Duties of the Situated Judge - Epistemic Restraint and the Ethics of Judging Persons for Others

Transcript

Abstract

Someone who has worked with a person is asked to say what they are like, by a party that will decide something about them. The request is ordinary and the act is not. This paper asks what is owed by the party who issues such a judgment, and by the party who uses it. It argues, first, that the ethics usually brought to this act is the wrong ethics. Norms of assertion govern saying, and the act at issue does something other than say: it opens or closes the relations available to the person judged, which is why an account signed by someone who observed nothing still works. Duties appropriate to an exercise of discretionary power over another’s conditions therefore govern it, and the obligation runs primarily to the person judged, who is absent from the transaction, rather than to the party who asked. The paper then develops the ethics from six principles of relational reproduction belonging to the framework in which it works: that reproduction preserve the possibility of continued becoming; that a shared past not be interpreted unilaterally by one of its authors; that value return to those who generated it, a principle inherited from work on generative justice and identified as such; that the authority conferred by an act of reproduction be met by a capacity to contest it, generated at the same moment; that no relation achieve its result by eliminating the other as a subject; and that no interpretation become historically irreversible, which requires that the conditions of later revision be preserved and not merely permitted. Applied to an account of a person, these yield results that a norm of assertion does not reach: that an ordinary confidential reference breaches the second and fifth principles by construction, that foreclosing a person’s field is a graver wrong than describing them inaccurately, and that a duty of contestation must be built into the instrument when it is issued rather than supplied afterwards as an appeal. Two further claims follow. Refusing to write is not the safe course but the act requiring the stronger justification, since it forecloses invisibly and leaves nothing to appeal against. And because those excluded by such judgments are never observed, the party exercising this power can never establish that they exercised it well, so the ethics cannot be an ethics of correctness and must be an ethics of conduct under permanent unverifiability. The paper’s duties are correlative: a restraint binding only the issuer penalises the scrupulous issuer and has, on the available evidence, already produced silence.

Keywords: ethics of assessment; norms of assertion; testimony; generative justice; references; epistemic restraint.

Discussion Paper Note

This paper is a preliminary discussion paper intended to share an evolving idea and invite further dialogue, criticism, revision, and independent development. Its definitions, distinctions, and constructions remain provisional. Circulation across scholarly and practical communities is part of the purpose of releasing the manuscript at this stage.

The author treats the viewpoints, concepts, and lines of reasoning presented here as contributions to a shared field of inquiry. Similar or related ideas may have appeared in other intellectual, cultural, and disciplinary traditions. The manuscript therefore states its known antecedents, separates the researcher-origin proposal from later formal reconstruction, and leaves historical priority open pending a systematic originality review.

The arguments should be understood as provisional and historically situated. Readers are encouraged to question, test, revise, extend, reinterpret, or independently develop the ideas presented here. Where appropriate, acknowledgment of this paper as one point of encounter in the development of a related idea is appreciated. Such acknowledgment records an intellectual route; the ideas themselves remain available for criticism, revision, and independent development.

Responsible Use and Rights Reservation

This section separates requested scholarly conduct from the legal permissions stated on the following page. It records an ethical request for responsible use and then defines the narrower scope of retained legal rights.

The author encourages good-faith discussion, criticism, independent inquiry, and responsible use of the material in this work. Separately from the licence’s terms, the author asks users to consider foreseeable harms when adapting or applying the arguments made here. This paper argues that refusing to give an account of a person requires a stronger justification than giving one. An argument of that shape can be read as pressing people to speak where they have good reason for silence, including where speaking would expose them or the person judged to harm. The paper accordingly states the conditions under which refusal is owed, and the author asks that the argument and those conditions be applied together. This ethical request leaves the licence’s permissions and legally authorized uses unchanged.

The author retains the rights preserved under CC BY-NC 4.0 and may pursue remedies to which the author is legally entitled for breach of the licence or violation of the author’s independently applicable rights. Reuse remains independent from authorial endorsement. Third-party rights require authorization from their respective holders where applicable. Copyright exceptions and limitations, including applicable forms of fair use or fair dealing, remain fully available.

Notices

This page consolidates the manuscript’s publication status, licence, development disclosure, research-programme relation, declared interest, and suggested citation.

Publication status of this draft.

This working draft records an evolving stage of the author’s position and is circulated for discussion. Definitions, section structure, statements, and numbering remain subject to revision. The legal material is used to establish what duties exist and are enforced, not to state the law of any jurisdiction, and the author is not a lawyer. Specialist review of the legal and philosophical material, empirical work on the proposals, and a systematic originality audit remain future research stages.

Licence.

Except where otherwise indicated, copyright 2026 Wanhong Huang. This work is made available under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). Subject to its terms, the licence permits sharing and adaptation for noncommercial purposes with appropriate attribution, a link to the licence, an indication of changes, and attribution that preserves the licensor’s independence from the reuse. Reuse is governed solely by that licence; the responsible-use request on the preceding page remains separate from its terms. The licence deed and legal-code link are available at https://creativecommons.org/licenses/by-nc/4.0/. The licence governs in case of conflict with this summary. Third-party material remains subject to the rights held by its respective rights holders.

Statement on the use of language models.

The exploratory discussions and preparation of this paper involved Anthropic’s Claude. The model supported exploratory dialogue, source discovery followed by verification against publisher, journal, governmental, and institutional pages, argumentative criticism, and drafting in LaTeX. Every source cited here was verified before it was written into the manuscript rather than after. The author selected the research question, directed and approved the theoretical commitments and the epistemic status of the claims, and bears sole responsibility for the manuscript, including its definitions, constructions, arguments, conclusions, and errors. Authorship credit remains with the human author. The access level and claim limit for every cited source are recorded in the accompanying literature audit.

Declared interest.

The author is subject to procedures of the kind examined here and has an interest in the conclusions reached about them. The paper is written as a general ethics of judging persons for others, of which academic recommendation is one case, no individual process or person is described, and the arguments are assessed against criteria stated in advance. The interest is declared because a paper assigning duties to parties who judge, and to parties who rely on judgments, should say where its author stands with respect to both.

Related research programme.

This paper is project P005 and the fifth in a series on trust, neutrality, and the transmission of shared experience. Project P001 takes responsibility for the account of neutrality as the governance of a field of relational conditions; P002 for the individual-scale credibility problem; P003 for the procedure of reference-based selection; and P004 for the ontology of the judged subject, for the finding that such instruments transmit a field of relational conditions rather than an account, and for the scope condition on what may be claimed of a person. The present paper takes responsibility for the duties of the issuing and receiving parties, for the ethics of refusal, and for the account of conduct under unverifiability. Later papers take the distribution of the conditions under which judgments are generated, the authority to interpret a shared experience, and the political economy of these instruments; those questions are marked where they arise and are not argued here.

Suggested citation.

Huang, Wanhong. “Duties of the Situated Judge: Epistemic Restraint and the Ethics of Judging Persons for Others.” Working discussion paper, 2026.

Introduction

A supervisor is asked what a former student is like. A manager is asked to rate someone who will be promoted by another department. A clinician’s formulation follows a patient to a service that did not produce it. A colleague is asked whether a candidate should be appointed, by people who will never see the work the colleague saw.

In each case a person who stood in a relation to another is asked to say what that other is like, by a party that will act on the answer. The request is ordinary. What is done in answering it is extraordinary, and this paper is about what is owed in doing it.

The obvious way to approach the question is through the ethics of speech. The party who answers is making a statement; there are norms governing when one may make statements; the duties follow from those norms. On that approach the issuer must have adequate grounds, must not assert beyond what they know, and must say what they believe. The recipient’s position is barely a subject at all, since a hearer who is told something has no obvious duty except to weigh it.

This paper argues that the approach is mistaken, and the argument begins from a practice that the speech account cannot explain. A referee is asked for an account and replies that the candidate should draft it and the referee will sign; or asks for a curriculum vitae and writes from that. No relation is being reported. The instrument functions exactly as it would otherwise. If what such instruments did were to state something, an instrument stating nothing would do nothing, and this one does something.

What it does is alter the relations available to the person it concerns. An account of this kind opens a position, or closes one, and what the person subsequently becomes is generated in the configuration that follows. The account of the relation is the justification offered for opening or closing, and not the thing conveyed. That finding belongs to the preceding paper in this series and is taken here as given.

Three consequences organise what follows.

The first is that the governing ethics is an ethics of power rather than of speech. Norms of assertion regulate what may be said; they have nothing to say about opening or foreclosing what a person can become, which is a different act and answers to different constraints. The duties that govern it are those of any actor exercising discretionary power over another’s conditions: to have grounds, to state them, to proportion the act to what the grounds support, and to leave the exercise open to challenge. Section 5 develops this, and draws a consequence that the speech account obscures: the obligation runs primarily to the person judged, who is absent from the transaction, rather than to the party who asked.

The second is that the framework within which this paper works already has an ethics of acts of this kind, and it has not been used. That framework holds six principles governing the reproduction of shared experience: that reproduction preserve the possibility of continued becoming; that a shared past not be interpreted unilaterally by one of its authors; that value return to those who generated it; that the interpretive authority conferred by an act of reproduction be met by a capacity to contest it; that no relation achieve its result by eliminating the other as a subject; and that no interpretation become historically irreversible. An account of a person is exactly such an act. Section 6 states the principles, identifying the inheritance from work on generative justice where one of them carries it, and Section 7 applies them.

The application yields results that no norm of assertion reaches. An ordinary confidential reference is a unilateral interpretation of a co-produced past by one of its two authors, made without the other and withheld from the other, which breaches the second and fifth principles at the moment of issue rather than failing some standard it might have met. Foreclosing a person’s field is a graver wrong than describing them inaccurately, which inverts the ordering that an ethics of assertion imposes. And a capacity to contest must be built into the instrument when it is issued, since the authority is conferred then, rather than supplied afterwards as an appeal.

The third consequence concerns the shape an ethics for this act can have at all. Where a judgment governs whether a person is admitted, the performance of those excluded is never observed. The party exercising the power can therefore never establish that they exercised it well, and no accumulation of experience will supply the missing evidence. An ethics of correctness is unavailable to them. What remains is an ethics of conduct under permanent unverifiability: declare the basis, bound what is claimed, keep the exercise reopenable, and ensure the person affected can contest it. Section 11 develops this and Section 13 shows that the law of evidence reached a related conclusion long ago, restricting the use of general judgments of persons rather than attempting to improve their accuracy.

Two claims in the paper will be resisted and both are stated plainly here.

Refusing to write is ordinarily treated as the safe course, and this paper holds that it is the act requiring the stronger justification. Section 10 argues that a refusal forecloses a field as effectively as an unfavourable account, does so invisibly, leaves no artefact against which anything can be appealed, and is cheap for the party refusing while expensive for the person refused. The burden of justification therefore falls on silence.

And the duties in this paper bind the recipient as well as the issuer. Section 9 argues that they are correlative, and that a restraint binding the issuer alone is self-defeating: an issuer who properly limits a conclusion has no control over a reader who treats the limited conclusion as unlimited, so unilateral restraint penalises the scrupulous issuer and rewards the unscrupulous one. There is evidence that this has already happened. Where duties were imposed on those who give references and none on those who use them, the reported response was withdrawal from disclosure, and Section 14 treats that evidence as the paper’s principal counter-argument rather than as a difficulty to be noted.

The paper concedes what it borrows. The idea that assertability may be indexed to a speaker’s access rather than to their state of belief is developed in recent work on norms of assertion, and Section 3.1 states what this paper adds to it. The idea that a speaker who tells something assumes a responsibility to the hearer belongs to the assurance view of testimony, and Section 3.2 states that the paper’s move is to convert the hearer’s entitlement into a duty. The principle that value should return to those who generated it is inherited from work on generative justice, and Section 6.3 identifies what is inherited and what is not. The law has already imposed duties on those who give references, and Section 2.2 treats that as the paper’s premise rather than its finding.

Section 2 supplies the practice and the vocabulary. Section 3 locates the account. Section 4 states the method. Section 5 and Section 6 develop the two routes to the duties and Section 7 joins them. Section 8 and Section 9 state the duties of each party. Section 10 treats refusal, Section 11 the shape of an ethics under unverifiability, and Section 12 the case of an account signed without observation. Section 13 examines the law of evidence, Section 14 the evidence against the paper’s own proposals, and Section 15 the institutions that would have to carry them. Section 16 states what the analysis returns to the wider framework, Section 17 records the limits, and Section 18 consolidates the position.

Background and Preliminaries

This section describes the occasions on which such judgments are solicited, records the duties already imposed on those who give them, and states the vocabulary carried from the preceding papers. Analysis begins in Section 5.

Occasions on Which a Judgment of a Person Is Solicited

Four features recur across the occasions this paper treats, and the argument uses each.

The request comes from a third party.

The person who will act on the judgment asks for it. The party who gives it ordinarily has no independent reason to produce it and does so because it was requested, which makes the act one performed within a relation to the requesting party as well as about a relation to the person judged.

The person judged is absent.

They stand outside the exchange, ordinarily remain unaware of what is said, and in many settings have waived in advance any right to see it. They are the subject of the transaction and not a party to it.

The judgment is acted on.

Something follows: an appointment, an admission, a promotion, a transfer of care. The act is consequential in a way that distinguishes it from describing a person to someone with no power over them.

