If we follow the discussion as it centres about the terms "nominal" and "real," it stands as follows: Leibniz objects to the use of the term "essence" in this connection, but is willing to accept that of "definition;" for, as he says, a substance can have but one essence, while there may be several definitions, which, however, all express the same essence. The essence is the _possibility_ of that which is under consideration; the definition is the statement of that which is supposed to be possible. The "nominal" definition, however, while it implies this possibility, does not expressly affirm it,--that is to say, it may always be doubted whether the nominal definition has any possibility (or reality) corresponding to it until experience comes to our aid and makes us know it _a posteriori_. A "real" definition, on the other hand, makes us know _a priori_ the reality of the thing defined by showing us the mode of its production, "by exhibiting its cause or generation." Even our knowledge of facts of experience cannot be said, therefore, to be arbitrary, for we do not combine ideas just as we please, but "our combinations may be justified by reason which shows them to be possible, or by experience which shows them to be actual, and consequently also possible." To take Locke's example about gold, "the essence of gold is that which constitutes it and gives it its sensible qualities, and these qualities, so far as they enable us to recognize it, constitute its nominal essence, while a real and causal definition would enable us to explain the contexture or internal disposition. The nominal definition, however, is also real in one sense,--not in itself, indeed, since it does not enable us to know _a priori_ the possibility or production of the body, but empirically real." It is evident from these quotations that what Leibniz understands by "possibility" is the condition or cause of a given fact; and that, while Locke distinguishes between particular, accidental and demonstrative, general knowledge as two opposed kinds, concerned with two distinct and mutually exclusive spheres, with Leibniz they are distinctions in the aspect of the same sphere of fact. In reality there is no combination of qualities accidental, as Locke thought that by far the greater part were; in every empirical fact there is a cause or condition involved that is invariable, and that constitutes the reason of the fact. The "accidental" is only in the relation of our ideas to objects, not in the objects themselves. There may be accidental mental associations; there are no accidental relations. In empirical, or _a posteriori_, knowledge, so-called, the reason is there, but is not known. _A priori_ knowledge, the real definition, discovers and explicitly states this reason. Contingent knowledge is therefore potentially rational; demonstrative knowledge is the actual development of the reasons implicitly contained in experience.