Bacardi rum

In Language and Theology, Gordon Clark says: To perceive a thing, these “sensations” must be combined. Note that no one ever sees a dog or a tree. A dog is not just black; he is also soft, fuzzy, and perhaps has an odor. But before one perceives a dog, he must choose black, fuzzy, and odor, combine them, and only then has he the perception of his pet. Yet there is nothing in the single qualities that forces him to select these particular ones and discard the dozens of others he also has at the same time. Why does he not select the fuzzy, the sound B-flat, and the taste of Bacardi rum, all of which he senses at the same moment, and combine them into the perceived object? Is there anything in a person’s fifty or more sensations that compels the selection of these few rather than another few?

On the face of it, this generates a dilemma for Scripturalist epistemology. How can we grasp what Clark's statement even means unless we already know the difference between a a dog, the sound of B-flat, and the taste of Bacardi rum? Clark's examples appear to be self-refuting, because his illustration is premised on knowing something that his conclusion denies. Given his epistemology, how does he know, or come to know, what's a dog or what's Bacardi rum? Surely he didn't deduce Bacardi rum from a Biblical proposition. What proposition would that be?