#Status labels

#The four substantive layers

A heavily weighted state is not necessarily quickly reachable. A unique admitted candidate is not automatically selected by efficiency. A stationary distribution need not describe a process at a standstill.

#One working definition of efficiency

Efficiency is the comparison of costs chosen in advance for the same effect fixed in advance, under the same hard constraints and system boundary. Y is more efficient than X in the Pareto sense when no cost component is higher and at least one is lower. Alternatives with a cost trade-off remain incomparable without an additional preference. Reachability and time window limit which alternatives may be compared in practice.

This is layer II. A statistical weight belongs to layer III; a minimum rule as a model of motion belongs to layer IV. The historical label “100% efficient” for zero active resistance is a model convention, not a universal thermodynamic efficiency. Where efficiency in that sense is measured, the separate useful output/input ratio remains necessary.

The third row is deliberately limited: a more extensive closed model can also be Hamiltonian. A decreasing fuel supply can temporarily drive circulation, but does not by itself turn a perpetual cycle into an autonomous gradient flow.

#The marble: the same intuition, different process laws

A marble in a landscape with friction gives off energy and can come to rest in a valley. Without friction it can keep moving; an exact equilibrium also remains possible. With continuous drive, circulation can arise. With noise, a distribution of positions and transitions becomes relevant; in the quantum regime, amplitudes and a measurement model are needed. If the marble changes the hill through its motion, the process changes its own future possibilities: that is the back-reaction EEML explicitly wants to track. The regress goes further: the outflow requires coupled degrees of freedom and their state, the landscape requires material laws and interactions, and an arrow of time also requires preparation and possibly cosmological initial conditions. Multiplicity helps describe that outflow statistically, but does not derive it without dynamics. This picture orders the questions; it does not resolve the origin of those laws and initial conditions.

#What EEML prohibits: historical claim and current limits

Revision history: the author initially chose the general dominance prohibition as EEML’s own claim. On request, that status has now been revised to the conditional D1 test. The strong original formulation remains readable below; it has not been proven as a universal law by Pareto or the literature. A law must operationally exclude possible outcomes or probability distributions, but writing down a prohibition does not yet make that true.

#1. Dominated realizations and self-undermining operation: conditional claims

Before the assessment, fix effect E, hard constraints C, preparation z₀, admitted transitions and cost vector r. Let \(\mathcal R_C(z_0)\) be the reachable realizations determined by these. Then Y dominates a candidate X when:

E-equality can mean an effect class, described in advance, with tolerances. Equality, costs and differences must be assessable together with their measurement uncertainties.

H_D, the historical strong hypothesis about nature: if such a Y exists, X does not continue to exist in the named process regime. The Pareto definition by itself says only that X is dominated. That X disappears, or is not persistently realized, is the new content; it does not follow automatically from reachability or existing conservation laws. Nor does it follow from this that there is only one winner: several undominated alternatives can remain.

Process version, likewise to be investigated conditionally: an operation that destroys the conditions for its own continuation does not continue to exist if a reachable alternative carries the same effect and preserves those conditions. If that continuation is part of the effect, constraints and costs from the outset, this can be worked out as the same dominance rule on trajectories. Otherwise the process version adds a separate preference for continuation. Preserving conditions does not by itself guarantee actual continuation.

The recurrent laryngeal nerve was named by the author as a candidate counterexample. No anatomical or evolutionary comparison has been carried out here. A comparison requires a set of developmental and evolutionary transitions delimited in advance; a shorter connection drawn after the fact does not prove that it was reachable with the same package of functions. Nor may that delimitation be adjusted after the fact to exclude a counterexample.

Standing requirements for every version of the test; limitations are recorded visibly:

• Persistence: do we mean one uninterrupted structure, a recurring structure, a persistently positive occupation or a stable solution? These statements are different.
• Time: a claim of disappearance “at some point” gives no finite observation limit. For a finite test, a time limit, survival probability or other observable consequence must be fixed independently. A chosen time limit is an additional testable sharpening, not an already derived EEML timescale.
• Reachability: a path from z₀ to Y is not necessarily a path from the X that has meanwhile been realized to Y. Possibility with positive probability also does not mean an inevitable choice. Fix in advance which meaning is tested.
• Complete comparison: preparation, transition barriers, required external control and maintenance are included according to system boundaries and costs chosen in advance. New hidden costs may not be added after every miss.
• Deterministic or statistical claim: a strict prohibition is not replaced after the fact by “usually less likely”. A probabilistic version requires a bound specified in advance.

At finite ε, dominated candidates can retain positive stationary occupation. That is a concrete objection to a general prohibition on persistent occupation, not something that disappears by itself through “temporary visits”. D1 therefore distinguishes structure preservation, individual residence time, return and occupation. Under named replicator conditions an elimination theorem does hold; with noise, a Kramers waiting time can be relevant. Both are worked out below. P1 tests a separate concrete process law.

#2. Additional postulate: no complete blockage

The idea remains: no state without outgoing transitions, unless a hard constraint named in advance enforces this. What counts as a state and as a transition must be fixed as part of this.

Zero-point fluctuations, tunneling and the third law of thermodynamics are retained as proposed connections, not as proof of this general postulate. A quantum energy state can be stationary; tunneling depends on the dynamics, and the unattainability of absolute zero is not a general theorem about outgoing transitions. For stationary states, see Feynman, The Dependence of Amplitudes on Time. The postulate remains separately testable within a chosen level of description.

#3. What is not prohibited by EEML’s own claim

• Effect at zero active dissipative resistance: ongoing motion can be compatible with this. Absence of friction does not mean absence of all forces or support.
• Input without the chosen effect: allowed; the effect definition determines what counts as a result. Standstill in the presence of forces is possible and does not automatically perform mechanical work.
• Progress without continuous external supply: not generally prohibited by conservation laws; initial energy established earlier can be conserved.
• Producing net energy without a corresponding change or transfer of energy, where energy conservation holds: already excluded by that conservation law. The broader statement “every effect demonstrably requires support somewhere in its entire past” does not automatically follow from it. EEML does not claim this as its own result.

Current working agreement: use D1 for process conditions determined in advance and their testable consequences. The self-undermining variant likewise requires a time limit, closed bookkeeping and a reachability test. Without those conditions it is not a proven prohibition. H_R is a separate hypothesis about the conservation of fine-grained information. Neither has been newly confirmed empirically here.

#Closure and saturation: operational and per quantity

Constraint closure here means that a package of constraints fixes the requested outcome. Saturation means that a proven lower bound or upper bound is attained. These are the main focus of the current exact selection benchmarks. They are not the only exact results in the dossier: for example, the log-sum-exp identity and the Lyapunov derivation are likewise exact within their conditions. None of these derivations by itself confirms a universal EEML law of nature.

First fix the model class, constraints C, physical equivalences ∼ and the quantity f to be predicted. Define:

f must be invariant under the chosen equivalences. Equivalences remove duplicate descriptions, not different physical outcomes. The following then holds as a definition:

Uniqueness of the full configuration requires, more strongly, \(|\widehat\Omega_C|=1\). An empty solution set is an inconsistent package, not a successful closure.

Report as a closure profile:

• the dimension per solution component, where a suitable dimension is defined;
• the number of inequivalent solutions if the set is discrete, otherwise the components that are still free;
• the possible values or band for the f chosen in advance.

Dimension is a diagnostic of continuous freedom, not a complete closure measure. For \(\Omega=\{-1,+1\}\) the dimension is zero, but f(x)=x has two outcomes. For f(x)=x², however, the outcome is unique. Conversely, f can be fixed while other parameters remain continuously free. A free volume modulus therefore does not automatically refute closure for a coupling that is independent of that volume.

For regular equality constraints \(F:\mathbb R^n\to\mathbb R^m\), constant Jacobian rank r gives dimension n−r locally. At singular points, for inequalities, discrete constraints or infinite-dimensional spaces, a separate analysis is needed. A rank-deficient Jacobian is not in itself proof of free solutions; for example, x²=0 has one solution despite F′(0)=0.

Discrete closure and discrete minimality differ. If the given constraints leave one value of f, no additional minimum rule is needed for it. If several values remain and we let a cost minimum choose, that minimum rule is an extra model condition. A strict minimum can be a proven consequence of the chosen cost; that nature minimizes that cost does not automatically follow from it. Even a uniquely closed result remains dependent on the chosen class and constraints.

For the benchmarks, this means:

The operational research action is therefore: look for compatible saturation of independent proven bounds, determine all equality cases, and then check whether the requested quantity really has one value in them. Repeated formulations of the same constraint do not count as independent bounds. The concrete steps are in the test protocol.

#Eight retained working rules

These rules preserve the original direction with their current status; not all eight are proven laws of nature.

1. Ongoing operation (postulate and research question). EEML looks for processes that carry and continue effect. That complete standstill is universally impossible remains a postulate. A constant visible distribution can contain ongoing events.
2. The whole counts (consistency requirement). Include relevant relations, environment, history and system boundaries. Parts can fit separately and be jointly incompatible.
3. Development changes possibilities (process hypothesis). Use can change habits, conductance and future accessibility. The adaptation law must be specified or measured.
4. Compare at equal effect (protocol). Fix quality, resources, duration and tolerance. Less loss through doing nothing is not a better execution of a required effect.
5. Possibility, coherence, weight, success and motion differ (definition). A possible route need not be used; a probable final state need not be reached or maintained in time.
6. Consistent redescription (consistency requirement). Hidden structure continues to count through relevant consequences: effective couplings, memory, noise correlations or other terms.
7. Close the bookkeeping (domain-dependent requirement). Include supply, output, storage, maintenance and damage. Conservation laws are used where they physically hold. There is no demonstrated universal conservation law for ‘resistance’.
8. Reconstruct and test, including EEML itself (protocol). Compare alternatives, remove constraints as a test, record failures and predict new circumstances without retuning.

#Notation

Symbols are defined more precisely per example. Statistical cost, mechanical action and fluctuation rate functional are not the same quantity without a derivation. F is thermodynamic free energy only under additional physical identifications.

#Mechanical concretization: effort as force, effect as motion

Proposal by the author, 28 September 2026: “effort = force, effect = motion”. This gives a concrete mechanical research branch. For a test, we operationalize effort as the required driving force, measured in newtons, and effect as a motion agreed in advance with initial conditions, load and duration. The magnitude of the force, its maximum and its time integral are different possible comparison measures; the choice must be fixed in advance. This does not tacitly replace the broader historical concepts with an already proven universal identification.

In classical mechanics with constant mass, in an inertial frame, the net force determines the acceleration: \(\sum\mathbf F=m\,d\mathbf v/dt\). Driving force, opposing force and their sum must therefore be kept separate. At zero net force the velocity stays constant; that can be ongoing motion or rest. A force can also change only the direction. See Feynman, Newton’s Laws of Dynamics.

An explicit model is a single rectilinear direction of motion with linear friction:

For the same constant velocity \(v_*>0\), the following holds:

Less friction here requires less driving force for the same motion effect. At γ=0 the velocity already present can be maintained without driving force; from rest, no acceleration follows without net force. The preparation that supplied the initial velocity belongs in the bookkeeping when we compare the full process.

Force is not the same as energy use. The transferred mechanical work is \(W=\int\mathbf F_{\rm drive}\cdot d\mathbf r\); the power is \(\mathbf F_{\rm drive}\cdot\mathbf v\). A smaller force over a longer path need not mean less work. These quantities may be tracked alongside force without swapping the definition of effort after the fact. See Feynman, Work and Potential Energy.

The relation above is a consequence of Newton plus the chosen friction law. An EEML claim of its own then asks which route, structure or adaptation lowers the required force, and which new outcome this predicts in advance. The identification “force/motion” does not by itself yet derive the costs \(c_i\), preferences or process parameters of P1.

#EEML-W: operation in its own terms

CAN means that the required context is present; DOES names the executed transition; LEAVES names the context that remains. A seed is preparation, a cycle an executable sequence that restores a named context, and residual demand the additional support during execution.

A simple discrete translation uses a stock \(m\in\mathbb N^k\). An operation has required input u and output v:

A shortfall is componentwise \((u-m)_+\). Replenishment must be booked explicitly; the formula does not make it available by itself. This is related to a Petri net. A cycle must actually be executable. If every transition requires a missing seed, it does not start by itself.

Example: a direct operation consumes a and delivers one effect pulse. A cycle \(a\to b\to c\to a\) delivers one effect pulse at the last step. With one initial a and one time unit per step, ten direct pulses require nine replenishments and ten steps; the cycle requires no new a, but thirty steps. This is a trade-off between support and duration. Fuel or energy that is not included in the stock may not be regarded as free.

The kinship with Petri nets and chemical organization theory does not automatically make a cycle optimal, reachable or thermodynamically closed. See Dittrich and Speroni di Fenizio.

#Founding ideas and zero point

Retained as postulates: nothing does not exist; effortless effect is underlying; motivations, habits and goals co-determine the context; EEML must also be applied to its own formulation. The convention ‘100% effortless’ denotes zero active opposition with actual effect. It is not automatically the thermodynamic output/input efficiency or a proof of energy for free.

The original argument remains visible: assume total absence; then resistance is absent; add the founding rule that absent resistance necessarily implies an existing effortless state; that state is something; hence total absence conflicts with that founding rule. The transition from absence to the existence of a state is additional content. The argument does not independently prove the founding rule or a cosmic zero point.

‘Nothing’ as total absence and ‘naught’ as something present that has no effect within a context remain different concepts. Couplings that are present can be inactive; this is made operational further on with a damping matrix. The concrete models can be investigated without a theory of origin.

#Coherent work blocks

The four substantive layers remain admissibility, efficiency comparison, weight and process. The work blocks below add their context, mutual consistency and test protocol; they do not introduce six alternative fundamental laws.

#A. Context and effect

Fix the system boundary, observables, initial conditions, resources, effect criterion, duration and tolerance. Distinguish the physically allowed space Ω_C from the subset Ω_E,C that carries the effect.

A state space may not be called an “arbitrary set” and at the same time be given, without further structure, a tangent bundle, derivatives and Euler–Lagrange equations. Discrete operations, smooth trajectories and stochastic paths require their own formal objects. A diffusion path is usually not differentiable; the old notation ρ(γ,γ̇,t) is not directly usable there.

#B. Operation and progression

Describe what can happen now, what actually happens and what context it leaves behind. For this, EEML-W provides the distinctions CAN, DOES and LEAVES. A seed is preparation; a cycle restores a named context; residual demand counts additional support during execution.

For a sufficiently completely described Markov system, a transition kernel K or generator L can specify the progression. The initial distribution and the environment are also needed. If the chosen visible state has memory, an ordinary time-homogeneous Markov generator on that state does not suffice: extend the state or use explicit history dependence.

On a graph, a cycle is only a possible continuation. Reachability, positive transition rates, stability and any exit routes determine whether it actually keeps running. The older tropical cycle calculations remain models of repeated execution at specified costs and duration; they do not yield a constant of nature.

#C. Input, support and durability

Keep a cost vector as long as there is no substantiated conversion between kinds of input. Compare at equal effect and with the full relevant context. A cheap intermediate step that makes the next operation impossible must not win against a slightly more expensive preparation on the basis of its low local cost.

Where appropriate, a path cost can be built up from:

This is an accounting possibility, not a requirement that every kind of effort adds up locally. Hidden memory must have been made explicit, and contributions must not be counted twice. Where needed, final reserves, stored damage and future possibilities also belong in the comparison criterion.

#D. Weight

Where the positive statistical description is justified:

The measure, S, ε, normalization and the regime of application are explicit. When macro-classes are compared, the macro-measure ν is added, as worked out in Part IV. This does not yet give a free model parameter a physically derived probability distribution.

Under independent composition, both costs and measure must have the required factorization. Under dependence, the couplings remain. Positive weighting does not replace coherent quantum amplitudes; Born is not derived from the multiplicative Cauchy rule.

#E. Coherence between the components

Weight and process are tested jointly. If π is meant to be the stationary distribution of the chosen dynamics, then L*π = 0. That is a consistency condition, not a unique determination of L. Detailed balance is a stronger, non-universal requirement.

Where a stationary small-noise limit exists, the exponent in the specified weights must moreover match the global rate function of the dynamics, with the same measure and noise convention. This is the delimited Freidlin–Wentzell bridge. Equality of the quasipotential with S does not prove reversibility: Part V gives an exact rotating Gaussian process with the same S and nonzero cycle current. Hard physical constraints are retained; a chosen goal need not automatically be reached permanently.

Exact reduction must yield not only the same static weights but also the same relevant visible trajectory probabilities. Write P for the full trajectory law and O for the observation map. Then, for every relevant visible event A, the following must hold:

The resulting visible process can have memory. Static log-sum-exp regrouping is one particular component of this broader gluing rule. The existing relation between reduction, memory and fluctuations has been worked out, for example, by Mori (1965).

#F. Reconstruction and testing

Measure or specify several aspects of the same process: occupation, directed transitions, waiting times, response and resource use. Look for a jointly fitting model and keep alternatives that cannot yet be distinguished. Then test a different load, blockage or initial condition without recalibrating.

Keep minimal closing sets of conditions, negative controls, errors and undecided outcomes. Self-application means that EEML's cost choices and explanatory freedom are also tested in this way. It is not yet a proven unique metamathematical fixed point.

#Three especially valuable components from the old versions

#Genuine minimal resistance, with fixed effect

The old water/current intuition has a precise realization. For two fixed linear resistors R₁,R₂ > 0 and fixed total current I, it follows from current conservation and equal voltage that:

For a current-conserving deviation δ, the power difference is:

Here the minimum actually follows from the local laws and their coherence. The current uses both routes. This is Thomson's well-known principle in a simple case; see Doyle and Snell, §1.3.5.

The limit of this statement becomes apparent with nonlinear response: for different nonlinear responses, the stationary network minimizes a potential derived from the response, not necessarily the heating power. Precisely these results, both positive and negative, make it a useful benchmark.

#First concrete retrodiction: water through two parallel tubes

Known mechanical benchmark, added on 28 September 2026. Two rigid round tubes connect the same inlet and outlet, at equal height. Their inner radius r is the same; tube A has length L and tube B length 2L. We assume stationary, fully developed laminar flow of the same Newtonian fluid, no slip at the wall and negligible losses in the connections. The geometry is measured independently.

The well-known Hagen–Poiseuille law gives \[ R_i=\frac{8\mu L_i}{\pi r^4},\qquad Q_i=\frac{\Delta p}{R_i}. \] Therefore \(R_B=2R_A\). The common pressure drop and flow conservation fix the distribution: \[ \boxed{Q_A=\frac23Q_{\rm tot},\qquad Q_B=\frac13Q_{\rm tot}.} \] For a total throughput of 30 ml/min chosen in advance, we therefore predict 20 ml/min through A and 10 ml/min through B. That is the concrete comparison with an existing law of nature, not a new measurement. The local flow law and its conditions of validity are physical input; they are not derived from the word “effortless”. See OpenStax, Viscosity and Turbulence.

Direct connection to effort = force. Let an ideal piston with fixed area \(A_p\) deliver the same total throughput. The required pressure contribution to the driving force is \(F=A_p\Delta p\). With both tubes open, \(R_{\rm eq}=2R_A/3\). If we close only the longer tube B, then \(R_{\rm eq}=R_A\), so that in the new stationary state: \[ \boxed{\frac{F_{\rm B\ closed}}{F_{\rm both\ open}}=\frac32.} \] The same displaced quantity per unit time therefore requires 50% more piston force. Piston friction, connection losses and acceleration during switching are not included here. At fixed pressure instead of fixed throughput, the total flow would decrease; that is a different experiment.

