The Principles of PT

The Idea

PT is not just a collection of formulas that return good numbers. Its strength comes from a harder discipline: a formula is admissible only if the structure forces it.

That means a PT result must not depend on a coefficient chosen after the fact, a regression, a best fit, or a local adjustment to save an observable. It must follow an explicit route.

In one sentence: PT tries to replace free parameters with constraints.

1. No continuous fitted parameter in the core

The PT core contains no continuous knob that can be turned to match data. The symmetry constant $s = 1/2$ is derived from the mod 3 structure of the sieve; it is not chosen.

The rule is simple: if a quantity can be moved continuously to improve agreement with experiment, it is not a canonical PT ingredient.

2. Every number must have a reason

A number cannot appear because it is convenient. It must be:

• an integer or prime forced by structure;
• a rational fraction coming from counting or geometry;
• an algebraic or trigonometric function of already derived objects;
• an observable obtained through the full PT chain.

The editorial test is severe: why this number, and not a nearby one?

3. Numerical bricks come from arithmetic geometry

Typical PT numbers are not free constants. They are small integers, prime fractions, angles, holonomies, or anomalous dimensions justified geometrically.

Examples: $s = 1/2$, $\{3,5,7\}$, $\mu^* = 15$, $C_F = 4/3$, $Q_{\rm Koide} = 2/3$, $\gamma_p$, $\sin^2\theta_p$.

The logic is not: “which coefficient works?” The logic is: “which coefficient does the geometry force?”

4. An observable must follow one route

A PT observable must not come from a catalogue of candidate formulas. It must follow a route:

If several routes coexist, they must be separated by status: theorem, identity, derivation, validation, or prediction. A route does not become canonical merely because it gives a smaller numerical error.

5. Every coefficient must be justified

An admissible coefficient must come from an entropy principle, an equilibrium condition, a symmetry, an algebraic identity, or a geometric constraint.

This also applies to precision corrections. A PT correction must have a sign, an amplitude, a domain of activation, and a structural trigger. Otherwise it is only a numerical patch.

6. Bifurcations must be closed

PT meets natural bifurcations:

• which sieve?
• which divergence?
• which metric?
• which fixed point?
• how many active primes?
• which product form?

The rigidity rule is that these bifurcations must not remain free. They must be closed by theorem, identity, structural derivation, or a clear falsification test.

7. The discrete marks points of persistence of the continuum

PT does not say that the continuum magically comes out of the discrete. It reads discrete points as the remarkable points of a continuous mechanics of phase, Fisher geometry, holonomy, and persistence.

Discreteness is therefore the marking of positions where the continuum persists, resonates, closes, or becomes readable.

This formulation matters: it prevents PT from becoming a naive atomistic theory. The deeper mechanics is continuous; the sieve reveals its points of persistence.

8. Rigidity matters more than an isolated precision

An isolated numerical match is not enough. A good PT formula must remain coherent inside a cascade.

If one brick appears in couplings, chemistry, nuclear physics, or cosmology, it must not change value from one domain to another. The same structure must carry multiple observables.

That is what separates a rigid theory from a clever fit: a fit improves one line; rigidity holds an architecture together.

9. Negative predictions count

An overfitted theory often tries to absorb everything. PT also does the opposite: it forbids things.

Examples of negative predictions: no dominant WIMP, no low-energy SUSY, no dominant axion, no proton decay in the usual channels.

These exclusions matter because they reduce escape routes. They give the theory clean ways to be wrong.

10. Epistemic status must remain visible

PT does not have the same level of certainty everywhere. It distinguishes:

• theorem;
• identity;
• derivation;
• bridge;
• validation;
• prediction;
• open hypothesis.

This discipline does not weaken the theory. It makes it readable. An honest PT page must say what is proved, what is derived, what is validated, and what remains exposed to future measurements.

11. Zero parameter does not mean zero mechanism

This is the subtle point.

In applied domains, PT may use corrections: relativistic, radial, inter-channel, topological, molecular, nuclear. But an admissible PT correction is not a fitted parameter. It must be a closed expression, derived from PT bricks, with a structural justification.

The right formulation is therefore:

zero fitted parameter does not mean zero physical layer; it means zero free coefficient inside the added physical layers.

A correction is acceptable if it answers four questions:

1. Which PT mechanism forces it?
2. Which closed amplitude does it give?
3. On which domain does it activate?
4. Which measurement could falsify it?

If those four answers exist, the zero-parameter logic has not been left behind. The cascade has simply gained one more PT mechanism.

Why These Principles Matter

Without these rules, PT could become a machine for producing coincidences. With them, it becomes much riskier: every number must be defensible, every correction traceable, every bifurcation closed, every prediction allowed to fail.

That rigidity is exactly what makes the theory interesting. PT does not only ask: “is the number good?”

It asks:

And that is a much stronger question.

Sources

This page synthesizes the methodological principles in the monograph preface, the “What 0 fitted parameters means” section, the audit chapter on structural overfitting, and the PP1–PP5 principles from the research journal. Point 11 follows the monograph’s nuance: applied corrections may be numerous, but they must remain closed, derived, and non-fitted.

← All essays