Science

Complete reference. Axioms, formulas, predictions, tests, data, and falsification criteria.

Core claim

Every dissipative system — any system that maintains itself by spending energy — runs on exactly two irreducible operations (hold and cross), follows a mandatory four-regime cycle, and produces a self-similar hierarchy at every depth level. From this single structural claim, the entire framework is derived.

The claim is falsifiable. A single counter-example — a third irreducible operation, a dissipative system that does not follow the sequence, or a prediction that fails at sufficient sample size — collapses the framework.

Complete formula set

1. Effective potency

M_eff = M × (1 − D) / (1 + D)

M = raw potency (0–1). D = drain (0–1). The (1−D) reduces the agent's force. The (1+D) represents the opponent strengthened by what drains you. Hyperbolic: small drain costs little; past D = 0.5, the agent is effectively neutralised.

2. Phase drain

D_next = M_dominant × φ

φ = 0.3 (same-axis transition: hold→hold or cross→cross). φ = 0.6 (cross-axis transition: hold→cross or cross→hold). The prior regime's dominant agent's strength becomes the next regime's specific drag.

3. Observer cascade

M_eff_sub = M_sub × (1 − D_sub) / (1 + D_sub) × (1 − D_observer)

Observer drain reduces EVERY subordinate agent's effective potency. The only drain that cascades. Structural consequence of information topology: the Observer is the single point through which all coordination passes.

4. Coverage (system health)

Health = blockade(SP1) × blockade(SP2) × blockade(SP3) × blockade(SP4) blockade(sp) = 1 − ∏(1 − M_eff_i) for all agents i covering sub-phase sp

An uncovered sub-phase (blockade = 0) collapses the entire product. 91% of gains come from adding coverage to uncovered functions, not from increasing potency of covered ones. Validated across 21 disease domains with 258 agents.

5. Profile distance

The profile distance is agent-specific, not a fixed schedule. Each agent has a four-dimensional impulse profile — and the four dimensions are the four threshold events of the cycle itself:

Threshold	Dimension	Hold pole	Cross pole
Release (Potentiality → Construction)	Direction	Inward — energy toward self	Outward — energy toward environment
Exposure (Construction → Encounter)	Resolution	Concrete — what exists	Abstract — what could be
Integration (Encounter → Conservation)	Logic	Systematic — universal rules	Contextual — situation-specific
Dissolution (Conservation → Potentiality)	Structure	Structured — closes options	Emergent — keeps options open

Each threshold is a binary choice: hold or cross at that transition. An agent's profile IS their pattern of hold-or-cross decisions at the four transition points. The four dimensions are not borrowed from psychology — they are produced by the framework's own structure. Two operations, four thresholds, each threshold binary. 2&sup4; = 16 possible profiles, one per position.

Each of the 16 positions requires a specific binary profile. An agent's effectiveness at any position is the average closeness between their threshold responses and the position's requirements. From these four values, you can predict how well any agent performs at each of the 16 positions and each of the 5 functions.

effectiveness(agent, position) = mean( 1 - |target_d - agent_d| ) for d in [direction, resolution, logic, structure] profile_distance = (effectiveness_current / effectiveness_home - 0.5) × 2.0

Maps to: 1.0 at home position, 0.0 at average distance (neutral), negative when actively opposing. The decay rate is a property of the agent's specific profile, not a fixed regime-based schedule.

When impulse dimensions are not available, the regime-based approximation applies:

0 regimes past: × 1.0 (full strength) 1 regime past: × 0.5 (diminished) 2 regimes past: × 0.0 (neutral — present but useless) 3 regimes past: × −0.5 (pathological — actively harmful)

Fallback when agent-level impulse data is unavailable. The real penalty is computed from the agent's specific profile distance.

Profile distance and personality types

The four dimensions are the four threshold events of the cycle. 2&sup4; = 16 profiles, one per position. This is the same structure discovered independently by MBTI from the psychology side — the same sixteen types because the same 2&sup4; geometry.

The framework makes a specific prediction that MBTI does not: the effectiveness of a profile is regime-dependent. The same profile that produces high performance in Construction positions should produce low performance in Encounter positions. MBTI treats the type as fixed. The framework says the type is fixed but its value changes with the system's position.

