What features of knowledge have an impact on its reliability?

What this prompt is really asking

Each of the twelve TOK concepts gives you a different way into a prompt. These are the ones that do the most work here.

• Certainty

  Status of Knowledge · primary lens

  Reliability and certainty are easily conflated. A reliable method can leave a knower unsure, and a confident knower can be using an unreliable one.

  The degree to which doubt about a claim has been closed off. Certainty is a state of the knower, not of the claim, which is why a claim can be true without anyone being certain of it and why confidence and reliability can come apart badly.

• Evidence

  Grounds of Knowledge

  Most features that make knowledge reliable are features of how its evidence was gathered rather than of the claim itself.

  What is offered in support of a claim. Evidence is not the same as proof: it raises or lowers how likely a claim is, and its force depends on how it was gathered, what it was gathered for, and what else it is consistent with. Asking what would count as evidence against a claim is usually more revealing than listing evidence for it.

Where the knowledge framework bites

Reached through this prompt’s concepts, in the framework’s canonical order.

Scope

Certainty: The certainty available in a field is a feature of the field, not a measure of its rigour. Mathematics can offer proof because of what it is about; the human sciences cannot, and reading that as a deficiency mistakes scope for weakness.

Evidence: What a field can find evidence for marks the outer edge of what it can be about. Questions whose evidence is unobtainable in principle do not become open questions within a discipline; they fall outside its scope, and often get reassigned to another area of knowledge or to no one.

Methods and tools

Certainty: Methods do not only produce results, they produce a warranted level of confidence in them. Confidence intervals, error bars, source triangulation and the standards of proof are all machinery for saying how certain a knower is entitled to be.

Evidence: Every discipline's methods are, in effect, a definition of what it will accept as evidence. A controlled trial, a dated document and a proof step are all evidence, but the practices that produce them are not interchangeable, and a method that cannot generate a certain kind of evidence will systematically miss the claims that would have needed it.