the English consonants [ f, v, , ] in various syllabic positions and sequences of the consonants [ w, , t

The Production of a More Detailed Description of Problems Experienced by Native Japanese Speakers When Articulating Certain English

Speech Sounds (1) : Contrastive Analysis.

Timothy Ashton

^＊

Abstact: Various published works make reference to speech sounds in English that native Japanese speakers find problematic. Such texts are used by EFL teachers worldwide. However, teachers would no doubt benefit from a more detailed determination of which speech sounds and sequences of sounds can be more or less problematic to learners. Such information would help determine how much attention these sounds should receive during class tuition. Twenty speech sounds and sequences were selected for investigation:

the English consonants [ f, v, , ] in various syllabic positions and sequences of the consonants [ w, , t

^h

] with various vowels. The aim of this study was to investigate the degree of difficulty surrounding each sound in the above syllabic conditions. An English-Japanese CA was conducted, based on the psychological notions of L1 Transfer and Interference. The above English sounds and sequences were described in articulatory terms and common conditions of occurrence. The Japanese speech sounds that

＊

Foreign Language Instructor, Fukuoka University

most closely correspond were described in the same fashion. Predictions of error were posited, based on contrasts in the phonetic descriptions of the English and Japanese sounds. The consonants [ f, v, , ] (where applicable) were predicted as most problematic in final position. Interference from Japanese syllable-timing was predicted as the main source of error in the sequences [ w, , t

^h

] plus vowel.

1. Terminology

Many EFL writers use terminology that is often ambiguous and obscure, resulting in the confusion of many linguistic categorisations. This terminology is generally used to provide explanation in terms familiar to the reading EFL teacher. In order to maintain clarity and consistency throughout this study, the terms used shall be listed and defined at the beginning of each section. The following terms appear in sections 1-6:

'Error': systematic and consistent deviation from a given norm, representa- tive of the state of the learner's L2 system at a given stage of acquisition or development. As opposed to 'mistakes' or 'lapses': random deviations which, when pointed out, can easily be corrected by the learner. This presupposes that the learner is already aware of the desired form (Sridhar 1981: 224).

'Contrastive Analysis (CA)': the systematic comparison of specific linguistic characteristics of two or more languages (Van Els et al 1977: 38).

'Error Analysis (EA)': the study of systematic deviations from a given norm produced by learners of a target language (Sridhar 1981: 224).

'Interlanguage (IL)': the learner's observable (through performance)

linguistic system perceived as an approximation of both L1 and L2, also

featuring unique elements non-attributable to either language (Sridhar 1981: 227, 230, James 1980: 3, Richards and Sampson 1974: 29-30). Alterna- tive terms for this phenomenon are 'approximative system' (Nemser 1974), 'idiosyncratic dialect' and 'transitional dialect' (Corder 1981).

'Intralingual errors': those which reflect the general characteristics of L2 rule learning, e.g. overgeneralisation of L2 target language rules, semantic features, etc. For example, a learner may assume that all syllables containing a vowel sound, but that do not carry main/primary stress, involve a reduction of the vowel to 'schwa' [ ] (which is not the case in words such as ' ´ heating' [

I

] or 'consti ´ tution' [ ,

I

]

¹

. (Selinker 1972: 37, 49, Richards and Sampson 1974: 29, Sridhar 1981: 229.)

'Syllable': the organs associated with speech contract and relax (at approximately five times per second) producing a pulse-like flow of air from the lungs. These are termed 'chest pulses', which in speech largely corre- spond to perceived single syllables (Abercrombie 1965: 16-17, 1967: 35). In reality however, the chest pulse does not correspond to the broad phonetic transcription of syllables as the perception and production of these 'groups of sounds' varies from language to language. For the purposes of this study, a 'syllable' will follow the definition of combinations of consonant and vowel sounds perceived and produced as groups by native speakers of English and Japanese.

'Speech sound' or 'phone': any continuous fraction of a phrase that is perceived as coextensive, and in turn representative of an articulation (an articulation is defined in terms of movement and positioning of the organs

１´= syllable carrying main or primary stress within a word.

associated with speech) (Bloch 1950: 89). The notion of 'single' speech sounds exists for practical purposes only and is those fractions which can be represented by a single IPA symbol (or equivalent). Diphthongs (included by some writers as a 'courtesy') are not 'single' sounds by definition but involve a single distinct articulatory movement.

'Phonetic Skills':

i. 'Oral skills': production, pronunciation, articulation of speech sounds.

ii. 'Aural skills': listening, perception, reception, discrimination of speech sounds.

'Speech sound' vs. 'Phoneme': Kenworthy (1987), along with numerous other EFL writers, uses the term 'phoneme' when in fact referring to single speech sounds in broad transcription. The term 'phoneme' however, introduces the concept of communicative function, which is not used or systematically applied in any way by the author. Phonemic analysis is problematic and in dispute (as well as not being the purpose of this study) therefore only phonetic terms will be used unless otherwise specified.

Examples of 'phonemes' listed by Kenworthy (and other EFL sources) are / a

I

/ and / a /. These sounds are clearly complex: /

I

/ and / / commute, / a / in / a

I

/ commutes with / / in /

I

/ and / / in / /. For example:

/ ba

I

t / / ba

I

l / / ba / / ba t / / b

I

l / / b /

/ a

I

/ and / a / are therefore not minimum successive units which is a

requirement (and part of the definition of) a 'phoneme'. Note that

Kenworthy also lists / a / and / / as separate phonemes (as in / kat / and

/ n /) throughout her text. It can therefore be concluded that much of the

information in Kenworthy (1987) is useful provided the misleading

terminology is stripped away.

2. Materials Used To Teach Native Japanese Learners of English

EFL publishers in Japan continue to produce new titles aimed at classroom use each year. Sadly, however, many such publications ignore areas of phonetic skills essential to English communication, despite proclaiming this as a common goal. Instead, teachers tend to refer to 'English pronun- ciation texts' when preparing lessons. A number of these publications have been available to teachers over the last three decades. Amongst other nationalities, such works typically list the most common oral and aural problems that Japanese learners experience when learning and practising English as a foreign language. Teaching and learning strategies which can be applied to the EFL classroom are also often provided. These are quite clearly based on the common sense, intuition and research of teachers and linguists. In this section two such works are discussed: Joanne Kenworthy's "Teaching English Pronunciation" (first published 1987;

Longman) and "Pronunciation Contrasts in English" by D. L. F. Nilsen and

A. P. Nilsen (1971; Regents Publishing Co.). These texts have been selected

as they are still to date recommended for classroom use. Nilsen and Nilsen's

older work was favourably reviewed by Frank Rowe (a university teacher

in Tokyo) in the September 1994 issue of "The Language Teacher", the

monthly publication of the Japan Association for Language Teaching. In

his article, Rowe goes on to explain how the contents of the book may be

used to implement regular English pronunciation practice to classes of

university students over the academic year. Kenworthy's work is more recent and can be found on the shelves of staff rooms in English language schools throughout Britain and Japan. The aim of this study is to provide further, supplementary information to works such as the above two. It is hoped that this will help to facilitate teachers' knowledge of Japanese learners' perceptions of, and pronunciation problems with certain English speech sounds.

3. Summaries of Nilsen and Nilsen's and Kenworthy's Pub- lished Texts

In their text, Nilsen and Nilsen set out to list and describe the sounds in English that prove difficult for native speakers of various languages to master. Their system evolved from the concept of 'contrastive pairs'.