The occasions differ in almost everything else.

A reference is written by one party who chose to accept the request. A performance rating is produced under an institutional schedule. A clinical formulation is made for one purpose and travels to another. A political endorsement is public and contested. This paper treats them together because of the four features above and leaves open whether they are alike in other respects.

Duties Already Imposed on Those Who Judge for Others

Duties of the kind this paper discusses already exist, are stated with some precision, and are enforced. The paper’s premise is their existence; its argument concerns their shape and their one-sidedness.

Professional codes.

The governing ethical standard for psychological assessment requires that opinions rest on information and techniques sufficient to substantiate the findings; that opinions about the psychological characteristics of individuals be given only after an examination adequate to support them, failing which the psychologist must document the efforts made, clarify the probable impact of the limited information on the reliability and validity of the opinion, and limit appropriately the nature and extent of the conclusions; and that where a record review is conducted without examination, this be explained and the sources stated (American Psychological Association, n.d.). Psychiatric ethics go further for a particular case, providing that it is unethical to offer a professional opinion about an individual who has not been examined and for whom proper authorization has not been granted (American Psychiatric Association, n.d.).

Assessment standards.

The standards governing educational and psychological testing locate validity in the interpretations of results for proposed uses rather than in instruments, so that a validation argument is bounded by the use for which it was constructed (American Educational Research Association, American Psychological Association, and National Council on Measurement in Education 2014).

The law of references.

In England and Wales an employer who gives a reference owes the subject a duty of care in its preparation (“Spring v Guardian Assurance Plc” 1994); a reference must be true, accurate and fair and must not give a misleading impression, whether by omission or by nuance, though it need not be comprehensive (“Bartholomew v London Borough of Hackney” 1998); and reasonable care must be taken that it is so, which was not satisfied where unproven and uninvestigated allegations were conveyed (“Cox v Sun Alliance Life Ltd” 2001). In the United States a writer of a recommendation owes third parties a duty not to misrepresent facts where doing so creates a substantial and foreseeable risk of physical injury, so that an unreservedly favourable account given by parties who knew of serious complaints was actionable (“Randi w. V. Muroc Joint Unified School District” 1997). The general principle behind the civil liability is that one who supplies information for the guidance of others in their business transactions is liable for pecuniary loss caused by justifiable reliance on it where reasonable care was not exercised, within limits fixed by the persons for whose benefit the information was supplied (“Restatement (Second) of Torts, Section 552: Information Negligently Supplied for the Guidance of Others” 1977).

The asymmetry these establish.

Every duty above binds the party who issues a judgment. None binds the party who uses it. A referee who correctly limits a conclusion has no control over a reader who treats the limited conclusion as unlimited, and no provision above addresses the reader at all. Section 9 argues that this asymmetry is a structural defect making the issuer-side duties self-defeating, rather than an oversight to be corrected at the margin, and Section 14 examines evidence that the defect has already produced its predictable effect.

Vocabulary Carried from the Preceding Papers

The account uses a small vocabulary established across this series.

A relation is an ongoing process between parties rather than a state obtaining at a moment. The generativity of a relation is its capacity to continue producing outcomes, including outcomes no party can specify in advance; it is a property of the relation and not of either party.

A relational condition is an arrangement whose presence or absence changes which relations can be formed or continued, without determining which of them are. The field of a person is the set of relations available to them under the conditions then in place. Foreclosure is the removal of conditions such that relations formerly available become unavailable.

Co-experience is what a relation generates between its parties: the undertakings, difficulties, judgments and understandings that arose in it and belong to neither party alone. A relational position is the standpoint from which a party stands in a relation, which fixes what of the relation is available to that party.

A situated judgment is a judgment of a person formed within a relation, on the basis of what that relation made available from the position the judging party occupied.

Two results of the preceding paper are carried here as premises and are taken as given. The first is that a situated judgment carries warrant with respect to the relation in which what it reports was exhibited, so that extending it to relations supplying different conditions is an error about scope rather than about accuracy. The second is that instruments of this kind transmit a field of relational conditions rather than an account: they alter the relations available to the person in the receiving system, and they do so whether or not any account of a prior relation underlies them, the account being the justification offered for the alteration rather than the thing transmitted. Section 5 draws the ethical consequence of the second, and Section 12 treats the case that established it.

Literature Review

This section locates the account. It concedes at the outset that two of the paper’s central moves have near neighbours, states what the paper adds to each, and identifies the one position on which the survey found the literature genuinely thin.

Norms of Assertion and the Position to Know

The received treatment of when one may say something about another proceeds from a norm on assertion.

The baseline holds that one must assert a proposition only if one knows it, supported by the observation that strong statistical evidence does not license assertion and that conjoining an assertion with a denial of knowledge is defective (Williamson 1996, 2000). Competing formulations substitute truth (Weiner 2005), rational credibility (Douven 2006), or the requirement that the assertion express the knowledge rather than merely accompany it (Turri 2011). A prominent objection holds that a speaker may properly assert what she withholds belief from, where she possesses evidence of the appropriate kind, which motivates a norm of reasonable belief rather than of knowledge (Lackey 2007).

One recent formulation is close enough to this paper’s position that the concession must be made before anything is claimed. On that account one should assert a proposition only if one is in a position to know it, and the norm is defended as foregrounding the public and other-regarding character of assertion (Willard-Kyle 2020). What matters here is the form of the norm rather than its defence: assertability is indexed to the speaker’s access rather than to the speaker’s state of belief, which is the structure this paper’s scope norm shares.

Two additions distinguish the present norm and both are stated in Section 8.3. (1) The index is a relation rather than a proposition. Being in a position to know is a relation between a speaker and a fact; the present norm indexes what may be asserted to the relation within which the conduct reported occurred, so that the same speaker may be in a position to know a great deal about a person under one set of conditions and nothing about them under others. (2) The conditions of access form part of what must be asserted. A speaker who is in a position to know satisfies the earlier norm by asserting the proposition alone; the present norm is satisfied only if the conditions under which the speaker was so positioned are conveyed with it.

If either addition collapses into the earlier norm, the scope norm is an application rather than a new norm, and Section 17 records this as the paper’s principal exposure in this area.

Testimony, Assurance, and the Responsibility of a Speaker

A second literature bears on the paper’s correlativity claim and partly anticipates it.

The epistemology of testimony treats reliance on what others say as a basic source of warrant rather than as a derivative one (Coady 1992), and the dualist development holds that testimonial knowledge is a joint product requiring positive contributions from the speaker, whose statement must be reliable, and from the hearer, who must possess appropriate positive reasons and lack defeaters (Lackey 2008). That is the strongest existing two-sided structure and it is conceded here.

The assurance view goes further in the direction this paper takes. On that account, telling someone something goes beyond producing evidence for them: it gives an assurance, on the model of a promise, whereby the speaker assumes a responsibility to the hearer and the hearer acquires a standing to complain if the assurance fails (Moran 2005, 2018); the parallel formulation treats telling as an invitation to trust that commits the teller (Hinchman 2005). The standard objection holds that assurance is neither necessary nor sufficient for a testimonial reason, since an overhearer may gain warrant without being assured of anything (Schmitt 2006).

The concession and the difference can now be stated precisely. The assurance view already establishes that a speaker who tells assumes a responsibility running to the hearer. What it gives the hearer in return is an entitlement: a standing to complain, to rely, to hold the speaker to what was assured. This paper’s claim is that in the case at issue the hearer also incurs a duty, and Section 9 argues that without it the speaker’s restraint is self-defeating.

Obligations of the Hearer in Testimony

The survey underlying this paper found the hearer’s side thin, and the thinness is worth stating exactly, since the paper’s most novel claim sits here.

The classical position holds that a hearer who accepts testimony uncritically is culpably gullible and must monitor for signs that the speaker is untrustworthy (Fricker 1994). Whether that monitoring is a standing requirement or is engaged only where something prompts it has been argued directly (S. Goldberg and Henderson 2006), and a related programme locates part of a hearer’s warrant outside the hearer altogether, in the reliability of the processes and the surrounding epistemic community (S. C. Goldberg 2010).

Each of these concerns whether the speaker is to be believed. None concerns the scope of what is being said: whether the statement, taken as entirely sincere and entirely competent, warrants what the hearer proposes to do with it. A hearer who has satisfied every duty in this literature may still rely on a scrupulously limited account as though it were unlimited, and nothing in the literature identifies that as a failure. Section 9.1 states the duty that would.

Intellectual Humility and the Owning of Limitations

The virtue-theoretic literature supplies the disposition the issuer’s duties would express.

Responsibilist virtue epistemology treats intellectual virtues as acquired excellences of character bearing on the regulation of belief and its expression (Zagzebski 1996; Roberts and Wood 2007), and the vice opposed to restraint has been analysed as a form of self-indulgence (Battaly 2010). The account nearest to this paper defines intellectual humility as the owning of one’s epistemic limitations, requiring proper attentiveness to them and an appropriate response (Whitcomb et al. 2017).

The relation to the present argument is one of expression rather than of derivation. A person who owns their limitations will, other things equal, state the conditions of their access and decline where the access will not bear what is asked. But the duties in Section 8 are requirements on the act rather than on the character of the actor, and they hold of an issuer who lacks the virtue entirely. The paper takes the virtue literature as describing what discharging the duties would look like in a person, and not as grounding them.

Correlativity of Rights and Duties

The architecture the paper needs for its central structural claim is long established in jurisprudence.

The analysis of jural relations distinguishes claim-rights, privileges, powers and immunities, and pairs each with a correlative occupied by the other party, so that a claim-right in one party is a duty in the other, a privilege in one is the absence of a right in the other, a power in one is a liability in the other, and an immunity in one is a disability in the other (Hohfeld 1913, 1917). The relevant feature is that positions come in pairs: to specify one party’s position is to specify the other’s.

Section 9.3 uses this to state what is wrong with a duty resting on the issuer alone. The assurance view pairs the speaker’s responsibility with the hearer’s entitlement, which is a coherent pairing. The paper’s argument is that in this case the pairing is the wrong one: an entitlement in the hearer leaves the hearer free to rely as they choose, and it is the freedom to rely without limit that makes the issuer’s restraint worthless.

Lying, Misleading, and Withholding

The case examined in Section 12 requires a distinction that has been analysed at length and contested.

The relevant treatment distinguishes lying, which involves saying what one believes false in a context warranting the truth of what is said, from merely misleading, which involves conveying a falsehood without saying it; and it argues, against a common intuition, that there is no general moral difference between the two (Saul 2012). A subsequent development refines the boundary by reference to what is put on the conversational record (Stokke 2018).

The bearing on this paper runs against it and is recorded here. Section 12 argues that an account signed without observation is assessed differently under an ethics of reporting than under an ethics of authorising, and that the two assessments select different remedies. If lying and misleading are morally on a par, then the difference between misdescribing a relation and authorising without having observed one may matter less than the paper claims, and Section 12.4 engages the point rather than passing it.

Generative Justice and the Return of Value to Its Producers

One of the principles stated in Section 6 is inherited, and the inheritance is identified here so that Section 6.3 can demonstrate exactly what is taken and what is not.

The account of generative justice holds that the destructive force of extractive arrangements lies in the alienation of value from those who generate it, and defines the position in three limbs: a universal right to generate unalienated value and to participate directly in its benefits; the right of value generators to create their own conditions of production; and the right of communities of value generation to nurture self-sustaining paths for its circulation (Eglash 2016). The account is developed across ecological, labour and social value, and locates the difficulty in extraction as such rather than in who does the extracting, so that centralised redistribution and private appropriation are treated alike.

Two features bear on this paper and the second is the one that does the work. The first limb supports the familiar claim that those who generated something should share in what it yields. The second limb is stronger and less often noticed: those who generate value have a claim over the conditions under which production occurs, and not only over its proceeds. Applied to an account of a shared past, the person judged co-generated what is being reported and has no say whatever in the conditions under which the report of it is produced: who writes it, when, in what form, and under what confidentiality. Section 6.3 states which of the framework’s principle is inherited from this account and which part of it is not.

Boundary of the Present Contribution

Table [tab:antecedents5] records what each literature licenses and where the contribution begins.

@P0.22YY@ Literature & Licensed role & Boundary of the present contribution
Norms of assertion & Assertability indexed to access rather than to belief (Willard-Kyle 2020); the knowledge norm and its rivals (Williamson 1996; Lackey 2007) & The index is a relation rather than a proposition, and the conditions of access form part of what is asserted.
Assurance and testimony & The teller assumes responsibility to the hearer (Moran 2005; Hinchman 2005); two-sided production of testimonial knowledge (Lackey 2008) & The hearer’s entitlement is converted into a duty, on the ground that an entitlement leaves the issuer’s restraint worthless.
Hearer’s obligations & Monitoring for untrustworthiness (Fricker 1994); the scope of that requirement (S. Goldberg and Henderson 2006) & Concerns whether to believe rather than what the statement warrants; the scoping duty is not found there.
Intellectual humility & Owning one’s epistemic limitations (Whitcomb et al. 2017) & Describes the disposition that discharging the duties expresses; the duties bind an issuer lacking the virtue.
Correlativity & Jural positions come in pairs (Hohfeld 1913, 1917) & Supplies the architecture; the claim that the existing pairing is the wrong one is made here.
Lying and misleading & The distinction and the argument against a general moral difference (Saul 2012) & Bears against the paper’s divergence claim and is engaged rather than avoided.
Generative justice & Value returns to its generators; generators hold a claim over the conditions of production (Eglash 2016) & Inherited, and identified as inherited where the framework’s principle carries it.
Law of references and of hiring & Enforced duties on issuers (“Spring v Guardian Assurance Plc” 1994; “Bartholomew v London Borough of Hackney” 1998; “Cox v Sun Alliance Life Ltd” 2001; “Randi w. V. Muroc Joint Unified School District” 1997); a proportionate duty of inquiry on employers (“Ponticas v. K.m.s. Investments” 1983; “Connes v. Molalla Transport System, Inc.” 1992) & The paper’s premise. The hiring duty runs to injured third parties and not to the person judged.

