If the signification of the names of mixed modes be uncertain, because there be no real standards existing in nature to which those ideas are referred, and by which they may be adjusted, the names of SUBSTANCES are of a doubtful signification, for a contrary reason, viz. because the ideas they stand for are supposed conformable to the reality of things, and are referred to as standards made by Nature. In our ideas of substances we have not the liberty, as in mixed modes, to frame what combinations we think fit, to be the characteristical notes to rank and denominate things by. In these we must follow Nature, suit our complex ideas to real existences, and regulate the signification of their names by the things themselves, if we will have our names to be signs of them, and stand for them. Here, it is true, we have patterns to follow; but patterns that will make the signification of their names very uncertain: for names must be of a very unsteady and various meaning, if the ideas they stand for be referred to standards without us, that either cannot be known at all, or can be known but imperfectly and uncertainly. 12. Names of Substances referred, I. To real Essences that cannot be known. The names of substances have, as has been shown, a double reference in their ordinary use. First, Sometimes they are made to stand for, and so their signification is supposed to agree to, THE REAL CONSTITUTION OF THINGS, from which all their properties flow, and in which they all centre. But this real constitution, or (as it is apt to be called) essence, being utterly unknown to us, any sound that is put to stand for it must be very uncertain in its application; and it will be impossible to know what things are or ought to be called a HORSE, or ANTIMONY, when those words are put for real essences that we have no ideas of at all. And therefore in this supposition, the names of substances being referred to standards that cannot be known, their significations can never be adjusted and established by those standards. 13. Secondly, To co-existing Qualities, which are known but imperfectly.