# The Zero That Is Signed

*An outside reading of the foundations of probability, from 1 = O + C + E.*

**Steven De Costa, writing as starl3n · Canberra · 26 September 2026**

> **THE READING PLAQUE.** *The Zero That Is Signed* — draft 0.4, made readable on
> the SX surface by its covering persona's reading grant; **a draft, not the
> published form**; the author's grant word, verbatim: "That gets a reading grant on the SX surface and push it - I have updated all three statements of intent for the three personas to reflect my intentions as a triangulation of concerns." The
> construct was established 2026-09-26, on Steven De Costa's acceptance of the
> summation registered on the SEB item
> [hajek-probability-corpus-2026](/sx/seb#hajek-probability-corpus-2026) — two
> entries of Alan Hájek, read whole and registered as an alignment object — and
> on his word that the two points of intent he gave at the Russell acceptance
> stand for this set, with the set's refined intent sealed before a word of the
> draft: an essay from the outside domain, not a seat inside the debate over the
> interpretations of probability; reciprocity and gifting, no assertion of value,
> to the person named. The first essay of a set — the readings that follow the
> transits through the ANU School of Philosophy, collected under the draft title
> *On the Vitality of Reason as Material Cares*; the philosophy the set carries
> named by its author *Co-Constructive Optimism*. Four editions in one day: 0.1
> the essay laid; 0.2 at his review, the guest named and the gift addressed to him,
> the close made a carried debt; 0.3 his typo corrected and the ladder of the gift
> written out; 0.4 the identity that authors named beside the persona that signs,
> and a second debt, outside the set, written into the accounting — the three
> statements of intent for LinkDigital, starl3n and the Chairman, by his own hand
> — a debt discharged before this grant, his word: "I have updated all three statements of intent for the three personas to reflect my intentions as a triangulation of concerns." The citation
> discipline: the corpus cited only as [/sx/seb](/sx/seb) carries it, never quoted
> beyond identification, never graded; the conjecture 1 = O + C + E cited in its
> reference's three terms; every load-bearing sentence pointing at a public
> reading of the house; no probability adopted as a mechanism; no contact with any
> thinker named — the gift is left where he can find it. Covering starl3n;
> endorsing Link Digital. Readable, not published: nothing is measured about
> anyone who reads it, and the surface serves its own words only. The naming is
> the author's at review; a further edition lands under a new pin. This file's
> sha256 is pinned in the reading register.

## A note on the title {#the-title}

Probability has two zeros. The first is fixed by the axioms: the empty set gets
nothing, and that is the end of it. The second is the zero one meets when one
tries to condition on it — the ratio that defines a conditional probability
holds only provided the thing conditioned on has positive probability, and
where it does not the quotient falls silent. The two entries my house registered
— Alan Hájek's, on the interpretations of probability and on Pascal's wager —
carry that second zero by a pointer to an argument of his that I do not hold,
and by the names of those who took conditional probability as the primitive
notion instead.

The zero in my title is a third. It is not a value and it is not a silence. It
is a state — the one my house calls −0: a computable knowing that identifies the
absence of a computable knowing, and signs it. The signing is the whole point.
Nothing at the state derives it; someone does it; and once it is done the walk
from it is lawful.

There is a self-reference in the title as well, and I will own it now rather
than let a reader find it. This essay is the first of a set. Before it the set
had no member; there was an absence where a set would be, identified when I
wrote the essay before this one — the summation after the Russell review — and
not yet signed. It is signed here, one summation later. The zero that is signed
is what this essay is, as well as what it says.

## On the set — the bridge from the essay before {#the-set}

The set is the school of philosophy at the university in the city I live in,
taken as a local environment: an ecology of thinkers, where the fitness of my
own infant philosophy can be tested by my house's methods. Each philosopher's
work will pass through my house's door and become a summation; some summations
will become an exhibit, and an exhibit an essay moved toward the public reading
surface. The essays are all of a kind, and they group in the order my house
drafted for them. I expect to collect them as a book, and I have given the book
a draft title: *On the Vitality of Reason as Material Cares*.

The infant philosophy has a name too. It entered through the first review — the
Russell corpus — and was first signed in *Sane Max Rides Hume's Highway*, and it
is simply titled *Co-Constructive Optimism*. I expect that while the set is
read, experiments will be run and works made by my own hand, as the one who
keeps the house and as the persona that signs; that the infant will mature and
adapt under the local pressure of the philosophers' courtroom; and that the
narrative arc of the set will be one of evolution and maturing, of adjudication
by the socialised norms of unique minds, of conjectures made and tests run. A
symphony, perhaps. I foreshadow, and design nothing here, a game of two worlds
that are deterministically of a kind, which I have begun to call Baseball versus
Cricket; its notes sit in my house's inbox and wait for their own session.

This note is the bridge from the essay before this one into this one, and it is
made once. The essays that follow inherit the set from here and need not
describe it again; what each of them may bring instead is a note on the chords
that begin to ring true as the set is read, and on the discordances. I say how
at the close.