²

A contrastive pair may be defined as two 'single' speech sounds with one or in some cases more, articulatory differences. For example, contrast can range from the maximum in 'boy' vs. 'gate' via the single speech sound 'boy' vs.

'toy', to the absolute minimum of one articulatory feature e.g. 'bat' vs.

'bad'. Even in the latter case, vowel length varies contextually however. The concept is then extended to 'contrastive sentences' for the student to practise. These contain the given contrastive pairs of sounds and are useful in that they provide partial communicative contextualisation. The authors

２

Nilsen and Nilsen originally use the term 'minimal pairs', defined as "two words

pronounced alike except for a single phonemic difference ... responsible for radical

changes in meaning." (p.xiv). Also the term 'minimal sentences'. These terms have not

been used here as a true minimal pair of sounds only differs in a single articulatory

feature. This is true for some of the cases listed in their text, but others display

several articulatory differences.

claim that a system of contrastive pairs makes a valuable aid to teaching and testing both oral and aural phonetic skills (p.xiv). As justification, the further claim is made that contrastive pairs encourage the student to make distinctions between the sounds of English rather than those of their native language (p.viii). According to Nilsen and Nilsen there are 27 contrastive pairs of English speech sounds that Japanese learners find problematic.

These are listed below:

1. [ i ] and [

I

] as in beat and bit 2. [

I

] and [ ] as in bait and bet 3. [ a ] and [ ] as in bat and bet 4. [

I

] and [ a ] as in aid and add 5. [ a ] and [ ] as in cat and cot 6. [ a ] and [ ] as in bat and but 7. [ ] and [ ] as in cut and cot 8. [ ] and [ ] as in buck and book 9. [ ] and [ ] as in cot and coat 10. [ ] and [ ] as in dot and caught 11. [ ] and [ a ] as in dot and doubt 12. [ ] and [ ] as in but and bought 13. [ l ] and [ ] as in lack and rack 14. [ w ] and [ hw ]

³

as in wet and whet 15. [ v ] and [ f ] as in vat and fat

３

[ hw ] as in

whet is included here to faithfully reproduce Nilsen and Nilsen's list.

However, some teachers may not feel its inclusion necessary as it is generally a

feature of British Received Pronunciation. Many speakers of Standard Southern

Spoken British English articulate both 'wet' and 'whet' using [ w ].

16. [ h ]

⁴

and [ f ] as in hat and fat 17. [ b ] and [ v ] as in ban and van 18. [ v ] and [ z ] as in veal and zeal 19. [ s ] and [ ] as in seat and sheet 20. [ ] and [ ] as in thank and shank 21. [ t ] and [ ] as in team and theme 22. [ s ] and [ ] as in sink and think 23. [ ] and [ z ] as in then and zen 24. [ d ] and [ ] as in dare and there 25. [ d ] and [ ] as in din and jin 26. [ n ] and [ ] as in fan and fang 27. [ ] and [ ] as in gag and gang

In part II of her handbook for language teachers, Kenworthy contrasts the sounds of English with those of nine other languages, including Japanese.

These are discussed in terms of the articulatory problems learners tend to have. In a 3

¹

/

2

page summary dedicated solely to native Japanese speakers, problems with consonants, vowels, stress and sentence rhythm, intonation and weak forms all receive due attention. Selected references from Kenworthy's text are reproduced below :

i) Problems with consonants: The following consonants do not occur in Japanese: [ v ] as in 'van', [ f ] as in 'fan,' [ ] as in 'soothe', [ ] as in 'run'.

ii) 'Sound substitution' errors: The sounds [ z ] or [ d ] will be used for [ ].

４

The symbol [ h ] is used for convenience here, but the example 'hat' is correctly

transcribed as [ at ] , as [ h ] is merely the voiceless counterpart of the following

vowel.

iii) Problematic consonant-vowel sequences: The following consonants occur in Japanese and can be transferred to English with little or no need for modification:

[ p, b, t, d, k, , , , s, z, , m, n, , w, j, h ]. However, some of these consonants only occur before particular vowels in Japanese, so learners may have problems when they try to pronounce an 'unfamiliar' consonant- vowel sequence. Here are some potential trouble spots:

a) The only vowel which commonly follows [ w ] as in 'wet' in Japanese is [ a ], so whereas 'wax' will not be a problem, 'win', 'wait', 'would', 'we', etc.

may be.

b) The sequences [ t

^h

i ] and [ t

^h

uw ] do not occur in Japanese. Watch out for problems with words like 'team', 'two', etc.

c) [ ] does not occur before [ ] as in 'bed' or [

I

] as in 'made'; possible problems will be 'shade', 'shell', 'shame' etc.

From the above, it can generally be concluded that further research could uncover far more detail that could be added to Kenworthy's text. In general, there appears to be a lot of scope for additional comment, especially concerning references to consonantal groups and stress patterns (etc.).

The authors of both texts claim that these areas of difficulty are in many

instances predictable as they result from the interference of deeply

established language habits from the individual's native language. These

tend to predominate until new English speech patterns have been mastered

(Nilsen and Nilsen 1971: viii). It should be noted that this does not necessar-

ily have to be the case. It is clear that communication requires a conven-

tional system of distinctions and that conventions differ from language to

language. Problems might be expected where differences of convention occur.

These beliefs reflect notions of Behaviourist psychology that have provided the foundations for Contrastive Studies since the 1940s (see 5.3). Nilsen and Nilsen's justification for the use of contrastive pairs (mentioned earlier) also appears to be based upon these assumptions, albeit more crudely. As this piece of research is a contrastive analysis of English and Japanese speech sounds, the relevance of these concepts to the present study is clear.

4. Comments and Criticisms

Numerous parallels can be detected upon comparison of the two texts where the authors have reached the same conclusions, for example:

i) Perception of L2 speech sounds: Japanese speakers often cannot distinguish English [ l ] from [ ], [ b ] from [ v ], [ i ] from [

I

], etc.

ii) Substitution of L2 sounds with approximative sounds: Learners generally tend to use (in production) the nearest available articulation to the L2 sounds causing difficulty. This may be an L1 sound that also articulatorily resembles another L2 sound. For example, [ z ] or [ d ] will often be used instead of English [ ]; similarly [ s ] or [ t ] will be used for [ ]. This is of course, a corollary of the view that the native system interferes, and therefore the above are examples of interlingual errors.

Alternatively, the sound produced in error may simply be the result of

incorrect assumption (on the part of the learner) as to the correct articula-

tion of the desired sound, in which case the error could be intralingual

(amongst other possibilities; see 5.3).

iii) Production of L2 speech sounds with no close L1 approximation:

There are sounds which are not common to the pronunciation of a given pair of languages. 'Schwa' [ ], for example, does not occur in Japanese.

Possibly due to the syllable-timed nature of Japanese, learners tend to replace [ ] with 'full' vowel sounds. Again, the cause of the error could be either interlingual or intralingual.

However, it may be felt that there is room for further detail within descriptions of problematic items such as the above. There is little discus- sion by Kenworthy, and none at all by Nilsen and Nilsen, on the following aspects concerning conditions of occurrence of speech sounds:

a) Generally syllabic structure is not looked at. The texts lack comment as to which sounds may be more or less difficult to perceive or articulate in their initial, medial or final positions, for example the aforementioned [ v ] or [ ]. Nilsen and Nilsen merely list examples of problem sounds in these positions. No comment is made on the comparative difficulty of each.

b) Which sound substitutions are more common, e.g. [ z ] or [ d ] in the case of [ ].

c) Which combinations of consonant and vowel sounds can be more or less problematic. For example, the sequence [ wa ] occurs in Japanese and so learners should have little difficulty in articulating or perceiving English words such as 'wag', or 'wax'. Other sequences of vowels succeeding [ w ] however could be troublesome, e.g. [ w

I

, w , wi, w ]. Kenworthy lists some of these combinations but gives no indication as to the possible degree of difficulty surrounding each one.

