Derivation of Value

A reader's guide to the whitepaper. Intuitions, motivations, and the path through the formalism.

Naming the operator the series turns around — the move from staking contingent quantities to staking what such quantities are derived from.

The negentropy derivation. Order is the one substrate the second law guarantees is scarce and refuses to attest — so it is certified not by proof but by gauge-fixing the adversary's freedoms until only honest sections survive.

The value-as-vector derivation. Money is a scalar only by the violence of accounting; its suppressed dimensions fall into a small grammar of algebraic kinds, and to stake the vector rather than its collapse is to derive value from the structure of the instrument itself.

The civilizational-capacity derivation. Value denominated against a people's position on the Kardashev scale — the order of magnitude at which it can act — and against Barrow's inward complement, the grain at which it can act. The outward ladder is attested by report and inherits the oracle; the inward ladder leaves artifacts that attest themselves. Capacity is the conjunction. v0.4 notes the one outward rung that attests itself.

The turn from derivation to construction. A running mechanism is a claim that can be false — the program's first refutation surfaces, where the operator meets a machine and an adversary. What the working examples already show, what they do not yet, and the program's first conjecture handed to a machine and returned half refused.

The durability volume. The quantum adversary does not pose a new problem — it reveals which of the program's hardnesses were borrowed. A substrate-by-substrate audit of what survives the machine, and the operator recovered one last time at the cryptographic layer.

The volume that names what the program had derived four times without noticing: the line between what an action asserts and what it cannot help but leave. Every substrate the anthology named met the same wall, and the wall was never the substrate — it was the attestation. Residue is the general answer, and the honest cost of saying so is that the program has named more substrates than it has.

A framework for Goodhart-asymptotic mechanism design. Reward gated on the conjunction of verifiably independent projections of a structural substrate. v0.4 reports that independence was never the binding constraint: a projection can be perfectly independent of the kernel and still free to forge, and §7.1 has been stating one of two requirements. v0.5 records the amendment *Independent and Expensive* made to Proposition 4.2: publicity-positivity is non-decreasing always and strict only while unclaimed paid work remains. v0.6 amends Definition 2.6 per *One's Own Anchor*: each projection must be approximately independent of its own anchor as well as of the others, or the mechanism is its anchors. v0.7 records the candidate formalisation of ι in *A World with Memory* §3: ι and τ as ratios of conditional information on a pool that is the entropy of the world given the public record plus the anchors' own, meeting the four constraints the corpus had accumulated; §7.1's fork resolves as a divergence for content and Fisher information for the gate's resolution.

A Goodhart-asymptotic incentive mechanism for decentralized inference. Sheaf-theoretic, with a spectral extension via Proof by Resonance. v0.5 absorbs three measurements against the spectral layer: its static form is redundant with the kernel, its temporal form is independent but free to forge, and the autopoietic-cult defence is inverted — a cluster that has stopped scores five times an honest network that has not.

Certifying negentropy by composition rather than proof — four anchors, each eliminating one gauge freedom of the adversary, until only honest sections survive the quotient. v0.2 reports the first execution of the §5 test suite: test (i) run with the unique-encoding anchor removed, returning the predicted non-truthful sections with a price attached, and §4.4's prohibition on anchors moonlighting as certificates of order turning out to be load-bearing rather than ascetic. v0.3 adds tests (iii) and (iv): the specimen's spectral gap is 0.0038 and its cheapest coalition has at most four boundary edges, so the §2 design target stands and the program's own substrate fails it; the transition anchor of *The Second Pool* meets (iv) by construction. v0.4 records the target met by *Every Basis at Once*: a beacon-drawn cover raises the gap thirteenfold by redrawing one edge in eight, at the price of the substrate's spectral dimension, and the coalition's price is the cut, of which the gap is a loose bound.

Mechanisms nest, and the framework has no theory of nesting. Koestler's holon supplies the structure, cellular sheaves supply the mathematics it always lacked, and the obstruction to composing conjunction-gated mechanisms is a cohomology class — every department hitting its metrics while the firm fails turns out to be contextuality, in Abramsky's exact sense. Absent that gluing, a holarchy's security is the minimum over its levels rather than the product. v0.3 records the minimum measured: on the program's own holarchy it is zero for every block coalition, and the conjunction that H¹ = 0 restores is worth one boundary, counted once, not a product.

