origin of each idea has been verified by the author, who takes full responsibility for the content.

§1 Introduction

Preliminary note. This paper is a preliminary discussion. Its argument rests on a primitive that is asserted rather than derived; it introduces a category of constraint that the author has not found named in the existing literature and cannot yet defend against every objection; and it leaves open, at the point where the account reaches its limit, a question it does not know how to settle. Each of these is marked in the text at the point at which it arises.

1.1 The case that tests the framework

The generative relational framework holds that value is generated in relation and is not possessed by a subject in advance of relation. The present paper subjects that commitment to the case in which it is hardest to sustain, namely the case in which no generation is measured at all.

The case is not a marginal one. It is occupied by the subject whose illness has ended their work, the subject whose age has ended it, the subject who after some rupture cannot act, and the subject who after some rupture will not act again. Theories of distributive justice concern the allocation of goods among participants. Theories of exploitation concern the appropriation of value from those who have produced it. Both presuppose that the subject stands in some determinate relation to production, whether as a claimant upon what has been produced or as a producer from whom what was produced has been taken. Neither addresses the subject who produces nothing, from whom no value has been appropriated because none has entered circulation.

The question the paper takes up is accordingly the following:

On what condition does a subject who generates nothing remain a participant in generative relations?

The question is not raised in order to defend a class of subjects. It is raised because the framework’s central commitment entails it. If value were a property that subjects carried into their dealings, then a subject who generates nothing would be a subject possessing less of it, and the question of their standing would reduce to a question of what is owed to those who possess little. If value is generated in relation, then the observation that nothing is generated is an observation about a relation and is not yet an observation about any subject. The subjects named above are therefore the conditions under which the ontology is tested, and are not the subject matter of the paper. What is at issue is the ontology, and the principal result is a result about it.

Claim 1 (principal result). The absence of measured generation cannot be identified with the absence of generative being.

1.2 Terms

Three terms are used throughout and are defined here.

Definition 1 (generative relation). A generative relation is a relation between a subject and its conditions within which value is generated. Participation in a generative relation consists in standing as a term of such a relation, and does not consist in the production of any output.

Definition 2 (productivism). Productivism is the ontological position that value is a property of outputs, together with the derived position that a subject’s standing within a system of value is conditional upon that subject’s production of such outputs. The position is not attributed here to any particular author.

Definition 3 (abandonment). Abandonment is the withdrawal of a subject’s standing as a participant in generative relations, on the ground that the subject’s generation is not registered by the system’s measure.

Abandonment so defined is not a synonym for neglect, and does not denote an emotional disposition. It denotes an inference, together with the withdrawal of standing that the inference is taken to license. The paper’s account of that inference is given in §5.

1.3 The position refused

One position must be set aside at the outset, since it is the position a reader is most likely to attribute to the paper, and since it is weaker than the position defended here.

The humane response to abandonment holds that the excluded subject possesses a worth that has not been perceived, that the injustice consists in a failure of perception, and that the remedy is a more generous assessment. This response concedes the productivist ontology and disputes only the accuracy of its inventory: it accepts that value is a property lodged within subjects and quarrels with the bookkeeping. It is therefore without resources at precisely the case that requires them, namely the subject who will not recover, will not subsequently contribute, and of whom no honest observer can predict a return. Against that case an appeal to unrecognized potential has nothing to offer, since there is no such potential.

The position defended here requires nothing to be found within the subject. Value is not located within the subject at all. It is generated in relation, and whether it appears is therefore a fact about the relation rather than a fact that the relation conceals. This is not a more lenient verdict upon the abandoned subject; it is a rejection of the premise that the subject was the proper object of a verdict.

That case governs what follows. A framework may be assessed by the case around which it is constructed, and the case around which this one is constructed is not the student who will graduate, the patient who will recover, or the researcher who lacks supervision and would flourish upon obtaining it. Those cases are real and the framework must address them, but they are not difficult, since a sufficiently patient productivism addresses them also: it need only extend its accounting period and await the return. The difficult case is the subject from whom no return will come. A framework that accommodates only the temporarily constrained has not departed from productivism but has deferred the settlement.

1.4 Plan of the paper

§2 sets out the inversion of the productivist order that the generative relational account requires, and declares the primitive upon which the account rests. §3 defines the constrained generative being and distinguishes the concept from those of weakness and vulnerability. §4 distinguishes three topologies of constraint, of which the third is introduced here for the first time. §5 establishes the epistemological result. §6 draws the normative consequence, which concerns where a remedy is to be applied.

Adjacent work on generative justice concerns the circulation of value among those already engaged in generating it [2, 3]. The question addressed here is a different and prior one, and nothing offered below is intended as a correction of that work.

§2 From Productive Capacity to Relational Capacity

2.1 The productivist order

Productivism, as defined in §1.2, orders capacity, production, and recognition as follows.

Capacity → Production → Recognition

On this ordering, capacity is a property of the subject, production is the exercise of that property, and recognition is what production earns. The ordering renders recognition conditional: what is awarded at the terminus of a sequence may be withheld at its origin.

Three familiar positions dispute the terms of this ordering while accepting the ordering itself. The first holds that the sequence should be entered more inclusively, on the ground that capacity is more widely distributed than is commonly supposed. The second holds that it should be entered more patiently, on the ground that capacity develops under favourable conditions. The third holds that recognition should be conferred on independent grounds, as a matter of human status or of right, alongside the sequence rather than through it. The three differ in their prescriptions and agree that the sequence describes the structure of the problem.

2.2 The relational inversion

The generative relational account rejects the ordering at its first term. Capacity is not a property that a subject brings to a relation; it is an outcome of relation.

The claim may be illustrated by a case. A researcher without supervision, without an intellectual community, and without time is described by productivism as possessing a capacity that circumstances prevent her from exercising. The relational account gives a different description: the capacity does not exist in an unexercised state. There is no reserve of unsupervised insight held within the subject and awaiting favourable conditions. Supervision does not release a capacity; it participates in constituting one. In the absence of the relation, the capacity is not concealed but absent.

This claim carries a cost that should be recorded. It follows that the paper cannot appeal, on behalf of a subject who generates nothing, to what that subject latently contains. Where a subject generates nothing, the paper is not entitled to reply that the capacity is nonetheless present. It must give a different reply, and the remainder of the paper is an attempt to state one.

2.3 The primitive

The account requires a foundational commitment which is stated here without derivation.

Definition 4 (the primitive). Generation is relational: value is generated in the relation between a subject and its conditions, and is not a property possessed by the subject in advance of that relation.

No derivation of this commitment is offered, and none is required. Every framework rests upon a primitive that it does not derive. Productivism rests upon the primitive that value is a property of outputs; it assumes this, constructs a metric upon the assumption, and reports the readings of the metric as findings about the world. The proper question to put to a framework is not whether it has demonstrated its primitive, since no framework has, but what follows from its primitive that does not follow from the alternative.

The consequence is the following. Under the generative relational account, the measure by which generation is assessed is itself an element of the relation, and is therefore available to doubt. Productivism cannot subject its own metric to doubt, since for productivism the metric constitutes the definition of value rather than a reading of it. The proposition this subject generates nothing, and the measure may nonetheless be at fault is, within productivism, not a caution but a contradiction. Within the generative relational account the proposition is well formed. §5 develops this consequence, which constitutes the warrant for holding the primitive.

