Whence Aristotle's famous definition of nature as "the principle or cause of motion and rest in that thing in which it exists primarily and in virtue of that thing, and not accidentally."
Which is, by itself, a bit cryptic. He explains "in that thing in which it exists" clause like this: the principle can't be outside as in the case of a house. The house itself does not arrange its own bricks, but the builder does, and the builder is not part of the house.
As for the "and not accidentally" clause, I think his explanation (again, of what he does not mean) goes like this: Let the principle be "having the ability to heal". And let the thing be the patient, that is, some sick guy. Now the healing arts are outside the patient--unless of course the sick guy also happens to be a doctor. In case the patient has the principle of healing inside himself, but only accidentally insofar as he is a patient.
Personally, I think this is a lousy example given his above examples, because now we're left trying to figure out what doctors have in common with, like, fire. Plus, the doctor is a person, which is a kind of animal, and so already has a nature. So does he now have two natures, an animal nature and a doctor nature? Or does his doctor nature clobber his animal nature? And isn't there also a human nature somewhere here along the rest?
Aristotle continues, specifying where in his categorial scheme we might find such natures: All things with a nature are substances, and nature always exists in a subject. (The substance-subject connection arising from the fact that you cannot predicate a substance of something else.) Counterexample: locomotion does not have a nature. An upwards motion is not a nature itself, but it can pertain to fire, which does tend upwards. (Now if you've been programmed by your educators to find that notion risibly medieval, just swap in boiling and water: boiling is not a nature, but boiling pertains to the nature of water, which does boil.)
Now Aristotle feels that the existence of natures is simply evident and that it would be folly to try to argue their existence, as this would inevitably involve using the less evident in order to prove the more evident. Sure, blind people can reason about colors, but that does not bring them any closer to knowing what color is. (Aristotle's black-and-white-Mary moment?)
At this point, Aristotle starts to line up the notion of substance with that of matter and form. He takes on his opponent, Antiphon, who apparently thinks that nature is the same as the matter (as opposed to form), or the stuff that makes up your x. If you plant a bed, Antiphon argues, and something sprouts up, it will be a tree--that is, wood--and not a bed, and this shows that what a bed really is is wood in bed-wise arrangement. But why stop there, Aristotle asks. What about the wood itself, doesn't it further reduce to an element such as earth, air or fire? And doesn't this, in turn, lead us straight back to the doctrine the "physicists" (Thales, Anaximines, Heracleitus)? That what there really is is just such elements (i.e., matter), which are modified by this or that disposition (i.e., form)?
No, Aristotle says, it's the other way around. The form, not the matter, of a thing is its nature, and we can argue this by reference to potentiality. A potential bed is not a bed, just as potential flesh is not yet flesh. Only when it takes on the form in question does it qualify as a thing of that type. (I think the missing argument here is that the potential bed is also a potential boat, if potentiality is nature, then something can have more than one nature at a given time.)
What about the matter-form composite, say man, who is made up of both form and matter? It is not the case that man is a nature, Aristotle asserts, but that he has a nature. He goes on to argue this by some strange argument from bed reproduction vs. human reproduction, in order to show that it is the form towards which a thing progresses that defines its nature.
There are at least two boundary conditions that stand in the way. The first is the status of this smallest possible thing, as given in Plank units or whatever: if it's extended then it has a center and it has extremities, and these are parts. So eventually you have to face up to the question of how your smallest possible thing exhibits unity, lest you go on forever. And this means to have to allow some immaterial principle to hold sway.