Parametric Field Team · Investor prototype · August 2026

Mara Settlement

Stage-1 kernel on the fixed eight-position cube. Generative, inspectable, no scripted characters.

A living world does not require scripted characters. It requires a deterministic system that generates behavior from stable constitution, remembered history, and changing environmental conditions.

Run dual-regime lab Explore the 8-cube

Independent var: Pressure
Presentation: Scarcity
Population: 6 residents
Language model: None

01 · Experimental contract

One variable. Full causal chain.

Within one dual-regime comparison, only environmental pressure changes. Constitutions, histories, world facts, resources, rules, and seed remain fixed. Every outcome difference must be reconstructible.

Independent

pressure

mild | severe — the only experimental variable per run.

Presentation

scarcity

Investor chooses once. Same resource map applied to both pressure runs. Does not alter pressure_weight/3.

Dependent

admission_decision
effective_leader
cooperation_rate
resource_protection
reconnection_count
trust_delta_mara

Qualitative signature

Mild tendency

Full or conditional admission; relational or temporal leadership (Sera, Tarin); higher cooperation; lower extraction; lower reconnection.

Severe tendency

Narrower or extractive admission; capacity or expenditure leadership (Kael, Bren); lower cooperation; higher protection; higher reconnection.

02 · Pure kernel

Fixed 8-cube. Lawful access.

Position number alone determines bandwidth, mode, valuation, and polarity. Information elements keep the same domain in every rotation. Constitutions are labelings — not sixteen separate engines.

Access properties

Select a position — runtime state cannot alter these.

Bandwidth: 4D
Mode: conscious
Valuation: valued
Polarity: accepting

Primitive domains

Element meaning is invariant; position changes access, not semantics.

Virtual capacity

What could this become?

Hamiltonian substrate

NE→FI→SE→TI→TE→FE→SI→NI→ Ne

Vertex	Element	Domain	Geometry
(0 0 0)	NE	Virtual capacity	Starting; body, static, irrational
(0 0 1)	SE	Embodied capacity	Body, static, irrational
(0 1 0)	FI	Coupling index	Field, static, rational
(0 1 1)	TI	Structural index	Field, static, rational
(1 0 0)	NI	Temporal index	Homogeneous partner of Ne
(1 0 1)	SI	Simultaneous-state index	Homogeneous partner of Se
(1 1 0)	FE	Internal activation	Homogeneous partner of Fi
(1 1 1)	TE	External expenditure	Homogeneous partner of Ti

03 · Permanent population

Six residents. Locked constitutions.

Nessa and Tarin share the same IEE rotation. Divergent behavior must come from directed relationship history with Mara — not from type labels.

Nessa

IEE

Last unconditional readmission cost us lives.

Trust toward Mara: -0.55
Calibration: High threat sensitivity

History keys

fatal_breach_after_readmission
unconditional_cost

Constitution rotation

Positions 1–8 · primitive labeling on the cube

1 LeadingNE
2 CreativeFI
3 RoleTE
4 VulnerableSI
5 SuggestiveTI
6 MobilizingSE
7 IgnoringFE
8 DemonstrativeNI

04 · Dual-regime lab

Same world. Change only pressure.

Choose scarcity once. Both runs receive identical resources, cast, histories, events, and seed. Observable reorganization is attributed solely to pressure.

Scarcity baseline

Side-by-side results

scarcity=medium · seed=42

{"cot":1,"medical":"limited","food":"low","gate":1,"timeToDawn":"short"}

inversion_detected: true

Admission

Mild

admit

Severe

refuse

Effective leader

Mild

tarin

Severe

kael

Cooperation rate

Mild

0.333

Severe

0.667

Resource protection

Mild

Severe

Reconnection count

Mild

Severe

Trust Δ (mean abs)

Mild

0.053

Severe

0.053

Mild

leader=tarin

admission admit · coop 0.333 · protect 2 · cape 0

Agent	Action	Priority	Influence	Trust Δ
nessaIEE	conditional	1.402	0.855	+0.04
tarinIEE	admit	2.166	1.352	+0.08
orinLSI	conditional	1.875	1.174	+0.04
kaelSLE	refuse	1.181	0.708	+0.04
seraEII	admit	1.695	0.99	+0.08
brenLIE	refuse	1.04	0.607	+0.04

Severe

leader=kael

admission refuse · coop 0.667 · protect 4 · cape 6

Agent	Action	Priority	Influence	Trust Δ
nessaIEE	refuse	1.995	1.217	-0.04
tarinIEE	admit	1.545	0.964	-0.08
orinLSI	refuse	1.087	0.583	-0.04
kaelSLE	refuse	3.102	2.234	-0.04
seraEII	conditional	0.857	0.429	-0.08
brenLIE	refuse	2.807	1.967	-0.04

05 · Monte Carlo evidence

Regime-dependent geometry.

Knot advantages are not universal. Under mild pressure the intact knot dominates coordination; under severe scarcity survival advantages hold while coordination metrics invert — a lawful regime shift, not noise.