Two derivations of the primitive are available in the literature and both are declined here. The first holds that difference gives rise to interaction, interaction to novel configuration, and novel configuration to generation. This derivation is teleological: it installs novelty as the end toward which generation is directed, and a framework in which generation has an end is one in which the question whether a given subject serves that end may be raised. That question is the instrumental question this paper is written to exclude. The second derivation holds that no isolated entity can generate relational properties, and that generation therefore requires relation. The inference is valid and its conclusion is trivial: relational properties require relations by the definition of the terms, and nothing follows concerning value.

It should be recorded that no non-circular justification of the primitive appears to be available. Any argument that a measure may fail to register value already presupposes that the measure is internal to the relation whose output it reports, which is the primitive itself. An argument of that form establishes what follows from the primitive and does not establish the primitive. The consequence stated above is accordingly exhibited as a yield and is not offered as a proof.

The primitive is therefore asserted and its consequences exhibited. A reader who finds the consequences insufficient has thereby been given the grounds on which the primitive should be rejected.

§3 Constrained Generative Beings

3.1 The definition

Definition 5 (constrained generative being). A constrained generative being is a locus of generative possibility whose participation in generative relations is blocked, redirected, or rendered unrecognizable by relational conditions.

Three elements of the definition require comment.

The term locus of generative possibility is to be distinguished from the notion of a container of latent value. To designate a subject a locus of generative possibility is to assert that generation may occur in the relation of which that subject is a term. It is not to assert that anything is stored within the subject. The distinction is the one drawn in §1.3 and is maintained throughout.

The three verbs of the definition, blocked, redirected, and rendered unrecognizable, do not denote three degrees of a single failure. §4 establishes that they denote structurally distinct situations, and that the third differs in kind from the first two.

The phrase by relational conditions identifies the site of the constraint. The site is not the subject.

3.2 The concept distinguished from weakness and vulnerability

The concepts of the weak subject and the vulnerable subject locate a deficiency within the subject and treat that deficiency as explanatory of the subject’s exclusion.

The alternative to locating a deficiency within the subject is not to locate an unrecognized sufficiency within the subject. Both operations accept that the subject is the site at which the explanation is to be found. The position advanced here, which may be termed constraint ontology, rejects that assumption. It is not a revised verdict upon the subject; it is a rejection of the subject as the object of the verdict.

Constraint ontology accordingly replaces the question whether a subject possessed a capacity, a question whose two available answers are alike productivist, with the question why generation failed to occur. That question admits of more than one form of answer, and the forms are distinguished in §4.

3.3 The levels of constraint

Constraint occurs at three levels and the analysis is uniform across them.

At the individual level, constraint characterizes the position of the student, the patient, and the subject recovering from injury.

At the social level, it characterizes the position of groups excluded from the institutions within which generation would otherwise occur.

At the historical level, it characterizes ideas and technologies for which the conditions of uptake do not obtain. The description of such cases as ahead of their time is rejected in §4, on the ground that it reintroduces at the level of propositions the notion of latency that has been rejected at the level of subjects.

§4 Three Topologies of Constraint

Let the generative system be written

$$S_{t+1} ;=; F(S_t,,E_t),$$

where $S$ denotes the subject’s internal dynamics, $E$ the environment, and $F$ the generative rule, that is, the structure by which a subject and an environment jointly determine what is generated. Let $M$ denote the measure, that is, the criterion by which the surrounding system identifies an output as generation.

The notation is introduced as a notation and not as a model. Its function is to make explicit a distinction that ordinary language does not enforce, namely that a failure of generation may be located in $F$, in $E$, or in $M$, and that these locations are not degrees of one another. The three cases are termed endogenous, exogenous, and epistemic constraint respectively.

4.1 Endogenous constraint: the altered grammar

Endogenous constraint obtains where $F \rightarrow F’$: the generative rule is itself altered. The subject’s dynamics are not reduced in magnitude but changed in kind. The parameter space differs from what it was, and the reachable states differ accordingly.

Illness, neurodegeneration, disability, and trauma exemplify this case. Trauma exemplifies it most clearly, since the alteration is not a subtraction. Where the rule was

$$\text{Subject} ;\longrightarrow; \text{World} ;\longrightarrow; \text{Action},$$

it becomes

$$\text{Subject} ;\longrightarrow; \text{Threat} ;\longrightarrow; \text{Defense}.$$

No term has been removed; one term has been substituted for another. The resulting system remains fully dynamical and fully generative, and generates other outputs than it did.

Claim 2. Endogenous constraint consists in $F’ \neq F$ and does not consist in $F’ < F$. The altered rule is not a diminished rule. It is a different rule, generative of different outputs, and its outputs are not on that account fewer.

The distinction is not terminological. A framework that construes endogenous constraint as diminution is committed to ordering generative rules by magnitude, and is thereby committed to the proposition that the constrained subject generates less. The account developed here is not so committed. What is asserted is that the subject generates otherwise, and the question whether the system’s measure registers what is otherwise generated is deferred to §4.3.

Endogenous constraint admits of two temporal phases, the acute and reversible and the permanent. The permanent phase is the difficult case identified in §1.3, and the standard responses to constraint fail against it.

4.2 Exogenous constraint: the unexecutable grammar

Exogenous constraint obtains where $F$ is unaltered and $E \rightarrow E’$ such that $F(S,E’)$ admits of no generation.

In this case the subject’s dynamics are unimpaired. The generative rule is complete and would generate, and no environment obtains within which it may be executed.

The Marxian case is paradigmatic: the capacity to labour is present, the means of production are absent, and their absence is not a property of the labourer [15]. Further instances include armed conflict, in which two parties retain their generative capacities and the relation between them forecloses co-production; institutional exclusion; and historical conditions that do not yet obtain.

The last of these requires the caution announced at the close of §3.3. The description of an idea as ahead of its time implies that the idea possessed a completeness which its environment failed to register, and thereby attributes latency to a proposition after latency has been denied of subjects. The relational description is that $F$ and $E$ co-evolve, and that a generative rule for which no environment obtains is not a frustrated rule. It is a rule for which the question what it would have generated has no determinate answer, since generation is not performed by a rule in isolation.

4.3 Epistemic constraint: the unread generation

Epistemic constraint obtains where $F$ is executed, $E$ permits execution, generation occurs, and the output lies outside the range of $M$.

The case is formally distinctive. Under endogenous constraint the rule is altered, and under exogenous constraint the environment is altered. Under epistemic constraint nothing is altered. $F$, $E$, and $M$ each stand as they stood; no transition has occurred; and the failure consists in a mismatch between the range of $M$ and the outputs that $F$ generates, a mismatch that obtained before any constraint was observed and would obtain in its absence. Nothing is blocked and nothing is redirected.

Care, the sustaining of another subject’s capacity to act, and the repair of damaged relations are instances. Each is generated continuously, in relation, and none is registered.

The third case is not a weaker instance of the first two, and the asymmetry between them is the central structural result of this section.

Claim 3. Under endogenous and exogenous constraint, generation does not occur, and the system’s observation that it has not occurred is correct. The system is not in error concerning the world; it is in error in what it infers from a correct observation, and the error is normative.

Under epistemic constraint, generation occurs, and the system’s observation that it has not occurred is false. The system is in error concerning the world, and cannot establish this from within, since the error is located in the instrument rather than in the inference drawn from it.

The two cases therefore require distinct responses. Their conflation, which the vocabulary of vulnerability encourages, results in the second being treated as an instance of the first: provision is made for a subject who requires no provision, and to whom what is owed is instead the acknowledgement that the system was unable to register what the subject was generating.

4.4 Sedimentation: the exogenous becoming endogenous

The three topologies are analytically distinct and empirically entangled. The most consequential entanglement is the following.

