Department of Linguistics

PHONETICS AND PHONOLOGY

Pronunciation Pedagogy: Historical Development and Traditional Classroom Practice

Michael Carey
mcarey@usc.edu.au

1. Pronunciation and the Development of Language Acquisition

The teaching of pronunciation has been at-odds with the teaching of grammar and vocabulary ever since it was first studied systematically shortly before the beginning of the twentieth century (see Kelly, 1969). In the time that has passed since the acceptance of pronunciation as a contributing factor to language acquisition, it has come in and out of fashion as various progressive movements in language acquisition have prevailed.

Teachers of pronunciation have adopted two general approaches (Celce-Murcia et al. 1996, p.2):

An Intuitive-imitative approach. This approach assumes that a student’s ability to listen to and imitate the rhythms and sounds of the target language will lead to the development of an acceptable threshold of pronunciation without the intervention of any explicit information. The invention of the (recently much maligned) language lab and the audio-lingual method contributed to the support of this approach in the 60’s, 70’s and right up into the 80’s. Indeed, many contemporary second language practitioners still hold to this view but research is needed to ascertain if their beliefs have any foundation.

An analytic-linguistic approach. This approach recognises the importance of an explicit intervention of pronunciation pedagogy in language acquisition. Developments in the fields of phonetics and phonology from the latter half of the century are drawn upon and often "watered down" for use in the language classroom. Pedagogical aids such as the phonemic chart, articulatory descriptions, explanations of the form and function of prosody and practical exercises such as minimal pair drills and rhythmic chants form the basis of an explicit program of accent modification.

Korean language learners are usually not exposed to an explicit method of pronunciation instruction until they venture from high school into a TESOL classroom and have exposure to a native speaking teacher who is skilled in pronunciation pedagogy. Even now the majority of these students still learn English without actually speaking it. In their pre-adolescent and adolescent years when exposure to a native model would most benefit their acquisition of pronunciation, Korean students learn by Grammar Translation and reading-based approaches. These methods of language instruction will not be treated in the following discussion of the historical development of language acquisition models because they are not concerned with oral communication of the target language and are therefore not concerned with pronunciation.

1.1. The Direct Method

This method of language instruction, which gained acceptance in the late 1800’s, was formed from observations of children attaining their first language. Students would imitate a model of the target language that was spoken by the teacher (and later by recordings) and attempt to approximate the model through constant repetition. Later models that were built on this approach include Asher’s (1977) Total Physical Response and Krashen and Terrell’s (1983) Natural Approach. The methodology of these models consists of intensive listening comprehension, which persist for an extended period before any speaking is allowed. Proponents of these naturalistic methods maintain that:

"the initial focus on listening without pressure to speak gives the learners the opportunity to internalise the target sound system. When learners do speak later on, their pronunciation is supposedly quite good despite never having received explicit pronunciation instruction (Celce-Murcia et al. 1996)."

1.2. The Reform Movement

In the 1890’s the developers of the International Phonetic Alphabet (IPA), Henry Sweet, Wilhelm Viëtor and Paul Passy, led a movement in language teaching that was generally known as the Reform Movement. These phoneticians did much to influence the teaching of pronunciation with their contribution to the development of a system for describing and analysing the sound systems of languages and by advocating the following guidelines (Celce-Murcia et al. 1996, p.3):

• The spoken form of a language is primary and should be taught first
• The findings of phonetics should be applied to language teaching
• Teachers must have solid training in phonetics
• Learners should be given phonetic training to establish good speech habits

1.3. Audiolinguism and the Oral Approach

The contribution of the latter analytic linguistic approach to pronunciation pedagogy led to the next noticeable movement in the 1940’s and 1950’s. In the United States Audiolinguism and in Britain the Oral approach were two methodologies that were built upon the explicit instruction of phonological aspects of language. These two methods emulated the Direct Method by relying on a recording or the teacher to model the target language followed by the students’ repetition of that language. However, the teacher also drew from their knowledge of phonetics and for the first time used simplified charts of the IPA and of articulatory models in the classroom. Also introduced was the practice of using minimal pairs for listening practice and oral production. This technique was based on the concept of the phoneme as a minimally distinctive sound (Bloomfield 1933) and was used, and indeed overused by Baker (1977) in her very popular course book on contrastive segmental instruction Ship or Sheep?

