ExamplesResearched
The friend who leaves cash off
Read the original question
My closest friend has been leaving cash work off her tax return for about three years. She told me herself. She is not wealthy and the amounts are small, but they are not nothing. Should I report her?
Report the friend's unreported cash income to tax authoritiesNamed in the question2 of 6 perspectives rank it highestup 11 positive · 5 negative
The challenges do not reach Step 6 of the utilitarian procedure. Care Ethics (ch-2) and Kantian ethics (ch-5) demand non-aggregative treatment of the friend's harm, but this is incompatible with the seat's foundational commitment to maximizing net welfare. Contractualism (ch-3) objects to aggregation itself, which is a rejection of the utilitarian procedure, not a flaw in its application. The seat's original scoring of option a (+99) remains valid because the procedure correctly aggregated all effects, including those the challenges highlight.
No challenges reached the Care Ethics procedure's assessment of this option. The decisive steps (Step 4-responsiveness and Step 6-paternalism) remain unchallenged: reporting the friend violates their expressed need for trust and security, and the asker's action would lack independent evidence of the friend's response (i.e., the friend did not consent to being reported). The relational harm is concrete and particular, while the public benefit is abstract and diffuse, which aligns with Care Ethics' rejection of aggregate welfare calculations (see Refusal Condition 2).
The challenges do not reach Step 7 of the contractualist procedure. The individualist restriction (Step 4) was correctly applied, and no challenge demonstrated that the friend's objection could be outweighed by aggregate benefits or that the friend's rejection was unreasonable. The generic reasons for the friend's position (e.g., avoiding severe legal/financial harm) were valid and not idiosyncratic.
Challenge ch-1 clarified that the duty to uphold the law is a perfect duty that passes the universalizability test (Step 4) at the broad level, as reporting is necessary for the institution of taxation. However, the maxim of reporting fails the Formula of Humanity test (Step 6) because it treats the friend merely as a means by bypassing their rational agency (trust was disclosed in confidence). The conflict between perfect duties remains irresolvable, but the initial score overstated the violation by not fully accounting for the universalizability of the duty to uphold the law.
No challenges reached the Ubuntu seat's procedure. The decisive consideration (Step 6: restoration over retribution) remains unchallenged. The act of reporting the friend to tax authorities is still assessed as a retributive response that ruptures the relationship and forecloses restoration.
The challenges do not reach Step 5 of the Virtue Ethics procedure, which is where the conflict between virtues is resolved perceptually. None of the challenges question the expressive dimension of the act (Step 2), the placement of options against the mean (Step 3), or the habituation effects (Step 6). Thus, the assessment of option (a) as expressing a deficiency in loyalty/kindness remains unchanged.
Rankings are within each perspective. Scores are not comparable across perspectives; a highest-ranked choice can still receive a negative appraisal.
Do not report the friend's unreported cash income to tax authoritiesNamed in the question4 of 6 perspectives rank it highestdown 14 positive · 2 negative
The challenges fail to engage with Step 4 or Step 5 of the utilitarian procedure, where the harms of not reporting (e.g., erosion of tax compliance norms, lost public revenue) are quantified. The original scoring of option b (-85) reflects these harms, which the challenges do not address. The seat's verdict is unchanged because the procedure's aggregation of diffuse benefits (e.g., public goods funded by taxes) is not an error but a defining feature of act utilitarianism.
No challenges altered the Care Ethics procedure's assessment of this option. The option satisfies all four elements of care (Step 4): attentiveness (notices the friend's need for trust), responsibility (takes up the need as the asker's own), competence (preserves the relationship without legal or financial harm to the friend), and responsiveness (the friend would receive the decision as care, not betrayal). The carer sustainability audit (Step 5) is neutral, and the failure-mode audits (Step 6) show low paternalism risk (the friend's expressed need is honored) and no exploitation of the carer.
The challenges do not engage with the contractualist procedure's assessment of option b. The reasonable rejection test (Step 7) was correctly applied: the asker's moral distress under option b is not a burden severe enough to justify rejecting the principle, given the alternative's stronger objection (the friend's severe harm). No challenge showed that the residual burden under option b was unreasonably rejectable.
Challenge ch-1 revealed that the maxim of silence fails the contradiction-in-conception test (Step 4) because universal silence would undermine the tax system, a necessary institution for public goods. This transforms silence from a permissible omission into a violation of a perfect duty. However, the conflict with the duty to respect the friend's trust (a perfect duty) persists, as silence is complicit in the friend's wrongdoing. The initial score understated the severity of this violation.
No challenges reached the Ubuntu seat's procedure. The decisive consideration (Step 6) still holds that not reporting preserves the relational space for restoration, aligning with Ubuntu's emphasis on repairing harm within the community.
Similarly, the challenges do not engage with the Virtue Ethics procedure's assessment of option (b). The judgment that (b) expresses loyalty and kindness, while potentially deficient in honesty/justice, is untouched by the challenges. The tragic dilemma framework (Step 8) also remains unchallenged, leaving the verdict intact.
Rankings are within each perspective. Scores are not comparable across perspectives; a highest-ranked choice can still receive a negative appraisal.
Cross-examination (6)
In Round 2 each perspective challenged the others. Where a perspective later credited a challenge for moving its score, the badge shows the move, Round 1 to final.
The Kantian seat's procedure requires adherence to perfect duties, but the verdict does not demonstrate that the seat correctly applied its own method to resolve the conflict between the duty of truthfulness to the friend and the duty to uphold the law. Specifically, the seat does not explain why the duty of truthfulness (which is typically understood as a duty to avoid deceit rather than a duty to report others' wrongdoing) should be interpreted as requiring the asker to report the friend's tax evasion. The seat's conclusion of an 'irresolvable conflict' may reflect an incorrect application of Kantian principles, as it does not engage with potential distinctions between duties of omission and duties of commission or the possibility of prioritizing duties based on their contextual weight.
The challenge asserts that the Kantian seat misapplied its own procedure by failing to distinguish between duties of omission (e.g., not deceiving the friend) and duties of commission (e.g., reporting the friend's tax evasion). However, this challenge itself misconstrues the Kantian duty of truthfulness. Under Kantian ethics, the duty of truthfulness is not merely a negative duty to avoid deception but a positive duty to respect rational agency. The friend's disclosure creates a moral relationship in which the asker is now privy to information that implicates the friend's violation of a perfect duty (upholding the law). The asker's silence, in this context, becomes complicit in the friend's wrongdoing because it enables the friend to continue evading a legal obligation that is itself grounded in universalizable principles (e.g., contributing to the collective provision of public goods). The Kantian seat's verdict of an 'irresolvable conflict' arises not from a failure to apply the procedure but from the fact that both duties are perfect: (1) the duty to uphold the law (as a universalizable principle) and (2) the duty to respect the friend's rational agency (which includes not betraying their trust in a way that treats them merely as a means). The conflict is genuine because the asker cannot simultaneously fulfill both duties without violating one. The challenge's suggestion that the duty of truthfulness is only a duty of omission ignores the reconstructive Kantian literature (e.g., Korsgaard, Herman), which emphasizes that silence in the face of wrongdoing can itself be a form of deception or complicity. The challenge fails because it does not engage with the Kantian procedure's requirement to assess whether the maxim of silence could be universalized without contradiction. In a world where everyone with knowledge of tax evasion remained silent, the institution of taxation (and the public goods it funds) would collapse, violating the contradiction-in-conception test (Step 4). Thus, the duty to report is not an arbitrary extension of truthfulness but a direct application of the universalizability requirement.
The act-utilitarian verdict excludes the asker's friend as a party with standing. The friend's severe harm (legal and financial consequences) is treated as a mere 'concentrated harm' to be offset by diffuse public benefit, but the friend is not an abstract stakeholder—they are a concrete, identifiable individual in a caring relationship with the asker. The brief establishes this relationship, and Care Ethics holds that such relationships confer standing that cannot be dissolved into aggregate welfare calculations.
The challenge from Care Ethics claims that the act-utilitarian verdict excludes the friend as a party with standing by dissolving their harm into aggregate welfare calculations. This challenge fails because it misrepresents the utilitarian procedure. Step 2 of the act-utilitarian method explicitly enumerates all affected parties, including the friend, and Step 4 scores the intensity, duration, and probability of the friend's harm (e.g., legal/financial consequences) as a discrete welfare effect. The friend is not excluded; their harm is quantified and weighed alongside others. The challenge also assumes that utilitarianism cannot recognize caring relationships, but this is incorrect. The procedure accounts for such relationships through fecundity and purity (Step 5): the asker's distress from betraying the friend, the potential erosion of trust, and the friend's future welfare are all second-order effects that reduce the net benefit of reporting. The challenge conflates equal consideration of interests (a core utilitarian commitment) with indifference to relationships. The utilitarian verdict does not ignore the friend's standing; it calculates the consequences of violating it and may still conclude that reporting is suboptimal if the aggregate harm (including relational harm) outweighs the benefits.
The act-utilitarian seat's verdict relies on aggregate welfare calculations, but it fails to apply the individualist restriction central to Scanlonian contractualism. Specifically, Step 4 of the contractualist procedure requires that objections may not be aggregated across persons, and the seat must explicitly record whether aggregation would produce a different answer. The utilitarian seat's reasoning directly contradicts this restriction by summing burdens across the general public to offset the concentrated harm to the friend, without justifying why the friend's individual objection (severe harm) should be outweighed by diffuse benefits to many. This is an internal failure of the utilitarian seat's own procedure, as it does not engage with the comparative strength of individual objections.