Exogenous constraint, sustained over time, alters the generative rule. Where a subject is deprived, for a sufficient duration, of the environment within which its rule may be executed, the rule changes. Learned helplessness, the atrophy of an unexercised skill, the interruption of a child’s development, and the cognitive load imposed by chronic scarcity are documented instances [16, 7]. In each, $F$ was unaltered, the environment was withheld, and $F$ was subsequently altered.

Claim 4 (sedimentation). A system that withholds from a subject the environment required for the execution of the subject’s generative rule thereby alters that rule, and may subsequently cite the altered rule in justification of the withholding. The judgement was false when it was issued and true when it was reviewed. Abandonment does not merely misjudge a subject; it produces the condition that the misjudgement had asserted.

An objection arises at this point and is here met rather than conceded. If constraint may sediment, by what criterion is an originally endogenous constraint to be distinguished from a sedimented one?

In many cases it cannot be. The demand that it must be is itself a productivist demand. Its function is to sort subjects into those whose constraint is attributable to themselves and those whose constraint is attributable to the system, and to calibrate obligation according to the sort. The account developed here does not perform that sort, since eligibility is not conditioned upon the etiology of the constraint. The topology governs the remedy and does not govern the eligibility. Diagnosis determines what is to be done and does not determine for whom it is to be done.

4.5 The relational character of severity

A further consequence follows, and it qualifies the account of endogenous constraint given above.

The severity of an endogenous constraint is not determined by the alteration of the rule alone. It is determined jointly by the alteration and by the degree to which the environment accommodates the altered rule:

$$\text{severity} ;=; f!\left(\text{alteration of } F,\ \text{non-accommodation by } E\right).$$

An identical impairment constitutes a constraint of different severity in an environment constructed to accommodate it than in an environment that is not. This is the social model of disability restated in generative terms [18], and the generative restatement yields an additional result.

The category purely endogenous is a limit concept. No internal alteration is met by an environment that is neither accommodating nor unaccommodating, and the severity that results is therefore jointly determined in every case. The case that appears most clearly internal to the subject, namely permanent and irreversible bodily alteration, is not on that account internal to the subject alone.

§5 The Epistemic Structure of Abandonment

The account developed above is ordered as follows: an ontology, from which an epistemology follows, from which normative consequences follow in turn.

Ontology → Epistemology → Ethics → Justice

The present section establishes the second register. Without it, the normative consequences of the account reduce to an appeal for more generous treatment of the excluded. With it, the account asserts something stronger: that the assessment by which exclusion is justified is itself defective.

5.1 The invalid inference

The system observes that no generation has been measured and concludes that no generation has occurred. The inference is invalid, and its invalidity is logical rather than normative. It proceeds from a proposition concerning a measurement to a proposition concerning the object measured, and the transition requires a premise that the system does not possess, namely that the instrument is neutral with respect to what it measures.

Claim 5. Abandonment commits an epistemic error prior to committing a normative one. It is in error concerning the world before it is in error concerning the subject.

A consequence follows which the paper takes to be among its principal results. Abandonment cannot be refuted by improved prediction. It cannot be answered by the assertion that the subject will recover, or will contribute at a later time, or will generate what was not anticipated. In some cases the subject will not. The refusal must be directed at the conditionality itself, and the reason it must be so directed is that the failure was never predictive in character. No refinement of forecasting corrects an instrument that does not register the object.

5.2 The measure as an element of the relation

The ground of the preceding claim is the following. Under the generative relational account, the criterion by which a system identifies an output as valuable is constituted by the same relational structure that the system purports to assess. The measure is not an instrument external to the relation and directed upon it. It is an element of the relation, produced by the same conditions and carrying the same commitments.

This is the consequence announced in §2.3, and it establishes Claim 1. Productivism cannot assert it. Under productivism the metric constitutes the definition of value rather than a reading of it, and the proposition the metric may be failing to register value is therefore not a caution but a contradiction in terms; the proposition cannot be formed. Under the generative relational account the proposition is well formed. The difference is that between a framework whose central claims are revisable in light of what it fails to observe and a framework whose central claims are not.

5.3 The generation that falls outside the measure

The outputs identified in §4.3 as unregistered include care, the sustaining of another subject’s capacity to act, the repair of damaged relations, and the production of the conditions under which a relation remains habitable.

Two interpretations of their non-registration are available. On the first, the failure is contingent: these outputs are in principle measurable, existing metrics are inadequate to them, and the remedy is the construction of better metrics. On the second, the failure is constitutive: outputs of this kind are not measurable by a metric of the form productivism requires, such that to render them measurable is to have altered them, and the demand that they be rendered measurable is therefore not a neutral demand.

The paper does not adjudicate between these interpretations. The second is stated as a question and not as a claim, since the author does not possess an argument sufficient to establish it. It is recorded here because the first interpretation, if adopted without argument, concedes what the account disputes, namely that the productivist form of measurement is the appropriate form and requires only refinement.

5.4 The dissolution of the humanist reply

The position set aside in §1.3 may now be rejected in full.

The humanist reply asserts that the excluded subject possesses a worth that has not been perceived. The assertion concedes that worth is a property lodged within subjects and either perceived or overlooked, and disputes the accuracy of the perception rather than the ontology that underwrites it. It is therefore without resources at the case where the perception is accurate, where nothing is lodged within the subject, and where the subject will generate nothing that any metric would register.

The account developed here requires no such possession.

Claim 6. Value is not located within the subject. The question what the subject contains does not arise, and the abandonment that turns upon that question is not refuted but dissolved: it is shown to have taken the wrong object.

5.5 The epistemic condition upon recognition

It follows that recognition cannot consist in the perception of an unrecognized worth. So construed, recognition is the humanist reply supplied with a superior instrument, and it fails whenever the instrument registers nothing.

Definition 6 (recognition). To recognize a constrained generative being is to hold open the possibility that the system’s measure is at fault. Recognition is an epistemic posture prior to being a normative one, and consists in the refusal to treat the range of one’s instrument as the boundary of what exists.

Recognition so defined is necessarily unconditional. Were it granted upon the belief that the subject possesses something of value, it would be revocable upon the finding that the subject does not. Recognition would then constitute a prediction concerning the subject, and predictions are adjudicated. Recognition is unconditional not because the outcome of the adjudication is known in advance, but because recognition does not consist in an adjudication.

5.6 Relation to existing accounts of epistemic injustice

The general form of the claim advanced in this section is established in the existing literature. Epistemic injustice [6], standpoint epistemology, the sociology of valuation [4], and the philosophy of measurement have each established that observation is not neutral and that instruments carry substantive commitments.

The present account specifies a location for the failure that these treatments do not. The claim is not that the observer is biased, nor that the observed is denied testimonial standing, but that the measure is an element of the relation whose output it reports. A system that assesses a subject’s generativity is, on this account, reading a criterion that the relation itself has produced, and treating that criterion as an independent report upon the world.

These connections are not developed here. The ontology is unaffected by them; only its epistemic consequences have been drawn.

§6 The Normative Consequence: Where the Remedy Is Applied

6.1 The location of the constraint determines the location of the remedy

The topologies established in §4 locate constraint in the conditions of generation. Endogenous constraint is jointly determined by the alteration of the generative rule and by the environment’s non-accommodation of the altered rule, as §4.5 established; exogenous constraint is located in the environment; and epistemic constraint is located in the measure. In no case is the constraint a property of the subject considered apart from its relations.

