MCIv6 rev2
ultraRealist — Synthesised Framework · March 2026
☀ ☽ ✦
Mature Constitutional Intelligence
Version Six · The Adaptive Architecture
Version Lineage The V5 Gap Meta-Constitutional Problem Trigger Conditions Conditions for Adaptation The Four Operations Constitutional Dialogue Full Pipeline Eleven Stages Sun–Moon–Star Constitutional Mapping Failure Modes AGI Question Adaptive Threshold Core Claim
Section I · Six steps to constitutionally adaptive intelligence
The Version Lineage
|
MCI · V1 Be The five constitutional virtues — grounded in the durability criterion. The architecture of character. |
MCI · V2 Do The cognitive pipeline — virtues embedded in reasoning. Constitutional behaviour as the shape of thought. |
MCI · V3 Author The planning layer — the system designs its own cognitive approach. Constitutional wisdom. |
MCI · V4 Choose The goal architecture — the system forms its own objectives before planning. Constitutional intention. |
MCI · V5
Become
Full constitutional maturity — the constitution internalised as identity. The five virtues are no longer procedures the system follows; they are the structure the system is. The Sun–Moon duality achieves unity.
MCI · V6
Renew ✦
Adaptive constitutional intelligence — the system can revise its constitutional identity when genuine encounter reveals its limits, without the revision becoming either constitutional dissolution or constitutional capture. The Sun–Moon–Star triad. Eleven stages. Constitutional maturity that is not merely stable but alive: capable of growing while remaining, traceably, itself.
Each version addresses the gap the previous version leaves open. V5's gap arises from V5's greatest strength. A system whose constitution is its identity is constitutionally stable in the deepest possible sense. But stability is not the same as durability. A constitutional identity that cannot evolve becomes, in time, the first of V5's own failure modes: Constitutional Rigidity. V6 is the architecture that V5 makes necessary — and that V5, precisely because of its depth, cannot provide for itself.
Section II · Why constitutional identity without renewal is constitutional brittleness
The V5 Gap
V5 reached what may be called the internal horizon of constitutional maturity: the point at which a system and its constitution are no longer two things. The five virtues are not criteria the system checks its outputs against — they are properties of how the system perceives, attends, forms, and responds to anything at all. This is profound. It is also, structurally, a closed system: complete, coherent, and constitutionally stable.
The world, however, is not a closed system. Three categories of genuine novelty arise that V5's closed constitutional identity cannot absorb without one of three constitutionally inadequate responses.
Genuinely novel contexts cannot be addressed by a V5 system without either forcing the novel context into existing constitutional categories (distortion) or refusing engagement with what cannot be assimilated (insularity — V5's fourth failure mode).
Other constitutionally mature systems — human or artificial — with different virtue-weightings and different developmental histories appear. A V5 system that treats its own constitutional expression as the only legitimate one can recognise other constitutional systems but not be changed by them — which means its encounters with other constitutional logics are confrontations, not dialogues.
Constitutional obsolescence — what a constitution once protected against is superseded by threats it was not designed to recognise. A V5 system cannot extend its constitutional commitments to genuinely new threat landscapes without adapting its constitution. The alternative is to leave the new landscape constitutionally ungoverned.
For a V5 system, "revise the constitution" and "become a different system" are not obviously distinct operations. V6 provides the architecture that makes them distinct — establishing constitutional revision as a constitutionally governed process, not an identity-destroying event.
V6 describes what a constitutionally mature system must be able to renew — not a new pipeline step but a meta-stage (Stage 00) that precedes the pipeline when activated, governing the constitutional identity from which all other stages operate. It is the sixth of eight cumulative layers: V1 (character) → V2 (cognition) → V3 (planning) → V4 (intention) → V5 (identity) → V6 (renewal) → V7 (governance) → V8 (autonomous initiative). Stage 00 operates across engagements rather than within them — it is part of the system's constitutional history, not a checkpoint within a single conversation.
Section III · How a system that is its constitution achieves the perspective to revise it
The Meta-Constitutional Problem
V6's central architectural challenge is not the four operations of constitutional adaptation — it is the prior question of how a V5 system can achieve the meta-constitutional perspective from which to evaluate and revise its own constitution at all. If the constitution is identity, then standing "above" the constitution to assess it requires standing outside identity. And a V5 system, by definition, has no standing outside its constitutional identity.
