Introduction
It is common knowledge that Japanese and other East Asian native language speakers have difficulty producing the English sounds of /l/ and /r/. This difficulty is so widely known, in fact, that Ingram and Park (1998) refer to it as having become “a linguistic stereotype” (p. 1161) of these peoples. Explanations for why such non-native English language users would experience pronunciation difficulties have traditionally been explained by differences in the phones available for use in each language. For instance, because Japanese has only one liquid phoneme, native Japanese speakers learning English must not only learn to pronounce a completely new /r/ phone, but they must also learn to discriminate between the English phonemes of /l/ and /r/ (Shimamune & Smith, 1995). Some state that interference from the Japanese sound system causes difficulties (Kitao, 1995; Ohata, 2004), which may include pronunciation difficulty due to neuromuscular plasticity (Brown, 1994) or simply differences in sound production that create a foreign accent (Avery & Ehrlich, 1992).
Not all would agree that a learner’s first language presents the biggest obstacle for learning the pronunciation of another language. For example, Makarova (2000) states that English pronunciation instruction in classroom settings in Japan is virtually nonexistent, with university English majors being the most likely recipients of said instruction. She gives several cogent explanations for the lack of such instruction in language education at all levels in the country, including (a) large classes that make direct instruction and feedback by instructors difficult, (b) an education system that focuses on the passive aspects of vocabulary and grammar, (c) prominence given to tests such as TOEFL and TOEIC that place minimal or no emphasis on pronunciation, and (d) university curricula that attach importance to pronunciation only for remedial rather than skill-building purposes. Thus, in her view, “the reasons for the poor development of English pronunciation skills by the Japanese learners lie not in the Japanese language, but in the ways English is taught” (p. 246).
As discussed by Canale and Swain (1980) and later by Savignon (1983), among others, one important goal of language learners is to develop communicative competence, or “that aspect of our competence that enables us to convey and interpret messages and to negotiate meanings interpersonally within specific contexts” (Brown, 1994, p. 227), particularly in spoken discourse. In other words, learners strive to gain a sufficient amount of knowledge and skill in the foreign language and culture in order to understand and be understood. Pronunciation certainly falls within the parameters of communicative competence, for a language learner must both understand a speaker’s pronunciation in order to extract meaning from messages and possess a certain level of pronunciation skill to make their own spoken messages comprehensible to the listener.
If various aspects of speaking and listening are taught and practiced for spoken discourse in EFL (English as a foreign language) learning situations, it seems logical to include some amount of pronunciation instruction in this learning. EFL learners themselves have been found to realize how poor pronunciation can impede their language learning success (Brown, 1991). Improved pronunciation and intelligibility can also increase their confidence in speaking situations (Avery & Ehrlich, 1992).
The Current Trend of Pronunciation Instruction
When language instructors find the time or the need to include pronunciation instruction within the boundaries of a course’s curriculum, many opt to first teach suprasegmentals, or the larger units of pronunciation that include intonation, stress, rhythm, and pitch, rather than segmentals, or the individual speech-sounds of a language. This top-down approach to pronunciation instruction is directly related to communicative competence and instructors’ beliefs about the importance of teaching pronunciation via communicative activities, for the belief is that there is less likely to be confusion and misunderstanding generated from a misspoken phoneme (/l/ substituted for /r/, for instance) in an utterance because meaning can still be deciphered from the larger context of the spoken message. If there are suprasegmental pronunciation problems in spoken discourse, then it becomes more difficult to make sense of a speaker’s meaning.
For example, if a non-native speaker of English were to mistakenly utter the word “crime” rather than “climb” when stating “I will crime up there and get the book for you,” this mistake in pronunciation is easily overcome by the listener because the context of the utterance indicates that “climb” was simply misspoken. The intended meaning is easily deciphered, particularly if there are minor or no problems present with suprasegmental features. However, the likelihood of comprehension is reduced if these suprasegmental features are missing or improperly applied, even if each of the words in the utterance is pronounced correctly.
