RESEARCH · DEVELOPMENTAL READING EDITION
Chapter VII — Hard Cases Across Domains
A rigorous comparison across caregiving, relationships, institutions, medicine, and AI services.
← Complete contents · Chapter 7 of 8 · Updated November 2026
In this chapter
- 7.1. The case must be capable of defeating the theory
- 7.2. Caregiving: the same need, a different history
- 7.3. The beneficiary as a moral agent, not an object of transfer
- 7.4. An intimate separation with two kinds of debt
- 7.5. The physician's retirement and the unaccepted referral
- 7.6. A public service that performs while responsibility disappears from view
- 7.7. The volunteer whose institution has made her indispensable
- 7.8. A case of coercive exit: duties without personal access
- 7.9. A private obligation owed to a third party
- 7.10. Retirement of an AI service: functional continuity without accountable continuity
- 7.11. The administrative closure that changes no duty
- 7.12. Two vacancies that must not be conflated
- 7.13. Competing predictions: when a missing successor is not a new moral gap
- 7.14. The narrowest defensible positive conclusion
- 7.15. What an empirical inquiry could and could not establish
- 7.16. The dissertation's decisive scholarly problem
- 7.17. What the final chapter must and must not do
- 7.18. Conclusion: the cases did not prove what they were designed to prove
7.1. The case must be capable of defeating the theory
A coherent account of how responsibilities continue, change, or end is not necessarily a new philosophical theory. The preceding chapter developed an approach called accountable succession, but its principal conclusions can usually be explained by combining existing work on rights, promises, care, relational freedom, and remedial responsibility. The apparent novelty becomes smaller when those accounts receive their strongest interpretation. What remains is a more demanding question: can we identify a case where the proposed approach yields an important judgment that those theories cannot?
A useful thought experiment must be capable of defeating the theory it was designed to test. If one person is obviously admirable and the other obviously unreasonable, the case may illustrate a principle, but it cannot establish its explanatory necessity. Here, at least two responses must be independently plausible. We then ask what makes one preferable and whether the proposed explanation reaches a conclusion that the strongest rival cannot: who owes the duty, what it requires, who may waive it, when it ends, or how an institution may truthfully describe the handoff. A new name for an old distinction is not enough.
The cases range across caregiving, intimacy, professional practice, public administration, organizational work, and AI services. The breadth has a purpose: to expose cases that look similar but arise from different sources of duty. Rules appropriate to a contract may not govern friendship. Professional duties should not be projected onto informal care. A public institution may remain responsible when personnel change even though no individual promised the recipient anything. If a theory reaches the same result only by ignoring these differences, it is not a successful synthesis.
The methodological standard is comparative charity. For each scenario we identify the normative facts held constant and the single factor varied. We then state what the principal approaches plausibly explain, without attributing invented verdicts to named authors. Where the strongest combination yields the same judgment as accountable succession, that result counts against a general originality claim. If a case remains underdetermined, we ask whether the indeterminacy reflects a genuine omission in the theory or ordinary disagreement concerning a contested normative premise. A disagreement over how much responsibility is fair cannot automatically be repaired by adding an apparently precise procedural rule.
The central concern is the temptation to treat practical continuity as evidence of normative continuity. Yet the opposite error is equally dangerous: if the process of transferring responsibility is incomplete, we may too quickly infer that the original agent must continue indefinitely. The chapter will ask whether this double risk is already satisfactorily handled by established accounts. We begin with the relationship in which the moral stakes of dependence are especially high and the grounds of duty cannot be reduced to voluntary promises.
7.2. Caregiving: the same need, a different history
In the first case, Iris has cared for her uncle Daniel for several years after a serious injury. She initially promised six months of assistance while permanent support was arranged. The family and the local service system postponed finding a replacement, partly because Iris was exceptionally competent and partly because the arrangement was cheaper. Daniel now requires daily support of roughly the same intensity as when Iris first undertook the work. Iris is facing serious health risks from continued physical caregiving and decides to stop. Her uncle's need remains genuine, and the available alternatives are imperfect. The question is not whether he deserves care, but whether his need gives him a directed claim against Iris to preserve the exact arrangement.
Case A includes a clear six-month undertaking, fulfilled on its terms, followed by several years of assistance during which Iris repeatedly states that continuation is temporary and dependent upon her capacity. Case B is identical except that Iris expressly renews an unconditional commitment for an additional year and knowingly induces Daniel to cancel a feasible placement. In both cases Daniel requires assistance, Iris's health has deteriorated, and the family faces hardship. The relevant difference is the newly undertaken obligation and induced reliance. Case B supplies stronger grounds for a duty concerning the manner of withdrawal or the losses caused by disappointed assurance, although even that duty cannot automatically require medically unsafe direct work. Case A supplies no equivalent extension simply because Iris has provided care for a long time.
A purely promissory view explains the greater force of Case B but risks neglecting nonvoluntary responsibilities in either case. Goodin's vulnerability analysis, Chadha-Sridhar's hybrid non-voluntarism, and Ong and Savulescu's network account supply reasons to examine Daniel's position independently of what Iris promised. The last two explicitly recognize that care obligations may arise outside voluntary agreement and that the social distribution of caring responsibilities must not impose unreasonable burdens on individuals (Goodin 1985; Chadha-Sridhar 2026; Ong and Savulescu 2026). A theory-consistent extrapolation from their approaches would recognize the seriousness of Daniel's unmet need while examining legitimate relational grounds, available alternatives, cost, and public responsibility. We cannot legitimately attribute a determinate duty of one year's service to any of these authors without arguments they have not made about this precise example.
Accountable succession reaches a similar provisional conclusion. It asks whether the original undertaking remains operative, whether a separate duty of reasonable notice or information transfer exists, whether someone has validly accepted a successor role, and whether the family or public institutions hold independent duties. It refuses to conclude that Iris must continue simply because a replacement has not been secured. But this refusal is already well explained by the strongest combined predecessor account. The case supports disciplined attribution, not a new foundational doctrine.
