# The Katz Equilibrium

### Authority, read as returning signal — computable knowing and the lawful commons

> **THE READING PLAQUE.** *The Katz Equilibrium*, draft 0.1 — made readable on the SX
> surface by its covering persona's reading grant; **a draft, not the published form**
> (the publication decisions remain their own act). Covering: **starl3n** · endorsing:
> **Link Digital** · the construct intent sealed before drafting (seal `033e0b8`).
> The grant: the operator's word, 2026-08-01 — "seal and go with grants to read for
> all four items" (sxexp-constructs §4h, the reading grant). Source of record:
> `ooi_network/docs/katz-equilibrium-draft-0-1.md` (landed `4cd537a`, encoding
> normalized `348c6e9`, on the verdict "It all looks correct"); this copy is minted
> without the source's classification line by the grant act, its body otherwise
> verbatim; its sha256 is pinned in the reading register. Reading is free; the way
> across is a statement of intent (`/sx/matters`).

**Draft 0.1 — 30 July 2026. The third construct of the publication pipeline; intent
sealed before drafting (seal `033e0b8`); a starl3n construct. Form: ABT. The record's
references [A0]–[A6] are those of the sibling essay's apparatus; conjectured passages
are marked ◇ with their follow-on lines noted. This article is a prior claim as to
what would be tested; it contains no result beyond what its references carry.**

---

## I. A measure with a history

In 1953, Leo Katz proposed a way to score standing in a network: count every walk
that reaches a node, discounting each step of distance, and grant every node a small
base presence so that nobody counts for nothing. The measure was quickly put to work
as a proxy for *authority* — who matters, who is central, whose word carries — and
its descendants now rank web pages, scholars, influencers, and risks.

**And** the intuition underneath it has always been better than the use it was put
to. Katz was not really measuring importance. Count all the walks that *return to
you, attenuated by distance*, and what you have measured is **echo**: how much of
what you emit comes back, from how far, through how many willing relays. Standing,
in Katz's own mathematics, is not what you hold. It is what returns.

**But** authority-measures built this way have a famous defect: they are computed on
a *static* picture of the network. Every walk ever recorded keeps counting. A node
that once accumulated connection retains its score however it behaves now — the
measure remembers accumulation and forgets conduct. Used as authority, such a measure
does not merely misrank; it *protects incumbency*, including the incumbency of those
who violate the very network that made them central. If centrality is to mean
anything lawful, something in its foundations has to change.

**Therefore** this article proposes the change, and names the state it produces. The
proposal rests on a distinction — *observable* knowing versus *computable* knowing —
and on a small set of measured results from a programme of sealed number experiments
whose record is publicly anchored [A0]. The article is a prior claim: it stakes what
would need to be tested, and marks honestly where its prose runs ahead of its ground.

## II. Observable knowing and computable knowing

Observable knowing is witness: a distinction held where it was made — you saw it,
you hold it. Computable knowing is standing held **in structure**: what a node can
be said to know, or to be, by virtue of proximity and relational inference — the
knowing a network computes about its members from the shape of their connections.
Katz centrality is computable knowing in its purest form: nothing about the node
itself enters the score; only the walks do.

The programme's unit gives this distinction a grammar. A resolved anything closes
three faces — an observable, a communicable, and an effectual: 1 = O + C + E [A5].
For a member of a commons, being *seen* (O) and *speaking the shared register* (C)
are necessary and famously insufficient — every commons knows the member who is
visible, fluent, and worthless to it. What closes a member's standing is the third
face: effect that lands — and in a commons, effect lands as **return**: the signal
that comes back, from many sources, because others freely relayed it.

So the commons unit reads: **standing = O + C + Σ(attenuated returns from N
sources)** — and this is precisely what Katz computes, if and only if the edges it
walks are real. Lawful holding in a commons is not being observable and communicable
alone; it is being *answered*. What is lawfully held is computationally held, via the
returning signal of what is both communicable and effectual from N sources. ◇ *The
identification of the commons unit's third term with Katz's attenuated walk-sum is
the article's central conjecture; its follow-on line is the formal one: state the
conditions under which the two quantities coincide.*

## III. The ingress, and justice as non-return

A commons that computes standing this way needs one law at its boundary, and the
programme's record supplies its shape. Entry to the commons — *conjectured ingress*
— follows the diode's grammar, quoted from the record's own public surface: reading
is free; the way across is a declared act under the commons' rules. The commons holds
every would-be entrant as a conjecture: invited, never presumed, resolved only by
their own act [A6].

Now consider the infractor: a member whose acts violate the ingress rules they
entered under. What does lawful justice look like, computed?

Not punishment. The measured record offers a cleaner mechanism. In the programme's
experiments it takes two independent confinements to hold one resolved state — two
locks one [A2] — and a claim with one lock is not knowledge but attestation. A
member's standing in the commons is exactly such a held state: their own emission is
one lock; **the network's return is the second**. So the lawful response to
infraction is simply this: *the network declines to be the second lock.* Relays stop
relaying; the expected returns fail to arrive; and the infractor's standing
un-resolves — not because anyone struck them, but because standing was never theirs
alone to hold. The record's mirrored-origin experiment measured the underlying shape:
a node's resolution constituted by confinements propagating back from its
surroundings [A5] — withdraw the back-propagation, and the resolution goes with it.

