It seems certain to me now that students must learn Latin grammar inductively, within the context of reading or speaking. Grammar occurs always within the flawed and fluid process of articulation, and only there does it make sense. In math abstraction is the key to understanding. y=2x+5 is somehow more real than the line you graph by it. But to learn grammar is not to uncover the unifying system beneath a concrete superstructure, it is rather to make observations about an indefinite and mutable phenomenon. Humans spoke thus, and usually in such patterns—who knows why? Only our imagination can utter the formula. Why would I have spoken thus? How does my partaking in the eternal act of speech assent to the peculiar speech of the ancients? The same goes for modern language, of course, since we are almost always merely partakers and not innovators. The ever-changing nature of our race changed language, and so what we observe as we walk in the forests of articulation tells us about ourselves individually and communally. No morphological forewarning spares me the strangeness of that self-discovery. Anyhow, the point is that, in language, abstraction puts a veil over the eyes rather than enlightening them. The student doesn’t see what is simply there in the words. Those of us who have been here before should guide through the woods, pointing out the features of the trees and noting the patterns—but we cannot guard against encounters with the inexplicable and the bizarre. These are the groves and hollows of speech, as they are of history.