(2) These relationships are of a nature which keeps them from becoming the object of attention during dormal language use.
(3) No one of these relationships ordinarily gets repeated several times in a row in normal conversation, while the consecutive reiteration of such a relationship is essential to the successful completion of a drill.
For these reasons, it might be well to replace 'stimulus' (or 'cue') and 'response' as terms for the two halves of a line in a drill, calling them instead 'the first and second terms of the relationship' that the drill is about.
How and why drills work is a much-discussed question, which we considered in Chapter 1 (p. 19). Some authorities seem to believe that constant reiteration of samples of the desired effect of a neurological potential will produce that potential in the minds of their students. It is quite possible that students' minds do work this way, if only in self-defense. It may also be the case, however, that drills are valuable first for exploring and elucidating the relationships that they exemplify, and second in establishing a short-term memory of the relationship, which is then lengthened (Carroll in valdman 1966, p.99) by repeated real or realistic application (Chapter 2, pp. 29-31).
The two principal kinds of manipulative drill are substitution, which deals primarily with 'enate' relationships (Chapter 1, p. 12), and transformation, which deals with 'agnate' relationships. The purpose of a
394
