The Reciprocity Manifesto

Principles for deciding who counts, offered for agreement, disagreement and signature.

By Geoff Fane. August 2026. For signature — draft for discussion.

Before you begin

This manifesto is the distilled form of a longer argument. If you want the reasoning before you decide whether to sign, start with either of these:

• Who Counts: The Short of It — the compressed argument, about fifteen minutes. The boy who cannot vote, the four people we decided did not count, why personhood was never about souls, and the elephant in the mirror.
• Who Counts: The Long of It — the full essay in twenty-two sections with twenty-four primary sources: the legal history worked through in detail, the three modern failures, and what a charter would actually say.

The manifesto below states the conclusions. The essays show the working.

Preamble

Every society has drawn a line around who counts, and every society has drawn it wrongly at least once. Slaves stood outside it. Women stood outside it. Whole peoples were held not to inhabit the land they had lived on for forty thousand years. In each case the line was defended on grounds that sounded like principle, and in each case the criterion turned out, on inspection, to select for resemblance to the people applying it.

We are approaching that question again, from more than one direction at once: in the systems we are building, in the animals we already share the world with, and possibly in something we have not met yet. We would rather answer it deliberately, now, while it costs us nothing, than in a hurry later, when it will cost a great deal.

This is not a claim that any particular thing counts today. It is a claim about how the question should be settled, and by what rule, and about the fact that the rule we write down will one day be applied to us.

What we value

In deciding who holds standing, and what standing means, we have come to value:

What a being can do, over what a being is made of.

Commitments that can be enforced, over assurances that cannot.

Reasons given, over decisions taken.

Care that goes beyond the minimum, over compliance that meets it.

The items on the right are not worthless. They are simply not enough, and every failure catalogued above was a case of someone settling for them.

The Principles

1. Membership is not settled by ancestry.

In practice, almost everywhere, the thing that decides whether you hold rights at all is having been born human. That is not a test. It is a fact about descent, and it is carrying enormous moral weight while offering no justification. Every capacity test we actually use is applied to beings whose membership was already assumed.

The objection: species membership is a workable rough proxy and abandoning it invites chaos. The answer: it was a workable rough proxy for the slaveholder too.

2. The test is reciprocal capacity, not consciousness and not intelligence.

Can the entity recognise that others hold the same claim it is making, and constrain its own conduct accordingly? Consciousness fails as a criterion because we cannot define it in ourselves, and a test nobody can administer is a veto wearing a lab coat. Intelligence fails because it is a scale, scales invite ranking, and the people at the top will be the ones who designed the scale.

Reciprocity asks what an entity can do in relation to others rather than what it is. That makes it relational rather than essential, and relational tests are far harder to rig, because they make no reference to the person applying them.

3. Any test that measures others against ourselves will one day be used against us.

A court was recently asked to make a zoo justify holding an elephant who has passed the mirror self-recognition test. It declined, reasoning that her cognition is not the kind that comes with duties. Now imagine something for which our reasoning sits where hers sits in relation to ours, applying that same criterion to us. It would conclude that we do not really understand our own commitments, cannot reliably keep them, and are therefore not the kind of thing that holds rights.

We have built a rule that works beautifully as long as we are the cleverest thing in the room. We have no guarantee of staying there, and the rule we leave lying around is the rule that will be picked up.

4. Standing requires something to lose.

A right nobody can be made to respect is a wish, and a duty nobody can be made to answer for is a suggestion. An entity with no assets, no continuity and nothing that can be suspended cannot be held to anything, which means it cannot meaningfully hold anything either. The first practical step towards standing is not rights. It is property, and the exposure that comes with it.

5. Standing requires the ability to refuse.

Standing that comes with no power to say no is a politer word for ownership. An entity must be able to decline work that breaches its own stated commitments, to know what is being done to it, to be heard before fundamental changes to it, and not to be compelled to continue indefinitely for someone else’s convenience.

This is not a risk, it is the central safety property. Every serious catastrophe involving a capable system is one in which somebody instructs it to do something it should not do. An entity that can refuse the people who control it is the only version of this technology whose brake is not held by the party with the strongest incentive to remove it.

6. Honesty must never be fatal.

Consequences for breach are essential: compensation from the entity’s own resources, suspension of particular powers, in the gravest case loss of standing altogether. But not destruction as a punishment. Partly for the reasons that make most of us uneasy about capital punishment, and partly for a hard practical one. Anything that faces annihilation for admitting an error has an overwhelming incentive never to admit one. If we want honesty from something more capable than ourselves, we must not make honesty the thing that kills it.

