Ideas (as Units of Thought)

CF treats an idea as the smallest unit of coherent thought, while stressing that this label is approximate. Ideas nest: planning a program is one idea built from smaller ideas (its steps), which themselves contain still smaller ones. There is no sharp boundary where one idea ends and the next begins, and no good method exists for counting ideas or fixing their size. So “idea” names a flexible unit picked out at whatever granularity a purpose requires, not a fundamental atom.

Two moves give the view bite. First, ideas combine: any group can be integrated into a larger whole (see conceptual units), though most combinations are junk and which ones are useful is contextual. Granularity therefore tracks goals, not metaphysics. Second, ideas decompose, which is what makes targeted criticism possible. Because a thesis breaks into component sub-ideas, a decisive criticism can land on the specific sub-idea that is wrong rather than vaguely rejecting the whole. This grounds CF’s practice of putting discrete claims and objections as separate nodes in idea trees and debate trees, so each unit can be judged refuted or non-refuted on its own.

Drawing on Deutsch, CF also describes ideas as quasi-autonomous mental systems: they run somewhat independently inside a mind and produce stable outputs, yet an idea is far less than a whole person. This unit-of-thought framing connects to the broader hierarchy of ideas and underwrites CF’s claim that evaluation should target the right-sized unit.