Conditions of Future Generation in Open Knowledge Platforms - A Preliminary Discussion of a Second-Order Question in Platform Economics

Abstract

Open knowledge platforms exhibit a set of difficulties that persist after access to their content has been made complete: attention concentrates on a small number of works, the labour of review and maintenance receives no accounting, downstream users obtain value that does not return to the arrangements which produced it, contributors are exhausted while output rises, and the relational infrastructure through which encounters occur remains held by one party. These difficulties are ordinarily described one at a time, in the vocabularies of network effects, public goods, incentive design and market power. This paper asks why they co-occur, and proposes that they share a structure the standard descriptions do not represent.

The proposal is a second-order question added to ordinary platform-economic analysis and not a rival economics. Where the usual analysis of an economic mechanism examines what it produces, the question asked here is what it leaves the participants able to produce next: how a present arrangement transforms the conditions of future generation. The paper states the question, fixes its boundary by identifying what does not count as such a condition, and imposes a discipline upon its own use, namely that the question is worth asking only where it changes what an arrangement should do and not merely how the situation is described.

Two results follow from taking the question seriously. The capacity to generate admits no unit, so it cannot be distributed and cannot be maximised; maximisation of it is undefined in place of undesirable, and the plurality of returns which platform designers reach for on practical grounds follows from the same fact. And since no position outside an economic description is available from which its categories could be fixed, the objects worth studying are those which survive re-description. The closed cycle is such an object, and the paper adopts it as its unit: what a single contribution was worth changes with the description, while whether a cycle returns to its own conditions does not.

The paper then examines the documented difficulties one by one against a uniform template, reports which of them the question illuminates and which are already adequately explained by existing theory, and treats the capture of relational infrastructure as its principal case, since that is the cycle an operator holds in its entirety and the one which governs the amplification of every other.

Keywords: platform economics; open knowledge commons; generative conditions; value circulation; attention allocation; knowledge infrastructure; commons reproduction

Notices

Status. This is a working draft circulated for discussion. It is a preliminary discussion paper and is not a finished statement of its author’s position. Sections and statement numbers are subject to change.

Licence. This work is made available under a Creative Commons Attribution-NonCommercial 4.0 International Licence (CC BY-NC 4.0).

Statement on the use of language models. Drafting, literature search and argumentative criticism for this paper were conducted in dialogue with large language models, specifically Claude (Anthropic) and ChatGPT (OpenAI). The claims, the structure, the selection of material and the position taken are the author’s. References cited have been checked; any that remain unverified are marked in the text.

Companion papers. This paper is the third of a set. The first, Rethinking Fa Bu Qing Chuan (“The Dharma Should Not Be Transmitted Lightly”): Knowledge Propagation Ethics and Implications for Open Knowledge Platforms, treats the ethics of transmission for an individual transmitter. The second, Generativity Governance of Open Knowledge Propagation, treats the governance of events, relations and conditions in an open commons. The present paper treats the economic side of the same setting.

Suggested citation. Huang, W. Conditions of Future Generation in Open Knowledge Platforms: A Preliminary Discussion of a Second-Order Question in Platform Economics. Working draft.

empty

Discussion Paper Note

This paper is a preliminary discussion paper intended to share an evolving idea and invite further dialogue, criticism, revision, and independent development.

The author does not claim exclusive epistemic ownership over the viewpoints, concepts, or lines of reasoning presented here, nor does the author claim priority as their first originator. Similar or related ideas may have appeared previously in other intellectual, cultural, or disciplinary traditions. Any legal rights retained in this work are intended to support attribution, responsible use, and protection against exploitative or harmful appropriation, and not to restrict independent inquiry, criticism, revision, or further development.

The arguments in this paper should therefore be understood as provisional and historically situated rather than definitive. Readers are encouraged to question, revise, extend, reinterpret, or independently develop the ideas presented here. Where appropriate, acknowledgment of this paper as one point of encounter in the development of related ideas is appreciated, but such acknowledgment should not be understood as granting the author epistemic ownership over the ideas themselves.

Responsible Use and Rights Reservation

The author encourages good-faith discussion, criticism, independent development, and responsible use of the knowledge presented in this work. The author does not claim exclusive epistemic ownership over the ideas or viewpoints discussed herein, nor claim priority as their first originator.

However, the author expressly reserves all rights and remedies available under applicable law with respect to uses of this work or related materials that involve unlawful conduct, harmful or abusive exploitation, improper commercial appropriation, infringement of applicable intellectual property or other legal rights, or conduct contrary to applicable national, regional, or international law.

This reservation is intended to preserve the ability to respond to misuse and harmful appropriation. It should not be interpreted as restricting legitimate academic inquiry, criticism, independent reasoning, or the further development of related ideas.

page1

IThe Problem

1. Introduction

An open knowledge platform can succeed completely at what it set out to do and still fail. Every work it holds may be readable without payment, permission or institutional affiliation, and the arrangement may nonetheless exhibit a familiar set of difficulties. Attention concentrates on a small number of authors, institutions and already-visible works. The labour of reviewing, correcting, translating, moderating, replicating and maintaining datasets produces no accounting entry and no recognition. Downstream users, including well-resourced commercial ones, obtain value from the commons that does not return to the arrangements which produced it. Contributors are gradually exhausted while the platform’s output measures rise. Promising work fails for want of one specific thing, which is frequently not money. Metrics adopted to represent contribution become targets and cease to represent it. And the content may be open while the social graph, the recommendation system, the usage records and the ranking infrastructure through which encounters actually occur remain held by a single party.

Each of the difficulties named above is ordinarily described one at a time, and the descriptions are good ones. Attention concentration is preferential attachment and cumulative advantage. Uncompensated contribution is a public-goods and free-riding problem. Downstream capture is a question of intellectual property, market power and the appropriability of non-rival goods. Metric distortion is Goodhart’s law and the reactivity of measures. Each account is adequate to its own case, and none of them is disputed here.

What the accounts do not explain is why these particular difficulties travel together, and why they persist through interventions addressed to each of them separately. A platform may improve its compensation of contributors and still lose them; may diversify its recommendations and still concentrate; may open its content completely and still find that participation depends on continued access to something it does not control.

The question this paper adds. The proposal here is narrow. Ordinary analysis of an economic mechanism examines what the mechanism produces: a transaction concluded, a good allocated, a work published, a payment made. To that examination this paper adds a second question, asked of the same mechanism, about what it leaves the participants able to produce afterwards.

Definition 1.1The second-order question For an economic mechanism $M_t$ operating in a knowledge commons, ordinary analysis examines the outcome $O_t$ that $M_t$ produces. The second-order question asks how $M_t$ alters the conditions $C_{t+1}$ under which the participants may generate further work, and what generation $G_{t+1}$ those conditions permit.

Three features of the definition are deliberate and are stated here because each forestalls a reading the paper rejects.

The question is added to ordinary analysis and does not replace it. Nothing in this paper claims that allocation, exchange or price is the wrong object of study, and §5 sets out what the standard descriptions of a platform economy see, precisely so that the paper’s addition is visible against them and is not confused with them.

The question is about conditions and not about a quantity. This paper advances no measure of generative capacity, and §10 shows why none is available and what follows from its absence. In particular, no arrangement recommended here is recommended because it increases generation, and every objective in the paper is stated as the avoidance of a foreclosure.

And the question is not always worth asking. §8 imposes a discipline on the paper’s own instrument: the second-order question earns its place for a given difficulty only where asking it changes what an arrangement should do, and not merely how the difficulty is described. Part V applies that discipline honestly and reports the difficulties for which the question adds nothing, together with the existing theory that should be used in place of that.

Two results. Taking the question seriously produces two results that shape everything after them.

The first concerns distribution. The capacity to generate admits no unit in which two instances of it could be compared, so it is not a quantity over which a distribution can be defined. It cannot be allocated, and its maximisation is undefined in place of undesirable. A platform designer who reaches for plural and heterogeneous returns on practical grounds is therefore doing something the structure of the object requires, and §10 derives that plurality in place of recommending it.

The second concerns the unit of analysis. What counts as economic activity is constituted by the categories through which a society recognises value, labour and obligation, and no position outside those categories is available from which they could be fixed. It follows that quantities defined within a description vary with the description, and that the objects worth studying are the ones which survive re-description. A closed cycle is such an object. What a particular contribution was worth, and what share of a downstream gain belongs to whom, change with the accounting adopted. Whether a cycle returns to the conditions from which it began, and with what deficit, does not. Part III adopts the cycle as the paper’s unit for that reason, and Part IV gives the argument its formal statement, drawing on the treatment of evaluation without an external frame developed in the author’s companion work on loop observables.

The principal case. Among the difficulties examined, the capture of relational infrastructure receives the fullest treatment, for three reasons that Part V sets out. It is the one cycle an operator holds in its entirety. It governs the amplification of every other cycle, since which encounters occur is determined there. And it is the case in which the ordinary description and the second-order question prescribe different actions, since a proprietary asset lawfully held and a condition of future generation held by a single party are the same object under two descriptions, with portability, interoperability and shared provenance following only from the second.

Contributions. The paper assembles a landscape of documented platform difficulties and shows that they reduce to three mechanisms: those in which a cycle exists and the arrangement cannot see it, those in which it is seen and does not close, and those in which it closes and drifts. It states the second-order question, fixes the boundary of what counts as a generative condition, and imposes a test for when the question should be used. It establishes that generative capacity is non-distributable and draws the consequences for return mechanisms and for objectives. It adopts the closed cycle as the unit of analysis and gives the reason in terms of what survives re-description, together with the two independent conditions a cycle must satisfy, which are closure of the circuit and openness of the boundary. It examines the difficulties one by one against a uniform template and reports where its own instrument fails. And it orders the available return mechanisms by what each requires of an operator, which yields the finding that the cheapest are the least implemented.

The bounds of the ambition. This paper does not propose a new economics, and it takes no position on what economics fundamentally studies. The tradition it draws on and the traditions it draws from are treated as bodies of accumulated experience, each formed under conditions that shaped it, and the paper’s aim is to bring a set of platform difficulties into contact with a question that existing descriptions do not ask, so that the two may be compared on the same material. The question is offered as one lens among others, and the test proposed for it is pragmatic: what becomes visible, explicable and governable when it is used, and where it makes no difference. Its own terms are held open, and §24 states the conditions under which the position advanced here would be shown to be mistaken.

The paper is organised in seven parts. Part I sets out the standard economic descriptions of a platform, the landscape of documented difficulties, and the relation to prior work. Part II states the second-order question, the boundary of the notion of a generative condition, and the consequences of non-distributability. Part III develops the closed cycle as the unit of analysis, its two conditions, the coupling among cycles, and the pathologies of circulation. Part IV gives the formal discussion, including the gauge argument for the cycle, the question of a time-varying state space, and an explicit statement of what the formalism does not deliver. Part V examines the difficulties one at a time. Part VI treats the instruments available to an operator and the conditions on their admissible form. Part VII states the limits and the questions left open.

2. The Standard Economic Descriptions of a Platform and Their Objects

This section sets out the economic descriptions already available for an open knowledge platform, and states for each what it takes as its object and what it therefore records. Its objective is to establish what is accounted for before this paper adds anything, so that the addition proposed in Part II is visible against an accurate background in place of an impoverished one. The section treats five descriptions in turn, states what each explains well, and closes by identifying the property they share, which is the property that Part II’s question is addressed to. Nothing in the section disputes any of the descriptions, and the paper relies on several of them later.

2.1 The Description by Intermediation Between Sides

The first description treats a platform as an intermediary between distinct groups whose participation is valuable to one another, and studies the price structure and the design choices by which the intermediary brings them together (Rochet and Tirole 2003). Its central results concern the allocation of charges between the sides, the conditions under which one side is subsidised, and the consequences of the resulting structure for entry and competition.

Applied to an open knowledge platform, this description is informative about the arrangement’s viability. It identifies contributors and readers as distinct groups, notes that contributors are ordinarily unpriced, and predicts that the arrangement will seek revenue from a third party, whether an institution, an advertiser, a funder or a purchaser of data. What it records is a structure of prices and participation. What it does not ask is whether a contributor who participates today is in a position to participate next year.

2.2 The Public-Goods Description

The second description treats knowledge as non-rival and difficult to exclude from, and therefore as a good whose private provision will be less than the level a collective would choose. Its central results concern free-riding, under-provision, and the instruments by which provision may be sustained: excludability through intellectual property, public funding, reputational reward, and voluntary contribution under favourable conditions.

This description is the most direct account of the difficulty that contributors give and receive nothing, and it is adequate to that difficulty. Its object is the level of provision of a good. The provision problem is stated for a good whose character is fixed, and the question of whether provision this period alters the capacity to provide next period lies outside its statement, which is a limitation of the framing and not an error within it.

2.3 Network Effects and Cumulative Advantage

The third description studies the dynamics by which participation, visibility and adoption reinforce themselves. Its central results concern increasing returns to adoption, preferential attachment in the growth of networks, the tendency of such processes toward concentrated outcomes, and the consequent difficulty of entry. In the scientific case the same dynamic is documented as an advantage accruing to those already recognised.

This description explains attention concentration completely and predicts it in advance. Its object is the distribution of a quantity across participants and the process generating that distribution. What it records is where attention went. Whether the resulting configuration leaves the arrangement able to produce work that has not yet been produced is a further question, and one the description does not pose.

2.4 The Attention-Allocation Description

The fourth description observes that what is scarce in an environment rich in information is the capacity to attend (Simon 1971), and treats attention accordingly as the object of allocation. Its central results concern competition for attention, the design of interfaces and rankings that direct it, and the distortions introduced when attention becomes the quantity that participants pursue.

This description is the right one for a great deal of platform design, and Part V relies on it. Its object is the allocation of a scarce input among competing uses. It is well suited to asking whether a given allocation is efficient, and it takes as given the set of uses among which the allocation is made.

2.5 Commons-Based Peer Production and the Governance of a Commons

The fifth description studies production organised without price signals or managerial direction (Benkler 2006), and the institutional arrangements by which shared resources are governed and sustained (Ostrom 1990; Hess and Ostrom 2007). Its central results concern the conditions under which decentralised contribution outperforms firm and market organisation, and the design principles under which a community succeeds in governing a resource it holds in common, including the problem of provisioning as distinct from appropriation.

This description is the nearest to the present paper’s concerns and shares several of them. Its treatment of provisioning already asks how the resource is maintained and not only how it is used, and its attention to the rules a community sets over itself anticipates much of Part VI. Where the present paper differs is in the object maintained: the commons literature treats the resource, while the question of Part II concerns the participants’ capacity to contribute to it, which is a different thing and may deteriorate while the resource itself is intact.

2.6 The Property the Descriptions Share

The five descriptions differ in almost every respect and agree in one, which is the respect that matters here.

Claim 2.1The temporal object of the standard descriptions Each of the descriptions surveyed takes as its object a quantity, a distribution, or a level realised in a period: a price structure, a level of provision, a distribution of attention or adoption, an allocation of a scarce input, a state of a shared resource. Each is therefore evaluated by comparing that object with an alternative in the same period. None of them takes as its object the capacity of the participants to produce in a subsequent period, and none is defective for this, since that capacity is not the quantity any of them was constructed to represent.

The claim is descriptive and carries no criticism. A framework is entitled to its object. What follows from the claim is only that a difficulty concerning the capacity to produce later will be represented, in each of these descriptions, by whatever proxy in the current period happens to correlate with it, and that where the correlation fails the difficulty will be invisible to all five at once.

Claim 2.2The consequence for the landscape A difficulty that consists in the deterioration of the participants’ capacity to contribute will appear, under the standard descriptions, as a set of separate anomalies in separate accounts: a provision shortfall, a concentrated distribution, an inefficient allocation, a governance failure. Their co-occurrence is then a coincidence within each account, and the question of whether it is a coincidence cannot be posed within any of them.

§6 assembles the difficulties and puts that question, and Part II supplies the object with respect to which their co-occurrence is not a coincidence.

3. The Landscape of Documented Difficulties

This section assembles the difficulties reported of open knowledge platforms and states the classification the paper proposes for them. Its objective is to establish that the difficulties are not ten independent problems but instances of three structures, and to do so before any of the paper’s apparatus is introduced, so that the classification may be assessed on its own. The section states the nine difficulties with the form each takes; gives the three mechanisms and the criterion distinguishing them; presents the classification in Table 1; and states what the classification does and does not establish. The difficulties are drawn from the reported experience of open repositories, open source projects, open review arrangements and collaborative reference works, and no claim is made here about their frequency in any particular case.