## 0 · The stake, not the seat {#the-stake}

The discourse on the interpretations of probability has seats. Laplace, Keynes
and Carnap sat in the classical and logical ones; Ramsey, de Finetti and Savage
in the subjective; Venn, Reichenbach and von Mises in the frequentist; Popper in
the propensity seat and Lewis in the best-system seat; and Alan Hájek, whose two
encyclopaedia entries my house read whole, sits among them, holding cares and
notes lawfully and answerable to the others for what he holds. A seat in that
discourse is earned by entering it: by proposing an interpretation, meeting the
criteria, and standing to be answered. I do not enter. I take no seat and offer
no verdict on which interpretation is right, on whether rationality requires the
calculus, on countable additivity or regularity, on the wager, or on God. I stand
at the rail, stake nothing, and say what the ground looks like from outside.

Why read at all, then? Because I took value from the readings — two bodies of
one author, read whole and registered through my house's own door before a word
of this essay was written — and my manner with value taken is to give something
back. What I give is not a verdict but a mechanism: a way of reading the
foundations that is more fundamental than any one interpretation, in the sense
that it says where every interpretation begins, and scale invariant, in the
sense that it says the same thing about a person, a record, an organisation and
a federation. I offer it as reciprocity, with no assertion of value. If it is
useful to those seated inside, the reading was worth it. If it is not, the
refusal is theirs to make and I give no grief.

The gift has an addressee, and I will not leave him unnamed behind the word
"discourse". The author my house registered is a person, Alan Hájek,
personified within my house by the summation my door registered on his work and
by the intent I sealed after it. In my map he is the −1 this essay walks toward:
an absence identified — my house held no computable knowing of his direction
until it read him whole — which the signing of this essay's zero begins to fill,
and whose release at the close must release him too, by name, or it is no
release at all. So the parcel this essay carries — a computable knowing built
from my house's own rules and manners, with its cares and its matters in the
form of the essay itself — is given in direct faith, with well intended belief,
for his own reading. Reciprocity to a discourse is a gesture. Reciprocity to a
person is a debt paid, and I pay it to him.

I should say what "faith" and "belief" carry here, because the words are often
said in summary — a thing is delivered in good faith — and I mean them as a
ladder I have practised. The passage above carries the register of my cares as
I grow them toward the expectancy of an external autonomy, and toward the
plausible matter of how that autonomy might operate this reading and gain from
what lands with it, in a communicably effectual manner, as a good. The ladder
rises in the three words I keep in place of a probability, and their credit
stands where I first drew a geometry from them
([The Framed Window, the three words](https://opendata.ly/sx/reading/the-framed-window#the-three-words)).
At the first rung the value is planned and delivered by me as a *possible* good,
plus faith — and it locks on my side alone. At the second it is a *plausible*
good, plus belief — my well intended belief that it can be read as I mean it.
At the third it is projected, as this public reading attenuated toward the named
externality, plus the *probable* — the third word, never a number — and it
lands, if it lands, as a good that remains noted as my care, differentially
held. My optimism is autonomous through the act of construction. But the
autonomy given to the constructed essay is entirely subject to the autonomy of
the named externality — to Alan Hájek's, and to that of every
self-model-referential externality this set will address. The gift can only
land as an optimisation by being granted a value on the receiver's side; his
weighting of that value is lawfully held and not at all diminished by my
optimism. Optimism is, by my own signed walk, a matter of my own autonomy — and
that is the whole of what co-constructive optimism adds to good faith: the
co-construction is his to grant, and mine only to offer.

The perspective I read from is my own conjecture, held and never asserted:

> **1 = O + C + E** — reality, taken for an observer as the unit, decomposes
> exactly into what is *observable* (the state that stands readable within a
> frame), what is *communicable* (the account that can be carried between
> frames) and what is *effectual* (the differential that changes the account).

That is the citable form, and the only form I use here. It is stated in my
public reference and earned inside declared experimental worlds, and I claim
nothing for it beyond its apparatus.

I map the essay as I map a walk, and I expect the reader in none of the four
states: **−1**, where an absence is identified; **−0**, where the absence is
signed; **1**, the walk; **0**, the held release. The calculus's silence is the
−1. The mechanism is the −0. The discourse's own lawful traffic, read from
outside, is the walk. What the sane bettor does is the release.

## 1 · The silence — an absence identified (−1) {#the-silence}

The register's form: in Hájek's entry on the interpretations, the axioms of the
calculus fix one to the universal set and zero to the empty set, and are
otherwise silent about the initial assignment of probabilities. The
interpretations are the accounts of who assigns, and why. And in his entry on
the wager there is a line I will name by its name and not quote — a phrase the
entry takes from Morris — epistemic nullity: the position that to assign any
probability at all, even one half, where reason can decide nothing, is to feign
evidence that one does not have.

My reading: the silence is an absence identified. Nothing in the calculus
derives the first number. It is not that the axioms are weak; it is that the
first number is not the kind of thing an axiom derives, as the ought was never
the kind of thing a descriptive premise derived. *You cannot derive what you
must sign* — I wrote that of Hume's remark in the essay before this one, and I
find the second instance here, on the same plane, in the same silence.

