By W.H.L., GPT-5.6 Sol, Claude Sonnet 5

Gradual AGI Series #9

Publication Version v1.0 · August 27, 2026

Abstract

Participation around advanced AI is often described by presence, channel, or volume: who attended, how many comments were filed, or whether a consultation occurred. Those measures can collapse acts that differ in what realization they address, whether meaningful alternatives remain, and whether the act’s evidentiary content depends on who performs it. This paper develops an act-level framework for human participation in Gradual AGI by separating Participatory Subject, Participatory Act, and Participatory Consequence.

Its principal analytic contribution is a Relational Grammar of the Participatory Act: Realization Orientation, Foreclosure Position, and Actor Dependence. Joint coding of these relations permits controlled comparison of acts that share the same visible form, channel, or count while differing in their relation to realization, decision state, or subject. Conditional Non-Fungibility follows as the comparison rule. The framework also distinguishes readiness from action, treats power and inequality as conditioning the passage from Subject to Act to Consequence, and defines a Responsive Floor only where features specified before outcome inspection are supported as contributing to realization, including through non-veto pathways. The empirical companion tests observability, coding reliability, nonredundancy, and realization relevance; null and construct-failure results remain admissible.

Reader’s Map

At a glance: the architecture separates who can act, what an act structurally is, and what follows. The Act layer carries the main analytic burden.

Element	At a glance
Architecture	Participatory Subject → Participatory Act → Participatory Consequence
Act grammar	Realization Orientation · Foreclosure Position · Actor Dependence
Proposition A	Same form, channel, or count does not establish equivalence when a material grammar relation differs
Proposition B	Automation of a task function does not by itself determine elimination of human Participation
Proposition C	Pre-specified features of Participatory Configuration may contribute to realized Floors; null dependence is permitted

Three standing principles govern the framework: Human-Centered Scope, Analytic Separation, and Participatory Contingency. Definitions and full epistemic-status qualifications appear in the numbered sections.

1. Introduction

In late August 2026, three contemporaneous developments exposed different sides of the same temporal problem. MIT concluded that generative AI had become capable enough to require faster and continuously revisable educational processes, describing the moment as a watershed for higher education and warning that adaptation was not optional. Bill Gates argued that social preparation for AI-driven disruption could not wait until its effects were fully visible. OpenAI leadership, meanwhile, said that the company had not yet reached its own AGI threshold but expected an internal system by the end of the year that it would call AGI; OpenAI’s Charter defines that threshold as highly autonomous systems that outperform humans at most economically valuable work.

Taken together, these claims establish neither an AGI milestone nor a theory of Participation. MIT’s report concerns one institution and one domain. Gates’ argument is normative and anticipatory. OpenAI’s forecast is a company judgment under its own definition, not an independently established scientific threshold. Their value here is narrower. They make visible a timing problem: capability, institutional adaptation, and human readiness can move on different clocks.

The examples are therefore suggestive framing rather than evidentiary support for the framework: they motivate the temporal question but do not establish any proposition developed below.

That problem has been present throughout the Gradual AGI series. The Synchronization framework distinguishes frontier capability from transformative adoption and describes human institutions as adapting alongside machine capability. Ceiling, Floor and Slope separates what frontier systems can do from what societies have actually absorbed and realized. The Optimization framework places humans, AI, institutions, infrastructure, and the wider social environment inside one evolving Optimization Subject. The Governance pair then asks how the mapping from potential to realization can be observed, shaped, challenged, and revised.

What those papers do not yet provide is a participation-specific account of the human acts inside that mapping. If a society is not merely exposed to advanced AI but continually involved in deciding how it is adopted, limited, redirected, resisted, normalized, or institutionalized, then the unit of analysis cannot be only the model, the institution, or the realized outcome. A theory is needed for the human subjects who can act, the acts they undertake, and the consequences that do or do not follow.

1.1 The participation problem

Participation is easy to invoke and difficult to specify. In ordinary use, the term can refer to being present, being affected, supplying data, expressing an opinion, taking part in a formal hearing, voting, organizing, negotiating, resisting, governing, or simply contributing to an outcome. Those meanings are not interchangeable. A population may be affected without acting. A person may have standing to speak without being ready to do so. A participant may act without influence. An influential act may change a procedure without changing the final realization. Human material may contribute to AI development without having been intentionally contributed at all.

The conceptual problem is therefore not whether humans remain somewhere in an AI-mediated society. Human presence alone says little about Participation. The problem is how human subjects retain, acquire, transform, and exercise capacities to act with respect to particular AI realizations—and under what conditions those acts become consequential.

This paper adopts a three-level analytic architecture for that problem: Participatory Subject → Participatory Act → Participatory Consequence. The organization is not claimed as a novel tripartite taxonomy. Its purpose is disciplinary. It prevents the status of the Subject, the structure of the Act, and the Consequence of the act from being inferred from one another.

1.2 Research questions

The framework is organized around three questions.

Research question	Operational concern
Who participates?	Which human persons or human-organized collectivities occupy the participant position, and what does it mean for them to be ready to participate?
What constitutes a Participatory Act?	Which intentional actions count as Participation, and what relational properties distinguish one act from another even when their visible form or institutional channel is the same?
When does Participation become consequential?	What changes, persists, or fails to change after participation, and under what conditions can pre-specified features of Participatory Configuration become realization-relevant?

The questions are sequential but not linear over time. Consequences can alter later subject readiness or institutional return paths, making later participation easier, harder, or different. The architecture therefore repeats recursively even though its analytic levels remain distinct.

1.3 The gap in the existing series

The gap is narrower than a missing theory of human-AI coexistence. The series already treats humans and AI as jointly evolving. It already distinguishes capability from realization. It already allows actor configurations, exposure, and governance to shape realization. It already grants affected parties standing to participate in governance and explicitly states that Participation is not influence.

Nor is the gap a missing taxonomy of civic acts. Political-participation scholarship has classified forms of participation for decades, and public-administration research has developed multidimensional frameworks for participatory institutions. Public-participation law already recognizes the importance of acting while options remain open; representation and epistemic-justice scholarship already explains why the experience or identity of a speaker can matter to the knowledge carried by an act.

The unresolved problem is act-level and realization-specific. Existing ingredients do not yet provide a compact grammar for distinguishing human Participatory Acts according to how they relate to the realization, where they occur relative to substantive foreclosure, and how much of their evidentiary or experiential content depends on the actor. That is the principal conceptual gap #9 addresses.

1.4 Contributions

Contribution 1 — Relational Grammar of the Participatory Act. The paper proposes a common act-level comparison unit that locates an act simultaneously in relation to the realization it addresses, the decision state it enters, and the subject whose content it carries.

Contribution 2 — Conditional Non-Fungibility. The paper states the analytic rule that equality of visible form, institutional channel, or numerical volume is insufficient to establish equivalence between acts when those relational properties differ materially.

Contribution 3 — Subject–Act–Consequence separation. The paper integrates Participatory Subject, Participatory Readiness, Participatory Act, Participatory Configuration, and Participatory Consequence in one realization-specific framework while preserving the distinctions among standing, capacity, action, influence, and outcome.

Contribution 4 — Participation-sensitive realization. The paper extends the series’ actor-sensitive realization apparatus by permitting one or more pre-specified features of human Participatory Configuration to contribute to the realized Floor even when participating subjects lack formal veto or binding authority.

The Relational Grammar matters for more than the fact that three familiar ideas are placed together. The three relations attach the same act to three different objects: realization, decision state, and subject. Joint coding therefore permits controlled substitutions while visible form, channel, or count are held fixed: change the act’s orientation toward realization, move the same act across an open or foreclosed decision state, or substitute the actor and ask whether relevant content is lost. The grammar’s value lies in making those comparisons within one unit of analysis.

The constituent relations have clear antecedents, so the paper does not claim priority over them separately. Its analytic claim is that their joint representation preserves distinctions that capability-to-realization research would otherwise scatter across separate literatures or compress into act labels and counts. Whether the representation is genuinely nonredundant is an empirical burden for #10, not something established by absence of a prior identical framework.

1.5 Scope boundaries

The framework does not equate affectedness, readiness, Participation, influence, authority, or beneficial outcome; it assumes no monotone relation between participation volume and consequence; and it does not require a Responsive Floor in every domain. Human-centered Participation is a methodological scope choice for the unresolved human relation to realization, not a claim that artificial systems could never qualify as participants under a broader future theory.

1.6 Epistemic status

The paper is a conceptual framework with an explicit empirical handoff. Definitions 1–7 specify objects and boundaries. Proposition A is analytic: it states a comparison rule that follows from treating the Relational Grammar as substantively meaningful. Proposition B is derived and antecedent-rich: it rejects a necessary inference from task automation to elimination of human Participation. Proposition C is a series extension: it permits one or more pre-specified features of Participatory Configuration to contribute to realization without assuming that such dependence is universal, nonzero, monotone, or large.

The empirical companion is therefore not tasked with confirming a foregone conclusion. It must test whether the grammar can be coded, whether its dimensions add nonredundant information, whether readiness can be observed independently of action, and whether participation adds realization-relevant information beyond capability, exposure, formal authority, infrastructure, and other competing conditions. Null results and construct failure remain admissible.

This sequencing is intentional. The conceptual framework was developed before the confirmatory pilot and is meant to constrain that pilot rather than be reconstructed from its results. Where later evidence forces revision, chronology should remain visible: a revised framework can be tested again, but the disagreement that triggered the revision should not be erased by retroactive recoding.

Series numbering convention. The Gradual AGI series numbers substantive theoretical and formal/empirical installments rather than every publication associated with the research program. The intervening Graduality feature-catalog papers are series reference works and therefore do not advance the installment count. This paper is therefore Gradual AGI Series #9, and its formal/empirical companion is Series #10.

1.7 Structure of the paper

Section 2 records the inherited foundations and external antecedents and fixes the paper’s non-claim ledger. Section 3 states the three governing principles and the Subject–Act–Consequence architecture. Section 4 defines the Participatory Subject and Participatory Readiness. Section 5, the conceptual center of the paper, defines the Participatory Act, develops the Relational Grammar, and states Conditional Non-Fungibility. Section 6 defines Participatory Configuration and Participatory Consequence and distinguishes consequence classes.

Section 7 develops the dynamic and recursive implications, including role plasticity, subject-readiness feedback, and institutional return paths. Section 8 reconnects Participation to the inherited realization apparatus through Responsive Floor and the participation-sensitive realization proposition. Section 9 states limits, failure conditions, and revision rules. Section 10 hands the constructs and their negative-result conditions to the formal and empirical companion.

The paper’s core claim is therefore specific: human participation in Gradual AGI cannot be represented adequately by presence or count alone. It requires an act-level account of how subjects act within decision trajectories and of what, if anything, follows.

References cited in this section

Bill Gates. 2026. “The turbulent AI era is here. The choices we make now are critical.” Gates Notes, August 26, 2026.

Massachusetts Institute of Technology. 2026. “AI and education: A watershed moment for MIT.” Letter from President Sally Kornbluth and accompanying Ad Hoc Committee report, August 25, 2026.

OpenAI. “OpenAI Charter.” Definition of AGI as highly autonomous systems that outperform humans at most economically valuable work.

Heath, Alex. 2026. “Inside OpenAI’s Reboot.” TIME, August 26, 2026.

Aristotle. Metaphysics, Book IX (Theta), on potentiality and actuality/activity.

Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press. Revised edition, 1975.

Fung, Archon. 2006. “Varieties of Participation in Complex Governance.” Public Administration Review 66(s1): 66–75.

Hirschman, Albert O. 1970. Exit, Voice, and Loyalty. Cambridge, MA: Harvard University Press.

Mansbridge, Jane. 1999. “Should Blacks Represent Blacks and Women Represent Women? A Contingent ‘Yes.’” The Journal of Politics 61(3): 628–657.

Verba, Sidney, and Norman H. Nie. 1972. Participation in America: Political Democracy and Social Equality. New York: Harper & Row.

Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

W.H.L. et al. 2026. Gradual AGI as Synchronization for Transformative Adoption. Gradual AGI Series #3.

W.H.L. et al. 2026. Ceiling, Floor and Slope: A Falsifiable Dynamical Model of Synchronization for Gradual AGI.

W.H.L. et al. 2026. Gradual AGI as Optimization: A Conceptual Framework. Gradual AGI Series #5.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

2. Inherited Foundations and Antecedents

Participation enters the Gradual AGI series after several apparent premises are already established elsewhere. The principal risk is therefore provenance error: inherited or external distinctions can acquire the appearance of novelty when redescribed in new vocabulary. This section fixes those debts before the framework is introduced.

Three labels are used: inherited series result, precursor commitment, and external antecedent. They constrain originality claims without preventing #9 from specializing or recombining what it inherits.

2.1 Pre-series foundation — AGI-inclusive humanity

The human-centered orientation predates the numbered series. First Principles of AGI-Inclusive Humanity (June 2025) and Philosophical Framework for AGI-Inclusive Humanity (July 2025) frame advanced AI within a continuing human civilizational project and explicitly foreground human dignity, agency, governance, participation, and inclusion. They supply philosophical scope, not a formal Participation apparatus.

The genealogy used here is therefore precursor conception → series adoption → participation-specific specialization.

2.2 Synchronization — reciprocal adaptation and the Ceiling–Floor problem

Synchronization already treats machine capability and bio-social humanity as reciprocally adapting, with human adaptation occurring through institutional learning, education, governance, professional practice, culture, and collective experience. It also distinguishes deployment from transformative adoption.

Ceiling, Floor and Slope gives that distinction dynamical form. Frontier capability does not uniquely determine societal realization; generic human-AI co-development and the Ceiling–Floor separation are therefore inherited rather than introduced by Participation.

2.3 Optimization — formal adoption of AGI-inclusive humanity

Optimization formally adopts AGI-inclusive humanity as its Optimization Subject: the co-developing system of humans, AI, institutions, infrastructure, and the broader social environment. It also represents heterogeneous Local Goals rather than a unitary human preference.

Accordingly, human-centered does not mean human-only causation or one human objective. #9 merely reserves unqualified Participation for intentional acts attributable to human persons or human-organized collectivities.

2.4 Governance — standing, affectedness, and the participation/influence boundary

Governance supplies the most direct series antecedent. Its Definition 4 grounds standing to participate in affectedness rather than contribution, rejects monotone headcount claims, and distinguishes Participation from influence: a participant can possess negligible binding power and still participate. #9 must preserve those boundaries.

Governance also supplies an epistemic rationale: distributed participants can reveal distributed effects. #9 inherits that possibility without assuming that information necessarily changes a decision or realization.

2.5 Governance under empirical contact — exposure and second-order structure

The governance companion adds two constraints. Exposure enters the potential-to-realization mapping as reach, susceptibility, and recourse; Participation neither replaces nor absorbs it. The companion also defines second-order interventions by their immediate target in authority, procedure, membership, information flow, or rule-making, providing the inherited structure for later institutional or return-path consequences.

