and patterns other than as discourse conventions (such as greetings and attention-getters) and in limited contexts (for example, commercial and sports domains) encourages very complex input that may be useless to the performer or that may place him in an awkward situation. Consider what often happens when the traveler attempts to use his phrase book to ask "Where is the opera?" The answer may be a barrage of incomprehensible input, for example, "You go two blocks until you see a big church, then turn left until you come to a traffic light...". Use of strategies such as employing yes/no questions instead of WH-questions may alleviate this problem. Still, responses to adult routines and patterns may be highly unpredictable and complex.

Conclusions

The available evidence indicates that routines and patterns are essentially and fundamentally different from creative language. Most studies support the correctness of position 2, that the creative construction process is independent of routines and patterns. It appears to be the case, however, that the use of the gestalt mode is possible and may be encouraged by certain environmental conditions. This alternative route may be the acquirer's reaction to demands for early production before sufficient competency has been built up. Routines and patterns may be very useful for establishing social relations and encouraging intake. They could conceivably serve as intake for the creative construction process. This intake, however, is probably insufficient for successful language acquisition. Adult teaching systems, such as the audio-lingual, seem to imply that positions 1 or 3 are correct. But these systems enjoy the limited success they do because they also provide at least some intake for the creative construction process. More successful teaching systems expose second language acquirers to input in which routines and patterns play a minor, though significant role.

99