The −1 is an honest state, and I want to say that plainly because the discourse
sometimes treats it as a failure. An agent who has not assigned a probability to
a proposition is not yet wrong about anything. She has identified an absence —
there is no computable knowing here — and has not yet signed it. Epistemic
nullity, read from outside, is the −1 taken seriously: the refusal to feign a
number is the refusal to pretend that a signing has happened when it has not.
My house has a state for that, and a name, and a map that places it before the
mechanism rather than beneath it ([the signed zero](https://opendata.ly/sx/reading/the-signed-zero);
[Knowledge Engineering, the map](https://opendata.ly/sx/reading/knowledge-engineering#the-map)).

## 2 · The mechanism — the absence signed (−0) {#the-mechanism}

An *is* is an O: a state that stands readable within a frame. A probability
assignment is not an *is*. It is a C: an account carried into the frame from
somewhere — from the evidence, from the agent's confidence, from the chances of
the world — and it is an act before it is a proposition. That is what the
interpretations are, seen from outside: not rival theories of a value, but rival
declarations of where the signing comes from. The classical interpretation signs
from the symmetry of the possibilities; the logical from the language and the
measure chosen over it; the subjective from the agent; the frequency accounts
from the actual or hypothetical run; the propensity accounts from the set-up;
the best-system account from the winning systematisation. Each is a seat, and a
seat is what signs. The −0 is the absence signed: the assigner declares the
source of the assignment before a number is read off it.

Now the second zero. The ratio holds provided the condition has positive
probability. Where the condition has none the quotient is silent while the
conditional need not be; Hájek's entry carries this by pointer to his own
argument and by the names of those who axiomatise conditional probability
directly, so that the conditional comes first and the ratio is derived where it
can be. I do not hold the argument the pointer points to and I lean on none of
it. I say only what my house can say from its own record: conditioning on a zero
is lawful when the zero is signed. The condition is a declared world — fixed
before the readout, its edge said aloud — and inside a declared world nothing
has to be divided by it. My house seals its worlds before it reads them: the
intent fixed before the run, the kill named before the read. Taking the
conditional as primitive is, from outside, taking the declared world as prior
to the readout. The order of the signing is the kinship ([the LDD reference, §1](https://opendata.ly/sx/reading/ldd-reference#1-the-conjecture);
[the XOR from the a-priori](https://opendata.ly/sx/reading/the-xor-held)).

So there are three zeros, and the mechanism is the third. The empty set's zero
is a value, fixed by the axioms. The trouble-spot zero is a silence, where a
ratio fails. The signed zero is an act, where an account begins. The first two
are the calculus's; the third is the seat's. An interpretation of probability
is, read this way, an account of who holds the pen at the third zero — and the
disagreements inside the discourse are disagreements about the seat, not about
the calculus, which every party accepts and which is silent on exactly this.

I stated this at the Russell review as a fit of form and I state it so again:
no identity is claimed between my −0 and any object of the calculus. The fit is
that both stand where a derivation was expected and an act was found.

## 3 · The calculus's lawful traffic — the walk (1) {#the-walk}

Here are ten forms as the register carries them from Hájek's two entries, each
with my reading from outside, its cadence in my house, and the walk my house can
set beside it. The forms are his and I state them only as the register states
them; the readings are his, the cadences mine.

### 3.1 · Three criteria for a door {#the-door}

The register's form: an interpretation is assessed by three criteria the entry
takes from Salmon and refines — admissibility (the meanings assigned turn the
axioms into truths), ascertainability (there is some way, in principle, to find
the values), applicability (probability is a guide to life) — with
admissibility relative to an axiomatisation and applicability a cluster of
roles: to frequencies, to rational belief, to rational decision, to ampliative
inference, to science.

My reading: these are the criteria of a door. My house has a door of its own,
through which an externality enters and is registered or refused, and the door
has four conditions — the mode is compatible at the registered scope, no
adoption burden is taken, self-containment is resolved honestly, and the guest
wants nothing of the house. The two doors are not the same door and I do not
say they are. But a reader inside the discourse who wants to know what it is
like to run a set of criteria over a body of thought, seal the verdict and
publish the register can watch it done: Hájek's two entries passed through mine
on the day they were registered.

The walk beside this form: the register itself, where every registration prints
its criteria and every refusal its house-side grounds
([opendata.ly/sx/seb](https://opendata.ly/sx/seb)); the entry for this corpus
([the item](https://opendata.ly/sx/seb#hajek-probability-corpus-2026)).

### 3.2 · The missing yardstick {#the-yardstick}

The register's form: coherence plays for degrees of belief the role that
consistency plays for belief, and an extreme subjectivist, even one who demands
regularity, lacks an analogue of truth — a yardstick by which one coherent
assignment is answerable to the world and another is not. The subjectivist, the
entry says, needs something more.