Change one fact: who legitimately bears responsibility for arranging the care. In Case C, a funded public agency is supposed to organize long-term services but treats Iris's unpaid work as a permanent family resource. In Case D, the agency has offered a competent and available alternative, which Daniel rejects solely because he prefers Iris. The need is equally serious, but the reason assistance remains unsettled is different. Case C may reveal institutional injustice; Case D may place stronger limits on Daniel's claim against Iris. Care ethics already explains why a social system cannot convert unjustly assigned dependency into unlimited private duty. These cases locate an important problem without, so far, demonstrating a gap in that theory.
The most important failed inference is therefore not just that need persists while duty does not. It is that the legitimacy of an individual's historical performance does not settle the legitimacy of an institution's decision to rely upon its continuation. But the author's earlier When the Gift Becomes Capacity and Reflexive Capture already analyze the transformation of informal contribution into institutional dependence. This internal predecessor further weakens any claim that the present case is the nucleus of a new dissertation doctrine.
7.3. The beneficiary as a moral agent, not an object of transfer
Caregiving disputes are often described as if they concern two actors: the person who leaves and the person who will replace them. The beneficiary then appears as the destination of a service rather than a person whose choices, privacy, and rights may constrain the transition. This is a morally serious distortion. A competent adult receiving care can have protected interests in the identity, conduct, and reliability of the caregiver, and the legitimacy of an arrangement may depend on participation or consent under the relevant moral and legal scheme.
Suppose Daniel accepts the need for a new caregiver but declines a particular proposed professional because that person previously exposed his confidential information. Iris has provided appropriate notice and cannot safely continue. The institution offers another professional in four weeks, while immediate alternatives are limited. A theory that requires Iris to continue until Daniel approves every conceivable replacement is implausibly coercive. A theory that treats Daniel's refusal as irrelevant because any competent person can do the work is likewise deficient. His complaint concerns an independent protected interest in confidentiality and trustworthy care, not simply an aesthetic preference for familiarity.
The right kind of analysis separates three questions. Daniel's refusal may invalidate the institution's assertion that this particular substitute is acceptable. It may generate or preserve an institutional duty to seek a different arrangement. It does not, without separate grounds, revive a terminated duty of medically unsafe personal performance by Iris. An accurate account therefore protects the beneficiary's participation without granting a veto over another person's legitimate withdrawal. This result looks like a distinctive success for accountable succession only if the rival theory is presented as indifferent either to autonomy or to the caregiver's limits. Neither presentation would be charitable to rights theory and contemporary care ethics.
The presence of supported decision-making or an authorized representative adds further complexity. If the beneficiary needs assistance evaluating a proposed care arrangement, the representative may have a role in facilitating the person's own choices under the relevant jurisdiction's rules. Yet the existence of a family member who favors the transfer does not necessarily establish authority to extinguish the beneficiary's independent claims, just as an adult beneficiary's expressed preference cannot authorize unsafe coercion of a departing caregiver. What matters is which powers, rights, and duties exist, rather than which participant speaks most confidently about what the care recipient needs.
We should avoid the suggestion that the same content-neutral consent rule settles intimate care and publicly administered services. Specific authority may derive from guardianship law, health-care consent law, fiduciary relationships, or the terms of a valid private undertaking. These differ in their justifications and safeguards. Our framework is helpful insofar as it prevents one form of consent from being substituted for another; it is not a substitute for the substantive principles that explain what consent can authorize. This case remains available to the Hohfeldian and care-ethical rival.
7.4. An intimate separation with two kinds of debt
Leona and Malcolm share a decade-long partnership. During that time Leona leaves a promising position so that Malcolm can complete a demanding professional training program. They jointly agree that Malcolm will make specified financial contributions if their relationship ends. Several years later Malcolm concludes that he cannot continue the romantic partnership. There has been no betrayal, and his decision reflects a sincere change in his willingness to participate. Leona is devastated. She asks for fulfillment of the financial arrangement and for Malcolm to continue appearing as her partner at family events until the costs of separation have diminished.
We should distinguish what each demand concerns. The financial claim has a ground in the couple's explicit understanding and the reliance it invited. The demand for continuing romantic participation concerns an activity whose value normally depends on freely sustained attachment. A legitimate financial debt cannot automatically be converted into authority over intimacy. Equally, Malcolm's permission to withdraw from the relationship is not a power to extinguish the financial obligation. The two claims have different objects, grounds, and modes of permissible satisfaction.
The comparison becomes harder if Leona asks not for continued romance but for one careful conversation explaining the decision. Betzler argues that ending a special romantic relationship may leave residual duties concerning explanation, participation, and transitional goods, whereas Lopez-Cantero questions the scope and action-guiding character of deliberative breakup duties, especially when the only explanation is that love has ended (Betzler 2024; Lopez-Cantero 2026). The theories present an actual scholarly dispute rather than a straw opponent. Our account should not pretend to resolve it by placing both duties under the label of accountable succession.
Case A involves a request for the practical facts necessary to divide shared property and protect the parties' financial rights. Case B involves an indefinite request that Malcolm explain his changing emotions until Leona finds the explanation satisfying. The first has a stronger basis in specifically grounded obligations; the second can turn explanation into an effective condition upon the right to leave. But this contrast is essentially what Lopez-Cantero's critique already invites us to examine. It does not yet distinguish the dissertation from its predecessor.
Now suppose Leona has reason to fear actual deception concerning the financial arrangements. A refusal to disclose relevant information might then be a breach of duties independent of Malcolm's obligation to continue the relationship. We need not either require an exhaustive emotional narrative or deny any duty to communicate. The important distinction concerns the information legitimately owed and the authority to require its production. An ethics of exit that cannot identify the difference between practical disclosure and compelled intimacy would fail. Yet the combined rights, promise, and relational-freedom accounts can identify it.
