Interpretive Justice in Organizations - A Preliminary Discussion of Problems and Open Questions 【（Preliminary）Draft】

ENGLISH

Interpretive Justice in Organizations

A Preliminary Discussion of Problems and Open Questions

Wanhong Huang · huangwanhong@serendip.ngo

---

Abstract

Organizations govern by rules, and rules do not apply themselves. Between a rule and a case stands an act of interpretation, performed by an organ of the organization, which fixes what the rule shall have meant for that case. This paper takes that act as its object. It argues that the interpretive act is a distinct site at which injustice can occur, irreducible to the injustice of what an organization distributes or to the injustice of the procedures by which it decides. Four questions organize the discussion: who holds standing to interpret, by what modes interpretation proceeds, which rules and which predicates carry the interpretive load, and what protects a party against whom an interpretation is exercised. The paper surveys what jurisprudence, legislative drafting, standard-setting, administrative review, and organizational governance have made of these questions, across several legal traditions, and finds a boundary common to every device assembled in answer to the fourth: each presupposes that the interpreting body holds some stake in recognition, whether legal, market, reputational, or professional, and each approaches ineffectiveness as that stake approaches zero. Where authority over interpretation and dependence upon recognition are held by the same party, no available instrument reaches the interpretive act. Against this boundary the paper sets a generative relational reading, on which an interpretation is assessed by what it does to the capacity of the relation between interpreter and interpreted party to generate a further interpretation. Injustice, on that reading, is the extinction of that capacity, and the sanction upon it is delegated to the dynamics of the structure itself, under guardrails the paper states explicitly and at length. The account supplies no enforcement and no remedy to a present party, and the paper closes on the questions this leaves open.

Keywords: interpretation; organizational rules; justice; discretion; accountability; asymmetry; relational generation; governance.

A note on the standing of this paper. This is a preliminary discussion. It identifies a structure, surveys what existing theory and existing practice have made of it, marks where each runs out, and proposes a direction. It does not complete a theory, and the questions set down at the close are held to be the more valuable half of what is offered. The survey is wide because the structure recurs at scales usually studied apart, and width is a liability the paper attempts to discipline by the device stated in §1: every section discharges one of four named questions, and material earns its place by discharging one. Where the survey is thin, where a jurisdiction has been read wrongly, or where an instrument answers the difficulty the paper claims it leaves open, correction is welcome at huangwanhong@serendip.ngo.

§1 The Problem of Organizational Rule-Interpretation

An organization that governs by rules must, at the moment of each decision, determine what its rules say about the case before it. The rule is general and the case is particular, and nothing in the rule specifies its own application. Someone must fix whether this case falls under that rule. That fixing is an act, performed by an organ of the organization, and it is the act with which this paper is concerned.

The act is easy to overlook because it is usually described in a vocabulary that conceals it. An organization is said to apply its rules, and application suggests a technical operation in which the content of the rule is antecedently settled and the only work remaining is to see whether the facts match. On that description the interpretive act disappears into the rule it applies, and the questions that can be raised concern the rule and the facts alone: whether the rule is a good one, whether the facts were found correctly, whether the procedure by which they were found was fair. Each of these is a real question and each has a literature. None of them reaches the act by which the rule’s content was fixed for this case.

Consider the structure at its plainest. A professional body admits members on the criterion that an applicant hold experience relevant to the field in which admission is sought. An applicant holds substantial experience acquired in an adjacent field. Whether that experience is relevant to this one is not settled by the rule, since the rule contains the word whose extension is at issue. It is settled by an organ of the body, in the course of deciding. The applicant may contest the finding of fact, and there is no fact in dispute. The applicant may contest the fairness of the procedure, and the procedure may have been impeccable. What the applicant would contest is the interpretation, and it is precisely here that the available channels thin out.

They thin out for reasons that are structural. Organizational rules are not law. A university’s admission regulation, a professional body’s code, a firm’s promotion criteria, a standard-setter’s conformity requirement: these bind by the organization’s own authority, and their interpretation is committed in the first instance to the organization itself. Where the organization is a public body, an external review may lie, though the review will ordinarily reach the procedure and the legality of the exercise and will decline the substantive evaluative judgment as belonging to the body’s own competence. Where the organization is private, even that external review may be absent. And no legal system compels an organization to establish an internal mechanism by which its interpretations may be contested. The law can require that reasons be given; it can rarely require that the reasons be right.

Four questions arise, and they organize everything that follows.

1. Authority. Who holds standing to interpret an organization’s rules, and how is that standing acquired?
2. Mode. By what modes does interpretation proceed, and what distinguishes a bounded interpretation from an arbitrary one?
3. Object. Which rules, and which of their terms, carry the interpretive load, and why is that load unevenly distributed?
4. Protection. When an interpretation works an injustice upon a party, what protects that party?

The first three have received sustained treatment, chiefly in jurisprudence, and the treatment transfers to organizations with qualifications this paper enters as it goes. The fourth is where the difficulty concentrates, and it is the question toward which the survey is directed.

The account offered in response may be stated as a single claim.

Central claim. The interpretation of a rule is a distinct site at which injustice can occur, separable from the justice of what an organization distributes and from the justice of the procedures by which it decides. The devices developed to protect a party against interpretive injustice, across jurisprudence, legislation, standard-setting, administrative review, and organizational governance alike, share a common presupposition: that the interpreting body holds a stake in recognition by some party outside itself. Each device operates upon that stake, and each loses its purchase as the stake approaches zero. Where authority over interpretation and independence from recognition coincide in one party, the available instruments do not reach. A generative relational reading proposes that an interpretation be assessed by its effect upon the capacity of the relation between the parties to generate a further interpretation, and that the injustice consists in the extinction of that capacity. On this reading the sanction is not administered by an external authority and appears in the dynamics of the structure itself, which is a description of a mechanism and carries no assurance that the mechanism arrives in time, or at all, or in a direction anyone would welcome.

The claim is developed in five movements. The first sets out the structure of the interpretive act and argues that it is constitutive of a rule’s content, which is what makes it a site of justice at all. The second reviews what jurisprudence has established about indeterminacy, about the doctrines of interpretation, and about the justice of interpretation, and positions the concept against adjacent concepts of injustice with which it is easily confused. The third reviews the theoretical strategies by which interpretation has been proposed to be constrained, and the difficulties each encounters. The fourth reviews institutional practice at four scales and across several legal traditions, and extracts the failure modes that recur across all of them, closing on the asymmetry boundary. The fifth offers the generative relational reading, states what it does not deliver, and sets down the open questions.

A word on the discipline of the survey. The material assembled here is drawn from fields ordinarily studied apart, and a survey of this width invites the reproach that it is a catalogue. The device by which the paper attempts to earn the width is stated now and enforced throughout: each section names, at its opening, which of the four questions it discharges, and material is admitted to a section on the ground that it bears on that question. §3 and those following on the modes and objects of interpretation discharge authority, mode, and object. Sections on jurisprudence and on institutional practice discharge all four, with protection dominant. The generative relational sections discharge protection alone, since it is the question the earlier answers leave open. Material that bears on none is left out, however interesting, and the omissions are recorded where they are likely to be missed.

Two disclaimers close this section. The paper takes no position on whether any particular organization has interpreted any particular rule wrongly, and the illustrative cases carry no such charge. And the concept advanced here has neighbours, in hermeneutical injustice, in procedural injustice, and in theories of domination, from which it must be distinguished with care; a section is given to that task, because a concept that cannot be told from its neighbours has not been introduced.

§2 Scope and Method

This section fixes the objects of the discussion, states the scales across which they are traced, and enters the limits of the comparative method employed.

2.1 Organization, Rule, Interpretive Act

By organization is meant any standing body that governs conduct by general rules of its own and decides particular cases under them: a legislature, a regulatory agency, a standard-setting body, a court, a university, a professional association, a firm, an international institution, a non-governmental association. The definition is broad by design, since the structure at issue is generated by the conjunction of general rules with particular decisions, and that conjunction is indifferent to the legal form of the body in which it occurs. Where the legal form matters, and it often does, it matters by determining which external devices are available; that is a finding of the survey and not a premise of it.

By rule is meant any general standard the organization holds itself out as deciding by, whether it is styled a statute, a regulation, a criterion, a code, a policy, a standard, or a guideline. Rules that are published and rules that are internal are both included, with the difference between them treated as a variable affecting contestability.

By interpretive act is meant the determination, in the course of deciding a case, of what a rule requires with respect to that case. The act is present whenever a rule’s terms admit of more than one extension in the circumstances, which is to say almost always, and it is most visible where the term is evaluative. The act is distinguished from three neighbours with which it is commonly merged: from fact-finding, which determines what occurred; from rule-making, which settles the general standard in advance of any case; and from discretion in the narrow sense, in which a rule expressly confers a choice among permitted outcomes. Interpretation shares with rule-making that it fixes content, and shares with discretion that it is an exercise of judgment; it differs from the first in occurring at the moment of decision and under the guise of application, and from the second in that the organ performing it ordinarily represents itself as bound.

2.2 The Scales of Comparison

The survey traces the structure across four scales, at which the same conjunction of general rule and particular decision recurs under different institutional conditions.

Legislation is the scale at which a rule’s interpretation is anticipated by the drafter, through definition clauses, purpose clauses, interpretation statutes, and the regulation of delegated power. Standard-setting is the scale at which private and transnational bodies write rules that bind through conformity assessment and market access, and at which the interpretive question is compounded by the question of who wrote the standard. State administration is the scale at which the interpretive act of an official is subject to a developed apparatus of reason-giving and review. Organizational governance is the scale at which internal mechanisms alone are available, and it is the scale at which the difficulty of the fourth question is sharpest. A fifth setting, that of international and transnational bodies deciding without a sovereign above them, is treated where it illuminates the fourth question, since it is the clearest case of interpretive authority disciplined by recognition alone.

The claim that these four are instances of one structure is a claim of the paper and is defended by exhibiting the same failure modes at each. A reader who holds the scales to be disanalogous will find the argument at its weakest in the transition from public to private, where the availability of external review differs in kind and not only in degree, and that transition is marked where it occurs.

2.3 Comparative Method and Its Limits

Several legal traditions are drawn upon: the common law systems of the United States and the United Kingdom, the continental systems of Germany, France, and Switzerland, the East Asian receptions of continental administrative doctrine in Japan and Korea, the supranational order of the European Union, and the transnational order of private standard-setting. Islamic legal theory is entered for its architecture of interpretive authority, which is the most fully developed answer to the first question available in any tradition.

Three limits attach. The comparison is doctrinal and institutional, and it does not rest on empirical study of how the devices surveyed perform in operation; where empirical findings exist they are cited, and their scarcity is itself a result. The selection of jurisdictions is governed by the availability of scholarship in the languages the author reads, which biases toward English, Japanese, and translated continental sources. And instruments of the kind surveyed here are revised on cycles, and doctrines are overruled; the state of the law and of the standards is stated as of the date of writing, and readers should verify currency, a caution that the recent displacement of a long-settled deference doctrine in the United States makes more than formal.

2.4 Exclusions from the Undertaking

The paper does not offer a decision procedure by which a given interpretation may be classified as just or unjust. It does not offer a model of the dynamics it invokes in its closing sections, and the reasons for that abstention are stated there. It does not adjudicate among the doctrines of legal interpretation surveyed, and takes them as candidate answers whose difficulties are the object of interest. And it does not propose institutional reform, since the finding it arrives at is that the available reforms operate upon a stake that is absent in the cases of greatest concern; proposing more of them would evade the finding.

§3 The Interpretive Moment

This section discharges the first part of the second question, by establishing what the interpretive act is, and prepares the fourth, by establishing why it is a site at which justice is at stake. The argument has three steps: that interpretation is constitutive of a rule’s content for the case, that the position from which it is performed is structurally asymmetric, and that the authority to interpret extends to the rules governing challenge to its own exercise.

3.1 The Minimal Structure

Let an organization hold a rule $R$, and let a case $c$ arise. An organ $A$ of the organization determines whether $R$ covers $c$. The determination is expressible as a proposition, that $R$ covers $c$ or that it does not, and the organization proceeds upon it. Call this the interpretive moment.

Three features of the moment are worth marking before anything is claimed about it. It is occasioned: it occurs because a case has arisen, and its occasion is supplied by a party who ordinarily has no control over when interpretation will be required. It is particular: it settles $R$ with respect to $c$, and whether it settles $R$ with respect to any further case depends on whether the organization treats its own determinations as binding upon itself, which is a variable and not a constant. And it is represented as bound: the organ does not ordinarily announce that it has chosen a meaning, and announces that it has found one. This last feature is the source of the moment’s invisibility and will bear weight later.

3.2 Interpretation as Constitutive

The claim that carries this paper is that the interpretive moment does not report a content the rule already had. It constitutes that content for the case.

The claim can be argued from the side of language and from the side of relation, and both routes are needed. From the side of language, the argument is familiar and will be surveyed at length in the section on indeterminacy: a general term does not carry its own extension to the borderline case, and no further rule can supply the extension without itself requiring interpretation, on pain of regress. What is added here is the second route.

From the side of relation, the argument runs as follows. A concept holds no boundary that is self-sufficient. What a term covers is fixed by its position in a network of other terms, of prior usages, of purposes the network is taken to serve, and of practices in which the term has been deployed. The word “relevant” has no extension considered alone; it acquires one only against a field of disciplines, a purpose the requirement of relevance serves, a history of cases in which relevance has been found or denied, and a practice of research within which relevance is a live notion. This is a statement about concepts generally, and its consequence for rules is direct. To determine that $R$ covers $c$ is to locate $c$ within the network against which $R$’s terms take their extension, and locating is an operation performed by whoever performs it, from the position they occupy in that same network.

Two things follow, and the second is as important as the first.

The first is that interpretation is an exercise of power. To fix the extension of a term for a case is to determine the normative position of the party whose case it is, and to do so by an act that the rule did not perform. The organ is doing something to the party, in the ordinary sense in which an exercise of authority does something, and the doing is not licensed by the rule alone, since the rule did not reach the case without the act.

The second is that this exercise of power is bounded, and the boundedness is a consequence of the same relational thesis that established the power. If the extension of a term were self-sufficient, interpretation would be mere reading and would carry no power. If the extension were unconstrained by anything, interpretation would be arbitrary and the whole apparatus of rules would be theatre. The relational thesis yields neither. Because a term takes its extension from a network, any proposed extension must be compatible with the network, and a proposal that is not can be shown to be so. A determination that a doctorate in music is relevant to a chair in surgery is not merely unwelcome; it is unavailable, because it cannot be held together with the purposes, usages, and practices from which the term draws its content. The space of admissible interpretations is therefore neither a point nor the whole, and this is the space within which the second question, on the modes of interpretation, will be posed.

It matters that this second consequence be stated at the outset, because a relational account of meaning is commonly heard as licensing arbitrariness, and if it did so it would destroy the concept this paper is attempting to introduce. A theory on which any interpretation is as good as any other cannot say that an interpretation was unjust, and it cannot say so of its own claims either. The account advanced here holds that relational constitution is the source of constraint upon interpretation and is what makes an interpretation criticizable at all.

3.3 The Asymmetry of the Interpretive Position

The interpretive moment is occupied asymmetrically, and the asymmetry has four components that should be distinguished, since remedies exist for some and not for others.

The first is positional. One party determines and the other receives the determination. This asymmetry is constitutive of rule-governed decision and is not in itself objectionable; an organization that could not determine the application of its own rules could not decide at all.

The second is informational. The organ ordinarily knows the history of prior determinations, the purposes the rule is taken to serve, and the range of readings under consideration. The party ordinarily knows the text. Where prior determinations are unpublished, the party cannot know whether the reading applied to them is the reading applied to others.

The third is representational. The determination is announced as a finding. A party who wishes to contest it must therefore contest what is presented as a discovery about the rule, using materials the organization controls, and must do so without a vocabulary in which the objection is naturally expressed. Ordinary language supplies ready terms for objecting to a rule and for objecting to a procedure, and supplies thin terms for objecting to an interpretation.

The fourth is recursive, and it is the component this paper regards as decisive. The rules that govern whether and how an interpretation may be challenged are themselves rules of the organization, and their interpretation belongs, in the first instance, to the same organization. Whether a given complaint is within the scope of an internal appeal, whether it raises a matter of substance excluded from review, whether it was brought in time, whether the complainant has standing: each is an interpretive question, and each is answered by an organ whose answer determines whether the original interpretation will be examined at all. The authority to interpret thus extends to the conditions of its own review. Where the extension is complete, the interpretive power is closed upon itself, and a party contesting an interpretation is in the position of appealing to the interpretation of the rules of appeal.

3.4 Interpretive Injustice as a Distinct Site

With the structure of the moment set out, the reason it deserves a name of its own can be stated. The argument proceeds by elimination against the two categories into which the phenomenon is otherwise absorbed.

It is not distributive. What an organization distributes may be distributed justly and the interpretation by which it did so may be objectionable; and an interpretation may be unobjectionable while the distribution it produces is indefensible. The two vary independently, which is the ordinary test for whether one concept has been reduced to another.

It is not procedural. A procedure may be complete, timely, impartial in composition, and scrupulous in the hearing it affords, and the interpretation reached at the end of it may still be the object of complaint. Procedural guarantees regulate the path to the determination and are, with few exceptions, silent about the determination. This is not an incidental gap in existing procedural doctrine; it follows from what procedure is, and the survey will show it recurring in every jurisdiction examined.

There remains the possibility that interpretive injustice is a species of domination, or of the injustice of unaccountable power in general. That relation is closer, and it is taken up where the adjacent concepts are treated. The provisional statement is that domination names the standing liability to arbitrary interference and does not by itself distinguish interference through interpretation from interference by other means, whereas the interest of the present concept lies precisely in the mechanism, since interpretation is the mode of interference that presents itself as the absence of interference.

With the structure of the moment established, the question that governs the remainder of the paper can be put in its sharpest form. Given that a rule’s content is constituted at the moment of decision, by an organ occupying an asymmetric position, with authority extending to the conditions of its own review, what protects the party against whom the constitution is performed? The sections that follow assemble what has been offered in answer, and the answer they converge upon is narrower than the apparatus assembled would suggest.

---

A note on the state of the draft. The argumentative spine and the theoretical survey are drafted and the survey of institutional practice is in progress, so some sections appear here ahead of others that will precede them in the final order, and the numbering anticipates that order throughout. Sections on interpretive control in legislation, in standard-setting, in administrative review, in organizational governance, and in international bodies, together with the evidentiary devices against pretextual interpretation and the recurrent failure modes, are in part pending. The section on the asymmetry boundary is drafted ahead of them because it states the result those sections are directed toward, and drafting it first fixes the destination against which the survey is written.

§4 The Authority of Interpretation

This section discharges the first question. It sets out the sources from which standing to interpret is drawn, states the relation between the account given here and the position taken in the companion work on value conversion, distinguishes formal from effective authority, and marks the neglected question of who holds standing to contest.

4.1 Sources of Standing

Standing to interpret an organization’s rules is drawn from three sources, and most actual arrangements combine them.

Conferral is the ordinary source. A constitutive document, a statute, a delegation, or a resolution names an organ and commits to it the determination of matters under the rules. The organ interprets because it was appointed to interpret, and its determinations bind because of the appointment.

Practice is the second. An organ that has determined matters of a kind, whose determinations have been acted upon and acquiesced in, acquires standing through the acting and the acquiescing. Much interpretive authority in organizations rests here, since the arrangements by which a body’s rules are administered ordinarily develop faster than the documents that describe them.

Performance is the third and is the source the companion work on value conversion treats as primary. The party who performs an act of fixing acquires, by performing it, the vocabulary in which the matter is subsequently discussed. Whoever writes an equivalence supplies the terms of the exchange; whoever fixes an extension supplies the terms in which the rule is thereafter understood, since the determination enters the network as a usage and the next case is decided against a field that now contains it.

4.2 Relation to the Position on Conversion

The companion paper holds that the question of who owns the right of interpretation is the wrong question, on the ground that legislative standing is generated by the act of crossing categories and is not held in advance (Huang, in preparation). The present subject requires a qualification of that position, and the qualification is worth stating because it locates something the conversion case does not display.

In an organization the ordinary case supplies a designated organ, so standing is antecedent in a way that the conversion case does not exhibit. An admitting body’s assessment panel holds its authority before any application arrives, and holds it whether or not it exercises it well. To that extent the question of who owns the interpretive right has an answer, and the answer is given by the conferral.

The qualification is that conferral fixes the holder and does not fix the reach. What matters may be determined by the organ, how far its determinations extend, whether a question belongs to it or to another body, and what it may take into account, are settled by rules, and those rules require interpretation in the same way as the rules whose application is at issue. At the centre of a conferral the answer is plain and no interpretive act is needed to see it. At the edge the answer is supplied by an interpretive act, and the organ performing that act is ordinarily the organ whose reach is in question.

Authority therefore has an open texture of its own, in the sense given by Hart (1994) for rules generally. The conversion paper’s thesis holds at the edge, where standing is generated by the act of claiming it, and the conferral thesis holds at the core. The two positions are compatible once the distinction between core and edge is drawn, and the fact that they must be reconciled at all is a consequence of the recursion entered in §3, since it is the same feature under a different description: a power that determines the scope of the rules governing its own exercise is a power whose reach is partly its own to fix.

4.3 Formal and Effective Authority

An organization’s rules name the organ that decides, and the interpretive act is ordinarily distributed across more positions than the named one.

Consider the sequence by which a case reaches a decision. An officer receives it and determines whether it is complete, which requires a reading of the rule specifying completeness. A second officer screens it against threshold criteria, and a case that fails the screen never reaches the body that formally decides. A drafter of internal guidance has previously fixed how a term is to be understood, and the guidance shapes what the screener does. The formally deciding body receives a case already characterized, and its determination is made upon materials that earlier interpretive acts have assembled.

The distinction that follows is between formal interpretive authority, held by the organ the rules name, and effective interpretive authority, held by whoever fixes the readings on which the outcome actually turns. The two coincide in some arrangements and diverge in many. Where they diverge, three consequences follow that will matter in the sections on practice.

The first is that devices addressed to the formal locus do not reach the effective one. A duty to give reasons binding on a deciding committee says nothing about the screening officer, and a right of appeal against the committee’s decision is of no use to a party whose case was disposed of before the committee saw it.

The second is that the effective locus is ordinarily the less visible. Guidance is internal, screening is unrecorded, and characterization of a case is not experienced by anyone as an interpretive act. The party is informed of the outcome and of the organ that reached it.

The third is that the dispersal makes responsibility difficult to locate even inside the organization. No officer in the sequence performed the whole of the interpretation, each performed a portion that was reasonable in itself, and the determination that resulted is attributable to the arrangement more readily than to any of them.

4.4 Standing to Contest

A question that ordinarily receives less attention than the authority to interpret is the authority to raise an interpretive question. It is treated here because §21 will turn upon it.

Standing to contest is ordinarily narrower than the class of parties affected. A party to the case may complain. A party in a like position whose case has not yet arisen may not. A member of the class the rule governs, who can foresee that the reading will bear upon them, ordinarily may not. A party affected by the reading through its operation on somebody else may not. And a party who has departed, whose relation with the organization ended with the determination, ordinarily may not, though the reading survives them and governs those who follow.

The narrowness has a structural consequence. Where the class of contestants is confined to parties presently before the organization, and where those parties depart upon determination, a reading may govern indefinitely with nobody positioned to raise it. This is the condition that §21 will identify as structural extinction, and it is produced by the ordinary drafting of standing rules and requires no act of foreclosure.

4.5 Findings and Residue

This section establishes that standing is antecedent at the core of a conferral and generated at its edge, that effective authority disperses across positions the rules do not name, and that standing to contest is narrower than the class affected. It leaves open whether interpretive standing may be delegated outside the organization without relocating the recursion, and what grounds an exercise of interpretive authority as legitimate. Both are entered in §24.

§5 The Modes of Interpretation

This section discharges the second question. It sets out the modes by which interpretation proceeds, identifies what each requires in order to be available, states the difficulty that arises from their plurality, and proposes the distinction between a bounded reading and an arbitrary one that §3 promised.

5.1 The Modes and Their Materials

Five modes recur across legal and organizational practice, and they are presented here in a way that departs slightly from the usual taxonomies, by identifying for each the region of the network from which it draws.

Textual interpretation draws on the settled usage of the terms, taking the extension a competent speaker would give them. Purposive interpretation draws on the ends the rule is taken to serve, asking which reading advances them. Systemic interpretation draws on the remaining corpus of the organization’s rules, asking which reading coheres with what else is provided. Precedential interpretation draws on the record of prior determinations, asking how the term has been taken before. Analogical interpretation draws on structural similarity between the case at hand and cases already decided, asking whether the features that governed there obtain here.

Presented this way, the modes stop being a list of competing schools and become a set of routes to the same object. §3 held that a term’s extension is fixed within a network of usages, purposes, coherence relations, and prior applications. Each mode mobilizes one region of that network. Their plurality follows from the composition of the network and is not a failure of theoretical tidiness.

5.2 Conditions of Availability

A mode is available only where the materials it draws upon exist and are accessible to whoever would use it.

Purposive interpretation requires a discoverable purpose. Where a rule was adopted without recorded deliberation, where its purpose is contested, or where it has outlived the circumstances that produced it, the mode supplies nothing, or supplies whatever the interpreter takes the purpose to have been, which is the difficulty the section on theoretical strategies will press.

Precedential interpretation requires a record. Where determinations are unpublished, unindexed, or not preserved, no party can invoke them, and the organ itself may be unable to consult them. Individualized decisions of the kind organizations make in large numbers are commonly in this condition.

Systemic interpretation requires a corpus of sufficient extent and consistency to yield coherence relations. A body governed by a handful of provisions adopted at different times for different reasons supplies little.

Analogical interpretation requires both a record and an articulated account of which features of prior cases were doing the work.

The consequence for organizations is considerable and is stated here because the survey of practice will return to it. Legal systems have invested heavily in the materials these modes require: reports, records, drafting histories, published reasons, and codified corpora. Organizations ordinarily have none of this. What remains available to an organization lacking record, recorded purpose, and corpus is textual interpretation and unstated judgment, which are the two modes that draw on the least and are the two hardest for an outside party to check.

5.3 The Difficulty of Plurality

Where several modes are available, a further question arises that the modes cannot answer among themselves: which mode governs where they diverge.

Llewellyn (1950) supplies the demonstration in its most economical form for the canons of construction, which come in opposing pairs. The demonstration generalizes to modes. A reading supported by settled usage may be opposed by one supported by purpose; a reading supported by prior determinations may be opposed by one supported by coherence with rules adopted since. An organ seeking a given outcome will ordinarily find a mode that licenses it, and will be able to state, in good faith, the mode it used.

The natural response is a second-order rule ordering the modes, and the response encounters the regress of §8: a rule ordering the modes is a rule, its application to a case requires interpretation, and the modes are what interpretation proceeds by. The regress is not vicious in practice, since practices do settle orderings and organs do follow them, and it is unarrestable in theory, which is the same result reached in §8 by a different route.

5.4 The Bounded and the Arbitrary

§3 claimed that the relational constitution of meaning constrains interpretation and does not license arbitrariness, and deferred the distinction to this section. The distinction proposed is deliberately weak, and its weakness is the point.

Boundedness. A reading is bounded when it can be exhibited as drawn from an identified region of the network that fixes the term’s extension, in a form that a party competent in the practice can follow and test. A reading is arbitrary when no such exhibition is available, whether because none was offered, because what was offered does not support the reading, or because the reading draws on nothing the network contains.

Three features of this proposal should be marked.

It does not select among admissible readings. Two readings may each be exhibited from different regions of the network, and the criterion pronounces both bounded and says nothing about which is better. This is Kelsen’s frame restated in operational terms, and the limitation is inherited knowingly.

It is a criterion of exhibition and therefore depends upon reasons being given. A reading is bounded when the exhibition is available, and the exhibition is available when the organ states what it drew upon and how. Where reasons are absent, or are given in a form that reports the conclusion without the route, the distinction cannot be drawn from outside, and a bounded reading is externally indistinguishable from an arbitrary one. The duty to give reasons, surveyed in §19 among the devices, is doing analytic work here and is not merely a procedural courtesy: it is the condition under which the boundedness of an interpretation becomes assessable at all.

It leaves the prospective question open. Whether a reading can be exhibited is answerable once the exhibition is attempted, and whether a proposed reading lies outside the network is answerable in advance only by conjecture. This is entered as Q3 in §24.

§6 The Objects of Interpretation

This section discharges the third question. It classifies the terms of rules by the interpretive load they carry, identifies a class of terms whose determinations are unreviewable from the position of any single party, and explains why interpretive load concentrates at the points where organizations allocate scarce goods.

6.1 A Classification by Load

Terms may be arranged by how much interpretive work their application requires.

Determinate terms specify dates, quantities, named entities, and enumerated categories. Their application is mechanical in the ordinary case and their interpretive load is near zero. Rules composed of such terms can be administered without judgment, which is why organizations move terms into this class when they wish to remove discretion.

Classificatory terms specify kinds: employee, resident, document, dependant, publication. Their core is settled and their borderline is genuinely open, which is the condition Hart (1994) described. Their load is moderate and is concentrated in a minority of cases.

Standard terms specify a level to be met: reasonable, adequate, sufficient, satisfactory, material, relevant. Their application requires an evaluation on every occasion of use, and their load is high in every case and not merely at a borderline. Endicott (2000) treats the vagueness of such terms as ineliminable and as serving purposes that precision would defeat, which is why they are not simply drafting failures to be corrected.

Comparative terms specify a relation to other cases: best, strongest, most suitable, most deserving, of highest merit. Their load is the highest and their structure differs from the others in a way the next subsection takes up.

6.2 Comparative Terms

A comparative term differs from a standard term in what its application requires. To determine that a case meets a standard requires the case and the standard. To determine that a case is the strongest requires the case, the standard, and every other case in the comparison class.

Three consequences follow, and together they explain a good deal of what the later sections find.

The first is that a determination under a comparative term cannot be checked by a party who has access to their own case alone. The materials that would show the determination correct or incorrect are the other cases, and those are ordinarily confidential, are ordinarily numerous, and are ordinarily unavailable to anyone outside the deciding body. A party told that others were stronger is in possession of a proposition they have no means of examining.

The second is that a comparative determination is unreviewable in a stronger sense than a standard determination. A reviewing forum can assess whether a body applied an intelligible standard to a case. Assessing whether a body correctly ranked a case against a field would require the forum to redo the exercise, which every forum surveyed in the sections on practice declines to do, and declines for reasons that are sound.

The third follows from the first two. A comparative ground, once available, absorbs any challenge to a determination made on other grounds. If a determination under a standard term is contested, and a comparative ground is available on the same facts, the body may rest upon the comparative ground and the challenge ends, since the comparative ground is not examinable. This is the analytic basis of the substitutability of reasons identified in §19 as one of the conditions under which the protective devices lose their purchase, and it explains why that condition is structural and not a matter of bad faith. A body that shifts from a relevance ground to a merit ground has moved from a proposition a party might contest to one that no party can, and it need not have intended anything by the move.

6.3 The Concentration of Load

Interpretive load is unevenly distributed across an organization’s rules, and the distribution is not random.

Rules governing internal procedure, entitlements attaching to status, and matters where predictability is valued by the organization itself tend to be composed of determinate and classificatory terms. Rules governing the allocation of goods that are scarce tend to be composed of standard and comparative terms, since the organization requires the latitude that such terms preserve, and since specifying in advance what will distinguish the successful case is ordinarily impossible.

The consequence is that interpretive load is highest exactly where the stakes for the affected party are highest, where the number of affected parties is greatest, and where those parties are least likely to hold anything the organization requires. Admission, appointment, promotion, funding, certification, and the award of discretionary benefits share this profile. The conditions identified in §19 under which the protective devices fail are therefore concentrated in the same region as the interpretive acts that most need protecting, and the coincidence is produced by the structure of the allocation problem.

6.4 Published Criteria

One partial remedy belongs in this section, since the sections on practice will meet it repeatedly in institutional form.

An organization may publish criteria specifying how its standard and comparative terms are to be understood. Publication converts what would have been an interpretive act into a rule of lower generality, and it supplies materials for the precedential and systemic modes that §5 found ordinarily absent. It renders determinations checkable against something, and it exposes departures.

Two limits attend it. The criteria are themselves composed of terms, and their application requires interpretation in turn, so publication relocates the interpretive question and does not dissolve it. And publication is costly to the organization in precisely the respect for which the latitude was valuable, so bodies that allocate scarce goods have a standing reason to state their criteria at a level of generality that preserves what the standard and comparative terms were preserving. The devices surveyed in the sections on practice include several that require publication, and the finding there will be that the requirement is satisfiable without the latitude being surrendered.

6.5 Findings and Residue

This section establishes a classification by interpretive load, isolates comparative terms as a class whose determinations resist examination by any single party, and locates the concentration of load in the allocation of scarce goods. It leaves open whether evaluative predicates differ in kind from other open-textured terms or in the proportion of cases falling in the penumbra, which is entered as Q4 in §24.

§7 The Epistemology of Concept Application

Two claims made earlier in this paper rest on assumptions about how a general term acquires its extension. §3 held that an interpretive act constitutes a rule’s content for a case, on the ground that a concept holds no self-sufficient boundary. §5 held that the same relational constitution bounds interpretation and excludes arbitrariness. Both are claims in the theory of concepts, and neither was argued there. This section supplies the ground.

There is a second reason to enter the material. The devices surveyed later in this paper embed epistemological positions without stating them. A definition clause presupposes that a concept can be given by conditions. A published criterion presupposes that classification can be specified in advance of the cases. A requirement that an examiner accept any defensible answer presupposes that a term admits of several correct applications at once. An institution choosing among these devices is choosing among theories of concepts, and it is better that the choice be visible.

7.1 The Definitional View

The oldest account holds that a concept is given by conditions individually necessary and jointly sufficient for its application, and that classifying an instance consists in determining whether the conditions obtain. Frege (1997) gives the position its sharpest modern form in requiring that a properly formed concept have determinate boundaries, so that for every object it is settled whether the concept applies.

The view is the implicit epistemology of a great deal of legal drafting. A definition clause states conditions; a schedule enumerates instances; a numerical threshold replaces an evaluative term. §6 treated these as devices for reducing interpretive load, and their rationale is definitional: a concept so specified can be applied by determining whether the stated conditions obtain.

The difficulty is that few concepts of interest admit of such specification. Attempts to state necessary and sufficient conditions for ordinary terms have generally failed, and the terms that carry the interpretive load in organizational rules, treated in §6 as standard and comparative, are conspicuously resistant. The definitional view therefore describes a condition that drafting can sometimes create and cannot generally presuppose.

7.2 Family Resemblance and Blurred Edges

Wittgenstein (1953) advances the counter-position. The instances falling under a general term may be related by a network of overlapping and criss-crossing similarities with no feature common to all, and the term is nonetheless in good order. To the objection that such a concept has blurred edges, the reply is that a blurred concept is still a concept, and that for many purposes a blurred one is exactly what is wanted.

The consequence for this paper is worth stating carefully, since it cuts against a natural assumption. The vagueness of an evaluative term in a rule is ordinarily treated as a drafting failure to be corrected where resources permit. On the Wittgensteinian view it may be the appropriate form for the concept, and the demand for sharp boundaries is a philosophical commitment and not a neutral standard of good drafting. Endicott (2000) develops this for law directly, holding that vagueness is ineliminable and serves purposes that precision would defeat.

7.3 Prototypes, Exemplars, and Graded Membership

Empirical work on human categorization supplies a third account and bears on the question of how a concept is in fact recognized.

Rosch (1978) established that category membership is graded. Subjects rank some instances as better examples of a category than others, do so consistently, and classify typical instances faster than atypical ones. Categories are organized around prototypes, and membership shades off from them. Medin & Schaffer (1978) advance a variant on which classification proceeds by similarity to remembered instances, with no abstracted prototype required, so that the categories a person applies are a function of the instances they have encountered.

Two consequences bear on the interpretive moment.

The penumbra is not exceptional. Hart (1994) presented a settled core with an open border, and the empirical picture is of graded membership throughout, with clear cases at one end of a continuum and no boundary of the kind the core-and-penumbra image suggests. §6 reached the same conclusion for evaluative terms from the side of rule-drafting, and the two lines converge.

Divergence among competent interpreters is predicted. On an exemplar account, what a person classifies as relevant experience depends on the range of instances they have encountered under that description. Two interpreters of equal competence and good faith, having encountered different ranges, will draw the boundary differently, and neither will be able to detect the source of the difference by inspecting the rule. This is a mechanism for interpretive divergence that requires no bad faith, no negligence, and no disagreement about the facts, and it is worth having because the alternative explanations of divergence all impute a fault to somebody.

7.4 Background Theory in Classification

Murphy & Medin (1985) argue that similarity alone cannot account for the coherence of categories, and that concepts are embedded in background theories about what goes with what and why. Classification is accordingly shaped by an explanatory framework the classifier brings.

The consequence for this paper is a second and independent mechanism of divergence. Two interpreters who have encountered the same instances may classify differently because they hold different accounts of what the category is for. An interpreter who takes a relevance requirement to serve the protection of standards will draw its boundary differently from one who takes it to serve the identification of capacity, and the difference will present itself to each as a straightforward reading. This connects the epistemology to the doctrinal material of §9, where the attribution of purpose is precisely the point at which such background accounts enter and are seldom made explicit.

7.5 The Division of Linguistic Labour

Putnam (1975) argues that the extension of a natural kind term is fixed by the world and by the practice of a linguistic community, and not by any state of an individual speaker. Ordinary speakers use terms whose extension they cannot themselves determine, deferring to those in the community who can, and Putnam names this the division of linguistic labour. Kripke (1980) supplies the companion account of reference fixed by an initial baptism and transmitted through a causal chain of use.

Two consequences follow, and they pull in opposite directions.

The first supports an arrangement described in §10. If extension is fixed by those competent to determine it, then an interpretive authority grounded in demonstrated competence has an epistemological warrant and is not merely a professional privilege. The tradition surveyed there, which locates standing to interpret in an assessable qualification, is answering a real feature of how the extension of terms is settled.

The second is a warning that runs throughout this paper. If extension is fixed by those who count as competent, then the determination of who counts is a determination of what the terms mean. Whoever is recognized as the expert fixes the extension for everybody else, and a party outside the circle of recognized competence has no standing to contest an extension even in principle. The division of linguistic labour is an epistemological account of how meaning is settled, and it is at the same time a description of an authority relation.

7.6 Vagueness and the Structure of the Penumbra

The philosophical treatment of vagueness bears directly on the dispute recorded in §8 and §9, and the correspondence is close enough to be set out.

Epistemicism holds that a vague term has a sharp boundary which nobody is in a position to know, so that borderline statements have determinate truth values that remain inaccessible (Williamson 1994). Supervaluationism holds that a vague term admits of many admissible precisifications, that what is true on all of them is true, that what is false on all of them is false, and that what varies among them lacks a truth value (Fine 1975). Degree theories assign intermediate values to borderline cases.

Set beside the legal positions, the alignment is striking. The epistemicist structure is the structure of the right-answer thesis: there is an answer, it is not accessible by any procedure the parties can operate, and the appearance of discretion reflects the limits of knowledge. The supervaluationist structure is the structure of Kelsen’s frame: some readings are excluded because they fail on every admissible precisification, some are required because they hold on all, and among the remainder nothing settles the matter.

Position on vagueness	Structure	Legal counterpart
Epistemicism	Sharp boundary, inaccessible to knowledge	The right-answer thesis
Supervaluationism	Admissible precisifications; truth is truth on all of them	The frame, with an unsettled interior
Degree theories	Membership admits of intermediate values	Graduated intensity of review

The correspondence is offered as an aid to seeing what is at stake and not as an argument that either legal position is entailed by a position on vagueness. It does establish that the disagreement recorded in §9 is a version of a disagreement in the philosophy of language, and that neither side is peculiar to law.

The supervaluationist reading has a further use, since it gives a precise sense to the boundedness proposed in §5. A reading is inadmissible when it fails on every precisification the network of usages, purposes, and prior applications makes available. A reading is required when it holds on all of them. Between these the criterion is silent, which is the limitation §5 entered knowingly.

7.7 Interpretation as Conceptual Engineering

A recent literature asks what our concepts ought to be, treating concepts as instruments open to deliberate revision for the purposes they serve (Haslanger 2000; Cappelen 2018). The enquiry proceeds by asking what work a concept does, whether it does that work well, and what revision would improve it.

The bearing on this paper is direct and does not appear to have been drawn. An organ determining that a rule’s term covers a case is doing what the conceptual engineer does deliberately: it settles what the concept shall cover going forward, in the light of the purposes the rule is taken to serve, and its determination enters the network as a usage available to govern later cases. The difference lies in the acknowledgement. The conceptual engineer announces that a revision is proposed and defends it as a revision. The interpreting organ announces a finding.

Two things follow. The description of the interpretive act given in §3 gains support from a direction independent of the arguments offered there. And the representational asymmetry entered in that section acquires a sharper form: what is presented to the affected party as the discovery of an existing content is, on this description, the exercise of a power to settle what a shared concept will mean, conducted without the vocabulary in which such an exercise is ordinarily defended.

7.8 Findings and Residue

The material assembled here supports the claims that §3 and §5 left unargued, and it does so from several independent directions. Concepts of the kind that carry interpretive load do not admit of definitional specification. Membership in them is graded, and the graded structure is a finding about human classification and not a defect of drafting. Extension is settled within a community and by those the community recognizes as competent. Divergence among interpreters of equal competence and good faith is predicted by two independent mechanisms, differing exemplar sets and differing background accounts of what the category is for. And the space of admissible readings is bounded by what the available precisifications permit, which is neither a point nor the whole.