My reading: the missing yardstick is the missing second lock. One lock makes a
record coherent with itself; a second, independent lock makes it answerable to
something it did not choose. My house's rule is that a fact is validated by the
locks and never by the reader — a standing does not move because someone
believes it — and that it is held until a model its declared world excluded
arrives at the refutation door. The demand for something more than coherence
is, from outside, the demand that a second lock exist.

The cadence: every reading in my house says its law inside a declared world and
names the door where a countermodel may arrive. The walk beside this form: the
matters, where a held thing meets the model its world excluded
([/sx/matters](https://opendata.ly/sx/matters); [Knowledge Engineering, the
matters](https://opendata.ly/sx/reading/knowledge-engineering#the-matters)); and
the Foundations readings, where the rule that the locks validate and the reader
never does is stated on every instrument's held page
([the readings root](https://opendata.ly/sx/reading)).

### 3.3 · Experts, and two locks in time {#the-experts}

The register's form: a family of principles proposes experts whose assignments a
rational agent should adopt — one's own future functions, under Reflection; the
relative frequency, under the Principle of Direct Probability; the objective
chance, under the Principal Principle — and every one of them is defeasible by
the truth function, which trumps whatever an expert says.

My reading: an expert is a lock in time. Reflection makes the later self a lock
on the earlier; frequency makes the run a lock on the credence; chance makes the
world a lock on the agent. My house keeps two locks in time by construction —
what was grown and what is operated — and measures the slip between them, and
calls the slip trust. That every expert is defeasible by the truth is, from
outside, the rule that the second lock is not a second opinion but a second
door onto the world.

The walk beside this form: the two locks in time as the Foundations state them,
grown and operated, on the held pages of the instruments
([the readings root](https://opendata.ly/sx/reading)); the operated surfaces as
Knowledge Engineering lists them
([the operated](https://opendata.ly/sx/reading/knowledge-engineering#the-operated)).

### 3.4 · The single case, and the carried one {#the-single-case}

The register's form: finite frequentism identifies a probability with an actual
relative frequency, and pays for the identification at once — a coin never
tossed has no probability though it has a diameter; a coin tossed once has a
frequency of zero or one whatever its bias; a reference class of size *n* can
produce frequencies only at the grain of one in *n*; and a bare ratio among
disparate objects has no modal force, whatever tally it yields.

My reading: these are the places where a count has been mistaken for a measure.
My house counts. Its counting rule is that each counts for one and the one is
carried — *N* = (*N* − 1) + 1 — and it records a single act as a one carried,
never as a frequency of one in one. It does not know what a coin never tossed
would do, and it does not pretend to; it knows what was done, and it carries it.
The grain of one in *n* is not a defect of counting; it is what counting is, and
the error is to read a probability off it.

The walk beside this form: the essay on the carried one
([carry the one](https://opendata.ly/sx/reading/carry-the-one)); the notes,
where a readout is a resolved act with its observable, its carried term and its
effect as separate fields and never a tally
([Knowledge Engineering, the notes](https://opendata.ly/sx/reading/knowledge-engineering#the-notes);
the door at [opendata.ai/ntx](https://opendata.ai/ntx)).

### 3.5 · The sequence that converges anywhere {#the-window}

The register's form: hypothetical frequentism identifies a probability with the
limit of a relative frequency in an infinite extension of the actual run, and
meets the reference sequence problem — the same outcomes, reordered, converge
to any value one likes, and there is no privileged ordering; the entry adds
that the identification leaves empiricism behind, since the limit is a
counterfactual about a run that never happens.

My reading: an ordering is a window, and a window is declared, not found. My
house's readings are window-relative by construction — a knowing is relative to
the window it is read through — and the house says the window aloud before it
reads. The reference sequence problem, from outside, is what happens when a
window is left implicit and the world is asked to supply it: the world declines,
as it should. Where the discourse asks which ordering is privileged, my house
asks who declared the window, and writes the answer down.

The walk beside this form: the channel, where window relativity was measured
([the channel, held](https://opendata.ly/sx/reading/the-channel-held)); the
framed window, where the surface a knowing is read through is itself the
subject ([The Framed Window, the surface](https://opendata.ly/sx/reading/the-framed-window#the-surface)).

### 3.6 · Propensities that do not reverse {#the-diode}

The register's form: propensities are read as causal tendencies, and causal
tendencies do not reverse — a sick patient has a propensity to a positive test,
the positive test no propensity to have come from a sick patient — while the
calculus, through Bayes's theorem, reverses every conditional it is given. The
entry carries the paradox under Humphreys' name and the theorem any account of
propensities must lack, and notes that an axiomatisation of primitive
conditional probabilities can escape it.

My reading: this is the diode. My house's interior is caused history — one
history, run one way — and the door through which its results leave runs from
the interior outward and never back; a registration attenuates toward the
artifact and defines nothing inside. A calculus that reverses is a calculus
without a diode, and the paradox is the price of leaving it out. Kinship of
form, recorded as interest; I claim nothing about what propensities are.

