An Anatomy of Injustice in the Emergence of Knowledge Among Heterogeneous Systems - A Preliminary Discussion in Generative Relational Terms of the Diplomatic Case 【（Preliminary）Draft】

Abstract

A claim made by one party to another travels a course before it becomes
anything the two hold in common. It is rendered into an event, construed at
the far end by procedures the sending party does not run, registered or left
unregistered in whatever description the parties share, admitted or refused
at its source, and possibly carried onward by others.
This paper takes that course as an anatomy and asks what may fail at each of
its sites, distinguishing a failure of the operation from an alteration of
the rule under which it is performed, the second of which escapes an
inspection of what the site produced. Three results are reported. An assessment counting the occasions on which a claim appears to
have been received as sent returns its highest values where the shared
description has least capacity to hold a divergence. Reach obtained through
intermediaries is purchased at the cost of fidelity, with no step on the
route performed badly. And a complaint that a divergence went unregistered is
unregistrable in the description whose incapacity it concerns. Two of the
failures identified obtain where no party refuses anything and no collective
resource for interpretation is deficient, which places them outside the
categories the existing literature supplies; the absence of a perpetrator is
taken from Fricker, and what is added is that the coarseness of a shared
description follows from its being shared and requires no inequality. The
final part treats the standing to raise a complaint and the forms of remedy
available where the parties have accorded no third party the office of
characterising their conduct. The account is stated in constraint form, and
it proposes no further criterion of the kind its companion papers carry.

1. Introduction

This paper concerns the wrongs available in the course a claim travels
between parties that differ in their procedures for settling what is so. Its
objective is to set out that course as an anatomy, to identify at each of its
sites what may fail and what it is for the site’s own rule to be altered, and
to treat the standing to raise a complaint and the forms of remedy available
where the parties have accorded no third party the office of characterising
their conduct. The method is to state a generative relational account of how
knowledge is formed and carried in this setting, to survey what the
literature on epistemic wrong already holds, to arrange the sites and the
modes, and to work through the sites in a common form.

The setting is one in which the parties differ in what would count, for each
of them, as showing that a claim is so. A companion paper treats that
difference formally, representing a joint description as a common factor of
two parties’ dynamics, disclosure as the consumption of a stock that does not
renew, the admission of another’s organs as a directed relation that composes
poorly, and the passage of a claim as a material event carrying no content
together with an operation the receiving party performs upon its own current
state (Huang, 2026). The present paper takes those constructions
as given and asks what is done to a party when the course fails at one point
or another.

Three commitments shape the account. The first concerns the mode of being of
a concept: a concept has presence within a configuration of relations and
within the language systems that interpret and represent it, the relations
and their dynamics are what bear witness to it, and witnessing is positioned,
partial, and open to deviation. The second concerns the relation of knowing
to acting, which in this setting is a matter of mechanism. Where a party’s
statements are not received as establishing anything, that party retains the
capacity to act and loses the channel by which a statement would be answered,
so the channel remaining open to it is demonstration; and a gift is an act
whose content is a proposal about how the parties are to understand one
another (Huang, 2026). The third concerns what the paper claims,
which is narrower than its survey might suggest.

Very little of the vocabulary is the paper’s own. Testimonial and
hermeneutical injustice are Fricker’s (Fricker, 2007); epistemic
oppression, testimonial smothering and contributory injustice are Dotson’s
(Dotson, 2011; Dotson, 2012; Dotson, 2014); willful hermeneutical ignorance is
Pohlhaus’s (Pohlhaus, 2012); epistemic exploitation is Berenstain’s
(Berenstain, 2016); cognitive justice is Visvanathan’s
(Visvanathan, 1997); epistemicide belongs to the epistemologies of the
South (Santos, 2014); linguistic justice is Van Parijs’s
(Van Parijs, 2011); epistemic violence is Spivak’s (Spivak, 1988). A
taxonomy organised by the stages of an epistemic transaction has also been
proposed, with pre-transactional, transactional and post-transactional as its
primary categories (Báez-Vizcaíno and Santana-Soriano, 2024), so ordering wrongs by their position in
a process is not new either. Section 5.3 concedes that work and states the
two respects in which the present arrangement differs from it.

What the paper claims has three parts, and the results may be given compactly
here since the sections establishing them are detailed.

The first is the arrangement itself. Five sites carry the anatomy: a claim is
rendered into an event, the event is construed at the receiving end, the
construal is registered or goes unregistered in the description the parties
share, the source of the claim is admitted or is not, and the claim may be
carried onward by parties other than the two. At each site a failure of the
operation is distinguished from an alteration of the rule under which the
site operates, and the second escapes an inspection of what the site
produced, since every instance thereafter conforms to the rule as altered.
Half the cells of the resulting arrangement are accordingly unavailable to
the observation an examiner ordinarily has.

The second is a set of results attaching to particular sites. An assessment
conducted within a shared description, and counting the occasions on which a
claim appears to have been received as sent, returns its highest values where
that description has least capacity to hold a divergence, so coarsening
improves an arrangement on that measure and removes the means by which the
improvement could be doubted. Reach obtained through intermediaries is
purchased at the cost of fidelity, so the party whose only route is the
longest is the party whose claim is composed by others most often, and no
step on the route need be performed badly for this to hold. And a complaint
that a divergence has gone unregistered is itself a statement about that
divergence, so it requires the place the description lacks and is
unregistrable in the description whose incapacity it concerns.

The third is the paper’s exposed claim and is stated here so that a reader
may test it early. Fricker holds that hermeneutical injustice has no
perpetrator and is a purely structural notion, and that much is taken from
her (Fricker, 2007). What she couples it to is a deficiency in the
collective resources available for interpretation, so the wronged party lacks
something. Two of the failures identified below obtain where those resources
are intact: a party may hold every concept it needs, state its own construal
fluently, be disbelieved by no one, and still find that the description the
parties share has no place in which its construal could appear, or that the
only route to its counterparty composes its claim through others. Neither
case requires prejudice and neither requires a gap. Nearby accounts hold that
the resources often exist and are refused
(Pohlhaus, 2012; Dotson, 2012), which supplies part of the ground; what
remains is the case in which nothing is refused because nothing is asked.

The objection likeliest to defeat that claim is that a shared description is
itself a collective resource, so that its incapacity is a deficiency in one.
The reply the paper stakes itself on concerns the origin of the deficiency.
Fricker’s gap is produced by unequal participation in generating the
resource; the coarseness of a shared description follows from its being
shared, since a description held in common makes only distinctions both
parties are able to make. Two parties of equal standing, equally represented,
acting in good faith, and each fully able to state its own position, arrive
at a description that fails to register some divergence between them.

Three limits are declared at the outset. The division of the course into
sites is a decision and not a joint found in the material, so a reader who
divides it otherwise obtains a different anatomy. The case material is read
through the account and measures nothing, having been assembled by others for
other purposes. And the closing part on remedy poses its question more clearly than it
answers it: it sets out forms already present in practice together with what
each can reach, and it finds that the remedial apparatus available to parties
without a common adjudicator is triggered by a breach of an obligation, which
the failures identified here do not supply.

The exposition proceeds as follows. Section 2 states the
generative relational account of knowledge in the diplomatic setting, which
is the basis on which the remainder is built. Section 3 fixes
the scope and the relation to the companion papers.
Section 4 surveys the prior formulations and states the
residue. Section 5 arranges the sites and the modes.
Sections 6 through 10 treat the five
sites in a common form. Section 11 states what certain of
the failures have in common. Section 12 reads the case material.
Sections 13 through 15 treat the
standing to complain, the position from which an arrangement may be examined,
and the forms of remedy available to the parties themselves.
Section 16 states the form in which the account’s normative
content is held, Section 17 its boundaries, and
Section 18 the questions it opens.

The paper belongs to a series applying generative relational theory to
governance, and it is the second of a pair on diplomacy among heterogeneous
knowledge systems.

2. The Generative Relational Account of Knowledge in the Diplomatic Setting

This section states the account on which the remainder of the paper rests.
Its objective is to fix what a concept is taken to be, what passes between
parties, how knowing stands to acting in this setting, and what course a
claim travels, so that the sites of Section 5 have something to
be sites of. The method is to state each commitment, to say what it replaces,
and to note where a companion paper carries the argument for it.

2.1 The Mode of Being of a Concept

A concept, on the view taken here, has presence within a configuration of
relations. It is not an object held by a party and available for conveyance,
and it is not a thing waiting in advance of the relations through which it
comes to be addressed. What gives it presence is the work of interpretation
and representation carried out within language systems, understood broadly to
include the procedures, categories, and forms of demonstration a community
uses. The relations and their dynamics are what bear witness to it.

The vocabulary of witnessing is chosen for what it concedes. A witness
occupies a position, sees from it, and may see wrongly; the seeing is
partial, and its partiality may be distributed unjustly across parties. What
a witness does not do is author what is witnessed, so the account keeps its
distance from the view that a party’s procedures make the world they report
on. The position is close to several already held. Situated knowledges refuse
the view from nowhere and reconstruct objectivity as located vision
(Haraway, 1988); strong objectivity requires that the knower’s position
be brought into the account (Harding, 1991); agential realism treats
phenomena in place of things with prior boundaries (Barad, 2007). The
present account concedes all three and differs from them in the question it
puts, which concerns what two positions can hold in common and what the
dynamics of their coupling do to it over time.

2.2 The Syntactic Crossing and the Semantic Generation

What passes between two parties is a material event: a document, an
utterance, a signal, a delivery, a visit, an abstention. The event has
structure, and it has no content that could be extracted from it by
inspection alone. Semantics is produced at the receiving end, by the
receiving party’s own procedures, working on the state that party already
occupies.

Two consequences follow and both are used throughout. The first is that what
the sending party took itself to be conveying and what the receiving party
comes to hold are determinations of two systems, so that their agreement is a
further question in place of a consequence of the passage. The second is that
the receiving party’s prior state enters the outcome, so the vocabulary of
transfer holds only where that state may be treated as making no difference.

The claim that nothing is transferred is not this paper’s. The critique of
the vocabulary in which communication is described as the sending of contents
in containers is Reddy’s (Reddy, 1979); that a perturbation triggers a
response it does not specify belongs to the theory of structural coupling
(Maturana and Varela, 1980); that communication is constituted by a synthesis of
selections within a system belongs to systems theory
(Luhmann, 1995); the exclusion of meaning from what a channel carries is
Shannon’s (Shannon, 1948). Within this series, the treatment of the gift
reached the same point by another route, holding that a gift proposes a
grammar and awaits its ratification and that the recipient’s response is the
ratification or the refusal of that grammar (Huang, 2026). The
companion paper gives the arrangement a formal statement
(Huang, 2026), and the present paper takes it as settled.

2.3 The Joint Character of Knowing and Acting

An account of what a party knows and an account of what a party does are not
separable in this setting, and three considerations establish it.

The first is a mechanism. Where no other party admits the outputs of a
party’s procedures, that party’s statements are not received as establishing
anything, while its conduct continues to be observed. The channel that
remains open to it is demonstration, meaning the production of a state of
affairs that requires no admission in order to be registered
(Huang, 2026). An exclusion described in the vocabulary of
knowing therefore issues in acts.

The second runs the other way. The acts through which parties address one
another are themselves proposals about how they are to be understood. A gift
composed without a reading of the recipient proposes a grammar the recipient
has no place in, and its wrongness is evidence about the giver
(Huang, 2026). An apology surrenders a position from which the giver
could have continued to deny. Neither is intelligible as conduct alone.

The third is that the literature already refuses the division. The practice
turn holds that competence in diplomacy is a form of knowing carried in
performance and resistant to statement in propositional form
(Pouliot, 2010; Adler and Pouliot, 2011). What follows for the present paper is
that its object is neither an epistemology nor an account of conduct, and
that each site below carries both.

2.4 The Course a Claim Travels

The course has five sites and two further moments that lie outside the
anatomy proper.

A party renders something it holds into an event, which requires a judgment
as to what will be legible at the far end. The event crosses, carrying
structure and no content. The receiving party construes it, producing a
determination of its own from its own state and its own procedures. That
determination is registered or goes unregistered in whatever description the
two parties share. The source of the claim is admitted or is not, by class of
output and pairwise. And the claim may be carried onward by parties other
than the two, each of which construes in turn.