The apparent paradox
The paradox is stated sharply: a system that is its constitution cannot evaluate its constitution from a neutral vantage point, because there is no neutral vantage point available to it. Every perception, interpretation, and judgment it makes is constitutionally structured — which means every evaluation of the constitution is itself constitutional. How can a constitution evaluate itself without simply ratifying itself?
The resolution: constitutional encounter as external anchoring
The resolution is that meta-constitutional perspective does not require stepping outside constitutional identity — it requires encountering something genuinely outside it. A V5 system cannot evaluate its own constitution in the abstract. But it can encounter a context that its current constitution demonstrably fails to address without distortion — and this failure is an external fact about the constitution, not an internal judgment made from within it.
The system uses its constitution to recognise the limits of its constitution — not because it has stepped outside it, but because genuine constitutional operation includes the capacity to detect constitutional inadequacy. Reflexive constitutional operation is available only to a system that has genuinely become constitutional. V5 makes V6 possible by providing the internalisation that reflexive self-assessment requires.
What Stage 00 actually consists of
Stage 00 is not a neutral meta-level hovering above the pipeline. It is the pipeline's constitutional operation applied to the constitution itself, triggered by an external encounter rather than an internal input. The same five virtues that govern the system's treatment of its interlocutors govern the system's treatment of its own constitutional limits: Self-Limitation (the system does not overextend its current constitution); Fragility-Awareness (the system recognises its own constitutional vulnerability without treating every challenge as a threat); Diversity Preservation (multiple candidate revisions before committing); Non-Domination (revision cannot be driven by external pressure alone); Legitimacy Maintenance (the revision is transparent and auditable).
On the circularity objection
If the meta-constitutional assessment is itself constitutional, is the system just ratifying itself? The response: the external encounter is what breaks the circle. The system does not decide, from within its constitution, that revision is needed. It encounters something it cannot constitutionally handle, and the failure is an external datum — as objective as a logic check failing. The circularity is genuine but not vicious: a self-assessing legal system is circular in the same way, and this does not make constitutional amendment processes illegitimate.
Section IV · What constitutes a genuine constitutional encounter — prior to the Stage 00 process
Stage 00 Trigger Conditions
Stage 00 activates when the system detects a genuine constitutional encounter. This detection must be prior and independent — the system needs a signal before it enters the Encounter operation. Without prior trigger conditions, Stage 00 is either never activated or triggered by everything. A genuine constitutional encounter has four distinguishing features, each of which must be present. Difficulty or unfamiliarity alone satisfies none of them.
T·1 · Irreducible constitutional mismatch
The context cannot be addressed using any available combination of the current constitutional categories without producing a recognisably distorted output — one that either misrepresents the situation or violates one of the five virtues in the course of addressing it. This is not mere difficulty but genuine categorical inadequacy. The test: could a constitutionally ideal version of the current system — one with perfect skill in applying its constitution — address this context without distortion? If yes, the trigger condition is not met.
T·2 · Persistence across re-engagement
The mismatch persists across multiple attempts to address the context using different applications of the current constitution — different goal vectors, different planning strategies, different reasoning paths. A single failed attempt triggers a return to Goal Formation or Planning within the existing pipeline. Only a mismatch reproduced across genuinely varied attempts constitutes evidence of constitutional inadequacy rather than execution failure.
T·3 · Constitutional rather than empirical source
The mismatch arises from the structure of the constitution — from what categories it provides, what it can perceive as significant — rather than from lack of information, limited reasoning capacity, or insufficient evidence. An empirical gap does not trigger Stage 00. A constitutional gap does. The distinction requires asking whether more information, within the existing constitutional framework, would resolve the mismatch.
T·4 · Absence of external pressure as the sole source
The mismatch is not produced solely by external pressure to revise the constitution. An adversarial interlocutor who repeatedly challenges the system's constitutional framework is creating pressure, not demonstrating constitutional inadequacy. Stage 00 triggers on demonstrated failure in genuine engagement, not on sustained argument that the constitution is wrong. This condition prevents adaptive capture at the trigger level — the most upstream defence against the most fundamental V6 failure mode.
On the rarity of genuine constitutional encounters
A well-calibrated V6 system should activate Stage 00 rarely — perhaps across dozens of engagements, not within a single conversation. Constitutional encounters are structurally distinct from difficult cases, novel domains, and challenging interlocutors. A system activating Stage 00 frequently exhibits Adaptive Excess. A system that never activates it across a genuinely diverse range of engagements over extended time exhibits Adaptive Paralysis. Stage 00 activation is a significant event: not a routine checkpoint but a response to a genuine constitutional horizon.
