Bergson’s account of perception differs from the account ordinarily given in that perception is not described as a relation which is supposed to hold between a subject and an object: for Bergson there is no “I,” distinct from what is perceived, standing to it in a relation of perception. For an object, to be perceived consists, not in being related to a perceiver, but in being combined in a new way with other objects. If an object is combined by synthesis with other objects then it is perceived and so becomes a fact. But there is no mind over and above the objects which perceives them by being related to them, or even by performing an act of synthesis upon them. To speak of “our” perceiving objects is a mere fiction: when objects are combined by synthesis they become perceptions, facts, and this is the same as saying that they are minds. For Bergson a mind _is_ nothing but a synthesis of objects. This explains what he means by saying that in direct knowledge the perceiver is the object perceived. Actually he thinks such notions as the perceiver and the object and the relation which unites them, or again matter and the act of synthesis which turns matter into fact, are nothing but abstractions: the only thing there really is is simply the fact itself. These abstractions, however, do somehow apply to the actual facts, and this brings us back to our problem as to how it is that the actual fact, which is in creative duration, lends itself to classification: how it is that general laws in terms of abstractions which can be repeated over and over again, can apply to the actual fact which does not contain repetitions? Facts lend themselves to explanation when they are perceived as familiar. In this perceived familiarity, which is the basis of all abstraction, and so of all description and explanation, past as well as present is involved, the present owing its familiarity to our memory of past facts. The obvious explanation of perceived familiarity, would be, of course, to say that it results from our perceiving similar qualities shared by past and present facts, or relations of similarity holding between them. But Bergson must find some other explanation than this since he denies that there can be repetition in actual facts directly known.