Prototype theory [...] permits us to state the general principles that provide the semantic basis of syntactic categories. In a classical theory of categories, one would be forced to say that there is no semantic basis at all. The reason is that classical categories have a homogeneous structure-there are no prototypes-and everything that is not completely predictable must be arbitrary. Since syntactic categorization is not completely predictable from semantic categorization, a classical theory of categories would be forced to claim, incorrectly, that it is completely arbitrary.
- Lakoff (1987), Pag. 280