The audio lingual approach however focuses attention on speaking rather than listening. While the audio-lingual approach places listening first in the sequence of language skills, it does not appear to place it first in importance. Indeed in the audio-lingual classroom a student "listens" not so much to comprehend as to mimic (in repetition drill) or to respond with the correct form (in pattern response drill ). In practice, the audio-lingual approach has consistently stressed the lingual at the expense of the audio. Active oral drill and expression has received far more attention than aural comprehension. Audio-lingual teachers are admonished to maximise each student's active oral opportunities in class recitation. In a good audio-lingual class the students do the talking, not the teacher.

More recently, some studies are indicating that this focus on speaking in the early stages is actually detrimental to learning. Not only should more emphasis be placed on listening, but far less emphasis should be placed on speaking during the early stages. The requirement to respond orally imposes listening for speaking and results in impaired comprehension. This problem was treated in a short experiment by James Asher involving the performance of drill movements by commands in Russian. Asher found that the students who merely performed the commands did so with a greater degree of reliability than those who first repeated the command before performing the necessary movement. He concluded that "... the stress of trying to pronounce the alien utterance may retard listening fluency.... The optimal strategy may be serial learning in which one achieves listening fluency just before one attempts to speak." ⁵

There is other experimental support for this position. In a study on the associative reaction time in language acquisition, Ley and Locascio state: "Our research suggests that one must make associations to verbal materials during learning in order that the material can be later recalled, and that some procedure such as repeatedly saying the material aloud interferes with the association process, and therefore has a detrimental effect on learning."⁶Part of the explanation for these results may be the role that task overload and anxiety play in the learning process.

A number of studies indicate the detrimental effect of task overload to the learning process. In a study which reviewed a large number of experimental reports within the aerospace industry, Greer et al., point out that when a subject is task overloaded, errors increase and there is a tendency to revert to previously learned generalisations rather than the specific distinctions involvedat that point.⁷James Miller refers to this process as the tendency to stereotypy.⁸In foreign language learning, this would mean that as a person became task overloaded (trying to both speak and understand before ready) he reverts back to the habit formations he knows best. This results in speaking the new language with sound habits and grammatical patterns of the old language. Heavy accents and poor word order are common with the task overloaded student. As Gaier has expressed it: ". . . it leads to an impairment in the ability to improvise in an unstructured and/or new situation. This results in stereotyped, habitual, and familiar approaches that may be maladaptive in the situation."⁹

Anxiety, often brought on by task overload, can be a major deterrent to listening comprehension. The effect of anxiety on learning in general has been widely studied. Eysenck points out that, "... conditioning is related to anxiety for the simple reason that it has been shown conclusively that the ease with which a conditioned reflex is formed depends very much on the anxiety of the person on whom the experiment is being performed. What is more, there is a good deal of what is called ' stimulus generalisation' in the anxious person.''^l0Any discrimination learning is much more difficult with the anxious person. Therefore any of the receptive or decoding skills such as listening comprehension and reading, which are primarily discrimination learning, is made more difficult through task overload and the corresponding anxiety which is created.

There is evidence that the delay of oral response in the initial stages actually improves the proficiency of speaking once it is commenced, provided that the pre-vocal period is devoted to training in aural comprehension. Postovsky challenged the assumption that the initial phase of instruction should be based on intensive oral practice, that the more vocally active the student is from the beginning, the faster he learns the foreign language. Instead, he proposed that the motor skill involved in the production of speech output is an end result of complex and mostly covert processes which constitute linguistic competence. He further proposed that: ". . . the linguistic competence includes at least two reciprocally correlated events: capacity to process auditory input and capacity to generate speech output,....''^llThe former he indicates is concerned with decoding capability which requires the development of recognition knowledge, while the latter is concerned with encoding capability which requires development of retrieval knowledge. He felt that it was logical to assume that the development of recognition knowledge should precede, not follow, the development of retrieval knowledge.

Starting with this assumption, Postovsky then set about to test the effect of replacing oral response with written response during the first four weeks of the Defense Language Institute's Russian course.¹²All told, there were about 100 experimental students and 100 control students, none of whom had any previous exposure to Russian. The sole and only difference in the experimental procedure was to delay the introduction of oral practice until after the fourth week or approximately 120 contact hours with the language in this intensive course. Instead of pronunciation practice, instead of memorising daily dialogues for oral recitation, and instead of engaging in oral pattern response drills, the experimental students responded during the four-week pre-vocal phase entirely in writing.

After four weeks the experimental group commenced oral drill and the methodology for both the experimental and the control groups was identical. At the end of six weeks and twelve weeks tests were administered to the