Idea Comparison by Purpose
also Superiority for a specific reason · Evaluating ideas by goals
One idea is better than another only when it succeeds at a specific purpose the other fails at, never because it has a higher overall rating.
Idea comparison by purpose is CF’s answer to “how do you choose between ideas?” without scoring them. The conventional approach evaluates an idea’s overall goodness — summing strong and weak considerations into a verdict that one option is “better on the whole.” CF rejects this. When one idea genuinely beats another, the difference is always specific: the winner succeeds at some purpose the loser fails at, in a given context.
This follows from CF’s binary epistemology. Because errors are reasons an idea fails at a goal, a comparison reduces to whether each rival is refuted or non-refuted for the purpose at hand. The same idea can be non-refuted for one goal and refuted for another (a typo is fine in a text, fatal in a published paper). So “better” is meaningless until you name the goal; superiority is a relation between an idea, a purpose, and a context — not a quality the idea carries around.
To compare, you make the goal precise enough that some outcomes count as success and others as failure, then find a decisive consideration: a reason one idea fails at that goal while the other does not. If you cannot state such a reason, the ideas are tied for that purpose and you may pick either — even by coin flip.
This opposes weighted-factor analysis, argument strength, and credences, which all rank ideas by accumulated degree of goodness. CF’s complaint is that a high aggregate score hides the one error that causes failure: fifty things going right plus one decisive flaw still equals failure. Comparing by purpose forces that decisive difference into the open instead of averaging it away.