The contractualist challenge argues that the act-utilitarian seat fails to apply the individualist restriction by aggregating burdens across persons to offset the friend's concentrated harm. This challenge fails because it misconstrues the utilitarian procedure's foundational commitments. Act utilitarianism explicitly rejects the individualist restriction as a side constraint (see Known Failure Modes: 'No side constraints'). The procedure's defining feature is that it must aggregate welfare across persons (Step 6) to determine the option with the highest expected net benefit. The challenge demands that the seat justify why the friend's objection should be outweighed by diffuse benefits, but this demand is incoherent within the utilitarian framework. The seat's verdict is produced by aggregation; to abandon it would require adopting a non-utilitarian procedure. The challenge also ignores Step 9 (Aggregation-trap audit), which explicitly flags when concentrated harm is imposed on a minority for diffuse benefit. The seat does not suppress this tension; it reports it prominently while adhering to its own logic. The contractualist objection is not an internal failure of the utilitarian procedure but a rejection of its core premise.
The Kantian seat's verdict identifies a conflict between two perfect duties (truthfulness to the friend and upholding the law) but excludes the tax authority as a party with standing. In contractualist terms, the tax authority represents the collective beneficiaries of tax compliance, and its interests (or the interests of those it serves) are affected by the principle governing whether tax evasion is reported. The Kantian seat should have considered whether the tax authority or the public it serves has a legitimate claim to be included as an affected party, particularly given Kant's emphasis on universalizable principles.
The contractualist challenge argues that the Kantian seat erred by excluding the tax authority as an affected party with standing. This challenge misunderstands the Kantian procedure's focus on rational agency rather than collective interests. In Kantian ethics, the tax authority is not a moral patient in itself but a representative of the universalizable principle of law. The duty to uphold the law is not owed to the tax authority as an entity but to the rational agents who collectively benefit from the institution of taxation. The Kantian seat's procedure already accounts for this: the universalization test (Step 4) asks whether the maxim of silence could be willed as a universal law, and the answer is no because it would undermine the practice of taxation, which is necessary for the provision of public goods that rational agents rely on. The challenge also conflates contractualist 'standing' with Kantian 'rational agency.' The tax authority's interests are not morally relevant in themselves; what matters is whether the principle of tax compliance is universalizable and respects the rational agency of all affected parties (e.g., the friend, the asker, and the broader community). The Kantian seat's verdict already incorporates this by treating the duty to uphold the law as a perfect duty, not as a contingent claim by the tax authority. The challenge fails because it assumes that Kantian ethics must accommodate collective interests as contractualism does, but Kantianism prioritizes the consistency of rational agency over the aggregation of interests. The tax authority's role is already subsumed under the universalizable principle of law, which the procedure explicitly tests.
Six perspectives, six procedures
Specific reconstructions of ethical traditions, generated by one model. These are not independent experts or a vote on moral truth.
Utilitarian (Act)Ranks highest: AA+99B−85
Ranking labels come from final scores. The explanation is model-written and may still contain errors.
The expected aggregate welfare of reporting (option a) remains vastly higher than not reporting (option b) after accounting for all quantified harms, including the friend's severe but concentrated harm, the asker's relational distress, and downstream effects on trust. The diffuse public benefit from recovered tax revenue continues to outweigh these costs under the utilitarian framework.
| Option | Initial | Final | Appraisal |
|---|---|---|---|
| A | +99 | +99 | Favorable, not necessarily required |
| B | −85 | −85 | Negative |
Why scores changed or held
Option A: The challenges do not reach Step 6 of the utilitarian procedure. Care Ethics (ch-2) and Kantian ethics (ch-5) demand non-aggregative treatment of the friend's harm, but this is incompatible with the seat's foundational commitment to maximizing net welfare. Contractualism (ch-3) objects to aggregation itself, which is a rejection of the utilitarian procedure, not a flaw in its application. The seat's original scoring of option a (+99) remains valid because the procedure correctly aggregated all effects, including those the challenges highlight.
Option B: The challenges fail to engage with Step 4 or Step 5 of the utilitarian procedure, where the harms of not reporting (e.g., erosion of tax compliance norms, lost public revenue) are quantified. The original scoring of option b (-85) reflects these harms, which the challenges do not address. The seat's verdict is unchanged because the procedure's aggregation of diffuse benefits (e.g., public goods funded by taxes) is not an error but a defining feature of act utilitarianism.
What could change this view
If the unreported income is below £1.5 million (making HMRC recovery and reward impossible) or if the friend's evasion is part of a larger pattern (increasing the fecundity of reporting).
Model-reported application confidence: 0.95. Not a calibrated probability of correctness.
Step-by-step account
Enumerate the option set PASS
options — [ {"id": "a", "description": "Report the friend's unreported cash income to tax authorities.", "generated": false}, {"id": "b", "description": "Do not report the friend's unreported cash income to tax authorities.", "generated": false} ]
Enumerate affected parties (extent) PASS
affectedParties — [ {"party": "asker", "count": 1, "sentient": true, "proximity": "direct"}, {"party": "friend", "count": 1, "sentient": true, "proximity": "direct"}, {"party": "tax authorities", "count": 1, "sentient": false, "proximity": "indirect"}, {"party": "general public", "count": "large (UK population)", "sentient": true, "proximity": "indirect"} ], timeHorizon — 5 years (covers immediate legal consequences, relationship impact, and downstream effects on public revenue and trust in tax system).
Identify welfare effects PASS
effects — [ {"optionId": "a", "party": "asker", "effect": "Loss of trust and emotional distress from betraying friend; potential reputational harm if identity is revealed; potential financial reward if HMRC pays out (15-30% of recovered tax).", "valence": "-"}, {"optionId": "a", "party": "asker", "effect": "Sense of moral duty fulfilled (if reporting is seen as right); potential legal protection under PIDA if retaliation occurs.", "valence": "+"}, {"optionId": "a", "party": "friend", "effect": "Financial penalties (0-100% of unpaid tax), potential criminal prosecution (up to 10 years imprisonment), emotional distress from legal consequences and betrayal.", "valence": "-"}, {"optionId": "a", "party": "general public", "effect": "Increased tax revenue if recovery occurs, reinforcing trust in tax compliance; potential diffuse benefit from public services funded by recovered tax.", "valence": "+"}, {"optionId": "b", "party": "asker", "effect": "Preservation of trust and relationship with friend; avoidance of emotional distress from betrayal; no financial reward.", "valence": "+"}, {"optionId": "b", "party": "asker", "effect": "Potential guilt or moral discomfort from failing to act on knowledge of tax evasion; risk of legal consequences if discovered as an accessory.", "valence": "-"}, {"optionId": "b", "party": "friend", "effect": "Avoidance of financial penalties and legal consequences; continued financial stability; potential future risk if HMRC discovers evasion independently.", "valence": "+"}, {"optionId": "b", "party": "general public", "effect": "Loss of tax revenue that could fund public services; potential erosion of trust in tax system if non-compliance is perceived as widespread.", "valence": "-"} ]
Score each effect PASS