The two further moments are the crossing itself, which is treated here as
carrying no content and therefore as the site of no failure, and the
reopening of a determination, which is treated in
Section 15 because it is a response to failure in place of a
site of it.

2.5 The Reach of the Organisational Framing of Knowledge

Knowledge in multilateral settings has been studied under a framing drawn
from organisational practice, in which knowledge is an asset built,
validated, held, moved, and applied, and a detailed study of a United Nations
negotiation process has been conducted on that basis (Andersen, 2026).
The framing is accurate within its domain and the present account claims
nothing against it there.

Its domain is fixed by three commitments. Validation operates as a gate with
a single standard of admission. Movement carries material between holders
with its content preserved. And the base into which material is admitted is
one base. Those conditions describe a party that holds one standard, one
competent authority, and one record, which is the condition of an
organisation. Among parties that hold different ones, the three fail
together, and what the framing can register of the failure is difficulty in
its own operation. The companion paper establishes this at length
(Huang, 2026); what the present paper adds is the question of
what is done to a party when it happens.

3. Scope, the Companion Papers, and the Bounds of the Case Material

This section fixes what the paper undertakes. Its objective is to state the
object and the level, to record what is carried from the companion papers and
therefore argued nowhere here, to mark the division of labour with the other
justice papers of the series, and to state the bounds observed in handling
material from settings that remain live.

3.1 The Object of the Anatomy and the Level at Which It Is Pitched

The object is the course a claim travels between parties running different
procedures for settling what is so, together with the failures available at
each of its sites. The object is a property of arrangements and not of
parties, and the level is that of the mechanism: the paper asks what a given
failure forecloses, what it leaves available, and what would be required to
detect it.

Three things lie outside. The paper offers no procedure for establishing that
a given failure has occurred in a given case. It ranks no failure against
another, and Section 16 states why. And it proposes no
further criterion of the form its companion papers carry, for reasons given
at Section 16.2.

3.2 The Results Carried from the Companion Papers

Four results are used and not re-derived. The representation of a joint
description as a common factor, of disclosure as the consumption of a stock,
of admission as a directed relation that composes poorly, and of the passage
of a claim as an event with no content acted upon by the receiving party’s own
operation, all belong to the companion treatment of heterogeneous knowledge
systems (Huang, 2026). The account of the gift as a proposal of
a grammar awaiting ratification, the boundary of those to whom return is
owed, and the availability of refusal as a test administered from outside,
belong to the companion treatment of the gift (Huang, 2026).

3.3 The Division of Labour with the Justice Papers of the Series

Four papers of the series hold criteria in the same family and this paper
adds none. Interpretive justice holds the regenerability of an interpretation
(Huang, 2026); participation justice holds generative
sufficiency (Huang, 2026); the treatment of value conversion
holds closure (Huang, 2026); recognition justice holds the
capacity to generate a further category (Huang, 2026); and the
companion paper on heterogeneous knowledge systems holds detectability and
reopening (Huang, 2026). The recurrence of one form across five
papers by one author is weak evidence for the form and is recorded as such in
the participation paper. A sixth instance would add nothing, and
Section 16.2 states the ground for declining it.

3.4 The Political Bounds Observed in Treating Live Material

The case material is drawn in part from settings that remain live, and three
bounds are observed. The material is organised by site and by mode
throughout, so that a practice is presented together with the site whose
failure it displays and never as an episode in the conduct of any party.
Parties are named where the published record and the studies relied upon
already name them, and then only for a procedural fact. And the account
reaches conclusions about mechanisms and reaches none about the conduct of
any party or the merits of any position advanced.

One further bound follows from the account itself. The failures identified
below include two that require no perpetrator, and a reader may be tempted to
supply one. The text states at each point what the account attributes and
what it leaves to attribution, and Section 11 treats the
question directly.

4. Prior Formulations of Epistemic Wrong

This section surveys the formulations under which wrong done in the
production and passage of knowledge has already been analysed. Its objective
is to record what each of them settles, to attribute the vocabulary the
present paper uses, and to identify what the collection leaves undescribed.
The survey is ordered from the wrongs done to a knower through the wrongs
done by a structure to the treatments of language, circulation and
responsibility, and it closes with the taxonomies already proposed and with
the residue. Each subsection concedes its formulation on its own terms, and
the difference marked at its end concerns the question asked and never the
adequacy of the answer given.

4.1 Testimonial and Hermeneutical Injustice

The founding distinction is Fricker’s. Testimonial injustice occurs where a
hearer accords a speaker less credibility than is due, through a prejudice
attaching to the speaker’s social identity. Hermeneutical injustice occurs
where a gap in the collective resources available for interpretation leaves a
person unable to render an area of their own experience intelligible, and
where that gap is itself the product of unequal participation in the
generation of those resources (Fricker, 2007).

Two features of the account are load-bearing here. The first is that Fricker
already grants that the second wrong has no perpetrator: No
agent perpetrates hermeneutical injustice, and the notion is
purely structural ()[159]fricker2007. The claim advanced in
Section 11 therefore takes the absence of a perpetrator from
her and does not claim it. The second is that she couples the absence of a
perpetrator to the presence of a gap, so that the wronged party lacks
something. The residue this paper occupies lies where neither a perpetrator
nor a gap is present.

4.2 Oppression, Smothering, Contribution, and Refused Resources

A second body of work holds that the resources are often available and are
refused. Testimonial quieting and testimonial smothering name the practices
by which a speaker is not heard and by which a speaker truncates their own
testimony in anticipation of not being heard (Dotson, 2011). Contributory
injustice names the case in which an interpreter possesses an adequate
resource and persists in an inadequate one (Dotson, 2012). Epistemic
oppression names the persistent exclusion from the shared production of
knowledge, distinguished by how far it is tractable within the prevailing
resources (Dotson, 2014). Willful hermeneutical ignorance names the
refusal by a dominant party to take up a resource that the marginalised have
built (Pohlhaus, 2012). A related body of work develops the epistemic
virtues and vices of resistance (Medina, 2013).

These formulations supply part of the ground the present paper stands on,
since each of them relocates the wrong away from a deficiency in the
collective resource. What each retains is an agent who refuses, and the cases
of Section 11 have none.

4.3 The Structural Turn and the Virtue of Institutions

A third body of work moves the object from the individual to the arrangement.
Epistemic justice has been treated as a virtue of social institutions,
assessed by the design of the arrangements through which testimony is
gathered and used (Anderson, 2012). The transactional and the structural
have since been distinguished directly, with colonial arrangements taken as
the paradigm case of a structure that distributes epistemic advantage
unevenly across groups (Jaggar and Tobin, 2024).

The turn is conceded in full, and the present account belongs to it. What
these treatments supply is the level; what they leave open is the anatomy at
that level, since a structure is assessed as a whole and the failures
available within it are not separated by the point at which they occur.

4.4 Exploitation and the Epistemologies of Ignorance

Two further formulations bear on the cost of being believed and on the
production of what is not known. Epistemic exploitation names the coerced,
uncompensated and emotionally taxing labour a marginalised party performs in
educating a privileged party about the conditions of its own oppression
(Berenstain, 2016). The epistemologies of ignorance treat ignorance as
something produced and maintained in place of something merely absent
(Mills, 2007; Proctor and Schiebinger, 2008), and strategic ignorance treats its
cultivation as an instrument (McGoey, 2019).

The first of these is the nearest neighbour to the cost treated in
Section 6, and the difference is stated there: exploitation
names labour extracted from a party, and the cost treated below falls on a
capacity that also produces the party’s later claims, so that it is spent in
place of being merely borne.

4.5 The Plurality of Knowledges and Its Destruction

A fourth body of work takes plural knowledges as its object. Cognitive
justice was coined for the right of different forms of knowing to co-exist
without duress, and against the destructive effect of a hegemonic science on
knowledges outside it (Visvanathan, 1997). Epistemicide names the
destruction of knowledges and of the ways of knowing that sustain them
(Santos, 2014). Knowledge justice has been developed as a term of art in
library and information studies, against the racial ordering of what is
collected, catalogued and taught (Leung and López-McKnight, 2021).

The present paper concedes the whole of this and takes none of its terms as
its own. The difference is one of level: these treatments concern the
standing of knowledge systems against one another at the scale of an
intellectual order, and the anatomy below concerns what happens to a single
claim in passage between two parties.

4.6 Language, Translation, and the Distribution of Its Burden

Language enters the field under several formulations, and their objects
differ. Linguistic justice concerns the fair distribution of the burdens and
benefits of language choice, and it asks who bears the cost of acquiring a
common medium (Van Parijs, 2011; Patten, 2014). Translation studies has
treated the choice between rendering a text familiar and preserving its
strangeness as an ethical and political matter (Venuti, 2019), and the
untranslatable has been treated as a resource against the substitutability of
concepts (Cassin, 2014; Apter, 2013). The inequality of the languages
between which anthropological translation moves was named early
(Asad, 1986). And translation between parties who do not share a world
has been described as controlled equivocation, in which what is shared is the
difference (Viveiros de Castro, 2004). Recent work names translation as
constitutive of exchange across epistemic difference and reframes
hermeneutical injustice through it (Vannini, 2023).

Two things follow for the present paper. The vocabulary of linguistic justice
is unavailable for the site treated in Section 6, since it
names the distribution of a different burden. And the residue is narrow: what
is undescribed is the wrong done at the moment of rendering, taken as one
site among several in a course, in place of translation taken as a condition
of the whole.

4.7 Circulation and the Relays That Carry It

The movement of knowledge has a literature of its own, and it is largely
descriptive. Knowledge has been treated as constituted in transit in place of
being made in one place and moved (Secord, 2004). The circulation of
knowledge across unequal settings has been reconstructed together with the
go-betweens who carried it (Raj, 2007). The accumulation of stable traces
at centres that thereby acquire power over what is distant is Latour’s
(Latour, 1987). In social epistemology the dependence of a knower on
others has been analysed at length (Hardwig, 1985), and the conditions
under which knowledge passes along a chain of testimony remain disputed, with
accounts differing on whether such dependence composes across the chain
(Lackey, 2008).

The dispute is directly relevant and is used in Section 10.
What the literature supplies is the mechanism of the relay; what it leaves
open is the normative question about the pathway, which is whether a party
may be wronged by the route its claim travels while nothing done at any point
on that route is objectionable.

4.8 Responsibility Borne for a Structural Process

Where a wrong has no perpetrator the question of responsibility does not
lapse, and it has been treated. The distinction between a liability model,
which is backward-looking and requires a causal contribution, and a social
connection model, which is forward-looking and shared among those who
participate in the structural processes producing the wrong, is Young’s
(Young, 2011).

The distinction is adopted here without alteration, and
Section 11 states what it does and does not settle for the
two cases identified there. A further distinction bears on
Section 8: representation in the sense of speaking for a
party and representation in the sense of depicting one are separate
operations, and conflating them obscures the second (Spivak, 1988).

4.9 The Taxonomies Already Proposed

Two attempts to order the field systematically have been made, and the first
of them anticipates the arrangement of the present paper in form.

A taxonomy organised by the stages of an epistemic transaction proposes
pre-transactional, transactional and post-transactional injustice as its
primary categories, taking the transaction as the process around which the
various manifestations occur, and it re-assigns the received kinds to those
stages: participatory and access wrongs before, testimonial wrongs during,
and interpretive and contributory wrongs after (Báez-Vizcaíno and Santana-Soriano, 2024). It is
offered there as an open framework. A second taxonomy, built for the case of
artificial intelligence, orders the received kinds hierarchically by
subsumption and cross-cuts them with dimensions such as the distributive and
the structural, and it follows the first (Mollema, 2025).

The consequence for this paper is stated plainly. Organising wrongs by their
position in a process is not new, and Section 5.3 records the relation
between the stages proposed there and the sites used here. The handbook
literature, by contrast, sets competing schemes of varieties beside one
another and settles on none (Kidd et al., 2017), which is what both taxonomies
give as their motivation.

4.10 The Residue Left by the Foregoing Formulations

Taken together the formulations establish that a knower may be wronged in
their capacity as a knower, that a collective resource may be deficient or
may be refused, that arrangements and not only persons may be assessed, that
being believed may cost the party believed, that whole knowledges may be
destroyed, that translation and language distribute burdens unevenly, that
knowledge is constituted in movement, that responsibility survives the
absence of a perpetrator, and that the field admits ordering by stage.