A normative consequence follows directly, and it does not require any independent premise concerning the subject’s worth.

Claim 7. Because constraint is located in the relational conditions of generation, any intervention that removes it is necessarily performed upon those conditions and is not performed upon the subject. The obligation to intervene follows from the location of the constraint and does not follow from an assessment of what the subject deserves.

The claim is materialist in form. What is altered is a set of conditions, and the alteration of conditions is an operation upon a system rather than upon a person. The account therefore yields the obligation to act without at any point requiring the judgement that the subject merits action.

6.2 The forms of intervention

The three topologies do not admit of a single form of intervention, and the differences are not differences of degree.

Exogenous constraint. The generative rule is intact and no environment permits its execution. The intervention consists in the provision of the environment. The removal of a barrier is the appropriate description of this case, and it is the case for which that description was devised.

Endogenous constraint. The generative rule is altered, and by Claim 2 the alteration is not a diminution. The intervention does not consist in the restoration of the prior rule, which in the permanent case is not available, and which in any case presupposes that the prior rule was the correct one. It consists in the reconstitution of the environment about the altered rule, so that $F’$ becomes executable in an environment that was constructed for $F$. By §4.5 there is in every case an environmental term upon which to act, since severity is jointly determined and no alteration is met by an environment that neither accommodates nor fails to accommodate it.

Epistemic constraint. There is no barrier. The generative rule is executed, the environment permits its execution, and generation occurs. Nothing is to be removed, and an intervention framed as removal is not merely ineffective but misdescribes the situation: it supplies assistance to a subject who requires none. The intervention consists in the revision of the measure.

Claim 8. The ethical task is not in general the removal of barriers to generation. It is action upon the relational conditions of generation, of which the removal of barriers is the form appropriate to the first two topologies and the revision of the measure is the form appropriate to the third.

6.3 The obligation to the disabled subject

The account yields the obligation to act on behalf of the disabled subject without the two premises that such an obligation is ordinarily thought to require.

It does not require the premise that the disabled subject possesses a capacity that has not been recognized. By Claim 6, value is not located within the subject, and the appeal to unrecognized capacity is unavailable and unnecessary.

It does not require the premise that the disabled subject will generate at a later time. By Claim 5, the inference from unmeasured generation to absent generation is invalid, and the obligation does not rest upon a prediction about what will subsequently be measured.

What it requires is the location of the constraint. Disability, on the account developed here, is a constraint jointly constituted by an altered generative rule and an environment constructed for a rule the subject does not instantiate. The environment is a term of the constraint, and the environment is alterable. The obligation to alter it follows.

The resulting relation between the parties is not that of donor and recipient. To act upon the environment is to act upon a relation of which the acting party is already a term, and the alteration alters the conditions under which that party generates as well. The asymmetry that renders assistance a gift, and the recipient of assistance a debtor, does not arise, because no transfer of value from a possessor to a non-possessor has occurred. What has occurred is a change in the conditions of a relation.

6.4 The relation to care ethics

The conclusion converges with the conclusions of care ethics, and the route to it differs [22, 14]. Care ethics establishes the moral standing of dependency and the moral significance of the practices that sustain it, and it does so, for the most part, by defending the value of what care produces and the standing of those who receive it. The account developed here does not defend care by exhibiting its productivity, and does not defend the dependent subject by exhibiting a worth that has been overlooked. It holds that the question of what the subject possesses does not arise, and that the practices in question are alterations of relational conditions rather than transfers to a deficient party.

The convergence is therefore not an appropriation. Care ethics and the present account arrive at overlapping prescriptions from incompatible ontologies, and the difference in ontology is not inconsequential: it determines whether the obligation survives the case in which the dependent subject will generate nothing that any measure registers. The present account holds that it does.

References

[1] Anderson, E. (1993). Value in Ethics and Economics. Harvard University Press.

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

[3] Eglash, R., Bennett, A., Babbitt, W., Lachney, M., Reinhardt, M., & Hammond-Sowah, D. (2020). Decolonizing posthumanism: Indigenous material agency in generative STEM. British Journal of Educational Technology, 51(4), 1334–1353.

[4] Espeland, W. N., & Stevens, M. L. (1998). Commensuration as a social process. Annual Review of Sociology, 24, 313–343.

[5] Fraser, N., & Honneth, A. (2003). Redistribution or Recognition? A Political-Philosophical Exchange. Verso.

[6] Fricker, M. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford University Press.

[7] Herman, J. L. (1992). Trauma and Recovery. Basic Books.

[8] Honneth, A. (1995). The Struggle for Recognition: The Moral Grammar of Social Conflicts (J. Anderson, Trans.). Polity Press.

[9] 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

[10] Huang, W. (2026). The Relational Mode of Being of Concepts: Towards a Generative Relational Epistemology. Serendip Commons Society. GRB Series.

[11] Huang, W. (2026). A Theory of Relational Reproduction: Definition and Ethical Principles. Knowledge Commons. https://doi.org/10.17613/qncwr-zbt82

[12] Huang, W. (2026). Relational Value under the Symbolic Crisis. Knowledge Commons. https://doi.org/10.17613/bm32z-aqt14

[13] Huang, W. (2026). Relational Wealth: Definition and Characteristics. Knowledge Commons. https://doi.org/10.17613/nq7js-vz283

[14] Kittay, E. F. (1999). Love’s Labor: Essays on Women, Equality, and Dependency. Routledge.

[15] Marx, K. (1988). Economic and Philosophic Manuscripts of 1844 (M. Milligan, Trans.). Prometheus Books. (Original work written 1844.)

[16] Mullainathan, S., & Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much. Times Books.

[17] Nussbaum, M. C. (2006). Frontiers of Justice: Disability, Nationality, Species Membership. Harvard University Press.

[18] Oliver, M. (1990). The Politics of Disablement. Macmillan.

[19] Raz, J. (1986). The Morality of Freedom. Clarendon Press.

[20] Sen, A. (1999). Development as Freedom. Oxford University Press.

[21] Shiffrin, S. V. (1996). Wrongful life, procreative responsibility, and the significance of harm. Legal Theory, 5(2), 117–148.

[22] Tronto, J. C. (1993). Moral Boundaries: A Political Argument for an Ethic of Care. Routledge.

中文

受约束的生成性存在者

论被度量的生成之缺席

一份初步讨论

黄万宏 · huangwanhong@serendip.ngo

摘要

生成性关系框架主张，价值是在关系中被生成的，而不是主体先于关系就已拥有的。本文把这项承诺置于它最艰难的案例之下，即根本没有任何生成被度量到的案例。分配正义的诸理论关切的是善物在参与者之间的分配，而剥削的诸理论关切的是价值自其生产者手中被占有；二者都不处理那个什么也不生产的主体。本文的主要结论是：被度量的生成之缺席不可被等同于生成性存在之缺席；而一个由前者推出后者的系统，在它就主体犯错之前，已经就世界犯了错。本文引入”受约束的生成性存在者”这一概念，用以指称一个生成可能性的场址，其对生成性关系的参与被关系性条件所阻断、改道，或弄成不可辨认。该概念既区别于软弱与脆弱的诸概念，后者把一项欠缺定位于主体之内；也同样区别于那个人道主义的回应，即被排斥的主体拥有一份未被察觉的价值，该回应承认价值寓于主体之内，而只对清单的准确性提出争议。本文区分约束的三种拓扑。在内源性约束之下，生成规则被改变；在外源性约束之下，规则完好无损，而没有任何环境容许它被执行；而在本文所引入的认识性约束之下，规则被执行，环境容许执行，生成发生了，而系统的度量并不登记它。第三种情形并不是前两种的一个较弱的实例：在前两种之下，系统观察正确而推论错误；而在第三种之下，系统的观察为假，并且无法从内部得到纠正，因为错误位于仪器之中，而不在推论之中。由此本文导出它的规范性后果。由于约束位于生成的关系性条件之中，补救便必然是施行于那些条件之上，而绝不施行于主体之上；而行动的义务出自约束的位置，而不出自对主体价值的任何评估。

