Framework State of the Commons

Vital Intelligence Model · Bridging Spiral SR2 · TKGPT Design Foundation

Humanity++ · Karen Doore · March 2026 Draft for partnership conversations and GitBook documentation


Purpose of This Document

This document serves as a navigation instrument for the TKGPT project — a proposed AI literacy companion grounded in the Vital Intelligence Model (VIM) and Bridging Spiral framework. It has three functions:

  1. Epistemic map — what the framework claims, at what confidence level, and on what evidence

  2. Architectural rationale — why specific technical concepts from distributed computing, operations research, and information theory are load-bearing in the framework, not decorative

  3. Design constraints — what the TKGPT can and cannot do, derived from the MPCM boundary and the framework's own claims about human learning

It is written for two audiences simultaneously: academic and research partners encountering the framework for the first time, and practitioner communities (DEAL, AI governance, education) who need to evaluate its practical claims. The framework's epistemic architecture is itself an expression of its values — Ostrom's Principle 4 (monitoring) applied to the framework's own knowledge claims.


Part I: The Core Claim and Its Evidence Base

The Problem the Framework Addresses

The human information environment has undergone a phase transition. AI systems now mediate the majority of information flows in economically developed societies at speeds, scales, and levels of personalization that exceed the adaptive capacity of most existing educational frameworks. The result is a widening gap between the technical sophistication of information systems and the cognitive, somatic, and relational capacities humans require to engage with those systems wisely.

This gap is not primarily a technical literacy problem. It is a mental model problem: most people — including most educators and policymakers — are navigating a dynamically complex, distributed, adversarial information environment using cognitive frameworks optimized for stable, centralized, high-trust conditions. The mismatch is structural, not individual.

The V-Dem Institute's Democracy Report 2026 provides the most recent empirical anchor for the macro-scale consequences of this mismatch. As of 2025: global democracy has returned to 1978 levels, effectively erasing the gains of the third wave of democratization that began with Portugal's Carnation Revolution in 1974. Seventy-four percent of the world population now lives in autocracies. Freedom of expression is the most attacked democratic indicator, declining in 44 countries. Media censorship is the most common autocratization tactic, deployed by 73% of autocratizing governments. The United States has experienced the most rapid executive aggrandizement in modern history, with legislative constraints reaching their lowest point in over 100 years.

This is not background context for the framework. It is what the framework's central metaphor — the Giant Pumpkin attractor — looks like at civilizational scale when the values vector has been pointing inward across AI-amplified information systems for a sustained period. The Ising model's frozen ordered phase made geopolitical.

The framework's claim: this trajectory is not inevitable. It is a design outcome, produced by specific architectural choices in information system design, and it can be redesigned. The educational precondition for redesign is the development of navigational capacities — cognitive, somatic, and relational — that allow individuals and communities to distinguish generative information dynamics from extractive ones, and to act from the distinction.

The Central Architectural Claim

Humans have been solving the problem of reliable decision-making under stress in distributed systems with faulty components for decades — not in psychology or education, but in computer science, operations research, and distributed systems engineering. The framework's core claim is that this formally validated knowledge has never been systematically translated into a cognitive and social literacy framework, and that this translation is both possible and necessary.

This is not metaphor. It is structural correspondence.

The Byzantine Generals Problem (Lamport, Shostak & Pease, 1982) establishes formally that: to reach reliable consensus in a distributed system where up to f components may be sending corrupted messages, a minimum of 3f + 1 independent nodes is required. With fewer nodes, corrupted signals cannot be distinguished from authentic ones — the system cannot achieve fault-tolerant consensus regardless of the quality of its algorithms.

Applied to human cognition: rationalization under stress is not an occasional failure of character. It is a structurally predictable behavior of the cognitive system under cortisol-driven regression (PSI Theory: Kuhl et al., 2020) — the equivalent of a Byzantine faulty node, generating internally consistent but corrupted messages that serve the threat-response attractor rather than genuine inquiry. The four instruments of the VIM framework — ♠ Somatic Gyroscope, ♦ Cognitive Radar, ♥ Relational Compass, ♣ Dimensional Integration — are the minimum viable panel for fault-tolerant human discernment. With fewer than four independent signal sources, a single captured instrument cannot be identified and isolated. Single-channel certainty under stress is a diagnostic signal of Byzantine failure, not a virtue.