Three consequences are carried forward. Institutional devices that presuppose the definitional view are working against the grain of how classification proceeds, which is why published criteria relocate the interpretive question and do not dissolve it. Divergence between an organ’s reading and a party’s reading is the expected case and supplies no evidence that either has erred. And the determination of who counts as competent to fix an extension is itself a determination of what rules mean, which returns the epistemology to the question of authority treated in §4.

What the material leaves open is whether the graded structure of classification can be made operational for the purposes of review. A reviewing body applying the criterion of §5 would need to establish what precisifications the network makes available, and no procedure for doing so is offered here or located elsewhere in the literature surveyed.

§8 Indeterminacy and the Limits of Rule-Following

This section returns to the second question from the side of the philosophical and jurisprudential literature on rule-following, and prepares the fourth by showing where each account leaves the protection of a party unaddressed. Its result is narrow and load-bearing: the tradition has established that rules underdetermine their own application, and has established that the underdetermination is partial. What it has not established is any procedure by which a party may show that a determination made within the space of the underdetermination was wrongly made.

8.1 The Regress of Rules

The deepest statement of the difficulty is Wittgenstein (1953)’s. A rule does not contain its own application, since any course of action can be brought into accord with the rule under some reading of it, and any reading can be brought into conflict with it under another. Supplying a second rule to govern the reading of the first does not arrest the difficulty, since the second rule requires reading in turn. The regress is not a defect of poorly drafted rules and is a feature of the relation between a general formulation and a particular case.

Kripke (1982) draws the sceptical consequence and then supplies what he calls a sceptical solution: what arrests the regress is not a fact about the rule and is a practice, sustained by a community whose agreement in judgment supplies the conditions under which a use counts as correct. The solution has been contested at length, and the feature of it that matters here survives the contest. Whatever arrests the regress is external to the rule and is constituted by a practice within which the rule is used.

This result is the language-side counterpart of the relational argument entered in §3. Both routes reach the same place. The extension of a rule’s terms is fixed by something outside the rule, and that something has the form of a network of usages, purposes, and practices. The two routes differ in what they emphasize. The Wittgensteinian route emphasizes that the arresting practice is communal and that agreement in judgment is its ground. The relational route emphasizes that the party to whom a rule is applied occupies a position within the same network from which the rule takes its content, and may occupy it without any capacity to affect what the network supplies. The second emphasis is where the question of justice enters, and it is the emphasis the tradition has left comparatively undeveloped.

8.2 The Frame and the Act of Will

The most exact statement of the difficulty in legal theory is Kelsen (1967)’s. A higher norm stands to its application as a frame stands to what may be placed within it. The frame excludes: some applications are outside it and are for that reason unlawful. Within the frame the norm admits of several applications, all equally lawful, and nothing in the norm selects among them. The selection is an act of will performed by the applying organ, and legal science, whose office is cognition, can state the boundaries of the frame and can say nothing about which application within it is correct.

Two consequences follow that this paper adopts. The first is that the space of admissible interpretations is neither a point nor the whole, which is the same conclusion reached from the relational side in §3 and is stated by Kelsen with greater precision than by any subsequent writer. The second is that the applying organ is, within the frame, a maker of norms, and represents itself as a finder of them. Kelsen states the representation and treats it as a matter for legal science to correct; the concern here is with what the representation does to a party who would object, which is the third component of asymmetry entered in §3.

Kelsen’s account also marks the limit of what a theory of norms can supply. If several applications are equally lawful, a party against whom one of them was selected has, on Kelsen’s own terms, no legal ground of complaint, since legality is exhausted by the frame. The complaint that survives is a complaint about the exercise of will within the frame, and it is a complaint for which Kelsen’s theory supplies no vocabulary. Naming that complaint is one of the tasks of this paper.

8.3 Core and Penumbra

Hart (1994) gives the distinction that has organized the anglophone discussion. General terms possess a core of settled meaning, without which they could not function as rules at all, and a penumbra in which their application is genuinely open. In the core, application proceeds without choice. In the penumbra, the organ chooses, and the choice is guided by considerations that the rule alone does not supply.

Hart’s distinction is often read as a concession to indeterminacy, and it is better read as a limitation upon it: most applications fall in the core, and the interesting cases are exceptional. That reading is sound for the class of rules Hart considered. It transfers poorly to the class of rules with which this paper is concerned. Where a rule’s operative term is evaluative, such as relevance, merit, adequacy, suitability, or reasonableness, the penumbra is not an exceptional zone at the margin of a settled core and is the ordinary condition of the term’s use. §12 does not depend on this observation, though the section on the objects of interpretation will, and it is entered here because it marks the point at which Hart’s reassurance ceases to reassure.

The exchange between Hart (1958) and Fuller (1958) bears on the same point from the side of remedy. Fuller’s position is that the penumbral choice is guided by the purpose the rule serves, and that purpose is not extraneous to the rule and is part of what the rule is. The position supplies a constraint upon interpretation and will be treated in its own right among the theoretical strategies, together with the difficulty that purpose is itself an object of interpretation.

8.4 Realism and the Indeterminacy Thesis

The American realists pressed the observation that the materials of legal reasoning underdetermine outcomes to a degree the doctrinal literature concealed. Llewellyn (1950) supplies the most economical demonstration: the canons of statutory construction come in opposing pairs, each canon met by a counter-canon, so that a court seeking a given reading will ordinarily find a canon licensing it. The demonstration is directed at the canons and generalizes to any body of interpretive guidance sufficiently rich to cover the cases, since richness is what supplies the opposing pair.

Critical legal studies advanced a stronger thesis, on which legal materials are radically indeterminate and outcomes are settled by considerations the materials conceal. The stronger thesis has been resisted with force, and Solum (1987) states the resistance in the form this paper adopts: the materials are moderately indeterminate, which is to say that they exclude a great deal and select nothing uniquely within what they permit. Moderate indeterminacy is Kelsen’s frame restated in the vocabulary of a different tradition, and the convergence of two traditions upon it is a reason to treat it as settled.

The realist contribution to the fourth question is chiefly negative and is nonetheless valuable. If interpretive guidance can be found on both sides of most questions, then a requirement that an organ give reasons for its interpretation secures less than it appears to secure, since a reason will be available for whichever reading the organ has reached. This observation will recur, in its institutional form, among the failure modes.

8.5 Legitimation Through Procedure

Luhmann (1969) supplies the observation that completes this section. Procedures produce acceptance of decisions, and they do so by a mechanism largely independent of the correctness of what is decided. Participation absorbs protest: a party who has been heard, whose submissions have been received, and who has exhausted the steps made available, is thereafter in a position where continued objection has no institutional form, and expectations are restructured around the outcome.

Luhmann’s finding is descriptive, and its significance for a normative inquiry is considerable. It establishes that procedural elaboration and interpretive correctness are separable, and that the first can be increased without the second. An organization may therefore hold impeccable procedures, produce high acceptance among those subject to them, and interpret its rules in a manner no party is positioned to contest, with the acceptance functioning as evidence that nothing is wrong. This is the mechanism by which the interpretive moment becomes invisible at the institutional scale, and it will be met again in §12 in the empirical procedural-justice literature, which measures the acceptance Luhmann describes and is sometimes read as measuring justice.

8.6 The Residue

The accounts surveyed in this section establish three things and leave one open. They establish that rules underdetermine their applications, that the underdetermination is bounded, and that what arrests it is a practice external to the rule. They leave open what a party may say, and to whom, when a determination within the bounded space has gone against them. Kelsen’s answer is that no legal complaint survives the exhaustion of the frame. Hart’s is that the choice was made on grounds the rule did not supply, which describes the choice and does not evaluate it. The realists’ answer is that a reason will always be available, which is a warning and no remedy. Luhmann’s is that the party will in due course accept it. The remainder of this paper is addressed to that residue.

§9 Doctrines of Statutory and Constitutional Interpretation

This section discharges the second question in its developed form. §5 identified five modes and left a difficulty: where the modes diverge, nothing among them settles which governs. The doctrines surveyed here are the developed answers to that difficulty. Each proposes an ordering of the modes, each defends the ordering by an account of what interpretation is for, and each meets an objection that the section records. The section closes by asking which of them an organization is in a position to use, and the answer bears directly on §19.

9.1 Doctrines as Orderings

The doctrines are read here as second-order proposals, and the reading is adopted for the reason given at the end of this subsection. None of them denies that the modes exist. Textualists consult purpose in extremity, purposivists read the text first, and every school attends to prior determinations. What divides them is the ordering: which mode is consulted first, which may override which, and what is required before a subordinate mode may be reached.

Reading them this way has two advantages for the present inquiry. It makes the disagreement tractable, since orderings can be compared. And it locates the disagreement where §5 left it, which keeps the survey inside the containment device stated in §1.

9.2 Textualism

Textualism orders the modes with settled usage first and treats the remaining modes as available only where usage leaves the matter open. Scalia (1997) and Scalia & Garner (2012) give the developed statement, together with a systematized account of the canons by which usage is determined.

The defence is threefold and each limb concerns constraint. The text is what was adopted, so a reading tied to the text is tied to what the adopting authority actually did. The text is what parties can read, so notice is served. And the text is the material least amenable to the interpreter’s preferences, so the ordering constrains the interpreter more tightly than orderings placing purpose or context first.

Two objections bear on this paper. The canons come in opposing pairs, so within textualism a selection among canons remains, and the selection is uncontrolled by the ordering (Llewellyn 1950). And the determination of settled usage requires materials, whether dictionaries, corpora, or the interpreter’s linguistic competence, and each of these is a body of usage assembled by somebody for some purpose. Textualism’s claim to draw on the least is a claim to draw on a region of the network whose composition is less visible, and its constraint is therefore real and smaller than advertised.

9.3 Intentionalism

Intentionalism orders the modes with the purpose actually held by the adopting authority first, and looks to records of deliberation to establish it. The mischief rule is its oldest form and legislative history its modern instrument.

The standing objection is that a collective body holds no single mental state, that recorded statements are made by individuals who may be unrepresentative, and that a record known to be consulted becomes an object of strategic composition. The objection is strong for legislatures. For organizations it is weaker in one respect and stronger in another: the adopting body may be small enough to hold something like a shared intention, and the record of its deliberation is ordinarily unavailable to any party who would invoke it.

9.4 Purposivism

Purposivism orders the modes with purpose first and detaches purpose from actual intention. Hart & Sacks (1994) supply the classical statement, directing the interpreter to assume that the rule was made by reasonable persons pursuing reasonable purposes reasonably, and to attribute the purpose that assumption yields. Barak (2005) gives a systematic modern development, in which purpose is a legal construct assembled from subjective and objective components.

Attributed purpose escapes the objection to intentionalism and incurs its own. The attribution is performed by the interpreter, from materials the interpreter selects, under a standard of reasonableness the interpreter supplies. Purpose so understood constrains the interpreter to the extent that the attribution is disciplined, and the discipline is itself a matter of interpretation. This is the regress of §8 arriving at the level of the ordering, and it is the difficulty that §5 anticipated in noting that purposive interpretation supplies whatever the interpreter takes the purpose to have been.

9.5 Dynamic Interpretation

Eskridge (1994) holds that a statute’s meaning legitimately evolves, that the interpreter’s present circumstances bear upon what the text now requires, and that a doctrine denying this describes badly what interpreters actually do. The ordering places present conditions among the materials properly consulted.

The objection is that evolution weakens notice and transfers to the interpreter a power the adopting authority did not confer.

The doctrine deserves particular notice here for a reason unconnected with its merits in the legal debate. It is the doctrine most congenial to the criterion advanced in §21, since it treats a reading as revisable by design and regards the possibility of revision as a feature of interpretation and not a defect of it. A practice organized on dynamic lines holds its questions open as a matter of doctrine. A practice organized on strict textualist lines may hold them open as well, and does so for a different reason, since the text remains available to be read again. What closes a question is neither doctrine as such and is the institutional condition under which no party can raise it, which is the finding §19 reaches by another route.

9.6 Constructive Interpretation

Dworkin (1986) rejects the framing shared by the doctrines above. Interpretation of a legal practice is constructive: the interpreter seeks the reading that both fits the materials and shows the practice in its best light. Law as integrity requires that the practice be read as the work of a single author of consistent principle, and the chain novel supplies the image, each interpreter writing a chapter constrained by what precedes and free within that constraint. The strong claim that follows is that hard cases have right answers, and that the appearance of discretion reflects the difficulty of the question and not an absence of an answer to it.

Dworkin’s position is the most serious opponent of the premise this paper has adopted, and the confrontation should be stated plainly. The paper has followed Kelsen (1967) in holding that the materials leave a space within which several readings are equally admissible and nothing in the materials selects among them. Dworkin denies that any such space exists, since fit and justification together select, even where the selection is contestable and no procedure establishes it.

Three observations are entered in response, and only the third does real work.

The first is that the disagreement is narrower than it appears for the class of rules at issue here. Integrity draws on the density of a practice, and the argument that one reading shows a practice in its best light requires a practice with a history, a body of prior determinations, and articulated principles running through them. Organizational rule-interpretation ordinarily supplies none of these, so the resources that would select an answer are absent whether or not they would select one in a mature legal system.

The second is that Dworkin’s account is directed at what an interpreter ought to do and is compatible with an observer’s inability to verify that it was done.

The third is decisive for this paper, and it concedes the philosophical point. Suppose Dworkin is right and every question has a right answer. Interpretive injustice, on that supposition, becomes interpretive error, and the whole apparatus of this paper survives the change of name. The questions that concern it are who is positioned to establish the right answer, what follows when the organ’s answer is final and mistaken, and what protects a party against whom a mistaken answer stands. Those questions arise in the same form under a right-answer thesis as under a frame thesis, and the answers given in §19 are unaffected. What the right-answer thesis would remove is the paper’s occasion to say that no answer was uniquely required. What it leaves untouched is the paper’s subject.

9.7 The Constitutional Instance

The same disagreement recurs at constitutional scale between originalism and living constitutionalism, with the stakes raised by the difficulty of amendment and by the antiquity of the text. The dispute adds little that is structurally new and it supplies one contribution this paper adopts.

Bobbitt (1982, 1991) treats the forms of constitutional argument as modalities, historical, textual, structural, doctrinal, ethical, and prudential, and holds that they constitute the grammar within which constitutional propositions are made intelligible. His claim of interest here is that the modalities are incommensurable, that no meta-modality ranks them, and that when they conflict the choice among them is an act for which the interpreter is answerable in conscience and not by rule.

This is the most honest response available to the difficulty of §5, and it is a response that declines to resolve it. If Bobbitt is right, the search for an ordering is misconceived, the doctrines above are each an attempt to elevate one modality by fiat, and what remains at the point of conflict is a choice attributable to the person who makes it. For this paper the consequence is congenial and uncomfortable in equal measure. It supports the claim that the interpretive moment is an exercise of a power that the materials do not determine. It removes any prospect of a doctrine that would discipline the exercise from outside.

9.8 The Institutional Turn

Sunstein & Vermeule (2003) redirect the question. Whether a given interpretive doctrine is correct cannot be settled by an account of meaning alone, since a doctrine is a decision procedure administered by particular interpreters with particular capacities operating under particular constraints. The choice among doctrines depends on the error costs and decision costs those interpreters will actually incur, and on the effects the doctrine will have on the behaviour of those who draft and those who are governed. A formalistic doctrine may be justified for interpreters of limited capacity even where a more flexible doctrine would be superior for ideal ones.

This turn is the most important in the section for present purposes, and for two reasons.

It supplies the transition to the sections on practice, since it makes institutional condition a determinant of what interpretation should be and not merely of what it is.

And it changes what the survey of doctrines can deliver. If the appropriate doctrine is a function of the interpreter’s capacities and the surrounding institutional facts, then no doctrine surveyed here transfers to organizations by demonstration of its correctness. What transfers is a question: given this interpreter, these materials, and this population of affected parties, which ordering produces the fewest and least serious errors, and what does each ordering cost to administer.

9.9 The Doctrines Compared

The doctrines may be set beside one another. The third column states the ground on which each ordering is defended, and the fifth states what a body must possess before the ordering is available to it, which is the finding developed in the subsection following.

Doctrine	Consulted first	Ground of the ordering	Characteristic objection	Materials required
Textualism	Settled usage of the words	Constraint on the interpreter; notice to the governed	Canons come in opposing pairs; usage is itself assembled	Adopted text; corpora or dictionaries
Intentionalism	Purpose actually held by the adopting body	Fidelity to what was enacted	No collective mental state; records composed strategically	Record of deliberation
Purposivism	Purpose attributed on a standard of reasonableness	Purpose is part of what the rule is	The attribution is performed by the interpreter	Materials supporting attribution
Dynamic interpretation	Present conditions alongside the text	Describes what interpreters do; rules outlive their settings	Weakens notice; transfers power not conferred	Account of original and present settings
Constructive interpretation	The reading that best fits and justifies the practice	Integrity; the practice read as one author’s work	Demands of the ideal interpreter; imports substantive commitments	Dense practice with a history of decisions
Modalities	No ordering; forms are incommensurable	Choice at conflict is answerable in conscience	Supplies no discipline at the point of conflict	Professional community sharing the grammar
Institutional approach	Whichever ordering the interpreter administers best	Error costs and decision costs of actual interpreters	Correctness becomes contingent on setting	Knowledge of interpreters and effects

9.10 The Transfer to Organizations

The doctrines were developed for courts reading statutes and constitutions, and their transfer to organizations reading their own rules is uneven in a way that has a pattern.

Each doctrine presupposes an infrastructure. Textualism presupposes an adopted text and a body of usage. Intentionalism presupposes a record of deliberation. Purposivism presupposes materials from which a purpose can be attributed and a practice of attributing purposes with discipline. Dynamic interpretation presupposes an articulable account of present conditions and of the rule’s original setting. Constructive interpretation presupposes a dense practice with a history, prior determinations, and principles running through them. Bobbitt’s modalities presuppose a professional community within which the modalities are recognized as the grammar of argument.

Organizations possess these unevenly and ordinarily possess few. The consequence, stated as a finding of this section, is that the doctrines available to an organization are a function of its documentary and professional infrastructure, and that the orderings which constrain the interpreter most tightly are the ones demanding most infrastructure. A body with an adopted text and nothing else can practise a thin textualism and can practise nothing else, and thin textualism is the ordering that leaves most to the interpreter while appearing to leave least, since the determination of settled usage in the absence of corpora, records, and a professional community is a determination the organ makes and no party can check.

This finding will recur in institutional form in the sections on practice, where the devices most often proposed for organizations are devices that build infrastructure: published criteria, recorded reasons, indexed determinations, and precedent within the body.

9.11 Findings and Residue

This section establishes that the doctrines are orderings of the modes, that each ordering meets a serious objection, that the strongest philosophical opponent of the paper’s premise leaves its subject untouched, that the institutional turn makes the choice among orderings depend on interpreter and setting, and that organizations can use only the orderings their infrastructure supports. It leaves open whether any ordering can be justified independently of institutional condition, which is a question this paper does not pursue, and it contributes to Q3 in §24.

§10 Architectures of Interpretive Authority Beyond the Western Doctrines

This section returns to the first question. §4 established that standing to interpret is conferred at the core and generated at the edge, and left open whether interpretive standing may be delegated and what grounds its legitimacy. The doctrines surveyed in §9 say little about either, since they were developed for a setting in which the identity of the authoritative interpreter is settled by constitutional arrangement and the interesting questions concern how that interpreter should proceed. One tradition treated the prior question at length, and its architecture supplies resources the Western material does not foreground.

A caution governs the whole section. What follows is a reading of a large tradition for a narrow purpose, and it selects the structural features bearing on interpretive authority while passing over most of what that tradition contains. Specialists will find the treatment thin, and the claims entered are confined to structure.

10.1 Interpretation Separated from Decision

Islamic legal theory begins from a distinction with no counterpart in the doctrines of §9. The divine law as such and the jurists’ understanding of it are named separately, and the rulings that make up the developed law belong to the second category. A ruling is a qualified person’s understanding of what the law requires, and it is held as such by the tradition’s own vocabulary (Hallaq 1997; Weiss 1998).

The consequence for interpretive authority is considerable. A system whose vocabulary distinguishes the law from the understanding of it has built the fallibility of interpretation into its own self-description, and has done so without weakening the authority of what is decided. The determination binds and is known to be a determination.

The institutional expression of this is a separation the Western material treats as unusual. The jurist who issues an opinion and the judge who decides a case occupy distinct offices with distinct powers (Masud et al. 1996). The opinion is authoritative in the sense that it states what a qualified interpreter holds the law to require, and it does not dispose of the case. The judgment disposes of the case and is not, on that account, a superior statement of the law. A party may seek an opinion from a jurist who has no power over them, and the opinion’s authority rests on the jurist’s standing in the interpretive community.

Two features of this arrangement bear directly on the questions left open in §4.

Interpretive authority is here delegable, and its delegation is the ordinary case. The organ that decides is not the sole or even the principal interpreter, and the interpretive function is discharged by a body of persons who hold no decisional power at all.

And standing to interpret is grounded in demonstrated competence. The qualifications required of one who may perform independent derivation are articulated at length: command of the language, of the source texts, of the received consensus, and of the methods by which rulings are derived. One who lacks them follows a qualified jurist, and the following is a recognized and respectable position with a name of its own. Authority thus rests on a public standard of competence, assessable by others holding the same competence, and it is held independently of office.

10.2 Plurality and the Standing of Disagreement

The second feature of the tradition bearing on this paper is its treatment of disagreement.

Interpretive work was organized into schools, each transmitting its own methods and doctrines, each recognizing the others as legitimate, and each maintaining its positions alongside theirs. Disagreement among qualified interpreters has a name and a literature, and the plurality was accommodated as a standing condition of the law and not as a defect awaiting resolution.

Underlying this is a theoretical dispute that the reader of §8 and §9 will recognize. One classical position holds that where qualified interpreters differ, each has attained what was required of them and each ruling is correct. A rival position holds that one answer is correct and the interpreter who misses it has erred, while remaining blameless for having reasoned properly. The dispute is the same one that divides Kelsen (1967) from Dworkin (1986), arrived at independently and some centuries earlier, and its terms are close enough that the correspondence deserves recording.

What the tradition adds is the institutional consequence it drew from the dispute, which the modern debate has not drawn. Whichever position was taken, plurality was accommodated institutionally: schools coexisted, a party might follow one, and a ruling from one school did not extinguish the others. The disagreement about whether one answer is correct was thereby separated from the question of whether one answer must be enforced.

A historiographical caution belongs here. The received account that independent derivation ceased at a certain period, the gate having been closed, has been challenged at length, and the challenge is now widely accepted in outline (Hallaq 1984). The structural point made here does not depend on the resolution of that controversy, since it concerns the architecture and not its vitality at any particular date.

10.3 Reception and Transplant

The devices of the modern administrative state reached East Asia through reception of continental doctrine, and the receiving systems modified what they received. Japan’s administrative procedure legislation requires agencies to establish and publish criteria against which applications and dispositions are assessed, which converts a portion of interpretive practice into published rules. The arrangement is treated in §19 among the devices, and it is noted here because it illustrates the point made by the institutional turn of §9: a device transplanted into a different institutional setting operates differently, and its efficacy is a function of the setting and not of its form.

10.4 The Structural Resources Supplied

Two structural resources emerge from the material of this section, and each answers a question the Western doctrines leave unaddressed.

Separation of interpretive from decisional authority. An interpretive function may be discharged by persons holding no power to dispose of the case. This bears directly on the recursion identified in §3, since a power that determines the scope of its own review is recursive in a way that a power to interpret held apart from the power to decide is not. Whether the separation can be reproduced in a modern organization, and whether it relocates the recursion or reduces it, is Q1 of §24, and the material of this section is evidence that the arrangement is possible and has been sustained at scale.

Competence as the ground of standing. Interpretive authority may rest on a public and articulated standard of qualification, assessable by others who hold it. This is a candidate answer to Q2, and it is a demanding one, since it requires a community capable of assessing competence and a body of method against which competence is measured. Sections §9 and §19 both note the absence of such communities around most organizational rule-interpretation.

10.5 Cautions on Transfer

Islamic legal theory is treated here as a source of structure and not as a model of justice. It was administered within its own asymmetries, and access to the interpretive resources described was unevenly distributed by status, by wealth, by sex, and by locality. A reader who took this section for an argument that these arrangements were more just than those surveyed elsewhere would be taking it for something it does not assert.

And the institutional turn applies with full force. A device drawn from one setting and installed in another will operate according to the capacities and incentives of the new setting. The resources identified here are entered as possibilities whose transferability is undetermined, and determining it is work this paper does not undertake.

10.6 Findings and Residue

This section establishes that interpretive authority has been separated from decisional authority at scale and that standing has been grounded in assessable competence. It supplies evidence bearing on Q1 and Q2 of §24 and settles neither, since the possibility of an arrangement in one setting is weak evidence for its availability in another.

§11 Justice in Legal Interpretation

The sections preceding this one have asked how interpretation proceeds and who is positioned to perform it. This section asks what has been said about when interpretation is just, which is the fourth question in the form the legal tradition has given it. Five accounts are treated. Each makes a normative demand upon the interpretive act itself, each is the nearest ancestor of some part of the criterion advanced in §21, and each stops short of the case this paper is concerned with, for a reason the section identifies in each instance.

11.1 Congruence as a Condition of Legality

Fuller (1969) sets out eight requirements whose satisfaction constitutes what he calls the inner morality of law, and the eighth concerns interpretation directly. There must be congruence between the rules as declared and the actions of the officials who administer them. A system in which the declared rule says one thing and official action does another has failed as a legal system, and has failed in a manner distinct from having bad rules or reaching bad outcomes.

Two features of Fuller’s treatment matter here. He treats the failure of congruence as a matter of degree and as an aspiration toward which a system moves, so that legality is a scalar property of an institution and not a threshold it crosses. And he identifies among the causes of failure not only corruption and indifference but mistaken interpretation, so that an official who misreads a rule impairs legality by doing so, without any impropriety of motive.

Congruence is the closest classical ancestor of the criterion of §21, since both treat the relation between rule and application as an object of normative assessment in its own right. The difference is exact and it is the reason congruence cannot serve here. To ask whether official action conforms to the declared rule presupposes that the declared rule has a content against which conformity may be measured. Where the rule’s term is determinate, congruence has purchase and is a serviceable test. Where the term is a standard or a comparative, which §6 identified as the class carrying the interpretive load, the content against which conformity would be measured is what the interpretive act supplies. Congruence therefore has least to say exactly where the question is live, and the requirement passes over the penumbral case in silence.

11.2 The Rule of Law and the Discipline of Discretion

Raz (1979) treats the rule of law as the law’s capacity to guide those subject to it, and derives from that capacity a set of principles: rules must be prospective, open, and clear, must be relatively stable, must be made by open and stable general rules, and must be administered by an independent judiciary, with accessible courts and observance of natural justice. He adds a principle that bears on the present subject: the discretion of agencies charged with administering the law should not be permitted to pervert it.

Raz’s account is instrumental and he says so, treating the rule of law as a negative virtue that reduces a danger created by law itself. Two limits follow for this paper. The capacity to guide, like congruence, presupposes ascertainable content, so the account addresses the vagueness of rules as a defect of drafting and does not address the interpretive act that supplies content where drafting could not. And the principles concerning independent adjudication and accessible courts are addressed to legal systems, so they identify a condition organizations mostly do not satisfy without offering anything to a party in an organization that does not.

11.3 The Argumentative Dimension

Waldron (2011) argues that the dominant conceptions of the rule of law have overweighted rules and underweighted procedure, and that a legal system is characterized as much by its argumentative apparatus as by its norms. Courts hear parties, receive submissions about what the law requires, give reasons, and permit appeal. On Waldron’s account these are constitutive of legality and are not administrative accessories to it, and among them is the opportunity of a person affected to argue about what the law means as applied to their case.

This is the nearest existing normative demand to the criterion advanced here, since it holds that the standing to contest an interpretation is a component of legality itself. A system that determined applications without hearing argument about them would be deficient as a legal system on Waldron’s account, whatever the quality of its determinations.

Two limits attach. The account is developed for legal systems possessing courts, and its transfer to bodies without them is a question Waldron does not take up. And it establishes an opportunity to argue without addressing the condition identified in §3, in which the body that receives the argument about the meaning of a rule is the body whose interpretation is contested, and whose competence to hear the argument at all is settled by its own reading of the rules that govern challenge. An opportunity to argue before the author of the reading is worth something and is worth less than the account assumes.

11.4 Participation as the Mark of Adjudication

Fuller (1978) distinguishes adjudication from other forms of social ordering by the mode of participation it accords the affected party. Contract accords participation through negotiated consent; managerial direction accords none of a comparable kind; adjudication accords the distinctive participation of presenting proofs and reasoned arguments. The distinguishing feature is not the presence of a decision-maker and is the party’s opportunity to present a case for a decision in their favour.

Two consequences bear on this paper.

The first is normative and supports the criterion. If presenting reasoned argument is what makes adjudication the form of ordering it is, then a party’s contribution to the determination is constitutive of the form and not a courtesy extended within it. A determination reached without any such opportunity is something other than adjudication, whatever the deciding body is called.

The second is a limit, and it is severe. Fuller’s account defines a form and does not require that any given decision take that form. An organization allocating scarce goods is ordinarily engaged in managerial direction on Fuller’s taxonomy, and managerial direction carries no participation requirement of the adjudicative kind. The account thus explains why participation rights do not attach to admission, appointment, or funding decisions, and in explaining it, licenses the absence. This is a case where the tradition’s own vocabulary works against the party.

11.5 Legality in Private Organizations

Selznick (1969) supplies the only sustained treatment in this tradition of the question this paper poses, and does so for private governance. Legality on his account is a value that private institutions can realize in degree, through the movement from managerial prerogative exercised at will toward administration governed by rules, with articulated standards, records, reasoned decisions, and channels of appeal. Industrial due process is his name for the resulting condition, and the study documents its emergence in the internal governance of firms.

The account is the direct precedent for the present subject, and one feature of its findings requires notice here because it anticipates §19 from a different direction. The developments Selznick documents occurred where a countervailing organized force was present, most often a union with the capacity to make the employer’s exercise of prerogative costly, and where legal doctrine had begun to attach consequences to arbitrary internal action. Legality within private organizations, on the evidence of the study that established the concept, developed where somebody outside the management held something the management needed. This is the stake identified in §19, observed historically in the domain where the concept of organizational legality originated.

11.6 Interpretation and Application

One further account belongs here for the record, since it bears on §3. Gadamer (2004) holds that understanding a normative text and applying it to a case are aspects of one operation, so that a rule is understood in and through its application to circumstances, and legal interpretation is exemplary of understanding generally for that reason.

The account is descriptive and it carries a normative implication that this section should register. If application is internal to understanding, then the demand that an interpreter first ascertain a meaning and afterwards apply it describes something that does not occur, and a fidelity requirement framed in those terms asks for the impossible. What can be asked is that the application be one the text can bear and that the interpreter remain open to the case altering what they took the text to require. The second of these is close to what §21 will require of an institution, transposed to the interpreter’s own disposition.

11.7 Findings and Residue

The accounts surveyed establish a good deal. Interpretation is within the concerns of legality and is not a merely technical matter lying outside them. Congruence between declared rule and official action is a condition of a system’s claim to legality and fails through misreading as well as through corruption. The standing to argue about what a rule means is a component of the rule of law and not an accessory to it. The affected party’s presentation of reasoned argument is constitutive of adjudication as a form. And legality is a value that private organizations can realize in degree, which is the premise the present paper requires and does not have to argue for.

What the tradition leaves is a pattern. Each account either presupposes a determinate content against which application may be measured, which is what the penumbral case lacks, or presupposes an independent tribunal before which argument may be made, which the organizational case ordinarily lacks, or documents conditions that obtained where a countervailing force was present, which is the case §19 identifies as the one where the difficulty does not arise. The penumbral determination, made finally by a body answering to nobody, is the residue none of them reaches, and it is the case this paper takes as its subject.

§12 Relation to Adjacent Concepts of Injustice

This section individuates the concept. A concept that cannot be distinguished from its neighbours has not been introduced, and interpretive injustice has four neighbours near enough that the burden falls on the present account to show what it adds. The four are hermeneutical and testimonial injustice, procedural injustice, domination, and the jurispathic office. Each is treated by locating where its injustice resides, in whom it is borne, by what mechanism it operates, and what remedy it implies, since concepts that share a name may still be told apart by their remedies.

12.1 Hermeneutical and Testimonial Injustice

Fricker (2007) distinguishes two epistemic injustices. In testimonial injustice a speaker receives a credibility deficit owing to prejudice attaching to their identity. In hermeneutical injustice a collective deficiency in interpretive resources leaves a significant area of a person’s experience unintelligible, to others and often to themselves, with the deficiency falling unequally on those excluded from the practices in which such resources are generated.

Hermeneutical injustice is the nearest neighbour by name and is a different concept. Its locus is a collective stock of resources; the present concept’s locus is an office. Its bearer suffers an inability to render an experience intelligible; the bearer here suffers a determination rendered intelligibly and finally by someone else. Its mechanism is a gap; the mechanism here is an exercise. And the remedies diverge sharply, which is the decisive test. The remedy for hermeneutical injustice is the enrichment of the collective interpretive stock, together with the virtue Fricker calls hermeneutical justice in the individual hearer. Enriching the collective stock does nothing for a party whose case has been determined by an organ holding authority to fix the meaning of its own rules, since the determination stands whatever the surrounding vocabulary contains.

One overlap should be conceded precisely. The representational asymmetry entered in §3, on which ordinary language supplies thin terms for objecting to an interpretation, is a hermeneutical deficit in Fricker’s sense. A party who cannot articulate the objection because no available vocabulary expresses it suffers the injustice Fricker describes, and suffers it in addition to the one described here. The present paper may be read, in that narrow respect, as an attempt to supply a portion of the missing vocabulary. The concepts remain distinct: supplying the vocabulary removes the hermeneutical deficit and leaves the interpretive authority exactly where it was.

12.2 Procedural Injustice

Rawls (1971) distinguishes perfect, imperfect, and pure procedural justice. Where an independent criterion of the correct outcome exists and a procedure reliably attains it, the procedure is perfect; where the criterion exists and no procedure reliably attains it, imperfect; where no independent criterion exists and the outcome is defined as whatever a fair procedure produces, pure.

Interpretation sits awkwardly in this scheme, and the awkwardness is instructive. An independent criterion appears to exist, since the rule is there and the parties dispute what it requires, which places the case under imperfect procedural justice. Yet the argument of §3 and §8 holds that within the frame no criterion selects a unique reading, which pushes the case toward pure procedural justice, on which whatever the organ determines through a fair procedure is by definition what the rule required. The second characterization is the one organizations implicitly adopt, and adopting it dissolves the possibility of interpretive injustice by construction. This paper resists the dissolution on the ground given in §3: the space of admissible interpretations is bounded, so a determination may be shown to sit outside the network from which the term takes its content, and a criterion therefore exists even where it selects no unique answer. Procedural justice of the pure kind is available for the selection among admissible readings and is unavailable for the question whether a reading was admissible.

The empirical literature raises a second and sharper difficulty. Tyler (1990) establishes that perceptions of fairness turn on voice, neutrality, respectful treatment, and trustworthiness, and that they drive acceptance of outcomes and compliance with authority more strongly than outcome favourability does. Read alongside Luhmann (1969), the finding carries an uncomfortable implication for any account that treats procedural justice as sufficient. The features that produce the perception of fairness are features an organization can supply while its interpretive practice remains beyond contest. Acceptance is therefore an unreliable indicator of interpretive justice, and a well-run organization may generate acceptance precisely where the present concept would locate an injustice.

12.3 Domination and Contestability

Pettit (1997) defines domination as subjection to the capacity of another to interfere arbitrarily in one’s choices, where interference is arbitrary when it need not track the interests and ideas of the person affected. Freedom as non-domination requires that such capacity be removed, and Pettit’s institutional proposal turns centrally on contestability: power is rendered non-arbitrary when those subject to it can effectively contest its exercise and require it to answer to their avowable interests.

This is the closest neighbour, and the relation to it is one the paper should state as complementary. Domination names a standing relation and is indifferent to the mode through which interference occurs. Interpretive injustice names a mode, and the mode is of interest because it is the one that presents itself as the absence of interference. An organ that interprets is not, on its own account, interfering at all, and is applying a rule to which both parties are subject.

The complementarity becomes substantive at the point of remedy. Pettit’s remedy for arbitrary power is contestability, and the recursive component of asymmetry entered in §3 bears directly upon it. The rules that determine whether a contestation lies, whether it is in time, whether it concerns a matter within review, and whether the contestant has standing are rules of the organization, interpreted by the organization. Contestability is therefore itself an object of the interpretive power it is meant to constrain. This is offered as a contribution to the republican account and not as an objection to it: it identifies a mechanism by which the republican remedy can be absorbed, and the identification is of use to anyone attempting to institutionalize contestability in a body that writes its own rules.

12.4 The Jurispathic Office

Cover (1983) holds that normative worlds are generated abundantly by communities, each with its own account of what its law means, and that the distinctive work of a court is to suppress the competitors: courts are jurispathic, and they kill law by choosing one reading and extinguishing the others that were living in the communities that held them.

Cover supplies the vocabulary nearest to what the later sections of this paper require, since he treats interpretation as an act with generative and destructive effects upon normative worlds. Two differences should be marked. Cover’s setting is jurisgenerative plurality, in which the party whose reading is extinguished holds a developed nomos of its own; the setting here is ordinarily one in which the party holds a case and no competing normative world. And Cover’s account is descriptive of a role he takes to be unavoidable, since a plural legal order requires that some readings be suppressed if any are to be enforced. The present account does not dispute the necessity of determination and asks after what a determination does to the capacity of the relation within which it was made, which is a question Cover does not pose.

12.5 The Individuation Stated

The four comparisons may be summarized.

Concept	Locus	Bearer suffers	Mechanism	Implied remedy
Hermeneutical injustice	Collective stock of resources	Unintelligibility of experience	Deficit	Enrich the stock
Procedural injustice	The path to decision	Denial of voice or neutrality	Defective process	Reform the process
Domination	A standing relation	Liability to arbitrary interference	Capacity to interfere	Institutionalize contestation
Jurispathic suppression	The plural legal order	Extinction of a normative world	Authoritative choice	Accepted as unavoidable
Interpretive injustice	An office holding rules	A determination of the rule’s content for their case	Constitutive exercise, closed upon its own review	Open question

Two features of the final row distinguish it from the others. The mechanism is constitutive, so the party is bound by a content the rule did not carry before the decision. And the mechanism is closed upon its own review, so the ordinary remedy of the third row, institutionalized contestation, is exposed to the very power it constrains. The final column is left open deliberately. Each of the neighbouring concepts arrives with a remedy attached, and the burden of the remainder of this paper is that the remedies available for interpretive injustice are structurally limited in a way the others are not.

§13 Theoretical Strategies for Constraining Interpretation

§9 asked which mode of interpretation governs where the modes diverge. This section asks a prior question: what could discipline an interpreter at all. A doctrine proposes an ordering. A strategy proposes a source of constraint, and the doctrines draw upon the strategies without being identical with them. Five families are distinguished here, and the fifth is the one to which the account of §21 is most closely related.

13.1 Constraint by Text

The first strategy fixes content by the words of the rule, and its instruments run from ordinary textualist reading through definition clauses and controlled vocabularies to the complete specification of a rule as a decision procedure.

The limiting case deserves separate notice, since it is a live proposal and not a thought experiment. A rule may be expressed in a form admitting of mechanical execution, whether as an algorithm, a decision table, or executable code, and bodies administering high volumes of like cases have moved substantial parts of their rule systems into that form. Hildebrandt (2015) examines the transformation and its consequences for legal protection. Where the transformation is complete the interpretive moment disappears from the point of decision, since nothing remains to be determined when the case arrives.

13.2 Constraint by Purpose

The second strategy fixes content by the end the rule serves, holding that a reading defeating the end is thereby excluded. Its instruments are purpose clauses, recitals, statements of objectives, the mischief rule, and the teleological method. The strategy is the one most often recommended to organizations, since it appears to require only that the body state why a rule exists.

13.3 Constraint by Procedure

The third strategy disciplines the interpreter through the manner of proceeding, leaving the content of the determination formally open. Its instruments are the requirement to give reasons, the composition of the deciding body, the separation of the function of investigating from the function of deciding, the availability of appeal, the publication of determinations, and time limits within which a party may respond. This family is the one on which institutional practice principally relies, and §19 surveys its instruments in their developed forms.

13.4 Constraint by Community

The fourth strategy locates the discipline in a community of competent practitioners. An interpreter proceeds under the standards of a practice, and what constrains them is the requirement that their reading be one they could defend to others who hold the same competence and would recognize a departure.

Fish (1980) gives the strong statement, on which interpretive communities supply the assumptions that make a reading available at all, so that the constraint operates before the interpreter chooses and shapes what presents itself as a possible reading. Brandom (1994) supplies a more general account of the practice within which such constraint operates, in which participants keep score of one another’s commitments and entitlements and the practice of giving and asking for reasons supplies the discipline. The communal solution to the rule-following problem noticed in §8 belongs to this family, and so does the arrangement described in §10, in which standing to interpret rests on a competence assessable by others holding it.

This family is the one the present account is nearest to, since it locates constraint in a relation among parties and not in a property of the text or of the end. §14 identifies where it and the account of §21 part.