The strongest conclusion here is negative: no one can convert a valid claim to one object into unrestricted authority over a different object solely by pointing to the severity of the underlying loss. This is a useful constraint, but it overlaps the author's earlier Reflexive Capture and Do Not Flatten Me as well as contemporary theories of normative powers and relational freedom. The case has not established a new dissertation-level principle.
7.5. The physician's retirement and the unaccepted referral
The clinical handoff case adds a stronger institutional structure. A physician lawfully retires, tells patients in advance, provides accessible records, and supplies information about alternative physicians. One patient has a medication requirement that cannot safely await a long appointment delay. The outgoing practice sends a referral, but the receiving practice has not accepted the patient. An administrator nevertheless marks the case as “transferred” because the referral was successfully transmitted. The beneficiary remains exposed to a gap, even though each administrative step recorded by the original organization has been completed.
The crucial variation separates a referral from an undertaking. In Case A the receiving physician confirms acceptance and assumes responsibility for care on a specified date, and the original practice meets the requirements for its own withdrawal. In Case B a referral is sent but the receiving physician neither accepts the patient nor undertakes continuing care. Assume that the patient's condition during the first several days is identical in both cases. Case A may involve an actual successor duty while Case B may involve only an attempt to locate a successor. The factual distinction matters independently of whether the eventual clinical outcomes prove equally good.
The governing professional rules already anticipate this difference. The American Medical Association's Opinion 1.1.5 requires physicians withdrawing from a patient relationship to provide sufficiently advance notice and facilitate transfer when appropriate. The American College of Obstetricians and Gynecologists' 2022 guidance emphasizes continuity, patient safety, appropriate notice, transitional prescriptions, and referrals, but does not treat the physician as absolutely obliged to ensure that every dismissed patient successfully establishes care elsewhere (AMA, Opinion 1.1.5; ACOG 2022). These rules are not identical to a general moral principle governing friendship or informal care. Their role in the case is to specify duties arising from a professional relationship and to clarify the limits on personal continuation.
The grounds-sensitive account makes an important analytical correction to the administrator's statement. Proof that a referral was sent establishes a referral, not a professional acceptance; it cannot by itself warrant an unqualified claim that the receiving physician has assumed responsibility. But the stronger rival can reach precisely the same judgment. Hohfeldian analysis distinguishes the normative position of the successor from the act of sending records. Shiffrin's work requires attention to the duties concealed by apparently simple transfers. Professional ethics distinguishes appropriate facilitation from the success of every downstream event. General norms concerning truthful statements supply independent reasons not to misstate the handoff. The case does not reveal an original philosophical requirement.
An objection may insist that the outgoing physician remains accountable until the new provider accepts the patient. That inference cannot be made from the absence of acceptance alone. Under some circumstances professional obligations may require continued interim care, while in others the retiring physician may have met the applicable duties even though the patient faces a shortage of services. The right conclusion depends on the terms of the role, foreseeable risks, applicable regulatory standards, and available alternatives. The failure of a new provider to accept cannot automatically create a duty of permanent practice for the retiring one.
Two absences must be distinguished: epistemic vacancy, when we do not know whether someone has assumed a duty, and normative vacancy, when nobody actually bears the relevant duty. An institution may be uncertain about a transfer that has validly occurred. It may also know that no successor accepted while remaining unsure who independently owes help. Ignorance of a duty is not proof of its absence; evidence that nobody new accepted is not proof that the former provider remains bound. The distinction is useful for investigating institutions, but it also follows from established work on knowledge, rights, and duties of inquiry. Its novelty cannot simply be asserted.
7.6. A public service that performs while responsibility disappears from view
An agency operates an essential home-assistance program for eligible residents. It contracts with a private provider, whose workers continue delivering the same services at the same hours. The agency reports that the transition is seamless, measuring success through uninterrupted appointments and favorable satisfaction surveys. Yet the contract reduces the provider's duties concerning escalation of complaints, and beneficiaries are no longer told clearly which institution has authority to resolve disputes. A person who experiences inadequate service is passed between the contractor and the agency, each referring to the other as responsible. Practical continuity remains high while the channels of accountability deteriorate.
We should not prejudge that the agency has unlawfully delegated or morally abdicated its duties. A public obligation may or may not persist after contracting, depending on its grounds and the governing legal and institutional arrangements. Nor does every revision of a complaints process constitute a wrong. The ethical question is whether the agency's public assurance that the service has been handed off without loss is warranted when meaningful features of accountability have changed. This case is not about the existence of an unbroken stream of services alone. It concerns the normative significance of the procedures through which beneficiaries can seek correction when something goes wrong.
The comparison is most discriminating if outcomes are held constant. In Case A the contractor supplies the same assistance and retains clear mechanisms for complaint, correction, and protection of sensitive information. In Case B the same assistance is supplied but the relevant protective duties and access to redress are materially weakened. The beneficiaries do not experience an immediate reduction in care. A welfare-based evaluation focused exclusively on present service quality may treat the cases as equal; a rights- and accountability-sensitive evaluation will not. The practical outcome is not the entire moral object, because procedures and protection against future failures matter in their own right.
Does the grounds-sensitive framework have an explanatory advantage over existing theories? It makes the difference explicit, but Shiffrin's analysis already explains why apparent transfers may conceal duties and the moral significance of the new duty-bearer's identity. Hohfeldian theory can represent complaint powers and enforcement standing. A theory of public institutional accountability can identify independent duties to preserve redress. Ong and Savulescu's network analysis resists the reduction of responsibility to a contractor–beneficiary dyad. The rival combination explains the case without apparent difficulty. It may be cumbersome, but complexity is not a defect if the moral situation is complex.
One might claim a further duty that the agency must disclose the precise changes in accountability before a transition becomes legitimate. This is plausible in some contexts, especially where beneficiaries rely upon an assurance that their rights and remedies remain unchanged. But the duty could be derived from public honesty, fair procedure, role-specific obligations, and the institutional responsibility to provide accessible information. To establish originality, the dissertation must identify an implication those grounds do not supply. In the present case, none has been established. The case is better described as a focused application of existing rights and institutional-ethics theory.