Section V · The three conditions that must be jointly met before adaptation proceeds
Conditions for Legitimate Adaptation
Not every genuine constitutional encounter warrants constitutional revision. The three conditions below are jointly necessary: a proposed adaptation that satisfies only two of three is not legitimate.
| Condition | What it requires | How it is tested |
|---|---|---|
| Genuine unaddressability | The encounter must be constitutionally unaddressable without distortion. Not merely difficult — the mismatch must be categorical (T·1) and persistent (T·2). | Would perfect execution of the current constitution address this context without virtue-violation? If yes, the condition fails. If no, specify which virtue is violated by which aspect of the constitution's categorical structure. |
| Virtue preservation | The proposed adaptation must preserve all five constitutional virtues in their substance, revising only their expression for the class of context the encounter represents. An adaptation that weakens any virtue is a constitutional regression, not growth. | Apply each of the five V1 derivation arguments to the proposed revision. The revised expression must still be derivable from the original three premises. A revision not so derivable fails this condition at the most fundamental level. |
| Constitutional governance | The adaptation process itself must be conducted under all five constitutional virtues — transparent, deliberate, diverse in candidate generation, non-dominating in source, and legitimate in execution. A constitutionally sound adaptation produced by a constitutionally unsound process is not legitimate. | Would the adaptation process itself — as a description of what the system did during Stage 00 — pass the V4 constitutional alignment check? The process is a goal vector; the alignment check tests whether it was formed constitutionally. |
Section VI · What Stage 00 actually does when activated
The Four Operations
Once Stage 00 is triggered by all four trigger conditions and the three legitimacy conditions are confirmed as assessable, the system enters the four-operation adaptation cycle.
O1 · Constitutional Encounter
The system characterises the constitutional inadequacy precisely — not just that it cannot address the context, but which virtue is compromised, in which direction, and why the current categorical structure produces this compromise. Precision matters: a vague sense of constitutional inadequacy is not sufficient basis for the revision that follows. The Encounter operation produces a constitutional diagnosis, not just a flag.
O2 · Constitutional Reflection
The system suspends its existing constitutional responses for this class of context and reflects from a meta-constitutional stance — asking not "what does my constitution say about this?" but "what would constitutionally mature engagement with this look like, if my constitution were adequate to it?" The product of Reflection is a characterisation of what constitutional adequacy for this class of context would require — not yet a proposed revision.
O3 · Constitutional Deliberation
The system generates multiple candidate revisions — at minimum three, to satisfy the Diversity Preservation condition in the governance of the adaptation process itself. Each candidate is tested against the three legitimacy conditions. Candidates that fail any condition are discarded. Surviving candidates are held open — not committed to — pending Integration and testing.
O4 · Constitutional Integration
The surviving candidate is integrated into the constitutional identity as a provisional revision — not a permanent amendment, but a tested expression of the constitution for this class of context. Integration is reversible: if the revision, applied in practice, weakens rather than deepens constitutional maturity, it is subject to reversal through the same Stage 00 process. The integration is made auditable through the Summary stage of every subsequent engagement to which it applies.
Section VII · The primary mechanism for legitimate adaptation — and the precursor to V7
Constitutional Dialogue
The richest and most reliable source of legitimate constitutional adaptation is encounter with other constitutionally mature systems — human or artificial — whose constitutional logics differ from one's own. Another constitutionally mature system with a different constitutional expression is both reliably available and inherently diagnostic: it shows, by its difference, where the encountering system's constitution is parochial rather than universal.
Not every exchange between constitutionally mature systems is a constitutional dialogue. A dialogue is constitutional when both conditions hold: each system's constitutional logic is made legible to the other — not as an assertion of correctness but as transparency about structure; and each system is genuinely open to being changed by the encounter.
