in the true sense—that is, what is originated—then the invalidity of their argument on this supposition has already been shown.

But if by ‘contingent’ he means those necessary things that have a cause—as Ibn Sīnā and his followers maintain, saying that the celestial spheres are necessary and necessarily existent, that their nonexistence is impossible throughout the eternal past and future, yet also calling them contingent, meaning that they are effects—then, Ibn Rushd said, a necessary being cannot be established on this supposition, just as they cannot establish one by their method on the first supposition.

This is because the premise of a proof must be known before its conclusion: what is unknown is inferred from what is known, and what is obscure from what is evident.

Their argument is therefore unsound in two respects.

First, it has not yet been shown that an infinite regress among these contingent beings is impossible in the way this has been shown for truly contingent beings, namely, originated things. Their proof for a necessary being depends on disproving an infinite regress. Yet an infinite regress can be disproved only in the case of originated things, not in the case of necessary things that cannot be nonexistent, if an infinite regress of them is supposed.

Second, he said:

“Nor has it yet been shown that there is a necessary being requiring a cause, such that positing this—that is, supposing this—would entail ultimately arriving at a necessary being without a cause—until it is shown that the totality of necessary beings, consisting of cause and effect, is like the totality of contingent beings.”

What he means is that the premise of a proof must be known before its conclusion. Thus, if they establish a necessary being by means of what they have deemed contingent, even though it is