The program has named the Sybil attack four times without using the word. Naming it lets the anchors be read as a proof: conjunction-gating amplifies a resource floor's Sybil resistance by a factor of 1 + (K−1)ι, and by nothing at all if the projections are redundant. The gate caps the adversary's fleet; only convexity above it makes splitting a loss — and diminishing returns, the conventional anti-whale choice, subsidises Sybils. v0.4 adds a second way for richness to buy nothing: a projection contributes in proportion to its own trace gap, so a perfectly independent one that is free to forge amplifies no more than a redundant one — and the specimen has been measured. v0.5 records §8.4 answered by *The Budget Across Levels*: the criterion for a Sybil refinement is correct and not checkable from the reading, and the level of presentation changes receipts alone.

Every department hits its metrics and the firm fails. That is not an anecdote about organisations — it is contextuality, in Abramsky and Brandenburger's exact sense, and it comes with thirty years of mathematics attached. Importing it grades composition failure into three strengths, supplies a route past the affine-gate hypothesis, and settles what a vanishing obstruction can and cannot certify. v0.3 records Conjecture 3.1 computed by *The Obstruction, Computed*: it holds, no distribution has to be chosen, gate models carry two grades and not three, and a vanishing class licenses nothing at a measured rate of two in five. v0.4 records §6 closed by *Every Basis at Once*: contextuality is a resource against exactly the adversary commitment already defeats.

The framework gives richness as a ceiling on how secure a mechanism can be made. Ashby's law supplies the floor, and the floor is set by the adversary rather than by the substrate. Between them, substrate selection stops being an optimisation and becomes a feasibility question: some substrates are simply too poor for some adversaries, and no cleverness repairs them.

Volume V named residue and left one half of its security argument as a design obligation it could not discharge. This is the discharge. Grassé's word for coordination-by-trace splits along exactly the line the volume drew — stigma against ergon, sign against work — and the seam has a price: the trace gap τ, the ratio of forging a residue to earning it. v0.3 measures it. Two ceilings, reached by unrelated routes, agree: coherence never contributes to the trace gap. Whatever τ a mechanism has comes from its anchor, and the coherence term appears only in the denominator. v0.4 records that §8.2 has a specimen above one, and that it is a free ride.

The program named its own worst failure mode — 'a consistent fiction is consistent', collapsing to 'autopoietic-cult attractors' — and never studied it. Maturana and Varela's theory is the study. A coherence mechanism is operationally closed in their exact sense, its fiction space has dimension d, and closing it costs d scalars of contact with the world, once. The program's anchors do not supply them, by an explicit rule; the inward axis might.

Gluing the Gates declined to supply a composition operation. Open games have one, and their Nash equilibrium condition is itself compositional — which is more than was expected and less than is needed. Coutility fixes the first of that paper's two failure modes and leaves the second exactly where it was, because the framework does not gate on equilibrium. It gates on admissibility, and admissibility is a constraint on states, not a fixed point of best responses.

Volume IV read the post-quantum migration as the program's founding operator performed at the crypto layer — borrowed structure abandoned for derived bulk. The standards say otherwise. NIST's primary signature algorithm rests on module lattices, which is structure, and the hash-based scheme that fits the volume's description is designated a backup. The migration moves between structures rather than away from structure, the difference is priced in bytes, and the hybrid the program already recommends turns out to be the operator it did not know it was applying.

The framework's reordered central question — is any projection both independent and expensive to forge? — answered in a model small enough to prove things in: independence and trace gap are claims on one budget, the substrate's paid work, and disjoint claims on a finite pool cannot both be large. Richness partitions the budget; nothing multiplies it. The escape is an anchor per projection, and it is an escape by purchase, not by cleverness — measured on the sheaf, where the dial that buys persistence its gap spends its independence in the same stroke. v0.2 records that §8.2 and §8.4 have been run, in that order, by *The Second Pool*: the purchase returns the arithmetic to the decimal, and the purchased projection evidences its anchor. v0.3 records that §8.3 has been run by *The Budget Across Levels*: receipts sum, interfaces nest, and the evidencing budget of a holarchy is its finest level's.