Evidently, a more detailed account of the phonetic properties of these

sounds (and other similar cases) with their possible subsequent effects on

communication is required. The above points have been dealt with during this study for the purpose of providing further detail to comments made in both publications (if any at all).

5. An Overview of the History and Applications of Contrastive Analysis

'Contrastive' theories and methodologies have traditionally been applied to three main areas of research (Van Els et al 1977: 38):

(i) Providing insights into similarities and differences between lan- guages.

(ii) Predicting and explaining problems in L2 learning.

(iii) Developing course materials for language teaching.

5.1 Providing Insights into Similarities and Differences between Lan- guages

Comparative studies in linguistics have a surprisingly long history

spanning several fields of research. For example, linguists have compared

different but related languages, and stages of the same language, using

known sources, in order to describe the history of a single language

(Historical Linguistics) at given stages of development, and to reconstruct

(unattested) proto-languages (Comparative Linguistics). Languages as they

are used today have also been compared in order to classify them into

groups (Typological Linguistics). Finally, two or more languages may be

compared to ascertain their similarities or differences (Contrastive

Analysis or Contrastive Study). This area of research is conducted for the purposes of academic understanding of the variety of linguistic structure or more often, as a tool for improving language teaching and learning. The latter is the sole concern of this study.

Typological and Contrastive Studies can also be said to belong to a common branch of linguistics, Synchronic Comparative Studies, as they both involve the comparison of languages synchronically. Synchronic Linguistics was defined by De Saussure (1959) as "everything that relates to the static side of our science" and is strongly related to the consideration of languages as systems of convention for communication considered at a given point in time. These studies can be termed synchronic when languages are grouped typologically or simply contrasted for the purposes mentioned above, according to (e.g.) their characteristics at a certain point in time. No reference is made to the history of the languages or possible historical relatedness. In contrast, De Saussure calls studies concerning the evolution of languages (e.g. Historical Linguistics) 'diachronic studies'. (Fisiak 1981:

1, James 1980: 2).

5.2 Interlingual Psychological Theories of Second Language Learning Processes That Underpin CA

Practitioners of CA (labelled 'Contrastivists' by James 1980: 1) base their

studies firmly on the behaviourist notion of 'Transfer': L1 habits predomi-

nate in the performance of the L2 until new language habits can be acquired

(Corder 1981: 5). Along with grammatical, lexical, syntactic and morpho-

logical (etc.) systems, the individual tends to transfer the whole L1 sound

system, including stress rhythm and intonation patterns (Lado 1957: 11-24).

It is also believed that elements of the L2 that are similar to the L1 will be easy to acquire and that those of a different nature will be difficult. For example, English [ b ] should not prove problematic to Japanese learners as there exists a sound in Japanese bearing the same broad phonetic descrip- tion (also transcribed as [ b ]). Japanese has no sounds corresponding to the descriptions of English [ , ] however, and that makes these sounds potential problem areas. This presupposes that before instruction the learner is ignorant of these easy and difficult items and that hence they are taught with equal intensity (James 1980: 24-25).

Corder (1981: 96) however, dismisses notions of "comparative difficulty". He states that the L2 learning process involves the restructuring of L1 habits and that the issue is the magnitude of the task, not the degree of difficulty.

This would hardly seem to make any difference to the end result but shows that there are various ways of interpreting the facts. It also provides a complementary aspect to the 'developmental' aspect of IL; the restructuring of language habits parallel the IL continuum from the L1 to the L2.

Sometimes L1 transfer has been seen to have a facilitating effect on

occasions when there is resemblance between the L1 and L2 systems. This

'positive transfer' has even been encouraged as a learning strategy but

unfortunately can result in unnatural L1-type sentences in the L2, e.g.,

when learners attempt to substitute English diphthongs with Japanese

'double vowels' (see 7.1, 7.4.1, 7.4.3). These are consequently viewed as

'negative transfer', i.e. erroneous. Corder (1981: 99) prefers to view this as a

case of "failure to facilitate" rather than facilitation or non-facilitation,

creating a dichotomy of 'facilitation' and 'zero effect' (James 1980: 144).

Numerous attempts have been made to redefine Transfer as James (1994:

180-185) reports, but they (as with Corder's comments above) merely serve to explain the L2 error in a different way. One of the most recent has been Cook's (1992) dichotomy of 'diachronic' and 'synchronic' beliefs. Diachronic transfer is 'transfer over time' whereas Synchronic transfer is 'transfer at a particular moment in time'.

⁵

The results of Transfer in the L2 performance of learners are termed 'Interference' (DiPietro 1971: 6). More recently, the Cognivist views of 'Ignorance' and 'Avoidance' have become more credible as explanations for errors. The degree of a learner's ignorance of the L2 preconditions interference, resulting in production through the use of whatever L2 knowledge can be mustered (James 1980: 22-25).

Duskova (1969) observed that 'ignorance-without-interference' can occur, when the learner does not know the correct L2 item and so of course will not use it in spontaneous speech. For example, a Japanese learner may not be aware that voiceless English [ ] also has a voiced counterpart [ ], especially as orthographically both sounds are represented as 'th'.

'interference-without-ignorance' also occurs when the learner, despite systematic drilling in class to an error-free standard, produces it incor- rectly minutes later! Alternatively, due to previous experience (or fear) of failure or greater difficulty, the learner may avoid use of the item altogether. This has been referred to as 'Avoidance Strategy' (Schachter 1974, Kleinman 1977).

５

This particular (re)definition of Transfer is of no particular relevance to this study

but illustrates that these concepts still very much play a part in contemporary

linguistic research and debate.

It would seem that Interference, if genuine, may have various causes and that the relation of Interference to error is not as simple as was first supposed. Later in this study, these notions will be related initially in a predictive capacity through comparison of English and Japanese systems.

5.3 The 'Strong' and 'Weak' Claims of CA

The original 'strong' claim of CA was that L1 Interference was the prime, if not sole, cause of L2 error due to the differences between the native and target languages: the greater the difference, the greater the learning difficulty and consequently the greater severity of error. CA was needed to predict these errors and determine what was needed to be taught in the L2 (Sridhar 1981: 211 citing Lee 1968). Even as far back as 1957 however, Lado called for "final validation" of such claims by checking them against the speech of students (Lado 1957: 72). The advent of Error Analysis (EA) from the late 1960s cast disappointing shadows on the CA 'strong claim'.

Research generally estimated that around

¹

/

3

-

¹

/

2

of errors could be interlingual with a further third intralingual. Some statistics from that period are listed below:

Results such as the above, although far from being conclusive, and in Interlingual Intralingual

Richards (1971) 53% 31%

Tran Thi Chau (1974) 51% 29%

Grauberg (1971) (L1 English learners of German) 36% Unknown

Mukattash (1977) (L1 Arabic learners of English) 23% Unknown

themselves rather unclear, have generally led to the more widely accepted 'weak' claim of CA; that CA can account or provide explanation for some aspects of learner performance a posteriori, i.e. after actual L2 errors have been made (Nemser 1971: 60-61, Fisiak 1981: 7, Sanders 1981: 22-23). CA, when applied predictively, could not after all account for all L2 performance errors as it did not take the actual performance of the learner into account at all. Even with the additional applications of intralingual theories and EA testing however, up to 20% of total errors were left unexplained. This problem still leaves the future of CA/EA open. Upon reflection, some of the failed predictions were too unsubtle. More detailed linguistic analysis, supported by more effective empirical testing, would have served purposes better.