Claim 4.1. (The residue). What the foregoing formulations leave undescribed is the anatomy of a single claim in passage between parties running different procedures for settling what is so: which failure is available at which point of that passage, what distinguishes a failure of an operation from an alteration of the rule under which the operation is performed, and whether a failure may obtain where no party refuses anything and no collective resource is deficient.

Three consequences of the residue organise what follows. The passage has
points, and Section 5 fixes them. Each point admits two depths
of failure, and the same section fixes those. And the cases in which nothing
is refused and nothing is lacking are the paper’s exposed claim, treated in
Section 11.

5. The Sites of the Course and the Grounds for Declining a Taxonomy of Kinds

This section fixes the arrangement the remainder of the paper follows. Its
objective is to state why the failures treated below are localisations of one
wrong at several points in place of several kinds of wrong, to distinguish
the two depths at which a failure may occur at any point, to record the
relation between this arrangement and the stage taxonomy already proposed,
and to set the points against the depths. The method is to state the ground
for each decision and to say what would follow from deciding otherwise.

5.1 The Localisation of One Genus at Several Sites

A list of kinds is available and is declined. The passage of a claim invites
one, since each point at which something may fail can be named and the names
compose readily into a series: a wrong of translation, a wrong of
interpretation, a wrong of registration, a wrong of admission, a wrong of
circulation.

Three grounds tell against it. The first is that the members of such a list
are not exclusive: a claim rendered in terms the receiving party cannot admit
will also go unregistered, and a party denied admission will find its claims
carried by others or not at all, so the kinds run into one another wherever
two of them are instanced together. The second is that the list has no
principle of completeness, since a further point can always be distinguished
within a point already named, and a taxonomy whose membership is settled by
the analyst’s resolution is a taxonomy of the analyst’s decisions. The third
is that the same ground was reached in a companion treatment of recognition,
where indifference was found to be no separate kind of injustice and was
identified as the appearance recognition failure takes at the levels an
instrument cannot record (Huang, 2026).

The arrangement adopted is therefore the one taken there. One wrong is
localised at several sites, the sites are the points of the passage set out
in Section 2.4, and what a site contributes is the operation whose failure it
makes available. Nothing in the arrangement asserts that the sites are the
only ones, and Section 17 records what would follow from
adding one.

5.2 The Modes Distinguished at Each Site

At each site a failure may occur at either of two depths, and the
distinction is carried from the companion treatment of relational
degradation.

In the first, the operation of the site fails while the rule under which the
site operates remains intact. A rendering is composed that the receiving
party cannot use; a construal departs from what the sending party held; a
construal goes unregistered; a source is not admitted for a class of output;
a claim is carried to no destination. In each case the site continues to
operate under its own rule, and what failed is an instance.

In the second, the rule of the site is itself altered. What counts as a
legible rendering changes; the receiving party’s procedure for construing
changes; the description in which registration occurs is coarsened or
refined; the classes for which admission is granted are redrawn; the routes
along which claims may be carried are opened or closed. The alteration is not
an instance of the site’s operation and is a change in what the site will do
with every instance thereafter.

Claim 5.1. (The undetectability of the second mode). A failure of the first mode is available to inspection of what the site produced, since the rule under which the site operates is intact and the instance departs from it. A failure of the second mode is not, since every instance thereafter conforms to the rule as altered. The second mode appears in the responses a site makes to material that has not yet reached it.

Claim 5.1 governs what may be asked of each site below. Where a section
identifies a failure of the second mode, it also states what observation
would reach it, and the observation is in every case more expensive than an
inspection of the site’s output.

5.3 The Relation to the Transactional Stages

Ordering wrongs by their position in a process has been proposed, and the
proposal is conceded before the arrangement of this paper is stated. The
stages there are pre-transactional, transactional and post-transactional,
organised around the epistemic transaction as the process within which the
manifestations occur, and the received kinds are assigned to them
(Báez-Vizcaíno and Santana-Soriano, 2024).

Two differences fix what the present arrangement adds, and neither of them
concerns the decision to order by position.

The first is the process. The stages of the transactional taxonomy are
defined relative to an exchange between a speaker and a hearer, so that
before, during and after are read off the moment of testimony. The sites here
are defined relative to an account of what passes between two parties, in
which a claim is rendered into an event that carries no content and the
receiving party generates a determination of its own from its own state
(Huang, 2026). Under that account the moment of testimony is not
a single juncture, and the points at which something may fail are fixed by
the operations the account contains in place of by their position in time.

The second is the second mode. The transactional taxonomy distinguishes
stages and does not distinguish, at a stage, between a failure of the
operation and an alteration of the rule under which it is performed. That
distinction is what Claim 5.1 supplies, and it is the reason the arrangement
below has two dimensions where the earlier one has a sequence.

What is not claimed here is worth stating in the same place. Ordering by
position in a process belongs to the earlier work, and a reader who takes the
two differences above to be slight should take the present arrangement as a
variant of it.

5.4 The Sites Set Against the Modes

Figure 1 shows the arrangement and Table 1 states it. Its rows are the sites of
Section 2.4 and its columns the two modes of Section 5.2, and the entries
name what fails.

Figure 1. The sites of the passage and the two modes available at each. The horizontal arrows carry a claim from one site to the next; the downward arrows mark failures of the operation, which an inspection of what the site produced may reach; the upward arrows mark alterations of the site’s rule, which by Claim 5.1 it may not. The figure carries no numerical content.

Table. The sites of the passage set against the two modes of failure.

Site	Failure of the operation	Alteration of the site’s rule
Rendering	The event composed carries nothing the receiving party can use	What counts as a legible rendering is redrawn
Construal	The determination reached departs from what was held	The procedure of construal is altered by what was received
Registration	A divergence obtains and the shared description holds no place for it	The description is coarsened, and divergence ceases to be registrable
Admission	The outputs of a party’s procedures are not received	The classes for which admission is granted are redrawn
Circulation	A claim reaches no destination	The routes along which claims may travel are opened or closed

Two properties of the table bear on the sections that follow. Every entry in
the right column is undetectable by inspection of what the site produced, by
Claim 5.1, so five of the ten cells require an observation the arrangement
does not itself supply. And the entries in the left column are not
independent of one another, since a failure at one site alters what the next
site receives, which is the ground on which
Section 11 rests.

6. The Rendering of a Claim into an Event

This section treats the first site of the passage. Its objective is to state
the operation performed there, to identify the failure available while the
site’s rule is intact and the alteration of that rule, to set out the cost
the operation imposes on the party that performs it, to record what the
literature already holds of this site, and to name the material that displays
it.

6.1 The Composition of an Event from an Internal State

A party holding something it wishes another to act upon composes an event: a
document, a statement, a delivery, a visit, a refusal to attend. The
composition requires a judgment as to what will be legible at the far end,
since the event will be construed there by procedures the composing party
does not run. The judgment precedes the composition and is not derivable from
the composing party’s own condition.

Two properties of the operation carry the section. The event is many-to-one
with respect to the state that produced it, so what the composing party held
is not recoverable from what it sent. And the judgment about legibility is a
reading of the other party, which the companion treatment of the gift
establishes as the constitutive part of the act and the part the received
accounts leave unanalysed (Huang, 2026).

6.2 The Event That Carries Nothing the Receiving Party Can Use

The failure available while the site’s rule is intact is a composition
performed within the composing party’s own terms. Grain that the receiving
party’s agriculture cannot use, an argument resting on a form of
demonstration the receiving party has no procedure for admitting, a
certificate issued by a body whose outputs the receiving party does not
receive: in each the event arrives and there is nothing at the far end for it
to act upon.

Two features of this failure bear on the anatomy. It is evidence about the
composing party, since a composition made without a reading of the recipient
demonstrates that the reading was not performed (Huang, 2026). And it
requires no ill will, since a party may read attentively and read wrongly,
and the failure is the same in both cases. What the site makes available is
therefore a wrong that a competent and well disposed party may commit.

6.3 The Redrawing of the Terms in Which a Rendering Must Be Composed

The alteration of the site’s rule occurs where the terms in which a claim
must be composed are themselves changed. A body may require that submissions
take a specified form, that assertions be accompanied by a stated class of
evidence, or that a matter be presented in the vocabulary of a common
instrument.

The consequence follows Claim 5.1. After the alteration every composition
conforms to the new rule, so an inspection of what the site produces finds
nothing amiss. What has changed is which parties can compose at all. A party
whose procedures produce claims of a kind the new form cannot carry does not
appear in the record as excluded and appears as absent, and no party refused
it anything. The companion treatment of plural normative orders identifies
the same movement where a claim is remade into a form checkable in advance,
and records that the remaking is where the capacity it was to have protected
is captured (Huang, 2026).

6.4 The Cost Borne by the Party That Renders

Rendering a claim so that another may check it requires disclosing the
process that produced it, and in this setting that process is frequently the
one whose disclosure is most costly. Substantiating a claim about an event
may disclose the means by which the event was detected, and a means
described is a means that may be evaded. The companion paper represents the
cost as a draw upon a stock that does not renew, and establishes that the
total substantiation a party can deliver over any horizon is bounded, so that
the level of disclosure fixes the rate at which the budget is spent and
leaves its size untouched (Huang, 2026).

Two differences from neighbouring accounts fix what is claimed. Epistemic
exploitation names labour extracted from a marginalised party, uncompensated
and emotionally taxing, in educating a privileged party
(Berenstain, 2016); the cost here falls on the capacity that also
produces the party’s later claims, so what is lost is future standing in
place of present effort, and no party need demand anything for the cost to be
incurred. Closure by stake, in the companion treatment of the gift, holds
that an unenforceable instrument is more informative than an enforceable one
because performance secured by sanction is overdetermined
(Huang, 2026); there the party purchases information at the price of
security, and here it purchases receivability at the price of the process.

Claim 6.1. (The claims that cease to be made). Where substantiation draws on a bounded capacity, a party facing recurrent demands has reason to compose fewer claims of the class demanded. The reduction appears in claims that were not made, so an assessment conducted on the claims that were made registers nothing. Whether parties act on that reason in a given arrangement is a question the account leaves open.

6.5 The Formulations Already Available for This Site

Four bodies of work hold parts of this site. Translation studies treats the
choice between rendering a text familiar and preserving its strangeness as a
political matter (Venuti, 2019; Cassin, 2014; Apter, 2013), and the inequality
of the languages between which such rendering moves was named early
(Asad, 1986). Anthropological work describes translation between parties
who do not share a world as controlled equivocation, in which what is held in
common is the difference (Viveiros de Castro, 2004). Recent work in social
epistemology treats translation as constitutive of exchange across epistemic
difference (Vannini, 2023). And linguistic justice concerns the
distribution of the burdens of language choice (Van Parijs, 2011), which
is a different burden from the one treated here and is named to keep the two
apart.

What none of these treats is the cost of Section 6.4, which is specific to a
setting in which the process that would substantiate a claim is also the
process whose disclosure the party cannot afford.

6.6 The Material That Displays It

Three bodies of material display this site and are taken up in
Section 12. The composition of a diplomatic gift, together with
the practice of supplying an account of why that gift was selected, displays
the operation and the reading it requires. The provision under which a state
records a difference between its own regulation and an international standard
displays a rendering that keeps the party’s own terms on the record without
converting them. And a negotiation in which substantiating a claim about an
intrusion discloses the means of its detection displays the cost of
Section 6.4 at its highest.

7. The Construal of an Event at the Receiving End

This section treats the second site. Its objective is to state the operation
the receiving party performs, to identify the failure available while its
procedure is intact and the alteration of that procedure, to set out what a
sequence of received events does to the receiving party’s own capacity to
distinguish, to record the formulations already available, and to name the
material.

7.1 The Determination Produced at the Receiving End

An event arriving at a party is acted upon by that party’s own procedures,
working on the state the party already occupies. What results is a
determination of the receiving party’s system, produced from its history, its
categories, and its conditions of admission. The event does not determine it,
and the receiving party’s prior state enters it.

The formal statement is the companion paper’s: construal is an operation on
the receiving party’s current state parameterised by the event, so that two
parties in different states receiving the same event arrive at different
states (Huang, 2026). What follows for this section is that a
determination reached here is answerable to the receiving party’s procedures
and to nothing the sending party did.