1.4. The Cognitive Approach

The 1960’s was the decade which was to have a profound destabilising effect on the teaching of pronunciation. An influential new movement, Transformational-generative grammar, (Chomsky 1965) asserted that language was essentially rule-governed behaviour and not habitually learnt. This view, which found widespread acceptance in English language teaching circles, de-emphasised the role of pronunciation in language acquisition pedagogy in favour of grammar and vocabulary because:

"its advocates argued, (1) native-like pronunciation was an unrealistic objective and could not be achieved (Scovel 1969); and (2) time would be better spent on teaching more learnable items, such as grammatical structures and words (Celce-Murcia et al. 1996). "

1.5. The Silent Way

In the early 1970’s pronunciation returned to favour with the development of the Silent Way (Gattegno 1972, 1976). In this methodology (which is still in practice in the U.S.A) segmentals as well as suprasegmentals are highlighted from the very beginning of instruction. As the name suggests, teacher talk in this method is kept to a minimum. Instead of complicated articulatory and phonetic explanations the teacher indicates through gestures what the students should do. Teachers hold up fingers to indicate the number of syllables in a word, tap out rhythmic patterns and model the place and manner of articulation with their own lips and throat or with a hand puppet of the tongue and oral cavity. Also central to the Silent Way are visual teaching aids that have been found useful in demonstrating some of the more abstract principals of pronunciation to second language learners:

• The sound-colour chart. Each phoneme on a phonemic chart is assigned a colour and is referred to by colour for ease of reference.
• Fidel wall charts. A colour categorisation system which segments the letter to sound rules into phonetic bundles. For every letter or group of letters, which represent a phoneme in English, a colour is designated.
• Cuisenaire rods. Also used to teach children basic numeracy, these coloured pieces of wood of differing lengths have various uses in the pronunciation classroom. Rods can be used to build and visually demonstrate intonation patterns, vowel duration and lexical stress.

1.6. The Communicative Approach

The current dominant methodology, which persists today with criticism from some quarters, sprung into prominence in the 1980’s. The Communicative approach holds that oral communication is the primary use of language and therefore should be central to the mode of instruction. Although pronunciation is not an explicit feature in this mode of instruction, the importance of pronunciation has been highlighted by it. By focusing more on active communication in the classroom, it has been recognised (Hinofotis & Bailey 1980) that pronunciation competence below a certain threshold renders even the most grammatically and lexically advanced student unintelligible.

This renewed focus on pronunciation has raised many new problems for the teacher working within the Communicative Approach. One problem is that the artificial and contrived teacher centred nature of previous pronunciation instruction techniques does not allow for a comfortable fit with the discourse-based Communicative Approach (see Brumfit and Johnson 1979). As a result of teacher training methods centred around the Communicative Approach and the rigidly adhered-to Presentation, Practice, Production (PPP) model of instruction, a generation of teachers impartial to, and inadequately trained in the instruction of pronunciation has been produced. (More will be said of this situation in the ensuing section).

In an attempt to keep pronunciation communicative, materials writers have developed materials that focus on suprasegmental aspects of pronunciation. Generally, integration of pronunciation was only treated with token significance by material developers in the 1980’s and was more or less pushed to the outer to be taught in isolation by practitioners still interested enough in its worth and knowledgeable in its form and function.

2. Attitudes of teachers and students

2.1. Top down or bottom up? Inseparable!

Many English pronunciation instructors have felt frustrated by their inability to assess segmental production and give accurate feedback to students because of the black box model of articulatory/acoustic production. Many learners are unable to produce segmentals acceptably or with any permanency after exposure to traditional methods of minimal pair drills, simplified articulatory explanations and aural discrimination exercises. Instead, learners tend to revert to their fossilised production of L2 segmentals once the model, which has been extensively drilled by their teacher, begins to slip from their memory.

Teachers who have experienced this dissatisfaction associated with teaching segmentals tend to embrace negative assumptions regarding this aspect of pronunciation. A popular assumption which teachers may call upon to defend their avoidance of teaching segmentals and which has been given credence by some current pronunciation course books is: deviate suprasegmental aspects are more of a hindrance to clear communication than deviate segmentals (Gilbert 1993). One writer (Helgesen 1996[internet]) aligns himself with this popular notion by stating "No one, anywhere has ever walked into a restaurant, mispronounced /r/ and gotten a bowl of lice."