Part II: Epistemic Tiering

The framework draws on a wide range of sources with varying degrees of empirical validation. Being explicit about this is an Ostrom Principle 4 commitment — the framework monitors its own epistemic status.

Tier 1 — Empirically Grounded

Peer-reviewed, replicated, or established through sustained empirical research programs.

Domain
Key Claims
Primary Sources

Commons governance

Eight design principles for sustainable commons management

Ostrom (1990); Wilson & Ostrom (2019)

Multilevel selection

Group-level selection conditions outcompete individual-level dynamics when Ostrom principles are implemented

Wilson (2019)

Transformative learning

Perspective transformation is rare, requires disorienting dilemma + safety conditions

Mezirow (1991, 2000)

Active inference

Living systems minimize prediction error through model-updating or environmental control

Friston (2010, 2013)

PSI Theory

Stress-driven regression suppresses higher-order motivational systems

Kuhl et al. (2020)

Self-organized criticality

Systems at the critical state exhibit maximum sensitivity to perturbation and cascade dynamics of any magnitude

Bak (1996)

Democracy data

Global democratic backsliding documented across 202 countries 1789–2025

V-Dem Institute (2026)

Cardiac autonomic dynamics

High MAD of HRV during creative engagement predicts anxiety reduction and creative output

Bellaiche et al. (2025)

Relational neuroscience

Compassionate concern (vs. empathic resonance) increases with contemplative practice; neuroplastic

Singer (2025)

Byzantine fault tolerance

3f+1 minimum node requirement for fault-tolerant consensus

Lamport et al. (1982)

Tier 2 — Theoretically Coherent

Formally argued, internally consistent, not yet fully empirically validated or in active peer debate.

Domain
Key Claims
Primary Sources
Notes

Natural Born Intelligence

Two-layer engine (co-applicability + meta-level attenuation) distinguishes biological from artificial intelligence

Gunji (2025)

Published in Biosystems; physiological signatures empirically supported

Thousand Brains Theory

Neocortex as heterarchy of semi-independent cortical columns, each building complete world models

Hawkins et al. (2025)

Preprint; theoretically significant, empirically ongoing

Punctuated geometry

Holonomy as measure of genuine vs. apparent transformation; crack mechanism

Smarandache (2026)

Used as design metaphor with explicit epistemic caution, not biological claim

Causal emergence

Macro-level organization can have greater causal power than micro-level components

Hoel et al. (2013)

PNAS publication; ongoing debate in philosophy of science

Ego-development stages

Developmental arc from egocentric through ecocentric world models

Kegan (1994); Cook-Greuter (2004)

Empirically supported in adult development research; stage models contested in detail

Tier 3 — Speculative / Generative

Productive hypotheses whose epistemic status is explicitly uncertain. Included because they generate useful questions, not because they are established.

Domain
Key Claims
Sources
Epistemic Flag

Toroidal consciousness models

Toroidal bipolaron energy flows as substrate of consciousness and coherent field effects

Meijer & Geesink (multiple)

Not mainstream; draws on real physics but makes consciousness claims that exceed current empirical consensus. Useful as generative metaphor for toroidal DIKW simulation design. Do not cite as established.

Edge of chaos and consciousness

SOC as substrate of conscious experience

Sbitnev (2024)

Preprint; speculative extension of SOC theory. Interesting, not validated.

Technosocial phase transformation

Current global information system is approaching a critical transition point analogous to thermodynamic phase transition

Framework synthesis

Theoretical claim grounded in convergent analogs, not directly empirically demonstrated. Honest framing: this is a working hypothesis, not a finding.

Tier 4 — Researcher Positionality

The framework has been developed through sustained cross-domain research, somatic healing practice, and iterative educational design over multiple years. This developmental history is part of the provenance chain — it shaped which connections were made and which analogies were pursued. It is documented in the Humanity++ repository and Substack archive. It is present implicitly throughout the framework as a shaping force, not cited as evidence for specific claims.


Part III: State Machines as the Cognitive Bridge

Why State Machines Are Load-Bearing

A finite state machine (FSM) is a formal model consisting of: a finite set of states, a set of transitions between states triggered by defined inputs, and an initial state. Every component has a named function. The machine's behavior at any moment is entirely determined by its current state and the input it receives.

