He said:
“Rather, the answer is that a relational attribute belonging to something in relation to something else depends only on the existence of its subject. Temporal precedence, for example, is a relational attribute belonging to something in relation to what comes after it, even if much time separates them, although the earlier and the later cannot exist together.”

I say: Al-Urmawī’s statement, “One might say that the infinite regress here is a regress of effects, because being a cause is a relational attribute whose relation depends upon the cause and the effect. It is therefore posterior to the effect and requires another instance of being a cause after the effect, so that every instance of being a cause is followed by another,” may be challenged as follows:

This contradicts his subsequent statement: “Rather, the answer is that a relational attribute belonging to something in relation to something else depends only on the existence of its subject.” For if this statement is correct, it would not follow from the cause; and if the attribute belonging to something is not dependent, then the realization of being a cause alone suffices.

He may, however, answer this by saying: “My intention is to hold another person to the implications of his position: if he says that being a cause depends upon the cause and the effect, then this is a regress of effects, not of causes.” This is a valid argument on his own premises.

But he may be told: k It would have completed this argument to say:
“If, according to you, being a cause is a relational attribute whose conception depends upon the cause and the effect, then it necessarily entails the existence of the effect,