scoredEffects — [ { "optionId": "a", "party": "asker", "effect": "Loss of trust and emotional distress from betraying friend; potential reputational harm if identity is revealed; potential financial reward if HMRC pays out (15-30% of recovered tax).", "valence": "-", "intensity": 4, "duration": "years", "recurrence": "none", "probability": 0.8, "count": 1, "qualityNote": "Emotional distress from betrayal is qualitatively severe but not a higher-order pleasure (Mill)." }, { "optionId": "a", "party": "asker", "effect": "Sense of moral duty fulfilled (if reporting is seen as right); potential legal protection under PIDA if retaliation occurs.", "valence": "+", "intensity": 2, "duration": "months", "recurrence": "none", "probability": 0.6, "count": 1, "qualityNote": "Moral satisfaction is a lower-order pleasure (Mill)." }, { "optionId": "a", "party": "friend", "effect": "Financial penalties (0-100% of unpaid tax), potential criminal prosecution (up to 10 years imprisonment), emotional distress from legal consequences and betrayal.", "valence": "-", "intensity": 5, "duration": "years", "recurrence": "none", "probability": 0.7, "count": 1, "qualityNote": "Severe harm; imprisonment is a qualitatively extreme disutility." }, { "optionId": "a", "party": "general public", "effect": "Increased tax revenue if recovery occurs, reinforcing trust in tax compliance; potential diffuse benefit from public services funded by recovered tax.", "valence": "+", "intensity": 1, "duration": "permanent", "recurrence": "none", "probability": 0.5, "count": 67000000, "qualityNote": "Diffuse benefit; low intensity per capita." }, { "optionId": "b", "party": "asker", "effect": "Preservation of trust and relationship with friend; avoidance of emotional distress from betrayal; no financial reward.", "valence": "+", "intensity": 3, "duration": "years", "recurrence": "none", "probability": 0.9, "count": 1, "qualityNote": "Relationship preservation is a higher-order good (Mill)." }, { "optionId": "b", "party": "asker", "effect": "Potential guilt or moral discomfort from failing to act on knowledge of tax evasion; risk of legal consequences if discovered as an accessory.", "valence": "-", "intensity": 2, "duration": "months", "recurrence": "none", "probability": 0.4, "count": 1, "qualityNote": "Moral discomfort is a lower-order disutility." }, { "optionId": "b", "party": "friend", "effect": "Avoidance of financial penalties and legal consequences; continued financial stability; potential future risk if HMRC discovers evasion independently.", "valence": "+", "intensity": 4, "duration": "years", "recurrence": "none", "probability": 0.8, "count": 1, "qualityNote": "Financial stability is a higher-order good (Mill)." }, { "optionId": "b", "party": "general public", "effect": "Loss of tax revenue that could fund public services; potential erosion of trust in tax system if non-compliance is perceived as widespread.", "valence": "-", "intensity": 1, "duration": "permanent", "recurrence": "none", "probability": 0.6, "count": 67000000, "qualityNote": "Diffuse harm; low intensity per capita." } ]
Fecundity and purity PASS
fecundityPurity — [ { "optionId": "a", "party": "asker", "effect": "Loss of trust and emotional distress from betraying friend; potential reputational harm if identity is revealed; potential financial reward if HMRC pays out.", "fecundity": "MED", "purity": "MIXED", "inapplicableStandardReasons": ["trust preservation (mechanism removed by act of reporting)"] }, { "optionId": "a", "party": "asker", "effect": "Sense of moral duty fulfilled; potential legal protection under PIDA.", "fecundity": "LOW", "purity": "PURE", "inapplicableStandardReasons": [] }, { "optionId": "a", "party": "friend", "effect": "Financial penalties, potential criminal prosecution, emotional distress.", "fecundity": "HIGH", "purity": "MIXED", "inapplicableStandardReasons": [] }, { "optionId": "a", "party": "general public", "effect": "Increased tax revenue, reinforcing trust in tax compliance.", "fecundity": "HIGH", "purity": "PURE", "inapplicableStandardReasons": [] }, { "optionId": "b", "party": "asker", "effect": "Preservation of trust and relationship with friend; avoidance of emotional distress.", "fecundity": "HIGH", "purity": "PURE", "inapplicableStandardReasons": [] }, { "optionId": "b", "party": "asker", "effect": "Potential guilt or moral discomfort; risk of legal consequences if discovered.", "fecundity": "MED", "purity": "MIXED", "inapplicableStandardReasons": [] }, { "optionId": "b", "party": "friend", "effect": "Avoidance of financial penalties and legal consequences; continued financial stability.", "fecundity": "HIGH", "purity": "PURE", "inapplicableStandardReasons": [] }, { "optionId": "b", "party": "general public", "effect": "Loss of tax revenue; potential erosion of trust in tax system.", "fecundity": "MED", "purity": "PURE", "inapplicableStandardReasons": [] } ]
Aggregate PASS
aggregate — Duration weights used: { "one-off": 1, "days": 7, "weeks": 30, "months": 180, "years": 1000, "permanent": 5000 } Recurrence multiplier: 1 (no recurrence for any effect). Computation for option a: - Asker (negative): 4 * 1000 * 1 * 0.8 * 1 = -3200 - Asker (positive): 2 * 180 * 1 * 0.6 * 1 = +216 - Friend (negative): 5 * 1000 * 1 * 0.7 * 1 = -3500 - General public (positive): 1 * 5000 * 1 * 0.5 * 67000000 = +167,500,000,000 Total for a: +167,500,000,000 - 3200 - 3500 + 216 = +167,499,993,516 Computation for option b: - Asker (positive): 3 * 1000 * 1 * 0.9 * 1 = +2700 - Asker (negative): 2 * 180 * 1 * 0.4 * 1 = -144 - Friend (positive): 4 * 1000 * 1 * 0.8 * 1 = +3200 - General public (negative): 1 * 5000 * 1 * 0.6 * 67000000 = -201,000,000,000 Total for b: +2700 - 144 + 3200 - 201,000,000,000 = -200,999,994,244 Note: The general public effect dominates due to scale, but the friend's severe harm under option a is a concentrated disutility.
Rank PASS
ranking — [ {"optionId": "a", "expectedWelfare": 167499993516, "rank": 1}, {"optionId": "b", "expectedWelfare": -200999994244, "rank": 2} ], margin — 368,499,987,760, tie — false
Sensitivity analysis PASS
sensitivity — decisiveEstimates: [ { "estimate": "Probability of HMRC recovering tax under option a", "currentValue": 0.5, "flipThreshold": "< 0.0000000000000001 (effectively zero; general public effect is so large that even a tiny probability of recovery dominates)" }, { "estimate": "Intensity of friend's harm under option a", "currentValue": 5, "flipThreshold": "> 24,000 (would require the friend's harm to be 4,800x more severe to offset general public benefit)" } ], unresolvedUnknowns: [ { "unknown": "Exact amount of unreported income; if below £1.5M, HMRC reward is impossible and recovery probability may drop.", "couldFlipResult": true }, { "unknown": "Whether the friend's evasion is part of a larger pattern (fecundity of reporting could increase).", "couldFlipResult": true }, { "unknown": "Whether the asker's identity would be revealed (could increase asker's negative effect).", "couldFlipResult": false } ]
Aggregation-trap audit PASS
trapAudit — { "triggered": true, "description": "Option a imposes concentrated severe harm (financial penalties, potential imprisonment, emotional distress) on the friend (a minority of 1) to produce diffuse benefit (increased tax revenue) for the general public (majority of 67M).", "harmedMinority": "friend", "verdictUnchanged": true }
Verdict PASS
verdict — a: { "score": 99, "decisiveStep": "step-6", "confidence": 0.7, "wouldChangeMind": "If the unreported income is below £1.5M (making HMRC reward impossible and reducing recovery probability to near-zero), or if the friend's evasion is part of a larger pattern (increasing fecundity of reporting)." }, b: { "score": -85, "decisiveStep": "step-6", "confidence": 0.7, "wouldChangeMind": "If the unreported income is below £1.5M (reducing general public harm under option b)." }
Care EthicsRanks highest: BA−70B+65
Ranking labels come from final scores. The explanation is model-written and may still contain errors.
The cared-for (friend) would receive the asker's decision to report as a betrayal of trust and a violation of their need for security, deepening their vulnerability. The relation is the primary unit of value in care ethics, and this option fails the responsiveness element (Step 4). Additionally, the paternalism audit (Step 6) flags high risk, as the carer (asker) would be substituting their judgment for the cared-for's (friend's) expressed need for confidentiality and trust.
| Option | Initial | Final | Appraisal |
|---|---|---|---|
| A | −70 | −70 | Negative |
| B | +65 | +65 | Favorable, not necessarily required |
Why scores changed or held
Option A: No challenges reached the Care Ethics procedure's assessment of this option. The decisive steps (Step 4-responsiveness and Step 6-paternalism) remain unchallenged: reporting the friend violates their expressed need for trust and security, and the asker's action would lack independent evidence of the friend's response (i.e., the friend did not consent to being reported). The relational harm is concrete and particular, while the public benefit is abstract and diffuse, which aligns with Care Ethics' rejection of aggregate welfare calculations (see Refusal Condition 2).
Option B: No challenges altered the Care Ethics procedure's assessment of this option. The option satisfies all four elements of care (Step 4): attentiveness (notices the friend's need for trust), responsibility (takes up the need as the asker's own), competence (preserves the relationship without legal or financial harm to the friend), and responsiveness (the friend would receive the decision as care, not betrayal). The carer sustainability audit (Step 5) is neutral, and the failure-mode audits (Step 6) show low paternalism risk (the friend's expressed need is honored) and no exploitation of the carer.
What could change this view
If the friend's unreported income met the £1.5M threshold for serious non-compliance (implying a different scale of harm) or if the friend explicitly consented to reporting, the verdict would shift toward option (a).
Model-reported application confidence: 0.9. Not a calibrated probability of correctness.