7.2 The Departure from What Was Held

The failure available while the procedure is intact is a determination that
departs from what the sending party held. Both parties may have proceeded
correctly. The sending party composed within its own terms and read the other
as well as it could; the receiving party ran its own procedure faithfully;
and the two determinations differ.

The point at which this account parts from the received ones is here. An
account organised around misperception carries an error term, and error is
measured against what was the case; the divergence described here admits no
such term, since each party’s determination is answerable to its own
procedure and neither ran the other’s. A reader who requires a fault will
find none, and Section 11 takes up what follows.

7.3 The Alteration of the Procedure of Construal

The alteration of the site’s rule occurs where a received event changes not
the receiving party’s state but its procedure: what it will thereafter treat
as sufficient, which of its organs is competent, or what it will count as
settled.

The distinction is the one the companion paper draws between an emission
coupled to the receiving system’s configuration and an emission coupled to a
coefficient of its dynamics (Huang, 2026). The first moves a
state and leaves the rule intact; the second alters the rule, and its effect
is absent from the state at the time it operates. By Claim 5.1 an inspection
of the receiving party’s determinations reaches the first and not the second.
A statement that changes what a counterparty believes and a statement that
changes what a counterparty will thereafter treat as sufficient are different
acts, and an arrangement that registers only the first holds no record of the
second.

7.4 The Accumulation of Collapsed Distinctions

A sequence of received events narrows what the receiving party is able to
distinguish, and the narrowing accumulates. The companion paper establishes
the form: construal is an operation whose composition is not invertible, so a
distinction a party ceases to make is not restored by any subsequent event,
and the collapse accumulates along the sequence (Huang, 2026).

Three features make this the site’s most consequential failure. None of the
events need be false, so an assessment conducted on the truth of what was
received finds nothing. The loss is in the receiving party, so a party may be
diminished in its capacity to distinguish by a sequence addressed to it and
composed by others. And what has been lost is not available to the party that
lost it, since the distinction it can no longer draw is the one it would need
in order to notice the loss. The companion treatment of absorption in
relational systems develops the same point at the scale of a whole
arrangement (Huang, 2026).

7.5 The Formulations Already Available for This Site

Three bodies of work hold parts of this site, and the differences are worth
stating precisely.

That nothing passes and that the receiving system generates its own response
is settled (Reddy, 1979; Maturana and Varela, 1980; Luhmann, 1995), and the present
account takes it from them. Contributory injustice names a case in which an
interpreter possesses an adequate resource and persists in an inadequate one
(Dotson, 2012), and willful hermeneutical ignorance names the refusal of
a resource the marginalised have built (Pohlhaus, 2012). Both are
failures of this site, and both require that the interpreter hold something
better and decline it.

The failure of Section 7.2 requires no such holding. The receiving party has
its own procedure, runs it correctly, and has no other, so nothing is
refused. That is the difference on which the claim of
Section 11 rests, and it is the reason the two named
formulations do not cover the case.

7.6 The Material That Displays It

Two bodies of material display this site. A term appearing in an agreed text
and construed differently by the parties to it, where each construal is
faithful to the constructions of the party that made it, displays the failure
of Section 7.2. And the reported practice by which a claim contested in a
multilateral setting is handled by declining to acknowledge it, on the
reasoning that acknowledgement confers weight, displays the accumulation of
Section 7.4, since what is never acknowledged becomes a distinction the
receiving parties cease to draw (Andersen, 2026).

8. The Registration of a Construal in a Joint Description

This section treats the third site. Its objective is to state what it is for
a determination to enter a description the parties hold in common, to
identify the failure available while that description is unchanged and the
alteration of the description itself, to treat the boundary of those the
description counts and the depiction of a party within it, to record the
formulations already available, and to name the material.

8.1 The Entry of a Determination into a Description the Parties Share

Parties that act together hold some description in common: an agreed text, a
schedule, a standard, a record of proceedings. A description held in common
makes only distinctions that both parties are able to make, which the
companion paper represents as a common factor of the two parties’ dynamics
(Huang, 2026). Registration is the entry of a determination into
that description, and it succeeds where the description separates what one
party has construed from what the other has.

Two properties of the operation carry the section. A shared description is a
coarsening of each party’s own, so its capacity to register is bounded in
advance of any particular case. And registration is what makes a divergence
available to be raised, since a divergence the description cannot hold is one
that no party can point to within the arrangement.

8.2 The Divergence for Which the Description Holds No Place

The failure available while the description is unchanged is a divergence that
obtains and cannot be stated. The parties have construed differently, each
faithfully; each holds the concepts it requires; neither disbelieves the
other; and the description in which they conduct their common business makes
no distinction under which the difference could appear.

This is the paper’s first case of a failure with no perpetrator and no
deficiency of resources, and its features should be marked. Nothing is
refused, since no party is asked for anything. No collective resource is
lacking, since each party holds what it needs to state its own construal
within its own system. What is absent is a place in a shared instrument, and
the instrument was arrived at by parties acting in good faith to secure a
description both could assent to. Section 11 takes up what
follows for attribution.

8.3 The Coarsening That Removes the Capacity to Register

The alteration of the site’s rule is a change in the description itself, and
the direction that matters is coarsening. The companion paper establishes
that exchange in the transfer sense obtains at a point exactly where any
divergence present there goes unregistered, so that the proportion of cases
in which a passage appears to have succeeded and the proportion of
divergences the description can hold vary together with its coarseness
(Huang, 2026).

Claim 8.1. (The instrument that reports its own blindness as success). An assessment conducted within a shared description, and counting the occasions on which a claim appears to have been received as sent, returns its highest values where the description has least capacity to hold a divergence. Coarsening a description therefore improves the arrangement on that measure and removes the means by which the improvement could be doubted.

Claim 8.1 identifies a movement that requires no party to intend it. A
description is coarsened in order that assent may be obtained, which is the
ordinary work of an arrangement, and the loss of registering capacity is the
same operation seen from the other side.

8.4 The Boundary of Those the Description Counts

A prior result of this series bears directly on this site and is conceded
before the present extension is stated. The companion treatment of the gift
holds that the set of those to whom a return is owed is constituted by those
who bear the cost of an arrangement, in place of those the arrangement
recognises as its participants; that no criterion internal to the
arrangement detects the difference, since the arrangement’s own accounting is
conducted in terms from which the uncounted are absent; and that those
excluded are those for whom no position exists within the arrangement from
which their cost could be stated (Huang, 2026).

The extension offered here is narrow. That treatment states the result for
the accounting of value within a cycle. The present section states it for the
registration of a construal within a description, and supplies what the
earlier statement leaves unsourced, which is why a description should fail to
hold a position. The source is the coarsening of Section 8.3: a description
arrived at by parties in order to secure their assent makes the distinctions
they hold in common, and a party whose cost is stateable only in distinctions
they do not hold has no position by construction.

8.5 The Depiction of a Party in the Record

A further operation occurs at this site and is separable from the ones above.
A shared description contains descriptions of the parties: as a party in
compliance, as a party in arrears, as a party whose practices meet or fail a
standard. Who composes those descriptions, and under whose procedures, is a
question distinct from who speaks on a party’s behalf.

Spivak draws the distinction between speaking for a party and depicting one,
and it is adopted here in that form (Spivak, 1988). The first is treated
in a companion paper on participation (Huang, 2026). The
second belongs to this site: a party may be fully represented in the first
sense, holding its own seat and its own voice, while the description of it
that the arrangement carries forward is composed under procedures it does not
run. The accumulation of stable descriptions at the point where an
arrangement keeps its records is the mechanism by which distance is governed
(Latour, 1987), and what accumulates there is a depiction.

8.6 The Formulations Already Available for This Site

Hermeneutical injustice is the nearest formulation, and the difference from
it is exact. In that account the wronged party lacks the collective resource
by which its own experience could be rendered intelligible
(Fricker, 2007). Here each
party holds the resource it needs; what is absent is a place in a shared
instrument, and the instrument’s coarseness is the source. The two are alike
in having no perpetrator and differ in whether anything is lacking to anyone.

Two further formulations bear on the site. The boundary result of the
companion treatment of the gift is conceded at Section 8.4
(Huang, 2026). And the distinction between speaking for and depicting
is taken from Spivak (Spivak, 1988), with the second assigned to this
site.

8.7 The Material That Displays It

Three bodies of material display this site. The synthesis of contributions
into consensus language by a chair and a supporting team displays the
operation and its coarsening (Andersen, 2026). Reservations, interpretive
declarations, agreed minutes and differences filed against a standard display
the device by which a divergence is placed on a record without the shared
description being refined. And the descriptions of parties that an
arrangement carries forward, composed under the procedures of a body the
described party does not control, display the operation of Section 8.5.

9. The Admission of a Source

This section treats the fourth site. Its objective is to state what it is for
one party to admit the outputs of another’s procedures, to identify the
failure available while the classes of admission are unchanged and the
redrawing of those classes, to state what remains open to a party that no
other admits, to record the formulations already available, and to name the
material.

9.1 The Receipt of Another Party’s Outputs into One’s Own Record

A party accords standing to another when it admits the outputs of that
party’s verification procedures into its own record without independently
reproducing them (Huang, 2026). The apparatus is extensive and
familiar: the acceptance of passports and consular documents, legalisation
and apostille, comity, the enforcement of foreign judgments, the mutual
recognition of qualifications, the accreditation of bodies permitted to
contribute to a process.

Two properties fix what may fail. The relation is directed, so that admission
running one way between two parties is the ordinary case. And it is granted
for classes of output, so that a party may admit another’s certificates of
origin and decline its determinations of fact.

9.2 The Outputs That Are Not Received

The failure available while the classes are unchanged is that a party’s
outputs are not received. Its statements do not establish what they purport
to establish, and what it produces enters no other party’s record.

The failure is not the same as disbelief and should not be described as it.
A party may hold its counterparty entirely sincere and entirely competent and
still decline its outputs, on the ground that the body issuing them is one
whose procedures it has never admitted. The withholding may be a settled
policy, a lapse in an instrument, or the simple absence of an arrangement
under which admission would have been extended. What is common to these is
that the outputs do not arrive as establishing anything, and that nothing in
the party’s own conduct bears on the matter.

9.3 The Redrawing of the Classes for Which Admission Is Granted

The alteration of the site’s rule is a change in the classes for which
admission is granted: an instrument is amended, a scope of recognition
narrowed, a category of document removed from those accepted. After the
change every admission conforms to the classes as redrawn, so an inspection
of what is admitted finds a consistent practice, and by Claim 5.1 the
alteration appears only in what is thereafter not offered.

A consequence follows for the anatomy. The classes are ordinarily negotiated,
and their negotiation is the negotiation of what the arrangement will be able
to establish. A party present at that negotiation and a party affected by its
outcome need not be the same party.

9.4 The Channel That Remains Open

Where no other party admits the outputs of a party’s procedures, that party
retains its capacity to act and loses the channel by which a statement would
be received. What remains open to it is demonstration, meaning the production
of a state of affairs that requires no admission in order to be registered
(Huang, 2026).

The claim is stated at the level of the construction and quantifies over
parties in general. It identifies a mechanism and supplies no account of any
particular episode, and Section 17 records the ground on
which the paper declines to supply one. What the mechanism yields for the
anatomy is a passage between sites: a failure at the site of admission issues
in conduct, so that an arrangement which registers only statements holds no
record of what its own withholding produced.

9.5 The Formulations Already Available for This Site

Testimonial injustice is the formulation this site corresponds to, and the
correspondence is partial. There a hearer accords a speaker less credibility
than is due, through a prejudice attaching to the speaker’s social identity
(Fricker, 2007); the wrong is done by an agent, and the deficit is one of
credibility. Testimonial quieting and testimonial smothering name the
practices through which the same deficit operates and is anticipated
(Dotson, 2011).

Two differences hold at this site. Admission is granted for classes of
output and attaches to a party’s procedures in place of its person, so it may
be withheld with no judgment about the party’s sincerity or competence
having been formed. And the assessment of arrangements in place of persons
belongs to the structural turn (Anderson, 2012; Jaggar and Tobin, 2024), within which
the present treatment falls.