This study attempts to draw its own conclusions concerning the validity of CA claims such as those above. CA-based interlingual hypotheses are first drawn in order to predict errors in Japanese learners phonetic skills, which should then in the future be tested empirically by an EA survey, then combined with intralingual theories in an explanatory capacity for the errors observed during testing. In other words, the 'weak claim' of CA is generally followed methodologically.

5.4 Developing Course Materials for Language Teaching

Fries (1945) was the first to endorse this potential application of CA,

stating that "the most effective materials (for teaching an L2) are those

that are based on a scientific description of the language to be learned,

carefully compared with a parallel description of the native language of the

learner." Attempts have since been made to convert descriptive data from CA (with or without the support of empirical testing) into teaching methods, programmes and course materials. It would seem that these have not always been successful. Furthermore, research has never proven whether such courses and materials are more effective than those based on different principles. The 'demotion' of the strong claim of CA to its weak claim has also largely deterred further ventures (Van Els et al 1977: 46).

6. Description of the Present Study 6.1 Aims and Purpose

Three main aims formed the basis of this study. These were to find out:

i. Whether certain English consonant sounds can be more or less troublesome to Japanese learners in their initial, medial or final positions.

ii. Which sound substitutions occur more or less often and in which syllable positions in the L2 English production of learners.

iii. Whether sequences of a given consonant and a selection of vowel sounds can be more or less problematic to learners than each other.

These aims will be later addressed (see 7.7) after the implementation of the

English - Japanese CA. Nilsen and Nilsen (1971) and Kenworthy (1987)

however make very little or no reference to the above issues. Thus, a

further aim to this study is to provide commentary on these points that

could be used as supplementary information to that already provided in the

existing works. It is hoped that this information may in turn be used by

EFL teachers in the instruction of Japanese learners' phonetic skills.

6.2 Selection of Test Items and Methodology

Twenty English consonant speech sounds and combinations of consonants and vowels were selected. These were:

i. [ f, v, , ] in initial, medial and final positions, including when orthographic 'r' is represented as [ ] in syllable final.

ii. [ w ] followed by [

I

,

I

, , i, ] iii. [ t

^h

] followed by [ i, uw ] iv. [ ] followed by [ ,

I

]

These sounds and sequences of sounds have been selected as they are listed by Nilsen and Nilsen (1971) and Kenworthy (1987), as well as regarded by many EFL teachers, as highly problematic to Japanese learners.

As mentioned in 5.3, the 'weak claim' of CA influenced the methodology

involved in this study. The general premise was held that CA has some

predictive power in the study of learners' L2 errors but ultimately, any

hypotheses drawn from a CA-based study must be empirically validated by

means of a future EA survey. A contrastive analysis was implemented, first

involving a description and comparison of the English and Japanese

syllabary systems. The English sounds listed above were listed and

described in articulatory (phonetic) terms along with the Japanese sounds

that most closely correspond. The sounds from both languages were then

examined and contrasted. Conjectures were drawn (based on the contrasts

in syllabic and articulatory description) as to which of the English sounds

are likely to prove most or least troublesome to native Japanese speakers.

7. A Contrastive Analysis of English and Japanese Speech Sounds

The study reported in this section is a CA of the English and Japanese syllable systems and 'single' speech sound units. The aim of this CA was to form hypotheses concerning learners' L2 English performance that could be used to address the issues outlined in 6.1. The sounds from both languages are represented in broad IPA transcription. Japanese words are ortho- graphically represented first in the Roman alphabet and then using the Hiragana syllabary system. All assumptions regarding difficulties for Japanese learners in English L2 production arising through differences between the two sound systems are based on the notions of L1 Transfer and Interference (see 5.2).

7.1 Terminology

The following terms appear in this section:

'Dialect': a manner of speaking, sharing pronunciations, words, expres- sions and grammatical constructions used more or less uniformly through- out an area or a group of speakers, which manner differs from those of other speakers of the same language. As a result, dialects are possibly mutually incomprehensible (Lado 1957: 22).

'Accent': a distinctive variation in pronunciation of the same linguistic

system. As only the pronunciation differs but grammatical, lexical,

syntactic (etc.) structures remain the same, accents are usually mutually

comprehensible.

'Phonotactics': the ways in which (consonant and vowel) sounds combine together in a particular language to form syllables and groups of syllables (O'Connor 1973: 229).

Japanese 'double vowels': two identical vowels in succession pronounced as a single 'long vowel' equal in length to two vowels of normal length, but still within the confines of a single syllable (Kenworthy 1987: 150).

'Sound Quality': any aurally distinguishable single component of the total auditory impression made by a given part of a phrase, e.g. a particular vowel colour, position, movement or manner of articulation, voicing, etc.

Qualities are identified by ear but defined in terms of their assumed production by the organs associated with speech (Bloch 1950: 89).

'Sound Quantity': the length, amplitude, energy involved in production (etc.); properties of a speech sound that could conceivably be measured.

7.2 Choices of Dialect

As a prerequisite to any CA, a specific dialect from both languages needs to be selected (Lado 1957: 23). The logical choices are those that are perceived on the whole as being 'standard' representations of the languages involved.

For the purposes of this study, Standard Southern Spoken British English

⁶

and 'Kokugo' based on the Edo Standard Conservative Spoken Japanese of Tokyo, were chosen. This variety of English has been selected as it is the most widely taught to foreign learners throughout Britain and featured in British English EFL textbooks. 'Kokugo' is generally taught as 'standard

６

Standard Southern Spoken British English could arguably be referred to as an

accent, rather than a dialect.

Japanese' and as part of the Japanese national curriculum. Throughout the CA, both dialects have influenced the selection of speech sounds for comparison in both languages. Bloch (1950: 87-88) pointed out long ago that the ever-growing number of loanwords in modern Japanese, particularly those taken from (in general, British and North American) English, has had a profound effect on the sound system of the language. His observation is even more valid now. He therefore urges the distinction of 'conservative' dialect, in which "English loanwords have been fully assimilated to the pronunciation of other words." In contrast, an 'Innovating' dialect contains special sound types and combinations (for foreign loanwords). It is spoken mainly by people with (e.g.) a good command of English.

7.3 A Methodological Framework for CA

Lado (1957: 12) proposes a three-stage standard framework by which a contrastive analysis of sound systems may be carried out. This framework was used throughout the present study:

1. Linguistic description and comparison of sound systems: the aim is to find or prepare an overall linguistic description of the sound systems of the two languages. The two systems are then contrasted and potentially problematic areas to the learner noted.

2. Description and comparison of sound units, to involve three checks:

i. Does the native language contain a sound similar in phonetic realisation?

ii. Are the variants of the speech sounds similar in both languages?

iii. Are the sounds and the variants similarly distributed?

3. Description of troublesome contrasts: based on the psychological concepts of L1 transfer, interference and comparative difficulty.

Articulatory differences between the sounds of the two languages will result in Interference in L2 performance.

