Prior Art and Inherited Limitations¶
Labour-currency schemes have died of the same two or three wounds for almost two hundred years. A framework that can say exactly what it does differently from Robert Owen is far better founded than one that cannot.
This page is the audit. For each ancestor: the path taken, what killed it, and what this framework does instead — with an honest verdict on whether the wound is closed, improved, or still open.
This is an open research project, not a finished model, and the verdicts are written as a progress report rather than a defence. Where a front is unexplored the page says so and says what would settle it. Its companion, What Anchors a Unit of Account, audits the anchors in use today rather than the ancestors.
Status — author sign-off, 2026-08-15
This page publishes three positions that are intellectual commitments rather than repo maintenance: the price-as-floor reframing, the demotion of system-wide inflation-impossibility to a within-collective floor property with an ε→1 asymptote, and the objectivity → transparency language pass (workstreams C and F of the reconciliation). They are adopted on the grounds that they are the more defensible case, which is the reconciliation's own argument: a weaker true claim beats a stronger false one, and every ancestor that reached for physical objectivity died on that rock. The original inflation-impossibility result is not deleted — it is documented here as the ε→1 limit case.
This rests on two pieces of unpublished working research: an adversarial historical autopsy of labour-currency schemes, and the reconciliation that autopsy forced. Their conclusions are stated here; the working documents themselves are not part of the repository.
The seven recurring failure modes¶
| # | Failure mode | Killed |
|---|---|---|
| 1 | Valuation — no neutral way to set the hour-content of a good a priori | Owen; the trudoden; and the target of Marx and Hayek |
| 2 | Skill differential — equal hours drive out skilled labour; unequal hours decree a hierarchy of human worth | Owen; the limit time banking survives by avoiding |
| 3 | Founder dependence — the system secretly runs on one trusted human | Ithaca HOURS; most LETS |
| 4 | Circulation sinks — currency pools where earning and spending are unbalanced | Ithaca HOURS |
| 5 | Adverse selection / free-riding — participants dump the bad and extract the good | Owen; LETS |
| 6 | Sovereign suppression — a complementary money that works is crushed to protect the state monopoly | Wörgl; the Wära |
| 7 | The objectivity mirage — anchoring to a physical unit confers measurability on the anchor, not on the human valuations it must carry | Technocracy |
The direct ancestors¶
Josiah Warren's Time Store — Cincinnati, 1827–1830¶
The path. Goods priced at the labour time embodied in them plus a markup for materials; payment in "labour notes" promising hours of the holder's own work. Warren ran it about three years, considered the point proven, and closed it.
The limit. It never had to compare a baker's hour to a physician's, because it stayed small enough that Warren's own judgement did the work a system would later have to do by rule.
What this framework takes from it. The honest reading is a warning about scale: the labour-hour works at the scale of a village with a trusted arbiter, and the hard problems begin exactly when you remove the arbiter. This framework proposes to remove the arbiter at civilizational scale, so every mechanism that was once a person's judgement has to become a rule that survives inspection.
Robert Owen's National Equitable Labour Exchange — London, 1832–1834¶
The true ancestor and the sharpest warning. Opened 3 September 1832 on Gray's Inn Road. Roughly 445,501 labour-hours deposited in the first seventeen weeks. Every branch closed within two years.
The path. Deposit goods, receive labour notes denominated in the hours the goods were said to embody.
What killed it — three mechanisms, all still live questions:
- Valuation. No reliable way to determine how many hours a good "contained", and a chronic shortage of competent valuators. The price a market would have discovered had to instead be declared, and every declaration was contestable.
- The skill differential. The promoters could not resolve the skilled hour against the unskilled one. Contemporaries named this as a direct cause of collapse.
- Adverse selection. Tradesmen dumped unsaleable stock for notes, then used the notes to extract genuinely saleable goods. Bad inventory drove out good.
What this framework does instead. All three are addressed structurally rather than by better administration — see the verdicts below. Owen is not a distant cautionary tale; he is a controlled experiment on two of this framework's load-bearing mechanisms, and both failed for him.
