Concept study / Tests & research methodsConceptual illustration. Shapes, colors, and motion are not measured data.
A broad theory project attracts papers from many fields, but citation volume is not a measure of confirmation. The relevant question is what each source supports. An experimental paper can establish a local memory effect; a mathematical paper can prove a theorem under assumptions; a simulation can explore a model; a philosophical argument can clarify interpretation. None automatically supplies observations of Alpha 1. The proposed evidence scheme therefore records both the kind of result and its distance from the claim. It should help readers identify a strong established foundation, a plausible analogy and an untested extrapolation without flattening these distinctions into one percentage.
Use descriptive dimensions
Record whether evidence is direct, indirect or analogical for the exact claim. Separately record design quality, uncertainty treatment, availability of data and code, and independent checking. A theorem’s proof status differs from an experiment’s replication status; neither is well represented by the same star rating. A preprint is not automatically wrong and peer review is not a guarantee of correctness. The project should show publication type and known limitations plainly, then assess the actual method. Where assessment has not been performed, label it unreviewed rather than implying validation.
Avoid counting dependent evidence repeatedly
Several papers may analyze the same observations, use the same calibration or repeat one citation. They are not equivalent to several independent measurements. Link shared datasets and methods in the evidence map. Separate replication from reanalysis, and distinguish a review’s synthesis from a new observation. If studies disagree, describe differences in conditions, observables and uncertainty before averaging conclusions. A hypothesis can remain promising while its evidence is limited; the useful output is a list of what would most reduce uncertainty, rather than an accumulation of vaguely related supportive references.
Make labels reversible and inspectable
Suggested claim states are established within a stated domain, supported local result, open research question, project hypothesis, and conceptual illustration. These are editorial labels for the atlas, not a recognized universal grading standard. Each label needs a short rationale and date. A source update or failed replication can change the label without erasing prior versions. The website should make counterevidence as accessible as supporting material. For the Hill framework, this means that familiar geometry effects can be presented confidently while the nested-universe and universal-learning extensions remain explicitly unresolved.
The mathematical lens
A list of separate assessment dimensions, not an additive numerical score or probability that a theory is true.
Turn it into a question.
- Question
- Does each source directly support the sentence attached to it?
- Method
- Have reviewers independently map claims to source findings and identify extrapolations.
- What would count against it?
- A source supporting only an analogy cannot justify labeling the broader hypothesis as demonstrated.
Judge evidence at the level of the claim it actually addresses.
Follow the original work.
These sources support the mechanisms and limits discussed here. They do not endorse Hill’s framework. Notes distinguish research proposals, reviews, and observations.
Continue the investigation
Edition 02 · 18 October 2026 · AI-assisted synthesis. See editorial method.