The Architecture of Agency Volume 7 Rights Are Forged — and Reciprocal

Rights Are Forged — and Reciprocal

Power makes boundaries real; reciprocity makes them legitimate

This chapter is a review — it is readable but still changing.

Rights are often called natural, divine, or state-created. Nature supplies consequences, not permissions. Revelation supplies authority only after its moral premise is accepted. State creation would make every lawful atrocity legitimate by definition. Institutions can recognize, embody, enforce, or simulate rights; none of those acts alone supplies moral justification.

Rights are forged in two senses that must not be collapsed. A moral right is a defensible protection or claim around agency whose justification survives reciprocal application. A legal right is an institutionally recognized claim. Enforcement makes a boundary effective; neither recognition nor power alone makes it legitimate.

Forged does not mean arbitrary. A sword and a shackle are forged too. This chapter makes both halves explicit: power makes a boundary real; reciprocity makes it a right.

The Forge

A justified moral claim can precede its protection, but no boundary remains stable without resistance. Enforcement may be personal, social, or institutional and may use persuasion, coordination, refusal, threat, or force. Each mechanism needs its own classification and justification. The claim and its reciprocal principle make a candidate right; embodiment makes it effective.

Patri Friedman’s compression — “there are no rights, only mechanisms”1 — captures the need for rules, detection, enforcement, and consequence. It cannot distinguish a rights regime from well-mechanized slavery, apartheid, or tyranny. A standard beyond effectiveness is required even to ask whether the mechanism is just.

Declarations can also precede enforcement by coordinating allies and norms. The Declaration of Independence was ink on parchment until people chose to act on it. Rights without mechanisms are toothless; mechanisms without a justifying principle are blind.

Slaveholders Had Preferences

Here is where the forging thesis, left on its own, fails. If rights are merely preferences we are willing to enforce and consider ethical to enforce, then every successful ideology can manufacture them. Slaveholders had preferences. Censors have preferences. Mobs have preferences. Patriarchs, priesthoods, platforms, and states all have preferences, and many of them enforce those preferences with complete moral confidence.

A preference, willingness to enforce it, and belief that enforcement is ethical still do not yield a right. Masters, censors, and mobs can each redescribe domination through necessity, custom, divine command, or collective approval. Moral confidence does not distinguish their boundary from a defensible one.

A right requires more. A right is a defensible boundary around agency. It marks a domain where coercion requires justification, and where the justification must survive reciprocal application to the person against whom it is used — a principle that applies to every agent exposed to the same kind of agency-invasion, with the reference class set by the coercive act and the agency it invades, not by the coercer’s preferred hierarchy. That condition does the work. Without it, rights collapse into moralized power.

Reciprocity Is Not Mysticism

The reciprocity constraint invites an obvious suspicion: haven’t I just smuggled natural law back in through the side door? No. Reciprocity is not a supernatural axiom or a moral particle hidden inside the universe. It is a constraint internal to justification among agents.

A coercer can reject reciprocity. He can say: I have power, you do not, and that is the end of the matter. That position is coherent as threat. It is not coherent as justification. The moment someone offers a reason for coercion, he has entered the space of justification — no longer merely grabbing, striking, or imprisoning, but claiming authority: saying the coercion is not just something he can do but something he may do. And that claim has a cost. A reason must be more than a disguised index of the speaker’s position; it must be capable of application beyond the speaker’s immediate advantage. Otherwise it is not a principle. It is a costume worn by power.

Reciprocity is the discipline that separates principle from costume. It asks whether the rule being invoked survives when applied from the standpoint of the person being coerced. An agent who refuses that discipline has not refuted rights; he has withdrawn from justification and returned to force, declaring himself outside the only practice that could legitimate his coercion. Rights are not force fields. They do not physically restrain the psychopath, the conqueror, or the machine that recognizes no reason beyond power. What they mark is the difference between authority and predation, and therefore between obedience and resistance. Against the agent who will not justify, the answer is not argument alone. It is defense.

Because domination is intelligent — it adapts, and it rarely announces itself as domination; it announces itself as order, tradition, safety, compassion, or emergency — the test needs an explicit anti-evasion rule: the coercer may not define the relevant class by the hierarchy the right is meant to constrain. This blocks the standard escape route. The master cannot define the enslaved as property and then observe that property has no liberty rights. The censor cannot define dissenters as dangerous and then note that dangerous people have no speech rights. The patriarch cannot define dependents as possessions, nor the state define citizens as subjects who hold only the permissions the state grants. Classification may still matter — children are not adults, guilt can justify bounded punishment — but the burden runs against the coercer. Whoever asserts coercive authority must identify the agency being invaded, the harm being prevented, the boundary being defended, and the reciprocal principle that authorizes the intervention. Administrative convenience is not enough. Disgust is not enough. Majority approval is not enough. The victim’s low status is not enough.

The Protected Domain

What kind of boundaries pass the test? Agency requires a protected domain of control. A person must have practical command over body, speech, movement, tools, and associations; remove those protections and the person remains biologically alive while becoming operationally subordinated to another will.

Bodily liberty protects against physical commandeering; property protects external means of action; speech protects judgment and error correction; privacy protects interior life; association protects cooperation and intimacy; due process protects against punishment without proof. These recurring rights protect conditions under which agency can survive conflict.

