One can know a theorem is true if it is contained in the body of propositions validly deducible from an axiom which yields a self-attesting, consistent philosophical system in which the ground and means of knowledge are explicated. Hence, while axioms by definition cannot be proven, there is nevertheless a mutual dependency inherent in the relationship between an axiom and its respective theorems. “By the systems they produce, axioms must be judged.” As a theorem can be discredited if it does not follow from a purported axiom, an axiom can be falsified if it bears contradictories.
Gordon Clark, Clark and His Critics, 53.
I’m saying an axiom must be self-attesting, not that it is. Empiricism isn’t self-attesting because nothing one discovers empirically could ex hypothesi attest to the idea that empirical procedures alone are a means to knowledge: the claim is arbitrary. For this reason Clark is concerned in the early part of God’s Hammer to show that Scripture claims to be God’s word. There must be a mutual dependency in an epistemic system: an axiom which prescribes knowledge as coming through certain means yet cannot by those means prescribe the axiom itself is self-defeating. This is illogical and, hence, cannot be a system of knowledge.
Of course the question may be asked: how do I know the criteria of knowledge? Ultimately, by my own axiom, Scripturalism. Proximately, by necessary inference which, since such is accounted for in Scripture, refutes your charge of rationalism. I don’t begin with logic, I begin with Scripture which, since logic is accounted for in it, allows me to use logic to come to these conclusions.
Sunday, May 6, 2012
Epistemic Justification and Axioms
Among some interesting things I have been discussing with Drake Shelton, a fellow Scripturalist, one of the main subjects is the nature [and role] of axioms in [the justification of] an epistemic system (link). Drake has said that he is skeptical with regards to the truth of his axiom, which I argue undercuts his whole system of "knowledge." Rather, I have proposed the following: