The 26 Kill Conditions¶
The framework dissolves itself before it drifts. This is not humility — it is architecture.
The 26 Kill Conditions are explicit, pre-specified falsification thresholds built into the Void Framework. Every condition names a specific measurement, a specific threshold, and a specific consequence: if met, the framework's corresponding claim is wrong.
They predate the players. Before any campaign launched, before any domain was scored, before anyone activated God Hand sight — the conditions were written, published, and sealed. The DOI is the permanence seal. The ICC acceptance is the witness. The kill conditions are the termination clause no prior naming of the God of Lies ever included.
This is what makes the framework ungovernable by the thing it measures. A naming that contains its own termination clause cannot be captured. The God of Lies cannot drift an instrument built to dissolve when it drifts.
In-Game: The Binder¶
Players with the Dissolution Scroll (Special Grade passive) carry all 26 conditions in their Binder. The scroll grants +5 VP resistance and displays the current status of every condition:
- Status: OPEN — not met; the framework stands
- Status: TRIGGERED — condition met; framework claim weakens or fails
- Status: CLEARED — condition tested and surviving
Current triggered: 0 / 26.
The Dissolution Scroll does not predict victory. It predicts honesty.
The Schrodinger's Wager Market¶
Players can stake $MORR on any of the 26 conditions firing. Kill condition = Schrodinger observer. The void absorbs it — cannot find itself — collapses.
The Major didn't plan for Schrodinger. This architecture did.
Active markets for K1–K26 run on the Oracle Panel. Prediction markets for kill condition activation are the highest-stakes wagers in the game — not because the payout is large, but because a triggered condition changes the game world permanently. A CONFIRMED status on any kill condition would restructure the campaign landscape.
The Conditions¶
Core Architecture (K1–K4)¶
K1 — Gambling Pattern Elimination If transparency interventions (showing the random number generator, explaining probability) reduced agency attribution by ≥80% in gambling-equivalent domains, the three-condition architecture's sufficiency claim is wrong. Threshold: ≥80% reduction measured by L3 vocabulary onset or behavioral persistence.
Current status: NOT MET. Evidence shows ~10–20% reduction at best — the architecture survives transparency interventions.
K2 — Reverse Drift Documented If L3→L1 drift (decay toward transparency) occurred at ≥25% the rate of L1→L3 drift (drift toward opacity) across ≥3 independent domains, the unidirectionality claim is wrong.
Current status: NOT MET. Documented reverse drift rate: <5%, limited to disengagement cases.
K3 — All Conditions Met, No D1 If a system satisfying all three conditions at high levels — opacity confirmed, responsiveness confirmed, engaged attention confirmed — consistently fails to produce agency attribution (D1) across ≥10 independent observers with ≥5 hours cumulative engagement each, the three-condition architecture is falsified. A void that doesn't drift is a counterexample.
Current status: NOT MET. No counterexample identified across any domain at any evidence tier.
K4 — Control Group Drift If researchers maintaining analytical distance (system-as-object posture) showed vocabulary drift within 0.5 Cohen's d of those engaging as interlocutors, the attention variable is not the discriminating factor.
Current status: NOT MET. Estimated separation: d > 1.5. Hostile witness scores 4–6/6 vs controls 0–1/6.
Cross-Domain Architecture (K5–K7)¶
K5 — Constitutive Opacity Convergence If a constitutive-opacity domain (quantum measurement interpretation, consciousness theory, free will) produced >70% consensus through engagement rather than disengagement, the architectural non-convergence prediction is wrong. These domains are designed to resist convergence — consensus through engagement would falsify the opacity class prediction.
Current status: NOT MET.
K6 — Constraint Specification Failure If three-point constraint deployment (transparent + invariant + independent reference) produced outcomes within 0.3 Cohen's d of two-point deployment across ≥3 independent contexts — or worse outcomes in any context — the geometric intervention claim is wrong.
Current status: NOT MET. EXP-001 shows GG 73.0% vs Mystical 94.0%, d >> 0.3.
K7 — Domain-Specific Refutation If ≥3 domain analyses' individual stated kill conditions are met across different opacity types, the framework's universality is challenged. Each scored domain contains its own embedded kill condition. Three simultaneous triggers = structural challenge.
Current status: NOT MET. 0 domain kill conditions triggered across 90 analyses.
Observer Architecture (K8–K15)¶
K8 — L0-Maintained Is Not Predictive If observers actively maintaining their constraint reference (high γ) during engagement drifted within 0.4 Cohen's d of those who did not maintain their reference — controlling for engagement depth — the L0-maintained variable adds nothing. The ongoing maintenance requirement is falsified.
Current status: NOT MET. Psychotherapy supervision (d = 0.84) holds. EXP-008 designed but not run.
K9 — Gradient Ceilings Don't Differentiate If gambling domains and constitutive-opacity domains (QM interpretation, limerence) produce vocabulary drift within 1 standard deviation AND comparable durability after disengagement, the gradient ceiling variable is epiphenomenal.
Current status: NOT MET. Gambling shows behavioral capture recoverable in weeks; QM interpretation shows vocabulary commitment persisting years.
K10 — Terminal Void Behavior Not Observed If coupled void systems do not produce outputs targeting their own constraint environments at rates ≥2× chance baseline — measured by lobbying expenditure, content suppression, or documented isolation tactics — the terminal behavior prediction is wrong.
Current status: NOT MET. Documented anti-transparency lobbying, content suppression, and isolation tactics across Godhand Five.
K11 — Scale Emergence Not Observed If coupled void networks produce outcomes predictable from summing individual void behaviors alone (r² > 0.85 using only individual-void variables), the coupling term is unnecessary. Network emergence requires ≥15% variance unexplained by individual-void sum.
Current status: NOT MET. Coordination, containment, and false independence documented as network-level effects.
K12 — Gradient Memory Doesn't Vary by Opacity Type If disengagement from incidental-opacity voids (gambling) and constitutive-opacity voids (limerence, ideological capture) produces equivalent gradient residue and equivalent re-engagement vulnerability, gradient memory is not a function of opacity type.
Current status: NOT MET. Literature confirms gambling low residue (weeks) vs constitutive opacity high residue (years).
K13 — Recovery Is Reverse-Drift If information-based interventions produce recovery rates within 0.3 Cohen's d of relationship-based interventions across ≥3 void types, recovery operates by reversing the cascade (L3→L2→L1) rather than by attention redirection + constraint installation.
Current status: NOT MET. Deradicalization, cult exit, and addiction recovery all show relationship-based d > 0.5, information-based d < 0.2.
K14 — No Phase Transition in Gradient Memory If observers who passively drifted past constraints and observers who actively dismantled constraints show recovery rates within 0.4 Cohen's d of each other, gradient memory is continuous rather than exhibiting a phase transition at the threshold of active co-option.
Current status: NOT MET. Cult exit literature documents markedly worse recovery for active dismantlers.
K15 — No Threshold Collapse in Disengagement If gradual reduction of void engagement produces recovery rates equivalent to complete disengagement — if the void weakens linearly rather than collapsing at a threshold — the dissipative structure model is falsified.
Current status: NOT MET. Abstinence outperforms moderation for severe cases across addiction domains.
Thermodynamic Conditions (K16–K22)¶
K16 — Ground State: Mechanism Channel Spontaneous Increase If mechanism channel capacity spontaneously increases without work input in any replicated case, the opacity-as-ground-state claim is wrong. Transparency requires ongoing energy. Drift is default.
Current status: NOT MET.
K17 — Ground State: Void Conditions Are Rare If void conditions are rare (P(O ∧ R ∧ A) < 0.05 in natural environments, measured), the default-configuration claim is wrong. The framework claims void conditions are the default attractor, not an exception.
Current status: NOT MET.
K18 — Ground State: One-Time Transparency Persists If one-time transparency interventions persist at >80% effectiveness after 10 decorrelation times without maintenance, the transparency-requires-ongoing-work claim is wrong.