Ithaca HOURS — New York, 1991–c.2015¶
The path. Paul Glover pegged one HOUR to \(10**, the local living wage. About ninety businesses at the start, growing into the hundreds; at peak over **\)100,000 in circulation. The longest-running modern local currency in the US.
What killed it. Two structural causes, not incidental ones. Founder dependence — Glover worked as a full-time "networker", personally troubleshooting circulation and buying back HOURS from businesses holding too many; when he left town the system withered. And circulation backlog — HOURS pooled at co-ops and cafés that consumers paid but that could not spend locally. One manager's phrase: "drowning in Hours."
What this framework does instead. A standing, public EOH register means there is always a registered obligation to spend against, which removes the "nowhere to spend it" boundary that created Ithaca's sinks. Founder dependence is answered by design — sortition-selected assessors, a minimum assessor count, adversarial review, and a revalidation clock — but see the verdict: designed against is not the same as modelled.
Time banking / Time Dollars — Edgar Cahn, 1980s–present¶
The path. Rigorously egalitarian: one hour given equals one hour earned, whatever the hour contained. An hour of legal advice trades one-for-one with an hour of raking leaves.
Why it survives. It refuses to do what this framework wants to do. It stays out of the commercial economy and never prices skilled against unskilled labour — so it never confronts failure mode 2 at all. In the US it also survives because the IRS treats it as volunteering.
The honest reading. The equal-hour rule is viable precisely to the degree it stays non-commercial and small. This framework wants the equal-hour's moral clarity and a skill multiplier and civilizational scope — the one combination history has not delivered. That is the bet, and it should be named as one.
The velocity cousin¶
Wörgl — Tyrol, 1932–1933¶
The path. Following Gesell, scrip issued 31 July 1932 that lost about 1% of value per month unless revalidated with a stamp — a demurrage punishing holding. It circulated perhaps 8–14× faster than the national shilling, funded public works, and local unemployment fell about 16% while it rose about 19% nationally.
It did not fail. It was killed. The Oesterreichische Nationalbank shut it down on 1 September 1933 to protect the currency monopoly, and unemployment promptly returned. The Wära in Bavaria had been outlawed the same way in November 1931.
Two lessons, both taken. An anti-hoarding rule can genuinely raise velocity — mild support for the spirit of Condition III (Zero Interest). And far more importantly: a money that works and begins to scale threatens the sovereign, and the sovereign will use law. A calibration detail is worth keeping: American stamp-scrip copies that set demurrage at 2% per week were simply not accepted — the anti-hoarding penalty has a ceiling above which people reject the money.
The survivors, and what durability costs¶
LETS — from 1983¶
Mutual-credit ledgers with no issued notes; members run positive or negative balances as they trade. Many persist. Those that died, died of volunteer coordinator burnout, members who ran permanent negative balances and never reciprocated, and an inability to grow past the point where everyone roughly knows everyone.
WIR Bank — Switzerland, 1934–present¶
The great survivor, and the most instructive case in the lineage. Ninety years of durability from being everything the failures were not: complementary (never trying to replace the franc), business-to-business (participants extend each other credit and have real reasons to both earn and spend), and countercyclical (used more in downturns, serving a function the franc cannot).
The uncomfortable conclusion. The survivors are the humble ones that supplement the dominant money. The ones that tried to be the money all died. This framework belongs, by ambition, to the second category. That is not a reason to abandon it. It is a reason to treat the transition as the central engineering problem rather than a closing flourish — and the transition is this framework's thinnest area — it is an open front, modelled nowhere, and the historical record is most directly predictive here.
The objectivity mirage¶
Technocracy's energy certificates — 1930s¶
The path. Howard Scott's Technocracy Inc. proposed distributing goods via certificates denominated in physical energy units, allocated by energy accounting rather than price. Never implemented.
The flaw, and it is pointed at this framework. Energy is measurable. How much energy a human need is worth is not. The certificate scheme never bridged that gap, and the same seduction operates whenever this framework leans on entropy and thermodynamics for prestige.
What this framework does instead — and this one is checkable. Rather than
asserting the difference in prose, the framework publishes a machine-checked
count of how little of itself is physics: only a handful of constants are
tagged physics, and every other constant says what kind of claim it makes.
