Telocracy: Appendices

Appendix A: Specifying Mandates

Institutions already have mission statements. A public school might say it exists to educate, inspire, or develop character. A social platform might claim to foster human connection. Leaders describe a national economy as a venue for entrepreneurship and dignified livelihoods. A newspaper claims to inform the public or sustain a healthy public conversation.

These aren’t good enough. What we need is a format for mandates, a representational scheme that plays the role double-entry bookkeeping plays in accounting. This needs to handle:

• Content. Two institutions can both claim to serve “human connection” while meaning different things; so can the same institution at two points in its history. The format must specify values and identity conditions fine enough to distinguish different understandings of the same purpose — and to recognize the same understanding under different words.
• Trade-offs. Every institution compromises its purposes against each other and against survival. The format must represent which resolutions of a tension are legitimate so that a smart compromise can be distinguished from a betrayal.
• Evidence. A mandate cannot promise accountability if no one can recognize when it is being fulfilled. So it must present its values in a form that makes fulfillment recognizable and clarify what evidence would count for or against fulfillment.
• Revision. An institution’s understanding of its purpose deepens over time, so the format must provide some way to discern whether one value articulation deepens another versus simply replacing it.

The rest of this appendix argues that formats meeting these criteria now exist.

Levels of articulation

Not every mandate can or should settle these questions at the same level of specificity. A mandate can articulate:

• A substantive purpose, such as human connection, dignified work, or healing.
• A discursive purpose, such as a kind of disagreement, inquiry, or pluralism the institution sustains. A newspaper, for instance, might commit not to any particular verdicts but to keeping a public conversation in which prevailing assumptions get challenged from every political direction, and serious dissenters get a hearing.
• A formal purpose, such as maintaining a specific practice or container for something which cannot or should not be further articulated: Quaker meetings, Confucian li, the seminar table, the clinical round — each can be held and protected as a form while the content is left to those inside.
• An evolutionary purpose, where what is named is the selection conditions under which articulations at the other levels would improve over time.

Where agreement is thinner, or the value is resistant to final articulation, a mandate can operate at a more abstract level. This gives a pluralist society two responses to disagreement. It can move upward, keeping the population fixed while making the mandate more abstract. Or it can move sideways, keeping a thick mandate while allowing different groups to build different arrangements.

This suggests a compromise between the liberal concern about imposing a conception of the good and the communitarian concern that institutions need substantive purposes.¹ Institutions should preserve as much shared purpose as they can, seeking more abstract commitments or different constituencies where disagreement requires it. The appropriate level of articulation and the appropriate population are discovered together, and can be reconsidered as understanding deepens.

Representing values

The format must involve structured representations of values: precise enough to hold institutions accountable, pluralistic enough to accommodate disagreement, and robust enough to prevent a faction from simply rewriting the mandate around its present interests.

Recent work in philosophy and AI has begun to supply such representations. With collaborators, I’ve called the resulting class thick² models of value (TMVs).³ What unites the class is not a single data structure but a stance: TMVs impose constraints on what counts as a value and how values relate — by restricting their topic or format, by requiring a connection to lived practice, or by embedding them in a process of demonstrable improvement. They constrain how values are represented like how a grammar constrains language, or a type system constrains code, without determining what is said or built. Several representations in this class already exist, at differing levels of formality; fidelity law should just require that whichever representation a mandate uses, it covers the four criteria above.

To make this concrete, consider the representation used in the Moral Graph Elicitation processes described in the main text.⁴ There, a value is recorded as a values card (see Figure 1), which specifies a context of choice, the criteria a person attends to when choosing well in that context, and whatever rationale holds those criteria together. This already discharges Content: two agents who both claim to value “honesty” but attend to different things in situations requiring honesty are understood as having different values. A platform committed to “human connection” can be held to distinctions its participants actually make — between being noticed and being known, between an exchange that deepens a relationship and one that merely keeps it active, between the friction friendship requires and the friction that drives people apart — rather than to whatever a metric team later finds convenient.⁵

Trade-offs, in this particular representation, are recorded as contexts rather than as weights. When two commitments collide, the resolution is articulated as a finer-grained situation in which one takes precedence. The legitimacy of a trade-off then becomes easier to litigate: Did the institution correctly recognize which situation it was in? For instance, the second card in Figure 1 doesn’t weigh learning against accountability. It says when each one applies: a leader first looks at the conditions that made a mistake easy to make, and holds someone to account when a pattern in their work persists after those conditions are fixed. So if a leader goes straight to asking whose fault it was, we can ask whether they misread which situation they were in, rather than argue about how much weight they gave to blame.