Reading within EEML: both routes contribute to the same effect, with the largest contribution via the route with the lowest resistance. The longer route nevertheless lowers the required driving force of the whole. Under these conditions, “everything exclusively takes the easiest way” is therefore incorrect. At fixed total throughput, the quadratic loss \(P=R_AQ_A^2+R_BQ_B^2\) has its minimum precisely at this distribution, as the network identity above shows.

Evidential strength: an existing law is correctly reconstructed under its own hypotheses and read in EEML terms. The 2:1 ratio and the force factor 3/2 are testable, but they do not thereby yet support EEML independently of Poiseuille and flow conservation. This benchmark does not investigate spontaneous tube growth, memory or the process law of P1.

#Electrical benchmark: the same charge current with less heat generation

Known retrodiction, added on 28 September 2026. Take two parallel, ohmic resistors at constant temperature: \(R_A=10\,\Omega\) and \(R_B=20\,\Omega\), connected to a source that keeps the total current at 1 A. Connection losses and the switching transient are left outside this ideal stationary example. The effect is the same net charge transport per second, over the same measurement duration.

Ohm and current conservation predict: \[ I_A=\frac23\,{\rm A},\quad I_B=\frac13\,{\rm A},\quad U=\frac{20}{3}\,{\rm V},\quad R_{\rm eq}=\frac{20}{3}\,\Omega. \] The heat generation is: \[ P=R_AI_A^2+R_BI_B^2=\frac{20}{3}\,{\rm W} \simeq6.67\,{\rm W}. \] With only route A open, the same effect requires \(U=10\,{\rm V}\) and \(P=10\,{\rm W}\). The extra route therefore lowers the required voltage and heat generation by one third. Over 60 seconds, the conversion into heat is 400 J versus 600 J, with 60 C of net charge passed through in both cases. See OpenStax, Resistors in Series and Parallel.

The minimum follows exactly: for \(I_A=2/3\,{\rm A}+\delta I\) and \(I_B=1/3\,{\rm A}-\delta I\), we have \[ P-P_{\min}=30\,\Omega\,(\delta I)^2\ge0. \] This compares current distributions on the same network at the same total current. It does not prove that nature designed the network itself or spontaneously lays down a new wire.

Particles and force: in metals, electrons carry the current as a net drift. Their electric force is \(\mathbf F=-e\mathbf E\). Current, voltage and force are different quantities: voltage is energy per charge, while the local field determines the force. The energy flow is accounted for here as electrical power and converted into heat in the resistors; energy does not disappear. A current distribution is no proof that individual electrons plan ahead. See Feynman, Linear Systems and Review.

Condition that determines the outcome: at a fixed voltage of 10 V, adding B would instead raise the current from 1 to 1.5 A and the power from 10 to 15 W. Then the effect is not held equal. The energy saving of one third belongs explicitly to the comparison at fixed total current.

Status for EEML: an exactly known case of less electrical drive and less dissipation for the same effect. The local laws are input. This is a concrete benchmark for EEML and not an independent derivation of Ohm, electron physics or a universal minimum principle.

#Zero active resistance without empty structure

For a symmetric positive semidefinite damping matrix Γ:

Γ need not be zero. With P = γ(v₁−v₂)², the channel does not slow down a joint motion v₁ = v₂, but it does slow down a relative motion. This preserves a substantive part of the zero-point idea without presupposing an origin of the universe.

A cost shift S → S + c creates a zero on a score axis; it does not switch off any physical dissipation channel. Likewise, ε → 0 is a different operation. The operational zero point, the zero point of an accounting cost and the sharp-selection limit remain separate.

Under its conditions, the older thermal realization K1 connects motion and weight:

For constant J,D, suitable boundary conditions and finite Z, e^(−H/ε)/Z is invariant. H, mobilities and thermal context remain input; uniqueness and attraction require more. Conservative motion can continue without active dissipation, but the same class can also allow rest. See the related matrix framework of Ma, Chen and Fox.

#Habits and structural memory

EEML-T 0.3 uses route shares x, learned preferences h and the current target distribution p, each on the probability simplex. The parameters κ, η and ε are positive; Φ is a differentiable potential of x. The chosen adaptation law reads:

With \(D_{\rm KL}(x\|h)=\sum_i x_i\log(x_i/h_i)\), positive shares, fixed conditions and Φ independent of h and time, Ψ = Φ + εD_KL(x‖h) gives:

Under an imposed blockage, the conditions change; Ψ may then jump between phases. The decrease holds within each phase that respects the hypotheses. A concrete restriction to open routes, with memory retained over all routes, is worked out in P1; an arbitrary renormalization or change of costs does not automatically fall under the proof. This is a precise example of “use changes what is effortless later”. Changing preference is not the same as changing physical resistance. Fixed preference, learning preference and changing conductance must be compared separately. The formula is model input with a derived consequence, not yet a validated universal learning law.

#Restored success probability: a necessary addition

Choosing the effect in advance and comparing only successful outcomes is useful for design. But it does not predict how often success occurs. For that, physically possible failures must remain in the probability space.

For trajectory space Γ_C and success event E:

When a justified positive path weighting is used, this can be written as Z_E,C/Z_C. The path measure and any path cost must then have been specified independently. Normalizing only over Γ_E,C gives P(E|E,C) = 1, not the unknown success probability. With four equally weighted paths of which one succeeds, that probability is 1/4.

The following control example makes the difference concrete. Take A ↔ B with rate k in both directions. All k > 0 have the same stationary occupation (1/2,1/2). Start in A. Success means: the first arrival in B no later than T, then remaining there without interruption for Δ. A failed first arrival is not repaired by later attempts. Then:

For T = 1 and Δ = 0.5:

Moving fast helps with arrival, but here it harms the subsequent stay. The maximum within this family lies at k = ln((T+Δ)/Δ)/T. That is a mathematical optimum for the chosen criterion; no mechanism has been added that makes k move to k by itself.

Why this must return: equal weights and even the same possible transitions do not yet give an equal probability of a usable effect that is sustained for long enough. The latest statistical layer alone would miss this difference.

#Corrections on measure, minima and stability

The following four corrections delimit the statistical formulation. MPP stands for the Multiple Point Principle, discussed in Part IV.

#The convexity theorem holds under fixed conditions

Take a fixed space, a λ-independent measure μ, fixed ε > 0 and Sλ = S₀ + λI. If Z is finite and may be differentiated twice under the integral:

With a flat macro-measure on a closed bounded interval, a maximum of Z is attained at an endpoint; an interior stationary point is a minimum, possibly not strict. The crossing example in Part IV is a concrete illustration of this.

The conclusion “MPP can come only from ν in this formalism” is, however, too general. For nonlinear parameter dependence, with otherwise the same conditions:

The first term can be positive. A simple counterexample is S₁ = λ², S₂ = 2λ² on [−1,1]. At λ = 0 the minima are equal and Z = e^(−λ²/ε) + e^(−2λ²/ε) is maximal, with flat ν. F''(0) = 3. This does not prove a physical MPP, but it refutes the universal exclusion of a preference from Z. A changing measure or support can also invalidate the affine theorem.

The recorded position remains: MPP does not follow from the current core; a separately justified parameter selection mechanism is needed. The affine canonical route with fixed measure has the above limitation.

#Refinement: genuine coefficients, parameter maps and a negative control

The additional review rightly notes that \(S_1=\lambda^2,\ S_2=2\lambda^2\) becomes affine in the coefficient \(g=\lambda^2\). On \(g\in[0,1]\), the preferred value \(g=0\) is a boundary point. This example therefore does not refute the affine coefficient theorem. It does, however, limit statements about arbitrary parameter maps. That a mathematical reparametrization is possible does not by itself prove which parameter or measure is physically fundamental.

For fixed space and measure, finite normalization and \(S_g=S_0+\sum_a g_a I_a\), the following holds:

The Hessian is positive semidefinite. On a compact convex coefficient range, a maximum can also be found at an extreme point; in flat cases, interior points can also be maximizers. This is not a statement about all coupled, nonconvex parameter sets, parameter-dependent measures or infinite partition functions.

Negative control: one positive cost contribution. If \(S_g=S_0+gI\), \(I\ge0\), \(g\ge0\), and the fixed measure and differentiation conditions hold, then:

With finite Z at g=0 and a flat macro-measure in g, the density is maximized there. The finite example \(S=(0,g)\), ε=1, gives \(Z(g)=1+e^{-g}\). If an application independently needs a nonzero g, this bare rule therefore does not select it. This is a negative model control, not a new measurement.

For \(I\le0\), the same formula instead gives a nondecreasing Z. ‘The maximum lies on a boundary’ requires that a suitable maximum exists: for example on a closed bounded interval with finite continuous Z. For an operator without fixed sign, \(\langle I\rangle\) can change sign; with fixed measure and affine costs, log Z remains convex, so that an interior stationary point is not a strict maximum.

On an unbounded or open range, there may be only a supremum at a missing boundary, or the normalization over couplings may diverge. For example, \(Z(g)=1+e^{-g}\) is finite for every \(g\ge0\), but \(\int_0^\infty Z(g)\,dg=\infty\). A flat macro-measure therefore yields no normalized probability distribution at all there. If the effect criterion or the allowed set itself depends on g, the derivative at fixed space does not automatically hold either. A non-flat ν can change the probability density without changing Z.

It does not follow from this that all canonically weighted theories are free and symmetric. Other interactions can remain in \(S_0\); a zero coefficient can give a divergent Z; and zero in an action coefficient is not automatically zero in every customary physical coupling convention. The regularized positive \(\phi^4\) contribution is a suitable example only if the other assumptions, including normalization at the boundary, also hold.

The macro-measure remains essential. A flat measure on \(\lambda\in[-1,1]\) becomes, under \(g=\lambda^2\) with both branches included, proportional to \(g^{-1/2}dg\), not to dg. Silently taking a flat measure in both coordinates changes the model. Moreover, for a continuous distribution, a highest density is not a probability of one at that point.

Decision for the core: \(P(dg)\propto\nu(dg)Z_g\) remains a possible, conditional macro-ensemble description. It does not become a general law of nature that chooses couplings. An effect requirement can restrict the allowed range, but does not by itself determine a single nonzero coupling. The measure, the effect and the selection process must be justified independently.

#Width is not the same as instability

A positive quadratic minimum S = κx²/2 is stable under the chosen dynamics ẋ = −κx for every κ > 0. Small κ gives a wide minimum and slow return, not negative stability. Escape under noise also requires barriers, mobility and an observation time.

There can be a trade-off between statistical weight, recovery rate and reliability. There is no necessary conflict between wide minima and stability. Therefore, use separate quantities for curvature, return time, escape probability and measure volume. The control example confirms this for κ = 0.01 and κ = 100.

#Noncompact moduli do not automatically mean infinite measure

A noncompact space can have finite or infinite measure. Even with S = 0 on the whole real line, ∫dx diverges, while ∫dx/[π(1+x²)] = 1. The cost and the space are the same; the measures differ.

Therefore, “infinite basin measure” does not follow merely from the number of K3 moduli or from noncompactness. For that, the precise space, quotients, volume treatment, measure and any potential must be specified. The Cauchy measure from the example is not proposed as a physical K3 measure. Normalizability remains a condition to be checked, not an impossibility proven in advance.

#Measure volume, reachability and efficiency remain separate

“Largest measure-weighted accessible volume” can describe statistical weight. It is not automatically maximal efficiency, and “accessible” does not follow from an equilibrium integral. The actual initial preparation and transition law determine whether that volume is reached within the relevant time. An ordinary generator K or L, too, contains memory only if the state has been extended sufficiently for it.

#What remains as a founding idea, and what does not return as proof

The original ideas remain visible: nothing does not exist; effortless effect as the underlying idea; motivations, habits and goals as context; self-application; and the convention that “100% effortless” means zero active opposition with actual effect. Continuing operation remains a proposed universal property. These statements are not a consequence of the partition function and need not silently disappear in order to use the concrete models.

Not returning as a proven core:

• “Zero resistance necessarily causes motion.” A drive or other dynamical structure remains necessary; zero active dissipation also allows rest.
• “Everything always reaches the global minimum” or “every standstill is unstable”. Existing model counterexamples still stand.
• “Existence is positive resistance” as a necessary and sufficient condition; this has not been shown and does not fit the retained zero-active-resistance states.
• A universal conservation of resistance, a unique resistance density per domain or a self-proven unique meta-law.
• Topological stationarity as moduli stabilization: a topological invariant is constant within a fixed sector and does not choose a volume or coupling there.
• Multiple minima as a derivation of superposition, or a “zero-effort limit” as a derivation of the zero point.
• The old 137, 18, Higgs, dark-sector and LISA numbers without the missing valid calculations.

K3 and the particle benchmarks remain instructive case studies with their checked conditions. They do not determine the general architecture in advance.

#What now yields the most in substance

The most concrete connection is EEML-W for executability and support + a tested process law + the gluing rule for the same observations at multiple scales. EEML-T 0.3 and K1 are different available model realizations; they must not be merged into one universal equation without derivation.

The appropriate research direction is: reconstruct transition counts and residence times, and then predict a new blockage. Besides occupation, the next physical test must also measure events, waiting times and support. For a new EEML claim, one then needs to know which extra relation EEML predicts on top of the reference model.

Retained core: carrying effect, continuing operation, comparing required input and support, safeguarding coherence and predicting a new consequence. Open heart: which independently motivated relation between these components nature actually follows and existing separate domain models do not yet impose.

#What is retained

The distinction between possibilities, joint consistency, costs and weight remains the core. The following elements can also stay:

• An effect and the model class under study are described in advance.
• Hard constraints are checked jointly.
• With several cost components, the Pareto frontier shows which candidates have avoidable overhead, without immediately choosing an exchange rate between components.
• A proven and attained lower bound is an exact minimum within the stated class.
• Algebraic or discrete closure can determine a ratio or choice without minimizing costs.

Two formulations need precision. Pareto is independent of a scalar weighting for a fixed cost vector, not independent of the choice of costs. And closure by constraints is not in itself proof of efficiency: a class can leave a single candidate without any cost criterion doing any work.

The K3 result remains the already verified conditional example: for closed, smooth, connected, oriented spin four-manifolds with a Riemannian Einstein metric and nonzero signature, one has χ ≥ 3|σ|/2 ≥ 24. K3 attains the bound. The physical choice of this class and of χ as selection criterion is still needed separately. The proof and the sources are in Part VI.

#Multiplicity: counting real alternatives

For separate, physically distinct possibilities with positive base measures μᵢ, finite costs Sᵢ and ε > 0:

S and ε have the same units; Z must be dimensionless relative to a fixed reference measure. For k states counted with equal weight and with the same cost S*, one has exactly:

for μᵢ = 1. Two such states against one yield a ratio of 2:1 for equal macro base measures. The outcome with two states gets probability 2/3, not certainty.

This lowers the effective cost of the combined outcome, not the physical cost S* of each individual process. No energy appears for free. If S is energy and ε = kBT, this correction has the familiar entropic meaning. In other applications the meaning of S and ε has to be established.

A second name for the same physical possibility must not increase the weight. If one possibility with measure μ is split into seven descriptions, their measures must still add up to μ. Giving each copy the full measure again artificially adds a factor of seven. The computational check confirms that measure-preserving relabeling preserves F exactly.

We therefore replace “equivalent realizations” with: distinguishable physical realizations of the same macro-effect, counted with a specified measure. Descriptive redundancy and gauge equivalence are not extra physical chances.

#The macro-measure belongs in the full formula

A precise statistical working version is:

Here ν is the base measure over macro-outcomes and μλ the measure within each outcome. The joint measure must be fixed in advance; if you normalize μλ separately, the removed total measure must reappear in ν. Otherwise you change the prediction.

The formula requires a finite, positive total normalization. Macro-classes must form a consistent partition. Overlapping classes must not silently double-count. For continuous macro-variables a suitable decomposition of the joint measure is needed; simply integrating over a geometric surface of measure zero is not enough.

Moreover, a model parameter that can be chosen freely is not automatically a random variable of nature. A partition function at a given λ is no proof of a physical ensemble in which λ varies. That requires an extra ensemble, dynamics or an explicit model prior. For continuous λ, what matters are probability densities relative to ν and probabilities of regions. Under a continuous distribution a single point has probability zero; the highest density is not defined without a reference measure either.

Numerical example: two equal minima against one give, for equal ν, a probability of 2/3 for the first class. With macro base measures 0.1 and 0.9 this becomes 2/11 ≈ 0.182. So the measure demonstrably does substantive work.

#How far does the relation with the Multiple Point Principle reach?

The physical Multiple Point Principle (MPP) concerns the hypothesis that couplings allow different vacuum phases with equal, approximately zero vacuum energy density. That is more than counting equally costly possibilities. See Froggatt et al., The Two Higgs Doublet Model and the Multiple Point Principle.

The conditional calculation is correct within a restricted class. For finitely many discrete states, a common minimal cost S*, positive measures and a positive gap Δλ to higher costs, one has:

Hence:

With equal individual measures, Mλ is proportional to the number of minima kλ. If only these minima are included, the counting rule is exact. If higher states are present, it is an asymptotic approximation, not a general formula at finite ε.

The usable conclusion is: with equal minimal costs, equal measures per state, equal macro base measures and negligible contributions above the minimum, the class with more minima weighs more. That is not yet a mechanism that first makes the vacua equal in energy, fixes their energy at zero, or lets the couplings evolve toward such a point.

#Check A: one broad minimum beats two narrow ones

Take dimensionless x and measure dx. Model A consists of two disjoint copies of the real line, each with S(x) = 100x²/2. Model B consists of one copy with S(x) = x²/2. All minima have zero cost. The Gaussian integrals give exactly:

With equal macro base measures, one broad minimum weighs five times as much as two narrow minima. This does not refute the discrete counting theorem: it shows that “equal measure per minimum” has additional content for continuous landscapes. The neighborhood of each minimum also counts.

More generally, for small ε, isolated, nondegenerate minima in the same dimension d yield the following Laplace approximation, provided the measure is sufficiently regular there and the integral is dominated by the neighborhoods of these minima:

H is the positive Hessian of S and ρ the density of the measure. Flat directions or continuous sets of minima require a different treatment. Counting k alone is then insufficient.

#Check B: the point with equal minima can in fact be unfavorable

Take two states with positive costs:

At λ = 0 both minima are equal. Yet:

F has its maximum at λ = 0. With a flat macro-measure, the probability density is smallest there. For ε = 0.5 the density at the endpoints is 3.762 times as large.

This example deliberately violates the condition that Smin is independent of λ. It therefore does not refute the conditional formula above. It shows why that condition must be derived physically before the formula can serve as an MPP mechanism.

#The limit ε → 0 means sharp selection

For a fixed finite set with fixed positive measures:

That minimum can be 3 or any other number. ε → 0 does not mean zero effort, zero resistance or the philosophical zero point. Likewise, “minimally sufficient structure” follows only if S actually measures that structure. Nor does a non-sharp lower bound need to be saturated.

With equal minima, the multiplicity can remain visible in the probabilities while it disappears from the limiting value of F. For two minima of cost 3 against one:

Looking only at the limiting value of F therefore loses information. For infinite systems, moreover, the measure, the growth of state counts and the order of limits matter, for example when the number of states grows exponentially with system size N. Under suitable hypotheses the free-energy density can then take the form \(f(\varepsilon)=\inf_e[e-\varepsilon s(e)]\), with s the entropy density. The limits \(\varepsilon\to0\) and \(N\to\infty\) must not be silently interchanged.