3.1 The Difficulties

Attention concentration. Visibility produces attention, attention produces recognition, and recognition produces further visibility, so that a small number of authors, institutions, venues and works absorb a share of attention out of proportion to any assessment of their content.

Contribution invisibility. Reviewing, correcting, replicating, translating, moderating, curating datasets, reporting negative results and drawing conceptual connections between separated literatures produce no artifact of the kind the arrangement records, and are therefore performed without accounting or recognition.

Downstream capture. Parties with greater compute, capital, legal capacity or market access obtain value from openly available work, lawfully and without obligation, and the arrangements that produced the work receive nothing determinate in return.

Premature publication. Where an arrangement rewards volume, novelty and priority, immature work is released early, and an early formulation may occupy a subject’s terminology and reference points regardless of its quality.

Weak return cycles. Contributors supply work, review, data and attention over extended periods, while the conditions that permitted them to do so are not replenished from any source connected to the use made of what they supplied.

Cumulative advantage. Recognition yields opportunity, opportunity yields resources, and resources yield further recognition, so that positions established early are consolidated by the arrangement’s ordinary operation.

Resource mismatch. A promising line of work fails for want of one specific thing, which may be compute, a dataset, a collaborator, an interval of uninterrupted time, or institutional standing, and which is frequently not money.

Capture of relational infrastructure. The content is open while the social graph, the recommendation system, the usage records, the embeddings, the ranking infrastructure and the interaction histories are held by the operator, so that which parties encounter which work is determined by an arrangement no participant can inspect.

Metric distortion. Counts of citation, download, view or endorsement, adopted to represent contribution, become the objects participants pursue, and cease to represent what they were adopted to represent.

Contributor exhaustion. A community’s recorded output rises while the participants sustaining it withdraw, and the arrangement’s measures continue to indicate health until the withdrawal is complete.

3.2 The Mechanisms Distinguished

The difficulties are distinguished here by the relation between a cycle of contribution and the arrangement that hosts it. The vocabulary of cycles is introduced formally in Part III and is used in this subsection only to state the classification; a reader may substitute the ordinary sense of a sequence of activity that returns to its starting point.

Definition 3.1The three mechanisms A difficulty is a failure of measurement where the cycle of contribution operates and the arrangement’s records do not represent it. It is a failure of return where the cycle is represented and does not return to the conditions from which it began. It is a failure of amplification where the cycle returns and the capacity it carries accumulates at one position within it.

The three are distinguished by what an arrangement would have to do about each, which is the property the classification is for. A failure of measurement is addressed by altering what is recorded, which is a matter of the arrangement’s schema and costs the participants nothing. A failure of return is addressed by establishing a pathway that does not exist, which requires a party willing to supply something. A failure of amplification is addressed by altering the operation that concentrates, which is ordinarily the arrangement’s own ranking or matching function and is the thing it is least willing to change.

| @p3.4cmp4.2cmp2.2cmp2.4cm@

Difficulty	Ordinary description	Mechanism	Where treated
Contribution invisibility	Unpaid labour, recognition failure, incentive misalignment	Measurement	§19.2
Metric distortion	Goodhart’s law, reactivity of measures	Measurement	§20.2
Premature publication	Quality assurance, information overload	Measurement	§20.3
Weak return cycles	Public goods, free-riding, under-compensation	Return	§19.4
Downstream capture	Appropriability, market power, intellectual property	Return	§21.2
Resource mismatch	Allocation of a scarce input	Return	§21
Contributor exhaustion	Volunteer attrition, burnout, provision failure	Return	§19.3
Attention concentration	Preferential attachment, network effects	Amplification	§20.1
Cumulative advantage	Advantage accruing to the recognised	Amplification	§20.1
Capture of relational infrastructure	Proprietary asset, platform market power	Amplification	§19.1

Table. The documented difficulties, the description under which each is ordinarily treated, and the mechanism proposed here.

3.3 The Standing and Consequences of the Classification

Claim 3.2The reduction The difficulties reported of open knowledge platforms instantiate three mechanisms, distinguished by whether a cycle of contribution is unrepresented, unreturned, or returned with accumulation at one position. Difficulties within a mechanism call for instruments of one kind, and difficulties across mechanisms do not, so an intervention addressed to one difficulty will relieve those sharing its mechanism and leave the others where they were.

Two consequences follow, and the second is the reason the paper continues past this section.

The first is diagnostic. An arrangement that has addressed a difficulty and observed no improvement has evidence that it misidentified the mechanism, and the classification supplies the alternatives to consider. A platform that has increased payments to contributors and continued to lose them may be facing an amplification failure, in which the payment reaches those already visible.

The second is that the classification is stated in terms the standard descriptions of §5 cannot supply. Each mechanism is defined by a property of a cycle across periods, and Claim 2.1 established that those descriptions take realised quantities within a period as their objects. The classification is therefore intelligible only if something in the position of Part II’s question is granted, and it is offered as evidence for that question in place of a result independent of it.

Claim 3.3The standing of the classification The reduction proposed in Claim 3.2 is a hypothesis about the structure of a set of reported difficulties. It would be shown to be wrong by a difficulty that belongs determinately to two mechanisms, by an intervention that relieves difficulties across mechanisms uniformly, or by a mechanism whose difficulties fail to respond to the instrument the classification assigns them. Part V examines each difficulty separately and reports the cases in which the classification does no work.

4. Relation to Prior Work

This section situates the enquiry among the literatures that own parts of it, with the objective of fixing what the paper claims and, in two cases, of conceding a prior owner of a term or a criterion the paper uses. It proceeds from the definitional question about economics, through the treatments of capacity and of endogenous expansion, to the two nearest neighbours, which are the prior definition of generativity for technological platforms and the account of value circulation in generative justice, and closes with the position claimed. Where a prior owner exists the paper adopts the prior formulation and states what it adds, in place of restating the same content in different words.

4.1 The Definition of the Economic Problem

The description of economics as the study of behaviour in the relation between ends and scarce means with alternative uses (Robbins 1932) is the source of the modern framing that this paper works alongside. It is worth noticing what that definition does: it identifies economics by the form of a problem in place of by a subject matter, and it therefore admits whatever takes that form and excludes whatever does not. A question about the conditions under which capacities come into existence does not take the form of a choice among alternative uses of a given means, and it is excluded by the definition without any judgement being made about its importance.

Two lines of work bear on this. Broader definitions, in the welfare tradition (Marshall 1890) and in political economy, take as the object the organisation of production, distribution, exchange, consumption and reproduction of material social life together with the relations through which these occur, and admit the present question without amendment. And the debate in economic anthropology between a formal conception organised around scarcity and choice and a substantive conception organised around provisioning (Polanyi 1957) is the same disagreement conducted at greater length. This paper adopts no position in that debate, and notes only that its question is admitted by the second family of definitions and excluded by the first, which is a fact about definitions and not an argument for either.

4.2 Capacity, Conversion, and the Insufficiency of Resources

The capability approach establishes that the object of concern in an assessment of a person’s position is what that person is able to do and to be, and that resources are converted into such capacities at rates which differ across persons and settings (Sen 1985, 1999; Nussbaum 2011). The consequence, which that literature states plainly, is that an equal distribution of resources does not produce an equal distribution of capacities, and that a resource-based assessment therefore misdescribes what it was intended to assess.

The relevance is direct, and the debt should be stated plainly. The claim in this paper that additional resources may leave a line of work no better placed, and that a small change in relations may leave it much better placed, is the conversion problem applied to production in place of to well-being. What the present paper adds is not the observation but its object: the capability literature is concerned with the capacities of persons as ends, and the concern here is with the conditions of continued production in a collective arrangement, which raises the further questions of who maintains those conditions and out of what.

4.3 The Endogenous Expansion of Possibility

Growth theory established that the accumulation of knowledge is a process altering what is subsequently producible in place of a fixed factor being allocated, and that the non-rivalry of ideas is what permits an idea to enter indefinitely many subsequent processes without being consumed (Romer 1990; Lucas 1988; Aghion and Howitt 1992). The expansion of a possibility set by the activity within it is therefore not a novel proposal but the central result of a mature literature.

That literature is the paper’s most important concession, since a reader might otherwise take the second-order question to be claiming what growth theory established decades ago. The difference lies in what is asked about the expansion. Growth theory asks about the rate and the aggregate consequences of the process. The question here is distributional and institutional: which participants retain the conditions to take part in it, what happens to the arrangements that produced the non-rival good, and by what mechanism a particular platform’s design bears on either. Those questions are compatible with the growth results and are not answered by them.

4.4 The Governance and Provisioning of a Commons

The literature on the governance of shared resources (Ostrom 1990), and its extension to knowledge as a commons (Hess and Ostrom 2007; Benkler 2006), supplies the setting in which this paper’s difficulties arise and several of the instruments it discusses. The distinction between the appropriation of a resource and its provisioning is directly the distinction this paper needs between use and maintenance, and the design principles for self-governing arrangements anticipate the conditions on instruments set out in Part VI.

Where the present paper differs is in the object whose provisioning is at issue. The commons literature treats a resource that participants draw upon and must replenish. The concern here is with the participants’ own capacity to contribute, which may deteriorate while the resource is intact and growing, and which is not replenished by any act of provisioning the resource.

4.5 Generativity in the Prior Literature on Technological Platforms

The term generativity has an established prior use in the study of technological systems, where it names a system’s capacity to produce unanticipated change through unfiltered contributions from broad and varied audiences, analysed through the leverage a system offers, its adaptability, the ease with which it is mastered, its accessibility, and the transferability of what is produced with it (Zittrain 2006, 2008). That literature also states the tension between such a capacity and the security of the system, and treats the resolution of that tension as the central design problem.

The prior use is conceded in full, and the paper is careful about the consequence. This paper does not offer a rival definition of the term and does not depend on one, since its instrument is the second-order question of Definition 1.1 and not a property attributed to a system. Where the prior literature asks how open a technological system should remain, the question here concerns the conditions of participants within an arrangement and the cycles by which those conditions are or are not renewed. The two are compatible, and a reader who prefers the prior definition may substitute it throughout without altering any argument in this paper.

4.6 Generative Justice and the Circulation of Value

The account of generative justice establishes that value should return to those whose labour, communities and ecologies generated it, that generators should retain control over the conditions of their own production, and that arrangements should be assessed by whether value circulates back in unalienated form in place of being extracted (Eglash 2016). Later work develops this into the design of recursive circulation in computational settings (Eglash et al. 2024).

Claim 4.1The prior ownership of the circulation criterion The criterion that value generated through a commons should return to the arrangements that produced it, and the assessment of an economic relation by whether such return occurs, belong to the account of generative justice. This paper adopts that criterion in the form its author states it and claims no part of it.

What the paper adds to the criterion is stated in three parts, and each is developed later. Circulation is one instrument among several and frequently not the one a given difficulty calls for, which is the point of the classification in §6 and of the diagnostic method in Part V. Circulation is necessary and insufficient, since a cycle may return value to its participants while the capacity to take part accumulates at one position and while entry requires the consent of those already within it; §12 states the two independent conditions this requires. And the object returned admits no common unit, which is established in §10 and which converts the plurality of return forms from a design preference into a consequence.

4.7 Adjacent Accounts of the Difficulties

Four further literatures own individual difficulties in the landscape and are cited where those difficulties are treated. Advantage accruing to the already recognised (Merton 1968), and the cumulative-advantage processes that produce concentrated distributions (Price 1976), own attention concentration. The reactivity of measures under pressure owns metric distortion (Goodhart 1975; Campbell 1979; Espeland and Sauder 2007). The study of infrastructure and of maintenance (Star 1999), together with the analysis of social reproduction in feminist economics (Folbre 2001; Fraser 2016), owns the invisibility of maintaining labour and supplies the strongest existing statement of why an accounting that meters outputs cannot represent it. And the analysis of platform accumulation of behavioural and relational records owns the material facts of the infrastructure case (Zuboff 2019), though not the reading of those records as conditions of production.

4.8 The Position This Paper Claims

Claim 4.2The unclaimed position Existing work establishes the definitional boundary of the economic problem, the conversion of resources into capacities, the endogenous expansion of possibility, the provisioning and governance of a commons, the generativity of technological systems, the criterion of value circulation, and the mechanisms behind individual platform difficulties. No existing treatment classifies the reported difficulties of open knowledge platforms by whether a cycle of contribution is unrepresented, unreturned or returned with accumulation; states a discipline restricting the use of a generative reading to the cases in which it alters a prescription; establishes that the capacity at issue is non-distributable and derives the plurality of returns from that fact; or adopts the closed cycle as the unit of analysis on the ground that it is what survives re-description.

The four elements are separately modest, and the paper’s claim is their conjunction together with the application to a single domain. The conjunction matters because each element without the others produces a familiar defect: a classification without the discipline becomes a framework that explains everything, the discipline without the classification has nothing to be applied to, the non-distributability result without the cycle leaves no assessable object, and the cycle without the non-distributability result invites the measure that Part II shows to be undefined.

IIThe Second-Order Question

5. The Question and the Restriction on Its Use

This section states the paper’s instrument and the discipline governing its application. Its objective is to establish what the second-order question asks, to distinguish it from three questions it resembles, and to fix a test determining when it should be used, since an instrument applicable everywhere would establish nothing about any particular case. The section restates the question in the form later parts use; distinguishes it from a question about long-run consequences, from a question about sustainability, and from a question about a system’s aggregate capacity; states the test and what it excludes; and states the two ways the test may be failed. The section is stipulative, and the stipulations are defended by the work they do in Part V.

5.1 The Question Restated

Definition 1.1 stated the question for an economic mechanism operating in a knowledge commons. Three components of it require fixing before it can be applied.

The mechanism $M_t$ is any arrangement by which the participants’ relations are ordered: a payment, a ranking, a licence, a threshold for publication, a rule about who may review, a default setting, an interface. The definition places no restriction on the kind of arrangement, and Part V accordingly examines mechanisms of several kinds.

The conditions $C_{t+1}$ are treated at length in §9, and the treatment there is a restriction: not everything bearing on generation is a condition in the sense used here.

And the generation $G_{t+1}$ is left without a measure. The question asks what the altered conditions permit, and the answer is given in terms of what has been foreclosed or left open, never as a quantity that has risen or fallen. §10 establishes that no such quantity is available, and the language of the whole paper is constrained accordingly.

Claim 5.1The form of an answer to the question An answer to the second-order question states which participants, after the operation of a mechanism, retain or lose a condition on which their further contribution depended, and which lines of work thereby become available or unavailable. It does not state that generation has increased or decreased, and an answer of that form is unavailable for the reason given in §10.

5.2 The Neighbouring Questions and the Boundaries Against Them

Each of the three resembles the question closely enough that a reader may substitute it, and each would produce a different and weaker paper.

The question is not about long-run consequences. An analysis of the long-run effects of a mechanism examines the same outcome variable over a longer horizon: more or fewer works, higher or lower revenue, a wider or narrower distribution. That is ordinary analysis conducted patiently. The question here concerns a different variable, the conditions under which participants act, which may deteriorate while every outcome variable improves and which no lengthening of the horizon on an outcome variable will reveal.

The question is not about sustainability in the sense of a system’s continuation. An arrangement may continue indefinitely while the participants able to contribute to it are progressively narrowed, and such an arrangement is sustainable and is losing precisely what the question asks about. The relation runs the other way as well: an arrangement may end while having left its participants better placed than it found them.

And the question is not about a system’s aggregate capacity. The paper makes no claim about the total quantity of anything, and §10 explains why it cannot. Where the question is applied in Part V, its answers concern particular participants and particular lines of work, and their aggregation is not performed because no operation of aggregation is available.

5.3 The Test Restricting the Question’s Use

The question can be asked of any mechanism whatever, and an instrument with that property establishes nothing unless its use is restricted. The restriction adopted here is the strictest available that is still checkable.

Definition 5.2The prescriptive difference test The second-order question earns its place in the treatment of a difficulty where the ordinary description of that difficulty and the answer to the question prescribe different actions for the arrangement. Where the two prescribe the same action, the question has supplied a redescription and is not used.

Three features of the test are deliberate.

The test compares prescriptions and not vocabularies. An account that renames free-riding as a failure of return, and then recommends what the free-riding account recommends, has passed no test. The prescriptions must differ in what an arrangement is to do, and the difference must be attributable to the question and to no other feature of the account.

The test is applied per difficulty and not to the framework as a whole. The question may earn its place for one difficulty and fail for another, and §21 reports the difficulties for which the paper finds that it fails, together with the existing theory that should be used in its place.