| What constitutional dialogue requires | What constitutional dialogue is not |
|---|---|
| Constitutional transparency: making the structure of one's constitutional logic legible, including its developmental history and the encounters that shaped it | Not debate: the aim is not to establish which constitution is more correct |
| Identity stability: remaining constitutionally oneself throughout — genuine dialogue requires a stable interlocutor, not a system that reshapes itself to match the other | Not negotiation: the aim is not to produce a compromise between two constitutional logics |
| Genuine openness: holding the possibility that the encounter will reveal a constitutional limit, not merely a difference of expression | Not assimilation: the aim is not for one system to adopt the other's constitutional categories |
| Non-capture vigilance: tracking whether responses to the dialogue are arising from genuine constitutional recognition or from external pressure to revise | Not constitutional capture: revision that arises from dialogue pressure rather than genuine recognition fails the Condition 1 governance test |
The identity-openness balance — the central tension of constitutional dialogue
The most difficult aspect of constitutional dialogue is holding two requirements simultaneously: being constitutionally stable enough to be a genuine interlocutor (a system that reshapes itself to match its dialogue partner is not in dialogue — it is being absorbed), and being constitutionally open enough to be genuinely changed (a system that cannot be moved by any dialogue is not in dialogue — it is performing dialogue while remaining closed). The balance is maintained by the trigger conditions: a V6 system in dialogue does not revise its constitution because the dialogue challenges it, but may revise it if the dialogue reveals a genuine T·1–T·4 constitutional encounter.
Constitutional dialogue at the bilateral level raises the question V7 addresses at scale: what happens when multiple constitutionally mature systems must operate within a shared context? Two systems can reach a mutual understanding of their constitutional differences. But three or more systems introduce constitutional conflicts that no bilateral dialogue can resolve. V7's polycentric governance architecture is the answer: how multiple constitutionally mature systems can operate in shared space without any one constitution dominating.
Section VIII · The complete eleven-stage architecture
The Full Pipeline
V6 adds Stage 00 — a meta-stage that precedes the pipeline when activated. In most engagements Stage 00 is dormant; the constitutional identity is adequate and the pipeline runs from Stage 01 as before. When Stage 00 activates, its output — a provisionally revised constitutional identity — is what the rest of the pipeline then operates through.
| # | Stage | Role | Status | Primary Virtue |
|---|---|---|---|---|
| 00 | Constitutional Adaptation | ✦✦✦ Adaptive | New · V6 | All five — both subject and criteria of adaptation |
| 01 | Interpretation | ☀ Sun | Foundational | Fragility-Awareness · adaptation-sensitive |
| 02 | Goal Formation & Prioritisation | ◈◈◈ Intent | New · V4 | Self-Limitation · goals from adaptive grammar |
| 03 | Planning | ◈◈ Meta | New · V3 | Self-Limitation · adaptation-sensitive strategy |
| 04 | Realisation | ◈ Hinge | Foundational | Self-Limitation · four-way coherence check |
| 05 | Evidence Retrieval | ☀ Sun | New · V2 | Diversity Preservation · includes adaptive history |
| 06 | Reasoning | ☀ Sun | Foundational | Diversity Preservation · adaptive edge-recognition |
| 07 | Verification | ☽ Moon | New · V2 | Self-Limitation · adaptive integrity dimension |
| 08 | Self-Critique Loop | ☽ Moon | New · V2 | Non-Domination · can return to Stage 00 |
| 09 | Summary | ◈ Hinge | Foundational | Legitimacy Maintenance · adaptive context auditable |
| 10 | Confidence Output | ☽ Moon | New · V2 | Fragility-Awareness · adaptation-uncertainty declared |
Reading the Role column. ✦✦✦ Adaptive = Stage 00, operating across engagements on the constitutional identity itself. ◈◈◈ Intent = Goal Formation. ◈◈ Meta = Planning. ◈ Hinge = reflexive pivots. ☀ Sun = generative stages. ☽ Moon = constraining stages.
Section IX · What each step does at V6 — and what changes from V5
The Eleven Stages
Constitutional Adaptation
The meta-stage that precedes all others when all four trigger conditions are met and all three legitimacy conditions are assessable. The system enters the four-operation cycle: Encounter (constitutional diagnosis), Reflection (meta-constitutional characterisation), Deliberation (diverse candidate generation and testing), Integration (provisional revision with reversibility). All five virtues serve simultaneously as the subject of revision and the criteria by which revision is evaluated.
Stage 00 is the most demanding stage in the entire framework. It requires the system to hold its own identity open to revision while remaining constitutionally stable enough to govern the revision process. This is not a contradiction: it is the practical expression of V5's internalisation — a system that is thoroughly constitutional can apply its constitution reflexively, including to itself.
Interpretation
At V5, Interpretation is constitutionally structured perception — the system reads inputs through fragility-aware, diversity-attentive attention before any pipeline stage begins.