关键词： 生成性关系存在；受约束的生成性存在者；约束本体论；内源性约束与外源性约束；认识性约束；遗弃；承认；价值度量的关系性；生产主义；残障。

AI 使用声明： 本文的思想、主张与理论框架均出自作者。AI 工具被用作起草与修订过程中的辅助。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

预先说明。本文是一份初步讨论。它的论证依赖于一个被断言而非被推导出来的原初设定；它引入了一个作者未曾在既有文献中发现被命名过、并且尚不能抵御每一项反驳的约束范畴；并且，在本文的说明抵达其界限之处，它留下了一个它不知如何了结的问题。这三者都在文中它们各自出现之处被标出。

1.1 检验该框架的那个案例

生成性关系框架主张，价值是在关系中被生成的，而不是主体先于关系就已拥有的。本文把这项承诺置于它最难以维持的案例之下，即根本没有任何生成被度量到的案例。

这个案例并不边缘。占据它的是：疾病终结了其工作的主体，年岁终结了其工作的主体，在某次断裂之后无法行动的主体，以及在某次断裂之后不会再行动的主体。分配正义的诸理论关切的是善物在参与者之间的分配。剥削的诸理论关切的是价值自其生产者手中被占有。二者都预设主体与生产处于某种确定的关系之中，或者作为对已被生产之物的请求者，或者作为一个其所生产之物已被拿走的生产者。二者都不处理那个什么也不生产的主体；没有任何价值自他手中被占有，因为没有任何价值进入过流通。

因此，本文所处理的问题如下：

在何种条件之下，一个什么也不生成的主体仍然是生成性关系中的一个参与者？

提出这个问题并不是为了替某一类主体辩护。提出它，是因为该框架的核心承诺蕴涵它。如果价值是主体带入其往来之中的一项属性，那么一个什么也不生成的主体，就是一个拥有较少价值的主体，而关于他的地位的问题，就化约为”欠那些拥有得少的人什么”的问题。如果价值是在关系中被生成的，那么”没有什么被生成”这一观察，就是一个关于某段关系的观察，而尚不是关于任何主体的观察。因此，上文所指名的那些主体，是该本体论受到检验的条件，而不是本文的题材。所争的是那套本体论，而主要结论是关于它的结论。

主张 1（主要结论）。 被度量的生成之缺席，不可被等同于生成性存在之缺席。

1.2 术语

有三个术语被贯穿使用，此处加以定义。

定义 1（生成性关系）。 一段生成性关系，是主体与其诸条件之间的一种关系，价值在其中被生成。对一段生成性关系的参与，在于作为这样一种关系的一个项而存在，而不在于任何产出物的生产。

定义 2（生产主义）。 生产主义是这样一种本体论立场：价值是产出物的一项属性；连同由之而来的派生立场：一个主体在价值系统之内的地位，以该主体对此类产出物的生产为条件。此处并不把该立场归于任何特定作者。

定义 3（遗弃）。 遗弃是撤回一个主体作为生成性关系之参与者的地位，其理由是该主体的生成未被系统的度量所登记。

如此定义的遗弃，不是忽视的同义词，也不指称一种情感倾向。它指称一项推论，连同该推论被认为许可的那种地位之撤回。本文关于该推论的说明见于 §5。

1.3 被拒绝的立场

有一个立场必须在开篇就予以搁置，因为它是读者最可能归给本文的立场，也因为它比此处所辩护的立场更弱。

对遗弃的人道主义回应主张：被排斥的主体拥有一份未被察觉的价值，不义在于知觉的失败，而补救之道是一份更慷慨的评估。这一回应承认了生产主义的本体论，而只对其清单的准确性提出争议：它接受价值是寓于主体之内的一项属性，而只与记账方式争执。因此，恰恰在最需要资源的那个案例上，它毫无资源可用，即：那个不会康复、此后不会有所贡献、并且没有任何诚实的观察者能够预言其回报的主体。面对那个案例，诉诸未被承认的潜能无话可说，因为并不存在这样的潜能。

此处所辩护的立场，不要求在主体之内找到任何东西。价值根本不位于主体之内。它是在关系中被生成的；因此，它是否显现，是一个关于该关系的事实，而不是一个该关系所隐藏着的事实。这不是对被遗弃的主体作出的一份更宽大的裁决；它是对”主体本是裁决的恰当对象”这一前提的拒绝。

那个案例统辖以下的一切。一个框架可以由它被围绕着构造的那个案例来评估；而这一个框架被围绕着构造的案例，不是那个将会毕业的学生，不是那个将会康复的病人，也不是那个缺乏指导、一旦获得指导便会兴盛的研究者。那些案例是真实的，而框架必须处理它们，但它们并不困难；因为一种足够有耐心的生产主义同样能够处理它们：它只需延长它的会计期间，并等待回报。困难的案例是那个不会有任何回报到来的主体。一个只容纳暂时受约束者的框架，并未离开生产主义，而只是推迟了结算。

1.4 本文的计划

§2 陈述生成性关系的说明所要求的对生产主义次序的颠倒，并宣告该说明所依赖的原初设定。§3 定义受约束的生成性存在者，并把该概念与软弱和脆弱的诸概念区别开来。§4 区分约束的三种拓扑，其中第三种是在此首次被引入的。§5 确立认识论上的结论。§6 导出规范性后果，它关乎补救应当被施行于何处。

关于生成正义的相邻工作，关切的是价值在那些已然从事生成的人之间的流通 [2, 3]。此处所处理的是一个不同且更为在先的问题，而以下所提出的一切，都无意作为对那项工作的纠正。

§2 从生产能力到关系能力

2.1 生产主义的次序

如 §1.2 所定义的生产主义，把能力、生产与承认排列如下。

能力 → 生产 → 承认

依这一排序，能力是主体的一项属性，生产是该项属性的行使，而承认是生产所换得的东西。这一排序使承认成为有条件的：凡在一个序列的终点被授予的东西，都可能在它的起点被扣留。

有三种熟悉的立场对这一排序的条款提出争议，同时接受该排序本身。第一种主张，进入该序列的门径应当更具包容性，其理由是能力的分布比通常所设想的更为广泛。第二种主张，进入它应当更有耐心，其理由是能力在有利的条件之下会得到发展。第三种主张，承认应当基于独立的根据而被授予，即作为人的身份问题或权利问题，与该序列并行，而不是经由它。这三者在其处方上彼此不同，而一致认为该序列描述了问题的结构。

2.2 关系性的颠倒

生成性关系的说明在该排序的第一项上就拒绝它。能力不是主体带入一段关系的一项属性；它是关系的一个结果。

这一主张可以用一个案例来说明。一个没有指导、没有智识共同体、也没有时间的研究者，被生产主义描述为拥有一项其境遇阻止她行使的能力。关系性的说明给出一个不同的描述：该能力并不以一种未被行使的状态存在着。并不存在一份未经指导的洞见的储备，被持存于主体之内，等待着有利的条件。指导并不释放一项能力；它参与构成一项能力。在该关系缺席之处，那项能力不是被隐藏了，而是不存在。