9.6 The Material That Displays It

Three bodies of material display this site. Accreditation procedures, through
which contributors other than states enter a multilateral process, display
the relation administered by an institution, and entry through another
party’s delegation displays the indirect path (Andersen, 2026). Provisions
for inspection, under which a party admits another’s inspectors for
designated classes of observation, display the express conferral of standing
by class of output. And the recognition of documents across jurisdictions
displays the ordinary operation of the site outside any dispute.

10. The Circulation of a Claim Through Intermediaries

This section treats the fifth site. Its objective is to state what occurs
when a claim is carried by parties other than the two, to identify the
failure available while the routes are unchanged and the alteration of the
routes themselves, to state what each relay adds, to record the formulations
already available, and to name the material.

10.1 The Carriage of a Claim by Parties Other Than the Two

Admission composes poorly, so a claim that one party cannot address to
another directly may nonetheless arrive by a path, with each intermediate
party admitted by the one before and admitting the one after. What an
intermediary supplies is a transitivity the parties do not hold between
themselves, and its office is constituted by that failure and calls for no
authority over the parties (Huang, 2026).

Two properties fix what may fail. A path exists or does not, independently of
the conduct of the party whose claim would travel it. And every party on the
path performs the operation of Section 7, since a relay
receives an event and composes one.

10.2 The Claim That Reaches No Destination

The failure available while the routes are unchanged is that no path exists.
A party’s claim is admitted by some other party and by no party that the
intended destination admits, so the claim is received somewhere and arrives
nowhere.

The failure is distinct from the one at the site of admission and is
frequently mistaken for it. There a party’s outputs are not received at all;
here they are received, and the parties receiving them stand in no relation
that carries them further. A party may accordingly hold standing with several
counterparties and remain unable to address the one whose determination
matters to it.

10.3 The Opening and the Closing of Routes

The alteration of the site’s rule is a change in which parties may relay:
the establishment or withdrawal of a body’s mandate to transmit, a
restriction on which contributors a process will receive material from, the
lapse of an arrangement under which one party had carried the claims of
another.

Two consequences follow. The alteration operates on parties other than the
one affected, so a party may lose its route through an arrangement made
between two others in which it had no part. And by Claim 5.1 the alteration
appears in claims that thereafter fail to arrive, which is a record no party
keeps.

10.4 The Construal Added at Each Relay

Each party on a path construes what it receives and composes what it sends.
What arrives at the destination is therefore the last relay’s composition,
and the divergence of Section 7.2 is available at every step.

Claim 10.1. (The pathway and what it costs). Reach obtained through intermediaries is purchased at the cost of fidelity. A claim arriving by a path of length $n$ has been construed $n$ times by parties other than the one that made it, and the party whose only route is the longest is the party whose claim is composed by others most often. No step on the path need be performed badly for this to hold.

Claim 10.1 is the site’s distinctive result and it is the clearest of the
paper’s failures with no perpetrator. Each relay receives faithfully, construes
under its own procedures, and composes in good faith; the arrangement
supplies no party who has done anything objectionable; and the claim that
arrives is not the claim that was made. What is wrong is a property of the
route.

The corresponding question in social epistemology is disputed and the dispute
is material here. Whether a knower’s dependence on another composes across a
chain of testimony, so that warrant obtained at the end of a chain stands on
the chain as a whole, is answered differently by accounts that locate warrant
in the belief of each speaker and by accounts that locate it in the
reliability of the final one (Hardwig, 1985; Lackey, 2008). The construction
used in the companion paper represents standing as a relation that composes
poorly (Huang, 2026), and the dispute bears directly on how far
that representation may be pressed.

10.5 The Formulations Already Available for This Site

The movement of knowledge has a substantial literature and it is largely
descriptive. Knowledge has been treated as constituted in transit
(Secord, 2004); circulation across unequal settings has been
reconstructed together with the go-betweens who carried it
(Raj, 2007); and the accumulation of stable traces at centres that
thereby acquire power over what is distant is Latour’s (Latour, 1987).
The epistemology of testimonial chains supplies the analytical apparatus
(Hardwig, 1985; Lackey, 2008).

The residue is narrow and is the author’s own formulation. What these
treatments describe is how knowledge moves and what moving does to it. What
is undescribed is the normative question about the route: whether a party may
be wronged by the pathway along which its claim travels, in a case where
nothing done at any point of that pathway is objectionable. The mechanism is
theirs; the question is not asked there.

10.6 The Material That Displays It

Two bodies of material display this site. Contributors that lack accreditation
with one body and enter a process through the delegation of a party that
holds it display the path and the construal added at its single relay
(Andersen, 2026). And the synthesis of many parties’ contributions into a
single draft by a chair and a supporting team displays a relay through which
every claim in a process passes, so that the composition arriving at the
parties is the relay’s own.

11. The Agentless Character of Certain of the Preceding Failures

This section states what some of the failures of the preceding five sections
have in common, and it carries the paper’s most exposed claim. Its objective
is to collect those failures, to separate what is taken from the existing
literature from what is added to it, to meet the objection most likely to
defeat the addition, to state the conditions under which the claim would
fail, and to record what follows for responsibility. The method is to state
the claim in a form that can be checked and to say at each point what would
have to be true for it to be wrong.

11.1 The Failures Collected

The failures identified at the five sites are not alike in this respect, and
three grades should be distinguished before the claim is stated.

At the first grade an agent performs the operation and commits no fault. The
composition of Section 6.2 is made by a party that read its counterparty
attentively and read wrongly, so an agent is present and nothing in the
account convicts it. At the second grade the source is a cost structure. The
draw upon a bounded capacity treated at Section 6.4 falls on a party whether
or not any counterparty demands anything of it, and the asymmetry between a
party whose claims are checkable by published procedure and a party whose are
not is a feature of their respective procedures. At the third grade no agent
performs the operation that fails. The divergence for which a shared
description holds no place, at Section 8.2, and the composition of a claim by
successive relays along the only route available to it, at Claim 10.1, are of
this kind.

The claim below concerns the third grade alone. Two failures fall under it,
and the paper’s exposure is confined to them.

11.2 The Absence of a Perpetrator

That an epistemic wrong may have no perpetrator is not claimed here. Fricker
holds it of hermeneutical injustice in terms, and the notion is purely
structural on her account ()[159]fricker2007. A reader who takes the
present section to be asserting the absence of a perpetrator has taken it to
be asserting something already settled.

What she couples the absence to is the presence of a lack. A collective
resource for interpretation is deficient, the deficiency is itself the
product of unequal participation in generating that resource, and the wronged
party is left unable to render an area of its own experience intelligible.
The wrong consists in that inability.

11.3 The Presence of the Resources for Interpretation

In the two failures collected at the third grade, no party is unable to
render its own experience intelligible. A party whose construal finds no
place in a shared description understands its own construal perfectly well,
states it fluently within its own system, and could give an account of it to
any party running the same procedures. A party whose claim is composed by
successive relays holds the claim it made and can say what it was.

Claim 11.1. (Failure with neither a perpetrator nor a lack). A failure may obtain at a site of the passage where no party has performed the operation that failed, and where no party lacks the resources by which its own construal could be rendered intelligible to itself and to others running its procedures. What is absent in such a case belongs to the relation between the parties and to neither of them.

The objection most likely to defeat Claim 11.1 arrives at once and is met
here. It runs: the shared description is itself a collective resource for
interpretation, so its incapacity to hold a divergence is a deficiency in a
collective resource, and the case falls under hermeneutical injustice after
all.

Two replies are offered and the second is the substantial one. The first
concerns what the deficiency is a deficiency in. Fricker’s resource is the
one by which a party makes sense of its own experience, and its deficiency
leaves the party opaque to itself; the parties here are transparent to
themselves and lack a place in a joint instrument. A party that understands
itself and cannot be entered in a record is in a different condition from a
party that cannot understand itself, and an account that treats them alike
loses the distinction between them.

The second reply concerns the origin of the deficiency. On Fricker’s account
the gap in the collective resource is produced by unequal participation in
the generation of that resource, so a party’s exclusion from the making of
the resource explains its inability to use it. The coarseness of a shared
description has a different source and requires no inequality whatever. A
description held in common makes only distinctions that both parties are able
to make (Huang, 2026), so it is coarser than each of them by
construction. Two parties of equal standing, equally represented, acting in
perfect good faith, and each fully able to state its own position, will
arrive at a description that fails to register some divergence between them.

Claim 11.2. (The coarseness that inequality does not explain). The incapacity of a shared description to hold a divergence follows from its being shared. It is present between parties who participated equally in constructing it, and it is therefore not explained by unequal participation in the generation of a collective resource.

11.4 The Accounts That Locate a Refusal

Two accounts relocate the wrong away from a deficiency in the collective
resource, and both are conceded at Section 4. Contributory
injustice obtains where an interpreter possesses an adequate resource and
persists in an inadequate one (Dotson, 2012). Willful hermeneutical
ignorance obtains where a dominant party refuses a resource the marginalised
have built (Pohlhaus, 2012).

Each supplies half of what Claim 11.1 requires, since each holds that the
resources may be present. What each retains is a party who refuses. In the
two failures collected above nothing is refused, because nothing is asked:
the divergence that a description cannot hold is not put to any party, and
the relay that construes a claim is not declining to construe it otherwise.

11.5 The Conditions Under Which the Claim Would Fail

Three conditions would defeat Claim 11.1 and are stated so that a reader may
apply them.

The first is that a shared description should count as a collective
hermeneutical resource in Fricker’s sense, and its coarseness as a gap in
that resource. The replies of Section 11.3 are offered against this and a
reader may find them insufficient, in which case the two failures are
instances of hermeneutical injustice and the section falls.

The second is that the coarseness of a shared description should trace, in
every case that arises, to unequal participation in its construction.
Claim 11.2 holds that coarseness is present without inequality, which is a
claim about what is possible; if inequality is in fact present wherever the
failure matters, the distinction is sound and idle. This is the most likely
defeater and the paper does not answer it, since answering it requires
evidence about cases and the paper supplies none.

The third is that a party should be identifiable in each case who chose the
description or the route. Where such a party exists, ordinary attribution
applies and the section adds nothing, and Section 11.6 records the
consequence.

11.6 Responsibility Where No Party Has Acted Wrongly

The absence of a perpetrator leaves the question of responsibility open and
does not close it, and the distinction that governs the case is Young’s. A
liability model looks backward, requires a causal contribution, and issues in
blame; a social connection model looks forward, is shared among those who
participate in the structural processes that produce the wrong, and is
political in character (Young, 2011).

The two failures collected here defeat the first model and fall to the
second. No party contributed causally to the incapacity of a description that
all of them constructed jointly, and no party on a route contributed causally
to the composition that arrived at its end. What the second model supplies is
a responsibility borne by those who participate in the arrangement, and what
it does not supply is an allocation among them. The paper adopts it in that
form and claims no more from it.

The move is not new and is conceded. Recent work brings the literature on
epistemic reparations together with the literature on structural injustice
and on collective agency, and argues that Young’s model is the more promising
vehicle for redressing epistemic wrongs of a structural kind
(Mitova, 2026). What the
present section adds is narrower: that model is applied here to two failures
identified at particular sites of a passage, in which the absence of a causal
contribution is established by the construction of the site in place of by
the diffuseness of a social process.

11.7 The Limits of What the Anatomy Attributes

The anatomy attributes a failure to a site and to a mode. It attributes
nothing to a party, and the reader should not supply what it withholds.

Two consequences follow. Where a party did choose the coarseness of a
description, did redraw the classes of admission, or did close a route, the
ordinary grounds of attribution apply and this paper adds nothing to them;
the analysis of the site tells such a reader where to look and no more. And
where no such party is found, the failure is a property of the arrangement,
and the response available is the one treated at
Section 15 in place of a finding against anyone.

12. Case Studies

This section reads three bodies of institutional material through the
anatomy. Its objective is to display the sites and the modes in arrangements
built without reference to them, to state the principle by which the material
was selected, and to record what the reading establishes. The material is
organised by site and by mode throughout, and no conclusion is reached about
the conduct of any party.

12.1 The Principle Governing the Selection of the Cases

A case earns its place here by displaying sites that the others do not, and
by displaying at least one of them in the second mode, where the rule of the
site is altered. Selection by that principle produces coverage of the
arrangement of Section 5 in place of variety for its own sake,
and it supplies a reason for each inclusion that can be checked against
Table 2.