Four positions are left unoccupied by the literatures surveyed. No treatment located here states a duty on the party who relies on a judgment to establish its scope before relying on it. None argues that duties on the issuer and the recipient are correlative in the sense that a duty on the issuer alone is self-defeating. None holds that refusing to give an account requires a stronger justification than giving one. And none develops an ethics for a power whose good exercise cannot in principle be established, which Section 11 argues is the situation of every party who decides whom to admit.

The claim is that these four are unoccupied, not that their components are unprecedented; the components are conceded above.

A terminological caution belongs here. Reliance, silence, and the duties of a recipient all carry established and entirely different meanings in the law of evidence and procedure, where reliance is an element of liability for a statement, silence bears on inferences that may be drawn against a party, and the duties of a recipient concern what a tribunal may admit. This paper uses the terms in their ordinary senses and claims no continuity with those doctrines, except where a legal source is cited for a proposition, as in Section 13.

A bounded search for prior use of the paper’s own formulations, covering a duty on the relying party to establish the scope of a judgment, the correlativity of issuer and recipient duties, and the treatment of refusal as the act requiring the stronger justification, returned no related scholarly use. A bounded search establishes that a formulation was not found rather than that it does not exist, and a systematic originality audit remains outstanding and is recorded in Section 17.

Three questions arising here belong to later papers in this series and are marked where they arise rather than argued: how the conditions under which judgments are generated are distributed, which is the subject of the injustice paper and which carries the treatment of testimonial injustice; who holds authority to interpret a shared experience, which is the subject of the jurisprudence paper; and what accumulates to those who issue such judgments, which is the subject of the political-economy paper.

Method and Selection of Cases

This section states the form of argument, why these cases were chosen, and what would count against the account.

Independent Routes to the Same Duties

The paper argues for its duties twice, from independent premises, and the independence is deliberate.

The first route runs from the mechanism. If an instrument of this kind authorises rather than reports, then the duties governing it are those of an actor exercising discretionary power over another’s conditions, and norms governing what may be said are addressed to a different act. Section 5 develops this. It requires the finding of the preceding paper in this series and nothing from the framework’s ethics.

The second route runs from the framework’s ethical principles governing the reproduction of shared experience. Section 6 states them and Section 7 applies them to an account of a person. This route requires the framework and nothing from the mechanism.

That the two converge on substantially the same duties is a result rather than a redundancy, and it has a practical consequence: a reader who rejects the framework retains Section 5, and a reader who rejects the authorisation finding retains Section 6. Where the two routes diverge, Section 7.5 says so.

Selection of the Cases Examined

Two cases are examined at length and each was chosen because it discriminates between positions rather than because it illustrates one.

An account signed without observation is examined in Section 12 because the ethics of reporting and the ethics of authorising deliver different verdicts on it. Under the first it is a species of deception; under the second it is an exercise of power without grounds. Both condemn it, so the case does not test whether the act is wrong; it tests which description of the wrong is correct, and the descriptions select different remedies.

The restriction of character evidence is examined in Section 13 because it is an instance in which an institution faced this paper’s problem and responded by limiting what a judgment of a person may be used for rather than by improving how such judgments are made. It is a precedent rather than an illustration.

Two absences are recorded. The unfavourable account is treated within Section 10 rather than as a separate case, since the decision to report unfavourably and the decision to decline are the same decision seen from two sides. And no case is examined in which the duties proposed here have been implemented, because none was located; the paper’s proposals are accordingly untested.

Conditions of Disconfirmation

The account should be narrowed or withdrawn under any of the following conditions.

First, if the additions distinguishing the scope norm from an access-indexed norm already in the literature (Willard-Kyle 2020) collapse into that norm, the scope norm is an application rather than a contribution, and Section 8.3 should be rewritten to say so.

Second, if a duty on the issuer alone can be shown to produce the intended restraint without a corresponding duty on the recipient, the correlativity claim of Section 9.3 is false and the paper’s central structural argument fails.

Third, if refusal can be shown to leave a person’s field unchanged, the inversion of Section 10 is false and declining reverts to the safe course.

Fourth, if the good exercise of this power can be established after the fact, the argument of Section 11 is unnecessary and the ethics may be an ethics of correctness after all.

Fifth, if the duties proposed here are shown to worsen what they govern, the proposals should be withdrawn rather than qualified. Evidence that issuer-side duties have already produced withdrawal from disclosure exists and is treated in Section 14, and evidence that codes of this kind do not alter conduct is treated in Section 15.

Judging for Others as an Exercise of Power over a Field

This section develops the first route. It argues that the act at issue is misdescribed as speech, states what it is instead, and derives the duties that attach to acts of that kind.

Authorising Rather Than Reporting

The finding is taken from the preceding paper in this series and restated because everything in this section rests on it.

A referee is asked for an account and replies that the candidate should draft it and the referee will sign; or asks for a curriculum vitae and writes from that. No relation is reported, because either none exists or nothing of it enters the document. The instrument functions as it would otherwise: it is accepted, weighed, and produces the same effect on the outcome.

If what such instruments did were to state something about a relation, an instrument stating nothing would accomplish nothing. It accomplishes something. What it does is alter the relations available to the person it concerns: it opens a position, or closes one, and what the person subsequently generates is produced in the configuration that follows. The account of the relation is the justification offered for the alteration and not the thing transmitted.

Two clarifications belong here. The finding leaves the content of such accounts relevant. Content bears on whether the alteration is justified, and most of this paper’s duties concern content. It holds that the alteration is what the instrument effects, and that the alteration occurs whether or not the content supports it.

And the finding concerns the act rather than what those who write such accounts believe themselves to be doing. They ordinarily believe they are conveying an assessment, and Section 5.2 explains why the belief is natural and why it persists.

Inapplicability of Norms Governing Assertion

The consequence for the ethics is immediate and it disqualifies the approach that would otherwise be taken.

Norms of assertion regulate saying. Whether the norm requires knowledge (Williamson 1996), reasonable belief (Lackey 2007), or a position to know (Willard-Kyle 2020), what it governs is the conditions under which a speaker may put a proposition forward as true. Every such norm presupposes that the act under evaluation is the putting forward of a proposition.

The act evaluated here is the alteration of the relations available to a person. Nothing in any norm of assertion addresses whether one may open or close a position for another, on what grounds, or subject to what review. A speaker who satisfies the strictest norm of assertion available has said nothing improper and may still have exercised a power over another person without grounds, without notice, and without recourse.

Claim 1 (Misdescription of the act). An ethics of judging persons for others that proceeds from norms of assertion evaluates the wrong act. It evaluates the statement offered in justification and leaves unevaluated the alteration the statement is offered to justify.

Claim 1 explains a feature of existing practice that is otherwise puzzling. The professional codes surveyed in Section 2.2 are constructed almost entirely as constraints on what may be asserted: have adequate grounds, limit the conclusions, name the sources (American Psychological Association, n.d.). They contain no provision about the duration of the effect, about notice to the person judged, or about review. If the act were assertion, that would be the right shape. Since the act is the alteration of a person’s conditions, the codes regulate the justification and leave the exercise itself untouched.

Two objections are recorded and answered in part.

The first holds that the alteration is effected by the receiving institution and not by the issuer, so that the issuer really does only speak. The reply is that the issuer’s act is a necessary component of an exercise that neither party performs alone, and that a participant in a joint exercise of power over another does not escape the duties of the exercise by pointing at the other participant. Section 9 distributes the duties between them rather than assigning them all to either.

The second holds that norms of assertion should not be extended to institutional practices at all, since they were developed for individual speech acts. The reply concedes the premise and turns it: this paper declines to apply them rather than extending them, and the burden falls on an account that would apply norms designed for individual speech to an institutional exercise of power.

Duties Attaching to Discretionary Power over Another’s Conditions

If the act is an exercise of discretionary power over another’s conditions, then the duties that attach to it are those attaching to such exercises generally. Four are stated here and developed in Section 8 and Section 9.

(1) Grounds. The exercise must rest on something, and the party exercising must have established that it does before exercising rather than afterwards.

(2) Statement of grounds. The grounds must be stated, and stated in a form that permits another party to assess whether they support what was done. An exercise whose grounds are unstated cannot be assessed at all, which is a defect independent of whether the grounds were good.

(3) Proportion. What is done must not exceed what the grounds support. This is the duty that Section 8.3 states as a scope condition, and it is the point at which the two routes of Section 4.1 first meet.

(4) Review. The exercise must remain open to challenge by the party affected. An exercise that is final at the moment it is made is merely unreviewable, and is no better made for being final.

None of these is novel as a constraint on discretionary power, and the paper claims none of them as such. What is claimed is that these are the duties governing this act, and that the duties ordinarily discussed in connection with it govern a different one.

Direction of the Obligation Owed

A consequence follows that inverts the direction of obligation assumed in the literature surveyed in Section 3.2.

On the assurance view, the party who tells assumes a responsibility to the party told (Moran 2005). That is the correct account of an assurance. It is the wrong account of this act.

Claim 2 (Primary obligation to the person judged). Where the act alters the relations available to a person, the primary obligation runs to that person, who is absent from the transaction and does not consent to it. The obligation to the party who asked is real and derivative: they are owed a justification adequate to the decision they must make, and they are owed it because they will exercise the power jointly.

Claim 2 reorders the duties in Section 8. On an assurance account the issuer’s principal wrong is to mislead the recipient. On the present account the issuer’s principal wrong is to alter a person’s conditions on grounds that will not bear it, and misleading the recipient is one way of doing that rather than the wrong itself.

The reordering also explains why the person judged is absent from every code surveyed in Section 2.2. Those codes were built on the assumption that the duty runs to the party who asked, which is what an ethics of assertion yields. Where the obligation runs primarily to the person judged, their absence from the transaction is the first thing requiring justification, and Section 7.2 argues that it cannot be justified.

The Ethical Principles of Relational Reproduction

This section develops the second route. The framework within which this paper works holds six principles governing acts that reproduce shared experience, and an account of a person given to a third party is such an act. The principles are stated here; Section 7 applies them.

Two preliminary remarks. The first five are ordered so that each presupposes the one above it, and the sixth stands outside that ordering as a principle governing how the others change in history. And one of the six carries an inheritance from an existing account, which Section 6.3 identifies limb by limb rather than absorbing.

Preservation of the Possibility of Continued Becoming

Principle 3 (Generativity). An act that reproduces shared experience should preserve the possibility of continued becoming for those whose experience it reproduces. The requirement is openness, and not the continuation of any particular relation or arrangement.

The distinction in the second sentence carries most of the principle’s content and is easily lost. It leaves relations free to end, arrangements free to lapse, and everything free to change. A relation that ends may leave every possibility open; an arrangement scrupulously maintained may close them.

What the principle forbids is foreclosure: the removal of conditions such that relations formerly available become unavailable, without a route by which they might become available again. Applied to the act at issue, an account that closes a person’s field is assessed under this principle, and Section 7.4 argues that the assessment is graver than the assessment of an account that describes them inaccurately.

Shared Authorship of a Shared Past

Principle 4 (Co-reproduction). What a relation generated belongs to its parties jointly. It may not be interpreted, defined, or revised unilaterally by one of them. Authorship of a shared past is shared.

Principle 4 presupposes Principle 3: a party whose account of the shared past is imposed on them has had a possibility closed, namely the possibility of that past being understood otherwise.

The principle concerns interpretation rather than fact. Two parties may agree entirely about what occurred and differ about what it was an instance of, and the principle governs the second. It is at the second that an account of a person operates, since what is reported is a characterisation of occurrences rather than a list of them, and Section 7.1 argues that the ordinary practice violates the principle at exactly this point.

Non-Exploitation and the Return of Value to Its Producers

Principle 5 (Non-exploitation). No party may be instrumentalised or sacrificed for another’s benefit. Where value is generated jointly, it should return to those who generated it, and those who generate it hold a claim over the conditions under which it is produced.

This principle is the floor: it presupposes the two above and states the limit below which an arrangement is not merely deficient but exploitative.

Inheritance from the account of generative justice.

The second sentence is not the framework’s own and is stated here as inherited. The account of generative justice holds that the destructive force of extractive arrangements lies in the alienation of value from those who generate it, and defines the position in three limbs: a universal right to generate unalienated value and to participate directly in its benefits; the right of value generators to create their own conditions of production; and the right of communities of value generation to nurture self-sustaining paths for its circulation (Eglash 2016).