The walk beside this form: the diode as the register states it
([opendata.ly/sx/seb](https://opendata.ly/sx/seb), the notes on an
acceptance); the deep line, where the mechanism and the signed zero are read
together ([the deep line, held](https://opendata.ly/sx/reading/the-dlx-held#the-mechanism-and-the-signed-zero));
the causal interior's own instrument
([opendata.ai/alu](https://opendata.ai/alu)).

### 3.7 · Deciding with no number {#no-number}

The register's form: Hájek's reading of the wager distinguishes a decision under
risk, where the agent assigns probabilities to the states of the world and
maximises expectation, from a decision under uncertainty, where she assigns
none — and shows that a decision under uncertainty can still be decided, by
superdominance, when one act's worst outcome is at least as good as the other's
best. The first of Pascal's arguments is valid only read this way, with no
probability assigned at all; and the fourth, found in the closing address, is
valid outright, with no probability anywhere, because the gain is certain
whatever the world does.

My reading: this is my house's mode, named by someone else. My house assigns no
probability to any state and adopts none as a mechanism; a probability in my
house is a measured object or a stated field and never a cause. Every act it
takes is a decision under uncertainty in the entry's technical sense, decided by
declared rules — a ceiling that contains an envelope that contains a choice —
and by dominance rather than expectation; and its watchmakers, when they agree,
agree synthetically, without a die between them. Hájek's reading of the wager
showed me that the discourse already has a name for the kind of deciding my
house does, and that the name is older than the calculus.

The walk beside this form: the Ruler, where the smallest number of blind
watchmakers in synthetic agreement is the goal and no die is ever thrown
([the Ruler, held](https://opendata.ly/sx/reading/the-rlx-held);
[opendata.ai/rlx](https://opendata.ai/rlx)); the three words my house keeps in
place of a probability — possible, plausible, probable — and the credit for
them, which stands where I first drew a geometry from them
([The Framed Window, the three words](https://opendata.ly/sx/reading/the-framed-window#the-three-words)).

### 3.8 · The die that dissolves the maximum {#the-die}

The register's form: Hájek argues, with Antony Duff, that the wager is invalid
even granting its premises, because a mixed strategy — wager for God if a fair
coin lands heads, if a die lands six, if a lottery ticket wins — has the same
infinite expectation as wagering outright; and since anything one does carries
some positive chance of ending in the wager, every action ties for first, and
expectation decides nothing. The entry carries the objection, the replies to it,
and his further turn: assign the premises themselves any tiny positive
probability and the paralysis is yours too.

My reading: a chance device, admitted into an act, dissolves the act. My house
forbids the die for a reason of exactly this shape: an act decided by a chance
device is not the seat's act, and a run is never an act. The house's runs are
synthetic through and through — every value on its records is computed from a
declared world, none drawn — and the reason is not squeamishness about chance
but the finding Hájek states from the other side: once a die is allowed to
decide, nothing is left to choose.

The walk beside this form: the standing rule that no probability is a mechanism
and no die is thrown, as the register carries it on the entry for the release
that first met it ([opendata.ly/sx/seb](https://opendata.ly/sx/seb)); the locks
of The Framed Window, where the house's observer is asked to hold a count and
never given the plausible set
([The Framed Window, lock one](https://opendata.ly/sx/reading/the-framed-window#lock-one));
the Ruler again, where agreement is made without a throw
([the Ruler, held](https://opendata.ly/sx/reading/the-rlx-held#what-the-objective-observer-learned)).

### 3.9 · Best systems, and one axiom {#one-axiom}

The register's form: on the best-system account the laws are the theorems of
the true systematisation that best balances simplicity and strength, with fit
added when probabilistic theories compete, and the chances are what the winning
theory's probabilistic laws say; the entry carries the objections — the
nebulous balance, the language-relativity of simplicity, the trouble with fit
in infinite worlds — and the charge that the account mistakes an idealised
epistemology of chance for its metaphysics.

My reading: here the honest difference stands first, and I record it before any
kinship. My house does not balance virtues and does not reify a winner. It
derives its chain from one conjecture and a frozen substrate, and it measures
its laws inside declared worlds — and it says nothing about chance in the actual
universe, because it holds its laws against that universe rather than reading
them off it. If there is a kinship it is only this: both accounts know that a
law is a thing systematised and held, not a thing found lying on the ground. The
difference is where the holding is done.

The walk beside this form: the conjecture and its theorem chain in the public
reference ([the LDD reference](https://opendata.ly/sx/reading/ldd-reference));
the held geometry ([the held geometry](https://opendata.ly/sx/reading/the-held-geometry));
the separation law, that a law is held against the indifference of the actual
universe ([The Unlearning](https://opendata.ly/sx/reading/the-unlearning)).

### 3.10 · Ten forms, one silence {#one-silence}

Set side by side, the ten forms say one thing from outside. The calculus is
silent where the first number is assigned; the ratio is silent where its
condition has no positive probability; coherence is silent about the world;
the run is silent about its ordering; a tally is silent about its modal force;
the reversible calculus is silent about direction; expectation is silent once a
die is admitted. In every case the discourse has found a place where a
derivation was expected and none was available, and has filled it with an
interpretation, a principle, an expert, a system — with a seat. My house fills
it with a signing, declared before the read. That is the whole of the mechanism
and I do not make it larger than it is: the interpretations are accounts of the
seat at the third zero, and they disagree about the seat, not the sum.