FSMs are not merely a useful analogy. They are the minimal formal structure that makes three critical AI literacy capacities simultaneously trainable:

1. Present-moment state awareness The FSM's most important contribution to cognitive literacy: the question what state am I in right now? Before any other analysis, the learner must read their current state register. This is somatic awareness made formal — the ♠ Somatic Gyroscope as state detector. In an FSM, behavior is fully determined by current state plus input. In human cognition: the same information, received in a threat state versus a regulated state, produces categorically different outputs. Teaching learners to ask "what is my current state?" before engaging with information is the FSM insight applied to self-regulation. It is also the entry point to the DIKW stack — you cannot locate yourself on the stack without first reading your state.

2. Counterfactual thinking Running the state machine backward: what state would we be in if a different transition had fired? This is the formal definition of counterfactual reasoning — and it is the cognitive capacity most directly suppressed by threat-state tunnel vision (PSI regression). Under stress, the Cognitive Radar operates in error-detection mode, scanning the current state for threats rather than modeling alternative state trajectories. Counterfactual thinking requires Extension Memory (PSI Theory) — the holistic associative capacity that activates under safety conditions. The FSM provides a scaffold: by making transitions explicit and named, it allows the learner to reason about alternatives without the cognitive load of holding the full causal chain in working memory.

In VUCA conditions, counterfactual thinking is the specific capacity that distinguishes adaptive from maladaptive responses to disruption. Authoritarian information systems suppress it systematically — media censorship (V-Dem's most documented autocratization tactic) is, at the cognitive level, the elimination of counterfactual information that would allow citizens to model alternative state trajectories for their society.

3. Abductive inference Running the machine forward from incomplete information: given the current state and available evidence, what is the most coherent account of how we got here and where we're going? Abduction (Peirce, 1903) is inference to the best explanation — the reasoning mode that generates hypotheses rather than verifying them. Distinguished from deduction (necessary conclusions from premises) and induction (generalizations from instances), abduction is the creative, defeasible inference that produces new mental models rather than confirming existing ones.

Abductive inference is the cognitive operation at the K→U threshold — the moment when the learner's existing model is insufficient and they must generate a new explanatory framework rather than assimilating the new information into old categories. It is also the operation suppressed most directly by the closed representational space: a system that has sealed its categories cannot generate genuinely new explanatory hypotheses, only rearrange existing ones.

The FSM as curriculum structure: The four instruments map onto the FSM architecture:

  • ♠ Somatic Gyroscope = state register (reads current state)

  • ♦ Cognitive Radar = transition function (pattern-matches inputs to state changes)

  • ♥ Relational Compass = acceptance condition (what states are we trying to reach? what transitions are impermissible?)

  • ♣ Dimensional Integration = the tape (the full history of states and transitions, integrated over time)

The DIKW stack is an FSM: each level is a state, transitions are triggered by specific conditions (kindness field established, aperture open, disequilibration tolerated), and the machine can run forward (learning) or backward (repair and integration).


Part IV: Advanced Operating Systems Concepts as Cognitive Literacy

The following AOS concepts are proposed as core curriculum elements for the TKGPT — not as technical content but as cognitive scaffolds. Each names a failure mode that is both formally understood in distributed systems and operationally present in human cognition and social dynamics.

Concurrency and Parallel Processing

Multiple processes active simultaneously, none waiting for a central controller. The ecocentric world model is concurrent — multiple stakeholder perspectives held in parallel, meaning emerging from local interactions rather than hierarchical resolution. The egocentric world model is sequential — one central process (the narrative self) that must finish before the next can begin.

Cognitive literacy application: The capacity to hold multiple simultaneous valid perspectives without requiring immediate hierarchical resolution is the cognitive analog of concurrent processing. It corresponds to co-applicability (NBI Level 1) and is a prerequisite for genuine empathy. Its absence — the demand for sequential resolution, one perspective at a time — is a diagnostic signal of control-flow mental architecture under stress.

Deadlock

Two processes each waiting for the other to release a resource before proceeding. Neither can move. The system freezes at a stable non-productive state.

Cognitive/social application: Political and relational deadlock occurs when neither party will risk vulnerability without first receiving it from the other. Neither will release the resource (trust, recognition, concession) because doing so feels like unilateral exposure. The kindness field is the deadlock-breaking protocol — not because it forces one party to yield but because it changes the resource calculus: in a sufficient kindness field, releasing first is no longer experienced as unilateral exposure because the field itself provides the safety that was previously withheld. Ostrom's Principle 6 (conflict resolution mechanisms) is the institutional deadlock-breaking architecture.

