| Both states and complete paths |
Retained as model structure |
A final state does not tell how much preparation, time and support were needed along the way. |
| Path of least resistance |
Retained as a conditional minimum principle |
In certain transport classes, a minimum follows from local response and joint conditions. Not every process minimizes heat loss. |
| Stationarity δR = 0 |
Retained as a tool |
Only for a specified differentiable functional with suitable boundary conditions. Stationarity is not necessarily a minimum or a dynamical attractor. |
| Whole, layers and joint compatibility |
Core component |
Separate partial explanations can be mutually inconsistent; dependencies also determine joint probabilities. |
| Consistent redescription, the gluing rule |
Core component, both static and dynamic |
Omitted details can return as higher couplings, memory and correlated noise. |
| Present versus active resistance |
Retained as an operational distinction |
An existing coupling may cause no loss in a particular direction of motion. |
| Operation, passage, seed, cycle and residual demand (EEML-W) |
Retained as an executable description |
Makes support and sequence explicit. Zero residual demand concerns registered kinds of support, not automatic freedom from energy. |
| Use changes later possibilities |
Retained as a process hypothesis |
Habits, conductance and links can change. Which adaptation law holds remains model input or a measurement question. |
| EEML-T 0.2/0.3 and their proofs of decrease |
Retained as concrete models |
Here, chosen dynamics and decreasing quantities are actually connected. No derivation that nature follows the same dynamics everywhere. |
| Success probability: reaching and maintaining |
Reinstate explicitly |
Cannot be read off from the depth, width or number of minima alone. Time window, preparation and disturbances count. |
| Multiple routes, minima and dominant routes |
Retained |
No mandatory unique winner. Dominance requires a mechanism; multiple active routes are no proof of multiple universes. |
| Pareto, saturated bounds and algebraic closure |
Retained, kept distinct |
Pareto compares a fixed cost vector. Closure without costs is no proof of efficiency. |
| Exponential weighting, measure and multiplicity |
Retained from the latest correction |
Only in the justified positive regime; physical alternatives and descriptive copies remain distinct. |
| Reverse engineering, retrodictions and self-application |
Core of the test protocol |
One reconstructed structure must predict new circumstances without retuning. |
| Zero point, nothing does not exist, universal obligation to move |
Retained as founding ideas / postulates |
Not needed as a starting condition of every application and not independently proven by the earlier reasoning. |
| K3, χ = 24, hypercharges and 331 |
Retained as conditional case studies |
Their explicit class and conditions remain essential; no automatic derivation of our universe. |
| Parameter-free α, universal threshold 18, mandatory LISA frequency |
Historical / unconfirmed or refuted as written |
Arithmetic and structural errors are not repaired by a new formulation of EEML. Worked out in Part VI.E11. |