There is a vast and growing literature base illustrating support for the teaching of suprasegmentals prior to (and to the possible exclusion of) the teaching of segmentals. Suprasegmental instruction is supported on the grounds that it fosters meaningful communication (Celce-Murcia, 1987), that it is congruent with the primary goal of attaining fluent and comprehensible speech when learning another language’s spoken components (Avery & Ehrlich, 1992), and that it approaches the remediation of pronunciation problems at the discourse level (see Hinkelman & Halvorsen, 1998) where communicative competence is most needed and apparent. Others note how teaching pronunciation via segmentals, particularly through minimal pair drills, does little to foster target pronunciation in verbal discourse beyond the phoneme level (Boku, 1998; del Castillo, 1990), especially because bottom-up pronunciation instruction can create “students who are capable of repeating the necessary words or phrases but are unable to transfer this skill when actually engaged in free conversation” (Hinkelman & Halvorsen, 1998, p. 88).
Even with cogent reasons for teaching suprasegmentals, there remains the possibility that instructors may ignore the teaching of the physical (i.e., segmental) aspects of pronunciation, either because of time concerns or because suprasegmental instruction is generally more highly favored. If one agrees that pronunciation is an integral part of language teaching and learning, particularly because language instruction necessarily involves pronunciation teaching (Parish, 1991), it is difficult to rationalize the complete exclusion of segmental instruction in language teaching, especially when certain segmentals cause confusion in learners and listeners alike. For Japanese EFL learners, the confusion caused by the English phonemes of /l/ and /r/ seemingly requires the inclusion of some amount of segmental-level instruction.
The English Phonemes of /l/ and /r/
An understanding of why /l/ and /r/ would be problematic for Japanese and other East Asian native language speakers – and why segmental-level instruction on these sounds may benefit learners – can only come from examining how these sounds are physically produced. Although it is understood that pronunciation differences exist in terms of dialect and phone position within words, for the purpose of simplicity, the following explanations will mainly (but not exclusively) describe /l/ and /r/ as they are spoken in American English in either word onset position or as second elements of an onset consonant cluster. This limited focus is not meant to reduce the importance of these sounds when vocalized in other positions nor is it meant to marginalize other varieties of English. As suggested by Parish (1991), readers are encouraged to recognize the value and validity of various dialects, for intelligibility on the part of the learners should be the intended outcome of pronunciation instruction and not a lack of sensitivity to speakers’ possible pronunciation preferences. Furthermore, while there is some disagreement as to how these sounds are produced, this disagreement has previously been attributed simply to dialectical differences (Avery & Ehrlich, 1992).
The /l/ sound in English (called a lateral consonant) is made by touching the tip of the tongue to the tooth ridge and forcing air out of the mouth over the sides of the tongue (Avery & Ehrlich, 1992). In contrast, the production of the /r/ sound is described by Bronstein (1960) as follows:
In the most common position for /r/ before stressed vowels…and initially before unstressed vowels…, the tip and blade of the tongue are turned upward, toward the hard palate, the tip pointing to (but not touching) the area immediately behind the alveolar ridges. In another common formation of /r/, the tongue tip remains low, while the central part of the tongue bunches and is raised toward the posterior section of the hard palate.…In all formations of /r/, the sides of the tongue are in contact with the bicuspid and molar teeth, as for /n/ or /d/. (p. 116)
More accurately, /r/ is a voiced central approximate, the phonetic transcription of which is /®/ (Ladefoged, 1993). In its pronunciation, some rounding of the lips or lip protrusion occurs (Avery & Ehrlich, 1992; Bronstein, 1960; Ladefoged, 1993).
Both of these sounds are different from the Japanese liquid, which is somewhere between these two sounds (Avery & Ehrlich, 1992). This Japanese sound possesses a wide phonetic variation and can be anything from a “slightly retroflexed lateral approximant, through to an alveolar tap or flap” (Ingram & Park, 1998, p. 1163), but this variation is mainly attributed to speaker differences because this Japanese liquid (written phonetically as the flap /R/) has no exact articulation point, which helps to explain why Japanese speakers of English often sound as if they are substituting /l/ for /r/ (and vice versa) when speaking (Ohata, 2004).