The inheritance is limb by limb. The first limb is taken in full and supplies the clause requiring that value return to those who generated it. The second limb is taken in full and supplies the clause about the conditions of production, which is the part of the principle that does most of the work in Section 7.3 and which the framework did not previously state. The third limb, concerning the rights of communities to sustain paths of circulation, is left aside, because the present paper concerns an act between particular parties rather than the design of a circulatory system, and taking it would commit the paper to claims about institutional economies that it does not argue.

Elements originating outside that account.

The first sentence of the principle, prohibiting instrumentalisation, comes from elsewhere and is older. And the framework’s application differs in domain: the account cited concerns ecological, labour and social value in systems of production, whereas the principle is applied here to what a relation between two persons generated. The transposition is the framework’s, and any error in it rests with the framework rather than with the account inherited from.

Simultaneous Generation of Counter-Power

Principle 6 (Counter-power). Every act that reproduces shared experience confers interpretive authority on the party performing it. A just reproduction must therefore generate, at the same moment, the capacity to question, revise, and renegotiate what it established.

Two features distinguish this principle from a right of appeal and both are load-bearing.

The first is its ground. The authority is conferred by the act itself, in the moment of performing it, rather than by an institution afterwards, because to fix an interpretation of a shared past is to occupy the position from which that past is understood. What must be met is therefore generated by the same act that requires meeting.

The second is the simultaneity. A capacity to contest supplied afterwards, as an appeal against a determination already acted on, arrives after the authority has been exercised and after its effects have propagated. The principle requires that the capacity be generated with the act, which is a requirement on the design of the instrument rather than on the availability of a forum. Section 8.5 states what this asks of an issuer and Section 9.2 what it asks of a recipient.

The principle’s purpose is stated in the framework as preventing history from freezing. An interpretation that confers authority without generating the means to contest it accumulates authority in one direction, and repetitions of the act compound rather than correct.

Preservation of the Other as a Subject

Principle 7 (Subject preservation). No relation achieves a good result by eliminating the other as a subject. Union falls short of identity, and the goal lies elsewhere than in agreement; what must survive the relation is the other’s standing as a party who can speak, differ, and answer.

Principle 7 presupposes the preceding four and states what they jointly protect. A party who cannot speak in a matter concerning them has been eliminated as a subject in it, whatever else is preserved.

The principle bears on a feature of the practice that is ordinarily treated as administrative. Where an account of a person is confidential to that person, they cannot see what is said, cannot answer it, and cannot be heard in the proceeding it governs. Section 7.2 argues that this is a breach of this principle, rather than a procedural defect attaching to an otherwise sound practice, in the very relation whose product is being reported.

Historical Revisability and the Conditions That Sustain It

Principle 8 (Historical revisability). No interpretation of what a relation generated, and no interpretation of what justice required in it, should become historically irreversible. Revisability must obtain at three layers: interpretations must remain contestable, institutions must remain alterable, and today’s understanding of what is just must remain open to revision by future relations. And the conditions under which revision is possible must themselves be preserved, since a right to revise is empty where what would be needed to exercise it has not been kept.

Three features of the principle are used later.

Position of the principle outside the axiomatic ordering.

The first five principles constrain relations. This one governs how those constraints change in history, and it therefore applies to itself. It is a meta-principle and not a sixth constraint of the same kind.

Asymmetry in who retains the standing to reopen.

The principle is stated in terms of who retains the standing to reopen, and the standing rests with the party bearing the cost of the interpretation rather than with the party who made it or the party who acted on it. A remedy on this account is a path by which the party who bore the cost re-enters the process and contests the interpretation that stands, rather than a correct judgment, which is continued participation in the historical generation of justice rather than compensation for its failure.

Preservation of the conditions of revision.

This is the extension the final sentence of the principle carries, and it is what distinguishes the principle from a right of appeal that is formally available and practically empty. An interpretation may be nominally open to revision while everything needed to revise it has been discarded: the record of what was said, the accounts of parties who observed differently, the divergences that were resolved before the decision was recorded. Preserving those is not a courtesy owed to a future contestant; it is what makes the right to contest a right to anything. Section 9.2 states this as a duty of the party who receives such judgments, and it is the point at which this principle bears most heavily on the paper’s practical proposals.

Application of the Principles to an Account of a Person

This section applies the six principles to the act at issue and joins the two routes. Four results follow, and three of them are unavailable to an ethics built on norms of assertion.

Unilateral Interpretation of a Co-Produced Past

An account of a person given to a third party reports what a relation generated. Under Principle 4 what a relation generated belongs to its parties jointly and may not be interpreted unilaterally by one of them.

The ordinary practice satisfies neither condition. One of the two parties characterises the shared past; the other neither participates in the characterisation nor sees it; and the characterisation is then given institutional force. The act is a unilateral interpretation of a co-produced past by one of its two authors.

Claim 9 (Breach by construction). An account of a person given without the participation of the person judged breaches Principle 4 in virtue of how it is produced, and not in virtue of anything it says. An account that is entirely accurate, carefully limited, and generously worded breaches it in the same way as one that is none of these.

Claim 9 states something stronger than any scope norm can, and the difference is worth marking. A scope norm holds that an account asserting more than the relation affords has exceeded what it may claim; it is satisfied by an account that stays within its scope. Claim 9 holds that the practice is in breach at the point of production, so that staying within scope is necessary and insufficient.

Two qualifications bound the claim. It leaves open whether such accounts should be given, and Section 10 argues that they should. What follows is that the practice carries a standing defect requiring justification, and that the justification cannot be that the account was accurate.

And the breach admits of degree in one respect: the further the account moves from occurrences toward characterisations of the person, the more of the co-produced interpretation is being appropriated. An account confined to what happened appropriates less than one stating what the happenings show about the candidate, which is the distinction Section 8.4 states as a duty.

Confidentiality and the Elimination of the Subject

Under Principle 7 no relation achieves a good result by eliminating the other as a subject, where what must survive is the other’s standing as a party who can speak, differ, and answer.

Where an account of a person is confidential to that person, they cannot see what is said of them, cannot answer it, and cannot be heard in the proceeding it governs. In the matter that concerns them most directly they have no standing to speak at all.

This is ordinarily treated as an administrative arrangement adopted for good reasons, and the reasons are real: candour is thought to depend on it, and Section 14 takes that argument seriously. Under Principle 7 it is something more than an administrative arrangement. It is the elimination of a party as a subject in the very relation whose product is being reported, and the fact that it is adopted for good reasons makes it a justified breach at best rather than no breach.

Two features sharpen the result.

The elimination is effected by the party who was the other author of the shared past. It is the referee, not the institution, whose account cannot be seen, and the referee ordinarily requires the confidentiality or accepts it as a condition of writing.

And the elimination is ordinarily consented to in form. The person judged waives the right to see what is said, and the waiver is offered as a choice. What is offered is a choice between an account they may read and an account that will be believed, which is a choice made under a condition the person judged did not set. Section 9 treats the waiver as a matter for the receiving party rather than for the issuer, since it is the receiving party that attaches the weight.

Accumulation of Standing from Others’ Co-Production

Principle 5 holds that value generated jointly should return to those who generated it, and that those who generate it hold a claim over the conditions under which it is produced. Both limbs apply and the second applies more sharply.

Return of accumulated standing to the issuer.

A referee whose accounts are accepted accumulates something: a record of endorsements taken up, a standing as a party whose judgment carries, and a set of former candidates with reason to regard them favourably. What is accumulated was generated in relations the candidate co-produced, and it returns to the referee. This paper records the point and leaves its development aside, since the accumulation and circulation of such standing is the subject of the political-economy paper in this series.

Absence of any say in the conditions of production.

The person judged has no say whatever in the conditions under which the account of their shared past is produced. They nominate but do not choose who writes it, and the timing, the form, the length, the confidentiality and the questions are all settled elsewhere. Every condition of production is set by the receiving institution and the referee between them.

Under Principle 5 this is a distinct wrong from receiving none of the value, and it is the one that admits of a remedy the paper can state. A person who cannot share in the standing an account generates may still hold a claim to participate in setting the conditions under which the account of their own shared past is produced: to know what questions were asked, to supply the conditions under which the conduct reported occurred, and to know that an account has been given at all. Section 8.2 and Section 9.2 state what follows.

Foreclosure Weighed against Inaccuracy

Principle 3 requires that an act reproducing shared experience preserve the possibility of continued becoming. The requirement is openness rather than continuation, and the consequence for this paper inverts an ordering that an ethics of assertion imposes without argument.

On an assertion account the gravity of a wrong tracks the divergence between what was said and what is the case. An inaccurate account is the paradigm wrong; an accurate account that closes a person’s options counts as no wrong at all, since nothing improper was said.

Claim 10 (Priority of foreclosure over inaccuracy). Under Principle 3 an accurate account that forecloses a person’s field is a graver wrong than an inaccurate account that leaves it open. Accuracy bears on whether the foreclosure was justified, and leaves the gravity of what was done undetermined.

Claim 10 will be resisted and three clarifications are offered.

It leaves inaccuracy unlicensed. An inaccurate account that forecloses is worse than either, and accuracy remains a condition on the justification of any foreclosure. What the claim denies is that accuracy is the measure of the wrong.

It concerns foreclosure rather than adverse outcome. An account that leads to a person’s not being appointed has not thereby foreclosed anything, since the positions remain available and the person may be appointed elsewhere or later. Foreclosure is the removal of conditions such that what was available becomes unavailable without a route back, and Section 11 argues that the absence of a route back is the feature that distinguishes it.

And it explains why the duties of Section 8 are weighted as they are. If foreclosure is the graver wrong, then a duty to leave the exercise reopenable is more important than a duty to be right, which is the opposite of the weighting that professional codes adopt.

Convergence of the Two Routes

The duties reached from the mechanism in Section 5 and the results reached from the principles in this section coincide at three points and diverge at one, and both facts are worth stating.

Points at which the two routes agree.

The duty of proportion in Section 5.3 and the scope condition following from Principle 4 require the same restriction on what may be claimed. The duty of review in Section 5.3 and Principle 6 require the same provision for contestation, though the second requires it earlier. And Claim 2, that the obligation runs primarily to the person judged, is reached independently by Principle 7, which makes their standing as a party the thing to be preserved.

The convergence matters because the routes share no premises. One requires the finding that the instrument authorises; the other requires the framework’s ethics. A reader may reject either and retain the duties.

The result available only from the principles.

The principles yield one result the mechanism does not. Claim 9, that the practice is in breach at the point of production regardless of content, follows from Principle 4 and does not follow from the account of the act as an exercise of power. An exercise of power over another’s conditions may be entirely legitimate where the grounds support it; what makes this act defective on the second route is that the grounds concern a past the other party co-authored and is not consulted about.

A reader persuaded only by Section 5 therefore obtains the duties of Section 8 and Section 9 without Claim 9, and Section 17 records which of the paper’s results depend on which route.

Duties of the Party Who Issues a Judgment

This section states what is owed by the party who gives an account of a person to a third party. The duties follow from the four requirements on exercises of discretionary power stated in Section 5.3 and from the principles applied in Section 7, and they are ordered by the weighting Claim 10 imposes rather than by the weighting professional codes adopt.

Establishing the Grounds before Asserting

Duty 11 (Grounds). The issuer must establish that what is to be asserted rests on something, before asserting it. Where an assertion would convey that conduct occurred, the issuer must have established that it occurred and not merely that it was alleged.

This duty is already imposed and enforced. Professional standards require that opinions rest on information sufficient to substantiate the findings (American Psychological Association, n.d.); and the law of references has held an employer liable for conveying allegations of dishonesty that were unproven and had not been investigated (“Cox v Sun Alliance Life Ltd” 2001). The paper adds nothing here and records the duty because the duties that follow depend on it.

Declaring the Position from Which the Judgment Is Made

Duty 12 (Declaration of position). The issuer must state the position from which the account is given: the capacity in which they knew the person, the period, the conditions under which the conduct reported occurred, and the respects in which their view was partial.

The professional analogue is close and is conceded. Where an adequate examination is impracticable, the psychologist must document the efforts made, clarify the probable impact of the limited information on the reliability and validity of the opinion, and limit the conclusions accordingly (American Psychological Association, n.d.). What the present duty substitutes for the examination is the relation: what must be declared is not how much of an assessment was completed but what the relation afforded.

The duty is a requirement on content and not on candour. It leaves the issuer’s doubts, reservations and private view undisclosed. It asks for the conditions of access, which are facts about the arrangement rather than about the issuer’s state of mind, and which a reader cannot recover from the account itself.

Restricting the Assertion to the Conditions of Access

Duty 13 (Scope). The issuer may assert of a person only what the relation they stood in afforded, and the conditions of that access form part of what is asserted.

Duty 13 is the paper’s scope norm and its two departures from the nearest existing norm were stated in Section 3.1. The index is a relation rather than a proposition, so that the same issuer may be well placed regarding a person under one set of conditions and not at all under others. And the conditions form part of what is asserted, so that an account satisfying Duty 12 is a component of satisfying this one rather than a separate courtesy.

Distinguishing Conduct Reported from Disposition Inferred

Duty 14 (Conduct and disposition). The issuer must distinguish what occurred from what the occurrences are taken to show about the person, and must mark the second as an inference drawn from the position stated.

Two grounds support this duty and they come from the paper’s two routes.

From the mechanism: what occurred is a fact about a relation the issuer was party to, and what it shows about a person is a claim extending beyond that relation, which by Duty 13 requires its own support.

