the essential attribute and the necessary accidental attribute when both necessarily belong to the subject. Instead, they proposed three points of difference, yet all three are baseless, a fact conceded even by their own leading thinkers. They say, for example, that the essential attribute is affirmed of the subject without any intermediary, whereas the necessary accidental attribute is affirmed only through an intermediary. Their more astute writers explain that this “intermediary” is simply a proof, as Ibn Sīnā makes clear. Others—al-Rāzī among them—take the intermediary to be an attribute residing in the subject itself; these people failed to grasp what the former group meant and thus only deepened their error. What the first group intended by “intermediary” was a proof, just as by “middle term” they mean the element joined to the particle li- in statements such as “because it is….” Accordingly, in their usage a “necessary accidental” attribute is nothing more than one whose connection to the subject is known by means of a proof. This, however, does not correspond to any fixed reality in and of itself; it concerns only the knower and his awareness of the attributes. Some of them possess so complete a conception that they recognize the attribute’s necessary connection to the subject without recourse to any proof.