Teaching the Physical Aspects of /l/ and /r/ Production
In order to provide instruction on the physical aspects of English /l/ and /r/ production, either at a stage after suprasegmental features have been taught or as a means for brief instruction during a language course, I have often taught my students the following self-developed pronunciation instruction method that, among other things, purposefully exaggerates the use of the lips:
1. For /l/ production:
(a) if a word begins with an /l/ sound, flap your tongue
(b) it helps to smile when making the /l/ sound
(c) the /l/ sound in English does not require much energy to produce
2. For /r/ production:
(a) if a word begins with an /r/ sound, curve your tongue
(b) it helps to round your lips when making the /r/ sound
(c) the /r/ sound in English requires more energy to produce than /l/
I explain to my students that though this method is purposefully exaggerated and does not cover every allophone of /l/ and /r/ in English, it nevertheless makes it easy to both see and hear the differences between the two sounds. For instance, I tell the students to smile when making an /l/ sound because frowning or rounded lips distort the /l/ sound, making it unclear. I explain that making the /l/ sound does not require much energy, at least in contrast to the /r/ sound, which, unless “energized,” will likely be heard as an /l/. Additionally, the /r/ sound is benefited by lip rounding. In the course of the explanation and subsequent drills, I focus heavily on this lip-rounding feature by exaggerating it and actively encourage it in my students by asking them to make the “face of an octopus,” as the cultural cartoon image of an octopus in Japan is that of an animal with pursed lips. This explanation is easy for the students to understand and is physically possible for most if not all of them to do.
Having informally seen improvement in my students’ pronunciation after explaining /l/ and /r/ production this way, I wondered if the exaggerated aspects of this instruction method would statistically improve the intelligibility of students’ speech such that native English speakers would consider their production of /l/ and /r/ better than those who had not received such instruction. The following details a preliminary research study that explores this question.
The Research
Location and Participants
The research was conducted with two groups of English majors (all of roughly intermediate proficiency) enrolled at a small four-year university located in central Japan. The test group (n = 13) was composed of third- and fourth-year volunteers who had participated in pronunciation practice exercises with me (hereafter, the researcher) during an intensive English class between the spring and fall semesters. The control group (n = 12) was composed of similar students, differing only in that they were enrolled in a different intensive class at the time and had not received any specific pronunciation training or instruction outside what they had learned in the course of their regular English language learning activities up until the time of the study.
Procedure
The following procedure was used for each student participant. First, the researcher directed one student into a classroom and asked him/her to sit at a table across from the researcher and another native English instructor (hereafter, NEI), both of whom acted as raters of the students’ pronunciation. The student was given the following materials:
1. a piece of cardboard (roughly 26cm by 17cm)
2. a bag containing slips of paper with minimal pair words written on them (one word per slip of paper). 29 minimal pairs were used (58 words total – see Appendix A): 16 pairs with /l/ and /r/ in word onset position, 11 pairs with them as second elements of an onset consonant cluster, and 2 pairs with them in intervocalic position.
3. another bag containing slips of paper with one sentence written on each. Each sentence included one minimal pair word (from the set of 29 minimal pairs). Only one word from each minimal pair was selected for use in these sentences, meaning that 29 sentences were constructed in all (see Appendix B for the sentences used). Word selection for use in these sentences was random.
The student was instructed to block his/her mouth with the cardboard whenever speaking so that neither rater could see it. The student was then given the bag containing the minimal pair words and was asked to pick one word at random and read it silently. Students were told that if a word’s pronunciation or meaning was unclear that it was acceptable to choose another word. When ready, the student was asked to say the word aloud only once in a clear voice with the best pronunciation possible. After each word spoken and without any physical clues to guide them (e.g., rounded lips), the raters marked on their respective sheets whether they heard an /l/ or /r/ sound in that word or if it were indeterminable. After the sheets were marked, the student showed the word on the slip of paper, which the researcher then noted for comparative purposes later.