Testable predictions:

Prediction	Test	Status
MBTI scores correlate with job performance only when regime is controlled for	Meta-analysis of MBTI-performance studies with regime coding	Testable
Adding regime context significantly improves MBTI-performance correlation	Compare r² with and without regime variable	Testable
Teams with all 16 profiles covered outperform teams with concentrated profiles	Team composition study with cycle-position controls	Testable
"Personality clashes" resolve when agents are placed in their home regime	Conflict resolution study with regime-based reassignment	Testable
The 16 MBTI types map 1:1 to the 16 position profiles	Compare MBTI binary coding with position profile matrix	Testable

If confirmed, this would mean MBTI did not discover personality types. It discovered the threshold structure of the dissipative cycle applied to human psychology. The same geometry, found from a different starting point.

6. Generative property

Φ = ∮ (dI/dS) · dθ > 0

The closed integral of informational order with respect to entropy, over one complete cycle, is strictly positive. Each cycle produces what was not there before. Confirmed in nanocrystal recrystallisation data (Brunner et al., 2020).

7. Baryon asymmetry

η = (2/π) × J²

Derived from the cycle geometry. Five Standard Model constants derived from integers and π, with zero free parameters.

8. Observer state cascade

State 1 (Attend): M_eff via standard formula State 2 (Drain): M_eff via standard formula (D elevated) State 3 (Withdraw): M_eff_cascade = 0.25 (discrete) State 4 (Confusion): M_eff_cascade = 0.10 (discrete)

The Observer degrades through four states in fixed order. States 1–2 follow the continuous M_eff formula. States 3–4 produce discrete cascade values. Recovery must reverse the sequence. Cannot skip states. Validated across 8 medical domains.

9. Operation matching

R_sp = (1 − M_eff × ω) × f(τ) × g(p) ω = 1.0 if agent operation matches cancer strategy at sub-phase ω = 0.7 if mismatch

Each agent operates as either HOLD (destroy, block, maintain) or CROSS (infiltrate, redirect, reconnect). Cancer runs a default strategy per sub-phase: SP1=CROSS, SP2=HOLD, SP3=CROSS, SP4=HOLD. Matching the operation to the cancer's strategy at each sub-phase improves prediction accuracy.

Predictions and tests

Physics

Prediction	Status
Structural axis correspondence (gravity/QCD beta functions differ by one power)	Confirmed
No proton decay (GUT leptoquark bosons do not exist)	Consistent with all data
Electroweak = complete unification of cross axis	Consistent
Phase transition taxonomy (4 types by cycle position)	Testable
Fractal regime structure at quantum/molecular/stellar/cosmological scales	384/384 positions
Cycle integral measurable (Φ > 0)	Confirmed (nanocrystals)
Non-uniform Φ distribution by regime	Testable
Process universality (16-step sequence across all substrates)	24 systems, 8 domains
Dark energy evolving (universe in Conservation drift)	4.2σ (DESI DR2)
Fine-structure constant drift	At precision edge (ESPRESSO)

Universal sequence

Metric	Result
Systems tested	24
Scientific domains	8
Positional matches (strict relational definitions)	384 / 384

Quantitative tests

System	Clean	Partial	Partial positions
Erythropoiesis	14/16	2/16	14–15 (Surveillance/Compensation)
Adaptive immune response	14/16	2/16	14–15 (Surveillance/Compensation)
Cancer hallmarks (Hanahan & Weinberg)	14/16	2/16	14–15 (Surveillance/Compensation)

The two partial scores consistently appear at positions 14–15 — the coupled Surveillance/Compensation pair — across three independent biological systems, consistent with the conservation degeneracy predicted by the framework.

Drug Response Simulator

Metric	Result
Evidence entries	184
Disease domains	21
Agents in database	258
Calibrated MAE	0.2 pts
Free parameters per trial	0
Patient-level AUC (NSCLC)	0.808

Climate intervention

System	MAE
EU Power sector	2.6 pts
Amazon Deforestation	0.3 pts
Urban Road transport	4.9 pts

Four-State Observer Protocol

Metric	Result
Medical domains tested	8
States confirmed per domain	4/4
Domains	Oncology, MS, T1D, Alzheimer’s, Sepsis, Heart failure, Liver cirrhosis, TB
Non-immune proof case	Liver cirrhosis (fibrotic cascade follows same four states)

The Observer function degrades through a fixed sequence: Attend → Drain → Withdraw → Confusion. Each state requires a different intervention. Recovery must reverse the sequence and cannot skip states. Most dangerous error: treating Confusion as Withdrawal.