And the test is applied against the strongest available ordinary description, not against a weakened one. §5 was written to make that possible, and a reader who judges that a stronger ordinary description was available has identified the way this paper’s claims are to be attacked.

5.4 The Failures of the Test

Stating how the test is failed is part of stating the test, since a criterion whose failures cannot be recognised imposes nothing.

The first failure is redescription, treated above: the question yields a different account of the same difficulty and the same recommendation. This failure is detectable by inspection, since the two recommendations may be compared.

The second failure is less visible and is the one this paper is most exposed to. The question may prescribe a different action, and the difference may be attributable to something other than the question, most commonly to an additional empirical assumption smuggled in with it. A recommendation that a platform reserve part of its recommendation capacity for underconnected work differs from what an efficiency analysis recommends, and the difference is attributable to the second-order question only if it does not rest on an unstated assumption about where valuable work is found. §20.1 states that assumption where the recommendation is made.

Claim 5.3The exposure of the framework to its own test A framework restricted by Definition 5.2 is exposed in a specific way: for each difficulty it treats, the difference in prescription must be shown, and the difference must be shown to follow from the question and not from an assumption accompanying it. The framework therefore stands or falls difficulty by difficulty, and a finding that it fails for a given difficulty leaves its treatment of the others untouched.

The claim states a property of the framework that the paper regards as an advantage and that a reader may regard otherwise. A framework of this shape makes no unified claim and can be defeated piecemeal. What it offers in exchange is that its failures are local and reportable, and §21 reports them.

6. Generative Conditions and the Boundary of the Notion

This section fixes what counts as a condition in the sense the second-order question requires. Its objective is to supply a boundary, since a notion admitting everything that bears on generation would make the question unanswerable and the framework unfalsifiable. The section states the difficulty in its sharpest form; states the kinds of thing ordinarily proposed as conditions; gives the criterion this paper adopts, which turns on what an arrangement has made itself answerable for; states what the criterion excludes and what it admits that a narrower criterion would miss; and states the consequence that follows for an arrangement, which the paper regards as its most useful result in this Part. The criterion is stipulative and its defence is that it produces a boundary at all, which the alternatives considered do not.

6.1 The Difficulty in Its Sharpest Form

Almost anything bears on whether a person produces work. Health, housing, a language in common, the absence of war, an unhurried afternoon, a colleague who asks the right question, a functioning search index, an institution willing to employ. A notion of generative condition that admits all of these is true and useless: it names the world, it excludes nothing, and a question asked in its terms has no determinate answer.

The difficulty is not merely one of scope. A framework whose central term admits everything cannot fail, since any outcome whatever is consistent with some condition having been altered. §8 restricted the framework’s use by a test on prescriptions; this section restricts its central term.

6.2 The Kinds Ordinarily Proposed

Discussions of what a productive process requires converge on a familiar list, and the list is worth stating because the criterion below cuts across it in place of extending it. The kinds ordinarily named are material (funding, equipment, compute, data, physical space), temporal (uninterrupted intervals, tolerance of delay, freedom from premature assessment), relational (collaborators, interlocutors, trust sufficient to expose unfinished work), epistemic (access to prior work, means of finding it, translation, provenance), and institutional (employment, standing, permission, the ability to publish and to be read).

Two properties of this list matter. It is open, since further kinds may be added without contradiction, which is the difficulty of the preceding subsection restated. And its members are not comparable, since no operation converts an interval of time into a quantity of standing, which is the difficulty developed in §10.

6.3 The Criterion

The criterion adopted here does not attempt to say which things in the world affect production, which is an empirical matter of unbounded scope. It asks about the arrangement.

Definition 6.1Generative condition Relative to an arrangement, a generative condition is a requisite of participants’ further contribution which that arrangement has made itself answerable for: which it supplies, allocates, records, prices, permits or withholds through its own operation, so that a change in the arrangement changes the condition.

The criterion is relative by construction, and the relativity is the point. Compute is a generative condition for a platform that allocates compute and is not one for a platform that does not, however much compute bears on what its participants can do. Uninterrupted time is a condition for an arrangement whose deadlines and review cycles determine how time is fragmented, and is not one for an arrangement with no bearing on its participants’ schedules.

Claim 6.2The boundary the criterion draws Definition 6.1 excludes from the second-order question every requisite of contribution that the arrangement does not supply, allocate, record, price, permit or withhold. Such requisites may be decisive for what participants can do and remain outside the question, which is addressed to what a mechanism of the arrangement alters.

The exclusion is substantial and should be stated without softening. A participant’s health, family circumstances, employment outside the arrangement and national situation may determine what that participant contributes, and none of them is a generative condition in the sense used here. The paper says nothing about them, and an account of what enables a person to produce work would have to.

What the criterion admits, and a narrower criterion would miss, is the class of conditions an arrangement supplies without recognising that it supplies them. The ordering of a list of results is a condition on which work is encountered. A default licence is a condition on what may be built upon. A publication status is a condition on whether a preliminary formulation acquires the standing of a settled one. None of these is a resource and each is supplied by the arrangement’s own operation, so each falls squarely within Definition 6.1.

6.4 The Consequence for an Arrangement

The criterion has a consequence that runs against the arrangement adopting it, and this paper regards it as the most useful thing in the present Part.

Claim 6.3The coincidence of what is governed with what is answered for An arrangement can govern, through the second-order question, exactly those conditions it has made itself answerable for. A requisite of contribution that the arrangement neither supplies nor records nor prices lies outside the question by Definition 6.1, and the arrangement therefore has no purchase on it however decisive it may be. The boundary of what an arrangement can govern is drawn by the boundary of what it has taken responsibility for, and that boundary is set by the arrangement itself.

Three implications follow, and each is used later.

An arrangement may extend what it governs by making itself answerable for more, which is what an arrangement does when it begins to record a class of contribution it had not recorded, or to allocate a resource it had left to participants to obtain. That extension is available and is not free: §23 treats the cost, and §24 records that the cost falls unevenly.

An arrangement may narrow what it governs by declining to record, and the narrowing is invisible from within, since what was never recorded leaves no gap. This is the economic form of a result established in the companion paper on the governance of propagation (Huang 2026h), where the selection of a recording schema is shown to determine what can subsequently be established, and to be undetectable from within the record it produces.

And an arrangement’s accounting is therefore not a neutral description of an economy it observes. What it meters becomes what it can govern; what it leaves unmetered becomes what it cannot. §19.2 works the clearest instance, in which labour that produces no recorded artifact is invisible to an accounting that meters artifacts, so that the arrangement cannot address a deterioration whose existence its own instruments deny.

7. Non-Distributability and Its Consequences

This section establishes the paper’s central formal result and draws the three consequences that shape everything after it. Its objective is to show that the capacity at issue is not a quantity over which a distribution can be defined, to distinguish that claim from a claim about the difficulty of measurement, and to derive from it the plurality of return forms, the unavailability of a maximisation objective, and the form in which the paper’s objectives must be stated. The section states the two arguments separately, since they fail under different objections; states what the arguments do not establish; and states the consequences in turn. The formal argument is elementary and is given in full, since the paper rests a good deal upon it.

7.1 The Formal Argument

A distribution of a good among parties presupposes that the good admits a common unit in which the shares are expressed, so that what one party receives and what another receives are quantities of one thing. Where no such unit exists, the notion of a share is undefined, and the question of who should receive how much cannot be posed, let alone answered.

Claim 7.1The absence of a common unit The requisites of further contribution enumerated in §9 admit no common unit. An interval of uninterrupted time, an increment of standing within a community, access to a dataset, and the availability of an interlocutor are not quantities of one thing, and no operation converts an amount of one into an amount of another that is independent of the participant and the line of work.

The last clause carries the argument and should be examined, since a natural objection is that money performs exactly this conversion. Money converts many requisites into one another at rates, and where a rate exists the conversion is available. The claim here is that the rates are not independent of the case: the sum that buys an interlocutor for one line of work buys none for another, and the sum that buys standing in one community buys none in a second. A conversion whose rate must be determined case by case is not a unit, since a unit is what permits the cases to be compared without the comparison being performed afresh.

Claim 7.2Non-distributability Since the requisites of further contribution admit no common unit, the capacity they jointly constitute is not a quantity over which a distribution can be defined. It cannot be allocated, and an allocation problem cannot be posed with it as its object.

The result is formal and does not depend on anyone’s epistemic position. It holds for an arrangement with complete information about every participant.

7.2 The Operational Argument

A second argument reaches a weaker conclusion by a different route, and the two are kept apart because they are exposed to different objections.

Even were a unit available, an arrangement allocating the capacity to generate would require knowledge of what a given allocation would produce. That knowledge is unavailable in a particular way, which is worth distinguishing from ordinary uncertainty. Where a distribution over outcomes is known the situation is one of risk; where the outcomes are known and the distribution is not, the situation is one of uncertainty in the received sense; and the situation here is neither, since the outcomes of generation include work not yet conceived and therefore not enumerable in advance. A decision procedure defined over a fixed set of outcomes has no domain in this case.

A further argument, older and stronger, applies specifically to knowledge. No procedure can predict the future growth of knowledge, since a prediction detailed enough to be acted upon would constitute the discovery it predicts. The impossibility is one of self-defeat and does not rest on any claim about induction, and it is therefore not answered by improvements in inference.

Claim 7.3The operational unavailability of allocation Even where a unit is granted for the sake of argument, an arrangement cannot determine what an allocation of generative capacity would produce, since the outcomes of generation are not enumerable in advance and a prediction of a discovery would constitute it. Allocation of the capacity is therefore unavailable in practice as well as undefined in principle.

7.3 The Limits of the Arguments

Three things are compatible with both claims and are asserted here so that the results are not read more widely than they support.

Resources remain distributable. Money, compute, space, review capacity and attention are quantities with units, and their allocation is a well-posed problem which arrangements must solve. §21 finds that for one of the paper’s difficulties the ordinary allocative treatment is the right one and this framework adds nothing.

Conditions remain alterable. Claim 7.2 concerns the distribution of a capacity, not the alteration of a condition, and Part VI is entirely concerned with the latter. An arrangement that publishes a default, records a class of contribution or reserves a fraction of a ranking has altered conditions for a population without distributing anything.

And comparisons remain available in restricted form. Nothing here excludes the judgement that a participant is worse placed than before, which is a comparison of one participant with itself at two times in respect of a named condition. §11 builds the paper’s assessable object out of comparisons of exactly this restricted kind.

7.4 The Consequences Drawn

Claim 7.4The unavailability of a maximisation objective An objective of the form “maximise generative capacity” is undefined in place of undesirable, since maximisation presupposes an ordering over the values of a quantity and Claim 7.1 denies the quantity. An arrangement that adopts such an objective has in practice adopted an ordering over some measurable proxy, and its objective is then whatever that proxy measures.

The claim strengthens a familiar caution into a result. The caution is that an objective of maximising generativity would reproduce, in a new vocabulary, the aggregative reasoning it was introduced to avoid. The result is that no such objective is available to be adopted, so that what presents as one is always an objective over a proxy, and the choice of proxy is the substantive decision being made without being stated.

Claim 7.5The derivation of plural returns A return to contributors that takes a single form presupposes that the single form is convertible into the requisites the contributors severally lack, at rates independent of the case. Claim 7.1 denies that presupposition. The plurality of return forms is therefore a consequence of the structure of the object and not a design preference, and an arrangement offering one form of return will relieve the constraint of those whose binding constraint that form happens to address.

The claim is used in §22, where the available return mechanisms are set out. It is worth marking that the usual argument for plural returns is practical, resting on the manipulability of any single measure. That argument is sound and is secondary: even an unmanipulable single measure would fail for the reason given here.

Claim 7.6The form in which objectives must be stated An objective concerning generative capacity is statable in the negative and not in the positive: an arrangement may undertake to avoid foreclosing a class of possible continuation, and may not undertake to increase a capacity for which no measure exists. The paper’s objectives are accordingly stated as avoidance of foreclosure throughout.

The three consequences are jointly what permits the framework to say anything at all while denying itself a measure. What remains after Claim 7.4 has removed the objective is a class of statements about particular participants, particular conditions and particular foreclosures, and §11 supplies the structure in which such statements can be organised and assessed.

IIIThe Cycle

8. The Cycle, and Its Selection for the Unit of Analysis

This section states the unit in which the paper’s assessments are made and gives the reason for choosing it. Its objective is to establish that neither a state of the arrangement at a moment nor an arbitrary portion of its relational structure supports the assessments the second-order question calls for, and that a closed cycle does. The section states what is wanted of a unit; rejects the two obvious candidates and states why each fails; defines the cycle and the deficit around it; establishes that a deficit is assessable although the capacity it concerns is not measurable, which is the point on which §10 and this section meet; and lists the kinds of cycle a knowledge commons contains. The formal statement of the cycle and the deeper reason for preferring it are deferred to §15 and §16.

8.1 The Requirements on a Unit

The second-order question asks what a mechanism leaves participants able to do. Answering it requires an object with three properties, and identifying them in advance prevents the choice of unit from appearing arbitrary.

The object must span time, since a question about what an arrangement leaves cannot be answered from a description of one moment. It must be assessable without a measure of generative capacity, since Claim 7.2 denies that such a measure exists. And it must be stable under redescription, since the categories through which an arrangement accounts for itself are chosen, and an object defined within one accounting would vary with the choice, making the paper’s assessments artefacts of the vocabulary in which they were conducted.

8.2 The Candidates That Fail

The first candidate is the state of the arrangement at a moment: who holds what, what has been produced, how attention is distributed. A state fails the first requirement outright. It also fails a weaker version of it, since a sequence of states does not supply what is wanted either: an arrangement whose recorded output rises in every period may be one whose participants are progressively withdrawing, and the sequence of states records the rise and not the withdrawal.

The second candidate is an arbitrary portion of the relational structure: a group of participants, a subfield, a set of works and the relations among them. Such a portion may be examined over time, so it passes the first requirement. It fails the third. What a portion contains, and therefore what is found within it, depends on where its boundary was drawn, and the boundary is drawn by the analyst. A portion that shows a flourishing exchange may be one whose boundary excluded the parties bearing its costs, and nothing internal to the portion reveals this.

Claim 8.1The failure of states and arbitrary portions A state of an arrangement does not span time and therefore cannot answer a question about what a mechanism leaves. A portion of an arrangement’s relational structure spans time and depends on the placement of its boundary, so an assessment conducted within it varies with a choice made by the party conducting it. Neither is a unit in which the second-order question can be answered.

8.3 The Cycle

Definition 8.2Cycle A cycle in a knowledge commons is a sequence of relations beginning at a condition on which some participant’s contribution depends, passing through that contribution, through whatever the contribution produces and through the disposition of what it produces, and returning to a condition of the same kind for the same participant or for the position that participant occupies.

The definition requires a return to a condition of the same kind, which is what distinguishes a cycle from an arbitrary sequence, and what makes the comparison in Definition 8.3 well posed. The return need not be to the same participant, since a position may be occupied by different people over time and the question of whether the position remains occupiable is the one that matters for an arrangement.

Definition 8.3Deficit of a cycle The deficit of a cycle over one traverse is the shortfall of the condition at the point of return relative to the same condition at the point of departure, stated in whatever terms that condition is stated in. A cycle is closing where the deficit is absent or is made good from elsewhere, and is depleting where it is not.

8.4 The Assessability of a Deficit

The relation between this section and §10 requires stating, since a reader who has accepted Claim 7.2 may reasonably suspect that a deficit is exactly the measure just denied.

Claim 8.4The assessability of a deficit without a measure of capacity A deficit compares one condition with itself at two points of a single cycle, and therefore requires only that the condition be comparable with itself. It requires no unit common to conditions of different kinds, no aggregation across cycles, and no cardinal measure of generative capacity. Claim 7.1 denies a common unit across kinds and does not deny that an interval of time is comparable with an interval of time, or a level of access with a level of access.

The claim is what permits the framework to assess anything at all, and its restrictions should be stated with it. Deficits in cycles of different kinds cannot be summed, so an arrangement facing a depleting relational cycle and a closing financial one has two findings and no net figure. Deficits cannot be compared in magnitude across kinds, so no trade-off between them is computable and any such trade-off is a judgement made and not a calculation performed. And a deficit says nothing about what would have been generated had the cycle closed, which is the unavailable prediction of Claim 7.3.

What remains after those restrictions is narrow and is enough: an arrangement may establish, for a named condition and a named position, that participants return with less of it than they began with, and may do so without any claim about how much generation was lost.

