FIELD / 05 · EVIDENCE, UNCERTAINTY & EVALUATION
Read generously.
Test rigorously.
A respectful standard for independent creative and research-informed work: neither applause nor dismissal is a substitute for examining what is actually claimed.
Five labels, five different kinds of claim
Something directly noticed or documented, limited by how it was measured and by who observed it.
An explanation of observations; plausible explanations can compete and can be mistaken.
A proposition that could be investigated or disproved through appropriately designed tests.
Systematically gathered data with described methods and limitations, not just a vivid example.
A prototype, essay, narrative or simulation designed to explore a possibility; meaningful, but not inherently a scientific test.
A five-minute reader's protocol
- Identify the strongest claim. Is the author reporting an observation, stating a value, proposing a mechanism or promising an effect?
- Find the original source. Does the linked paper actually support this particular statement—or only something adjacent?
- Generate an alternative. What else could explain the same story or simulated result?
- Ask what would change the conclusion. Does the claim admit a counterexample or a possible failure?
- Separate utility from truth. An essay can help someone think without establishing that its framework is scientifically novel or generalisable.
How to provide better criticism
'I disagree' starts a conversation but gives little to work with. A more useful contribution names the exact passage, the kind of error, the better alternative and the supporting source. For example: 'This passage treats motivation as sufficient, but the cited model also requires ability and a prompt. Could the wording be narrowed?'
Fair criticism is not the enemy of independent work. It is one method by which work can become clearer, more reliable and sometimes less grandiose.
What visitors can and cannot establish
Personal reflections can illuminate how a reader understood a claim or experienced an interface. They cannot, by themselves, validate a psychological construct or establish efficacy across populations. A real evaluation might involve preregistration, explicit outcomes, comparison conditions, adequate sampling, ethics review where appropriate, independent analysis and replication. The specifics depend on the question.
The independent frameworks and essays collected here have not been presented as peer reviewed or empirically validated. No diagnostic, therapeutic or professional outcome is promised.
A small act of epistemic freedom
Read one argument you expect to like and identify its weakest assumption. Then read one you expect to resist and identify its strongest point. You do not have to change your mind. You only have to let evidence have a fair hearing.