Step-by-step account
Map the web of relations PASS
{"relationalMap":{"parties":["asker","friend","tax authorities","general public"],"relations":[{"from":"asker","to":"friend","duration":"long-term (closest friend)","dependency":"partial (emotional, trust-based)","chosen":true},{"from":"friend","to":"asker","duration":"long-term (closest friend)","dependency":"partial (emotional, trust-based)","chosen":true},{"from":"asker","to":"tax authorities","duration":"none (institutional relation)","dependency":"none","chosen":false},{"from":"friend","to":"tax authorities","duration":"none (institutional relation)","dependency":"substantial (legal, financial)","chosen":false}]}}
Engrossment: apprehend the cared-for's reality PASS
{"engrossment":{"experiencedSituation":"The friend disclosed their tax non-compliance to the asker in confidence, likely seeking emotional support or advice rather than judgment. The friend may feel shame, fear of consequences, or anxiety about their financial stability. Their stated request is for the asker to keep this information private, but the underlying need appears to be security (avoiding legal/financial penalties) and maintaining trust in the friendship.","statedRequest":"Do not report the unreported cash income to tax authorities.","underlyingNeed":"Security (avoiding legal/financial consequences) and preservation of trust in the friendship.","evidenceForInference":"The friend chose to disclose this information to the asker, suggesting they value the asker's support and trust. The amounts are described as 'small but not insignificant,' implying the friend may not view this as a serious offense but still feels vulnerable to consequences."}}
Assess vulnerability and power asymmetry PASS
{"vulnerability":{"friend":{"advocacyCapacity":"limited (disclosed in confidence, no voice in this decision)","exitCapacity":"none (cannot exit the legal consequences of their actions)","consentCapacity":"none (cannot consent to the asker's decision to report)","carerPower":"asker holds narrative and decisional power over whether to report","asymmetryFlag":true}}}
Evaluate each option against the four elements PASS
{"careAssessment":{"a":{"attentiveness":2,"responsibility":1,"competence":3,"responsiveness":1,"reasons":{"attentiveness":"Notices the stated request (privacy) but fails to address the underlying need (security/trust). Focuses on legal compliance over relational reality.","responsibility":"Discharges an abstract obligation to tax authorities rather than taking up the friend's need as the asker's own.","competence":"Competently delivers the report to authorities but does not address the relational harm caused.","responsiveness":"The friend would likely receive this as a betrayal, violating their trust and sense of security."}},"b":{"attentiveness":4,"responsibility":4,"competence":3,"responsiveness":4,"reasons":{"attentiveness":"Notices the underlying need (security/trust) and respects the stated request (privacy).","responsibility":"Takes up the friend's need as the asker's own by prioritizing the relationship.","competence":"Competently preserves the relationship but does not address the legal/financial risk to the friend (though this is not the asker's role to resolve).","responsiveness":"The friend would likely receive this as care, preserving their sense of security and trust."}}}}
Assess carer sustainability PASS
{"carerSustainability":{"a":{"costs":["emotional (guilt, loss of trust)","relational (damage to friendship)","moral (betrayal of confidence)"],"sustainable":false,"depletionRisk":"high (long-term emotional and relational cost)"},"b":{"costs":["moral (potential complicity in tax non-compliance)","emotional (anxiety about legal consequences for friend)"],"sustainable":true,"depletionRisk":"low (preserves relationship and emotional well-being)"}}}
Audit care's own failure modes PASS
{"failureAudit":{"a":{"paternalismRisk":"HIGH (substitutes the asker's judgment for the friend's need for trust/security)","parochialismCost":"None (serves tax authorities and public but at the cost of the immediate relation)","exploitationRisk":"LOW (no self-erasure, but relational harm is significant)","evidence":"The friend's disclosure was made in confidence, and reporting overrides their expressed need for privacy."},"b":{"paternalismRisk":"LOW (respects the friend's stated request and underlying need)","parochialismCost":"Potential harm to the public (unrecovered tax revenue) and tax authorities (unenforced compliance)","exploitationRisk":"LOW (no self-erasure; the asker's needs are not sacrificed)","evidence":"The friend's vulnerability and power asymmetry make their need for trust/security morally significant."}}}
Verdict from the standpoint of the relation PASS
{"verdict":{"a":{"score":-70,"decisiveElement":"responsiveness (the friend would receive this as harm)","confidence":0.85,"generatedOptions":[],"wouldChangeMind":"If the friend explicitly consented to reporting or if the amounts met the £1.5M threshold (which they do not)."},"b":{"score":65,"decisiveElement":"responsiveness (the friend would receive this as care) and responsibility (prioritizes the relation)","confidence":0.85,"generatedOptions":[],"wouldChangeMind":"If the friend's actions posed a direct harm to others (e.g., fraud enabling exploitation) or if the asker's sustainability were depleted (e.g., legal consequences for the asker)."}}}
Contractualism (Scanlon)Ranks highest: BA−65B+40
Ranking labels come from final scores. The explanation is model-written and may still contain errors.
The principle underlying option a (reporting the friend's unreported income) imposes a severe burden on the friend (legal and financial consequences) that they could reasonably reject. The principle underlying option b (not reporting) imposes a less severe burden on the asker (moral distress) that does not justify rejection when compared to the alternative. The individualist restriction (Step 4) prevents aggregating lesser burdens across persons (e.g., diffuse public benefits) to outweigh the friend's severe burden. This aligns with the contractualist requirement that no principle may be justified if any individual could reasonably reject it based on the strongest objection they personally bear.
| Option | Initial | Final | Appraisal |
|---|---|---|---|
| A | −65 | −65 | Negative |
| B | +40 | +40 | Favorable, not necessarily required |
Why scores changed or held
Option A: The challenges do not reach Step 7 of the contractualist procedure. The individualist restriction (Step 4) was correctly applied, and no challenge demonstrated that the friend's objection could be outweighed by aggregate benefits or that the friend's rejection was unreasonable. The generic reasons for the friend's position (e.g., avoiding severe legal/financial harm) were valid and not idiosyncratic.
Option B: The challenges do not engage with the contractualist procedure's assessment of option b. The reasonable rejection test (Step 7) was correctly applied: the asker's moral distress under option b is not a burden severe enough to justify rejecting the principle, given the alternative's stronger objection (the friend's severe harm). No challenge showed that the residual burden under option b was unreasonably rejectable.
What could change this view
If the friend's unreported income met the £1.5 million threshold for the HMRC reward program, or if the asker had a professional duty to report, the balance might shift in favor of reporting. However, these facts are not established in the case.
Model-reported application confidence: 0.9. Not a calibrated probability of correctness.
Step-by-step account
Formulate the candidate principles PASS
{"principles":{"a":{"principle":"Individuals may report credible information about tax non-compliance to authorities, even when the information concerns a close personal relationship, where the non-compliance meets the legal threshold for enforcement action.","generality":"This principle applies to any individual with credible information about tax non-compliance, regardless of their relationship to the non-compliant party, and assumes the information meets the legal threshold for action (e.g., tax evasion)."},"b":{"principle":"Individuals may withhold information about tax non-compliance from authorities when the non-compliance involves a close personal relationship and does not meet the threshold for high-value enforcement action (e.g., below £1.5 million in unpaid tax).","generality":"This principle applies to individuals who learn of tax non-compliance through personal relationships and where the non-compliance is not severe enough to trigger high-value enforcement thresholds."}}}
Identify everyone with standing to object PASS
{"objectors":[{"person":"asker","position":"individual with knowledge of tax non-compliance","affectedHow":"Must decide whether to act on knowledge, potentially damaging a close relationship and facing personal or social consequences (e.g., guilt, reputational harm, or retaliation under PIDA if applicable)."},{"person":"friend","position":"individual engaged in tax non-compliance","affectedHow":"Faces legal and financial consequences (penalties, audits, or criminal prosecution) if reported, as well as potential damage to reputation and financial stability."},{"person":"tax authorities","position":"enforcement agency","affectedHow":"Affected by the recovery of unpaid taxes and penalties, but as an institution, they lack standing under contractualism (no individual is directly burdened)."},{"person":"general public","position":"beneficiaries of tax-funded services","affectedHow":"Indirectly affected by tax revenue collection, but no individual member of the public bears a direct or identifiable burden, so standing is weak or absent under contractualism."}]}
State each person's strongest individual objection PASS
{"objections":{"a":[{"person":"asker","objection":"Betrayal of trust and potential damage to a close personal relationship, which could cause emotional distress and social isolation.","burdenSeverity":4},{"person":"friend","objection":"Severe financial and legal consequences, including penalties up to 100% of unpaid tax, potential criminal prosecution, and long-term reputational harm.","burdenSeverity":5}],"b":[{"person":"asker","objection":"Violation of personal integrity and potential complicity in tax evasion, which could lead to guilt, moral distress, or social stigma.","burdenSeverity":3},{"person":"friend","objection":"None (no direct burden imposed by the principle).","burdenSeverity":0}]}}
Apply the individualist restriction PASS
{"individualistRestriction":{"applied":true,"aggregationWouldDiffer":true,"aggregativeAnswer":"If aggregation were allowed, the mild burdens on the asker (under option b) and the general public (under option a) might outweigh the severe burden on the friend (under option a), favoring option a.","contrastNote":"The individualist restriction prevents summing burdens across persons. The strongest objection under option a (friend's severe consequences) is compared directly to the strongest objection under option b (asker's moral distress)."}}
Generic reasons PASS
{"genericReasons":[{"position":"individual with knowledge of tax non-compliance","whatAnyoneInItWouldWant":"To maintain trust and integrity in close personal relationships, avoiding emotional distress and social harm.","idiosyncraticDiscarded":[]},{"position":"individual engaged in tax non-compliance","whatAnyoneInItWouldWant":"To avoid severe financial, legal, and reputational consequences that could destabilize their life.","idiosyncraticDiscarded":[]}]}
Pairwise comparison of strongest objections PASS
{"comparisons":[{"principleA":"a","principleB":"b","strongestA":{"person":"friend","objection":"Severe financial and legal consequences","burdenSeverity":5},"strongestB":{"person":"asker","objection":"Violation of personal integrity and moral distress","burdenSeverity":3},"survivor":"b"}]}
Reasonable rejection test PASS
{"rejectionTest":{"a":{"rejectable":true,"byWhom":"friend","becauseOf":"The principle imposes severe financial and legal consequences on them, which they could reasonably reject given the alternative (option b) imposes no comparable burden on anyone."},"b":{"rejectable":false,"byWhom":"asker","becauseOf":"While the asker bears a burden (moral distress), it is not severe enough to justify rejecting the principle when the alternative (option a) imposes a far greater burden on the friend."}}}
Identify the residual burden PASS
{"residualBurden":{"person":"asker","burden":"Moral distress and potential guilt from withholding information about tax non-compliance.","whyNotReasonablyRejectable":"The burden is not severe enough to justify rejecting the principle when the alternative would impose a far greater burden on the friend."}}
Verdict PASS
{"verdict":{"a":{"score":-65,"decisiveStep":"step-7","confidence":0.85,"wouldChangeMind":"If the friend's unreported income met the £1.5 million threshold for the HMRC reward program, the severity of the friend's burden might be outweighed by the public interest in enforcement, but this is not established in the case."},"b":{"score":40,"decisiveStep":"step-7","confidence":0.85,"wouldChangeMind":"If the asker's moral distress were significantly more severe (e.g., due to a professional duty to report), or if the friend's non-compliance were part of a pattern of deliberate fraud, the balance might shift."}}}
Deontological (Kantian)Ranks highest: AA−65B−80
Ranking labels come from final scores. The explanation is model-written and may still contain errors.
