3. I have said very little about causation. The fact is, that in Logic the cause necessarily fades away into the reason, that is, the explanation. If we follow Mill’s account, we see how this takes place. I will put the stages very briefly.
Cause
(a) We start, no doubt, by thinking of a cause as a real event in time, the priority of which is the condition of another event, the effect. Pull the trigger—cause—and the gun goes off—effect.
Complete conditions
(b) The moment we look closer at it, we see that this will not do, and we begin to say with Mill, that the cause is the antecedent which includes all the conditions of the effect. The plurality of alternative causes breaks down, through the conditions defining the effect. Pull the trigger?—yes, but the cartridge must be in its place, the striker must be straight, the cap must be in order, the powder must be dry and chemically fit, and so on, and so on, till it becomes pretty clear that the cause is a system of circumstances which include the effect.
Law
(c) But then our troubles are not ended. Only the essential and invariable conditions enter into the cause, if the {165} cause is invariable. This begins to cut away the particular circumstances of the case. You need not use the trigger, nor even the cap; you may ignite powder in many ways. You may have many kinds of explosives. All that is essential is to have an explosion of a certain force and not too great rapidity. Then you will get this paradox. What is merely essential to the effect, is always something less than any combination of real “things” which will produce the effect, because every real thing has many properties irrelevant to this particular effect. So, if the cause means something real, as a material object is real, it cannot be invariable and essential. If it is not something real, and is essential, it fines down into a reason or law—the antecedent in a hypothetical judgment.
Ground, or real system with known laws
(d) We can only escape this by identifying both cause and reason with the complete ground; that is, the nature of a system of reality within which the cause and effect both lie. But even then, though the ground is real, it is not antecedent in time. We see, indeed, that the conditions of an effect must be continuous through the effect. If the process were taken as cut in two at any point, its connection would be destroyed. If a cause and b effect were really detached events, what difference could it make if, instead of a, c preceded b?
Postulate of Knowledge