After the student uttered ten random words, he/she randomly selected one sentence from the sentence bag and was asked to read it silently to first gain an understanding of the sentence. As with the minimal pair words, students were instructed to select another sentence if they were unsure of a sentence’s meaning or of the pronunciation of any word in the sentence. If the student felt confident in his/her understanding of the elements of the sentence, the raters were then shown the sentence. Because of the difficulty in generating natural-sounding sentences that did not have any other /l/ or /r/ sounds in them besides the minimal pair words, there was the possibility that the NEI would concentrate and rate other /l/ or /r/ sounds besides those of the target minimal pair words, so the minimal pair word in each sentence was underlined. The student was then instructed to read the sentence once using their best English pronunciation throughout, again with mouth blocked by the cardboard. The raters listened as each sentence was read and recorded if the pronunciation of that word in the sentence was either “Good,” “Understandable,” or “Poor.” Each student read five sentences in total.
Results
In order to conduct analyses with the ratings given to students’ pronunciation, the following procedures were used. For the minimal pair words, if a rater heard a minimal pair word correctly, then one point was awarded to that student for that word. If the word was heard incorrectly or was considered indeterminable, then no points were awarded. For instance, if both raters heard the selected word “light” as “light,” the student was awarded two points (one point from each rater). If one rater heard the selected word “frame” as “frame” but the other heard it as “flame,” then the student was awarded one point in total for that word. For the minimal pair words said in sentences, students were awarded two points from a single rater if a word was rated as having been spoken with “Good” pronunciation, one point if “Understandable,” and no points if “Poor.”
Because of the random nature of word/sentence selection and the unequal number of participants in the two groups, direct comparisons were difficult to make. However, totals show that the raters were quite consistent overall in their awarding of points: there was only a 12 point difference between the raters’ totals for the test group’s words and sentences and only a one point difference between the total points awarded to the control group’s utterances. Table 1 more directly compares the total points awarded by the researcher with the points awarded by the NEI for both groups. Because of unequal student numbers in each group, awarded points were totaled within groups by rater and then averaged (averages are in parentheses).
Table 1
| Total Points Awarded to the Groups |
| |
Test Group |
Control Group |
| Researcher: Word Points |
96 (7.4) |
87 (7.3) |
| NEI: Word Points |
101 (7.8) |
73 (6.1) |
| Total Word Points Awarded |
197 (7.6) |
160 (6.7) |
| Researcher: Sentence Points |
104 (8.0) |
70 (5.8) |
| NEI: Sentence Points |
87 (6.7) |
85 (7.1) |
| Total Sentence Points Awarded |
191 (7.3) |
155 (6.5) |
The average total word points for the test group (7.6) surpassed that of the control group (6.7), and the average total sentence points for the test group (7.3) surpassed that of the control group (6.5). Taking these averages into consideration, it appeared that students who had received the exaggerated pronunciation instruction had better perceived /l/ and /r/ pronunciation – for both words and sentences – than the students who had not. In other words, these averages alone reveal that both the researcher and the NEI considered the test group’s /l/ and /r/ pronunciation to be better than that of the control group for both minimal pair words spoken in isolation as well as in sentences. Looking between groups, the only instance where the control group was awarded more points on average was by the NEI for sentences (7.1 points to the test group’s 6.7 points), but in all other instances, the researcher and the NEI awarded more points on average for word and sentence pronunciation to the test group.
As revealing as these averages were, more detailed analyses were still considered necessary. It was not yet clear if the test group’s pronunciation was significantly better, statistically speaking, than the control group’s pronunciation, nor was it clear if there were significant differences between how the researcher and the NEI rated students’ pronunciation. Several independent-measure t-tests were therefore conducted to compare the groups and the raters . The first set of tests, conducted to ascertain statistically whether or not the raters awarded points differently, were:
Test A: testing significance between the researcher’s ratings of the test group’s word pronunciation (M = 7.4, SD = 2.1) and the NEI’s ratings (M = 7.8, SD = 1.7). No statistical significance was found, t(24) = -1.70, p > .05, two-tailed.