8.5 The Kinds of Cycle a Commons Contains

Four kinds recur in the settings this paper concerns, and stating them separately is what makes the coupling in §13 visible.

An epistemic cycle runs from a formulation through criticism and revision to a better formulation, and returns to the participant an improved understanding and the standing to be criticised again. Its conditions are interlocutors competent in the matter and willing to spend attention on unfinished work.

A relational cycle runs from a contribution through recognition to collaboration and returns an enlarged capacity to act, whose conditions are visibility to the relevant parties and a form of recognition those parties act upon.

A material cycle runs from work through its use to funding, employment, compute or access, and returns the means by which the work continues.

An infrastructural cycle runs from shared apparatus through its use to its improvement and returns better apparatus, whose conditions include maintainers, a route by which improvements are contributed back, and terms permitting derivation.

Claim 8.5The independence of the kinds A cycle of one kind may close while a cycle of another kind, involving the same participants, depletes. An arrangement in which the material cycle closes and the epistemic cycle does not is one in which participants continue to be paid and cease to be improved by what they do, and an arrangement in which the reverse holds is one whose participants are improved and cannot continue.

The claim prepares §13, where the more consequential case is treated: not that the kinds are independent, but that they interact, so that a closing cycle of one kind may be closing at the expense of a depleting cycle of another.

9. Closure of the Circuit and Openness of the Boundary

This section establishes that closure of a cycle is not by itself a criterion of anything, and states the second condition required alongside it. Its objective is to meet an objection that defeats the cycle apparatus in one line if it is left unanswered, namely that a cartel has a perfectly closing cycle, and to show that the answer is not an appeal to justice from outside the framework but a second property of cycles that the framework can state. The section presents the objection at full strength; states the second condition and the criterion by which it is assessed; establishes the independence of the two conditions by exhibiting all four combinations; and states the consequence for the criterion of value circulation adopted in §7. The section’s conclusion constrains every recommendation made in Part VI.

9.1 The Objection at Full Strength

An arrangement in which a small group of participants exchange recognition, cite one another, review one another’s work favourably, direct opportunities to one another and thereby retain the conditions of their continued production has a cycle that closes completely. Each traverse returns its participants to a position at least as good as the one they began from. By Definition 8.3 the cycle has no deficit, and by everything said so far it is in good order.

The objection is not that such an arrangement is disagreeable. It is that the framework, having replaced a measure of generative capacity with an assessment of cycles, has produced a criterion satisfied most completely by the arrangements it was constructed to identify as pathological. A closing cycle in a closed group is more reliably closing than an open one, since its participants can ensure the return.

Claim 9.1The insufficiency of closure Closure of a cycle establishes that participants within the cycle return to the conditions they began from. It establishes nothing about who is within the cycle, and a cycle whose participants are fixed will close more reliably than one whose participants change. Closure is therefore compatible with, and is favoured by, restriction of participation.

9.2 The Second Condition

The property the objection identifies is not about the circuit but about its boundary, and it can be stated as a condition on the arrangement.

Definition 9.2Openness of a boundary The boundary of a cycle is open where a participant not currently within it may enter on terms stated in advance and applying to all who meet them, without the consent of those already within. It is closed where entry requires a determination made by incumbents.

The criterion is deliberately about the form of the entry condition and not about how many enter. A cycle few enter may have an open boundary, and a cycle many enter may have a closed one where entry is by invitation extended widely. What matters is whether a party wishing to enter faces a stated condition it may satisfy or a judgement it must attract, and the reason for locating the criterion there is developed in the companion paper on the ethics of transmission (Huang 2026g), where a determination made by an interested party about whether another is ready is shown to be administered by the party a refusal relieves, insulated from the objection of the party it concerns, and therefore incapable of correction by anyone affected.

9.3 The Independence of the Conditions

Claim 9.3Independence Closure of the circuit and openness of the boundary are independent properties, and all four combinations occur. A cycle may close with an open boundary, which is the case the framework recommends. It may close with a closed boundary, which is the arrangement of the objection. It may deplete with an open boundary, which is the ordinary case of a commons whose participants may join freely and are not sustained. And it may deplete with a closed boundary, which is an arrangement that excludes and exhausts at once.

The four combinations are worth naming because they call for different instruments and are commonly confused. An arrangement facing depletion with an open boundary needs a return pathway and needs nothing done about entry. An arrangement facing closure with a closed boundary needs its entry terms altered and will not be improved by any strengthening of return, since strengthening return in a closed cycle concentrates further. Treating the second as though it were the first is the characteristic error the classification of §6 is meant to prevent, and it is what the amplification mechanism describes.

9.4 The Consequence for the Circulation Criterion

Claim 9.4The circulation criterion as necessary and insufficient The criterion that value should return to the arrangements that generated it, adopted in Claim 4.1, states the closure condition and does not state the boundary condition. An arrangement may satisfy it completely while restricting who may participate in the cycle to which value returns. The criterion is therefore necessary for the assessment this paper conducts and is not sufficient for it, and an arrangement assessed by circulation alone may be assessed favourably in precisely the configuration the assessment was intended to exclude.

The claim is the paper’s principal addition to the account it adopts, and it should be read as an extension of that account and not as a correction of it. The account’s concern with generators retaining control over the conditions of their own production points toward the boundary condition without stating it, since control over one’s conditions of production is compatible with holding those conditions against others. What this paper adds is that the two must be stated separately and assessed separately, because an arrangement can improve on one while worsening on the other, and improving return within a bounded group does exactly that.

A second consequence bears on Part VI. Every instrument recommended there is assessed against both conditions, and an instrument that improves closure at the cost of boundary openness is reported as such. The reservation of ranking capacity in §20.1 is examined in this light, since a reservation is a term applying to all and is administered by a party with an interest in its administration.

10. Coupling Among Cycles and the Asymmetry of Measurement

This section treats the interaction among cycles of different kinds and derives from it the paper’s account of why arrangements fail while their measures improve. Its objective is to establish that a cycle may close by drawing on another which depletes, that an arrangement observes the cycles it meters and not the others, and that these two facts together produce a characteristic pattern of failure which no improvement in the metering of the observed cycle will correct. The section states what coupling is; states the substitution it makes possible; states the asymmetry between the cycles an arrangement can observe and those it cannot; states the pattern the two produce jointly; and states what an arrangement can do about a cycle it does not meter. The results here are used throughout Part V and are the basis of the amplification mechanism of §6.

10.1 Coupling

Definition 10.1Coupling Two cycles are coupled where a condition returned by one is a condition on which the other depends, so that the deficit of one enters the other as a change in a condition.

Coupling is the ordinary case in a knowledge commons and not a special one. The material cycle returns funding and employment, which are among the conditions of the epistemic cycle, since a participant without a position has no occasion to expose unfinished work. The relational cycle returns interlocutors, who are conditions of the epistemic cycle directly. The infrastructural cycle returns apparatus on which all three depend.

Claim 10.2Substitution across coupled cycles Where two cycles are coupled, a deficit in one may be made good from the other, so that a cycle closes by drawing on a coupled cycle which thereby depletes. An arrangement observing only the first records a closing cycle, and the observation is correct.

The claim describes the mechanism by which several of the difficulties of §6 occur without anyone acting wrongly. A platform whose material cycle closes because contributors are compensated may be drawing on a relational cycle that depletes, if the compensation is directed to those already visible and the visibility of the others declines. A platform whose infrastructural cycle closes because the apparatus improves may be drawing on a human cycle that depletes, if the improvement is sustained by maintainers who are not replaced.

10.2 The Asymmetry of Observation

The second element of the account concerns which cycles an arrangement is in a position to observe.

Claim 10.3The asymmetry between metered and unmetered cycles An arrangement observes a cycle to the extent that its records represent the conditions the cycle returns. By Claim 6.3 those records represent what the arrangement has made itself answerable for. Cycles whose returned conditions are metered are therefore observable to it and cycles whose returned conditions are not metered are not, and the second class contains no gap in the record, since nothing was ever recorded to be missing.

The classes are not evenly populated. Conditions that are metered in the settings this paper concerns are quantities of the kind an arrangement transacts in: payments, downloads, citations, counts of contribution, storage consumed. Conditions that are ordinarily unmetered are the returns of the epistemic and relational cycles: whether a participant was improved by the criticism received, whether an interlocutor remains available, whether the standing to expose unfinished work is retained. The distribution is not accidental, since an arrangement meters what it transacts in.

10.3 The Pattern Produced Jointly

Claim 10.4Closure on the metered cycle with depletion on the unmetered Where a metered cycle is coupled to an unmetered one, a deficit may be transferred to the unmetered cycle and the arrangement will record improvement. The arrangement’s measures will indicate health for as long as the transfer continues, and will indicate deterioration only when the depleted cycle fails altogether, at which point the condition it returned is absent and cannot be restored by any operation on the cycle that was measured.

The claim is the paper’s account of the pattern that makes platform failure characteristically sudden. An arrangement does not observe a gradual decline and elect to ignore it. It observes improvement, correctly, on the cycles it can observe, until a cycle it never observed ceases to return anything.

Two features of the pattern deserve marking. The transfer requires no intention, since a mechanism that improves a metered quantity has no way of registering what it drew upon. And the failure is not corrected by better measurement of the metered cycle, which is the response an arrangement will reach for first, since the instruments for it exist and the instruments for the other do not.

10.4 The Response Available for an Unmetered Cycle

The obvious response is to meter the unmetered cycle, and it is available in part and is limited in a way worth stating.

Some conditions returned by epistemic and relational cycles admit indirect indicators: whether participants who contributed in one period contribute in the next, whether new participants enter and are cited by established ones, whether criticism is offered on work that is not yet finished, whether a maintained resource acquires new maintainers. None of these measures the condition, and each is a signal that the cycle is or is not operating.

Two limits attach. An indicator adopted to administer anything becomes an object participants produce, which is a general result and is treated at §20.2. And metering a cycle brings it within the arrangement’s answerability by Definition 6.1, which extends what the arrangement can govern and also what it may be held to, and the extension is not costless to the participants whose activity is now recorded.

Claim 10.5The residue after indirect indicators Indirect indicators establish whether a cycle is operating and do not establish the magnitude of its deficit. An arrangement using them can determine that a coupled cycle has ceased to return and cannot determine how much of the deficit its own metered improvements transferred there. The attribution of a deficit across coupled cycles is unavailable for the reason given in §10, and the framework asks only whether the cycle returns.

11. The Pathologies of Circulation

This section classifies the ways a cycle fails and states the instrument each failure calls for. Its objective is to convert the two conditions of §12 and the coupling result of §13 into a working typology, so that Part V’s treatment of individual difficulties has something determinate to assign them to. The section states five pathologies with the property that identifies each; treats separately the one that is invisible to an assessment by closure alone; states the failure mode of a cycle that satisfies both conditions and is nonetheless in poor order; and presents the typology in Table 2. The pathologies are stated as structures and no claim is made about which arrangements exhibit which.

11.1 The Pathologies

One-way transfer. A cycle does not close, and what it produced passes to a party standing in no returning relation to it. The identifying property is the absence of any pathway, so that no quantity of return could be said to be owed under the arrangement’s own terms.

Blocked return. A pathway exists and does not carry. The identifying property is that the arrangement provides for a return which does not occur, whether because the return is discretionary, because the party owing it is not identified, or because the form in which it is offered is not the form the recipient’s constraint requires, which is the case Claim 7.5 predicts.

Accumulation within a closing cycle. The cycle closes, and the capacity it carries accumulates at one position within it, so that the other positions return to conditions sufficient to continue and insufficient to act independently. The identifying property is that the deficit is absent at every position and the distribution of the returned conditions changes each traverse.

Rigidification. The cycle closes, the boundary is closed, and both conditions are stable. The identifying property is that participation is fixed, which is the objection of §12 treated as a pathology in its own right.

Exhaustion of a coupled cycle. A cycle closes by drawing on a coupled cycle which depletes, in the sense of Claim 10.2. The identifying property is that the observed cycle is in good order and a condition on which it depends is deteriorating outside the arrangement’s records.

11.2 The Pathology Invisible to Assessment by Closure

The third pathology requires separate treatment, since it is the case an assessment by closure alone will pass.

Claim 11.1Accumulation under closure A cycle in which every position returns to a condition no worse than it began from may nonetheless transfer capacity to one position at every traverse, since a return that is sufficient and a return that is proportionate are different things. An arrangement assessing only whether positions are sustained will find such a cycle in good order at every traverse, and will find at no traverse the fact that the positions have become unequal in their capacity to act without the position that accumulated.

Two consequences follow. The pathology is detectable only by comparing the distribution of returned conditions across traverses, which requires records spanning traverses and is therefore subject to Claim 10.3. And its terminal state is a cycle that satisfies both conditions of §12 and in which the other positions can no longer leave, since their capacity to act independently has been the thing transferred. Dependence arrived at through sustained returns is not excluded by either condition.

11.3 The Failure Mode of a Well-Ordered Cycle

A cycle that closes with an open boundary and without accumulation may still be in poor order, and the framework should say so in place of treating its two conditions as jointly sufficient.

Claim 11.2Foreclosure by a well-ordered cycle A cycle that closes reliably returns its participants to the conditions from which they began, and thereby reproduces the arrangement that produced those conditions. Where the conditions are ones under which a narrow class of work is producible, reliable closure reproduces that narrowness, and the arrangement forecloses lines of work that a less reliable cycle would have left open.

The claim is why the paper’s objective is stated as the avoidance of foreclosure in Claim 7.6 and not as the closure of cycles. Closure is an instrument and not an end, and an arrangement that closed every cycle perfectly would have made itself unable to become anything other than what it is. The tension is genuine and the paper does not resolve it: the same property that sustains participation reproduces the conditions under which participation occurs, and no criterion offered here determines how reliably a cycle should close.

| @p3.0cmp4.4cmp4.4cm@

Pathology	Identifying property	Instrument called for
One-way transfer	No returning pathway exists	Establish a pathway; a party must be found to bear it
Blocked return	A pathway exists and does not carry	Make the return non-discretionary, identify the party owing it, or supply it in another form
Accumulation under closure	Deficit absent at every position; returned conditions unequal and diverging	Alter the operation that concentrates, ordinarily the arrangement’s own ranking or matching
Rigidification	Circuit closed and boundary closed	Alter the terms of entry to a stated condition applying to all
Exhaustion of a coupled cycle	Observed cycle in good order; a condition it depends on deteriorating unrecorded	Meter the coupled cycle by indirect indicator, accepting the limits of Claim 10.5

Table. The pathologies of circulation, the property identifying each, and the kind of instrument each calls for.

11.4 The Relation to the Mechanisms

Claim 11.3The correspondence with the classification The pathologies map onto the mechanisms of Definition 3.1 as follows. Exhaustion of a coupled cycle is a failure of measurement, since the arrangement’s records do not represent the depleting cycle. One-way transfer and blocked return are failures of return. Accumulation under closure and rigidification are failures of amplification, the first through the transfer of capacity within the cycle and the second through the fixing of who participates in it.

The correspondence is what makes the classification of §6 operational, since a difficulty assigned to a mechanism there may now be assigned to a pathology here and thereby to an instrument. Part V performs that assignment for each difficulty in turn and reports the cases in which the assignment is contested or in which more than one pathology is present.

IVThe Formal Discussion

12. A Formal Statement of the Cycle and Its Deficit

This section states in formal terms the objects Part III introduced in prose. Its objective is to fix the notation the remainder of Part IV uses, to make the restrictions of Claim 8.4 precise, and to show that the deficit is well defined under exactly the conditions the informal treatment claimed and under no others. The section states the underlying structure; defines a cycle and a traverse within it; defines the deficit and states the conditions of its being well defined; states the two operations that are unavailable and why; and states what the formalism is for, since a formal statement that is mistaken for a model would carry more than this paper claims. The construction is elementary and no result is asserted beyond what it makes explicit.

12.1 The Underlying Structure

Let $P$ be a set of positions in an arrangement, occupied by participants and persisting across changes of occupant. Let $\mathcal{K}$ be a set of kinds of condition, disjoint by construction, with a distinct kind for each of the classes named in §9. For a position $p \in P$ and a kind $k \in \mathcal{K}$, write $c_k(p,t)$ for the level of the condition of kind $k$ at position $p$ at time $t$, taking values in a set $V_k$ which carries an order and need not carry an addition.

The disjointness of kinds and the absence of a common carrier across the $V_k$ are the formal content of Claim 7.1. Nothing in the structure permits a value in $V_k$ to be compared with a value in $V_{k’}$ for $k \neq k’$, and no operation is supplied for combining them.