Race Conditions

Unpredictable outcomes when the system's behavior depends on the relative timing of events that cannot be controlled. A program that works correctly when events occur in one sequence fails when they occur in another — and in complex concurrent systems, the sequence cannot be guaranteed.

Cognitive application: The control-flow world model assumes that the narrative self can sequence events. VUCA conditions are race conditions at civilizational scale: the sequencing assumption fails. Anxiety about VUCA is often specifically the experience of a control-flow mind encountering race conditions — the mental model predicts that if I do A then B then C, outcome D will follow; VUCA conditions mean that A, B, and C may all be happening simultaneously in different subsystems, producing emergent outcomes that were not in the original specification. The data-flow mental model does not eliminate unpredictability; it expects it and designs for emergence rather than control.

Fault Tolerance and Graceful Degradation

The system continues to function — at reduced performance — when components fail, rather than failing catastrophically. Designed redundancy allows the system to lose components without losing function.

Cognitive/social application: Resilient communities and individuals are not those that never encounter failure but those whose architectures allow graceful degradation. The repair cycle (♣ Dimensional Integration: Action → Repair → Integration) is the personal fault-tolerance protocol. Ostrom's Principle 5 (graduated sanctions) is the community fault-tolerance protocol — it allows violations to be addressed proportionally rather than triggering catastrophic social collapse or enabling indefinite exploitation.

The absence of fault tolerance in brittle control-flow systems manifests as: perfectionism (the system cannot function at reduced performance), shame spirals (a single failure cascades to total system collapse), and authoritarian over-control (external control substituting for internal resilience). All three are failure modes of insufficient fault tolerance architecture in human cognitive and social systems.

The Mixture-of-Experts (MoE) Implication

Modern AI systems use mixture-of-experts architectures because no single model can be trusted for all inputs — routing mechanisms distribute queries to specialized sub-networks and aggregate outputs. The four instruments are the human MoE architecture. The critical insight: the routing mechanism is the values vector. A miscalibrated router (corrupted values vector) will consistently over-weight the Cognitive Radar's rationalization capacity regardless of what the other instruments report. This is why values vector specification (Ostrom's eight principles) is not a moral add-on to the framework — it is the technical specification of the routing function.


Part V: Design Constraints for the TKGPT

These constraints are derived from the framework's own claims. They are not limitations to be overcome — they are design specifications that make the tool honest.

What the TKGPT can do

  • Create pre-symbolic encounters with formal structures (simulation, metaphor, FSM visualization) that anchor subsequent symbolic understanding

  • Scaffold TIF checks: invite the learner to assign T, I, and F values independently before acting on a proposition

  • Support present-moment state awareness by consistently asking "what state are you in right now?" before engaging with content

  • Model counterfactual thinking by explicitly naming alternative state trajectories: "what transition would have produced a different outcome?"

  • Provide provenance-traceable responses: every framework claim linked to its Tier 1/2/3 epistemic status

  • Support the learner's documentation of their own learning journey — the GitBook as personal FSM trace

  • Implement Ostrom's Principle 3 (collective choice): invite the learner to participate in shaping the interaction rather than receiving prescribed content

What the TKGPT cannot do

  • Cross the MPCM boundary: Context and Meaning require a living system embedded in time, relationship, and embodied experience. The tool can scaffold; it cannot ground.

  • Perform genuine Understanding or Wisdom: these require somatic engagement with lived experience that no language model can provide

  • Replace the kindness field: the tool can describe and scaffold; it cannot maintain the relational safety that makes genuine transformation possible

  • Diagnose or treat legacy trauma: the tool should never attempt to identify a learner's trauma history. It can create conditions for ToM of the past self; clinical work is outside its scope

  • Guarantee cascade: transformative learning cannot be forced. The tool maintains SOC conditions; it cannot manufacture the avalanche

The provenance protocol

Every significant claim in the TKGPT knowledge base should be tagged with:

  • Epistemic tier (1/2/3)

  • Primary source with full citation

  • Zotero key (for provenance chain continuity)

  • GitBook page reference (for learner follow-up)

This implements Ostrom's Principle 4 (monitoring) at the knowledge architecture level and directly addresses the confabulation problem that made ChatGPT inadequate for this purpose.