Measurement debt — roughly two constants in five — is declared on the front of
the provenance page with the live figure, and a
test with no allowlist fails the build if a constant is untagged or a tag
overclaims. Technocracy claimed physical objectivity and could not be checked;
this framework claims transparency and can be. Run eoh provenance check.
Soviet labour accounting — War Communism 1918–21; the trudoden 1930–1966¶
The largest real experiment in administering an economy in labour and natural units. War Communism collapsed into hyperinflation and barter; the NEP restored money in 1921. The trudoden, or "labour-day", became a byword for arbitrariness — precisely because the conversion between a day of labour and the goods it earned was set administratively rather than discovered. When the state sets labour-value by decree at scale, the result is not justice but a new opacity, and the people on the wrong end of the multiplier have no market to appeal to.
The two theoretical autopsies¶
Marx against John Gray, 1859 — the critique from inside the labour theory¶
Devastating precisely because Marx agreed that labour creates value. Gray had proposed labour-time as the direct measure of money. In A Contribution to the Critique of Political Economy, Marx took it apart: under private production each person's labour is private, and becomes social only through exchange. Socially necessary labour time cannot be measured in advance; it is discovered retroactively in the totality of exchanges. Gray's scheme tries to treat every private hour as immediately social by decree while keeping the commodity production that makes labour private in the first place.
Why this is the sharpest item on the page. Registering an obligation, certifying the labour that fulfils it, and minting a unit is an attempt to declare labour socially valid up front. Marx's argument says that is impossible unless you have genuinely replaced private commodity production with conscious planning of what labour is needed.
Mises (1920) and Hayek (1945) — the calculation problem¶
A central body cannot compute an economy's allocation, not from stupidity or corruption but because the relevant knowledge is dispersed across millions of people and exists nowhere in aggregate. Prices are not a veil over "real" labour values; they are the only known mechanism for compressing dispersed knowledge into an actionable signal.
The convergence. Marx from the left and Hayek from the right arrive at the same wall from opposite directions: the social value of an hour cannot be set a priori. Any system that tries — Owen's valuators, the trudoden, an assessment body — must either fail or claim an impossible quantity of knowledge. The answer cannot be a better committee.
What this framework does differently¶
The reconciliation found that the framework ran on two engines pulling against each other: objectivity (price is measured labour content; inflation impossible by definition) and transparency plus discovery (an auditable floor with a market above it). Every place the framework scored well rode the second engine; every place it scored badly rode the first.
The decision was to commit to the second and demote the first to a floor-level, within-collective property. Three consequences drive everything below.
- The computed price is the floor price, not the price. It is the price
below which the collective guarantees work is always available and always
paid. Discovery sets everything above it. In code:
core/prices.floor_price()with an explicitmarket_premiumseam; the CLI column readsfloor_price. - The system is a federation of Coasean collectives (Coase, 1937) — islands
of conscious planning in a sea of exchange. Inside a firm there is no internal
market and none is needed; between firms, exchange discovers value. The
collective count is emergent from ε, not set:
coasean_collective_count()falls from many collectives at subsistence to one near post-scarcity. - Contestability is the invariant held constant across the whole arc.
