4. After defining the name of substance, and making a clear and distinct idea correspond to it, it is still necessary to show how far the idea represents objects really existing, or, whether it belongs to the class of ideas expressing only the relation of different ideas, without our having any means of ascertaining whether this relation is found in the positive world or not; that is to say, whether the idea of substance is only the work of our understanding, a mere result of the combination of certain ideas, or is furnished us by experience itself. I shall try not to fall into any of these faults; I know not, however, whether I can escape them. For this purpose, I shall first analyze the word, with respect to its etymological sense, and then examine the various meanings which have been given to it. The analysis of words is very useful for the analysis of ideas: words often contain a great deal of truth, which we lose by not attending to their common meaning. 5. The word substance, _substantia_, implies something which is under, _substat_, which is the subject on which other things are placed; just as its correlative, accident or modification, expresses something which happens to the subject, _accidit_; something which modifies it, which is in it, as a mode of being, _modus_. 6. By substance we seem to understand something constant in the midst of variation, something which, although it is in various ways successively, according to the variety of modifications which affect it, remains constant and identical under different transformations. When we say that the substance has received any new modification, although we understand by this that the substance is, in a new mode, we do not mean that it is different in itself, that it has lost its internal primitive being, and taken a new being; but we only consider this change as external, and as leaving untouched a certain base, which is what we call substance. If it were not so, if we did not conceive something constant and identical under modifications, we could not distinguish substance from its modifications. The modification passes from not-being to being, and from being to not-being; now it is, and now it resigns its post to another and very different modification. But the substance is the same under different modifications; it does not pass from not-being to being with the succession of its modifications. From the moment that we attribute to substance the instability which belongs to its modifications, it ceases to be distinguishable from them.