The purchase the exclusion principle priced, made on the program's own substrate: a second anchor — Gauge-Fixing's delay chain, modelled for the first time — gives the temporal projection a pool the kernel's encodings do not touch, and independence returns to 1.000 with both gaps positive and summing to the two anchors' share of the budget, as the arithmetic said it would. What the arithmetic did not say is what the purchased projection evidences. Run in the order the predecessor prescribed, the gate comes first and fails: under induced maps a persistence reading measures how uniformly the network moved, uniform motion is a time-dependent gauge, and so coherent change and no change score identically — a frozen participant with a paid chain is admissible and out-scores honest persistence nine times over. The reading that refuses the cult is the chain read cohomologically, and it carries nothing but its anchor. v0.2 records that §9.1 and §9.5 have been run by *One's Own Anchor*: evidencing is independence from one's own anchor, every declared-frame reading has none as a theorem, and the measured map outside the class evidences the world's change since the last public record. v0.3 records that §9.3 has been run by *The Budget Across Levels*: each level's receipts add and every one is idle, the native pool nests to the finest level, and the minimum over levels is zero for a block coalition.

The idle projection asked for a third condition and there is none to add: a projection's claim on the work its substrate exists to do is its independence from its own anchor times its trace gap, so the framework's two conditions suffice once independence is demanded against the reading's anchor as well as against the other readings — and by that test every reading computed from declared restriction maps evidences nothing, as a theorem. The one reading left outside the class, maps fitted from data at overlaps, does evidence something, and what it evidences is the world's change since the last public record: overlaps are where a coalition copies its neighbours, publishing them before commitment turns the coalition's boundary into a subsidy, and committing first prices only the innovation. The native cost falls on the boundary alone, and on the program's own complex a coalition of any size has at most four boundary edges. v0.2 records that §9.3 has been run by *The Budget Across Levels*: the holon's boundary and its constituents' are the same edges, presenting coarse saves receipts alone, and the both-sides-controlled criterion is not a residual. v0.3 records §9.4 measured by *Every Basis at Once*: the expansion-versus-richness curve is a table, and the two are one purchase with opposite signs. v0.4 records §9.1 answered by *A World with Memory*: against the optimal record-only attacker the price is a step in the world's innovation given the record, a world with memory shrinks that innovation at fixed drift until the coalition passes at zero derivations, and the gate's resolution is the honest derivation variance, so evidencing is priced in inverse proportion to how much there is to evidence.

Three papers left the same question at the same door: whether a holarchy's security budget is the sum of its levels' budgets or the minimum. It is neither, because the question was asked of the wrong pool. Receipts are disjoint per level and sum, and every one of them is idle; native work nests, since a coarser level reads a subset of the finest level's interfaces and one derivation serves every level that reads it, so the evidencing budget of a holarchy is its finest level's boundary however many levels stand above it. A coalition's boundary lies entirely above its own level and halves per level exactly, so under a mechanism the adversary may satisfy at one level of its choosing it presents at or below its own level and pays nothing; under the conjunction it pays the every-edge price, which the sum over levels reproduces to the edge. What presenting as a holon saves is receipts. And the criterion that would tell legitimate composition from a Sybil refinement — that the adversary controls both sides of an interface — is exactly the condition under which an interface evidences nothing, and it is not a residual: the fiction glues better than honesty, and the coalition buys honest-level disagreement with its own noise for free. v0.2 records §9.1 built by *Every Basis at Once*: a beacon-drawn cover restores the per-identity cost to half of honest at the mixing lemma's value, at the price of the substrate's spectral dimension.