Structure reference

The hierarchy

Depth	Count	Name	What is visible
1	4	Regimes	Which mode (hold/cross)
2	16	Positions	Which challenge
3	64	Agents	Who holds each position (4 phased functions)
4	256	Impulses	What drives each function

The 16 positions

#	Name	Regime	Mirror	Requires	Produces
01	Signal	Potentiality	P of P	Prior equilibrium	Recognised perturbation
02	Accumulation	Potentiality	C of P	Perturbation	Concentrated potential
03	Configuration	Potentiality	E of P	Potential	Testable configuration
04	Threshold	Potentiality	Co of P	Configuration	Committed system
05	Constraint	Construction	P of C	Commitment	Defined constraint
06	Architecture	Construction	C of C	Constraint	Route from source to form
07	Testing	Construction	E of C	Pathway	Tested configurations
08	Selection	Construction	Co of C	Tested configs	Defined identity
09	Manifestation	Encounter	P of E	Identity	Mutual visibility
10	Discovery	Encounter	C of E	Visibility	Encounter-dependent knowledge
11	Exchange	Encounter	E of E	New info	Sustained relationship
12	Equilibrium	Encounter	Co of E	Exchange	Winning formula
13	Differentiation	Conservation	P of Co	Equilibrium	Internal boundaries
14	Surveillance	Conservation	C of Co	Differentiation	Self-observation
15	Compensation	Conservation	E of Co	Observation	Active maintenance
16	Continuation	Conservation	Co of Co	Maintenance	Next perturbation

The four transitions

Symbol	Name	Direction	What activates	What dissolves
\|	Release	Potentiality → Construction	Hold	All other possibility
+	Exposure	Construction → Encounter	Cross	Builder's control
○	Integration	Encounter → Conservation	Hold deactivates	Novelty
−	Dissolution	Conservation → Potentiality	Cross deactivates	The form itself

Five agent functions

Function	Operation	Hold action	Cross action	Structural drain
Sentinel	Hold × same level	Slow	Consolidate	Off-target holding
Miner	Hold × deeper	Prevent	Provoke	Extraction depletion
Architect	Cross × same level	Transform	Accelerate	Off-target crossing
Catalyst	Cross × deeper	Regain control	Catalyse	Collateral acceleration
Observer	Surveillance loop	Withdraw	Attend	Cascades to ALL subordinates

Opposition pairs — the price of function

Each function is structurally opposed by the function on the same axis but the other operation. They compete for the same resource at the same depth. The opponent does not attack — it absorbs.

Function	Opponent	Axis	Price paid	In plain words
Architect	Sentinel	Same	Time	The sentinel holds continuously. The architect’s window is finite.
Sentinel	Architect	Same	Relevance	The architect makes the sentinel’s structure obsolete.
Catalyst	Miner	Deeper	Access	The miner has locked the deeper level shut.
Miner	Catalyst	Deeper	Adaptability	The catalyst forces a transition the miner cannot absorb.
Observer	—	Meta	Commitment	Breadth costs depth. Watching everything means seeing nothing.

Three observable levels

At any observation point, exactly three levels are accessible. Each has four parameters, one per regime. The same structure applies at every fractal depth.

System profile		Fixed per system type (e.g. per cancer type)
Delta	dt	Doubling time	“How fast do I cycle?”
Raido	ε_sub	Internal independence	“How do I build — redundantly or efficiently?”
Gebo	ε_top	Cross-phase independence	“How coupled are my defences?”
Sigma	β	Adaptation rate	“How fast do I learn to resist?”

System state		Fixed per clinical stage / situation
Delta	depth	Fractal depth	“How deep have I built?”
Raido	τ	Temporal maturity	“How long have I been building?”
Gebo	prior_lines	Treatment history	“How many times have I been attacked?”
Sigma	γ	Cross-resistance	“How much did I learn from each attack?”