7.4 Linguistic Description and Comparison of Sound Systems 7.4.1 Phonotactic Analysis

The first step in this stage of the CA framework involves an overall statement and comparison of the English and Japanese sound systems in terms of syllabic structure (i.e. a phonotactic analysis).

The most general simplified formula that can be given to possible sound sequences within an English syllable is (ccc)v(cccc)

⁷

. The vowel may occur alone, e.g. the Indefinite Article 'a' [ ], but may also be preceded by clusters of one to three consonants and followed by clusters of one to four.

Clusters after the vowel are more complex than those preceding. This is partly because of inflectional endings incurred through grammatical structuring, e.g. plurals, past tenses and ordinal numbers, otherwise there would be only three-consonant clusters. Due to grammatical complexity a total of 153 clusters could occur: 51 two consonant clusters and 7 three consonant clusters (O'Connor 1973: 200, 229-231). See fig. 1 for a full summary of Standard Southern Spoken British English based on the above model.

Japanese generally follows a far simpler (in comparison to English) syllabic

７

c = consonant, v = vowel.

system which can be listed in its entirety as below:

Note that there are restrictions in consonant-vowel distribution within these models: [ w ] only occurs before [ a ]; [ Φ ] only occurs before [ ] ; [ , , ] only occur before [

I

].

Presupposing that L1 Transfer is a genuine phenomenon, the above syllabic restrictions imply that certain sequences of English consonants and vowels can be problematic to Japanese learners; even those involving sounds that closely approximate Japanese sounds.

1. v e.g. 'e' え 'picture'

2. vv

⁸

'ii' いい 'good'

3. v(c) 'en' えん 'circle'

4. (c)v 'jidai' じだい 'era'

5. (c)v(c) 'shinjitsu' しんじつ 'fact'

6. (c)vv 'doushi' どうし 'verb'

7. vv(c)

⁹

accidental gap

8. (cc)v 'myaku' みゃく 'pulse'

9. (cc)v(c) 'o-kyan' おきゃん 'tomboy'

10. (cc)vv 'kyoudai' きょうだい 'brother, sister'

11. (cc)vv(c) accidental gap

８

vv = Japanese double vowels, articulated as a single syllable, not two.

９

vv(c) and (cc)vv(c) are conceivably possible but more likely accidental gaps. Japanese

double vowels (vv) occur as syllables in their own right but could theoretically be

followed by the generalised nasal 'ん' [ m, n, ] (see fig. 2) perhaps in foreign

loanwords assimilated to the pronunciation of 'Conservative Kokugo' (see 7.2).

(CCC)V(CCCC) 1.

Al l cons o nan ts o ccu r si ngl y b ef or e the vo w el exc ep t [ ]. [ ] isr a r eb u t a p p ea r si n r ec en t loanw o r d s, e. g. '

g

ig ol o' .

V O W E L

1.

A ll cons o nan ts ca n oc cur si ngl y a ft er a v owel . 2.

(i) (ii)

2-consonantclustersbeforeavowel(44inall):

[s ]+Ce .g . '

st

ay,

sw

im,

sl

ee p' et c. but n ot e. g . [ s ,s b, s ]+o th er s. C+[ w , j, ,l ] e. g .

tw

in ,

b

eau ty,

cr

ea m ,

pl

ai n et c. but n ot e. g . [ fw , j, h ,t l ] +o th er s. [ ] d o es n ot o ccur in final cluste rs. [v]o n ly o cc u r s w it h [ j ]a s in '

v

ie w' . [h ] o n ly o cc u r s w it h [j , w ]a s in '

h

ug e,

wh

ich .' [w , j, ] a re not fou nd as th e fi r st tw o cons o - nan ts ; n ei the r is [ l ] ex cep t for the p o ss ib le d is tin cti o n o f [ lj ] a s in '

l

ut e. '

10

2. (i) (ii) (iii) (iv)

2-consonantclustersafteravowel:

C + [ t,d ,s ,z ,] e. g . 'a

pt

,b e

gged

,s i

nce

,c le a

nse

'e tc . [l ] + C e. g . 'b u

lk

,e

lf

,b u

lge

,h e

lp

'e tc . [m , n , ]+C e. g . 'n y

mph

,p lu

nge

,i

nk

'e tc . C+ [ ]e .g . 'w i

dth

,f i

fth

'e tc .

N.B.3.3-consonantclustersafteravowel(69inall):

A n y o f the tw o -co nso n an t se q u enc es as ab ov e, p r ece d ed or fo ll owe d b y anot he r con son a nt , e. g .: [p t ] +[ s ] a s in 'c r y

pts

'. [ n z ] + [ d ] a s in 'cle a

nsed

'. [k s ] + [ ]a s in 's i

xth

'.

3.3-consonantclustersbeforeavowel:

[s ]=f ir stc o n so n a n t. [ p ,t ,k ,f ] = m id d le co n so n a n t. [w , j, ,l ] = la st co n so n a n t. e. g. '

spl

as h,

st

ew ,

sq

ua r e,

sph

rag id .'

4. (i) (ii)

4-consonantclustersafteravowel(8inall):

3 -c o ns onan t cl u st er s a s a bov e + [s ] (7 ca se s) e. g . 't w e

lfths

,e x e

mpts

,t e

xts

'e tc . [z ] (1 ca se ): 'g li

mpsed

'.

Fig.1:

Ta bl e sum m a r isi ng th e sy ll a bi c st r u ct u r e o f St an dar d Sout h er n S p ok en B r it is h E ngl is h , b a se d o n th e m o de l (ccc) v (cc cc) (O 'C o nno r 197 3: 229 -2 31) .

10

[ h w ] and [ lj ] a r e m o r e fe at ur es of RP th an St an dar d So ut he r n Sp oke n Br it is h E ng li sh .

Fig. 2:

Chart of English consonant sounds (adopted from Ladefoged 1975: 33)

¹²

Fig. 3:

Chart of Japanese consonant sounds (adopted from Bloch 1950: 107)

11

[ w] is shown in two places on the chart as it is articulated both with a narrowing of the lip aperture (bilabial) and the back of the tongue towards the soft palate (velar).

12

[ h ] does not appear on the chart as it is merely the voiceless counterpart of the following vowel, e.g. 'hat' is correctly transcribed as [ at ].

Manner of Articulation bilabial labio-dental inter-dental alveolar palatal-alveolar palatal prevelar (front) medio-velar (back) glottal

Place of Articulation

nasal stop

(central) fricative affricate lateral fricative (central) approximant lateral approximant flap semi-vowel

m n Č

p^h t^h k^h Ɲ

p b t d k Ū

f vθ s zƌ ƛ

ts ư ƭ

Ŵ

(w)¹¹ Ƅ j w

l

Manner of Articulation bilabial labio-dental inter-dental alveolar palatal-alveolar palatal prevelar (front) medio-velar (back) glottal

Place of Articulation

nasal stop (central) fricative affricate lateral fricative (central) approximant lateral approximant flap semi-vowel

m n Č

p b t d k Ū Ɲ

ě s zƌ ƛ h

ts ư ƭ

Ɔ

j w

Fig. 4:

Chart of English vowel sounds. N.B. The sounds [ , a, ] occur as the first elements of diphthongs (Ladefoged 1975: 34).

Fig. 5:

Chart of Japanese vowel sounds (Ladefoged 1975: 200).

back

lax tense

lax tense

lax central

front

high

mid

low

tense i

I

e ƛ

Ť

a

ƕ Ţ

u

Ɠ

ŝ

ś

back

lax tense

lax tense

lax central

front

high

mid

low

tense

i

e o

a

For example, combinations of English [ w, , , ] with [ e, ], and other sequences such as [ t, s, z, d ] with [

I

].

