Meta-Framework： The Foundation, Sieve, Bottom Line, and Interface of the Cognitive Underlying Protocol— A Non-Ultimate, Dynamic, Compatible Framework for Evaluating Cognition

This paper proposes and constructs a cognitive framework whose goal is not to “explain the world” but to “evaluate all explanations.” The framework consists of four layers: **Foundation** (the presuppositions of cognition), **Sieve** (seven criteria and the “three-fold test” for evaluating theories), **Bottom Line** (the ethical requirements of cognitive activities: honesty, non-silence, non-packaging, and non-blame-shifting), and **Interface** (future-oriented openness: leaving room to be transcended). The core claims of this paper are: 1. Any system of thought that claims to be a “meta-theory” but fails to simultaneously satisfy the three-fold test—retrospection, present relevance, and future projection—does not qualify as “meta.” 2. Compatibility and generativity are more fundamental than correctness. 3. The first ethic of cognition is honesty, not success. 4. The ultimate value of a theory lies not in being unfalsifiable, but in allowing itself to be surpassed. The very presentation of this paper follows the principles of the framework it proposes—making no ultimate assertions, imposing no closed boundaries, and actively inviting scrutiny, revision, and transcendence.