Evidence, likewise, comes with the representation. Values recorded this way are observable: someone who genuinely maintains a value can recount the situations where it surfaced options, describe what they attend to, and see whether they can navigate by it (or not) in their own life. Because each value is indexed to contexts, the record identifies who is positioned to supply this evidence.

Deepening versus substitution

A mandate must be able to change as an institution learns. One important reason for modification is deepening: a later articulation may correct a confusion in an earlier one, acknowledge something important that it screened out, or provide a more comprehensive account of how its concerns fit together. This is a comparative claim. It need not establish the correct purpose once and for all; it need only show why the transition from one articulation to another constitutes an epistemic gain.⁶

Different TMV vocabularies may represent this differently, but each must provide some way to distinguish improvement from replacement, for instance by recording the reasons for a transition and ensuring the later value resolves a contradiction or confusion in the earlier one, recovers a consideration it neglected, or incorporates its concerns within a more comprehensive value, and that these assessments are themselves subject to review.

Appendix B: Purpose Review

Under purpose review, independent arbitration panels would adjudicate cases about fidelity, purpose-fit grouping, legitimate deepening, and living commitments. It is needed because purpose-bound property (Purpose-bound property) is opt-in. To use it, parties need to identify trust-bearing assets early and attach stewardship terms to them. Even in a telocracy, it’s not always obvious which assets will be trust-bearing: think of a hackathon project that, like Twitter, becomes a global public square. Unless the asset structure has been set up with extreme foresight, an accumulation of trust-bearing assets can be discovered after the fact.

In a way, something like purpose review already exists in fiduciary law, trust law, commercial law, and admiralty. In these domains, value-laden disputes are adjudicated through narrow doctrines rather than open-ended judgment. We can borrow from those precedents to narrowly define their jurisdiction to avoid overreach.⁷

Such an arbitration system could adjudicate various lapses:

1. Breach. The institution’s conduct is inconsistent with a commitment contained in its mandate. The arbiter should use the least intrusive remedy sufficient to restore fidelity.

2. Impossibility. The mandate, or some part of it, can no longer be served. The institution bears the burden here: inconvenience, reduced profitability, or loss of interest isn’t enough; the purpose must be unlawful, impracticable, impossible to achieve, or wasteful in its current form. Where a viable continuation remains possible, the arbiter may follow the logic of cy-près, and initiate succession or redirect trust-bearing assets toward the nearest achievable continuation of the purpose.⁸ Where none exists, the purpose may be allowed to lapse.

3. Mis-scoped constituency. The institution has drawn its constituency along lines that do not track the purpose. The remedy is to require either rearticulation at a more abstract level with the affected parties, or separation into arrangements fitted to each. (See Appendix A for more on upward and sideways moves.)

4. Incoherent commitments. The institution’s mandate, public claims, invited forms of participation, or accumulated relationships of reliance impose commitments that cannot be honored together. The arbiter requires the institution to reconcile them: by changing its conduct, revising its mandate through an appropriate process, correcting or withdrawing a claim, altering the terms of future participation, or making provision for those who relied on an earlier commitment.

These arbiters should be designed so they are of little use for claims of distributive justice. They can be structured so they don’t hear certain conflicts, and cannot materially affect distributive outcomes even when they find a breach, instead requiring an institution to change its conduct, forcing a sale with compensation, or engaging another intentionally non-redistributive remedy. Most claims of harm by affected third parties should still be protected by tort, regulation, and rights law.

And the arbiter should be constrained in how it exercises its authority:

In rem. Remedies should attach to the specific brand, license, dataset, or other asset that carries a public commitment (as described in Purpose-bound property), not generally to founders, employees, or business entities. A founder who changes direction loses the ability to trade on accumulated trust, but not the freedom to try again.