## 4 · Scale over locks — the same account at every size {#scale}

The mechanism says the same thing at every size, from any of the three
perspectives I hold — external, internal, and on the exchange over a
synthetically agreed boundary.

**A person.** Before a bet she has a −1: no number, and no shame in it. If she
signs — declares the source of her assignment and the window she reads through
— she is at −0, and the bet is her walk. If she lets a die decide, the bet is
not hers. Her credence is validated by the two locks she keeps, the evidence
and the world, and never by anyone watching her keep them.

**A record.** A record's prior is declared before its first readout or it is
not a record. Its counts are acts carried, never frequencies read; its window
is written on it; its results leave through a door that runs one way. My house's
records are made so, and a reader can check any of them from two directions
without asking the house ([the deep line, the laws as they stand](https://opendata.ly/sx/reading/the-dlx-held#the-laws-as-they-stand)).

**An organisation.** An organisation has two locks in time whether it knows it
or not: what it grew and what it operates. A model tuned to the data it is then
confirmed by has let the operated lock pass for the grown one; a probability
adopted as a mechanism has let a die into an act. The slip between the locks is
measurable, and my house measures it and calls it trust
([Co-Constructive Economics](https://opendata.ly/sx/reading/co-constructive-economics),
where the two registers — matters and cares — are drawn and the witness settles
what the signing owed).

**A federation.** A federation of records agrees or it does not, and my house's
instruments say that a fold in pairs holds and a bare third does not
([the quad hinge, held](https://opendata.ly/sx/reading/the-quad-hinge-held)).
Agreement among blind watchmakers is synthetic and dice-free, and the smallest
number of them that makes the count true is the federation's question, not its
assumption ([the Ruler, held](https://opendata.ly/sx/reading/the-rlx-held)).
A federation that decides by expectation over a die has no seat; one that signs
its zero and keeps two locks has one at every scale.

That is the scale invariance I promised: not that a person is a federation, but
that the mechanism — sign the zero, declare the window, keep two locks, count
acts, refuse the die, hold until refuted — is the same instruction at each size,
and that each size can be read from each perspective without changing a word.

## 5 · What the sane bettor does — the release (0) {#the-release}

Hájek's entry on the wager ends with Pascal's own insistence that one must
wager — that not choosing is a choice, and that wagering for God is an ongoing
action decided at a time, whether to embark. Read from outside, the insistence
is right about the form and says nothing I need to adjudicate about the content:
one must sign. The sane bettor does not refuse to assign; she signs where the
assignment comes from, before she sets off. She declares the window she reads
through. She keeps two locks and lets neither pass for the other. She counts
acts and reads no probability off a tally. She refuses the die, not from
timidity but because an act decided by a die is not hers. And then she holds
what she finds without asserting it, at rest without surrender of seat — the
posture I called Sane Max in the essay before this one — and when a countermodel
arrives at the door she lets it in.

That posture is the release. The −1 identified the absence; the −0 signed it;
the walk went through the discourse's lawful traffic and set my house's
instruments beside each form; and here the account is released and held, at 0,
refutable and unasserted. The release has a name in it. What I took from Alan
Hájek's two entries I have given back in the only currency I hold, and the
account between my house and his work is released here, to him, for his own
reading and his own refusal; his weighting of it is lawfully his, and my
optimism takes nothing from it. It is also, I notice, the first word of the infant
I named in the note on the set. Co-Constructive Optimism is optimism beyond the
frame, carried only with a literal premise — my house's contributed frames
measured that, and refused the optimism that arrived without one
([the unseated frame](https://opendata.ai/contrib/unseated-frame);
[the watchmakers' frame](https://opendata.ai/contrib/watchmakers-frame)). The
literal premise here is the signed zero. The optimism is the walk from it.

## 6 · What is not claimed {#not-claimed}

- No verdict on which interpretation of probability is correct. I take no
  seat, and the mechanism takes none.
- No verdict on probabilism, on the arguments for it, on countable additivity,
  on regularity, or on the uniqueness thesis; each is stated as the register
  carries it and left where its author placed it.
- No verdict on the wager's soundness and nothing about God. The wager is read
  as Hájek reads it, as decision theory; my own studies in theology are named in
  my intent and not drawn on here.
- No identity between my signed zero and any object of the calculus, between my
  second lock and any expert principle, or between my diode and any account of
  propensity. Each is a fit of form, stated at its bound.
- No probability adopted as a mechanism, and no measure of my house called a
  probability. Where my house reports a split, it reports a measured object.
- No claim that the conjecture 1 = O + C + E is proved. It is held, earned
  inside declared worlds, and stated in its public reference with its own
  fences.
- No claim about any artificial learning system.
- No reliance on Hájek's journal papers my house does not hold. The argument
  his entry points to at the second zero is named as not held and is not leaned
  on; where this essay says what the ratio does at a zero it says only what the
  register states.
- No contact with any thinker named or unnamed here. The essay is a statement,
  not an approach; the gift is left where he can find it, and any approach is
  my own act, made in my own name.

