It may be said in reply:
This argument is valid as a counterargument that holds them to the consequences of their own position. It exposes the inconsistency of those who affirm that He knows some things but not others, such as Ibn Sīnā and his like. What they affirm necessarily entails something comparable to what they sought to escape by their denial. If that objection is valid, it follows that He knows nothing, not even Himself; if it is invalid, it follows that He knows everything.
There is no doubt that this latter conclusion is correct. He is saying to them:
“If knowledge is an attribute of perfection and nobility, then clearly His knowledge of everything is more perfect than His knowledge of only some things. If it is not an attribute of perfection, then none of it should be affirmed of Him.”
He supported this by pointing out: “You have denied Him hearing, sight, and knowledge of changing things. If denying these constitutes an imperfection, then you have described Him as imperfect. If it does not, then denying other knowledge would not constitute an imperfection either.”
Ibn Sīnā distinguishes between His knowledge of unchanging things and His knowledge of changing things, to avoid the implication that He changes as His knowledge changes.
Abū Hāmid subsequently explained that there is nothing objectionable in His knowledge of changing things; rather, he is bound to affirm His knowledge of them. Thus, what must be affirmed is knowledge without restriction, rather than its denial. Whatever necessarily follows from affirming knowledge is not objectionable at all. At most, the issue comes back to the denial of attributes, acts, and choices. Their denial of knowledge rests on nothing beyond what arises from these two negations.
The position of those who deny the attributes and the realities that subsist in His essence and are subject to His will has already been shown to be unsound.
If His knowledge—glory be to Him—of everything necessarily entails affirming these two principles, then this is evidence of their validity, as is also indicated by