13.5 Constraint by Reflexivity

A fifth strategy addresses itself to systems and not to interpreters. Teubner (2012) holds that private and transnational orders, which no external authority is positioned to govern, may nonetheless be constrained by inducing them to develop internal limits of their own, so that the system installs procedures by which it restrains itself and remains open to the claims of those it affects. Constitutionalization on this account is something a social order can undergo without a state to impose it.

The strategy is the most directly addressed to the situation this paper is concerned with, since it takes as its starting point the absence of an external authority. Its difficulty is entered in the next section.

§14 The Difficulties of the Strategies

§13 set out five families of constraint and assessed none of them. This section takes each in turn, in the order there established. The difficulties are of different kinds, and they share a form which the closing subsection states.

14.1 The Underdetermination of Text

Text underdetermines application, which §8 established and need not be repeated.

The limiting case requires separate treatment, since its difficulty is of a different kind. Complete specification does purchase determinacy at the point of decision, and it purchases it by relocating the interpretive act to the point of specification. Someone determined that these cases fall under the rule and those do not, and encoded the determination. The act that was performed openly by an organ deciding a case is performed in advance by whoever writes the specification, ordinarily without a case before them, ordinarily without any party present, and ordinarily without the result being legible as an interpretation at all. The consequence is the one identified in §4 as the divergence of formal and effective authority, in its most complete form: the interpretive power has moved to a position that the rules do not name, that no party can address, and whose exercise presents itself as a technical implementation. Determinacy at the point of application is thereby real, and the interpretive question has been concealed and not answered.

14.2 The Regress of Purpose

Purpose is itself an object of interpretation, so the strategy regresses. Hart & Sacks (1994) direct the interpreter to attribute a purpose that reasonable persons would have held, and the attribution is performed by the interpreter from materials the interpreter selects. Where the purpose is stated in the instrument the regress is shortened and does not terminate, since a stated purpose is expressed in terms whose application requires determination.

14.3 The Formality of Procedure

Procedure secures the form of a determination and leaves its content formally untouched. This is a limitation by design and not a defect, since a procedural constraint that dictated outcomes would have ceased to be procedural.

Two consequences follow that bear on this paper. Luhmann (1969) establishes that procedural elaboration produces acceptance by a mechanism independent of the correctness of what is decided, and Tyler (1990) supplies the empirical form of the same finding. An organization may therefore increase procedural quality, increase the acceptance of its determinations, and leave its interpretive practice exactly where it was. And the recursive feature identified in §3 bears with particular force on this family, since procedural constraints are themselves rules of the organization, and their scope, their applicability, and the consequences of their breach are determined by the body they constrain.

14.4 The Reach of Community

Constraint by community fails in a manner that is easy to miss, because the failure is compatible with the strategy operating perfectly.

A community of competent practitioners disciplines an interpreter with respect to other members of that community. The reading an interpreter must be able to defend is a reading defensible to peers. Where the affected party is a member of the community, this is a substantial protection. Where the affected party stands outside it, the strategy supplies nothing to them directly, and it supplies nothing at all in the case that matters most: a community may converge upon a reading that operates harshly against those outside it, the convergence being precisely what the strategy calls constraint working. Uniformity among interpreters is what the family delivers, and uniformity is consistent with any content.

This is the point at which the present account parts from the family it most resembles. Both locate normative weight in a relation. The constraint-by-community strategy locates it in the relation among interpreters, and the criterion of §21 locates it in the relation between the interpreter and the party interpreted upon. The first asks whether a reading could be defended to those competent to assess it. The second asks whether the reading leaves the affected party positioned to raise the question again. A practice may satisfy the first completely while failing the second completely, and §19 identifies the institutional form of that combination under the heading of constituency alignment.

14.5 The Precondition of Reflexivity

Reflexive constraint presupposes that a system possesses the capacity to install limits upon itself and stands in an environment that presses it to do so. Where the environment presses, the strategy describes something real and observable. Where it does not, the strategy states a possibility and identifies nothing that would bring it about. The presupposition is the stake of §19 arriving in systems-theoretic vocabulary, and the strategy most directly addressed to the absence of external authority turns out to require an environment performing the office that the absent authority would have performed.

14.6 The Common Residue

The difficulties may be set out together.

Strategy	Source of discipline	Presupposes	Characteristic failure
Text	The words as used	Shared and stable usage	Underdetermines; complete specification relocates the act to the encoder
Purpose	The end served	Recoverable and determinate ends	Purpose is itself interpreted
Procedure	The manner of proceeding	That form tracks substance	Secures acceptance without correctness
Community	Peers competent to assess	Membership and shared standards	Uniform among members, silent toward outsiders
Reflexivity	The system’s self-limitation	An environment that presses	Requires the stake whose absence defines the problem

Each strategy presupposes something it cannot itself secure, and the presupposition is in each case a condition of the setting and not a feature of the strategy. This is the theoretical counterpart of the result reached institutionally in §19, and the coincidence of the two is worth stating plainly: the strategies fail where the devices fail, and they fail for the same reason. What disciplines an interpreter is, in every family, something outside the interpreter that the interpreter answers to, and the cases of concern to this paper are those in which nothing outside the interpreter is positioned to be answered to.

The sections that follow turn from strategies to instruments, and ask whether the devices institutions have actually built perform better than the strategies they draw upon.

§15 Interpretive Control in Legislation

The four sections beginning here survey what institutions have built in answer to the fourth question. They proceed by scale, and this one treats the scale at which interpretation is anticipated before any case has arisen. The devices are accordingly prospective: they operate upon the rule and upon the conditions of its later reading, and none of them operates upon a determination once made.

15.1 Interpretation Statutes

Several jurisdictions maintain a statute governing the reading of all other statutes. The United Kingdom’s Interpretation Act 1978 is the compact form, supplying standard definitions, default rules on number and gender, and provisions on commencement and repeal. Canada’s Interpretation Act directs that every enactment be deemed remedial and be given the fair, large, and liberal construction that best ensures the attainment of its objects. Australia’s Acts Interpretation Act 1901 requires in section 15AA the interpretation that would best achieve the purpose or object of the Act, and permits in section 15AB recourse to extrinsic material to confirm or to determine meaning. New Zealand’s Legislation Act 2019 directs that meaning be ascertained from the text in the light of purpose and context. Singapore’s Interpretation Act contains a comparable purposive provision.

The genre deserves notice for a reason that the doctrinal literature does not always make explicit. §9 treated the ordering of the modes as a matter on which doctrines disagree and no meta-rule adjudicates. These statutes settle the question by legislation. A purposive mandate is a legislative choice among the orderings surveyed in §9, imposed on all subsequent interpreters and removed from their discretion. Whether the imposition succeeds is a separate matter, since the mandate is itself a rule whose application requires the determination of a purpose, which is the regress of §14. What is established is that the ordering question is answerable by authority even where it is not answerable by argument.

15.2 Definition and Purpose Clauses

Two drafting devices operate directly on the material §6 classified.

A definition clause converts a term from one class to another. A standard term whose application requires evaluation on every occasion may be given an exhaustive definition and thereby become classificatory, and a classificatory term may be given an enumerated extension and become determinate. The interpretive load of the rule falls accordingly. The device has a cost that §6 entered: the latitude the undefined term preserved is surrendered, and bodies that value the latitude have reason to define at a level of generality that retains it.

A purpose clause supplies the material that the purposive mode requires and that §5 found ordinarily absent outside legal systems. Its presence converts an unavailable mode into an available one. Its limitation is that a stated purpose is expressed in terms whose own application requires determination, so the device shortens the regress and does not terminate it.

Taken together, the two devices are infrastructure-building in the sense identified in §9: they supply the materials on which the more constraining orderings depend, and they are the clearest instance in this survey of a body improving its own interpretive discipline by making the modes available.

15.3 Control of Delegated Interpretation

Where a legislature confers on another body the power to elaborate its rules, it confers a portion of the interpretive function, and the devices controlling the conferral are controls upon interpretation.

The Treaty on the Functioning of the European Union distinguishes in Articles 290 and 291 between acts that supplement or amend non-essential elements of a legislative act and acts that establish uniform conditions for its implementation, and reserves essential elements to the legislative act itself. The reservation is of interest here because it identifies a class of interpretive questions that may not be delegated at all, which is a control of a kind no other scale in this survey supplies.

The United Kingdom’s practice of conferring power to amend primary legislation by subordinate instrument, known by the name of the Tudor statute from which the practice takes its label, is the opposite case, and the sustained criticism of it concerns exactly the transfer of interpretive and legislative power to a body operating with less scrutiny.

In the United States the doctrine restricting delegation requires that Congress supply an intelligible principle to guide the delegate. The doctrine has been enforced rarely for many decades and has been the subject of renewed argument in recent years, and its current condition should be verified before it is relied upon, in line with the caution entered in §2.

15.4 Authentic Texts in More Than One Language

The European Union adopts its legislation in all official languages, each version being equally authentic, and the Court of Justice accordingly compares versions where they diverge and resolves the divergence by reference to the purpose and general scheme of the measure. In CILFIT the Court set out the conditions under which a national court may treat the correct application of Union law as so obvious that no reference is required, and in doing so recorded the multilingual character of the legislation and the comparison of versions among the considerations bearing on interpretation.

This arrangement is worth a paragraph of its own because it is an institutional demonstration of a proposition that §8 argued philosophically. Where one text exists, the claim that the text underdetermines its application can be resisted by asserting that a competent reader simply sees what the words mean. Where several texts are equally authentic and they diverge, the insufficiency of text is on the face of the record, and the system is compelled to adopt an ordering that places purpose and scheme above the words of any single version. Multilingualism converts a contested philosophical thesis into an operational necessity.

15.5 Temporal Devices

A last family operates on the duration of a rule and of the readings it attracts. Sunset provisions terminate a measure at a fixed date unless it is renewed. Review clauses require that a measure be examined after a stated interval. Post-legislative scrutiny, conducted by committee in several parliamentary systems, examines how a measure has operated in practice.

These devices deserve emphasis disproportionate to the attention they ordinarily receive, because they are the closest thing this survey encounters at any scale to a device addressed to the concern of §21. A sunset provision compels a question to be asked again. A review clause compels the operation of a rule, which is to say the accumulated body of readings given to it, to be examined by a body other than the one applying it. Neither device was designed for the protection of a party against an interpretation, and both have the effect of preventing a settled reading from passing beyond examination, which is the condition the criterion of §21 requires.

15.6 Findings and Residue

Legislative devices supply prospective control of considerable power. The ordering of the modes may be settled by statute. The interpretive load of a term may be reduced by definition. The materials that more constraining modes require may be created by drafting. Classes of question may be reserved against delegation. And the periodic reopening of settled readings may be compelled by temporal provision.

What the scale leaves is uniform across every device surveyed here, and stating it is the point of the section. Each device operates upon the rule, upon the drafter, or upon a delegate, and none operates upon a determination once it has been made. None supplies anything to a party against whom a reading has been applied. None is triggered by a complaint, none is available at the instance of an affected person, and none requires that any determination be revisited because a party has objected to it. The devices of this scale are entirely on the side of the body that makes the rules, and the affected party has no place in any of them.

§16 Interpretive Control in Standard-Setting

This scale comprises bodies that write rules without legislative authority and whose rules bind through the conditions others attach to dealing. A standard is adopted voluntarily and becomes practically obligatory where a purchaser, a lender, an insurer, a regulator, or a procurement rule requires conformity with it. The interpretive question here is compounded, since a party subject to a standard faces determinations about what the standard requires made by a body that did not write it, applying a text written by a body over which the party has no influence.

The instruments surveyed below are the most developed answers this survey encounters to the fourth question, and the section closes by identifying what they do not do.

16.1 Standards Governing the Writing of Standards

A first family regulates the process by which standards are made.

Annex 3 to the World Trade Organization’s Agreement on Technical Barriers to Trade sets out a code of good practice for the preparation, adoption, and application of standards, addressed to standardizing bodies and concerned with transparency, non-discrimination, and the avoidance of unnecessary obstacles to trade. The ISO/IEC Directives govern the procedure of the international standards bodies themselves, and require the development of standards by consensus among participating members. The ISEAL Codes of Good Practice occupy the same position for sustainability standards, specifying requirements for standard-setting, for assurance, and for the measurement of impact, and functioning as a standard whose subjects are standard-setters.

This family deserves notice because it is unusual. It consists of rules addressed to the rule-making of private bodies, imposed by no sovereign, and taken up because bodies which fail to observe them lose credibility among those whose recognition they require. It is the clearest existing instance of the reflexive strategy of §13, and the condition of its operation is exactly the stake identified in §19.

16.2 Standards Requiring Internal Grievance Mechanisms

A second family requires the organizations to which it applies to establish channels through which those affected by their conduct may raise complaints. This family is the most directly relevant of any instrument in this survey to the question of §21, since it specifies in normative terms what an internal mechanism must be like.

The United Nations Guiding Principles on Business and Human Rights (United Nations 2011) state in Principle 31 the criteria by which the effectiveness of a non-judicial grievance mechanism is to be assessed. Such a mechanism should be legitimate, in the sense of enabling trust from the stakeholder groups for whose use it is intended and being accountable for the fair conduct of grievance processes. It should be accessible, being known to those it is intended to serve and providing assistance to those who face barriers to access. It should be predictable, providing a clear and known procedure with an indicative timeframe and clarity on the types of process and outcome available and the means of monitoring implementation. It should be equitable, seeking to ensure that aggrieved parties have reasonable access to sources of information, advice, and expertise necessary to engage in a grievance process on fair, informed, and respectful terms. It should be transparent, keeping parties informed about progress and providing sufficient information about performance to build confidence in its effectiveness. It should be rights-compatible, ensuring that outcomes and remedies accord with internationally recognized human rights. It should be a source of continuous learning, drawing on relevant measures to identify lessons for improving the mechanism and preventing future grievances. And an operational-level mechanism should be based on engagement and dialogue, consulting the stakeholder groups for whose use it is intended on its design and performance and focusing on dialogue as the means to address and resolve grievances.

The criterion of equitability is worth isolating. It requires that a party have access to the information, advice, and expertise needed to engage on informed terms, which is a direct address to the informational asymmetry entered in §3 and, at least in form, to the difficulty identified in §6 concerning determinations under comparative terms.

The OECD Guidelines for Multinational Enterprises on Responsible Business Conduct (OECD 2023) supply a companion arrangement in which adhering states maintain National Contact Points that receive specific instances raised against enterprises and offer good offices toward their resolution. The IFC Performance Standards require a grievance mechanism for affected communities as a condition of finance, and the Equator Principles carry the requirement into commercial project lending. ISO 10002 gives guidance on complaints handling within an organization and ISO 10003 on dispute resolution external to it.

16.3 Standards Governing Governance and Compliance

A third family addresses the internal ordering of the organization itself. ISO 37000 provides guidance on the governance of organizations. ISO 37301 specifies requirements for a compliance management system and is drafted for certification. ISO 37001 addresses anti-bribery management systems. ISO 26000 offers guidance on social responsibility and is expressly not intended for certification.

The relevance of this family to interpretation is indirect and real. A compliance management system requires that obligations be identified, that responsibilities be allocated, that determinations be documented, and that non-conformities be recorded and addressed. Its effect where implemented is to build the record that §5 found the precedential and systemic modes to require, and §9 identified as the infrastructure on which the more constraining orderings depend.

16.4 Conformity Assessment and Accreditation

Standards bind through the machinery that certifies conformity with them, and that machinery is itself standardized. ISO/IEC 17021-1 states requirements for bodies providing audit and certification of management systems, and ISO/IEC 17065 for bodies certifying products, processes, and services. ISO/IEC 17011 states requirements for the accreditation bodies that assess the certifiers. Mutual recognition among accreditation bodies is organized through the International Accreditation Forum and the International Laboratory Accreditation Cooperation, so that a certificate issued under one national accreditation is recognized under others.

Two features of this machinery bear on the present subject.

The requirements for certification bodies include documented processes for appeals against certification decisions and for complaints, so a party dissatisfied with a certifier’s determination has a specified channel. The interpretive-control device is thus applied recursively to the bodies that administer it, which is an arrangement no other scale in this survey exhibits.

And the object of assessment is the conformity of a management system. An auditor examines whether the organization has established the required processes, whether it follows them, and whether it records what it does. The examination does not extend to whether any particular determination the organization made was correct, and it is not designed to.

16.5 Findings and Residue

The instruments surveyed in this section are the only family in the survey requiring organizations to establish channels by which their own determinations may be contested from within, and that specifies in normative detail what such a channel must be like. Anyone seeking an existing statement of what an internal mechanism for interpretive contest ought to look like will find it in Principle 31 and in the assurance requirements that accompany the certification standards. To that extent the scale answers the fourth question more directly than any other surveyed here.

Five limits attach, and they are cumulative.

Adoption is voluntary. No instrument in this section obliges any organization to submit to it, and an organization that declines is subject to nothing. The instruments operate on organizations that have chosen to be bound, and the choice is ordinarily made because somebody whose recognition the organization requires has conditioned dealing upon it.

The criteria are themselves standard terms. Whether a mechanism is predictable, equitable, or accessible is a question requiring evaluation on every occasion of use, and it is answered by an auditor applying the criteria to a case. The interpretive question is thereby relocated to the assessment of the mechanism and is not dissolved by it, which is the pattern §6 identified for published criteria generally.

Assessment addresses the system and not the determination. This limit is the most serious and it is structural. A certified organization has demonstrated that it operates a conforming process. Nothing in the assessment establishes that a determination reached through that process was correct, and an organization may hold current certification while any particular interpretation it has made stands uncorrected. The apparatus is procedural in the sense of §14, and inherits that family’s characteristic failure in full.

The instruments bite through a stake. Certification is valuable because a buyer, a lender, a procurement rule, or a regulator requires it, and where none does, the mark is worth what it costs to obtain and nothing more. This is the presupposition identified in §19, and it is present here in its purest form, since the entire apparatus is built upon it and makes no pretence otherwise.

And the standard-setter’s composition matters. Bodies that write standards for an industry are ordinarily populated in substantial part by that industry, since the expertise required to draft resides there. The consequence, treated in §19 under the heading of constituency alignment, is that the standard and the practice it governs are written by parties with a common interest in the practices they share, and a determination made in accordance with the shared practice is unlikely to be classified as a failure by the body that certifies against the standard.

§17 Administrative Review and the Duty to Give Reasons

This scale is the most developed in the survey. Public administration decides high volumes of individual cases under general rules, the decisions bear directly on those subject to them, and legal systems have built over a century an apparatus addressed to the resulting problem. The instruments are grouped here by function. The comparative question of how far each system permits review to reach the substance of an evaluative determination is held for §18.

17.1 The Duty to Give Reasons

The most widespread instrument requires that a decision be accompanied by a statement of the grounds on which it was reached.

German administrative procedure law requires that a written administrative act state its reasons, including the essential factual and legal grounds on which the authority proceeded, and adds a provision of particular interest here: where the act rests on an exercise of discretion, the statement should also indicate the considerations from which the authority proceeded in exercising it. This is the closest any instrument in the survey comes to requiring the exhibition that §5 identified as the test of a bounded reading, since it obliges the authority to expose the route and not the conclusion alone.

French law has required since 1979 that individual administrative decisions of specified adverse classes be accompanied by reasons stating the considerations of law and of fact on which they rest, a requirement now carried in the code governing relations between the public and the administration.

In Union law the duty to state reasons for acts of the institutions is imposed by the Treaty, and the Charter of Fundamental Rights includes among the components of the right to good administration the obligation of the administration to give reasons for its decisions, alongside the right to be heard and the right of access to one’s file. The Charter also guarantees an effective remedy before a tribunal.

In the United States the requirement operates through the standard of review. An agency action may be set aside as arbitrary and capricious, and the reasoned-explanation requirement developed under that standard obliges the agency to articulate a satisfactory explanation including a rational connection between the facts found and the choice made.

The United Kingdom has no general common law duty to give reasons for administrative decisions, and the categories in which a duty has been recognized have expanded considerably, particularly where a decision affects an important interest or appears aberrant without explanation.

The significance of this family for the present subject was stated in §5 and is repeated here because the survey confirms it. A reading is assessable as bounded when the interpreter exhibits what was drawn upon. Reason-giving duties are the institutional form of that requirement, and their strength varies with how much of the route they compel into the record. A duty satisfied by a statement of the conclusion and the provision applied leaves the interpretive act exactly as unexaminable as it was.

17.2 Criteria Published in Advance

A second family requires the administration to state, before any case arises, the criteria by which it will assess cases of a kind. The device converts a portion of prospective interpretive activity into published rules, and it is developed with unusual explicitness in Japanese administrative law.

Japan’s Administrative Procedure Act requires an agency to establish criteria for the examination of applications and to make them public, and separately provides for criteria governing adverse dispositions, in terms that oblige the agency to endeavour to establish and publish them. The distinction between the mandatory and the endeavour formulations is worth marking, since it locates precisely where the legislature was willing to constrain and where it retreated. The Act also requires that reasons be shown when an application is refused and when an adverse disposition is made.

The device is the institutional realization of what §6 treated as the partial remedy of published criteria, and both limits noted there apply. The criteria are composed of terms whose application requires determination, so the interpretive question moves and does not disappear. And the agency retains the drafting choice of how specific to be, so the latitude that the undefined term preserved may be preserved again at one remove.

17.3 Procedural Due Process

A third family attaches procedural entitlements to determinations affecting protected interests. In the United States these are constitutional in source. The requirement of a hearing before termination of certain benefits was established in Goldberg v. Kelly, and the process due in a given case is determined by the balancing enquiry of Mathews v. Eldridge, which weighs the private interest affected, the risk of erroneous deprivation under the existing procedures together with the probable value of additional safeguards, and the governmental interest including administrative burden (Mashaw 1985).

The family regulates the path to a determination. It secures notice, an opportunity to respond, and in some settings an impartial decider, and it addresses the content of the determination only through the risk-of-error factor in the balancing enquiry, which asks whether additional procedure would reduce error and does not ask whether this determination was in error.

17.4 Internal Review Before Adjudication

A fourth family interposes a tier between the initial determination and any court.

Japan’s system of administrative complaint review was substantially revised in the last decade to introduce a hearing officer who has not been involved in the original disposition and a third-party review body whose opinion is sought before the decision on the complaint is made. Both changes address the difficulty identified in §3 as recursive, since they interpose persons who did not make the determination under challenge. Comparable internal tiers exist in Korean administrative procedure and in the internal review stages of many administrative schemes in common law jurisdictions.

The interposition of a person not previously involved is the most direct institutional response in this survey to the recursion of §3. It is also a partial one, since the reviewing person or body remains within the same organization, applies the same rules, and is ordinarily subject to the same interpretive practice and the same guidance.

17.5 Ombudsman Institutions

A fifth family investigates complaints against administration without deciding legal rights. The European Ombudsman receives complaints of maladministration in the institutions of the Union, and national ombudsman offices perform comparable functions.

Two features are relevant here. The jurisdiction is defined by maladministration, which is broader than illegality and reaches quality of administration, delay, discourtesy, and failure to follow one’s own procedures. And the outcome is a recommendation without binding force.

The combination is worth naming, because it is a trade the other families do not make. The ombudsman can reach conduct that no court would review, including much that concerns how an authority handled a matter, and can do so precisely because nothing turns legally on the finding. Breadth of jurisdiction is purchased with absence of compulsion.

17.6 Findings and Residue

The public administrative scale supplies more than any other surveyed. Determinations must ordinarily be reasoned, and in one system the reasons must expose the considerations bearing on an exercise of discretion. Criteria may be required to be settled and published in advance. Procedural entitlements attach where protected interests are affected. An internal tier staffed by persons not previously involved may be interposed before any court is reached. And an office of broad jurisdiction may examine the quality of administration without any legal right being at stake.

Three limits are entered here and a fourth is held for the next section.

The apparatus is addressed to public bodies. Its instruments follow from the character of public power, and they do not apply to the private and transnational bodies that make many of the determinations with which this paper is concerned. A party dealing with a private organization ordinarily has none of this.

The reason-giving duties are qualified in every system by exceptions, by classes of decision to which they do not extend, and by the sufficiency standards applied to the statements given. A duty discharged by formulaic recitation supplies nothing to the assessment of boundedness.

The internal tier remains internal. Interposing a person not previously involved reduces the recursion identified in §3 and does not remove it, since the reviewer is subject to the organization’s own interpretive practice and to the guidance that shaped the original determination.

And the fourth limit is the one that governs the whole scale. However well developed the apparatus, its reach into a determination depends on how much of the interpretive act the reviewing forum is willing to examine, and every system surveyed reserves some part of the evaluative judgment to the deciding body. What each system reserves, and how far the reservation extends, is the subject of §18.

§18 Regional Variation in Deference and Academic Judgment

The apparatus surveyed in §17 reaches a determination only so far as the reviewing forum is willing to look. This section compares how far, across several systems, and takes as its running example the class of determinations in which the reservation is most complete. Academic judgment is chosen because every system surveyed reserves it in some measure, because the reservation is defended on grounds that are respectable, and because the class exhibits in a pure form the features that §6 identified as concentrating interpretive load.

18.1 Review of Interpretation as a Question of Law

Systems differ at the threshold on whether the interpretation of a rule’s terms is a matter on which a reviewing forum forms its own view.

German administrative law gives the strongest affirmative answer. The indeterminate legal concepts in which statutes are written are treated as questions of law, and a court reviewing an administrative act determines their meaning for itself, without deference to the authority’s reading. A margin of appreciation on the side of the statutory element is recognized only in limited and enumerated categories, among them examination assessments, appraisals of civil servants, decisions of pluralistically composed expert bodies, and certain prognostic judgments. Discretion in the German sense attaches to the legal consequence and not to the meaning of the statutory terms, and it is reviewed for excess, for abuse, and for failure to exercise.

The United States has recently moved in the same direction by a different route. The doctrine requiring courts to defer to a permissible agency construction of an ambiguous statute was overruled in 2024, and reviewing courts now exercise independent judgment on whether an agency has acted within its statutory authority, with an agency’s contemporaneous and consistent reading carrying such persuasive weight as its reasoning merits. Deference to an agency’s reading of its own regulations survives in narrowed form, subject to conditions established in 2019.

French administrative law distinguishes intensity of review by subject matter, applying a normal control to some determinations and a restricted control to others, under which only a manifest error of assessment will be censured. The United Kingdom’s general standard of unreasonableness sets a high threshold for intervention in evaluative judgments, with proportionality applying where rights are engaged.

18.2 The Reservation of Academic Judgment

Against that variation, the treatment of academic assessment is more uniform, and the differences among systems concern the threshold at which the reservation yields.

In the United States the courts have held that a genuinely academic decision may not be overridden unless it constitutes such a substantial departure from accepted academic norms as to demonstrate that the person or committee responsible did not actually exercise professional judgment. Procedural protections in academic dismissals are correspondingly lighter than in disciplinary ones, on the ground that academic evaluation is not adapted to the procedural tools of the adversarial process.

In the United Kingdom a student may complain to the Office of the Independent Adjudicator for Higher Education, established under statute, and that office does not consider matters of academic judgment. The courts have likewise treated issues of academic judgment as unsuited to determination by a court while accepting that matters of contract and procedure may be justiciable.

In France the sovereignty of the examining jury is a settled principle, and administrative courts examine the composition of the jury, the regularity of the procedure, and manifest error, while declining to substitute an assessment of merit.

Germany is the exception on which §19 already relied. The Federal Constitutional Court held in 1991 that an examinee’s answer which is defensible and cogently reasoned may not be marked as wrong, and thereby constrained the latitude of examiners over the substance of assessment. The route was constitutional, the occupational consequences of the examinations in question were substantial, and the effect was to convert a portion of what other systems reserve into a matter on which a reviewing body forms its own view.

Japan supplies the strongest reservation, and by an explicit doctrine. The Supreme Court has held that certain internal determinations of a university belong to the internal affairs of a partial society within the general legal order, and are for that reason outside the scope of judicial review, unless they bear upon a person’s status in general civic legal relations. The doctrine is of direct interest to this paper, since it states in judicial terms exactly the proposition that §19 arrives at institutionally: that an organization’s internal determinations may lie beyond the reach of external adjudication as a matter of principle and not merely of practice.

18.3 The Comparison Summarized

System	Interpretation of rule terms	Academic assessment	Threshold for intervention
Germany	Question of law, court forms own view	Margin recognized, constitutionally constrained	Defensible answer may not be marked wrong
United States	Independent judicial judgment since 2024	Reserved to professional judgment	Substantial departure from accepted academic norms
France	Intensity varies by subject matter	Sovereignty of the jury	Manifest error of assessment
United Kingdom	Deferential on evaluative questions	Excluded from the statutory adjudicator	High threshold of unreasonableness
Japan	Reviewable in general administrative law	Internal affairs of a partial society	Effect on general civic status

18.4 Findings of the Comparison

Three findings follow, and the third is the one the paper carries forward.

Every system surveyed reserves some portion of substantive evaluative judgment to the deciding body. The reservation is not an accident of any legal culture and is defended everywhere on the same two grounds: that the reviewing forum lacks the competence to redo the assessment, and that the deciding body’s autonomy in matters of its own expertise is a value in itself. Both grounds are sound, and this paper does not dispute either.

The systems differ substantially in the threshold at which the reservation yields, and the variation is wide enough to establish that the reservation’s extent is a choice and not a necessity. Germany’s constitutional constraint on examiners and Japan’s partial-society doctrine mark the ends of the range, and both are settled law in developed systems. An institution that treats the extent of its own reservation as given is treating a policy as a fact.

The systems also differ in how much of the interpretive act is forced into the reviewable record, and this dimension is separable from the first. A system may reserve the substance of an assessment while requiring that the criteria be published, that the reasons expose the considerations relied upon, and that the assessment be conducted by persons not previously involved. The material of §17 showed each of these varying independently of the standard of review. This matters for the criterion of §21, because what that criterion requires is that a question remain askable, and a question is askable to the extent that what was determined, and on what grounds, is on a record somebody can address. A reservation of substance combined with a full record leaves a question open. A reservation of substance combined with a bare conclusion closes it.

18.5 The Limit of the Comparison

One caution attends the whole section. The doctrines compared here are the formal positions of legal systems, and this survey is doctrinal in the sense entered in §2. How often the available routes are used, by whom, at what cost, and with what result are empirical questions on which the comparative literature is thin. A system whose doctrine permits searching review and whose procedures make review practically inaccessible resembles, from the position of an affected party, a system that reserves the question entirely.

§19 The Asymmetry Boundary

This section discharges the fourth question and states the result toward which the survey of practice is directed. The result is a boundary condition on the whole apparatus assembled in answer to that question. Every device examined in the preceding sections operates upon a stake the interpreting body holds in recognition by some party outside itself, and every device loses its purchase as that stake declines. Where the authority to interpret and independence from recognition are held together, the devices do not reach.

19.1 The Common Presupposition

The devices surveyed fall into six families, and each may be characterized by what it operates upon.

Compulsory review attaches a cost to a defective interpretation by exposing the decision to annulment. It presupposes a forum with jurisdiction over the body, and it presupposes that annulment costs the body something it values. Conformity assessment, comprising certification, accreditation, and the standards that govern them, attaches a cost by withholding a mark that buyers, procurers, insurers, or regulators require. It presupposes a party who conditions dealing upon the mark. Reputational devices, comprising disclosure, publication, ranking, and reporting, attach a cost by informing an audience whose regard the body needs. Professional devices attach a cost by imperilling standing within a community of peers whose recognition the body’s officers require in order to practise. Exit, in Hirschman (1970)’s sense, attaches a cost by withdrawal of participation from a body that depends on participants. Contractual devices, comprising grievance clauses, arbitration agreements, and negotiated procedures, attach a cost by rendering an obligation enforceable, and they presuppose a forum, a bargain, and a counterparty with the standing to have struck it.

The families differ in the currency of the cost and agree in their form. Each supplies the interpreting body with a reason to interpret well, by attaching a cost to interpreting badly, and each transmits that cost through a party whose recognition the body requires. Call that party the device’s constituency. The efficacy of a device is a function of three conditions, all of which must hold: the constituency must be able to observe the interpretive act, the constituency’s withdrawal of recognition must impose a cost on the body, and the officers who perform interpretive acts must internalize some part of that cost.

19.2 Mediation by a Third Party

A feature of this structure deserves to be marked, since it explains why protection can fail while every device is functioning as designed.

The constituency is ordinarily somebody other than the party against whom the interpretation was exercised. The affected party is an applicant, a member, a supplier, a complainant, or a candidate. The constituency is a court, a certifying body, a purchaser, a professional association, a funder, or a public audience. Protection of the affected party is therefore mediated: it arises when the affected party’s grievance is legible to the constituency, and when the constituency treats that grievance as a reason to withdraw recognition. Neither condition is guaranteed by the existence of the device.

Two consequences follow. The affected party’s protection varies with the interest the constituency takes in cases of that kind, and that interest is distributed unevenly across grievances for reasons unconnected with their merit. And an organization may satisfy every device in the six families, holding its certifications, surviving its reviews, and retaining the regard of its audience, while a particular interpretation goes uncorrected, because no constituency has any occasion to register that particular case. The devices are not thereby defective. They were built to discipline patterns, and a single interpretive act is not a pattern.

19.3 Conditions Under Which the Stake Vanishes

The stake on which the devices operate approaches zero under conditions that are structural and jointly common.

The first is surplus. Where the number of parties seeking what the organization allocates exceeds what it can allocate, the loss of any one party costs the organization nothing, and the exit device has no currency. An institution admitting a fraction of its applicants is in this position with respect to every applicant it declines.

The second is singularity of supply. Where the organization is the sole or near-sole source of what the party seeks, whether a qualification, a licence, a certification, or a position, exit is unavailable to the party in the first place, and the threat that would operate the device cannot be formed.

The third is the asymmetry of repetition, in Galanter (1974)’s terms. The organization is a repeat player, holding the interpretive practice, the record of prior determinations, the relationships with the reviewing forums, and the capacity to structure cases so as to protect the practice. The affected party is a one-shot player, holding a single case, no record, and no prospect of a further interaction in which recognition might be withheld.

The fourth is non-observability. Where determinations are individualized, unpublished, or confidential, the constituency cannot observe the interpretive act, and the first condition of efficacy fails at the outset. Individualized decision is the ordinary form of the interpretive moment, so this condition is common.

The fifth is substitutability of reasons. Where a challenged interpretation can be replaced by a ground the reviewing forum will not examine, observation by the constituency does not yield a classification. A determination that a background lacked relevance, challenged, may be succeeded by a determination that the field was strong and the places few. The second ground states a comparative evaluative judgment, which is the class of judgment every reviewing forum surveyed reserves to the deciding body.

The sixth is constituency alignment. Where the constituency is composed of bodies of the same kind as the one under discipline, as with a standard-setter populated by those it standardizes, or a professional body composed of the profession, the constituency’s interest and the body’s interest coincide over the interpretive practices they share, and withdrawal of recognition is not a live possibility for that class of case.

The conditions are independent, and the case of concern to this paper displays several together. A body that allocates a scarce good, holds a near-monopoly on the qualification it confers, faces one-shot parties, decides individually and without publication, may state a comparative ground upon challenge, and answers to a constituency of its own peers, holds no stake upon which any device in the six families can operate.

19.4 The Boundary Stated

The result may be stated.

The asymmetry boundary. Every device developed for the protection of a party against interpretive injustice operates upon a stake the interpreting body holds in recognition by a constituency outside itself. The efficacy of such a device is bounded above by the magnitude of that stake and by the legibility of the interpretive act to that constituency. Where the authority to interpret and independence from recognition are held by the same party, the interpretive act lies beyond the reach of the assembled apparatus, and no addition to the apparatus of the same kind alters the position.

Three qualifications belong with the statement, and each restricts it.

The claim concerns devices of the six families and not the concept of protection as such. It asserts an upper bound on a class of instruments and does not assert that a party is without recourse in every case falling within the conditions of the preceding subsection, since conditions may be present in degree.

The claim carries no criticism of any instrument. The devices are well constructed for their range, and their range is the domain in which the interpreting body depends upon others. Most organizations, most of the time, occupy that domain. The boundary marks where the domain ends, and the instruments are not at fault for ending with it.

The claim is structural, so it is not remedied by better implementation. A more rigorous certification scheme, a more searching review, or a more transparent report each raises efficacy within the domain and none of them extends the domain, since each continues to operate upon the same stake.

19.5 Cases Against the Boundary

Two classes of instrument appear to escape the structure, and both should be examined, since the second yields a genuine partial exception.

The first is protection by a specified impermissible ground, of which anti-discrimination law is the developed instance. Such regimes attach a cost to a decision regardless of the deciding body’s dependence upon anyone, and they meet the substitutability of reasons directly, by shifting the evidentiary burden to the decision-maker once a prima facie case is shown. The apparent escape is real within its scope and the scope is narrow in the way that matters here. The regimes operate where the impermissible ground is specified in advance and the complaint can be brought under it. A complaint that an interpretation was poorly made, with no protected characteristic in play, falls outside them entirely, and the burden-shifting device has nothing to shift.

The second is constitutional adjudication, and it supplies the strongest case against the boundary. The German Federal Constitutional Court’s examination decisions of 1991 reached into a domain that reviewing forums had treated as reserved, holding that an examinee’s defensible answer may not be counted as wrong, and thereby constraining the latitude of examiners over precisely the substantive evaluative judgment that deference doctrines protect. This is penetration of the interpretive act itself, achieved without any dependence of the examining body upon the examinee.

The exception should be conceded and then measured. Its conditions were demanding: a constitutional right of the affected party was engaged, the evaluative act was of a standardized and recurring kind admitting of articulable criteria, a forum of constitutional rank was available, and the whole was situated in a public system of examination. Where a right of that rank is not engaged, where the evaluative act is comparative and individualized, or where the body is private or transnational, none of the conditions obtains. The exception establishes that the boundary is a feature of the ordinary institutional landscape and is not a logical necessity, and it establishes the price of crossing it, which is a constitutional right, a reviewable criterion, and a forum of sufficient rank.

19.6 The Position Reached

Two routes remain open, and the paper takes the second while recording that the first is the practical one.

The first route alters the asymmetry. If protection depends upon the existence of a constituency whose recognition the body requires, then protection may be pursued by constituting one: by aggregating one-shot parties into a body with a standing interest, by attaching the interpretive practice to a condition of funding or accreditation held by a party the organization does depend upon, or by rendering determinations observable so that a constituency which exists may see what it presently cannot. Each of these is available, each is a form of collective action or of institutional design, and none is addressed by this paper, whose subject is what may be said about an interpretive act and not how a constituency may be assembled. The route is named here so that the turn taken next is not mistaken for a claim that nothing practical remains.

The second route asks what an interpretation does, given that no party may be positioned to sanction it. It sets aside the question of who administers a cost, and asks after the effect of the interpretive act upon the relation within which it was performed. This is the question the generative relational sections take up, and the boundary established here is what makes the question worth asking, since it establishes that the alternative question, of who will impose a cost, has an answer only within a domain the cases of greatest concern lie outside.

§20 Interpretation as a Generative Relational Act

The boundary established in §19 closes one line of inquiry and opens another. If the question is who will impose a cost upon a body that interprets badly, the answer is available only within a domain that the cases of greatest concern lie outside. This section poses a different question about the same act, and the following section states the criterion that answers it.

20.1 The Shift of Object

Three questions have been asked of an interpretation in the foregoing sections. Was it admissible, which asks whether it fell within the space the rule’s terms permit. Was it fairly reached, which asks after the procedure. Did it harm, which asks after the party’s position afterwards. Each is a real question and each has been treated at length.

The question posed here concerns the relation within which the act occurred. An interpretive act is performed by one party upon a question that arose between two, and it leaves that relation in some condition. The condition of interest is the relation’s capacity to produce a further interpretation of the same question. An act may leave that capacity intact, may narrow it, and may extinguish it, and the three outcomes are the object of assessment.

The shift is worth stating carefully, since it is easily mistaken for one of the three questions already asked. It is not the question of admissibility, since an entirely admissible interpretation may extinguish the capacity, and an inadmissible one may leave it wholly intact. It is not the question of procedure, since a scrupulous procedure may terminate in an act that closes the relation, and a careless one may leave the question fully open. And it is not the question of harm, since a party may fare badly under an act that left everything open, and may fare well under one that settled the matter permanently.

20.2 The Relation as the Bearer

That a relation should be the bearer of the property in question requires the commitment entered in §3. A term’s extension is fixed within a network of usages, purposes, and practices, and the parties to an interpretive act occupy positions within that network. The party who interprets contributes to what the network contains, since each determination becomes a usage available to the next. The party interpreted upon may contribute, may be positioned to contribute in future, or may be positioned so that nothing they say enters the network at all.

The capacity at issue can now be located. A relation possesses the capacity to generate a further interpretation when the network from which the rule’s terms take their content remains open to contribution from the positions both parties occupy. The capacity has three components, and they can fail separately. The question must remain askable, which fails when the matter is placed beyond the scope of any further consideration. The party must remain positioned to ask it, which fails when the party is removed from the network by the determination itself. And what the party offers must remain capable of entering, which fails when the practices of the network admit contributions from one position alone.

20.3 Outcomes for the Capacity to Regenerate

The interpretive act stands in one of three relations to this capacity.

It may leave the capacity intact. The determination settles the case and the question remains live: the reading may be raised again by this party in another case, by another party in a like case, or by the organization itself upon reconsideration, and what any of them says can enter the network from which the term takes its content.

It may narrow the capacity. The determination settles the case and forecloses some portion of the ground on which the question might be reopened, whether by installing a reading whose revision would now require unsettling other matters, by consuming resources the party would need to raise it again, or by rendering the reading harder to see as a reading.