From the principles: Section 7.1 recorded that the breach of Principle 4 admits of degree, and that the further an account moves from occurrences toward characterisations of the person, the more of a co-produced interpretation is appropriated by one of its authors. A report of conduct appropriates least; a characterisation of the person appropriates most.

The duty accordingly permits characterisation and requires that it be marked. An issuer who writes that a candidate abandoned a project has reported conduct; one who writes that the candidate lacks persistence has characterised a person, and owes the reader the information that this is what has been done.

Building Contestation into the Instrument at Issuance

Duty 15 (Contestation at issuance). The issuer must provide, in the instrument itself, the means by which what it establishes may later be questioned: what was asked, what was answered, from what position, and how the account may be located again.

This duty follows from Principle 6 and from its simultaneity requirement. The interpretive authority is conferred by the act, so the capacity to meet it must be generated by the act, and an appeal available afterwards arrives after the authority has been exercised.

What the duty asks is modest and it is worth stating exactly, since the principle behind it is strong. It leaves the issuer free of any obligation to establish a forum, to guarantee a hearing, or to undertake anything toward a future contest. It asks that the instrument carry what a contest would need: the question the issuer was answering, the position from which they answered, and a record that persists. An instrument that carries none of these is not merely unappealed but unappealable, since there is nothing for a challenge to engage.

Section 9.2 states the corresponding duty on the receiving party, and the two together are what Principle 8 requires when it holds that the conditions of revision must themselves be preserved.

Authorising without Observation

Duty 16 (Observation). The issuer may not authorise an alteration in a person’s conditions on the strength of a relation they did not stand in. Where the issuer’s access is null, the ground of the exercise is absent, whatever the account says.

Duty 16 is a special case of Duty 11 and is stated separately because the case it governs is common and is not ordinarily regarded as a breach. Section 12 examines it, and the statement here anticipates the finding: on an ethics of reporting the wrong is that something false is conveyed, while on an ethics of authorising the wrong is that a power is exercised with no ground at all, which is available even where every sentence in the account is true.

Duties of the Party Who Receives a Judgment

This section states the paper’s most novel claim and the argument that makes it necessary. The duties above bind the party who gives an account. This section argues that a corresponding set binds the party who uses it, and that without them the first set is worse than useless.

Establishing the Scope before Relying

Duty 17 (Scope establishment). Before relying on an account of a person, the receiving party must establish its scope: the position from which it was given, the conditions under which the conduct reported occurred, and therefore what the account warrants. An account whose conditions of access are unstated is one whose scope cannot be established, and it must be treated as such.

The survey underlying this paper found no existing statement of this duty, and Section 3.3 recorded why the nearest neighbours are not it. The hearer’s duty to monitor for untrustworthiness (Fricker 1994) asks whether the speaker is to be believed. The requirement of positive reasons (Lackey 2008) asks whether the hearer is justified in believing. Neither asks what a statement, taken as entirely sincere and entirely competent, warrants the hearer in doing.

The second sentence of the duty is what gives it force. Without it, an account that omits its conditions is simply read as unrestricted, which rewards the issuer who declares nothing. With it, the omission is a defect in the account rather than a licence to the reader, and Section 9.3 argues that this reversal is the whole point.

A legal analogue exists and is instructive in its reach and in its limits. An employer owes a duty of reasonable care in hiring persons who, because of the employment, may pose a threat of injury to members of the public (“Ponticas v. K.m.s. Investments” 1983), and the depth and scope of the inquiry required tracks the degree of risk the position carries, so that exhaustive investigation is required where the post exposes the public and minimal inquiry suffices where it does not (“Connes v. Molalla Transport System, Inc.” 1992; “Di Cosala v. Kay” 1982). That is a duty of inquiry on the relying party and it is proportionate to what is at stake, which is the structure Duty 17 adopts. It is owed, however, to third parties who may be injured rather than to the person judged, and it arises independently of the reference. The law therefore gives the relying party a duty toward strangers and none toward the person whose account they are relying on.

Preserving the Conditions of Later Revision

Duty 18 (Preservation). The receiving party must preserve what a later revision would require: the account as given, the conditions declared with it, the accounts of other parties, and the respects in which they diverged. Divergence resolved before the decision is recorded, and not preserved, is not available to be revisited.

This duty follows from Principle 8 and specifically from the extension it carries: a right to revise is empty where what would be needed to exercise it has not been kept. The receiving party is the only party in a position to discharge it, since it is the party that holds the file.

Three observations bound the duty. It asks for preservation and not for publication, so it is compatible with confidentiality regimes even though Section 7.2 finds those regimes independently objectionable. The distinction matters for a further reason: where a measure concerning persons is made public, those measured alter their conduct in response to being measured, so that the measure changes what it reports (Espeland and Sauder 2007). Preservation without publication avoids that consequence, and forgoes with it the contestation that publicity would enable. It asks that divergence be recorded rather than resolved, which is the requirement the preceding paper in this series argued for on reconstructive grounds and which is reached here on ethical ones. And it is cheap: what it asks is a note in a file that already exists, made at a moment when the material is in front of the decision-maker.

Correlativity of the Two Sets of Duties

The argument for this section’s duties rests on something other than a claim that recipients ought also to behave well. It rests on the issuer’s duties failing without them.

Claim 19 (Correlativity). A scope condition binding the issuer alone is self-defeating. An issuer who states the conditions of access and limits the conclusions accordingly produces an account that is weaker on its face than one that states nothing and claims everything. Where the recipient is free to rely as they choose, the scrupulous issuer’s account is discounted and the unscrupulous issuer’s is not, so the duty penalises compliance.

The jural analysis of correlative positions supplies the form of the point (Hohfeld 1913, 1917). To specify one party’s position is to specify the other’s, and the pairing matters. The assurance view pairs the speaker’s responsibility with the hearer’s entitlement (Moran 2005), which is a coherent pairing for the case it addresses. It is the wrong pairing here: an entitlement leaves the hearer free to rely without limit, and it is that freedom which makes the issuer’s restraint worthless.

Two consequences follow.

The first is that the duties in Section 8 should not be imposed without those in this section. A regime that regulates issuers and leaves recipients unconstrained produces exactly what Claim 19 predicts, and Section 14 presents evidence that this has already occurred.

The second concerns where the paper’s principal charge lies. If the recipient is free to treat a properly scoped account as unscoped, then the recipient commits the error that the preceding paper in this series identified as the central one, namely extending a warrant beyond the relation that supports it. The issuer’s duties are about supplying what the recipient needs; the recipient’s duties are about using it, and the second is where the failure occurs.

Absence of a Duty Running to the Person Judged

One further asymmetry is recorded, since it bears on what the duties above can achieve.

Every duty examined in this paper, whether professional or legal, runs to a party other than the person judged. The professional codes run to the profession and to those who rely on its members (American Psychological Association, n.d.). The law of references runs to the subject in one jurisdiction (“Spring v Guardian Assurance Plc” 1994) and to foreseeably injured third parties in another (“Randi w. V. Muroc Joint Unified School District” 1997). The hiring duty runs to members of the public (“Ponticas v. K.m.s. Investments” 1983). In no case does a duty run from the party who uses an account to the person the account concerns.

Under Claim 2 that is the wrong direction. The party whose conditions are altered is the party to whom the primary obligation is owed, and they are the only party with no standing anywhere in the arrangement. Section 17 records that this paper states the duty and does not supply the mechanism by which it would be enforced, since a duty owed to a party who cannot see the account, was not told it was given, and has no forum in which to raise it is a duty in name.

Refusal as the Act Requiring the Stronger Justification

Declining to give an account of a person is ordinarily treated as the safe course. Nothing is asserted, so nothing can be asserted wrongly; no duty in Section 8 is engaged; and the party who declines has, on the received view, done nothing requiring justification at all. This section argues that the received view follows from the misdescription identified in Claim 1, and that on a correct description the burden of justification falls on refusal rather than on speech.

Foreclosure Effected by Declining

If the act is the alteration of the relations available to a person, then declining to perform it is not abstention from the alteration. Where an institution requires accounts and a person cannot obtain one, the position is closed to them as effectively as it would be closed by an unfavourable account, and more reliably, since an unfavourable account may be discounted while a missing one cannot be supplied.

Under Principle 3 the relevant question is whether the possibility of continued becoming has been preserved. A refusal that closes a position the person could otherwise have reached forecloses in exactly the sense the principle identifies. That the party who refused did nothing is true of their conduct and false of its effect.

Claim 20 (Refusal as exercise). Declining to give an account, where an account is required for a position and the person cannot obtain one elsewhere, is an exercise of the same power as giving one. It is not abstention from the exercise.

Two limits belong with the claim and neither is small.

The claim is conditional on the account being required and unobtainable elsewhere. Where a person has several potential issuers and one declines, the refusal has cost them an option rather than a position, and the analysis does not apply with the same force. Section 10.5 states what follows.

And the claim does not hold that refusal is wrong. It holds that refusal requires justification, which is a different matter and which the remainder of this section develops.

Invisibility and Unappealability of a Refusal

Three features distinguish a refusal from an unfavourable account, and all three run against the refusal.

Absence of any artefact produced by a refusal.

An unfavourable account exists. It has content, a source, and a date; it can be examined, contested, and in principle revised. A refusal produces nothing. Under Duty 15 an instrument must carry what a contest would need; a refusal carries nothing, and there is no object for a challenge to engage.

Reasons withheld and explanations substituted.

An account states something, and what it states can be assessed against what occurred. A refusal states nothing and is ordinarily accompanied by an explanation that is not the reason: pressure of time, unfamiliarity with the field, a policy. The person refused cannot distinguish a refusal grounded in what the relation contained from one grounded in preference, inconvenience, or prejudice, and neither can anyone else.

Visibility of a refusal to the person refused alone.

An unfavourable account is seen by the receiving institution, which may weigh it, discount it, or find it inconsistent with others. A refusal is seen by nobody. Where an institution requires three accounts and receives two, it observes an incomplete file and not a refusal, and the party who refused does not appear in the record at all.

Under Principle 6 the authority exercised must be met by a capacity to contest it. A refusal exercises authority and supplies nothing that could be contested, which is a more complete failure of that principle than any account, however damaging, can achieve.

Reluctance to Report Unfavourably

The empirical setting in which these decisions are made is documented and it bears on the argument in both directions.

Parties are reluctant to transmit unfavourable information, a reluctance established as a general phenomenon rather than as a feature of any particular setting (Rosen and Tesser 1970), and shown to vary with the definiteness of the consequences and with the relationship between the parties (Weenig, Groenenboom, and Wilke 2001). In performance rating, the variance attributable to the individual rater substantially exceeds that attributable to the person rated (Scullen, Mount, and Goff 2000), which is consistent with ratings reflecting the rater’s disposition toward reporting as much as the ratee’s conduct. And those who voice complaints against parties with power over them face documented retaliation, in forms that depend on the relative positions of the parties (Cortina and Magley 2003).

Two consequences follow and they pull against each other.

The reluctance is real and its causes are creditable. An issuer who declines rather than reporting unfavourably is responding to costs that exist: a continuing relation damaged, a dispute invited, an attribution made. The resulting register is uniform: of a corpus of nearly seven thousand reference ratings, ninety-six per cent rated candidates above average and fewer than one per cent below (Aamodt 2006), which is what an equilibrium of this kind produces.

But the same body of work records that silence has costs of its own, borne by those who remain silent (Cortina and Magley 2003), and the present section adds that it has costs borne by the person about whom silence is kept, which that literature does not treat.

Distribution of the Costs of Silence

The distribution is the section’s central point and it is stated plainly.

The cost of refusing falls on the issuer as the loss of nothing. They retain the relation, avoid the dispute, and incur no attribution, since no artefact exists to be attributed. Their position after refusing is their position before.

The cost of being refused falls on the person judged as the loss of a position, and it is compounded by the features of Section 10.2: they cannot establish what the refusal rested on, cannot contest it, and cannot show that it occurred.

Claim 21 (Asymmetry of the costs of silence). Refusal is cheap for the party refusing and expensive for the person refused, and the asymmetry is greater than that of an unfavourable account, which is costly to both. A duty to decline assessed without regard to this distribution is a duty designed from the position that bears no cost.

Claim 21 bears on the professional codes surveyed in Section 2.2. Those codes require an issuer to limit conclusions where access is inadequate (American Psychological Association, n.d.) and, in one formulation, prohibit opinions about persons not examined altogether (American Psychiatric Association, n.d.). Read as guidance to an individual issuer, they are sound. Read as a distribution of costs, they place the entire burden of the issuer’s epistemic limitation on the person judged, who did not create the limitation and cannot remedy it.

Conditions under Which Refusal Is Owed

The argument above does not hold that one must always write, and this subsection states when refusal is owed rather than merely permitted.

Duty 22 (Refusal). An issuer must decline where the grounds required by Duty 11 are absent and cannot be supplied, and where declining does not close a position the person cannot otherwise reach. Where the second condition fails, the issuer owes an account limited to what the relation afforded, together with a statement of how limited that is.

Three cases are distinguished by the duty.

Where the issuer’s access is null, as in the case of Section 12, Duty 16 governs and refusal is required regardless of the cost, since an exercise without grounds is not made permissible by the consequences of declining.

Where the issuer’s access is real but narrow, refusal is not owed and a limited account is. The issuer who declines on the ground that they cannot say enough has converted their own limitation into the person’s loss, when the alternative was to say what they could and say how little that was.