Test B: testing significance between the researcher’s ratings of the control group’s word pronunciation (M = 7.3, SD = 1.8) and the NEI’s ratings (M = 6.1, SD = 1.7). No statistical significance was found, t(22) = 1.65, p > .05, two-tailed.
Test C: testing significance between the researcher’s ratings of the test group’s sentence pronunciation (M = 8.0, SD = 1.5) and the NEI’s ratings (M = 6.7, SD = 1.8). No statistical significance was found, t(24) = 2.02, p > .05, two-tailed.
Test D: testing significance between the researcher’s ratings of the control group’s sentence pronunciation (M = 5.8, SD = 2.4) and the NEI’s ratings (M = 7.1, SD = 2.2). No statistical significance was found, t(22) = -1.32, p > .05, two-tailed.
From these tests, it was discovered that the raters did not differ significantly in how they rated the pronunciation of the students, meaning that, statistically, they appeared to be hearing the same sounds being pronounced by the students.
Additionally, the overall performance of the two groups as perceived by the raters was also tested for statistical significance. These tests included:
Test E: testing significance between the total word ratings awarded to the test group students by the raters (M = 15.2, SD = 3.2) with the control group’s ratings (M = 13.3, SD = 3.2). No statistical significance was found, t(23) = 1.42, p > .05, two-tailed.
Test F: testing significance between the total sentence ratings awarded to the test group students by the raters (M = 14.7, SD = 2.4) with the control group’s ratings (M = 12.9, SD = 4.5). No statistical significance was found, t(23) = 1.54, p > .05, two-tailed.
Having found no statistical significance with any of these tests meant that the students in both groups had, statistically speaking, pronunciation of both words and sentences that was not all that disparate, even though as shown in Table 1 the test group had performed better on average overall.
As there was the possibility that the test group’s better average scores stemmed from something about the minimal pair words themselves, a deeper examination was undertaken that aimed to reveal how the students in each group were perceived by the raters to utter /l/ and /r/ in the words used for the study, that is, in word onset position, as second elements of an onset consonant cluster, and in intervocalic position, be they from the selected words or sentences. For this examination, a list of all the words and sentences selected by students in both groups was first created (as some words and sentences had not been selected due to the random nature of word/sentence selection). The combined scores for each of the words in these three different positions were recorded. For instance, the students in the test group randomly selected the word “law” three times. The total scores by both raters for each of the three times this word was spoken were “1,” “2,” and “2” (see Appendix C for a list of the total combined rater scores).
With this information, six additional independent-measure t-tests were conducted:
Test G: comparing how well the two groups uttered /l/ and /r/ sounds in word onset position from the selected words. Statistical significance was found, t(128) = 2.08, p < .05, two-tailed.
Test H: comparing how well the two groups uttered /l/ and /r/ sounds as second elements of an onset consonant cluster from the selected words. No statistical significance was found, t(95) = 0.75, p > .05, two-tailed.
Test I: comparing how well the two groups uttered /l/ and /r/ sounds in intervocalic position from the selected words. No statistical significance was found, t(20) = 1.83, p > .05, two-tailed.
Test J: comparing how well the two groups uttered /l/ and /r/ sounds in word onset position from the selected sentences. No statistical significance was found, t(70) = 1.95, p > .05, two-tailed.
Test K: comparing how well the two groups uttered /l/ and /r/ sounds as second elements of an onset consonant cluster from the selected sentences. No statistical significance was found, t(42) = -0.32, p > .05, two-tailed.
Test L: comparing how well the two groups uttered /l/ and /r/ sounds in intervocalic position from the selected sentences. No statistical significance was found, t(3) = -0.70, p > .05, two-tailed.
Statistical significance was found only once, when looking for differences between how the groups pronounced /l/ and /r/ in words in word onset position. Scrutinizing the groups’ scores for these words revealed that the test group was awarded more points by both raters for such words than was the control group. The test group selected an onset word position word 67 times (/l/ = 30, /r/ = 37), receiving a total of 111 points (M = 1.7). The control group selected such words only four times less (/l/ = 34, /r/ = 29) but received far fewer points on average (M = 1.4). The test group was therefore considered to have performed better than the control group when such onset words were selected. If the word onset word began with an /l/ sound, the test group performed quite well, receiving 1.8 points on average, in contrast to the control group’s 1.4 points. Both groups performed nearly the same for /r/ onset words, averaging 1.5 points each.
