(1) If both Kantian critique of metaphysics and James’s limitation of metaphysics are true, then metaphysics can offer no proof for God’s existence and no arguments for God’s existence are at all possible.
(2) If metaphysics can offer no proof for God’s existence and no arguments for God’s existence are at all possible, then James’s Will to Believe argument offers the only plausible reason to explain why religious beliefs are rational, but not conclusive.
(3) If both the Kantian critique of metaphysics and James’s limitation of metaphysics are true and no arguments for God’s existence are at all possible, then James’s Will to Believe argument offers the only plausible reason to explain why religious beliefs are rational, but not conclusive.
You could posit the consequent of (2) with other fideistic options, but in the end that’s the only real plausible argument to make about God’s existence since metaphysics is very limited–if possible at all. If we opt for Kant, then there is no speculative metaphysics. However, a Jamesian might defend further speculation about beliefs from the initial leap of faith into both aspects of nature and the divine. In this argument, faith is its own form of justification given that there are no other types of justification for metaphysical beliefs (apart from their conceivable effect on our experience). This also embraces the fact that scientific beliefs rest on pragmatic assumptions but that those assumptions can never be taken to be metaphysically conclusive.
However, let’s think from the other side. This restriction would also hold in some interesting ways from the naturalistic side of things. A fideism about naturalism? I imagine that we might posit the regular likelihood of future congruent beliefs cohering about our interactions and beliefs we discover about the natural world. Let me explain since I think this holds for both sides of how religious and metaphysical beliefs might work given that neither religion nor metaphysics can be definitively proven.
Naturalism, like James’s WtB argument, makes it such that there are clear cases of belief (namely options and genuine options when put together) that do not map onto reality independently of how we experience it. One could call this the pragmatic restriction of belief formation since again, all beliefs are really dispositions to respond habitually because of how belief (B) coheres and facilitates the set of future experiences (F1- Fn). So there are two cases that cut along naturalism and religion:
1. Epistemic Agent E accepts naturalistic beliefs (B) such that habits form to anticipate and explain (F1-Fn)
In case of 1., E represents the world as if one can conceive of nature independently of one’s experience and that helps conceive of likely consequences of action. Moreover, these consequential benefits might not come back for some time because it might not be that clear how naturalistic beliefs foster future consequences to our practices. I would readily admit, however, that naturalistic beliefs probably return with greater occasion than speculation merely because scientific beliefs often engender technological innovation.
2. Epistemic Agent E accepts speculative metaphysics B such that habits form to anticipate and explain (F’ to F’n).
In the case of 2., E’s speculation may be either concrete or removed from afield than naturalistic beliefs though it’s very possible that the set of F’ to F’n and F1-Fn may concern the same objects of concrete experience. Speculation serves to connect disparate threads and gaps of our more mundane knowledge, and a great deal of speculation serves to connect various pieces of the natural world together with the unseen order. In the same way, someone might have faith in science, a type of scientism that combines naturalism and belief in science such that they act as if science is their religion.
Habits embody the practices of the relationality between beliefs and action. This is the heart of the proposal such that 2. means that we can accept arguments for God’s existence on pragmatic grounds, and what that might mean. However, these same arguments can only be regarded as 2, but never independent truths that map onto a world without practice.