§ 7. Considerations of this kind compel us to forego the attempt to find the substance or being of a thing in the mere sequence of its different states considered as an aggregate. To make Phenomenalism workable, we are forced to say at least that the thing or substance to which the various attributes are assigned is the “_law_ of its states,” or again is “the _mode of relation_ of its various qualities.” Such a definition has obviously a great advantage over either of the two we have just rejected. It is superior to the conception of the thing as an unknown substratum of qualities, since it explicitly excludes the absurd notion of a world of things which first _are_, without being in any determinate way, and then subsequently set up determinate ways of existing among themselves. For a law, while not the same thing as the mere collection of occurrences in which it is realised, has no existence of its own apart from the series of occurrences which conform to it. Again, every law is a statement of possibilities, a formula describing the lines which the course of events will follow _if_ certain conditions are operative; no law is a mere register of actually observed sequences.[85] Hence, in defining the thing as the “law of its states,” we avoid the difficulty dealt with in the last paragraph, that the collection of the thing’s states never actually exists as a “given” collection. Thus for ordinary practical purposes the definition is probably a satisfactory one. Yet it should be evident that in calling the thing the “law” of its states, we merely repeat the metaphysical problem of the unity of substance without offering any solution of it. For, not to dwell on the minor difficulty that we might find it impossible to formulate a _single_ law connecting _all_ the ways in which one thing reacts upon others, and thus ought more properly to speak in the plural of the laws of the states, we are now left with two distinct elements or aspects of the being of the thing, namely, the successive states and the law of their succession, and how these two aspects are united the theory fails to explain. We have the variety and multiplicity on the one hand in the states or qualities of the thing, its unity on the other in the form of the law connecting these states, but how the variety belongs to or is possessed by the unity we know no better than before. Thus the old problem of substance returns upon us; the many qualities must somehow be the qualities of a single thing, but precisely how are we to conceive this union of the one and the many?[86]