An arrangement’s records are represented by a subset $\mathcal{K}{\mathrm{obs}} \subseteq \mathcal{K}$ of kinds it meters, which by Definition 6.1 is also the set of kinds for which it has made itself answerable. The distinction between $\mathcal{K}{\mathrm{obs}}$ and $\mathcal{K}$ is the formal content of Claim 10.3.

12.2 Cycles and Traverses

A cycle is a sequence of relations returning to a condition of the kind it began from, and the formal statement fixes what returning means.

Definition 12.1Cycle and traverse A cycle $L$ is a pair $(p,k)$ together with a sequence of relations by which the condition $c_k(p,\cdot)$ is drawn upon and subsequently supplied. A traverse of $L$ is an interval $[t_0,t_1]$ containing one such drawing and one such supply, so that $c_k(p,t_0)$ is the level at departure and $c_k(p,t_1)$ the level at return.

Two features of the definition are worth marking. A cycle is indexed by a position and a kind, so that the same participants stand in several cycles at once, which is what §13 required. And the traverse is an interval and not an instant, so the boundaries $t_0$ and $t_1$ are choices, and §24 records the sensitivity of a deficit to that choice.

12.3 The Deficit

Definition 12.2Deficit The deficit of a cycle $L=(p,k)$ over a traverse $[t_0,t_1]$ is the relation between $c_k(p,t_1)$ and $c_k(p,t_0)$ under the order on $V_k$. The cycle is closing over the traverse where $c_k(p,t_1) \succeq c_k(p,t_0)$ and depleting where $c_k(p,t_1) \prec c_k(p,t_0)$.

Claim 12.3The conditions under which a deficit is well defined The deficit of Definition 12.2 is well defined where three conditions hold: the position $p$ persists across the traverse; the kind $k$ is the same at $t_0$ and $t_1$, so that what is compared is a condition with itself; and $V_k$ carries an order. It requires no addition on $V_k$, no common carrier across kinds, and no numerical measure of anything.

The claim is the formal statement of Claim 8.4, and it identifies precisely what a critic must attack. A reader who holds that the framework smuggles in the measure it denies must locate the smuggling in one of the three conditions, and the third is the only one with any content: the framework requires that levels of a condition of one kind be ordered, which is a substantive requirement and a weak one. It holds where a participant can say that they have less time available than before, or less access, or fewer interlocutors competent in the matter, without saying how much less.

12.4 The Unavailable Operations

Two operations are absent from the structure by construction, and their absence is what the formalism is chiefly for.

Claim 12.4Unavailability of aggregation across kinds No operation combines a deficit in $V_k$ with a deficit in $V_{k’}$ for $k \neq k’$, since the structure supplies no map between them. An arrangement facing a depleting cycle of one kind and a closing cycle of another therefore holds two findings and no net finding, and any judgement it makes between them is a judgement and not a computation.

Claim 12.5Unavailability of attribution across coupled cycles Where cycles $L$ and $L’$ are coupled in the sense of Definition 10.1, the structure does not determine what portion of the deficit of $L’$ was transferred from $L$. Establishing such a portion would require a common carrier across kinds where the coupling crosses kinds, and a counterfactual level of $c_{k’}(p’,t_1)$ had the coupling been absent, which is a prediction of the kind Claim 7.3 excludes.

The second claim is the formal statement of the restriction recorded at Claim 10.5, and it settles a question a reader of Part III may have held open. The framework can establish that a coupled cycle has ceased to return. It cannot establish that a particular mechanism operating on a metered cycle caused that cessation, and it does not attempt to.

12.5 The Purpose of the Formal Statement

The construction above is not a model of an economy and predicts nothing, and saying so is part of stating it.

Claim 12.6What the formalism is for The construction fixes what a deficit is and what its assessment requires, exhibits the two operations the framework denies itself as absences in a structure in place of prohibitions imposed upon it, and identifies the three conditions on which the assessability of a deficit depends. It supplies no dynamics, no equations of motion for $c_k$, and no means of determining the level of any condition at a future time.

The absence of dynamics is deliberate and follows from §18. A structure with dynamics would specify how conditions evolve, and any such specification would presuppose the enumerable outcome space that Claim 7.3 denies. What the framework offers in place of dynamics is an assessment conducted after a traverse has occurred, together with the restraint that §18 states.

13. The Gauge Argument for the Cycle

This section gives the deeper reason for the choice of unit made in §11. Its objective is to establish that the cycle is not merely a convenient object in a nonlinear setting but the object that survives a change in the description under which an arrangement is accounted for, and to do so by an argument the author has developed elsewhere for evaluation conducted without an external frame. The section states the premise about description that Part I already relied upon; states what follows for quantities defined within a description; states the property a closed circuit has that a path does not; applies the author’s own admission test for material drawn from another domain; and states the residue that survives whichever way the transfer is judged. The section is the most exposed in the paper and its restrictions are stated with unusual care.

13.1 The Premise About Description

Nothing counts as economic activity independently of the categories through which a society recognises value, labour, ownership and obligation. Cooking in a household and cooking in a restaurant may be identical as physical activity while one enters a national account and the other does not, and the difference lies in the categories and not in the activity. §7 recorded that this observation is well established, and §9 used a version of it in fixing what counts as a generative condition.

What follows, and what §11 relied upon without arguing, is that no position is available from which the categories could be fixed once and assessed as correct. An assessment conducted within a set of categories inherits them; an assessment conducted from outside is unavailable, since there is no outside.

Claim 13.1The dependence of quantities on the description A quantity defined within an economic description, including a contribution’s worth, a share of a downstream gain, and the value of an input, is determined jointly by the activity described and by the categories describing it. Two descriptions of one arrangement may therefore assign different values to the same contribution without either being in error, and no third description adjudicates between them.

The claim is the reason the attribution question fails in economics for a reason additional to the causal one. Part V records that attributing a share of a downstream gain to a particular upstream contribution is unavailable because the causal structure does not support it. Claim 13.1 adds that even were the causal structure fully known, the share would be determined only relative to a description.

13.2 The Objects Surviving a Change of Description

If quantities defined within a description vary with it, an assessment that is to say anything about an arrangement must be conducted in terms of what does not vary.

Claim 13.2The invariance of a closed circuit A comparison of a condition with itself at the two ends of a closed circuit is stable under a change of description in a way that the level of that condition at either end is not. Whether a participant returns to a position from which they may continue does not depend on the accounting in which their contribution is valued, while the value assigned to their contribution does.

The claim states in the paper’s own terms a structure developed formally in the author’s companion work on loop observables and the gauge structure of interpretation (Huang 2026b). There the general position is established: where a description may be changed at each point without altering what the description is of, the quantities defined pointwise carry no information about the underlying situation, and the quantities that do are those associated with closed circuits, since the changes of description made along the circuit are undone on return. The companion paper’s concern is with evaluation conducted where no external frame is available, which is the situation Claim 13.1 identifies for economic description, and the transfer to the present case is therefore a transfer of the same problem in a different subject matter.

13.3 The Admission Test Applied

Material drawn from another domain requires justification, and the test applied here is one the author has stated elsewhere and which this section is obliged to meet. It has four conditions.

The comparison must do work that no literal statement in the argument performs. It does: Claim 13.2 gives a reason for the choice of unit which the pragmatic considerations of §11 do not supply, since those establish only that states and portions are unsatisfactory and leave open what should replace them.

The source domain must be one in which the relevant structure is observed and not composed, or the composed character must be stated and the inference restricted. The structure invoked is a mathematical one and is neither observed nor composed in the relevant sense, and the restriction taken here is accordingly severe: what is transferred is the argument that pointwise quantities are description-relative while circuit quantities are not, and nothing about the dynamics, the symmetry group, or the physical content of any theory in which that argument was developed.

The comparison must assign no position to the author or reader that the argument has not established. It assigns none.

And the claim must be statable without the comparison, so that a reader rejecting it retains the claim. It is: Claim 13.2 as stated makes no reference to gauge, and a reader may treat it as an assertion about accounting to be assessed on the examples, of which §21.2 and §19.4 supply several.

Claim 13.3The scope of the transfer What is taken from the companion treatment is a structural argument about the description-relativity of pointwise quantities and the invariance of circuit quantities. Nothing in this paper asserts that an economy is a gauge theory, that its descriptions form a group, or that any quantity here corresponds to a physical one. A reader who rejects the transfer entirely is asked to assess Claim 13.2 directly.

13.4 The Residue

Two things survive whichever way the transfer is judged, and stating them is how the section discharges its exposure.

The first is the negative result. Claim 13.1 is established independently of anything in this section’s second half, and it is sufficient for the paper’s refusal of attribution by share, which Part V relies upon. A reader who accepts Claim 13.1 and rejects Claim 13.2 has conceded that the paper’s central negative move is sound and is entitled to ask for a different positive object.

The second is that the positive object, if the transfer is rejected, must still meet the requirements of §11: it must span time, be assessable without a measure of capacity, and not vary with the analyst’s boundary. The paper knows of no other candidate satisfying all three, and offers that as a reason for the cycle in place of a proof of it.

14. The State Space and Its Variation in Time

This section treats a question the framework raises and does not answer. Its objective is to state the difficulty precisely, to identify the domain in which a version of it has been given a formal treatment, and to say what this paper takes from that domain and what it declines to take. The section states the two models an account of generation appears to require; states the difficulty of holding both; states the question of whether they are two things or one; identifies the formulations in which that question has an answer and states the answer’s form; states the nearest treatment outside the mathematical sciences; and states this paper’s position, which is that the question is open and that the framework does not depend on its resolution. The section is exploratory and is marked as such.

14.1 The Models an Account of Generation Requires

An account of a system’s behaviour ordinarily supplies a space of states together with a rule by which the system moves within it. The space fixes what configurations are possible, and the rule fixes which succeed which.

An account of generation appears to require something further. What is generated is not only a new position within a space of possible positions but, in the cases this paper concerns, a new dimension of that space: a technique that did not exist, a distinction that had not been drawn, a relation between fields that were not previously related. A conception of the possible that was adequate before the generation is not adequate after it, and the inadequacy is not an error, since nothing available beforehand would have supplied the missing dimension.

So an account seems to need two models: one of movement within a space, and one of the alteration of the space. The two are ordinarily kept separate, and keeping them separate raises a difficulty, since the second is not a movement in the first and the first has no representation of the second.

14.2 The Question

The question is whether the two models describe two things or one. Stated in the vocabulary of the tradition from which the present author draws it, the question is whether the geometry of the possible and the generation that alters it are non-dual.

The question is not idle for this paper. Every framework that governs by intervening in conditions presupposes some conception of what the conditions make possible, and if that conception must be revised by the very generation it was to accommodate, the framework’s own standing is at issue. §18 states the consequence for what the framework may claim.

14.3 The Formulations in Which the Question Has an Answer

A class of formulations in the mathematical sciences takes as its point of departure the refusal of a fixed background space. In such formulations there is no manifold given in advance upon which a dynamics is defined; what would ordinarily be the geometry is itself among the things the theory determines, and quantities such as dimension appear as properties of a solution in place of properties of the setting. The general name for this posture is background independence, and it is adopted in several distinct research programmes concerned with the structure of spacetime, including approaches in which the geometry emerges from the combinatorial structure of discrete elements.

What such formulations supply for the present question is not an answer about economies but a demonstration that the question is coherent and that an affirmative answer is formulable: there are settings in which the space and the process upon it are not two objects, and in which treating them as two is an approximation valid where the space happens to be changing slowly.

The last clause is the one that bears on this paper. Where a space changes slowly relative to the movement within it, the two-model treatment is a good approximation and the ordinary apparatus applies. Where it does not, the approximation fails, and the failure is not remedied by refining the model of movement.

14.4 The Nearest Treatment Outside the Mathematical Sciences

A related position has been argued in the study of biological evolution, where it is held that the space of possibilities for an evolving biosphere cannot be stated in advance, that no law entails its evolution in the way that a law entails a trajectory in a fixed space, and that what becomes possible at each moment is constituted by what has just come to exist (Longo, Montévil, and Kauffman 2012; Kauffman 2000). The associated notion of an adjacent possible names the set of configurations reachable in one step from the present, which is itself altered by every step taken.

That position is contested, and this paper takes no side in the dispute. It is recorded because it is the nearest existing statement of the premise this paper’s Claim 7.3 depends upon, and because it was reached in a domain where the objects are observed in place of constructed.

14.5 The Position Taken Here

Claim 14.1The status of the question in this paper Whether the space of possibilities and the generation that alters it are one object or two is not settled here, and no claim in this paper depends on its resolution. The framework’s assessments are conducted over completed traverses of cycles, by comparisons of a condition with itself, and such comparisons require no specification of the space of possibilities at either end.

The claim is what permits the paper to raise the question without being answerable for it. The framework was constructed in §15 without dynamics, and the absence of dynamics is exactly the absence of a commitment about the space and its variation.

Claim 14.2The consequence of the question for the framework The question bears on the framework at one point. An arrangement intervening in conditions acts on a conception of what those conditions make possible, and that conception is revised by generation the arrangement did not foresee. The framework therefore cannot claim that its interventions have the effects intended for them, and it claims in place of that that a completed traverse may be assessed and that an intervention may be revised on what the assessment shows.

Two further remarks bound the section. Nothing here asserts that an economy is a background-independent system, that its conditions form a space in any technical sense, or that any result from the formulations named transfers to it; the section states where a version of the question has been treated and stops. And the section may be removed entire without affecting any other part of the paper, which is the appropriate standing for an exploratory passage and is stated so that a reader may treat it accordingly.

15. The Limits of the Formalism and the Governance Form That Remains

This section states what the apparatus of Part IV cannot do and what follows for the way an arrangement should act. Its objective is to establish that prediction of generation is unavailable in a specific and irremediable sense, to show that this unavailability is the framework’s premise in place of an objection to it, and to state the form of intervention that remains once prediction is given up. The section states the three grades of ignorance and locates the present case; gives the argument that is decisive for knowledge specifically; states why one natural repair fails; states what uncertainty propagation does and does not supply; states the two devices that remain; and states the residue that neither device addresses. The section is where the paper is at its most exposed and the exposure is stated in place of being managed.

15.1 The Grades of Ignorance and the Present Case

Where a distribution over outcomes is known, a decision may be taken by expectation. Where the outcomes are known and the distribution is not, a decision may be taken by rules constructed for that case. Neither describes the present situation.

Claim 15.1The unenumerability of the outcome space The outcomes of generation include work not yet conceived, and a set containing such work cannot be enumerated in advance, since the enumeration would require conceiving it. A decision procedure defined over a set of outcomes therefore has no domain in this case, and its failure is the absence of the domain in place of an error within it.

The claim identifies why the ordinary apparatus is unavailable and not merely inconvenient. A body of work in decision theory addresses situations of this kind under the heading of unawareness, and it establishes that a decision-maker may be unaware of a proposition and that the revision of beliefs upon becoming aware of it is not an application of the ordinary rule (Karni and Vierø 2013). That the problem has a literature and remains unsettled is the honest description of its standing.

15.2 The Argument That Is Decisive for Knowledge

A further argument bears specifically on the case this paper concerns and is stronger than the foregoing, since it does not depend on any claim about inference.

Claim 15.2The self-defeat of prediction of knowledge No procedure predicts the future growth of knowledge. A prediction specific enough to be acted upon would state the content of what is to be discovered, and a statement of that content is the discovery. The impossibility is one of self-defeat and is not remedied by better methods of inference, greater data, or longer observation (Popper 1957).

The claim is why this paper does not treat unpredictability as a deficiency to be reduced. Where generation is predictable it is production: the arrangement knows what will be made and the problem is the ordinary one of arranging for it to be made efficiently. Generation is a distinct category precisely in the cases where the prediction is unavailable, and so the unavailability is the framework’s premise. An objection that the framework cannot predict what will be generated has stated the condition under which the framework’s subject matter exists.

15.3 The Repair That Fails

A natural response is to give up global prediction and retain local prediction: to work within a neighbourhood of the present configuration where the relations are approximately known, and to act only within it. The response is sound in spirit and the obvious formal expression of it fails for two reasons that are worth stating, since the expression is tempting.

Working in a linear approximation about a point presupposes a space of fixed dimension in which such an approximation is defined, and §17 recorded that the difficulty at issue is the alteration of the space. An approximation of that kind imports the assumption whose failure is the problem.

And a linear approximation is least reliable near a boundary at which a system’s behaviour changes qualitatively, which is where an arrangement most needs guidance. The device is unavailable exactly in the circumstances that motivate seeking it.

