Aristotle sees the peculiar feature of substance as its ability to remain through contraries, which makes substance uniquely responsible for motion and change. Motion, which is of itself sensible and a defining characteristic of the physical world, is therefore only possible because of something that is not of itself sensible.
Objection: a surface changes both in quality, position, and place, but a surface is not a substance. Surfaces, moreover, are per se sensible. Therefore we do not need to posit substances to explain the motion of the physical or sensible world.
Response: A physical surface is the limit or totality of a quantity, but things can change from one physical quantity to another, and so one and the same thing has more than one definite quantity. But no actual quantity has more than one definite quantity.
Notice that this account of substance captures it in two ways: on the one hand, we see it in its indetermination or potentiality, and so in its dependence. It is neither this quantity nor another. Considered in this way, the actual quantity can be seen as replacing the substance, and so far as we study merely changes in quantity, we can overlook substance without consequence. On the other hand, substance is not this sheer possibility of things but an actuality prior to the flux, and is therefore independent of it. To the extent that we forget about the reality of substance in this latter sense, change in quantity becomes either arbitrary or impossible, and so – even within the limits of a quantitative study – such forgetfulness will lead to either the belief that that the science is essentially subjective or that nothing it studies really changes. Physical science has arguably reached both extremes in the Copenhagen interpretation (which allows for indeterminism only by making the science essentially dependent on the willed act of measurement) and in Einstein’s block universe.