Initial two-consonant clusters in Japanese, as in (cc)v, comprise of c + [ j ] as in 'byouin' (びょういん: 'hospital'). Consonant sounds succeeding the vowel are few but occur in words such as 'des(u)' (です: 'be'). Note also that the aforementioned 'byouin' features a final nasal consonant [ n ]. In Japanese orthography, such 'closed syllables' do not exist but are repre- sented as separate syllables hence the bracketed 'u' in 'des(u)'. Final [ n ] (actually a generalised nasal, often transcribed as [ n ], [ m ] or conceivably [ ]) is also represented orthographically as a syllable in its own right.

Some native Japanese speakers however, transfer these orthographic conventions to their Japanese pronunciation, which is in turn perceived as correct in (e.g.) some formal Japanese situations and when reading poetry.

When speaking English, Japanese learners tend to insert vowels between consonants (to break up clusters) and at the end of final consonants (to eradicate closed syllables) due to transfer of the generally cvcv-sequenced Japanese syllable system. This habit can result in the production of too many syllables.

The implications of transfer of some prosodic features that differ between English and Japanese should also be noted:

i) English is a stress-timed language whereas Japanese is syllable-timed.

Japanese learners often misplace 'heavy stress' on syllables in English L2 production. This can result in, for example, the reduced vowel [ ] substituted with 'full vowels' such as [ a, ] etc.

ii) Higher pitched intonation patterns in English indicate politeness

whereas they indicate femininity in Japanese. Male learners have been

known to comment on 'feeling feminine' when expressing politeness in English (Kenworthy 1987: 151-152).

7.4.2 Description and Comparison of Consonant Systems

Fig. 2 and Fig. 3 display charts of the English and Japanese consonant systems respectively presented in the standard manner following two axis.

The categories of place of articulation are arranged along the top of the chart, starting from the right (glottal) and finishing on the left (bilabial).

This represents the order in which an egressive air-stream would meet each place of articulation. The categories of manner of articulation are listed from top to bottom. A stricture

¹³

of complete closure (i.e. the greatest degree of restriction of the air-stream through the vocal tract by the active and passive articulators) is listed at the top and an 'open' stricture at the bottom. Voiced units (therefore lenis) are placed on the right and voiceless units (fortis) on the left of each cell in the charts (Abercrombie 1967: 54, Ladefoged 1975: 33).

The sounds [ f, v, , , ] occur in English but not in Japanese. Learners will tend to substitute [ f ] with [ Φ ], [ v ] with [ b ], [ ] with [ z ] or [ d ] and [ ] with [ s ] or [ t ]. Similarly, English [ l ] and [ ] will be substituted with [ ] or a sound approximating [ d ] to the English ear. The aspirated [ t

^h

, p

^h

, k

^h

] can be produced (in error) with aspiration when syllables medial and final (medial and final [ t, p, k ] are not aspirated in English). In the case of [ t

^h

] for example, this can result in production of a sound resembling [ ]

13A 'stricture' is the position taken up by the active articulator in relation to the passive one (Abercrombie 1967: 44).

to the English listener (Kenworthy 1987: 149-150).

7.4.3 Description and Comparison of Vowel Systems

Fig. 4 and fig. 5 display the complex system of vowel sounds that occur in English and the five vowels of Japanese respectively. The charts follow the standard presentation of vowels in two dimensions. The horizontal axis of the charts indicate the horizontal position of the tongue; front, central and back. The vertical axis represents the height of the tongue; high, mid and low. Sounds shown at the top of each cell in the charts may be termed 'tense', those at the bottom 'lax'. No indication is made as to the degree of lip rounding although generally front vowels may be regarded as unrounded/spread and back vowels rounded. The degree of spreadedness and roundedness increases from low to high positions. The degree of vowel length is not indicated (Ladefoged 1975: 34).

As Japanese only has a five-vowel system, learners can have difficulty with the much larger English vowel system. Sounds in English which do not occur in Japanese such as [

I

] may be substituted with [ i ], [ ] with [ ] and so on. [ ] also does not occur in Japanese. The syllable in which it occurs is often overstressed by learners in English L2 production (see 7.4.1), possibly becoming substituted with any of the five Japanese vowels.

English diphthongs can also prove problematic to learners. However,

transfer of Japanese 2-vowel sequences in place of English diphthongs can

be used as a learning strategy (Kenworthy 1987: 150). For example, [

I

] as

in 'gate' may be substituted with Japanese [ ei ] as in 'eiga' (えいが：'movie,

film') as an initial step in the mastery of [

I

]. Also, the vowel sequence in

'Hai' (はい：'yes') [ ai ] may be said to approximate the diphthong [ a

I

] as in 'Kite' ('Hai' is usually articulated as a single syllable although ortho- graphically represented as two syllables). Positive transfer of this Japanese vowel sequence may therefore facilitate production of the English diph- thong. Care must be taken not to allow these strategies to become habitual or fossilise in English production however.

7.5 Description and Comparison of Sound Units

The second phase of the CA framework was carried out in accordance with the three checks outlined in 7.3. The English and Japanese speech sounds chosen for analysis are described and contrasted side-by-side. Comments are made on articulatory differences between the sounds after the descrip- tions of each pair. English and Japanese words containing these sounds are also given as examples of their distribution.

For descriptions of the English sounds, texts by Abercrombie (1967), Brosnahan and Malmberg (1970), O'Connor (1973) and Ladefoged (1975) have been referred to. Descriptions of the Japanese sounds are based on studies by Bloch (1950) and Miller (1967). The phonetic terminology used in this study reflects that used in the above texts.

English Japanese

1. Initial [ f ] as in 'funny, fault'.

2. Final [ f ] as in 'thief, Jeff'.

= voiceless labiodental fricative (longer in syllable final.)

[ Φ ] as in 'furui' (ふるい : 'old') or 'kangofu' (かんごふ : 'nurse').

= voiceless bilabial fricative.

The difference lies in the positive articulator (i.e. place of articulation): [ f ] involves the upper teeth, [ Φ ] the upper lip.

The differences are in place of articulation (the passive articulator): [ v ] uses the upper teeth, [ b ] the upper lip. Also the manner of articulation:

[ v ] is a fricative, [ b ] is a stop.

The place of articulation differs: [ ] involves the tongue protruding between, or simply behind, the upper and lower teeth but with [ d ] the tongue tip touches the upper teeth and alveolar ridge. Also the manner of articulation: [ ] is a fricative, [ d ] is a stop.

3. Initial [ v ] as in 'vowels'.

4. Medial [ v ] as in 'rover'.

5. Final [ v ] as in 'curve'.

= voiced labiodental fricative (only 'partially voiced'

¹⁴

and longer in syllable final).

[ b ] as in 'basho' (ばしょ: 'place') or 'konban' (こんばん : 'tonight').

= voiced bilabial stop (plosive).

6. Initial [ ] as in 'those', then'.

7. Medial [ ] as in 'heather, breathing'.

8. Final [ ] as in 'loathe, smooth'.

= voiced inter-dental fricative (only 'partially voiced' and longer in syllable final).

[ d ] as in 'doko' (どこ : 'where') or 'fude' (ふで : 'writing brush').

= voiced apico-denti-alveolar stop (plosive).