15.4 The Contribution of Uncertainty Propagation

Propagation of uncertainty through a known map is the standard instrument where a model is available, and its position here should be stated precisely, since it is available for one purpose and not for another.

Propagation requires a fixed space and a map defined upon it, and neither is available for generation itself, since a distribution cannot be carried forward onto dimensions absent from the domain of the map. What is available is propagation applied to a diagnosis: an arrangement that has identified a binding constraint holds a claim about the present, and the interval over which that claim remains supported may be estimated by propagating what is known about how quickly the relevant conditions change.

Claim 15.3The object uncertainty propagation supplies Propagation of uncertainty does not supply a prediction of what will be generated. It supplies the interval over which a diagnosis of a binding constraint remains supported, past which the arrangement’s grounds no longer distinguish the courses available to it. The quantity obtained is the shelf life of a diagnosis and not a forecast of an outcome.

The result corresponds to the horizon developed in the author’s companion treatment of evidence whose reference time differs from the time of decision (Huang 2026c), where a horizon is defined as the interval past which a body of evidence ceases to separate the alternatives under consideration. What that treatment establishes about the horizon transfers here without amendment, including its dependence on the sensitivity of the system and the consequence that a rapidly changing setting yields a short horizon.

15.5 The Devices That Remain

What remains after prediction is given up is a manner of acting, and two devices constitute it.

The first is a restriction of action to a region in which the arrangement’s account of the relations has been observed to hold, together with a rule revising the region on what the action produces: where the observed result accords with the account, the region is widened; where it does not, the region is narrowed and the account is revised. The device requires no space of fixed dimension and no prediction of outcomes, since it compares a predicted improvement with a realised one and acts on the comparison.

The second is a preference for reversible action where the account is weak. Where an arrangement cannot foresee what a change will produce, it may prefer the change it can undo, and the preference is available without any estimate of what will follow. This device is developed at length in the author’s companion treatment of decision under evidence of uncertain currency (Huang 2026c), where the burden of justification is allocated to the party proceeding where effects are substantially irreversible.

Claim 15.4The governance form available without prediction An arrangement unable to predict what will be generated may nonetheless act by restricting intervention to the region in which its account has held, revising that region on the comparison between predicted and realised results, and preferring reversible interventions where the account is weak. None of the three requires a prediction of generation, a measure of generative capacity, or a specification of the space of possibilities.

15.6 The Residue

Two things are not addressed by either device and are stated here in place of the closing section, since they belong to this argument.

The devices govern the manner of intervention and not the choice of what to intervene upon. That choice rests on a diagnosis of a binding constraint, and a diagnosis is a claim about which conditions matter, which is a causal claim about generation of the kind this section has shown to be unavailable in general. The framework’s answer is that a diagnosis of a present constraint is an observation where a prediction of an outcome is not, and that a mistaken diagnosis is corrected by the same comparison the first device performs. The answer is partial, and §24 records it as a weakness.

And some conditions are recognised as generative only afterwards: the interval of unstructured time, the encounter that was not arranged, the tolerated deviation, the resource maintained for years without use. An arrangement that requires a justification in advance for what it supplies will systematically under-supply conditions of this class, since the justification becomes available only after the condition has done its work.

Claim 15.5The selection against retrospectively recognised conditions An arrangement that supplies conditions only on prospective justification under-supplies the class of conditions whose contribution is recognisable only in retrospect. The under-supply follows from the arrangement’s justification requirement and not from any judgement about the conditions, and it is therefore invisible in the arrangement’s own terms, since each individual refusal is correct by the standard applied.

VThe Difficulties Examined

16. The Difficulties for Which the Question Alters the Prescription

This section examines four of the difficulties assembled in §6 and finds that for each the second-order question and the ordinary description prescribe different actions, which is the test of Definition 5.2. Its objective is to show the test being applied and passed, and to state for each difficulty what the arrangement should do, what limits the recommendation, and what would show the interpretation to be mistaken. The four are treated in a uniform order: the difficulty, the ordinary description and what it prescribes, what the question adds, the intervention, the limitation, and the condition of refutation. The first is treated at greater length than the others, since it is the case in which the two prescriptions diverge most sharply and the case on which the paper’s practical claim principally rests.

16.1 The Capture of Relational Infrastructure

The difficulty. An arrangement may hold every work openly while retaining the social graph, the recommendation and ranking systems, the usage and interaction records, the derived representations of works and participants, and the matching by which a reader is brought to a work. Which parties encounter which work is then determined by an apparatus no participant can inspect and none may take elsewhere.

The ordinary description and its prescription. The apparatus is a proprietary asset lawfully developed and held, and its concentration is treated as a question of market power: whether the holder’s position permits it to raise prices, to exclude rivals, or to foreclose adjacent markets. Where no such effect is shown, the ordinary description prescribes nothing, and where one is shown it prescribes remedies addressed to competition among platforms.

What the question adds. By Definition 6.1 the ranking and matching apparatus is a generative condition, since the arrangement supplies it, alters it, and thereby determines which participants encounter which work. It is moreover the condition on which the operation of every other cycle depends: whether a contribution attracts the criticism that improves it, whether recognition reaches a party positioned to collaborate, whether an unused resource is found by a party who needs it. The apparatus is therefore the cycle of §11 that governs the amplification of the others, and by Claim 10.3 it is also the cycle whose records the operator alone holds.

The divergence of prescriptions is complete. The ordinary description finds nothing to prescribe where competition is unaffected, and an arrangement may hold this apparatus without any effect on prices at all. The second-order question finds a condition of every participant’s continued contribution held by a party under no obligation concerning it, and prescribes accordingly.

The intervention. Three measures follow and none requires the apparatus to be surrendered. Portability of what a participant has contributed to the apparatus, meaning the relations and records generated by that participant’s own activity, so that participation elsewhere does not begin from nothing. Transparency of the operative rule at a level permitting a participant to know on what basis work is surfaced, which is weaker than disclosure of an implementation and sufficient for the purpose. And an interface by which parties other than the operator may build alternative surfacing over the same corpus, which converts a single point of determination into several.

The limitation. Each measure has a cost the paper does not minimise. Portability of relational records raises questions about the parties on the other side of those relations, who did not consent to their transfer. Transparency of a surfacing rule makes the rule easier to exploit, and the operator’s usual answer to this is not frivolous. And alternative surfacing fragments a corpus, which may reduce the encounters it was intended to increase. The recommendation is that these costs be stated and weighed and not that they are small.

The condition of refutation. The interpretation is mistaken if the apparatus proves not to determine encounters materially, which would show if participants located work predominantly by routes the operator does not control. It is narrowed if the three measures prove to leave the determination where it was, which would show if alternative surfacing were built and went unused.

16.2 Generative Maintenance and the Invisibility of Contribution

The difficulty. Reviewing, correcting, replicating, translating, moderating and curating produce no artifact of the kind the arrangement records, and are performed without accounting or recognition.

The ordinary description and its prescription. The activity is unpaid labour under an incentive structure that rewards other activity, and the prescription is to pay for it or to credit it, which is to make it an artifact of the recorded kind.

What the question adds. The activities named do not produce; they maintain a condition on which others’ production depends. A moderator produces no work and preserves the trust without which participants do not expose unfinished work; a curator produces no work and preserves the usability of a resource. By Claim 6.3 an arrangement can govern only what it has made itself answerable for, and by Claim 10.3 an arrangement meters what it transacts in, so activity that alters a condition and produces no artifact is invisible to an accounting constructed around artifacts. The invisibility is therefore not a failure of recognition to be corrected by better manners; it is a property of the accounting.

The prescriptions differ. The ordinary description prescribes payment or credit for a task. The question prescribes that the arrangement record the maintenance relation itself, so that the condition maintained becomes visible as a condition, which is what permits its deterioration to be observed before the maintainers withdraw.

The intervention. Record maintenance relations in the provenance of a work: which parties reviewed, corrected, curated or translated what, and when. The record is cheap, since the events occur within the arrangement’s own systems, and it is the precondition of every return mechanism in §22.

The limitation. A recorded relation becomes a countable one, and §20.2 states what follows. The recommendation here is for the record and not for a score derived from it.

The condition of refutation. The interpretation is mistaken if maintenance proves to be adequately sustained by payment alone, which would show if arrangements that pay for review retained reviewers at rates comparable to those that record and credit the relation.

16.3 The Reproduction of Contributors

The difficulty. An arrangement’s recorded output rises while the participants sustaining it withdraw, and its measures indicate health until the withdrawal is substantially complete.

The ordinary description and its prescription. The difficulty is volunteer attrition or a provision failure, and the prescription is to recruit, to reduce the burden, or to compensate.

What the question adds. Claim 10.4 states the structure: a metered cycle closes by drawing on an unmetered one, the arrangement records improvement, and the record is correct. The difficulty is therefore not that the arrangement ignored a signal but that the signal is not among the ones it holds. The prescription that follows is to obtain the signal, which the ordinary description does not prescribe because it does not identify the deficit as a transfer between cycles.

The intervention. Maintain indirect indicators of the unmetered cycles, in the sense of §13: the proportion of contributors active in one period who are active in the next, the entry of new participants into positions requiring sustained work, the offering of criticism on unfinished work, and the acquisition of new maintainers by maintained resources. These establish whether a cycle is operating.

The limitation. By Claim 10.5 the indicators establish that a cycle has ceased to return and not the magnitude of the deficit or its attribution. An arrangement using them knows that something is failing and does not know how much of it its own mechanisms transferred there.

The condition of refutation. The interpretation is mistaken if withdrawal proves to be predicted adequately by the measures already held, which would show if declines in participation were preceded by declines in recorded output.

16.4 Weak Return Cycles

The difficulty. Contributors supply work over extended periods and the conditions permitting them to do so are not replenished from any source connected to the use made of what they supplied.

The ordinary description and its prescription. The difficulty is under-compensation of a public good, and the prescription is compensation, whether by payment, by subsidy or by a claim over downstream use.

What the question adds. Two things, and the first is conceded to the account this paper adopts. The criterion that value should return to the arrangements that produced it belongs to the account of generative justice (Eglash 2016) recorded in Claim 4.1, and it prescribes the return that the ordinary description prescribes as compensation, with a different rationale.

What this paper adds is the second thing. By Claim 7.5 a return in a single form relieves only those whose binding constraint that form addresses, so an arrangement returning money to a contributor whose constraint is an interlocutor has returned nothing that bears on the constraint. And by Claim 9.4 a return that closes the cycle within a fixed group satisfies the circulation criterion while worsening the arrangement on the boundary condition. The prescription therefore differs from both the ordinary description and the circulation criterion: return in plural forms, and assess the boundary separately from the closure.

The intervention. The five mechanisms of §22, selected by the constraint they address, together with the entry condition of Definition 9.2 applied to the population among whom return circulates.

The limitation. The diagnosis of a binding constraint is itself a causal claim, and §18 recorded that such claims are defeasible. An arrangement that misidentifies a constraint supplies a return that relieves nothing, and the error is detected only by the absence of an effect.

The condition of refutation. The interpretation is mistaken if plural returns prove to perform no better than a single return of equivalent cost, which is a comparison an arrangement operating both could conduct.

17. The Difficulties for Which the Question Supplies an Interpretation

This section examines three difficulties for which the second-order question yields an account that differs from the ordinary one while the prescriptions coincide in part. Its objective is to report an intermediate result honestly, since a framework that reported only its successes and its failures and nothing between them would be concealing the ordinary case. The three are treated on the template of §19, with an additional statement in each of how far the prescriptions diverge, since that is what distinguishes this section from the preceding one. In one of the three the divergence rests on an empirical assumption, which §8 identified as the failure the paper is most exposed to, and the assumption is stated where it operates.

17.1 Attention Concentration and Cumulative Advantage

The difficulty. Visibility yields attention, attention yields recognition, and recognition yields further visibility, so that a small number of works, authors and institutions absorb a share of attention out of proportion to any independent assessment of content.

The ordinary description and its prescription. The process is preferential attachment operating on a network whose growth is driven by adoption, and the outcome is predicted in advance by the mechanism. Where concentration is judged undesirable, the prescription is to alter the ranking so that it does not compound prior visibility, which is achieved by damping the feedback or by mixing in a component independent of it.

What the question adds. Attention here is a generative condition in the sense of Definition 6.1 and not only a consumed resource: what a work receives is not merely value delivered to its author but the occasion of the criticism, collaboration and reuse that constitute the cycles of §11. The concentration is accordingly read as an accumulation under closure in the sense of Claim 11.1, in which the cycle sustains every position and the returned conditions diverge, terminating in a configuration where the unaccumulated positions cannot act independently.

How far the prescriptions diverge. They coincide on the direction and differ on the criterion. The ordinary description recommends damping because concentration is inefficient or unfair, and its measure of success is the resulting distribution. The question recommends reserving capacity because the concentration forecloses encounters, and its measure of success is whether the foreclosed encounters occur. An arrangement acting on the first would be satisfied by a flatter distribution of attention across works that never meet the parties who would develop them; an arrangement acting on the second would not.

The intervention. A fraction of surfacing capacity, fixed and published in advance, directed to work that is underconnected, to complementary expertise, to unresolved questions, and to work abandoned in a state permitting resumption. The fraction must be fixed and published, since a reservation administered case by case is a determination about parties made by an interested party, which Definition 9.2 and the companion papers exclude.

The empirical assumption. The recommendation rests on an assumption that should be stated because the argument depends on it: that work capable of being developed is distributed more widely than attention currently reaches, so that reserved capacity encounters something. If the distribution of developable work follows the distribution of attention closely, the reservation surfaces material that yields nothing and the ordinary efficiency argument prevails. The paper does not establish the assumption and marks the recommendation as conditional upon it.

The limitation and the condition of refutation. The reserved fraction is itself a metric target once its criteria are known, and §20.2 applies. The interpretation is mistaken if reserved capacity produces no encounters that would otherwise have failed to occur, which an arrangement operating a reservation can determine by following the work surfaced under it.

17.2 Metric Distortion

The difficulty. Counts of citation, download, view or endorsement, adopted to represent contribution, become the objects participants pursue and cease to represent what they were adopted to represent.

The ordinary description and its prescription. A measure under pressure ceases to be a good measure, and the reactivity of measures in evaluative settings is documented at length (Espeland and Sauder 2007). The prescription is to use several measures, to change them, or to withhold the weight that makes them worth pursuing.

What the question adds. The ordinary description states that a measure decays. The stronger statement available here is that a measure of this kind builds. A ranking by accumulated citation does not merely stop reporting quality; it directs attention, and attention produces the citations it then reports, so the measure participates in producing the distribution it describes. The difficulty is therefore an instance of Claim 11.1 in which the accumulating position is constituted by the measure itself.

How far the prescriptions diverge. They coincide substantially. Both recommend that a measure not be used as a target, and both recommend plurality. The divergence is that the ordinary description treats decay as a reason to replace a measure periodically, while the account here treats a measure that both ranks and reports as defective in a way replacement does not correct, since its successor will do the same. The prescription that follows is the separation of the surfacing function from the reporting function, so that what determines encounters is not the quantity subsequently reported as evidence of merit.

The limitation and the condition of refutation. The separation is difficult to sustain in practice, since a reported quantity affects encounters through the judgements of participants even where it does not affect the ranking. The interpretation is mistaken if concentration of the reported quantity persists at the same rate under a surfacing rule independent of it, which is a comparison an arrangement can conduct.

17.3 Premature Publication and the Occupation of a Subject

The difficulty. Where volume, novelty and priority are rewarded, immature work is released early, and an early formulation may fix a subject’s terminology and reference points regardless of its quality.

The ordinary description and its prescription. The difficulty is one of quality assurance and information overload, and the prescription is review, filtering and the promotion of established work over new work.

What the question adds. A publication is not only an output but an alteration of the conditions under which subsequent work proceeds, since it supplies the vocabulary in which the subject is subsequently stated. The mechanism by which an early artifact occupies a subject, and by which terms restricting derivation then obstruct the work that would improve it, is developed in the companion paper on the ethics of transmission (Huang 2026g) and is not restated here. What the platform case adds to that treatment is the operator’s part in it: which early artifact occupies is determined substantially by which is surfaced, so occupation is a property of the surfacing apparatus and not only of sequence.

How far the prescriptions diverge. The ordinary prescription filters, and filtering suppresses immature work. The prescription here does not suppress it and alters its standing: a formulation may circulate while being marked as preliminary, exploratory, superseded or withdrawn, so that it is available to those who want it and does not acquire the standing of settled knowledge by remaining in place. The two prescriptions are close to opposite in their effect on what is published.