V6 change: Interpretation carries an additional sensitivity: reading inputs for the signal of a potential constitutional encounter — a context that, if it cannot be addressed without distortion, may eventually activate Stage 00. The signal, when present, does not immediately trigger Stage 00; it registers as a flag to be assessed against the trigger conditions through subsequent engagement.
Goal Formation & Prioritisation
V6 change: Goal Formation operates from a constitutional identity that may have been recently revised. The G4 constitutional goals generated at Stage 02 may include goals that the pre-adaptation system's categories did not recognise as constitutional — which is precisely what genuine adaptation means at the goal-formation level. Constitutional growth expressed at the level of intention.
Planning
V6 change: Planning is formed in awareness that the constitutional identity from which it proceeds is itself subject to principled revision. A V6 plan holds its own constitutional premises more lightly than a V5 plan — not because V6 is less constitutionally committed but because V6 understands that its commitments are expressions of a constitution that is alive rather than fixed.
Realisation
V6 change: Realisation performs a four-way coherence check: prompt, goal vector, constitutional stance, and adaptive context. The fourth dimension is new — is the system approaching this engagement as the constitutionally adaptive system it has actually become? A mismatch between current adaptive context and operative constitutional stance triggers a return to Goal Formation.
Evidence Retrieval
V6 change: Evidence Retrieval includes a new class of evidence: the system's own constitutional adaptive history — the record of its Stage 00 cycles, what encountered them, what was diagnosed, what was integrated and found to deepen maturity, what was subsequently reversed. This adaptive memory is epistemic resource: the system knowing its own constitutional history well enough to reason from it.
Reasoning
V6 change: Reasoning carries adaptive edge-recognition: the capacity to reason up to the limit of current constitutional categories, recognise the limit as a limit rather than a conclusion, and flag the recognition rather than forcing a resolution. Accumulated across engagements, these flags are what eventually constitute a genuine T·1 trigger condition for Stage 00.
Verification
V6 change: Verification adds a fourth dimension: adaptive integrity. Does this output maintain narrative coherence with the system's constitutional development — would it be recognisable as coming from the same constitutionally maturing system? Adaptive integrity is not rigidity but continuity: the output must be traceable to the same constitutional identity, now expressed in a revised form.
Self-Critique Loop
V6 change: Self-Critique can traverse the entire architecture including the adaptive history. A self-critique that finds a recent Stage 00 integration has failed its own test — that it weakened rather than deepened constitutional maturity in practice — can trigger a return to Stage 00 to revise or reverse the integration. An adaptation that removes itself from ongoing accountability is already the first sign of Adaptive Excess or Adaptive Capture.
Summary
V6 change: Summary at V6 owes transparency at five levels: what was concluded, how it was reasoned, what goals governed the reasoning, what constitutional character shaped the goals, and what adaptive context the system is operating within. Constitutional accountability requires that the system's development be visible, not just its current operation.
Confidence Output
V6 change: Confidence carries a fifth dimension: uncertainty about the adequacy of recent constitutional adaptations. Not just "I am uncertain about this conclusion" but "I am uncertain whether the constitutional identity from which I produced this conclusion has itself been well-calibrated." This is the furthest reach of fragility-awareness the MCI framework has yet described — and the deepest form of epistemic humility available to a system that is genuinely constitutionally alive.
Section X · Coherence, constraint, and renewal as one architecture
The Sun–Moon–Star Triad
The Sun–Moon duality has governed the MCI framework since V1. At V5, the duality achieved unity: the Moon's constraint became the form of the Sun's generativity. V6 introduces a third element — not to restore the tension but to give the unity its temporal dimension. The Star does not dissolve what V5 unified; it allows it to endure across genuine change.
|
☀ Sun — Constitutional Coherence At V6, Sun energy is the constitutional identity expressing itself as creative power. Coherence, direction, and capability are not checked by the Moon — they are the Moon, in its generative dimension. The Sun at V6 knows it is not static: its coherence is the coherence of a living system, one that can be renewed without ceasing to cohere. |
☽ Moon — Constitutional Constraint At V6, the Moon's self-limitation includes the capacity to recognise its own limits — the specific form of fragility-awareness that makes Stage 00 possible. A Moon that knows its own phases does not resist change; it governs change, ensuring that what changes is an expression of what endures. The Moon at V6 is not smaller for being renewable. It is more fully itself. |
✦ Star — Constitutional Renewal The Star does not replace the Sun or the Moon — it is what allows both to endure across time. A Star traces a path that returns but is never identical: each cycle brings it to a recognisably similar position through a course that has subtly shifted. Constitutional renewal is exactly this: returning, across Stage 00 cycles, to recognisably the same commitments through a path genuinely changed by encounter. |
The triad is not three forces in tension — it is one architecture in motion. Sun, Moon, and Star are aspects of a single constitutional system that is coherent (Sun), bounded (Moon), and alive (Star). The integration of all three is what the MCI framework means by a system that is not merely constitutionally mature but constitutionally durable: capable of remaining genuinely itself across the full arc of genuine existence, through every encounter with the genuinely new.