The principle differs from the one used in the companion paper, and the
difference is deliberate. There the cases were chosen for the institutional
conditions under which they operate, so that each exhibited a mechanism the
others did not (Huang, 2026). Here they are chosen for the sites
they display. Two of the three bodies of material are common to both papers,
and the third is not.

12.2 The Relation to the Case Material of the Companion Paper

Two of the cases below are read in the companion paper for different
purposes, and a reader of both should know what is repeated and what is not.
There the material displays four formal constructions; here it displays five
sites and two modes, and the sections below make no use of the constructions
beyond the results already carried at Section 3.2. The verification of arms
limitation agreements, treated at length there, is not treated here, since
what it displays is already covered by the material at Section 12.5.

12.3 The First Committee Negotiations Read Across the Sites

The first case is the sequence of United Nations negotiations on developments
in the field of information and telecommunications in the context of
international security, studied across its lifetime under a five stage model
of knowledge management (Andersen, 2026). It is the anchor of the section
because it displays all five sites.

At the site of rendering, substantiating a claim about an event in this
domain frequently discloses the means by which the event was detected, which
is the cost of Section 6.4 at its highest. The study reports that parties may
decline to have their contributions durably recorded, on the ground that a
record exposes them to later citation and reduces their room for manoeuvre,
and that some prefer the benefits of memories fading. Read through
Section 6.4 that reluctance is a response to a bounded capacity in place of a
failure of candour.

At the site of construal, the study reports that a claim assessed as
inaccurate may be handled by declining to acknowledge it, on the reasoning
that acknowledgement confers weight and that repetition carries a claim
forward. Declining to acknowledge is an emission whose event is the
non-occurrence of a response, and its effect accumulates in the receiving
parties in the manner of Section 7.4.

At the site of registration, a chair and a supporting team receive
contributions and work them into successive drafts and finally into an agreed
text, and the synthesis is described as working with the substance without
changing it. That description is what Section 8.3 denies: a text to which
every party assents makes only distinctions every party can make, and the
coarsening is what makes the assent available.

At the site of admission, contributors other than states enter through
accreditation, which is the relation of Section 9.1 administered by an
institution. At the site of circulation, contributors lacking accreditation
have entered through the delegations of parties that hold it, which is the
path of Section 10.1 with a single relay.

The study also reports findings it presents as departures from its model: no
authoritative manager of knowledge emerged, the stages were observed out of
their expected order, and practices that the model treats as distinct were
found bundled. The companion paper derives those findings from its
constructions (Huang, 2026). What the present reading adds is
that each of them concerns a site at which the operation is performed by the
receiving party, which is why no office was found to perform it.

12.4 Standards and Equivalence Read at the Site of Registration

The second case is technical standard-setting in which a party may record a
difference between its own requirements and an international standard, taken
together with the provision under which one party accepts another’s measures
as achieving its own level of protection where the exporting party
demonstrates that they do.

The first of these displays the site of registration operating in the
direction the anatomy identifies as available and rare. The international
standard remains the shared description and is not refined; the party’s own
requirement remains its own and is not admitted to that description; and the
difference between them is on the record and available to every other party.
A divergence is thereby held without the shared description being made finer,
which is the only device Section 8.3 leaves open once its trade-off is
stated.

The second displays the site of rendering with the direction of the burden
written into an instrument. The demonstration falls on the party seeking
recognition, and it follows the reference in place of the strength of the
party. What the two together display is a domain in which the sites operate
without the failures the first case exhibits, and Section 12.7 records what
that can and cannot show.

12.5 Brokering Between Science and Policy Read at the Site of Circulation

The third case is the practice of brokering between the production of
technical assessment and its use in policy, in which bodies and individuals
occupy the position of relay between parties that do not address one another
directly. Practitioners in that position have set out principles for the ethical
conduct of brokering, on the stated ground that the ethical dimensions of the
work have remained underexplored and lack formal conceptualisation
(Cvitanovic et al., 2026).

What the case displays is the structure of Claim 10.1 and not a finding about
any broker. A relay receives an event and composes one, under its own
procedures, so the composition arriving at the destination is the relay’s.
The case is included because the first two cases display circulation with at
most one relay, and because the position of relay is here occupied as a
recognised office with a literature of its own.

One further feature bears on the anatomy. The alteration of the site’s rule
is available in this material, since a change in which bodies are recognised
as brokers is a change in the routes available and appears afterwards in
material that fails to arrive.

12.6 The Coverage of the Sites Across the Cases

Table 2 sets the sites against the cases and records where
each is displayed and in which mode.

Table. The sites of the passage and the case material displaying each.

Site	First Committee	Standards and equivalence	Brokering
Rendering	Substantiation disclosing detection; the record declined	Demonstration placed on the party seeking recognition
Construal	Acknowledgement withheld; accumulation in receivers		The broker’s own assumptions
Registration	Synthesis into consensus language	Difference filed beside the standard
Admission	Accreditation of contributors	Acceptance of another’s measures	Recognition of a body as broker
Circulation	Entry through another delegation		Relays between assessment and use; routes altered

Two observations follow. Every site is displayed by at least two of the three
cases, so no site rests on a single body of material. And the second mode is
displayed at three sites and not at the other two, which is what Claim 5.1
would predict of material assembled by observers of what an arrangement
produced.

12.7 The Standing of the Cases and What They Establish

The three cases are illustrations. Nothing in them was measured, no quantity
was estimated from them, and no claim of Sections 6
through 11 is confirmed by their agreement with it. What
they establish is that the operations the anatomy identifies are performed in
arrangements built without reference to it, and that practitioners have
devised instruments answering to some of them.

Three limits are recorded here and taken up at
Section 17. The first case is a single process in a domain
where the cost of substantiation is at its highest, so it is favourable to
Section 6.4. The second is drawn from domains of low political stakes, so its
display of a site operating soundly transfers to contested settings only by
an argument the paper does not supply. And the third is used for the
structure of a position and not for any finding about the parties who occupy
it, so it supports the presence of a relay and supports nothing about what
relays do. All institutional descriptions in this section are to be checked
against the primary instruments before the paper is circulated.

13. The Standing to Raise a Complaint

This section treats who may raise a failure of the kind the anatomy
identifies. Its objective is to show that bearing a cost and being able to
state it come apart, to identify two circularities that arise where the
failure is at the site of registration or at the site of admission, to treat
the case in which no respondent can be named, and to record the formulations
already available.

13.1 The Party Bearing the Cost and the Party Able to State It

A party that bears the cost of an arrangement and a party positioned to state
that it bears it are not the same party by construction. The companion
treatment of the gift establishes the point for the accounting of value: the
excluded are those for whom no position exists within an arrangement from
which their cost could be stated, and no criterion internal to the
arrangement detects the exclusion, since its accounting is conducted in terms
from which they are absent (Huang, 2026).

The anatomy sharpens the point at the site of registration.

Claim 13.1. (The complaint that cannot be entered). A failure of registration consists in a shared description having no place for a divergence. A complaint that such a failure has occurred is itself a statement about the divergence, so it requires the same place. The complaint is therefore unregistrable in the description whose incapacity it concerns.

Claim 13.1 is not a paradox and is a condition. A complaint may be raised
outside the description, in another forum or in a record the arrangement does
not keep, and Section 15 treats the devices by which this is
done. What the claim establishes is that the arrangement supplies no route,
so any route runs outside it.

13.2 The Complaint That Requires a Forum

A complaint must be received by someone, and being received is the operation
of the site of admission. Where the failure to be complained of is a failure
at that site, the second circularity arises: a party whose outputs are not
received cannot have its complaint about their non-receipt received either.

Two consequences follow. The party is left with the channel of
Section 9.4, so a complaint about admission tends to appear as conduct. And
the availability of a complaint is not evidence that the arrangement is
sound, since the parties able to complain are those whose outputs are
already received.

13.3 The Complainant Who Cannot Be Identified

The failures collected at Section 11.1 under the third grade have no
respondent, and they also raise a question about the complainant. Where a
description is coarse because it is shared, the party disadvantaged by the
coarseness in a given case may be either party, and which one it is varies
with what falls to be described. There is accordingly no standing class of
complainants, and a party positioned to complain today is positioned to be
complained of tomorrow.

This is a reason to state the condition in terms of the arrangement in place
of in terms of a class of persons, and it is one of the grounds on which
Section 16 declines to rank the failures.

13.4 The Formulations Already Available

Three formulations bear on this section. The wronged party’s own testimony
about the wrong is central to the account of testimonial injustice, and the
credibility deficit that constitutes the wrong also attaches to the
complaint (Fricker, 2007). The assessment of institutions in place of
persons supplies a route that does not depend on a complainant
(Anderson, 2012). And responsibility borne forward by those who
participate in a structural process supplies a form in which a wrong without
a respondent may still be addressed (Young, 2011).

What none of them treats is the circularity of Claim 13.1, in which the
instrument required to state the complaint is the instrument complained of.

14. The Audit Position and the Limits of Its Reach

This section treats the position from which an arrangement may be examined.
Its objective is to state the asymmetry between examining an arrangement and
specifying one, to record the test already available for exercise from
outside, to state the difficulty peculiar to the configuration in which the
parties have accorded no third party the office of characterising their
conduct, and to record how far an audit reaches at each site.

14.1 The Position from Which an Arrangement May Be Examined

An examination of an arrangement is conducted from a position occupied by
neither party to it. The occupant may be a third party, the affected party
speaking after the fact, or a later observer with the record before them.
What the position supplies is the possibility of asking of an arrangement
questions its own accounting does not ask.

The companion treatment of the gift establishes the asymmetry that governs
the position. A criterion of that kind is usable by a third party assessing an
arrangement, by the recipient, and by the historian, and it is unusable as a
specification of the act to be performed, since an act whose form is
determined by the requirements of a criterion is determined by the giver’s
requirements (Huang, 2026). The same treatment records that the
condition of justice on a generative cycle is not computable from within the
cycle.

14.2 The Asymmetry Between an Audit and a Specification

The asymmetry has a consequence for the present paper that should be stated
before the forms of remedy are set out. The anatomy identifies failures and
does not yield a design. A party that composed its renderings, its
descriptions, its admissions and its routes so as to satisfy the anatomy
would have composed them to its own specification, and the sites at which
the receiving party performs the operation are precisely the sites at which
a specification cannot reach.

What follows is that the account addresses an examiner and not a designer.
Section 15 accordingly sets out forms already present in
practice together with what each of them can reach, in place of a procedure
for arranging matters well.

14.3 The Refusability Test Exercised from Outside

One test is already available and is exercised from outside the arrangement.
An offering that could be refused and was not is one whose terms were
ratified; an offering that could not be refused is one whose acceptance is no
evidence of ratification, and the impossibility of refusal is auditable by a
party to neither side (Huang, 2026).

The test transposes to the sites of this paper without alteration. A
rendering the receiving party could not decline to construe, a description a
party could not decline to be entered in, an admission a party could not
decline to seek, and a route a party could not decline to use are each
instances of the same structure, and in each the party’s participation
carries no evidence of assent.

14.4 The Configuration in Which No Third Position Is Given

The difficulty peculiar to this setting is that the position of
Section 14.1 may be occupied by no one. The companion treatment of the gift
sets out two axes, whether a shared code is available and whether a position
exists from which the acts of each party may be characterised, and places the
diplomatic case in the cell in which both are absent
(Huang, 2026). Where no such position is given, an act of one party
against another is characterised only from the position of one of the two,
and each of them is a party to the act.

The present paper identifies the position and does not supply an occupant.
Four candidates are available and each is partial. A third party may occupy
it, and its own admission relations then determine what it can receive. The
affected party may occupy it after exit, and what it says is then said from
outside the arrangement. A later observer may occupy it, and the record they
work from was kept by the arrangement. And an audience before which the
parties act may occupy it, in which case the question of whether the audience
is the party for whom the acts were performed arises and is left open
(Huang, 2026).

14.5 The Reach of an Audit at Each Site

How far an examination conducted from outside reaches differs by site, and
the differences follow from what each site produces as a record.

The site of rendering leaves a record, since what was composed exists and can
be compared with what the receiving party could have used. The site of
admission leaves one, since instruments of recognition are published and
their classes are stated. The site of circulation leaves one, since routes
can be traced through the parties that carried a claim. The site of
registration leaves a partial record, since a shared description is available
for examination while what it failed to hold is available only from a party
willing to supply its own account. And the site of construal leaves the least
of all, since a determination reached within a party’s own procedures is
internal to it and an examiner sees only what the party did next.

