Okay, this certainly helps. Thanks.
Given the third condition, I think it's illegitimate to combine the probability of the premises to determine the probability of the conclusion.
Well, I don't think that the probability/plausibility of the premises determine the probability/plausibility of the conclusion. I only think that the probability/plausibility of the premises establish a lower bound for the probability/plausibility of the conclusion. I think Craig agrees with me here.
I don't know if the rules of inference map 1:1 with such a probabilistic approach. Particularly, I don't know if, in a deductive argument, the premises are meant to be additive.
Consider the Kalam argument:
1) Everything that begins to exist has a cause. (Premise)
2) The universe began to exist. (Premise)
3) Therefore, the universe has a cause.
Now consider the Kalam* argument:
4) Everything that begins to exist has a cause, and the universe began to exist. (Premise)
5) Everything that begins to exist has a cause. (From 4)
6) The universe began to exist. (From 4)
7) Therefore, the universe has a cause.
The Kalam* argument is very similar to the Kalam, but the Kalam* argument has only one premise. I think, and hope that you'll agree, that the Kalam and the Kalam* are equally good arguments, should be equally persuasive, and should equally affect how plausible one finds the conclusion. And since the only premise in the Kalam* is just the conjunction of the two premises of the Kalam, we can see that the plausibility of the conjunction of the two premises of the Kalam is what matters in assessing the plausibility of the conclusion.