The verdicts¶
| # | Failure mode | Verdict | The mechanism, and its residual |
|---|---|---|---|
| 1 | Valuation | Improved, not closed | The framework no longer claims to compute the price. It computes a floor and lets exchange discover above it, which is what Marx and Hayek say is the only available move. Where the floor is measured it is currency-free: the infrastructure floor is doctrine-invariant — its spread is exactly 1.000 — where the monetized path spreads by an order of magnitude. Residual: most of the personal floor now carries a price from physical quantities, but much of that price rests on declared bounds rather than measurements, and the remaining components are either the collective's to declare or not yet measurable. The architecture answers the critique; the measurement is well begun and far from done. |
| 2 | Skill differential | Improved substantially | Owen had to set the price and could not. This framework sets a floor wage, which need only be fair-enough, and reads the differential from data rather than decree: a measured O*NET/BLS registry over 751 occupations, 94.2% of US employment. Governance is explicit — sortition, adversarial review, inter-rater reliability, a revalidation clock. Residual: the band [1.8, 2.1] is a constitutional decision, and is now tagged normative rather than dressed as physics. It needs argument, not data. |
| 3 | Founder dependence | Designed against, partly modelled | Sortition-based assessor selection, a minimum assessor count, adversarial review and tier sunset are all anti-single-point-of-failure by construction. Capture of the register is now modelled (scenarios/register_capture.py), and its volume is bounded by structure: registering cannot create obligation, and the mint cannot exceed what was served. Residual: who a captured register admits and excludes is not represented, and succession is not simulated. Those halves are still a design claim, not a result. |
| 4 | Circulation sinks | Closed | A standing public register of obligations means there is always somewhere to spend. Levies are circulatory; destruction is explicit and enumerated (D1–D6). Ithaca's sink required a boundary this design does not have. |
| 5 | Adverse selection / free-riding | Split, and both halves stated | Creation side: closed. No unit exists without a verified labour record against a registered obligation (Condition I), so minting by free-riding is structurally impossible — this is exactly Owen's third wound, and it is the one cleanly answered. Consumption side: a values choice, not a mechanism. The sufficiency floor is deliberately unconditional and tolerates free-riding by design, bounded by a minimum-hours obligation and by membership terms (research/membership.py). |
| 6 | Sovereign suppression | Strategy adopted, not modelled | The WIR lesson learned and the Wörgl mistake avoided: become the issuer, run parallel with an open exchange rate, let adoption be voluntary. Residual: none of this is in the codebase, and the international dimension — reserve status, capital flight, arbitrage — is the framework's single largest acknowledged gap. |
| 7 | Objectivity mirage | Retired, and checkably so | Objectivity language demoted to floor language; only a handful of constants claim physics; the measurement debt is published with its live figure; the gate has no allowlist; and a verdict may not outrank the weakest input it rests on, which caps every headline result at "possible". Residual: "entropy" still does organising work in the EOH core that no exergy accounting backs. The thermal layer is genuine radiative physics; the four-domain core is not thermodynamics and should not trade on the word. |
| — | Marx / Hayek convergence | Answered structurally; one domain still open | The Coasean fork is the answer: no collective computes the whole economy, and discovery happens at the boundaries. Residual: the desire economy — novelty, art, status, play — is dispersed private knowledge that stays polycentric by nature, and it is an explicit NotImplementedError stub (research/desire.py). The autopsy named it the weakest area; it is still the weakest area. |
The question prior art never asked¶
Every scheme above asked whether the currency worked. None asked whether you could leave.
That is this framework's own addition, and it is the invariant held constant while everything else varies with ε. A single collective is legitimate if anyone could genuinely exit and found or join a viable alternative; it is a cage if leaving is legally permitted but practically impossible. Both are "mono". Only one is acceptable.
The lineage is real and mostly outside the labour-currency tradition: contestable markets (Baumol, Panzar and Willig, 1982 — a market served by one firm still behaves competitively if entry and exit are free); the social dividend (Meade, 1964); commonly-held capital paying a portable per-capita dividend (Alaska Permanent Fund, 1976/1982; Norway's GPFG, 1990); non-transferable individual capital accounts (Mondragon); and the indivisible reserve (Italian cooperative law), which prevents the last members standing from liquidating the commons.
The design constraint is Piketty inverted. Zero interest already kills the
passive return; contestability extends the same logic to automated returns by
requiring the commonly-held share not to erode — dτ/dε ≥ 0. At the shipped
defaults, over the canonical arc: exit is financeable at every point, the
commons share φ rises across the arc, dτ/dε ≥ 0 holds throughout, and
the financing channel moves labour → underwritten → self as automation rises.
Reproduce with eoh contestability recal.
The honest caveat is load-bearing. Which regime obtains is empirically unknown. If automated capital is cheap and replicable, contestability is nearly free. If it has increasing returns and concentrates, the conditions bind hard. The framework designs for the adversarial case on the grounds that a generous portable dividend is cheap insurance against the expensive failure — robustness, not optimism. And the bare χ = P/K_entry test was retired as a flow/stock mismatch; it is kept visible as a stress reading rather than deleted, because a superseded negative result is still a result.