Where the issuer’s access is real and what it affords is unfavourable, refusal is not owed and is the more damaging course. An unfavourable account can be answered, weighed against others, and shown to concern conditions that no longer obtain. Silence cannot. Under Principle 7 the person judged retains standing as a party who can speak and answer, and only an account gives them something to answer.

Ethics under Permanent Unverifiability

This section states the constraint within which the whole account operates and which determines the shape it can take. It is short, and it is the paper’s hardest claim.

Impossibility of Establishing That the Power Was Well Exercised

Where a judgment governs whether a person is admitted, outcomes are generated only for those admitted. The performance of those excluded is never observed, because they were not admitted, and the excluded may differ systematically from the included on characteristics no one recorded.

The formal result is established. Outcomes are observed only for the cases a decision-maker passed through, so a candidate decision rule cannot be evaluated against the outcomes of cases it would have treated differently (Lakkaraju et al. 2017).

Claim 23 (Unverifiability of the exercise). A party exercising this power cannot establish that they exercised it well. The evidence that would establish it is generated only for those they admitted, and no accumulation of experience supplies the missing half. The constraint is structural and is not remedied by care, by volume, or by time.

Two consequences are worth stating before the section draws its conclusion.

Favourable outcomes among those admitted establish nothing about the judgments, since they are equally consistent with the judgments having been accurate and with the judgments having opened conditions under which the outcome became attainable.

And the absence of adverse outcomes among the excluded is not evidence, because there are no outcomes among the excluded. An institution that has never admitted a person of a certain description has no basis for its belief about how such persons perform, and its confidence in that belief grows with the length of time it has held it.

Displacement of Correctness by Conduct

An ethics for such a power cannot be an ethics of getting it right, and the reason is not that correctness is hard to achieve. It is that correctness cannot be established by the party required to achieve it, and a standard whose satisfaction is in principle unknowable to the party bound by it cannot guide them.

What remains is an ethics of conduct: not whether the exercise was correct, but whether it was conducted in a way that a party who could not know whether it was correct ought to conduct it. That is a familiar structure in settings where outcomes are unavailable or arrive too late, and it is the structure the duties of Section 8 and Section 9 already have.

Three of those duties survive the displacement without alteration, and it is worth recording which and why.

(1) Declaring the position and the conditions of access (Duty 12) is discharged or not at the moment of writing, and whether it was discharged is knowable then.

(2) Restricting the assertion to what the access affords (Duty 13) concerns the relation between what was available and what was claimed, both of which the issuer knows.

(3) Carrying what a contest would need (Duty 15) and preserving it (Duty 18) are requirements on the instrument and the file, and are met or unmet on inspection.

A duty to be right, by contrast, cannot be discharged knowingly, and a party told to satisfy it is told nothing they can act on.

Requirements That Survive the Displacement

The section closes by stating what an ethics of this shape requires, since the list is the paper’s practical residue.

Claim 24 (Conduct under unverifiability). Where a power’s good exercise cannot be established, what may be required of the party exercising it is that the basis be declared, that what is claimed not exceed the basis, that the exercise be bounded in time, and that the party affected be able to contest it. These are requirements on conduct and each is satisfiable knowingly.

Claim 24 explains why Principle 8 carries the weight it does in this account. Where correctness cannot be established, the capacity to revise is not a remedy for error but the only available substitute for knowing whether error occurred. An interpretation that stands indefinitely does not become more likely to be correct with the passage of time; it becomes less likely to be examined.

Section 13 examines an institution that reached a related conclusion, and responded by restricting what such judgments may be used for rather than by attempting to improve them.

An Account Signed without Observation

This section examines the case that motivated the paper’s central distinction. It is chosen because the two ethics deliver different verdicts on it while agreeing that it is wrong, so the case does not test whether the act is objectionable but which description of the objection is correct.

Description of the Practice

A person asks a senior figure for an account. The reply is that the person should draft it and the senior figure will sign; or that the person should send a record of their work and results, from which the account will be written.

Four features are recorded and none is contested by those who engage in the practice. The signatory has not observed the conduct the account reports, and in the first variant has not composed the sentences. The instrument functions as any other: it is accepted, weighed, and produces the same effect. The parties do not regard the practice as fraudulent, and describe it as an accommodation to the demands on senior people’s time. And the receiving institution cannot distinguish such an instrument from any other, since nothing in the document records how it was produced.

Assessment under an Ethics of Reporting

On an ethics of reporting the act is a species of deception, and the identification is straightforward.

The signatory puts forward as their own a set of claims they are not in a position to make. Under any norm of assertion the act fails: they do not know what is asserted, are not in a position to know it, and lack reasonable belief formed on the appropriate basis (Williamson 1996; Lackey 2007; Willard-Kyle 2020). The receiving institution is misled about the provenance of what it reads and would weigh it differently if it knew.

The remedy this description selects is disclosure. If the wrong is that the recipient is misled about who observed what, then the wrong is removed by stating who observed what. An account marked as compiled from records supplied by its subject, and signed by a party who did not observe the conduct, would misdescribe nothing. On this description the practice would then be unobjectionable.

An adjacent finding cautions against expecting much of disclosure. Where conflicts of interest were disclosed, advisers gave more distorted advice and recipients failed to discount sufficiently, so that disclosure worsened outcomes for those it was intended to protect (Cain, Loewenstein, and Moore 2005). The finding concerns advice rather than accounts of persons, and it indicates that a remedy resting on disclosure carries a burden of proof rather than being obviously sufficient.

Assessment under an Ethics of Authorising

On the description defended in Section 5 the act is something else, and the difference is visible precisely in the disclosed case.

What the instrument does is alter the relations available to the person it concerns. Under Duty 11 an exercise of that power must rest on something the party exercising has established, and under Duty 16 the ground of this exercise is the relation the issuer stood in. Where there is no such relation the ground is absent, and the exercise proceeds on nothing.

The wrong on this description is not that the recipient is misled. It is that a power over a person’s conditions is exercised by a party who has no basis for exercising it, and disclosure does not remedy this. An account marked as compiled from records and signed by a non-observer is a fully disclosed exercise of power without grounds. Every party knows what is happening, and what is happening is that a person’s field is being altered by someone with no warrant to alter it.

The remedy this description selects is therefore not disclosure but abstention, and Duty 16 states it as such.

Convergence and Divergence of the Two Assessments

The two descriptions agree that the practice is objectionable and disagree about what is objectionable, and the disagreement is not verbal, since the remedies differ.

Claim 25 (Divergent remedies). Under an ethics of reporting the defect is in the provenance of the account and is cured by disclosing it. Under an ethics of authorising the defect is in the absence of grounds for the exercise and is not cured by disclosure. The disclosed case therefore separates the two descriptions, and on the second it remains a wrong.

One objection must be met and it comes from the literature surveyed in Section 3.6. It has been argued that there is no general moral difference between lying and merely misleading (Saul 2012), and a reader may extend the point: if the distinction between saying something false and conveying it without saying it carries no moral weight, perhaps the distinction between misdescribing a relation and authorising without one carries none either.

The reply is that the two distinctions are not of the same kind. The lying-and-misleading distinction concerns two ways of bringing about the same result, namely a false belief in the hearer, and the argument that they are on a par is an argument about routes to one outcome. The present distinction concerns two different results: a recipient misled, and a person’s conditions altered without warrant. Claim 25 does not rest on a difference between ways of misleading; it rests on there being a wrong that persists when nobody is misled at all.

Two further consequences are recorded.

The practice is more common than an ethics of reporting predicts, and the authorising description explains why. Those who engage in it are not deceiving anyone in their own understanding: the requesting party often knows, the candidate certainly knows, and the receiving institution’s ignorance is incidental rather than the point. What is being traded is the authorisation, and the account is the form the authorisation takes.

And the practice is a limiting case rather than an anomaly. If what such instruments do is authorise, then every account lies somewhere on a range between one fully grounded in a relation and one with no relation behind it, and Duty 12 exists so that a recipient can tell where on that range a given account lies.

Restriction of Use in the Law of Evidence

An institution has already faced this paper’s problem and responded to it. This section examines that response, because it is a precedent for the shape of the remedy rather than an illustration of the difficulty.

Exclusion of Character Evidence to Prove Conduct

The rules of evidence in one system provide that evidence of a person’s character or character trait is not admissible to prove that on a particular occasion the person acted in accordance with the character or trait (“Federal Rules of Evidence, Rule 404: Character Evidence; Other Crimes, Wrongs, or Acts,” n.d.). Evidence of other acts is likewise inadmissible to show that a person acted in accordance with a propensity, though admissible for purposes that do not run through propensity, such as motive, opportunity, intent, preparation, plan, knowledge, or identity (“Federal Rules of Evidence, Rule 404: Character Evidence; Other Crimes, Wrongs, or Acts,” n.d.). Where character is admissible, the method of proof is ordinarily reputation or opinion rather than specific instances (“Federal Rules of Evidence, Rule 405: Methods of Proving Character,” n.d.), and attacks on a witness’s credibility are confined to character for truthfulness and generally exclude extrinsic evidence of particular acts (“Federal Rules of Evidence, Rule 608: A Witness’s Character for Truthfulness or Untruthfulness,” n.d.).

The structure is a restriction on use. The evidence is not said to be unavailable, nor its makers to be unreliable, nor its accuracy to be improvable. It is admissible for some purposes and inadmissible for one.

Rationale Given for the Exclusion

The reason given is recorded in the drafting commentary and it is close to this paper’s own. Character evidence is said to be of slight probative value and capable of great prejudice; to distract the trier of fact from the question of what actually happened on the particular occasion; and to permit the trier subtly to reward the good man and punish the bad because of their respective characters, despite what the evidence shows about the occasion (“Federal Rules of Evidence, Rule 404: Character Evidence; Other Crimes, Wrongs, or Acts,” n.d.).

Three elements of that rationale bear on the present account.

The complaint is about inference from a general judgment to a particular occasion, which is the inference Duty 14 requires to be marked and which the preceding paper in this series identified as an error about scope rather than about accuracy.

The complaint is about displacement: the general judgment does not merely add little, it draws attention from the question at issue.

And the complaint concerns the person rather than the act, in the observation that such evidence permits reward and punishment of persons for being what they are.

Restriction of Use as an Alternative to Improvement of Accuracy

The response is the feature this paper takes.

Faced with a form of evidence about persons that is weakly probative, prejudicial, and difficult to evaluate, the system did not require that such judgments be better made, that those who give them be trained, or that their reliability be established. It restricted what they may be used to prove.

Claim 26 (Precedent for restriction). Restricting the use of a judgment of a person is an available and established response to the difficulties this paper identifies, and it is a response that does not require the judgment to be improved. It is therefore available where, by Claim 23, improvement cannot be verified.

Claim 26 supports Duty 17 directly. A duty on the receiving party to establish what an account warrants before relying on it is a restriction on use, imposed by the party who would otherwise rely without limit, and the law of evidence shows that restrictions of this shape are workable and enforceable.

Limits of the Analogy

Three limits are recorded and the third is substantial.

The setting differs. A trial has an adjudicator, an adversary, and rules of exclusion enforced by a party with power to exclude. Selection has none of these, and Section 15 treats the question of who would enforce the duties proposed here.

The purposes differ. The rules govern proof of what a person did on an occasion; selection concerns what a person is likely to do in a setting they have not yet entered, which is a prediction rather than a finding, and the propensity inference that the rules exclude is what selection is for.

And the restriction is contested in its application. Commentary on the operation of the exception for non-propensity purposes reports that it is applied unpredictably and that propensity reasoning re-enters through it. The paper takes the rules as a precedent for the shape of a remedy and not as evidence that remedies of that shape succeed.

Evidence That Duties on Issuers Suppress Disclosure

This section states the principal argument against the paper’s proposals. It is given its own section because the evidence is real, because it concerns exactly the intervention the paper recommends, and because the reply is also the paper’s best argument for its central claim.

Liability of the Issuer in the Law of References

The duties described in Section 2.2 were imposed by courts over the objection that they would deter the giving of references. In the leading case establishing a duty of care owed by a referee to the subject, the objection was made in dissent (“Spring v Guardian Assurance Plc” 1994); and in the decision holding a recommender liable for an unreservedly favourable account given despite knowledge of serious complaints, the court acknowledged the concern that broader liability might inhibit the giving of references and subordinated it to the protection of foreseeable victims of physical harm (“Randi w. V. Muroc Joint Unified School District” 1997).

Withdrawal from Disclosure

What followed is the difficulty. Legislative findings supporting model legislation to immunise employment references record that recent court decisions had a substantial chilling effect on employers providing good-faith factual information about employees, and cite a survey from 1995 reporting that sixty-three per cent of responding employers would no longer give information about a former employee for fear of a lawsuit. The figure is reported here as it is reported there: a survey of that date, cited in the findings of a body advocating the immunity, and not a statistic this paper has independently verified.

The associated practice is described in the professional and legal literature as a retreat to confirming dates of employment and job title. Whatever the precise magnitude, the direction is what matters for this paper: duties were imposed on those who give accounts, and the reported response was withdrawal from giving them.

If that is what duties on issuers produce, the proposals in Section 8 would produce more of it, and Section 10 would then be arguing against a consequence the paper’s own recommendations create.

