Within cognitive experience, traditional philosophy and common sense distinguish two polar epistemological components: those called data and the rational elements called concepts. The former are used as protocol (P) experiences and are taken as authoritative, as the last instances of appeal of any theory; the latter owe their genesis to reason rather than observation, we feel a rational responsibility concerning them and modify them with some freedom. They are constructed as counterparts to data in order to provide a larger measure of coherence and intelligibility than P-experiences alone contain. Whether concepts, originally constructed to satisfy inductive suggestions coming from observations, are ultimately accepted and retained as valid explanations of datal experience depends on the way in which they meet a set of complicated requirements. Primary among these requirements is empirical confirmation, the need for conclusions drawn from theory (which is itself a set of constructs joined by logical and mathematical relations) to be confirmed within a certain tolerance by observations. Other requirements, more vague in their formulation, demand simplicity, mathematical elegance, extensibility, causality in the theoretical transcription of data.