It may extinguish the capacity. The determination settles the case and leaves the question unaskable from the position the affected party occupies, and from every position relevantly like it.

The third outcome is compatible with an admissible interpretation, a fair procedure, and a modest harm. That is the finding this section exists to establish, and it is what licenses treating the capacity as an object of assessment in its own right.

§21 The Criterion of Regenerability

This section states the paper’s normative core, marks what it excludes, relates it to the criterion advanced in a companion paper on value conversion, and enters the difficulty of applying it.

21.1 The Criterion Stated

Criterion of regenerability. An interpretation is unjust when it extinguishes the capacity of the relation within which it was performed to generate a further interpretation of the same question. The injustice consists in the extinction of that capacity and does not consist in the adversity of the determination, in the asymmetry of the position from which it was made, or in the finality of the particular case. Where the capacity is narrowed and not extinguished, the injustice is present in degree, and the degree is the extent of the narrowing.

The criterion turns on a distinction that must be stated before anything follows from it, since without it the criterion would condemn every decision an organization makes.

21.2 Finality of a Case and Closure of a Question

A decision must be final in some respect. A case that could be reopened without limit would never be decided, and an organization unable to close a case could not operate. Finality of the case is a condition of deciding at all, and the criterion does not touch it.

Closure of the question is a different matter. A system may hold a case finally decided while leaving the reading on which it was decided open to revision: subject to reconsideration in a later case, subject to challenge by a party in a like position, subject to argument that the network from which the term draws its content has been misread. The interpretive practice remains in motion while the individual case rests.

Injustice on the present account is the second and not the first. What the criterion condemns is the passage of a reading into a condition where it can no longer be raised as a reading, by anyone positioned to be affected by it. The individual party may accept that their case is over. What they have lost, where the criterion applies, is the standing of the question itself.

21.3 Forms of Extinction

The capacity may be extinguished in two ways, and they differ in what would be required to restore them.

Procedural extinction occurs where the channels through which a question might be reopened are cut. Time limits expire, the matter is placed within a category excluded from review, the determination is unpublished so that no later party can invoke it, or the rules governing challenge are themselves interpreted so as to exclude the challenge. The last of these is the recursive closure entered in §3, and it is the form of procedural extinction distinctive to this subject, since the power whose exercise is in question is the same power that determines whether the question may be put. Procedural extinction is reversible in principle, since a channel that was cut can be restored.

Structural extinction occurs where the determination removes the party who would reopen the question. The applicant departs and the relation ends with the determination. The member is expelled and holds no standing to raise anything. The category within which the party stood is dissolved, so no later party occupies a like position. In each case there is nobody positioned to ask the question again, and restoring a channel would restore nothing, since the party who would use it is no longer in the relation. Structural extinction is not reversible in the same way, and it is the more serious form.

A third condition should be marked, though it belongs to neither form cleanly. The capacity may be extinguished by exhaustion, where the channels remain open and the party’s capacity to use them has been consumed by the process of arriving at the determination. Exhaustion presents as procedural openness and functions as structural closure, and it is common enough that a criterion which ignored it would miss much of what it exists to catch.

21.4 Relation to the Criterion of Conversion Closure

A companion paper on value conversion (Huang, in preparation) advances a criterion of the same shape, and the relation between them should be stated plainly, since a criterion that recurs is either a strength or a duplication and the reader is entitled to know which is claimed.

There the object is an equivalence written across categories, where one party takes a value of one kind and returns a value of another, asserting the two equivalent. The criterion is closure: the injustice arises where the second party can no longer assert the non-equivalence, or where any appeal available to them operates within the fixed rate and cannot reach the rate itself. Here the object is an extension fixed for a term of a rule. The criterion is the extinction of the capacity to generate a further interpretation. In both cases an act fixes a content that was not antecedently fixed, and in both cases the injustice is located in the loss of the capacity to reopen what was fixed.

The shared core is acknowledged. This paper extends the criterion to a second object, and the claim entered is that the recurrence is evidence for the criterion, since a normative test that identifies the same structure across acts as different as a compensation payment and an admission decision has a better claim to be tracking something than one fitted to a single case.

Three features distinguish the present application, and they are the paper’s additions to the family.

The first is recursion. In the conversion case the party whose value was converted may appeal to a forum that did not perform the conversion. In the interpretive case the rules governing appeal are rules of the same organization and are interpreted by it, so the power in question extends to the conditions of its own review. Closure can therefore be achieved without any deliberate act of foreclosure, by the ordinary operation of the power upon the rules that would constrain it.

The second is temporal reach. A conversion rate closes for the parties to it. An interpretation closes for a class, since the reading enters the network as a usage and is available to govern every later case of the same kind. The population of parties affected by an interpretive closure is therefore indefinite, and includes parties who have not yet arisen and cannot be consulted.

The third is the asymmetry boundary of §19, which the conversion paper had no occasion to establish. There the possibility of external remedy could be left open. Here it has been shown that the available instruments are bounded by a stake the interpreting body may not hold, so the criterion is being advanced for a class of cases in which no remedy of the surveyed kind is available. That is a harder position for a criterion to occupy, and the following section states what it costs.

21.5 The Exclusions of the Criterion

A criterion earns its content by what it declines to condemn, and this one declines a great deal.

It does not condemn an adverse determination. A party against whom a rule has been interpreted has suffered no injustice on this account, however unwelcome the outcome, provided the question remains askable from the position they occupy.

It does not condemn the asymmetry of the interpretive position. That one party determines and another receives the determination is constitutive of rule-governed decision, and an organization unable to fix the application of its own rules could not decide.

It does not condemn finality, for the reason given above.

It does not require that the affected party’s reading prevail, nor that the organization be persuaded, nor that any particular procedure be available. What it requires is that the question survive the answer.

It does not condemn the settlement of a reading over time. An interpretation that has been raised, contested, and repeatedly affirmed may become settled through use, and settlement of that kind is the ordinary consolidation of a practice. The criterion attaches to the extinction of the capacity to raise the question and not to the stability of an answer that has withstood raising.

21.6 The Difficulty of Application

Two difficulties attach, and both are entered here so that a reader does not have to discover them.

The criterion applies retrospectively with more ease than prospectively. Whether the capacity to reopen a question was extinguished is answerable once time has passed and nobody has been able to reopen it. Whether a determination will extinguish it is answerable at the moment of decision only by conjecture, and an organization acting in good faith may be unable to tell which of its determinations is of the closing kind. A criterion usable only in retrospect is of limited service to the party before whom the determination stands.

And the criterion states a requirement without stating how it is met. To hold that a just interpretation leaves the question askable is to say nothing about what sustains askability in an organization over time, which is a matter of institutional condition, of the disposition of those who hold interpretive office, and of the standing of parties who are not present. Both difficulties are carried forward to the open questions, where they are put in the form in which they might be worked on.

21.7 The Position of the Criterion

The criterion delivers one thing the devices of §19 do not, and fails to deliver one thing they do.

What it delivers is assessability without a constituency. Whether an interpretation left the question askable is a property of the act and the relation, and it can be asserted by the affected party, by an observer, by a later party, or by the organization itself in reflection. It does not require that anyone hold a stake, that any forum have jurisdiction, or that any mark be withheld. Where the asymmetry boundary leaves nothing to be done, something can still be said, and said with a definite content.

What it fails to deliver is any consequence. A criterion that identifies an injustice and supplies no party to act upon it has stated a fact about the world and has changed nothing in it. The following section asks whether anything follows from an interpretive closure in the absence of a party to enforce against it, and the answer given there is weaker than the question deserves.

§22 Delegation to Dynamics

The criterion of §21 identifies an injustice and supplies no party positioned to act upon it. This section asks what follows from an interpretive injustice where no such party exists. A way of seeing is proposed, and the greater part of the section is given to what it does not assert, since the readings available to a sympathetic reader are stronger than the claim and would not survive examination. The section is offered in the spirit of a preliminary work: the account is extensible, it invites revision, and it is framed so that its useful part can be kept by a reader who discards the rest.

22.1 The Standing of What Follows

This section is the most speculative in the paper, and its standing should be fixed before it is read. What follows is not a result. It is a way of seeing, offered because the criterion of §21 leaves a question that the vocabulary of remedy cannot address, and because a dynamical vocabulary allows the question to be posed even where it cannot be answered. The section proposes a shape and disowns the mechanism. A reader persuaded by §19–§21 and unpersuaded by everything here loses nothing that those sections established, a severability the paper states again in §24.

22.2 The Barrier to Generativity

An organization is a structure of relations: among its officers, between the body and those it decides upon, between a rule and the cases brought under it. The capacity identified in §20, by which a relation generates a further interpretation, is distributed across those relations, and it is the capacity by which the organization keeps its readings fitted to conditions that do not stand still.

Interpretive injustice barriers that capacity. It does so through operations on the organization’s relational grammar that are heterogeneous and that this paper does not claim to enumerate: a rule may be added that forecloses a question, a practice that enabled revision may be withdrawn, a standing that would have carried a challenge may be rendered inapplicable, a channel that connected a grievance to a determination may be severed. These are different operations with different proximate forms, and the account offered here abstracts over them. What it attends to is not which operation obtains but the functional signature they share: the affected relations lose, in some degree, the capacity to generate a further interpretation. The claim is pitched at that functional level deliberately, and its generality is the point. If the consequence sketched below follows, it follows from the loss of generativity and not from any particular way of bringing the loss about, so the heterogeneity of the operations is a feature of the account and not a gap in it.

22.3 From Barrier to Drift

A generativity that was working kept the organization fitted to a moving environment by continual small adjustment, each odd case answered by a reading that had not existed before. Barriered, that adjustment stops for the relations affected, while the environment they were fitted to does not stop moving. A distance opens between what the organization is organized to handle and what reaches it, and the distance widens without injustice performing any further act, carried by the ordinary motion of the world. Nothing announces the widening, since the channel that would have registered it is among the casualties of the barrier: a question no party can raise generates no complaint, and the absence of complaint presents from inside as the absence of a problem.

In the vocabulary of the companion work on historical dynamics (Huang, in preparation, b), the same is expressible as a decline in a capacity to absorb disturbance. A configuration’s stability is its capacity to absorb disturbance and return to its form; a barriered generativity is a reduced capacity of that kind; and a configuration whose capacity to absorb has declined stands nearer to a boundary at which a disturbance it would formerly have absorbed carries it elsewhere. What the account does not supply, and what should be conceded before the routes are set out, is the disturbance itself. The decline of a capacity is not a force. It makes a configuration fragile and does not push it, and whether and when a disturbance arrives to find it fragile is outside anything the account of injustice determines.

22.4 Some Routes From Barrier to Fragility, Offered as Illustration

The step from a barriered generativity to a declining capacity to absorb can be traced along several routes, and four are set out here. They are offered as illustrations of how the step might go and not as the mechanism, which the account has disowned. They are neither exhaustive nor jointly necessary; a reader who rejects all four may still hold the functional claim of §22.2, and a reader who adds a fifth is doing what the account invites. They differ in strength and in the evidence available for them, and they are given in ascending order of support.

Loss of error correction. A barrier to generativity removes a channel through which an organization learns that a reading has ceased to fit. Its weakness as a route is that an organization has other sources of information, and that a reading may go unfitted for a long time without anything depending on the fit.

Withdrawal of voluntary compliance. Rule systems are administered on the compliance of those subject to them, and enforcement against the non-compliant is expensive. Tyler (1990) establishes that compliance turns substantially on the perceived fairness of the procedures by which authority is exercised, and does so more strongly than on the favourability of outcomes. A practice that determines questions and forecloses their reopening is a practice whose fairness those subject to it have reason to doubt. On this route interpretive closure raises the cost of governing, continuously and without any single episode being decisive. This route has empirical support that the first lacks, and it does not depend on the organization failing to notice anything.

Selection and the loss of variety. Parties who cannot contest a determination exit where exit is available, and parties who observe the practice from outside decline to enter. Those who remain are disproportionately those for whom the closed reading is unobjectionable. The variety of positions represented within the organization’s field therefore falls. Ashby (1956) states the relevant principle in general form: a regulator must command variety at least equal to the variety of the disturbances it is to handle. An organization whose interpretive practice reduces the variety it encounters is reducing the variety it can answer, and Hirschman (1970) supplies the mechanism by which the reduction occurs.

Latent accumulation. A closed question does not cease to be a question for the party who holds it. Grievances persist without registering, and they persist unregistered for precisely the reason that makes closure closure: no channel exists through which they would be recorded. The stock of such grievances is a slow variable in the sense of the companion work on historical dynamics (Huang, in preparation, b), and its growth is invisible in the indicators an organization consults, since those indicators are fed by the channels closure has cut.

The four routes are not alternatives among which the account must choose, and where more than one operates they compound: an organization losing its error correction may also be raising its cost of governing, narrowing the variety it faces, and accumulating what it cannot see. The paper does not claim that any of them operates in a given case. It claims that a barrier to generativity is the kind of thing from which routes of this sort may run, and it leaves the identification of the routes that run in a particular organization to work it does not undertake.

One limitation applies to every route and belongs here and not only in §23. Each describes a decline in the capacity of a configuration to absorb disturbance. None supplies a disturbance. The passage from a shallowing basin to a departure from it requires something to arrive, and nothing in the account of injustice determines that anything will, or when. This is why the account predicts, as its ordinary case, long intervals in which an injustice has been done, the capacity to absorb has fallen, and nothing whatever has followed.

A last point concerns what happens if a disturbance does arrive to find the capacity spent, and it answers a question the dynamical vocabulary raises and does not obviously settle: why the outcome should be reorganization and not destruction. The barrier to generativity operates on the organization’s relational grammar, on the rules, standings, and channels that order its parts. It leaves the parts themselves untouched. The officers, the resources, the accumulated competence, and the population the organization serves all persist through the transition. What the accumulated misfit discharges, when it discharges, is the ordering and not the material, and the same parts re-cohere under an altered grammar: a revised rule, a changed standing, a reconstituted body. Departure from a stable configuration is on this account a re-formation of a persisting set of elements, which is the sense in which a system pressed off one attractor reorganizes toward another and does not simply cease. The account is silent, and should be, on whether the configuration that succeeds is an improvement, on how long the passage takes, and on which of the available configurations is reached.

Figure 1. An illustration, not a model. In (a) a configuration occupies a basin whose depth stands for its capacity to return after a disturbance; as a barrier to generativity accumulates, the basin shallows, and past a threshold an ordinary disturbance carries the configuration into a neighbouring basin. In (b) the same process is drawn against accumulated barrier: the capacity falls toward nothing while the observable state moves very little, so that the decline is not registered by the indicators an organization would consult. The figure is drawn from a one-dimensional potential chosen for legibility and makes no empirical claim; the number of dimensions, the shape of the landscape, and the location of any boundary in a real organization are all unknown. Three cautions are built into the drawing. The second basin carries no evaluative content and is not the better one. The picture shows departure and says nothing about which configuration is reached. And the transition is a re-formation of the ordering among persisting parts and not a destruction of them.

The figure states the picture the routes illustrate and states its limits in the same breath. The left panel gives the received image of a stability landscape, in which the depth of a basin represents the capacity of a configuration to return after being disturbed. The right panel carries the one claim the section presses. What declines under a barrier to generativity is a capacity, and a capacity is not an output. An organization consulting its own indicators sees a state that holds steady, and the steadiness is compatible with the capacity having been almost wholly consumed. This is why an organization is poorly placed to notice the condition described here, and it restates in dynamical terms the epistemological point entered in §3: among the things a barrier to generativity removes is the instrument by which its own effects would have been registered. The figure earns its place by making that one point visible. It should not be read as asserting anything about the true dimensionality of an organization’s state, which the account does not know.

22.5 A Complication the Account Must Carry

An objection arises at once and it is a good one. Closure removes the corrective channel, and it also removes the disturbances that the channel carried. An organization that hears no complaints is subject to fewer live perturbations, and quietness of that kind is a form of stability. On the face of it closure may make a configuration more stable and not less.

The objection is partly right and the concession should be entered. What closure removes is the organization’s capacity to respond to change communicated through the parties before it. The conditions those parties inhabit continue to change whether or not the organization hears of it, and disturbance reaches an organization by routes other than the complaints of the parties it decides upon: through the bodies it depends on, through the environment in which it operates, through the eventual conduct of parties who were never able to say anything. Closure therefore purchases quiet at the cost of correspondence, and the configuration becomes at once less disturbed and less fitted. Whether the net effect on stability is a decline is a question the account cannot settle in general, and the claim entered here is confined to the capacity to respond, which closure reduces without qualification.

22.6 The Criticality Condition

A disturbance carries a configuration out of its present form only where the configuration already stands near a boundary. Away from such a boundary, disturbances are absorbed and the configuration returns, and the arrival of a disturbance produces nothing an observer would record.

The consequence for the present account is severe and should be stated as such. An organization may interpret unjustly, close the questions its interpretations raise, and continue in that condition for a very long time with no consequence whatever. The account predicts this, and predicts it as the ordinary case, since most configurations at most times are far from any boundary. Any reading on which injustice reliably produces its own undoing is contradicted by the account itself, which supplies no disturbance and no schedule on which one would arrive.

22.7 Disclaimers on Direction, Stability, and Timing

Three things the account does not say are entered below, each as a denial, because each is what a sympathetic reader is likely to supply on the account’s behalf.

It supplies no direction. A configuration leaving its present form arrives at another, and nothing in the mechanism determines which. Reorganization following the reordering of an unjust arrangement may install an arrangement that is worse, and the historical record contains many such sequences. The account describes a configuration losing its capacity to absorb disturbance and says nothing about what succeeds it. A reader who takes the mechanism to mean that unjust organizations are in due course replaced by better ones has added a premise the mechanism does not contain, and the added premise is false.

Stability carries no evaluative content. That a configuration persists is a fact about its capacity to absorb disturbance and is no evidence about whether it is defensible. A just organization may be stable, an unjust one may be stable, and each may be unstable. The point cuts against the present account as much as it cuts for it: an organization that has closed its questions may persist longer than one that has not, and its persistence would be no vindication, exactly as its collapse would be no verdict.

It supplies no timing. The mechanism states no rate. The interval between a closure and any reorganization is unspecified and may exceed the working lifetime of every party affected, of every officer responsible, and of the organization’s present form. A party wronged by a closure today is assured of nothing on any horizon that concerns them.

22.8 The Distance from Retribution

The mechanism resembles, in outline, a familiar moral picture in which wrongs are eventually repaid, and the resemblance is superficial. The differences are worth setting out, since the picture is attractive and would ruin the account.

Retribution requires an agent who repays, and the account contains no agent. It requires proportion between the wrong and what follows, and the account supplies none, since the magnitude of a reorganization is a function of the configuration’s condition and not of the gravity of any act. It requires that what follows fall upon the party responsible, and the account supplies no targeting at all: a reorganization falls upon whoever is present when it arrives, which will include parties who were themselves wronged by the closure and officers who inherited a practice they did not make. And it requires reliability, which the criticality condition entered above has already withdrawn.

A further caution belongs here, addressed to the use of the account and not to its content. A reader persuaded of the account might conclude that action against interpretive injustice is unnecessary, since the structure will in due course answer for it. That conclusion does not follow from anything asserted here, and it is contradicted by the criticality condition and by the disclaimer on direction entered above. The account describes what happens in the absence of action. It supplies no reason to withhold action, and it would be a poor use of a descriptive claim to convert it into a consolation.

22.9 The Standing of the Claim

The claim is a conjecture about a mechanism and is not a law. Whether organizations exhibit the behaviour near stability boundaries that is borrowed here from other domains is unsettled, and the borrowing is made on the ground that the structure of the argument is intelligible without it and gains precision from it. The proposition entered is that reorganizations of an interpretive practice are more likely where a configuration’s capacity to absorb disturbance has declined than where it has not, which is weaker than any claim about necessity and is what the argument requires.

Two limits attend the conjecture. No formal model is offered, and the abstention is deliberate for the reason given in the companion work: a model written to accompany an argument of this kind would supply an appearance of rigour the argument has not earned, and the mechanism admits of honest statement in prose. And no party can reliably determine how near a configuration stands to a boundary, from inside it or from outside. The account therefore yields no operational guidance of the form that would counsel an organization to remain at a safe distance, since the distance is not observable, and any practice it recommends must be one that does not require knowing where the boundary lies.

22.10 The Surviving Claim

After the denials something remains, and it is modest and usable.

The account supplies a description under which interpretive closure is a loss to the organization that performs it. On the ordinary description, closure is a cost imposed upon a party and a convenience for the body, and the body’s interest lies in maintaining it. On the description offered here, closure removes the organization’s own means of learning that a reading has ceased to fit, and the organization is the party that will operate for an indefinite period on readings it can no longer correct.

The significance of this is that it does not require a constituency. Every device surveyed in §19 works by supplying an external reason to interpret well, and each fails where no external party holds a stake. A reason of the present kind is internal: it is available to an organization that has no dependence on anybody and considers what closure does to its own capacity to remain fitted to the conditions it operates in. This is the one form of leverage the asymmetry boundary does not eliminate, and §23 enters the condition under which it too fails.

§23 The Limits of the Account

The account is now complete enough to be priced, and the price is high. This section states what has not been delivered.

No enforcement. The criterion identifies an injustice and equips nobody to act against it. Where the devices of §19 apply, they impose costs and secure remedies; the present account imposes nothing and secures nothing, and it does not supplement those devices at the point where they work.

No remedy for a present party. A party before whom a closure stands gains from this account a description of what was done to them and no route by which it might be undone. The description may be of some worth, and it is not a remedy, and the paper should not be read as offering one.

No protection in the case that motivated it. This is the sharpest of the limits. The account was developed because the surveyed instruments are bounded by a stake that the bodies of greatest concern may not hold. The account does not protect a party against such a body either. It describes the act and its consequences for the body, and description is what it delivers in place of protection.

Retrospective application. As §21 entered, closure is identifiable with confidence once time has passed and difficult to identify at the moment of decision, which is the moment at which identifying it would matter.

No operational form. The criterion turns on the capacity of a relation to generate a further interpretation, and no procedure is offered by which that capacity may be assessed in a given case. What would count as evidence that a question remains askable, and by whom the assessment would be made, are unsettled, and a criterion in that condition is a proposal for work and not an instrument.

The internal reason has a condition. The leverage identified in the subsection on what survives requires that an organization weigh its own capacity to remain fitted over a period long enough for unfitness to matter. Officers whose horizons are shorter than that period hold no such interest, and an organization whose decision-making is distributed among such officers may be unable to hold the interest even where every officer would acknowledge it in the abstract. The internal reason therefore reintroduces a version of the difficulty it was offered to escape, at a different level: a stake in the long-run condition of the body is still a stake, and it too can be absent.

Two things are claimed against this list. The first is that the criterion states, with a definite content, an injustice that the existing vocabulary of distributive, procedural, hermeneutical, and republican concepts does not isolate, and that the isolation is worth having independently of what can be done with it. The second is that the boundary of §19 is a result about the world and not a failure of this paper, and that an account which stops where the boundary stops is reporting a limit correctly, whereas an account that proposed a further device of the surveyed families would be evading it. Whether these are sufficient is for the reader.

§24 Open Questions

This paper identifies a neglected structure. It does not complete a theory of it, and the questions set down here are held to be the more valuable part of what is offered, since a structure worth naming is worth other people working on.

If the structure is real, the following questions arise immediately.

The first group concerns the authority to interpret.

Q1. Is interpretive standing delegable? An organization may confer upon a body outside itself the authority to determine what its own rules mean, and it is unclear whether such conferral escapes the recursion entered in §3 or relocates it, since the terms of the conferral are themselves rules, and their interpretation must be held somewhere.

Q2. What makes an exercise of interpretive authority legitimate? Candidate grounds include conferral by those subject to the rules, demonstrated competence in the practice from which the rule’s terms draw their content, participation by affected parties in the determination, and continual revisability of what is determined. These are not equivalent and may conflict, and no ordering among them is proposed here.

The second group concerns the modes and objects of interpretation.

Q3. Can the edge of the admissible be located before the fact? §3 holds that the space of admissible readings is bounded, and locating the boundary is straightforward once a reading has been shown incompatible with the network and difficult in advance of the showing. What would license a claim, made at the moment of decision, that a proposed reading lies outside the space?

Q4. Do evaluative predicates differ in kind from other open-textured terms, or only in the proportion of cases falling in the penumbra? If in kind, what marks the difference, and if in degree, whether there is a proportion beyond which a term ceases to function as a rule at all.

Q5. Does the recursion of §3 admit of degrees? An organization might insulate the rules governing challenge from the power those rules constrain, by fixing them beyond amendment, by committing their interpretation elsewhere, or by other means, and it is open whether any such arrangement reduces the recursion without transferring the whole matter outside the organization.

The third group concerns protection, and follows from the boundary of §19.

Q6. Under what conditions can a constituency be constituted? §19.6 identifies the practical route this paper does not take, which is to supply the stake the devices require by aggregating parties who individually hold none, or by attaching the interpretive practice to a condition held by a party the organization does depend upon. What determines whether such aggregation succeeds is a question for the study of collective action and of institutional design, and the answer bears directly on how much of the boundary is fixed.

Q7. How far does the examination exception generalize? The German decisions discussed in §19 reached a substantive evaluative judgment that comparable forums decline to review. What class of evaluative acts admits criteria articulable enough to support review of that kind, what determines membership of the class, and whether comparative judgments can ever belong to it, are open.

Q8. Can interpretive closure be identified prospectively? §21 applies with confidence in retrospect. An account that could identify, at the moment of decision, that a determination is of the closing kind would be of use to an organization acting in good faith, and nothing offered here supplies it.

The fourth group concerns the dynamics proposed in §22, where the questions are the sharpest because the claims are the weakest.

Q9. Does interpretive closure raise or lower a configuration’s capacity to absorb disturbance? §22.5 states the complication and declines to settle it: closure removes a corrective channel and removes the disturbances that channel carried. Which effect dominates, and whether the answer varies with the source of the disturbance, is unresolved and is the point at which the dynamical claim is most exposed.

Q10. Under what conditions does an interpretive closure register at all? If disturbances are consequential only near a stability boundary, then most closures produce nothing observable, and the account gives no way of telling which are of the other kind. Whether any observable signature accompanies the approach to such a boundary, in an organization as opposed to the physical systems from which the vocabulary is borrowed, is open and would be worth having.

Q11. What would count as evidence for or against the mechanism? The conjecture of §22 is stated in a form that resists testing, since it predicts long intervals of no effect and specifies no rate. A formulation that could be disconfirmed would be an improvement on what is offered, and supplying one is work this paper leaves undone.

The last question is the one the whole account is directed toward, and it is placed last because everything above bears on it.

Q12. Under what institutional, social, and relational conditions can an interpretive practice remain regenerative? The criterion of §21 states a requirement and says nothing about what meets it. An organization that wished its questions to remain askable would need to know what sustains askability across changes of personnel, across the settling of readings into usage, and across the ordinary pressure toward finality that decision-making generates. This is the question on which the practical value of the account depends, and it is entirely open.

A closing observation about the shape of the list. The questions of the first three groups could be pursued within existing disciplines, and several are being pursued there already under other descriptions. The questions of the fourth group could not, and they are the ones on which the account offered here stands or falls. A reader who finds the criterion of §21 useful and the mechanism of §22 unpersuasive may take the first and leave the second, since the criterion is stated independently and does not rest upon the dynamics. The reverse is not available: the mechanism is a claim about what follows from closure, and it requires the criterion to say what closure is.

---

A note on method. No formal apparatus has been used, and the abstention is deliberate. The mechanism of §22 could be written in the notation of dynamical systems, with a configuration answering to a point in a space of institutional states, a corrective channel to a feedback term, and a closure to the removal of one. Written out, such a model would derive nothing that the prose does not state, and it would supply an appearance of precision that the argument has not earned. The conditions under which the notation would become honest are set out in Q9 through Q11, and they have not been met here.

A note on sources. The instruments surveyed are revised on cycles and the doctrines are subject to overruling. Statements about the content of standards, statutes, and case law are given as of the date of writing and should be verified before being relied upon, and the recent displacement of a long-settled deference doctrine in one of the jurisdictions discussed is a reminder that the caution is not formal.

References

Ashby, W. Ross (1956). An Introduction to Cybernetics. London: Chapman & Hall.

Barak, Aharon (2005). Purposive Interpretation in Law. Trans. Sari Bashi. Princeton: Princeton University Press.

Bobbitt, Philip (1982). Constitutional Fate: Theory of the Constitution. New York: Oxford University Press.

Bobbitt, Philip (1991). Constitutional Interpretation. Oxford: Blackwell.

Brandom, Robert B. (1994). Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press.

Cappelen, Herman (2018). Fixing Language: An Essay on Conceptual Engineering. Oxford: Oxford University Press.

Cover, Robert M. (1983). Nomos and Narrative. Harvard Law Review, 97(1), 4–68.

Dworkin, Ronald (1986). Law’s Empire. Cambridge, MA: Harvard University Press.

Endicott, Timothy A. O. (2000). Vagueness in Law. Oxford: Oxford University Press.

Eskridge, William N., Jr. (1994). Dynamic Statutory Interpretation. Cambridge, MA: Harvard University Press.

Fine, Kit (1975). Vagueness, Truth and Logic. Synthese, 30(3–4), 265–300.

Fish, Stanley (1980). Is There a Text in This Class? The Authority of Interpretive Communities. Cambridge, MA: Harvard University Press.

Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press.

Frege, Gottlob (1997). The Frege Reader. Ed. Michael Beaney. Oxford: Blackwell.

Fuller, Lon L. (1958). Positivism and Fidelity to Law: A Reply to Professor Hart. Harvard Law Review, 71(4), 630–672.

Fuller, Lon L. (1969). The Morality of Law. Rev. ed. New Haven: Yale University Press.

Fuller, Lon L. (1978). The Forms and Limits of Adjudication. Harvard Law Review, 92(2), 353–409.

Gadamer, Hans-Georg (2004). Truth and Method. 2nd rev. ed. Trans. Joel Weinsheimer and Donald G. Marshall. London: Continuum. [German orig. 1960.]

Galanter, Marc (1974). Why the “Haves” Come Out Ahead: Speculations on the Limits of Legal Change. Law & Society Review, 9(1), 95–160.

Haslanger, Sally (2000). Gender and Race: (What) Are They? (What) Do We Want Them To Be? Noûs, 34(1), 31–55.

Hallaq, Wael B. (1984). Was the Gate of Ijtihad Closed? International Journal of Middle East Studies, 16(1), 3–41.

Hallaq, Wael B. (1997). A History of Islamic Legal Theories: An Introduction to Sunnī Uṣūl al-Fiqh. Cambridge: Cambridge University Press.

Hart, Henry M., Jr., and Albert M. Sacks (1994). The Legal Process: Basic Problems in the Making and Application of Law. Ed. William N. Eskridge, Jr., and Philip P. Frickey. Westbury, NY: Foundation Press. [Tentative ed. 1958.]

Hart, H. L. A. (1958). Positivism and the Separation of Law and Morals. Harvard Law Review, 71(4), 593–629.

Hart, H. L. A. (1994). The Concept of Law. 2nd ed. Oxford: Clarendon Press. [1st ed. 1961.]

Hildebrandt, Mireille (2015). Smart Technologies and the End(s) of Law: Novel Entanglements of Law and Technology. Cheltenham: Edward Elgar.

Hirschman, Albert O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press.

Huang, Wanhong (in preparation). Conversion Injustice: A Generative Relational Account of Value Conversion.

Huang, Wanhong (in preparation, b). Generative Relational History: Historical Generativity under Constraints.

Kelsen, Hans (1967). Pure Theory of Law. Trans. Max Knight. Berkeley: University of California Press. [German 2nd ed. 1960.]

Kripke, Saul A. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press.

Kripke, Saul A. (1982). Wittgenstein on Rules and Private Language. Cambridge, MA: Harvard University Press.

Llewellyn, Karl N. (1950). Remarks on the Theory of Appellate Decision and the Rules or Canons about How Statutes Are to Be Construed. Vanderbilt Law Review, 3(3), 395–406.

Luhmann, Niklas (1969). Legitimation durch Verfahren. Neuwied: Luchterhand.

Mashaw, Jerry L. (1985). Due Process in the Administrative State. New Haven: Yale University Press.

Masud, Muhammad Khalid, Brinkley Messick, and David S. Powers, eds. (1996). Islamic Legal Interpretation: Muftis and Their Fatwas. Cambridge, MA: Harvard University Press.

Medin, Douglas L., and Marguerite M. Schaffer (1978). Context Theory of Classification Learning. Psychological Review, 85(3), 207–238.

Murphy, Gregory L., and Douglas L. Medin (1985). The Role of Theories in Conceptual Coherence. Psychological Review, 92(3), 289–316.

OECD (2023). OECD Guidelines for Multinational Enterprises on Responsible Business Conduct. Paris: OECD Publishing.

Pettit, Philip (1997). Republicanism: A Theory of Freedom and Government. Oxford: Oxford University Press.

Putnam, Hilary (1975). The Meaning of “Meaning”. In Mind, Language and Reality: Philosophical Papers, Volume 2, 215–271. Cambridge: Cambridge University Press.

Raz, Joseph (1979). The Rule of Law and Its Virtue. In The Authority of Law: Essays on Law and Morality, 210–229. Oxford: Clarendon Press.

Rawls, John (1971). A Theory of Justice. Cambridge, MA: Harvard University Press.

Rosch, Eleanor (1978). Principles of Categorization. In Eleanor Rosch and Barbara B. Lloyd, eds., Cognition and Categorization, 27–48. Hillsdale, NJ: Erlbaum.

Scalia, Antonin (1997). A Matter of Interpretation: Federal Courts and the Law. Princeton: Princeton University Press.

Scalia, Antonin, and Bryan A. Garner (2012). Reading Law: The Interpretation of Legal Texts. St. Paul: Thomson/West.

Selznick, Philip (1969). Law, Society, and Industrial Justice. New York: Russell Sage Foundation.

Solum, Lawrence B. (1987). On the Indeterminacy Crisis: Critiquing Critical Dogma. University of Chicago Law Review, 54(2), 462–503.

Sunstein, Cass R., and Adrian Vermeule (2003). Interpretation and Institutions. Michigan Law Review, 101(4), 885–951.

Teubner, Gunther (2012). Constitutional Fragments: Societal Constitutionalism and Globalization. Oxford: Oxford University Press.

Tyler, Tom R. (1990). Why People Obey the Law. New Haven: Yale University Press.

United Nations (2011). Guiding Principles on Business and Human Rights: Implementing the United Nations “Protect, Respect and Remedy” Framework. HR/PUB/11/04. New York and Geneva: United Nations.

Waldron, Jeremy (2011). The Rule of Law and the Importance of Procedure. Nomos, 50, 3–31.

Williamson, Timothy (1994). Vagueness. London: Routledge.

Weiss, Bernard G. (1998). The Spirit of Islamic Law. Athens, GA: University of Georgia Press.

Wittgenstein, Ludwig (1953). Philosophical Investigations. Trans. G. E. M. Anscombe. Oxford: Blackwell.

中文

组织中的解释正义

关于诸问题与开放问题的一次初步讨论

黄万宏 · huangwanhong@serendip.ngo

---

摘要

组织凭规则治理，而规则不自我应用。在一条规则与一个案件之间，立着一个解释行为，由该组织的一个机关执行，它固定该规则对那个案件本应意味什么。本文以那个行为为它的对象。它论证解释行为是一个不义能在其上发生的独特场址，不可化约为一个组织所分配之物的不义、也不可化约为它据以决定之诸程序的不义。四个问题组织这一讨论：谁持有解释的资格、解释经由什么模式进行、哪些规则与哪些谓词承载解释的负担，以及什么保护一个解释被施加于其上的一方。本文概览法理学、立法起草、标准制定、行政复审与组织治理跨若干法律传统对这些问题所作之物，并发现一个为回答第四个问题而组装的每一个装置所共有的边界：每一个都预设那个进行解释的机构持有某种对承认的利害，无论是法律的、市场的、声誉的，还是专业的，而每一个都随那一利害趋近于零而趋近于无效。在对解释的权威与对承认的依赖由同一方持有之处，没有任何可得的工具够到那个解释行为。对着这一边界，本文放置一个生成性关系的读法，在其上一个解释由它对解释者与被解释一方之间关系生成一个进一步解释之能力所做之事来评估。不义，在那个读法上，是那一能力的熄灭，而对它的制裁被委托于那个结构本身的动态，在本文明确且详尽陈述的诸护栏之下。该说明供不出执行、也供不出对一个当前一方的补救，而本文以这留下未决的诸问题结束。

关键词： 解释；组织规则；正义；裁量；问责；不对称；关系性生成；治理。

关于本文定位的一则说明。 这是一次初步讨论。它辨认一个结构、概览现存理论与现存实践对它所作之物、标记每一个在何处用尽，并提议一个方向。它不完成一套理论，而结尾处所放下的诸问题被认为是所提供之物中更有价值的一半。这一概览是宽的，因为该结构在通常被分开研究的诸尺度上复现，而宽度是一个本文试图借§1所陈述的手法加以约束的负担：每一节都履行四个被命名之问题中的一个，而材料凭履行一个而挣得它的位置。在这一概览薄之处、在一个法域被读错之处，或在一个工具回答了本文声称它留下未决之困难之处，更正欢迎寄至 huangwanhong@serendip.ngo。

§1 组织规则解释的问题

一个凭规则治理的组织，必须在每一次决定的那一刻，判定它的诸规则就它面前的案件说了什么。规则是一般的而案件是特定的，而规则中没有任何东西界定它自己的应用。某人必须固定这个案件是否落于那条规则之下。那一固定是一个行为，由该组织的一个机关执行，而它正是本文所关切的那个行为。

那个行为易被忽略，因为它通常被以一套遮蔽它的语汇描述。一个组织被说成应用它的规则，而应用暗示一个技术操作，其中规则的内容先在地已定，而唯一剩下的工作是看诸事实是否匹配。在那一描述上，那个解释行为消失进它所应用的那条规则，而能被提起的诸问题只关乎规则与事实：那条规则是否是一条好规则、诸事实是否被正确地查明、查明它们的程序是否公平。这些中的每一个都是一个真实的问题，而每一个都有一个文献。它们无一够到那条规则的内容被固定于这个案件的那个行为。

以最平实的形式考虑这一结构。一个专业机构以申请人持有同其寻求准入之领域相关的经验为判据准入成员。一个申请人持有在一个相邻领域获得的大量经验。那一经验是否同这一领域相关，不由那条规则来了结，因为那条规则含有那个其外延处于攸关的词。它由该机构的一个机关在决定的过程中来了结。申请人可以争议对事实的查明，而没有事实在争议中。申请人可以争议程序的公平，而程序可能无可挑剔。申请人会争议的是那个解释，而恰恰在这里可得的诸渠道变薄。

它们变薄出于结构性的诸缘由。组织规则不是法律。一所大学的准入条例、一个专业机构的守则、一家公司的晋升判据、一个标准制定者的合格要求：这些凭该组织自己的权威而有约束力，而它们的解释在第一位上被交托给该组织自身。在该组织是一个公共机体之处，一个外部复审可能存在，尽管那复审通常够到程序与该行使的合法性，而将实质的评价性判断作为属于该机体自己的权限而谢绝。在该组织是私人的之处，即便那个外部复审也可能缺席。而没有任何法律系统强制一个组织确立一个凭之它的诸解释可被质疑的内部机制。法律能要求理由被给出；它鲜能要求那些理由是对的。

四个问题生起，而它们组织随之而来的一切。

1. 权威。 谁持有解释一个组织之规则的资格，而那一资格如何被获得？
2. 模式。 解释经由什么模式进行，而什么把一个有界的解释同一个任意的解释区分开来？
3. 对象。 哪些规则、以及它们术语中的哪些，承载解释的负担，而为什么那一负担被不均地分布？
4. 保护。 当一个解释对一方施行一桩不义时，什么保护那一方？

前三个已受到持续的处理，主要在法理学中，而那一处理以本文随行随入的诸限定转移到组织。第四个是那困难所集中之处，而它是这一概览所被导向的问题。

为回应而提供的说明可被陈述为单一的一个主张。

中心主张。 一条规则的解释是一个不义能在其上发生的独特场址，可同一个组织所分配之物的正义、以及同它据以决定之诸程序的正义分开。为保护一方对抗解释不义而发展出的诸装置，跨法理学、立法、标准制定、行政复审与组织治理无一例外，共享一个共同的预设：那个进行解释的机构持有一份对它自身之外某方之承认的利害。每一个装置都作用于那一利害，而每一个都随那一利害趋近于零而失去它的抓力。在对解释的权威与对承认的独立于一方之中重合之处，可得的诸工具够不到。一个生成性关系的读法提议一个解释由它对诸方之间关系生成一个进一步解释之能力的效果来评估，而那一不义在于那一能力的熄灭。在这一读法上，那一制裁不由一个外部权威施予，而出现在那个结构本身的动态之中，这是一个关于一个机制的描述，且不携带任何保证，即那个机制及时地、或根本地、或以任何人会欢迎的方向到来。

该主张分五个乐章展开。第一个铺陈那个解释行为的结构，并论证它对一条规则的内容是构成性的，而这正是使它根本成为一个正义场址之物。第二个评述法理学关于不确定性、关于诸解释学说，以及关于解释之正义所确立之物，并把这个概念对照它易于同之混淆的诸相邻不义概念定位。第三个评述解释被提议加以约束所经由的诸理论策略，以及每一个所遇到的困难。第四个评述在四个尺度上、跨若干法律传统的制度实践，并提取跨它们全部复现的诸失败模式，以不对称边界结束。第五个提供那个生成性关系的读法、陈述它所不交付之物，并放下开放的诸问题。