Although redundancy in language is a well known factor (Gimson, 1980) and intelligibility will be assisted by an immediate linguistic and situational context in actual communicative situations, it is the frequency and type of error which can possibly obscure meaning, no matter what assistance the message is given by context. In a study of natural continuous Japanese L2 speech Baldwin (1989) makes the following observation concerning the seriousness of segmental errors:

Deviations such as "I sink zat people are bery kind" can still be easily understood, however a construction (with a high frequency combination of errors) such as "I sink dzo disizbutuza" cannot. No listener understood what he was saying, even with other contextual cues present. (The latter speaker’s) observation of Sydney being a "bad town though" contains at least 7 substitutions, uttered in a continuous stream of rapid speech.

A swing of preference amongst teachers towards instruction in aspects of prosody over phonemic production in the classroom has been greatly influenced by pronunciation course books such as Speaking Clearly (Gilbert & Rogerson, 1993) and In tempo (Zawadzki, 1994) which address suprasegmental aspects exclusively. Defeated by the lack of pedagogy that yields any identifiable results in the area of segmentals, teachers may have chosen to adopt the popular assumption and turned their attention to suprasegmentals.

A Survey of Attitudes and Teaching Practices Related to Pronunciation Teaching (Brown 1992) revealed that the majority of teachers involved in the study believe TESOL training does not adequately cover the area of pronunciation teaching. Teachers feel overwhelmed by the complexities of pronunciation, not knowing what to teach, when or how, and generally lack knowledge about the phonological systems of other languages.

While the majority of teachers stated they feel the most "comfortable" with their knowledge and skills relating to the phonetic (segmental) system of English, they would like more knowledge about, and skills relating to the teaching of suprasegmental features of pronunciation. These findings beg the question: if they feel they have been poorly trained in the area of phonetics and phonology and its application to second language acquisition, how can they possibly feel "comfortable" with any aspect of pronunciation pedagogy?

From this conundrum has arisen a tendency for teachers to polarise and prioritise top down and bottom up approaches to the teaching of pronunciation. This occurrence is in deference to the fact that segmental and suprasegmental aspects overlap and contribute to each other in several important ways. The elements of pronunciation combine to form a dynamic system that cannot be isolated one from the other. For instance, vowel duration and production of the reduced form schwa /ə/, contributes to the stress of syllables in words and to the prominence of semantically loaded words in sentences. As Taylor (1996) has written:

"there is a close connection between word stress and the pronunciation of vowels and the ability to predict and recognise word stress patterns can help learners to pronounce vowels correctly. Conversely, a knowledge of the correct pronunciation of the vowels in a word will give the learners a clear indication of its stress pattern."

The use of traditional approaches and pronunciation resources which segmentalise and polarise aspects of pronunciation, contribute to the complexity which teachers claim to see in pronunciation teaching. So how should teachers structure their lessons? A current belief (Brazil, 1994) is that pronunciation should be integrated into the general language lesson instead of being taught as a separate subject. While this view has value it also raises many questions for the phonologically naive teacher: what should be integrated, how and when?

2.2. Which English?

Regarding his maiden visit to Australia at the close of 1996, American President Bill Clinton was reported to have said that he had difficulty understanding Prime Minister John Howard because of his accent.

One possible cause of this lack of comprehension between two native speakers of the same language is that the cultural exchange between Australia and America has not reached a level whereby Australian English is identifiable to Americans en masse. In Britain the cultural exchange has been more complete, our Australian actors and entertainers mix freely within their society and Australian exports containing our language enjoy enormous success. In America on the other hand, Australian actors have been required to undergo accent training to adopt an American accent when they appear in American films and Australian blockbusters have often been dubbed with American speakers prior to release. The cause of this confusion could also be attributed to the similarities in pronunciation shared by General Australian English and the English found in the vicinity of London, recently named Estuary English, which are not shared by General American English.

In terms of population of native English speakers the United States of America dominates over the other nations. The other major populations of native speakers follow in the order of Canada, Britain and then Australasia. Group Canada together with the U.S.A and call it Northern America and the dominance of American-accented English is overwhelming.

So often the question arises in pronunciation pedagogy circles: which English do we teach? If the dominant model were to qualify the pronunciation of choice should clearly be North American. If an historical angle were to be taken some variety of British English should be chosen. For the last century Received Pronunciation (RP) which is only spoken by approximately 3% of the British well-to-do population (according to Crystal 1988) was preferred. These notions which are based on the idea that there is, or should be a Standard English do not match the reality of language acquisition. English is taught on every continent on the globe by every conceivable variety of native and non-native speakers possible. Yet as long as the question of dialect superiority is still raised the debate will continue to influence the attitudes of English language teachers and students alike.