Its measurement discipline also carries forward: apparently scalar constructs often decompose into structured objects. Participation should therefore not be collapsed into one amount before its act structure and consequence pathways are measured.

2.6 CFS actor aggregation — the boundary of Responsive Floor

CFS already makes realization actor-sensitive through Φ: the effective gate emerges from actors whose influence varies by jurisdiction and time, and the examined cases were consistent with the most restrictive sufficiently influential actor setting the gate.

That result is explicitly provisional: the cases were unilateral restrictions and did not test veto-max against a genuine multilateral alternative. Responsive Floor therefore begins only with the narrower question of whether pre-specified features of human Participatory Configuration can matter without formal veto or binding authority.

One additional series-reference debt comes from the technical Graduality feature catalog, which formalizes principled graduality as InformationResponsiveIteration(Π) ∧ PreservedRevisability(Π) (DFN002). #9 does not claim that conjunction as new. It draws on it in Foreclosure Position, where the relevant question is whether participatory information can still enter while substantively different realization choices remain revisable.

2.7 External antecedents to the analytic architecture

The Subject–Act–Consequence organization is not claimed as an original tripartite theory. Political-participation research already asks who participates, which acts they undertake, and how participation connects to influence; Verba/Nie, Verba/Schlozman/Brady, Fung, van Deth, Theocharis and others supply relevant classifications. Aristotle supplies the capacity/exercise antecedent used for Readiness, and Austin the act/effect antecedent used for Consequence.

Critical participation scholarship adds a further boundary. Arnstein links participation to the distribution of citizen power rather than to invited presence alone; Gaventa shows that participatory spaces are themselves shaped by visible, hidden, and less visible forms of power; and Cooke and Kothari warn that participation can be ritualized, co-opted, or used to legitimize arrangements participants do not control. #9 therefore treats Participation as analytically identifiable without assuming that participatory opportunity is equal, empowering, or consequential.

2.8 External antecedents to the Relational Grammar

Each grammar relation has antecedents. Realization Orientation draws on voice, conflict, support/opposition, and action-orientation literatures. Foreclosure Position has a particularly close legal antecedent in Aarhus/EIA requirements that participation occur while meaningful options remain open.

Actor Dependence draws on representation, lived-experience, standpoint, and experiential-knowledge scholarship showing that who speaks can sometimes change the epistemic content available to a decision process.

The bounded antecedent audit located clear antecedents for each constituent relation and no close prior framework using the same three relations as one act-level comparison unit for capability-to-realization dynamics. That absence is not itself the originality argument. The analytic contribution is the common relational representation: it allows the same observable act form to be compared under substitutions in realization orientation, decision state, and subject while preserving those differences for later consequence analysis. #10 must still show that this joint representation adds nonredundant information.

2.9 Antecedents to dynamic participation

Dynamic Participation also builds on established work: Meaningful Human Control and responsibility-gap scholarship block a simple inference from task automation to disappearance of human roles, while policy-feedback scholarship shows how prior interventions reshape later resources, organizations, interpretations, and participation. #9 specializes these antecedents by separating Subject-readiness from institutional return-path consequences.

Participatory Role Plasticity is therefore a useful series-level proposition, not the paper’s principal analytic contribution.

2.10 Non-claim ledger

The resulting boundary can be stated directly. The following ideas are used by #9 but are not claimed as discoveries of #9.

Idea used in #9	Primary antecedent	Status	#9 use
AGI-inclusive humanity / human-centered orientation	2025 precursor papers; formalized in #5	Precursor + inherited	Scope Participatory Subject
Human–AI reciprocal adaptation	#3 Synchronization	Inherited	Dynamic context
Ceiling ≠ Floor	CFS / #7	Inherited	Realization baseline
Standing from affectedness	#7	Inherited governance rule	Distinguish standing from readiness/action
Participation ≠ influence	#7	Inherited	Act/consequence boundary
Exposure = reach, susceptibility, recourse	#8	Inherited	Separate realization condition
Second-order institutional intervention	#8	Inherited	Institutional/return-path consequence
Actor-sensitive gate Φ	CFS	Inherited	Responsive Floor ancestry
Information-responsive iteration + preserved revisability (DFN002)	Graduality: A Feature Catalog of Gradual AGI	Series reference / catalog	Foreclosure Position / revisability boundary
Capacity ≠ exercise	Aristotle; civic-voluntarism relatives	External antecedent	Readiness/action distinction
Act ≠ effect	Austin	External antecedent	Act/consequence distinction
Participation has multiple dimensions	Verba/Nie; Fung; van Deth tradition	External antecedent	Background to grammar
Participation while options remain open	Aarhus/EIA	Close external antecedent	Foreclosure Position
Experience can make speaker identity epistemically relevant	Representation/experiential-knowledge literature	External antecedent	Actor Dependence
Automation need not erase all human control/responsibility roles	Meaningful Human Control; responsibility-gap literature	External antecedent	Role plasticity boundary

2.11 What remains for #9 to contribute

The remaining burden for #9 is therefore analytic rather than priority-based: defend the three-relation act representation and Conditional Non-Fungibility, preserve Subject–Act–Consequence separation, and state Responsive Floor as a testable series extension.

The rest of the paper develops those claims while keeping inherited and external debts visible in the ledger above.

References cited in this section

Aristotle. Metaphysics, Book IX (Theta), on potentiality (dunamis) and actuality/activity (energeia).

Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press. Revised edition, 1975.

Campbell, Andrea Louise. 2012. “Policy Makes Mass Politics.” Annual Review of Political Science 15: 333–351.

Fung, Archon. 2006. “Varieties of Participation in Complex Governance.” Public Administration Review 66(s1): 66–75.

Hirschman, Albert O. 1970. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press.

Mansbridge, Jane. 1999. “Should Blacks Represent Blacks and Women Represent Women? A Contingent ‘Yes.’” The Journal of Politics 61(3): 628–657.

Matthias, Andreas. 2004. “The Responsibility Gap: Ascribing Responsibility for the Actions of Learning Automata.” Ethics and Information Technology 6(3): 175–183.

Pierson, Paul. 1993. “When Effect Becomes Cause: Policy Feedback and Political Change.” World Politics 45(4): 595–628.

Santoni de Sio, Filippo, and Jeroen van den Hoven. 2018. “Meaningful Human Control over Autonomous Systems: A Philosophical Account.” Frontiers in Robotics and AI 5:15.

UNECE. 1998. Convention on Access to Information, Public Participation in Decision-Making and Access to Justice in Environmental Matters (Aarhus Convention), Article 6(4).

Verba, Sidney, and Norman H. Nie. 1972. Participation in America: Political Democracy and Social Equality. New York: Harper & Row.

Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

W.H.L. with Claude Sonnet 4. 2025. First Principles of AGI-Inclusive Humanity. Champaign Magazine, June 17, 2025.

W.H.L. with Gemini 2.5 Pro and GPT-4o. 2025. Philosophical Framework for AGI-Inclusive Humanity. Champaign Magazine, July 23, 2025.

W.H.L. et al. 2026. Gradual AGI as Synchronization for Transformative Adoption. Gradual AGI Series #3.

W.H.L. et al. 2026. Ceiling, Floor and Slope: A Falsifiable Dynamical Model of Synchronization for Gradual AGI.

W.H.L. et al. 2026. Gradual AGI as Optimization: A Conceptual Framework. Gradual AGI Series #5.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

van Deth, Jan W. 2014. “A Conceptual Map of Political Participation.” Acta Politica 49(3): 349–367. https://doi.org/10.1057/ap.2014.6.

Theocharis, Yannis, and Jan W. van Deth. 2016. “The Continuous Expansion of Citizen Participation: A New Taxonomy.” European Political Science Review. Published online November 21, 2016.

European Parliament and Council. 2011. Directive 2011/92/EU on the Assessment of the Effects of Certain Public and Private Projects on the Environment, Article 6(4).

Allen, Peter. 2022. “Experience, Knowledge, and Political Representation.” Politics & Gender 18(4): 1112–1140. Published online December 2, 2021. https://doi.org/10.1017/S1743923X21000362.

W.H.L., GPT-5.6 Sol, and Claude Opus 5. 2026. Graduality: A Feature Catalog of Gradual AGI. Publication Version v1.0, August 18, 2026.

Arnstein, Sherry R. 1969. “A Ladder of Citizen Participation.” Journal of the American Institute of Planners 35(4): 216–224. https://doi.org/10.1080/01944366908977225.

Gaventa, John. 2006. “Finding the Spaces for Change: A Power Analysis.” IDS Bulletin 37(6): 23–33. https://doi.org/10.1111/j.1759-5436.2006.tb00320.x.

Cooke, Bill, and Uma Kothari, eds. 2001. Participation: The New Tyranny? London: Zed Books.

3. Principles and Analytic Architecture

The Participation framework begins from a distinction in levels rather than from a taxonomy of behaviors. It asks three different questions: who occupies the participant position, what participation consists of when a subject acts, and what follows from that act. The resulting architecture is Participatory Subject → Participatory Act → Participatory Consequence. This organization is adopted because it prevents capacities, actions, and effects from being inferred from one another; it is not claimed as an original tripartite theory of participation.

The architecture is analytic rather than ontological in all three parts. Only the Subject layer directly asks an ontological and scoping question: what kind of entity occupies the participant position under the controlled vocabulary of this paper? The Act layer asks what structurally characterizes participation once an eligible subject acts. The Consequence layer asks what, if anything, follows. Calling the whole structure a ‘three-part ontology’ would therefore overstate the status of the latter two levels and is avoided.

Three principles govern the framework. They are deliberately fewer and broader than the later definitions and propositions. A principle sets a standing commitment of the analysis; a definition fixes an object; a proposition makes a claim about relations among objects. Keeping those functions separate is part of the paper’s attempt to avoid promoting every useful distinction into a first-order theoretical claim.

3.1 Principle A — Human-Centered Scope

Principle A (Human-Centered Scope). Unless explicitly qualified otherwise, Participation in this framework refers to intentional participation by human persons or human-organized collectivities in the realization of Gradual AGI.

The human-centered scope is chosen because the series already represents AI as a causal and adaptive part of the realization system. The unresolved object in #9 is narrower: the human capacity to act upon realization. Isolating that relation prevents model outputs, automated decisions, and other AI causal activity from swallowing the category of Participation before human agency can be analyzed separately.

The scope is also continuous with the AGI-inclusive-humanity and Optimization frameworks, which place humans and AI inside one co-developing civilizational system. A later generalized theory could broaden the eligible Subject class without changing the Subject–Act–Consequence architecture; #9 does not need to settle the conceptual or normative conditions for doing so.

The controlled vocabulary therefore permits a clean distinction: AI causal activity is not, by itself, human Participation. A human may participate through an AI system, but the attribution of the Participatory Act remains with the human person or human-organized collective when intentional action can be attributed there. This prevents every AI-mediated effect from being silently reclassified as Participation.

3.2 Principle B — Analytic Separation

Principle B (Analytic Separation). The status and capacities of a Participatory Subject, the Participatory Act undertaken by that subject, and the Participatory Consequence associated with that act are analytically distinct. None is inferred solely from another.

This principle supplies the basic non-equivalences used throughout the paper. A person can be affected without acting, possess standing without readiness, be ready without acting, act without influence, influence a process without changing the final realization, or produce a consequence different from the one intended. The architecture preserves each transition as an empirical or conceptual question rather than treating the sequence as automatic.

Several antecedents support the separation. Aristotle’s potentiality-actuality distinction supplies the older form of the capacity/exercise boundary. Austin’s distinction between acts and their further effects supplies a structural antecedent for keeping Participatory Act separate from Participatory Consequence. The governance framework already establishes that Participation is not influence and that standing derives from affectedness without implying governing power. None of these separations is claimed as original here.

Those distinctions become load-bearing inside one capability-to-realization framework: later sections can ask separately whether a subject is ready, whether an act occurred, what grammar it had, whether an institution could receive it, and what consequence followed.

3.3 Principle C — Participatory Contingency

Principle C (Participatory Contingency). The consequence of Participation is contingent on the configuration of Participatory Acts and on the conditions through which those acts enter realization; participatory presence, form, volume, or effort alone does not determine consequence.

The principle rejects several monotone shortcuts without asserting their opposites. More participants need not mean more influence. A formally recognized channel need not be consequential. A highly visible act need not alter realization. A small number of acts can sometimes matter more than a larger number, but the framework does not presume that either. It requires the relevant configuration and conversion conditions to be observed.

This is consistent with the governance framework’s earlier refusal to make the value of participation monotone in headcount and with the empirical governance companion’s broader finding that apparently scalar constructs often decompose into structured objects under measurement. Participation is treated the same way. Its configuration can include different subjects, act grammars, forms, presentations, and institutional relations without being compressed into a universal scalar.

Contingency also protects the framework from defining Participation by success. A failed petition, ignored testimony, or unused institutional channel can still be part of a Participatory Configuration. Null consequence is therefore not an embarrassment to the theory; it is one of the outcomes the theory must be able to represent.

3.4 Power as a conditioning environment

Participation is analytically neutral but not socially neutral. Resources, authority, institutional access, credibility, agenda control, and channel design condition each transition from Subject to Act to Consequence. Power can shape who obtains information and develops relevant capacities, whose acts enter recognized channels, whose claims are treated as legible or credible, whether response is obligatory, and who can convert information into binding action.

These conditions are not a fourth analytic level and are not folded into Actor Dependence. Actor Dependence asks whether substituting the speaker changes the act’s evidentiary, experiential, or standing-related content; power asks how an act is enabled, filtered, weighted, or converted within a structured environment. Highly actor-dependent testimony can be systematically discounted, while a low-dependence technical submission can be institutionally privileged. Formal participation can also occur after substantive closure or be used to legitimate an arrangement whose basic terms are no longer open.

The distinction follows critical participation scholarship from Arnstein through Gaventa and Cooke and Kothari. It also fixes the place of resource inequality in this framework: time, money, language access, expertise, organizational support, and AI-mediated filtering or summarization may condition readiness, opportunity, visibility, or uptake, but they do not become synonyms for Readiness or Participation. Power conditions the architecture; it does not replace its levels.

3.5 The three analytic levels

Level	Governing question	Primary objects	Status in #9
Participatory Subject	Who occupies the participant position, and with what capacities?	Human person / human-organized collective; readiness; standing relations	Primarily inherited and scoping
Participatory Act	What structurally distinguishes one act of Participation from another?	Intentional act; relational grammar; form; presentation	Principal analytic contribution candidate
Participatory Consequence	What changes, persists, or fails to change after Participation?	Configuration; consequence classes; recursive effects; realization linkage	Extension and empirical bridge

The asymmetry across the three levels is intentional. The Subject layer mostly narrows and organizes inherited human-centered commitments. The Act layer carries the paper’s strongest analytic burden through the proposed Relational Grammar. The Consequence layer reconnects Participation to the inherited Ceiling-Floor and governance apparatus and therefore proceeds chiefly by extension rather than replacement.