这一主张带有一项应当被记录的代价。由此可得：本文不能代表一个什么也不生成的主体，去诉诸那个主体潜在地含有什么。凡一个主体什么也不生成之处，本文无权回应说该能力仍然在场。它必须给出一个不同的回应；而本文的其余部分正是陈述这样一个回应的尝试。

2.3 原初设定

本文的说明要求一项基础性的承诺，此处不加推导地陈述它。

定义 4（原初设定）。 生成是关系性的：价值是在主体与其诸条件之间的关系中被生成的，而不是主体先于那段关系就已拥有的一项属性。

本文不提供对这项承诺的推导，也不需要提供。每一个框架都依赖于一个它并不推导的原初设定。生产主义依赖于”价值是产出物的一项属性”这一原初设定；它假定这一点，在该假定之上构造一套度量，并把该度量的读数当作关于世界的发现来报告。向一个框架提出的恰当问题，不是它是否已证明了它的原初设定，因为没有任何框架做到过；而是：由它的原初设定所推出、而由那个替代设定所不推出的东西是什么。

后果如下。依生成性关系的说明，据以评估生成的那套度量，其本身就是该关系的一个要素，因而是可供怀疑的。生产主义无法把它自己的度量置于怀疑之下；因为对生产主义而言，度量构成的是价值的定义，而不是对价值的一次读数。命题”这个主体什么也不生成，而那套度量仍然可能有误“，在生产主义之内不是一句告诫，而是一个矛盾。在生成性关系的说明之内，该命题是良构的。§5 展开这一后果，而它构成了持守该原初设定的依据。

文献中有两种对该原初设定的推导，而二者在此都被谢绝。第一种主张：差异引起互动，互动引起新颖的格局，而新颖的格局引起生成。这一推导是目的论的：它把新颖性安置为生成所朝向的目的；而一个生成具有目的的框架，正是一个可以提出”某个给定主体是否服务于那个目的”这一问题的框架。那个问题正是本文所要排除的工具性问题。第二种推导主张：没有任何孤立的实体能够生成关系性属性，因此生成要求关系。这一推论是有效的，而其结论是平凡的：依这些术语的定义，关系性属性本就要求关系，而关于价值则什么也推不出来。

应当记录的是：似乎不存在对该原初设定的任何非循环的证成。任何一个”度量可能未能登记价值”的论证，都已然预设了度量内在于它所报告其产出的那段关系，而这正是该原初设定本身。这种形式的论证确立的是由该原初设定所推出的东西，而并不确立该原初设定。因此，上文所陈述的后果是作为一项产出而被展示的，而不是作为一项证明而被提出的。

因此，该原初设定被断言，而其后果被展示。一位认为这些后果不足的读者，也就因此被给予了拒绝该原初设定的根据。

§3 受约束的生成性存在者

3.1 定义

定义 5（受约束的生成性存在者）。 一个受约束的生成性存在者，是一个生成可能性的场址，其对生成性关系的参与被关系性条件所阻断、改道，或弄成不可辨认。

该定义中有三个要素需要评注。

生成可能性的场址这一术语，须与”潜在价值的容器”这一观念区别开来。把一个主体指定为一个生成可能性的场址，是断言生成可能发生在那个主体作为其一项的关系之中。它不是断言任何东西被储存在该主体之内。这一区分正是 §1.3 所作的区分，并被贯穿维持。

该定义中的三个动词，即阻断、改道与弄成不可辨认，并不指称同一种失败的三个程度。§4 确立它们指称结构上彼此不同的处境，并且第三种在种类上不同于前两种。

被关系性条件这一短语指认出约束的场址。该场址不是主体。

3.2 该概念与软弱和脆弱相区别

软弱的主体与脆弱的主体这两个概念，把一项欠缺定位于主体之内，并把该项欠缺当作对主体之被排斥的解释。

“把一项欠缺定位于主体之内”的替代方案，并不是”把一份未被承认的充足定位于主体之内”。这两种操作都接受：主体是解释应当于其中被找到的场址。此处所推进的立场，可以被称作约束本体论，它拒绝那个假定。它不是对主体作出的一份经修订的裁决；它是对”主体是裁决之对象”这一点的拒绝。

因此，约束本体论以”生成为何未能发生”这一问题，取代了”一个主体是否拥有某项能力”这一问题；后者的两个可用答案同样都是生产主义的。那个问题容许不止一种形式的答案，而诸种形式在 §4 中被区分开来。

3.3 约束的诸层级

约束发生于三个层级，而分析在这三个层级上是一致的。

在个体层级上，约束刻画学生、病人以及自伤病中恢复的主体的处境。

在社会层级上，它刻画那些被排除在生成本会于其中发生的诸制度之外的群体的处境。

在历史层级上，它刻画那些其接纳条件尚不成立的思想与技术。把此类案例描述为超前于其时代，在 §4 中被拒绝，其理由是：这一描述在命题的层级上重新引入了那个已在主体的层级上被拒绝的潜在性观念。

§4 约束的三种拓扑

设生成性系统写作

$$S_{t+1} ;=; F(S_t,,E_t),$$

其中 $S$ 记主体的内部动力学，$E$ 记环境，而 $F$ 记生成规则，即一个主体与一个环境据以共同决定何物被生成的那个结构。设 $M$ 记度量，即周围系统据以把某一产出物辨识为生成的那个判准。

该记法是作为一套记法、而不是作为一个模型被引入的。它的功能是把日常语言并不强制的一项区分弄得明确，即：生成的失败可能位于 $F$ 之中，位于 $E$ 之中，或位于 $M$ 之中；而这些位置并不是彼此的程度。这三种情形分别被称为内源性约束、外源性约束与认识性约束。

4.1 内源性约束：被改变的语法

内源性约束成立于 $F \rightarrow F’$ 之处：生成规则本身被改变了。主体的动力学并不是在量级上被削减，而是在种类上被改变。参数空间不同于它先前所是，而可达的状态也相应地不同。

疾病、神经退行、残障与创伤例示了这一情形。创伤最清楚地例示它，因为那项改变不是一次减法。原先的规则是

$$\text{主体} ;\longrightarrow; \text{世界} ;\longrightarrow; \text{行动},$$

它变成

$$\text{主体} ;\longrightarrow; \text{威胁} ;\longrightarrow; \text{防御}.$$

没有任何项被移除；只有一个项被另一个项所替换。所得的系统仍然是完全动力学的、完全生成性的，并且生成着与先前不同的产出物。

主张 2。 内源性约束在于 $F’ \neq F$，而不在于 $F’ < F$。被改变的规则不是一条被削减的规则。它是一条不同的规则，生成着不同的产出物；而它的产出物并不因此就更少。

这一区分不是术语上的。一个把内源性约束理解为削减的框架，就承诺了按量级对生成规则进行排序，因而承诺了”受约束的主体生成得更少”这一命题。此处所发展的说明并无此项承诺。所断言的是：主体生成得不同；而系统的度量是否登记那被不同地生成的东西，这一问题被推迟至 §4.3。

内源性约束容许两个时间相位，即急性且可逆的相位与永久的相位。永久的相位正是 §1.3 所指认的那个困难案例，而对约束的标准回应在它面前失效。

4.2 外源性约束：无法执行的语法

外源性约束成立于这样的情形：$F$ 未被改变，而 $E \rightarrow E’$，以致 $F(S,E’)$ 不容许任何生成。

在这种情形下，主体的动力学未受损伤。生成规则是完备的，并且本会生成；而不存在任何它可以在其中被执行的环境。

马克思式的案例是范式性的：劳动的能力在场，生产资料缺席，而它们的缺席不是劳动者的一项属性 [15]。进一步的实例包括：武装冲突，其中双方都保有其生成能力，而它们之间的关系封闭了共同生产；制度性排斥；以及尚不成立的历史条件。