And conflict is the operating environment. People will disagree, compete, offend, deceive, exclude, organize, and accuse. A society of agents needs boundaries that prevent disagreement from becoming capture, and a right is exactly that: a declaration that some domains cannot be invaded merely because another person, faction, corporation, or state finds invasion useful. Rights are not guarantees of comfort, status, or success. They are constraints against capture — protections for the space in which an agent can continue to act as an agent among other agents.

Needs and Property

The next chapter owns two consequences. Need can motivate provision without by itself identifying a debtor or authorizing coercion; an enforceable duty needs a bridge such as causation, consent, contract, guardianship, or fiduciary responsibility. And property remains subject to reciprocity: control across time can protect agency, while titles descending from theft, conquest, or fraud can preserve an earlier capture. Restitution still requires an identifiable wrong, claimant, responsible party, and bounded remedy.

Hard Cases

Rights are not identical across all beings and all institutions. Reciprocity does not erase relevant differences in agency, capacity, dependence, knowledge, or responsibility; it forbids using them as pretexts for capture. The interests rights protect form a profile — perception, memory, anticipation, learning, self-direction, vulnerability — even though the book separately distinguishes basal agency from the threshold it proposes for sovereign authorship. Rights track the domains in which those capacities or interests can be captured; they need not turn every difference of degree into a different kind of agent.

Children lack some adult capacities while retaining rights against assault, abuse, enslavement, and exploitation. Parental authority is fiduciary, justified by dependence and development rather than ownership.

A rights-violator may be restrained or required to make restitution, but guilt does not erase personhood. Proof, proportionality, and constraints against cruelty remain.

Many animals provide strong evidence of sentience through pain behavior, anticipation, bonding, goal pursuit, and physiological continuity. That supports welfare protections fitted to the experience or agency at risk, with precaution under uncertainty, without requiring a human rights profile.

For artificial minds, tribal membership cannot settle status. The structural question is what kind of agency can be captured, frustrated, confined, or destroyed; Sapientism and Sentience Without Sovereignty carry the criteria.

Institutional powers and immunities can protect persons through pluralism, exit, and limits on concentrated power. Their claims remain derivative: the analysis must pass through to the agents protected or invaded.

Hard cases show why rights require judgment. They do not reduce rights to preference.

The Seven Failed Claims

Now run the claimants from earlier — each armed with power, tradition, scripture, law, or consensus — through the test.

The master’s claim fails because ownership cannot be reciprocal among the same class of agents. The censor’s rule makes speech permission from whoever holds power; the mob’s outrage creates no jurisdiction; the priesthood cannot turn one person’s metaphysics into another’s cage; the patriarch cannot convert fiduciary care into ownership. Each claim defines a hierarchy and then uses that hierarchy to excuse the invasion under review.

The platform case is usually misdescribed, and it is worth disaggregating. A private forum does not violate my rights by refusing to host me; its servers, brand, and audience are not my property, and scale does not by itself convert access into entitlement. The rights issue arises when exclusion is backed by fraud, breach of contract, state pressure, collusion, or legal privilege that blocks exit and competition. Then the problem is not that I possess a positive right to another person’s platform — it is that coercion or privileged exclusion has invaded the conditions of agency and exit. What a free-speech order requires of its infrastructure is the subject of Civilizational Infrastructure.

The state’s claim fails whenever legal classification is treated as moral cleansing: taxation called revenue, imprisonment justice, surveillance security, war policy. The labels do not settle who is coerced, which boundary is defended, or whether the authorizing principle survives reciprocal application.

Parchment Barriers

Law, custom, and consensus can embody rights without creating their moral legitimacy. Otherwise lawful abuse would be conceptually impossible and social recognition would become a moral oracle. A right needs embodiment to become effective and reciprocity to remain distinct from domination.

The practical diagnosis asks what the state cannot do. If it may arbitrarily seize property, extinguish livelihood, and criminalize resistance, citizenship collapses toward subjection no matter what the charter says.

Bastiat called law used for expropriation legal plunder. Constitutions constrain only through institutions, incentives, and people willing to honor and defend them. Emergency, security, or public-health language cannot substitute for those mechanisms.

Enforcement Without Monopoly

Rights need enforcement; it does not follow that they need a monopoly. Credible defense, adjudication, and consequence may be personal, reputational, contractual, communal, market-based, or state-based.

None is self-validating. Private agencies can become gangs, communities mobs, markets privilege wealth, and protocols freeze bad rules into code. The point is not that polycentric enforcement is automatically just, but that no institution becomes the source of rights by claiming final jurisdiction. The Archist Axiom takes up whether order can be kept without monopoly.

Rights, then, are forged by agents living under conflict — argued into form, defended under pressure, revised through experience, embodied in institutions when institutions can be made to serve them. They need force, because a boundary that cannot be defended becomes a request; they need moral discipline, because force without reciprocity becomes domination. A claim does not become a right merely because someone is willing to fight for it. It becomes a right when the fight defends agency under a principle that applies to every agent exposed to the same kind of invasion. Rights are forged. They are also tested. Power makes boundaries possible; reciprocity makes them legitimate.


  1. Patri Friedman (@patrissimo), post on X, September 2025, https://x.com/patrissimo/status/1964773590731296777.↩︎