#Making composition and self-application precise

A positive, nontrivial weight function that multiplies under independent composition and satisfies suitable regularity has an exponential form. The decreasing choice e^(−S/ε) additionally requires that higher S receives less weight. The additive costs and the independence are hypotheses that must be justified for each application.

For independent systems with additive costs, a product measure and the same ε, it follows exactly that:

There is also a precise form of consistent redescription. For a disjoint partition into blocks Bⱼ, define:

Then, with the measure already included in Fⱼ:

Bundling first and weighting afterwards gives the same total as weighting directly. This follows from the associativity of addition; it is a checkable consistency requirement, not a new law of nature.

“EEML(EEML) = EEML” remains a program for now, not a defined mathematical fixed-point equation. It lacks an operator with a domain and a meaning of equality. The usable self-application rule reads:

Apply consistency checking, ablation and the removal of demonstrably superfluous assumptions to EEML itself as well. In doing so, do not silently change the effect that is to be explained.

Nor is the positive counting rule automatically a quantum rule. Coherent alternatives require the addition of amplitudes and can cancel each other out. Born and the physical choice of a coherent or positive weighting remain additional structure.

#The course of the process remains a separate layer

A weight does not say how quickly a state is reached, which barriers lie along the way, or whether there is a net current. Even the same stationary distribution does not determine that current.

Check on a ring with three states:

Events occur in both models; zero net current is not an absence of events. The statistical landscape is the same; the process differs. These Markov models are mathematical counterexamples, not a thermodynamic explanation of the drive.

That is why the process idea of EEML is explicitly retained: reachability, transition rates, memory, adaptation and drive must be added or derived. A static sum over possibilities does not replace that layer.

#The two editorial corrections

Costs and weight stay separate. Measure-weighted volume belongs to statistical weight. The effective free-energy-like cost is its negative logarithm; volume and free energy are therefore not literally the same quantity. Pareto and proven cost bounds belong to the efficiency comparison. Closure without costs is a separate consistency result, not a proof of efficiency.

The identity is called exact regrouping. It holds for a disjoint partition with the measure correctly included. An exact marginalization can likewise preserve the weights of the retained variables exactly, even though the omitted variables cannot be reconstructed. That does not make the reduced description fully information-preserving.

The formulation “genuine coarse-graining produces entropy and the identity does not hold for it” is too general. Changing a description is not in itself a physical act of dissipation. In an equilibrium distribution, two hidden equilibrium states can be merged into one visible label without this giving rise to heat production. An approximation or Markov closure can, however, lose predictions; effective entropy balances must be derived separately. That is why we keep both exact regrouping and the broader dynamical gluing rule.

#The first consistency requirement is stationarity

When a model claims that π is the stationary distribution of L, the following must hold:

with the specified boundary conditions. This requirement is exact, but it does not determine L uniquely. Away from stationarity, the distribution and the dynamics must respect the full evolution equation. If S is only a design or performance cost, it need not be a stationary potential: that interpretation must not be added without a model hypothesis.

For a finite irreducible Markov chain with links allowed in both directions and positive rates, the cycle criterion reads:

Ordinary detailed balance then holds precisely when all cycle affinities are zero. Stationarity alone allows nonzero cycle currents. Time reversal of, for example, momentum variables requires the formulation appropriate to that case. See Joshi, Theorem 4.8 and Schnakenberg's network formulation.

#Freidlin–Wentzell with the correct normalization

Choose the convention explicitly:

Under suitable regularity and positive diffusion, small-noise theory describes rare paths via the rate functional:

The quasipotential relative to a given attractor is obtained by minimizing this cost over paths and durations starting from that attractor. This is a minimal fluctuation cost, not automatically physical heat, human effort or a new universal resistance. For the applications and their conditions, see Bouchet and Reygner, §2.

With a single suitable attracting minimum, a = I and b = −∇S, the quasipotential in this convention is S−Smin. With suitable confinement and boundary conditions, the Gibbs distribution exp(−S/ε)/Z is even exactly invariant.

Another common convention uses √ε instead of √(2ε). In that notation the exponent is exp(−2S/ε) and the quasipotential 2(S−Smin). That difference is a noise normalization, but it must not disappear while ε keeps the same name. Direct example: dX = −Xdt + √ε dW has stationary variance ε/2; with √(2ε) it is ε.

With several attractors, the global stationary rate function also requires the structure of transitions between them. A single local quasipotential from one attractor is not simply the global stationary exponent. Nor are prefactors and measure factors determined by the exponent alone.

Valid coupling: when a family π_ε ∝ exp(−S/ε)μ_ε is meant to describe the stationary weights, the measure carries no extra exponential ε-dependence, and the required stationary large-deviation limit exists, S−min S must agree with that global rate function. At finite ε, prefactors and the full equation L*π = 0 remain necessary. Small-noise theory does not yet choose a unique generator.

#Exact counterexample: the same potential, yet with a cycle current

Take, in two ordinary position coordinates:

The drift is not a Euclidean gradient when ω ≠ 0. Yet for every ω the same Gaussian distribution is invariant:

A direct check of the stationary covariance gives:

The stationary probability current is:

It has zero divergence, but for ω ≠ 0 it is not zero everywhere. The quasipotential remains S. The corresponding Hamilton–Jacobi identity also holds:

Thus “S is the quasipotential” does not imply “the process is gradient-like and all affinities are zero”. These are different hypotheses. This example is the simplest concrete relative of the older EEML-K1 structure with conservative and dissipative parts. Non-reversible processes with a preserved Gibbs measure have also been studied explicitly by Rey-Bellet and Spiliopoulos.

#What hard constraints must preserve

If Ω_C is the physically admissible space for the full process, the process must respect it. For a transition kernel this means, where applicable:

This does not automatically hold for Ω_E,C. Reaching a target and then losing it can be physically allowed. The reaching-and-maintaining example in Part III shows this explicitly. E must be a separately proven invariant or an explicitly absorbing criterion before the effect set is declared invariant.

For an open subsystem, the environment or a suitable time-dependent constraint must be included. An apparent violation of the subsystem bookkeeping is not automatically a violation of conservation in the full system.

Likewise, a transition between abstract model classes is not automatically a transition in spatial topology. The Geroch/Tipler route may therefore not be used as a general requirement “every transition between closure classes requires a singularity or quantum mechanics”. Which topological theorem is relevant depends on the geometric object and on the theorem's hypotheses; that extra identification is missing from the abstract EEML architecture.

#Ricci flow and self-organised localisation

Ricci flow. Perelman gives a gradient interpretation with an additional function f and a suitable treatment of diffeomorphisms. Along the coupled flow, his functional satisfies:

In that convention it is therefore not a decreasing positive resistance. Moreover, the deterministic geometric flow does not yet specify a noise measure, stationary ensemble or Freidlin–Wentzell problem on the space of metrics. “Perelman's F is exactly the quasipotential” is not shown by this. See Perelman, §1.

Self-organised localisation. The cited proposal contains inflation, stochastic fields, phase transitions and a volume-weighted Fokker–Planck description. The volume growth is extra structure; this is not simply an ordinary probability-conserving generator with L*π = 0. It can be investigated as a separate model route, not as an already proven EEML bridge. See Giudice, McCullough and You.

#Registering hypotheses H1 and H2 more sharply

H1 is split.

• H1a: the independently determined S yields, in the designated small-noise regime, the correct stationary rate function.
• H1b: the process is reversible under the specified time reversal; for the ordinary finite Markov description, all cycle affinities are zero.

The rotating Gaussian example satisfies H1a and violates H1b. Cycle currents test H1b, not automatically H1a. A driven biological experiment that nevertheless yields no visible current is not by definition a faulty measurement: there may be resolution limits, hidden currents or genuinely different conditions. The test outcome must be able to surprise.

H2 as supplied is not simply Prigogine's principle. Under suitable near-equilibrium conditions, that principle characterizes a stationary state for given kinetics and external constraints. It does not state in general that nature chooses the least dissipative generator out of all possible ones. The precise Markov derivation and its approximate character can be found in Maes and Netočný.

The check example uses a three-state ring with rates a = 3c and b = c. Then the stationary probabilities stay equal and the cycle affinity 3 ln 3 stays constant, but:

Making c small lowers the entropy production per unit time, while transport also slows down. Fixing only the drive does not prevent this trivial reduction. A selection hypothesis about generators must therefore also fix the effect, time frame, kinetic constraints and resources. That ties in exactly with the older EEML rule of equal effect.

#Corrected process rule for the integrated version

Course and coherence. Specify the process law, initial preparation and environment. When memory is visible, use a sufficiently extended state or an explicitly history-dependent law. The process respects the applicable hard physical constraints; a target set is not automatically invariant. When statistical weights are given as a stationary distribution, they must satisfy L*π = 0 together with the process law. Where a stationary small-noise limit exists, the leading exponent must also agree with the global rate function determined by the dynamics, with the same noise normalization and measure convention. Detailed balance remains a separate, testable property. Which extra rule EEML imposes on the physically realized dynamics remains open.

This is a substantive connection between the layers. It preserves more of the old EEML than a universal gradient requirement: both relaxation and ongoing circulation can be investigated, and static agreement can no longer pass for a complete process explanation.

These examples illustrate different kinds of closure. A correct outcome within a class does not prove that nature realizes that class.

#A. K3 and χ = 24

Status: conditional mathematical theorem. Let 𝒜 be the class of smooth, closed, connected, oriented spin four-manifolds that admit a Riemannian Einstein metric and for which σ≠0. Closed means compact without boundary; Riemannian means a positive-definite metric.

Within precisely this class, the following chain is central:

The two steps have different foundations:

1. Einstein geometry: Hitchin–Thorpe gives 2χ≥3|σ|.
2. Smooth spin topology: Rokhlin gives σ∈16ℤ. With σ≠0, it follows that |σ|≥16.

Together they yield χ≥(3/2)·16=24. The number 24 is not put into a functional to be fitted. It is the sharp, reachable lower bound that follows from these conditions. See Furuta, introduction, for Rokhlin and LeBrun, Lemma 1, for Hitchin–Thorpe and its rigidity.

#Why equality yields K3

For K3 with its complex orientation, σ=−16 and χ=24. A Ricci-flat K3 metric is Einstein and reaches both bounds. Conversely, within 𝒜, χ(M)=24 forces |σ(M)|=16 and 2χ(M)=3|σ(M)|.

The equality case of Hitchin–Thorpe is Ricci-flat and is finitely covered by a flat torus or a Ricci-flat K3. The torus possibility is ruled out by σ≠0. For the remaining covering K3→M of degree d, the following holds:

Hence:

possibly with reversed orientation. This identifies the underlying smooth four-manifold, not one unique complex structure, metric or size.

EEML meaning: within 𝒜, K3 is the realization for which both structural bounds are saturated simultaneously. Therefore minₘ∈𝒜χ(M)=24. The choice that nature uses this class and realizes this minimum, and the process by which it reaches it, remain separate physical research questions. The derivation combines existing mathematical theorems; it is not a new empirically confirmed law of nature.

#Conditional cost derivation without a pre-imposed Einstein condition

We vary both the smooth space M and its metric g. M is closed (compact without boundary), connected, oriented, four-dimensional and spin. The metric is Riemannian, hence positive-definite. We impose σ(M) ≠ 0.

We do not assume in advance an Einstein condition, Ricci-flatness, Kähler structure, simple connectedness, string theory, K3 topology or the value 24.

All norms are full tensor norms, for example |Rm|² = RᵢⱼₖₗRⁱʲᵏˡ. Norms of curvature as an operator on two-forms sometimes use different factors. W⁺ and W⁻ are the two orthogonal parts of the Weyl curvature; Ric₀ is the trace-free Ricci tensor and R the scalar curvature.

The basic identities are:

See Marigliano, chapters 3–4 for these conventions and identities. The optimization below is built explicitly from these identities here.

#The minimum theorem

Define the non-negative numbers

For arbitrary fixed a,b,c > 0 we choose

This describes a family of costs with positive, orientation-symmetric weights. The original full curvature norm is the case a=1, b=2, c=1/6. This is not the most general possible physical action.

From the signature formula it follows that U₊−U₋ = 3σ/2. Since x+y = |x−y|+2 min(x,y) for x,y ≥ 0:

For closed smooth spin four-manifolds, Rokhlin's theorem gives σ ∈ 16ℤ. Together with σ ≠ 0 this yields the exact identity

See Walpuski, Spin Geometry, Theorem 30.9 for Rokhlin. The four terms on the right show exactly where costs above the lower bound come from.

For equality, the following are necessary and, for an existing metric, sufficient:

1. |σ| = 16;
2. one of W⁺, W⁻ vanishes;
3. Ric₀ = 0;
4. R = 0.

It follows that Ric = 0. The Einstein condition and Ricci-flatness are conclusions here, not assumptions. From Gauss–Bonnet it then follows that χ = 24.

The classification of compact Ricci-flat half-conformally flat four-manifolds gives a finite covering by a flat torus or a Ricci-flat K3. The torus case has signature zero. In the other case, the covering degree d satisfies 16 = d|σ(M)|. For |σ(M)|=16, d=1. So M is K3, possibly with reversed orientation. The classification result and its explanation are in LeBrun, Lemma 1.

Ricci-flat K3 metrics exist and reach this lower bound. As a result we prove not just an infimum: the minimum is actually attained. The result selects a smooth topology and a class of metrics, not a unique shape or size.

Within this positive cost family, a, b and c need not be tuned to K3. However, this cost family, four internal dimensions, spin, closedness, Riemannian signature and a nonzero geometric signature remain model conditions. A full physical action can have extra contributions. Without the nonzero-signature requirement, the flat torus has cost zero. The static minimum theorem gives no formation process, moduli stabilization or prediction of a constant of nature.

#Physical connection: quadratic gravity, without a solved compactness problem

The invariants used also appear in local gravitational actions with terms that are quadratic in curvature. That is an existing physical connection for the choice of a model, not an independent derivation of the positive coefficients, the candidate class or the full action from EEML. See Stelle, Renormalization of higher-derivative quantum gravity.

With \(U=U_++U_-\) and the full tensor norms used here, Gauss–Bonnet gives:

On a fixed closed topology, the last term is constant. When comparing topologies, it is not. The three positive terms from the minimum problem may therefore not be replaced by a ‘standard action’ in which the Euler contribution has been silently omitted. Einstein–Hilbert, cosmological and matter terms are likewise not automatically included in this cost family.

#Completed comparison with entanglement equilibrium

Check of the proposed bridge, 28 September 2026. Alonso-Serrano and Liška, arXiv:2212.03168, equation 37 write the action that corresponds to their linearized equations as \[ S=\frac1{16\pi G}\int\sqrt{-g}\, \left[R+d\left(W_{\mu\nu\rho\sigma}W^{\mu\nu\rho\sigma} -\frac13R^2\right)\right]d^4x,\qquad d=\frac{5\mathcal C\,l_P^2}{168\pi}. \] The derivation uses, among other things, entanglement equilibrium, a logarithmic entropy correction and conformal matter in the linearized regime. It does not provide a full nonlinear derivation of our internal geometry costs.

Compare the quadratic coefficients in the same formal invariant basis, after separating out an Euler term: \[ \begin{array}{c|cc|c} & W^2 & R^2 & R^2/W^2\text{ coefficient}\\ \hline \text{EEML cost} & a+b/2 & c+b/12 & (c+b/12)/(a+b/2)>0\\ \text{action from the paper} & d & -d/3 & -1/3 \end{array} \]

Outcome: the signs do not match for \(a,b,c>0\) and \(d\ne0\). A common normalization or sign flip does not change the relative ratio. A free Euler term does not resolve this either. Write \[ d(U-Q/3)+\tau\chi=aU+bP+cQ. \] Then \(a=d+\tau,\ b=-2\tau,\ c=-d/3+\tau/6\). Positive b requires \(\tau<0\); positive a requires \(d>-\tau>0\); positive c requires \(d<\tau/2<0\). That is contradictory. The general exclusion follows from this algebra. In addition, 81 basic identities have been recomputed exactly as an implementation check.

This refutes the proposed direct identification, not the conditional K3 minimum theorem or the paper. A Lorentzian spacetime action is not automatically a positive Riemannian cost for internal geometry. A different reduction or physical cost derivation may yield a different comparison, but must be carried out explicitly. On a Ricci-flat K3, P and Q are zero; comparing only the value on that minimum family would hide the contradictory signs outside the minimum. The result is therefore recorded as connection via the same invariants; coefficient bridge fails in this form.

#Lesson for EEML: robust families and scale-dependent visibility

Refinement of 28 September 2026. The breadth of a family and its physical grounding are different properties. That the same K3 minima work for all positive a, b, c is a genuine robustness result: within that family no fine-tuning is needed. It does not determine the weights and does not prove that nature belongs to that family. “The larger the family, the less it says” is therefore not a general rule. A statement that is proven for a larger class can in fact be stronger; a prediction that leaves more possibilities open can be less sharp.

Also, the entanglement derivation mentioned has not yet empirically identified the realized member. It provides a conditional candidate action. The sign difference found shows that this candidate is not directly our positive cost. The absolute statement “Euclidean negative R²” moreover requires an explicit continuation and action convention; the opposite relative signs are the convention-independent core of our comparison under a common sign flip.

For a characteristic variation length L, the derivative counting in this candidate action gives, when the leading Einstein contribution does not vanish, schematically: \[ \frac{\text{quadratic correction}}{\text{Einstein contribution}} \sim \frac{|d|}{L^2} =\frac{5|\mathcal C|}{168\pi} \left(\frac{l_P}{L}\right)^2. \] This is an order-of-magnitude estimate within the regime of validity, not a ratio of action integrals that would also hold for a zero denominator. At large L, differences are suppressed; they do not vanish exactly. Large coefficients, additional light fields and other integrated-out mass scales can change the regime. Close to the cutoff scale, it is precisely the truncation to quadratic terms that is insufficient. The leading gravitational description can moreover contain Λ and matter. “Corrections only play a role close to the Planck scale” and “exactly one theory remains” are therefore too strong. See the power counting in Burgess, §4.2.2.

EEML reading: different microscopic theories can describe the same leading operation at a given scale, within a chosen accuracy. That is effective universality, not a proven selection of the cheapest fundamental theory. The original members and their parameters do not disappear from the candidate space because their differences become hard to measure.

The physical constraints must then be justified separately:

• Freedom from tachyons, absence of ghosts, Euclidean boundedness from below and a controlled effective expansion are different requirements. One of them does not prove the others; the earlier higher-derivative warning still applies.
• If positive a and c are taken to mean the conformal anomaly coefficients, they are called \(a_{\rm CFT},c_{\rm CFT}\) here. They are not our cost weights a and c. Positivity conditions can bound ratios under their own hypotheses, but do not by themselves identify the full gravitational action. See Hofman–Maldacena.
• Local counterterms and renormalization conditions remain relevant: the \(\Box R\) anomaly term is sensitive to the local R² counterterm. Positivity of anomaly coefficients does not by itself fix that freedom. See Asorey, Gorbar and Shapiro.

Working rule: keep all members that are compatible with the physical information fixed in advance. Narrow that class only with independently motivated conditions, not with the wish to prove a minimum. Record the scale, conventions, sign conditions, uncertainties and the translation from action to any costs. A remaining theorem is then physically better motivated, but remains conditional until class and prediction have been tested independently. For the candidate with ratio −1/3, that cost translation is needed first; simply minimizing its action does not yield a new K3 selection.

