Why a uniform standard

Why a uniform standard matters

Consciousness studies has long operated with different evidentiary expectations depending on the ontological family to which a theory belongs.

Mechanistic claims are typically expected to meet experimental standards. Metaphysical claims are often permitted to rest on argument and coherence. Anomalous claims have sometimes been excluded from serious consideration before their internal structure has even been measured.

This uneven arrangement protects no one over time. It allows weak physicalist claims to borrow authority from neuroscience, weak non-physicalist claims to borrow authority from philosophical sophistication, and potentially strong anomalous results to remain outside ordinary comparison.

A single, ontology-neutral standard removes these exemptions at once. It tells every position the same thing: state what would count against you, expose your predictions to procedures capable of failure, and report your standing under rules fixed before the result is known.

The five requirements

CSTP evaluates theories according to five explicit requirements, each drawn from established principles in the philosophy of science and contemporary scientific practice.

A theory must generate at least one falsifiable prediction. A claim that can accommodate every possible outcome cannot be tested against alternatives on empirical grounds and therefore lies outside the scope of the standard.

Predictions must be assessed according to the severity of the tests to which they have been subjected, where severity refers to the likelihood that a test would have detected error if the prediction were false.

Those tests must be independent, relying on data or assumptions separate from those used to construct the claim.

Key results must be reproducible and supported by independent lines of evidence or investigators.

The program's later development must be judged according to its willingness to maintain empirical risk rather than insulate its core claims from potential refutation. A program is progressive when it continues to generate novel predictions that are subsequently supported by new evidence. It becomes protective when it survives primarily by narrowing its empirical scope or by making its central claims more difficult to test.

These five requirements are the principles. The six levels and the gate are how they are applied to a published record: the gate asks for requirement 1, levels 2 and 3 for prospective exposure, level 4 for blinding, level 5 for reproducibility, and level 6 for severity in its strongest available form — a test designed by the position's own critics.

Applied under the governing condition that the scoring rules are fixed in advance, they provide a consistent basis for comparing theories that differ substantially in content and method.

Offered provisionally

The standard proposed here is offered in a deliberately provisional form. Its requirements are stated precisely enough to be applied, contested, and revised. The grading presented in this paper is the first such application rather than a finished product. This is the ordinary and respectable way an evaluative instrument enters a field.

The GRADE system for rating the quality of evidence began as a working proposal from a small group, was adopted provisionally, tested in practice, and revised as its weaknesses became visible. The CONSORT guidelines for reporting randomized trials followed the same course. The PRISMA guidelines for systematic reviews developed through the same iterative process.

In each case the instrument earned its authority not by arriving complete, but by being explicit enough to be tested and corrected in the open.

What it changes

A uniform standard also changes the incentives. It tells a physicalist theory that correlation alone does not settle ontology. It tells a non-physicalist theory that philosophical coherence alone does not establish reality. It tells an anomalous-state program that testimony must be converted into structured, reproducible data. And it tells the critic that dismissal must identify a failed rule, a better null, a broken prediction, or a superior explanation, rather than resting on the location of a claim on the map.

It changes the burden of debate as well. The immediate questions become procedural rather than metaphysical: Were the claims fixed before the relevant tests? Are the data independent of the theory's construction? Do the nulls preserve what they should preserve? Can another operator reproduce the coding? Does a competing explanation make predictions at least as precise?

None of this rests on CSTP being final. The instrument is provisional; the principle is not. What should not survive is the present arrangement, in which theories are admitted, dismissed, or protected by different rules and no shared standard is applied at all. The remedy for an incomplete standard is a better standard, tested in the open. It is not the absence of one.