Two groups of English speakers received either auditory or articulatory instruction in learning to produce exotic sounds. Performance on production and discrimination tests indicated a striking superiority for the subjects who received systematic training in the production of exotic sounds as opposed to those subjects who received only discrimination training in listening to these sounds. The results of this study suggest that what is effective in the teaching of sound production and discrimination is the systematic development by small steps from known articulatory postures and movements to new and unknown ones. The possession of a scientific knowledge of articulatory phonetics by the teacher was shown to be extremely successful in leading students to the correct production of foreign sounds and thereafter to facilitate the discrimination of these sounds. The latter finding was taken as support for some carry-over from productive competence to auditory discriminatory competence.
Some seven years ago, John B. Carroll wrote as follows on the teaching of foreign language phonology:
Speculation among linguists seems to run to an almost schizoid indecision as to which of two diametrically opposed theories to accept: (1) that there is an automatic capacity to form the correct modes of sound production simply by careful and repeated listening--as if the learner is already "wired" to pronounce sounds correctly if he will only give full rein to this automatic capacity, or (2) that (except possibly for the young child) the learning of a foreign phoneme occurs as a result of conscious attention to the articulatory processes involved in its production, and that a scientific knowledge of articulatory phonetics is a positive aid (Carroll, 1963, p. 1070).
At that time there were few relevant research results available, and Carroll added: "We have a rather neat experimental problem which urgently needs exploration." The problem has not so far received this exploration.
...
A lesson possibly to be learned from this is the inadequacy of any purely auditory tape-recorded pronunciation-training program which relies entirely on mimicry of vowels without supplying explicit information at least on lip positions. This, after all, is one of the very simplest phonetic features to describe and teach, even to people with no phonetic training.
In general, the results clearly vindicate the view that if you want people to produce sounds you must accurately train them to do just that. This would seem to be a truism, but the fact is that, as Carroll implies in the quote at the beginning of this report, there apparently is a current belief that you can teach people to produce sounds by merely making them listen to them. Our results certainly indicate that auditory methods are significantly less effective than teaching production by means of systematic application of articulatory phonetic knowledge. However, this point must be emphasized: what is effective in the teaching of sound-production is the systematic development by small steps from known articulatory postures and movements to new and unknown ones. That is to say, the application of phonetic knowledge by the teacher enables the student to pick up some knowledge of "phonetic theory" inductively as a result of experiencing phonetic activitie in his own vocal tract.
If, as we have said, it is not surprising that subjects learn to produce sounds through being taught to produce them, it may indeed appear a little surprising that they thereby also learn to identify them by ear. Our results show that Group A, taught by exclusively articulatory techniques were significantly more successful at identifying sounds by ear than the group taught by purely auditory techniques. This obviously implies some kind of carry-over from productive competence to auditory discriminatory competence, and may, indeed be taken to be some support for a "Motor theory of speech perception" (Liberman, Cooper, Harris, & MacNeilage, 1963), As a matter of fact, it has been the experience of one of the investigators in a lifetime of teaching phonetics and analyzing languages, that "exotic" sounds can generally be more readily and unerringly identified after one has learned to produce them.
Be that as it may, our investigation indicates that "ear-training" and mimicking alone are less effective than articulatory training in teaching both the auditory discrimination and the production of exotic sounds. It seems to us that these preliminary findings are worthy of additional investigation preferably with a larger group of subjects and with speakers of several different languages.