Audio-lingual proponents stressed that language is primarily speech. Therefore, initial and continued emphasis should be placed on listening and speaking. The ear was to be trained before the students were permitted to see the written representations of these sounds. Seeing the printed page prevented the students from developing native-like pronunciation due to the negative influence of the written word.
Listening training Tezza (1962) found that students who received listening training did not do better than those who read or practiced audio-lingual drills. A related topic is auditory discrimination training. Based on the results of his research, Henning (1966) concluded that students who had been given discrimination training without repetition practice had better pronunciation than those who had imitated a model without discrimination practice. The implications of Henning's findings were expanded by Cook (1965). He hypoth-
esized that in order to prevent errors in the initial learning stages, the teacher should not permit the students to "encode," i.e., to pronounce the sounds they hear. Asking them to pronounce the sounds that they are unable to process mentally necessitates their giving native-language interpretations to these sounds. The author extends this principle to include syntax as well as phonology. He suggests that it may be that students should be given much practice in "decoding" syntax before requiring them to "encode."
The effect of the written word on student pronunciation In 1956, Richards and Appel reported the results of their study focusing on this aspect of second-language teaching. In both the experimental and control groups, the instruction was the same except that the students in the experimental class saw no written Spanish. They found that the experimental group had higher oral production and pronunciation. The outcome of Muller's study (1964) in Portuguese was similar. He found that being exposed to the written word adversely affected the students' pronunciation.
Two other studies reached an opposite conclusion. Sawyer et al. (1963) reported that having a text was a "slight benefit" in the development of student pronunciation. Estarellas and Regan (1966) contended that the simultaneous presentation of the sound and its graphemic representation helped the students in learning and probably in retention and recall.
Transfer of learning across sense modalities Related to the insistence upon oral presentation prior to visual, i.e., the natural order of first-language learning, is the problem of transfer of learning across sense modalities. Dunkel (1948) examined a large number of studies and concluded that results of studies comparing transfer as a result of visual versus auditory presentations did not seem to vary greatly, but he indicated that visual presentation was superior in teaching grammar. Later studies on the subject have seemed to divide fairly evenly. Pimsleur, et al. (1964) found that the amount of transfer depended upon the language being learned, the learner, and her stage of development in the process of learning the language. Asher (1964) concluded that, even though there was transfer in both directions in the case of the so-called phonetic languages, in general transfer was greater from visual to auditory.
Prereading period Lange (1966) studied the effects of a prereading period in beginning French in secondary schools. There were two groups, prereading and ordered introduction to all four skills, in two different schools. After twelve weeks, the prereading group was significantly better in speaking. There was no difference between the two groups in listening, reading, and writing. At the end of a year, there were no differences.
Motivation in classes stressing oral skills The contention that students are more highly motivated in classes that emphasize listening and speaking has been widely accepted. Agard and Dunkel (1948) stated that, notwithstanding claims to the contrary, the "relentless regularity" of the drills often tires the students. In two studies on the problem of attrition in audio-lingual classes, Mueller and Leutenegger (1964) and Mueller and Harris (1966) concluded that the "drop-out" rate is higher in classes in which the students are forced to rely on only one sense modality in learning and who "have to talk so soon."