A lower bound is not coercivity. Coercivity would, in a specified space of metrics and after treatment of symmetries, have to prevent geometries with bounded costs from escaping uncontrollably. Here the scaling direction in fact provides an explicit counterexample. In four dimensions, for \(g_\ell=\ell^2g\):

Quadratic curvature densities scale as \(\ell^{-4}\), the volume element as \(\ell^4\). A fixed Ricci-flat K3 can therefore keep the minimum cost 24a for every \(\ell>0\) while its volume becomes arbitrarily large or small. As a result, the unfixed sublevel sets are not compact in the usual smooth geometric sense. This is not resolved by the lower bound. A volume normalization removes this specific scaling direction, but does not by itself prove compactness of all remaining moduli.

The degeneracy is concrete even at fixed topology: modulo diffeomorphisms, the space of smooth Ricci-flat K3 metrics locally has 58 real parameters, of which 57 at fixed volume and one scale parameter. The 57 come from positive three-planes in a space of signature (3,19): \(3\cdot19=57\). Singular loci and discrete identifications require separate treatment. Every smooth Ricci-flat K3 metric in this family reaches the same cost 24a. See Tripathy and Trivedi, Compactification with flux on K3 and tori, §2.3.

The result therefore reads: the cost family is bounded below, but fixes no volume, shape or unique metric. The explicit scaling direction prevents compact sublevel sets in the unfixed metric space. Incidentally, the existence of flat cost directions is not by itself a general proof of non-compactness; for that, their geometry is also relevant.

Within a separately defined ensemble that consists solely of these minima, the exponential cost factor is constant and the chosen measure determines their relative weights. That measure is not derived from EEML. However, an ensemble over all metrics cannot simply be restricted to the set of minima: this set can have measure zero in the original measure. Integration over nearby fluctuations can produce extra determinant and measure factors. Reachability also remains a process question. ‘The measure does the work’ thus holds for the specified minimum ensemble, not as a general replacement for the other layers.

The existence of a minimum is nevertheless known in this example: existing Ricci-flat K3 metrics reach the derived bound. That is a different argument from coercivity. Choice of scale, moduli stabilization, normalization of an integral over metrics and dynamical reachability remain separate questions.

Euclidean positivity does not prove a healthy Lorentzian quantum theory either. Generic local higher-derivative additions to Einstein gravity contain, in the usual linear analysis, an extra spin-2 mode with the wrong energy sign; under standard quantization, the well-known ghost/unitarity problem arises. Which degrees of freedom the precise chosen theory has must be checked separately. See Stelle, Classical gravity with higher derivatives.

Status: a robust application of existing geometric identities within a chosen cost family. Here the dossier shows no new general mathematical theorem, no solved quantum gravity model and no empirical topology selection.

K3 and the torus remain different candidates. Ricci-flatness alone does not make K3 more favorable than \(T^4\): both have \(\mathrm{Ric}=0\), while the torus is completely flat. A cost that uses only \(\int\|\mathrm{Ric}\|^2\) does not distinguish them. A tie remains a valid result.

The Kummer connection: the quotient of a complex two-torus under inversion, \(T^4/\mathbb Z_2\), has sixteen fixed points. The Euler characteristic of the quotient is \((0+16)/2=8\). Resolution of the sixteen singularities replaces each point by an exceptional two-sphere: net \(8+16=24\). The result is a Kummer K3. This is a construction relation, not a proof of cosmic evolution, and not every K3 is a Kummer surface. See Huybrechts, Lectures on K3 Surfaces.

String-theoretic dualities connect specifically chosen theories and compactifications. They do not say that K3 and \(T^4\) are always the same. Theory, dimensions, bundles, fluxes, scales and observables remain necessary. The class in the proof of the bound has not yet been derived independently from our universe. Without further index and bundle analysis, four-dimensional chirality does not imply the required internal signature. There is no derivation here that K3 is our four-dimensional spacetime, or a valid parameter-free derivation of α.

#Conditional physical reading: six-dimensional (2,0) supergravity

In ordinary 6D (2,0) supergravity with one gravity multiplet and, as extra massless matter, only abelian tensor multiplets, cancellation of the gravitational anomaly fixes their number at \(n_T=21\). This statement concerns that field content and consistency conditions; extra sectors or anomaly inflow must be treated separately. See Riccioni, Tensor Multiplets in Six-Dimensional (2,0) Supergravity.

Type IIB compactification on a smooth K3 realizes this low-energy theory with 21 tensor multiplets. The local scalar moduli space is \(SO(5,21)/(SO(5)\times SO(21))\). See Heydeman, Schwarz, Wen and Zhang.

This is a useful connection between geometry, spectrum and anomaly cancellation. It is not an observation of 6D (2,0) supergravity in our universe, not a unique inverse reconstruction of K3 and not a prediction of 21 dark matter channels.

#B. Hypercharges: what exactly do the conditions close?

#What exactly is closed?

Fixed: one generation, the usual non-abelian representations, one Higgs doublet, no right-handed neutrino and real hypercharges. The goal is to determine the labeled fermion charges relative to the Higgs charge uniquely. We take h≠0 and choose the computational unit h=1. This only fixes the unit of hypercharge; in this unit, the electromagnetic generator is \(T_3+Y/2\).

The seven conditions are:

“Minimal” here means minimal under omission of conditions. This must be distinguished from the smallest number of conditions in some arbitrary other formulation. The result also depends on what is counted as one condition.

#The 2-of-3 derivation is correct

With the three Yukawa conditions fixed, q,l,h remain. Then

For h≠0, each pair enforces the known ratios. No single condition does so. For h=1, for example, \((q,l)=(0,0)\), \((0,-1)\) and \((1/18,0)\) are counterexamples for only \(A_2\), only G and only C, respectively.

The motivation for h≠0 must remain precise: it concerns the usual electroweak-electromagnetic embedding. Merely saying that “a U(1) remains” after breaking does not exclude h=0; after all, an uncoupled spectator U(1) can remain. The nontrivial Higgs hypercharge remains a named input.

#The full ablation gives nine sets

There is also a second replaceability:

as a polynomial identity. Each pair from \(\{Y_u,Y_d,A_3\}\) therefore implies the third condition.

All minimal closing sets have exactly the form

That makes \(1\times3\times3=9\) sets, each with five conditions. They give

The check uses exact rational elimination of the linear conditions and an analysis of the real zeros of the remaining cubic. It examines all 128 subsets, not just a numerical sample. The known solution is used as a reference point to examine uniqueness; this is an ablation of a known derivation, not a blind prediction.

The previously stored exact check reports 128 examined subsets and 256 additional point checks. This compilation does not rerun that calculation.

#Why the route without Yukawas is different

The four anomalies without Yukawas give

There are thus several branches. The branch \(q=0,u=-d,l=e=0\) is vector-like for the U(1) charges, but not automatically vector-like for the full gauge group: the quark doublet, for example, has no partner in its conjugate full representation. General “chirality” therefore does not exclude that branch. Specifically requiring that the U(1) assignment be chiral is a different, additional condition.

Even if q≠0 is imposed, the ratio to h remains free: h does not appear in any of these anomalies. Take

All four anomalies vanish and the fermion ratios are standard, but q/h=1 instead of 1/3. The Higgs interactions do not close. Moreover, the interchange of the two quark singlets must be dealt with if one wants to determine their labels.

Conclusion: without Yukawas, fermion ratios can be constrained under additional conditions. That is not the same closure of the full Higgs-plus-fermion question. This distinction belongs explicitly in the ablation protocol.

#C. Positive weighting and quantum amplitudes

Under suitable regularity, the multiplicative Cauchy relation gives an exponential form. In addition, positive decreasing weights and complex phase factors require different conditions. The constant ℏ, the measure, the Born rule and the choice of the physical action do not follow from this functional equation.

For a suitable Hamiltonian, analytic continuation shows the well-known operator connection:

In field theory, Euclidean correlation functions that satisfy the full applicable Osterwalder–Schrader conditions yield a reconstruction of a Lorentzian quantum field theory. Reflection positivity is essential, but not the only condition: regularity/growth bounds, covariance and symmetry, among others, also play a role. This concerns correlation functions and their reconstruction, not merely replacing i in an arbitrary exponent. Osterwalder and Schrader, 1975.

Moreover, reflection positivity is not the same as a pointwise positive classical weight function. Nor is every Gibbs distribution automatically the analytic continuation of a physical quantum field model. A thermal quantum state also comes with the imaginary-time period \(\beta\hbar\) and the associated boundary condition.

For the naive Euclidean Einstein–Hilbert integral, the conformal direction makes the action unbounded below. That is a real problem for that definition of the integral. It is not a proven impossibility of every Euclidean-Lorentzian connection in quantum gravity. Research on measures and contours addresses precisely this issue. Dasgupta and Loll, A proper-time cure for the conformal sickness in quantum gravity.

Formulation to retain: positive Euclidean and coherent Lorentzian descriptions can be connected where a suitable analytic reconstruction exists. The conditions and the chosen definition of the integral must be checked per model. This does not replace a Born postulate with an EEML derivation.

#C1. Stationary action, free fall and Jacobson's derivation

Status (addition of 28 September 2026): known physics, with a possible EEML interpretation. The statement “gravity and interactions are based on EEML” is not demonstrated by this.

Stationarity is not a general efficiency minimization. The usual classical descriptions of gravity and the gauge fields have action principles. There, \(\delta S=0\) is the stationarity condition; a stationary point can also be a saddle point. The Lorentzian action is not a generally non-negative measure of effort or dissipation. It therefore does not follow from its stationarity that every process has the lowest costs. The specific action and couplings are input as long as EEML does not derive them separately.

Free fall makes one meaning of effortlessness concrete. An ideal freely falling test particle without non-gravitational forces has zero proper acceleration:

That does not rule out coordinate acceleration or tidal effects on extended bodies. A timelike geodesic locally maximizes proper time in a suitable convex normal neighborhood; this is not an unrestricted global statement. Here, the absence of a required propulsive force is the useful connection to “effortless effect”. See Tong, geodesic motion and Reall, General Relativity, §2.

The coherent branch explains stationarity conditionally. In a suitable path-integral description we write schematically:

In a valid semiclassical approximation, rapidly varying phases largely cancel each other out. Neighborhoods of stationary paths can supply the leading contributions, even when there are several. This uses a given action, measure and quantum theory; it does not derive those ingredients from EEML. At finite \(\hbar\), non-stationary contributions are not generally exactly zero. See Feynman, the action principle and quantum amplitudes.

Jacobson is a concrete conditional reconstruction. His derivation assumes, among other things, horizon entropy \(S_{\rm hor}=\eta A\), the Unruh temperature, local Rindler horizons in local equilibrium and \(\delta Q=T\,dS_{\rm hor}\). Using the Raychaudhuri equation and the local conservation law, one obtains, in units \(c=k_B=1\):

The unknown entropy density \(\eta\) thus determines G; the derivation does not predict its value. Likewise, \(\Lambda\) remains an integration constant. This is more than multiplicity alone: geometry, causality, temperature and energy flow are involved. See Jacobson, 1995, Thermodynamics of Spacetime.

What EEML can do with it. This example fits the protocol of having independent descriptions close jointly. A standalone extension would have to derive a required relation or a measurable deviation that was not yet input. The assertions “this is the only lead” and “nothing comparable exists for gauge interactions” are not adopted: this check does not substantiate such exclusionary claims about the literature.

Connection to EEML's own claim about nature: neither stationary phase, nor free fall, nor Jacobson's reconstruction proves the prohibition of persistently dominated realizations. That prohibition compares reachable alternatives with the same effect and a predetermined cost vector. That comparison structure is not automatically present in \(\delta S=0\) or \(\delta Q=T\,dS_{\rm hor}\). For that, EEML-D1 remains a separate test proposal.

#C2. The “effortless member”: information conservation and emergent dissipation

Status (proposal of 28 September 2026): an additional hypothesis, designated H_R. It is separate from the historical H_D and the conditional D1 test. The connection to unitary quantum theory is known; EEML has not derived unitarity independently.

#What the pattern does and does not say

Under closed unitary evolution, \(\rho(t)=U(t)\rho(0)U(t)^\dagger\), the eigenvalues of \(\rho\) are conserved. Hence the fine-grained von Neumann entropy is conserved as well:

This is a consequence of the assumed dynamics. It does not show a competition in which nature chooses the most efficient member of every family of theories. The entropy of a subsystem can increase at the same time, and fine-grained entropy is not simply the thermodynamic entropy or local entropy production. In a complete Hamiltonian description, a related conservation of the fine-grained Gibbs entropy holds, given a suitable measure and boundary conditions.

Three concepts remain separate:

T violation has been measured directly in B-meson transitions. That is not proof of intrinsic loss of the complete quantum state. CPT is a theorem under the applicable quantum-field-theoretic conditions and is tested experimentally; “exact in every possible model of nature” is stronger. See BABAR, T violation and PDG, CPT tests and hypotheses of the theorem.

The claim that reversible microdynamics “has so far been reconstructed for everything” is not adopted. The models discussed do not provide a complete, proven microtheory of measurements, black holes and the universe. Classical deterministic equations, too, give a backward calculation only on their valid, well-posed domain of evolution; singularities require separate treatment.

#Two known mechanisms, without mixing up the layers

Statistical reduction. Multiplicity helps explain why some macrostates are typical. For the course of events, dynamics, the chosen macrovariables and preparation are needed as well. A static count produces no transition rate or arrow of time. In the usual statistical explanation, a special low-entropy preparation plays an essential role; a universal cosmological explanation of it is not thereby provided. See Lebowitz, microscopic dynamics and the macroscopic arrow of time.

The exact Mori–Zwanzig redescription brings hidden dynamics back as, among other things, memory and fluctuation terms. A practically irreversible description, for example a Markovian one, requires further conditions or approximations. Merely making a description coarser is not physical heat production. Dissipation is therefore not exclusively “layer III that counts”: layer IV, preparation and the chosen system boundary also play a part. See Mori, 1965.

Effective field theory. Within a controlled expansion, contributions from higher-dimension or higher-derivative operators can be small at energies well below the relevant heavy scale. Symmetries, light degrees of freedom, coefficients and the valid power counting remain input. The leading law is then an approximation; the other allowed terms are not necessarily absent. Einstein–Hilbert is the leading two-derivative term in the corresponding metric expansion, but the zero-derivative cosmological term may not in general be omitted. This does not explain a universal preference for zero dissipation and does not select a unique fundamental theory. See Burgess, effective field theory.

#H_R: conservation of fine-grained information

Additional hypothesis: the complete physical dynamics does not intrinsically and irretrievably erase state information. In the quantum formulation, a truly closed whole evolves unitarily. It must be possible to account for effective information loss through an explicit dynamics of omitted degrees of freedom, correlations and the preparation.

“Fundamental” is not defined here as “reversible”. H_R is precisely a claim about the complete dynamics. In this specific hypothesis, the name “effortless” refers to conservation of information; the hypothesis does not prove 100% thermodynamic efficiency, minimal costs, automatic motion or the dominance prohibition. The universal extension is open, even when a concrete closed quantum experiment agrees with unitarity.

#Negative control: information loss without heating

A simple dephasing model with fixed \(H=(\hbar\omega/2)\sigma_z\) and \(\gamma>0\) reads:

Because H commutes with \(\sigma_z\), \(\operatorname{Tr}(H\rho)\) remains constant. Yet the coherence decreases. Starting from \(|+\rangle=(|0\rangle+|1\rangle)/\sqrt2\), the eigenvalues are:

The entropy rises from zero to \(k_B\ln2\), without any change in the mean energy. For \(t>0\), the channel has no physical inverse on all states: formally calculating backward would yield negative eigenvalues for some valid input states. This distinguishes physical reversibility from mere algebraic invertibility.

Nor does observed dephasing prove fundamental loss. A unitary controlled-Z on two qubits initially prepared in \(|+\rangle\) makes each one individually maximally mixed, while the whole remains pure. This coupling commutes with the system energy \(H\otimes I\). The same complete dephasing of the system is thus compatible with information in correlations. Both examples have been checked analytically and numerically; they are not measurements of nature.

More generally, a finite-dimensional quantum channel has a unitary extension with a suitable environment. That mathematical representation does not prove that the environment in question physically exists, nor that it yields a local microtheory over all times without further conditions. It does show why a measured reduced channel alone does not determine the fundamental interpretation. See Kretschmann, Schlingemann and Werner, Stinespring representation.

#Collapse as a test direction, with a delimited interpretation

GRW/CSL-type theories add objective stochastic collapses. Common CSL variants predict, among other things, suppression of coherence, momentum diffusion and spontaneous heating or radiation. Not every variant has the same heating signal: dissipative extensions change the energy dynamics. Heating and fundamental non-unitarity are therefore not synonyms. See Carlesso et al., experimental collapse tests and Smirne and Bassi, dissipative CSL.

Checked existing exclusion: in March 2026, XENONnT published stronger bounds on the investigated Markovian CSL and Diósi–Penrose models via an X-ray search. They cover values from originally proposed CSL parameter regions. It is not an exclusion of all collapse models and not an EEML prediction made in advance. See XENON Collaboration, PRL 136, 120201.

A test must fix in advance the collapse model, parameter region, noise spectrum, environment model, observables, uncertainties and decision rule. A single unexpected heat signal is insufficient. Mutually consistent mass, distance and time dependences must distinguish between the alternatives laid down. A convincingly demonstrated intrinsically non-unitary process would refute H_R; excluding a limited environment model does not yet demonstrate that universal character.

Ever smaller upper bounds leave H_R standing for now within the tested range, but do not prove an exact zero and do not distinguish EEML from ordinary unitary quantum theory. An unbounded appeal to an unknown environment must not make the hypothesis irrefutable after the fact. The universal H_R is therefore not yet a fully operational test; concrete model comparisons can be.

#Registration of H_R: “effortless member”

Dossier registration: 28-09-2026; hypothesis with a test direction, no external preregistration. “Fundamental” here means an assumed complete process law of the physical degrees of freedom within an explicitly specified domain of validity, distinguished from a reduced description by separate information about couplings and environment. Reversibility is not part of that definition. That a laboratory system is truly completely isolated is not a directly measurable fact; the environment boundary remains part of the model package to be tested.

The test coverage is limited: as \(\lambda\to0\), CSL approaches the reference model, so finite precision does not exclude every positive collapse rate. The link to an existing experimental program makes the hypothesis open to investigation, but gives EEML no exclusive prediction.

#D. Three routes to family counting: different evidential strength

1. 331: checked closure within a specific class. Take the gauge group \(SU(3)_C\times SU(3)_L\times U(1)_X\), \(N_f>0\) lepton families as antitriplets, n quark families as triplets and the rest as antitriplets. The cubic \(SU(3)_L\) anomaly gives:

With three colored Dirac flavors per family, without extra colored matter or scalars, the additional requirement of QCD asymptotic freedom above the relevant thresholds gives \(b_0=11-2N_f>0\), so \(N_f\le5\). Combined, exactly \(N_f=3\) remains. The representation assignment and asymptotic freedom of this extension are input. Other lepton/quark assignments change the count. See Pisano and Pleitez and Calabrese et al.. This is discrete closure, not a new efficiency minimization.

2. Supersymmetric 6D route via Bershadsky–Vafa: proposed, not yet independently reconstructed. The historical work imposes congruence conditions on specific six-dimensional gauge and matter sectors. On this basis, a supplied in-house application claims \(N_f\equiv0\pmod3\), but its full spectrum, anomaly polynomial, tensor charges and Green–Schwarz quantization are missing here. It is therefore not counted as a second successful derivation. See Bershadsky and Vafa. The modern treatment shows why the global definition of the Green–Schwarz mechanism must also be checked: Lee and Tachikawa.

