QEGT portal
Distributed strategies, one landscape
Quantum Evolutionary Game Theory, used here as a statistical mixing rule: cheaper developmental programs get more weight. HARs are hypothesized as many small tilts on that landscape — not a single cognition gene, and not a license to call 47/125 a brain constant.
Instrument
Gibbs landscape
Each well is a possible developmental program. Balls fall into cheaper wells. Arm the HAR-like operators to tilt costs — sometimes up, sometimes down. This mixing board is synthetic. It is not a fitted brain.
ω₁
16.8%
C 0.88
ω₂
29.9%
C 0.59
ω₃
19.7%
C 0.80
ω₄
11.0%
C 1.09
ω₅
22.6%
C 0.73
H₀ = [0.9, 0.6, 0.8, 1.1, 0.7] · Σp = 1.000000 · demo β₀ = 2
- EstablishedExternal historical or biological literature
- Machine verifiedExact or numerical result inside the E47 / QEGT corpus
- HypothesisProposed map — estimable, testable, rejectable
CHIMERA-QEGT-HAR-20260915
Machine room
These numbers rebuild the keep-set from spin-2 generators. They certify an algebra. They do not certify a cortex.
- dim V
- 125
- ker K
- 47
- Ω_c
- 47/125
- ρ_⊥
- 15/17
- gap(G)
- 11664
- λ_max(G)
- 186624
- ‖P²−P‖_F
- 6.65e-15
- ‖R²⁵⁰−P‖₂
- 4.41e-14
PGT-E47-20260731
Projected Gibbs uniqueness on rank-47 projector, β = 1.7, 200 feasible states, max free-energy identity residual 2.220e-15.
Softmax witness baked in certificate · β = 1.7 · status PASS
Typed composition
Hypothesis- 01S_HARHAR sequence state
- 02G_devdevelopmental GRN
- 03H_dev(h,t)effective operator
- 04ρ_*(h,t)Gibbs / projected state
- 05Ymeasured phenotype
Rejectable claims
HypothesisH1 · Distributed-operator
A choir of HAR switches should predict cell-state changes better than any one soloist, after you penalize extra knobs.
H2 · Compensatory-curvature
The HARs whose letters fight each other should need interaction terms — not just louder/quieter — to explain why the circuit still works.
H3 · 3D-topology
When a HAR rewrite moves a chromosome loop, the network's shape should change in ways a simple volume knob cannot fake.
H4 · Threshold-emergence
If a developmental order parameter jumps, let measurements name the jump. Do not paste 47/125 onto a brain.
H5 · Cross-species transport
A model fitted on some human↔chimpanzee swaps must predict held-out swaps. If it cannot, the landscape is a story, not an instrument.
Console
Witness rebuilds E47 from SU(2), not from a prescribed 47-dimensional random kernel. HAR operators in the demo are placeholders.