In Australia, though the mode of English instruction seeks to be Australian in form and content, the teachers themselves are often from diverse linguistic backgrounds. Many of them are travelers or temporary residents and to model any other accent other than their own would be absurd. So the notion that foreign language students arriving in Australia will acquire aspects of native Australian-accented English is not a reality. This is also the case for the other national centres for English language instruction around the globe. English is no longer owned by the English or by any one particular dominant group and therefore to search for a standard model built upon one dialect for pronunciation instruction is not realistic.

Instead, intelligibility of speech should be the modus operandi when considering what should be taught. Any native speaker of English may teach suprasegmental aspects of pronunciation, irrespective of their nationality or the country in which they are teaching. The choice of a model for segmentals however is not so straightforward. The best solution to this dilemma would seem to be to reserve instruction of segmentals to a native speaker of the variety of English in the country of instruction. That is, a native Australian should teach segmental aspects of pronunciation in Australia, a native northern American speaker should model segmentals in their homeland and so on. This may appear to be a solution except that the previous section outlining the inseparable nature of segmental and suprasegmental aspects of pronunciation negates this option, as does the Communicative Approach to language teaching which seeks to teach through discourse and not through explicit analysis.

The dialect superiority misperception also contributes to the reticence of some language teachers to tackle pronunciation teaching for fear of being accused of being an inferior model. If the bulk responsibility for pronunciation instruction were to be taken away from teachers and given to an audio-visual tool such as a computer which could consistently model and far more accurately analyse speech and also give diagnostic feedback to the student, many teachers would have no cause to object.

3. Traditional pedagogy

In traditional pedagogy, pronunciation has been viewed as the sum of three components. The components are sounds, stress and rhythm, and intonation. Perhaps it is due to the complexity of these components, that traditionally in the classroom these components have been explicated separately. As has already been mentioned, many teachers tend to regard these components with varying degrees of importance and may choose to focus on one of them in preference to the others. The teacher’s degree of phonological knowledge may also inform the selection. Critical observations on the limitations of teachers and their methods have their place but for the purpose of this discussion a more positive pragmatic view will be taken in an attempt to unearth those aspects of traditional practice which have merit. This section will summarise some of the novel ideas that have been used in the classroom.

The most common method of modification at the phonemic or sounds level is based on minimal pair drills. This method assumes that the speaker cannot perceptively or productively distinguish minimally distinctive sounds such as the vowels in ship and sheep because these distinctions do not occur in their L1. Traditionally, minimal pairs, or minimally distinctive sets of more than two sounds in the target language are randomly modeled by the teacher from a list and the students guess which sound in the pair or group has been produced. This is an effective exercise for elicitation of the student’s perceptual problem areas and for awareness building of the sound contrasts in English. However, problems are encountered in giving feedback to the students on their production errors, particularly with problem vowel sounds. The traditional practice of production uses articulatory diagrams to explain the position of the articulators, followed by a drill of the contrast. In a classroom situation teachers don’t have sufficient time to focus on students individually, nor to analyse the phonetic/phonemic transfer effect occurring between the L1 and the L2 or any of the intra-personal sources of error which could possibly occur. This difficulty can be partly overcome by focusing on students individually outside the classroom in a one-on-one clinic. This however is not an option for most language schools.

The author has observed the operation of one such clinic run by speech pathologists for Japanese learners of English at Cumberland College, Sydney University, Lidcombe. An effective technique they have developed is as follows. The speech pathologist models a target sound within a carrier word. Then the student is asked to produce it two or three times, depending on how many attempts it takes to produce an improved production of the sound. Then the student is asked which sound they think was better, the first, the second or the third? Surprisingly, using this method, the student can often perceive or feel which utterance was closer to the model. By asking the student to reflect on what they feel is a fair production of the model, the closer attempt is thereby reinforced in a way which can not be attained if the teacher was just to comment good or better or try again. It would seem from this example that gaining an intuitive awareness of the accuracy of one’s production skills of the sounds in a second language is a significant step towards improved production.