Part VI: The Velocity Vector Problem

The framework's deepest pedagogical challenge: how do you make transmissible a trained alignment between personal decision-making velocity and prosocial values — when that alignment was produced by years of cross-domain integration, somatic healing, contemplative practice, and iterative learning that cannot be compressed into curriculum content?

The honest answer, derived from the framework's own claims, is: you don't transmit it. You create conditions in which each learner begins growing it from their own current state.

The FSM provides the most useful framing: the velocity vector is a transition function that has been tuned through repeated practice to favor prosocial attractor states even under perturbation. You cannot install a transition function in another person's state machine. You can:

  1. Make the concept of a transition function visible — show learners that their current responses are not inevitable reactions but the outputs of a tunable system

  2. Show what prosocial attractor states look like (Ostrom's principles, Commitment Pool mode) so the learner has a target state to navigate toward

  3. Create repeated low-stakes encounters with the K→U threshold — titrated disequilibration within sufficient kindness field — so the learner's own transition function gradually updates toward greater resilience and prosocial orientation

  4. Model the process: the researcher/educator's own trajectory (Tier 4 positionality) is present implicitly as evidence that the transition function can be trained

This is the TKGPT's core design constraint and its core design gift: it cannot give learners the velocity vector. It can show them that they have one, that it was shaped by specific conditions, that those conditions can change, and that the shaping is ongoing.


Part VII: Integration Map — What Exists, Where, and What's Next

Current knowledge commons state

Asset
Location
Epistemic status
Next action

Project files / GitBook

Draft for circulation

Add V-Dem 2026 citation; add BFT section; add speculative tiering note

Glossary (VIM Bridging Spiral v3)

Project files / GitBook

Version 3, March 2026

Add new entries (BFT, TIF practice, ToM+Trauma, Ecocentric Arc)

Speculative glossary

Does not exist yet

Planned

Create Tier 3 entries: Meijer toroidal models, SOC+consciousness

Zotero library

Zotero (local/cloud)

Research Curation

Map Zotero keys to framework Tier 1/2 citations

Substack

Public record

Link to GitBook as researcher positionality documentation

TKGPT system prompt

Does not exist yet

Next phase

Build after framework documentation is stable

The one-sentence orientation for a new partner

The Vital Intelligence Model is a framework for AI literacy that uses formally validated concepts from distributed systems engineering, evolutionary biology, commons governance, and transformative learning theory to help individuals and communities distinguish generative information dynamics from extractive ones — and develop the cognitive, somatic, and relational capacities to act from that distinction.


References (Tier 1 anchors for this document)

Bak, P. (1996). How Nature Works: The Science of Self-Organized Criticality. Copernicus Books.

Bellaiche, L., et al. (2025). Selective emotion regulation in creative art production. Psychology of Aesthetics, Creativity, and the Arts.

Friston, K. (2010). The free-energy principle: A unified brain theory. Nature Reviews Neuroscience, 11, 127–138.

Gunji, Y.P. (2025). Natural Born Intelligence Manifesto. Biosystems. https://doi.org/10.1016/j.biosystems.2025.105398

Kegan, R. (1994). In Over Our Heads. Harvard University Press.

Kuhl, J., Quirin, M., & Koole, S.L. (2020). The functional architecture of human motivation. Advances in Motivation Science, 7, 1–63.

Lamport, L., Shostak, R., & Pease, M. (1982). Byzantine generals problem. ACM Transactions on Programming Languages and Systems, 4(3).

Mezirow, J. (1991). Transformative Dimensions of Adult Learning. Jossey-Bass.

Nord, M., Altman, D., Fernandes, T., Good God, A., & Lindberg, S.I. (2026). Democracy Report 2026: Unraveling The Democratic Era? V-Dem Institute, University of Gothenburg.

Ostrom, E. (1990). Governing the Commons. Cambridge University Press.

Singer, T. (2025). A neuroscience perspective on the plasticity of the social and relational brain. Annals of the New York Academy of Sciences, 1547, 52–74.

Wilson, D.S. (2019). This View of Life. Pantheon Books.

Wilson, D.S. & Ostrom, E. (2019). Prosocial. Context Press.


Humanity++ · Bridging Spiral SR2 Repository: https://kdoore.github.io/HumanityPlusPlus License: CC BY-SA 4.0 Draft: March 2026 — TKGPT Development Cycle

Last updated