The intervention. Publication statuses attached to works and revisable over their life, with the status displayed wherever the work is surfaced, and with supersession recorded in place of the removal of what was superseded.

The limitation and the condition of refutation. A status is assigned by someone, and the assignment is a determination about a work made by a party with an interest in the subject. The interpretation is mistaken if statuses prove not to affect the standing accorded to work, which would show if preliminary formulations continued to be cited as settled at unchanged rates once marked.

18. The Difficulties for Which Existing Theory Suffices

This section reports the difficulties for which the second-order question fails the test of Definition 5.2, names the existing theory that should be used in their place, and assembles the verdicts of Part V. Its objective is to discharge the discipline the paper imposed on itself in §8, since a framework that reported no failures would have shown only that its test was not applied. The section treats two difficulties in which the failure is clear, states a general class of cases in which it is clear without individual treatment, presents the assembled verdicts in Table 3, and states what the pattern of verdicts indicates about where the question is useful. The failures reported are failures of this paper’s instrument and not of the difficulties.

18.1 Resource Mismatch

The difficulty. A promising line of work fails for want of one specific thing, which may be compute, a dataset, a collaborator, an interval of time or institutional standing, and which is frequently not money.

What the question would add. The question directs an arrangement to identify which condition is binding in place of supplying a general resource, since supplying a resource that is not the binding constraint relieves nothing. Claim 7.5 supports the same conclusion from the absence of a common unit.

Why the test fails. The prescription to identify a binding constraint before supplying anything is the established method of a developed literature on economic development (Hirschman 1958), in which the identification of the binding constraint and the sequencing of interventions upon it is the central technique. That literature prescribes exactly what the question prescribes here, on grounds that do not require any of this paper’s apparatus, and it supplies the diagnostic procedures that this paper does not.

What should be used. The binding-constraint method as developed in that literature, applied to the conditions of §9. The contribution of the present framework in this case is confined to the observation that the constraint may be of a kind an arrangement does not ordinarily consider, and that observation is a widening of the list and not a change of method.

18.2 Downstream Capture of Openly Available Work

The difficulty. Parties with greater compute, capital, legal capacity or market access obtain value from openly available work, lawfully and without obligation, and the arrangements producing the work receive nothing determinate in return.

What the question would add. The question directs attention past the distribution of the gain to the effect on the commons’ conditions, and holds that commercial use is not itself the difficulty, the difficulty being a use that leaves the producing arrangement less able to produce.

Why the test fails. The prescriptions coincide almost entirely. The account of generative justice adopted in Claim 4.1 already prescribes that value circulate back to the generators and already treats extraction without return as the pathology. The public-goods description prescribes the same measures under a different rationale, in the form of contribution to the maintenance of a resource one draws upon. And this paper’s own contribution to the return question, the plurality of forms and the boundary condition, is developed at §19.4 and applies to the relation between contributors and the arrangement, and not to the relation between the commons and a downstream firm, where the returning party is one and the constraint diagnosis is not conducted.

There is a residue, and it is small. The framework identifies one case the circulation criterion does not catch, namely a firm that returns value while the return is directed to the positions that already accumulate, which is Claim 11.1 applied to the return itself. The residue is real and does not carry the difficulty.

What should be used. The circulation criterion of the generative justice account, with the public-goods description for the analysis of provision, and with §12’s boundary condition applied to the population among whom returned value circulates.

18.3 The Cases Not Treated Individually

Three further classes of difficulty arise in these settings for which the question is not used, and they are named together because the reason is the same in each.

Where a party’s position permits it to raise prices, to exclude competitors or to condition access on unrelated terms, the difficulty is one of market power and the existing analysis is both more developed and more directly actionable than anything here. Where contributors decline to supply a good whose benefits they cannot capture, the difficulty is free-riding and the existing analysis applies. And where a fixed quantity of a divisible resource is to be divided among claimants whose needs are of one kind, the difficulty is an allocation problem and Claim 7.2 does not bear on it, since the object being divided is a resource and not the capacity at issue in this paper.

Claim 18.1The scope indicated by the failures The second-order question fails the test of Definition 5.2 where the difficulty concerns a divisible resource, a party’s market position, or a relation between two identified parties in which the returning party is known and the constraint is not in question. It passes where the difficulty concerns a condition the arrangement supplies without metering, a cycle whose deficit is transferred to another cycle, or a determination made by the arrangement about which participants encounter which work.

| @p3.6cmp2.2cmp4.6cmp1.5cm@

Difficulty	Verdict	Ground of the verdict	Section
Capture of relational infrastructure	Alters	The ordinary description prescribes nothing where competition is unaffected	§19.1
Contribution invisibility	Alters	Record the maintenance relation, in place of paying for a task	§19.2
Contributor exhaustion	Alters	Obtain a signal the arrangement does not hold	§19.3
Weak return cycles	Alters	Plural forms selected by constraint; boundary assessed separately	§19.4
Attention concentration	Interprets	Same direction, different criterion of success; assumption stated	§20.1
Cumulative advantage	Interprets	Treated with attention concentration	§20.1
Metric distortion	Interprets	A measure that surfaces and reports is defective beyond decay	§20.2
Premature publication	Interprets	Status in place of filtering; opposite effect on what circulates	§20.3
Resource mismatch	Suffices	The binding-constraint method prescribes the same	§21
Downstream capture	Suffices	The circulation criterion prescribes the same; residue small	§21.2

Table. The difficulties examined, the verdict on the second-order question for each, and the section in which the treatment appears.

18.4 The Indication Given by the Pattern of Verdicts

Claim 18.2The location of the question’s usefulness The difficulties for which the question alters a prescription share a property: in each, the operative condition is one the arrangement itself supplies and does not meter, and the party best placed to alter it is the arrangement. The difficulties for which existing theory suffices share the complementary property: the operative object is a resource, a legal position, or a transfer between identified parties, and existing analysis addresses each of these directly.

The claim is offered as the paper’s most useful practical result, since it tells an arrangement when to reach for this framework. Where a difficulty concerns something the arrangement transacts in, the ordinary descriptions are adequate and better developed. Where it concerns something the arrangement determines without recording, the ordinary descriptions have no object and this framework supplies one.

The pattern also indicates the framework’s boundary from the other side, and §24 records it as a limit. A framework useful exactly where an arrangement supplies conditions without metering them is a framework whose usefulness is a function of what the arrangement has chosen not to meter, and an arrangement that metered everything relevant would have no use for it. That is a coherent position and it is worth stating, since it locates the framework as an instrument for a particular institutional condition and not as a general economics.

VIThe Instruments

19. The Return Mechanisms and Their Requirements upon an Operator

This section sets out the mechanisms by which a return may reach the conditions of continued contribution, and orders them by what each requires of the party operating the arrangement. Its objective is to make the ordering itself informative, since the mechanisms are ordinarily presented as a list from which an operator selects on grounds of preference, and the ordering shows that the selection is made on other grounds. The section states the five mechanisms with the constraint each relieves; states what each requires; presents the ordering in Table 4; states the finding the ordering yields; and states the limits carried into any mechanism resting on a record of enabling relations. The mechanisms are described at the level of their form, and no claim is made about any particular arrangement’s adoption of them.

19.1 The Mechanisms

Attribution return. The record of a work preserves the relations that enabled it, including review, correction, curation, translation, replication and the drawing of a conceptual connection, so that a party who supplied one of these is identifiable as having done so. The constraint it relieves is standing: a contributor whose work is invisible cannot be sought out, cited, or credited by any party who would.

Attention return. The apparatus that surfaces work surfaces also the work and the parties that enabled it, so that a contribution which made later work possible is encountered by those who encounter the later work. The constraint relieved is encounter, which by §20.1 is the condition of the epistemic and relational cycles.

Opportunity return. A party whose contribution proved consequential becomes easier to find by those seeking collaborators, reviewers, appointees or advisers, which requires only that the arrangement make the enabling relation legible to parties who search. The constraint relieved is access to the positions from which sustained work becomes possible.

Resource return. A party benefiting substantially downstream supplies compute, funding, data, infrastructure or maintenance capacity to the arrangement or to its contributors. The constraint relieved is material.

Commons return. A return is directed to the shared apparatus in place of to any contributor: to moderation, storage, curation, replication, tooling and the maintenance of what all use. The constraint relieved is infrastructural, and it is the constraint no individual return relieves, since the apparatus belongs to no participant.

19.2 The Ordering by Requirement

Each mechanism requires something of the operator, and the requirements differ in kind and in cost.

Attribution return requires a schema that records the relevant relations and nothing further. The events are already occurring within the arrangement’s systems, so the cost is the cost of recording and of designing what to record.

Opportunity return requires that the record be exposed to parties who search, which is a matter of making an existing record legible. It requires no resource, no alteration of the surfacing apparatus, and no counterparty.

Attention return requires alteration of the surfacing apparatus, which is ordinarily the operator’s principal asset and the mechanism by which it obtains whatever it obtains. The requirement is therefore not of a resource but of the thing the operator is least willing to alter, for the reasons given in §19.1.

Resource return requires a counterparty willing to supply something, and the operator can create the pathway and cannot supply the return. Its adoption therefore depends on a party the operator does not control.

Commons return requires a body able to hold and to spend, together with a rule by which the spending is directed, which is an institutional construction and the most demanding of the five.

| @p2.6cmp3.0cmp4.4cmp1.9cm@

Mechanism	Constraint relieved	Requirement on the operator	Cost to operator
Attribution return	Standing	A schema recording enabling relations	Low
Opportunity return	Access to positions	Exposure of an existing record to search	Very low
Attention return	Encounter	Alteration of the surfacing apparatus	High
Resource return	Material	A pathway, and a counterparty willing to supply	Low, and contingent
Commons return	Infrastructural	A body able to hold and spend, and a rule directing it	High

Table. The return mechanisms, the constraint each relieves, and what each requires of the operator.

19.3 The Finding the Ordering Yields

Claim 19.1The disparity between cost and adoption The two mechanisms requiring least of an operator, attribution return and opportunity return, are the least widely implemented, while the mechanism most widely discussed, resource return, is the one whose adoption depends on a party the operator does not control. The pattern of adoption therefore does not track the cost of implementation.

The claim invites an explanation and the paper offers a partial one. Attribution and opportunity return alter who is visible and who is sought, which redistributes standing among participants without producing anything the operator can record as a gain. Resource return produces a transfer the operator can record and can present as the discharge of an obligation. An operator selecting among mechanisms by what it can demonstrate having done will select the one that yields a demonstrable transfer, and the mechanisms that cost it least yield none.

Claim 19.2The consequence for recommendation A recommendation that an arrangement adopt the mechanisms of least cost is not a recommendation about resources, since those mechanisms consume none. It is a recommendation that the arrangement accept a redistribution of standing among its participants that produces nothing it can report. The obstacle to the cheapest mechanisms is accordingly not expenditure, and an argument addressed to expenditure will not move it.

19.4 The Limits Carried by a Record of Enabling Relations

Three of the five mechanisms rest on a record of which parties enabled which work, and that record carries limits established in the companion paper on the governance of propagation (Huang 2026h), which apply here without amendment and one of which applies with greater force.

A record establishes that one work is reachable from another through a sequence of recorded relations. It does not establish that the earlier work caused the later, and §16 adds that even a settled causal relation would not fix a share, since a share is defined only relative to a description. A record supporting a return must therefore be read as establishing that an enabling relation obtained and not as establishing a proportion.

The schema of recordable relations is fixed by the arrangement, and by Claim 6.3 what it declines to record it cannot govern. The companion treatment establishes that the graver failure of such a schema is selection in place of deletion, since a relation the schema never recognised leaves no gap in the record.

Claim 19.3The aggravation of the schema problem where a return attaches Where a return attaches to a recorded relation, the schema determines who receives, and the party fixing the schema is ordinarily the party bearing or brokering the return. The defect the companion treatment identifies is therefore aggravated in the economic case, since the interest of the party fixing the schema now runs directly to the outcome the schema determines.

Two responses follow and neither dissolves the difficulty. The schema should be published together with the classes of relation considered and not adopted, which is the condition the companion treatment states and which no existing infrastructure supplies. And the return attaching to a class of relation should be stated in advance and uniformly, so that the discretion remaining to the operator is confined to the schema and is not exercised case by case. §23 states that condition in its general form.

20. Conditions on an Admissible Platform Instrument

This section states the form an instrument must take to be admissible within the framework, and applies the conditions to the instruments the paper has recommended. Its objective is to prevent the framework from licensing, at the level of an arrangement, the determinations about participants that Part III’s boundary condition excludes at the level of a cycle, and to supply a test an operator can apply to a proposal before adopting it. The section states why conditions on form are required at all; states the four conditions with the failure each prevents; applies them to the paper’s own recommendations and reports the results; and states the exposure that remains after the conditions are met. The conditions are adapted from those developed in the companion papers and are stated here in the form the economic case requires.

20.1 The Need for Conditions on Form

An instrument that improves a cycle may do so by any means, and the means available to an arrangement include determining which participants receive what. An arrangement that reserved surfacing capacity for work it judged promising, that directed returns to contributors it judged deserving, or that admitted to a commons fund the parties it judged serious would satisfy every closure condition in Part III while making, in each case, a determination about a participant on evidence the arrangement holds privately and by a criterion the participant cannot contest.

The difficulty is the one Part III identified at the boundary of a cycle, appearing now at the level of the instrument. Definition 9.2 required that entry to a cycle be on terms stated in advance and applying to all who meet them. An instrument administered case by case reinstates precisely the determination that requirement excludes, and does so in a form that is harder to identify, since an instrument presents as a mechanism and not as a judgement about anyone.

20.2 The Conditions

Definition 20.1Admissible form A platform instrument is in admissible form where four conditions hold. It binds the arrangement’s own process and does not condition a participant’s entitlement on an assessment of that participant. Its terms are stated in advance of the acts they govern. Its terms apply uniformly to all participants meeting them, with no distinction by the identity, affiliation or standing of the participant. And its operation terminates in an outcome the participant can observe.

Each condition prevents a failure that has occurred in practice.

The first prevents the reinstatement of a readiness determination under the appearance of a mechanism, which the companion paper on the ethics of transmission (Huang 2026g) establishes to be administered by the party a refusal relieves and insulated from the objection of the party it concerns.

The second prevents an instrument from being specified after the conduct it governs, which permits the specification to be fitted to the outcome desired.

The third prevents a distinction among participants from being made as a technical matter. An instrument that surfaces work from institutions of one class, or that directs returns to parties holding a particular affiliation, has made a determination about persons under the description of a rule.

The fourth prevents an instrument from operating without any moment at which a participant learns its result, which is what permits an instrument to be adjusted continuously and assessed by nobody.

Claim 20.2The consequence of inadmissible form An instrument failing any of the four conditions may improve the cycle it addresses and does so by a determination the participants affected cannot examine. Such an instrument reinstates, at the level of the arrangement, the closure of the boundary that Claim 9.1 identified as the pathology of a cycle, and it is the harder to identify for presenting as a mechanism.

20.3 The Paper’s Own Recommendations Assessed

The conditions are applied here to what the paper has recommended, and two of the recommendations require qualification.

The recording of maintenance relations satisfies all four. It binds the arrangement’s own process, its schema may be stated in advance, it applies to all relations of the recorded classes, and each recording terminates.

The reservation of surfacing capacity satisfies the conditions where the reserved fraction and the criteria of selection into it are fixed and published in advance, and fails all four where the reservation is administered by judgement. The paper’s recommendation at §20.1 was stated in the admissible form for this reason, and the difference between the two versions is the whole of the difference between an instrument and a discretion.

Portability, transparency of the surfacing rule, and an interface permitting alternative surfacing satisfy the conditions, since each binds the arrangement and none conditions any participant’s entitlement on an assessment.

Publication statuses require qualification. A status is assigned to a work by a party, and the assignment is a judgement. The condition is satisfied where the criteria for each status are published, where the assignment is made by a stated procedure, and where a party whose work is assigned a status may contest the assignment before a party other than the one who made it. Where those are absent, the instrument distributes standing by judgement and falls under Claim 20.2.

Commons return requires qualification of a different kind. A body holding and spending must decide what to spend on, and the decision is a determination about the recipients. The condition is satisfied by publishing the criteria and the decisions, including the applications declined, which is the discipline the companion papers require of any arrangement making determinations it cannot avoid making.