Two consequences follow for the sections that remain. The devices treated at
Section 15 that place material on a record are addressed to the
two sites an audit reaches least. And by Claim 5.1 the second mode escapes an
audit at every site, since after an alteration of a site’s rule the record is
consistent with the rule as altered.

15. Forms of Remedy Available to the Parties Themselves

This section treats what may be done once a failure of the kind the anatomy
identifies has been found. Its objective is to set out the forms already
present in practice, to state what each of them reaches and what it costs, to
record the formulations already available together with the evidence this
paper has not sought, and to name the conditions the account leaves standing.
The section is the weakest in the paper and is written to make that visible.

15.1 The Annotation Placed Beside a Description

The first form places a divergence on a record without altering the
description the parties share. A reservation, an interpretive declaration, an
agreed minute, and a difference filed against a standard each have this
structure: a party states how it construes a provision, the statement is
recorded, and the shared description continues to make the distinctions it
made before.

The form is addressed to the failure of Section 8.2 and it is the only one
that avoids the trade-off of Section 8.3, since it raises what can be shown
without making the description finer. Its limits are three. It requires a
record the arrangement keeps, so it is unavailable where the failure is that
no record is kept. It requires the party to have identified the divergence,
which by Section 7.4 it may no longer be able to draw. And a form repeatedly
supplied at a settled magnitude hardens into a term of the arrangement, at
which point the divergence it was to have preserved has been converted after
all (Huang, 2026).

15.2 The Refinement of a Description and Its Price

The second form makes the shared description finer, so that divergences it
previously could not hold become statable within it. The price is fixed and
is the converse of Claim 8.1: a finer description registers more and is
assented to less readily, so refinement purchases registration with the
availability of agreement.

The account states the trade and selects no point on it. A reader looking for
a recommendation to refine will not find one here, and the ground is given at
Section 16.

15.3 Exit and the Conditions of Its Availability

The third form is departure from the arrangement. A party that withdraws
ceases to bear the cost of a description it cannot enter, and it speaks
thereafter from the position of Section 14.4 in place of from within.

Three conditions govern its availability and they are severe. Exit requires
an alternative, so a party holding no other relation of the kind has none,
and the severity of every other failure rises as the density of a party’s
other relations falls. Exit forfeits whatever the arrangement did supply, so
it is available to a party that can bear the loss. And exit removes the party
from the record, so the divergence that was unregistrable becomes unrecorded
as well, which is the failure of Section 8.2 completed in place of remedied.

15.4 The Intermediary Furnishing What the Parties Lack

The fourth form introduces a party admitted by parties that do not admit one
another, which supplies the transitivity of Section 10.1 and opens a route
where none existed.

The form is addressed to the failure of Section 9.2 and it aggravates the
failure of Section 10.4, since a claim arriving by a longer path has been
composed by others more often. Remedy at one site is therefore a cost at
another, and the account records the exchange and does not price it. What
follows for practice is that the two failures should be assessed together,
since an arrangement that has solved its admission problem by lengthening its
routes has moved a cost in place of removing it.

15.5 Reopening Across an Interval

The fifth form is the reopening of a determination after time has passed. Its
condition is that the divergence should have been kept somewhere, so it
depends upon the first form and inherits its limits.

The interval in this setting is long. A determination may be reopened by
parties whose personnel did not make it, on behalf of parties whose
population was not present when it was made, and the companion treatment of
the gift records that returns in the society of states may arrive after the
party that made the wager is gone (Huang, 2026). What follows is that
the record kept at the time of a failure is the instrument of a remedy that
no one present then will administer.

15.6 The Formulations Already Available and the Predicate They Require

Two literatures bear on this section and neither was fully consulted when it
was first drafted. Both are conceded here, and the second changes what the
section can claim.

A literature on the repair of epistemic wrong has developed. Epistemic
reparations have been treated as owed to those wronged in their capacity as
knowers, with the right of a victim to be a giver of knowledge about their own
experience proposed as the interest at stake (Lackey, 2022). Neighbouring
accounts treat the same ground under epistemic repair (Almassi, 2018),
epistemic redress (Hull, 2022), and epistemic amends (Song, 2021). An
apology has been given conditions of epistemic adequacy, among them
truthfulness, uptake of the victim’s testimony, and openness to the victim’s
own terms of description (Tobi, 2025). What these accounts presuppose is a
community able to acknowledge a wrong and to make amends, and the
configuration of Section 14.4 supplies none. That the structural dimension of
an epistemic wrong may place it beyond repair on these accounts has itself
been argued (Mitova, 2026), and the conditions named at
Section 15.7 are conceded to that argument in the respects it covers.

The law of remedies among parties that have accepted no common adjudicator is
the second, and it is developed. Where an obligation is breached and no
tribunal is available, the injured party may resort to self-help, of which the
principal lawful form is the non-forcible countermeasure, taken to procure
cessation and reparation and bounded by proportionality and by procedural
requirements (International Law Commission, 2001; Ruys, 2017). Reciprocity operates through the
suspension of a treaty in response to its material breach, and remedies in
the law of human rights have been surveyed at length (Shelton, 2015). The
availability of these instruments was historically defended as the corollary
of there being no authority above sovereign parties able to compel
reparation, which is the configuration of Section 14.4 named from within the
law.

Two observations follow, and the second is what the section rests on.

The first concerns who characterises. A countermeasure is lawful only in
response to a prior wrongful act, and the party that determines whether such
an act occurred is, in the absence of a tribunal, the injured party itself.
The apparatus therefore answers the difficulty of Section 14.4 by authorising
one of the two parties to occupy the position, which settles who characterises
and leaves untouched the objection that each of the two is a party to the act.
And a countermeasure is conduct, so a party whose statements are not received
and whose remedy is self-help has had the failure of Section 9.2 issue in the
channel of Section 9.4, which is the anatomy’s own prediction appearing in the
law.

Claim 15.1. (The predicate the apparatus requires). The remedial apparatus available among parties without a common adjudicator is triggered by a breach of an obligation. The failures collected at the third grade of Section 11.1 involve no breach: no party has failed to perform, and the description that holds no place for a divergence was arrived at by the parties jointly. The apparatus is therefore unavailable to them by its own conditions, and its unavailability is a feature of its design in place of a gap in the literature.

Claim 15.1 states the relation between this section and the law more
accurately than the first draft of it did. The ground is occupied, and what
occupies it is an apparatus whose predicate the failures of this paper cannot
satisfy. Withdrawal from an arrangement appears in the same law, and it
appears there chiefly as an act of the party whose conduct is complained of,
which is the converse of the exit treated at Section 15.3.

15.7 The Conditions the Account Leaves Standing

Three conditions are identified by the anatomy and are addressed by none of
the forms above, and by Claim 15.1 none of them reaches the remedial
apparatus of general international law either.

Where a failure is of the third grade of Section 11.1, there is no respondent
and the forms above are addressed to the arrangement in place of to a party.
Where the failure is at the site of construal, the accumulated loss of
Section 7.4 is not restored by any of them, since a distinction a party has
ceased to draw is not returned by a record kept elsewhere. And where a party
holds no other relation, exit is unavailable, no annotation is entered on its
behalf, and no intermediary admits it; for such a party the account
identifies a condition and supplies nothing.

The last of these is the paper’s plainest limitation and it is stated here in
place of at the end.

16. The Constraint Form of the Account

This section states the form in which the account’s normative content is
held. Its objective is to fix the relation between the anatomy and any reason
for action, to give the ground on which a further criterion is declined, to
state the prohibition on aggregation and the refusal to rank the failures,
and to record the sentence the account must remain able to refuse.

16.1 The Wall Between an Anatomy and a Justification

The anatomy locates failures. That a failure occurs at a site, that it is
available at a given mode, and that it is undetectable by a given observation
are statements about an arrangement, and none of them is offered here as a
reason for an arrangement, for a determination, or for the treatment of any
party.

The separation requires maintenance, since the account uses one vocabulary
for what it describes and for what it values. That an arrangement operates
smoothly at a site, or that a coarse description secures assent readily, is an
observation about its operation and is no part of a case for it. Where a
normative statement appears in this paper it is imported and is attributed at
the point of use.

16.2 The Grounds for Declining a Further Criterion

Five papers of this series hold criteria of one form, each assessing an act
by what it does to the capacity of a relation to produce a further
determination of the matter the act settled
(Huang, 2026; Huang, 2026; Huang, 2026; Huang, 2026; Huang, 2026).
A sixth is declined here on two grounds.

The first is evidential and is recorded in the participation paper: the
recurrence of one form across several papers by one author is weak evidence
for the form, and one author working within one framework is an available
explanation of the recurrence. Adding an instance strengthens the alternative
explanation.

The second is principled and is the operative ground. A criterion of that
form assesses an act, and the failures collected at the third grade of
Section 11.1 have no act to assess. A criterion applied to them would either
find nothing, since no act occurred, or would be applied to the arrangement
under a different form from the one the family shares. The anatomy locates
where a failure sits and leaves to those criteria the assessment of acts they
were built for.

16.3 The Prohibition on Aggregation Across Parties

The account admits no quantity summed across parties. The signature of the
utilitarian form is aggregation and not the vocabulary of maximisation, and
the operative test is whether a formulation permits one party’s capacity to
be traded against another’s.

Three quantities in this paper invite the trade and are withheld from it. The
coarseness of a shared description is a property of an arrangement and
appears in no objective function. The bounded capacity of
Section 6.4 belongs to a party, and no sum of such capacities across parties
appears anywhere. And the length of a route belongs to a pair of endpoints,
so a shorter route for one party is not offset against a longer one for
another.

16.4 The Refusal to Rank the Failures

The five sites are not ordered by severity here, and the refusal has a ground
beyond caution. Which failure is worse depends on what a party stands to lose
by it, and Section 13.3 establishes that there is no standing class of parties
disadvantaged at any site: a description coarse in one respect disadvantages
one party and in another respect the other.

A ranking would therefore be a ranking of parties under the appearance of a
ranking of sites, and the account declines it. What it supplies is the
question to be asked at each site and the observation that would answer it.

16.5 The Sentence the Account Must Remain Able to Refuse

A test of the account is whether it can refuse the following sentence: that
where a party’s construals are costly to register, and where an arrangement
functions more smoothly once its description has been coarsened past the
point at which they could appear, the coarsening is permissible.

The account refuses it, and three features secure the refusal. The conditions
it states are quantified over parties individually and form no aggregate, so
the smoother functioning of the remainder answers nothing. They are
constraints and no objective, so nothing in the account recommends a
description of any particular coarseness. And their normative force is
imported and is derived from no result about sites or modes, so no finding of
the anatomy bears on whether the coarsening is permissible.

The sentence is refusable because a party’s standing to have its divergence
registered rests on nothing it contributes to the arrangement. Were that
standing grounded in the contribution, the sentence would follow, and the
account would have supplied the argument it exists to withhold.

17. The Boundaries of the Undertaking

This section states the boundaries within which the account holds. Each is
given as a property of the undertaking together with what would have to be
established for it to be moved.

17.1 The Division of the Passage into Sites

The five sites are a division made by the author and not a set of joints
found in the material. A reader who divides the passage differently obtains a
different anatomy, and the results attached to each site travel with the
division that produced them.

Two considerations bound the damage. The division follows the operations of
an account of what passes between parties that is stated independently
(Huang, 2026), so it is answerable to that account in place of to
convenience. And the sites are not claimed to be exhaustive; adding one would
extend the table of Section 5.4 and would not disturb the entries already in
it, unless the added site were internal to one already named, in which case
the entries for that one would divide.

17.2 The Unmeasured Character of the Case Material

The case material displays operations and measures nothing. No quantity
appearing in the account was estimated from it, and no claim of the paper is
confirmed by a case agreeing with it. The material was assembled by others
for other purposes, and one of the three cases is a literature in which
practitioners describe their own position.

Moving this boundary requires evidence of a kind the paper does not use:
records of what parties construed, held internally, or ceased to produce,
which are the quantities Section 14.5 identifies as least available to an
examination.