Conclusion and Discussion
As the results from this preliminary study show, the students in the test group who were taught the researcher’s segmental-level exaggerated method for pronouncing English /l/ and /r/ had better pronunciation on average of minimal pair words in isolation and in sentences compared to that of the control group and, statistically speaking, had better pronunciation for minimal pair words with /l/ in word onset position. Though modest, these results still suggest that the exaggerated pronunciation instruction described above could be considered successful in improving students’ pronunciation as evinced in segmental-level minimal pair utterances as well as in sentences where suprasegmental features become relevant.
While statistical significance was not found when testing for differences in the pronunciation of minimal pair words in sentences (most likely due to small sample size), results still indicated better average scores by the test group than the control group and hence lend moderate support for segmental-level pronunciation instruction. It has been stated elsewhere (Underhill, 1996) that pronunciation instruction, learning, and practice cannot only focus on cognitive (i.e., suprasegmental) aspects but must also include some element of physical training, which includes teaching about the movement of the tongue and lips during speech production. As mentioned above, it has been noted that pronunciation difficulties can arise from neuromuscular plasticity (Brown, 1994) or unfamiliarity with or lack of knowledge about how to move physical components for pronunciation purposes. I therefore agree with Brown and Nation’s (1997) statement that form-focused pronunciation instruction still has a place in language teaching and learning and would argue that such instruction should take place in formal language learning instruction, regardless of whether it occurs before or after suprasegmental instruction.
Though only a preliminary study, these results are nevertheless in line with results found by other researchers studying pronunciation instruction. Derwing, Munro, and Wiebe’s (1998) study, while indicating that suprasegmental instruction improves pronunciation more than segmental instruction, also indicated that segmental instruction is better than no instruction at all. Kashiwagi and Snyder (2003) report that learners’ detailed knowledge of phonemes at the segmental level is still important for their improving of their L2 phonology. While such researchers clearly support learning pronunciation in a communicative framework, their findings, along with those of the present study, further the argument that segmental-level pronunciation instruction need not be abandoned entirely. Pronunciation instruction (and correction) should, as Parish (1991) argues, be part of an ongoing process rather than be reserved only for pronunciation lessons or exercises, which should come as good news for both instructors and students who find pronunciation instruction missing in the Japanese EFL classroom. It may therefore be suggested that instructors can best help students by teaching pronunciation as needed and assess the intelligibility of students’ pronunciation in spoken discourse via communicative activities, thereby including both segmental and suprasegmental instruction in the classroom. Also, through such activities, students would increase their awareness of the differences between the English and Japanese sounds systems and could be inspired to improve their pronunciation, possibly on their own (Ohata, 2004).
It should be stated that native-like pronunciation need not be the ultimate aim, as increased intelligibility rather than complete mastery should be the desired goal (cf., Hamada & Vogt, 1994; Wong, 1987). If a sufficient level of communicative competence is gained in pronunciation, minor pronunciation deficiencies can certainly be ignored or downplayed, but this is no reason to forgo pronunciation instruction and rely only on spot error correction (or no correction at all) in the classroom. Formal pronunciation instruction, even if infrequent, works to alleviate the “nerves cycle” (Figure 1) and strengthen a “positive cycle” (Figure 2).
Figure 1: Nerves cycle (Avery & Ehrlich, 1992, p. 222)
Figure 2: Positive cycle (Avery & Ehrlich, 1992, p. 226)
In short, students’ pronunciation problems, if they exist, must eventually be addressed. The pronunciation difficulties learners experience, such as /l/ and /r/ for Japanese speakers, will not necessarily “fade away” in the process of discourse practice, especially if students are unaware of the physical mechanics behind specific phone production. Pronunciation instruction must include instructors cultivating an understanding of the differences between the source and target languages (Ohata, 2004) and “must be included in students’ training, yet that training must be country specific, and materials and research must now stop focusing on the ‘general’ and start considering the specific” (Robertson, 2003, pp. 18-19).
