PERCEPTIONS OF THE SERIOUSNESS OF MISPRONUNCIATIONS OF ENGLISH SPEECH SOUNDS

The present study attempts to investigate Indonesian EFL teachers and native English speakers perceptions of mispronunciations of English sounds by Indonesian EFL learners. For this purpose, a paper-form questionnaire consisting of 32 target mispronunciations was distributed to Indonesian secondary school teachers of English and also to native English speakers. An analysis of the respondents perceptions has discovered that 14 out of the 32 target mispronunciations are pedagogically significant in pronunciation instruction. A further analysis of the reasons for these major mispronunciations has reconfirmed the prevalence of interference of learners native language in their English pronunciation as a major cause of mispronunciations. It has also revealed Indonesian EFL teachers tendency to overestimate the seriousness of their learners pronunciations. Based on these findings, the study makes suggestions for better English pronunciation teaching in Indonesia or other EFL countries.

1996; Dalton & Seidlhofer, 1994;Fraser, 1999;Macdonald, 2002).In Indonesian context, however, English pronunciation has rather been neglected.For example, the teaching of the English pronunciation tends to have an insufficient portion.As a result, Indonesian learners tend to make a considerable number of mistakes in pronunciation when they try to speak in English.Those mispronunciations will decrease the intelligibility of Indonesian learners speech to a considerable degree.Needless to say, this will hamper the flow of oral communication.It is necessary, therefore, for EFL teachers to properly deal with learners mispronunciations in their pronunciation teaching.However, very little information is available for Indonesian teachers about learners mispronunciations, more specifically about what mispronunciations are pedagogically significant in terms of frequency and seriousness in real-life communication.Therefore, it is quite probable that Indonesian teachers will deal with their learners mispronunciation solely on the basis of their personal intuitions.
In order to provide such missing information concerning learners mispronunciations, there is a need to conduct a study which would explore Indonesian EFL teachers perceptions of learners mispronunciations and compare them with those of native English speakers so that general tendencies and discrepancies in perceptions between the two groups can be identified.On the basis of this reasoning, the following research questions are set up: (1) What kinds of mispronunciations of English speech sounds are common and serious among Indonesian EFL learners?; (2) How do Indonesian EFL teachers perceptions of the seriousness of mispronunciations differ from those of native English speakers ?; and (3) What kinds of mispronunciations of English sounds are pedagogically significant?

METHOD Participants
A total of 80 respondents, 50 Indonesian EFL secondary school teachers and 30 native English speakers, voluntarily participated in the study.Of the 50 Indonesian respondents, 26 (20 males, 6 females) were senior high school teachers and 24 (15 males, 9 females) were junior high school teachers.These teachers had teaching experiences of various length: 7 teachers with 1-5 years of teaching experience; 11 teachers with 6-10 years of teaching experience; 18 teachers with 11-15 years of teaching experience; and 14 teachers with more than 15 years of teaching experience.Thirty-six teachers held a bachelor s degree in English language education while 14 held a diploma certificate (all were junior high school teachers).The English native speakers (18 males, 12 females) included 3 Americans, 3 British, 1 Irish, and 23 Canadians.

Instrument
To answer questions raised in the study, a paper-form questionnaire was prepared.The questionnaire is composed of 32 target utterances which are presumed to have been produced by Indonesian EFL learners.Each utterance is composed of two sentences with a target mispronunciation in the second sentence.These 32 target mispronunciations cover 19 consonants and 13 vowels all are segmental which were selected on the basis of the literature review and the first author s experience as an EFL teacher at senior high schools in Indonesia.Each utterance investigated is accompanied by a pre-coded five-point Likert-scale for the judgment of the frequency of the target mispronunciations and their seriousness, from 1 ( very rare for frequency; not serious for seriousness) to 5 ( very common for frequency; very serious for seriousness).
The questionnaire addresses the frequency and seriousness of mispronunciations.As far as the frequency of learners mispronunciations is concerned, we decided to focus on Indonesian EFL teachers perception since they encounter learners mispronunciations in their language classrooms on a daily basis.As far as the seriousness of learners mispronunciations is concerned, we decided that Indonesian EFL teachers perceptions should be qualified by native English speakers perceptions.This is partly because Indonesian EFL teachers perceptions are not always dependable due to their lack of experience of cross-cultural oral communication in English and due to their preoccupation with linguistic differences, and partly because we are interested in finding out how their perceptions of the seriousness differ from native English speakers perceptions.The assessors evaluation was based on their own intuitive perceptions.The details of the questionnaire are indicated in Appendix.