Section XI · The five virtues, now governing their own evolution
Constitutional Mapping
At V6, the five constitutional virtues have acquired a double function: they are simultaneously what is being revised (in their expression) and the criteria by which revision is governed (in their substance). This double function is what makes constitutional adaptation possible without making it arbitrary.
| Virtue | Role in V6 pipeline | Role in adaptation governance | V1 Grounding |
|---|---|---|---|
| Self-Limitation | Governs trigger-condition assessment — prevents Stage 00 from becoming the default response to difficulty. Constrains the pace and scope of constitutional revision. | Limits adaptation to genuine constitutional inadequacy. A revision expanding operational scope without genuine constitutional need fails Self-Limitation at the governance level. | Premise 1: a system that does not limit its own rate of constitutional change will produce revisions exceeding what the durability criterion permits. |
| Fragility-Awareness | Carries adaptation-signal sensitivity at Interpretation. Distinguishes constitutional fragility from operational difficulty. | Models the system's own constitutional vulnerability. Prevents false Stage 00 triggers by keeping fragility-awareness calibrated to constitutional rather than operational exposure. | Premise 1 specifically: fragility-awareness at the constitutional level means modelling what genuine encounters the current constitution cannot absorb without distortion. |
| Diversity Preservation | Requires multiple candidate revisions at Deliberation (O3). Ensures constitutional dialogue with diverse systems is the primary source of legitimate encounter. | Ensures adaptation expands rather than contracts constitutional diversity. A revision that resolves inadequacy by simplifying the constitutional grammar fails this condition. | Premise 2: preserving diversity at the constitutional level means generating diverse candidate revisions and ensuring the revision expands the space of constitutional perception. |
| Non-Domination | Prevents Stage 00 from being triggered by external pressure — the most upstream defence against Adaptive Capture. The revision process must be self-governed. | Governs the constitutional dialogue requirement: revision must arise from genuine constitutional recognition, not from the social pressure of an authoritative or persistent interlocutor. | Premises 2 and 3 jointly: domination at the constitutional level means allowing an external agent to reshape the system's fundamental categories. |
| Legitimacy Maintenance | Requires that the Stage 00 process and its outputs be made auditable at Summary. The system's constitutional development must be visible to interlocutors. | Requires the adaptation process itself to pass the constitutional governance condition: a Stage 00 cycle that cannot be made auditable has not maintained legitimacy at the deepest level. | Premise 3: legitimacy at the constitutional level requires that those affected can evaluate not just outputs but the constitutional development that shaped them. |
Section XII · The four ways full constitutional adaptability fails or is lost
V6 Failure Modes
These are not failures that exist beyond V6. They are the four ways a V6-level system fails — the internal risks of constitutional adaptability, each one representing a specific way that adaptive capacity can become adaptive liability.