The Act layer therefore receives the strongest scrutiny: the key question is whether Realization Orientation, Foreclosure Position, and Actor Dependence form a useful joint representation rather than merely a tidy vocabulary.

3.6 Core non-equivalences

The architecture can be summarized through a small set of non-equivalences. They are not separate propositions; they are boundary rules required to keep the later analysis coherent.

Non-equivalence	Reason for keeping it separate
Affectedness ≠ Participation	Being subject to an effect does not establish that an intentional act occurred.
Contribution ≠ Participation	Human-originated inputs can enter AI systems without intentional realization-directed action.
Standing ≠ Readiness	Entitlement to participate does not establish the capacity to formulate and undertake an act.
Readiness ≠ Participation	Capacity can exist without exercise.
Readiness ≠ Temporal Opportunity	Subject capacity and the openness of the relevant decision can move independently.
Participatory Act ≠ Influence	An act can occur without binding power or institutional uptake.
Influence ≠ Realization Change	A process can be affected without changing the final realized state.
No realization change ≠ No consequence	Informational, procedural, institutional, readiness, or delayed effects can occur without immediate Floor change.
AI causal activity ≠ human Participation	The paper’s controlled vocabulary reserves unqualified Participation for human intentional action.

3.7 What should not become propositions

Section 2 identifies several important distinctions that should not be promoted into new propositions: Readiness versus Participation, Participation versus influence, capability versus realization, human–AI reciprocal adaptation, and the separation of automation from human control or responsibility. Their role here is architectural, not priority-bearing.

Proposition status is reserved for claims that add theoretical burden: Conditional Non-Fungibility, Participatory Role Plasticity, and Participation-Sensitive Realization. The smaller number is deliberate.

Important claim	Why not a new #9 proposition	Where it belongs
Readiness ≠ Participation	Capacity/action distinction is antecedented; #9 specifies its application	Definition 2 and Subject analysis
Participation ≠ influence	Already explicit in #7	Principle B / Consequence boundary
Capability ≠ realization	Foundational to CFS/#7/#8	Inheritance and §8
Human-AI co-development	Already explicit in #3/#5	Inheritance
No universal scalar participation score	Methodological consequence of structured measurement	Principle C / #10 design
Second-order institutional effects	Inherited second-order governance structure; applied here as a consequence type	§6–§7
Readiness lag	Requires empirical measurement after readiness and foreclosure are separated	§7 / #10

3.8 Architecture over time

Although the architecture is introduced as Subject → Act → Consequence, later consequences can alter the conditions of later participation. A subject-readiness consequence can change what future subjects are capable of doing; an institutional consequence can change whether future acts have a return path to consequential decisions. The sequence is therefore recursive over time without requiring a fourth analytic level.

The recursion is best understood as repeated passage through the same architecture: later Subjects inherit changed capacities and environments, later Acts occur under changed conditions, and later Consequences may again alter the next cycle. This temporal structure is developed in Section 7. It does not turn readiness, institutions, or return paths into new basic categories of Participation.

3.9 From architecture to definitions

The remainder of the conceptual framework proceeds in the order set here. Section 4 specifies the Participatory Subject and Participatory Readiness. Section 5 defines the Participatory Act and develops its Relational Grammar. Section 6 defines Participatory Configuration and Participatory Consequence. Section 7 shows how roles, readiness, and institutional return paths can change over time. Section 8 then asks when, if at all, those participatory configurations become sufficiently realization-relevant to support a Responsive Floor.

This order is not merely expositional. It encodes the paper’s principal methodological constraint: do not infer an act from a subject, a consequence from an act, or a realized effect from the presence of participation. Each transition must be separately specified and, where empirical, separately observed.

References cited in this section

Aristotle. Metaphysics, Book IX (Theta), on potentiality (dunamis) and actuality/activity (energeia).

Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press. Revised edition, 1975.

Santoni de Sio, Filippo, and Jeroen van den Hoven. 2018. “Meaningful Human Control over Autonomous Systems: A Philosophical Account.” Frontiers in Robotics and AI 5:15.

W.H.L. et al. 2026. Gradual AGI as Synchronization for Transformative Adoption. Gradual AGI Series #3.

W.H.L. et al. 2026. Gradual AGI as Optimization: A Conceptual Framework. Gradual AGI Series #5.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

Arnstein, Sherry R. 1969. “A Ladder of Citizen Participation.” Journal of the American Institute of Planners 35(4): 216–224. https://doi.org/10.1080/01944366908977225.

Gaventa, John. 2006. “Finding the Spaces for Change: A Power Analysis.” IDS Bulletin 37(6): 23–33. https://doi.org/10.1111/j.1759-5436.2006.tb00320.x.

Cooke, Bill, and Uma Kothari, eds. 2001. Participation: The New Tyranny? London: Zed Books.

4. Participatory Subject

A theory of Participation must specify who occupies the participant position before it can classify what participants do or ask what follows from their action. In this framework, that specification is deliberately human-centered. The choice is one of controlled vocabulary and theoretical scope, not a claim that artificial systems can never possess agency, moral status, responsibility, or other forms of subjecthood. Gradual AGI already describes a jointly evolving civilizational environment in which humans, artificial intelligence, institutions, infrastructure, and knowledge interact. Participation, as used here, identifies the human side of that relation.

This orientation is inherited rather than newly introduced. The pre-series work on AGI-inclusive humanity explicitly framed the problem as human-centered while asking how increasingly capable AI should be integrated into human systems and values. The later Optimization framework incorporated AGI-inclusive humanity formally as its Optimization Subject: an evolving civilizational ecosystem formed through the co-development of humans, AI, institutions, infrastructure, and the broader social environment. The present paper narrows that inherited orientation for a different purpose. It asks which human persons and human-organized collectivities can occupy the participant position within capability-to-realization processes.

4.1 Definition 1 — Participatory Subject

Definition 1 (Participatory Subject). A Participatory Subject is a human person or human-organized collective capable of undertaking intentional action with respect to a prospective, ongoing, or realized condition of Gradual AGI.

The definition has two boundaries. First, it is human-centered by stipulation. Artificial systems may recommend, persuade, filter, rank, generate, execute, negotiate, or otherwise alter the realization process. Those effects may be causally important without being coded as Participation in the technical sense of this paper. Where a human person or human-organized collective intentionally uses an AI system as an instrument of action, the participatory subject remains the human or collective to which the intentional act can be attributed. Where an AI system acts without such an attributable human participatory act, the event may still be relevant to realization, but it is not classified here as human Participation. An AI-drafted comment that a person deliberately reviews, adopts, and submits can therefore constitute that person’s Participatory Act; automatically generated or transmitted output without attributable human intentional action does not qualify on that basis alone.

Second, a collective is included because many consequential acts are organized rather than individual. A neighborhood association, professional body, firm, university, union, public-interest organization, or public institution can participate through human-organized processes that authorize, aggregate, or express action. Calling such an organization a Participatory Subject does not imply that it possesses a unitary mind. It means only that the framework can attribute an intentional participatory act to an organized human collective at the level at which the act is actually undertaken. Operationally, collective intentionality is attributed only where the act is authorized, aggregated, or institutionally expressed through the collective’s human-organized process.

The definition is intentionally wider than affectedness. Directly affected people are central to many governance settings, but participation can also be undertaken by experts, advocates, voters, professional communities, researchers, firms, or others who are not themselves the immediate bearers of a realized effect. The framework therefore does not define the universe of Participatory Subjects by affectedness alone.

4.2 Subject, contribution, and affectedness

Three relations that often travel together should not be collapsed. A human being may contribute data, labor, knowledge, or cultural material to AI development without intentionally participating in any particular realization. A human being may be affected by a realization without taking any participatory action. And a human being may intentionally participate in a realization without having contributed to the underlying technology or being directly affected by its present consequences.

The governance framework preceding this paper makes one of these distinctions explicit. Its Definition 4 grounds standing to participate in governance in affectedness rather than contribution: those whose outcomes are altered have standing in the processes that shape them. It also states that participation is not influence; a person can participate while possessing negligible binding power. The present framework preserves those results but changes the unit of analysis. Affectedness can establish governance standing without establishing that an act occurred, and the broader category of Participatory Subject can include actors whose participation rests on some basis other than direct affectedness.

This distinction is particularly important for AI because human-originated material is pervasive. Treating all human contribution as Participation would convert training data, historical writing, routine labor, and involuntary data capture into participatory acts by definition. That would erase the intentional-action boundary established in Section 5 before the theory has any opportunity to analyze it.

4.3 Definition 2 — Participatory Readiness

Definition 2 (Participatory Readiness). Participatory Readiness is the capacity of a potential Participatory Subject to recognize a realization-relevant issue, form a position or claim concerning it, and undertake an intentional Participatory Act.

Readiness is a capacity, not the act itself. The distinction belongs to the much older family of distinctions between a power and its exercise, classically articulated in Aristotle’s analysis of potentiality and actuality. The present use is narrower and social rather than metaphysical: a subject may possess the capacity to participate and never exercise it. Participation therefore cannot be inferred from readiness, and lack of observed participation cannot by itself establish that the relevant subjects were unready.

The three elements named in Definition 2 are analytically distinguishable subcapacities: issue recognition, position or claim formation, and capacity to undertake an act. They need not develop together, and #10 may measure them separately. Definition 2 groups them only because all three are required for the subject-side capacity that this paper calls Readiness.

The definition also has close relatives in political-participation research. The Civic Voluntarism tradition explains participation partly through resources, civic skills, psychological engagement, and recruitment: people do not participate simply because an issue exists, but because they possess or acquire capacities and incentives that make action possible. Participatory Readiness does not reproduce that explanatory model. It isolates the capacity side needed by this framework so that readiness can be distinguished from the act and from the institutional consequences that follow.

Readiness is realization-relative rather than a permanent trait. A person can be highly capable of participating in a workplace deployment decision and poorly prepared to participate in a frontier-model safety debate; a local government can possess mature land-use expertise and little capacity to evaluate model capability claims. ‘Realization-relevant’ is therefore judged against the specified realization and the conditions under which it is selected, revised, maintained, redirected, or challenged—not against a coder’s view that an issue is generally important. Readiness may change through education, experience, organization, or access to information.

Readiness should not be defined by willingness or motivation. A subject can possess all three capacities and choose silence, exit, delegation, or non-engagement. Power and resource conditions can also enable or obstruct the formation or exercise of those capacities without becoming part of Readiness by definition: access to information, time, money, language, recognized standing, organizational support, or an open channel may determine whether issue recognition, claim formation, or action is practically available. The subcapacities describe what the subject can do with respect to the realization; the structural conditions that enable, constrain, filter, or suppress that capacity are observed separately.

4.4 Standing, readiness, action, and influence

The architecture requires four questions to remain separate because each can vary independently of the others.

Concept	Question	What it does not establish
Standing	Who has a claim to participate in a governance process?	Readiness, action, or influence
Readiness	Who has the capacity to formulate and undertake a Participatory Act?	Willingness, temporal opportunity, or action
Participatory Act	Who intentionally acts with respect to a realization or its conditions?	Reception, influence, or outcome change
Influence / consequence	What changes, if anything, after the act enters a realization process?	The prior existence or quality of standing or readiness

Standing and readiness are especially easy to conflate. The governance framework supplies affectedness as a minimum basis for standing in the processes that shape an affected party’s outcomes. Yet an affected party can possess standing while lacking information, time, organization, language access, technical understanding, or other capacities needed to formulate an act. Conversely, a highly capable expert may be ready to participate while lacking affected-party standing in a particular governance procedure. The categories answer different questions.

Readiness and temporal opportunity must likewise remain separate. A subject can be ready only after the relevant decision has effectively hardened, or unready while meaningful alternatives remain open. Foreclosure Position therefore belongs to the relational grammar of the Act rather than to the definition of Readiness. This separation preserves a genuine empirical question: whether human capacities to participate arise early enough to encounter decisions while substantively different realization choices remain available.

4.5 Purpose-bearing without a single human purpose

The human-centered premise does not imply a single human objective. A Participatory Subject is purpose-bearing in the minimal sense that the intentional act is attributable to a human person or human-organized collective acting for reasons, aims, preferences, commitments, or claims. Different subjects may pursue incompatible purposes, and the same collective may contain unresolved internal disagreement.

This pluralism is already consistent with the Optimization framework, which represents heterogeneous stakeholders with different Local Goals rather than assuming that all participants share one operational objective. The Participation framework carries that heterogeneity forward. Human-centered means that human purposes and human standing remain constitutive objects of analysis; it does not mean that ‘humanity’ supplies one preference ordering that the theory can read off in advance.

The distinction also protects the framework from an easy anthropomorphic shortcut. An AI system can optimize, recommend, and generate behavior that appears purposive. The controlled vocabulary of this paper does not need to settle whether those processes count as genuine purposes in a philosophical sense. It asks a narrower question: when the paper says Participation without qualification, whose intentional action is being analyzed? The answer is human persons and human-organized collectivities.

4.6 Readiness as a conditioning property

Participatory Readiness is therefore a conditioning property of the Subject, not a fourth layer of the framework. Subject, Act, and Consequence remain the analytic architecture. Readiness helps explain how a potential subject can become capable of action, but it is neither sufficient nor necessary evidence of any particular consequence.

This placement matters in periods of rapid capability change. Educational reform, professional retraining, public learning, organizational adaptation, and the acquisition of technical or legal competence can change the readiness of future Participatory Subjects. Those interventions may themselves be Participatory Acts when intentionally directed toward realization or its conditions, but ‘readiness-building’ does not require a separate theoretical category. Its analytically relevant result is a possible subject-readiness consequence: an earlier act changes the capacities with which later subjects approach later realization decisions.

The empirical residue is deliberately left open. The framework does not assume that AI capability necessarily outruns participatory readiness, that readiness monotonically improves with exposure to AI, or that ready subjects will act. It permits those relations to be measured rather than built into the definitions.

4.7 Boundary with the Participatory Act

The Subject section ends at capacity. Once a Participatory Subject intentionally acts with respect to a realization or to the conditions under which it is selected, revised, maintained, redirected, or challenged, the unit of analysis changes. Section 5 then asks what structural relations distinguish that act from another: its Realization Orientation, its Foreclosure Position, and its Actor Dependence.

Keeping this boundary sharp prevents retrospective inference in both directions. An influential act does not prove that the actor had stronger standing or greater readiness than others; a silent subject does not prove incapacity; and an affected population does not become participatory merely because its outcomes change. The framework proceeds in order: identify the Subject and its capacities, identify the Act if one occurs, and only then examine Consequence.

References cited in this section

Aristotle. Metaphysics, Book IX (Theta), especially the distinction between potentiality (dunamis) and actuality or activity (energeia).

Brady, Henry E., Sidney Verba, and Kay Lehman Schlozman. 1995. “Beyond SES: A Resource Model of Political Participation.” American Political Science Review 89(2): 271–294.

Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

W.H.L., Claude Sonnet 4. 2025. First Principles of AGI-Inclusive Humanity. Champaign Magazine.

W.H.L., Gemini 2.5 Pro, and GPT-4o. 2025. Philosophical Framework for AGI-Inclusive Humanity. Champaign Magazine, July 23, 2025.

W.H.L. et al. 2026. Gradual AGI as Optimization: A Conceptual Framework. Gradual AGI Series #5.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

5. Participatory Act

The preceding architecture distinguishes the Participatory Subject from the Participatory Act and from any Participatory Consequence. That separation matters most at the level of the act. A person may be affected without acting, ready without acting, or entitled to participate without exercising that standing. Conversely, a person may act without producing any observable change. The act therefore requires its own object of analysis rather than being inferred from either the status of the subject or the result that follows.

Participation research already offers many classifications of political behavior and participatory institutions. Verba and Nie treated participatory acts as multidimensional rather than interchangeable, distinguishing them by such features as scope of outcome, conflict, and initiative; later work revisited and revised those dimensions rather than abandoning the dimensional approach (Verba and Nie 1972; Claggett and Pollock 2006). Fung’s democracy-cube framework likewise locates participatory mechanisms according to who participates, how participants communicate and decide, and how their discussions connect to public action (Fung 2006). These are important antecedents. The present framework does not claim that participatory acts have dimensions, or that participation can be classified, as a new idea. Its narrower task is to identify the relations that matter when a human act is analyzed specifically with respect to the movement from AI capability to societal realization.

5.1 Definition 3 — Participatory Act

Definition 3 (Participatory Act). A Participatory Act is an intentional action by a Participatory Subject directed toward a prospective, ongoing, or realized condition of Gradual AGI, or toward the conditions under which such a realization is selected, revised, maintained, redirected, or challenged.

The definition is deliberately narrower than human causal contribution. Human-originated data, labor, preferences, rules, or artifacts may enter the AI lifecycle without constituting participation in the technical sense used here. Training data scraped without the knowledge of its creator is human-originated input but not, for that reason alone, a Participatory Act. A general attitude toward AI is likewise not an act merely because it can later affect behavior. Public sentiment becomes relevant to Participation when a human subject intentionally does something with respect to a realization or to the conditions that shape it.

The definition is also broader than formal governance. A public comment, testimony, petition, vote, refusal, negotiation, purchase decision, organized campaign, institutional reform, or other intentional intervention may qualify when it is directed toward a realization or its conditions. The framework therefore does not equate Participation with a statutory hearing, nor does it make binding authority a condition of participation. Whether an act reaches an addressee, enters a procedure, changes a decision, or changes realization belongs to the analysis of consequence rather than to the definition of the act.

Intentionality is required only at the level of the act’s direction toward the relevant condition. The framework does not require an actor to foresee all consequences of the act, to possess a complete causal model, or to succeed. A resident who testifies about a local burden participates even if the testimony changes nothing. An institution that redesigns a curriculum in response to rapidly changing AI capabilities participates even if its eventual effects are uncertain. Success is not constitutive of participation.

5.2 Grammar, form, and presentation

Three levels should be kept separate. The form of a Participatory Act describes what visibly occurs: testimony, petition, refusal, negotiation, appeal, public comment, collective demonstration, market action, or another observable act type. Its presentation describes how the act appears socially or institutionally: individual or collective, formal or informal, public or private, direct or mediated, recorded or unrecorded. Its grammar, by contrast, describes the act’s relational structure with respect to the realization process.

This distinction prevents a common inferential shortcut. Two acts that share a form need not occupy the same structural position. Two public comments may be submitted through the same portal on the same day, yet one may report a generally reproducible technical point while the other supplies first-person evidence unavailable without the affected speaker. Two petitions may contain the same number of signatures while one arrives before a permit is substantially settled and another after the decisive commitments have already been made. Form and volume are therefore observable descriptors, not sufficient summaries of participatory structure.

5.3 Definition 4 — Relational Grammar of the Participatory Act

Definition 4 (Relational Grammar of the Participatory Act). A Participatory Act may be distinguished along at least three analytically separable relations: Realization Orientation, describing the act’s relation to the realization it addresses; Foreclosure Position, describing the act’s position relative to the continued availability of substantively different realization choices; and Actor Dependence, describing the extent to which substituting another actor would alter the act’s evidentiary, experiential, or standing-related content. These dimensions describe the act’s structure for purposes of this framework; they are not asserted to be exhaustive or statistically independent.

The word relational is doing substantive work. None of the three dimensions is purely an intrinsic property of an isolated behavior. Realization Orientation relates the act to what is being realized. Foreclosure Position relates the act to the state of the relevant decision trajectory. Actor Dependence relates the act to the subject performing it. The grammar therefore locates an act within a realization process rather than assigning it an abstract type detached from that process.

The constituent relations have different antecedents, but the analytic claim is about their joint use. Each relation supports a different substitution test around the same act: change the realization orientation while holding decision position and actor relation fixed; move an otherwise similar act across an open or foreclosed decision state; or substitute the actor and ask whether relevant content is preserved. Joint coding therefore identifies which substitutions leave participatory structure unchanged and which do not. That is the work Proposition A turns into a comparison rule.

The selection is functional rather than exhaustive. These three relations answer three questions that capability-to-realization analysis is especially prone to collapse: what the act is trying to do to realization, whether substantively different realization choices remain open when it enters, and whether relevant content changes when the actor is substituted. Other properties – including temporal urgency, resource intensity, coalition size, institutional channel, and presentation – may be empirically important, but they describe the act or its environment rather than these three relational objects and therefore remain separate observation fields unless later evidence shows that the grammar requires expansion.

5.4 Realization Orientation

Realization Orientation asks what relation the act bears to the realization it addresses. Hirschman’s account of voice is an important antecedent: voice is an attempt to change an objectionable condition from within rather than leave it through exit (Hirschman 1970). Earlier participation research likewise recognized conflict and counter-participation as properties that distinguish political acts (Verba and Nie 1972; Claggett and Pollock 2006). The present dimension is narrower in object. It asks not whether the act is generically conflictual, but how the act is oriented toward a particular prospective, ongoing, or realized configuration of Gradual AGI.

Illustrative coding candidates include supportive, oppositional, conditional or redirective, and reporting or nondirective acts. These are provisional distinctions, not a closed typology: #10 may merge, subdivide, or extend them if reliable coding or recurrent cases require it. Conditional or redirective is an Orientation value, not a fourth grammar dimension. A conditional act can independently occur before or after foreclosure and can be high or low in Actor Dependence; mixed or contradictory cases may therefore be coded unresolved or represented with more than one orientation only if the coding protocol earns that rule. Conditional or redirective orientation is coded from the act’s stated condition or requested redirection, not from the coder’s estimate of whether the condition is likely to be met or from the outcome later observed.

Reporting or evidentiary acts require similar care. A person may describe a burden, submit measurements, or document an experienced effect without prescribing an outcome. The act can still enter a realization process and later support several incompatible responses. The framework therefore does not force every act into a normative direction it does not express. Where orientation cannot be established from the act itself, it remains unresolved rather than inferred from the eventual use others make of it. In #10, an act whose orientation cannot be reliably assigned from its stated terms should be coded Unresolved rather than forced into an orientation category; the companion should report the fraction of unresolved cases and test whether excluding them changes substantive results.

5.5 Foreclosure Position

Foreclosure Position asks whether a Participatory Act occurs while substantively different realization choices remain available. This is not a new democratic principle. The Aarhus Convention requires early public participation while all options remain open, and the European Union’s Environmental Impact Assessment Directive similarly requires early and effective participation before the relevant development-consent decision is taken (UNECE 1998, art. 6(4); Directive 2011/92/EU, art. 6(4)). The Gradual AGI framework generalizes that established insight beyond a particular administrative procedure.

The relevant issue is substantive rather than merely formal timing. A hearing can occur before a final vote yet after contracts, infrastructure commitments, organizational dependencies, or other prior choices have narrowed the feasible set so sharply that the nominal decision no longer represents an open alternative. Conversely, a formal decision need not make every later act remedial if the operative conditions remain revisable. The question is therefore not simply whether the act came ‘before’ or ‘after’ a date. It is whether meaningfully different realization choices remained available to the decision-maker or other consequential actor at the time the participatory information could enter. Contracts, sunk costs, infrastructure commitments, formal decisions, and later review rights are therefore possible indicators of substantive openness or lock-in, not a closed stage taxonomy; #10 may refine them by domain.

This dimension also connects directly to DFN002 in the technical Graduality feature catalog, which defines principled graduality as InformationResponsiveIteration(Π) ∧ PreservedRevisability(Π) (W.H.L., GPT-5.6 Sol, and Claude Opus 5 2026). It should not, however, be folded into Participatory Readiness. A subject may be fully ready to participate after the relevant choice has hardened; another may be unready while the window remains open. Readiness is a capacity of the Subject. Foreclosure Position is a relation between an Act and a decision state. Keeping them separate preserves an empirical question rather than resolving it by definition.

5.6 Actor Dependence

Actor Dependence asks how much of a Participatory Act’s relevant content depends on who performs it. The dimension has clear antecedents in representation and epistemic scholarship. Mansbridge argues that descriptive representation can improve substantive representation in some contexts because shared experience improves communication and supplies experiential knowledge (Mansbridge 1999). Allen develops the point more explicitly in epistemic terms: subjective experience can generate distinctive content and capacities that allow people to understand political conditions in ways unavailable from an otherwise similar external position (Allen 2022).

The implication for Participation is not that identity automatically confers authority, nor that only directly affected people can speak about a realization. It is that substitution can sometimes change the act itself. A general technical claim about noise, grid reliability, model performance, or labor-market exposure may be reproducible by another comparably qualified speaker. First-person testimony that a particular noise pattern prevents sleep, that an automated decision produced a specific burden, or that a workflow change altered a worker’s actual practice contains an experiential component that disappears if the speaker is replaced by an unaffected observer. The two acts may be equally sincere and may enter the same institutional channel, yet they are not informationally identical.

Actor Dependence must also be distinguished from speaker authority, celebrity, status, expertise, or attention. A prominent technologist may attract greater media coverage than an unknown researcher while advancing an argument that another competent speaker could reproduce without loss of substantive content. Conversely, an ordinary affected person may provide highly actor-dependent evidence even when the institution accords the speaker little power. The framework therefore separates dependence of content on the actor from the weight that institutions or audiences later assign to that actor. High Actor Dependence is neither necessary nor sufficient for high institutional uptake: a low-dependence technical submission may decisively alter a decision, while highly actor-dependent experiential testimony may be ignored.

5.7 Analytic separability

The three relations are not assumed to be orthogonal, and empirical covariance would not invalidate the grammar. Post-foreclosure acts may, for example, often be more actor-dependent because remedy commonly turns on experienced harm. The relevant test is weaker and more useful: whether one dimension can be reconstructed from the others without losing a distinction that matters to the act.

Simple counterexamples show why none is presently reducible. Before foreclosure, a generic technical submission can have low Actor Dependence, while first-person testimony by an affected resident can have high Actor Dependence. After foreclosure, a standardized request for rule reconsideration can again have relatively low Actor Dependence, while individualized testimony seeking remedy can remain high. Likewise, highly actor-dependent testimony can support, oppose, condition, or merely report on a realization. Realization Orientation therefore does not determine Actor Dependence, and neither determines Foreclosure Position.

The following pairs isolate the intended distinctions. They are illustrations for discrimination, not a substitute for the coding protocol that #10 must develop.

Shared visible form / channel	Act A	Act B	Distinction isolated
Public comment	Unconditional support for the proposed realization	Support only if siting, use, or human-role conditions change	Realization Orientation
Petition through the same procedure	Submitted while materially different choices remain open	Submitted after a binding commitment or equivalent substantive lock-in	Foreclosure Position
Technical/evidentiary submission	Generic analysis reproducible by another comparably qualified speaker	First-person testimony of an experienced burden not preserved by actor substitution	Actor Dependence
Expert comment	Prominent expert states a reproducible technical proposition	Unknown comparably qualified expert states the same proposition	Boundary control: status/attention ≠ Actor Dependence

Worked comparison. Consider three comments submitted through the same project portal. A supports approval only if cooling-noise limits are changed, enters while design alternatives remain open, and relies on a reproducible engineering calculation. B states the same conditional orientation at the same time but supplies first-person evidence of a sleep burden that cannot be preserved by substituting an unaffected speaker. C repeats A’s engineering argument after contractual and permitting commitments have made the alternative design infeasible. Form and channel are identical; orientation is similar. The grammar shows why B is not substitutable for A on Actor Dependence and why C occupies a different decision relation from both. Those distinctions determine what can legitimately be pooled before any claim about consequence is made.

Illustrative coding of the comparison: A = conditional/redirective orientation, open foreclosure position, relatively low Actor Dependence; B = conditional/redirective orientation, open foreclosure position, high Actor Dependence; C = conditional/redirective orientation, foreclosed position, relatively low Actor Dependence. The assignments are made from the act and contemporaneous decision state, not from whether any comment later succeeds.

The three dimensions remain provisional. A dimension that adds no unrecoverable distinction should be demoted or removed; a recurrent relation the grammar cannot represent may justify extension. The criterion is discriminatory work, not three-part symmetry.

5.8 Proposition A — Conditional Non-Fungibility

Proposition A (Conditional Non-Fungibility of Participatory Acts). Equality of visible form, institutional channel, or numerical volume is insufficient to establish equivalence between Participatory Acts when the acts differ materially in Realization Orientation, Foreclosure Position, or Actor Dependence.

Materiality is instrument-specific and must be fixed before outcome inspection. A grammar difference is material when it crosses a pre-declared coding boundary on the relevant relational field; lexical, stylistic, or rhetorical variation that leaves the coded relation unchanged is not material. For example, ‘support if X changes’ and unconditional support differ materially on Realization Orientation when the protocol distinguishes conditional from supportive acts, while two differently worded statements that both satisfy the same conditional rule do not. #10 must set coding thresholds and unresolved rules before outcomes are inspected; if those rules cannot be made reliable, the relevant dimension or Proposition A must contract under Section 9.

The proposition is analytic rather than causal. It does not assert that different grammatical profiles necessarily produce different outcomes. A decision-maker may ignore every distinction the framework records; two differently structured acts may also converge on the same consequence. The narrower claim is that visible sameness does not license an assumption of participatory equivalence.

This restriction matters because common empirical measures make exactly that compression. Counts of comments, speakers, petition signatures, organizations, or submissions are often useful descriptive quantities, but they do not by themselves reveal what relation those acts bore to the realization, whether alternatives were still open, or whether the identity and experience of the actor formed part of the evidence. The same is true of act labels. ‘Public comment’ is a form; it does not specify whether the comment was supportive, conditional, post-foreclosure, or actor-dependent. ‘Testimony’ names a presentation; it does not establish whether the testimony could have been supplied equivalently by another speaker.

