HKB
dog is 'accidental' to the nature of the matter; and the use of this matter, rather than of some other equally suitable, is accidental to the nature of the dog.
No account of material events can dispense with accidental relations altogether. We must at least recognize that there are accidental relations between particular things. Accident in the sense of brute fact had to be acknowledged even by the tidiest and most dogmatic atomism of the last century. That atomism must allow it to be accidental, in this sense, that the space surrounding any given atom was occupied by other atoms in a given manner. It belonged neither to the nature of space to be occupied by just those atoms in just those places, nor to the nature of the atoms to be [17] distributed just like that over space; and so in a certain sense the environment of any atom was an accidental environment. That is, the particular arrangement of the environment was accidental. The nature of the environment was not accidental at all. It was proper to the nature of the atom to be in interaction with other atoms over a spatial field, and it never encountered in the fellow-denizens of space any other nature but its own. It was not subject to the accident of meeting strange natures, nor of becoming suddenly subject to strange or unequal laws of interaction. All interactions, being with its own kind, were reciprocal and obedient to a single set of calculable laws.
But the medieval philosophy had asserted accidental relations between distinct sorts of natures, the form of living dog and the form of dead matter, for example. No one could know a priori what effect an accidental relation would produce, and all accidental relations between different pairs of natures were different: at the most there was analogy between them. Every different nature had to be separately observed, and when you had observed them all, you could still simply write an inventory of them, you could not hope to rationalize your body of knowledge. Let us narrow the field and consider what this doctrine allows us to know about the wood of a certain kind of tree. We shall begin by observing the impressions it makes on our several senses, and we shall attribute to it a substantial form such as naturally to give rise to these impressions, without, perhaps, being so rash as to claim a knowledge of what this substantial form is. Still we do not know what its capacities of physical action and passion may be. We shall find them out by observing it in relation to different 'natures'. It turns out to be combustible by fire, resistant to water, tractable to the carpenter's tools, intractable to his digestive organs, harmless to ostriches, nourishing to wood-beetles. Each of these capacities of the wood is distinct; we cannot relate them intelligibly to one another, nor deduce them from the assumed fundamental 'woodiness'.