Agent profile		Fixed per agent per system type
Delta	sub-phase	Target position	“Where do I strike?”
Raido	function	Structural mode	“How do I operate?”
Gebo	M	Potency	“How hard do I hit?”
Sigma	action	Maintenance mode	“What do I maintain?”
—	operation	Hold or Cross	“Which attack plane?”

The language boundary

An observer can see one level down, act one level down, and observe results. It cannot see into the agent. It cannot modify the agent from inside. The agent is a closed box — deploy it or don’t, but you cannot change what it is.

Reason: depth creates language specificity. Each level down, the vocabulary becomes more specific to that level’s material. After approximately two levels, the observer’s language and the agent’s language share no vocabulary. The observer cannot instruct what it cannot address. A doctor cannot adjust a drug’s binding affinity — not because of ignorance, but because “binding affinity” is not expressible in the doctor’s action vocabulary.

Consequence: coverage beats potency. You cannot make one agent better. You can only add another agent that covers what the first one misses. The only lever is combination — because the agent itself is untouchable.

Looking up: the same constraint applies. Whatever system we are agents of cannot instruct us in our language. Its intent reaches us as circumstance, constraint, selection pressure — never as direct instruction. We infer the pattern (the geometry), not the intent.

Papers

The Code Truth — Physics derivation

Five SM constants from integers and π. 15 predictions, 0 free parameters. Axial force pairing.

DOI: 10.5281/zenodo.19023438

First Principles of Change

Complete derivation of the framework from the two-operation starting observation.

DOI: zenodo.org/uploads/18967781

There Is Only One Way to Grow — Universal sequence

24 dissipative processes, 384/384 positional matches across 8 domains.

DOI: 10.5281/zenodo.18929428

Quantitative Tests of the Universal Sequence

Erythropoiesis, adaptive immune response, and cancer hallmarks. 14/16 clean, 2/16 partial at positions 14–15 across all three systems.

DOI: 10.5281/zenodo.18956282

The Geometry of Intervention

Drug response formula. 21 disease domains, 258 agents. Coverage beats potency.

DOI: 10.5281/zenodo.19022091

Coverage Beats Potency — Strike Order

91% of gains from coverage not potency. SP4 missed in 7/11 standard-of-care cases.

Fractal Depth Resolves Conservation — Conservation Drain

M_eff = M×(1−D)/(1+D). Phase-locked potency. Observer drain cascades.

The Four-State Observer Protocol

Attend → Drain → Withdraw → Confusion. 8 medical domains, 5 treatment protocols, fractal Observer hierarchy.

DOI: 10.5281/zenodo.19130570

FULCRUM — Predicting Immunotherapy Response from the Structural Theory of Dissipative Systems

Patient-level AUC 0.808 (NSCLC, n=159). Pan-cancer HR 0.935 (p=0.0002, n=9,966). 12/13 ICI datasets. Two measurements, zero fitted parameters.

DOI: 10.5281/zenodo.19399587

Observer Degradation in Alzheimer’s Disease

Microglial surveillance as Observer function. DAM phenotype = State 3 Withdrawal.

DOI: 10.5281/zenodo.19119749

Falsification criteria

The framework makes strong claims. Each can be broken by a specific counter-example:

1. A third irreducible operation. Something that is genuinely neither maintaining coherence (hold) nor enabling interaction (cross). If found, the two-axis structure fails and everything derived from it collapses.

2. A dissipative system that does not follow the 16-step sequence. A system that maintains itself by spending energy but whose events do not map to the mandatory order. 24 systems tested. None found.

3. A drug combination prediction with >5% MAE at sufficient sample size. The blockade formula predicts response rates from structural position alone. If leave-one-out predictions consistently deviate by more than 5% from observed rates, the coverage model is wrong.

4. Proton decay. The framework predicts it does not occur (the strong force relates to gravity by scale, not to the electroweak forces by symmetry embedding). If proton decay is observed, the axial pairing is wrong.

5. A stable fine-structure constant. If ELT-ANDES (expected ~2032) confirms zero variation in α at 10x current precision, the vacuum-drift prediction weakens. Not fatal (G and Λ are independent tests) but significant.

6. Observer recovery that skips a state. The framework predicts that a system in State 4 (Confusion) must pass through State 3 (Withdrawal) before reaching State 2 (Drain). If a clinical intervention restores full Observer function from Confusion without passing through Withdrawal, the cascade model is wrong.