14The term 'partially voiced' is used for convenience here and throughout when in fact there is a reduction of the energy involved in articulation, which accounts for an apparent reduction in voicing.

In manner of articulation, [ ] is an approximant where the tongue is centralised. [ ] is a flap where the tongue is not centralised but extends towards and strikes the alveolar ridge.

There would appear to be very little difference in phonetic description between these pairs of sound sequences. Both English sequences however, involve the reduced vowel [ ] as a lightly stressed second syllable in [ a

I

] and the second part of a diphthong in [ ]. In contrast both Japanese sequences consist of two equally stressed syllables, where a fully articu- lated [ a ] takes the place of [ ].

9. Initial [ ] as in 'right'.

10. Medial [ ] as in 'pirate'.

= voiced lamino-alveolar (central) approximant.

[ ] as in 'rakuda' (らくだ: 'camel') or 'kore' (これ: 'this').

= voiced lamino-alveolar flap.

11. Final [ ] as in 'fire'.

= [ a

I

] low (moving to) high front unrounded vowel

+ [ (j) ] lax mid central vowel.

[ aja ] as in 'ayamaru' (あやまる:

'apologise').

= [ a ] lax low back unrounded vowel

+ [ j ] prevelar front vowel (semi- vowel) or 'yod' + [ a ] as above.

(or) Final [ ] as in 'bore'

= [ ] lax mid back rounded vowel

+ [ ] as above.

[ oa ] as in 'yoake' ( よ あ け : 'dawn').

= [ o ] mid back very rounded vowel

+ [ a ] as above.

Although [ w ] and [ ] are different in phonetic description (consonant and vowel respectively), in both cases the tongue is in the same position ([ u ]). [ ] therefore differs only in that it involves weaker lip-rounding.

[

I

] and [ i ] vary only in muscular tension. The main difference lies in syllabic structure: [ w

I

] occupies a single syllable but [ i ] is a two-syllable sequence.

Again, both sequences closely resemble each other in phonetic description but differ in syllabic structure. [ w

I

] is articulated as a single syllable [ ei ] potentially as three.

12. [ w

I

] as in 'witch', win'.

= [ w ] voiced bilabial approximant + [

I

] (checked) lax high front unrounded vowel.

[ i ] as in 'nuide' (ぬいで: 'taking off clothes').

= [ ] slightly tense high back weakly rounded vowel + [ i ] very tense high front unrounded vowel.

13. [ w

I

] as in 'wait'.

= [ w ] as above

+ [

I

] mid (moving to) high front unrounded vowel.

[ ei ] does not occur in Japanese but [ ei ] occurs in 'eisei' (えいせ い: 'sanitation').

= [ ] as above

+ [ e ] slightly lax mid front unrounded vowel + [ i ] as above.

14. [ w ] as in 'wood'

= [ w ] as above

+ [ ] lax high back rounded

[ ] as in 'uo' (うお: 'fish').

= as above.

Both sounds in this English sequence become substituted with the single Japanese vowel sound [ ].

[ i ] and [ ii ] vary only in muscular tension (Japanese double vowels closely resemble English diphthongs in length and positioning). Syllabic structure is the main difference: [ wi ] spans one syllable, [ ii ] two syllables.

[ ] and [ oo ] differ mainly in muscular tension and lip-roundedness, and insignificantly in positioning and length (see 15). The same syllabic difference is found in 12 and 15.

vowel.

15. [ wi ] as in 'wheat, weeds'.

= [ w ] as above

+ [ i ] tense high front unrounded vowel.

[ i ] (does not occur in Japa- nese.)

= [ ] as above

+ [ ii ] as above (doubled).

16. [ w ] as in 'warder'.

= [ w ] as above

+ [ ] lax mid back rounded vowel.

[ oo ]

¹⁵

(does not occur in Japa- nese.)

= [ ] as above

+ [ oo ] slightly lax mid back very rounded vowel (doubled).

17. [ t

^h

i ] as in 'team'.

= [ t

^h

] aspirated voiceless apico-

[ ii ] does not occur in Japanese but

15[ oo ] is an example of a Japanese double vowel, occupying a single syllable (see 7.4.1 / 7.4.3).

[ t

^h

] and [ ] differ in both place and manner of articulation. However, [ ] is the nearest post-dental sound to [ t

^h

] as the friction in [ ] approximates the aspiration in [ t

^h

]. [ i ] and [ ii ] vary only slightly in muscular tension.

[ t

^ｈ

] and [ ts ] share a common place of articulation but unlike [ ], [ ts ] has no quality that can be identified with the aspiration in [ t

^h

]. [ ] resembles the diphthong [ uw ] in positioning but is shorter and involves less muscular tension.

[ ] and [ s ] vary in place of articulation. [ ] is slightly laxer than [ e ] but denti-alveolar stop (plosive)

+ [ i ] as above.

[ i ] occurs as in 'kuchi' (くち:

'mouth').

= [ ] voiceless lamino-palatal affricate

+ [ ii ] as above (doubled).

18. [ t

^h

uw ] as in 'two'.

= [ t

^h

] as above

+ [ u(w) ] tense high back rounded vowel.

[ ts ] as in 'atsusa' ( あ つ さ : 'heat').

= [ ts ] voiceless apico-denti- alveolar affricate

+ [ ] as above.

19. [ ] as in 'shed', shell'.

= [ ] voiceless lamino-palato alveolar fricative

+ [ ] lax mid front unrounded vowel.

[ se ] as in 'binsen' (び ん せ ん : 'writing paper').

= [ s ] voiceless apico-alveolar fricative

+ [ e ] slightly lax mid front

unrounded vowel.

otherwise bears close resemblance.

The two sequences vary little in phonetic description but significantly in syllabic structure: [

I

] spans a single syllable whereas [ sei ] spans two.

7.6 Summary of Troublesome Contrasts

Based on the observations made during 7.5, it can be concluded that Japanese learners may experience the following problems when producing the English speech sounds and sound sequences examined. These conclu- sions in turn lead to the CA-based hypotheses stated in 7.7.

7.6.1 Contrasts in Phonetic Description

From phonetic description alone, no evidence can be found to suggest that any of the English fricatives [ f, v, ] or the approximant [ ] would prove more or less troublesome to Japanese learners when syllable initial or medial. Problems could be experienced with [ f, v, ] when syllable final however. Japanese has relatively fewer post-vocalic consonants than English (see 7.4.1); for example, Japanese [ Φ ] cannot occur in syllable final, only before [ ] as in 'kangofu' (かんごふ: 'nurse'). Consequently,

20. [

I

] as in 'shave'.

= [ ] as above

+ [

I

] mid (moving to) high front unrounded vowel.

[ sei ] as in 'eisei' (えいせい: 'sani- tation').

= [ s ] as above

+ [ e ] as above

+ [ i ] as above.

Japanese learners may add extra vowels after final [ f, v, ]. In absolute final position fricatives such as the above tend to be followed by a short parasitic vowel anyway but transfer of final vowels may result in an exaggeration of this tendency, particularly as a slight lengthening of these sounds is required without breaking into a distinct final vowel.

Alternatively, problems may arise with respect to the reduction in energy that accounts for a reduction in voicing that [ v, ] undergo when syllable final. Learners could also exaggerate this feature, resulting in the production of [ f ] in place of [ v ] and [ ] in the case of [ ]

¹⁶

.