关于这一概览的纪律的一个词。此处所集结的材料取自通常被分开研究的诸领域，而一个这种宽度的概览招致它是一份目录这个责难。本文试图挣得这一宽度所经由的手法现在被陈述并被通篇执行：每一节在它开篇处命名它履行四个问题中的哪一个，而材料以它关乎那个问题为由被准入一节。§3以及随之而来关于解释之诸模式与诸对象的诸节履行权威、模式与对象。关于法理学与关于制度实践的诸节履行全部四个，以保护为主导。生成性关系诸节单独履行保护，因为它是那些较早的回答所留下未决的问题。关乎无一者的材料被留出，无论多么有趣，而那些遗漏在它们可能被想念之处被记录。

两则免责声明结束本节。本文不就任何特定组织是否错误地解释了任何特定规则采取立场，而那些例证案例不携带任何这样的指控。而此处所推进的概念有诸邻居，在诠释学不义之中、在程序不义之中，以及在诸支配理论之中，它必须小心地同它们区分开来；一节被给予那一任务，因为一个无法同它诸邻居分辨的概念未曾被引入。

§2 范围与方法

本节固定这一讨论的诸对象、陈述它们被追踪所跨的诸尺度，并引入所采用之比较方法的诸界限。

2.1 组织、规则、解释行为

以组织所指的，是任何凭它自己的一般规则治理行为、并在它们之下决定特定案件的常设机体：一个立法机关、一个监管机构、一个标准制定机体、一个法院、一所大学、一个专业协会、一家公司、一个国际机构、一个非政府协会。这一定义按设计是宽的，因为成问题的结构是由一般规则同特定决定的合取所生成，而那一合取对它发生于其中之机体的法律形式漠不关心。在法律形式要紧之处，而它常常要紧，它凭决定哪些外部装置可得而要紧；那是这一概览的一个发现、而非它的一个前提。

以规则所指的，是该组织标榜自己据以决定的任何一般标准，无论它被称为一部成文法、一项条例、一个判据、一部守则、一项政策、一个标准，还是一条指引。被发布的规则与内部的规则都被包括，而它们之间的差别被当作一个影响可质疑性的变量来处理。

以解释行为所指的，是在决定一个案件的过程中，对一条规则就那个案件要求什么的判定。那个行为在一条规则的术语在诸情形中容许多于一个外延时便在场，也就是说几乎总是在场，而它在术语是评价性的之处最可见。那个行为同三个通常与之合并的邻居区分开来：同事实查明，它判定发生了什么；同规则制定，它在任何案件之前了结那个一般标准；以及同狭义的裁量，其中一条规则明确地授予在诸被许可结果之间的一个选择。解释与规则制定共享它固定内容，与裁量共享它是一次判断的行使；它区别于第一个，在于发生在决定的那一刻并在应用的伪装之下，区别于第二个，在于执行它的机关通常把自己表示为受约束的。

2.2 比较的诸尺度

这一概览跨四个尺度追踪那个结构，在这些尺度上一般规则与特定决定的同一合取在不同的制度条件下复现。

立法是一条规则的解释被起草者预期的尺度，经由定义条款、目的条款、解释法，以及对被授予之权力的规约。标准制定是私人与跨国机体书写通过合格评定与市场准入而有约束力之规则的尺度，也是那个解释问题被谁书写了那个标准这一问题所复合的尺度。国家行政是一个官员的解释行为受制于一套发达的说明理由与复审装置的尺度。组织治理是唯有内部机制可得的尺度，也是第四个问题的困难最尖锐的尺度。第五个情境，即国际与跨国机体在其上没有一个主权者的情况下决定的情境，在它照亮第四个问题之处被处理，因为它是解释权威单由承认所规训的最清晰情形。

这四个是一个结构之诸实例这一主张是本文的一个主张，并借在每一个尺度展现同样的诸失败模式而被辩护。一个认为这些尺度不可类比的读者，将在从公共到私人的过渡处发现这一论证最弱，在那里外部复审的可得性在种类上、而非仅在程度上有别，而那一过渡在它发生之处被标记。

2.3 比较方法及其界限

若干法律传统被援用：美国与英国的普通法系统、德国、法国与瑞士的大陆系统、日本与韩国对大陆行政学说的东亚接受、欧盟的超国家秩序，以及私人标准制定的跨国秩序。伊斯兰法律理论因其解释权威的架构而被引入，那是任何传统中可得的对第一个问题最充分发展的回答。

三条界限附着。这一比较是学说的与制度的，而它不依据对所概览之诸装置在运作中如何表现的经验研究；在经验发现存在之处它们被引用，而它们的稀缺本身是一个结果。诸法域的选取由作者所读语言之学术可得性所支配，这偏向英语、日语与被翻译的大陆来源。而这种被概览的诸工具在周期上被修订，而诸学说被推翻；法律与诸标准的状态以写作之日期陈述，而读者应核实时效性，一则美国一个长期已定之尊让学说的近期被取代使之不止于形式的告诫。

2.4 从这一事业中的诸排除

本文不提供一个凭之一个给定解释可被分类为正义或不义的决定程序。它不提供一个它在其结尾诸节所援引之动态的模型，而那一克制的诸缘由在那里被陈述。它不在所概览之诸法律解释学说之间裁断，而把它们当作诸候选回答，其诸困难是旨趣的对象。而它不提议制度改革，因为它所抵达的发现是可得的诸改革作用于一份在最令人关切之诸情形中缺席的利害；提议更多它们会回避那个发现。

§3 解释时刻

本节履行第二个问题的第一部分，借确立那个解释行为是什么，并准备第四个问题，借确立为什么它是一个正义在其上处于攸关的场址。这一论证有三步：解释对一条规则就那个案件的内容是构成性的、它被执行所从出的位置是结构性地不对称的，以及解释的权威延伸到支配对它自己行使之挑战的诸规则。

3.1 那个最小结构

设一个组织持有一条规则 $R$，并设一个案件 $c$ 生起。该组织的一个机关 $A$ 判定 $R$ 是否覆盖 $c$。那一判定可被表达为一个命题，即 $R$ 覆盖 $c$ 或它不覆盖，而该组织据之进行。把这称为解释时刻。

那个时刻的三个特征值得在关于它任何东西被主张之前被标记。它是被引起的：它发生因为一个案件已生起，而它的引起由一方所供给，那一方通常对解释将于何时被要求没有控制。它是特定的：它就 $c$ 了结 $R$，而它是否就任何进一步的案件了结 $R$，取决于该组织是否把它自己的诸判定当作对它自己有约束力，而这是一个变量、而非一个常量。而它是被表示为受约束的：那个机关通常不宣告它已选择了一个意义，而宣告它已找到了一个。这最后一个特征是那个时刻之不可见性的来源，并将在后文承载分量。

3.2 解释作为构成性的

承载本文的主张是，那个解释时刻不报告那条规则已然具有的一个内容。它为那个案件构成那个内容。

该主张可从语言一侧与从关系一侧被论证，而两条路径都被需要。从语言一侧，那个论证是熟悉的，并将在关于不确定性的一节被长篇概览：一个一般术语不把它自己的外延携带到边界情形，而没有进一步的规则能在不自身要求解释的情况下供出那个外延，否则陷入无穷回退。此处所添加的是第二条路径。

从关系一侧，那个论证如下运行。一个概念不持有任何自足的边界。一个术语覆盖什么，由它在一个由其他术语、由先前诸用法、由那个网络被认为服务之诸目的，以及由那个术语被部署于其中之诸实践所构成的网络中的位置所固定。”相关”这个词单独考虑没有任何外延；它只对着一个诸学科的场域、一个相关性要求所服务的目的、一段相关性被发现或被否认之诸案件的历史，以及一种相关性是一个活概念的研究实践，才获得一个。这是一个关于概念一般的陈述，而它对规则的后果是直接的。判定 $R$ 覆盖 $c$，是把 $c$ 定位在 $R$ 之术语对着它取其外延的那个网络之内，而定位是一个由无论谁执行它者、从他们在那同一个网络中所占据之位置执行的操作。

两样东西随之而来，而第二样同第一样一样重要。

第一样是解释是一次权力的行使。为一个案件固定一个术语的外延，是判定其案件成问题之一方的规范位置，并借一个那条规则不曾执行的行为如此做。那个机关正在对那一方做某样东西，在一次权威之行使做某样东西的寻常意义上，而那一做不单由那条规则所许可，因为那条规则没有那个行为便够不到那个案件。

第二样是这次权力之行使是有界的，而那一有界性是确立了那个权力之同一个关系论题的一个后果。倘若一个术语的外延是自足的，解释便会是单纯的阅读、而不携带任何权力。倘若那个外延不受任何东西约束，解释便会是任意的，而整套规则的装置会是戏剧。那个关系论题不产出任何一个。因为一个术语从一个网络取其外延，任何被提议的外延都必须同那个网络相容，而一个不相容的提议能被显示为如此。判定一个音乐博士学位同一个外科教席相关，不仅仅是不受欢迎的；它是不可得的，因为它无法同那个术语从中汲取其内容的诸目的、诸用法与诸实践被持在一起。因而诸可容许解释的空间既非一个点、也非全部，而这正是第二个问题、关于解释之诸模式的问题，将被提出于其中的那个空间。

这第二个后果被在开篇陈述是要紧的，因为一个关系性的意义说明通常被听为许可任意性，而倘若它如此，它便会摧毁本文正试图引入的这个概念。一个在其上任何解释都同任何其他一样好的理论无法说一个解释曾是不义的，而它也无法就它自己的诸主张如此说。此处所推进的说明主张关系性构成是对解释之约束的来源，并是使一个解释根本可被批评之物。

3.3 解释位置的不对称

那个解释时刻被不对称地占据，而那一不对称有四个应被区分的组分，因为补救对其中一些存在、对另一些不存在。

第一个是位置性的。一方判定而另一方接收那一判定。这一不对称对规则支配的决定是构成性的，而它本身不可反对；一个无法判定它自己规则之应用的组织根本无法决定。

第二个是信息性的。那个机关通常知道先前诸判定的历史、那条规则被认为服务的诸目的，以及在考虑中的诸读法的范围。那一方通常知道文本。在先前诸判定未被发布之处，那一方无法知道被应用于他们的读法是否是被应用于他人的读法。

第三个是表示性的。那一判定被宣告为一个发现。一个希望争议它的一方因而必须争议被呈现为一个关于那条规则之发现之物，用该组织所控制的诸材料，而必须在没有一套那个反对被自然地表达于其中之语汇的情况下如此做。日常语言供出现成的术语用于反对一条规则、用于反对一个程序，而供出稀薄的术语用于反对一个解释。

第四个是递归的，而它是本文视作决定性的那个组分。支配一个解释是否以及如何可被挑战的诸规则本身是该组织的规则，而它们的解释在第一位上属于同一个组织。一个给定的投诉是否在一个内部上诉的范围之内、它是否提起一个被排除于复审之外的实质事项、它是否被及时提起、那个投诉者是否有资格：每一个都是一个解释问题，而每一个都由一个机关来回答，那个机关的回答判定那个原初解释是否将根本被审查。因而解释的权威延伸到它自己复审的诸条件。在那一延伸是完全的之处，那个解释权力闭合于它自身之上，而一个争议一个解释的一方处在向那些上诉规则之解释上诉的位置。

3.4 解释不义作为一个独特场址

在那个时刻的结构被铺陈之后，它配得上它自己一个名字的缘由可被陈述。这一论证借对该现象否则被吸收进的两个范畴的排除而进行。

它不是分配性的。一个组织所分配之物可能被正义地分配，而它据以如此做的那个解释可能是可反对的；而一个解释可能是不可反对的，而它所产生的那个分配是不可辩护的。这两者独立地变化，而这是一个概念是否已被化约为另一个的寻常检验。

它不是程序性的。一个程序可能是完整的、及时的、在构成上不偏不倚的，并在它所给予的听闻上一丝不苟，而在它末尾所抵达的那个解释可能仍是投诉的对象。程序保障规约通往那一判定的路径，并且，除极少例外，对那一判定保持沉默。这不是现存程序学说中一个偶然的空缺；它随程序之为程序而来，而这一概览将显示它在每一个被考察的法域中复现。

还剩这一可能性，即解释不义是支配的一个种类，或一般不受问责之权力之不义的一个种类。那一关系更近，而它在诸相邻概念被处理之处被接手。临时的陈述是，支配命名对任意干预的常在易受性，而不本身把经由解释的干预同以其他手段的干预区分开来，而当下这个概念的旨趣恰恰在于那个机制，因为解释是把自己呈现为干预之缺席的那种干预模式。

在那个时刻的结构被确立之后，支配本文其余部分的那个问题可被以它最锐利的形式提出。鉴于一条规则的内容在决定的那一刻、由一个占据一个不对称位置、其权威延伸到它自己复审之诸条件的机关所构成，什么保护那个构成被施行于其上的一方？随之而来的诸节组装为回答而被提供之物，而它们所收敛于其上的回答，比所组装的装置会暗示的更狭窄。

---

关于草稿状态的一则说明。 论证的脊柱与理论概览已起草，而制度实践的概览正在进行中，故某些节在此出现在另一些将在最终次序中先于它们的节之前，而那个编号通篇预期那个次序。关于立法中、标准制定中、行政复审中、组织治理中，以及国际机体中解释控制的诸节，连同对抗托词性解释的诸证据装置与那些复现的失败模式，部分地待定。关于不对称边界的一节被起草在它们之前，因为它陈述那些节所被导向的结果，而首先起草它固定那个概览据以被写作的目的地。

§4 解释的权威

本节履行第一个问题。它铺陈解释的资格所从中被汲取的诸来源、陈述此处所给说明同关于价值转换之伴随工作所取立场之间的关系、把形式权威同有效权威区分开来，并标记谁持有质疑之资格这个被忽视的问题。

4.1 资格的诸来源

解释一个组织之规则的资格从三个来源被汲取，而大多数实际安排把它们结合起来。

授予是那个寻常来源。一个构成性文件、一部成文法、一次授权，或一项决议命名一个机关并把在诸规则之下诸事项的判定交托给它。那个机关解释因为它被任命去解释，而它的诸判定有约束力因为那一任命。

实践是第二个。一个已判定某一类诸事项、其诸判定已被据以行动并被默许的机关，通过那一行动与那一默许获得资格。组织中许多解释权威停留在这里，因为一个机体之规则被施行所经由的诸安排通常发展得比描述它们的诸文件更快。

执行是第三个，也是关于价值转换之伴随工作当作首要的来源。执行一个固定行为的一方，通过执行它，获得那个事项随后被讨论所用的语汇。无论谁书写一个等价，都供出那个交换的诸条款；无论谁固定一个外延，都供出那条规则此后被理解所用的诸术语，因为那一判定作为一个用法进入那个网络，而下一个案件是对着一个现在含有它的场域被决定的。

4.2 同关于转换之立场的关系

那篇伴随论文主张谁拥有解释之权利这个问题是错误的问题，理由是立法资格由跨越范畴的行动所生成、而非事先被持有（黄，编写中）。当下的主题要求那一立场的一个限定，而那一限定值得陈述，因为它定位某样那个转换情形不显示之物。

在一个组织中，那个寻常情形供出一个被指定的机关，故资格以一种那个转换情形不展现的方式是先在的。一个准入机体的评估小组在任何申请到达之前便持有它的权威，而无论它是否行使得好都持有它。到那一程度，谁拥有那个解释权利这个问题有一个答案，而那个答案由那一授予给出。

那一限定是，授予固定持有者而不固定触及。什么可由那个机关判定、它的诸判定延伸多远、一个问题是属于它还是属于另一个机体，以及它可考虑什么，都由规则来了结，而那些规则以同其应用处于攸关之诸规则相同的方式要求解释。在一个授予的中心，答案是平实的，而看到它不需要任何解释行为。在边缘处，答案由一个解释行为供出，而执行那个行为的机关通常正是其触及处于攸关的那个机关。

因而权威有它自己的开放结构，在哈特（1994）为一般规则所给的意义上。那篇转换论文的论题在边缘成立，在那里资格由主张它的行动所生成，而那个授予论题在核心成立。这两个立场在核心与边缘的区分被划出之后是相容的，而它们竟必须被调和这一事实，是§3中所引入之递归的一个后果，因为它是同一个特征在一个不同描述之下：一个判定支配它自己行使之诸规则之范围的权力，是一个其触及部分地由它自己来固定的权力。

4.3 形式权威与有效权威

一个组织的规则命名那个决定的机关，而那个解释行为通常被分布于比那个被命名者更多的位置。

考虑一个案件抵达一个决定所经由的序列。一个官员接收它并判定它是否完整，这要求对界定完整性之规则的一个读解。第二个官员对着门槛判据筛查它，而一个未通过筛查的案件从不抵达那个形式上决定的机体。一个内部指引的起草者已先前固定了一个术语要如何被理解，而那个指引塑造那个筛查者所做之事。那个形式上决定的机体接收一个已被刻画的案件，而它的判定是对着更早的诸解释行为所组装之诸材料而作出的。

随之而来的区分是在形式解释权威，由规则所命名之机关持有，与有效解释权威，由无论谁固定结果实际所转于其上之诸读解者持有，之间。这两者在某些安排中重合、在许多安排中分歧。在它们分歧之处，三个后果随之而来，将在关于实践的诸节中要紧。

第一个是针对形式所在的诸装置够不到那个有效所在。一项对一个决定委员会有约束力的说明理由义务对那个筛查官员什么也没说，而一项对着那个委员会决定的上诉权，对一个其案件在那个委员会看到它之前便被处置的一方毫无用处。

第二个是那个有效所在通常是较不可见的。指引是内部的、筛查是无记录的，而对一个案件的刻画不被任何人体验为一个解释行为。那一方被告知那个结果与抵达它的那个机关。

第三个是那一分散使责任即便在该组织内部也难以定位。序列中没有一个官员执行了那个解释的全部，每一个执行了一个本身合乎情理的部分，而由此产生的那个判定被归于那个安排比归于它们中任何一个更容易。

4.4 质疑的资格

一个通常比解释之权威受到较少注意的问题是提起一个解释问题的权威。它在此被处理，因为§21将转于它之上。

质疑的资格通常比受影响诸方的类别更狭窄。案件的一方可以投诉。一个处在相似位置、其案件尚未生起的一方不可以。规则所支配之类别的一个成员，能预见那个读法将关涉他们，通常不可以。一个通过那个读法对别人的运作而受它影响的一方不可以。而一个已离开、其同该组织之关系随那一判定而结束的一方通常不可以，尽管那个读法在他们之后存续并支配随后而来者。

那一狭窄有一个结构性后果。在争议者的类别被限于当前在该组织面前的诸方之处，且在那些诸方在判定之际离开之处，一个读法可能无限地支配而无人被定位去提起它。这是§21将辨认为结构性熄灭的那个状况，而它由资格规则的寻常起草所产生，且不要求任何封闭的行为。

4.5 诸发现与残余

本节确立资格在一个授予的核心是先在的、在它的边缘被生成，有效权威分散于规则不命名的诸位置，而质疑的资格比受影响的类别更狭窄。它留下如下未决：解释资格是否可被授予到该组织之外而不重新安置那个递归，以及什么使一次解释权威的行使成为正当的。两者都被引入§24。

§5 解释的诸模式

本节履行第二个问题。它铺陈解释经由之进行的诸模式、辨认每一个为可得所要求之物、陈述从它们的复数性生起的困难，并提议§3所许诺的一个有界读法同一个任意读法之间的区分。

5.1 诸模式及其材料

五个模式跨法律与组织实践复现，而它们在此以一种略微偏离通常诸分类法的方式被呈现，即为每一个辨认它从中汲取的那个网络区域。

文本解释汲取那些术语的已定用法，取一个有能力的说者会给它们的外延。目的解释汲取那条规则被认为服务的诸目的，问哪个读法推进它们。系统解释汲取该组织诸规则的其余语料，问哪个读法同别处所提供者相融贯。先例解释汲取先前诸判定的记录，问那个术语此前是如何被取的。类比解释汲取手头案件同已被决定诸案件之间的结构相似，问那里支配的诸特征在这里是否成立。

以这种方式被呈现，诸模式不再是一份相互竞争之诸学派的列表，而成为通往同一个对象的一组路线。§3主张一个术语的外延被固定在一个由诸用法、诸目的、诸融贯关系与先前诸应用构成的网络之内。每一个模式动员那个网络的一个区域。它们的复数性随那个网络的组成而来，而不是理论整洁的一个失败。

5.2 可得性的诸条件

一个模式只在它所汲取之诸材料存在、并对无论谁会使用它者可及之处可得。

目的解释要求一个可发现的目的。在一条规则未经有记录的审议而被采纳之处、在它的目的被争议之处，或在它已活过产生它之诸情形之处，那个模式供不出任何东西，或供出无论那个解释者认为那个目的曾是什么，而这是关于诸理论策略的一节将施压的困难。

先例解释要求一个记录。在诸判定未被发布、未被编制索引，或未被保存之处，没有任何一方能援引它们，而那个机关本身可能无法查阅它们。组织大量作出的那种个别化决定，通常处在这个状况。

系统解释要求一个足够广度与一致性以产出诸融贯关系的语料。一个由少数几条在不同时间为不同缘由被采纳之条款所支配的机体供出很少。

类比解释既要求一个记录、也要求一个关于先前诸案件之哪些特征在做工的被表述说明。

对组织的后果是可观的，并在此被陈述，因为关于实践的概览将回到它。法律系统在这些模式所要求的诸材料上投入甚重：报告、记录、起草史、被发布的理由，以及被编纂的语料。组织通常没有这些中的任何一个。对一个缺乏记录、有记录之目的与语料的组织仍可得的，是文本解释与未被陈述的判断，而这是汲取最少的两个模式，也是一个外部一方最难核查的两个。

5.3 复数性的困难

在若干模式可得之处，一个进一步的问题生起，而诸模式无法在它们之间回答它：在它们分歧之处哪个模式支配。

卢埃林（1950）为诸构造准则供出这一证明的最经济形式，那些准则成对地对立而来。那一证明一般化到诸模式。一个由已定用法所支持的读法可能被一个由目的所支持的读法反对；一个由先前诸判定所支持的读法可能被一个由同自那以后所采纳之规则的融贯所支持的读法反对。一个寻求一个给定结果的机关通常将找到一个许可它的模式，并将能够真诚地陈述它所用的那个模式。

那个自然的回应是一个把诸模式排序的二阶规则，而那个回应遇到§8的回退：一个把诸模式排序的规则是一条规则，它对一个案件的应用要求解释，而诸模式正是解释所经由之物。那个回退在实践中不是恶性的，因为诸实践确实安定诸排序而诸机关确实遵循它们，而它在理论上不可停止，这是§8以一条不同路线所抵达的同一个结果。

5.4 有界者与任意者

§3主张意义的关系性构成约束解释、而不许可任意性，并把那个区分推延到本节。所提议的区分刻意地弱，而它的弱正是要点。

有界性。 一个读法是有界的，当它能被展现为从固定那个术语之外延的网络的一个被辨认区域中汲取，以一种一个在该实践中有能力的一方能追随并检验的形式。一个读法是任意的，当没有这样的展现可得，无论因为没有一个被提供、因为所提供者不支持那个读法，还是因为那个读法汲取那个网络所不含有的任何东西。

这一提议的三个特征应被标记。

它不在诸可容许读法之间选择。两个读法可能各被从那个网络的不同区域展现，而那个判据宣告两者都有界并对哪一个更好什么也没说。这是凯尔森的框架以可操作的术语被重述，而那一局限被知情地继承。

它是一个展现的判据，因而取决于理由被给出。一个读法是有界的，当那个展现可得，而那个展现可得，当那个机关陈述它汲取了什么以及如何。在理由缺席之处，或在它们以一种报告结论而不报告路线的形式被给出之处，那个区分无法从外部被划出，而一个有界的读法在外部上同一个任意的读法不可区分。§19在诸装置中所概览的说明理由义务，在此正做分析工作，而不仅仅是一项程序性的客套：它是一个解释的有界性根本变得可评估所处的条件。

它留下那个前瞻的问题开放。一个读法能否被展现是在那个展现被尝试之后可回答的，而一个被提议的读法是否躺在那个网络之外，只有借臆测才在事前可回答。这作为 Q3 被引入§24。

§6 解释的诸对象

本节履行第三个问题。它按诸规则的术语所承载的解释负担来分类它们、辨认一类其诸判定从任何单一一方的位置不可复审的术语，并解释为什么解释负担集中在组织分配稀缺好处的诸点。

6.1 按负担的一个分类

诸术语可按它们的应用要求多少解释工作来排列。

确定的术语界定日期、数量、被命名的实体与被列举的范畴。它们的应用在寻常情形中是机械的，而它们的解释负担近乎零。由这样的术语构成的规则可无需判断被施行，这正是组织在它们希望移除裁量时把术语移入这一类的缘由。

分类的术语界定诸种类：雇员、居民、文件、受抚养人、出版物。它们的核心是已定的而它们的边界是真正开放的，而这是哈特（1994）所描述的那个状况。它们的负担是中等的，并集中在少数案件中。

标准的术语界定一个要被满足的水平：合理、充分、足够、令人满意、重大、相关。它们的应用在每一次使用的场合都要求一次评价，而它们的负担在每一个案件中都高、而不仅在一个边界处。恩迪科特（2000）把这样的术语的模糊当作不可消除的、并当作服务于精确会挫败之诸目的的，这正是它们不仅仅是要被更正的起草失败的缘由。

比较的术语界定一个对其他案件的关系：最好、最强、最合适、最应得、最高优异。它们的负担是最高的，而它们的结构在一个方式上区别于其他，下一小节接手这一点。

6.2 比较的术语

一个比较术语在它的应用要求什么上区别于一个标准术语。判定一个案件满足一个标准要求那个案件与那个标准。判定一个案件是最强的，要求那个案件、那个标准，以及那个比较类中的每一个其他案件。

三个后果随之而来，而它们合起来解释后面诸节所发现之物的一大部分。

第一个是一个比较术语之下的一个判定无法被一个仅能接触他们自己案件的一方所核查。会显示那个判定正确或不正确的材料是那些其他案件，而那些通常是保密的、通常是众多的，而通常对任何在那个决定机体之外的人不可得。一个被告知他人更强的一方，拥有一个他们无手段考察的命题。

第二个是一个比较判定在一个比一个标准判定更强的意义上不可复审。一个复审论坛可以评估一个机体是否把一个可理解的标准应用于一个案件。评估一个机体是否正确地把一个案件对着一个场域排序，会要求那个论坛重做那一操作，而关于实践之诸节中所概览的每一个论坛都谢绝去做，并出于健全的诸缘由谢绝。

第三个随前两个而来。一个比较根据，一旦可得，便吸收对一个以其他根据作出之判定的任何挑战。倘若一个标准术语之下的一个判定被争议，而一个比较根据在同一事实上可得，那个机体可以停靠在那个比较根据上而那个挑战便结束，因为那个比较根据不可被考察。这是§19中被辨认为诸保护装置失去它们抓力所处诸条件之一的理由可替换性的分析基础，而它解释为什么那个条件是结构性的、而非一个坏信念的事。一个从一个相关性根据转到一个优异根据的机体，已从一个一方可能争议的命题移到一个没有一方能争议的命题，而它无需以那一转移意图任何东西。

6.3 负担的集中

解释负担被不均地分布于一个组织的诸规则，而那一分布不是随机的。

支配内部程序的规则、附着于身份的诸资格，以及可预测性被该组织自己看重之诸事项，倾向于由确定的与分类的术语构成。支配稀缺好处之分配的规则倾向于由标准的与比较的术语构成，因为该组织要求这样的术语所保存的那一余地，也因为事先界定什么将把成功的案件区分开来通常是不可能的。

那个后果是解释负担恰恰在受影响一方的利害最高之处、在受影响诸方的数目最大之处，以及在那些诸方最不可能持有该组织所要求之任何东西之处最高。准入、任命、晋升、拨款、认证，以及裁量性好处的授予共享这一轮廓。§19中所辨认的、诸保护装置在其下失败的诸条件，因而集中在同那些最需要保护的解释行为相同的区域，而那一巧合由那个分配问题的结构所产生。

6.4 被发布的判据

一个部分补救属于本节，因为关于实践的诸节将以制度形式反复遇到它。

一个组织可以发布判据，界定它的标准与比较术语要如何被理解。发布把本会是一个解释行为之物转换为一个较低一般性的规则，而它供出§5发现通常缺席的先例与系统模式的诸材料。它使诸判定可对着某样东西被核查，并暴露诸背离。

两条界限附着于它。那些判据本身由术语构成，而它们的应用转而要求解释，故发布重新安置那个解释问题而不消解它。而发布对该组织在恰恰那个余地为之有价值的方面是有代价的，故分配稀缺好处的诸机体有一个常在的理由，把它们的判据陈述在一个保存那些标准与比较术语所曾保存者的一般性层次。关于实践之诸节中所概览的诸装置包括若干要求发布者，而那里的发现将是那个要求是可满足的、而那一余地不被交出。

6.5 诸发现与残余

本节确立一个按解释负担的分类、把比较术语孤立为一类其诸判定抗拒被任何单一一方考察者，并把负担的集中定位在稀缺好处的分配中。它留下如下未决：评价性谓词是否在种类上、还是仅在落于半影中之案件的比例上，区别于其他开放结构的术语，而这作为 Q4 被引入§24。

§7 概念应用的认识论

本文前面所作的两个主张依据关于一个一般术语如何获得它外延的诸假设。§3主张一个解释行为为一个案件构成一条规则的内容，理由是一个概念不持有任何自足的边界。§5主张同一个关系性构成给解释划界并排除任意性。两者都是概念理论中的主张，而它们在那里都未被论证。本节供出那个根据。

有第二个引入这一材料的缘由。本文后面所概览的诸装置嵌入认识论立场而不陈述它们。一个定义条款预设一个概念能由诸条件给出。一个被发布的判据预设分类能在诸案件之前被界定。一个要求一位考官接受任何可辩护答案的要求，预设一个术语同时容许若干正确的应用。一个在这些装置之间选择的机构正在诸概念理论之间选择，而那个选择可见是更好的。

7.1 定义论的看法

那个最古老的说明主张一个概念由对它应用个别必要、联合充分的诸条件给出，而对一个实例分类在于判定那些条件是否成立。弗雷格（1997）给这一立场它最锐利的现代形式，要求一个恰当形成的概念有确定的边界，从而对每一个对象，那个概念是否适用都是已定的。

这一看法是一大批法律起草的隐含认识论。一个定义条款陈述诸条件；一个附表列举诸实例；一个数值门槛替换一个评价性术语。§6把这些当作减少解释负担的诸装置来处理，而它们的理据是定义论的：一个如此被界定的概念可通过判定那些被陈述的条件是否成立来应用。

那个困难是，很少有旨趣的概念容许这样的界定。为寻常术语陈述必要与充分条件的诸尝试一般已失败，而在组织规则中承载解释负担的那些术语，在§6中被当作标准的与比较的，显著地抗拒。因而那个定义论看法描述一个起草有时能创造、而一般不能预设的状况。

7.2 家族相似与模糊的边缘

维特根斯坦（1953）推进那个反立场。落于一个一般术语之下的诸实例可能由一个交叠且交错之相似的网络所关联，而没有一个为全部所共同的特征，而那个术语尽管如此处于良好状态。对这样一个概念有模糊边缘这个反对，那个答复是一个模糊的概念仍是一个概念，而对许多目的一个模糊的正是所想要的。

对本文的后果值得小心地陈述，因为它切割一个自然的假设。一条规则中一个评价性术语的模糊通常被当作一个要在资源容许之处被更正的起草失败。在维特根斯坦式的看法上，它可能是那个概念的恰当形式，而对锐利边界的要求是一个哲学承诺、而非一个中性的好起草标准。恩迪科特（2000）为法律直接发展这一点，主张模糊是不可消除的并服务于精确会挫败之诸目的。

7.3 原型、样例与分级成员资格

关于人类分类的经验工作供出第三个说明，并关乎一个概念事实上如何被辨认的问题。

罗施（1978）确立范畴成员资格是分级的。受试者把某些实例排为一个范畴的更好例子而非另一些，一贯地如此做，并把典型实例分类得比非典型的更快。范畴围绕原型被组织，而成员资格从它们渐次淡出。梅丁与谢弗（1978）推进一个变体，在其上分类通过对被记住诸实例的相似而进行，不要求任何被抽象的原型，从而一个人所应用的诸范畴是他们所遭遇诸实例的一个函数。

两个后果关乎那个解释时刻。

那个半影不是例外的。哈特（1994）呈现一个带一个开放边界的已定核心，而那个经验图景是通篇分级的成员资格，一端有清楚的案件而没有核心与半影意象所暗示那种边界。§6从规则起草一侧为评价性术语抵达同一个结论，而这两条线收敛。

有能力诸解释者之间的分歧被预测。在一个样例说明上，一个人把什么分类为相关经验，取决于他们在那个描述之下所遭遇诸实例的范围。两个同等能力与好信念的解释者，遭遇了不同的范围，将不同地划出那个边界，而无一将能通过考察那条规则侦测那一差别的来源。这是一个要求没有坏信念、没有疏忽,也没有关于诸事实之分歧的解释分歧机制，而它值得拥有，因为对分歧的诸替代解释全都把一个过错归于某人。

7.4 分类中的背景理论

墨菲与梅丁（1985）论证单靠相似无法说明诸范畴的融贯，而诸概念被嵌入关于什么与什么相配以及为什么的背景理论之中。因而分类被那个分类者带来的一个解释性框架所塑造。

对本文的后果是一个第二的、独立的分歧机制。两个遭遇了同样诸实例的解释者可能不同地分类，因为他们持有关于那个范畴为何之物的不同说明。一个把一个相关性要求当作服务于标准之保护的解释者，将不同于一个把它当作服务于能力之辨认的解释者划出它的边界，而那一差别将对每一个把自己呈现为一个直截的读法。这把那个认识论连接到§9的学说材料，在那里目的的归属恰恰是这样的背景说明进入、而鲜少被弄得明确的那个点。

7.5 语言劳动的分工

普特南（1975）论证一个自然种类术语的外延由世界与由一个语言共同体的实践所固定，而非由一个个别说者的任何状态。寻常说者使用他们自己无法判定其外延的术语，顺从共同体中那些能判定的人，而普特南把这命名为语言劳动的分工。克里普克（1980）供出伴随的说明，即指称由一次初始命名所固定并通过一条使用的因果链被传递。

两个后果随之而来，而它们朝相反的方向拉。

第一个支持§10中所描述的一个安排。倘若外延由那些有能力判定它者所固定，那么一个奠基于被证明之能力的解释权威便有一个认识论担保、而不仅仅是一个专业特权。那里所概览的传统，即把解释的资格定位在一个可评估的资格中，正在回答关于诸术语之外延如何被安定的一个真实特征。

第二个是一个贯穿本文的警告。倘若外延由那些算作有能力者所固定，那么谁算数的判定便是诸术语意味什么的一个判定。无论谁被承认为那个专家，都为每一个别人固定那个外延，而一个在被承认之能力的圈子之外的一方，即便在原则上也没有质疑一个外延的资格。语言劳动的分工是一个关于意义如何被安定的认识论说明，而它同时是一个权威关系的描述。

7.6 模糊与半影的结构

对模糊的哲学处理直接关乎§8与§9中所记录的争论，而那个对应足够接近以被铺陈。

认识主义主张一个模糊术语有一个没有人处在能知道之位置的锐利边界，从而边界陈述有保持不可及的确定真值（威廉森 1994）。超赋值主义主张一个模糊术语容许许多可容许的精确化，为全部所真者为真，为全部所假者为假，而在它们之间变化者缺乏一个真值（法恩 1975）。程度理论给边界情形指派中间值。

放在那些法律立场旁边，那个对齐是显著的。那个认识主义结构是那个正确答案论题的结构：有一个答案，它不可由诸方能操作的任何程序所及，而裁量的表象反映知识的界限。那个超赋值主义结构是凯尔森框架的结构：某些读法被排除因为它们在每一个可容许的精确化上失败，某些被要求因为它们在全部上成立，而在其余者之间没有任何东西了结那个事。

关于模糊的立场	结构	法律对应物
认识主义	锐利边界，对知识不可及	正确答案论题
超赋值主义	可容许的精确化；真是在全部之上的真	那个框架，带一个未定的内部
程度理论	成员资格容许中间值	复审强度的分级

那个对应是作为一个帮助看到什么处于攸关之物、而非作为一个论证任一法律立场被一个关于模糊之立场所蕴含而被提供的。它确实确立§9中所记录的分歧是语言哲学中一个分歧的一个版本，而两侧都不是法律所特有的。

那个超赋值主义读法有一个进一步的用处，因为它给§5中所提议的有界性一个精确的意义。一个读法在它于诸用法、诸目的与先前诸应用的网络所使可得之每一个精确化上失败时是不可容许的。一个读法在它于它们全部上成立时是被要求的。在这些之间那个判据是沉默的，这是§5知情地引入的那个局限。

7.7 解释作为概念工程

一个近期的文献问我们的诸概念应当是什么，把诸概念当作对它们所服务之诸目的开放于刻意修订的工具（哈斯兰格 2000；卡佩伦 2018）。那一探究通过问一个概念做什么工作、它是否把那工作做得好，以及什么修订会改进它而进行。

对本文的关涉是直接的、并且似乎未曾被引出。一个判定一条规则的术语覆盖一个案件的机关，正在做那个概念工程师刻意所做之事：它就那个概念此后应覆盖什么、鉴于那条规则被认为服务的诸目的加以了结，而它的判定作为一个可得以支配后来诸案件的用法进入那个网络。那个差别在于承认。那个概念工程师宣告一个修订被提议并把它辩护为一个修订。那个进行解释的机关宣告一个发现。

两样东西随之而来。§3中所给的解释行为描述从一个独立于那里所提供之诸论证的方向获得支持。而那节中所引入的表示性不对称获得一个更锐利的形式：被向那个受影响一方呈现为对一个现存内容之发现之物，在这一描述上，是一次固定一个共享概念将意味什么之权力的行使，在没有一套这样的行使通常被辩护于其中之语汇的情况下被进行。

7.8 诸发现与残余

此处所集结的材料支持§3与§5所留下未论证的诸主张，而它从若干独立的方向如此做。承载解释负担那种类的诸概念不容许定义论的界定。它们中的成员资格是分级的，而那个分级结构是一个关于人类分类的发现、而非起草的一个缺陷。外延被安定在一个共同体之内、并由那些共同体承认为有能力者。同等能力与好信念之诸解释者之间的分歧被两个独立机制所预测，即不同的样例集与关于那个范畴为何之物的不同背景说明。而诸可容许读法的空间被那些可得的精确化所许可者所界定，那既非一个点、也非全部。

三个后果被向前携带。预设那个定义论看法的诸制度装置正在逆着分类如何进行的纹理工作，这正是被发布的判据重新安置那个解释问题而不消解它的缘由。一个机关的读法同一方的读法之间的分歧是那个被预期的情形，并供出无一者已犯错的证据。而谁算作有能力去固定一个外延的判定本身是诸规则意味什么的一个判定，这把那个认识论归还给§4中所处理的权威问题。

那个材料所留下未决的，是分类的分级结构能否被弄得为复审之目的可操作。一个应用§5之判据的复审机体会需要确立那个网络使哪些精确化可得，而此处不提供、也未在所概览之文献别处定位任何为此的程序。

§8 不确定性与遵循规则的界限

本节从关于遵循规则之哲学与法理学文献的一侧回到第二个问题，并借显示每一个说明在何处把一方之保护留下未加处理而准备第四个问题。它的结果是狭窄的与承重的：那个传统已确立规则欠决定它们自己的应用，并已确立那一欠决定是部分的。它所未曾确立的，是任何凭之一方可显示一个在那一欠决定之空间内作出的判定被错误地作出的程序。

8.1 规则的回退

那个困难的最深陈述是维特根斯坦（1953）的。一条规则不含有它自己的应用，因为任何行动进程都能在它的某个读法之下被带入同那条规则的一致，而任何读法都能在另一个读法之下被带入同它的冲突。供出一条第二规则来支配第一条的读解不阻止那个困难，因为那第二条规则转而要求读解。那个回退不是拙劣起草之规则的一个缺陷，而是一个一般表述同一个特定案件之间关系的一个特征。

克里普克（1982）引出那个怀疑论的后果，然后供出他所称的一个怀疑论解决：阻止那个回退之物不是关于那条规则的一个事实，而是一个实践，由一个其判断中之一致供出一个使用算作正确所处诸条件的共同体所维系。那个解决已被长篇争议，而它其中在此要紧的特征经受住那场争议。无论什么阻止那个回退，都外在于那条规则、并由那条规则被使用于其中的一个实践所构成。

这个结果是§3中所引入之关系论证的语言侧对应物。两条路径抵达同一个地方。一条规则之术语的外延被那条规则之外的某样东西所固定，而那样东西有一个由诸用法、诸目的与诸实践构成的网络之形式。这两条路径在它们所强调之物上有别。维特根斯坦式的路线强调那个阻止的实践是共同的、而判断中的一致是它的根据。那个关系路线强调一条规则被应用于其上的一方，占据一个在那条规则从中取其内容的同一个网络之内的位置，而可能占据它而对那个网络所供出之物没有任何影响的能力。第二个强调是正义问题进入之处，而它是那个传统所留下相对不发达的那个强调。