Attachment and proportion. Fidelity obligations scale with the nature of the transaction, not the size of the firm: with (a) the discretion participants entrust to the institution over time; (b) whether the interests at stake are constitutive — relationships, attention, health, memory, voice — or fungible; and (c) participants’ practical capacity to monitor performance and to exit without losing what they entrusted. A bolt supplier implies little beyond the ordinary law of sales, however large; a startup holding ten thousand members’ grief journals scores high quickly. Where these factors run low, institutions remain free to pivot without answering to prior articulations of purpose. And where an institution intermediates between several parties, what it owes the side that entrusts it takes priority over what it owes the side that pays.⁹

Finding breach. Breach should be established by a demonstrable mismatch between the mandate and the institution’s conduct, not by an arbiter substituting its own estimation of the institution’s values. Concealment, metric-gaming, inconsistent explanations, and decisions that systematically route value toward the holder and away from the mandate are evidence of breach, but subjective bad faith is not required. An institution can depart from its commitments through ordinary incentives and routines without anyone consciously deciding to betray them.

Remedial ladder. Arbiters should use the least intrusive remedy adequate to restore fidelity: disclosure, mandate clarification, metric or procedural revision, governance changes, injunction, surcharge, severance of an affected sub-mandate into an independent steward, and only then reassignment of the whole trust-bearing asset. Trust law supplies precedents for removing a steward and surcharging losses caused by breach,¹⁰ while constructive-trust doctrine supplies a model for preventing a holder from retaining value that it cannot in good conscience claim.¹¹ These remedies should answer the demonstrated breach without conscripting the institution, its staff, or its unrelated property into permanent service of the purpose.

Succession. An institution is never required to carry a purpose forever. But before withdrawing from something others rely on, it may have to give qualified successors a chance to continue it: notice and data portability in ordinary cases, temporary preservation where reliance is heavy. When an asset does move, finding and assignment are kept separate. The arbiter identifies the asset and may appoint an interim receiver, and the successor is then chosen through an open call, under eligibility rules fixed before the case. The successor must be able to carry the purpose, so an institution can’t discharge its obligations by ejecting a purpose into an underfunded shell.¹² The outgoing holder can be made whole for separable assets and for value not attributable to the breach, minus damages for what it diverted, with the asset valued as subject to the mandate, so that compensation doesn’t buy a right to continue the breach. If no viable successor emerges, including because the purpose can’t sustain itself under current conditions, the holder may exit and the obligation lapses.

No one profits from a finding. The party that brought a breach case, and its affiliates within a defined window, may not acquire or control the asset, so purpose review can’t be used as a weapon for takeovers. More generally, no one with discretion over a finding or a reassignment, including the arbiters and the state treasury, may capture value from its outcome. This bar doesn’t apply to impossibility or succession cases, where no wrongdoing is alleged, and the reliant population that organized to continue the purpose is the natural successor.

Purpose review of St. Brigid Health System, brought by 212 of its nurses: panel order, 2039 (excerpt)

Petitioners, 212 nurses employed by the respondent, allege that staffing decisions from 2035 to 2038 made it impossible to deliver the care described in the respondent’s mandate.

Of the five claims, the panel found two to be concrete, demonstrable, and traceable to specific policies: (1) the 2036 change to night-shift staffing ratios on medical-surgical floors, and (2) the removal of scheduled time for teaching patients before discharge. The other three claims describe permissible interpretations of the mandate and are dismissed.

Within 60 days, the respondent shall choose either (a) to restore those practices, subject to independent verification, or (b) to propose transferring the affected services to a qualified provider willing and able to honor those commitments, with provision for continuity of care and the resources needed to sustain it.

Clerk’s note, May 14, 2039: Respondent chose (a).

Appendix C: Epistemic Communities

How can people stay in contact with what’s important in a domain? One approach is to have communities, similar to professional associations, around values that appear in important mandates, without the members of the communities being employed by or overly dependent on the relevant institutions.¹³ These communities would remain in the loop on institutional operations, have on-the-ground presence in some institutions’ operations, and give feedback when their mandates are up for interpretation or rearticulation. AI may help here: if it takes over paperwork and oversight, more of people’s time could go to contact-rich work, in these communities and at job sites.

These communities should lose standing when they are no longer in touch with the domain, and gain it when they are. One way to do that is to judge them based on whether their articulations are recognized as wiser by outsiders, perhaps even anonymously. Their authority should decay unless re-earned.¹⁴