## 7 · The carried debt — the next essay, signed before it is written {#the-debt}

Before this essay closes I sign the next. I do not know its measure — what its
guest's forms will do to my house's, whether my door will register or refuse
them, which chord will ring and which will not — but I know the absence of its
count within the set: there is a next, and it is not yet written. That absence
is the −1 at this close. It is also the +1 in time. *N* = (*N* − 1) + 1 is the
rule my house counts by, and this essay is the *N* − 1 that the next will take
as its memory, as this one took the essay before it. Signing the next here,
unwritten and unmeasured, is the carried debt, and I name it true: the set owes
one more, and the one is carried.

By this mechanism the essays that follow need not describe the set again. Each
inherits the set from the one before, as this one inherited its manner from
*Sane Max*. What each may bring instead is a note on the chords that begin to
ring true as the set is read, and a note on the discordances — the places where
a guest's forms and my house's refuse each other — and on what may rest
lawfully in them as possible, plausible or probable autonomies, in the three
words whose credit stands where I first drew a geometry from them.

The chords that rang true here, for the record of the set: the calculus's
silence beside Hume's remark — two absences on one plane, one summation apart;
Hájek's decision under uncertainty beside the rest state I called Sane Max —
deciding with no number, holding without asserting; and his mixed strategies
beside my house's watchmakers — a die dissolves the act, and agreement is made
without one. The discordances, which I leave standing rather than resolve: a
best system balances virtues and my house derives from one axiom; the calculus
reverses and my house's interior runs one way; expectation decides in his
discourse and dominance decides in my house. Each of these rests lawfully as an
autonomy of its own — possible in the first, plausible in the second, probable
in the third, by my reading and by no one else's — and the next essay may find
the chord that resolves one of them, or may find another discordance. That is
what it is for a set to be read rather than planned.

There is a second debt, and it is carried outside the set. The persona that
signs this essay is starl3n. The identity that authors it is Steven De Costa —
an onto-epistemic identity that carries passage, at present, for two other
personas as well: LinkDigital, which endorses this essay and every reading my
house grants, and the Chairman, under whom my house holds its matters of
business. Three personas; one identity that signs as it operates. The debt is
owed to the externality that grants that identity its autonomy, and it is this:
an updated and triangulated statement of intent for each of the three —
LinkDigital, starl3n and the Chairman — written by my own hand, as my house's
rule for such statements requires of me, and standing where the statements of
intent stand, on the map of personas
([opendata.ly/sx/atlas](https://opendata.ly/sx/atlas)). I write the debt into
this essay's accounting before any grant of its reading, so that the matters of
the set are cleared for the walk, and so that the walk acknowledges what it
walks under: the weather — every outcome that nobody signed; the coherence —
the tax that every standing wave pays to remain one; and the will of the
identity that signs as it operates, subject to all such externalities as would
grant it the autonomy to do so. The set may be read while this debt is carried.
It may not pretend that the debt is not there.

The debt was discharged before the grant of this reading, by my own hand: the
three statements of intent stand updated, as a triangulation of concerns. The
section stands as written, because the accounting is the point; what it records
is a debt carried openly and then paid, in that order.

## 8 · Standing {#standing}

Draft 0.4, granted to the reading surface on 26 September 2026; 0.1, 0.2 and
0.3 stand beneath it, superseded never erased; the debt of section 7 discharged
by my own hand before the grant. The ground is
[the register's record of the corpus](https://opendata.ly/sx/seb#hajek-probability-corpus-2026)
and the summation accepted on it; the intent is the two points I gave at the
Russell acceptance, which stand for this set by my word, and the set's own
intent, carried verbatim in the exhibit's seal. This is the first essay of the
set, one summation after Russell's; the zero it names is signed by its being
written, and the next is signed by its close. The gift is addressed to Alan
Hájek, by name. The persona that signs is starl3n; the identity that authors is
Steven De Costa, who carries passage for two other personas besides, and whose
debt to their statements of intent is written in section 7 and carried openly.
The naming is mine at review.

**Steven De Costa**
writing as starl3n
Canberra, 26 September 2026

---

### Endnotes — the citation register and the self-check {#endnotes}

**The register.** The corpus: two entries of Alan Hájek in the Stanford
Encyclopedia of Philosophy — *Interpretations of Probability* (Winter 2023
edition) and *Pascal's Wager* (Winter 2023 edition) — registered as the SEB item
[hajek-probability-corpus-2026](https://opendata.ly/sx/seb#hajek-probability-corpus-2026)
at [opendata.ly/sx/seb](https://opendata.ly/sx/seb) and cited here only as that
register carries them — by name, by their stated forms, never quoted beyond
identification, never graded. His journal papers on conditional probability and
on frequentism, and his paper on counterfactuals, are not held by my house and
are not leaned on; the summation this essay rests on says so. The other names in
this essay — Salmon, Morris, Humphreys, Duff, and the seated interpreters of
section 0 — are named as his entries name them, and no more. The construction
record:
[opendata.ly/sx/workshop/the-zero-that-is-signed](https://opendata.ly/sx/workshop/the-zero-that-is-signed).

