After my little jest about the Romans hunting up first the subject and then the predicate as Cicero talked to them, or first the predicate and then the subject, whichever one thinks the Roman method may have been, I assure them that "what we have to do is to learn to understand a Roman sentence precisely as a Roman understood it as he heard it or read it, say in an oration, for example. Now the Roman heard, or read, first the first word, then the second, then the third, and so on, through sentence after sentence, to the end of the oration, with no turning back, with no hunting around. And in doing this he was so guided all the time, by indications of one kind or another in some way strown through each sentence, that, when the last word of that sentence had been spoken or read, the whole of the meaning had reached his mind. The process of detecting these indications of meaning was to him a wholly unconscious one. We moderns, however, of course cannot begin so far along. What we are to reach finally is precisely this unconsciousness of processes; but we shall be obliged, for the first few years, explicitly to study the indications, until we come to know them familiarly, one after another. We must for some time think out, at every point, as the sentence progresses (and that without ever allowing ourselves to look ahead), all those conveyings of meaning, be they choice of word, or choice of order, or choice of case, or choice of mode, or choice of tense, or whatsoever else which at that point sufficed for the Roman mind. And when these indications — which after all are not so many in number — have come to be so familiar to us that most of them are ready to flash before the mind without our deliberately summoning them, we shall be very near the point at which, in Latin graded to our growing powers, we shall interpret indications unconsciously. And the moment we do that, we shall be reading Latin by the Roman's own method."