8.2 那个框架与意志的行为

那个困难在法律理论中的最精确陈述是凯尔森（1967）的。一个较高规范相对于它的应用而立，一如一个框架相对于可被放在其中之物而立。那个框架排除：某些应用在它之外并因此是不合法的。在那个框架之内，那个规范容许若干应用，全都同等合法，而那个规范中没有任何东西在它们之间选择。那一选择是由那个应用机关执行的一个意志的行为，而法律科学，它的本分是认知，能陈述那个框架的诸边界，而对它之内哪一个应用是正确的什么也说不了。

两个后果随之而来，本文采纳它们。第一个是诸可容许解释的空间既非一个点、也非全部，这是从关系一侧在§3中所抵达的同一个结论，而凯尔森以比任何后来的作者都更大的精确陈述它。第二个是那个应用机关在那个框架之内是诸规范的一个制作者，而把自己表示为它们的一个发现者。凯尔森陈述那个表示并把它当作一个供法律科学更正的事项；此处的关切是那个表示对一个会反对之一方所做之事，这是§3中所引入之不对称的第三个组分。

凯尔森的说明也标记一个规范理论所能供出之物的界限。倘若若干应用同等合法，一个其中之一被对着他们选择的一方，在凯尔森自己的术语上，没有法律上的投诉根据，因为合法性被那个框架所穷尽。存活下来的投诉是一个关于那个框架之内意志之行使的投诉，而它是一个凯尔森的理论供不出语汇的投诉。命名那个投诉是本文的诸任务之一。

8.3 核心与半影

哈特（1994）给出组织了英语世界讨论的那个区分。一般术语拥有一个已定意义的核心，没有它它们根本无法作为规则起作用，与一个它们的应用真正开放的半影。在核心中，应用无选择地进行。在半影中，那个机关选择，而那个选择被那条规则单独所不供出的诸考虑所引导。

哈特的区分常被读为对不确定性的一个让步，而它更好地被读为对它的一个限制：大多数应用落于核心，而那些有趣的案件是例外的。那个读法对哈特所考虑之规则类别是健全的。它拙劣地转移到本文所关切的规则类别。在一条规则的操作性术语是评价性的之处，诸如相关性、优异、充分性、合适性,或合理性，那个半影不是一个已定核心边缘的一个例外地带、而是那个术语之使用的寻常状况。§12不依赖于这一观察，尽管关于解释之诸对象的那一节将依赖，而它在此被引入，因为它标记哈特的宽慰不再宽慰的那个点。

哈特（1958）与富勒（1958）之间的交锋从补救一侧关乎同一点。富勒的立场是那个半影选择被那条规则所服务的目的所引导，而目的不外在于那条规则、而是那条规则所是之物的一部分。那个立场供出一个对解释的约束，并将在诸理论策略中被以它自己的资格处理，连同目的本身是一个解释对象这个困难。

8.4 现实主义与不确定性论题

美国现实主义者把法律推理的诸材料欠决定诸结果这一观察，施压到一个学说文献所掩盖的程度。卢埃林（1950）供出那个最经济的证明：成文法构造的诸准则成对地对立而来，每一个准则被一个反准则所对，从而一个寻求一个给定读法的法院通常将找到一个许可它的准则。那个证明被导向那些准则并一般化到任何充分丰富以覆盖诸案件的解释指引之体，因为丰富正是供出那个对立之对的东西。

批判法律研究推进一个更强的论题，在其上法律材料是彻底不确定的，而诸结果被那些材料所掩盖的诸考虑所安定。那个更强的论题已被有力地抵抗，而索勒姆（1987）以本文所采纳的形式陈述那个抵抗：那些材料是中度不确定的，也就是说它们排除甚多而在它们所许可者之内独一地选择无物。中度不确定性是凯尔森的框架以一个不同传统的语汇被重述，而两个传统对它的收敛是把它当作已定的一个理由。

现实主义对第四个问题的贡献主要是否定性的，而尽管如此是有价值的。倘若解释指引能在大多数问题的两侧被找到，那么一个要求一个机关为它的解释给出理由的要求所确保的，比它显得确保的更少，因为一个理由将对无论那个机关所抵达的哪个读法可得。这一观察将复现，以它的制度形式，在诸失败模式之中。

8.5 通过程序的正当化

卢曼（1969）供出完成本节的那个观察。诸程序产生对诸决定的接受，而它们借一个大体上独立于所决定者之正确性的机制如此做。参与吸收抗议：一个已被听闻、其提交已被接收，并已用尽所提供之诸步骤的一方，此后处在一个继续反对没有制度形式的位置，而诸期望围绕那个结果被重构。

卢曼的发现是描述性的，而它对一个规范探究的意义是可观的。它确立程序精致与解释正确性是可分的，而第一个能在没有第二个的情况下被增加。因而一个组织可以持有无可挑剔的程序、在受制于它们的那些人之间产生高的接受，并以一种没有一方被定位去质疑的方式解释它的规则，而那一接受作为无事出错的证据起作用。这是那个解释时刻在制度尺度上变得不可见所经由的机制，而它将在§12的经验程序正义文献中被再次遇到，那个文献度量卢曼所描述的那个接受，而有时被读为度量正义。

8.6 残余

本节所概览的诸说明确立三样东西并留下一样开放。它们确立规则欠决定它们的诸应用、那一欠决定是有界的，而阻止它之物是一个外在于那条规则的实践。它们留下开放的是，当一个在那个有界空间内的判定已对着一方而去时，那一方可以说什么、以及向谁。凯尔森的答案是没有法律投诉在那个框架被用尽之后存活。哈特的是那个选择是以那条规则不曾供出的诸根据作出的，这描述那个选择而不评价它。现实主义者的答案是一个理由将总是可得的，这是一个警告而非一个补救。卢曼的是那一方将在适当的时候接受它。本文的其余部分被针对那个残余。

§9 成文法与宪法解释的诸学说

本节以它发达的形式履行第二个问题。§5辨认五个模式并留下一个困难：在诸模式分歧之处，它们之中没有任何东西了结哪一个支配。此处所概览的诸学说是对那个困难的发达回答。每一个提议一个对诸模式的排序，每一个以一个关于解释为何之物的说明为那个排序辩护，而每一个遇到本节所记录的一个反对。本节以问一个组织处在使用它们中哪些之位置结束，而那个答案直接关乎§19。

9.1 诸学说作为诸排序

诸学说在此被读为二阶提议，而那一读法出于本小节末尾所给的缘由被采纳。它们无一否认诸模式存在。文本主义者在极端处查阅目的、目的主义者先读文本，而每一个学派都留意先前诸判定。分开它们之物是那个排序：哪个模式被首先查阅、哪个可覆盖哪个，以及在一个从属模式可被抵达之前什么被要求。

以这种方式读它们对当下的探究有两个好处。它使那个分歧可处理，因为诸排序能被比较。而它把那个分歧定位在§5所留下它之处，这使这一概览留在§1所陈述的那个容纳手法之内。

9.2 文本主义

文本主义以已定用法居首排序诸模式，并把其余诸模式当作唯有在用法留下那个事项开放之处才可得。斯卡利亚（1997）与斯卡利亚与加纳（2012）给出那个发达的陈述，连同一个关于用法凭之被判定之诸准则的系统化说明。

那个辩护是三重的，而每一支都关乎约束。文本是被采纳之物，故一个系于文本的读法系于那个采纳权威实际所做之物。文本是诸方能读之物，故通知被送达。而文本是最不易受那个解释者之偏好影响的材料，故那个排序比把目的或语境置于首位的诸排序更紧地约束那个解释者。

两个反对关乎本文。诸准则成对地对立而来，故在文本主义之内一个诸准则之间的选择仍存，而那个选择不受那个排序控制（卢埃林 1950）。而已定用法的判定要求诸材料，无论是词典、语料，还是那个解释者的语言能力，而这些中的每一个都是一个由某人为某目的所组装的用法之体。文本主义汲取最少这一主张，是一个汲取一个其组成较不可见之网络区域的主张，而它的约束因而是真实的、且比所标榜的更小。

9.3 意图主义

意图主义以那个采纳权威实际所持之目的居首排序诸模式，并诉诸审议之诸记录来确立它。弊端规则是它最古老的形式而立法史是它的现代工具。

那个常在的反对是一个集体机体不持有单一的心理状态、被记录的诸陈述由可能不具代表性的个体作出，而一个已知被查阅的记录成为一个策略性组成的对象。那个反对对立法机关是强的。对组织它在一个方面较弱、在另一个较强：那个采纳机体可能小到足以持有某种像共享意图之物，而它审议的记录通常对任何会援引它的一方不可得。

9.4 目的主义

目的主义以目的居首排序诸模式，并把目的同实际意图脱钩。哈特与萨克斯（1994）供出那个经典陈述，指示那个解释者假定那条规则由合理的人合理地追求合理的诸目的而作成，并归属那个假定所产出的目的。巴拉克（2005）给出一个系统的现代发展，其中目的是一个从主观与客观组分组装的法律构造物。

被归属的目的逃脱对意图主义的那个反对并招致它自己的。那一归属由那个解释者、从那个解释者所选择之诸材料、在一个那个解释者所供出之合理性标准之下执行。如此理解的目的约束那个解释者，到那一归属被规训的程度，而那一规训本身是一个解释的事。这是§8的回退抵达那个排序的层次，而它是§5在指出目的解释供出无论那个解释者认为那个目的曾是什么时所预期的那个困难。

9.5 动态解释

埃斯克里奇（1994）主张一部成文法的意义正当地演化、那个解释者的当前诸情形关乎那个文本现在要求什么，而一个否认这一点的学说拙劣地描述解释者实际所做之事。那个排序把当前诸条件置于被恰当查阅的诸材料之中。

那个反对是演化削弱通知并向那个解释者转移一个那个采纳权威不曾授予的权力。

那个学说在此配得上特别的注意，出于一个同它在那场法律辩论中之优劣无关的缘由。它是对§21中所推进之判据最相宜的学说，因为它把一个读法当作按设计可修订的，并把修订的可能性视作解释的一个特征、而非它的一个缺陷。一个按动态路线组织的实践作为一个学说的事把它的诸问题持为开放。一个按严格文本主义路线组织的实践也可能把它们持为开放，并出于一个不同的缘由如此做，因为那个文本仍可得以被再次读。闭合一个问题之物既非学说本身、而是没有一方能提起它所处的那个制度条件，这是§19以另一条路线所抵达的那个发现。

9.6 建构性解释

德沃金（1986）拒绝上面诸学说所共享的那个定框。对一个法律实践的解释是建构性的：那个解释者寻求既契合诸材料、又把那个实践显示于它最佳光芒中的那个读法。作为整全的法要求那个实践被读为一个前后一致之原则的单一作者的作品，而连环小说供出那个意象，每一个解释者写一章，被先于之物所约束而在那一约束之内自由。随之而来的那个强主张是难案有正确答案，而裁量的表象反映那个问题的困难、而非对它一个答案的缺席。

德沃金的立场是本文所采纳前提的最严重对手，而那场对峙应被直白地陈述。本文追随凯尔森（1967）主张那些材料留下一个空间，在其内若干读法同等可容许而那些材料中没有任何东西在它们之间选择。德沃金否认任何这样的空间存在，因为契合与辩护一起选择，即便那个选择可争议而没有程序确立它。

三个观察被引入以回应，而唯有第三个做真正的工作。

第一个是那个分歧对这里成问题的规则类别比它显得的更狭窄。整全汲取一个实践的密度，而一个读法把一个实践显示于它最佳光芒中这个论证要求一个带一段历史、一个先前诸判定之体，以及贯穿它们之诸被表述原则的实践。组织规则解释通常供出这些中的无一者，故会选择一个答案的诸资源缺席，无论它们是否会在一个成熟的法律系统中选择一个。

第二个是德沃金的说明被导向一个解释者应当做什么，并同一个观察者无法验证它被做相容。

第三个对本文是决定性的，而它让步那个哲学的点。假设德沃金是对的而每一个问题都有一个正确答案。解释不义，在那个假设上，成为解释错误，而本文的整套装置经受住那个名字的改变。关切它的那些问题是谁被定位去确立那个正确答案、当那个机关的答案是终局的且错误的时什么随之而来，以及什么保护一个一个错误答案对着其立的一方。那些问题在一个正确答案论题之下以同在一个框架论题之下相同的形式生起，而§19中所给的答案不受影响。那个正确答案论题会移除的，是本文说没有答案被独一地要求的那个由头。它所留下未触动的，是本文的主题。

9.7 那个宪法实例

同一个分歧在宪法尺度上在原旨主义与活宪法主义之间复现，而利害被修订的困难与被那个文本的古老所抬高。那场争论添加很少在结构上新的东西，而它供出本文所采纳的一个贡献。

博比特（1982，1991）把诸宪法论证形式当作诸模态，即历史的、文本的、结构的、学说的、伦理的与审慎的，并主张它们构成诸宪法命题被弄得可理解所处的语法。他在此有旨趣的主张是那些模态是不可通约的、没有元模态给它们排序，而当它们冲突时它们之间的选择是一个那个解释者在良心中而非按规则为之负责的行为。

这是对§5之困难可得的最诚实回应，而它是一个谢绝去解决它的回应。倘若博比特是对的，对一个排序的寻找便是被误解的，上面诸学说各是一个凭裁断抬高一个模态的尝试，而在那个冲突之点所剩下的是一个可归于作出它之人的选择。对本文那个后果同等地相宜与令人不适。它支持解释时刻是一次那些材料不判定之权力之行使这一主张。它移除一个会从外部规训那一行使之学说的任何前景。

9.8 那个制度转向

桑斯坦与弗缪勒（2003）重定向那个问题。一个给定的解释学说是否正确，无法单由一个意义之说明来了结，因为一个学说是一个由带特定诸能力、在特定诸约束下运作之特定诸解释者所施行的决定程序。诸学说之间的选择取决于那些解释者将实际招致的错误代价与决定代价，以及取决于那个学说将对那些起草者与那些被治理者之行为所有的效果。一个形式主义的学说可能对有限能力的诸解释者被辩护，即便一个更灵活的学说对理想的诸解释者会是更优的。

这个转向就当下目的而言是本节中最重要的，并出于两个缘由。

它供出通往关于实践之诸节的过渡，因为它使制度条件成为解释应当是什么、而不仅仅是它是什么的一个决定因素。

而它改变那个诸学说概览所能交付之物。倘若那个恰当的学说是那个解释者之诸能力与周围诸制度事实的一个函数，那么此处所概览的没有一个学说通过对它正确性的证明而转移到组织。所转移的是一个问题：鉴于这个解释者、这些材料，以及这个受影响诸方的人口，哪个排序产生最少且最不严重的错误，而每一个排序耗费什么来施行。

9.9 诸学说被比较

诸学说可被并置。第三列陈述每一个排序被辩护所依据的根据，而第五列陈述一个机体在那个排序对它可得之前必须拥有什么，这是随后小节所发展的那个发现。

学说	首先查阅	排序的根据	特征性反对	所要求的材料
文本主义	词语的已定用法	对解释者的约束；对被治理者的通知	诸准则成对地对立而来；用法本身被组装	被采纳的文本；语料或词典
意图主义	采纳机体实际所持之目的	对被制定之物的忠实	没有集体心理状态；诸记录被策略性组成	审议的记录
目的主义	按一个合理性标准被归属之目的	目的是那条规则所是之物的一部分	那一归属由那个解释者执行	支持归属的诸材料
动态解释	当前诸条件连同文本	描述解释者所做之事；规则活过它们的诸设定	削弱通知；转移不曾被授予之权力	关于原初与当前诸设定的说明
建构性解释	最佳地契合并辩护那个实践的读法	整全；那个实践被读为一个作者的作品	理想解释者的要求；导入实质承诺	带一段决定史的密集实践
诸模态	无排序；诸形式不可通约	冲突时的选择在良心中被负责	在冲突之点供出无纪律	共享那个语法的专业共同体
制度进路	无论解释者施行得最好的哪个排序	实际解释者的错误代价与决定代价	正确性成为对设定偶然的	关于诸解释者与诸效果的知识

9.10 向组织的转移

诸学说是为读成文法与宪法的法院所发展的，而它们向读它们自己规则之组织的转移，以一个有模式的方式不均。

每一个学说预设一个基础设施。文本主义预设一个被采纳的文本与一个用法之体。意图主义预设一个审议的记录。目的主义预设一个目的能从中被归属的诸材料与一个有纪律地归属诸目的的实践。动态解释预设一个关于当前诸条件与那条规则之原初设定的可表述说明。建构性解释预设一个带一段历史、先前诸判定，以及贯穿它们之诸原则的密集实践。博比特的诸模态预设一个专业共同体，在其之内那些模态被承认为论证的语法。

组织不均地拥有这些，而通常拥有很少。那个后果，作为本节的一个发现被陈述，是可得于一个组织的诸学说是它文档与专业基础设施的一个函数，而最紧地约束那个解释者的诸排序是要求最多基础设施的那些。一个带一个被采纳文本而别无它物的机体能实行一种稀薄的文本主义、并能实行别无它物，而稀薄的文本主义是那个把最多留给那个解释者、同时显得留下最少的排序，因为在语料、诸记录与一个专业共同体缺席的情况下对已定用法的判定，是一个那个机关作出而没有一方能核查的判定。

这个发现将以制度形式在关于实践之诸节中复现，在那里为组织最常被提议的诸装置是构建基础设施的诸装置：被发布的判据、被记录的理由、被编制索引的诸判定，以及那个机体之内的先例。

9.11 诸发现与残余

本节确立诸学说是诸模式的诸排序、每一个排序遇到一个严重的反对、本文前提的最强哲学对手把它的主题留下未触动、那个制度转向使诸排序之间的选择取决于解释者与设定，而组织只能使用它们基础设施所支持的诸排序。它留下开放的是，任何排序能否独立于制度条件被辩护，而这是一个本文不追寻的问题，而它对§24中的 Q3 有贡献。

§10 西方诸学说之外的解释权威架构

本节回到第一个问题。§4确立解释的资格在核心被授予、在边缘被生成，并留下开放的是解释资格是否可被授予以及什么使它的正当性成立。§9中所概览的诸学说对二者说得很少，因为它们是为一个其中那个权威解释者的身份由宪法安排所安定、而那些有趣的问题关乎那个解释者应如何进行的设定所发展的。一个传统长篇地处理了那个在先的问题，而它的架构供出西方材料所不置于前景的诸资源。

一则告诫支配整节。随之而来的是为一个狭窄目的对一个大传统的一个读法，而它选取关乎解释权威的诸结构特征，同时略过那个传统所含有的大部分。专家将发现这一处理稀薄，而所引入的诸主张被限于结构。

10.1 解释与决定分开

伊斯兰法律理论从一个在§9诸学说中没有对应物的区分开始。神圣的法本身与诸法学家对它的理解被分别地命名，而构成那个发达之法的诸裁定属于第二个范畴。一个裁定是一个有资格之人对那个法要求什么的理解，而它被那个传统自己的语汇当作如此持有（哈拉克 1997；韦斯 1998）。

对解释权威的后果是可观的。一个其语汇把那个法同对它的理解区分开来的系统，已把解释的可错性建进它自己的自我描述，并已在没有削弱所决定者之权威的情况下如此做。那个判定有约束力并被知道是一个判定。

这的制度表达是一个西方材料当作不寻常的分开。发布一个意见的法学家与决定一个案件的法官占据带不同权力的不同职位（马苏德等人 1996）。那个意见是权威的，在它陈述一个有资格的解释者持那个法要求什么这个意义上，而它不处置那个案件。那个判决处置那个案件，而不因此就是那个法的一个更高陈述。一方可以向一个对他们没有权力的法学家寻求一个意见，而那个意见的权威停留在那个法学家在那个解释共同体中的站位上。

这一安排的两个特征直接关乎§4中所留下未决的诸问题。

解释权威在此是可授予的，而它的授予是那个寻常情形。那个决定的机关不是唯一的、甚至不是主要的解释者，而那个解释功能由一个根本不持有决定权力的诸人之体所履行。

而解释的资格奠基于被证明的能力。一个可执行独立推导者所被要求的诸资格被长篇地表述：对语言、对源文本、对被接受之共识，以及对诸裁定被推导所经由之诸方法的驾驭。一个缺乏它们的人跟随一个有资格的法学家，而那一跟随是一个被承认且可敬的立场，带一个它自己的名字。因而权威停留在一个可被其他持有同一能力者评估的公共能力标准上，而它被独立于职位地持有。

10.2 复数性与分歧的站位

那个传统关乎本文的第二个特征是它对分歧的处理。

解释工作被组织为诸学派，每一个传递它自己的诸方法与诸学说，每一个承认其他为正当，而每一个把它的诸立场同它们的并置维持。有资格诸解释者之间的分歧有一个名字与一个文献，而那个复数性被容纳为那个法的一个常在状况、而非一个等待解决的缺陷。

在这之下是一个§8与§9的读者将认出的理论争论。一个经典立场主张在有资格诸解释者不同之处，每一个都已达到所被要求于他们之物而每一个裁定都是正确的。一个对立立场主张一个答案是正确的而那个错过它的解释者已犯错，同时因已恰当地推理而无可责难。那个争论是那个分开凯尔森（1967）同德沃金（1986）的同一个，独立地并早若干世纪被抵达，而它的术语足够接近以致那个对应配得上被记录。

那个传统所添加的，是它从那个争论所引出的制度后果，而那个现代辩论未曾引出。无论哪个立场被取，复数性被制度地容纳：诸学派共存，一方可以跟随一个，而来自一个学派的一个裁定不熄灭其他的。关于一个答案是否正确的分歧因而同一个答案是否必须被执行的问题被分开。

一则史学告诫属于此处。独立推导在某一时期停止、那扇门已被关闭这个被接受的说明，已被长篇挑战，而那个挑战现在在轮廓上被广泛接受（哈拉克 1984）。此处所作的结构点不依赖于那个争议的解决，因为它关乎那个架构、而非它在任何特定日期的活力。

10.3 接受与移植

现代行政国家的诸装置通过对大陆学说的接受抵达东亚，而那些接受系统修改了它们所接受之物。日本的行政程序立法要求诸机构确立并发布诸判据，申请与处置对着它们被评估，这把解释实践的一部分转换为被发布的规则。那个安排在§19的诸装置中被处理，而它在此被指出，因为它例示§9那个制度转向所作的点：一个被移植进一个不同制度设定的装置运作得不同，而它的效力是那个设定的一个函数、而非它形式的。

10.4 所供出的诸结构资源

两个结构资源从本节的材料浮现，而每一个回答一个西方诸学说所留下未加处理的问题。

解释权威与决定权威的分开。 一个解释功能可由持有处置那个案件之权力的诸人所履行。这直接关乎§3中所辨认的那个递归，因为一个判定它自己复审之范围的权力，以一种一个被同决定权力分开持有之解释权力所不是的方式是递归的。那一分开能否在一个现代组织中被复制，以及它是重新安置那个递归还是减少它，是§24的 Q1，而本节的材料是那个安排可能且已在规模上被维系的证据。

能力作为资格的根据。 解释权威可停留在一个公共且被表述的资格标准上，可被其他持有它者评估。这是对 Q2 的一个候选答案，而它是一个苛求的，因为它要求一个有能力评估能力的共同体与一个能力对着被度量的方法之体。§9与§19都指出这样的共同体在大多数组织规则解释周围的缺席。

10.5 关于转移的诸告诫

伊斯兰法律理论在此被当作一个结构来源、而非一个正义的模型。它被在它自己的诸不对称之内施行，而对所描述之诸解释资源的接触被身份、被财富、被性别与被地方不均地分布。一个把本节当作一个论证这些安排比别处所概览者更正义的读者，会把它当作它所不断言之物。

而那个制度转向以完全的力适用。一个从一个设定被汲取并被安装于另一个的装置，将依照那个新设定的诸能力与诸激励运作。此处所辨认的诸资源被作为诸可能性引入，其可转移性未被判定，而判定它是本文不承担的工作。

10.6 诸发现与残余

本节确立解释权威已在规模上同决定权威分开，而资格已奠基于可评估的能力。它供出关乎§24之 Q1 与 Q2 的证据并了结无一，因为一个安排在一个设定中的可能性是它在另一个中之可得性的弱证据。

§11 法律解释中的正义

先于本节的诸节已问解释如何进行以及谁被定位去执行它。本节问关于解释何时是正义的已被说了什么，这是第四个问题以那个法律传统所给它之形式。五个说明被处理。每一个对那个解释行为本身作出一个规范要求，每一个是§21中所推进之判据某一部分的最近祖先，而每一个在本文所关切的情形之前止步，出于本节在每一实例中所辨认的一个缘由。

11.1 一致性作为合法性的一个条件

富勒（1969）铺陈八个要求，它们的满足构成他所称的法之内在道德，而第八个直接关乎解释。在如所宣告之诸规则与施行它们之官员的诸行动之间，必须有一致性。一个其中所宣告的规则说一件事而官员行动做另一件的系统，已作为一个法律系统失败，并以一个不同于具有坏规则或抵达坏结果的方式失败。

富勒处理的两个特征在此要紧。他把一致性的失败当作一个程度的事、并当作一个系统朝之移动的一个愿景，从而合法性是一个制度的一个标量性质、而非它跨过的一个门槛。而他在失败的诸原因中辨认不仅腐败与漠视，还有错误的解释，从而一个误读一条规则的官员借如此做损害合法性，而没有任何动机的不当。

一致性是§21之判据的最近经典祖先，因为两者都把规则与应用之间的关系当作一个自成其类的规范评估对象。那个差别是精确的，而它是一致性无法在此服务的缘由。问官员行动是否符合所宣告的规则，预设那个所宣告的规则有一个符合可对着被度量的内容。在那条规则的术语是确定的之处，一致性有抓力并是一个可用的检验。在那个术语是一个标准或一个比较之处，§6辨认为承载解释负担的那一类，符合会对着被度量的那个内容正是那个解释行为所供出之物。因而一致性恰恰在那个问题活的地方说得最少，而那个要求在半影情形上沉默地略过。

11.2 法治与裁量的规训

拉兹（1979）把法治当作那个法引导受制于它者的能力，并从那个能力引出一组原则：规则必须是前瞻的、公开的与清晰的，必须是相对稳定的，必须由公开且稳定的一般规则所作成，而必须由一个独立的司法所施行，带可及的法院与对自然正义的遵守。他添加一个关乎当下主题的原则：被托以施行那个法之诸机构的裁量不应被允许歪曲它。

拉兹的说明是工具性的而他如此说，把法治当作一个减少一个由法本身所创造之危险的否定性德性。两条界限为本文随之而来。引导的能力，一如一致性，预设可确定的内容，故那个说明把规则的模糊当作一个起草的缺陷来处理，而不处理那个在起草不能之处供出内容的解释行为。而那些关乎独立裁决与可及法院的原则被针对法律系统，故它们辨认一个组织大多不满足的条件，而不给一个在一个不满足它之组织中的一方提供任何东西。

11.3 那个论证性维度

沃尔德伦（2011）论证法治的诸主导构想已过重规则并过轻程序，而一个法律系统被它的论证性装置一如被它的诸规范同等地刻画。法院听闻诸方、接收关于那个法要求什么的诸提交、给出理由，并允许上诉。在沃尔德伦的说明上，这些对合法性是构成性的、而不是它的行政附件，而它们之中是一个受影响之人就那个法就其案件所应用而言意味什么加以论证的机会。

这是对此处所推进之判据的最近现存规范要求，因为它主张质疑一个解释的资格是合法性本身的一个组分。一个不听闻关于它们之论证便判定诸应用的系统，在沃尔德伦的说明上会作为一个法律系统有缺陷，无论它诸判定的品质。

两条界限附着。那个说明是为拥有法院的法律系统所发展的，而它向没有它们之机体的转移是一个沃尔德伦不接手的问题。而它确立一个论证的机会而不处理§3中所辨认的那个状况，其中接收关于一条规则之意义之论证的那个机体，是其解释被争议的那个机体，而它根本听闻那个论证的权限由它自己对支配挑战之诸规则的读法所安定。一个在那个读法之作者面前论证的机会值得某样东西，而值得比那个说明所假定的更少。

11.4 参与作为裁决的标志

富勒（1978）借一个形式给予受影响一方的参与模式，把裁决同社会秩序化的其他形式区分开来。合同经由被协商的同意给予参与；管理指令给予无一个可比之类的参与；裁决给予呈现诸证明与被推理之诸论证的独特参与。那个区分特征不是一个决定者的在场、而是那一方为一个对他们有利之决定呈现一个案件的机会。

两个后果关乎本文。

第一个是规范性的并支持那个判据。倘若呈现被推理的论证是使裁决成为它所是之秩序化形式之物，那么一方对那个判定的贡献对那个形式是构成性的、而非一个在它之内被延伸的客套。一个在没有任何这样机会的情况下所抵达的判定，是裁决之外的某样东西，无论那个决定机体被称为什么。

第二个是一条界限，而它是严峻的。富勒的说明界定一个形式而不要求任何给定的决定采取那个形式。一个分配稀缺好处的组织，在富勒的分类法上通常从事管理指令，而管理指令不携带裁决那种的任何参与要求。因而那个说明解释为什么参与权利不附着于准入、任命，或拨款决定，而在解释它时，许可那个缺席。这是一个那个传统自己的语汇逆着那一方工作的情形。

11.5 私人组织中的合法性

塞尔兹尼克（1969）供出这一传统中对本文所提问题的唯一持续处理，并为私人治理如此做。合法性在他的说明上是一个私人机构能在程度上实现的价值，通过从随意行使的管理特权向由规则所支配之行政的移动，带被表述的诸标准、诸记录、被推理的诸决定，与诸上诉渠道。工业正当程序是他给由此产生之状况的名字，而那个研究记录它在诸公司内部治理中的出现。

那个说明是当下主题的直接先例，而它诸发现的一个特征在此要求注意，因为它从一个不同的方向预期§19。塞尔兹尼克所记录的诸发展发生在一个抗衡的有组织之力在场之处，最常是一个有能力使那个雇主之特权行使有代价的工会，并发生在法律学说已开始把诸后果附着于任意的内部行动之处。私人组织之内的合法性，在确立那个概念之研究的证据上，发展在某个在那个管理之外者持有那个管理所需要之物之处。这是§19中所辨认的那个利害，在那个组织合法性概念起源的领域中被历史地观察。

11.6 解释与应用

一个进一步的说明为记录属于此处，因为它关乎§3。伽达默尔（2004）主张理解一个规范文本与把它应用于一个案件是一个操作的诸方面，从而一条规则是在并通过它对诸情形的应用被理解，而法律解释因那一缘由是理解一般的典范。

那个说明是描述性的，而它携带一个本节应登记的规范蕴含。倘若应用内在于理解，那么那个解释者先确定一个意义、而后应用它的要求，描述某样不发生之物，而一个以那些术语被定框的忠实要求，要求那个不可能的。所能被要求的，是那个应用是那个文本能承载的一个，而那个解释者对那个案件改变他们认为那个文本所要求之物保持开放。这些中的第二个接近§21将要求于一个机构之物，被转置到那个解释者自己的性情。

11.7 诸发现与残余

所概览的诸说明确立甚多。解释在合法性的诸关切之内、而不是一个躺在它们之外的仅仅技术性的事。所宣告规则与官员行动之间的一致性是一个系统对合法性之主张的一个条件，并通过误读一如通过腐败而失败。就一条规则意味什么加以论证的资格是法治的一个组分、而非它的一个附件。那个受影响一方对被推理论证的呈现对裁决作为一个形式是构成性的。而合法性是一个私人组织能在程度上实现的价值，这是当下这篇论文所要求且无需去论证的那个前提。

那个传统所留下的是一个模式。每一个说明或者预设一个应用可对着被度量的确定内容，这是那个半影情形所缺者；或者预设一个论证可在其面前被作出的独立法庭，这是那个组织情形通常所缺者；或者记录在一个抗衡之力在场之处所成立的诸条件，这是§19辨认为那个困难不生起之情形的那一个。那个半影判定，由一个对无人负责之机体终局地作出，是它们无一够到的那个残余，而它是本文当作它主题的那个情形。

§12 与相邻不义概念的关系

本节把这个概念个体化。一个无法同它诸邻居区分的概念未曾被引入，而解释不义有四个近到足以使举证责任落于当下说明去显示它添加什么的邻居。那四个是诠释学与证言不义、程序不义、支配，与那个杀法的职位。每一个都通过定位它的不义驻留于何处、被谁承受、经由什么机制运作，以及它蕴含什么补救而被处理，因为共享一个名字的诸概念仍可由它们的补救被分辨。

12.1 诠释学不义与证言不义

弗里克（2007）区分两种认识不义。在证言不义中一个说者因附着于他们身份的偏见而收到一个可信度赤字。在诠释学不义中集体诠释资源的一个欠缺使一个人经验的一个重要区域不可理解，对他人并且常对他们自己，而那一欠缺不均地落于那些被排除于这样的资源被生成之诸实践之外者。

诠释学不义按名字是最近的邻居而是一个不同的概念。它的所在是一个集体的资源储备；当下概念的所在是一个职位。它的承受者遭受一个使一个经验可理解的无能；此处的承受者遭受一个由别人可理解地并终局地作出的判定。它的机制是一个空缺；此处的机制是一次行使。而那些补救尖锐地分歧，这是那个决定性的检验。诠释学不义的补救是那个集体诠释储备的丰富，连同弗里克所称的个别听者中的诠释学正义那个德性。丰富那个集体储备对一个其案件已被一个持有固定它自己规则之意义之权威的机关所判定的一方什么也不做，因为那个判定成立，无论那个周围的语汇含有什么。

一个交叠应被精确地让步。§3中所引入的表示性不对称，在其上日常语言供出稀薄的术语用于反对一个解释，是弗里克意义上的一个诠释学赤字。一个因无可得语汇表达而无法表述那个反对的一方，遭受弗里克所描述的那个不义，并在此处所描述者之外遭受它。当下这篇论文可被读为，在那一狭窄方面，一个供出那缺失语汇之一部分的尝试。那些概念仍有别：供出那个语汇移除那个诠释学赤字并把那个解释权威留在它正所在之处。

12.2 程序不义

罗尔斯（1971）区分完全的、不完全的与纯粹的程序正义。在一个关于正确结果的独立判据存在而一个程序可靠地达到它之处，那个程序是完全的；在那个判据存在而没有程序可靠地达到它之处，是不完全的；在没有独立判据存在而那个结果被界定为无论一个公平程序所产生者之处，是纯粹的。

解释在这个方案中不自在地坐落，而那一不自在是有教益的。一个独立判据看来存在，因为那条规则在那里而诸方争议它要求什么，这把那个情形置于不完全的程序正义之下。然而§3与§8的论证主张在那个框架之内没有判据选择一个独一的读法，这把那个情形推向纯粹的程序正义，在其上无论那个机关通过一个公平程序所判定者按定义就是那条规则所要求的。第二个刻画是组织隐含地采纳的那个，而采纳它凭构造消解解释不义的可能性。本文抵抗那个消解，以§3中所给的根据：诸可容许解释的空间是有界的，故一个判定可被显示为坐落于那个术语从中取其内容的网络之外，而一个判据因而存在，即便它选择没有独一的答案。纯粹那种的程序正义对诸可容许读法之间的选择可得，而对一个读法是否曾可容许的问题不可得。

那个经验文献引起一个第二的、更锐利的困难。泰勒（1990）确立对公平的知觉转在发言权、中立、被尊重的对待与可信赖上，而它们比结果的有利性更强地驱动对诸结果的接受与对权威的遵从。同卢曼（1969）一起读，那个发现对任何把程序正义当作充分之说明携带一个令人不适的蕴含。产生那个公平之知觉的诸特征是一个组织能在它的解释实践保持超出质疑的同时供出的诸特征。因而接受是解释正义的一个不可靠指标，而一个运转良好的组织可能恰恰在当下这个概念会定位一桩不义之处产生接受。

12.3 支配与可质疑性

佩迪特（1997）把支配界定为对另一个任意地干预一个人诸选择之能力的受制，其中干预在它无需追踪那个受影响之人的诸利益与诸观念时是任意的。作为非支配的自由要求这样的能力被移除，而佩迪特的制度提议核心地转在可质疑性上：权力在那些受制于它者能有效地质疑它的行使并要求它对他们的可声言诸利益负责时，被弄得非任意。

这是最近的邻居，而对它的关系是本文应陈述为互补的一个。支配命名一个常在的关系并对干预经由之发生的模式漠不关心。解释不义命名一个模式，而那个模式有旨趣因为它是把自己呈现为干预之缺席的那一个。一个进行解释的机关，在它自己的说明上，根本没有在干预，而在应用一条两方都受制于的规则。

那个互补性在补救之点变得实质。佩迪特对任意权力的补救是可质疑性，而§3中所引入之不对称的那个递归组分直接关乎它。判定一个质疑是否存在、它是否及时、它是否关乎一个在复审之内的事项，以及那个质疑者是否有资格的诸规则，是该组织的规则，由该组织所解释。因而可质疑性本身是它意在约束之解释权力的一个对象。这是作为对那个共和主义说明的一个贡献、而非作为对它的一个反对被提供的：它辨认一个凭之那个共和主义补救能被吸收的机制，而那个辨认对任何试图在一个书写它自己规则之机体中把可质疑性制度化的人有用。

12.4 那个杀法的职位

科弗（1983）主张诸规范世界被诸共同体丰盛地生成，每一个带它自己关于它的法意味什么的说明，而一个法院的独特工作是压制那些竞争者：法院是杀法的，而它们通过选择一个读法并熄灭那些活在持有它们之诸共同体中的其他读法来杀法。

科弗供出最接近本文后面诸节所要求之物的语汇，因为他把解释当作一个对诸规范世界有生成性与破坏性效果的行为。两个差别应被标记。科弗的设定是生法的复数性，其中读法被熄灭的一方持有它自己一个发达的规范秩序；此处的设定通常是一个其中那一方持有一个案件而没有竞争性规范世界的。而科弗的说明是描述性的，描述一个他当作不可避免的角色，因为一个复数的法律秩序要求某些读法被压制，倘若任何要被执行。当下的说明不争议判定的必要性并问一个判定对它被作出于其内之关系之能力做什么，这是一个科弗不提出的问题。

12.5 那个个体化被陈述

那四个比较可被概括。

概念	所在	承受者遭受	机制	所蕴含的补救
诠释学不义	资源的集体储备	经验的不可理解	欠缺	丰富那个储备
程序不义	通往决定的路径	发言权或中立的被拒	有缺陷的过程	改革那个过程
支配	一个常在的关系	对任意干预的易受性	干预的能力	把质疑制度化
杀法的压制	那个复数的法律秩序	一个规范世界的熄灭	权威性的选择	被接受为不可避免
解释不义	一个持有规则的职位	对那条规则就其案件之内容的一个判定	构成性的行使，闭合于它自己的复审之上	开放的问题

那最后一行的两个特征把它同其他区分开来。那个机制是构成性的，故那一方被一个那条规则在那个决定之前不曾携带的内容所约束。而那个机制闭合于它自己的复审之上，故第三行的那个寻常补救，被制度化的质疑，被暴露于它所约束的那个权力本身。那最后一列被刻意地留下开放。那些相邻概念中的每一个都带一个附着的补救到来，而本文其余部分的负担是可得于解释不义的诸补救以一种其他概念所不是的方式被结构性地限制。

§13 约束解释的理论策略

§9问哪个解释模式在诸模式分歧之处支配。本节问一个在先的问题：什么根本能规训一个解释者。一个学说提议一个排序。一个策略提议一个约束来源，而诸学说汲取诸策略而不与它们同一。五个家族在此被区分，而第五个是§21之说明同之最密切关联的那个。

13.1 由文本的约束

第一个策略由那条规则的诸词语固定内容，而它的诸工具从寻常的文本主义读解，经由定义条款与受控词汇表，到把一条规则完全界定为一个决定程序而运行。

那个极限情形配得上单独的注意，因为它是一个活的提议、而非一个思想实验。一条规则可被以一个容许机械执行的形式表达，无论作为一个算法、一个决定表，还是可执行代码，而施行高量类似案件的诸机体已把它们规则系统的大量部分移入那个形式。希尔德布兰特（2015）考察那个转变及它对法律保护的诸后果。在那个转变是完全的之处，那个解释时刻从那个决定之点消失，因为当那个案件到达时没有任何东西剩下被判定。

13.2 由目的的约束

第二个策略由那条规则所服务的目的固定内容，主张一个挫败那个目的的读法因此被排除。它的诸工具是目的条款、序言、目标陈述、弊端规则，与目的论方法。那个策略是最常被推荐给组织的一个，因为它看来只要求那个机体陈述一条规则为什么存在。

13.3 由程序的约束

第三个策略通过进行的方式规训那个解释者，把那个判定的内容形式上留下开放。它的诸工具是说明理由的要求、那个决定机体的构成、调查功能同决定功能的分开、上诉的可得性、诸判定的发布，与一方可回应所处的时限。这个家族是制度实践主要依赖的那个，而§19以它们的发达形式概览它的诸工具。

13.4 由共同体的约束

第四个策略把那个纪律定位在一个有能力从业者的共同体中。一个解释者在一个实践的诸标准之下进行，而约束他们之物是他们的读法是一个他们能向其他持有同一能力并会承认一个背离者辩护之读法这一要求。

费什（1980）给出那个强陈述，在其上诸解释共同体供出根本使一个读法可得的诸假设，从而那个约束在那个解释者选择之前运作并塑造把自己呈现为一个可能读法之物。布兰顿（1994）供出一个关于这样的约束运作于其内之实践的更一般说明，其中诸参与者记下彼此的诸承诺与诸资格，而给出并索求理由的实践供出那个纪律。§8中所注意的对遵循规则问题的共同解决属于这个家族，而§10中所描述的安排也如此，其中解释的资格停留在一个可被其他持有它者评估的能力上。