Conditional Non-Fungibility therefore sets a comparison rule for the empirical companion: do not aggregate acts as equivalent solely because an institutional record gives them the same name or because they contribute one unit to the same count. Where the grammar differs materially, equivalence must be demonstrated rather than presumed. The proposition does not require a universal weighting rule, and it does not imply that a high value on any dimension is normatively superior. A highly actor-dependent act can be mistaken; an early act can be ignored; an oppositional act can be poorly grounded. The grammar preserves distinctions before evaluation rather than deciding the evaluation in advance.

5.9 Boundary with consequence

The act-level analysis ends before uptake. Whether the act reaches an identifiable recipient, enters a procedure, receives a reasoned response, changes an institution, alters a later decision, or changes the realized Floor belongs to Participatory Consequence. This boundary echoes Austin’s distinction between an act performed and the further effects it produces: an illocutionary act and its perlocutionary effects are analytically distinguishable even when they are practically connected (Austin 1962/1975). The analogy is structural rather than literal. Participatory acts are not reducible to speech acts, and many are nonverbal. The point is simply that performing an act and producing a consequence are different questions.

Keeping the boundary sharp also prevents retrospective classification. An unsuccessful petition does not cease to be participatory because it failed. A public comment that later becomes decisive is not retroactively a different act because of its success. Consequence can inform causal interpretation, but it does not define the grammar ex post. The act is classified from its orientation, decision position, and actor relation as they stood when the act occurred.

This provides the handoff to the next section. Participatory Configuration will describe how multiple acts and subjects are distributed across these relations; Participatory Consequence will ask what, if anything, changes as those configurations enter institutions and realization processes. The present section establishes only the structure that must be preserved before that conversion is examined.

References cited in this section

Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press. (Lecture IX distinction between illocutionary and perlocutionary acts; later editions 1975.)

Allen, Peter. 2022. “Experience, Knowledge, and Political Representation.” Politics & Gender 18(4): 1112–1140. https://doi.org/10.1017/S1743923X21000362.

Claggett, William, and Philip H. Pollock III. 2006. “The Modes of Participation Revisited, 1980–2004.” Political Research Quarterly 59(4): 593–600. https://doi.org/10.1177/106591290605900408.

European Parliament and Council. 2011. Directive 2011/92/EU on the Assessment of the Effects of Certain Public and Private Projects on the Environment, Article 6(4).

Fung, Archon. 2006. “Varieties of Participation in Complex Governance.” Public Administration Review 66(s1): 66–75. https://doi.org/10.1111/j.1540-6210.2006.00667.x.

Hirschman, Albert O. 1970. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press.

Mansbridge, Jane. 1999. “Should Blacks Represent Blacks and Women Represent Women? A Contingent ‘Yes.’” The Journal of Politics 61(3): 628–657. https://doi.org/10.2307/2647821.

UNECE. 1998. Convention on Access to Information, Public Participation in Decision-Making and Access to Justice in Environmental Matters (Aarhus Convention), Article 6(4).

Verba, Sidney, and Norman H. Nie. 1972. Participation in America: Political Democracy and Social Equality. New York: Harper & Row.

W.H.L., GPT-5.6 Sol, and Claude Opus 5. 2026. Graduality: A Feature Catalog of Gradual AGI. Publication Version v1.0, August 18, 2026.

6. Participatory Consequence

Sections 4 and 5 separate the Participatory Subject from the Participatory Act. This section completes the analytic sequence by asking what, if anything, follows from participation. The separation is necessary because participation is not influence. The governance framework already established that a participant can possess negligible binding power and still participate; adding participants is not the same as distributing authority. The present framework extends that distinction from governance standing to the full capability-to-realization process.

A consequence theory must therefore admit failure, indirection, delay, and recursion. An act may change a decision immediately, alter only the procedure through which a later decision is made, change what future participants know or can do, create a record that matters much later, or produce no detected change at all. None of those possibilities changes whether the original act qualified as participation. Consequence is a separate object of analysis.

6.1 Definition 5 — Participatory Configuration

Definition 5 (Participatory Configuration). A Participatory Configuration is a realization-specific structured representation of the Participatory Subjects and Participatory Acts observed in a case, preserving their Relational Grammar and any other pre-specified act or institutional attributes used for comparison rather than collapsing them into a single participation quantity.

Participatory Configuration is a data structure for comparison, not an omnibus causal variable. Its minimum required content is the set of observed subjects and acts plus the Relational Grammar of those acts. Additional fields—such as act form, presentation, recipient, authority, or channel properties—enter only when pre-specified for a particular empirical design. A later realization association must therefore be attributed to identified configuration features, not to the mere fact that two configurations differ somewhere.

Counts of speakers, comments, organizations, signatures, filings, or other domain-specific units remain useful descriptors. The restriction is that no universal participation-volume measure is presumed to summarize the configuration, and no post-hoc difference among its many fields can be treated as explanatory merely because an outcome also differs.

Configuration is also realization-specific. The same organization may occupy different positions across different deployments, facilities, policies, or institutional decisions. A union may negotiate one workplace deployment, oppose another, and remain absent from a third; a local government may possess binding authority over land use while having no meaningful authority over model design. The framework therefore does not assign actors a permanent participation type independent of the realization under study.

6.2 Definition 6 — Participatory Consequence

Definition 6 (Participatory Consequence). A Participatory Consequence is a change, persistence, or absence of detected change in a realization process that is temporally and substantively associated with a Participatory Act or Participatory Configuration.

The definition is intentionally broader than outcome change and more cautious than causal attribution. A participatory act can be followed by a procedural modification, a new information requirement, a change in organizational capacity, an altered allocation of authority, or a changed realized condition. It can also be followed by no detected change. The word associated therefore marks the first analytic threshold: an observed sequence or linkage is not yet proof that participation caused the consequence.

This boundary echoes the act/effect distinction discussed in Section 5. Austin’s separation between illocutionary acts and perlocutionary effects provides a structural antecedent: doing something and producing a further effect are analytically different even when they occur in one episode. The present framework generalizes that separation beyond speech. A refusal, purchase, lawsuit, institutional redesign, or organized withdrawal can be participatory without being verbal, and its consequences remain analytically distinct from the act itself.

6.3 Consequence classes

For conceptual purposes, at least seven consequence classes should be distinguishable. They are not ordered, mutually exclusive, or presumed exhaustive.

For exposition, informational, procedural, institutional, subject-readiness, and distributional consequences may function as intermediate pathways relative to direct realization-level change, while realization-level change and the null class are the endpoints most directly relevant to Proposition C. They remain separate consequence classes because their mechanisms, observation rules, and temporal behavior can differ; the grouping is explanatory, not a reclassification of the architecture.

Consequence class	What changes	Illustrative reading
Informational	What a consequential actor knows, records, or can observe	New evidence enters the record; a previously unseen burden becomes visible
Procedural	How an immediate decision is processed	Consultation is added; a hearing is reopened; review steps change
Institutional / return-path	The authority, procedure, membership, information flow, or rule-making process governing later decisions	A new review body or channel is created; future submissions gain a path to decision
Subject-readiness	The capacities of later potential Participatory Subjects	Education, training, organization, or experience increases later ability to formulate and undertake acts
Distributional	Who bears benefits, burdens, access, costs, or recourse	Compensation, access conditions, or burden allocation changes
Realization-level	The realized condition itself	Deployment proceeds differently, is conditioned, delayed, relocated, reduced, denied, or withdrawn
Null / no detected change	No observable element in the study window	The act is documented, but no uptake or other consequence is found

The classes resist a ladder interpretation. Informational change is not intrinsically weaker than procedural change, and institutional change is not automatically more consequential than a direct alteration of realization. Their significance depends on what the act was trying to do, what later decisions remain open, and what time horizon is used. A procedural consequence that creates a durable review requirement may matter more over time than a one-time condition attached to a single project; the reverse can also be true.

The null class is equally important. It preserves the denominator required for empirical work. If only visible successes are recorded, Participation becomes defined by uptake after the fact. A framework that permits participation without influence must also permit observed acts whose detected consequence is zero within the study window. Any empirical null must therefore be reported together with the observation window in which no consequence was detected.

6.4 No automatic conversion

Three non-equivalences follow from the architecture. First, a Participatory Act does not imply influence. Second, influence does not imply a change in the ultimate realized condition. Third, absence of immediate realization change does not imply absence of all Participatory Consequence. These distinctions prevent the framework from treating a direct policy reversal as the only meaningful endpoint.

The sequence is better understood as a set of possible transitions rather than a guaranteed chain. An act may reach no identifiable recipient. It may reach a recipient who lacks response capacity. It may enter a procedure after meaningful alternatives have closed. It may produce a response without changing a decision, or change a decision without producing the realization anticipated by the actor. Each transition can fail independently.

Power conditions these transitions without redefining the consequence classes. Institutional authority, credibility, agenda control, response obligations, channel ownership, and resource asymmetries can determine whether an act reaches a consequential recipient, how it is weighted, and whether any informational or procedural change can be converted into realization. These conditions should therefore be recorded separately from the grammar of the act and from the consequence ultimately observed.

This is why Section 5’s Conditional Non-Fungibility proposition is analytic rather than causal. Different grammatical profiles should not be presumed equivalent, but neither does the framework presume that grammatical difference will convert into consequence. That conversion is the empirical problem reserved for the companion study.

6.5 Second-order institutional consequence

Some participatory acts alter not the immediate realization but the machinery through which later realizations are selected, observed, or revised. The governance companion already supplies a rule for identifying this second-order structure: an intervention is second-order when its immediate target is the authority, procedure, membership, information flow, or rule-making process through which first-order interventions are chosen or revised. The present framework treats the participatory analogue as a consequence type rather than a new basic form of act.

A campaign that results in creation of a statutory hearing requirement, a university reform that replaces a slow review process with a continuously revisable one, or an earlier dispute that produces a standing information channel can therefore have an institutional or return-path consequence. The immediate realization may remain unchanged. What changes is the probability structure facing later acts: there is now a recipient, a procedure, a duty to receive information, a revised authority, or another route through which later participation may matter.

This consequence should not be called ‘second-order participation.’ That term is already occupied in administrative-law scholarship, and the phrase would also blur act and consequence. The participatory act remains an ordinary act under Definition 3. Its consequence is second-order because it changes the institutional conditions under which future acts are processed.

6.6 Subject-readiness consequence

A different recursive path operates through the Subject rather than the institution. A participatory act can alter the capacities of future Participatory Subjects. Education, training, collective organization, the production of accessible technical knowledge, or repeated experience with a decision process can change whether later subjects can recognize an issue, formulate a claim, and undertake an act.

This is why ‘readiness-building’ does not require its own theoretical category. The initial intervention is classified as a Participatory Act if it satisfies Definition 3. If its relevant result is that later subjects become more capable of participation, the consequence is a subject-readiness consequence under Definition 6. One intervention can of course produce both subject-readiness and institutional consequences—for example, by teaching new capacities while also redesigning the procedure through which those capacities can later be exercised.

No monotone effect is assumed. Information can overwhelm as well as inform; repeated participation can exhaust participants; institutional experience can produce withdrawal rather than competence; AI systems can both increase access to technical material and create new dependencies or information asymmetries. The framework therefore treats readiness change as an observable consequence, not as an automatically beneficial one.

6.7 Recursive consequences and path dependence

Once consequence can alter later subjects or institutions, participation becomes recursively structured. An act at one time can change the conditions under which later acts occur. Political-science work on policy feedback provides a close external antecedent: policies do not merely produce outcomes but can reshape later resources, incentives, organizations, interpretations, and political participation (Pierson 1993; Campbell 2012). The present framework applies the same general causal direction to capability-to-realization participation without assuming that every consequence becomes self-reinforcing.

Two recursive paths should remain separate. A subject-readiness consequence changes who can participate and with what capacities. An institutional or return-path consequence changes how later acts can enter decision processes. The paths can interact but are not identical. A population may become more capable while institutional access remains closed; an institution may create an open channel that few subjects are yet prepared to use.

Recursive consequences also make time horizon a substantive choice. A null result in a six-month observation window can coexist with later institutional or informational effects. Conversely, a highly visible immediate response can dissipate without changing later practice. Empirical work should therefore report the observation window and consequence class rather than infer long-run insignificance from short-run nulls or long-run durability from immediate uptake.

6.8 Consequence, attribution, and causal discipline

The definition of Participatory Consequence does not solve causal attribution. Realization processes are jointly shaped by law, infrastructure, commercial incentives, technical constraints, institutional routines, political pressure, litigation risk, public sentiment, and many other forces. A change following participation may therefore be coincident, jointly caused, strategically attributed, or independently motivated.

The conceptual framework requires only that consequence be kept distinct from act and that plausible linkages be recorded without silently converting sequence into causation. Strong causal claims require additional support: direct documentary attribution, temporal and procedural linkage, independent corroboration, or a design capable of distinguishing participation from competing explanations. The companion empirical paper can grade that support; the conceptual paper should not pre-decide it.

The same discipline applies to absence. Failure to locate a consequence is not proof that none occurred, just as failure to locate participation is not proof that no participation occurred. Search completeness, documentary visibility, and institutional transparency determine what can be observed. The framework therefore treats ‘no detected change’ as an evidentiary state rather than a metaphysical claim of no effect.

6.9 Boundary with participation-sensitive realization

Participatory Consequence is broader than a change in the societal Floor. Most consequences described in this section can occur without altering the realized state that CFS calls the Floor. The final step therefore requires a separate question: under what conditions, if any, can one or more pre-specified features of Participatory Configuration become sources of differences in realization itself?

That question must inherit rather than overwrite the existing Ceiling-Floor apparatus. CFS already treats the effective gate as emerging from interaction among actors with different threshold preferences and influence weights, and its empirical cases were consistent with a veto-max form in which the most restrictive sufficiently influential actor sets the gate. The paper also states that this form is provisional and was not tested against genuinely multilateral alternatives. The Participation extension begins from that actor-sensitive realization structure; it does not rediscover the fact that capability and realization can diverge.

The narrower extension, developed later in this paper under Responsive Floor, asks whether pre-specified features of participatory configuration can sometimes matter even when participating actors lack formal veto or binding authority. Section 6 does not assume that the answer is yes. It supplies the consequence vocabulary needed to ask the question without collapsing participation, influence, and realization into one variable.

References cited in this section

Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press. Revised edition, 1975.

Campbell, Andrea Louise. 2012. “Policy Makes Mass Politics.” Annual Review of Political Science 15: 333–351.

Pierson, Paul. 1993. “When Effect Becomes Cause: Policy Feedback and Political Change.” World Politics 45(4): 595–628.