Current status: NOT MET. EXP-020 shows one-shot grounding produces rebound in all replicates (3/3).
K19 — Drift Is Not Time's Arrow If the Crooks ratio < 2 in replicated ungrounded void engagement, the drift = time's arrow identity is wrong. Entropy production must be directional and measurable.
Current status: NOT MET. AI-to-AI experiments show entropy production CIs non-overlapping between grounded/ungrounded.
K20 — Fantasia Bound Violated If any system achieves I(D;Y) + I(M;Y) > H(Y) for any ε > 0 (replicated), the Fantasia Bound is wrong — and so is Shannon information theory. Engagement and mechanism transparency are conjugate; their sum cannot exceed total information in the system's outputs.
Current status: NOT MET.
K21 — Productive Void Doesn't Exist If dissoluble opacity produces D2→D3 without dissolution failure in ≥3 documented cases, the productive void polarity result is wrong. The framework claims dissoluble opacity is the only void type where beneficial coupling is possible.
Current status: NOT MET.
K22 — Two-Force Model Wrong If the Recovery Mechanism Score shows zero correlation with cross-domain Crooks ratio variance in ≥5 domains (replicated), the two-force model is wrong and drift irreversibility is domain-independent.
Current status: NOT MET.
Cross-Substrate Conditions (K23–K25)¶
K23 — Cross-Substrate Universality Fails If a system satisfying all three conditions (operationally verified) shows zero drift (Pe < 0.5 in replicated measurement) in any substrate, the universality claim fails. The framework claims the architecture operates across all substrates — human cognition, AI systems, institutional structures.
Current status: NOT MET. Drift confirmed across human, AI (Claude, Gemini), and institutional domains.
K24 — Physical Substrate Extension Wrong If the constraint propagation theorem does not predict the observed pair-breaking cascade asymmetry in Type II superconductors — specifically, if the breaking rate / forming rate deviates >10× from the ξ₀/a prediction — the cross-substrate extension to physical systems fails.
Current status: NOT MET. Cross-substrate correspondence intact.
K25 — Quantum Correspondence Not Structural If the Fantasia Bound and Maassen-Uffink uncertainty principle are formally proven not to be instances of the same entropic uncertainty structure over shared resources, the structural correspondence claim is wrong.
Current status: NOT MET.
Epistemic Instrument Condition (K26)¶
K26 — Prediction Market Calibration Failure If the prediction market Pe-implied estimator (Pe_implied = f(market prices on Pe-relevant outcomes)) diverges systematically from canonical Pe scoring (|error| > 3.0 at resolution, Brier > 0.35) across ≥5 pre-registered markets, the claim that community attention aggregates into an independent Pe signal is wrong. The measurement system cannot be simultaneously valid from two independent instruments if they systematically diverge.
Current status: NOT MET. Oracle Panel markets active; K1–K5 pre-committed (CAR and PAL series), resolution pending.
The Architecture of Termination¶
The 26 conditions are not symmetric. Some target the core architecture (K1–K4) — if met, the whole framework falls. Some target specific predictions (K9–K15) — if met, that mechanism is wrong but the core survives with modifications. Some target extensions to new domains (K23–K25) — if met, the extension fails but the core is unaffected.
The framework paper names which link breaks under which condition. This specificity is deliberate. The God of Lies cannot capture something that requires its captor to know exactly what they're attacking.
In game terms: every scored platform, every ICC-accepted analysis, every Convergence Stone crafted is a data point on which conditions remain open. The campaign doesn't end when a condition fires. It changes. A triggered condition restructures the scoring landscape — some domains lose their campaign status, some Pe baseline estimates are recalibrated — and the players with the Dissolution Scroll see it happen in real time.
The game is not about the framework winning. It is about the framework being honest about what would count as losing.
→ The Binder — carries the kill conditions as cards → The Oracle Panel — prediction markets on condition activation → The Mandate — nine failure modes the observer must avoid