3. Dobrescu–Poppitz: historical route, modern reassessment needed. The original non-supersymmetric 6D model links the chosen chiral content and a historical SU(2) anomaly condition to multiples of three: Dobrescu and Poppitz. According to the modern bordism analysis, the argument ‘non-trivial \(\pi_6\), hence a separate global anomaly’ does not suffice. For the spin theories treated with SU(2), SU(3) or \(G_2\), the seven-dimensional bordism groups in question vanish; relevant older congruences must be assessed via local anomaly cancellation and the full compensation structure. See Davighi and Lohitsiri and Lee and Tachikawa.

This correction does not prove that every multiple-of-three condition disappears. It does prevent the old homotopy step from being presented, without recalculation, as an independent conclusive proof. Even after six-dimensional consistency, a compactification/projection to four dimensions must yield the desired family spectrum.

In addition, the historical route from older EEML texts: the standard embedding. In heterotic compactifications with the standard embedding, the net number of families is |χ|/2 (Candelas, Horowitz, Strominger and Witten 1985). Three families therefore require |χ| = 6; the classic example is the Tian–Yau space with χ = −6, which is not simply connected. This is a conditional count: nothing in this route selects |χ| = 6, and spaces with χ = 0 exist too. The claim in those texts that 6 is the smallest possible value is wrong (see E11).

Combined status: one explicitly checked 331 count, one incompletely supplied supersymmetric calculation and one historical 6D route under reassessment. These are not three independent confirmations that EEML predicts three generations.

#E. Derivation attempt for α: from topology to a coupling strength

Question: can the current EEML derive the fine-structure constant? Here α refers to electromagnetism, ultimately to the low-energy value α(0). The curvature coefficient a from VI.A is a different quantity. This attempt does not use 137 to choose parameters or integers.

#E1. Which part must the derivation fix?

In natural, rationalized units, with the electron charge as the unit in the covariant derivative, we write the Euclidean kinetic term:

After canonical normalization of A, e appears in the matter coupling. Hence:

These normalization conventions are explained, for example, in Tong, Supersymmetry, §4.1. The electromagnetic sector used here need not be supersymmetric. The coefficient k is not an electrical resistance in ohms, nor is it a measure of dissipation.

The existing K3 theorem does not determine k. A minimal extension with an independent, positive Maxwell term leaves the same geometric minimum cost 24a in place on a configuration F=0, for every k>0. In this extension, a vectorlike charged Dirac sector can retain the same charges and anomaly cancellation. As soon as charged matter is used as a probe, different k do have different observable interaction strengths. So this is not a mere change of units.

Scope: this is an explicit freedom in a factorizable classical/low-energy model. It is not a proof that every possible complete theory of quantum gravity must leave k free. It does show that the already proven geometric bound and the stated charge conditions do not supply that missing relation. Quantum corrections or a non-factorizable action must be computed separately.

Nor is a itself determined: for t>0 we have \(\mathcal E_{ta,tb,tc}=t\mathcal E_{a,b,c}\), with the same minimizers. The value of an action relative to ℏ or ε can be physically relevant; that normalization does not follow from the minimizer.

#E2. A more concrete geometric bridge

For harmonic two-forms on K3 we must distinguish two matrices:

Q is the topological intersection form in a fixed integral basis. H is positive definite and contains metric information via the Hodge star. A fixed Q does not determine H.

This distinction has a physical application: under reduction of the M-theory field \(C_3=A^I\wedge\omega_I+\cdots\) on K3, H appears in the kinetic matrix of the resulting gauge fields. In the seven-dimensional Einstein frame there is moreover a volume factor \(V_{K3}^{-3/5}\). This is a 7D result; further compactification and identification of the electromagnetic field are needed for α in four dimensions. See Castellano, Herráez and Ibáñez, §5.2, equations 5.11–5.13.

A uniform rescaling of the internal four-dimensional metric does not change the Hodge star on two-forms. Changing the shape, however, can change H. The volume factor and any further reduction must therefore not be confused with the shape dependence.

For an exactly computable slice we use the hyperbolic lattice block U, which occurs in the K3 lattice. Define for x>0:

Then the following hold exactly:

This model therefore has one unchanged intersection form and yet a continuous family of positive Hodge norms. The relation to the lattice and the description in terms of positive three-planes are given in Tripathy and Trivedi, §2.3.

Limitation: the 2×2 matrices form an algebraic toy model of their own here. A fully smooth K3 realization requires an admissible positive three-plane, treatment of roots with square −2 and exclusion of singular loci. We have not constructed a smooth compact physical solution for every x.

#E3. The attempt with EEML: let two contributions bound each other

Fix a nonzero integral two-form class \((n,m)\), with n and m both different from zero. As an additional model cost, take its quadratic norm, with a fixed positive prefactor omitted:

This is the structure of a positive field-energy term at fixed flux. The chosen class is input: the effect or the flux sector must be fixed independently. The cost family of VI.A did not supply this flux.

Differentiating gives:

Hence:

The global lower bound also follows directly:

The conditional result: as soon as the two specified contributions are genuinely present, minimal total input fixes one parameter that the earlier geometric costs left free. No choice of the desired final number is needed to prove this minimum theorem.

The ablation is decisive. For m=0 and n≠0 the infimum shifts to x→0; for n=0 and m≠0, to x→∞. If both may be zero, the empty sector costs zero for every x. The interior minimum thus rests on keeping both contributions jointly, in accordance with the working rule ‘compare at equal effect’.

Moreover, an equal minimum cost does not automatically fix the same parameter in different sectors. In this marked basis, (n,m)=(1,6) and (2,3) both have minimum cost 12, but x=6 and x=3/2 respectively. This is a check within the toy model, not a classification of physically inequivalent K3 vacua; for that, basis identifications and the embedding of the measured gauge field must also be taken into account.

#E4. Static selection is not yet a law of motion

We can choose a process separately. Set z=log x and take:

Then:

For two nonzero contributions, \(\mathcal R(z)=n^2e^z+m^2e^{-z}\) goes to infinity at both ends and has a single minimum. This chosen relaxation converges to it. For n=1, m=2 and Γ=1, five initial states were computed; all reach x=2 with cost 4, with decreasing cost up to numerical rounding.

Status: this illustrates a possible process that realizes the minimum condition. The choice of the gradient flow, the coordinate mobility and any outflow is extra model input. A closed conservative system need not relax in this way. This process law is not derived from zero resistance or from EEML alone.

#E5. What does this yield for α?

Only if a consistent four-dimensional reduction with fixed charges actually makes the remaining electromagnetic coefficient \(k=\kappa x\) does it follow, at the chosen matching level, that:

After matching, mixing where needed and scale evolution, the schematic outcome would be:

This is a conditional formula with unknown input, not a computed constant of nature. κ must follow from the full action, scale factors and charge normalization. The flux sector (n,m), all other relevant moduli and the quantum corrections are not derived here.

Moreover, a two-form probe flux on K3 is not automatically an admissible background flux of the M-theory field C₃ mentioned above, whose field strength is a four-form. We have used the norm structure, but have not built a full compactification that fulfills both roles at once. A real background must respect its field equations, source/charge conditions and backreaction on the geometry. It can change the Ricci-flat background. The toy calculation is therefore not a proof of moduli stabilization in our universe.

#E6. Can spectral action or scale evolution fill in the missing number?

Spectral action. It can give relations between gauge couplings, but requires a spectral geometry and an action prescription. In the conventions of Chamseddine–Connes (1996), equation 3.18:

The spectral normalization f₄ affects the absolute strength. A fixed choice of the cutoff function can pin down that normalization within that model; EEML has not yet derived this choice or the required spectral data. Moreover, this historical calculation concerns a different geometric construction and is not a K3 calculation of α(0). See Chamseddine and Connes, The Spectral Action Principle.

Scale evolution. For the simple QED model with one active Dirac fermion of unit charge, well above its mass and without thresholds in the interval studied:

The one-loop coefficient is checked, for example, in Igarashi and Itoh, QED in the exact renormalization group. The equation transports an initial value. It does not fix it: at a scale ratio of 10, initial values 100 and 150 for α⁻¹ stay exactly 50 apart. This illustration is not a calculation through the electroweak or hadronic thresholds to α(0).

Outcome of the attempt: the current conditions do not yield a unique α. What has been found is an exact intermediate step: a specified double load can fix a shape parameter via a positive quadratic norm. To turn this into a derivation of α, a single physical model must supply that load as well as the normalization and the electromagnetic identification, and its other predictions must hold up independently.

#E7. Full shape space: a minimum is not yet closure for the coupling

The earlier x-model held the other shape directions fixed. Now we let them take part. At fixed volume, a positive three-plane Σ in a real space of signature (3,19) determines the Ricci-flat K3 metric, with the usual exclusions for singularities and discrete identifications. The local shape space has dimension 57.

Write a fixed flux vector v as v₊+v₋ according to the Hodge splitting. Then:

For Q(v,v)>0, equality means v∈Σ. The remaining two positive directions can still move in a space of signature (2,19): 38 shape parameters remain free. For Q(v,v)<0, equality means Σ⊂v⊥, and 54 shape parameters remain. These are local counts on admissible smooth branches, not a guarantee that every integral flux admits a smooth minimum. A vector of square −2 in the negative plane can in fact force a singular locus.

More generally, r independent positive and s independent negative directions, under compatible, mutually orthogonal saturation, locally leave \((3-r)(19-s)\) shape parameters free. Volume is not counted. The geometric parametrization is given in Tripathy and Trivedi, §2.3; the dimension count here follows by removing the fixed directions from the positive and negative planes.

A small exact counterexample makes the consequence for couplings visible. Take a real, not integrally normalized basis:

We have H_z>0 and H_zQ⁻¹H_z=Q. For the fixed flux v=(1,0,0), the minimum cost remains H_z(v,v)=1. Another fixed probe direction w=(0,1,0) has norm H_z(w,w)=cosh(2z), which does change. This is an algebraic demonstration of a free coupling direction; not a full K3 compactification.

Correction to a possible reading of E3: one fixed shape parameter does not mean full moduli stabilization. At the same time, not all moduli need to be fixed to predict one observable. The right test is whether the electromagnetic coefficient changes along the remaining minimum set.

#E8. A physical exclusion: does the intended photon stay massless?

A two-form flux has a concrete realization in heterotic line bundles on K3. There, the anti-self-duality condition stabilizes part of the shape space. But the same flux can give the corresponding U(1) gauge field a mass via coupled axion fields. The effective action moreover retains dilaton dependence in the gauge couplings. So this route does not automatically yield electromagnetism. See Louis, Schasny and Valandro, especially §4.4.

The relevant linear check is simple. With positive axion metric G and charge matrix N, the action contains \((db-NA)^{\mathsf T}G(db-NA)\). The quadratic mass block is NᵀGN. A combination A=vA_light escapes this mass exactly when Nv=0. Other mass mechanisms still have to be excluded separately.

For N=(1,−1), v=(1,1) is massless under this term. Its kinetic coefficient, however, remains vᵀKv and is not determined by Nv=0; for K=diag(2,5) that coefficient is 7. Charge normalization remains necessary for a physical coupling. Masslessness thus closes a condition, not a value of α.

#E9. From toy model to an explicit IIB flux family

We now choose one existing physical framework: type IIB on K3×T²/ℤ₂ with RR and NS three-form flux. In the leading-order supergravity description, the following apply: flux quantization, charge cancellation and, for the supersymmetric solutions studied here, the requirement that G₃=F₃−φH₃ is primitive of type (2,1). φ=C₀+i/g_s is the axio-dilaton; τ is the complex shape of the torus. This framework and the two-dimensional flux sector are given in Tripathy and Trivedi, §2 and §4.1.

Our own worked-out subfamily. Choose orthogonal integral K3 classes p and q with p²=q²=2, for example e₁+f₁ and e₂+f₂ in two hyperbolic lattice blocks. Let:

where F₃ and H₃, apart from the common unit factor, are a_x∧dx+a_y∧dy and b_x∧dx+b_y∧dy respectively. The even coefficients respect the flux quantization of this orientifold convention.

Set Ω=p−iq and choose the Kähler class orthogonal to p and q. The equations \((a_x-\phi b_x)^2=0\) and \(a_y-\phi b_y=\bar\tau(a_x-\phi b_x)\) give:

Then G₃ is proportional to Ω∧d\bar z: the required type (2,1). The stated orthogonality gives primitivity. The choice of an admissible smooth Kähler class and the remaining sources is part of the model; no new complete ten-dimensional solution has been constructed numerically here.

The positive norm gives the same balance. In the rectangular slice with C₀=Re τ=0, s=Im φ and t=Im τ, the normalized flux norm yields:

For BC=AD, s₀=A/B and t₀=C/A, the following holds exactly:

Both balances saturate simultaneously at (s₀,t₀). This is an explicit application of the sharpened protocol: the conditions have a joint realization, and the two parameters studied are then fixed within the chosen flux sector.

The sources bound the sectors. In the same convention:

Without anti-D3-branes and without the additional source contributions omitted here, this means AD≤3. The positive integer solutions of BC=AD≤3 have been fully enumerated: nine flux assignments before duality identifications. These are not necessarily nine physically distinct vacua. Five assignments have N_D3>0.

Three examples from the checked list:

In the subset with g_s≤1, the values 1, 1/2 and 1/3 thus occur. This is not a general bound on string couplings: only this small family has been computed. g_s=1 offers no small string-loop parameter; nor do the other examples justify a precision measurement without an error analysis.

Why ‘least flux’ does not yield a justified winner. The flux norm is not the full four-dimensional vacuum potential. After the consistent source bookkeeping, the leading-order no-scale description has a non-negative contribution that, in this slice, is proportional to \(\mathcal R-N_{\rm flux}\). For all the supersymmetric solutions above, this contribution reaches zero. Less flux charge here means more D3-branes; comparing only the flux norm leaves out a changing part of the whole. The no-scale structure and the role of sources are treated by Giddings, Kachru and Polchinski.

So there is closure for φ within a fixed sector, but not yet closure for φ across the sectors studied. The weights, reachability or extra corrections that could single out a sector are not derived here.

How close does this come to α? For a D3 gauge sector, the tree-level kinetic function depends on the axio-dilaton, whereas a D7 sector also uses the wrapped four-cycle. In the conventions of Conlon, Quevedo and Suruliz, equations 12–13, \(f_{\rm D3}=S/(2\pi)\), with Re S=1/g_s and g_YM⁻²=Re f. For that named, normalized gauge sector, this would then give \(\alpha_{\rm D3}^{-1}=2/g_s\) at tree level. For the first two table rows, that gives 2 and 4. The third row contains no mobile D3 sector; we do not add a probe there without changing the charge bookkeeping.

This is not α_em(0): the charged spectrum, a massless electromagnetic field, its generator normalization and the scale evolution have not been constructed. A free volume modulus does not directly block this D3 tree-level coupling, but it can affect scales and corrections. No complete prediction or exclusion of a realistic model has been achieved.

Progress: the norm balance from the toy model has now been applied in a named flux sector and yields a fixed string coupling there. The open problem has shifted to selection between sectors and the electromagnetic identification. Neither K3 nor 24 alone resolves those two questions.

#E10. What fixes the volume? A sign rule from topology

In E9 the volume stays free: the leading no-scale description does not fix it. The best-known mechanism that does is the large volume scenario (LVS) of Balasubramanian, Berglund, Conlon and Quevedo. It combines two known contributions: an α′³ correction to the Kähler potential that is proportional to the Euler characteristic χ of the Calabi–Yau threefold (Becker, Becker, Haack and Louis, eqs. 3.18 and 3.24), and a non-perturbative term on a small four-cycle:

Here \(\mathcal V\) is the Einstein-frame volume in units \(\ell_s=2\pi\sqrt{\alpha'}\); in the units of Becker et al. themselves, \(\xi=-\chi\zeta(3)/2\). For a volume with one small cycle, \(\mathcal V=\alpha(\tau_b^{3/2}-\gamma\tau_s^{3/2})\), the potential at large volume becomes:

Minimizing gives, for \(a_s\tau_s\gg1\):

The sign rule. A solution requires \(\hat\xi>0\), so χ < 0, that is \(h^{2,1}>h^{1,1}\). For χ ≥ 0 the potential, after minimizing over \(\tau_s\), is negative everywhere at large volume and increasing in \(\mathcal V\). There is then no minimum: the model slides to small volume, where this description no longer holds. Cicoli, Conlon and Quevedo, §2.2 give the general conditions: \(h^{2,1}>h^{1,1}>1\) and at least one blow-up cycle with a non-perturbative term. The minimum has negative energy (anti-de Sitter) and breaks supersymmetry.

Rechecked (1 October 2026). Own check in the example \(\mathbb P^4_{[1,1,1,6,9]}[18]\) (\(h^{1,1}=2\), \(h^{2,1}=272\), χ = −540, α = 1/(9√2), γ = 1); the code is available on request via the contact address.

• With the dilaton fixed and only the Kähler moduli, two exact identities hold for every volume that is homogeneous of degree 3/2: \(K^{i\bar\jmath}K_iK_{\bar\jmath}-3=3\hat\xi/(4\mathcal V-\hat\xi)\) and \(K^{i\bar\jmath}K_{\bar\jmath}=4\tau_iY/(Y-3\mathcal V)\), with \(Y=\mathcal V+\hat\xi/2\). Derived with Sherman–Morrison and the Euler relation, checked numerically to 10⁻¹⁵. The form of Becker et al. with the dilaton included differs only at relative order \(\hat\xi/\mathcal V\).
• Full N=1 potential with \(g_s\) = 0.1, \(W_0\) = 1, \(A_s\) = 1 and \(a_s\) = 2π/10: a minimum at \(\tau_s\) = 41.63 and \(\mathcal V\) = 1.34·10¹¹, with \(V_{\min}\) = −2.45·10⁻³⁴ in the units of the code. All eigenvalues of the Hessian are positive. The boxed formulas give \(\tau_s\) = 41.08 and \(\mathcal V\) = 0.98·10¹¹. Minimizing the three terms exactly gives 41.627 and 1.343·10¹¹: equal to five digits. The first correction in \(1/(a_s\tau_s)\) is in Conlon and Pedro, eqs. 2.16–2.18. The sensitivity is exponential: 1.3% in \(\tau_s\) becomes 30% in \(\mathcal V\).
• Scan over 216 combinations: \(g_s\) ∈ {0.02, 0.1, 0.2}, \(W_0\) ∈ {0.01, 1, 10}, \(a_s\) ∈ {2π/5, 2π/10, 2π/20} and χ ∈ {−540, −100, −20, −6, 0, 6, 100, 540}, within the window \(\max(10,\,10|\hat\xi|)<\mathcal V<10^{40}\) and \(\tau_s\) ≥ 1. For χ ≥ 0 there is no minimum in any case. For χ < 0 the minimum lies exactly where the formula places it. When that place is outside the window (small \(|\chi|/g_s^{3/2}\), or \(\mathcal V\) > 10⁴⁰), there is none in the window.

What topology does here, and what it does not. It does: the sign of χ decides whether a minimum at large volume can exist, before there is anything to compare. And |χ| sets the volume exponentially through \(\tau_s\propto(-\chi)^{2/3}/g_s\). It does not: it picks no Calabi–Yau, and \(W_0\), \(g_s\) and \(a_s\) remain input. The minimum has negative energy; a universe like ours needs a separate, disputed uplift. And it is a minimum of a chosen effective potential: that nature minimizes that cost does not follow.