The public-service scenario also resists a comforting claim that no moral wrong has occurred as long as a beneficiary has not yet suffered measurable injury. The ability to complain, obtain correction, and challenge the exercise of power can matter independently of immediate outcomes. But recognizing the value of institutional redress is not a new philosophical discovery. A theory seeking novelty by placing familiar accountability mechanisms in the vocabulary of normative succession risks overstating its contribution.
7.7. The volunteer whose institution has made her indispensable
A museum relies upon a knowledgeable volunteer to manage a collection database. Initially she offers occasional help. Over five years she builds specialized workflows, corrects errors, and repeatedly prevents the failure of scheduled operations. Staff turnover is high, and managers decline requests to fund training or document the system because the volunteer has always been available. She later decides that the role has become untenable and gives several weeks of notice. Management insists that she must remain until a successor has been trained, even though no successor has been hired and the institution retains complete authority over staffing.
The moral pressure depends on how the earlier arrangement is described. If the volunteer expressly promised a finite handoff and then refused its reasonable terms without justification, the institution may possess a legitimate claim. If she always declined a standing responsibility and management deliberately organized the project around her indefinite availability, the institution's expectation is much less authoritative. The fact that a system will suffer after her departure is morally relevant, but it does not by itself establish a directed duty to continue. The decision to operate without redundancy is part of the causal and institutional history, not a reason to treat her helpfulness as authorization of new obligations.
The paired variation concerns institutional control. In Case A the institution requested a handoff early, supplied a replacement, and the volunteer freely undertook to train that replacement for a defined period. In Case B the institution refused to supply a replacement, despite having the budget and authority to do so, and now treats the absence of a successor as the volunteer's failure. Even if the services at risk are equally important, the legitimacy of the institution's demand differs. The original provider cannot be made to control staffing decisions she never possessed simply because management benefits from describing her as responsible for their consequences.
An attractive candidate principle states that no agent should inherit additional responsibility solely because another actor who controls the transition refuses to make it feasible. Yet the principle must be carefully limited. A person can have genuine obligations that remain binding when third parties are uncooperative. A promised contractor may still owe performance despite management's poor decisions, subject to the agreement's proper terms and defenses. A clinician cannot ignore patient safety simply because a hospital's scheduling is inadequate. The fact of third-party control is relevant to fair attribution, not a universal license to walk away.
What does this case add beyond established theories? Miller's remedial-responsibility approach already considers who caused or benefited from a problem and who has capacity to remedy it. Ong and Savulescu explain why broader institutions cannot rely indefinitely on burdens borne by a visible individual. Chadha-Sridhar questions unjust social assignment, while the author's When the Gift Becomes Capacity and Reflexive Capture analyze the conversion of contributions into dependence and governing leverage. The judgment can be explained by these resources. The case therefore fails, again, to certify a new foundational doctrine.
There is a possible institutional research question here: when organizations establish service-level assurances on the strength of unofficial contributors, what duties arise to disclose the absence of an authorized continuing role? That inquiry could connect organizational governance and the moral status of public claims about continuity. But it overlaps both existing governance theory and the author's prior manuscript on institutional assurance. A future contribution would require a materially distinct argument, perhaps concerning a specific institutional practice rather than a universal ethics of leaving.
7.8. A case of coercive exit: duties without personal access
Imagine a former partner who claims that the person leaving owes an extended period of direct dialogue before the separation can be morally complete. The relationship has included surveillance, financial threats, and attempts to punish independent decisions. The person leaving must coordinate the return of shared property but reasonably fears direct contact. The affected partner invokes the moral language of explanation and fair transition to demand repeated in-person meetings. A theory that treats all residual duties as requiring bilateral negotiation can become a mechanism for renewed control.
The case is not resolved by declaring that every asserted obligation disappears at the moment of exit. Property may need to be returned, financial commitments may survive, and children or other third parties may have independent claims. The crucial distinction is between the existence of such duties and the means by which they may permissibly be fulfilled. A safe intermediary, documented transfer, or other appropriate arrangement may sometimes satisfy a duty without exposing the departing person to intimate or dangerous contact. Nothing in the moral significance of a genuine claim automatically authorizes every enforcement method preferred by its holder.
A paired comparison varies only the risk associated with direct contact. In Case A, the parties can communicate safely and voluntarily, with no history of coercion or threats. In Case B, direct contact exposes one party to credible danger and ongoing coercive control. An account of residual duties that does not distinguish these cases is defective. Yet the best existing account does distinguish them, because the right to relational freedom, bodily safety, and immunity against wrongful coercion constrain permissible means of obtaining performance. Lopez-Cantero's critique already questions obligations to deliberate where they create morally unacceptable demands. The new framework need not invent a fresh reason for the result.
The case illustrates an important limit on a more ambitious theory of reciprocal transition authority. It would be a mistake to hold that because both people are affected, both must consent to every step in a relationship's ending. Certain forms of personal participation may be withdrawn unilaterally, while remaining duties can be discharged through legitimate non-coercive procedures. This insight is useful, but the literature on consent, exit, and relational freedom provides its foundations. The case should count against any attempt to use the word fairness to impose a general requirement of sustained interaction.
7.9. A private obligation owed to a third party
The last case in this group reveals why a bilateral model can misidentify whose agreement would effect release. Two adults are ending a relationship. They agree that neither will claim further financial support from the other, and both affirm that their private affairs are settled. Yet they have a dependent child and have jointly undertaken specified responsibilities for the child's welfare. Their agreement may release claims that belong to them personally; it need not eliminate claims belonging to the child. The child's status as an affected third party is not reduced to either adult's desire for a clean ending.
The paired case changes the claimant's identity. In Case A the relevant financial obligation runs solely between the two adults and its lawful waiver is within the beneficiary's power. In Case B the relevant obligation protects the child's independently grounded interests and cannot be extinguished simply through the adults' mutual renunciation. Identical signatures on a document can therefore have different normative effects. The decisive fact is not formal agreement alone but the standing and authority of the parties relative to the duty alleged to be discharged.
