Gordon Clark begins chapter 8 of his "The Lord God of Truth" by affirming the necessity of universals for Christian theology. He reiterates his point that the history of physics has proved over and over again that there is no such thing as a universal law of physics. The problem with empiricism, however, for Clark, is that it cannot produce universal propositions. The reason for this is that experience itself is never universal. It is particular. No matter how many times we observe something, we can never infer a universal law from these observations, because the regularity we observe in these observations can always be broken by future deviations. This is called the problem of induction, and Clark notes that it makes a purely empirical epistemology impossible, because empiricism can never be used to make universal statements. Furthermore, one cannot use induction to justify induction itself.
In order to prove his point about the inability of empiricism to produce universal knowledge, he quotes Herbert Feigl:
Probably the most decisive division among philosophical attitudes is the one between the worldly and the other-worldly types of thought. Profound differences in personality and temperament express themselves in the ever-changing forms these two kinds of outlook assume. Very likely there is here an irreconcilable divergence. It goes deeper than disagreement in doctrine: at bottom it is a difference in basic aim and interest. Countless frustrated discussions and controversies since antiquity testify that logical argument and empirical evidences are unable to resolve the econflict. In the last analysis this is so because t he very issue of jurisdictive power of the appeal to logic and experience (and with it the question of just what empirical evidence can establish) is at stake(Herbert Feigl)
Clark approves of Feigl's statement here, and notes that such a sentiment is quite similar to that of Van Til:
Dr. Cornelius Van Til of Westminster Seminary has annoyed the empirical apologete by insisting that there is no common ground shared by believers and unbelievers - that is if both are consistent with their principles. The empirical aim is to discover some point of agreement which they can use in convincing any man of the truth of Christianity. Dr. Van Til denies that there is such an agreement. Well, there is an agreement of sorts: Van Til and Feigl agree that there is no agreement, no proposition held in common from which a Christian doctrine, or anything else, could be deduced"(Clark, p. 37).
What we have here is a basic incommensurability of worldviews. The Christian worldview and the unbelieving worldview, at bottom, have totally different standards of what constitutes evidence, rationality, consistency, intelligibility, etc. Clark points out that Feigl's choice to rely upon induction is entirely arbitrary and unjustifiable:
His choice of induction, as a choice, s hows that any system must have a starting point. If a system has no starting point, it cannot start, nicht? But a starting point cannot have been deduced or based on something prior to the start, for nothing is prior to the start, n'est-ce pas? Every system, th erefore, every attempted syste, must have an original undeduced axiom. Our dear friend Aristotle noted this, for he argued that if all propositions had to be deduced, they would regress to infinity, with the result that nothing could be deduced(Clark, p. 40)
Clark, Gordon. "Lord God of Truth." The Trinity Foundation, 1986. Hobbes, New Mexico.