Limits of the rule. “Required” holds within this truncated description. Branes and orientifold planes shift the relevant χ to an effective value (Minasian, Pugh and Savelli). With extra corrections (string loops, F⁴ terms, α′² logarithms), proposals have been made that also give a minimum for χ ≥ 0; their coefficients or signs are unknown or chosen, see for example Ciupke, Louis and Westphal. The original LVS paper left the case ξ < 0 open; the scan above finds no minimum in this model.

Back to E9. The E9 family uses K3×T², with χ = 24·0 = 0. There ξ = 0 and this mechanism gives no minimum: the volume stays free at this level. Nothing changes for α: the D3 coupling of E9 depends only on \(g_s\).

Lesson for EEML. This is a clean example of layer I before layer II: a chosen cost has a minimum only if a topological condition holds. It is existing physics and does not distinguish EEML from the existing model. Historical note: older EEML texts (April 2026) used a term −χ ln \(\mathcal V\) without a derivation and required χ = +6. In the real mechanism, that sign is exactly the one that does not work.

#E11. Looking back: the old α derivation was circular

In April 2026, EEML texts circulated with a “derivation” of \(\alpha^{-1}=137.036\) from K3 and a Calabi–Yau space with χ = 6. Part II already books that route as historical. Here is exactly why, rechecked with the PDG values at \(M_Z\) (MS-bar: \(\alpha^{-1}=127.951\), \(\sin^2\theta_W=0.23122\)); the code is available on request via the contact address.

So every “derived” number is a measured value, or the gap to a measured value: input is output. The oldest versions also wrote \(\alpha^{-1}(M_Z)=137.036\); at \(M_Z\) it is about 128.

Where 24 does appear in the couplings. Run the three measured couplings at one loop with the supersymmetric extension (MSSM) and they meet at \(1/\alpha\approx24.3\), near \(2\cdot10^{16}\) GeV. That holds if the superpartners sit near \(M_Z\). At 1 TeV it becomes 26.0, at 3 TeV 26.8; two loops and thresholds shift it further. This unification is classic (Amaldi, de Boer and Fürstenau 1991). That the value lies close to χ(K3) = 24 depends on an unknown scale: it is a coincidence, not a derivation. With the Standard Model alone, the couplings do not meet.

Three more errors from the same texts. Calabi–Yau spaces with χ = 0 exist, so 6 is not “the smallest”. A Calabi–Yau threefold has no elliptic fibration over K3. And K3 carries no SU(2) instanton of charge 1: the expected dimension of that solution space is \(8k-12\), so −4 at \(k=1\); the smallest charge is 2.

#F. EEML-PT 0.1: process topology as a physical candidate model

New proposal, 28 September 2026. This section is a constructed research model, not a derived law of nature. It does not replace the earlier benchmarks. The hypothesis reads:

The links along which operation is possible are themselves physical degrees of freedom. Matter and links change jointly. An observable topology can emerge as a long-lasting pattern in that mutual dynamics.

So here a process that can change topology is chosen before a topological winner is named. The hypothesis that this approach can also describe space itself is an additional one. The closest existing relatives are Quantum Graphity and the Bose–Hubbard model with a dynamical graph by Hamma et al.. There, matter and changeable interactions are already coupled. This dossier claims no original discovery of that general idea; it provides an explicit, checkable EEML trial variant.

#F1. What is physical, and what is assumed?

Take N possible carriers of an excitation and, for every pair, a switchable link. A link is a two-state system with occupation \(\hat n_{ij}=0,1\). At most one bosonic excitation can be present on each carrier. A fixed, measured excitation energy may be added; at a fixed total number it is a constant.

Three parameters are model input, to be determined independently:

• \(\Delta_b>0\): energy difference between a switched-on and a switched-off link when no excitations are present;
• \(J>0\): energy of coherent exchange via an existing link;
• \(g\ge0\): quantum amplitude for changing a link.

All three have units of energy. These are not arbitrary exchange rates between material, reliability and attention. In a possible quantum simulator they must be calibrated spectroscopically. That such a model describes the universe is not thereby established.

As a concrete transition rule, a link may change only when its endpoints already have a common neighbor. Call the corresponding projector \(\hat F_{ij}\). Explicitly: \[ \hat F_{ij}=1-\prod_{k\ne i,j}(1-\hat n_{ik}\hat n_{kj}). \] With this we choose the Hamiltonian \[ \boxed{ \hat H_{\rm PT}= \Delta_b\sum_{i<j}\hat n_{ij} -J\sum_{i<j}\hat n_{ij} (\hat a_i^\dagger\hat a_j+\hat a_j^\dagger\hat a_i) -g\sum_{i<j}\hat F_{ij}\hat\sigma^x_{ij}. } \]

The common-neighbor rule is an explicit new model choice, not derived from EEML. It restricts changes to proximity as determined by the current network. \(\hat F_{ij}\) does not contain \(\hat n_{ij}\) and therefore commutes with the corresponding flip. The Hamiltonian is Hermitian. The model already has an evolution parameter: \[ i\hbar\,\partial_t|\Psi\rangle=\hat H_{\rm PT}|\Psi\rangle. \] Time itself and Lorentz symmetry are not thereby derived.

#F2. Bookkeeping and connectedness

For time-independent parameters, the closed unitary evolution conserves the total energy expectation value and the number of excitations \(\hat N_{\rm ex}=\sum_i\hat a_i^\dagger\hat a_i\). Matter can exchange energy with the link sector; those sectors need not have constant energy separately.

The chosen flips preserve the number of connected components. After all, a removed link has an alternative route via the common neighbor; an added link connects points that were already connected. A connected initial configuration therefore stays connected. One connected whole is here an initial/sector condition, not a derivation that only one universe is possible. The model needs a seed; under this flip rule an empty graph develops no links.

Unitarity alone does not cause cooling to the ground state. For thermal selection, a named environment must be able to absorb energy. In a cosmological reading, such degrees of freedom must be accounted for within the entire whole. The heat-bath approximation below is a separate, effective assumption; a lower energy guarantees neither a transition nor a prescribed recovery time.

#F3. A genuine, computed topology switch

Take three physically identifiable carriers, one excitation and a connected network. There are three distinct labeled lines with two links and one closed triangle with three links. For now there are no filled faces: the triangle is a cycle, not a disk.

In the limit of small g, the lowest energies of the fixed graphs are: \[ E_{\rm line}=2\Delta_b-\sqrt2J,\qquad E_{\rm cycle}=3\Delta_b-2J. \] The exchange spectra are \((-\sqrt2J,0,\sqrt2J)\) and \((-2J,J,J)\), respectively. It follows that \[ E_{\rm cycle}-E_{\rm line} =\Delta_b-(2-\sqrt2)J. \]

The cycle becomes energetically favorable when \[ \boxed{\frac{J}{\Delta_b}>\frac1{2-\sqrt2} =1+\frac1{\sqrt2}\simeq1.70710678.} \]

The extra link raises the link energy but lowers the energy of the joint excitation state more strongly. So no rule has been introduced that directly rewards cycles or prescribes a desired Euler characteristic. The number follows from this chosen three-carrier spectrum; it is not a new universal constant of nature.

A cycle is not yet a cycle current. This gain concerns the coherent binding/exchange energy. The real ground states of the fixed graphs have no net particle current. Here the model explains the preference for a link structure, not compulsory ongoing motion. Drive or a differently prepared state requires a separate process calculation.

Topologically, under a graph measurement the number of independent cycles \(b_1\) changes from 0 to 1; \(b_0=1\) stays the same. The graph Euler characteristic changes from \(N-|E|=1\) to 0. At finite g the state can be a coherent superposition of graphs. Then, without a measurement/decoherence description, there is no single sharply classical topology.

#F4. One quantitative prediction, with temperature and counting

Assume a weakly coupled environment that actually thermalizes the connected one-excitation sector, with independently measured temperature T. Next, take g small compared with the relevant thermal and spectral scales. Set \(\beta=1/(k_BT)\). For one labeled line and the cycle: \[ Z_{\rm line}=e^{-2\beta\Delta_b} [1+2\cosh(\sqrt2\beta J)],\qquad Z_{\rm cycle}=e^{-3\beta\Delta_b} [e^{2\beta J}+2e^{-\beta J}], \] \[ \boxed{P_{\rm cycle} =\frac{Z_{\rm cycle}}{3Z_{\rm line}+Z_{\rm cycle}}.} \]

The factor 3 counts three distinct switching configurations of identifiable carriers. It must not be adopted blindly when node labels are purely gauge labels. In that case the physical state space and measure must be constructed anew.

At low temperature this gives approximately \[ \frac{P_{\rm cycle}}{1-P_{\rm cycle}} \simeq\frac13 \exp\!\left[\frac{(2-\sqrt2)J-\Delta_b}{k_BT}\right]. \] Equal ground energy therefore does not automatically mean equal total probability for a line and a cycle. The statistical layer here follows from the specified Hamiltonian, thermal assumption and counting; efficiency and weight are not equated.

Calculation performed, \(\Delta_b=1,\ k_BT=0.05\) in the same energy units:

The finite-g column was computed by diagonalizing the full Hamiltonian, not by refitting the small-g formula. For this finite system it is an energy crossing at g=0 and a rounded quantum/thermal transition, not a proven thermodynamic phase transition.

The equilibrium formula at \(g\to0\) first requires exchange and equilibration at \(g>0\). At exactly \(g=0\) the graph stays unchanged under the closed evolution. The checked counter-test starts in a line at \(J=2\Delta_b\): the cycle probability stays zero, although the cycle has a lower ground energy. The limit of an equilibrium ensemble and a dynamically blocked preparation are different questions.

#F5. What this tests, and what is still missing

A first laboratory route is a programmable system with three identifiable excitation modes and separate switching dynamics. That also requires implementing the common-neighbor condition; no already available setup is claimed here. Calibrate \(\Delta_b,J,g,T\) separately, fix the measurement basis and the conservation of one excitation, check equilibration, and then predict the cycle occupation at unused J values without readjustment.

A reproducible deviation outside the joint uncertainty model refutes this implementation/Hamiltonian package. Agreement shows that the mechanism works; a simulator that mimics our Hamiltonian does not prove that space is fundamentally built this way. Actual cosmological support requires independent, discriminating consequences for systems not designed by us.

The step to the geometry of the universe remains large and concrete:

1. More than a graph: these one-dimensional complexes have no higher cells and cannot represent K3. Faces, volumes and their attachment rules must come from additional physical interactions; automatically filling in a triangle in fact changes the topology.
2. A large-scale limit: show that the dynamics yields stable, local geometries, with a derivable dimension and measure. It can also yield dense cliques, irregular networks or persistent superpositions. Those outcomes must not be filtered out afterwards.
3. Spacetime and physics: causality, Lorentz behavior, gravity, the matter spectrum and their measurable couplings are missing. A changing discrete graph does not automatically evade the constraints on smooth Lorentzian topology change; the conditions for a possible continuum limit must be tested separately. See Borde, Topology Change in Classical General Relativity.

Verification log: four connected graphs, a full 32-dimensional Hilbert space within this connected sector for links plus hard-core bosons, and a 12-dimensional one-excitation sector have been computed in full. Hermiticity holds exactly and the commutator with the excitation number is exactly zero in the matrices used. Numerical errors in norm and energy conservation are smaller than \(1.4\cdot10^{-15}\); the spectral formulas agree to within \(4.5\cdot10^{-16}\). The small-g partition function and the blocked-graph check have also been verified. No data from nature have been used.

Significance for EEML: in this proposal, topology is a changeable carrier structure of operation. The calculation shows how more links can require less joint energy, while reachability and energy removal remain separate conditions. The new physical content lies in the chosen carriers, couplings and flip rule. EEML has not yet derived these uniquely; PT 0.1 is therefore a concrete candidate to test and extend, not a new proof claim about K3, χ=24 or the shape of the universe.

#G. Memory, sector probabilities and Hawking–Page: check of the additional report

Status, 28-09-2026: literature check and limited calculation. The submitted report uses T1–T5; their full rules had not been supplied at the time of this inclusion, nor were they found in this text. Therefore those labels are not introduced here as new axioms. This chapter checks the concrete claims from the report.

#G1. A 3D construction is not an experimental confirmation

Balasubramanian, Davydova and Lin, arXiv:2605.10943v1, dated 11 May 2026, present a local 3D stabilizer Hamiltonian with a long-lived logical qubit under specified thermal dynamics. The theorem concerns sufficiently low, nonzero temperature and a suitable encoding. The construction does not use a translation-invariant lattice; the authors discuss limitations of TQFT fixed points. A general prohibition for translation-invariant 3D codes is not thereby proven. The theorem and proof outline consulted here are not an independent verification of the full proof.

Roberts et al., Cored product codes report numerical memory times for up to 60,000 qubits. The work first appeared in October 2025, with a revision in January 2026. Within the models studied, it supports the importance of entropy suppression; it is not a measurement on a built memory.

The wording “T5 has been confirmed twice” is therefore not adopted. These model results do, however, support the familiar lesson that an energy barrier without control over the number of error paths gives insufficient information.

An elementary illustration: if a correctly derived error probability is bounded by contributions whose number is at most \(B^m\) and whose energy is at least \(m\Delta\), then for \(\beta\Delta>\ln B\):

Here energy is set against multiplicity. The required probability bound, local dynamics and conversion to a memory time must be proven separately. This geometric series does not replace a self-correction theorem.

#G2. What the lattice coefficient does and does not measure

Lucini, Rago and Rinaldi (2012) find, with Monte Carlo lattice calculations for pure SU(\(N_c\)) Yang–Mills:

These are numerical field theory data, not an observation of a Hawking–Page geometry. Their extrapolation first takes the thermodynamic and continuum limits and then large \(N_c\). A finite \(N_c\) does not rule out a sharp transition at infinite spatial volume. Moreover, this theory is not the same as \(\mathcal N=4\) SYM or minimal AdS₄ supergravity.

An additional algebraic check shows why \(0.458\) does not directly measure the shift of \(T_c\). Write \(T_c=T_0+t_1/N_c^2+\cdots\) and \(\sigma=\sigma_0+s_1/N_c^2+\cdots\). Then:

The measured coefficient mixes both corrections. So even \(t_1=0\) can give a nonzero correction to the ratio. The same power \(N_c^{-2}\) does not identify an Euler term, a log term or an EEML mechanism.

#G3. The shift that does follow

Take two saddle contributions in the same canonical theory, with the same boundary data and renormalization conditions. Use \(k_B=\hbar=1\) and define \(\Delta F=F_{\rm BH}-F_{\rm AdS}\). Assume a controlled expansion

Here \(N_c^2\) stands for the chosen large parameter; that scaling does not automatically hold for every AdS duality. First-order expansion of \(\Delta F(T_*)=0\) gives:

\(\Delta S_0=-N_c^2\Delta f_0'(T_0)\) is the leading entropy difference. With leading latent heat \(\mathcal L_0=T_0\Delta S_0\) and \(\Delta I_1=\Delta F_1/T_0\):

A positive extra black-hole free energy shifts the point of equality upward. If \(\Delta F_1=O(1)\) and \(\Delta S_0=O(N_c^2)\), then \(N_c^{-2}\) follows. With a surviving \(\log N_c\) term, the shift can contain \((A\log N_c+B)/N_c^2\). One-loop order and literally \(O(1)\) are therefore not always the same.

For a compact finite-volume problem, within a positive two-saddle approximation with prefactors included in \(I\), the following holds:

The width between 10% and 90% is then \(2\ln(9)T_0/\Delta S_0\). This is rounding around a transition of the large-\(N_c\) limit, not a finite-\(N_c\) jump in latent heat. Here too, equilibrium probabilities do not determine a switching time or reachability.

The topologies are those of competing Euclidean saddle points. Their competition does not prove a smooth Lorentzian topology change of one universe. The established AdS/CFT context is set out in Witten, Thermal Phase Transition and Confinement.

#G4. A checked AdS₄ partial result

David et al., JHEP 08 (2022) 043, table 1 and equations (3.11–3.15), use \((4\pi)^2a_4=-a_EE_4+cW^2+b_1R^2+b_2RF^2\). Their renormalized integrals give \(C_{\rm local}=0\) on thermal AdS₄. The local curvature invariants themselves are not all zero. For Schwarzschild–AdS₄, \(z=r_+^2/L^2\):

Our own substitution at the classical Hawking–Page point \(r_+=L\) gives:

for the minimal \(\mathcal N=2\) gravity multiplet, with \(a_E=-11/24,\ c=0\). The \(b_1\) contribution drops out there. This is a local log coefficient in the scaling convention of the paper. Zero modes, the choice of ensemble and finite determinant parts remain necessary for the full \(\Delta F_1\); extra fields change the result.

This is not a calculation of \(\delta T\). A log coefficient does not determine the whole function. In particular, \(11/6\) must not be substituted as \(\Delta F_1\), as \(0.458\) or as a universal EEML constant.

The thermodynamic conversion can, however, be done explicitly. For a neutral spherical black hole in AdS₄:

It follows that \(\delta T/T_0=G_4\Delta I_1/(\pi L^2)\). The suppression here is \(G_4/L^2\); the translation to \(N_c\) requires a duality dictionary.

For AdS₅, one has instead \[ T_0=\frac3{2\pi L},\quad \Delta S_0=\frac{\pi^2L^3}{2G_5}. \] With the leading AdS₅×S⁵ dictionary \(G_5=\pi L^3/(2N_c^2)\), it follows that \(\delta T/T_0=\Delta I_1/(\pi N_c^2)\). The AdS₄ coefficient from the previous paragraph cannot be inserted here.

#G5. Scope of the next genuine calculation

The smallest next step is a single AdS₄ benchmark: neutral spherical Schwarzschild–AdS₄ versus thermal AdS₄, fixed \(L,G_4\), the same field content, the same thermal boundary conditions, zero modes and one fixed renormalization prescription. Determine the complete relative one-loop free energy at \(T_0\), with error control, and insert it into G3. Any finite higher-curvature couplings must be fixed independently; otherwise this temperature correction, too, remains model-dependent.

A comparison with \(\mathcal N=4\) SYM on \(S^3\) requires a separate AdS₅×S⁵ calculation and matching at the same coupling. The thermal partition function is not automatically protected when going from weak to strong coupling. Kristensson and Wilhelm, From Hagedorn to Lee–Yang treat finite-\(N_c\) partition functions in the free theory; without further steps, that does not give the strong-coupling correction. The specific weak-coupling shift meant in the report has not been identified here as a fully suitable check.

Execution: 27 exact checks of the coefficient reduction; 16 numerical root checks of the temperature shift in AdS₄/AdS₅ with both signs of a synthetic constant \(\Delta F_1\); a check of \(dF_0/dT=-S_0\), the two-saddle width and the normalization counter-check. The largest linearization error is 0.236% at the smallest large parameter used and decreases by roughly a factor of four upon doubling. This checks the expansion, not a physical determinant. No new lattice data, memory experiment or full one-loop calculation has been carried out.

Consequence for EEML: a small sector contribution can determine the dominant contribution when the leading free energies are nearly equal. That is a conditional thermodynamic mechanism. To turn it into an EEML prediction of its own, EEML must supply an independently determined contribution or relation that the existing theory does not yet fix. The supplied T2/T4/T5 statuses are therefore not adopted as empirical confirmation.