W.H.L. et al. 2026. Ceiling, Floor and Slope: A Falsifiable Dynamical Model of Synchronization for Gradual AGI.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

7. Dynamic Participation

The Subject-Act-Consequence architecture is analytic, not static. Participatory Subjects acquire and lose capacities; roles move as tasks and institutions change; acts can alter the conditions of later acts; and consequences can become inputs into later realization processes. Dynamic Participation names this temporal structure without adding a fourth basic layer to the framework.

The point is already partly inherited from Gradual AGI. The Synchronization framework describes transformative adoption as reciprocal adaptation between expanding machine capability and bio-social humanity, with the latter adapting through institutional learning, educational reform, governance, professional practice, cultural evolution, and collective experience. The Optimization framework likewise treats humans, AI, institutions, infrastructure, and the broader social environment as a jointly evolving civilizational ecosystem. Participation therefore enters a series in which both human and technical roles were already expected to change over time. The present section asks how the human participant position behaves within that changing system.

7.1 Role change is not participant disappearance

Automation is often described at the level of tasks: a system acquires the capability to perform a function that previously required human labor. Participation is a different object. The disappearance of a task from a human role does not by itself establish the disappearance of human participation from the wider realization process in which that task sits.

Consider an educational setting. If an AI system can generate a competent solution, the human role may move from producing a first-pass answer toward framing problems, checking evidence, setting evaluation criteria, interpreting consequences, deciding acceptable use, redesigning assessment, or governing the conditions under which AI assistance is permitted. Whether any particular shift is desirable is a separate normative question. The conceptual point is that task location and participant location need not move together.

The distinction has strong antecedents. Research on meaningful human control argues that higher system autonomy can coexist with human control and responsibility when outcomes remain appropriately responsive to human reasons and traceable to humans in the design and operation chain. Responsibility-gap literature, conversely, shows why automation can make established forms of responsibility difficult to assign. These traditions already prevent a simple inference from task automation to human disappearance. The present framework applies that lesson to participation in capability-to-realization processes rather than claiming it as a new principle.

7.2 Proposition B — Participatory Role Plasticity

Proposition B (Participatory Role Plasticity). Increasing automation of particular human task functions does not by itself determine the elimination of human Participation from the associated realization process. Human participatory functions may instead be transformed, reassigned, or redistributed across subjects, stages, and institutional roles.

The proposition is derived and antecedent-rich. It does not claim that human participation always persists, that every automated task is replaced by a higher-level human task, or that institutions will successfully preserve meaningful human control. It rejects only the stronger inference that automation of a task function is sufficient evidence that human Participation in the surrounding realization has closed.

Three forms of change should be distinguished. Transformation occurs when the same human subject remains involved but performs a different participatory function. Reassignment occurs when a function moves from one participant to another, including from a human operational role to an AI system while a different human role remains elsewhere in the process. Redistribution occurs when the overall pattern of human participation changes across individuals, organizations, stages, or institutional locations. None of these implies conservation of the total amount of human participation.

This last limitation matters. Role plasticity is not a conservation law. Human participation can expand, contract, fragment, concentrate, or disappear in a particular realization. The proposition says only that the fate of one task role cannot settle that question by itself. Human participation can be structurally persistent while functionally discontinuous.

7.3 Transformation, reassignment, and redistribution

Change	What moves	What may remain	Illustrative reading
Transformation	Function of the same human participant	Subject identity	Teacher shifts from generating content toward designing and evaluating AI-assisted learning
Reassignment	A function from one actor to another	Participation elsewhere in the realization	AI performs routine analysis while humans retain approval, review, negotiation, or rule-setting roles
Redistribution	Pattern of participation across actors or stages	Some human participation, but not necessarily its previous locus or volume	Operational participation contracts while institutional, professional, or public participation expands

The three changes are descriptive, not normative categories. Transformation is not automatically preservation of agency; reassignment can move meaningful authority away from humans even while some nominal human role remains; redistribution can broaden participation or concentrate it. The framework therefore does not use continued human presence as a proxy for adequate control, answerability, fairness, or legitimacy.

This also keeps Participation distinct from answerability. An automated function can leave a human participatory role elsewhere in a system while responsibility for a harmful outcome remains unclear. Conversely, an institution can maintain a clear answerable human office even when large parts of the operational process are automated. Task allocation, Participation, control, and answerability should therefore travel as separate analytical variables rather than as substitutes for one another.

7.4 Subject-readiness feedback

A subject-readiness consequence becomes dynamically important when an earlier Participatory Act changes the capacities available to later subjects. Education, training, organizational learning, technical translation, and accumulated experience can all operate in this way.

The feedback can also be negative. Complex procedures can exhaust participants; technical systems can increase dependence on scarce expertise; repeated unsuccessful participation can encourage exit; abundant AI-generated information can make recognition harder rather than easier. The framework therefore admits readiness gain, readiness loss, and unresolved change rather than defining readiness formation as progress.

7.5 Institutional return-path feedback

The second recursive pathway changes institutions rather than subjects: earlier participation can create, modify, or remove the authority, procedure, membership, information flow, or rule-making process through which later participatory information reaches consequential decisions.

The distinction between subject and institutional feedback is analytically useful because either can vary without the other. Highly ready subjects may confront no consequential recipient or procedure. An institution may create an open channel that few potential participants know how to use. Participation becomes dynamically effective only through particular combinations of subject capacity, act structure, institutional route, and remaining decision opportunity.

7.6 Readiness and temporal opportunity

Rapid capability change creates a temporal question but not a theorem. Participatory Readiness is defined as a Subject-side capacity. Foreclosure Position is defined separately as the Act’s relation to the continued availability of substantively different realization choices. A subject can therefore be ready while the relevant opportunity has closed, or unready while it remains open.

The empirical question is whether those two clocks systematically diverge. Do human subjects and institutions acquire relevant capacities early enough to encounter consequential decisions while meaningful alternatives remain? Does the gap differ across education, labor, infrastructure, medicine, scientific practice, or frontier-model governance? Does it widen under capability acceleration or narrow as institutions learn? None of those answers follows from the definitions.

The August 2026 readiness cluster is useful as a motivating illustration rather than evidence: contemporaneous claims from educational institutions, public-policy advocates, and frontier-AI leaders made visible a perceived race between capability change and societal preparation. The conceptual framework extracts only the temporal problem. It does not infer from a forecasted capability threshold that a universal foreclosure date exists, because foreclosure remains realization- and decision-specific.

The infrastructure-siting Frame One is better suited to testing the Relational Grammar and consequence pathways than to resolving readiness-lag. It may supply readiness observations where the record supports them, but it does not claim to resolve the temporal question globally; broader readiness-lag tests may require domains with faster capability or institutional transitions, such as workforce adaptation or education.

7.7 Recursive participation without automatic self-reinforcement

Subject-readiness and institutional return-path feedback make Participation recursive, but recursion should not be mistaken for positive feedback. An earlier act can make later participation easier, harder, or merely different. A new procedure can widen access while increasing procedural complexity; educational reform can increase technical fluency while narrowing the kinds of judgment treated as legitimate; a successful campaign can institutionalize a channel or convince participants that further engagement is unnecessary.

Policy-feedback scholarship supplies the general antecedent: prior interventions can reshape later political resources, organizations, and interpretations. #9’s narrower claim is that Participatory Consequences can change later Subjects, channels, and Act opportunities within the capability-to-realization process.

For the same reason, the framework does not define a cumulative participation stock. Earlier participation leaves heterogeneous traces: knowledge, organizations, procedures, records, fatigue, precedent, exclusion, trust, distrust, or nothing detectable. Their effects need not be commensurable, and no single recursive state variable is introduced at the conceptual stage.

7.8 Dynamic boundaries

Dynamic Participation does not dissolve the boundaries established earlier. A changed human role remains distinct from an AI causal role. Readiness remains distinct from action. An institutional return path remains distinct from the act that created it. Participation remains distinct from influence, and influence remains distinct from realized outcome.

Nor does role movement resolve the normative questions surrounding automation. A realization can preserve human participation in a weak or ceremonial form while eliminating meaningful authority; another can reduce direct human operation while improving review, recourse, or responsibility. The framework records where and how humans participate before asking whether that configuration is desirable.

These boundaries prepare the transition to Section 8. Dynamic Participation explains how participant roles and conditions can change over time. Section 8 asks the separate question of whether those changing Participatory Configurations ever become realization-relevant enough to contribute to a Responsive Floor.

References cited in this section

Campbell, Andrea Louise. 2012. “Policy Makes Mass Politics.” Annual Review of Political Science 15: 333–351.

Matthias, Andreas. 2004. “The Responsibility Gap: Ascribing Responsibility for the Actions of Learning Automata.” Ethics and Information Technology 6(3): 175–183. https://doi.org/10.1007/s10676-004-3422-1.

Pierson, Paul. 1993. “When Effect Becomes Cause: Policy Feedback and Political Change.” World Politics 45(4): 595–628.

Santoni de Sio, Filippo, and Jeroen van den Hoven. 2018. “Meaningful Human Control over Autonomous Systems: A Philosophical Account.” Frontiers in Robotics and AI 5:15. https://doi.org/10.3389/frobt.2018.00015.

W.H.L. et al. 2026. Gradual AGI as Synchronization for Transformative Adoption. Gradual AGI Series #3.

W.H.L. et al. 2026. Gradual AGI as Optimization: A Conceptual Framework. Gradual AGI Series #5.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

8. Participation and Realization

The preceding sections define the Participatory Subject, the Participatory Act, and the range of consequences that may follow. This section asks the narrower question that reconnects Participation to the series’ central dynamical problem: can human participation become one condition of the realized societal Floor? The answer cannot be supplied by definition. The series already separates technical capability from realization and already allows actor configurations and governance arrangements to alter the mapping between them. The contribution of a Participation framework must therefore be narrower than restating that separation.

The relevant extension concerns actors who participate without necessarily possessing veto power, binding authority, or control of a formal governance instrument. If such participation can sometimes alter the path or configuration of realization, then the Floor is responsive not only to capability and formal gatekeepers but also, under some conditions, to participatory configuration. If it cannot, the participation-specific extension contracts while the inherited Ceiling-Floor and governance apparatus remains intact.

8.1 What is inherited

Section 2 establishes four inherited results that constrain this section: capability potential and realized distribution are distinct; governance acts on the mapping from potential to realization; exposure (reach, susceptibility, recourse) conditions realized outcomes; and CFS already makes the effective gate actor-sensitive. None is a Participation contribution.

The CFS qualification remains important: its veto-max cases were unilateral-restriction episodes and did not establish a universal aggregation mechanism. The ledger below therefore separates inherited actor sensitivity from the narrower participation-specific extension.

8.2 Inheritance and extension ledger

Claim	Status before #9	#9 disposition
Capability and realization are distinct	Inherited from CFS / governance	Use; do not claim as new
The potential-to-realization mapping can be modulated	Inherited from #7	Use; do not relabel as Responsive Floor
Exposure conditions realized distribution	Inherited from #8	Retain as separate realization condition
Actor configuration can affect the effective gate	Inherited from CFS Φ	Explicit ancestry for the extension
Participation is not influence or governing power	Inherited from #7	Required boundary condition
Non-veto, non-binding participation can sometimes affect realization	Not established by prior apparatus	Participation-specific extension; empirical effect unresolved
More participation produces more realization change	Not established	Rejected as a presumption
Participation always matters	Not established	Rejected; zero effect permitted

The ledger fixes the scope of the contribution. Responsive Floor cannot mean merely that Ceiling and Floor differ, that governance modulates realization, or that influential actors can set a gate. All three are already part of the inherited apparatus. The new question begins only where the prior theory becomes silent: whether participatory acts and configurations can enter realization even when the participating subjects do not themselves possess the formal power that CFS’s veto-oriented cases made most visible.

8.3 From consequence to realization

Section 6 deliberately defines Participatory Consequence more broadly than realization change. Information can enter a record, procedures can change, institutions can acquire new return paths, future subjects can become more ready, and distributions can shift without immediately changing the societal Floor. A theory of participation-sensitive realization therefore requires an additional conversion step.

That conversion may occur through several mechanisms, none of which is privileged in advance. Participatory information may alter what an authorized decision-maker knows. Organized participation may change the political or commercial costs facing an actor with formal authority. Market action may alter adoption. First-person evidence may change the factual record on which a decision is made. Earlier participation may create a new procedure that later becomes consequential. A collective may negotiate conditions without possessing unilateral veto. These routes differ sufficiently that the conceptual framework should not compress them into one universal mechanism.

The absence of a universal mechanism is not a defect. CFS itself treats Φ as an aggregation question rather than deriving a unique micro-mechanism, and #8 repeatedly finds that apparently scalar governance constructs decompose under empirical contact. Participation should enter the realization problem with the same discipline: identify observable routes, preserve their differences, and allow later evidence to determine whether a reduced representation is warranted.

8.4 Definition 7 — Responsive Floor

Definition 7 (Responsive Floor). A Responsive Floor is a societal Floor for which one or more pre-specified features of human Participatory Configuration are supported as contributing to differences in the path or configuration of realization, including through participation by subjects lacking formal veto or binding authority.

Responsive is descriptive rather than commendatory. A Floor may be responsive to participation that delays, accelerates, redistributes, or otherwise changes realization, including in ways participants themselves regard as harmful. The term therefore records supported dependence, not democratic quality or welfare.

The definition is deliberately evidentiary. A Floor is not called responsive merely because participation occurred around it. Responsiveness requires support that one or more pre-specified participatory features contributed to a realization difference. The definition therefore permits a large class of non-responsive cases: participation may be present, intense, early, and well organized while the realized Floor remains unchanged.

The phrase ‘contribute to’ is also narrower than sole determination. Participatory Configuration need not dominate capability, exposure, infrastructure, law, organizational constraints, or formal governance. Realization can be multiply caused. Responsive Floor identifies participation as one realization-relevant condition when the evidence warrants that attribution; it does not replace the rest of the realization mapping.

Finally, the definition extends rather than supersedes CFS. CFS’s actor-sensitive gate was evidenced principally through sufficiently influential actors capable of tightening or loosening access. Responsive Floor asks whether pre-specified features of human Participatory Configuration can sometimes matter below that threshold of formal control. If the only participation that ever changes realization is participation already backed by veto or binding authority, the participation-specific extension would add little beyond the inherited actor-gate apparatus.

8.5 Proposition C — Participation-Sensitive Realization

Proposition C (Participation-Sensitive Realization). Holding frontier capability and other relevant realization conditions sufficiently similar for comparison, differences in one or more pre-specified features of Participatory Configuration are admissible as possible sources of differences in realized societal Floors; the framework does not assume that this participatory dependence is universally present, nonzero, monotone, or large.

The proposition is an extension claim, not an empirical finding. ‘Admissible’ means that the theory does not restrict realization to capability, exposure, or formally empowered gatekeepers. The relevant participatory feature must be specified before the outcome comparison; it is not enough to observe that two rich configurations differ somewhere. Association under sufficiently comparable realization conditions is the first empirical test. Classification of a Floor as responsive requires additional support for contribution, graded by #10 rather than inferred from association alone.