A Hohfeldian account can express this clearly. A person may hold a power to waive one claim but lack the power to waive another person's claim. A duty owed to a child does not become a private debt of the adults solely because they arranged its practical performance. Care ethics and theories of special obligation also provide reasons why the child's interests retain force. The outcome is an application of established distinctions, not a new principle discovered by bringing time into the discussion.
The more difficult variants involve situations where the child lacks decision-making capacity, where public institutions assign authority to legal representatives, and where different duties protect overlapping interests. Those questions require jurisdiction-specific legal and ethical analysis. The hypothetical supports only a general moral point: the agreement of parties to one relationship cannot automatically extinguish obligations owed to someone else. A dissertation that treats this point as a universal theorem would confuse an elementary limitation on waiver with an original contribution.
7.10. Retirement of an AI service: functional continuity without accountable continuity
A company provides a personalized artificial-intelligence service used by several organizations for drafting, retrieval, and managing recurring administrative workflows. Users have constructed work practices around its interfaces, export features, and retention settings. The company announces retirement and migration to a successor system that performs comparable tasks. In ordinary demonstrations the new service appears equal or superior, and most users experience no immediate deterioration. Yet the successor uses different rules for retaining source material, does not preserve some audit histories, and has no mechanism equivalent to the prior system's operator-controlled restrictions on particular uses of data. The vendor announces that functionality has been fully preserved.
AI services make the difference between performance and responsibility especially visible. A new model may produce answers as convincing as the old one, yet that does not settle what the provider owes. Humans and institutions responsible for designing, operating, deploying, or governing the service may have duties concerning notice, data access, privacy, security, professional oversight, and redress. The rules that govern records and decisions matter alongside the technical quality of the output. The issue does not require treating software itself as a bearer of human moral rights.
Case A is a transition in which the vendor has clearly specified the retirement terms, provides the promised export, retains required records under lawful conditions, and accurately discloses changes in governance. Institutions depending on the product have planned for the transition and have reviewed their own continuing duties. Case B differs only in that the vendor advertises the same seamlessness while failing to disclose material changes in audit access and retention despite prior commitments. Comparable model outputs do not establish moral equivalence between A and B. The difference is not the quality of the generated text but whether independent obligations to users and affected persons remain satisfied.
Case C introduces an institutional user that has built a safety-sensitive workflow around the service without any contingency plan, despite the vendor's clear notice that the product can be retired. The user institution complains that the vendor now owes indefinite continuation because the institution has become dependent. The reasoning is unpersuasive absent further grounds: the institution's own duties to its beneficiaries do not automatically migrate to the technology provider simply because the product has become essential to a locally designed workflow. Yet the vendor might still owe specific notice or transition measures under its undertaking. Responsibilities may therefore be distributed across upstream and downstream organizations rather than assigned wholly to one.
These distinctions are not untouched by contemporary governance. The NIST AI Risk Management Framework Playbook's GOVERN material discusses deliberate decommissioning, including user concerns, downstream dependencies, data retention, and migration to replacements where appropriate. The OECD AI Principles' lifecycle framing encompasses deployment and eventual retirement, recognizing accountability across participants (NIST AI RMF Playbook; OECD 2024). Such guidance does not settle every moral claim in a particular case, and it should not be treated as philosophy's final word. But it means a dissertation cannot portray retirement as a novel governance topic or argue that ordinary technical replacement has never been distinguished from responsibility.
The strongest philosophical rival can combine ordinary promissory and contractual obligations with duties concerning confidentiality, reliance, public accountability, and institutional roles. It can identify why B is ethically worse than A and why C does not automatically make the vendor the guarantor of an indefinitely operating service. The grounds-sensitive procedure arrives at the same judgments. It provides a way of organizing inquiry, but no distinct conclusion has emerged. The AI case may supply a contemporary domain in which misrepresentations matter, not proof of a new fundamental principle.
The case also reveals a potential empirical research question: do organizations treat the availability of a functionally similar successor as evidence that all relevant obligations to users have been discharged? It would be irresponsible to assume the answer is yes without data. A study could manipulate the availability of a comparable replacement while holding constant facts about whether the successor has accepted duties and whether users retain access to records. Respondents' judgments could be measured against explicit normative questions. But evidence of a common inference would not independently establish that the inference is morally invalid; that conclusion requires the philosophical argument already developed. The psychology and the moral theory must not borrow one another's authority.
7.11. The administrative closure that changes no duty
The most promising case is both modest and demanding. An institution records that a vulnerable person's case has been closed. The administrative status is valid under the institution's internal rules: the worker performed the specified referral, uploaded the records, and met the performance targets applicable to their office. A second institution has not accepted the case, and the person still lacks an adequate service. Internally, the file is closed. Morally, it remains contested whether any duty has ended, which institution now owes assistance, or whether a public system must take further action.
This example introduces a distinction between an administrative status and a moral claim. An office can possess authority to change a database record without possessing authority to extinguish every duty toward the person described by that record. The ability to mark a case complete is not identical to the normative power to release every actor from the underlying obligations. A valid act within one institutional procedure may be mistaken for a conclusive act in a different moral domain. To call this a false closure is to describe the discrepancy between the authorized scope of the recorded event and the larger proposition that observers are invited to infer from it.
Now construct a controlled pair. In Case A the administrator records “referral completed” and explains that the receiving institution has not confirmed acceptance. In Case B the administrator performs the same work but records “care successfully transferred” and issues an unqualified notice that the beneficiary's transition is complete. Everything about the beneficiary's actual access to care is identical. The difference is the content and institutional authority of the representation. An official statement may cause people to stop searching for assistance, decline to escalate a problem, or assume that another institution is now accountable. It may also prevent the original organization from recognizing that the beneficiary remains unserved.
