Abstract

Testing the validity of knowledge requires formal expression of that knowledge. Formality of an expression is defined as the invariance, under changes of context, of the expression's meaning, i.e. the distinction which the expression represents. This encompasses both mathematical formalism and operational determination. The main advantages of formal expression are storability, universal communicability, and testability. They provide a selective edge in the Darwinian competition between ideas. However, formality can never be complete, as the context cannot be eliminated. Primitive terms, observation set-ups, and background conditions are inescapable parts of formal or operational definitions, that all refer to a context beyond the formal system. Heisenberg's Uncertainty Principle and Gödel's Theorem provide special cases of this more universal limitation principle. Context-dependent expressions, on the other hand, have the benefit of being more flexible, intuitive and direct, and putting less strain on memory. It is concluded that formality is not an absolute property, but a context-dependent one: different people will apply different amounts of formality in different situations or for different purposes. Some recent computational and empirical studies of formality and contexts illustrate the emerging scientific investigation of this dependence.