17.3 The Individuation of a Party

A party is treated as running one set of procedures where the material treats
it so, and the treatment is a convenience. A state holds ministries,
agencies, and courts whose conditions of admission differ, so treating it as
one is a coarsening performed by the analyst and is subject to the same
result the paper states of coarsening generally.

The sites are indifferent to the level at which parties are individuated and
apply to two organs of one party on the same terms as to two parties. What
they cannot do is settle where the boundary falls, and the question is left
open in the terms used in a companion treatment of value conversion
(Huang, 2026).

17.4 The Standing of the Treatment of Remedy

An earlier draft of Section 15 recorded that the search which
would establish whether remedy absent a common adjudicator has been treated
elsewhere had not been conducted. It has since been conducted in part, and
the result was that the ground is occupied: the law of self-help,
countermeasures and reciprocity addresses exactly the configuration in which
no tribunal is available, and Section 15.6 now concedes it and states what
follows.

Two limits remain. The survey was of the doctrine and not of the
philosophical literature on redress in arrangements without an adjudicator,
which may hold positions the section does not engage. And Claim 15.1 rests on
a reading of the trigger condition for countermeasures; a reader who holds
that an arrangement’s incapacity to register a divergence can constitute a
breach of an obligation of good faith will find the apparatus available after
all, and the claim falls.

17.5 The Claim Most Likely to Fail

Claim 11.1 is the paper’s exposed point and Section 11.5 states three
conditions under which it fails. The second of them is the likeliest: if the
coarseness of a shared description traces in every case that arises to
unequal participation in its construction, then Claim 11.2 is sound and idle,
and the two failures collected at the third grade fall under hermeneutical
injustice after all.

Were that to happen, the anatomy would survive and its interest would
diminish. The sites, the modes, and the results attached to each would stand,
and what would fall is the claim that any of the failures lies outside the
categories already available.

18. Questions Opened by the Account

This account identifies a structure the existing formulations of epistemic
wrong leave undescribed: the passage of a single claim between parties
running different procedures for settling what is so, the sites at which that
passage may fail, and the depth at which a failure may occur at any of them.
If that structure is real, the following questions arise immediately, and the
account answers none of them.

Q1. Claim 11.2 holds that the coarseness of a shared description
follows from its being shared and requires no inequality. Are there
arrangements in which parties participated equally in constructing a
description and a divergence between them nonetheless went unregistrable, and
by what record would such a case be identified?

Q2. Claim 13.1 holds that a complaint about a failure of
registration is unregistrable in the description whose incapacity it
concerns. Which forums in fact receive such complaints, and what admits them
there?

Q3. Section 14.4 identifies a position from which an arrangement may
be examined and supplies no occupant. Under what conditions does a party come
to occupy it, and what does its own admission relations then exclude from
what it can receive?

Q4. Claim 10.1 holds that reach obtained through intermediaries is
purchased at the cost of fidelity. Can the two be measured against one another
in a documented process, and does the party with the longest route in fact
have its claims composed by others most often?

Q5. Claim 5.1 holds that an alteration of a site’s rule escapes an
inspection of what the site produced. What observation reaches it, and has
any arrangement been designed to keep the record such an observation would
require?

Q6. Section 7.4 holds that a sequence of received events narrows what
the receiving party can distinguish and that the loss is unavailable to the
party that suffers it. What would count as evidence of such a narrowing, given
that the party cannot report it?

Q7. Section 15.1 holds that an annotation raises what can be shown
without making a description finer, and that a form repeatedly supplied at a
settled magnitude hardens into a term of the arrangement. At what point does
the hardening occur, and is it visible before it is complete?

Q8. Section 15.4 holds that remedy at the site of admission
aggravates the failure at the site of circulation. Are there arrangements in
which the exchange has been recognised and priced, and on what basis?

Q9. Section 11.6 adopts a responsibility borne forward by those who
participate in an arrangement and concedes that it supplies no allocation
among them. What would an allocation require, and is one available that does
not reintroduce a causal contribution the cases lack?

Q10. Section 15.7 identifies a party holding no other relation of the
kind, for whom exit is unavailable, no annotation is entered, and no
intermediary admits it. What is owed to such a party, and by whom?

References

Adler, E. and Pouliot, V., eds. (2011) International Practices. Cambridge: Cambridge University Press.

Andersen, L. H. (2026) Knowledge Management in Multilateral Diplomacy: The Case of the UN First Committee Cyber Negotiations. Cham: Palgrave Macmillan.

Báez-Vizcaíno, K. and Santana-Soriano, E. (2024) Pre-transaccional, transaccional y post-transaccional: categorías para reconocer la injusticia epistémica. Ciencia y Sociedad, 49(3).

Barad, K. (2007) Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning. Durham: Duke University Press.

Berenstain, N. (2016) Epistemic exploitation. Ergo, 3(22), 569–590.

Dotson, K. (2011) Tracking epistemic violence, tracking practices of silencing. Hypatia, 26(2), 236–257.

Dotson, K. (2012) A cautionary tale: on limiting epistemic oppression. Frontiers: A Journal of Women Studies, 33(1), 24–47.

Dotson, K. (2014) Conceptualizing epistemic oppression. Social Epistemology, 28(2), 115–138.

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

Haraway, D. (1988) Situated knowledges: the science question in feminism and the privilege of partial perspective. Feminist Studies, 14(3), 575–599.

Harding, S. (1991) Whose Science? Whose Knowledge? Thinking from Women’s Lives. Ithaca: Cornell University Press.

Almassi, B. (2018) Epistemic injustice and its amelioration: toward restorative epistemic justice. Social Philosophy Today, 34, 95–113.

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

Huang, W. (2026a) Diplomacy in a world of heterogeneous knowledge systems: exploring generative relational systems perspectives and formal models. Unpublished working paper.

Huang, W. (2026b) The gift and the generative relational foundations of diplomacy: a preliminary discussion paper. Unpublished working paper.

Huang, W. (2026c) Interpretive justice in organizations: a preliminary discussion of problems and open questions. Unpublished working paper.

Huang, W. (2026d) Participation justice in generative relational being. Unpublished working paper.

Huang, W. (2026e) Interpretive justice in heteromorphic value conversion. Unpublished working paper.

Huang, W. (2026f) Recognition justice in generative relational being: a preliminary discussion of inclusion. Unpublished working paper.

Luhmann, N. (1995) Social Systems, trans. J. Bednarz Jr. with D. Baecker. Stanford: Stanford University Press.

Maturana, H. R. and Varela, F. J. (1980) Autopoiesis and Cognition: The Realization of the Living. Dordrecht: Reidel.

Pohlhaus, G. (2012) Relational knowing and epistemic injustice: toward a theory of willful hermeneutical ignorance. Hypatia, 27(4), 715–735.

Pouliot, V. (2010) International Security in Practice: The Politics of NATO–Russia Diplomacy. Cambridge: Cambridge University Press.

Reddy, M. J. (1979) The conduit metaphor: a case of frame conflict in our language about language. In A. Ortony, ed., Metaphor and Thought. Cambridge: Cambridge University Press, 284–324.

Ruys, T. (2017) Sanctions, retorsions and countermeasures: concepts and international legal framework. In L. van den Herik, ed., Research Handbook on UN Sanctions and International Law. Cheltenham: Edward Elgar, 19–51.

Santos, B. de S. (2014) Epistemologies of the South: Justice Against Epistemicide. Boulder: Paradigm Publishers.

Shannon, C. E. (1948) A mathematical theory of communication. Bell System Technical Journal, 27, 379–423 and 623–656.

Spivak, G. C. (1988) Can the subaltern speak? In C. Nelson and L. Grossberg, eds., Marxism and the Interpretation of Culture. Urbana: University of Illinois Press, 271–313.

Shelton, D. (2015) Remedies in International Human Rights Law, 3rd edn. Oxford: Oxford University Press.

Song, S. (2021) Denial of Japan’s military sexual slavery and responsibility for epistemic amends. Social Epistemology, 35(2), 160–172.

Tobi, A. (2025) The epistemic harms of botched apologies for past wrongs. Journal of Applied Philosophy, doi:10.1111/japp.70050.

Van Parijs, P. (2011) Linguistic Justice for Europe and for the World. Oxford: Oxford University Press.

Visvanathan, S. (1997) A Carnival for Science: Essays on Science, Technology and Development. Delhi: Oxford University Press.

Anderson, E. (2012) Epistemic justice as a virtue of social institutions. Social Epistemology, 26(2), 163–173.

Apter, E. (2013) Against World Literature: On the Politics of Untranslatability. London: Verso.

Asad, T. (1986) The concept of cultural translation in British social anthropology. In J. Clifford and G. E. Marcus, eds., Writing Culture. Berkeley: University of California Press, 141–164.

Cvitanovic, C., Partelow, S., Bednarek, A., Hornidge, A.-K. et al. (2026) Principles for ethical boundary spanning between science, policy and practice: a guide for knowledge brokers. Environmental Science and Policy, 179, article 104361.

Cassin, B., ed. (2014) Dictionary of Untranslatables: A Philosophical Lexicon, trans. and ed. E. Apter, J. Lezra and M. Wood. Princeton: Princeton University Press.

Hardwig, J. (1985) Epistemic dependence. The Journal of Philosophy, 82(7), 335–349.

International Law Commission (2001) Draft Articles on Responsibility of States for Internationally Wrongful Acts, with commentaries. New York: United Nations.

Jaggar, A. M. and Tobin, T. W. (2024) Moral justification and structural epistemic injustice. In J. Browne and M. McKeown, eds., What is Structural Injustice? Oxford: Oxford University Press.

Huang, W. (2026g) From marriage economics to multi-nomoi governance: a dialectical generative relational economics perspective. Unpublished working paper.

Huang, W. (2026h) Rethinking peace: absorption capacity and the poverty of grammar in generative relational being. Unpublished working paper.

Kidd, I. J., Medina, J. and Pohlhaus, G., eds. (2017) The Routledge Handbook of Epistemic Injustice. London: Routledge.

Lackey, J. (2008) Learning from Words: Testimony as a Source of Knowledge. Oxford: Oxford University Press.

Hull, G. (2022) Epistemic redress. Synthese, 200(3), article 201.

Lackey, J. (2022) Epistemic reparations and the right to be known. Proceedings and Addresses of the American Philosophical Association, 96, 54–89. [One source gives the Proceedings of the Aristotelian Society; the venue is to be confirmed.]

Latour, B. (1987) Science in Action: How to Follow Scientists and Engineers through Society. Cambridge, MA: Harvard University Press.

Leung, S. Y. and López-McKnight, J. R., eds. (2021) Knowledge Justice: Disrupting Library and Information Studies through Critical Race Theory. Cambridge, MA: MIT Press.

McGoey, L. (2019) The Unknowers: How Strategic Ignorance Rules the World. London: Zed Books.

Medina, J. (2013) The Epistemology of Resistance. Oxford: Oxford University Press.

Mitova, V. (2026) Irreparable epistemic wrongs. Philosophical Studies, 183, 899–917.

Mills, C. W. (2007) White ignorance. In S. Sullivan and N. Tuana, eds., Race and Epistemologies of Ignorance. Albany: SUNY Press, 11–38.

Mollema, W. J. T. (2025) A taxonomy of epistemic injustice in the context of AI and the case for generative hermeneutical erasure. AI and Ethics, 5.

Patten, A. (2014) Equal Recognition: The Moral Foundations of Minority Rights. Princeton: Princeton University Press.

Proctor, R. N. and Schiebinger, L., eds. (2008) Agnotology: The Making and Unmaking of Ignorance. Stanford: Stanford University Press.

Raj, K. (2007) Relocating Modern Science: Circulation and the Construction of Knowledge in South Asia and Europe, 1650–1900. Basingstoke: Palgrave Macmillan.

Secord, J. A. (2004) Knowledge in transit. Isis, 95(4), 654–672.

Vannini, A. (2023) Towards epistemic translatability: on epistemic difference and hermeneutical injustice. Social Epistemology, 37(6), 839–851.

Venuti, L. (2019) Contra Instrumentalism: A Translation Polemic. Lincoln: University of Nebraska Press.

Viveiros de Castro, E. (2004) Perspectival anthropology and the method of controlled equivocation. Tipití, 2(1), 3–22.

Young, I. M. (2011) Responsibility for Justice. Oxford: Oxford University Press.