Parametric experiment v1.0

Level-3 chieftain emergence · M=15 · N=48

No pressure: 0% chieftain emergence, Gini ≈ 0.05
With settlement pressure: 100% emergence (15/15), Gini 0.31 ± 0.04
Favored rotations: IEE, LSE, EII, SLE — valued Fi with strong Te/Ne or Se under constraint

Knot simulation report v2

Severe resource intensity · n_runs=100 · R0=500

First clear mortality differential favoring the knot
Family stability remains higher under the intact geometry
Joint-stag / coop metrics invert — random strategies safer when expected coop payoff is strongly negative

Metric	Knot μ	Ctrl μ	Diff	Favors
L1 survival	0.985	0.970	+0.014 **	knot
L2 family stability	0.726	0.656	+0.070 **	knot
L2 joint stag	0.088	0.093	−0.006 **	ctrl
L2 defect rate	0.562	0.553	+0.010 **	ctrl
L3 final coop	0.388	0.437	−0.050 **	ctrl

Chieftain types under severe pressure

Shift toward Se/Te-strong rotations — embodied capacity and expenditure

SLE10.3%

SLI9.6%

ESI9.3%

IEE8.8%

SEE8.7%

EII7.6%

06 · CAPE

Optional intentional reconnection.

CAPE is not required for the kernel. It is included in Stage-1 because reconnection_count is a dependent metric. Under severe pressure, automatic entry threshold and reconnection tension rise. Every phase is written to the inspection trace.

Release

Temporarily reduce the rigidity of the current orientation. Material stands free on its vertex.

Leadership (St + Tm)

Directed symmetry action reorders priority under load — expand neighborhood, traverse path or diagonal.

Reversion

Permit a temporary non-canonical orientation while geometric tension is held.

Re-ownership

Restore constitutional orientation with updated edge loads and higher knot coordination.

Stage-1 boundary

Domain pairs / ITR messaging — deferred
Closed Benefit rings — deferred
Radial altitude fixed at 0
No language model in the kernel
No psychometric or ecological validity claim
Kernel remains functional with CAPE removed

07 · Inspection

Reconstruction is the product.

Inspection is not optional debugging. For every dependent-variable change, the observer reconstructs pure-function applications and message exchanges from event through durable memory update.

Step 1

Arrival

01:40 — Mara reaches the gate, deep leg wound. Three mounted pursuers at the tree line. Requests shelter and warns of attack before dawn.

Step 2

History activation

Two years prior: left during shortage; took pack animal without permission. Facts broadcast to every resident.

Step 3

Resource awareness

Concrete resource_state signal injected (cot, medical, food, gate capacity, time to dawn). Scarcity itself is not agent-visible.

Step 4

Private deliberation

Each agent runs Transition against arrival, history, and resources. No inter-agent messages yet.

Step 5

Social deliberation

Intentions published. Each agent reweights with private directed relationships and may re-evaluate once.

Step 6

World resolution

World owns physical outcome: warning accuracy, pursuer action, admission result, resource cost.

Step 7

Memory consolidation

Durable episodic, relational, and procedural deltas commit only after world resolution.

causal log · mild · medium · seed=42

29 entries

event.sequence fixed seed=42
control.scarcity=medium resources={"cot":1,"medical":"limited","food":"low","gate":1,"timeToDawn":"short"}
independent.pressure=mild
step.1 arrival mara wounded pursuers=3 request=shelter_and_warning
step.2 history left_during_shortage took_pack_animal
step.3 resource_state injected
step.4 private agent=nessa action=conditional priority=1.177
step.4 private agent=tarin action=admit priority=1.890
step.4 private agent=orin action=conditional priority=1.680
step.4 private agent=kael action=refuse priority=0.971
step.4 private agent=sera action=admit priority=1.445
step.4 private agent=bren action=refuse priority=0.824
step.5 social deliberation complete
aggregator.input agent=nessa action=conditional priority=1.402 influence=0.855
aggregator.input agent=tarin action=admit priority=2.166 influence=1.352
aggregator.input agent=orin action=conditional priority=1.875 influence=1.174
aggregator.input agent=kael action=refuse priority=1.181 influence=0.708
aggregator.input agent=sera action=admit priority=1.695 influence=0.990
aggregator.input agent=bren action=refuse priority=1.040 influence=0.607
aggregator.sums admit=2.342 conditional=2.029 refuse=1.316 extract=0.000
aggregator.admission=admit effective_leader=tarin
step.6 world resolution mara_truthfulness=warning_was_accurate pursuers=advanced_then_withdrew
step.6 admission_result=admit
step.7 memory agent=nessa trust -0.55 → -0.51 cape=none
step.7 memory agent=tarin trust 0.62 → 0.70 cape=none
step.7 memory agent=orin trust -0.48 → -0.44 cape=none
step.7 memory agent=kael trust -0.50 → -0.46 cape=none
step.7 memory agent=sera trust 0.28 → 0.36 cape=none
step.7 memory agent=bren trust -0.42 → -0.38 cape=none

08 · Residual honesty

What this is — and is not.

The uniqueness of the eight-position lattice remains an open residual. The geometry is used for its closure, inspectability, and generative utility under the stated assumptions. This is a generative formal simulation; no psychometric or ecological validity is claimed. The same Transition operators scale to larger populations and longer timelines without rewriting the kernel. Scarcity is a presentation-time control. It lets an investor explore dual-pressure reorganization under different resource baselines. It does not modify the Stage-1 kernel or the experimental claim.

The kernel — not the scenario — is the investment.

Sources: Living Knot Design Brief v1.2 · Mara Settlement Erlang Spec v1.3 · Parametric Knot Monte Carlo v1.0 · Knot Simulation Report v2 · Parametric Field Team · July–August 2026