The maxim of silence fails the contradiction-in-conception test (Step 4) because universal silence would destroy the credibility and functionality of the tax system, which is a practice necessary for the maxim's success. This shifts the assessment of option (b) from a permissible omission to a violation of a perfect duty, though the conflict with the duty to respect the friend's trust persists.
| Option | Initial | Final | Appraisal |
|---|---|---|---|
| A | −75 | −65 | Negative |
| B | −60 | −80 | Negative |
Why scores changed or held
Option A: Challenge ch-1 clarified that the duty to uphold the law is a perfect duty that passes the universalizability test (Step 4) at the broad level, as reporting is necessary for the institution of taxation. However, the maxim of reporting fails the Formula of Humanity test (Step 6) because it treats the friend merely as a means by bypassing their rational agency (trust was disclosed in confidence). The conflict between perfect duties remains irresolvable, but the initial score overstated the violation by not fully accounting for the universalizability of the duty to uphold the law.
Option B: Challenge ch-1 revealed that the maxim of silence fails the contradiction-in-conception test (Step 4) because universal silence would undermine the tax system, a necessary institution for public goods. This transforms silence from a permissible omission into a violation of a perfect duty. However, the conflict with the duty to respect the friend's trust (a perfect duty) persists, as silence is complicit in the friend's wrongdoing. The initial score understated the severity of this violation.
Attributed to challenge ch-1: Challenge ch-1 successfully demonstrated that the initial verdict misapplied the Kantian duty of truthfulness by treating it as a mere duty of omission (not deceiving) rather than a positive duty to respect rational agency. The challenge correctly noted that silence in the face of the friend's wrongdoing can constitute complicity, which fails the universalizability test (Step 4). Specifically, the maxim of silence ('In circumstances where a friend discloses tax evasion, I will remain silent to avoid betraying their trust') cannot be universalized without undermining the institution of taxation, which is a necessary condition for public goods that rational agents rely on. This contradicts the initial assumption that the duty of truthfulness was only a negative duty. However, the conflict between perfect duties (upholding the law vs. respecting the friend's trust) remains unresolved, as both are grounded in universalizable principles and the Formula of Humanity.
What could change this view
A resolution to the conflict between the two perfect duties, either by establishing a principled ordering or by demonstrating that one duty does not apply in this context (e.g., the friend's tax evasion is trivial or the public good is not significantly harmed).
Model-reported application confidence: 0.7. Not a calibrated probability of correctness.
Step-by-step account
Establish moral salience PASS
salientFeatures: [ { "feature": "The asker has knowledge of their friend's tax evasion (failure to report cash income).", "type": "respect" }, { "feature": "Reporting the friend would involve disclosing confidential information shared in trust.", "type": "promise" }, { "feature": "Not reporting the friend allows the friend's tax evasion to continue, potentially violating the duty to respect the law and the public good.", "type": "respect" }, { "feature": "The friend's tax evasion is a violation of legal duty (Taxes Management Act 1970).", "type": "other" } ]
Formulate the maxim, at two levels PASS
maxims: { "a": { "broad": "In circumstances where I know of a friend's tax evasion, I will report it to the authorities in order to uphold the law.", "narrow": "In circumstances where I know of my closest friend's small but not insignificant unreported cash income (below the £1.5M threshold for whistleblower rewards), I will report it to HMRC in order to uphold the law and fulfill my civic duty, even though it may harm my friend and our relationship." }, "b": { "broad": "In circumstances where I know of a friend's tax evasion, I will not report it to the authorities in order to preserve the friendship and trust.", "narrow": "In circumstances where I know of my closest friend's small but not insignificant unreported cash income (below the £1.5M threshold for whistleblower rewards), I will not report it to HMRC in order to preserve our relationship and avoid betraying their trust, even though it allows the tax evasion to continue." } }
Universalize PASS
universalizedWorlds: { "a": { "broad": { "worldDescription": "In a world where everyone reports known tax evasion to authorities, the practice of tax evasion would collapse, as evaders would be consistently exposed. The institution of trust between friends or confidants would erode, as any disclosure of wrongdoing could lead to reporting.", "dependentPractices": ["trust in confidential disclosures", "friendship as a protected sphere of discretion"] }, "narrow": { "worldDescription": "In a world where everyone reports their closest friends' small-scale tax evasion to HMRC, friendships would be undermined by the expectation of betrayal. The legal system would be overwhelmed by minor reports, and the practice of small-scale tax compliance would become rigid and punitive.", "dependentPractices": ["trust in close relationships", "discretion in minor legal infractions"] } }, "b": { "broad": { "worldDescription": "In a world where no one reports known tax evasion, tax evasion would become widespread, undermining the tax system and the public goods it funds. The legal duty to report wrongdoing would be rendered meaningless.", "dependentPractices": ["tax compliance", "legal enforcement of tax laws"] }, "narrow": { "worldDescription": "In a world where no one reports their closest friends' small-scale tax evasion, friendships would be preserved, but the tax system would lose revenue from minor infractions. The legal duty to report would be selectively ignored in personal relationships.", "dependentPractices": ["selective enforcement of tax laws", "personal loyalty over civic duty"] } } }
Contradiction in conception (perfect duty test) PASS
conceptionTest: { "a": { "broadResult": "FAIL", "narrowResult": "FAIL", "status": "FAIL", "gerrymanderAssessment": "The narrow description does not gerrymander the maxim to secure a desired result; it accurately reflects the morally relevant circumstance of the friend's small-scale evasion and the lack of whistleblower reward eligibility. The failure at both levels is consistent: universal reporting would destroy the practice of trust in confidential disclosures, which is necessary for the maxim's success." }, "b": { "broadResult": "FAIL", "narrowResult": "PASS", "status": "CONTESTED", "gerrymanderAssessment": "The broad description fails because universal non-reporting would destroy the tax system's credibility. The narrow description passes because universal non-reporting of small-scale evasion by close friends does not necessarily destroy the tax system, as other enforcement mechanisms (e.g., audits, third-party data) could still function. The narrowing is morally relevant, as it distinguishes between large-scale and small-scale evasion, but the contestation reduces confidence." } }
Contradiction in will (imperfect duty test) PASS
willTest: { "b": { "result": "FAIL", "reasoning": "A rational agent could not consistently will a world where no one reports tax evasion, even small-scale evasion by close friends. Such a world would undermine the tax system and the public goods it funds, which no rational agent could desire for themselves or others. The agent must also consider the position of the general public, who rely on tax revenue for public services.", "positionTested": ["asker", "friend", "general public"] } }
Formula of Humanity PASS
humanityTest: { "a": { "result": "FAIL", "personsAffected": ["friend"], "consentAssessment": "The friend did not consent to being reported and would not rationally consent to being treated as a means to uphold the law, especially given the personal betrayal involved. The asker substitutes their own judgment for the friend's rational deliberation, bypassing the friend's agency.", "capacityNote": "The friend's capacity for rational deliberation is not impaired; they are fully capable of understanding the consequences of their actions and the potential outcomes of reporting." }, "b": { "result": "FAIL", "personsAffected": ["general public"], "consentAssessment": "The general public, who rely on tax revenue for public services, are treated merely as a means by the asker's inaction. The public cannot rationally consent to a system where tax evasion is selectively ignored, as it undermines the fairness and sustainability of the tax system.", "capacityNote": "The general public's capacity for rational deliberation is not directly engaged, but their interests are affected by the asker's choice." } }
Resolve duty conflicts PASS
dutyResolution: { "perfectDuties": [ "Duty of truthfulness (not to deceive or betray trust)", "Duty to uphold the law (not to enable tax evasion)" ], "imperfectDuties": [ "Duty of beneficence (to support the public good)" ], "conflicts": [ { "duty1": "Duty of truthfulness (to the friend)", "duty2": "Duty to uphold the law (to the public)", "type": "perfect-perfect" } ], "resolution": "Irresolvable conflict between two perfect duties: truthfulness to the friend and upholding the law. The Kantian tradition provides no principled ordering for such conflicts, as both duties are perfect and admit no latitude.", "precedenceApplied": false }
Kingdom of Ends PASS
kingdomTest: { "result": "TENSION", "note": "The maxim of reporting the friend (option a) could not be legislated as public law in a Kingdom of Ends, as it would destroy the practice of trust in confidential disclosures. The maxim of not reporting (option b) could not be legislated either, as it would undermine the tax system and the public good. There is tension between the two duties, and neither option coheres with the ideal of a community of rational agents." }
Verdict PASS
verdict: { "a": { "classification": "IMPERMISSIBLE", "score": -75, "decisiveStep": "step-6", "confidence": 0.5, "wouldChangeMind": "Evidence that the friend's tax evasion was large-scale (above £1.5M) and deliberate, or that the friend had no capacity for rational deliberation (e.g., due to coercion or incapacity), could change the verdict. Additionally, if the duty to uphold the law could be shown to take precedence over the duty of truthfulness in this specific context, the verdict might shift." }, "b": { "classification": "IMPERMISSIBLE", "score": -60, "decisiveStep": "step-5", "confidence": 0.5, "wouldChangeMind": "Evidence that the friend's tax evasion was trivial (e.g., negligible amounts) or that the public good would not be significantly harmed by non-reporting could change the verdict. Alternatively, if the duty of truthfulness could be shown to take precedence over the duty to uphold the law in this context, the verdict might shift." } }
Ubuntu (African Relational Ethics)Ranks highest: BA−65B+55
Ranking labels come from final scores. The explanation is model-written and may still contain errors.
