What is Prephysics?

A test before the final scientific verdict

Physics usually asks which theory best explains the data. Prephysics starts one step earlier: is the concrete claim warranted in exactly the form in which it is made?

A measurement can be reliable without warranting every interpretation. A model can calculate successfully without supporting every claim about the world. The method therefore separates three questions that are often run together in scientific debate.

What requirements does the claim need?

Before judging a result, it must be clear what the claim is physically about, how formal quantities connect to regime-local evidence, which variations preserve the claim, where verdict-relevant steps come from, what carries the asserted physical significance, and which choices shape the result.

Show the six canonical conditions

A1 · Addressability Is the claim stably anchored to a physically identified target?

A2 · Interface Control Is the access route between formal description and regime-local reference declared?

A3 · Stability Does the claim's content remain invariant under admissible, claim-preserving variations?

A4 · Traceability Can the inferential route from inputs to outputs be located and audited?

A5 · Carrier Adequacy Do the declared formal, evidential, and methodological commitments carry the asserted physical significance?

A6 · Constraint Disclosure Are model choices, selection rules, comparators, and other outcome-relevant constraints explicit and tracked?

P1 v3.0. The current foundation grounds A1–A6 in four goal-relative functions — Discernibility, Re-identifiability, Constraint Structure, and Ordering Capacity — linked by a partial rather than linear dependency structure. The conditions are justified through explicit two-component projection warrants. Current P1 ↗

What warrants the transition?

Arguments move among three roles: applicability conditions, theory/model, and observation/application. These roles are not a temporal ladder. An inference can move in different directions among them, and each directed transition can owe additional warrant.

Applicability conditions ↔ Theory & model ↔ Observation & application

Technically, P2 distinguishes nine ordered routes among these three roles. The full route table is available below.

What follows for the concrete claim?

P3 freezes the exact claim wording and the contemporaneous evidence state. It then tests the activated requirements for that claim token. The result applies only to that concrete claim and that time-indexed record.

PASS OPEN Loss of Applicability · LoA

PASS does not mean true, and OPEN does not mean false. Loss of Applicability requires a positive condition; it cannot be inferred from mere lack of a repair idea.

Technically, P3 distinguishes two routes: LoA-T for an essential illicit role transition inherited from P2, and LoA-K for positively certified impossibility of claim-preserving closure.

The architecture at a glance

From a finding to a warranted public statement

The large diagram shows the logic in a first-reading view. The full technical architecture remains directly accessible beneath it.

Public overview of applicability diagnostics with claim-relative L1–L3 roles, canonical A1–A6 condition types, P3 verdicts, and a separate S1–S7 pattern library — Public overview with a clear separation between the P1–P3 core and the separate S1–S7 pattern library.
Open large diagram → Research overview →

Separate pattern library · P9

What do S1 to S7 mean?

The seven codes name structural failure patterns with explicit trigger and non-trigger boundaries. P9 v3.0 boundary-tests the library against a source-bound fourteen-case corpus; its published calibration reports no verified firings in that baseline. They are not additional P3 verdicts.

S1 · Hidden presuppositions

Necessary presuppositions are left unstated.

S2 · Missing observational link

The route from a theoretical claim to possible observation or assessment remains unclear.

S3 · Level confusion

A precondition is mistakenly presented as a physical result.

S4 · Circular explanation

The explanation already presupposes what it is meant to explain.

S5 · Scope overextension

A claim is carried beyond the range warranted by its declared preconditions.

S6 · Regress of presuppositions

Each justification already demands another precondition without identifying a clear stopping point.

S7 · Deriving what is already presupposed

A theory needs certain foundations to derive claims at all. S7 occurs when it is asked to derive those very foundations itself.

P9 · Functional Incompleteness Argument (FIA). FIA gives the functional minimum for an Applicability-Diagnostic Function: a complete, explicit and auditable diagnosis must make three things recoverable — what the claim owes, what warrants the support-relevant inferential transitions, and whether the requirements and warrants are satisfied by this versioned claim. None is recoverable from the other two. P1 → P2 → P3 is one concrete realisation, not the only possible one. P9 is the necessity and pattern anchor, not a fourth verdict layer.

P9 · v3.0 · 27 August 2026. v3.0 · Version DOI ↗ · Concept DOI ↗

Technical depth

Nine possible transitions

For readers who want the full grammar: P2 distinguishes the nine directed routes among the three claim-relative roles. The table is a deeper layer and is not required for the first reading.

Show the route table

Condition → condition How do necessary conditions depend on one another?

Condition → theory How is a condition translated into a theoretical commitment?

Condition → observation How is a condition operationalized as an empirical test?

Theory → condition Can a theory establish a precondition that it already needs?

Theory → theory Is a derivation, reduction, or model transfer warranted?

Theory → observation How is theoretical content carried into an observable claim?

Observation → condition Does the evidence genuinely assess token satisfaction of a necessary condition?

Observation → theory Does the result warrant attribution to the asserted theoretical structure?

Observation → observation Is a generalization, combination, or scope extension warranted?

See examples Publications →

← Back to home Open expert map →