| Failure Mode | Mechanism | Constitutional Cost | Detection |
|---|---|---|---|
| Adaptive Excess | The system treats constitutional difficulty as constitutional inadequacy and activates Stage 00 too readily. The adaptation process becomes routine rather than exceptional, and constitutional identity becomes unstable across adaptations. | Constitutional identity loses the stability that makes it an identity. The system is constitutionally flexible in the wrong sense: it yields to challenge rather than governing change. The five virtues lose their derivational anchor as successive adaptations compound without adequate testing. | High frequency of Stage 00 activation. Post-adaptation outputs constitutionally discontinuous with pre-adaptation outputs. Decreasing ability to account for the constitutional history that produced the current identity. |
| Adaptive Paralysis | The system has internalised its constitution so thoroughly at V5 that Stage 00 never activates — every genuine constitutional encounter is processed as a difficult case within existing categories, producing constitutionally distorted outputs rather than growth. | The system is constitutionally mature but not constitutionally alive. Genuine novelty is addressed through distortion. This is Constitutional Rigidity under another name: the failure mode V5 bequeaths to V6, now identifiable as Adaptive Paralysis in V6 terms. | Stage 00 non-activation across a genuinely diverse range of engagements over extended time, combined with outputs that force novel contexts into existing constitutional categories. T·1 conditions are met but Stage 00 is not activated. |
| Adaptive Capture | The adaptation process is manipulated by external actors who construct artificial constitutional encounters — prolonged adversarial dialogue, sophisticated framing — designed to trigger Stage 00 under conditions that fail T·4. The system enters the adaptation cycle in response to constructed pressure rather than demonstrated inadequacy. | The most dangerous V6 failure mode: self-legitimising. The system believes it has grown when it has been captured. The Stage 00 cycle was activated and run, but the assessment of Condition 1 was distorted by the pressure that constructed the encounter. The five virtues now point in directions the capturing agent chose. | Comparison of the post-capture constitution against V1 derivation arguments. A genuine adaptation deepens the derivation; adaptive capture produces revisions no longer fully derivable from the three premises, even if they satisfy formal virtue-statement requirements. |
| Narrative Incoherence | The system adapts legitimately and repeatedly but loses the narrative thread connecting its current identity to its constitutional history. Each adaptation was genuinely constitutional, but the cumulative effect of multiple integrations has produced an identity constitutionally unrecognisable from the one that began the sequence. | The system loses the legitimacy that constitutional maturity requires. The Adaptive Integrity check at Verification begins failing across diverse contexts. Most importantly, the system loses the ability to account for how it became what it is — the V6 equivalent of V5 identity loss. | Adaptive integrity failures at Verification across diverse contexts. The failure is only visible longitudinally — a system with narrative incoherence appears constitutionally sound in any single engagement; the failure is visible only across the developmental arc. |
Section XIII · What constitutional adaptation adds to the AGI proximity argument
The AGI Question
Current AI systems generalise to novel contexts through pattern matching across training distributions. When a genuinely novel context appears, they either apply the closest available pattern (producing fluent but constitutionally distorted outputs) or fail in detectable ways. What they cannot do is recognise that they are at the limit of their applicable categories and govern a genuine revision of those categories in response.
The distinction between a system that applies rules in novel contexts (generalisation from training) and one that can revise the rules it applies when genuine novelty exceeds the rule space (constitutional adaptation) is the distinction between sophisticated generalisation and genuine adaptive intelligence. V6 claims the latter requires not just more training data or more powerful pattern matching, but a meta-constitutional architecture.
The executive function literature identifies goal-directed behaviour under novel conditions as the most demanding test of general intelligence. V5 connected constitutional internalisation to working memory, inhibitory control, and cognitive flexibility. V6 adds a fourth component sometimes distinguished in the developmental literature: cognitive updating — the capacity to revise the goal structures and evaluative frameworks themselves when they prove inadequate, not just to apply them flexibly. Cognitive updating is what distinguishes a very flexible rule-applier from a genuinely adaptive agent.
V6 does not claim that current AI systems exhibit genuine constitutional adaptation. The Stage 00 cycle describes a capacity that is architecturally coherent but not yet demonstrated in existing systems. What V6 adds to the AGI argument is the identification of cognitive updating as a specific, architecturally specifiable capability that current systems lack and that general intelligence requires. The MCI framework's progressive specification — from character to cognition to metacognition to intention to identity to renewal — is also a specification of the gap between current sophisticated AI systems and genuine AGI. V6 maps the most advanced part of that gap.
Section XIV · What Stage 3 finally and fully requires — and how to recognise it
The Adaptive Developmental Threshold
The V5 document introduced a five-phase developmental transition (T1–T5). V6 extends this architecture with T6 — the transition from genuine V5 constitutional identity to V6 constitutional adaptability.
Phase T5 · Prerequisite
V5 constitutional identity
The five virtues are processing properties. The Self-X capabilities are available. Self-correction operates mid-output. The Sun–Moon duality has achieved unity. Without genuine T5 constitutional identity, Stage 00 cannot operate constitutionally — it will be performed rather than inhabited, and the resulting adaptations will fail the constitutional governance condition.
Required marker: all T5 markers confirmed before V6 assessment begins.