Scope: what is settled, and what is open¶
Two different things get called "limitations", and collapsing them is the error this page most wants to avoid. It would read every deliberate boundary as a failure, and every unexplored front as permanent.
So the split below is the same one the framework already applies to its own
constants. A normative constant is a decision that no dataset retires — it
is forbidden from even naming a resolves_by, because pretending data could
settle it would be a category error. A placeholder is real debt that must
name the evidence that would settle it. The same discipline applies to the
framework's claims.
Deliberate boundaries — these are not scheduled to close¶
These are commitments, not gaps. They are where the framework has decided what kind of thing it is, and further research does not move them.
- It does not compute the true price of anything. It computes a floor and makes the judgements above it explicit, auditable and contestable. This is the whole reconciliation: a floor need only be good enough, not optimal, which is precisely why it survives the critique that killed the ancestors. A future version claiming to compute the price would be a regression, not progress.
- It does not claim system-wide inflation impossibility. Demoted to a within-collective, floor-level property. Between collectives relative value floats, and that float is inflation carried honestly as an exchange rate. System-wide impossibility survives only as the ε→1 asymptote, if the need economy consolidates to one unit.
- It does not claim to be thermodynamics. Only a handful of constants are physics. "Entropy obligation" is a disciplined organising frame for recurring maintenance demand. Where real physics runs — the thermal radiative layer — it is confined to that layer and labelled. Adopting full exergy accounting is a live option, but it would be a different framework, not this one completed.
- It does not centrally solve the calculation problem. It relocates allocation to where discovery can operate: planning inside a collective, exchange between them. No amount of further work will produce a central computation, because the architecture is a rejection of that goal.
- Its shipped defaults are not evidence about your jurisdiction. The
constants tagged
instance— population frame, capital, land, the register's cadence, the distance to water, and more — are yours to supply, and the framework will never measure them for you. What it supplies is the bound your declaration is checked against. This is an interface property and it is permanent.
Open fronts — unexplored, and what would settle each¶
These are not verdicts. They are the work queue, stated with the route that would close each, so that progress is checkable and a gap cannot quietly become a permanent excuse.
| Open front | Where it stands | What would settle it |
|---|---|---|
| Retrodiction | Run for the US, by two instruments that share no data — a fixed-asset inventory (scenarios/capital_retrodiction.py) and time diaries (scenarios/labour_epsilon.py). They land adjacent, not overlapping; the capital route returns a grid rather than a number because three declared judgements set it. Both divide by the same obligation, so the machine/human split is checked and the level is not. |
A third instrument that does not divide by total_eoh, for which there is no candidate yet; and the same run for a second jurisdiction. A failed retrodiction is still a stronger result than another thousand tests. |
| Verification cost | Costed for the apparatus — the occupations that decide what counts and check it was done, at two scopes — and checked against the labour a population can supply rather than only against the obligation. The registrant side is priced by the register's declared cadence: on the shipped episodic default it fits with room to spare; a continuous register does not fit until mid-arc. Whoever certifies tiers and measures ecosystem health still holds real power. | A measured registrant multiple for a fulfilment register — hours per record — and verification carried inside other jobs, which is unbuilt. See What Anchors a Unit of Account §6. The term is reported beside the accounts, not inside them, until the author decides whether the registrant's hour enters the obligation. |
| The desire economy | An explicit NotImplementedError stub. The autopsy named it the weakest area and it remains so. |
research/desire.py states its own two questions: how want-goods price with no EOH register to derive a floor from, and whether the contestability machinery transfers to a domain whose entry cost is not entropy-resistance capacity. |