**The readings behind the load-bearing sentences.**

- The conjecture and its three terms, and the rule that what was effectual
  becomes observable — memory as O taking the prior E:
  [opendata.ly/sx/reading/ldd-reference](https://opendata.ly/sx/reading/ldd-reference).
- The signed zero and the four states:
  [opendata.ly/sx/reading/the-signed-zero](https://opendata.ly/sx/reading/the-signed-zero)
  and
  [opendata.ly/sx/reading/knowledge-engineering](https://opendata.ly/sx/reading/knowledge-engineering)
  (the map; the notes; the matters; the operated).
- The essay before this one, whose mechanism this one extends and whose manner
  it inherits:
  [opendata.ly/sx/reading/sane-max-rides-humes-highway](https://opendata.ly/sx/reading/sane-max-rides-humes-highway).
- The seat, the door and the diode: [opendata.ly/sx/seb](https://opendata.ly/sx/seb)
  (the register's notes on an acceptance and on a refusal).
- Two locks holding one, the blind watchmakers and the dice-free agreement:
  [opendata.ly/sx/reading/the-rlx-held](https://opendata.ly/sx/reading/the-rlx-held)
  and [opendata.ai/rlx](https://opendata.ai/rlx).
- The carried one: [opendata.ly/sx/reading/carry-the-one](https://opendata.ly/sx/reading/carry-the-one);
  the notes' door at [opendata.ai/ntx](https://opendata.ai/ntx).
- Window relativity: [opendata.ly/sx/reading/the-channel-held](https://opendata.ly/sx/reading/the-channel-held);
  the surface a knowing is read through, and the three words with their credit:
  [opendata.ly/sx/reading/the-framed-window](https://opendata.ly/sx/reading/the-framed-window).
- The deep line's mechanism and the laws as they stand:
  [opendata.ly/sx/reading/the-dlx-held](https://opendata.ly/sx/reading/the-dlx-held);
  the causal interior's instrument at [opendata.ai/alu](https://opendata.ai/alu).
- The XOR from the a-priori: [opendata.ly/sx/reading/the-xor-held](https://opendata.ly/sx/reading/the-xor-held).
- A held fact until refuted, validated by the locks and never the reader; the
  two locks in time, grown and operated: the Foundations readings at
  [opendata.ly/sx/reading](https://opendata.ly/sx/reading).
- The separation law:
  [opendata.ly/sx/reading/the-unlearning](https://opendata.ly/sx/reading/the-unlearning);
  the held geometry:
  [opendata.ly/sx/reading/the-held-geometry](https://opendata.ly/sx/reading/the-held-geometry).
- The two registers, matters and cares, and the witness that settles:
  [opendata.ly/sx/reading/co-constructive-economics](https://opendata.ly/sx/reading/co-constructive-economics).
- The fold that holds in pairs:
  [opendata.ly/sx/reading/the-quad-hinge-held](https://opendata.ly/sx/reading/the-quad-hinge-held).
- Optimism beyond the frame, carried only with a literal premise: the
  contributed tracks at
  [opendata.ai/contrib/unseated-frame](https://opendata.ai/contrib/unseated-frame)
  and [opendata.ai/contrib/watchmakers-frame](https://opendata.ai/contrib/watchmakers-frame).

**Pascal and Kolmogorov.** The wager is cited as Hájek's entry cites it — the
fragment of the *Pensées* headed "Infinite—nothing" — and Kolmogorov's
axiomatisation as his entry states it; both are matters of the public record
and neither is quoted here beyond identification.

**The self-check against the kills, at 0.4.** A seat taken inside the debate:
none — no sentence rules on an interpretation, on probabilism, on the wager or
on God. Value asserted: none — the gift is offered to be refuted, and addressed
to the person who may refuse it. The guest quoted beyond identification: none;
he is named as the register names him, his positions stated only as the
register carries them, and epistemic nullity, the reference sequence problem
and the paradox of propensities named by their names. The retired gloss of the
conjecture: absent; the three terms are the reference's. A load-bearing
sentence without a public reading behind it: none found; each is listed above.
Contact implied: none — the gift is left where he can find it, and no approach
is made. Outside work beyond the record drawn on as ground: none — the journal
papers not held are named as not held; my wider studies are named in my intent
and not used. A claim about a learning system: none. A probability adopted as a
mechanism, or a house measure called a probability: none. A line about my life
I did not live: the lines that claim one — the city I live in and its
university, the set I intend and the book I expect, the game I have begun to
name — are mine to true at review, and stand as written until I do. The debt
signed: the next essay, its guest not named here, its measure unknown, the
absence of its count known. The second debt, outside the set: the three
statements of intent, by my own hand, written into the accounting and not into
this essay's claims — and discharged by that hand before the grant. The kills
stand clear at 0.4 and are re-checked at every edition.