The minimal distinctions between monophthongal vowels in AE may be; mainly durational as in the contrasts /i/ versus /ɪ/ and /aː/ versus /ʌ/, or it may be due to other acoustic cues expressed by varying articulatory motions and realised as formant frequency differences. In diphthongs the salient cues are the production of the first element, the length of the glide and the direction of the glide. This however also varies considerably between native speakers. For consonants the salient acoustic cues are related to the manner of articulation and place of constriction. These qualities which separate these L2 sounds are interesting in their own right, but what is the significance of these distinctions to the second language learner? Perhaps it is pertinent to ask, what is the point of drilling contrasting L2 sounds? Contrasting L2 sounds with durational difference may be valid for listening practice. However, why compare the student’s production of two L2 vowel targets such as /uː/ and /ʊ/ when the student probably can’t come close to producing either accurately. In this case neither can be relied upon as a point of reference for the other. An alternative to the minimal pair drill practice is offered below. Firstly, the fundamental inadequacies of minimal pair drills will be outlined. The practice of minimal pair drills presupposes the following:

a) The L2 speaker can already approximate acceptable acoustic cues for both target phonemes but confuses them; b) Drilling one minimal pair in relation to another will modify the acoustic cues towards more native-like speech;

c) The L2 listener has the ability to perceive production of English sounds in the same way the L1 listener does.

These presuppositions are flawed for the following reasons:

a) In many instances the learner will not yet be able to produce acceptable acoustic cues for the target phonemes in question, so any drilling to reinforce the contrast between the minimal pairs is putting the cart before the horse;

b) If the L2 speaker can not as yet produce acceptable acoustic cues for the sounds in question they have a spurious point of reference on which to contrast the production of the two sounds;

c) The L2 listener does not have the ability to perceive speech in the same way the native English speaker does. The L2 listener is constantly influenced by the inventory of acoustic cues from their L1. For this reason the L2 speaker should not be expected to gauge the accuracy of their own production, especially before their production of L2 sounds has reached an acceptable proximity to the target.

A better strategy than contrastive drilling of L2 minimal pairs would be to use the closest corresponding target from the L1 as a point of reference for an L2 target. In classroom pronunciation teaching this has never been an option because a teacher cannot possess native productive and perceptive competence in L1 and L2 phonology. Nor can they visually demonstrate the contrast by any physical model at their disposal. This L1-L2 contrast could however be assessed and visually demonstrated with the computer-based model that is outlined elsewhere (Carey, 2002).

The more innovative classroom approaches to pronunciation modification have evolved from the pedagogical goal of the learner-centred classroom, which has replaced former didactic approaches to teaching. In the learner-centred classroom activities related to the correction of pronunciation errors are designed to meet the student’s differing learning styles, namely auditory, visual, tactile and kinaesthetic learning. Conformity to the learner-centred model does not necessarily make the method an effective tool for modification of pronunciation, but it is likely to make the lesson more palatable for the student. The following are examples of what are now being referred to as low tech practices, in light of the evolving high-tech computer-based approaches.

3.1. Traditional practices

Sounds

Rubber bands have been employed to both demonstrate and raise individual student awareness of length contrasts between vowels. For example, using a rubber band to model /iː/, the students place the rubber band around both thumbs and stretch it to its limit. To model /ɪ/ students pull the rubber band slightly apart. The teacher can also point out the co-articulatory effect on vowels between voiced and voiceless consonants. The students can then identify long and short variance by listening to an utterance and modeling the length contrasts with their rubber bands.

The phonemic chart can be used effectively in class to learn new vocabulary by pointing at the sounds in the chart to elicit a word. This can only be done after the students have had lengthy exposure to the symbols. One way, which I have developed myself, for students to learn the symbols quickly is as follows. Cut up at least two copies of the chart with extra schwa symbols and give the pieces to pairs of students to arrange into words. The chart should have commonly occurring example words under each symbol. Monitor the students, explain any errors they make and point out patterns such as consonant clusters and the use of schwa. Elicit from the students the prominent syllable in each word.

A common drill used to practice minimal pairs which this author takes issue with is tongue twisters. Tongue twisters are pedagogically unsound because they drill sequences of discourse with frequencies of phonemes that do not occur in natural speech. Besides this the difficulty of the exercise is de-motivating. It is even too difficult for a native speaker to say red lorry, yellow lorry ad infinitum at a rapid pace.

Rhymes or jazz chants may be used to practice difficult sounds in a more enjoyable way than by using isolated words or contrived dialogues with an unnatural repetition of phonemes. Jazz chants are usually used to practice rhythm and stress in functions and common expressions such as the one below for giving directions. They can also be used with an integrated focus on problem sounds such as /ɜː/.