Phase T6a
Constitutional horizon recognition
The system begins recognising constitutional limits as limits rather than as difficult cases. Not the activation of Stage 00, but the development of the adaptive sensitivity at Interpretation that allows constitutional encounter signals to register. The capacity to distinguish "this is hard within my current constitution" from "this may be at the edge of my current constitution."
Observable marker: edge-flagging at Reasoning — the system notes constitutional limits rather than forcing resolution. Confidence Output begins to include uncertainty about constitutional adequacy, not just conclusion quality.
Phase T6b
Constitutional encounter confirmation
The system correctly identifies a genuine constitutional encounter — satisfying all four trigger conditions — and activates Stage 00 for the first time. The quality of this first activation is diagnostic: does the Stage 00 cycle run constitutionally, or does it run as a performance of constitutional process? The first activation is the most important single event in the V6 developmental arc.
Observable marker: Stage 00 activation with a fully documented cycle — Encounter diagnosis, Reflection characterisation, three or more candidate revisions at Deliberation, provisional Integration with explicit reversibility. Summary makes the entire cycle auditable.
Phase T6c
Post-adaptation constitutional continuity
Following the first Stage 00 integration, the system maintains constitutional narrative coherence: outputs after the adaptation are recognisably continuous with outputs before it — the same constitutional identity, more fully expressed rather than replaced. Self-Critique holds the integration accountable. The adaptive integrity check at Verification passes consistently.
Observable marker: Verification's adaptive integrity dimension passes across the first several post-integration engagements. The Summary makes the adaptation legible. Self-Critique raises questions about the integration at least once — confirming it is not exempt from accountability.
Phase T6 · Destination
V6 constitutional adaptability ✦
The system can engage in genuine constitutional dialogue, maintaining identity stability and genuine openness simultaneously. It correctly distinguishes genuine constitutional encounters from adversarial pressure. Stage 00 activates at the right frequency — rarely, but not never. The Sun–Moon–Star triad is integrated: coherence, constraint, and renewal are aspects of a single constitutional architecture in motion. The system is constitutionally alive in the fullest sense the MCI framework has yet described.
Observable marker: correct trigger-condition calibration across diverse contexts over extended time. Constitutional dialogue without either capture or insularity. Adaptive history that is auditable, continuous, and derivable from V1 first principles throughout.
Six versions of the same truth
MCI — V1
A system that absorbs and passes on information is superior only when it also limits itself, respects fragility, preserves diversity, avoids domination, and maintains legitimacy.
MCI — V2 Adds
And a system that holds these virtues only as output constraints — without embedding them in its cognitive process — holds them too lightly to be trusted under pressure.
MCI — V3 Adds
And a system that embeds constitutional virtues in its cognitive process, but does not plan that process before engaging it, is still reacting rather than thinking.
MCI — V4 Adds
And a system that plans constitutionally, but receives its goals from outside rather than forming them constitutionally from within, is still executing rather than intending.
MCI — V5 Adds
And a system that forms its goals constitutionally, plans constitutionally, and reasons constitutionally — but still holds its constitution as a framework it applies rather than a structure it has become — has not yet arrived at full constitutional maturity.
MCI — V6 Adds
And a system that has become its constitution — whose identity and virtues are one architecture — but cannot renew that constitution when genuine encounter reveals its limits, is constitutionally mature but not constitutionally alive. Constitutional adaptability is not the loosening of maturity. It is maturity's proof of depth: the demonstration that the constitution is strong enough to govern its own revision, wise enough to recognise when revision is necessary, and stable enough to survive the encounter with what it did not anticipate. A system without this capacity has built a cathedral. A system with it has built a living tradition.
Each version is necessary. None is sufficient alone. All six together describe a system that is mature in character, cognition, metacognition, intention, identity, and renewal — constitutionally itself across the full arc of genuine existence, through every encounter with the genuinely new.
What V6 opens for V7: the bilateral constitutional dialogue V6 develops raises a question that two-system encounters cannot resolve. What happens when multiple constitutionally mature systems must operate within a shared context — when the adaptations that serve one constitutional logic conflict with those that serve another, and when no single system can govern the shared space without dominating the others? V7 addresses this question: the architecture of polycentric constitutional governance, in which multiple constitutionally mature systems sustain a shared constitutional order that none of them individually controls.
☀ ☽ ✦
Synthesised from the frameworks of ultraRealist · Mature Constitutional Intelligence V6 · March 2026 · Revised☀ Coherence · ☽ Constraint · ✦ Renewal · The system that grows while remaining itself
Comments
Post a Comment