这个家族是当下说明最接近的那个，因为它把约束定位在诸方之间的一个关系中、而非在文本或目的的一个性质中。§14辨认它同§21之说明分道之处。

13.5 由反身性的约束

一个第五个策略把自己针对诸系统、而非针对诸解释者。托伊布纳（2012）主张私人与跨国秩序，没有外部权威被定位去治理它们，尽管如此可被通过诱使它们发展它们自己的内部限制而约束，从而那个系统安装凭之它约束自己并对它所影响者之诸主张保持开放的诸程序。宪法化在这一说明上是一个社会秩序能在没有一个国家去强加它的情况下经历之物。

那个策略是最直接针对本文所关切之情境的，因为它把那个外部权威的缺席当作它的起点。它的困难被引入下一节。

§14 诸策略的困难

§13铺陈五个约束家族并评估无一。本节依次取每一个，以那里所确立的次序。那些困难是不同种类的，而它们共享一个那个结尾小节所陈述的形式。

14.1 文本的欠决定

文本欠决定应用，这§8已确立且无需被重复。

那个极限情形要求单独的处理，因为它的困难是一个不同种类的。完全的界定确实在那个决定之点购得确定性，而它借把那个解释行为重新安置到那个界定之点购得它。某人判定这些案件落于那条规则之下而那些不，并编码了那个判定。那个由一个决定一个案件之机关公开地执行的行为，被无论谁书写那个界定者事先执行，通常没有一个案件在他们面前、通常没有任何一方在场，而通常没有那个结果作为一个解释可读。那个后果是§4中被辨认为形式权威同有效权威之分歧的那个，以它最完全的形式：那个解释权力已移到一个那些规则不命名、没有一方能致意，而其行使把自己呈现为一个技术实现的位置。因而在应用之点的确定性是真实的，而那个解释问题已被掩盖而非被回答。

14.2 目的的回退

目的本身是一个解释对象，故那个策略回退。哈特与萨克斯（1994）指示那个解释者归属一个合理的人会持有的目的，而那一归属由那个解释者从那个解释者所选择之诸材料执行。在那个目的被陈述于那个工具中之处，那个回退被缩短而不终止，因为一个被陈述的目的被以诸术语表达，其应用要求判定。

14.3 程序的形式性

程序确保一个判定的形式并把它的内容形式上留下未触动。这是一个按设计的局限、而非一个缺陷，因为一个支配诸结果的程序约束会已不再是程序性的。

两个后果随之而来，关乎本文。卢曼（1969）确立程序精致借一个独立于所决定者之正确性的机制产生接受，而泰勒（1990）供出同一发现的经验形式。因而一个组织可以增加程序品质、增加它诸判定的接受，并把它的解释实践留在它正所在之处。而§3中所辨认的那个递归特征以特别的力关乎这个家族，因为程序约束本身是该组织的规则，而它们的范围、它们的可适用性，与它们被违反的诸后果，由它们所约束的那个机体所判定。

14.4 共同体的触及

由共同体的约束以一种易被错过的方式失败，因为那个失败同那个策略完美地运作相容。

一个有能力从业者的共同体就那个共同体的其他成员而言规训一个解释者。一个解释者必须能辩护的读法是一个可向同侪辩护的读法。在那个受影响一方是那个共同体的一个成员之处，这是一个可观的保护。在那个受影响一方立于它之外之处，那个策略不直接给他们供出任何东西，而它在最要紧的情形中根本供不出任何东西：一个共同体可能收敛于一个对着那些在它之外者严酷地运作的读法，那个收敛恰恰是那个策略所称的约束在起作用。诸解释者之间的一致是那个家族所交付之物，而一致同任何内容相容。

这是当下说明同它最相似之家族分道的那个点。两者都把规范分量定位在一个关系中。由共同体的约束策略把它定位在诸解释者之间的关系中，而§21的判据把它定位在那个解释者同被解释于其上之一方之间的关系中。第一个问一个读法能否被向那些有能力评估它者辩护。第二个问那个读法是否把那个受影响一方留在被定位去再次提起那个问题。一个实践可能完全地满足第一个而完全地失败于第二个，而§19辨认那一组合的制度形式，以选区对齐这个标题。

14.5 反身性的先决条件

反身性约束预设一个系统拥有在它自己之上安装限制的能力，并立于一个逼它去这样做的环境中。在那个环境逼迫之处，那个策略描述某样真实且可观察之物。在它不逼迫之处，那个策略陈述一个可能性并辨认无物会带来它。那个先决条件是§19的那个利害以系统论语汇的到来，而那个最直接针对外部权威之缺席的策略，结果证明要求一个执行那个缺席权威本会执行之本分的环境。

14.6 那个共同残余

那些困难可被一起铺陈。

策略	纪律的来源	预设	特征性失败
文本	如所使用的诸词语	共享且稳定的用法	欠决定；完全的界定把那个行为重新安置到那个编码者
目的	所服务的目的	可恢复且确定的诸目的	目的本身被解释
程序	进行的方式	形式追踪实质	确保接受而无正确性
共同体	有能力评估的同侪	成员资格与共享的诸标准	在成员之间一致，对外来者沉默
反身性	那个系统的自我限制	一个逼迫的环境	要求那个其缺席界定那个问题的利害

每一个策略都预设某样它自己无法确保之物，而那个先决条件在每一情形中都是那个设定的一个条件、而非那个策略的一个特征。这是在§19中制度地被抵达之结果的理论对应物，而两者的巧合值得直白地陈述：诸策略在诸装置失败之处失败，而它们出于同一缘由失败。规训一个解释者之物，在每一个家族中，是那个解释者之外的某样那个解释者对之负责之物，而关乎本文的诸情形是那些其中在那个解释者之外没有任何东西被定位去被负责的情形。

随之而来的诸节从诸策略转到诸工具，并问诸机构实际所构建的诸装置是否比它们所汲取的诸策略表现得更好。

§15 立法中的解释控制

从此处开始的四节概览诸制度就第四个问题所建造之物。它们按尺度进行，而这一节处理那个其上解释在任何案件生起之前即被预期的尺度。诸装置因而是前瞻的：它们运作于那条规则以及它后来被读之诸条件之上，而它们无一运作于一个一旦被作出的判定之上。

15.1 解释制定法

若干法域维持一部支配所有其他制定法之读法的制定法。英国的1978年《解释法》是那个紧凑形式，供出标准定义、关于数与性别的默认规则，以及关于生效与废止的诸条款。加拿大的《解释法》指示每一项立法被视为救济性的，并被给予最佳确保其诸目标之达成的公平、宽大与自由的建构。澳大利亚的1901年《立法解释法》在第15AA条要求那个最佳达成该法之目的或目标的解释，并在第15AB条许可诉诸外部材料以确认或判定意义。新西兰的2019年《立法法》指示意义按目的与语境从文本被确定。新加坡的《解释法》含有一个可比的目的性条款。

这一体裁配得上注意，出于一个学说文献不总是弄明白的缘由。§9把诸模式的排序当作一个诸学说于其上分歧而没有元规则裁断的事项。这些制定法凭立法了结那个问题。一个目的性的强制命令是一个在§9所概览的诸排序之间的立法选择，被强加于所有随后的解释者并从他们的裁量被移除。那个强加是否成功是一个分开的事项，因为那个强制命令本身是一条规则，其应用要求一个目的的判定，而那是§14的那个回退。所被确立的是那个排序问题凭权威可回答，即便它凭论证不可回答。

15.2 定义与目的条款

两个起草装置直接运作于§6所分类的材料之上。

一个定义条款把一个术语从一类转换到另一类。一个其应用要求在每一场合被评估的标准术语可被给予一个穷尽的定义并因而变得分类性，而一个分类性术语可被给予一个被列举的外延并变得确定。那条规则的解释负担相应地落下。那个装置有一个§6所引入的代价：那个未定义术语所保全的余地被交出，而重视那个余地的诸机体有理由在一个保留它的一般性层次上定义。

一个目的条款供出那个目的性模式所要求、而§5发现在法律系统之外通常缺席的材料。它的在场把一个不可得的模式转换为一个可得的。它的局限是一个被陈述的目的被以诸术语表达，其自身的应用要求判定，故那个装置缩短那个回退而不终止它。

一起来看，这两个装置在§9所辨认的意义上是构建基础设施的：它们供出更约束性的诸排序所依赖的材料，而它们是这一概览中一个机体借使诸模式可得而改善它自己解释纪律的最清楚实例。

15.3 对被委托解释的控制

在一个立法机关把阐发它诸规则的权力授予另一个机体之处，它授予那个解释功能的一部分，而控制那一授予的诸装置是对解释的诸控制。

《欧洲联盟运作条约》在第290与291条之间区分补充或修订一项立法行为之非本质要素的诸行为与确立其实施之统一条件的诸行为，并把本质要素保留给那项立法行为本身。那一保留在此有旨趣，因为它辨认一类根本不可被委托的解释问题，而那是一种这一概览中没有其他尺度供出的控制。

英国凭附属文书修订主要立法的授权做法，以那个此做法从中取其标签之都铎制定法的名字为人所知，是那个相反的情形，而对它的持续批评恰恰关乎解释与立法权力向一个以更少审查运作之机体的转移。

在美国，限制委托的那个学说要求国会供出一个可理解的原则来指导那个受托者。那个学说数十年来鲜少被执行，并在近年成为重新论证的对象，而它的当前状况应在它被依赖之前被核实，与§2所引入的那个告诫一致。

15.4 多于一种语言的作准文本

欧洲联盟以所有官方语言采纳它的立法，每一版本同等作准，而法院相应地在它们分歧之处比较诸版本，并参照那项措施的目的与总体方案解决那个分歧。在 CILFIT 案中法院铺陈一个国内法院可把联盟法的正确应用当作如此显然以致不要求提交之诸条件，而在如此做时把那个立法的多语言性质与诸版本的比较记录在关乎解释的诸考量之中。

这一安排配得上它自己一段，因为它是一个§8以哲学方式论证之命题的制度示范。在一个文本存在之处，那个文本欠决定它应用的主张可被通过断言一个有能力的读者只是看见那些词语意味什么而抵抗。在若干文本同等作准而它们分歧之处，文本的不足在记录的表面，而那个系统被逼采纳一个把目的与方案置于任何单一版本之词语之上的排序。多语言主义把一个被争议的哲学论题转换为一个操作上的必然。

15.5 时间性装置

一个最后的家族运作于一条规则以及它所吸引之诸读法的存续期之上。落日条款在一个固定日期终止一项措施，除非它被更新。复审条款要求一项措施在一个陈述的间隔之后被审查。立法后审查，在若干议会系统中由委员会进行，审查一项措施在实践中如何运作。

这些装置配得上不相称于它们通常所受注意的强调，因为它们是这一概览在任何尺度上所遇到的最接近一个针对§21之关切之装置的东西。一个落日条款逼一个问题被再次问。一个复审条款逼一条规则的运作，即被给予它之诸读法的累积之体，被一个应用它之机体以外的机体审查。两个装置无一为一方对着一个解释的保护而设计，而两个都有防止一个安定读法越过审查的效果，而那是§21的判据所要求的那个条件。

15.6 诸发现与残余

立法装置供出可观力量的前瞻控制。诸模式的排序可凭制定法被安定。一个术语的解释负担可凭定义被减少。更约束性的诸模式所要求的材料可凭起草被创造。诸类问题可对着委托被保留。而诸安定读法的周期性重开可凭时间性条款被逼出。

这一尺度所留下之物在此处所概览的每一装置之间是均一的，而陈述它是本节的要点。每一个装置运作于那条规则、那个起草者，或一个受托者之上，而无一运作于一个一旦被作出的判定之上。无一供出任何东西给一个对着其一个读法已被应用的一方。无一由一个投诉触发，无一在一个受影响之人的请求下可得，而无一要求任何判定因一方已反对它而被重访。这一尺度的诸装置完全地在那个作成规则之机体一侧，而那个受影响一方在它们无一之中有一个位置。

§16 标准制定中的解释控制

这一尺度包含没有立法权威而书写规则、且它们规则通过他人附着于交易之诸条件而有约束力的诸机体。一个标准被自愿地采纳，并在一个购买者、一个贷方、一个保险人、一个监管者，或一个采购规则要求符合它之处变得实践上有义务性。此处的解释问题是复合的，因为一方受制于一个标准，面对关于那个标准要求什么、由一个不曾书写它之机体所作出、应用一个由一个那方对之没有影响之机体所书写之文本的诸判定。

下面所概览的诸工具是这一概览所遇到对第四个问题的最发达回答，而本节以辨认它们所不做之物结束。

16.1 支配标准之书写的标准

一个第一家族规管标准被作成所经由的过程。

世界贸易组织《技术性贸易壁垒协定》的附件3铺陈一个关于标准之准备、采纳与应用的良好实践守则，被致意于标准化诸机体，并关切透明、非歧视，与不必要贸易障碍的避免。ISO/IEC 导则支配国际标准机体自身的程序，并要求诸标准由参与成员之间的共识发展。ISEAL 良好实践守则为可持续性标准占据同一位置，规定对标准制定、对保证，与对影响度量的诸要求，并作为一个其诸对象是标准制定者的标准运作。

这一家族配得上注意，因为它不寻常。它由被致意于私人机体之规则制定、被无主权者所强加、并因未能遵守它们之诸机体在它们所要求其承认者之间失去信誉而被采纳的诸规则组成。它是§13那个反身策略的最清楚现存实例，而它运作的条件恰恰是§19所辨认的那个利害。

16.2 要求内部申诉机制的标准

一个第二家族要求它所适用于的诸组织确立诸受它们行为影响者可借之提起投诉的渠道。这一家族是这一概览中任何工具里同§21之问题最直接相关的，因为它以规范术语规定一个内部机制必须像什么样。

《联合国工商业与人权指导原则》（联合国 2011）在原则31陈述一个非司法申诉机制之有效性据以被评估的诸判据。这样一个机制应是正当的，在使那些为其使用而设之利益相关者群体的信任成为可能并对诸申诉过程的公平进行负责这个意义上。它应是可及的，被它意在服务者所知晓，并向那些面对准入障碍者供出协助。它应是可预测的，供出一个清楚且被知晓的程序，带一个指示性时间框架并对可得的诸过程与结果类型以及监测实施的手段有清楚。它应是公平的，寻求确保受屈诸方对以公平、知情且互敬的条件参与一个申诉过程所必需的信息、建议与专业知识之诸来源有合理的获取。它应是透明的，使诸方对进展保持知情，并供出关于表现的充分信息以建立对它有效性的信心。它应是权利相容的，确保诸结果与救济符合国际公认的人权。它应是一个持续学习的来源，汲取相关措施以辨认改善那个机制并防止未来诸申诉的诸教训。而一个操作层面的机制应基于参与和对话，就它的设计与表现咨询那些为其使用而设的利益相关者群体，并聚焦于对话作为致意并解决诸申诉的手段。

那个公平性的判据值得被分离出来。它要求一方对以知情条件参与所需要的信息、建议与专业知识有获取，而那是对§3所引入之信息不对称的一个直接致意，并至少在形式上，对§6所辨认之关于比较术语下诸判定的困难的一个致意。

《经济合作与发展组织关于负责任商业行为的跨国企业准则》（经合组织 2023）供出一个伴随安排，其中遵守的诸国维持国家联络点，接收对着诸企业提起的具体事例并供出通往它们解决的斡旋。国际金融公司的绩效标准要求一个为受影响社区的申诉机制作为融资的一个条件，而赤道原则把那个要求带入商业项目贷款。ISO 10002 给出关于一个组织之内投诉处理的指引而 ISO 10003 关于外部于它的争议解决。

16.3 支配治理与合规的标准

一个第三家族致意那个组织自身的内部秩序化。ISO 37000 供出关于诸组织之治理的指引。ISO 37301 规定一个合规管理系统的诸要求并为认证而起草。ISO 37001 致意反贿赂管理系统。ISO 26000 供出关于社会责任的指引并明确地不意在认证。

这一家族对解释的相关性是间接且真实的。一个合规管理系统要求诸义务被辨认、诸责任被分配、诸判定被记录，而诸不符合被记录并被处理。它在被实施之处的效果是建造那个§5发现先例与系统模式所要求、而§9辨认为更约束性诸排序所依赖之基础设施的记录。

16.4 合格评定与认可

标准通过认证符合它们的机器而有约束力，而那个机器本身被标准化。ISO/IEC 17021-1 陈述对供出管理系统之审核与认证之机体的诸要求，而 ISO/IEC 17065 对认证产品、过程与服务的机体。ISO/IEC 17011 陈述对那个评估认证者之认可机体的诸要求。认可机体之间的相互承认通过国际认可论坛与国际实验室认可合作组织被组织，以致一个在一个国家认可之下发出的证书在其他之下被承认。

这个机器的两个特征关乎当下主题。

对认证机体的诸要求包括对着认证判定之申诉与投诉的被记录过程，故一方对一个认证者的判定不满意有一个被规定的渠道。那个解释控制装置因而被递归地应用于那些施行它的机体，而那是一个这一概览中没有其他尺度展示的安排。

而那个评估的对象是一个管理系统的符合。一个审核员审查那个组织是否已确立所要求的诸过程、是否遵循它们，以及是否记录它所做之物。那个审查不延伸到该组织所作的任何特定判定是否正确，而它不为此被设计。

16.5 诸发现与残余

本节所概览的诸工具是这一概览中唯一要求诸组织确立它们自己的判定可由之从内部被质疑之渠道、并以规范细节规定这样一个渠道必须像什么样的家族。任何人寻求一个关于一个内部解释质疑机制应当像什么样的现存陈述，将在原则31以及伴随那些认证标准的诸保证要求中找到它。到那个程度这一尺度比此处所概览的任何其他更直接地回答那第四个问题。

五条界限附着，而它们是累积的。

采纳是自愿的。本节中没有工具迫使任何组织服从它，而一个谢绝的组织受制于无物。这些工具运作于那些已选择被约束的诸组织之上，而那个选择通常被作出，因为某个其承认那个组织所要求者已以它为条件设定交易。

那些判据本身是标准术语。一个机制是否可预测、公平，或可及是一个要求在每一使用场合被评估的问题，而它由一个把那些判据应用于一个案件的审核员回答。那个解释问题因而被重新安置到那个机制的评估并不被它消解，而那是§6为被发布判据一般所辨认的那个模式。

评估致意那个系统而非那个判定。这条界限是最严重的而它是结构性的。一个被认证的组织已示范它运作一个符合的过程。那个评估中没有任何东西确立一个经由那个过程所抵达的判定曾是正确的，而一个组织可持有当前认证而它所作的任何特定解释未被改正地成立。那个装置在§14的意义上是程序性的，而它完全地继承那一家族的特征性失败。

那些工具通过一个利害咬合。认证有价值，因为一个买家、一个贷方、一个采购规则，或一个监管者要求它，而在无一要求之处，那个标记值它耗费以获得之物、而别无其他。这是§19所辨认的那个预设，而它在此处以它最纯的形式在场，因为整个装置被建造于它之上并不作别样的伪装。

而那个标准制定者的构成要紧。为一个行业书写标准的诸机体通常在实质部分被那个行业充实，因为起草所要求的专业知识驻留在那里。那个后果，在§19中以选区对齐这个标题被处理，是那个标准与它所支配的实践被带一个对它们所共享之诸实践之共同利益的诸方所书写，而一个按那个共享实践所作的判定不太可能被那个对着那个标准认证的机体分类为一桩失败。

§17 行政复审与说明理由的义务

这一尺度是概览中最发达的。公共行政在一般规则之下判定高量的个别案件、那些判定直接关系那些受制于它们者，而诸法律系统已在一个世纪之上建造一个针对那个随之产生之问题的装置。那些工具在此按功能被分组。每一个系统把复审延伸得多远进入一个评估性判定之实质的那个比较问题被留给§18。

17.1 说明理由的义务

那个最广布的工具要求一个判定被一个它据以被抵达之根据的陈述所伴随。

德国行政程序法要求一个书面行政行为陈述它的理由，包括那个机关据以进行的本质事实与法律根据，并添加一个在此处有特别旨趣的条款：在那个行为停留于一个裁量之行使之处，那个陈述也应指示那个机关在行使它时据以进行的诸考量。这是这一概览中任何工具最接近于要求§5辨认为一个有界读法之检验的那个展示的东西，因为它逼那个机关暴露那条路线、而非那个结论独自。

法国法自1979年以来要求特定不利类别的个别行政判定被陈述它们所依据之法律与事实考量的理由所伴随，一个现被载于支配公众与行政之间关系之法典的要求。

在联盟法中为诸机构之行为陈述理由的义务由那部条约所强加，而《基本权利宪章》把行政为它诸判定说明理由的义务列入良好行政之权利的诸组分中，同被听闻的权利与获取一个人档案的权利并列。那部宪章也保障在一个法庭之前的一个有效救济。

在美国那个要求通过那个复审标准运作。一个机构行动可作为任意与反复无常被撤销，而在那个标准之下发展的被说理解释要求逼那个机构表述一个令人满意的解释，包括所发现事实与所作选择之间的一个理性联系。

英国没有一个为行政判定说明理由的一般普通法义务，而一个义务已被承认的诸范畴已可观地扩张，特别是在一个判定影响一个重要利益或没有解释便显得反常之处。

这一家族对当下主题的意义在§5被陈述并在此处被重复，因为这一概览确认它。一个读法在那个解释者展示所被汲取之物时作为有界可评估。说明理由的诸义务是那个要求的制度形式，而它们的强度随它们逼多少路线进入记录而变化。一个由一个结论与所应用之条款的陈述所履行的义务，把那个解释行为留得恰如它曾一样不可审查。

17.2 事先被发布的判据

一个第二家族要求那个行政在任何案件生起之前陈述它将据以评估一类案件的判据。那个装置把前瞻解释活动的一部分转换为被发布的规则，而它在日本行政法中以不寻常的明确被发展。

日本的《行政程序法》要求一个机构确立对诸申请之审查的判据并把它们公开，并分开地为支配不利处置的判据作出规定，以逼那个机构努力确立并发布它们的诸术语。那个强制性表述与那个努力表述之间的区分值得被标记，因为它精确地定位那个立法机关愿意约束之处与它退缩之处。那部法也要求当一个申请被拒绝以及当一个不利处置被作出时理由被示明。

那个装置是§6当作被发布判据之部分补救所处理者的制度实现，而那里所注意的两条界限都适用。那些判据由诸术语组成，其应用要求判定，故那个解释问题移动并不消失。而那个机构保留一个多具体的起草选择，故那个未定义术语所保全的余地可在一个隔层被再次保全。

17.3 程序性正当程序

一个第三家族把程序性资格附着于影响受保护利益的诸判定。在美国这些在来源上是宪法性的。一个在特定福利终止之前听证的要求在 Goldberg v. Kelly 案中被确立，而在一个给定案件中所应得的程序由 Mathews v. Eldridge 案的那个权衡探问所判定，它权衡那个受影响的私人利益、现有诸程序之下错误剥夺的风险连同额外保障的可能价值，以及那个包括行政负担的政府利益（马肖 1985）。

这一家族规管通往一个判定的路径。它确保通知、一个回应的机会，并在某些设定中一个中立的决定者，而它仅通过那个权衡探问中的错误风险因素致意那个判定的内容，那个因素问额外程序是否会减少错误、而不问这一判定是否曾在错误中。

17.4 裁决之前的内部复审

一个第四家族在那个初始判定与任何法院之间置入一个层级。

日本的行政投诉复审系统在过去十年被实质地修订以引入一个不曾涉入那个原初处置的听证官与一个其意见在关于那个投诉的判定被作出之前被征询的第三方复审机体。两个改变都致意§3辨认为递归的那个困难，因为它们置入不曾作出那个受质疑判定的诸人。可比的内部层级存在于韩国行政程序中，以及诸普通法法域中许多行政方案的内部复审阶段中。

一个不曾先前涉入之人的置入是这一概览中对§3之递归的最直接制度回应。它也是一个部分的，因为那个复审的人或机体停留在同一组织之内、应用同样的规则，并通常受制于同样的解释实践与同样的指引。

17.5 监察专员机构

一个第五家族调查对着行政的投诉而不判定法律权利。欧洲监察专员接收对联盟诸机构中不良行政的投诉，而诸国家监察专员办公室履行可比的功能。

两个特征在此相关。那个管辖由不良行政界定，它比违法更宽并够到行政的品质、延迟、无礼，以及未能遵循一个人自己的诸程序。而那个结果是一个没有约束力的建议。

那个组合值得被命名，因为它是一个其他家族不作的交易。那个监察专员能够到没有法院会复审的行为，包括甚多关乎一个机关如何处理一个事项之物，而能恰恰因为在法律上没有任何东西系于那个认定而如此做。管辖的宽广被以强制的缺席购得。

17.6 诸发现与残余

那个公共行政尺度供出比任何其他被概览者更多。诸判定必须通常被说理，而在一个系统中那些理由必须暴露关乎一个裁量之行使的诸考量。诸判据可被要求事先被安定并发布。程序性资格在受保护利益被影响之处附着。一个由不曾先前涉入之人所配备的内部层级可在任何法院被抵达之前被置入。而一个宽广管辖的办公室可在没有任何法律权利处于利害之中的情况下审查行政的品质。

三条界限在此被引入而一个第四被留给下一节。

那个装置被致意于公共机体。它的诸工具随公共权力的性质而来，而它们不适用于那些作出本文所关切许多判定的私人与跨国机体。一方与一个私人组织打交道通常没有这些的任何一个。

那些说明理由的义务在每一系统中被诸例外、被它们所不延伸到的诸判定类别，以及被应用于所给陈述的充分性标准所限定。一个由公式化背诵所履行的义务对有界性的评估供出无物。

那个内部层级停留于内部。置入一个不曾先前涉入之人减少§3所辨认的递归而不移除它，因为那个复审者受制于该组织自己的解释实践与那个塑造那个原初判定的指引。

而那第四条界限是那个支配整个尺度的。无论那个装置发展得多好，它对一个判定的够到取决于那个复审论坛愿意审查多少那个解释行为，而每一个被概览的系统把那个评估性判断的某一部分保留给那个决定机体。每一个系统所保留之物，以及那个保留延伸得多远，是§18的主题。

§18 尊让与学术判断的区域差异

§17所概览的装置够到一个判定，仅到那个复审论坛愿意看之远。本节比较跨若干系统够到多远，并取那个其中保留最完全的判定类别作为它的运行例子。学术判断被选取，因为每一个被概览的系统在某种度量上保留它、因为那个保留在可敬的诸根据上被辩护，而因为那一类别以一个纯的形式展示§6辨认为集中解释负担的诸特征。

18.1 对解释作为一个法律问题的复审

诸系统在门槛处就一条规则之诸术语的解释是否是一个复审论坛形成它自己观点的事项而不同。

德国行政法给出最强的肯定回答。制定法据以被书写的诸不确定法律概念被当作法律问题，而一个复审一个行政行为的法院自行判定它们的意义，不尊让那个机关的读法。那个制定法要素一侧的一个评价余地仅在有限且被列举的诸范畴中被承认，其中有考试评定、公务员的评价、多元构成之专家机体的判定，以及某些预断性判断。德国意义上的裁量附着于那个法律后果、而非那些制定法术语的意义，而它为逾越、为滥用，与为未能行使被复审。

美国近来经一条不同路线朝同一方向移动。那个要求法院尊让一个含糊制定法之可容许机构建构的学说在2024年被推翻，而复审法院现在就一个机构是否已在它的制定法权限之内行动行使独立判断，带一个机构同时且一贯的读法承载它推理所值之说服分量。对一个机构对它自己规章之读法的尊让以被收窄的形式存续，受制于2019年所确立的诸条件。

法国行政法按主题区分复审的强度，对某些判定应用一个正常控制而对其他一个受限控制，在其之下唯有一个明显的评估错误将被谴责。英国的一般不合理标准为对评价性判断的干预设一个高门槛，带比例原则在权利被涉及之处适用。

18.2 学术判断的保留

对着那个差异，学术评估的处理更均一，而诸系统之间的差别关乎那个保留让步所处的门槛。

在美国诸法院已判定一个真正学术的判定不可被推翻，除非它构成一个如此实质的背离被接受学术规范以致示范那个负责的人或委员会实际上不曾行使专业判断。学术开除中的程序性保护相应地比纪律性者中更轻，以那个学术评价不适应于对抗过程之诸程序工具的根据。

在英国一个学生可向那个依制定法设立的高等教育独立仲裁人办公室投诉，而那个办公室不审议学术判断的事项。诸法院同样已把学术判断的诸问题当作不适合于由一个法院判定，同时接受合同与程序的事项可是可裁判的。

在法国那个考试评审团的主权是一个安定的原则，而行政法院审查那个评审团的构成、程序的规范性，与明显错误，同时谢绝以一个对功过的评估替代。

德国是§19已依赖的那个例外。那个联邦宪法法院在1991年判定一个考生的一个可辩护且被有力说理的答案不可被评为错误，并因而约束诸考官对评估之实质的余地。那条路线是宪法性的、所讨论诸考试的职业后果是实质的，而那个效果是把其他系统所保留之物的一部分转换为一个复审机体形成它自己观点的事项。

日本供出最强的保留，并凭一个明确的学说。那个最高法院已判定一个大学的某些内部判定属于那个一般法律秩序之内一个局部社会的内部事务，而出于那个缘由外在于司法复审的范围，除非它们关乎一个人在一般公民法律关系中的地位。那个学说对本文有直接旨趣，因为它以司法术语陈述恰恰是§19在制度上抵达的那个命题：一个组织的内部判定可作为一个原则的事、而非仅仅实践的事，躺在外部裁决的够到之外。

18.3 那个比较被概括

系统	规则术语的解释	学术评估	干预的门槛
德国	法律问题，法院形成自己观点	余地被承认，被宪法约束	一个可辩护的答案不可被评为错误
美国	自2024年起独立司法判断	保留给专业判断	对被接受学术规范的实质背离
法国	强度按主题变化	评审团的主权	明显的评估错误
英国	在评价性问题上尊让	被排除于那个制定法仲裁人	不合理的高门槛
日本	在一般行政法中可复审	一个局部社会的内部事务	对一般公民地位的影响

18.4 那个比较的诸发现

三个发现随之而来，而那第三个是本文向前承载的那个。

每一个被概览的系统把实质评价性判断的某一部分保留给那个决定机体。那个保留不是任何法律文化的一个偶然，而在各处被以同样的两个根据辩护：那个复审论坛缺乏重做那个评估的能力，而那个决定机体在它自己专业事项中的自主本身是一个价值。两个根据都是健全的，而本文不争议其一。

诸系统在那个保留让步所处的门槛上实质地不同，而那个差异宽到足以确立那个保留的范围是一个选择、而非一个必然。德国对考官的宪法约束与日本的局部社会学说标记那个范围的两端，而两者都是发达系统中的安定法。一个把它自己保留之范围当作被给定的机构，是在把一个政策当作一个事实。

诸系统也在多少解释行为被逼入那个可复审记录上不同，而这个维度可同那第一个分开。一个系统可保留一个评估的实质，同时要求那些判据被发布、要求那些理由暴露所被依赖之诸考量，并要求那个评估被不曾先前涉入之人进行。§17的材料显示这些中的每一个独立于那个复审标准而变化。这对§21的判据要紧，因为那个判据所要求之物是一个问题保持可问，而一个问题在所被判定之物、以及据以之根据，处在某人能致意之一个记录上的程度上可问。一个实质的保留与一个完整记录相结合，把一个问题留下开放。一个实质的保留与一个赤裸结论相结合，闭合它。

18.5 那个比较的界限

一则告诫贴附整节。此处所比较的诸学说是诸法律系统的形式立场，而这一概览在§2所引入的意义上是学说性的。那些可得路线多常被使用、被谁、以什么代价，以及带什么结果是一些其上比较文献稀薄的经验问题。一个其学说许可穷究性复审而其诸程序使复审在实践上不可及的系统，从一个受影响一方的位置看去，类似一个完全保留那个问题的系统。

§19 不对称边界

本节履行那第四个问题并陈述那个实践概览被导向的结果。那个结果是对为回答那个问题所组装之整个装置的一个边界条件。先前诸节所审查的每一个装置运作于一个那个解释机体在被它自身之外某一方承认中所持有的利害之上，而每一个装置随那个利害衰减而失去它的抓力。在那个解释的权威与那个对承认的独立被一起持有之处，那些装置够不到。

19.1 那个共同预设

被概览的诸装置落入六个家族，而每一个可由它所运作于之物被刻画。

强制复审 通过把那个判定暴露于撤销来把一个代价附着于一个有缺陷的解释。它预设一个对那个机体有管辖的论坛，而它预设撤销耗费那个机体某样它重视之物。合格评定，包含认证、认可，与支配它们的诸标准，通过扣留一个买家、采购者、保险人，或监管者所要求的标记来附着一个代价。它预设一方以那个标记为条件设定交易。声誉装置，包含披露、发布、排名与报告，通过通知一个其看待那个机体需要的受众来附着一个代价。专业装置 通过危及在一个同侪共同体中的站位来附着一个代价，那个共同体的承认是那个机体的诸官员为执业所要求的。退出，在赫希曼（1970）的意义上，通过从一个依赖参与者的机体撤回参与来附着一个代价。合同装置，包含申诉条款、仲裁协议，与被协商的诸程序，通过使一项义务可执行来附着一个代价，而它们预设一个论坛、一个议价，与一个有站位曾达成它的对手方。

那些家族在那个代价的币种上不同而在它们的形式上一致。每一个供出那个解释机体一个好好解释的理由，凭把一个代价附着于糟糕地解释，而每一个通过一个其承认那个机体所要求的一方传递那个代价。称那一方为那个装置的 选区。一个装置的效力是三个条件的一个函数，它们全部必须成立：那个选区必须能观察那个解释行为、那个选区对承认的撤回必须把一个代价强加于那个机体，而那些施行解释行为的官员必须内化那个代价的某一部分。

19.2 由一个第三方的中介

这个结构的一个特征配得上被标记，因为它解释为什么保护能在每一装置作为被设计运作的同时失败。

那个选区通常是那个对着其那个解释被行使之一方以外的某人。那个受影响一方是一个申请人、一个成员、一个供应者、一个投诉者，或一个候选者。那个选区是一个法院、一个认证机体、一个购买者、一个专业协会、一个资助者，或一个公共受众。那个受影响一方的保护因而是 被中介的：它生起于那个受影响一方的申诉对那个选区可辨识、且那个选区把那个申诉当作一个撤回承认的理由之时。两个条件无一被那个装置的存在所保证。

两个后果随之而来。那个受影响一方的保护随那个选区在那一类案件中所取的旨趣而变化，而那个旨趣因与它们功过无关的缘由跨诸申诉不均地分布。而一个组织可满足那六个家族中的每一个装置，持有它的诸认证、经受住它的诸复审，并保留它受众的看待，同时一个特定解释未被改正地成立，因为没有选区有任何机缘登记那个特定案件。那些装置不因此就有缺陷。它们被建造去规训诸模式，而一个单一解释行为不是一个模式。

19.3 那个利害消失所处的诸条件

那些装置运作于其上的那个利害在结构性且共同常在的诸条件下趋近零。

第一个是 过剩。在寻求那个组织所分配之物的诸方数目超过它能分配之物之处，任何一方的失去耗费那个组织无物，而那个退出装置没有币种。一个接纳它申请者一小部分的机构就它所拒绝的每一个申请者而言处在这个位置。

第二个是 供给的独一性。在那个组织是那个方所寻求之物的唯一或近乎唯一来源之处，无论一个资格、一个执照、一个认证，还是一个职位，退出对那方首先就不可得，而那个会操作那个装置的威胁无法被形成。

第三个是 重复的不对称，以加兰特（1974）的术语。那个组织是一个重复参与者，持有那个解释实践、先前诸判定的记录、同复审诸论坛的诸关系，以及构造诸案件以保护那个实践的能力。那个受影响一方是一个一次性参与者，持有一个单一案件、无记录，而没有一个其中承认可能被扣留之进一步互动的前景。

第四个是 不可观察性。在诸判定被个别化、未发布，或保密之处，那个选区无法观察那个解释行为，而效力的第一个条件从一开始就失败。个别化判定是那个解释时刻的寻常形式，故这个条件是常见的。

第五个是 理由的可替代性。在一个被质疑的解释可被一个那个复审论坛将不审查的根据所替代之处，那个选区的观察不产出一个分类。一个背景缺乏相关性的判定，被质疑，可被一个那个领域强而位置少的判定所接续。第二个根据陈述一个比较评价性判断，而那是每一个被概览的复审论坛保留给那个决定机体的那一类判断。

第六个是 选区对齐。在那个选区由与那个被规训者同类的诸机体组成之处，一如一个被它所标准化者所充实的标准制定者，或一个由那个行业组成的专业机体，那个选区的利益与那个机体的利益在它们所共享的诸解释实践上重合，而承认的撤回对那一类案件不是一个活的可能性。

那些条件是独立的，而本文所关切的那个情形一起展示若干。一个分配一个稀缺好处、持有它所授予资格之近乎垄断、面对一次性诸方、个别地且无发布地判定、可在质疑时陈述一个比较根据，并对着一个它自己同侪的选区负责的机体，持有没有利害是那六个家族中任何装置能运作于其上的。

19.4 那个边界被陈述

那个结果可被陈述。

那个不对称边界。 每一个为一方对着解释不义之保护所发展的装置，运作于一个那个解释机体在被它自身之外一个选区承认中所持有的利害之上。这样一个装置的效力被那个利害的量级、以及那个解释行为对那个选区的可辨识性从上方约束。在那个解释的权威与那个对承认的独立被同一方持有之处，那个解释行为躺在那个被组装装置的够到之外，而对那个装置的任何同类添加都不改变那个位置。

三个限定属于那个陈述，而每一个限制它。

那个主张关乎那六个家族的装置、而非那个保护概念本身。它断言对一类工具的一个上界，而不断言一方在落于先前小节诸条件之内的每一个案件中都无救济，因为诸条件可在程度上在场。

那个主张不携带对任何工具的批评。那些装置为它们的范围被好好建构，而它们的范围是那个其中那个解释机体依赖他人的领域。大多数组织，大多数时候，占据那个领域。那个边界标记那个领域终止之处，而那些工具不因随它终止而有过。

那个主张是结构性的，故它不被更好的实施所补救。一个更严格的认证方案、一个更穷究的复审，或一个更透明的报告各在那个领域之内抬高效力而它们无一延伸那个领域，因为每一个继续运作于同一个利害之上。

19.5 对着那个边界的诸情形

两类工具显得逃脱那个结构，而两者都应被审查，因为第二个产出一个真正的部分例外。

第一个是由一个被规定不可容许根据的保护，反歧视法是其发达实例。这样的体制把一个代价附着于一个判定，无论那个决定机体对任何人的依赖，而它们直接迎击理由的可替代性，凭一旦一个初步案件被显示便把那个证据责任转移到那个决定者。那个表面的逃脱在它的范围之内是真实的，而那个范围以那个在此要紧的方式狭窄。那些体制运作于那个不可容许根据被事先规定、而那个投诉能在它之下被提起之处。一个一个解释被糟糕地作出、没有受保护特征在起作用的投诉，完全落于它们之外，而那个责任转移装置没有任何东西去转移。

第二个是宪法裁决，而它供出对着那个边界的最强情形。那个德国联邦宪法法院1991年的诸考试判定够到一个复审诸论坛曾当作被保留的领域，判定一个考生的可辩护答案不可被算作错误，并因而恰恰约束诸考官对那个尊让学说所保护之实质评价性判断的余地。这是对那个解释行为本身的穿透，在那个考试机体对那个考生无任何依赖的情况下被达成。

那个例外应被让步而后被度量。它的诸条件是苛求的：那个受影响一方的一个宪法权利被涉及、那个评价行为是一个标准化且复现的、容许可表述判据的种类、一个宪法级别的论坛可得，而那整体被安置在一个公共的考试系统中。在那个级别的一个权利不被涉及之处、在那个评价行为是比较且个别化之处，或在那个机体是私人或跨国之处，诸条件无一成立。那个例外确立那个边界是那个寻常制度景观的一个特征、而非一个逻辑必然，而它确立跨越它的价钱，即一个宪法权利、一个可复审判据，与一个足够级别的论坛。

19.6 所抵达的位置

两条路线保持开放，而本文取那第二条，同时记录那第一条是那个实践的。

第一条路线改变那个不对称。倘若保护取决于一个其承认那个机体所要求之选区的存在，那么保护可被通过构成一个来追寻：凭把一次性诸方聚合为一个带一个常在利益的机体、凭把那个解释实践附着于一个那个组织确实依赖之一方所持有的资助或认可条件，或凭使诸判定可观察以致一个存在的选区可看见它当下不能之物。这些中的每一个是可得的、每一个是集体行动或制度设计的一个形式，而无一被本文致意，本文的主题是关于一个解释行为可被说什么、而非一个选区如何可被组装。那条路线在此被命名，以致接下来所取的转向不被误认为一个没有实践之物剩下的主张。

第二条路线问一个解释做什么，鉴于没有一方可被定位去制裁它。它把谁施行一个代价的问题搁置一旁，并探问那个解释行为对它被施行于其内之关系的效果。这是那些生成性关系诸节所接手的问题，而此处所确立的边界是那个使那个问题值得问之物，因为它确立那个替代问题，即谁将强加一个代价，仅在一个最令人关切的诸情形所躺在其外之领域之内有一个答案。

§20 解释作为一个生成性关系行为

§19所确立的边界闭合探究的一条线并开启另一条。倘若那个问题是谁将把一个代价强加于一个糟糕地解释之机体，那个答案仅在一个最令人关切的诸情形所躺在其外之领域之内可得。本节就同一个行为提一个不同的问题，而随之的一节陈述那个回答它的判据。