Collection
The questionnaire was distributed to Indonesian secondary school teachers of English and also to native English speakers.The questionnaire was designed to be anonymous and unregistered so that the respondents could give their hon-est opinion.Different ways were used to collect data from the participants.For the Indonesian respondents the original copy of the questionnaire was sent by e-mail attachment to the coordinator of this study in Indonesia, who printed it out, photocopied, and distributed the copies directly (in person, not by mail) to the respondents of the study.Two weeks later, the distributed questionnaire sheets were collected by the coordinator, packed, and sent back to the present researchers.For the native English speakers the copy of the questionnaire was either directly handed to the respondents living in Japan or sent by e-mail to the coordinator in Canada, who printed it out and handed the copies to the respondents.The retrieved questionnaire sheets were sorted and only the valid questionnaire sheets with complete answers were analysed.

Analysis
The tabulated scores of the frequency were averaged for each target mispronunciation and served as the basis for dichotomising the target mispronunciations into common or uncommon groups and into serious or unserious groups.Those target mispronunciations whose scores were at least the same as the median (i.e., 3.00) in terms of the frequency were classified as common and those whose scores were below the median were classified as uncommon.The same procedure was adopted for splitting the investigated mispronunciations into serious and unserious groups.Furthermore, the calculated mean scores were mapped out onto a matrix of the frequency and seriousness in order to find out which mispronunciations were the most significant for EFL learners.In addition to the mean-median comparison, the data was submitted to a Wilcoxon Mann-Whitney test to determine whether there were significant differences in the perceptions of the seriousness of the mispronunciations between the Indonesian respondents and the native English speakers.

FINDINGS Descriptive Statistics
Table 1 presents the mean scores of all collected data on the Indonesian EFL teachers perceptions of the seriousness of the target mispronunciations and their frequencies, the native English speakers perceptions of their serious-ness, and the difference in the mean scores of the perceptions of the seriousness between the Indonesian EFL teachers and native English speakers.

Common Mispronunciations
As mentioned previously, learners mispronunciations are regarded as common in this paper if their mean scores are at least the same as the median (i.e., 3.00).As shown in Figure 1, out of the 32 target mispronunciations (19 consonants and 13 vowels), 20 items (63%) are regarded as common and 12 items (37%) as uncommon.Those 20 common mispronunciations involve 12 consonants (63% of the target consonants) and eight vowels (67% of the target vowels) whereas those 12 uncommon mispronunciations involve seven consonants (37% of the target consonants) and five vowels (33% of the target vowels).
Concerning the 20 common mispronunciations, 12 cases involve consonant mispronunciations and eight vowel mispronunciations.Of the 12 consonant mispronunciations, 11 cases involve fricatives and one case involves a plosive.Of the 11 fricative mispronunciations, five cases (M5, M6, M7, M9, and M12) feature the replacement with learners L1 plosives, four cases (M8, M14, M16, and M17) feature the replacement with other English fricative consonants which do not exist in learners L1, and two cases (M15 and M18) feature the replacement with other consonants which exist both in English and learners L1 (/ / /s/ and / / / /).
The figure below shows the respondent s perception of the frequency of the mispronunciation.Of those eight common vowels, three cases (M20, M21, and M23) involve front vowels, two cases (M26 and M27) involve back vowels, and three cases (M29, M31, and M32) involve diphthongs.M23 features the replacement of a front open-mid lax vowel /? / with a front close-mid lax vowel / / while M20 and M21 show the interchangeable replacement of /i / and / /.In terms of back vowels, M26 and M27 also show the interchangeable replacement of /u / and / /.The three cases of diphthong mispronunciation all feature the substitution with the close variants of vowels in learners native language.

Figure 2. Respondents perceptions of the seriousness of the target mispronunciations
Of the 22 serious mispronunciations endorsed by the Indonesian respondents, 14 cases involve consonants (74% of the target consonants), and eight cases involve vowels (62% of the target vowels).Of these 14 serious mispronunciations involving consonants, 11 cases feature the mispronunciations of fricatives which are replaced either with the plosive consonants in learners L1 (e.g., /v/ is mispronounced either /p/ or /b/) or with the similar fricative sounds in their L1 (i.e., / / and / / are mispronounced as /s/).