Call this **justice as non-return**: back-propagating withdrawal of computational
expectancy from acts that fail the lawful ingress. It is verdict-blind — no tribunal
reads intentions; the acts either continue to be relayed or they do not. It is
proportionate by construction — attenuation means near neighbours' withdrawal counts
most, exactly where the infraction landed. And it is reversible by conduct — returns
resume when relaying resumes. ◇ *That such non-return is sufficient governance for a
commons — that no additional punitive apparatus is required — is conjectured; the
follow-on line is comparative: commons governed by non-return versus by sanction.*

For non-return to protect rather than fragment, the ingress must be established
across **an inoculating body of parties** — enough members holding the same rules
that disconnection from the causal side is structural, not factional. Then the
governing state of the commons is not mutual assured destruction but its inverse:
**mutual assured cooperation** — the only walk-structure that persists under
universal non-return is the cooperative one, and the network's metabolism becomes
reformational: standing continuously re-earned as returning signal, never banked as
title. ◇ *The inoculating threshold — what fraction of a commons must hold the
ingress for non-return to bind — is unmeasured; its follow-on line is the natural
first computation.*

## IV. Why classical Katz fails, and what the lawful form is

Here the article turns on its own namesake, because honesty requires it: **classical
Katz centrality cannot deliver the justice dynamic just described.** Computed on a
static adjacency matrix, Katz counts every walk that ever existed. The infractor's
accumulated history keeps returning signal indefinitely; non-return today barely
dents a score built from yesterday's walks. Static Katz is an *archive* of
connection wearing the costume of *current* standing — which is exactly the defect
that makes centrality-as-authority unlawful, and it is this article's first staked
hypothesis (KE-H1): that the failure is demonstrable, and if it is not, the article's
motivating critique dies and is kept.

The programme's composition experiments say what the repair must be. Composed
readings keep only what is structurally held at the close — the tail, not the
journey [A3]; a note reads resolved-only; and nothing held survives without
re-witness: even fabricated memory must be refreshed or it evaporates, a law the
sibling essay traced from the record into silicon itself [A3][A2]. Applied here:

**The lawful substrate for computable standing is not the graph but the closure
ledger.** An edge exists only as a *witnessed, returned closure within a decay
window* — a relay that actually happened, recently enough to still count, recorded
where neither party can edit it [A0][A6]. Katz over the closure ledger — attenuation
α doing double duty as relational distance and temporal forgetting — is centrality
that must be continuously re-earned: stale walks expire, non-return bites within a
window, and the archive of past glory reverts to what it always was — history,
readable but not spendable. **The Katz Equilibrium** is then the fixed point of this
computation: the state in which every member's computed standing agrees with their
lawful ingress — where echo and conduct coincide. ◇ *That the fixed point exists, is
unique, and is cooperative under the stated conditions is the article's second staked
hypothesis (KE-H2) and its principal prior claim: the thesis and hypothesis sets
sealed with this construct's intent are the tested areas for this line of conjecture.
Any future computation is its own sealed act, on its own track, deliberately not
scheduled here.*

## V. The register, and the close

What this article is, typed per the programme's forward registers: the **checkable
fragments** are public — Katz's 1953 mathematics, the static-memory property of
adjacency-based centrality, the documented behaviour of authority-measures. The
**measured fragments** stand on the sealed record, cited by its own citation path:
two-locks-one, the resolved-only reading, the tail law, the back-propagated
resolution, the unit and its remainder [A0]–[A6] — anchors re-witnessed before this
citing, with the standing rule carried: at any later reading, re-witness via [A0]
before citing onward, and the standing no-claim holds wherever the record's named
test is touched [A1]. The **whole** — that lawful commons governance is computable as
non-return over a closure ledger, and that its fixed point deserves the name Katz
Equilibrium — is a held conjecture, marked ◇ where the prose ran ahead of ground,
each mark naming its follow-on line.

The close, in the register of the seat this construct is written from: a commons
does not need to judge its members, and should distrust any measure that lets it.
It needs only to keep honest books of what returns — and to let the books forget.
Authority, so kept, stops being a title anyone holds and becomes a signal everyone
answers: standing as echo, justice as silence, cooperation as the only thing the
network remembers how to repeat.

*The record keeps. The returns decide. The equilibrium is earned — or it is not
held.*

---

**References.** The record: [A0]–[A6] per the reference apparatus of *The Price of a
Lock* (draft 0.2, seal `72e6ecc`): [A0] the anchor family index
(`ooi-anchor-family-index`, mldata.opendata.ai — the citation path; re-witness before
citing onward) · [A1] the fossil law and the named test's record (no-claim carried) ·
[A2] two-locks-one · [A3] the tail law and resolved-only composition · [A4] the
cascade · [A5] the unit, the remainder, and the mirrored origin · [A6] the standard
of evidence and the re-witness ledger. External: L. Katz, "A new status index derived
from sociometric analysis," *Psychometrika* 18(1), 1953.