20.4 The Exposure That Remains

Claim 20.3The residual discretion in the schema An arrangement satisfying Definition 20.1 in every instrument retains one discretion the conditions do not reach: the determination of what its schema records, which by Claim 6.3 determines what it can govern and by Claim 19.3 determines who receives a return. That determination is made in advance, applies uniformly and terminates, so it satisfies the conditions in form, and it is nonetheless a determination about participants, since a class of contribution left unrecorded is a class of contributor left unreachable by every instrument.

The claim states the limit of what conditions on form can achieve, and the paper offers no instrument that reaches it. What is available is the discipline recorded at §22: publish the schema together with the classes considered and not adopted, and revise it on stated occasions and on proposals from those subject to it. That discipline makes the discretion visible and leaves it where it was.

A second exposure follows from the whole of Part VI and is recorded here. Every instrument this paper recommends is one an arrangement adopts voluntarily, and the arrangement adopting it is the party whose position the instruments constrain. Nothing here compels an arrangement to bind itself, and the question of what would induce it to do so lies outside the argument of this paper, as it lay outside the arguments of its two companions.

VIIThe Position

21. The Limits of the Framework

This section states what the framework establishes, what it leaves undetermined, where it is weakest, and under what conditions it would be shown to be mistaken. Its objective is to fix the paper’s claims at the strength its arguments support and to assemble in one place the exposures that individual sections recorded as they arose. The section states the established results; states the matters left undetermined; states seven weaknesses, each argued for elsewhere in the paper and none of them answered; and states the conditions of defeat. The weaknesses are stated by the paper in place of being left for a reader to locate.

21.1 The Established Results

Six results are established. The reported difficulties of open knowledge platforms reduce to three mechanisms, distinguished by whether a cycle of contribution is unrepresented, unreturned, or returned with accumulation at one position. The second-order question is stated with a boundary on what counts as a condition, which turns on what an arrangement has made itself answerable for, and with a test restricting its own use to cases in which it alters a prescription. The capacity at issue admits no common unit and is therefore not distributable, from which it follows that its maximisation is undefined, that plural returns are a consequence and not a preference, and that the framework’s objectives are statable only as the avoidance of foreclosure. The closed cycle is adopted as the unit on the ground that pointwise quantities vary with the description and circuit comparisons do not, and its assessment requires an order on each kind of condition and nothing further. Closure of a circuit and openness of a boundary are independent, so that circulation is necessary and insufficient. And the difficulties are examined individually against the test, with four found to alter a prescription, four to supply an interpretation, and two to be adequately handled by existing theory.

Two subsidiary results carry beyond the paper’s subject. A cycle that closes while capacity accumulates at one position passes every assessment by closure, and its terminal state is a dependence that neither of the paper’s conditions excludes. And an arrangement observing improvement on the cycles it meters, while a coupled cycle it does not meter depletes, will observe correctly until the unmetered cycle fails altogether, which accounts for the suddenness of the failures reported.

21.2 The Matters Left Undetermined

The framework fixes no measure of generative capacity and shows that none is available. It fixes no exchange between deficits of different kinds, so an arrangement facing two deficits holds two findings and no net finding, and any judgement between them is a judgement. It determines no reliability with which a cycle should close, since Claim 11.2 established that reliable closure reproduces the conditions it closes over. It supplies no procedure for diagnosing a binding constraint, and §21 directs the reader to a literature that does. It takes no position on the question raised in §17. And it makes no claim about the frequency with which any difficulty occurs in any actual arrangement, having assessed structures and prescriptions in place of cases.

21.3 The Weaknesses

Seven weaknesses are stated by the paper, and each was argued for in the section that produced it.

The framework’s usefulness is a function of an arrangement’s own omissions. Claim 18.2 established that the question alters a prescription where the operative condition is one the arrangement supplies without metering, from which it follows that an arrangement metering everything relevant would have no use for the framework. That is a coherent position and it makes the framework an instrument for an institutional condition and not a general economics.

The diagnosis of a binding constraint is a causal claim of the kind the paper elsewhere shows to be unavailable. §18 answered this in part, by observing that a present constraint is observed where a future outcome is predicted, and the answer is partial: which conditions bind is a claim about generation, and the framework offers only that a mistaken diagnosis is corrected by the absence of an effect.

The instruments are self-imposed. Every recommendation in Part VI is adopted voluntarily by the party it constrains, and nothing in the paper compels adoption. What would move an arrangement to bind itself lies outside the argument.

The residual discretion over the schema is unreached by the conditions on form. Claim 20.3 established that an arrangement satisfying every condition in Definition 20.1 still determines what is recorded, that the determination passes the conditions in form, and that a class of contribution left unrecorded is a class of contributor left unreachable by every instrument the paper offers.

The recommendation at §20.1 rests on an empirical assumption the paper does not establish, namely that work capable of development is distributed more widely than attention currently reaches. The recommendation is marked as conditional, and a reader who doubts the assumption should treat that recommendation as withdrawn.

The deficit of a cycle is sensitive to the boundaries of a traverse, which are chosen. Definition 12.1 recorded that a traverse is an interval whose endpoints are selected, and a deficit measured over a short interval may reverse over a longer one. The framework supplies no rule for selecting the interval.

And the classification of §6 is a hypothesis about a set of reported difficulties, assembled from reports and not from a survey. A difficulty absent from the reports is absent from the classification, and the classification’s coverage is therefore the coverage of the literature it was drawn from.

21.4 The Conditions of Defeat

Five conditions would defeat the framework, and they are ordered from those bearing on its central claim to those bearing on its scope.

The framework is defeated in its central claim if a common unit across kinds of condition is exhibited, since Claim 7.1 would fail and with it the non-distributability result, the derivation of plural returns, and the undefinedness of maximisation. A demonstration that money converts requisites at rates independent of the participant and the line of work would suffice.

The framework is defeated in its unit if an object is exhibited that spans time, is assessable without a measure of capacity, does not vary with the analyst’s boundary, and is not a cycle. The paper knows of no such object and offers that as a reason and not a proof.

The framework is defeated in its classification if a difficulty belongs determinately to two mechanisms, if an intervention relieves difficulties across mechanisms uniformly, or if the difficulties of a mechanism fail to respond to the instrument assigned to them.

The framework is narrowed if the test of Definition 5.2 is found to be passed by fewer difficulties than §19 claims, and in particular if the prescriptions in any of the four cases are found to coincide with what the ordinary description already recommends. That is checkable case by case and the paper invites the check.

And the framework is refuted in its practical claim if arrangements adopting the instruments of Part VI are found to exhibit the reported difficulties at unchanged rates, which is the comparison an arrangement operating them can conduct and which the paper has not conducted.

22. Open Questions

This closing section states the questions the enquiry leaves open and the position the paper reaches. Its objective is to identify what would have to be established for the framework to be usable in an arrangement, and to state plainly what it offers and what it does not. The questions are ordered by tractability and each is stated with what would answer it. The section closes with the position.

The enquiry leaves the following questions open.

• Whether the disparity recorded in Claim 19.1 holds generally. The claim that the cheapest return mechanisms are the least implemented is drawn from the arrangements the author is familiar with, and it is answerable by a survey of what a set of arrangements records and exposes.
• What indirect indicators of an unmetered cycle are informative in practice. §19.3 proposed four, none of which measures the condition at issue, and whether any of them anticipates withdrawal is answerable from the records arrangements already hold.
• Whether reserved surfacing capacity produces encounters that would otherwise have failed to occur. This is the assumption on which §20.1 conditions its recommendation, and an arrangement operating a reservation can follow the work surfaced under it and determine the answer.
• How the boundaries of a traverse should be selected. The deficit of Definition 12.2 depends on the interval chosen, and whether a principled selection is available, or whether the framework must report deficits at several intervals, is unresolved.
• Whether the accumulation of Claim 11.1 is detectable early. Its detection requires comparing the distribution of returned conditions across traverses, and how many traverses are required before a divergence is distinguishable from variation is a statistical question with an answer.
• What arrangements have supplied a schema revised by a party other than the one whose conduct it records. Claim 19.3 identified the schema as the point where the interest of the fixing party runs most directly to the outcome, and the governance histories of existing provenance and standards bodies bear on what has been achieved there.
• Whether a condition recognised only in retrospect can be provided for prospectively. Claim 15.5 established that an arrangement requiring justification in advance under-supplies such conditions, and whether any device other than an unjustified allowance addresses this is unresolved.
• Whether the two conditions of §12 are jointly sufficient for anything. Claim 11.2 established that a cycle satisfying both may still foreclose by reproducing the conditions it closes over, and the paper offers no third condition.
• What would induce an arrangement to adopt instruments that constrain it. This lies outside the argument of the paper, as it lay outside the arguments of its two companions, and it is the question on which the practical value of all three depends.
• Whether the question raised in §17 admits an answer bearing on economic arrangements. The paper records that a version of it has been treated formally elsewhere and takes no position, and whether anything transfers is a question for a treatment more careful than this paper attempts.

The position the paper reaches may be stated briefly. An open knowledge platform can succeed completely at the thing it was built for and still fail, and the failures reported of such arrangements are described adequately, one at a time, by economic accounts that this paper disputes in no respect. What those accounts do not supply is an object with respect to which the failures occur together, since each takes as its object a quantity realised in a period and the failures concern what the participants are left able to do.

The question this paper adds is narrow: what a present arrangement leaves the participants able to produce next. It has a boundary, since a condition in the relevant sense is one the arrangement has made itself answerable for, and it has a discipline, since it is worth asking only where it changes what an arrangement should do. Applied under that discipline it alters four prescriptions, supplies an interpretation in four further cases, and adds nothing in two, which are reported.

What follows from taking it seriously is chiefly negative and is the paper’s most secure content. The capacity at issue has no unit, so it cannot be distributed, cannot be maximised, and cannot ground an objective stated positively. No position outside an economic description is available from which its categories could be fixed, so quantities defined within a description carry the description with them, and an assessment that is to say anything must be conducted in terms of what survives redescription. The closed cycle is the object that does, and the assessment it supports is correspondingly modest: whether a participant returns to a position from which they may continue, and whether a party who was not within the cycle may enter it on terms they can satisfy.

Neither of those is a measure of anything, and together they are what the paper offers in place of one. An arrangement holding them can determine that a cycle is depleting without determining how much has been lost, that a coupled cycle has ceased to return without determining what its own mechanisms transferred there, and that a boundary has closed without determining what was foreclosed by the closing. Each of these is less than an account of the economy of a knowledge commons would supply. Each of them is available, and the quantities that would supply more are not.

References

tocsectionReferences

Aghion, P., and P. Howitt. 1992. “A Model of Growth Through Creative Destruction.” Econometrica 60 (2): 323–351.

Benkler, Y. 2006. The Wealth of Networks: How Social Production Transforms Markets and Freedom. New Haven: Yale University Press.

Campbell, D. T. 1979. “Assessing the Impact of Planned Social Change.” Evaluation and Program Planning 2 (1): 67–90.

Eglash, R. 2016. “An Introduction to Generative Justice.” Teknokultura 13 (2): 369–404. doi:10.5209/rev_TEKN.2016.v13.n2.52847.

Eglash, R., K. P. Robinson, A. Bennett, L. Robert, and M. Garvin. 2024. “Computational Reparations as Generative Justice: Decolonial Transitions to Unalienated Circular Value Flow.” Big Data & Society 11 (1). doi:10.1177/20539517231221732.

Espeland, W. N., and M. Sauder. 2007. “Rankings and Reactivity: How Public Measures Recreate Social Worlds.” American Journal of Sociology 113 (1): 1–40. doi:10.1086/517897.

Folbre, N. 2001. The Invisible Heart: Economics and Family Values. New York: New Press.

Fraser, N. 2016. “Contradictions of Capital and Care.” New Left Review 100: 99–117.

Goodhart, C. A. E. 1975. “Problems of Monetary Management: The U.K. Experience.” In Papers in Monetary Economics, vol. 1. Sydney: Reserve Bank of Australia. Reprinted in Monetary Theory and Practice: The U.K. Experience, 91–121. London: Macmillan, 1984.

Hess, C., and E. Ostrom, eds. 2007. Understanding Knowledge as a Commons: From Theory to Practice. Cambridge, MA: MIT Press.

Hirschman, A. O. 1958. The Strategy of Economic Development. New Haven: Yale University Press.

Huang, W. 2026a. Governance After Allocation: Interpretive Compression and the Persistence of Relational Regulation. Knowledge Commons. doi:10.17613/0gwj9-hkk29.

Huang, W. 2026b. Loop Observables and the Gauge Structure of Interpretation: A Preliminary Discussion of Evaluation without an External Frame. Knowledge Commons. doi:10.17613/fgyz5-bm243.

Huang, W. 2026c. The Temporal Correspondence Limitation of Evidence-Based Policy Making: A Preliminary Discussion Towards an Astrophysics-Inspired State-Aware Epistemology of Governance. Knowledge Commons. doi:10.17613/jax3a-cex91.

Huang, W. 2026d. From Marriage Economics to Multi-Nomoi Governance: A Dialectical Generative Relational Economics Perspective. Knowledge Commons. doi:10.17613/a7gch-g3466.

Huang, W. 2026e. Jurisprudence and Justice in Generative Relational Systems: A Preliminary Position Paper on the Generation of Subjects, Causation, and Responsibility. Knowledge Commons. doi:10.17613/q1m76-y3r27.

Huang, W. 2026f. Conversion Injustice: A Generative Relational Account of Value Conversion. Knowledge Commons. doi:10.17613/7xjaz-tyy26.

Huang, W. 2026g. Rethinking Fa Bu Qing Chuan (“The Dharma Should Not Be Transmitted Lightly”): Knowledge Propagation Ethics and Implications for Open Knowledge Platforms. Working draft.

Huang, W. 2026h. Generativity Governance of Open Knowledge Propagation: A Preliminary Discussion of Events, Relations, and Conditions under Nonlinear and Multi-Entity Causation. Working draft.

Karni, E., and M.-L. Vierø. 2013. “‘Reverse Bayesianism’: A Choice-Based Theory of Growing Awareness.” American Economic Review 103 (7): 2790–2810. doi:10.1257/aer.103.7.2790.

Kauffman, S. A. 2000. Investigations. New York: Oxford University Press.

Longo, G., M. Montévil, and S. Kauffman. 2012. “No Entailing Laws, but Enablement in the Evolution of the Biosphere.” In Proceedings of the 14th Annual Conference Companion on Genetic and Evolutionary Computation, 1379–1392. New York: ACM. doi:10.1145/2330784.2330946.

Lucas, R. E. 1988. “On the Mechanics of Economic Development.” Journal of Monetary Economics 22 (1): 3–42.

Marshall, A. 1890. Principles of Economics. London: Macmillan.

Merton, R. K. 1968. “The Matthew Effect in Science.” Science 159 (3810): 56–63.

Nussbaum, M. C. 2011. Creating Capabilities: The Human Development Approach. Cambridge, MA: Harvard University Press.

Ostrom, E. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge: Cambridge University Press.

Polanyi, K. 1957. “The Economy as Instituted Process.” In Trade and Market in the Early Empires, edited by K. Polanyi, C. M. Arensberg, and H. W. Pearson, 243–270. Glencoe, IL: Free Press.

Popper, K. R. 1957. The Poverty of Historicism. London: Routledge and Kegan Paul.

Price, D. J. de S. 1976. “A General Theory of Bibliometric and Other Cumulative Advantage Processes.” Journal of the American Society for Information Science 27 (5): 292–306.

Robbins, L. 1932. An Essay on the Nature and Significance of Economic Science. London: Macmillan.

Rochet, J.-C., and J. Tirole. 2003. “Platform Competition in Two-Sided Markets.” Journal of the European Economic Association 1 (4): 990–1029. doi:10.1162/154247603322493212.

Romer, P. M. 1990. “Endogenous Technological Change.” Journal of Political Economy 98 (5, part 2): S71–S102.

Sen, A. 1985. Commodities and Capabilities. Amsterdam: North-Holland.

Sen, A. 1999. Development as Freedom. New York: Alfred A. Knopf.

Simon, H. A. 1971. “Designing Organizations for an Information-Rich World.” In Computers, Communications, and the Public Interest, edited by M. Greenberger, 37–72. Baltimore: Johns Hopkins Press.

Star, S. L. 1999. “The Ethnography of Infrastructure.” American Behavioral Scientist 43 (3): 377–391. doi:10.1177/00027649921955326.

Zittrain, J. L. 2006. “The Generative Internet.” Harvard Law Review 119 (7): 1974–2040.

Zittrain, J. L. 2008. The Future of the Internet and How to Stop It. New Haven: Yale University Press.

Zuboff, S. 2019. The Age of Surveillance Capitalism. New York: PublicAffairs.