It is difficult to comment as to which sequence in the cases of [ t

^h

i ] or [ t

^h

uw ] could be the most troublesome. Their Japanese counterpart sequences contain either a quality approximating aspiration (in the case of [ t

^h

i ]) or share a common place of articulation (in the case of [ t

^h

uw ]) but not both. Possibly the difference in degree of difficulty between [ t

^h

i ] and [ t

^h

uw ] is negligible.

7.6.2 Contrasts in Syllabic Structure

Transfer of the (syllable-timed) Japanese syllabary system to the English sound sequences analysed in 7.5 would appear to be a major potential problem area, for example:

(i) Final [ ] in both [ a

I

] and [ ] would become a separate and distinct syllable, substituted with the 'full' vowel [ a ] due to stressing equal to that

16The possibility of 'spelling pronunication' error should not be overlooked [ ] and [ ] are both orthographically represented in English as 'th', e.g. 'mouth' [ ] and 'smooth' [ ].

of the preceding syllable.

(ii) Of all five [ w ] + vowels sequences, [ we

I

] is potentially the most problematic as it could conceivably be articulated as three distinct syllables.

Second in succession of difficulty would be [ w

I

], [ wi ] and [ w ], articulated as two-syllable sequences. In terms of syllabic interference [ w ], should be the least problematic (one syllable) but in terms of contrast in phonetic description with its Japanese counterpart, the most problematic.

(iii) Syllable interference would make the sequence [

I

] more troublesome than [ ] as learners could articulate [

I

] as two equally stressed syllables. Phonetic description reveals potentially greater problems in the case of [ ] than [

I

] however.

7.7 Hypotheses Drawn From the Conclusions of The CA

From the discussion in 7.6 it is possible to formulate hypotheses regarding the aforementioned speech sounds in the English of Japanese learners:

(i) Initial and medial [ f, v, , ]: no difference in degree of difficulty will be experienced when these consonant sounds occur in syllables initial or medial. This would seem logical as after all, 'medial' tends to mean 'initial in the next syllable'.

(ii) Final [ f, v, ]: the most problems will be experienced when these sounds occur in syllable final.

(iii) Final [ ]: syllabic interference will cause the overstressing of syllables containing this sound resulting in its substitution with a 'full' vowel.

(iv) [ w ] + vowel: syllabic interference will lead to more frequent error in

the sequence [ we

I

] than [ wi ], [ w ] or [ w

I

]. Whether [ w ] is the most or least problematic remains inconclusive.

(v) [ ] + vowel: syllabic interference will cause more problems with the sequence [

I

] than [ ].

(vi) [ t

^ｈ

] + vowel: there will be negligible difference in the difficulty experienced between the sequences [ t

^h

i ] and [ t

^h

uw ].

7.8 Relation of CA-Based Hypotheses to the Aims of the Present Study

Some of the hypotheses stated in 7.7 may be related to the aims of this study (see 6.1):

(i) [ f, v, ]: the most problems will be experienced by Japanese learners when these consonants occur in syllable final than when initial or medial.

Whether there is potentially any difference in difficulty between initial or medial positioning remains inconclusive however.

(ii) Syllabic interference will be the main cause of difficulty between the consonant-vowel sequences examined. Unfortunately, a definite hierarchy of difficulty of sequences of a given consonant and various vowels cannot be drawn.

(iii) No conclusions can be drawn based on the CA alone concerning the

issue of which sound substitutions can occur more or less often and in

which syllable positions in the cases of the consonants examined.

7.9 Conclusions

From this CA alone it has been impossible to address all of the aims of this study, as originally stated in 6.1, as well as comment on whether more problems would be experienced in reception or production of the speech sounds and sound sequences examined. This is more than likely because the actual L2 performance of native Japanese learners of English has not been taken into account (a general drawback of CA; see 5.3).

The validity of the hypotheses listed in 7.7 should therefore be tested by means of an EA survey, the results and conclusions of which could be used to address the issues and aims of the present study left inconclusive by the CA.

Bibliography

Abercrombie, D. (1965), Studies in Phonetics and Linguistics. Oxford University Press.

Abercrombie, D. (1967), Elements of General Phonetics. Edinburgh Univer- sity Press.

Bloch, B. (1950), "Studies in Colloquial Japanese IV: Phonemics". Language 26: 86-125.

Brosnahan, L. F. and Malmberg, B. (1970), Introduction to Phonetics.

Cambridge University Press.

Corder, S. Pit (1981), Error Analysis and Interlanguage. Oxford University

Press.

DiPietro R. J. (1971), Language Structures in Contrast. Newbury House Publications.

Fisiak, J. E., Takagi, N. and Mann, V. (1995), "Japanese can Learn to Produce English [ ] and [ l ] Accurately." Language and Speech 38 (1):

22-55.

Flege, J. (1981), "Some Introductory Notes Concerning Contrastive Linguis- tics." In Fisiak, J. (ed.) (1981), Contrastive Linguistics and the Language Teacher. Pergamon Press Ltd.

Gimson, A. C. (1962), An Introduction to the Pronunciation of English.

Edward Arnold Ltd.

Henly, E. and Sheldon, A. (1986), "Duration and Context Effects on the perception of English [ ] and [ l ]: A Comparison of Cantonese and Japanese Speakers". Language Learning 36, 4: 505-521.

James, C. (1980), Contrastive Analysis. Longman Group Ltd.

James, C. (1994), "Don't Shoot My Dodo: On the Resilience of Contrastive and Error Analysis". IRAL XXXII/3: 179-200.

Kenworthy, J. (1987), Teaching English Pronunciation. Longman.

Ladefoged, P. (1975), A Course in Phonetics. Harcourt Brace Jovanovich, Inc.

Lado, R. (1957), Linguistics Across Cultures (Applied Linguistics for Language Teachers). University of Michigan Press.

Miller, R. A. (1967), The Japanese Language. University of Chicago Press.

Miura, I. (1996), "Discrimination of Segmental and Suprasegmental Phones by Japanese Students Learning English from an Early Age." IRAL XXXIV/2: 135-143.

Mulder, J. W. F. (1974), "Descriptive Adequacy in Phonology and the Vowel

Phonemes of the Scottish Dialects of Angus and Perthshire Compared with

the Southern English System." La Linguistique 1974/1/.

Mulder, J. W. F. (1978), "Phoneme Tables and the Functional Principle". La Linguistique 1978/1.

Nemser, W. (1971), "Approximative Systems of Foreign Language Learners."

In Richards, J. (ed.), (1974), Error Analysis: Perspectives on Second Language Acquisition. Longman Group Ltd.

Nilsen, D. L. F. and Nilsen, A. P. (1971), Pronunciation Contrasts in English. Regents Publishing Co. Inc.

O'Connor, J. D. (1973), Phonetics. Pelican Books.

Richards, J. and Sampson, P. (1974), "The Study of Learner English". In Richards, J. (ed.), (1974), Error Analysis: Perspectives on Second Language Acquisition. Longman Group Ltd.

Sanders, C. (1981), "Recent Developments in Constrastive Analysis and their Relevance to Language Teaching". In Fisiak, J. (ed.), (1981), Contrastive Linguistics and the Language Teacher. Pergamon Press Ltd.

Sridhar, S. N. (1981), "Contrastive Analysis, Error Analysis and Interlanguage: Three Phases of One Goal". In Fisiak, J. (ed.), (1981), Contrastive Linguistics and the Language Teacher. Pergamon Press Ltd.

Van Els, T. et al (1977), Applied Linguistics and the Learning and Teaching

of Foreign Languages. Edward Arnold.