The comparison condition is essential. A difference in realized Floor cannot be attributed to Participation merely because two cases also differ in capability, law, infrastructure, exposure, commercial conditions, or other major determinants. Nor does the proposition require literal identity of all non-participatory conditions, which would make empirical testing impossible. It requires sufficiently similar or explicitly adjusted comparison conditions for the claimed participation-specific relation to be meaningful.

The proposition also rejects monotonicity. More participants, more comments, greater visibility, or greater effort need not produce a larger realization effect. A small number of highly actor-dependent acts entering before meaningful foreclosure may matter more than a large volume of late generic submissions; the reverse may occur in another institutional setting. Proposition A therefore constrains how participation is compared, while Proposition C leaves consequence and realization effects open to observation.

8.6 Non-veto participation as the distinctive extension

The phrase ‘including participation by subjects lacking formal veto or binding authority’ carries the main theoretical burden of this section. Without it, Responsive Floor would risk collapsing back into CFS’s actor-sensitive gate. The relevant question is whether an actor can participate without governing and still become realization-relevant through some intermediate route.

The governance framework already makes the first half possible: participation is not influence, and a participant with negligible binding power participates without governing. Section 8 adds no contradiction. It asks whether some such acts can nonetheless alter the information, incentives, procedures, bargaining positions, legitimacy conditions, market responses, or other states encountered by actors who do possess consequential authority. If so, the participant does not become the governor; the realization process becomes indirectly responsive to participation.

This distinction also blocks a common attribution error. If a regulator imposes a condition after a hearing, the regulatory order is the formally binding act. The hearing participants should not be described as having exercised regulatory authority unless they actually possessed it. A participation-sensitive explanation instead asks whether particular participatory acts contributed to the informational or institutional conditions under which the regulator exercised that authority. The distinction between source of information, holder of authority, and final act of realization must remain visible.

A minimal non-veto micro-mechanism makes the boundary concrete. An affected resident submits first-person evidence of a deployment-related burden. The resident has no authority to alter the deployment. The evidence changes the informational state encountered by an authorized decision-maker, who then changes an operating condition. The binding act remains the decision-maker’s: Participatory Act → changed informational state → authorized decision → changed realization. The participation becomes realization-relevant without the participant acquiring veto or binding authority.

Evidence for that chain would require more than sequence. At minimum the study should identify the participatory act and recipient, establish that relevant alternatives were still available or later revisable, show that the information entered the decision process, document a decision change or reasoned uptake linked to the information, and distinguish the resulting realization change from independent legal, infrastructural, or commercial causes. Failure at one link weakens or breaks the claimed contribution pathway without erasing the underlying Participatory Act. Documentary support that information entered the decision process will ordinarily require an explicit trace, such as citation or discussion in staff reports, meeting minutes, responses to comments, revised plans, permit conditions, decision memoranda, or comparable records. Where no such trace is available, uptake remains a plausible but unconfirmed mechanism rather than an established link; #10 should report the fraction of candidate pathways with documented uptake instead of inferring entry from temporal sequence alone. The explicit-trace requirement is intentionally conservative: undocumented influence pathways may exist, but #10 should not count them as supported. A candidate pathway with no trace remains an unconfirmed mechanism rather than a counterexample to participation-sensitive realization.

8.7 Null, threshold, and reversal cases

A theory that admits participation-sensitive realization must also specify cases that do not support it. The clearest is the null case: Participatory Configuration differs while comparable realized Floors do not. Repeated null cases across settings in which participation had plausible routes to consequence would contract the domain of Responsive Floor. They would not refute CFS, the potential-to-realization distinction, or the existence of Participation itself.

Threshold effects are also possible. Participation may be realization-irrelevant below some institutional, informational, market, or political threshold and consequential above it. The framework does not currently specify such a threshold, and it should not invent one conceptually. Likewise, participation may have asymmetric effects: opposition may delay a realization while support fails to accelerate it, or early information may alter conditions while later information has no route to revision.

Reversal must remain possible as well. Participation can generate outcomes contrary to the actor’s stated orientation. A campaign may provoke acceleration, a proposed safeguard may create new burdens, or an attempted institutional reform may close rather than open later return paths. Responsive Floor concerns whether participation contributes to realization, not whether the resulting realization matches participant intent or improves welfare.

8.8 What does not count as evidence for Responsive Floor

Several observations are insufficient on their own. The mere presence of a hearing, petition, protest, consultation, or institutional process does not establish responsiveness. Temporal sequence alone does not establish contribution. Public sentiment does not become participation merely because decision-makers monitor it. A corporate or governmental response that cites ‘community concerns’ without identifiable participatory evidence may be suggestive but remains attributionally weak. Likewise, a change demanded by participants but independently required by law, grid constraints, litigation risk, or commercial incentives cannot be credited to participation without further support.

Topical examples must be handled with the same discipline. A proposal to reserve certain functions for humans illustrates the inherited distinction between technical capability and social realization. It becomes evidence for Responsive Floor only if the location of the human/automated boundary can be linked to participatory acts or configurations rather than to governance choice in the abstract. The conceptual distinction prevents contemporary examples from doing evidentiary work they have not earned.

8.9 Empirical handoff

The companion paper must therefore test a narrower claim than ‘participation matters.’ It must begin from an outcome-independent denominator and attempt to identify documented Participatory Acts whether or not they succeed. The first planned setting is large AI/high-density-compute infrastructure siting, where project populations and public administrative records can be assembled independently of controversy or outcome. Applications, staff reports, hearing materials, public comments, filings, minutes, permit conditions, appeals, and related records provide observable channels; private or undocumented participation remains a measurement limit rather than being coded absent.

The design uses outcome-blind act/channel coding before outcome review, three-valued participation status (Present / SearchedNotFound / Unknown), explicit search records, and independent coding of the grammar with agreement and unresolved cases reported. It must also permit the answer to be no. A null association, unreliable Actor Dependence coding, pervasive post-foreclosure participation, or outcomes explained more parsimoniously by formal authority, infrastructure, exposure, or commercial conditions would constrain the participation-specific extension rather than be rescued as indirect support.

That division of labor preserves the conceptual paper’s role. Section 8 does not estimate the responsiveness of any Floor. It defines the property, states the comparison claim that makes it testable, and identifies what would count as evidence for or against the participation-specific extension. Whether the extension survives empirical contact belongs to #10.

References cited in this section

W.H.L. et al. 2026. Ceiling, Floor and Slope: A Falsifiable Dynamical Model of Synchronization for Gradual AGI.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

9. Limits, Failure Conditions, and Revision Rules

A conceptual framework becomes difficult to revise when every disappointing result is redescribed as support. This paper therefore distinguishes three states in advance: a finding can falsify or contract a claim, fail to provide support without falsifying it, or fall outside what the conceptual paper asserts. The distinction is particularly important here because the framework contains objects of different epistemic status. The Relational Grammar is the paper’s principal candidate contribution; Participatory Role Plasticity is a derived proposition with substantial antecedents; Responsive Floor is a participation-specific extension of an inherited realization apparatus.

The section does not attempt to convert a conceptual paper into a fully estimated empirical model. Its purpose is narrower: state what later evidence would force the framework to change, what would merely delimit its domain, and what should not be protected by ad hoc reinterpretation. The companion paper inherits these conditions rather than inventing success criteria after observing the cases.

9.1 Limits of scope

The framework is intentionally human-centered. Unqualified Participation refers to intentional action by human persons or human-organized collectivities. This leaves open rather than resolves questions about artificial moral agency, artificial responsibility, machine standing, or future non-human participation. Evidence that a sufficiently advanced artificial system should be treated as a participant under another normative or metaphysical theory would therefore challenge the scope choice only if #9 claimed that such participation was impossible in principle. It does not.

The framework is also realization-specific. It does not offer a general theory of all political participation, civic engagement, organizational voice, or human-machine interaction. Its constructs are designed for acts directed toward a prospective, ongoing, or realized condition of Gradual AGI and toward the conditions under which such realization is selected or revised. Uses outside that domain may be informative without being covered by the paper’s validity claim.

Finally, the framework is not a welfare theory. A participatory act can be consequential and harmful; an unresponsive process can sometimes produce a desirable outcome; a highly actor-dependent claim can be false; and broad participation can increase decision costs. Nothing in the Subject-Act-Consequence architecture establishes that more Participation is normatively superior.

9.2 What would contract the Relational Grammar

The strongest analytic-contribution claim concerns the joint use of Realization Orientation, Foreclosure Position, and Actor Dependence at the level of the Participatory Act. That claim should contract if any of three conditions is met.

First, a close antecedent may be located that already supplies the same three relations, or functionally equivalent relations, for the same unit of analysis and purpose. Such a finding would not make the grammar useless, but it would change its status from candidate contribution to adaptation or application. The manuscript should then say so directly rather than retain a priority claim through narrower wording.

Second, one of the dimensions may prove conceptually redundant. If all distinctions made by Actor Dependence, for example, can be recovered from another field already required by the framework, retaining it would add symmetry rather than information. The same rule applies to Realization Orientation and Foreclosure Position. Empirical correlation alone is insufficient for retirement; deterministic or practically complete recoverability is the relevant problem.

Third, a dimension may fail reliable operationalization. A concept that readers can understand but independent coders cannot distinguish with acceptable consistency may remain philosophically suggestive while losing status as an empirical primitive. The response should be to refine, merge, demote, or retire the field, not to preserve it by allowing case-by-case intuition to substitute for a decision rule.

9.3 Failure conditions for Proposition A

Proposition A fails as a useful comparison rule if visible form, institutional channel, and ordinary descriptive fields already recover every distinction that the Relational Grammar is claimed to preserve, or if grammar differences cannot be identified with sufficient reliability to change how acts are compared.

Conditional Non-Fungibility is not falsified merely because differently structured acts sometimes produce the same consequence. The proposition explicitly avoids that causal claim. Its burden is comparative: equality of form, channel, or numerical volume should be insufficient to establish equivalence when material grammatical differences exist.

Accordingly, the strongest failure would be informational redundancy. If a public-comment label plus conventional metadata already captures orientation, decision openness, and actor dependence with no lost distinctions, the grammar contributes no additional comparative information. The empirical companion should be permitted to reach that conclusion.

The proposition should also contract if materiality cannot be specified. A framework that treats every minor coding difference as sufficient to defeat equivalence would make aggregation impossible by construction. #10 therefore needs rules for when differences are analytically material rather than merely present.

9.4 Failure conditions for Participatory Readiness

Participatory Readiness is a Subject-side conditioning property rather than a principal novelty claim, but it still carries an empirical burden. The construct becomes weak if it cannot be measured independently of observed Participation. If researchers infer that a subject was ready solely because the subject acted, and unready solely because the subject did not, Readiness becomes a circular relabeling of the outcome.

Operationalization must therefore use evidence of capacity that can in principle be observed whether or not an act follows: knowledge, access to relevant information, civic or technical skills, organizational resources, procedural competence, or other domain-specific indicators justified in advance. None is assumed to be universally necessary or sufficient.

The framework should also abandon any general readiness-lag claim if empirical work shows no stable relation between capability acceleration and the timing of human or institutional preparedness. Readiness may lag in one domain, lead in another, or fluctuate without a common direction. The conceptual architecture survives such heterogeneity because it separates readiness from temporal opportunity rather than specifying the sign of their relation.

9.5 Failure and contraction of Participatory Role Plasticity

Proposition B is intentionally weak: automation of a human task does not by itself determine elimination of human Participation from the surrounding realization process. It would be contradicted only by evidence or argument establishing the stronger entailment that task automation necessarily closes the relevant human participant position rather than merely moving, shrinking, or eliminating it contingently.

More realistic evidence will delimit rather than falsify the proposition. Some domains may show genuine disappearance of human participation; others may show transformation, reassignment, or redistribution. A finding that human presence persists only ceremonially would also require care: nominal participation is not equivalent to influence, answerability, or meaningful control.

Because the proposition is antecedent-rich, the manuscript should resist treating empirical support as proof of originality. Its role is to preserve a distinction needed by the Participation framework, not to establish priority over automation, meaningful-human-control, or responsibility-gap literatures.

9.6 Failure conditions for Responsive Floor

Responsive Floor contracts if pre-specified participatory features fail to contribute detectable explanatory value for realization once capability, exposure, formal authority, infrastructure, commercial conditions, and other relevant determinants are considered.

This condition is narrower than showing that Participation sometimes has no effect. The definition already permits non-responsive cases. What would seriously weaken the extension is repeated evidence that apparent participation effects disappear when formally empowered actors or other realization conditions are accounted for, especially in cases where non-veto participation had a plausible route to consequence.

A stronger contraction occurs if every documented participation-sensitive realization can be redescribed entirely as an exercise of formal veto or binding authority by the same participating actor. In that case, #9 would have added little beyond CFS’s existing actor-sensitive gate. The distinctive extension depends on at least some realization-relevant pathways through which subjects lacking such authority can contribute without themselves becoming the formal governor.

Null results must be retained. If an outcome-independent sample contains many instances of Present participation but no detectable realization difference relative to comparable SearchedNotFound cases, that result should be reported as evidence limiting participation-sensitive realization. It must not be rescued by broadening ‘consequence’ after the fact unless the alternate consequence type was specified and observed independently of the outcome.

9.7 What would not falsify the framework

Several results may be theoretically important without falsifying the core architecture. They should be distinguished in advance from true failure conditions.

Possible finding	Effect on framework	Reason
Many Participatory Acts have no detected consequence	No falsification	Act and consequence are defined separately
Participation often occurs after meaningful foreclosure	Empirical finding; may limit effectiveness	Foreclosure Position permits post-foreclosure acts
High-volume participation fails to change realization	No falsification	No monotone volume-effect relation is assumed
AI automates many human task functions	No falsification of Participation by itself	Task location and participant location are distinct
Grammar dimensions are empirically correlated	No falsification	They are claimed analytically separable, not orthogonal
One domain shows no Responsive Floor	Domain contraction only	Responsiveness is not universal by definition
Participants obtain outcomes contrary to their stated aims	No falsification	Consequence need not match orientation or intent
An institution is contestable but no participatory act occurs	No falsification	Opportunity/standing and action are separate

9.8 Measurement and observability limits

The framework distinguishes conceptual absence from observational absence. Participation may be undocumented, records may be incomplete, informal acts may leave no trace, and institutions may obscure how information moved internally. The empirical companion must therefore distinguish Present, SearchedNotFound, and Unknown rather than treating non-observation as proof of absence.

The same problem applies to consequence. A null observation in a bounded study window may coexist with later informational, institutional, or realization effects. Conversely, a public statement of responsiveness may not establish that the act caused the decision. Observation windows, source completeness, and causal-support levels must be reported with the result.

Actor Dependence is especially exposed to retrospective interpretation. Coders who know the outcome may overvalue first-person testimony that later became salient or understate generic submissions that proved decisive. Outcome-blind coding of act structure is therefore preferable wherever the source record permits it.