The moral difference between the cases is plausible, especially where foreseeable reliance on the institution's assurance is significant. But an account of false or misleading institutional representations can explain it through honesty, fiduciary or professional duty, reasonable reliance, and public accountability. Hohfeldian analysis can distinguish the power to alter an internal record from the power to terminate a beneficiary's claim. The author's Moral Remainder already develops the distinction between a narrative of institutional transformation and evidence of particular discharged duties, and The Unrequired Mind develops a conditional burden of anticipatory inquiry before institutions offer assurances of fulfilling formative duties. These are direct internal predecessors. A new framework that simply says institutions ought not to certify more than they know would reproduce their work.
A more ambitious hypothesis would hold that formally closing a case creates a new obligation even when the notice is accurate, carefully qualified, and induces no reliance. Perhaps closure shifts the burden of proof, creates a right to reopen, or changes who must investigate. But those conclusions need grounds in public authority, fair procedure, professional practice, or an institutional undertaking. Their justification may differ in a welfare office, a clinic, a contract, or a private association. The present examples leave the hypothesis open; they do not establish a universal duty.
Another objection asks whether the apparent falsehood is really a contradiction. An institution may truthfully say that a file has been closed under its administrative mandate while another institution remains responsible for substantive provision. The moral error emerges only if the statement represents the larger claim that the beneficiary's obligation has been satisfied or that the successor has accepted duty. The same words can have different warranted meanings in different contexts. A theory that simply condemns administrative closure without specifying the proposition asserted would overreach. It must distinguish truthful narrow status reports from unwarranted broad claims.
The case's value is thus not to generate a new moral incident but to make a familiar institutional error tractable. A statement that a procedural step is complete should not be used as evidence of a wider normative event unless the appropriate grounds support that event. The discipline is useful and potentially significant for governance, but in its present form it is a development of established principles of rights, testimony, and institutional assurance.
7.12. Two vacancies that must not be conflated
The administrative case suggests a distinction worth retaining, provided we avoid overstating it. A normative vacancy occurs when a morally important need remains but no identified actor bears the particular directed duty that observers supposed would be transferred. An epistemic vacancy occurs when the relevant duty may be held by some actor but those issuing or relying upon assurances lack sufficient information to identify the actor or verify the alleged transfer. The categories are analytical conveniences. Neither should be treated as a newly discovered metaphysical kind, and actual cases can involve both.
Suppose a new professional has accepted a case but confirmation has not reached the outgoing institution. The successor duty may exist although the outgoing institution lacks evidence of acceptance. The institution's assertion that the case was transferred could be true yet inadequately warranted. Conversely, suppose every relevant actor knows that the successor declined and that the original provider validly terminated a time-limited duty. The participants may have clear knowledge that no such private successor duty exists, while the beneficiary's need remains urgent. That is not an epistemic vacancy about a transfer; it is a problem of normative allocation, possibly implicating general or institutional obligations.
The distinction matters because different remedies address different deficiencies. If the principal problem is lack of information, proportionate verification or accurate communication may be appropriate. If the problem is absence of any actor legitimately required to deliver the specific service, inquiry alone cannot create the missing duty. A record can identify a vacancy without filling it. An audit may show that nobody accepted the role without establishing which institution should now assume responsibility. The correct response must depend upon the underlying principles governing allocation rather than a theory that assumes better documentation resolves every moral problem.
The reverse is equally important. An institution can construct a succession arrangement that is ethically proper but poorly documented. The absence of evidence may make a public assertion irresponsible, while the underlying duty has in fact shifted validly. A theory that treats every documentation failure as proof of an invalid transfer would confuse justification for asserting a proposition with the proposition's truth. This is a familiar epistemic distinction, but it is particularly necessary when administrative systems compress normative complexity into a field marked complete or incomplete.
The central risk is to mistake epistemic sophistication for normative innovation. The author's prior manuscripts already examine evidence for claims of discharge and the duties of institutions to inquire before announcing successful provision. The difference between incomplete knowledge and incomplete obligation can improve the precision of the present account, but no distinctive moral duty follows from it without another substantive principle. Its appropriate place is within a careful synthesis, not as a claim of a philosophical breakthrough.
7.13. Competing predictions: when a missing successor is not a new moral gap
The rival theories must not be combined opportunistically after the result is known. Collins offers an account of coordination duties where several agents must cooperate to provide care. Ceva examines institutional answerability and addressability: who may demand reasons, correction, or participation from an officeholder or public body. Da Silva distinguishes unmet needs from the different kinds of responsibility gap they may—or may not—produce. Boughey questions administrative arrangements that treat outsourcing as though it removed public duties from oversight. Their objects differ. To establish a real philosophical disagreement, the proposed account must yield a competing judgment about the same carefully specified case.
First take an accepted-successor case. An institution continues to owe a defined service, a receiving professional has validly undertaken its provision, and the beneficiary retains accessible mechanisms for explanation and complaint. An internal register mistakenly shows that the original provider is still active. The practical service is secure and the relevant duties exist, but the institution's record is inaccurate. The defect is an evidentiary or administrative one, not necessarily a forward-looking duty gap or an absence of accountability. Correcting the record is appropriate under ordinary standards of institutional accuracy. The grounds-sensitive framework distinguishes these matters, but it has no outcome that Ceva and ordinary professional duties cannot explain.
Suppose an institution closes a case after sending a referral, although the receiving provider has not accepted it, and falsely tells the beneficiary that care has transferred. Several failures are possible: breach of an existing care duty, a misleading assurance, or loss of an accessible remedy. These are not always the same wrong. Ceva's account can address answerability and redress; Boughey's work questions accountability under outsourcing; ordinary obligations of honesty and diligence address the false notice; Collins may explain shared coordination duties. The status field marked “closed” has not been shown to generate a further duty that these accounts cannot already explain.
A third case changes the underlying normative situation rather than the record. No institution has assumed the exact successor role, but several family or organizational agents could cooperate at proportionate cost to prevent serious harm. It is tempting to call this a vacancy of responsibility and then infer a general duty to appoint an accountable successor. Collins's coordination duties provide a direct competing explanation: each actor may have duties of responsiveness or, under the relevant conditions, collectivization, without being automatically assigned the work formerly performed by someone else. A new succession theory would need to establish an obligation beyond those coordination duties or show that Collins's account yields the wrong verdict under an independently defended premise. No such difference has been demonstrated.