Contested Status of the Chilling Argument

The first reply is that the chilling argument is weaker than its currency suggests, and this is established rather than asserted.

An examination of how the chilling effect functions as a justification in speech law concludes that, although the effect may be a real concern, it is unsatisfactory as a ground for the rules it is used to support: the empirical assumptions on which it rests have little demonstrable basis, and where such assumptions are the primary or sole justification for a rule, the legitimacy of the rule is itself in question (Kendrick 2013). The same treatment surveys the kinds of evidence adduced for chilling, including litigation data, quantitative comparisons, interviews and surveys, and economic models, and examines the limits of empirical inquiry into the phenomenon.

This bears on the figure in Section 14.2 in a specific way. That figure is an interview-and-survey datum reported by an advocacy body in support of a legislative proposal, which is the weakest of the categories that treatment examines, and it is a report of what employers say they would do rather than an observation of what they did.

Reply from the Correlative Duty on the Recipient

The second reply is the paper’s own and it is the reason Section 9 exists.

What was imposed in the cases described was liability for the content of accounts, without any corresponding constraint on how accounts are used. Under Claim 19 that arrangement is self-defeating in a specific way: an issuer who states less is exposed to less liability and produces an account that a recipient, free to rely as they choose, discounts. The issuer therefore faces a choice between exposure and uselessness, and withdrawal is the rational response to that choice.

Claim 27 (Diagnosis of the withdrawal). The withdrawal from disclosure is what Claim 19 predicts of a regime that binds issuers and leaves recipients unconstrained. It is evidence for the correlativity claim rather than against the duties, and it is evidence against imposing the issuer’s duties alone.

Two features of the paper’s proposals distinguish them from what was imposed and are stated here rather than assumed.

The duties in Section 8 are requirements on content and on form: state the position, restrict the claim to what the access affords, mark inference as inference. They are not additional grounds of liability, and an account satisfying them is not thereby more exposed than one that does not.

And Duty 17 removes the incentive that produces withdrawal. Where an account whose conditions are unstated is treated as one whose scope cannot be established, the scrupulous issuer’s limited account is worth more to the person judged than the unscrupulous issuer’s unlimited one, and the choice between exposure and uselessness does not arise.

Claim 27 is offered as a diagnosis and not as a prediction of success. Whether a regime of correlative duties would in fact avoid the withdrawal has not been tested, no jurisdiction has attempted it, and Section 17 records this as the largest untested element of the account.

Institutional Bearing of the Duties Proposed

The duties stated in Section 8 and Section 9 have so far been addressed to parties. This section argues that parties are the wrong bearers, that the evidence requires the argument, and that the duties must be carried by instruments and procedures instead.

Evidence on Whether Codes Alter Conduct

A paper that ends by stating duties invites the question whether stating them changes anything, and the available evidence is unfavourable.

A meta-analytic treatment of the sources of unethical decisions at work reports that the mere existence of a code of conduct bears weakly on such decisions, and that effects appear where codes are embedded in the practices and culture of an organisation and enforced, rather than where they are promulgated (Kish-Gephart, Harrison, and Treviño 2010). An experimental study of a professional code in one field found no measurable effect on practitioners’ decisions when the code was made salient (McNamara, Smith, and Murphy-Hill 2018).

The finding concerns what makes codes effective rather than showing them pointless. It is that the causal work is done by embedding and enforcement, and that a code standing alone leaves it undone.

Two consequences follow for this paper and neither is comfortable.

The duties in Section 8 would, if promulgated as guidance to individual issuers, be predicted by this evidence to change little. The existing codes surveyed in Section 2.2 are already promulgated, already state most of Duty 11 and Duty 12, and the practice examined in Section 12 persists alongside them.

And the same evidence bears on Section 14. What was effective in altering the behaviour of issuers was not a code but liability, and what it produced was withdrawal. The one intervention shown to change conduct changed it in the direction the paper wishes to avoid.

Institutional Rather Than Individual Bearers

The response is to relocate the duties.

Claim 28 (Institutional bearing). The duties stated in this paper are discharged by institutions and not by individuals. What an individual issuer or recipient can do is constrained by the instrument they are given and the procedure they work within, and a duty requiring them to act against both remains undischarged however clearly it is stated.

Three considerations support Claim 28, and the first is the strongest.

(1) The duties concern the conditions of an exchange that neither party sets. Whether an account states the conditions of access depends on whether it is asked for them. Whether divergence among accounts is preserved depends on whether the file has a place to record it. Whether the person judged knows an account was given depends on whether the procedure tells them. In each case the party bound by the duty is not the party who determines whether discharging it is possible.

(2) The correlativity argument requires simultaneous imposition. Under Claim 19 the issuer’s duties and the recipient’s must be imposed together or the first penalises compliance. Simultaneous imposition on two parties who do not deal with each other directly, and who often belong to different institutions, is something only an arrangement governing the exchange can achieve.

(3) An institution can bear a duty of inquiry that an individual cannot discharge. The legal analogue in Section 9.1 is instructive: the duty of reasonable inquiry in hiring is imposed on the employer, is proportioned to the risk the position carries, and is discharged through procedures rather than through the diligence of whoever happens to read the file (“Ponticas v. K.m.s. Investments” 1983; “Connes v. Molalla Transport System, Inc.” 1992), resting on the principle that a principal is subject to liability for harm resulting from a failure to exercise care in the employment of others (“Restatement (Second) of Agency, Section 213” 1958).

The wider warrant for constraining decision-makers in this way has been defended directly. Institutions may be justified in constraining what an inquirer may consider, for the sake of the inquiry itself and without the inquirer’s consent, and rules of evidence are advanced as an instance of such a constraint that is justified (Ahlstrom-Vij 2013). That is the form the present proposal takes: the receiving institution constrains what its own decision-makers may rely on and how, which is a constraint the decision-maker could not impose on themselves.

Design of Instruments That Carry the Duties

If the duties are borne by instruments, then the paper’s practical content is a specification of instruments, and this subsection states it. Nothing here requires an issuer or a recipient to be more scrupulous than they are.

The request carries the scope question.

Duty 12 and Duty 13 are discharged by an instrument that asks for the position, the period, the conditions under which the conduct occurred, and the respects in which the issuer’s view was partial. An issuer answering the questions asked discharges the duties without adopting any disposition; an issuer left unasked will rarely volunteer them.

The form separates conduct from characterisation.

Duty 14 is discharged by an instrument whose questions distinguish what occurred from what it is taken to show. This is the structural device the assessment literature already uses, and it converts a duty on the issuer’s judgment into a property of the form.

The record carries what a contest would need.

Duty 15 and Duty 18 are discharged by retaining the account as given, the questions it answered, the accounts of other parties, and the respects in which they diverged. The requirement is on the file rather than on either party’s conduct, and Section 9.2 recorded that it is cheap.

The procedure fixes what an unscoped account is worth.

Duty 17 is discharged by a rule, adopted by the receiving institution, that an account whose conditions of access are unstated is treated as one whose scope cannot be established. This is the single most consequential item in the specification, because it is what removes the incentive identified in Section 14.4: where the rule holds, the issuer who declares little produces an account worth little, and the issuer who declares much is not penalised for candour.

The procedure informs the person judged that an account was given.

Nothing in Section 8 or Section 9 can be enforced by a party kept ignorant that the exercise occurred. This is the minimum condition of the duty running to the person judged that Section 9.4 found nowhere imposed, and it is stated here as a requirement on procedure because no other party is positioned to supply it.

Elements absent from the specification.

It supplies no forum. A person informed that an account was given, holding the conditions under which it was given, and knowing that the record persists, has what a contest would require and nowhere to bring it. Section 17 records this as the largest gap between what the paper argues and what it delivers, and the question of who should hold authority to determine such a contest belongs to the jurisprudence paper in this series.

Implications for the Generative Relational Framework

Four results return to the framework and two of them are corrections.

An Ethics of Acts That Alter Another’s Field

The framework’s ethical principles govern acts that reproduce shared experience, and have been applied to relations in which the parties continue to deal with one another. This paper applies them to an act performed about a relation, by one of its parties, to a third party who was not in it.

Two features of that application should be carried forward. The principles hold in it: each yielded the results of Section 7 unmodified. And the act it governs is a class the framework has not identified, namely acts by which a party to one relation alters the conditions obtaining in another. Recommendation, endorsement, certification and clinical handover are members of the class.

Simultaneity as a Requirement on Counter-Power

Principle 6 was stated in this paper with an emphasis the framework had left implicit: that the capacity to contest must be generated at the moment the authority is conferred, and not supplied afterwards.

The ground for the emphasis is that the authority is conferred by the act itself. Where an institution supplies a channel of appeal later, the appeal engages an authority already exercised and effects already propagated, and the principle’s purpose of preventing history from freezing is served late or not at all. Section 8.5 showed that the requirement is modest in what it asks of an issuer and demanding in what it implies about instruments, which is a general shape the framework should expect where simultaneity is required.

The Conditions of Revisability as Part of the Requirement

Principle 8 was likewise stated with an addition: that the conditions under which revision is possible must themselves be preserved, since a right to revise is empty where what would be needed to exercise it has been discarded.

This converts the principle from a permission into a duty with a bearer. A right to reopen lies dormant until exercised; a requirement that the record, the divergent accounts, and the conditions declared with them be preserved imposes something on whoever holds the file, continuously, before anyone has asked. Section 9.2 states it as Duty 18, and the framework should carry the general form: wherever it requires revisability, it requires the maintenance of what revision would consume.

Limits of the Framework Where Correctness Cannot Be Established

The fourth result is a constraint the framework should adopt.

Claim 23 holds that a party exercising this power cannot establish that they exercised it well, because the evidence is generated only for those they admitted. Where that holds, the framework’s normative apparatus cannot be applied through outcomes. Judgments about whether generativity was preserved, whether a subject was eliminated, or whether foreclosure occurred must be made on the conduct of the exercise and on the conditions it left in place, and not on what became of the people affected. That restriction follows from the evidence rather than from caution.

The framework has tended to state its principles in terms of what an act achieves. This case indicates that where achievement is unobservable, the principles must be stated in terms of what an act does and leaves, and Claim 24 states the four requirements that survive the restatement.

Limits of the Account

Dependence of the Results on the Two Routes

Section 4.1 argued the duties twice from independent premises, and this subsection records which results depend on which.

Reached by both routes, and therefore surviving the rejection of either: Duty 12, Duty 13, Duty 15, Duty 17, Duty 18, Claim 2, and Claim 19.

Reached only from the framework’s principles: Claim 9, that the practice is in breach at the point of production regardless of content, and the finding in Section 7.2 that confidentiality eliminates the person judged as a subject. A reader who accepts the account of the act as an exercise of power but rejects the principles obtains the duties without these two.

Reached only from the account of the act: Claim 1 and the treatment of the signed blank in Section 12.3. A reader who accepts the principles but rejects the authorisation finding retains Claim 9 and loses the divergence result.

Conditions of Falsification

Section 4.3 stated five conditions and their status is as follows.

The scope norm’s distinction from an access-indexed norm already in the literature (Willard-Kyle 2020) rests on two additions stated in Section 8.3 and argued rather than demonstrated. If either collapses, Duty 13 is an application rather than a contribution. This is the paper’s principal exposure in the philosophy of assertion.

Whether a duty on the issuer alone can produce the intended restraint is untested. Claim 19 is a structural argument supported by one body of reported evidence (Kendrick 2013), and that evidence concerns liability rather than the content duties this paper proposes.

Whether refusal leaves a person’s field unchanged is untested, and is treated here as governed by the conditional in Claim 20. The conditional’s antecedent, that the account is required and unobtainable elsewhere, is an empirical matter that varies by field and has not been established for any.

Whether the good exercise of this power can be established after the fact is answered negatively by an established formal result (Lakkaraju et al. 2017), and this is the condition the paper meets most securely.

And whether the duties proposed would worsen what they govern is the condition the paper is most exposed on. Section 14 treats the evidence that an earlier intervention produced withdrawal from disclosure, and Section 15 treats the evidence that codes without enforcement change little. The reply in each case is argued and untested.

Claims Advanced Without Support

Three claims rest on argument alone.

Claim 21, on the asymmetric distribution of the costs of silence, is developed from the structure of the situation. The literature relied on documents the costs borne by those who remain silent (Cortina and Magley 2003) and leaves untreated the costs borne by the person about whom silence is kept.

Claim 25, that the two ethics select different remedies, rests on the disclosed case and on the reply to the objection from the parity of lying and misleading (Saul 2012). It has no empirical support and belongs to a kind of claim that admits of none.

And the specification in Section 15.3 has not been implemented anywhere, so its feasibility, its cost, and its effects are unknown.

Absence of a Forum

The largest gap between what the paper argues and what it delivers is stated here rather than left to be inferred.

The paper argues that the primary obligation runs to the person judged (Claim 2), that they must be able to contest what was established (Principle 6), and that the conditions of contest must be preserved (Duty 18). It supplies no forum in which a contest could be brought, no account of who should determine it, and no remedy that would follow from determining it.

Section 9.4 recorded that no duty in any of the codes or legal doctrines examined runs from the party who uses an account to the person it concerns. A duty stated without a bearer able to be held to it, and without a forum in which the holder of the correlative right can raise it, remains a duty in name. The paper states the duty because the analysis yields it, and records that the institution the duty would require remains undelivered. Who holds authority to determine such a contest is the subject of the jurisprudence paper in this series.