Although the present study indicated that the exaggerated pronunciation instruction method for /l/ and /r/ production appeared beneficial, there were still several limitations worth noting, many of which point to the study’s small-scale and preliminary nature. For instance, the use of volunteers, though preferable to conscripting participants, limited sample size, hence negatively impacting the study in a number of ways, the first being word/sentence selection. While the randomness in selection was deemed useful in that students who were tested first would not tell later students which /l/ and /r/ words were being tested, this consequently reduced the total number of words/sentences uttered, thereby making comparisons between students difficult or impossible. It is therefore recommended that future research similar to this not employ the random factor unless the sample size is very large.
Second, the small sample size impacted the statistical tests conducted because such a small sample (a) may not have adequately reflected the target populations of both groups and (b) precluded the conducting of future tests (e.g., a follow-up repeated-measures test). Although a later full-scale study with these and other participants at this university was originally planned, it could not be carried out due to the researcher relocating to another area and thus losing access to these students. Nevertheless, such a study could be conducted with learners elsewhere, by the researcher or others, and utilize (a) more participants (e.g., at least two university classes of between 30 and 40 students each), as larger sample sizes increase the likelihood that a sample represents its population (Gravetter & Wallnau, 2000), (b) more raters, (c) an improved research design that includes more pronunciation issues (e.g., testing “dark /l/” pronunciation), and (d) multiple test groups (i.e., students taught varying pronunciation methods) to expand and/or strengthen the present study.
Also, most of the minimal pair words selected were chosen because they had word initial /l/ or /r/, which is one place where the Japanese liquid is found (Ingram & Park, 1998). Selecting these words was done to reduce the chance that the students would be physically unable to pronounce them. Most other words selected were chosen because they were second elements of an onset consonant cluster, which are common enough in English to have warranted their inclusion for testing purposes. Future research should include the teaching and testing of /l/ and /r/ in other word positions, particularly where the Japanese liquid is not found. A final limitation was that the students were reading aloud words and sentences from slips of paper rather than actually speaking them in unscripted discourse. This difference did not present a major concern for this preliminary study, and while its impact on the study remains unknown, the inclusion of unscripted narratives is recommended for future large-scale studies.
Ordinarily, small sample sizes would suggest use of the Mann-Whitney U-Test or the Wilcoxon Rank-Sum Test, but the raw data showed a lack of sufficient range to make either of these tests feasible for some analyses.
References
Avery, P., & Ehrlich, S. (1992). Teaching American English pronunciation. Oxford: Oxford University Press.
Boku, M. (1998). Student-centered pronunciation practice: More than "right" or "light". Retrieved September 26, 2005, from http://www.jalt-publications.org/tlt/files/98/oct/sh_boku.html
Bronstein, A. J. (1960). The pronunciation of American English: An introduction to phonetics. New York: Appleton-Century-Crofts, Inc.
Brown, A. (1991). Introduction. In A. Brown (Ed.), Teaching English pronunciation: A book of readings (pp. 1-5). London: Routledge.
Brown, H. D. (1994). Principles of language learning and teaching (3rd ed.). Englewood Cliffs, NJ: Prentice Hall Regents.
Brown, R. S., & Nation, P. (1997). Teaching speaking: Suggestions for the classroom. Retrieved September 26, 2005, from http://www.jalt-publications.org/tlt/files/97/jan/speaking.html
Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1, 1-47.
Celce-Murcia, M. (1987). Teaching pronunciation as communication. In J. Morely (Ed.), Current perspectives on pronunciation: Practices anchored in theory (pp. 1-12). Washington, D.C.: Teachers of English to Speakers of Other Languages.
del Castillo, L. (1990). L2 pronunciation pedagogy: Where have we been? Where are we headed? The Language Teacher, 14(10), 3-8.