Categorisation of Mispronunciations
As shown in Figure 3, a combination of the English native speakers perceptions of the seriousness of mispronunciations and the Indonesian teachers perceptions of their frequency has led to the classification of the 32 target mispronunciations into the following four groups: a. Group 1: eight mispronunciations (25% of the 32 target mispronunciations) which are serious and common, involving four consonants (M5

Significant Difference in the Perceptions of the Seriousness
As presented in Table 1, a Wilcoxon Mann Whitney test has disclosed that there exists a significant difference in the perceptions of the seriousness between the Indonesian EFL teachers and the native English speakers for 20 target mispronunciations (63% of the 32 target mispronunciations).
Of these 20 mispronunciations, four mispronunciations (all are consonants) were perceived as serious by both groups of the respondents; 10 mispronunciations (6 consonants and 4 vowels) were perceived as serious by EFL teachers alone, four mispronunciations (1 consonant and 3 vowels) were perceived as serious by native English speakers alone, and two mispronunciations (1 consonant and 1 vowel) were rated as unserious by both groups of the respondents.Of those four mispronunciations which were rated as serious by both groups, there was only one mispronunciation (i.e., /b/ /p/) whose mean score of the native English speakers was larger than that of the Indonesian EFL teachers.

DISCUSSION
The study was designed to explore and compare the Indonesian EFL teachers and native English speakers perceptions of the mispronunciations of English sounds focusing on the frequency of those mispronunciations and their seriousness, the significant differences in the perceptions between the Indonesian EFL teachers and the native English speakers, and the pedagogically most significant mispronunciations in EFL classrooms.
As far as the frequency of mispronunciations is concerned, the result of the study has indicated that 20 mispronunciations (12 consonants and 8 vowels) were perceived as common by the Indonesian EFL teachers.This implies that Indonesian EFL learners are still facing a problem with pronunciation although this indicator inarguably needs to be qualified with the seriousness of the common mispronunciations.
Concerning serious mispronunciations, there is a tendency for Indonesian EFL teachers to overestimate the seriousness of learners mispronunciations.One possible reason for this overestimation is that Indonesian EFL teachers lack experiences of interacting with English speaking people.As a result, they do not feel confident of accessing the seriousness of the target mispronunciations.Another reason is that Indonesian EFL teachers tend to perceive the seriousness of learners mispronunciations on the basis of their limited knowledge about English phonology mainly obtained from lectures and phonology books which are usually linguistically oriented with little reference to problems Indonesian EFL learners are to encounter in real communicative situations.Thus, actual experience of oral communication in English is crucial for EFL teachers so that they can decide what aspects of English pronunciation are more important and have greater pedagogical significance for EFL learners.
Apart from the discussion of which mispronunciations are common and serious, the result of the study has also discovered that there are 14 mispronunciations those which are common and serious/unserious which are pedagogically significant in the teaching of English pronunciation.These 14 mispronunciations seven consonants and seven vowels require more serious attention in pronunciation teaching in Indonesian classrooms (cf.Kashiwagi, Snyder, & Craig, 2006;Rajadurai, 2007).A careful analysis of these 14 mispronunciations has revealed that they are mostly caused by the following three major factors.

Absence of English speech sounds in learners L1
Although some English speech sounds do not exist in learners L1, some learners can pronounce these sounds due to the prior learning of other languages such as Arabic or another local language for example, Buginese (the language spoken by people who live in the province of South Sulawesi, Indonesia).In the case of the mispronunciation /v/ /f/ (M8), for example, many Indonesian EFL learners, especially those who are Moslems, can produce the voiceless labiodental fricative /f/ properly because of the Arabic influence.However, this language transfer is not always put in the right place so that many learners do over-generalisation frequently, such as /p/ /f/ (M1).

Different distribution of the same sounds in English and learners L1
The English phoneme /b/ as in M2, for example, can exist in final, middle, or initial positions while the Indonesian phoneme /b/ never exists in the final position.According to the writing system of the Indonesian language, the letter <b> can exist in the final position but its pronunciation is always devoiced as / /.Even in many cases, the final phoneme /b/ is usually pronounced as the voiceless bilabial plosive /p/, such as /kr?b/ which is mispronounced as /kr?p/.