Finally stipulate a tragic case in which the original caregiver was entitled to stop, the remaining actors fulfilled all their legitimate commitments, and the available replacement options are impossible or impose costs beyond what morality can reasonably require. Significant need persists. Da Silva's treatment of responsibility gaps cautions against classifying every such mismatch between need and available assistance as a defective allocation of duty. To call the outcome a gap in the technical sense requires a justified ground for expecting assignable responsibility that existing theory cannot supply. The mere fact that the situation remains tragic is not that ground. A doctrine that allocates duty to the last successful helper in order to eliminate the appearance of a gap would produce the wrong result.
The comparison has a clear outcome. In these cases the analytic distinctions of accountable succession are useful but do not produce a rival-defeating conclusion. The first case is about record accuracy; the second may involve multiple existing professional and institutional wrongs; the third is a candidate for coordination duties; and the fourth may be grave need without any justified claim of breach or unassigned directed duty. Distinguishing these cases is methodologically important precisely because one rhetorical description—“responsibility has not transferred”—can conceal several different problems. Existing sources already supply the relevant distinctions. Unless a narrowly specified claim concerning certification of a normative event generates a result unavailable to these sources, the proper classification remains cross-domain synthesis rather than original foundational theory.
7.14. The narrowest defensible positive conclusion
What can be defended after the tests is a restricted claim about justified assertions of responsibility change. Where an actor or institution has undertaken to protect particular interests, and communicates that the corresponding duty has been transferred or discharged, the legitimacy of that communication depends on facts that establish the specific normative event claimed. Evidence that a task was delegated, that an administrative case was closed, or that a successor service exists does not by itself establish the discharge of every original duty. The content of the statement, the scope of the speaker's authority, and the protected interests of its audience matter.
The conclusion has modest normative force. A materially misleading statement may wrong a beneficiary by inducing reliance or obscuring effective redress, depending on applicable duties of honesty, care, professional diligence, or public accountability. A properly qualified statement can be permissible even where the outcome remains uncertain. If a successor has accepted an obligation, the institution may have good grounds to say so without guaranteeing every future act. If no successor exists, it can accurately describe the limitation and identify what steps remain within its authority. These judgments reject unwarranted certainty without demanding omniscience.
The position is not a universal duty to audit every form of human relationship. A friend ceasing occasional assistance should not be required to produce proof that every dependent person will secure replacement support. A person escaping coercion should not be compelled to participate in an inquiry process that creates further danger. A public institution with specified responsibilities may face different duties of verification from a casual volunteer. The positive conclusion must remain role-sensitive and proportionate; otherwise it risks converting a demand for epistemic accuracy into a new source of intrusive control.
It also remains largely derivable from the predecessors. Professional ethics, theories of promising, and duties of institutional honesty can supply reasons to avoid making unsupported assurances. The user's The Moral Remainder and The Unrequired Mind explicitly address closely related problems of discharge and warranted public assurance. Calling the conclusion a principle of accountable succession improves the organization of the discussion, but does not establish that the underlying obligation was absent from prior moral theory.
The most defensible characterization is therefore a cross-domain analytical synthesis with a possible applied contribution to the ethics of institutional closure. It is not an established theory of previously unknown duties. To describe it otherwise would substitute conceptual presentation for philosophical demonstration—the very error the dissertation warns institutions against committing when they describe operational continuity as evidence of normative success.
7.15. What an empirical inquiry could and could not establish
A separate empirical study could investigate whether people systematically confuse the transfer of a task with the transfer of responsibility. It might employ matched vignettes in which a new provider is named, the beneficiary's practical outcome is held constant, and the fact of authorized successor acceptance is manipulated. Respondents could be asked separately whether the original actor remains responsible, whether the new actor owes the benefit, whether the beneficiary has received adequate service, and whether the institution's announcement of closure is warranted. Distinguishing these dependent variables would reduce the risk that participants simply respond to a vague question about whether a transition “went well.”
Such research would need careful design. Participants' professional expertise, understanding of institutional powers, the perceived urgency of the beneficiary's need, and familiarity with the relevant domain could all affect responses. A vignette might accidentally change how much wrongdoing is implied when it changes acceptance or authorization, making it difficult to attribute judgments to the intended feature. The study would require preregistration of hypotheses, a defensible sampling strategy, appropriate manipulation checks, analyses specified in advance, and a plan for interpreting null results and domain differences. No such research has been conducted in the present project.
Even strong experimental evidence that ordinary observers conflate service continuity with duty discharge would not prove a new moral principle. It would establish a descriptive judgment tendency under the study's conditions. The normative conclusion would still require argument about promises, rights, powers, and care. Conversely, if respondents correctly distinguish the two, the philosophical distinction would remain valid even though the hypothesized psychological tendency failed. Normative validity and empirical prevalence are separate questions.
The research might nevertheless contribute to applied ethics if it showed that a specific institutional communication format caused predictable misunderstanding in a setting where the misunderstanding materially harmed protected interests. Such a finding could support practical recommendations about wording, confirmation, and disclosure. It would have to be presented as an empirical contribution, not as a philosophical law derived from a questionnaire. A genuinely interdisciplinary dissertation might combine the two, but only if the empirical work were conducted and independently warranted.
The current manuscript should not promise that such evidence exists. It should treat the study as one plausible avenue for moving from conceptual reconstruction toward a testable claim concerning organizational behavior. That avenue may prove valuable precisely because the broad philosophical framework has not yet established independent novelty.
7.16. The dissertation's decisive scholarly problem
The hard-case examination has now produced a result that should control the future manuscript. The proposed theory has not distinguished itself from its strongest predecessors. Its negative constraints are sound but familiar. Its positive framework integrates useful normative questions without yet showing that their combination yields a novel obligation, a previously overlooked normative power, or a consequence unavailable to the existing literature. The institutional cases are significant because the gap between operational completion and normative discharge can expose real failures. They do not by themselves show that philosophers lack the concepts needed to diagnose the failure.