| Personal floor coverage | Most of the obligation now carries a price from physical quantities — it was under a tenth until care and processing were priced as declared bounds. Water resolves by declaration: a collective states its distance to water and the floor reports whether it fits. | Shelter is instance data of the same kind. Sanitation has no named instrument. Health is undefined below the apparatus that delivers it, and is reported as undefined rather than imputed. The extraction wedge still needs the frontier panel (HETUS/MTUS) to identify it without a fitted residual. |
| Domain balance | Personal EOH is almost all of the obligation at ε=0, so ε is very largely a personal-domain number at the low arc. The recurring ecological cost was moved to the Ground Use Fee (Phase 4f), where it scales with land held; the ecological domain now carries stocks only. | Measure the personal base, which is the largest single share of what ε divides by. Read ecological results through the fee rather than through the domain's share — including the thermal layer's solvency margin, which is comfortable partly because the obligation in the domain is small. |
| Governance capture | Modelled, and half-bounded. scenarios/register_capture.py gives the register's failure model and a drift monitor that declares its own threshold. The volume a captured register can mint is bounded by structure — registering cannot create obligation, the mint cannot exceed what was served, and the exposure shrinks as automation arrives. |
The distribution: who a captured register admits and who it excludes. Every figure is a per-capita aggregate with no variable for it, so this needs a distributional layer, downstream of exchange work at scale. Succession is still unsimulated. Ithaca died of founder dependence; a design claim is not a result until something tries to break it. |
| Transition | Strategy, not code. The parallel-currency path, capital flight, arbitrage and reserve-currency dynamics are argued in prose and modelled nowhere. | Model the parallel-currency path. WIR survived by complementing the sovereign and Wörgl was killed for challenging it, so this is the front where the historical record is most directly predictive. |
What would change our mind¶
An open project should say what would falsify it. Three live examples, all already in the repo rather than hypothetical:
- Abatement runs the wrong way. Block II's
a(K)says capital reduces the personal obligation. The measured ATUS series runs the other way — food preparation rose +34% over 22 years of capital deepening while grocery shopping fell 26%, and shift-share attributes the rise to within-household behaviour, not composition. Ifa(K)cannot fit that, it is reported, not retuned. - The personal floor could falsify its own constant. At low coverage it
could not — the only ordering compatible with the evidence was the one
observed. With most of the floor now priced it can contradict
PERSONAL_EOH_BASE, which is the point of building it. The first time it crossed a standard, the crossing turned out to be a scope artefact — the standard covered less than the floor summed — and that is recorded rather than hidden. - A retrodiction could produce an implausible ε. It did, once: an early run saturated. Run over a properly declared grid it reads a band, and the condition re-opens the moment anyone quotes a single ε for a real economy.
Sources¶
Verify before citing in print. Each entry below is stated with its own dates and figures so it can be checked without the underlying autopsy.
- Warren's Time Store — Cincinnati, 1827–1830; "cost the limit of price".
- Owen's NELE — opened 3 Sept 1832, Gray's Inn Road; ~445,501 labour-hours in 17 weeks; closed by 1834.
- Ithaca HOURS — founded Nov 1991 by Paul Glover; 1 HOUR = $10; peak
$100,000 circulating.
- Time banking — Edgar Cahn, No More Throw-Away People (2000).
- Wörgl — 31 July 1932 – 1 Sept 1933; ~1%/month demurrage; ~8–14× velocity; ended by the Oesterreichische Nationalbank. Wära outlawed Nov 1931.
- LETS — Michael Linton, Comox Valley BC, 1983.
- WIR Bank — Switzerland, 1934–present; B2B mutual credit.
- Technocracy — Howard Scott, Technocracy Inc., 1930s; never implemented.
- Soviet labour accounting — War Communism 1918–21 → NEP 1921; the trudoden 1930–1966.
- Marx contra Gray — A Contribution to the Critique of Political Economy (1859), Ch. 2; John Gray, A Lecture on Human Happiness (1825).
- Calculation debate — Mises, "Economic Calculation in the Socialist Commonwealth" (1920); Hayek, "The Use of Knowledge in Society" (1945).
- Coasean boundary — Coase, "The Nature of the Firm" (1937).
- Contestable markets — Baumol, Panzar and Willig (1982).
- Social dividend — Meade (1964); Alaska Permanent Fund (1976/1982); Norway GPFG (1990); Mondragon internal capital accounts.