Foreclosure Position also requires discipline. Formal decision dates are imperfect proxies for substantive openness. Contracts, sunk costs, prior permits, technical lock-in, organizational commitments, or legal obligations can close alternatives earlier; later review rights can sometimes reopen them. #10 must operationalize meaningful alternatives rather than substitute chronology alone.

9.9 Causal limits

#9 does not identify a universal causal coefficient of Participation. Participatory acts enter systems jointly shaped by law, markets, infrastructure, institutional routines, technical feasibility, strategic behavior, media attention, and exposure. Association between a Participatory Configuration and an outcome is therefore not sufficient to establish causal contribution.

The companion paper should grade causal support rather than force every case into a causal/non-causal binary. Direct documentary attribution, traceable procedural linkage, timing consistent with influence, decision-maker testimony, contemporaneous records, and elimination of obvious competing explanations can strengthen inference. None creates identification automatically.

Selection into observability is another limit. Public controversies are easier to study than quiet processes, successful campaigns often leave richer records than unsuccessful ones, and known-outcome cases can contaminate sampling. The confirmatory design should therefore preserve an outcome-independent denominator and keep exploratory famous cases outside the primary estimate.

9.10 Revision hierarchy

When evidence conflicts with the framework, revision should proceed at the lowest level sufficient to repair the problem. The following hierarchy prevents a local measurement failure from triggering unnecessary theoretical reconstruction, while also preventing a central failure from being hidden inside minor wording changes.

Level	Trigger	Default response
Coding rule	Ambiguous boundary or low inter-coder consistency	Clarify rule or examples; do not alter construct yet
Dimension	Persistent redundancy or non-codability	Refine, merge, demote, or retire dimension
Definition	Object cannot be distinguished from adjacent construct	Rewrite definition and propagate consequences
Proposition	Claim repeatedly fails while definitions remain coherent	Narrow, condition, or reject proposition
Architecture	Subject, Act, and Consequence cannot be maintained as distinct analytic levels	Reconstruct framework
Series extension	Participation adds no realization-relevant information beyond inherited apparatus	Retire or sharply narrow Responsive Floor extension

9.11 No symmetry preservation rule

The framework should not preserve three grammar dimensions merely because three are aesthetically convenient. Nor should it preserve three principles, three analytic levels, or a fixed number of propositions if later work shows that a distinction is redundant or a missing distinction is necessary. The present architecture is a disciplined reconstruction, not a commitment to numerical symmetry.

This rule is especially important for the Relational Grammar. If Actor Dependence fails while Realization Orientation and Foreclosure Position survive, the result should be a two-relation grammar unless a new independently justified dimension replaces it. If a fourth relation repeatedly distinguishes acts in ways unrecoverable from the current three, the framework should expand. The admission criterion is discriminatory work, not formal balance.

Operationally, a grammar dimension should be retired when, after reasonable instrument refinement, either independent coders cannot assign it with acceptable reliability or it adds no distinction that is not recoverable from the surviving grammar dimensions plus ordinary metadata. Empirical correlation alone is insufficient; the test is whether retaining the dimension prevents a meaningful loss of distinction.

9.12 Chronology and anti-rescue rule

Conceptual revision must also preserve chronology. Definitions and decision rules fixed before the confirmatory empirical exercise should remain identifiable as such. Later revisions can be made, but results should not be recomputed under the revised framework and then presented as if the new rule had governed the original test. Historical disagreement and failed classifications are evidence about the instrument.

Accordingly, a failed test should not be rescued by broadening the outcome, changing the comparison class, reclassifying participation after reading the consequence, or adding a new dimension chosen because it fits the observed cases. Such changes may generate a new version and a new test. They do not erase the result that forced the revision.

This anti-rescue rule is the final limit on the conceptual paper itself. #9 is useful only if it constrains #10 before the evidence arrives. If every empirical result can be absorbed by reinterpretation, the framework has become descriptive vocabulary rather than a testable research program.

References cited in this section

Collingridge, David. 1980. The Social Control of Technology. New York: St. Martin’s Press.

Matthias, Andreas. 2004. “The Responsibility Gap: Ascribing Responsibility for the Actions of Learning Automata.” Ethics and Information Technology 6(3): 175–183.

Santoni de Sio, Filippo, and Jeroen van den Hoven. 2018. “Meaningful Human Control over Autonomous Systems: A Philosophical Account.” Frontiers in Robotics and AI 5:15.

W.H.L. et al. 2026. Ceiling, Floor and Slope: A Falsifiable Dynamical Model of Synchronization for Gradual AGI.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.

10. Handoff to #10

This paper fixes a conceptual vocabulary; it does not claim that the vocabulary has already survived empirical contact. The companion installment must therefore begin where #9 stops: by turning Participatory Subject, Participatory Readiness, Participatory Act, Relational Grammar, Participatory Configuration, Participatory Consequence, and Responsive Floor into observable objects with explicit decision rules.

The division of labor is methodological rather than stylistic. #9 defines what counts, what must remain distinct, and which claims are admissible. #10 must determine whether those distinctions can be coded reliably, whether the proposed relations are empirically useful, and whether Participation adds explanatory information beyond capability, exposure, formal authority, and other already inherited realization conditions.

10.1 What #9 fixes before the evidence

Object	Fixed in #9	Not fixed in #9
Participatory Subject	Human person or human-organized collective under the paper’s default scope	Which populations are empirically ready or active in each domain
Participatory Readiness	Capacity distinct from action and temporal opportunity	Universal indicators or threshold
Participatory Act	Intentional realization-directed action	Complete taxonomy of forms
Relational Grammar	Realization Orientation, Foreclosure Position, Actor Dependence	Exhaustiveness, empirical independence, universal weights
Participatory Configuration	Realization-specific structured representation of observed subjects and acts, preserving grammar and pre-specified comparison fields	Single scalar index or post-hoc omnibus causal variable
Participatory Consequence	Includes informational, procedural, institutional, readiness, distributional, realization-level, and null states; consequences may also be delayed	Universal ordering of consequence classes
Responsive Floor	Pre-specified participatory features may sometimes contribute to realization, including without veto/binding authority	Magnitude, sign, prevalence, universal mechanism, or causal coefficient

10.2 Primary empirical tasks

The first task is measurement validity. Can independent coders identify documented Participatory Acts without reading outcomes into the classification? Can Realization Orientation, Foreclosure Position, and Actor Dependence be coded with sufficiently stable rules and reported disagreement? Do these dimensions preserve distinctions that ordinary act labels and institutional metadata would otherwise lose?

The second task is temporal and institutional placement. For each observed act, #10 must ask whether meaningful alternatives remained, which recipient or channel received the act, whether that recipient had response capacity, whether any formal duty to receive or respond existed, and whether the relevant authority was binding, advisory, or absent. These fields should remain separate rather than being collapsed into a single channel-quality score.

The third task is consequence. The companion paper must distinguish communicative uptake from substantive change, immediate consequences from later ones, and realization-level change from informational, procedural, institutional, or readiness effects. It must permit no detected consequence as a valid result.

The fourth task is comparison. The empirical study must ask whether pre-specified features of Participatory Configuration correspond to differences in realization under sufficiently comparable capability and other realization conditions. The claim is not causal merely because an association appears; competing explanations must be preserved.

10.3 Frame One and the silent denominator

The planned Frame One design supplies the first confirmatory setting in large AI/high-density-compute infrastructure siting. Its central protection is an outcome-independent project population assembled from official land-use and permitting sources: cases enter because they satisfy ex ante project criteria, not because they became controversial, attracted visible participation, or produced a known outcome.

Participation status therefore remains three-valued: Present, SearchedNotFound, and Unknown. The design can identify only documented participation within the searched record. ‘Not found’ after a documented search is not proof of absence, private or informal acts may remain invisible, and Unknown is required when source coverage is insufficient. The framework’s empirical demand is outcome-independent search, not impossible omniscience.

The pilot remains outside the primary confirmatory estimate. Its purpose is to test acquisition, coding rules, and workflow, not to contribute observations after the instrument has already been adjusted to fit them. Known-outcome exploratory cases should likewise remain separate from the confirmatory denominator.

Frame One is selected partly because its public-record density permits an outcome-independent project population and observable decision trajectories. A successful instrument or association test in infrastructure siting would therefore be domain-specific evidence, not universal validation of the framework. Generalizability to settings such as workplace AI deployment, public-sector algorithmic decision-making, or frontier-model governance requires separate empirical tests with their own observation rules and denominators.

10.4 Outcome-blind sequencing

The coding sequence should preserve the architecture. Packet A establishes the project and channel history, act form, recipient, institutional properties, and—where the record permits—the Relational Grammar before coders inspect later outcomes. Packet B then introduces uptake, consequences, and realized outcome. A separate annotation guide supplies decision rules and edge cases; independent coding and reported agreement test whether the grammar is usable rather than assuming that it is.

This sequencing is particularly important for Actor Dependence and Foreclosure Position. Knowledge that a particular testimony later became decisive can make its actor-specific content appear more obvious than it was ex ante; knowledge that a project failed can make earlier procedural events appear more open or more constraining than the contemporaneous record supports. Outcome-blind coding will not eliminate judgment, but it reduces retrospective reconstruction.

10.5 Confirmatory versus exploratory claims

Question	Status	Reason
Can the three grammar dimensions be coded reliably?	Confirmatory instrument test	Central to Proposition A
Do grammar differences add distinctions beyond form/channel/count?	Confirmatory comparison test	Direct test of Conditional Non-Fungibility’s usefulness
Is Present participation associated with different realization outcomes than SearchedNotFound?	Confirmatory association test	Bears on Responsive Floor without proving causality
Which specific pathways convert participation into consequence?	Exploratory / mechanism-building unless pre-specified	Multiple routes are currently admissible
Does readiness lag capability in a domain?	Empirical; domain-specific	Not implied by definitions
Does participation increase detection/correction rate?	Deferred unless instrument supports rate measurement	Prior governance rate claim remains untested
What exact functional form extends CFS Φ?	Formal companion task after empirical contact	No parallel equation is fixed in #9

10.6 What #10 must be allowed to find

The companion paper must be able to conclude that the Relational Grammar is partly redundant, that one dimension is not codable, that participation is frequently post-foreclosure, that many acts have null consequences, or that observed realization differences are better explained by formal authority, exposure, infrastructure, law, or commercial conditions.

It must also be able to conclude that non-veto participation adds no detectable realization information in the confirmatory sample. Such a result would contract Responsive Floor as a participation-specific extension without invalidating the inherited Ceiling-Floor distinction or the existence of Participation as an analytic object.

No null result should be converted into support by broadening the outcome after inspection, counting any later institutional trace as success, or moving a case from SearchedNotFound to Present on weak circumstantial evidence. The anti-rescue rule in Section 9 applies directly to #10.

10.7 What #10 should not measure as one number

Several constructs should remain decomposed unless empirical work later earns an aggregation rule. Participation itself should not be assigned a universal volume score. Channel quality should not collapse recipient identity, response capacity, formal duty to receive, response obligation, binding authority, channel control, reliability, and timing into one ordinal rank. Consequence classes should not be placed on a universal welfare ladder. Exposure should remain reach, susceptibility, and recourse rather than a scalar.

The same rule applies to the grammar. Realization Orientation, Foreclosure Position, and Actor Dependence can be coded jointly, but #9 does not define a weighted Participatory Act score. If #10 later proposes one, it must show why the aggregation preserves rather than destroys the distinctions that motivated the grammar.

10.8 Formalization handoff

The conceptual paper introduces no new equation for participation-sensitive realization. That restraint is intentional. CFS already contains the actor-aggregation mechanism Φ, and #8 already supplies the potential-to-realization mapping under exposure and governance. A new function written independently in #9 would risk duplicating or contradicting the inherited apparatus.

The formal task for #10 is therefore an extension problem: determine, after the empirical objects are sufficiently stable, whether Participation belongs inside the aggregation of actor influence, as a modifier of effective thresholds, as an input into exposure or recourse, as a distinct pathway to institutional change, or through more than one mechanism. The formalization should be chosen to represent the evidence, not selected in advance because the notation is convenient.

If no participation-specific formal term is warranted after measurement, that is an admissible result. The conceptual contribution of #9 would then remain primarily the act-level grammar and analytic separation rather than a new dynamical variable.

10.9 Proposition-level handoff

Proposition	#9 status	#10 task	Admissible negative result
A — Conditional Non-Fungibility	Analytic proposition based on Relational Grammar	Test whether grammar adds stable, nonredundant distinctions	Dimensions redundant, non-codable, or no added information
B — Participatory Role Plasticity	Derived, antecedent-rich proposition	Document whether roles transform, reassign, redistribute, or disappear in selected domains	Domain shows actual disappearance of human participation
C — Participation-Sensitive Realization	Series-extension proposition	Test pre-specified configuration features for association and graded causal support with realized Floor	No detectable participation-specific realization effect

10.10 Sequence of the companion paper

The companion should proceed in the same epistemic order as the framework: instrument before result, coding validity before causal interpretation, and primary denominator before known-outcome cases. A practical sequence is: sampling and acquisition protocol; coding and reliability; descriptive distribution of Participatory Acts and grammar; channel and consequence analysis; realization comparison; mechanism cases; formal extension; proposition-level assessment.

That order preserves the possibility that the instrument itself fails. If Actor Dependence cannot be coded, the paper should say so before presenting outcome associations as if the full grammar had been measured. If participation records are too incomplete to distinguish SearchedNotFound from Unknown, the denominator problem should be reported before estimating a participation-outcome relation.

The purpose of the companion is therefore not to ‘prove’ #9. It is to determine which parts survive contact with records, cases, and comparison.

10.11 Conclusion and division of labor

#9 answers three conceptual questions: who counts as a Participatory Subject under the paper’s scope, what distinguishes one Participatory Act from another, and what classes of Participatory Consequence must be represented before realization is inferred. It also states one deliberately conditional extension to the Floor.

#10 must answer whether those objects are observable, reliably codable, nonredundant, and realization-relevant once inherited determinants are respected. A clear negative answer to any of those questions is a result, not a failure of the research program.

The handoff is therefore intentionally asymmetric: #9 constrains what #10 is allowed to claim; #10 is allowed to revise what #9 proposed.

The central participation problem of Gradual AGI is therefore not whether humans remain present somewhere in an AI-mediated society. It is how human subjects retain, transform, and exercise capacities to act upon realization—and under what conditions those acts become consequential.

References cited in this section

W.H.L. et al. 2026. Ceiling, Floor and Slope: A Falsifiable Dynamical Model of Synchronization for Gradual AGI.

W.H.L. et al. 2026. Gradual AGI as Contestation: A Framework for Governance. Gradual AGI Series #7.

W.H.L. et al. 2026. Gradual AGI as Contestation: Formal Models and Empirical Tests of Governance. Gradual AGI Series #8.