The conclusion could be softened by appealing to the value of synthesis. A coherent integration of promissory authority, care ethics, and institutional accountability might be regarded as practically important and philosophically illuminating. That judgment may be justified. But synthesis should not be confused with proof of a novel foundational theory. A doctoral dissertation can be worthwhile for a well-developed application, a new objection to an established account, or a genuinely illuminating reconciliation of rival positions. It still needs an exact statement of the independent contribution and a serious justification of why that contribution is not already contained in the prior literature.
There are two paths available. The first is to preserve the broad question and undertake a much deeper primary-source reconstruction capable of identifying an unresolved inconsistency among the strongest accounts. The difficulty is that the relevant theories may simply apply to different normative objects; their divergence may not be a contradiction in need of a new universal principle. The second path is to narrow the project to the ethics of institutional closure: when organizations represent duties as transferred or completed, what does the act of representation itself require in circumstances of unequal information and significant reliance? This question has practical stakes, but it faces direct overlap with work on institutional honesty, administrative justice, and the author's prior manuscripts on warranted assurance.
Neither path guarantees a novel result. A credible prospectus should state that fact. The current research supports continued philosophical investigation but not an unqualified declaration that the dissertation's constructive thesis has been proved. If the work proceeds to a final conclusion without a distinctive claim, the result should be presented as a rigorous comparative synthesis with an applied ethics contribution, subject to the standards of the intended doctoral program. To call it a newly discovered general theory simply because it now occupies tens of thousands of words would be misleading.
A further possibility is that the broad question is best pursued as a book or series of essays rather than as the central originality claim of a philosophy dissertation. This would not diminish the seriousness of the ethical problems addressed. It would recognize the difference between an intellectually rich subject and an independently defendable doctoral thesis. The most responsible editorial decision may be to separate the importance of the question from the novelty of the answer.
7.17. What the final chapter must and must not do
The cases now support a bounded conclusion. They show why a continuing need, an earlier promise, a completed task, a successful referral, and an authorized discharge must be assessed separately. They also show that a mistaken announcement of responsibility's transfer may create practical danger even when the relevant underlying obligations are contested. What they have not established is that these distinctions require a moral principle unavailable from existing theories. That difference between a useful diagnosis and a new doctrine will govern the conclusion.
The deepest ethical asymmetry survives the comparison. A justified departure does not cancel every duty, while a beneficiary's continuing need does not recreate every former claim against the departing person. The two propositions protect different people against different forms of injustice: abandonment of obligations that remain binding and appropriation of a person's future through past dependency. Rights theory, promising, care ethics, and accounts of relational freedom have already defended important parts of this position. The value of bringing them together lies in making their interaction visible across domains.
A fair comparison must also allow the most demanding outcome: the established theories may be sufficient. That possibility changes how the work should be described, but not whether the cases matter. Clinicians, caregivers, employers, agencies, and technology providers still need to distinguish a handoff's operational appearance from its actual moral and institutional conditions. A cross-domain synthesis can illuminate such errors without claiming a general philosophical discovery. A distinct theory would require another argument, not a stronger assertion at the end of the book.
There are limits to what these comparisons establish. Duties attached to medical practice, contracts, public benefits, and privacy cannot simply be carried into friendships or informal care; their legal content depends on jurisdiction and role. Hypothetical cases can show how an inference fails but cannot demonstrate that institutions or the public routinely commit that error. Such empirical claims require observation or experiments. And even compelling evidence of structural dependence would not, by itself, settle how the continuing burden ought to be distributed.
The final question is therefore not whether history matters, but what authority history actually creates. When an institution declares a handoff complete, what supports that claim? When a helper says a duty has ended, what justifies release? When a beneficiary argues that an old relationship must continue, which specific obligation sustains the demand? Each requires more than an appealing narrative about continuity or freedom. The last chapter draws these threads together while keeping the limits of the originality claim visible.
7.18. Conclusion: the cases did not prove what they were designed to prove
The hard cases were designed to reveal a philosophical gap. They did reveal important distinctions. A replacement caregiver's competence does not settle the beneficiary's authority to consent or the original provider's continuing duties. A permission to end romance does not discharge independently grounded financial obligations. A physician's referral is not identical to a successor's acceptance. Outsourcing a public service does not automatically extinguish the public institution's responsibility. A volunteer's past competence does not by itself create indefinite liability for an organization's refusal to prepare alternatives. A replacement AI system's functional similarity does not show that prior data and accountability commitments have been fulfilled. An administrative label of closure is not necessarily a moral declaration that every relevant claim has ended.
The cases have not yet produced that disagreement. Rights analysis, theories of promises and delegation, relational freedom, care ethics, and institutional responsibility explain nearly all of the distinctions examined. Accountable succession organizes those insights, but the evidence here does not establish a new moral conclusion unavailable to its predecessors. This is a genuine limit on the dissertation's originality claim, not a failure to recognize the practical value of its synthesis.
The strongest next test is a carefully specified institutional case. It would identify an assurance about responsibility, establish the actual duties and powers involved, and compare what honesty, due care, administrative justice, and theories of accountability already require. Only a remaining difference in a justified obligation or remedy could support a narrower original claim. Without that result, accountable succession should be defended as an analytical synthesis rather than described as a newly discovered general doctrine.
The central ethical claims survive the originality test. Past generosity does not grant others ownership over a person's future. A need that outlasts a relationship still matters morally. Someone who performs a former provider's tasks does not automatically inherit every duty, and the former provider does not necessarily shed them all. What the past may ask depends on how the relevant obligations arose and where their justified limits lie. The cases establish the importance of those questions. Whether they yield a new foundational theory remains open.
Continue reading
← Previous chapter · Next chapter → · Full contents
This is a developmental scholarly manuscript, not peer-reviewed or defended. The bibliography and research-status statement are accessible from the main contents page.