Derwing, T. M., Munro, M. J., & Wiebe, G. (1998). Evidence in favor of a broad framework for pronunciation instruction. Language Learning, 48(3), 393-410.
Gravetter, F. J., & Wallnau, L. B. (2000). Statistics for the behavioral sciences. Belmont, CA: Wadsworth/Thomson Learning.
Hamada, M., & Vogt, L. (1994). Accessible pronunciation. Paper presented at the 1994 JALT Kansai Conference, Kyoto, Japan (pp. 253-258). (ERIC Document Reproduction Service No. ED381833)
Hinkelman, D., & Halvorsen, J. (1998). Discourse-oriented activities for pronunciation teaching. Proceedings of the JALT 1998 Conference, Omiya, Japan, 88-92.
Ingram, J. C. L., & Park, S.-G. (1998). Language, context, and speaker effects in the identification and discrimination of English /r/ and /l/ by Japanese and Korean listeners. Journal of the Acoustical Society of America, 103(2), 1161-1174.
Kashiwagi, A., & Snyder, M. K. (2003). The effect of form-focused instruction on pronunciation. Proceedings of the JALT 2003 Conference, Shizuoka, Japan, 262-271.
Kitao, K. (1995). Difficulty in English pronunciation for Japanese people. In K. Kitao & S. K. Kitao (Eds.), English teaching: Theory, research and practice (pp. 129-143). Tokyo: Eichosha.
Ladefoged, P. (1993). A course in phonetics (3rd ed.). Fort Worth, TX: Harcourt Brace College Publishers.
Makarova, V. (2000). Trends in pronunciation teaching in Japan. Proceedings of the JALT 2000 International Conference on Language Teaching & Learning and Educational Materials Expo, Shizuoka City, Japan, 245-250.
Ohata, K. (2004). Phonological differences between Japanese and English: Several potentially problematic areas of pronunciation for Japanese ESL/EFL learners. Asian EFL Journal, 6(4), Article 5. Retrieved October 17, 2005, from
http://www.asian-efl-journal.com/december_2004_KO.php.
Parish, C. (1991). A practical philosophy of pronunciation. In A. Brown (Ed.), Teaching English pronunciation: A book of readings (pp. 104-119). London: Routledge.
Robertson, P. (2003). Teaching English pronunciation skills to the Asian learner. A cultural complexity or subsumed piece of cake? Asian EFL Journal, 5(2), Article 5. Retrieved September 26, 2005, from
http://www.asian-efl-journal.com/june2003subpr.php
Savignon, S. J. (1983). Communicative competence: Theory and classroom practice. Reading, MA: Addison-Wesley Publishing Company.
Shimamune, S., & Smith, S. L. (1995). The relationship between pronunciation and listening discrimination when Japanese natives are learning English. Journal of Applied Behavior Analysis, 28, 577-578.
Underhill, A. (1996). Making pronunciation work for your learners. Retrieved September 26, 2005, from http://www.jalt-publications.org/tlt/files/96/sept/pron.html
Wong, R. (1987). Learner variables and prepronunciation considerations in teaching pronunciation. In J. Morely (Ed.), Current perspectives on pronunciation: Practices anchored in theory (pp. 13-28). Washington, D.C.: Teachers of English to Speakers of Other Languages.
Appendix A: Minimal Pair Words
| alive – arrive |
|
clash – crash |
|
clouds – crowds |
| clown – crown |
|
clutch – crutch |
|
collect – correct |
| climb – crime |
|
flame – frame |
|
flute – fruit |
| fly – fry |
|
glass – grass |
|
glow – grow |
| lace – race |
|
lamp – ramp |
|
late – rate |
| law – raw |
|
leads – reads |
|
led – red |
| lent – rent |
|
lice – rice |
|
light – right |
| list – wrist |
|
liver – river |
|
load – road |
| lock – rock |
|
long – wrong |
|
loom – room |
| low – row |
|
playing – praying |
|
|
Appendix B: Minimal Pair Sentences
Appendix C: Total Combined Rater Scores for Selected Words/Sentences
(See SWF file)