Sources Not Yet Verified

This draft cites only sources verified against a publisher, journal, index, or institutional page before the section using them was written, and several literatures are consequently represented thinly.

The evidence on withdrawal from disclosure rests on a figure reported in the legislative findings of a body advocating an immunity, and the paper attributes it as such in Section 14.2 rather than treating it as established. A primary source has not been located.

The philosophy of silence and concealment is absent from Section 10, which develops its argument from the structure of the situation and from the empirical literature on reluctance rather than from the ethical literature on withholding.

Professional codes beyond those of one discipline are absent from Section 15, which weakens the claim that the pattern identified there is general.

And the treatment of testimonial injustice is deliberately absent, since it belongs to the paper on distribution in this series, with the consequence that Section 10.4 states an asymmetry of costs without connecting it to the distribution of credibility that would explain part of it.

Extensions

Four extensions are identified and none is attempted.

A trial of the specification in Section 15.3 in a single institution would establish whether requesting the conditions of access changes what issuers write, and is the cheapest test of the paper’s central proposal.

An empirical treatment of Claim 20 would establish how often refusals close positions rather than costing options, which the conditional in that claim leaves open.

A design for a forum, addressing the gap in Section 17.4, is what the paper’s own argument requires and does not supply.

And an account of the duties owed where the judging party is not a person is absent. Where an instrument scores or ranks persons without a party who stood in any relation to them, Duty 12 has no addressee, and whether the remaining duties survive that absence is unexamined.

Conclusion

Someone who has worked with a person is asked what they are like, by a party that will decide something about them. This paper has argued that the ethics usually brought to that request is addressed to the wrong act.

Norms of assertion govern saying. The act at issue does something else: it opens or closes the relations available to the person it concerns, which is why an account signed by someone who observed nothing still works. What follows is that the duties governing it are those of an actor exercising discretionary power over another’s conditions, and that the obligation runs primarily to the person judged, who is absent from the transaction and does not consent to it. The party who asked is owed a justification, and is owed it because they will exercise the power jointly.

The framework’s own principles yield the same duties by a different route and one result besides. An account of a person is an act reproducing shared experience, and the principles governing such acts hold that reproduction preserve the possibility of continued becoming, that a shared past not be interpreted unilaterally by one of its authors, that value return to those who generated it and that they hold a claim over the conditions of its production, that the authority conferred by the act be met by a capacity to contest it generated at the same moment, that the other survive as a subject, and that no interpretation become historically irreversible, which requires that the conditions of revision be kept and not merely permitted.

Applied to the practice, these hold that an ordinary confidential reference is in breach at the point of production and not in virtue of anything it says; that confidentiality eliminates the person judged as a subject in the very relation whose product is reported; that they have no say in the conditions under which the account of their own shared past is produced; and that foreclosing a person’s field is a graver wrong than describing them inaccurately, which inverts the ordering an ethics of assertion imposes without argument.

Two of the paper’s conclusions run against received practice. Refusing to write is treated as the safe course and is the act requiring the stronger justification: it forecloses as effectively as an unfavourable account, leaves no artefact for a challenge to engage, gives reasons that are not the reasons, is invisible to everyone but the person refused, and is cheap for the party refusing while expensive for the person refused. And the duties bind the party who uses an account as well as the party who gives one. A restraint on the issuer alone produces an account weaker on its face than one claiming everything, so where the recipient may rely as they choose the duty penalises compliance. Where duties were imposed on issuers and none on recipients, the reported response was withdrawal from disclosure, and this paper reads that as evidence for the correlativity claim rather than against the duties.

The shape an ethics of this act can take is constrained by something that cannot be remedied. Those excluded by such judgments are never observed, so the party exercising the power cannot establish that they exercised it well, and no accumulation of experience supplies the missing half. An ethics of correctness is therefore unavailable to them. What remains is an ethics of conduct: declare the basis, claim no more than the basis supports, bound the exercise in time, and leave the person affected able to contest it. Each is satisfiable knowingly, which a duty to be right can never be. The law of evidence reached a related conclusion long ago and responded by restricting what judgments of persons may be used to prove rather than by attempting to improve them.

What the paper does not supply is recorded. It states duties that would have to be borne by institutions rather than by individuals, since the evidence indicates that codes addressed to individuals change little. It specifies instruments that would carry them and has tested none. And it argues that the person judged is owed a duty by the party who relies on an account of them, while supplying no forum in which that duty could be raised. A person who knows an account was given, holds the conditions under which it was given, and knows the record persists has what a contest would require and nowhere to bring it.

Acknowledgments

The present definitions, constructions, arguments, conclusions, and errors remain the author’s responsibility. The principle stated in Section 6.3 carries an inheritance identified there. The interest arising from the author’s own position with respect to the procedures examined is declared in the front matter.

Aamodt, Michael G. 2006. “Validity of Recommendations and References.” Assessment Council News, February, 4–6.

Ahlstrom-Vij, Kristoffer. 2013. Epistemic Paternalism: A Defence. Basingstoke: Palgrave Macmillan. https://doi.org/10.1057/9781137313171.

American Educational Research Association, American Psychological Association, and National Council on Measurement in Education. 2014. Standards for Educational and Psychological Testing. Washington, DC: American Educational Research Association.

American Psychiatric Association. n.d. “The Principles of Medical Ethics with Annotations Especially Applicable to Psychiatry, Section 7.3.” American Psychiatric Association.

American Psychological Association. n.d. “Ethical Principles of Psychologists and Code of Conduct, Standard 9.01: Bases for Assessments.” American Psychological Association.

“Bartholomew v London Borough of Hackney.” 1998. [1998] EWCA Civ 1604, [1999] IRLR 246 (Court of Appeal).

Battaly, Heather. 2010. “Epistemic Self-Indulgence.” Metaphilosophy 41 (1–2): 214–34.

Cain, Daylian M., George Loewenstein, and Don A. Moore. 2005. “The Dirt on Coming Clean: Perverse Effects of Disclosing Conflicts of Interest.” Journal of Legal Studies 34 (1): 1–25. https://doi.org/10.1086/426699.

Coady, C. A. J. 1992. Testimony: A Philosophical Study. Oxford: Clarendon Press. https://doi.org/10.1093/0198235518.001.0001.

“Connes v. Molalla Transport System, Inc.” 1992. 831 P.2d 1316 (Supreme Court of Colorado, en banc, 29 June 1992).

Cortina, Lilia M., and Vicki J. Magley. 2003. “Raising Voice, Risking Retaliation: Events Following Interpersonal Mistreatment in the Workplace.” Journal of Occupational Health Psychology 8 (4): 247–65. https://doi.org/10.1037/1076-8998.8.4.247.

“Cox v Sun Alliance Life Ltd.” 2001. [2001] EWCA Civ 649, [2001] IRLR 448 (Court of Appeal).

“Di Cosala v. Kay.” 1982. 91 N.J. 159, 450 A.2d 508 (Supreme Court of New Jersey).

Douven, Igor. 2006. “Assertion, Knowledge, and Rational Credibility.” The Philosophical Review 115 (4): 449–85.

Eglash, Ron. 2016. “An Introduction to Generative Justice.” Revista Teknokultura 13 (2): 369–404. https://doi.org/10.5209/rev_TEKN.2016.v13.n2.52847.

Espeland, Wendy Nelson, and Michael Sauder. 2007. “Rankings and Reactivity: How Public Measures Recreate Social Worlds.” American Journal of Sociology 113 (1): 1–40. https://doi.org/10.1086/517897.

“Federal Rules of Evidence, Rule 404: Character Evidence; Other Crimes, Wrongs, or Acts.” n.d. United States.

“Federal Rules of Evidence, Rule 405: Methods of Proving Character.” n.d. United States.

“Federal Rules of Evidence, Rule 608: A Witness’s Character for Truthfulness or Untruthfulness.” n.d. United States.

Fricker, Elizabeth. 1994. “Against Gullibility.” In Knowing from Words, edited by Bimal Krishna Matilal and Arindam Chakrabarti, 125–61. Dordrecht: Kluwer.

Goldberg, Sanford C. 2010. Relying on Others: An Essay in Epistemology. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199593248.001.0001.

Goldberg, Sanford, and David Henderson. 2006. “Monitoring and Anti-Reductionism in the Epistemology of Testimony.” Philosophy and Phenomenological Research 72 (3): 600–617. https://doi.org/10.1111/j.1933-1592.2006.tb00586.x.

Hinchman, Edward S. 2005. “Telling as Inviting to Trust.” Philosophy and Phenomenological Research 70 (3): 562–87.

Hohfeld, Wesley Newcomb. 1913. “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning.” The Yale Law Journal 23 (1): 16–59.

———. 1917. “Fundamental Legal Conceptions as Applied in Judicial Reasoning.” The Yale Law Journal 26 (8): 710–70. https://doi.org/10.2307/786270.

Kendrick, Leslie. 2013. “Speech, Intent, and the Chilling Effect.” William and Mary Law Review 54 (5): 1633–91.

Kish-Gephart, Jennifer J., David A. Harrison, and Linda Klebe Treviño. 2010. “Bad Apples, Bad Cases, and Bad Barrels: Meta-Analytic Evidence about Sources of Unethical Decisions at Work.” Journal of Applied Psychology 95 (1): 1–31.

Lackey, Jennifer. 2007. “Norms of Assertion.” Noûs 41 (4): 594–626. https://doi.org/10.1111/j.1468-0068.2007.00664.x.

———. 2008. Learning from Words: Testimony as a Source of Knowledge. Oxford University Press.

Lakkaraju, Himabindu, Jon Kleinberg, Jure Leskovec, Jens Ludwig, and Sendhil Mullainathan. 2017. “The Selective Labels Problem: Evaluating Algorithmic Predictions in the Presence of Unobservables.” In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 275–84. https://doi.org/10.1145/3097983.3098066.

McNamara, Andrew, Justin Smith, and Emerson Murphy-Hill. 2018. “Does ACM’s Code of Ethics Change Ethical Decision Making in Software Development?” In Proceedings of the 2018 ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering.

Moran, Richard. 2005. “Getting Told and Being Believed.” Philosophers’ Imprint 5 (5): 1–29.

———. 2018. The Exchange of Words: Speech, Testimony, and Intersubjectivity. Oxford University Press.

“Ponticas v. K.m.s. Investments.” 1983. 331 N.W.2d 907 (Supreme Court of Minnesota, 25 March 1983).

“Randi w. V. Muroc Joint Unified School District.” 1997. 14 Cal.4th 1066, 929 P.2d 582, 60 Cal.Rptr.2d 263 (Supreme Court of California).

“Restatement (Second) of Agency, Section 213.” 1958. American Law Institute.

“Restatement (Second) of Torts, Section 552: Information Negligently Supplied for the Guidance of Others.” 1977. American Law Institute.

Roberts, Robert C., and W. Jay Wood. 2007. Intellectual Virtues: An Essay in Regulative Epistemology. Oxford: Clarendon Press.

Rosen, Sidney, and Abraham Tesser. 1970. “On Reluctance to Communicate Undesirable Information: The MUM Effect.” Sociometry 33 (3): 253–63. https://doi.org/10.2307/2786156.

Saul, Jennifer Mather. 2012. Lying, Misleading, and What Is Said: An Exploration in Philosophy of Language and in Ethics. Oxford University Press.

Schmitt, Frederick F. 2006. “The Assurance View of Testimony.” In The Epistemology of Testimony, edited by Jennifer Lackey and Ernest Sosa. Oxford University Press.

Scullen, Steven E., Michael K. Mount, and Maynard Goff. 2000. “Understanding the Latent Structure of Job Performance Ratings.” Journal of Applied Psychology 85 (6): 956–70.

“Spring v Guardian Assurance Plc.” 1994. [1994] UKHL 7, [1995] 2 AC 296 (House of Lords).

Stokke, Andreas. 2018. Lying and Insincerity. Oxford University Press.

Turri, John. 2011. “The Express Knowledge Account of Assertion.” Australasian Journal of Philosophy 89 (1): 37–45.

Weenig, Mieneke W. H., Anniek C. W. J. Groenenboom, and Henk A. M. Wilke. 2001. “Bad News Transmission as a Function of the Definitiveness of Consequences and the Relationship Between Communicator and Recipient.” Journal of Personality and Social Psychology 80 (3): 449–61.

Weiner, Matthew. 2005. “Must We Know What We Say?” The Philosophical Review 114 (2): 227–51.

Whitcomb, Dennis, Heather Battaly, Jason Baehr, and Daniel Howard-Snyder. 2017. “Intellectual Humility: Owning Our Limitations.” Philosophy and Phenomenological Research 94 (3): 509–39. https://doi.org/10.1111/phpr.12228.

Willard-Kyle, Christopher. 2020. “Being in a Position to Know Is the Norm of Assertion.” Pacific Philosophical Quarterly 101 (2): 328–52. https://doi.org/10.1111/papq.12305.

Williamson, Timothy. 1996. “Knowing and Asserting.” The Philosophical Review 105 (4): 489–523. https://doi.org/10.2307/2998423.

———. 2000. Knowledge and Its Limits. Oxford University Press.

Zagzebski, Linda Trinkaus. 1996. Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge. Cambridge University Press.