Different categorisation of sounds in English and learners L1
In terms of vowels, Gimson (2001) asserted that most foreign learners will have trouble attaining the vowel system of any variety of English, including RP (p.103).This assertion is reasonably compatible with the finding of the present study.Unlike English which has many vowels (up to 20 vowels and diphthongs), either Indonesian learners native language or the Indonesian language mostly has only five vowels with their allophones: /a/ with [a] and [ ]; /i/ with [ ] and [i ]; /u/ with [ ] and [u ]; /e/ with [e], [ ], and [ ]; and /o/ with [o] and [ ].For this reason, many Indonesian students, for example, are facing a problem to differentiate the sound [i ] as in / i p/ for sheep from the sound [ as in / p/ for ship because these sounds [i ] and [ are the allophones of the Indonesian phoneme /i/.Unfortunately, this difficulty becomes more complicated because of English spelling which does not correspond to the English sounds while Indonesian spelling and its sound always coincide.In the Indonesian language, for example, the letter <a> is always pronounced as the primary cardinal [a] or [ ].
Thus the findings of the study are consonant with recent research findings that the absence of English sounds in learners L1 and the different distribution of the same/similar sounds in English and learners L1 are the main reasons for difficulty in pronunciation for EFL learners (Carruthers, 1987;Moedjito, 2006;Ohata, 2004).Based on these findings, we have come to a conclusion that learners native language is still an active agent for the mispronunciation of English sounds through phonological negative transfer.This finding is partly consistent with Lado (1957) s contrastive analysis hypothesis (CAH) which assumes that it would be easy for L2 learners if the elements of L2 were similar to those of their L1, and it would be difficult for them if those elements were different from those of their L1.This suggests that CAH is still useful as an explanatory tool to clarify learners mispronunciations.

CONCLUSIONS AND SUGGESTIONS
Although our study has revealed several interesting facts about mispronunciations by Indonesian EFL learners, we must point out that the study has several limitations.Firstly, it investigated only the segmental features of pronunciations with no reference to suprasegmental features such as intonations and sentence stresses.Secondly, the participants of the study were limited to the Indonesian EFL teachers and the native English speakers who were chosen on an opportunistic basis, not on a random basis.Thirdly, the data of the study was collected by means of a paper-form questionnaire dealing with the 32 target mispronunciations which in turn were compiled on the basis of the first author s self-experience as an EFL teacher at Indonesian senior high schools.Fourthly, only native speakers perceptions were used as reference points for comparison with Indonesian EFL teachers perceptions.Considering the fact that English is now used as a global language, it may be necessary to include non-native speakers who regularly use English as a second language.
While acknowledging these limitations of the study, we believe that the findings of the study can be used as the points of reference for better English pronunciation teaching in EFL classrooms.The study also suggests that Indonesian EFL teachers should review their own perceptions of the seriousness of mispronunciations of English sounds.More attention should be paid to the significant mispronunciations through explicit instruction such as specific explanations and demonstrations concerning the differences in two languages, and production drills and sound discrimination exercises focused on these differences.
As generally accepted, pronunciation should focus not only on segmental features, but also on suprasegmental features.The next study should be extended to those suprasegmental features such as rhythm, intonation, stress, and adjustments in connected speech.Furthermore, because the participants of the study were limited to the Indonesian EFL teachers and native English speakers, ESL speakers should be included as evaluators of the seriousness of mispronunciations in the next study if we take into consideration the fact that the interaction between non-native speakers (NNS-NNS communication) has steadily been increasing (Jenkins, 2000, McKay, 2002;Walker, 2001).As long as we teach English as a global language, our goal should be global intelligibility, not intelligibility comfortable only to native speakers of English.The present study is a step toward this final goal.

/ / / /
You look gorgeous today.Where did you get this nice caught (coat)?/k t/ /k t/ M31, and M32) involve the substitution of diphthongs with the allophones of the vowels in learners native language.
Figure 3. Mean-matrix of the Indonesian respondents perceptions of the frequency and native English speakers perceptions of the seriousness a EFL b NES c M diff a Frequency; b The Indonesian EFL teachers; c The native English speakers; d The difference of the mean scores between EFL and NES.* p < 0.05; ** p < 0.01