#Specify in advance for every new application

1. Effect and target quantity: what is reached, for how long, with what tolerance? Which failures are possible? Specify in advance which quantity f is to be predicted.
2. Candidates and context: states, trajectories, preparation, system boundary, environment and equivalences.
3. Conditions: what comes from independent measurements, existing theory and new assumptions? For a cost/action family, also give the physical origin of signs and ratios, conventions, renormalization conditions, valid scale and uncertainties. Keep track of all members that are still admissible; an artificially small class is no additional evidence.
4. Input and weight: cost vector, any conversion, measure, weighting rule, normalization and ε. No hidden choice of a desired winner.
5. Process: transition law, time scale, drive, barriers, memory and adaptation. Check stated stationary weights against the dynamics.
6. Look for simultaneous saturation: collect proven bounds and their hypotheses; check mutual dependence; write out the equality conditions; investigate whether they are jointly realizable; determine all equality cases. An infimum that is not attained does not count as a realized minimum. State separately whether saturation is derived from the model or imposed as an additional selection hypothesis.
7. Determine closure: remove only physically equivalent descriptions; report the dimension per component, counts where discrete, and the set of possible f-values. Dimension zero is not the same as a single outcome. Also check whether remaining free directions affect the target quantity.
8. Ablation: determine minimal closing subsets for that same f. Also test which conditions are needed for existence and for saturation. ‘Minimal’ depends on what counts as one condition. Full enumeration can require an exponential amount of work.
9. Negative controls: record what the package does not determine. New hypotheses may turn a negative control into a test case only after visible registration.
10. Independent test: do not use the same outcome for model choice, tuning and confirmation. A retrodiction that has already been used for the design is not an independent prediction.
11. Decision rule: expected effect size, uncertainty, comparison model, falsification criterion and discriminating power. ‘Undecided’ remains a valid outcome.
12. Self-application: also check whether EEML introduces extra assumptions, redefines the effect after the fact, or merely renames known theory. Closure within a class is not a derivation that nature realizes that class.

#Register of mentioned tests and candidate hypotheses

Origin and date of inclusion: review of 28 September 2026; included on 28 September 2026. According to the latest clarification, EEML-C1 is an operational proposal from that review, not an external preregistration. CPH-ν remains the additional candidate hypothesis dated earlier on 28-09-2026. The date attached to a proposal, the reproducibility of a calculation, the registration of a decision rule and the execution of a test are different matters.

EEML-C1: fixed proxy. The supplied linearization of the absolute stability bound is:

Solving for the critical top mass gives:

This is the proxy chosen by the author, in line with the explicit linearization in Buttazzo et al., Investigating the near-criticality of the Higgs boson, equation (64). Using it is not a new NNLO fit or a new analysis of experimental events.

The supplied inputs, fixed for this proposal, are:

These values are input from the review, not current world averages re-established here. The statistic is:

The mass interpretation remains explicit. The direct mass measurement is not simply a precise pole-mass measurement. Write \(M_t^{\rm pole}=M_t^{\rm direct}+b_{\rm map}\). For the central numbers below, \(b_{\rm map}=0\) is chosen as a reference and its uncertainty is varied separately. That is a scenario assumption, not a derivation that the conversion is exactly zero. A nonzero central conversion also shifts the numerator of the statistic.

Substituting gives:

Error budget: checked calculation convention. For the following sensitivity calculation, the contributions are treated as independent and combined in quadrature. Here, interpreting the theory and conversion uncertainties as standard deviations is an additional statistical assumption.

Hence:

Visible arithmetic correction: the supplied summary ‘about 1.5–2σ’ was a rough indication. The full band from this error budget is, rounded, 1.41–2.24. This is a band across two chosen uncertainty scenarios, not a confidence interval for a single estimated parameter. Nor do the numbers prove that 0.5–1.0 GeV is the correct conversion uncertainty; that scale is model information supplied as input here.

When correlations are available, the calculation must be replaced by the corresponding covariance propagation. With \(f=M_t^{\rm crit}\), \(v=(1,-\partial_{M_H}f,-\partial_{\alpha_s}f)\) and input covariance Σ, the linear form is \(v^{\mathsf T}\Sigma v\), plus only the genuinely separate extra contributions. Non-Gaussian systematic errors require a suitable model. Without such a statistical foundation, there is no conversion of this normalized distance into a p-value or discovery significance.

Claim and decision rule: the canonical EEML core contains no generic mechanism for exact criticality. This is the previously delimited theoretical conclusion; C1 predicts no future value or band for the distance. Decision rule: none. A change in \(\delta_{\rm crit}\) therefore does not by itself refute EEML. A later hypothesis about proximity to the bound requires, in advance, its definition, treatment of uncertainties and exclusion limit. A proxy, mass conversion or error budget changed later must be recorded as a visible protocol change.

The comparison with the stronger MPP requirement \(\lambda(\mu_\star)=\beta_\lambda(\mu_\star)=0\) remains open. This requirement and the stability-line proxy are not the same executed test. The proposal starts no automatic monitoring and contains no new empirical dataset.

CPH-ν: Majorana as an additional hypothesis. The original idea is that, for the same observed effect, less remaining continuous charge freedom is preferred, so that a lepton-number-violating/Majorana description is expected rather than an exact Dirac scenario. That is extra model content. A Majorana term can remove one charge freedom and at the same time add other parameters; other interactions can likewise remove freedom. The freedom to be minimized and a necessary or quantitative prediction therefore still have to be specified.

A limited test can, however, be described precisely. Under three light Majorana neutrinos, dominance of their standard exchange and fixed remaining interaction and nuclear-physics assumptions:

With \(a_i=|U_{ei}|^2m_i\), the minimum over the free Majorana phases is \(\max(0,2\max_i a_i-\sum_i a_i)\). An established inverted ordering plus exclusion of the entire band fixed in advance, including experimental and theoretical uncertainties, would refute this limited model package. The sensitivity must be stated quantitatively; ‘ton-scale’ alone does not suffice.

This criterion does not automatically refute all Majorana theories, general CPH or EEML. With normal ordering, mutual cancellation and blind regions can exist. Alternative mechanisms or interference may not be added after the fact to the same tested model. A discovery of neutrinoless double-beta decay supports, under the usual conditions, lepton-number violation and a Majorana contribution, but does not automatically identify the dominant mass mechanism or CPH as the cause. See Duerr, Lindner and Merle.

Operational protocol still incomplete: there is as yet no fixed combination of isotopes, sensitivities, parameter intervals, nuclear matrix elements, statistical decision rule and quantitative CPH probability claim. Nor does the core of EEML predict a unique neutrino mass ordering. The independent test coverage must be reported; an unquantified ‘preference’ cannot be sharply refuted by a single outcome.

The wording ‘risk low as long as normal ordering holds’ remains recognizable in the registration history as a qualitative assessment of limited test coverage. It is not a calculated probability of discovery, exclusion or a false rejection. Normal ordering allows blind regions; inverted ordering can, for the limited standard-exchange model, give a band that can be fully covered. The stated assumptions and uncertainties remain necessary for that.

#Other components: registration and execution kept separate

‘Not yet registered’ does not mean that no substantive calculation has ever been done on it; nor does ‘already discussed or calculated’ mean that a test protocol existed in advance. This inventory preserves that distinction without claiming additional results.

#Four negative controls: actually carried out

Execution: 28 September 2026; constructive model checks, not an empirical test or a prediction registered in advance. The question is whether the current package of structural conditions fixes the intended quantity. For each control, two allowed inputs were constructed that give different outcomes. That refutes uniqueness within that package; it does not say that both models reproduce all measurements of nature.

For the particle controls we use three families with the fixed representations and Yukawa-invariant hypercharges, without CPH, imposed CP symmetry, axion mechanism or additional flavor law. In the left-chiral convention: \[ (q,u,d,l,e,h)=(1/3,-4/3,2/3,-1,2,1). \] The seven residuals \(q+u+h,\ q+d-h,\ l+e-h,\ 2q+u+d,\ 3q+l,\ 6q+3u+3d+2l+e\) and \(6q^3+3u^3+3d^3+2l^3+e^3\) are, in rational arithmetic, all exactly zero. Yukawa invariance fixes charges, not the coefficients of the allowed interactions. The three-family choice also has an even number of SU(2) doublets. This is not a full UV consistency test.

The mixing-matrix check gives a maximum unitarity error of \(2.22\cdot10^{-16}\); the singular values and determinant were checked numerically. The Λ residuals are exactly zero in rational arithmetic. For Λ, it is therefore the corresponding solution that was changed, not the constant within the same flat solution. The chosen units and numbers serve as a counterexample, not as a cosmological estimate.

This freedom is in line with the parameter structure of the Standard Model, PDG. The check concerns the packages explicitly named here, not two fully constructed string vacua that simultaneously satisfy every condition in this dossier. The layer-III measure, a process selection or a new additional law may later provide extra information, but is not available here.

Conclusion of the four controls: the current package closes hypercharges under conditions, but not Yukawa magnitudes, mixing angles, strong CP or Λ. This is a result of the protocol as executed. It is not a failed numerical fit, nor a physical refutation of an EEML prediction made in advance: for these quantities, no closed prediction had been derived in the first place.

#Concrete next step: reconstruction of a process

Start with a process whose events, waiting times, support and effect are separately measurable. For a fully observed time-homogeneous Markov chain, transition counts \(N_{ij}\) and total residence times \(T_i\) can be used to estimate rates:

This requires that observation and model justify this estimate. With limited data, an unobserved transition is not proven impossible.

Check whether a single reconstructed law simultaneously describes occupation, currents and waiting times. Then predict a new blockage or load, fixed in advance, without retuning. Keep track of alternatives that explain the same existing data. A successful reconstruction first supports that process model; for a result that is EEML's own, it must be clear in advance which extra relation EEML predicts.

H1a, H1b and H2 from Part V remain separate hypotheses. Quasipotential agreement is something other than reversibility. Minimal dissipation over all generators is not a proven general law. Setting a cost equal to \(-\varepsilon\log\pi\) after the fact makes agreement tautological and provides no independent test.

#EEML-D1: conditions for elimination and process classification

Revised test proposal, 28 September 2026. D1 operationalizes a conditional elimination claim, not a universal prohibition and not an automatic dichotomy based on the outcome. Physarum is the chosen development example. Literature and outcomes already examined are development information. The registration is frozen only after the measurement and calibration fields have been filled in, and before any inspection of independent test data.

#Revised status: testing conditions, no outcome as proof of class

The strong H_D formulation remains above as history. D1 now investigates which process conditions, fixed in advance, justify elimination, and then tests their quantitative consequence. Survival alone does not identify the dynamics. If a prediction fails, the process model, the coupling to the measured quantities and the hypothesis are examined; the conditions are not adjusted afterwards to make the outcome fit after all.

The connection to evolutionary game theory is useful but not literally identical:

• Pareto dominance compares different cost components for the same effect.
• Strict dominance in a game compares a single payoff for every possible composition of opponents: \(f_Y(z)>f_X(z)\) for all relevant z.
• The conversion of a cost vector into that payoff is additional domain input. “Nowhere worse and somewhere better” across opponents is weak dominance, for which elimination does not follow in general.
• An alternative that initially has a positive share is more than an alternative that is merely reachable. Pure replicator dynamics does not discover a strategy with zero initial share.

For regular dynamics \(\dot x_i=x_i g_i(x)\), payoff monotonicity means that \(f_i>f_j\) also gives \(g_i>g_j\). Convex monotonicity requires the stronger comparison with mixed strategies: \(\sum_i p_i f_i>f_j\Rightarrow\sum_i p_i g_i>g_j\). Within that class, Hofbauer–Weibull (1996) proves elimination of iteratively strictly dominated pure strategies from an interior initial state. The different dominance/monotonicity classes are explicitly distinguished by Viossat. “Every below-average strategy shrinks” is not a complete substitute for these conditions; nor does imitative mean gradient-like.

The negative claim from the literature is also limited. Under continuity, positive correlation, Nash stationarity and innovation, Hofbauer–Sandholm (2011) gives an existence result for games with surviving strictly dominated strategies. This does not mean that they survive in every innovative process. Moreover, there are imitation dynamics with such survival: Mertikopoulos–Viossat (2022). The dichotomy “imitation eliminates, innovation circulates” is therefore withdrawn.

#An exact, finitely testable reference

For the replicator law \(\dot x_i=x_i(f_i-\bar f)\), positive initial support on X and Y and a uniform advantage \(f_Y-f_X\ge\delta>0\), in the chosen unit of time:

For a detection tolerance \(\tau>0\), this yields a guaranteed upper bound \(x_X\le\tau\) as soon as \[ t\ge \max\!\left(0,\frac1\delta \log\frac{x_X(0)}{\tau x_Y(0)}\right). \] If X is initially absent, its share in this replicator reference remains zero. This is a conditional model result, not a universal EEML clock. The payoff normalization sets the time scale. For a time-dependent advantage, the integral of the advantage is decisive. With constant payoffs \(f_Y=1,f_X=0\) and equal initial shares, one has exactly \(x_X(5)=1/(1+e^5)=0.006693\), below the bound \(e^{-5}=0.006738\).

Two counter-checks carried out:

1. For the replicator with payoff matrix \[ A=\begin{pmatrix}0&0&1\\0&0&0\\-1&-1&-1\end{pmatrix} \] strategy 1 dominates strategy 2 only weakly. Starting from \(x=(1,1,1)/3\), one has \(f_2=0\) and \(\bar f=-x_3(1-x_1)\le0\), so \(x_2\) does not decrease. Numerically, \(x_2(30)=0.408248\). Halving the integration step changes the final distribution by less than \(2.4\cdot10^{-12}\). Applying a strict-dominance theorem here would be wrong.
2. With a constant advantage for Y, add symmetric inflow: \[ \dot x=x(1-x)+\mu(1-2x),\qquad x=x_Y. \] At \(\mu=0.1\), the stable equilibrium is \(x=0.909902\). X therefore retains \(9.01\%\), without a cycle: this is autonomous one-dimensional dynamics. This example accounts for inflow, but does not claim to satisfy all conditions of the Hofbauer–Sandholm theorem.

#What is actually measured?

The classification therefore has several axes. Selection monotonicity orders growth by payoff. H1b requires detailed balance of a specified stochastic process law. Replicator dynamics can circulate in non-potential games; conversely, a reversible noise process can keep visiting worse states. Neither axis is read off from a stationary distribution alone. Limited measurement data can exclude a class or be consistent with it, but do not prove all global conditions of a theorem.

#Kramers: a mean waiting time, not a hard end time

For the separate reference \[ dX=-V'(X)\,dt+\sqrt{2\varepsilon}\,dW \] with small noise, a suitable smooth potential, a nondegenerate local minimum \(x_m\) and a separating maximum \(x_s\), the one-dimensional Eyring–Kramers approximation with unit mobility gives: \[ \mathbb E\tau_{\rm escape}\sim \frac{2\pi}{\sqrt{V''(x_m)|V''(x_s)|}}\, e^{[V(x_s)-V(x_m)]/\varepsilon}. \] See Berglund, Kramers' law.

This is an asymptotic mean time for the chosen transition; it is not an upper bound for every realization. Barrier, noise and prefactor must be determined independently. A lower destination does not prove a low barrier. Return remains possible; with finite noise, a higher-cost state can have positive stationary occupation, even under detailed balance. The limits \(\varepsilon\to0\) and \(t\to\infty\) cannot be interchanged without conditions: removing the noise first can permanently lock in a local minimum. D1 therefore reports residence times as well as return and occupation.

#The general comparison form

This form determines effect, costs, dominance and reachability. A subsequent elimination claim additionally requires the independently tested process conditions described above. “Violation” means violation of that claim as named in advance, not automatically that every surviving dominated candidate breaks a law of nature.

Starting condition: choosing a process alone is not sufficient to fill in these fields numerically. The measurement setup and independent calibration determine resolution and error margins. These must not be chosen on the basis of the desired D1 outcome.

#Filled-in Physarum design D1-P (before a new measurement)

Status: concrete design, not a frozen measurement registration. The choices below are protocol input. The literature data were first assessed for usability; no suitable complete dataset has been verified yet. Therefore no measurements, physiological constants or numbers of organisms are invented.

A finite library of reachable Y's can yield a dominance witness, but cannot prove that X is globally undominated. For the replicator reference, a conservative advantage bound from calibration is used to fix the time prediction. A Kramers prediction requires a separately validated noise/barrier model; the exponent is not back-fitted from the same recovery time and then called “predicted”.

Data gate before execution: the complete time series of geometry, flow or tracer, drive, preparation and permitted interventions must be available, with independent replicates and error estimates. The checked Dryad descriptions do not provide all of these data. The mediaTUM repository remained inaccessible on rechecking. The literature survey has thus been carried out, but the analysis of literature data has not yet proved possible. Numbers of organisms, numerical material limit, measurement tolerances and statistical discriminating power are fixed only from independent calibration. Version freezing follows after that; until then, this remains a design.

#Costs: what the √k term really means

For a round tube with length L, radius r and viscosity μ:

D is the conductivity factor before division by length; k is the full hydraulic conductance of the tube. The material/fluid volume of this cylinder is:

“√k costs” therefore require a specified k convention and the correct length factor. This gives a volume proxy. The assumption that metabolic maintenance is proportional to it requires a separate justification. A model-based cost comparison is not a measurement of maintenance energy. Watanabe and Takamatsu (2014) use loss, construction/maintenance costs and vulnerability as different performance measures in a Physarum-inspired algorithm.

In the stationary or justified quasi-static Poiseuille approximation, the first component can be chosen as \[ R_{\rm diss}=\frac1T\int_0^T\sum_e\frac{Q_e(t)^2}{k_e(t)}\,dt. \] With changing tube volumes and active contractions, the full drive, storage and dissipation accounting must be checked. Flow derived from a tube model may not be presented as independently measured flow. For recovery, “not returned within the measurement window” is a censored outcome; it is not silently entered as recovery exactly at the end time.

#Reachability without circular reasoning

A purely optimized network Y is not yet a physically reachable Y. A witness can be an independently validated trajectory or a replicated realization within the same initial and environment class described in advance. Preparation differences, measurement uncertainty and required control are recorded explicitly.

A deterministic process law with one complete initial condition and one fixed input usually yields one trajectory. A set of alternatives arises only through variations, noise or control permitted in advance. Those variations must be included in C and z₀. Nor may the reachability law already assume that all dominated final states disappear: the test would then presuppose its own outcome.

If both Y’s reachability and its low costs are computed exclusively by the process model, the conclusion is conditional on that model. For a physical refutation, this combined measurement and model package must be sufficiently substantiated independently.

#Assessing dominance and persistence separately

Write \(\Delta_a=R_a(Y)-R_a(X)\). Let \([L_a,U_a]\) be joint uncertainty bounds, with the treatment, fixed in advance, of shared measurement errors and the number of candidates examined. A conservative witness of dominance satisfies:

\[ U_a\le0\quad\text{for every component }a,\qquad U_b<-\delta_b\quad\text{for at least one component }b, \] with \(\delta_b>0\) fixed in advance. Effect, constraints and reachability must be sufficiently certain at the same time. Not being significantly more expensive is not proof of not being more expensive. Introducing a small permitted deterioration means an explicit, relaxed variant of dominance and must not be done afterwards.

A found, confirmed Y is sufficient to demonstrate dominance. Not finding a Y with an incomplete search procedure does not prove that no Y exists. Therefore also record the coverage of the search space; without sufficient coverage, “no violation found” remains a limited outcome.