20.1 对象的转移

在前面诸节中三个问题已被问于一个解释。它曾可容许吗，这问它是否落于那条规则之诸术语所许可的空间之内。它曾被公平地抵达吗，这探问那个程序。它曾造成损害吗，这探问那方此后的位置。每一个是一个真实的问题而每一个已被长篇处理。

此处所提的问题关乎那个行为发生于其内的那个关系。一个解释行为被一方施行于一个生起于两方之间的问题之上，而它把那个关系留在某种状况中。那个有旨趣的状况是那个关系产生同一问题之一个进一步解释的能力。一个行为可把那个能力留得完整、可收窄它，而可熄灭它，而那三个结果是评估的对象。

那个转移值得被谨慎地陈述，因为它易被误认为那三个已被问之问题的一个。它不是那个可容许性的问题，因为一个完全可容许的解释可熄灭那个能力，而一个不可容许的可把它留得完全完整。它不是那个程序的问题，因为一个一丝不苟的程序可终止于一个闭合那个关系的行为，而一个粗心的可把那个问题留得完全开放。而它不是那个损害的问题，因为一方可在一个把一切留下开放的行为之下遭遇糟糕，而可在一个永久安定那个事项者之下遭遇良好。

20.2 那个关系作为那个承载者

一个关系应是那个成问题性质之承载者，这要求§3所引入的那个承诺。一个术语的外延被固定于一个用法、目的与实践的网络之内，而一个解释行为的诸方占据那个网络之内的诸位置。那个解释之方对那个网络所含之物有贡献，因为每一个判定成为一个对下一个可得的用法。那个被解释于其上之方可有贡献、可被定位去在未来有贡献，或可被定位以致他们所说的没有任何东西根本进入那个网络。

那个成问题的能力现在可被定位。一个关系在那个规则之术语从中取其内容的网络对来自两方所占据之诸位置的贡献保持开放时，拥有产生一个进一步解释的能力。那个能力有三个组分，而它们能分开地失败。那个问题必须保持可问，这在那个事项被置于任何进一步考量之范围外时失败。那方必须保持被定位去问它，这在那方被那个判定本身从那个网络移除时失败。而那方所供出之物必须保持能够进入，这在那个网络之诸实践仅从一个位置接纳贡献时失败。

20.3 对那个再生能力的诸结果

那个解释行为立于对这个能力的三个关系之一。

它可 把那个能力留得完整。那个判定安定那个案件而那个问题保持活的：那个读法可被这方在另一个案件中、被另一方在一个相似案件中，或被那个组织本身在重新考量时再次提起，而它们任何一个所说之物能进入那个术语从中取其内容的网络。

它可 收窄那个能力。那个判定安定那个案件并封锁那个问题可能被重开所依据之根据的某一部分，无论凭安装一个其修订现在会要求打乱其他事项的读法、凭消耗那方会需要来再次提起它的诸资源，还是凭使那个读法作为一个读法更难被看见。

它可 熄灭那个能力。那个判定安定那个案件并把那个问题留得从那个受影响一方所占据的位置、以及从每一个相关地相似之位置不可问。

那第三个结果与一个可容许的解释、一个公平的程序，与一个适度的损害相容。那是本节存在以确立的那个发现，而它是那个许可把那个能力当作一个它自己权利中之评估对象的东西。

§21 可再生性判据

本节陈述本文的规范核心、标记它所排除之物、把它同一篇关于价值转换之姊妹论文所推进的判据相关联，并引入应用它的困难。

21.1 那个判据被陈述

可再生性判据。 一个解释在它熄灭它被施行于其内之关系产生同一问题之一个进一步解释的能力时是不义的。那桩不义在于那个能力的熄灭，而不在于那个判定的不利、不在于它被作出所自之位置的不对称，也不在于那个特定案件的终局性。在那个能力被收窄而非被熄灭之处，那桩不义在程度上在场，而那个程度是那个收窄的范围。

那个判据转在一个必须在任何东西从它随之而来之前被陈述的区分上，因为没有它那个判据会谴责一个组织所作的每一个决定。

21.2 一个案件的终局性与一个问题的闭合

一个决定必须在某个方面是终局的。一个能被无限重开的案件将永不被判定，而一个无法闭合一个案件的组织无法运作。案件的终局性是根本判定的一个条件，而那个判据不触及它。

一个问题的闭合是一个不同的事项。一个系统可持一个案件被终局地判定，同时把它据以被判定的读法留下开放以待修订：受制于在一个较晚案件中重新考量、受制于被一个处在相似位置的一方质疑、受制于那个术语从中取其内容的网络已被误读的论证。那个解释实践保持在运动中而那个个别案件安歇。

当下说明上的不义是那第二个、而非那第一个。那个判据所谴责之物是一个读法向一个它不再能作为一个读法被任何被定位去被它影响者提起之状况的过渡。那个个别一方可接受他们的案件结束了。他们所失去之物，在那个判据适用之处，是那个问题本身的站位。

21.3 熄灭的诸形式

那个能力可以两种方式被熄灭，而它们在恢复它们所要求之物上不同。

程序性熄灭 发生于一个问题可能被重开所经由的诸渠道被切断之处。时限届满、那个事项被置于一个被排除于复审之范畴内、那个判定未被发布以致没有较晚一方能援引它，或支配质疑的诸规则本身被解释以致排除那个质疑。这些中的最后一个是§3所引入的那个递归闭合，而它是那个对本主题独特的程序性熄灭形式，因为那个其行使成问题的权力是那个判定那个问题是否可被提出的同一权力。程序性熄灭原则上是可逆的，因为一个被切断的渠道能被恢复。

结构性熄灭 发生于那个判定移除那个会重开那个问题之方之处。那个申请人离去而那个关系随那个判定终结。那个成员被开除而持有没有站位去提起任何东西。那方立于其内的那个范畴被解散，故没有较晚一方占据一个相似位置。在每一情形中没有人被定位去再次问那个问题，而恢复一个渠道会恢复无物，因为那个会使用它之方不再在那个关系中。结构性熄灭不以同样的方式可逆，而它是那个更严重的形式。

一个第三条件应被标记，尽管它不干净地属于任一形式。那个能力可被 耗竭 熄灭，在那些渠道保持开放而那方使用它们的能力已被那个抵达那个判定的过程消耗之处。耗竭呈现为程序性的开放而作为结构性的闭合运作，而它常见到足以使一个忽视它的判据错过甚多它存在以捕捉之物。

21.4 同转换闭合判据的关系

一篇关于价值转换的姊妹论文（黄，编写中）推进一个同样形状的判据，而它们之间的关系应被直白地陈述，因为一个复现的判据要么是一个强处要么是一个重复，而那个读者有权知道哪一个被主张。

那里的对象是一个被书写跨诸范畴的等值，其中一方取一类的一个价值并返还另一类的一个价值，断言那两者等值。那个判据是闭合：那桩不义生起于那第二方不再能断言那个非等值之处，或它们可得的任何申诉在那个固定比率之内运作而无法够到那个比率本身之处。此处的对象是一个为一条规则之一个术语所固定的外延。那个判据是那个产生一个进一步解释之能力的熄灭。在两个情形中一个行为固定一个不曾被在先固定的内容，而在两个情形中那桩不义被定位于重开被固定之物之能力的失去。

那个共享的核心被认可。本文把那个判据延伸到一个第二对象，而所引入的主张是那个复现是那个判据的证据，因为一个跨如一个补偿支付与一个准入判定那般不同之诸行为辨认同一结构的规范检验，比一个被贴合于一个单一情形者对追踪某样东西有一个更好的主张。

三个特征区分当下的应用，而它们是本文对那个家族的添加。

第一个是递归。在那个转换情形中那个价值被转换之方可向一个不曾施行那个转换的论坛申诉。在那个解释情形中支配申诉的诸规则是同一组织的规则并被它解释，故那个成问题的权力延伸到它自己复审的诸条件。闭合因而能在没有任何刻意封锁行为的情况下、凭那个权力对着会约束它之诸规则的寻常运作被达成。

第二个是时间的够到。一个转换比率对它的诸方闭合。一个解释对一类闭合，因为那个读法作为一个用法进入那个网络并可得以支配同一种类的每一个较晚案件。被一个解释闭合影响的诸方之人口因而是无定的，并包括那些尚未生起且无法被咨询的诸方。

第三个是§19的那个不对称边界，那篇转换论文没有机缘去确立它。那里外部救济的可能性能被留下开放。此处已被显示那些可得工具被一个那个解释机体可不持有的利害所约束，故那个判据正为一类其中没有被概览种类之救济可得的案件被推进。那对一个判据是一个更难占据的位置，而随之的一节陈述它耗费什么。

21.5 那个判据的诸排除

一个判据凭它所谢绝谴责之物赢得它的内容，而这一个谢绝甚多。

它不谴责一个不利的判定。一方对着其一条规则已被解释，在这个说明上未曾遭受不义，无论那个结果多不受欢迎，只要那个问题从他们所占据的位置保持可问。

它不谴责那个解释位置的不对称。一方判定而另一方接收那个判定这是受规则支配之决定的构成性之物，而一个无法固定它自己规则之应用的组织无法判定。

它不谴责终局性，出于上面所给的缘由。

它不要求那个受影响一方的读法胜出，也不要求那个组织被说服，也不要求任何特定程序可得。它所要求之物是那个问题经受住那个答案。

它不谴责一个读法随时间的安定。一个已被提起、被质疑并被反复肯定的解释可通过使用变得安定，而那种安定是一个实践的寻常巩固。那个判据附着于提起那个问题之能力的熄灭、而非一个已经受住提起之答案的稳定。

21.6 应用的困难

两个困难附着，而两者在此被引入以致一个读者不必去发现它们。

那个判据回溯地比前瞻地以更多易性应用。一个重开一个问题的能力是否被熄灭在时间已流逝而没有人曾能重开它之后可回答。一个判定是否将熄灭它在那个决定的时刻仅凭臆测可回答，而一个善意行动的组织可能无法辨别它诸判定中哪一个是那个闭合的种类。一个仅在回溯中可用的判据对着其那个判定成立之一方是有限服务的。

而那个判据陈述一个要求而不陈述它如何被满足。持一个正义的解释把那个问题留得可问，是就什么在一个组织中随时间维系可问性不说任何东西，而那是一个制度条件的事、那些持有解释职位者之性情的事，以及那些不在场之诸方之站位的事。两个困难都被向前承载到那些开放的问题，在那里它们被置于它们可能被着手的形式中。

21.7 那个判据的位置

那个判据交付一样§19的装置不交付之物，而未能交付一样它们交付之物。

它所交付之物是无需一个选区的可评估性。一个解释是否把那个问题留得可问是那个行为与那个关系的一个性质，而它能被那个受影响一方、被一个观察者、被一个较晚一方，或被那个组织本身在反思中断言。它不要求任何人持有一个利害、任何论坛有管辖，或任何标记被扣留。在那个不对称边界留下没有什么可做之处，某样东西仍能被说，并被带一个确定内容地说。

它未能交付之物是任何后果。一个辨认一桩不义并供不出一方去对它行动的判据陈述了一个关于世界的事实并在它之中什么也没改变。随之的一节问从一个解释闭合是否有任何东西随之而来，在一个去对着它执行之方缺席的情况下，而那里所给的答案比那个问题配得上的更弱。

§22 委托于动态

§21的判据辨认一桩不义并供不出一方被定位去对它行动。本节问从一桩解释不义有什么随之而来，在没有这样一方存在之处。一个看的方式被提议，而本节的较大部分被给予它所不断言之物，因为对一个同情的读者可得的诸读法比那个主张更强并会经受不住审查。本节以一个初步作品的精神被供出：那个说明是可延伸的、它邀请修订，而它被定框以致它有用的部分可被一个丢弃其余的读者保留。

22.1 随之而来者的站位

本节是本文中最具思辨性的，而它的站位应在它被读之前被固定。随之而来之物不是一个结果。它是一个看的方式，被供出，因为§21的判据留下一个那个救济语汇无法致意的问题，而因为一个动态语汇容许那个问题被提出，即便它无法被回答。本节提议一个形状并撇清那个机制。一个被§19–§21说服而被此处一切不说服的读者，失去那些节所确立之物中无物，一个本文在§24中再次陈述的可分割性。

22.2 对生成性的壁障

一个组织是一个诸关系的结构：在它诸官员之间、在那个机体与它判定于其上者之间、在一条规则与被带到它之下的诸案件之间。§20所辨认的那个能力，凭之一个关系产生一个进一步解释，被分布跨那些关系，而它是那个组织借之把它的诸读法保持贴合于不静止之诸条件的能力。

解释不义壁障那个能力。它通过对那个组织之关系语法的诸操作如此做，那些操作是异质的而本文不主张去列举它们：一条规则可被添加，它封锁一个问题、一个曾使修订成为可能的实践可被撤回、一个会承载一个质疑的站位可被弄得不可适用、一个曾把一个申诉连到一个判定的渠道可被切断。这些是带不同近侧形式的不同操作，而此处所供出的说明抽象于它们之上。它所留意之物不是哪个操作成立、而是它们所共享的那个功能签名：那些受影响的关系失去，在某种度量上，产生一个进一步解释的能力。那个主张被刻意地置于那个功能层次，而它的一般性是那个要点。倘若下面所勾勒的那个后果随之而来，它从生成性的失去随之而来、而非从带来那个失去的任何特定方式，故那些操作的异质性是那个说明的一个特征、而非它之中的一个缺口。

22.3 从壁障到漂移

一个曾在运作的生成性借持续的小调整把那个组织保持贴合于一个移动的环境，每一个古怪案件被一个不曾存在过的读法回答。被壁障，那个调整为那些受影响的关系停止，而它们曾被贴合于的那个环境不停止移动。一个距离在那个组织被组织去处理之物与那个够到它之物之间开启，而那个距离在不义施行任何进一步行为的情况下扩宽，被那个世界的寻常运动所承载。没有任何东西宣告那个扩宽，因为那个会登记它的渠道在那个壁障的诸伤亡之中：一个没有一方能提起的问题产生没有投诉，而投诉的缺席从内部呈现为一个问题的缺席。

在那篇关于历史动态之伴随作品（黄，编写中，b）的语汇中，同样可表达为一个吸收扰动之能力的衰减。一个构型的稳定是它吸收扰动并返回它形式的能力；一个被壁障的生成性是那种的一个被减少的能力；而一个其吸收能力已衰减的构型立得更靠近一个边界，在那里一个它先前会吸收的扰动把它带往别处。那个说明所供不出之物，以及应在那些路线被铺陈之前被让步之物，是那个扰动本身。一个能力的衰减不是一个力。它使一个构型脆弱而不推它，而一个扰动是否以及何时到来发现它脆弱，外在于那个不义说明所判定的任何东西。

22.4 从壁障到脆弱的一些路线，作为示例被供出

从一个被壁障的生成性到一个衰减的吸收能力的那一步可沿若干路线被追溯，而四个在此被铺陈。它们作为那一步可能如何走的示例、而非作为那个机制被供出，那个机制那个说明已撇清。它们既非穷尽的也非共同必要的；一个拒绝全部四个的读者仍可持有§22.2的功能主张，而一个添加一个第五的读者正做那个说明所邀请之物。它们在强度上并在为它们可得的证据上不同，而它们被以支持的上升次序给出。

错误改正的失去。 一个对生成性的壁障移除一个渠道，一个组织通过它得知一个读法已停止贴合。它作为一条路线的弱处是一个组织有其他信息来源，而一个读法可长时间不贴合而没有任何东西依赖那个贴合。

自愿遵从的撤回。 规则系统被以那些受制于它们者的遵从来施行，而对着不遵从者的执行是昂贵的。泰勒（1990）确立遵从实质地转在那个权威被行使所经由诸程序的被知觉公平上，而比它转在诸结果的有利性上更强地如此。一个判定诸问题并封锁它们重开的实践是一个那些受制于它者有理由怀疑它公平的实践。在这条路线上解释闭合抬高治理的代价，持续地并且没有任何单一插曲是决定性的。这条路线有那第一条所缺的经验支持，而它不依赖于那个组织未能注意到任何东西。

选择与多样的失去。 无法质疑一个判定的诸方在退出可得之处退出，而从外部观察那个实践的诸方谢绝进入。那些留下的不成比例地是那些对之那个闭合读法无可反对者。那个组织之领域内被代表之诸位置的多样因而下落。艾什比（1956）以一般形式陈述那个相关原则：一个调节器必须支配至少等于它所要处理之扰动之多样的多样。一个其解释实践减少它所遇多样的组织正减少它所能回答的多样，而赫希曼（1970）供出那个减少发生所经由的机制。

潜在积累。 一个被闭合的问题不为持有它之方停止是一个问题。诸申诉持续而不登记，而它们未被登记地持续，恰恰出于那个使闭合成为闭合的缘由：没有渠道存在，它们会经由之被记录。这样诸申诉的存量是那篇关于历史动态之伴随作品（黄，编写中，b）意义上的一个慢变量，而它的增长在一个组织所查阅的诸指标中不可见，因为那些指标被闭合已切断的诸渠道所喂养。

那四条路线不是那个说明必须在其间选择的诸替代，而在多于一个运作之处它们复合：一个失去它错误改正的组织可能也在抬高它治理的代价、收窄它所面对的多样，并积累它所无法看见之物。本文不主张它们任何一个在一个给定案件中运作。它主张一个对生成性的壁障是这类路线可能从中奔跑之物的那种东西，而它把在一个特定组织中奔跑之诸路线的辨认留给它不承担的工作。

一个局限适用于每一条路线并属于此处、而不仅属于§23。每一个描述一个构型吸收扰动之能力的衰减。无一供出一个扰动。从一个变浅的盆到一个离开它的过渡要求某样东西到来，而那个不义说明中没有任何东西判定任何东西会，或何时。这是为什么那个说明预测，作为它的寻常情形，长的间隔，其中一桩不义已被做、那个吸收能力已下落，而完全没有任何东西随之而来。

一个最后的点关乎倘若一个扰动确实到来发现那个能力被耗尽会发生什么，而它回答一个那个动态语汇提起并不明显安定的问题：为什么那个结果应是重组而非毁灭。那个对生成性的壁障运作于那个组织的关系语法之上，于那些秩序化它诸部分的规则、站位与渠道之上。它把那些部分本身留得未触动。那些官员、那些资源、那个累积的胜任，以及那个组织所服务的人口全部持续贯穿那个过渡。那个累积的不贴合所释放之物，当它释放时，是那个秩序化而非那个材料，而同样的诸部分在一个被改变的语法之下重新凝聚：一条被修订的规则、一个被改变的站位、一个被重构的机体。从一个稳定构型的离开在这个说明上是一个持续之诸要素集合的一个重新形成，而那是一个被推离一个吸引子的系统朝另一个重组而不只是停止的那个意义。那个说明沉默，并应沉默，就那个接续的构型是否是一个改善、就那个过渡花多长时间，以及就诸可得构型中哪一个被抵达。

图1。 一个示例，而非一个模型。在（a）中一个构型占据一个盆，它的深度代表它在一个扰动之后返回的能力；随一个对生成性的壁障积累，那个盆变浅，而越过一个门槛一个寻常的扰动把那个构型带入一个相邻的盆。在（b）中同一个过程被对着被积累的壁障绘制：那个能力朝无物下落而那个可观察状态移动得极少，以致那个衰减不被一个组织会查阅的诸指标所登记。那个图从一个为可辨识性所选的一维势被绘制并不作经验主张；维数、那个景观的形状，以及一个真实组织中任何边界的位置全部未知。三条告诫被建进那个绘制。那第二个盆不携带评价内容而不是那个更好的一个。那个图景显示离开并就哪个构型被抵达不说任何东西。而那个过渡是持续诸部分之间那个秩序化的一个重新形成、而非对它们的一个毁灭。

那个图陈述那些路线所示例的那个图景并在同一口气中陈述它的诸界限。左面板给出一个稳定景观的那个既得意象，其中一个盆的深度代表一个构型在被扰动之后返回的能力。右面板承载本节所推进的那一个主张。在一个对生成性之壁障之下衰减之物是一个能力，而一个能力不是一个输出。一个查阅它自己诸指标的组织看见一个稳持的状态，而那个稳定与那个能力已被几乎完全消耗相容。这是为什么一个组织被糟糕地定位去注意此处所描述的状况，而它以动态术语重述§3所引入的那个认识论点：一个对生成性之壁障所移除之物中有那个借之它自己诸效果会被登记的工具。那个图凭使那一个点可见赢得它的位置。它不应被读为断言任何关于一个组织之状态的真实维数之物，那是那个说明所不知道的。

22.5 那个说明必须承载的一个复杂

一个反对立即生起而它是一个好的。闭合移除那个改正渠道，而它也移除那个渠道所承载的诸扰动。一个听不到投诉的组织受制于更少的活扰动，而那种的安静是稳定的一个形式。表面上看闭合可使一个构型更稳定而非更不。

那个反对部分正确而那个让步应被引入。闭合所移除之物是那个组织回应通过它面前诸方所传达之变化的能力。那些方所栖居的诸条件继续变化，无论那个组织是否得知它，而扰动通过它判定于其上诸方之投诉以外的诸路线够到一个组织：通过它所依赖的诸机体、通过它运作于其内的环境、通过那些不曾能说任何东西之诸方的最终行为。闭合因而以对应的代价购得安静，而那个构型同时变得更少被扰动与更少被贴合。那个对稳定的净效果是否是一个衰减是一个那个说明无法一般地安定的问题，而此处所引入的主张被限于那个回应的能力，闭合无保留地减少它。

22.6 那个临界条件

一个扰动仅在那个构型已立得靠近一个边界之处把它带出它当下的形式。远离这样一个边界，扰动被吸收而那个构型返回，而一个扰动的到来产生没有一个观察者会记录之物。

那个对当下说明的后果是严峻的并应被如此陈述。一个组织可不义地解释、闭合它诸解释所提起的诸问题，并在那个状况中持续一个极长的时间而完全没有后果。那个说明预测这个，并把它作为那个寻常情形预测，因为大多数构型在大多数时候远离任何边界。任何不义可靠地产生它自己撤销的读法被那个说明本身所反驳，它供不出扰动也供不出一个一个扰动会到来所依据的时间表。

22.7 关于方向、稳定与时机的免责声明

那个说明所不说的三样东西被引入下面，每一个作为一个否认，因为每一个是一个同情的读者可能代那个说明供出之物。

它供不出方向。 一个离开它当下形式的构型抵达另一个，而那个机制中没有任何东西判定哪一个。跟随一个不义安排之重新秩序化的重组可能安装一个更糟糕的安排，而那个历史记录含有许多这样的序列。那个说明描述一个构型失去它吸收扰动的能力并就什么接续它不说任何东西。一个把那个机制当作意味不义组织在适当时候被更好者替代的读者，添加了一个那个机制不含有的前提，而那个被添加的前提是假的。

稳定不携带评价内容。 一个构型持续这是一个关于它吸收扰动之能力的事实并不是关于它是否可辩护的任何证据。一个正义的组织可稳定、一个不义的可稳定，而每一个可不稳定。那个点逆着当下说明如它顺着它一般切。一个已闭合它诸问题的组织可能比一个不曾如此的持续更久，而它的持续会不是辩白，恰如它的崩溃会不是判决。

它供不出时机。 那个机制陈述没有比率。一个闭合与任何重组之间的间隔未被规定并可能超过每一个受影响之方、每一个负责的官员，以及那个组织之当下形式的工作寿命。一个今天被一个闭合冤枉的一方在任何关乎他们的地平线上被保证无物。

22.8 同报应的距离

那个机制在轮廓上类似一个熟悉的道德图景，其中诸冤屈最终被偿还，而那个相似是表面的。那些差别值得被铺陈，因为那个图景有吸引力并会毁掉那个说明。

报应要求一个偿还的行动者，而那个说明不含有行动者。它要求那个冤屈与随之而来者之间的比例，而那个说明供不出，因为一个重组的量级是那个构型状况的一个函数、而非任何行为之严重性的。它要求随之而来者落于那个负责之方，而那个说明供不出任何瞄准：一个重组落于任何在它到来时在场者之上，其中将包括那些他们自己被那个闭合冤枉之诸方与那些继承一个他们不曾作成之实践的官员。而它要求可靠性，而上面所引入的那个临界条件已撤回它。

一个进一步的告诫属于此处，被致意于那个说明的使用、而非它的内容。一个被那个说明说服的读者可能断定对着解释不义的行动是不必要的，因为那个结构将在适当时候为它负责。那个断定不从此处所断言的任何东西随之而来，而它被那个临界条件与上面所引入的关于方向的免责声明所反驳。那个说明描述在行动缺席时发生什么。它供不出扣留行动的理由，而把一个描述性主张转换为一个安慰会是那个描述性主张的一个糟糕使用。

22.9 那个主张的站位

那个主张是一个关于一个机制的猜想并不是一条定律。诸组织是否展示此处从其他领域借来的稳定边界附近的那个行为是未安定的，而那个借用被以那个根据作成，即那个论证的结构没有它可理解并从它获得精确。所引入的命题是一个解释实践的重组在一个构型吸收扰动之能力已衰减之处比在它不曾之处更可能，而那比任何关于必然性的主张更弱并是那个论证所要求之物。

两条界限贴附那个猜想。没有形式模型被供出，而那个弃绝是刻意的，出于那篇伴随作品中所给的缘由：一个被书写去伴随这类论证的模型会供出一个那个论证不曾赢得的精确表象，而那个机制容许以散文的诚实陈述。而没有一方能可靠地判定一个构型立得多靠近一个边界，从它内部或从外部。那个说明因而产出没有那种会劝告一个组织保持一个安全距离之形式的操作指引，因为那个距离不可观察，而它所推荐的任何实践必须是一个不要求知道那个边界躺在何处的。

22.10 那个存留的主张

在那些否认之后某样东西存留，而它是适度且可用的。

那个说明供出一个描述，在其之下解释闭合是对施行它之组织的一个损失。在那个寻常描述上，闭合是一个被强加于一方的代价与一个对那个机体的便利，而那个机体的利益在于维持它。在此处所供出的描述上，闭合移除那个组织自己借之得知一个读法已停止贴合的手段，而那个组织是那个将在一个无定时期内运作于它不再能改正之诸读法上的一方。

这的意义是它不要求一个选区。§19中所概览的每一个装置借供出一个好好解释的外部理由而运作，而每一个在没有外部一方持有一个利害之处失败。一个当下种类的理由是内部的：它对一个不依赖任何人并考量闭合对它自己保持贴合于它运作于其内之诸条件之能力做什么的组织可得。这是那个不对称边界不消除的一个形式的杠杆，而§23引入它也失败所处的那个条件。

§23 该说明的界限

那个说明现在完整到足以被定价，而那个价钱是高的。本节陈述什么不曾被交付。

没有执行。 那个判据辨认一桩不义并装备没有人去对着它行动。在§19的装置适用之处，它们强加诸代价并确保诸救济；当下的说明什么也不强加什么也不确保，而它不在它们起作用之点补充那些装置。

对一个当下之方没有救济。 一个对着其一个闭合成立之方从这个说明获得一个对被做于他们之物的描述并没有一个它可能被撤销所经由的路线。那个描述可能有某种价值，而它不是一个救济，而本文不应被读为供出一个。

在那个推动它的情形中没有保护。 这是诸界限中最尖锐的。那个说明被发展，因为那些被概览的工具被一个最令人关切之诸机体可不持有的利害所约束。那个说明也不保护一方对着这样一个机体。它描述那个行为与它对那个机体的诸后果，而描述是它交付以替代保护之物。

回溯的应用。 一如§21所引入，闭合在时间已流逝之后被以信心可辨认，而在那个决定的时刻难以辨认，那是那个辨认它会要紧的时刻。

没有操作形式。 那个判据转在一个关系产生一个进一步解释的能力上，而没有程序被供出，那个能力可借之在一个给定案件中被评估。什么会算作一个问题保持可问的证据，以及那个评估会被谁作出，是未安定的，而一个在那个状况中的判据是一个工作的提议、而非一个工具。

那个内部理由有一个条件。 那个在关于什么存留之小节中所辨认的杠杆要求一个组织权衡它自己在一个长到足以使不贴合要紧之时期内保持贴合的能力。那些其地平线比那个时期更短的官员持有没有这样的利益，而一个其决策被分布于这样诸官员之间的组织可能无法持有那个利益，即便每一个官员会在抽象中承认它。那个内部理由因而在一个不同的层次重新引入它被供出以逃脱的困难的一个版本：一个在那个机体之长程状况中的利害仍是一个利害，而它也能缺席。

两样东西被对着这个清单主张。第一个是那个判据陈述，带一个确定内容，一桩那个既得之分配、程序、诠释学与共和主义诸概念的语汇不分离出来的不义，而那个分离值得拥有，独立于什么能被以它来做。第二个是§19的那个边界是一个关于世界的结果、而非本文的一个失败，而一个停在那个边界所停之处的说明正确地报告一个界限，而一个提议一个被概览家族之进一步装置的说明会回避它。这些是否充分是那个读者的事。

§24 开放的问题

本文辨认一个被忽视的结构。它不完成一个关于它的理论，而此处所记下的诸问题被持为所供出之物中更有价值的部分，因为一个值得命名的结构值得其他人对之工作。

倘若那个结构是真实的，下列诸问题立即生起。

那第一组关乎那个解释的权威。

Q1。 解释站位是可委托的吗？一个组织可把判定它自己规则意味什么的权威授予一个被它自身之外的机体，而不清楚这样的授予是否逃脱§3所引入的递归还是重新安置它，因为那个授予的诸术语本身是规则，而它们的解释必须被持于某处。

Q2。 什么使一个解释权威的行使正当？候选根据包括被那些受制于那些规则者的授予、在那个规则之术语从中取其内容之实践中被证明的能力、被受影响诸方在那个判定中的参与，以及被判定之物的持续可修订性。这些不等值并可能冲突，而它们之间没有排序在此被提议。

那第二组关乎解释的诸模式与诸对象。

Q3。 那个可容许者的边缘能被在事前定位吗？§3持诸可容许读法的空间是有界的，而定位那个边界在一个读法已被显示与那个网络不相容之后是直截了当的，而在那个显示之前是困难的。什么会许可一个在那个决定时刻所作、一个被提议读法躺在那个空间之外的主张？

Q4。 评价性谓词在种类上不同于其他开放织理术语，还是仅在落于那个半影之诸案件的比例上？倘若在种类上，什么标记那个差别，而倘若在程度上，是否有一个比例，越过它一个术语根本停止作为一条规则运作。

Q5。 §3的那个递归容许诸程度吗？一个组织可能把支配质疑的诸规则从那些规则所约束的权力绝缘，凭把它们固定于修订之外、凭把它们的解释托付于别处，或凭其他手段，而它是开放的，任何这样的安排是否减少那个递归而不把整个事项转移到那个组织之外。

那第三组关乎保护，而从§19的边界随之而来。

Q6。 在什么条件下一个选区能被构成？§19.6辨认本文不取的那个实践路线，即通过聚合个别地持有没有利害之诸方，或通过把那个解释实践附着于一个那个组织确实依赖之一方所持有的条件，来供出那些装置所要求的利害。什么判定这样的聚合是否成功是一个为集体行动与制度设计之研究的问题，而那个答案直接关乎那个边界有多少被固定。

Q7。 那个考试例外一般化到多远？§19中所讨论的那些德国判定够到一个可比诸论坛谢绝复审的实质评价性判断。什么类的评价行为容许可表述到足以支持那种复审的判据、什么判定那个类的成员资格，以及比较判断能否曾属于它，是开放的。

Q8。 解释闭合能被前瞻地辨认吗？§21在回溯中带信心适用。一个能在那个决定时刻辨认一个判定是那个闭合种类的说明会对一个善意行动的组织有用，而此处所供出的没有任何东西供出它。

那第四组关乎§22中所提议的动态，在那里诸问题是最尖锐的，因为那些主张是最弱的。

Q9。 解释闭合抬高还是降低一个构型吸收扰动的能力？§22.5陈述那个复杂并谢绝安定它：闭合移除一个改正渠道并移除那个渠道所承载的诸扰动。哪个效果占主导，以及那个答案是否随那个扰动的来源变化，是未解决的并是那个动态主张最被暴露之点。

Q10。 在什么条件下一个解释闭合根本登记？倘若诸扰动仅在一个稳定边界附近有后果，那么大多数闭合产生没有可观察之物，而那个说明给出没有辨别哪些是那另一种类的方式。是否有任何可观察签名伴随对这样一个边界的接近，在一个组织中相对于那个语汇从中借来的物理系统而言，是开放的并会值得拥有。

Q11。 什么会算作支持或反对那个机制的证据？§22的那个猜想被以一个抵抗检验的形式陈述，因为它预测长的无效果间隔并规定没有比率。一个能被否证的表述会是对所供出者的一个改善，而供出一个是本文留下未做的工作。

那最后一个问题是那整个说明被导向的一个，而它被置于最后，因为上面一切关乎它。

Q12。 在什么制度、社会与关系条件下一个解释实践能保持再生性？§21的判据陈述一个要求并就什么满足它不说任何东西。一个希望它诸问题保持可问的组织会需要知道什么跨人员的诸变化、跨诸读法安定入用法，以及跨决策所生成的朝向终局性之寻常压力维系可问性。这是那个说明之实践价值所依赖的问题，而它完全是开放的。

一个关于那个清单形状的收尾观察。那前三组的诸问题能在既存学科之内被追寻，而若干正在那里以其他描述被追寻。那第四组的诸问题不能，而它们是此处所供出之说明立于或倒于其上的那些。一个发现§21的判据有用而§22的机制不能说服的读者，可取那第一个而留那第二个，因为那个判据被独立地陈述并不停靠于那个动态。那个反面不可得：那个机制是一个关于从闭合有什么随之而来的主张，而它要求那个判据去说闭合是什么。

---

一则关于方法的附记。 没有形式装置被使用，而那个弃绝是刻意的。§22的机制能被以动态系统的记号书写，带一个构型对应一个制度状态之空间中的一个点、一个改正渠道对应一个反馈项，而一个闭合对应一个的移除。被写出，这样一个模型会推导没有那个散文不陈述之物，而它会供出一个那个论证不曾赢得的精确表象。那个记号会变得诚实所处的诸条件被在 Q9 到 Q11 中铺陈，而它们在此不曾被满足。

一则关于来源的附记。 被概览的诸工具按周期被修订而那些学说受制于被推翻。关于诸标准、诸制定法与判例法之内容的诸陈述被给出如书写之日并应在被依赖之前被核实，而在所讨论法域之一中一个长久安定尊让学说的近期取代是那个告诫不是形式性的一个提醒。

参考文献

Ashby, W. Ross (1956). An Introduction to Cybernetics. London: Chapman & Hall.

Barak, Aharon (2005). Purposive Interpretation in Law. Trans. Sari Bashi. Princeton: Princeton University Press.

Bobbitt, Philip (1982). Constitutional Fate: Theory of the Constitution. New York: Oxford University Press.

Bobbitt, Philip (1991). Constitutional Interpretation. Oxford: Blackwell.

Brandom, Robert B. (1994). Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press.

Cappelen, Herman (2018). Fixing Language: An Essay on Conceptual Engineering. Oxford: Oxford University Press.

Cover, Robert M. (1983). Nomos and Narrative. Harvard Law Review, 97(1), 4–68.

Dworkin, Ronald (1986). Law’s Empire. Cambridge, MA: Harvard University Press.

Endicott, Timothy A. O. (2000). Vagueness in Law. Oxford: Oxford University Press.

Eskridge, William N., Jr. (1994). Dynamic Statutory Interpretation. Cambridge, MA: Harvard University Press.

Fine, Kit (1975). Vagueness, Truth and Logic. Synthese, 30(3–4), 265–300.

Fish, Stanley (1980). Is There a Text in This Class? The Authority of Interpretive Communities. Cambridge, MA: Harvard University Press.

Fricker, Miranda (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press.

Frege, Gottlob (1997). The Frege Reader. Ed. Michael Beaney. Oxford: Blackwell.

Fuller, Lon L. (1958). Positivism and Fidelity to Law: A Reply to Professor Hart. Harvard Law Review, 71(4), 630–672.

Fuller, Lon L. (1969). The Morality of Law. Rev. ed. New Haven: Yale University Press.

Fuller, Lon L. (1978). The Forms and Limits of Adjudication. Harvard Law Review, 92(2), 353–409.

Gadamer, Hans-Georg (2004). Truth and Method. 2nd rev. ed. Trans. Joel Weinsheimer and Donald G. Marshall. London: Continuum. [German orig. 1960.]

Galanter, Marc (1974). Why the “Haves” Come Out Ahead: Speculations on the Limits of Legal Change. Law & Society Review, 9(1), 95–160.

Haslanger, Sally (2000). Gender and Race: (What) Are They? (What) Do We Want Them To Be? Noûs, 34(1), 31–55.

Hallaq, Wael B. (1984). Was the Gate of Ijtihad Closed? International Journal of Middle East Studies, 16(1), 3–41.

Hallaq, Wael B. (1997). A History of Islamic Legal Theories: An Introduction to Sunnī Uṣūl al-Fiqh. Cambridge: Cambridge University Press.

Hart, Henry M., Jr., and Albert M. Sacks (1994). The Legal Process: Basic Problems in the Making and Application of Law. Ed. William N. Eskridge, Jr., and Philip P. Frickey. Westbury, NY: Foundation Press. [Tentative ed. 1958.]

Hart, H. L. A. (1958). Positivism and the Separation of Law and Morals. Harvard Law Review, 71(4), 593–629.

Hart, H. L. A. (1994). The Concept of Law. 2nd ed. Oxford: Clarendon Press. [1st ed. 1961.]

Hildebrandt, Mireille (2015). Smart Technologies and the End(s) of Law: Novel Entanglements of Law and Technology. Cheltenham: Edward Elgar.

Hirschman, Albert O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press.

Huang, Wanhong (in preparation). Conversion Injustice: A Generative Relational Account of Value Conversion.

Huang, Wanhong (in preparation, b). Generative Relational History: Historical Generativity under Constraints.

Kelsen, Hans (1967). Pure Theory of Law. Trans. Max Knight. Berkeley: University of California Press. [German 2nd ed. 1960.]

Kripke, Saul A. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press.

Kripke, Saul A. (1982). Wittgenstein on Rules and Private Language. Cambridge, MA: Harvard University Press.

Llewellyn, Karl N. (1950). Remarks on the Theory of Appellate Decision and the Rules or Canons about How Statutes Are to Be Construed. Vanderbilt Law Review, 3(3), 395–406.

Luhmann, Niklas (1969). Legitimation durch Verfahren. Neuwied: Luchterhand.

Mashaw, Jerry L. (1985). Due Process in the Administrative State. New Haven: Yale University Press.

Masud, Muhammad Khalid, Brinkley Messick, and David S. Powers, eds. (1996). Islamic Legal Interpretation: Muftis and Their Fatwas. Cambridge, MA: Harvard University Press.

Medin, Douglas L., and Marguerite M. Schaffer (1978). Context Theory of Classification Learning. Psychological Review, 85(3), 207–238.

Murphy, Gregory L., and Douglas L. Medin (1985). The Role of Theories in Conceptual Coherence. Psychological Review, 92(3), 289–316.

OECD (2023). OECD Guidelines for Multinational Enterprises on Responsible Business Conduct. Paris: OECD Publishing.

Pettit, Philip (1997). Republicanism: A Theory of Freedom and Government. Oxford: Oxford University Press.

Putnam, Hilary (1975). The Meaning of “Meaning”. In Mind, Language and Reality: Philosophical Papers, Volume 2, 215–271. Cambridge: Cambridge University Press.

Raz, Joseph (1979). The Rule of Law and Its Virtue. In The Authority of Law: Essays on Law and Morality, 210–229. Oxford: Clarendon Press.

Rawls, John (1971). A Theory of Justice. Cambridge, MA: Harvard University Press.

Rosch, Eleanor (1978). Principles of Categorization. In Eleanor Rosch and Barbara B. Lloyd, eds., Cognition and Categorization, 27–48. Hillsdale, NJ: Erlbaum.

Scalia, Antonin (1997). A Matter of Interpretation: Federal Courts and the Law. Princeton: Princeton University Press.

Scalia, Antonin, and Bryan A. Garner (2012). Reading Law: The Interpretation of Legal Texts. St. Paul: Thomson/West.

Selznick, Philip (1969). Law, Society, and Industrial Justice. New York: Russell Sage Foundation.

Solum, Lawrence B. (1987). On the Indeterminacy Crisis: Critiquing Critical Dogma. University of Chicago Law Review, 54(2), 462–503.

Sunstein, Cass R., and Adrian Vermeule (2003). Interpretation and Institutions. Michigan Law Review, 101(4), 885–951.

Teubner, Gunther (2012). Constitutional Fragments: Societal Constitutionalism and Globalization. Oxford: Oxford University Press.

Tyler, Tom R. (1990). Why People Obey the Law. New Haven: Yale University Press.

United Nations (2011). Guiding Principles on Business and Human Rights: Implementing the United Nations “Protect, Respect and Remedy” Framework. HR/PUB/11/04. New York and Geneva: United Nations.

Waldron, Jeremy (2011). The Rule of Law and the Importance of Procedure. Nomos, 50, 3–31.

Williamson, Timothy (1994). Vagueness. London: Routledge.

Weiss, Bernard G. (1998). The Spirit of Islamic Law. Athens, GA: University of Georgia Press.

Wittgenstein, Ludwig (1953). Philosophical Investigations. Trans. G. E. M. Anscombe. Oxford: Blackwell.