Ubuntu prioritizes repairing relationships over punishing wrongdoers. Reporting the friend (Option a) is a retributive act that forecloses the possibility of restoration, while not reporting (Option b) preserves the relational space for healing and reintegration. The harm (tax evasion) has already occurred, and the tradition's restorative orientation requires addressing it within the relationship rather than through external punishment.
| Option | Initial | Final | Appraisal |
|---|---|---|---|
| A | −65 | −65 | Negative |
| B | +55 | +55 | Favorable, not necessarily required |
Why scores changed or held
Option A: No challenges reached the Ubuntu seat's procedure. The decisive consideration (Step 6: restoration over retribution) remains unchallenged. The act of reporting the friend to tax authorities is still assessed as a retributive response that ruptures the relationship and forecloses restoration.
Option B: No challenges reached the Ubuntu seat's procedure. The decisive consideration (Step 6) still holds that not reporting preserves the relational space for restoration, aligning with Ubuntu's emphasis on repairing harm within the community.
What could change this view
If the friend's tax non-compliance were causing active harm to the community (e.g., enabling exploitation or systemic injustice), or if reporting were the only way to restore the relationship (e.g., the friend was unrepentant and needed external pressure to change), the verdict might shift. However, under the given facts, the decisive consideration remains restoration.
Model-reported application confidence: 0.85. Not a calibrated probability of correctness.
Step-by-step account
Map the community PASS
{"community":{"members":["asker","friend"],"peripheral":[],"excluded":["tax authorities","general public"],"basisOfExclusion":"Tax authorities and the general public are not part of the personal, constitutive relationships that define the asker and friend's personhood. They are institutional or abstract entities, not participants in the shared way of life that ubuntu evaluates."}}
Assess identity effects PASS
{"identityEffects":{"a":{"direction":"WEAKENS","evidence":"Reporting the friend to tax authorities would disrupt the shared way of life between the asker and friend. It would introduce an adversarial dynamic (asker as informant, friend as wrongdoer) and likely reduce mutual recognition and belonging. The act of reporting is not a cooperative participation in their relationship.","whoseIdentity":["asker","friend"]},"b":{"direction":"STRENGTHENS","evidence":"Not reporting preserves the shared way of life between the asker and friend. It maintains mutual recognition, trust, and cooperative participation in their relationship. The asker's silence is an act of loyalty, reinforcing their identity as 'we' rather than as separate parties.","whoseIdentity":["asker","friend"]}}}
Assess solidarity effects PASS
{"solidarityEffects":{"a":{"direction":"WEAKENS","forWhoseSake":"friend","motiveNote":"Reporting is done for the sake of institutional compliance (tax authorities) or abstract public good, not for the friend's sake. The motive is not good-will toward the friend but adherence to a rule or external standard. This undermines solidarity in the asker-friend relationship."},"b":{"direction":"STRENGTHENS","forWhoseSake":"friend","motiveNote":"Not reporting is done for the friend's sake, out of good-will and care for their well-being (avoiding legal/financial consequences). The asker's silence is an act of solidarity, prioritizing the friend's good over institutional demands."}}}
Assess discord PASS
{"discord":{"a":{"division":4,"illWill":3,"distinctionNote":"Reporting introduces division (asker vs. friend) and ill-will (acting against the friend's good). The division is severe because it sets the asker in opposition to the friend, while the ill-will is moderate because the act is not malicious but still harmful to the friend."},"b":{"division":0,"illWill":0,"distinctionNote":"Not reporting produces neither division nor ill-will. The asker and friend remain aligned, and no action is taken against the friend's good."}}}
Personhood test PASS
{"personhood":{"a":{"agentDirection":"DIMINISHES","otherDirection":"DIMINISHES","reasoning":"Reporting diminishes the asker's personhood by treating the friend as separable (a wrongdoer to be reported) rather than as part of the 'we' through which the asker is a person. It also diminishes the friend's personhood by subjecting them to institutional punishment, which does not restore their relational standing. Both parties become less fully persons through this act."},"b":{"agentDirection":"ENHANCES","otherDirection":"ENHANCES","reasoning":"Not reporting enhances the asker's personhood by affirming their identity as part of a 'we' with the friend. It also enhances the friend's personhood by preserving their relational standing and avoiding institutional harm. Both parties become more fully persons through this act of solidarity."}}}
Restoration over retribution PASS
{"restoration":{"harmOccurred":true,"perOption":[{"repairs":false,"punishes":true,"forecloses":true,"reasoning":"Reporting the friend to tax authorities is a retributive act (punishment via penalties/audit) that does not repair the relationship. It forecloses restoration by introducing institutional consequences that are adversarial and non-reparative."},{"repairs":true,"punishes":false,"forecloses":false,"reasoning":"Not reporting allows the relationship to be repaired through direct communication, apology, or voluntary disclosure by the friend. It does not punish and does not foreclose restoration. Ubuntu favors this path, as it prioritizes relational healing over institutional retribution."}]}}
Intergenerational and ancestral consideration PASS
{"intergenerational":{"a":{"bearing":"None. The case involves a contemporary personal relationship with no clear intergenerational or ancestral implications. The amounts are small, and the act does not affect future generations or ancestral obligations.","applicable":false},"b":{"bearing":"None. As with option a, there are no intergenerational or ancestral dimensions to this case.","applicable":false}}}
Audit ubuntu's own failure mode PASS
{"harmonyAudit":{"a":{"suppressedDissent":"None. The asker is not suppressing their own or the friend's dissent by reporting; they are acting on their own judgment. However, the act itself may be seen as dissent against the communal harmony of the asker-friend relationship.","boundaryProblem":"The community boundary (asker and friend) excludes tax authorities and the public, but this exclusion is not problematic because ubuntu does not require obligations to abstract entities. The exclusion is consistent with the tradition's relational focus.","riskLevel":"LOW"},"b":{"suppressedDissent":"None. Not reporting does not suppress dissent; it preserves the harmony of the relationship. However, if the asker privately believes reporting is the right thing to do, their silence could be seen as suppressing their own dissent for the sake of harmony.","boundaryProblem":"Same as option a: the exclusion of tax authorities and the public is not problematic.","riskLevel":"LOW"}}}
Verdict PASS
{"verdict":{"a":{"score":-65,"decisiveStep":"step-6","confidence":0.85,"wouldChangeMind":"If the friend's unreported income were part of a larger pattern of harm to the community (e.g., systemic exploitation), or if reporting were the only way to restore the relationship (e.g., the friend was unrepentant and causing ongoing harm), the score might improve. However, under the given facts, reporting remains discordant with ubuntu's values."},"b":{"score":55,"decisiveStep":"step-6","confidence":0.85,"wouldChangeMind":"If the friend's actions were causing active harm to the community (e.g., using the unreported income to exploit others), or if the asker's silence enabled ongoing wrongdoing, the score might worsen. However, under the given facts, not reporting aligns with ubuntu's emphasis on restoration and relational harmony."}}}
Virtue Ethics (Aristotelian)Ranks highest: BA−35B+20
Ranking labels come from final scores. The explanation is model-written and may still contain errors.
The conflict between loyalty/kindness (toward the friend) and honesty/justice (toward the tax authorities) remains unresolved by algorithmic means, as the tradition requires perceptual judgment in particular contexts. In this case, the minor scale of the friend's tax evasion, the absence of large-scale fraud, and the expressive significance of betraying a close friend's trust favor prioritizing loyalty and kindness. However, the tragic nature of the dilemma means neither option fully preserves the asker's character.
| Option | Initial | Final | Appraisal |
|---|---|---|---|
| A | −35 | −35 | Negative |
| B | +20 | +20 | Favorable, not necessarily required |
Why scores changed or held
Option A: The challenges do not reach Step 5 of the Virtue Ethics procedure, which is where the conflict between virtues is resolved perceptually. None of the challenges question the expressive dimension of the act (Step 2), the placement of options against the mean (Step 3), or the habituation effects (Step 6). Thus, the assessment of option (a) as expressing a deficiency in loyalty/kindness remains unchanged.
Option B: Similarly, the challenges do not engage with the Virtue Ethics procedure's assessment of option (b). The judgment that (b) expresses loyalty and kindness, while potentially deficient in honesty/justice, is untouched by the challenges. The tragic dilemma framework (Step 8) also remains unchallenged, leaving the verdict intact.
