It is not, however, enough for a principle of demonstration to be true; it must also be necessary. A necessary conclusion follows from necessary premises (NLP I. 19, 110–113). The kind of universality required for demonstration is not the same as that of a universal acquired through abstraction, which is said of many (ut dicatur de multis)—the kind described in the Perihermeneias and in Porphyry’s Isasoge—but the universal that must be said of all and always (de quodlibet et semper et primo).[2