其中最后一项要求 §3.3 结尾处所宣告的那项审慎。把一个思想描述为超前于其时代，隐含着该思想拥有一份其环境未能登记的完备性，因而在潜在性已被否认于主体之后，又把潜在性归给了一个命题。关系性的描述是：$F$ 与 $E$ 共同演化；而一条没有任何环境成立于其上的生成规则，不是一条受挫的规则。它是这样一条规则：关于”它本会生成什么”的问题没有确定的答案，因为生成不是由一条孤立的规则所执行的。

4.3 认识性约束：未被读出的生成

认识性约束成立于这样的情形：$F$ 被执行，$E$ 容许执行，生成发生了，而产出物落在 $M$ 的量程之外。

这一情形在形式上是独特的。在内源性约束之下，规则被改变；在外源性约束之下，环境被改变。而在认识性约束之下，没有任何东西被改变。$F$、$E$ 与 $M$ 各自一如其先前所是；没有任何转变发生过；而失败在于 $M$ 的量程与 $F$ 所生成的产出物之间的一种失配，这种失配在任何约束被观察到之前就已成立，并且在它缺席时也仍将成立。没有任何东西被阻断，也没有任何东西被改道。

照护、维系另一个主体行动能力的工作，以及受损关系的修复，都是实例。它们各自都在关系中被持续地生成，而其中没有一项被登记。

第三种情形不是前两种的一个较弱的实例；而它们之间的不对称，正是本节的核心结构性结论。

主张 3。 在内源性约束与外源性约束之下，生成并未发生，而系统关于”它未曾发生”的观察是正确的。系统并未就世界犯错；它是在它由一个正确的观察所推出的东西上犯错，而那个错误是规范性的。

在认识性约束之下，生成发生了，而系统关于”它未曾发生”的观察为假。系统就世界犯了错，并且无法从内部确立这一点，因为错误位于仪器之中，而不在由它所作出的推论之中。

因此，这两种情形要求彼此不同的回应。它们的混同，正是脆弱性的词汇所鼓励的，其结果是第二种被当作第一种的一个实例来处理：为一个不需要供给的主体作出了供给，而欠他的东西其实是这样一份承认，即：系统未能登记他一直在生成的东西。

4.4 沉积：外源性者成为内源性者

三种拓扑在分析上彼此有别，而在经验上彼此纠缠。最有后果的一种纠缠如下。

外源性约束，若持续足够长的时间，会改变生成规则。凡一个主体在足够长的时段内被剥夺了其规则可以在其中被执行的环境之处，那条规则就会改变。习得性无助、一项未被行使的技能的萎缩、一个儿童发展的中断，以及长期匮乏所施加的认知负荷，都是有文献记载的实例 [16, 7]。在其中每一个之中，$F$ 原本未被改变，环境被扣留，而 $F$ 此后被改变了。

主张 4（沉积）。 一个系统若扣留一个主体执行其生成规则所需要的环境，它便因此改变了那条规则；而它此后可以援引那条被改变的规则，来为那次扣留辩护。该判断在它被作出之时为假，而在它被复核之时为真。遗弃不只是误判一个主体；它生产出那项误判曾经断言的那种状况。

此处出现一项反驳，而此处是回应它，而不是让步于它。如果约束可以沉积，那么依据什么判准，一项原本就是内源性的约束才能与一项沉积而来的约束区别开来？

在许多案例中，它不能被区别开来。而”它必须能够被区别开来”这一要求，其本身就是一项生产主义的要求。它的功能是把主体分拣为两类，即那些其约束可归咎于其自身的主体，与那些其约束可归咎于系统的主体，并据此分拣来校准义务。此处所发展的说明并不执行那种分拣，因为受助资格并不以约束的病因为条件。拓扑统辖补救，而不统辖资格。诊断决定应当做什么，而不决定应当为谁而做。

4.5 严重程度的关系性性质

由此推出一项进一步的后果，而它限定了上文关于内源性约束所给出的说明。

一项内源性约束的严重程度，并不单由规则的改变所决定。它是由那项改变、以及环境容纳被改变之规则的程度共同决定的：

$$\text{严重程度} ;=; f!\left(F \text{ 的改变},\ E \text{ 的不容纳}\right).$$

同一项损伤，在一个被构造来容纳它的环境之中，与在一个并非如此的环境之中，构成的是严重程度不同的约束。这就是残障的社会模型以生成性的术语被重述 [18]；而这一生成性的重述产出一项额外的结论。

纯粹内源性这一范畴是一个极限概念。没有任何内部改变，是由一个既不容纳也非不容纳的环境所迎接的；因此，所产生的严重程度在每一个案例中都是被共同决定的。那个看上去最清楚地内在于主体的案例，即永久且不可逆的身体改变，并不因此就仅仅内在于主体。

§5 遗弃的认识结构

上文所发展的说明按如下次序排列：一套本体论，由之推出一套认识论，再由之依次推出诸规范性后果。

本体论 → 认识论 → 伦理学 → 正义

本节确立第二个层域。没有它，该说明的规范性后果便化约为一份对被排斥者予以更慷慨对待的呼吁。有了它，该说明便断言某种更强的东西：据以证成排斥的那项评估，其本身就是有缺陷的。

5.1 无效的推论

系统观察到没有任何生成被度量到，并断定没有任何生成发生过。这一推论是无效的，而它的无效是逻辑上的、而不是规范上的。它从一个关于一次度量的命题，推进到一个关于被度量之对象的命题；而这一过渡要求一个系统并不拥有的前提，即：仪器相对于它所度量之物是中立的。

主张 5。 遗弃在犯下一项规范性错误之前，先犯下一项认识性错误。它在就主体犯错之前，先就世界犯了错。

由此推出一项后果，本文认为它属于其主要结论之列。遗弃不可能被改进的预测所驳倒。它不可能被这样的断言所回应：该主体将会康复，或将会在日后有所贡献，或将会生成未曾被预料的东西。在某些案例中，该主体不会。拒绝必须被导向那种条件性本身；而它之所以必须被如此导向，是因为那项失败在性质上从来就不是预测性的。任何预测的精细化，都无法纠正一个不登记对象的仪器。

5.2 度量作为关系的一个要素

前一项主张的根据如下。依生成性关系的说明，一个系统据以把某一产出物辨识为有价值的那个判准，是由它声称要加以评估的那同一个关系性结构所构成的。度量不是一件外在于关系并被指向关系的仪器。它是关系的一个要素，由同样的条件所产出，并携带同样的承诺。

这正是 §2.3 所宣告的那项后果，而它确立主张 1。生产主义无法断言它。依生产主义，度量构成的是价值的定义，而不是对价值的一次读数；因此命题”该度量可能未能登记价值“不是一句告诫，而是一个用语上的矛盾；该命题无法被构成。依生成性关系的说明，该命题是良构的。这一差异，就是一个其核心主张可依据它所未能观察到的东西而被修正的框架，与一个其核心主张不可如此被修正的框架之间的差异。