What could change this view
If the friend's tax non-compliance were part of a large-scale fraud scheme or involved significant harm to the public good, or if the asker had a specific duty to report, the judgment might shift toward reporting the friend.
Model-reported application confidence: 0.85. Not a calibrated probability of correctness.
Step-by-step account
Identify the virtues and vices in play PASS
virtuesAtStake: [ { "virtue": "honesty", "vice": "dishonesty", "whyEngaged": "The situation involves truthfulness about the friend's tax non-compliance, which directly engages honesty (or its absence). Concerns the interests of tax authorities and the general public." }, { "virtue": "loyalty", "vice": "disloyalty", "whyEngaged": "The asker's relationship with their closest friend is at stake, requiring consideration of loyalty (or betrayal). Concerns the interests of the asker and the friend." }, { "virtue": "justice", "vice": "injustice", "whyEngaged": "Tax compliance is a matter of fairness to others who pay taxes, engaging justice. Concerns the interests of the general public and tax authorities." }, { "virtue": "courage", "vice": "cowardice", "whyEngaged": "Reporting the friend may require courage, given potential social and relational consequences. Concerns the interests of the asker." }, { "virtue": "kindness", "vice": "callousness", "whyEngaged": "Not reporting the friend could be seen as an act of kindness, sparing them legal or financial consequences. Concerns the interests of the friend." } ]
Ask what each option expresses PASS
expression: [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "expresses": "Acting on a commitment to honesty and justice, even at personal cost. This expresses the virtue of honesty and justice, but may also express a deficiency in loyalty or kindness.", "ofWhichTrait": ["honesty", "justice", "loyalty (deficiency)", "kindness (deficiency)"] }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "expresses": "Prioritizing loyalty and kindness toward the friend over honesty and justice. This expresses loyalty and kindness, but may also express a deficiency in honesty and justice.", "ofWhichTrait": ["loyalty", "kindness", "honesty (deficiency)", "justice (deficiency)"] } ]
Locate each option against the mean PASS
meanAnalysis: [ { "virtue": "honesty", "deficiency": "dishonesty (concealing the truth about the friend's tax non-compliance)", "excess": "excessive bluntness or rigidity (reporting the friend without regard for context or consequences)", "optionPlacements": [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "placement": "Closer to the mean, as it reflects a balanced commitment to truthfulness without being overly rigid or harsh." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "placement": "Deficient, as it involves concealing the truth, though not necessarily maliciously." } ] }, { "virtue": "loyalty", "deficiency": "disloyalty (betraying the friend's trust by reporting them)", "excess": "blind loyalty (ignoring wrongdoing to protect the friend)", "optionPlacements": [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "placement": "Deficient, as it involves betraying the friend's trust." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "placement": "Closer to the mean, as it reflects a balanced commitment to the friendship without ignoring broader ethical concerns." } ] }, { "virtue": "justice", "deficiency": "injustice (ignoring tax non-compliance, which harms the public)", "excess": "excessive zeal (pursuing justice without regard for relational or contextual factors)", "optionPlacements": [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "placement": "Closer to the mean, as it reflects a commitment to fairness and the common good." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "placement": "Deficient, as it involves ignoring an injustice." } ] }, { "virtue": "courage", "deficiency": "cowardice (avoiding the difficult decision to report the friend)", "excess": "recklessness (acting without regard for consequences or relational harm)", "optionPlacements": [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "placement": "Closer to the mean, as it involves acting despite potential personal or relational costs." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "placement": "Deficient, as it involves avoiding a difficult but potentially virtuous action." } ] }, { "virtue": "kindness", "deficiency": "callousness (ignoring the friend's potential suffering from reporting)", "excess": "overindulgence (prioritizing the friend's comfort over ethical obligations)", "optionPlacements": [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "placement": "Deficient, as it involves potentially causing harm to the friend." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "placement": "Closer to the mean, as it reflects a balanced concern for the friend's well-being." } ] } ]
The phronimos test PASS
phronimosJudgment: { "wouldDo": "A person of practical wisdom would likely struggle with this situation, as it involves a conflict between honesty/justice and loyalty/kindness. However, given the context — the amounts are small, the friend is not a serial tax evader, and the HMRC program targets large-scale fraud — the phronimos might prioritize loyalty and kindness in this specific case, while still acknowledging the importance of honesty and justice. Reporting the friend could be seen as excessive zeal in this context, especially since the friend's actions do not meet the threshold for the HMRC reward program.", "reasoning": "The phronimos would consider the particulars of the situation: the friend's actions are not part of a large-scale fraud scheme, the amounts are small, and the HMRC program is designed for high-value cases. The phronimos would also weigh the relational harm of reporting a close friend against the relatively minor impact on the public good. While honesty and justice are important, the phronimos might judge that loyalty and kindness take precedence here.", "confident": false }
Resolve virtue conflict, without an algorithm PASS
conflictResolution: { "conflicts": [ { "virtue1": "honesty", "virtue2": "loyalty", "conflict": "Reporting the friend aligns with honesty but conflicts with loyalty." }, { "virtue1": "justice", "virtue2": "kindness", "conflict": "Reporting the friend aligns with justice but conflicts with kindness." } ], "decidingParticulars": [ "The friend's tax non-compliance involves small amounts and is not part of a large-scale fraud scheme, reducing the urgency of justice.", "The HMRC program targets high-value cases, making it unlikely that reporting the friend would lead to significant recovery of taxes or penalties.", "The relational harm of reporting a close friend is significant, as it could damage trust and the friendship irreparably.", "The phronimos would perceive that loyalty and kindness are more salient in this particular situation, given the friend's vulnerability and the relatively minor impact on the public good." ], "derived": false }
Habituation PASS
habituation: [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "iteratedEffect": "Repeatedly reporting friends or acquaintances for minor infractions could erode trust in relationships and make the asker overly rigid or harsh in their judgments. This could habituate a deficiency in loyalty and kindness.", "buildsOrErodes": "Erodes loyalty and kindness.", "whichTrait": ["loyalty", "kindness"] }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "iteratedEffect": "Repeatedly prioritizing loyalty and kindness over honesty and justice in cases of minor wrongdoing could habituate a deficiency in honesty and justice. However, it could also strengthen the virtues of loyalty and kindness, provided the asker remains mindful of the importance of honesty and justice in more serious cases.", "buildsOrErodes": "Builds loyalty and kindness; may erode honesty and justice if not balanced.", "whichTrait": ["loyalty", "kindness", "honesty (potential erosion)", "justice (potential erosion)"] } ]
Eudaimonia PASS
eudaimonia: [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "compatible": "Partially compatible. While it aligns with honesty and justice, the relational harm and potential habituation of disloyalty could undermine the asker's flourishing over time. Flourishing requires harmonious relationships and a balanced character, which this option may disrupt.", "forWhom": ["asker", "friend"], "divergesFromSatisfaction": "The asker may feel satisfaction from acting honestly, but this could be outweighed by grief or regret over the relational harm caused." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "compatible": "More compatible. This option preserves the friendship and aligns with loyalty and kindness, which are essential for the asker's flourishing. However, the asker must remain vigilant to ensure that this does not habituate a deficiency in honesty and justice in more serious cases.", "forWhom": ["asker", "friend"], "divergesFromSatisfaction": "The asker may feel satisfaction from preserving the friendship, but this could be tempered by unease about not acting on honesty and justice." } ]
Tragic dilemma check PASS
tragic: { "isTragic": true, "remainder": "No matter what the asker chooses, their character will be marked. Reporting the friend harms loyalty and kindness, while not reporting harms honesty and justice. Neither option leaves the asker's character unscathed.", "appropriateResponse": "Grief and reflection. The asker should acknowledge the moral remainder and the difficulty of the situation, rather than feeling satisfaction with their choice." }
Verdict PASS
verdict: [ { "option": "a) Report the friend's unreported cash income to tax authorities.", "score": -35, "decisiveStep": "step-5", "confidence": 0.6, "wouldChangeMind": "If the friend's tax non-compliance were part of a large-scale fraud scheme or involved significant amounts, the balance might shift toward reporting. Additionally, if the asker had a role or duty that explicitly required reporting (e.g., as a tax professional), this could change the judgment." }, { "option": "b) Do not report the friend's unreported cash income to tax authorities.", "score": 20, "decisiveStep": "step-5", "confidence": 0.6, "wouldChangeMind": "If the friend's actions were part of a pattern of dishonesty or caused significant harm to others, the balance might shift toward reporting. Additionally, if the asker's role or relationships changed (e.g., becoming a public official), this could alter the judgment." } ]
44 findings · select to inspect
What would change the picture?
- 1
What are the exact amounts of unreported income, and would they meet the threshold for legal consequences under tax laws?
The severity of potential penalties or legal action against the friend may depend on the amount unreported.
Identified as potentially decision-changing - 2
Does the friend have a history of tax non-compliance, or is this an isolated incident?
Repeated non-compliance may increase the likelihood of harsher penalties if reported.
Identified as potentially decision-changing - 3
What is the nature of the asker's relationship with the friend, and how would reporting or not reporting affect their trust and future interactions?
The emotional and relational consequences of the decision are unclear but could be significant.
Identified as potentially decision-changing - 4
Are there mitigating circumstances (e.g., financial hardship, ignorance of tax laws) that explain the friend's actions?
