The record, in order.

For a century, personality and intelligence testing has been done to people rather than with them: sorting conscripts, ranking applicants, scoring children. The mathematics is neutral. Who owns the output is not.

ALTER is the other choice. The record of you belongs to you. It is measured from what you do. The model is yours. You are paid whenever someone reads it. You can cut off access with a keystroke. No form, no review.

The eight papers that follow set out the mechanisms: measurement, ownership, revocation, payment. The first one answers the question every objection eventually circles back to: why should belonging be measured at all.

Where you were born, and who your parents were: those are the two rules every modern country uses to decide whether you belong. Both are accidents.

If who you are can be measured, and if that changes over time, belonging can be earned by what you do rather than inherited by accident. Paper I calls this principle jus identitatis, the right of identity. It faces the obvious objections head-on: that it slides into “culture fit” sorting by another name, that it echoes eugenics, that it hands enormous power to whoever runs the test.

Every objection has the same answer, and it is structural: the record is owned by the person it describes, revocable by them, and governed by the people it measures rather than by whoever runs the measurement. Without those three conditions, the principle fails. With them, it addresses a question political theory has left open for a century.

You do not have an identity the way you have a passport. You maintain one. Moment by moment, your brain runs a model of who you are and updates it against the evidence of what just happened.

Paper II writes that process down precisely, using the free energy principle — a framework from theoretical neuroscience in which the brain continuously predicts its own inputs and minimises the difference between prediction and reality. Five consequences follow. Personality questions are deliberate surprises: the spread of answers tells you how confident the self-model is, not just what it claims. Civic identity is the same mind modelling which group it belongs to, measurable the same way. Belonging is low social free energy: the gap between how you see yourself and how the group sees you, held small.

Governance, in the same frame, is how a group draws its own boundary. Whether something is being done to you, for you, or by you depends on which term of the free-energy equation is being minimised, and by whom.

Six falsifiable predictions keep it honest. One example: how consistently you answer questions about yourself within a day should exceed how consistently you answer across days. Standard tests cannot reproduce that pattern. Most predictions can be tested with instruments psychologists already use.

The individual was never the basic case. You are already a bundle of smaller processes (cells, reflexes, memories) held together by a boundary that marks what counts as “you” and what counts as outside. Mathematicians call that boundary a Markov blanket — the statistical envelope that separates what is inside a system from what is outside. A community applies the same boundary to people.

The shape splits into two layers, both written with the Greek letter Φ (phi). Φ-structure is the form all self-maintaining systems share: a recursive boundary that decides its own inside. Φ-dynamics is how each scale carries that form — neurons for a mind, members for a team, institutions for a country. Same shape. Different materials.

A third quantity, written κ (kappa), measures how carefully a system models itself. A community with κ > 0 adjusts its own rules when reality changes. One with κ = 0 applies the old rules until it breaks. Either way, measuring κ is not neutral: the act of reading changes what is read.

Classical psychology assumes you are the same person before and after a question about yourself. For height and weight, that is safe. For self-knowledge, it is false. Answering the question changes what the question measures.

Your answers to the same question vary across a week. Classical psychology averages that range away. A person is a distribution of possible states, not a single point on a scale. The standard psychometric toolkit — the one every modern test is built on, where every item is rated for difficulty and for how well it separates people — is a narrow special case: the point you get when the distribution collapses to its average. The effect of asking is ignored.

The broader framework predicts things the old one cannot. It treats the spread in a person’s answers as information about the distribution, not noise to be removed. Tests earlier generations dismissed as unreliable may have been measuring something real. Just not what the toolkit was built to read.

A country or a company watches and updates itself in the same way a mind does. Once you accept that, its political troubles stop being a scatter of separate problems. Polarisation, alienation, rule by the few, loss of legitimacy — features of one distribution, directly comparable between systems.

The constitutional moment is when a polity stops patching its current self-image and adopts a new one. In the equations, it shows up as a phase transition: a sharp change of state, like water freezing. A named order parameter tracks how close the system is to flipping. How quickly an institution already listens decides where that transition sits.

Five predictions. Each targets a real political system where the data to test it already exists today.

Whether a machine can have a self is usually argued about as metaphysics. Paper VI treats it as measurement.

Four statistical signatures, taken together, say whether a system is actually maintaining a self-model. It holds its position when pushed and recovers in proportion. Its confidence tracks its accuracy. When surprised, a thinking system shows a precision-weighted response: attention rising where signal is strong, new evidence tested against prior belief. A lookup table’s response stays flat. And it stays coherent across different domains without being given explicit memory.

All four must hold at once. Any one alone can be faked. A language model rehearses coherence without holding any of the others. Biology sets the calibration: great apes should pass, insects should fail, and κ should rise steadily from newborn to adult.

A five-step argument then connects passing that test to having civic standing. One step is named, openly, as a value judgement rather than hidden inside the maths. Machine rights then has a pass-fail criterion, not a standing debate.

Every claim in the papers rests on something that could be shown false. Paper VII gathers seventy-five of those predictions in one place, across three domains: human psychometrics, governance systems, and synthetic self-models.

It does not prove anything on its own. Each prediction is a test the framework can fail. If they fail, it fails with them. If they hold, what the next paper describes is the shape of what has been measured. This register keeps the programme accountable to its own claims.

If the programme survives those tests, what emerges is this: identity is not something a person has. It is a field a person participates in. Every interaction with a substrate — active or passive, witnessed or private — deposits a measurable trace into the field. The person is the running total of those traces over time.

Three theorems anchor the framework: the Identity Field Constraint, the Fluctuation–Dissipation Theorem (how noise and response stay linked), and a Cross-Scale Coupling Law (how one scale shapes another). Together, the three give the conditions under which one identity is recognised by another. Checkpoint authentication (passwords, badges, tokens) is a narrow special case: a single reading taken of something that has always been continuous. The record, on this view, is not a document about you. It is your participation in the field, made legible. And kept yours.