Aristotle’s second argument against an infinite regress of motions is the one that Thomas Aquinas relies upon in the First Way. It occurs in Physics VIII, chapter 5. In this argument, Aristotle makes a distinction between two kinds of causes: primary and intermediate. If we have a series of motions that are causally related, and the series has three or more members, then we can apply the distinction: the last motion is the ultimate effect, the first motion in the series is the primary cause, and any other members of the series are intermediate causes. Aristotle asserts that the intermediate causes are dependent for their causal efficacy on the primary cause. They are merely conduits through which the first mover causes the ultimate effect. Consequently, if we have an infinite regress of motions, all the causes would be merely intermediate, and there would be no primary cause upon which their causal efficacy could depend. Hence, none of the causes in the series would have any causal power at all, resulting in a contradiction. So, infinite regresses of this kind are impossible.
Aristotle and Thomas are sometimes accused of simply begging the question, assuming from the outset the impossibility of an infinite regress. However, that isn’t the case. As we’ve seen, both Aristotle and Thomas accept that infinite regresses are in fact possible. So, how are we to understand Aristotle’s argument? As I argued in an earlier post, we should take it that Aristotle accepts the possibility of an infinite regress of partial causes, but not an infinite regress of total causes. A partial cause is something that is a proper part of a total cause.
But if an infinite regress of partial causes is possible, wouldn’t that entail the possibility of an infinite regress of total causes? Couldn’t we just take the series of total causes of which each partial cause is a part, and thereby uncover a regress of total causes? No, because a partial cause can combine with an uncaused cause (the Prime Mover) to produce an effect. Suppose that P1 is a partial cause of P2, because P1 is part of T1, and T1 is a total cause of P2. And suppose that P2 is a partial cause of E, because P2 is part of a total cause T2 of E. Does it follow that T1-T2-E is a series of total causes? Not necessarily. Suppose that P2 combines with some uncaused cause U to produce E. So, T2 = (P2 + U). Since T2 includes an uncaused part, it is impossible for anything to cause T2 (since a cause of T2 would be a cause of all of its parts). So T1 is not a total cause of T2. Consequently, T1 is also not a total cause of E, and P1 is not a partial cause of E.
This case illustrates an important consequence of Aristotle’s systems: the relation of total cause is transitive, but the relation of partial cause is not. P1 is a partial cause of P2, P2 is a partial cause of E, but P1 is not even a partial cause of E. This is crucial in understanding Aristotle’s argument. Infinite regresses of total causes are impossible, because total causation is transitive. If such a regress were to exist, the ultimate effect would have an infinite number of distinct and separate causes. A series of total causes is possible only if all but the first one are merely intermediate causes– not causes in the strict sense. Consequently, an infinite regress, having no first member, is impossible.
Why is it impossible for a motion to have more than one total cause? In Aristotle’s system, there is an intimate connection between a thing’s total cause and its essence or definition. Aristotle is explicit about this in Posterior Analytics, Book 2, and Metaphysics Zeta 17. A motion is the kind of motion it is because of its cause. A motion with more than one total cause would have more than one essence, which is impossible. A motion with infinitely many total causes would be completely unintelligible, lacking any possible definition.
Next week I’ll cover Aristotle’s third and last argument against infinite regresses.
The Analytic Thomist 