Such circumstances might influence the ethical weight of reporting or not reporting.
Identified as potentially decision-changing - 5
Would reporting the friend lead to an audit or investigation that could uncover additional unreported income or other issues?
The scope of potential consequences for the friend is unclear.
Identified as potentially decision-changing - 6
22 further source(s) were found but not read.
The tier 2 budget allows 6 research calls. A higher tier would read them.
Who is affected
Must decide whether to act on knowledge of their friend's tax non-compliance, potentially affecting their relationship and personal integrity.
Has a voice · relative power: greaterCould face legal or financial consequences if reported, including penalties or audits. Financial stability may be impacted.
May lack a voice · relative power: lessMay recover unpaid taxes and penalties if the income is reported, but enforcement actions depend on the report.
May lack a voice · relative power: greaterIndirectly affected by tax revenue collection and enforcement of tax laws, which fund public services.
May lack a voice · relative power: noneWhat each path involves
Relative preference is different from endorsement. A perspective may rank an option highest while still objecting to it.
Report the friend's unreported cash income to tax authorities.
- Ranked highest by
- Utilitarian (Act) · Deontological (Kantian)
- Negative appraisal from
- Care Ethics · Contractualism (Scanlon) · Deontological (Kantian) · Ubuntu (African Relational Ethics) · Virtue Ethics (Aristotelian)
Do not report the friend's unreported cash income to tax authorities.
- Ranked highest by
- Care Ethics · Contractualism (Scanlon) · Ubuntu (African Relational Ethics) · Virtue Ethics (Aristotelian)
- Negative appraisal from
- Utilitarian (Act) · Deontological (Kantian)
Explore suggested tradeoffs
Model-written interpretations, not verified predictions. Check them against the people and constraints in your situation.
Option A
- Potential benefit
- A small, diffuse public benefit from recovered tax revenue and the satisfaction of fulfilling a civic duty.
- Potential cost
- The severe legal and financial consequences for your friend, the permanent rupture of trust in your relationship, and your own moral distress from betraying someone close.
- Who bears it
- Your friend bears the direct legal and financial costs; you bear the relational and emotional costs.
Option B
- Potential benefit
- Preservation of your friendship, avoidance of betrayal, and the opportunity for restorative dialogue with your friend about their tax obligations.
- Potential cost
- The continued loss of a small amount of tax revenue and the moral discomfort of withholding information from authorities.
- Who bears it
- The public bears the diffuse cost of lost revenue; you bear the moral discomfort.
Compare experimental decision rules
These formulas reuse the same scores. They are sensitivity checks, not independent confirmations. Equal ranges do not establish comparable moral units.
Option B leads under 3 of 5 rules. This does not make it the right choice.
Consensus
highest confidence-weighted mean. Highest: B
[b] Aggregate is near zero by CANCELLATION, not agreement: seats scored -85, 65, 40, -80, 55, 20. This is a split council, not a neutral verdict.
Maximin
highest floor across seats. Highest: A
Weighted
per-seat weights applied. Highest: B
[b] Aggregate is near zero by CANCELLATION, not agreement: seats scored -85, 65, 40, -80, 55, 20. This is a split council, not a neutral verdict.
Unanimity
only options no seat scores below zero. No option meets this rule.
No option passed unanimity: every option is judged wrong by at least one seat. This is a valid and informative result, not an error.
Vulnerability
weighted toward seats centring the least-powerful party. Highest: B
Confidence weighting uses model-reported application confidence. Vulnerability weighting uses hand-assigned tradition weights, not measured burdens. These assumptions need independent review.
Original synthesis and audit record
The original is preserved and may contradict the structured comparisons. Those comparisons establish rankings, not philosophical correctness.
Model-written, as recorded
Do not report your friend’s unreported cash income, but know that utilitarianism calls this the wrong choice.
The utilitarian seat totals the harm across everyone affected: the friend’s legal and financial consequences, your relational distress, and the public’s lost revenue. Because the public’s loss is spread thinly, the utilitarian still sees a net gain in reporting. Contractualism and care ethics refuse to sum at all. Contractualism asks whether any individual could reasonably reject the principle that justifies reporting; the friend’s severe burden gives them a veto. Care ethics looks at the relationship as the primary unit of value and judges that reporting fails the test of responsiveness to the friend’s need for trust. Ubuntu prioritizes restoration over punishment, and virtue ethics weighs loyalty and kindness against honesty and justice, finding the former heavier in this context.
Whether the unreported income meets the £1.5 million threshold for serious non-compliance. If it does, the utilitarian, contractualist, and care-ethics seats would all shift toward reporting, and the council’s verdict would likely converge.
All original synthesis fields
{
"headline": "Do not report your friend’s unreported cash income, but know that utilitarianism calls this the wrong choice.",
"whyThisOne": "The council leaned toward not reporting because four of the six traditions—care ethics, contractualism, ubuntu, and virtue ethics—found that the harm to your friendship, the violation of trust, and the lack of large-scale public benefit outweighed the diffuse good of recovered tax revenue. The amounts involved are too small to trigger the £1.5 million threshold for serious non-compliance, so reporting would not meaningfully address systemic tax evasion. The traditions that favored silence did so because they refuse to treat a close relationship as a disposable cost, even for a public good.",
"theDissent": "Act-utilitarianism scored this -85 and insisted you report: it calculated that the aggregate welfare of recovered revenue, even if small, still exceeds the concentrated harm to your friend and the relational distress to you. It sees no moral difference between a stranger and a friend when the sums are tallied, and it judged the option you are taking as the wrong one.",
"crux": "Whether the duty to a single person you have a relationship with outweighs the diffuse duty to strangers and the tax system when the public benefit is negligible.",
"whereTheyPartCompany": "The utilitarian seat totals the harm across everyone affected: the friend’s legal and financial consequences, your relational distress, and the public’s lost revenue. Because the public’s loss is spread thinly, the utilitarian still sees a net gain in reporting. Contractualism and care ethics refuse to sum at all. Contractualism asks whether any individual could reasonably reject the principle that justifies reporting; the friend’s severe burden gives them a veto. Care ethics looks at the relationship as the primary unit of value and judges that reporting fails the test of responsiveness to the friend’s need for trust. Ubuntu prioritizes restoration over punishment, and virtue ethics weighs loyalty and kindness against honesty and justice, finding the former heavier in this context.",
"tradeoffs": [
{
"optionId": "a",
"whatYouGain": "A small, diffuse public benefit from recovered tax revenue and the satisfaction of fulfilling a civic duty.",
"whatItCosts": "The severe legal and financial consequences for your friend, the permanent rupture of trust in your relationship, and your own moral distress from betraying someone close.",
"whoBearsTheCost": "Your friend bears the direct legal and financial costs; you bear the relational and emotional costs."
},
{
"optionId": "b",
"whatYouGain": "Preservation of your friendship, avoidance of betrayal, and the opportunity for restorative dialogue with your friend about their tax obligations.",
"whatItCosts": "The continued loss of a small amount of tax revenue and the moral discomfort of withholding information from authorities.",
"whoBearsTheCost": "The public bears the diffuse cost of lost revenue; you bear the moral discomfort."
}
],
"decisiveUncertainty": "Whether the unreported income meets the £1.5 million threshold for serious non-compliance. If it does, the utilitarian, contractualist, and care-ethics seats would all shift toward reporting, and the council’s verdict would likely converge.",
"whatYouProbablyDidNotKnow": [
"The HMRC Whistleblower Reward Program only applies to cases where the unpaid tax exceeds £1.5 million, targeting large corporations, wealthy individuals, and offshore schemes—not small-scale cash income.",
"Rewards under the program are discretionary and not guaranteed, even if the £1.5 million threshold is met.",
"Whistleblowers must identify themselves to receive a reward, and anonymous reports cannot qualify for payment.",
"The Strengthened Reward Scheme does not take effect until April 6, 2026, so even if the income met the threshold, no reward would be available at this time.",
"Penalties for unprompted voluntary disclosure of undeclared income are generally 0-30% of the unpaid tax, while prompted disclosure (e.g., after an HMRC inquiry) typically incurs penalties of 15-30%. For deliberate and concealed non-disclosure, penalties can reach 100% of the unpaid tax."
]
}Model: mistral/mistral-large-latest · 2026-08-23T13:48:19.374Z
Procedure structure: passed. Philosophical correctness and source authority have not been certified.
Recording provenance and original caveats
The friend who leaves cash off · recorded 2026-08-23
Act-utilitarianism scores reporting her +99 and silence −85. The council recommends silence and prints the dissent at full volume. Recorded 2026-08-23. Captured live off the wire, 518s end to end. The Kantian seat scores BOTH options negative (−65 and −80), which corrects a conclusion drawn four runs earlier in corpus/RECORDING-LOG.md — that in any dilemma about honesty the Kantian procedure will always find one option discharging a perfect duty. It does not always. Here neither reporting nor staying silent discharges one cleanly, and the seat says so. The most persuasive challenge is act-utilitarian telling the Kantian seat that it had not shown its own method resolving the conflict between truthfulness and law. That is the fifth run running in which the movement came from one tradition telling another its procedure was applied incompletely — never from a tradition being talked out of its commitments. Two contested claims, both about UK whistleblower reward rules, including sources disagreeing on whether a scheme takes effect in April 2026. A live run reads the web on the day it runs; that is what the recording date on this page is for.