At Isharat V.2, the shaykh makes the following claim:
Pointer: every temporally originated thing (hadith) is, before its existence, possible of existence (mumkin al-wujud).
He then goes on to argue that this possibility must exist in something that serves as its substrate. We can schematize the conclusion of V.2 like so:
[PM]: For all F, if F comes to be, the possibility of F’s existence exists before F does and exists in a substrate.
In his Sharh al-Isharat (ed. H. Amoli, Bustan-e Kitab: 1391, vol. 3, 660-663), Tusi has a fine exposition of the argument for [PM]. His commentary on the entire fasl can be divided into three parts. In the first, he explicates the argument in the matn (text) of V.2; in the second, he elucidates his own independent treatment of the points raised in the matn; and in the third, he, in light of part 2, responds to Razi’s objections. Here, I want to look at the second part, where independently investigates the issues V.2 raises.
Tusi begins by noting that possibility is always relative to some existence. And this existence is either (1) per accidens (bi’l-araz) or (2) per se (bi’l-dhat). As an example of per accidens existence, he gives ‘a white body’, and as an example of per se existence, he gives ‘whiteness’. (The examples will be clarified in what follows below).
With (1), the possibility belongs to a thing, e.g, x, in two ways; either (1.1) “in relation to the existence of something else, [F], for x” or (1.2) “with respect to x becoming another thing, [F].” According to Tusi, the first way, i.e., (1.1), is exemplified in statements like:
(A) x is possibly F (e.g, the body is possibly white)
(B) F possibly exists for x (e.g., whiteness possibly exists for the body)
And the second way, i.e., (1.2), is exemplified is statements like:
(C) X possibly becomes F (e.g., water possibly becomes air)
Clearly, insofar as all these are cases wherein the possibility for F is spoken of in terms of something else, x, possibly being/becoming F, the possibility for F requires x as the substrate in which it inheres.
With regards to (2), the possibility belongs to a thing with respect to its own existence and not, like in (1), with respect to the existence of something else. The difference between (2) and (1) is the difference between the way possibility is ascribed in statements (A)-(C) and the way it is ascribed in a statement like:
(D) F is possible (e.g., whiteness is possible)
(E) F possibly exists (e.g., whiteness possibly exists)
The possibility of F to be, in other words, is not related to something else, e.g., x to be F. It is just said of F in relation to its own existence. In this sense, the possibility is essential to F, as opposed to accidentally occurring to it in virtue of something else that bears the possibility. Now, the thing, F, about whose per se possibility we’re inquiring, Tusi states, must be one of two kinds. He writes (ibid., 662):
[Either F] is (2.1) among that which exists in a subject, or in matter, or with matter […]. [Or] (2.2) that is not the case and it is rather self-subsisting, having no connection to any subject and matter.
The contrast here is between something that has some connection with a subject/matter (= (2.1)) and something that has no connection with a subject/matter (= (2.2)). If the thing is of type (2.1), then it follows, given premise [PM], that the thing’s possibility must have something that bears it, i.e., the substrate/matter in or with which it will come to exist. As such, Tusi states, (2.1) falls under (1) – things the possibility of whose existence is relative to something else (= their subject/mattter).
However, if F is of type (2.2), then, Tusi urges, “it’s not possible for [F] to come to be”. Why not? Tusi’s reasoning is as follows: otherwise, given [PM], F’s possibility would exist before F itself actually exists. And since, on assumption (2.2), F is self-subsistent, its possibility to come to be can’t be relative to something else, i.e., can’t have something else as its subject. This entails that F is a (self-subsisting) substance (since a subject is that which, when it exists, does not exist in a subject). But a substance is not relative to anything else; possibility (for coming to be) though, as was said, is always relative to some existence, i.e., either per accidens or per se. It follows, Tusi states, that the possibility for F to come to be is not the very reality (haqiqa) of F qua substance. And that entails that it must be something that is accidental to F (given that anything belonging to substance F not identical to F is accidental to F). But if so, this contradicts the assumption that possibility for existence per se is essential to a thing, since it isn’t had by the thing in relation to something else which serves as its subject.
In other words, where F = self-subsistent substance, and it’s true that:
 F possible exists (i.e., at some t)
then F’s possibility P is for existence per se. As such, P is essential to F. However, given , P will, per [PM], obtain prior to F’s actual existence. But F’s P is not identical to F; for P, qua possibility, is a relative, but no substance (and hence F included) is a relative. Therefore, P will be something accidental to F, not essential. That is, F will have P because of something else, e.g., G, which bears P. This contradicts . Therefore, if F= self-subsistent substance, and F exists, then F can’t possibly come to be. Tusi concludes (ibid., 663):
Since it has become clear that something like this can’t possibly be a temporally generated thing, then it, if it exists, would be of permanent existence (da’im al-wujud), and if it doesn’t exist, then it would be impossible of existence.