and an effect, insofar as it is an effect, must have a cause. Thus, each of them must have a cause, and an endless succession of effects does not prevent each from needing a cause. If, then, there is no aggregate beyond these individual members, each of which is an effect in need of a cause, it follows that nothing exists except effects in need of a cause. Yet an effect that needs a cause necessarily cannot exist by itself.

On this assumption, nothing would exist by itself, and whatever does not exist by itself exists only through an agent that brings it into existence. If that agent does not exist by itself, it too belongs among those things that do not exist by themselves, and therefore would not exist. It follows that nothing would exist, yet existing things do exist.

This would entail affirming two contradictories: none of the existing things would exist, if it were assumed that none of them exists by itself, yet they all exist. Thus, there must be something that does not exist by itself.

The existing thing would therefore both exist by itself and not exist by itself, which is an affirmation of two contradictories.

[The third reply]

It may be said:

By the sum total of what the aggregate depends on, we meant the efficient cause, for the discussion concerns establishing an efficient cause for the aggregate of contingent things, not something more general than that.

As for his statement:

“If you mean the complete efficient cause, why do you say that this entails some of the parts being sufficient for the aggregate?”

The reply is:

We said this because the existence of the complete efficient cause necessarily entails the existence of the effect. By “cause,” we mean precisely the totality of that whose existence entails the existence of the effect. A contingent thing does not exist until the complete determining factor that necessarily entails its existence is present. Thus, if