5.3 落在度量之外的生成

§4.3 所指认的那些未被登记的产出物包括：照护，维系另一个主体行动能力的工作，受损关系的修复，以及使一段关系保持宜居的那些条件的生产。

对它们之不被登记，有两种解释可用。依第一种，该项失败是偶然的：这些产出物在原则上是可度量的，既有的度量对它们不敷使用，而补救之道是构造更好的度量。依第二种，该项失败是构成性的：此类产出物无法被生产主义所要求的那种形式的度量所度量，以致把它们弄成可度量的，就是已经改变了它们；因此，”把它们弄成可度量的”这一要求并不是一项中立的要求。

本文并不在这两种解释之间作出裁决。第二种是作为一个问题、而不是作为一项主张被陈述的，因为作者并不拥有足以确立它的论证。它之所以被记录于此，是因为第一种解释若不加论证地被采纳，便让渡了本文的说明所争议的东西，即：生产主义形式的度量是恰当的形式，而只需要精细化。

5.4 人道主义回应的消解

§1.3 所搁置的那个立场，如今可以被完全拒绝。

人道主义的回应断言：被排斥的主体拥有一份未被察觉的价值。这一断言承认了价值是寓于主体之内、并且或被察觉或被忽略的一项属性，而它所争议的是知觉的准确性，而不是为之作担保的那套本体论。因此，在知觉是准确的、在主体之内并无任何东西寓居的、并且该主体将不会生成任何度量所能登记之物的那个案例上，它毫无资源可用。

此处所发展的说明并不要求任何这样的拥有。

主张 6。 价值不位于主体之内。”主体含有什么”这一问题并不出现；而系于那个问题的遗弃，并不是被驳倒，而是被消解：它被表明取错了对象。

5.5 承认之上的认识性条件

由此可得：承认不可能在于对一份未被承认的价值的知觉。若如此理解，承认便是被配上了一件更优良仪器的人道主义回应；而每当仪器什么也登记不到时，它便失效。

定义 6（承认）。 承认一个受约束的生成性存在者，就是保持”系统的度量可能有误”这一可能性的敞开。承认首先是一种认识性的姿态，其次才是一种规范性的姿态；它在于拒绝把自己仪器的量程当作存在之物的边界。

如此定义的承认必然是无条件的。倘若它是基于”该主体拥有某种有价值的东西”这一信念而被给予的，那么它便可因”该主体并不拥有”这一发现而被撤销。承认于是将构成一项关于该主体的预测，而预测是要被裁定的。承认之所以是无条件的，并不是因为裁定的结果预先已知，而是因为承认根本不在于一次裁定。

5.6 与既有的认识不义诸说明的关系

本节所推进的主张，其一般形式在既有文献中已被确立。认识不义 [6]、立场认识论、估值社会学 [4] 以及度量哲学，都各自确立了观察并非中立，而仪器携带着实质性的承诺。

本文的说明为该项失败指定了一个这些处理所不指定的位置。所主张的不是观察者有偏见，也不是被观察者被剥夺了证言上的地位，而是：度量是它所报告其产出的那段关系的一个要素。依本文的说明，一个评估某主体之生成性的系统，正是在读取一个由该关系本身所产出的判准，并把那个判准当作一份关于世界的独立报告。

这些关联在此并未展开。本体论不受它们影响；此处所导出的只是它的认识性后果。

§6 规范性后果：补救被施行于何处

6.1 约束的位置决定补救的位置

§4 所确立的诸拓扑把约束定位在生成的条件之中。内源性约束，如 §4.5 所确立的，是由生成规则的改变与环境对被改变之规则的不容纳共同决定的；外源性约束位于环境之中；而认识性约束位于度量之中。在任何一种情形中，约束都不是脱离其诸关系而被考虑的主体的一项属性。

一项规范性后果直接随之而来，而它并不要求任何关于主体价值的独立前提。

主张 7。 由于约束位于生成的关系性条件之中，任何移除它的干预，都必然是施行于那些条件之上，而不是施行于主体之上。干预的义务出自约束的位置，而不出自对”主体配得什么”的评估。

这一主张在形式上是唯物主义的。被改变的是一组条件，而条件的改变是对一个系统、而不是对一个人的操作。因此，本文的说明产出了行动的义务，而在任何一点上都不要求作出”主体值得被行动”这一判断。

6.2 干预的诸形式

三种拓扑不容许单一形式的干预，而这些差异不是程度上的差异。

外源性约束。 生成规则完好无损，而没有任何环境容许它被执行。干预在于提供该环境。”移除一道障碍”是对这一情形的恰当描述，而它正是那一描述被创设来针对的情形。

内源性约束。 生成规则被改变了；而依主张 2，该项改变不是一次削减。干预不在于恢复先前的规则，因为在永久的情形中，先前的规则不可得，并且无论如何，这样做都预设了先前的规则是正确的那一条。它在于围绕被改变的规则重构环境，使得 $F’$ 在一个原本为 $F$ 而构造的环境之中变得可执行。依 §4.5，在每一个案例中都存在一个可供施行的环境项；因为严重程度是被共同决定的，而没有任何改变是由一个既不容纳也非不容纳它的环境所迎接的。

认识性约束。 并不存在障碍。生成规则被执行，环境容许它被执行，而生成发生了。没有什么可供移除；而一项以移除为框架的干预，不只是无效，而且误描述了处境：它向一个不需要援助的主体供给了援助。干预在于对度量的修正。

主张 8。 伦理任务一般而言不是移除生成的障碍。它是对生成的关系性条件的行动；其中，移除障碍是适用于前两种拓扑的形式，而修正度量是适用于第三种拓扑的形式。

6.3 对残障主体的义务

本文的说明产出了代表残障主体而行动的义务，而无需那两个通常被认为该项义务所要求的前提。

它不要求”残障主体拥有一项未被承认的能力”这一前提。依主张 6，价值不位于主体之内，而诉诸未被承认的能力既不可用也不必要。

它不要求”残障主体将在日后有所生成”这一前提。依主张 5，从”生成未被度量到”到”生成缺席”的推论是无效的；而该项义务并不依赖于一项关于”此后将有什么被度量到”的预测。

它所要求的是约束的位置。依此处所发展的说明，残障是一项由被改变的生成规则、与一个为该主体并不例示的规则而构造的环境所共同构成的约束。环境是该约束的一个项，而环境是可改变的。改变它的义务随之而来。

由此产生的各方之间的关系，不是捐赠者与受赠者的关系。对环境施行行动，就是对一段行动方本已作为其一项的关系施行行动；而该项改变同时也改变了那一方据以生成的条件。那种使援助成为一份礼物、并使援助的接受者成为一个欠债者的不对称，并不出现；因为并没有任何价值从一个拥有者转移到一个非拥有者手中。所发生的是一段关系的条件的一次改变。

6.4 与照护伦理的关系

本文的结论与照护伦理的诸结论趋同，而通向它的路径不同 [22, 14]。照护伦理确立了依赖的道德地位，以及维系依赖的那些实践的道德意义；而它这样做，大体上是通过为照护所产出之物的价值、以及接受照护者的地位作辩护。此处所发展的说明并不通过展示照护的生产性来为照护辩护，也不通过展示一份被忽略的价值来为依赖的主体辩护。它主张：”主体拥有什么”这一问题并不出现；而所论的那些实践是关系性条件的改变，而不是向一个有欠缺的一方所作的转移。

因此，这种趋同并不是一次挪用。照护伦理与本文的说明，从彼此不相容的本体论出发，抵达彼此重叠的处方；而本体论上的差异并非无关紧要：它决定了该项义务能否在”依赖的主体将不会生成任何度量所能登记之物”这一案例中存活下来。本文的说明主张它能。

参考文献