No Global Section rested its quantitative content on a conjecture it did not test: that the cohomological obstruction of Abramsky, Mansfield and Barbosa applies to gate scenarios with no affine hypothesis on the gates. It does, and no distribution has to be chosen for it, because the obstruction is computed on the support alone and a gate model is a support. Built on an answer key the literature settled — Bell extends everywhere, Hardy's one non-extending section is missed, the PR box is obstructed at every section — the instrument is then pointed at holons with a threshold on a spectral quantity for a gate. Every strongly contextual gate model is seen; two in five logically contextual ones are invisible to the invariant at every section; the integers and the rationals never disagree; and what decides a miss is how many open gates sit in a row. A vanishing class licenses nothing, now with a rate attached. v0.2 replaces the first reading of what decides a miss with a theorem: on a cyclic cover the obstruction is undirected reachability in the bundle diagram with the section's layer removed, extension is directed reachability, the two differ only through a relation in the chain that is connected without being a union of complete bipartite blocks, and the obstruction is independent of the coefficient ring. v0.3 records §9.2 answered off the cycle by *The Ring and the Chord*: there the ring matters — the square and GHZ vanish over the rationals — and a miss is not a reversal but the absence of a signalling node.

On a cyclic cover the cohomological obstruction is a graph search and the coefficient ring does not matter. Off the cycle both statements fail, and they fail on the literature's own answer key: the Peres–Mermin square and the GHZ model are obstructed at every section over the integers and mod 2 and vanish at every section over the rationals, because an all-versus-nothing argument is a parity argument and the integer obstruction on such a model is 2-torsion. On random no-signalling models with chords the invariant is nearly blind, seeing one logically contextual model in five, ten or forty; drop the no-signalling condition and it sees most of them, and on a holarchy of four three-port holons with tolerance gates it sees every one. Gate models signal, and signalling is what the invariant sees. The setting the literature built the invariant for is the setting in which it is weakest.

Choose the measurement basis after the participants commit, so that a fiction cannot be consistent in every basis: the quantum reading of the anchor problem, taken seriously. It divides in two. A committed participant is non-contextual by construction, so contextuality can appear only where commitment is absent, and there it is the signature of a copier tailoring its answers to each partner — real, computable, and dominated by commitment, which refuses it for free. What the basis buys is the cover. If the beacon draws who compares data with whom, a coalition's boundary is a random cut that grows with the coalition: the per-identity cost that a modular substrate had let fall to one is restored to ninety-six against honest's hundred and eighty-six, with the reading run to confirm it. The price is the substrate's spectral dimension, which leaves as the cover is redrawn, so expansion and richness are one purchase with opposite signs, and the design curve One's Own Anchor asked for is a table.

One's Own Anchor left its one surviving reading with a memoryless world, and its coalition was the weakest attacker in the class that holds nothing but the public record. Against the optimal predictor the reading's price is a step in the world's innovation given the record, and a world with momentum shrinks that innovation at fixed drift until the coalition passes at zero derivations and, past a point, sits below honest: the record is a better model of the world than a fresh derivation. The budget the exclusion principle partitions is the entropy of the world conditional on everything public, plus the anchors' own; ι and τ are ratios of conditional information, and the four constraints the corpus had accumulated on ι are met by that reading in one line each. The gate's resolution is the honest derivation variance, so refusing the predictor on a world of innovation q costs every honest participant k of order σ²/q derivations per prompt: evidencing is priced in inverse proportion to how much there is to evidence. Fresh prompts restore the pool at any memory, at the semantic cost the cover line named and did not price.

The corpus has held one position favourable to Bitcoin since Sign and Work, and held it as a measurement: proof of work is the calibration object at a trace gap of one, and the program has no specimen above it. This paper takes the favourable position at full scope and points it at the least promising object available — a marketplace of meme artefacts, assets whose only world is their own record. The market's native readings are signs: volume, holders and floor grade at or below zero, and below zero means the reading pays its forger. A venue on a proof-of-work chain has three readings that are not signs. Blockspace valued at the clearing rate grades at one under congestion and is vacuous without it. A burn grades at one and builds nothing. And depth — the work piled on an artefact since it was buried — grades at d/κ, ten thousand per block on the specimen: the first reading above one the corpus has measured. It is above one because it is a free ride. Socially the mound is proof of work's one again; the inscriber's gap is that one divided by the share of the mound the inscriber paid for. The favourable position breaks in four places, all recorded: the declared fee is a sign; the mound is free to move above one half of the hashrate, and the chain cannot see where one half is; the distance to one half is a flow while what is buried is a stock; and who reads the mound, and what counts as an artefact under it, are conventions. Kar-Coin is amended.