“X persists” gets its own graph/capacity tolerance, observation duration and decision rule. Under its conditions, the replicator reference yields the time bound derived above; the Kramers reference yields an asymptotic mean time. Where such a derivation is lacking, a finite period is additional hypothesis input. A tested \(H_D(T)\) is not equated afterwards with the unbounded “X eventually disappears”. For statistical claims, the permitted survival probability is also fixed in advance.

#Both a negative and a positive control

The negative control contains a cheap Y that demonstrably lies outside the reachable set fixed in advance. That Y must not count as a witness of a violation. The positive control contains a demonstrably reachable Y with the same effect and lower costs; the method must recognize this Y.

Correction to the supplied formulation: the fact that the method finds no violation anywhere does not by itself prove that reachability is too narrow. It can also be a correct outcome, insufficient search coverage or insufficient measurement precision. The positive control tests whether the method can find a known dominance. Moreover, a wrong control answer can be due to the implementation or the measurement model, not exclusively to the reachability definition.

Six synthetic controls have been carried out: unreachable Y, reachable dominating Y, a cost trade-off, uncertainty about not being more expensive, a Y outside the time limit and a Y with the wrong effect. They give, respectively, inadmissible, dominance demonstrated, non-dominating, undecided, inadmissible and inadmissible. The volume identities have also been checked on 36 combinations of radius, length and viscosity. This validates only those computational steps and the decision logic.

#Which literature data can help?

This literature survey uses paper content and data descriptions. No raw datasets have been downloaded or reanalyzed. Published results are already known and count as development/retrodiction for this dossier. An independent test requires unused data separated in advance, or a new measurement. The statement that the complete form can be filled in directly with existing data has not yet been demonstrated.

Evidential weight of an outcome: a valid counterexample refutes the universal prohibition formulation insofar as that formulation actually excludes this case. It does not automatically make all other EEML models and mathematical results incorrect. A domain claim restricted afterwards is a visible revision, not a rescue of the same universal claim. A successful domain test does not prove universality; it only supports the formulation specified in advance and tested, within its scope.

Current D1 status: the scientific status has been revised, the class features and reference predictions have been worked out, and the Physarum design and the data gate have been filled in. The synthetic controls have been carried out. Measurement calibration, numbers of independent organisms and sufficiently complete data are still missing for a frozen test against nature. A miss is not automatically relabeled as “another class”: first, the independently tested premises and measurement errors are examined.

#EEML-P1 · three adaptive routes: checked test proposal

Origin: submitted proposal of 28 September 2026, checked and sharpened in this master version. Status: proposal plus algebraic and synthetic checks; not yet a full preregistration, new dataset or physical confirmation. The existing T 0.3 theorem is retained. The experimental identification of its variables and the choice of an organism or experimental setup are additional hypotheses.

The central question is concrete: can a single pre-calibrated process law predict the course after a new blockage, including route use and structural memory, without retuning? Physarum is a candidate. The model below initially applies to three parallel routes with positive, unambiguously defined transport. A real Physarum network must be tested against that reduction.

#P1.1. Do not confuse measured quantities with effect

Measure \(Q_i(t)\), \(Q_{\rm tot}(t)=\sum_iQ_i(t)>0\), route capacities \(D_i(t)>0\), and a transport performance over a time window, fixed in advance. Define, where the measured flows are non-negative:

What D measures (for example a hydraulic conductivity factor) must be fixed in advance. Do not use a D back-calculated from x to then test the relation between x and h. D, route geometry, pressure/drive and flow must be measured separately where possible; shared measurement errors remain in the uncertainty budget.

Shares do not predict total performance. The same x and h fit both a total throughput of 1 and one of 10. The T 0.3 model contains no equation for \(Q_{\rm tot}\), failure probabilities or the absolute magnitude of capacity. P1 initially tests the distribution and memory development, conditional on the specified drive. Total throughput and effect are checked separately. A prediction of effect success probability first requires an extension with absolute capacity, drive, noise and possible failures. Failed trials must not be silently filtered out.

Physarum moreover exhibits back-and-forth, contraction-driven flow. A raw instantaneous flow is therefore not automatically a probability share. Direction, measurement cross-section, time averaging and, where applicable, tracer transport must be chosen in advance. Summing \(|Q_i|\) measures activity and is not by itself net transport from source to destination. See Alim et al., Random network peristalsis in Physarum polycephalum organizes fluid flows across an individual.

#P1.2. Fully specify the model and the blockage

For this first candidate, choose fixed, independently determined costs \(c_i\) and \(\Phi(x)=\sum_i c_ix_i\). Costs and ε have the same units. The costs must not be chosen after the fact to reproduce measured shares. Fixed length can be a route characteristic; actual hydraulic resistance can change during structural adaptation and is therefore not simply a constant c.

Let A be the set of open routes within one phase. Define:

x lies on the simplex of A; h stays on the simplex of all three routes. On the closed route x=0, and the memory decays as \(h_i(t)=h_i(t_b)e^{-\eta(t-t_b)}\). That memory is not set to zero.

As an explicit ideal intervention law upon closing, we choose \(x_i^+=0\) outside A and \(x_i^+=x_i^-/\sum_{j\in A}x_j^-\) inside A, provided the denominator is positive; \(h^+=h^-\). This is an imposed instantaneous redistribution, not a derived law of fluid flow. A real setup with storage, a different flow response or a changing throughput requires a different intervention law, described in advance. Upon reopening, only A changes, and x and h are not reset.

With the convention \(0\log(0/h)=0\), we keep \(\Psi=c\cdot x+\varepsilon D_{\rm KL}(x\|h)\). Within a fixed phase with positive shares on A:

The KL terms between x and p are taken over their common support A; the last sum covers all three routes. The same derivation works because the constant normalization term drops out against \(\sum_{i\in A}\dot x_i=0\). The jump upon closing is allowed. Upon reopening, Ψ is continuous for this intervention law, while a reopened route with x=0 and p>0 gives a one-sided infinite entropy slope. Use finite time differences there; do not add an artificial small flow.

Renormalizing memory changes the equation. For \(H_A=\sum_{i\in A}h_i\) and \(z_i=h_i/H_A\) on A, the following holds:

Replacing h by z with unchanged η is therefore a different model. With time-dependent costs, \(\dot c\cdot x\) is added to \(\dot\Psi\); with costs that depend on structure, the corresponding chain-rule term is added. “Within a phase” therefore also means that these hypotheses actually hold.

#P1.3. When is capacity really the T 0.3 memory?

Normalizing a capacity does not automatically make its evolution law \(\dot h=\eta(x-h)\). For the explicitly chosen reinforcement law \(\dot D_i=aQ_i^\mu-\gamma D_i\), with non-negative \(Q_i=Q_{\rm tot}x_i\), it follows that:

For μ=1 this becomes the T 0.3 form, but with \(\eta_{\rm eff}=aQ_{\rm tot}/D_{\rm tot}\). A constant η requires a constant ratio or a justified approximation. With fixed Q and preparation \(D_{\rm tot}=aQ/\gamma\), that ratio is indeed constant and η=γ. This is a conditional connection to known adaptive flow models, not a proof that every measured capacity memory follows this law.

There is moreover an exact overlap with ordinary flow physics. For three passive parallel channels with the same pressure drop and with conductivity factor D, the following holds: \[ Q_i=\frac{D_i}{L_i}\Delta P,\qquad x_i=\frac{h_i/L_i}{\sum_jh_j/L_j}. \] The algebraic logit target distribution has the same form if one chooses \(c_i=\varepsilon\log(L_i/L_0)\). That is a demonstration of possible indistinguishability, not permission to introduce this cost after inspecting the data. The inertia, capacity law and independently fixed cost remain testable. If D is already the full channel conductance including length, this simple hydraulics gives precisely x=h.

A hydraulic adaptation model is therefore a necessary reference as soon as P1 is applied to real tubes. Ito, Johansson, Nakagaki and Tero describe an existing combination of Kirchhoff flow and flow-reinforced conductivity. Recognizing a related memory law is not in itself a new result about nature.

#P1.4. Calibration requires change, not only a quiet baseline

In an equilibrium of T 0.3, x=h=p. For every route in use, the constants c_i are then equal. With a strictly cheapest route, the limit can lie on the boundary. With equal costs and x=h, all positive κ, η and ε are compatible with exactly the same resting measurement. A quiet baseline therefore does not identify these parameters; interior equilibria with different constant costs do not exist here.

Therefore use a calibration phase with sufficient transients, separated in advance: known preparations or small, predetermined perturbations. Keep the later blockage/reopening trials as test data. Before that test, check whether the time resolution can distinguish κ and η, and whether the cost contrasts can determine ε. Report the joint parameter uncertainty; freeze the estimation method and the uncertainty distribution, not just three rounded numbers.

Two additional relations make the content concrete:

\[ h_i(t_2)-h_i(t_1)=\eta\int_{t_1}^{t_2}(x_i-h_i)\,dt, \] \[ \log\frac{(x_i+\dot x_i/\kappa)h_j} {(x_j+\dot x_j/\kappa)h_i} =-\frac{c_i-c_j}{\varepsilon}\qquad(i,j\in A). \]

The second relation holds where the reconstructed p shares are positive; numerically differentiating noise requires a measurement model fixed in advance. The first can be tested directly with integrals. Both require the same parameters for all routes and test phases. Earlier, a nonlinear Φ was used in T 0.3; the interior equilibria found there are no justification for a flat baseline in this new linear variant.

#P1.5. References must be as concrete as EEML

“A freer memory model” does not yet specify a testable rival. A delimited candidate family is: \[ p_i^{A,\rho}\propto\mathbf1_{i\in A}h_i^\rho e^{-c_i/\varepsilon}, \qquad \dot x=\kappa(p^{A,\rho}-x),\qquad \dot h=\eta(x-h). \]

Here M0 is given a passive structural variable so that the same measurements can be assessed; h does not steer its route choice. A purely static formula without a time law is not a full-fledged rival for the recovery rate.

M1 contains M2. Therefore “M2 must always predict better than M1” is not a necessary consequence of M2. Equivalent prediction may in fact mean that the extra freedom is not needed, or that the experiment does not distinguish them sufficiently. Compare prediction errors and uncertainty on unused data, with a predetermined equivalence margin. A required statistical win over every broader model is not a valid automatic falsification rule.

The ablations are likewise precise: η=0 switches off learning but keeps a fixed route preference; ρ=0 switches off the feedback of h on choice; equal costs remove the cost contrast. “h constant” and “η=0” are the same ablation within this law, not two independent pieces of evidence.

#P1.6. Decision rules: the decrease is necessary, but not sufficient

A different positive time scale can make the same Ψ decrease and still give the wrong recovery rate. The check below doubles both rates only after reopening: even then Ψ decreases, but the route shares deviate from the prediction frozen in advance. Therefore assess the full course of both x and h, with Ψ as an additional model check. The final occupancy alone does not count as a process test.

The proposed test statistic \[ V_+=\sum_k\max\{0,\Psi(t_{k+1})-\Psi(t_k)-\delta_k\} \] can be useful, but “approximately zero” is not a decision threshold. Measurement times, windows, δ_k, shared measurement errors, parameter uncertainty and a threshold for V_+ must be fixed ahead of the test data. Pointwise error margins do not automatically protect against a false alarm over many time steps. Calibrate the overall decision rule with the measurement model chosen in advance and report its discriminating power.

The decrease theorem applies to the deterministic model state. Measurement noise or physical fluctuations need not give a monotonic measured trajectory. A stochastic extension requires its own derivation; noise must not be invoked without limit after the fact to explain every deviation. Compute Ψ from independent observations with their measurement model, not only from the curve predicted by M2.

Still missing for the experimental registration: concrete setup, operational D and c, direction/averaging rule, intervention response, number of independent networks, calibration and test windows, blockage durations, measurement frequency, detection limits, error model, parameter ranges, prediction score and numerical decision thresholds. Several blockage durations are useful for distinguishing mechanisms, but their choice must be frozen before the test. A recovery criterion must also be able to report “not recovered within the measurement window”.

Provisional test claim: a single fixed M2 parameter distribution predicts, after new interventions, the route shares and the measurably defined memory, within predetermined error bounds and alongside explicit rivals. A verified violation rejects this M2-plus-measurement model in the named domain. A successful test supports that process model; a tie with known hydraulics does not yet yield a distinctive EEML law of nature.

#P1.7. What has now actually been checked

Without new measurement data, the following were carried out: 1,000 states for the Lyapunov identity, spread across open and blocked phases; the conversion to normalized active memory; 27 parameter combinations that give the same quiet baseline; 300 normalizations of capacity laws; 300 examples of hydraulic-logit overlap; and a three-part synthetic trajectory with blockage and reopening, including halving of the integration step.

The largest numerical deviation in the Lyapunov identity is \(1.78\times10^{-15}\). Halving the time step changes the computed trajectory by at most \(1.74\times10^{-12}\). The two processes with different rates after reopening both have decreasing Ψ, but differ by at most 0.239 in total variation distance of the route shares. These are checks of chosen, dimensionless examples; not measurement precision or an effect size from an organism.

Furthermore, counterexamples were checked for “decrease determines the recovery rate”, “the same decrease identity holds for a freer memory coupling”, “changing costs automatically preserve decrease” and “shares determine total throughput”. The results test the proposal for internal consistency. The experiment itself has not been carried out.

#What remains open

• An independent physical derivation of relevant costs, measure, noise scale and process law per domain.
• An extra relation between the four layers that differs empirically from existing domain models.
• Independent measurement calibration and execution of D1-P, with class conditions measured in advance and the corresponding quantitative prediction; in addition, a concrete experimental model package for H_R.
• The regime of validity of an EEML selection or adaptation law; a universal minimum has not been proven.
• A new independent dataset with sufficient discriminating power.
• Topology selection and cosmology: the K3 bound does not replace the missing physical selection process.

For anyone building on this, researcher or AI, the priority is: treat EEML as a proposed theory of how processes work; keep the corrections, choose one delimited process and formulate one new discriminating test. Material laws and constants may be explicit input. K3, 24, 137, a unique cosmic zero point or a mandatory winner must not be fixed in advance as the outcome.

This compilation uses previously stored check outcomes from the local dossier of 28 September 2026. The existing models have not been changed. Older extensive simulations and empirical analyses have not been rerun; later additional checks are listed separately below. The model checks are not independent measurements of nature or proof of a universal EEML law.

#Nine statistical checks: previously passed

#Eight correction and process checks: previously passed

The hypercharge ablation is a separate, previously checked result: nine minimally closing sets among 128 subsets of seven conditions. The K3 bound rests on the geometric theorems mentioned; elementary computational checks do not replace their hypotheses.

Empirical status: this compilation analyzes no new dataset and confirms no universal EEML law. Old numerical derivations do not retroactively become valid through the new architecture.

#Additional review of 28 September: five targeted checks

In this revision, the following were checked again: the reparametrization under \(g=\lambda^2\) including \(F_{\lambda\lambda}(0)=3\) via central differences, the monotone finite positive-operator check \(Z=1+e^{-g}\), the rewriting of the curvature costs including the Euler term, the four-dimensional scaling factor and the 331 intersection with the QCD bound. All five pass. The scaling argument in fact shows why boundedness from below does not prove coercivity.

These are algebraic and finite model checks. The execution of the external recalculations has not been independently inspected. No new K3 metric, quantum gravity solution, 6D anomaly calculation or empirical test has been produced. The five checks therefore do not justify an upgrade to ‘proven universal law of nature’.

#Supplement to the registration inventory and operational C1 check

The 58-dimensional K3 metric family has been checked against a primary source. The sign conditions, maximum versus supremum and possible divergence of the macro-normalization remain explicit.

With the subsequently supplied C1 definition, the inverse stability line, the central mass distance, the four error contributions and both conversion scenarios have been recalculated directly. Back-substitution into the chosen forward line returns the input Higgs mass. The correct scenario distance is 1.41–2.24. This check confirms the arithmetic under the assumptions made, not the statistical model for the top-mass conversion and not EEML as a law of nature. No new NNLO calculation, empirical test or external preregistration has been carried out.

#Check of the derivation attempt for α

Newly carried out: eight exact rational Hodge block checks, 256 nonzero integer flux pairs for the minimum and the convexity, two examples of equal minimum cost with different x*, five relaxation trajectories and the one-loop check of the free initial value. All checks pass. The dynamical calculation uses the explicitly chosen gradient flow. These are checks of the described toy model, not a spectrum of a concrete K3, not a complete compactification and not a prediction of 137. The text on the spectral action and the geometric kinetic matrix has been checked against primary sources.

#Follow-up check: closure and physical flux sectors

Newly checked: five exact matrices with equal flux cost and different probe norms, a massless combination with a nontrivial axion charge matrix, and all nine positive integer flux assignments of the explicitly bounded IIB subfamily. For each assignment, even quantization, the two complex flux equations, charge cancellation, the norm minimum and the product-of-cosh identity have been checked. For the last identity, 25 rational ratios of s/s₀ and t/t₀ have additionally been recalculated for each assignment. All checks pass.

The dimensions 38 and 54 follow from the described geometric count, not from a computed K3 spectrum. The zero values of the normalized no-scale potential hold within the leading description, not including all string corrections. No duality classification, new 10D solution, realistic photon spectrum or α_em(0) calculation has been made. It is precisely the sharpened closure definition that preserves the distinction between those missing steps and the passed checks.

#Latest run: D1, H_R, negative controls and coefficient bridge

On 28-09-2026, the four structural negative controls were carried out, together with a strict replicator reference, the weak-dominance example and the inflow example without a cycle. For the integrated replicator trajectory, two time steps were compared. The dephasing and two-qubit checks for H_R were verified analytically and numerically. The K3 basis conversion was recalculated exactly; the impossibility of sign agreement follows from the given general algebra, not only from a finite parameter sample.

The new check outcomes appear in full alongside their models in this text. The literature survey for Physarum did not yield a fully verified dataset for D1-P. No Physarum experiment, independent 6D expert check or new collapse data analysis is claimed.

#Provenance

Additional check (3D memory and Hawking–Page): VI.G distinguishes between theoretical constructions, numerical lattice results and gravity calculations. The local coefficient \(11/6\) within the stated field content and the formula \(\delta T=\Delta F_1/\Delta S_0\) have been checked. The free-energy correction itself has not been fully calculated. The coefficient \(0.458\) has not been recorded as a measurement of an EEML or topology term.

Latest candidate elaboration (EEML-PT 0.1): a dynamical graph model has been added, with a Hermitian Hamiltonian, conserved excitation number, a conditional energy crossing at \(J/\Delta_b=1+1/\sqrt2\), thermal graph probabilities and a counter-check for blocked dynamics. The calculations are given in full in VI.F. This is a new proposal within the dossier, related to existing Quantum Graphity models; not a claim of priority in the literature or proof that the space of the universe is a graph.

Brought together here are the integrated working version, the multiplicity note, the process–weight check, the K3 bound check and the hypercharge/quantum consolidation. The language of operation and eight core rules are summarized explicitly. The additional review of 28 September has been incorporated with visible limitations and checked primary sources. Public sources appear directly alongside the claims. Their results support the named existing mathematics and physics, not automatically EEML as a new law.

Current conclusion: EEML is being developed as “a theory of how it works”. The current content is a four-layer framework with conditional results and explicit process tests. D1 investigates independently measurable conditions for elimination; H_R adds a separate information-conservation hypothesis that is testable to a limited extent. The four negative controls show concretely which quantities the current package does not fix. The comparison with entanglement equilibrium yields no derivation of the positive K3 costs. Empirical execution and a distinctive process prediction of its own remain the next substantive step; a universal EEML law has not been proven.