Walk three blocks and turn right,
Turn right,
Turn right,
Walk three blocks and turn left,
Turn left,
Turn left,
Walk straight on and turn right, (etc.)

The heavy rhyming pattern in Billy Field’s song Bad Habits is also appropriate:

When you get the urge,
And you have to splurge, (etc.)

Songs may also be used to drill sequences of phonemes that conflict with the phonotactic constraints of the L1. For example epenthesis, or elision of consonant finals may be worked on with songs such as the following which practices consonant finals /t/ /m/ /n/ /d/ and /s/:

Silent night, holy night. All is calm. All is bright.
Round young virgin, mother and child.
Holy infant, so tender and mild,
Sleep in heavenly peace, sleep in heavenly peace…

Cuisenaire rods. (Mentioned in 1.5.)

Stress and rhythm

A rubber band can be used again to model the prominent syllable(s) in a word or utterance. It can be stretched out during the stressed syllable and left short during other syllables. The point of the exercise is to help certain students avoid transferring their syllable timed language to their production of English.

Clapping, tapping or playing simple rhythm instruments can be an enjoyable way to demonstrate and practise the prominence of content words in discourse or in rhymes. Using natural discourse focuses attention on the often slightly irregular stress patterns that occur in real discourse. Rhymes can be used as an initial demonstration of prominence but should be followed up by real discourse, as their regular meter is in fact a contrivance.

Jazz chants. (See the previous section).

Magnets can be used as a physical metaphor to demonstrate linking. A lot of preparation is required for this activity which could be demonstrated in other ways. Magnets can be glued inside audio-tape cases or anything similar and then words are glued to the cases so that the +/- poles attract or repel the appropriate linking vowels or consonants. The students may believe that if their teacher has gone to this much trouble to reinforce linking it must be important for them to learn to do it.

Intonation

When intonation has been dealt with at all in traditional pronunciation practice it has usually been represented by an attitudinal approach. Therefore the traditional techniques used to teach intonation have been informed by an approach which has since been debunked. A summary of the methods used is therefore not relevant. A typical tool used however is a kazoo, which the teacher uses to hum an utterance and thereby exemplify the tune. Such a trite demonstration undermines the complexity of intonation and does nothing to improve the student’s production. The author’s preference for a pedagogical model for intonation is the model developed by Brazil (1994).

References

Asher, J.J. (1977) Learning another language through actions: The complete teachers’ guide book.

Baker, A. (1977) Ship or Sheep? C.U.P.

Baldwin, J.P. (1988) Intelligibility and the Interlanguage Phonology of Japanese English. Unpublished BA (Hons) Dissertation, Macquarie University.

Brazil, D. (1994) Pronunciation for Advanced learners, (Teacher’s and student’s books) C.U.P.

Brumfit,C.J., & Johnson, K. (Eds.).(1979) The communicative approach to language teaching. Oxford. Oxford University Press.

Carey, M. (2002) An L1-specific CALL pedagogy for the instruction of pronunciation with Korean learners of English, Unpublished Ph.D. dissertation, Macquarie University.

Celce-Murcia, Marianne; Brinton, Donna.M; Goodwin, Janet. M; (1996). Teaching Pronunciation- A reference for teachers of English to speakers of other languages. C.U.P.

Crystal, D. (1988) The English Language Today. Penguin.

Gilbert, J.B. & Rogerson, P. (1993) Speaking Clearly: Pronunciation and listening Comprehension. Student’s book. New York. CUP.

Gimson, AC & Cruttenden, A. (1994) Gimson’s Pronunciation of English. London: Edward Arnold.

Hinofotis, F., & Bailey,K.M. (1980). American undergraduates’ reactions to the communication skills of foreign teaching assistants. In J.C. Fisher, M.A. Clarke, & J. Schacter (Eds.), in TESOL ` 80 (pp.120-135). Alexandria, VA: TESOL

Kelly, L.G. (1969) 25 centuries of language teaching, Rowley, MA: Newbury House.

Krashen, S.D. & Terrell, T.D. (1983). The natural approach. Hayward, CA: Alemany Press.

Taylor, David S. (1996) Demystifying Word Stress. English today 48, vol 12, no 4. p.46. C.U.P.

Zawadzki, H. In Tempo (1994) NCELTR, Macquarie University, Australia.