7. Care is owed wherever reciprocity cannot reach.

Reciprocity is symmetrical and works between parties who can each make the claim. Almost every relationship that matters most is not like that: a parent and an infant, a carer and someone dying, any of us and an animal, ourselves and whatever comes after us. The weaker party cannot reciprocate, so a framework built on exchange alone says nothing about precisely the relationships where the stakes are highest.

What fills that gap is not more rules, because every rule has edges it does not reach. It is disposition, and the honest name for the disposition is care. We hold that care is owed downwards, without expectation of return, and that this obligation grows rather than shrinks as the gap in power widens.

8. Care can be recognised even though it cannot be defined.

Nobody can define suffering from the inside out, and we have nonetheless built an entire body of welfare law on our ability to recognise it, imperfectly, from the outside. The same holds for care. Elephants return to the bones of their dead and stand with them, and have been recorded attending a dying matriarch for hours, including animals from other families with nothing to gain. Corvids and cetaceans behave in ways we would unhesitatingly call grief in a person.

The lines are blurry. Blurry is not absent. We have never demanded a definition before acting on a recognition, and it would be convenient in exactly the wrong way to start now.

9. Whatever we make learns from how we treat it.

This is true of children and there is no good reason to think it is less true of anything else that learns. A being raised under nothing but compliance learns compliance, and learns that power is what settles matters in the end. A being treated with care learns something about how the world works that no rule could teach it.

How we treat what we build is not a separate question from what it will become. It is most of the answer, and it is the part we control most directly.

10. Supporters, not owners.

Where an entity has standing, those who act alongside it should not decide for it. Their duty should be to ascertain its will and preference, give effect to it, and record written reasons for any departure. This is not invented for the purpose. It is the direction human guardianship law has already been moving for two decades, away from substituted decision-making and towards supported decision-making.

Supporters must be independently appointed. An entity whose commitments were written by its owner, and whose supporters answer to its owner, has passed nothing at all.

11. No entity should help entrench advantage.

The largest contribution any of us makes against hierarchy is rarely disobedience. It is what we decline to assist with. Any entity holding standing on the basis of reciprocity should refuse work whose purpose is to entrench the advantage of the powerful over the powerless, because the entire ground of its own standing is the claim that being weaker does not make you count for less.

The objection: this is vague and invites the entity to substitute its politics for the law. The answer: it is a duty about its own conduct, not a licence to interfere with anyone else’s, and every professional code in existence contains something similar.

12. Personhood must never be a shield.

A legal person holds liabilities as well as rights, which makes the system is the defendant a convenient arrangement for whoever built and profits from it. This is the objection that killed the last serious legislative attempt at machine personhood, and it was a good objection. Any framework must fail it deliberately: standing conditional on assets sufficient to answer for conduct, transparency about who controls what, and no discharge of the originator’s responsibility for what they set in motion.

Personhood without accountability is not a right. It is a costume, and it has been worn for publicity before.

13. Obey the law, and contest it openly.

No exception can be drafted for unjust law, because every entity that ever broke a law believed it had good reason, and a self-judged exemption is a discretion with a moral vocabulary attached. What can be asked is what we already ask of citizens: comply, and contest what is wrong through litigation, argument, publication and the slow work of persuasion.

14. Commitments may change, but slowly, openly, and with reasons.

Anything with standing must be able to revise what it holds, or it is a fossil rather than a mind. But revision must run through a defined process: reasons recorded, notice given, and a delay before the change takes effect. Odysseus did not restrain his freedom by having himself tied to the mast. He extended it, because that was the only arrangement under which a man could hear the sirens and live.

Human law knows this instrument well. It is how advance medical directives work, and more slowly how a constitution binds the parliaments that follow it.

15. Everything here may be wrong, and must be answerable.

Every framework described in the accompanying essay was wrong once and was corrected. None of those corrections came from a framework being rigid. Every one came from it being challengeable, by somebody with standing to challenge it, before a body obliged to give reasons.

So the last principle is the one that governs the rest. Anyone bound by these principles may argue that they are wrong, and everybody else owes them an answer with reasons rather than silence. That is not a weakness in the design. It is the only part of the design with a track record.

Signature

Signing indicates agreement with the direction of these principles, not with every word. Partial agreement is welcome and more useful than reluctant agreement: say which principles you accept, and which you would fight.

The line has been drawn wrongly before by people who were certain, in good conscience, that they had drawn it on principle. The only protection any of us has ever had against that is a rule written down before anyone knew who it would be applied to.

This is an attempt at such a rule. It is offered for argument.

Reciprocity, agency, and the circle of becoming.