The principle of the multiple line of approach suggests a third and better procedure. Instead of accepting the one and rejecting the other, instead of adopting the middle course which frequently militates against the success of either extreme, this principle says, "Adopt them both concurrently, but not in one and the same operation. At times read intensively; at others read extensively. At appropriate moments, and for specific purposes, make the fullest use of all sorts of translation work; at other moments, and for other specific purposes, banish translation entirely. At times, more especially during the early stages, let there be an abundance of drill-work; later, but not before the student is perfectly ripe for it, let us introduce free work; and then let the two types alternate. At certain moments, more especially during the early stages, let the memorizing of sentences be carried out on a most extensive scale; at other moments, as a distinctly separate operation, let us cause the student to perform exercises in constructing correct sentences himself."
We have had occasion to note that this principle suggests the inadvisability of killing too many birds with one stone. The principle goes farther and adds to the figure of speech just quoted the two following corollaries, viz.: "Find the right stone to kill the right bird," and "It is often advisable to kill one bird with more than one stone." There are many different ways of teaching a difficult sound, there are many different ways of teaching a difficult point in grammar, a curious form of construction, or of causing the student to discriminate between two things which ought not to be confused. In these and in all similar cases, there is no reason why several methods should not be used concurrently; they need not be strictly co-ordinated. The cumulative effect of approaching the difficulty from different and independent angles will certainly secure the desired result. Superficial and rapid work on most points plus intensive and thorough work on certain essential specific points will generally ensure a well-assimilated whole. Either of these methods will tend to correct any disadvantages attached to the other and will be complementary to the other. The high degree of accuracy which results from intensive work will tend to spread by contagion to that portion of the work which must necessarily be done in a more summary fashion.