Based on the previous results in Section 4.2.3, several phenomena are observed.
First, towards the eight categories of sound patterns in the word pronunciation tests, the participants handled the sounds b, f, soft g, h, k, p in single consonant letters, silent gh and ch in consonant digraphs, long/short o in single vowel letters, and ou in diphthongs well. These letters didn’t seem to be problems to them.
Second, it is taken for granted that if participants handled the words well in the pre-test, they were apt to distinguish the words with the same sound patterns well in the post-test. Namely, they had internalised these sounds and could pronounce them accurately in most cases. Therefore, when the participants had good performances on the words yellow, down, game, Sunday, house, and new in the pre-test, they got a high percentage of correctness in the post-test, too, on the words bow, jade, culture, layout, and dew.
However, it is also interestingly found that the same sound patterns in different word positions or different word clusters did not necessarily get the same
pronunciation from the participants. For example, participants got a percentage of 87.14 on the word game and 80 on jade, but only 37.14 on shape. Similar results were found in the word pairs Chinese (84.29%)/concrete (37.14%), cute (68.57%)/fume (31.43%), thank (17.14%)/hank (92.86%), exercise (15.71%)/exchange (57.14%), and stood (34.29%)/wood (72.86%). Consequently, every feature of a word contributes to the success of its pronunciation. That is, whether the sound of a word can be correctly pronounced or not depends on its phonological environment. This supports Li’s (1998:646) viewpoint that “as for the factor of sound environment, the problem for the Chinese learners to pronounce a consonant may be not only simply related to its different word positions but what the other sounds it is grouped with.”
Fourth, the participants’ major errors in word pronunciation tests towards eight types of sound patterns echoed the previous studies (Chang, 1994; Ing, 1997; Chu, 1997; Li, 1998; Lin, 2001). As Li (1998:639) points out, Chinese EFL learners have deletion, insertion, and substitution problems towards pronouncing consonants. These problems derive from lack of Mandarin Chinese corresponding sounds to English sounds and L1 language interference (Chang, 1994). For example, Mandarin Chinese doesn’t have consonant clusters as commonly seen in English; therefore, participants either deleted certain consonants or inserted certain vowels like /D/ when they
produce words. It was common that doctor was read as dotor or docator, fresh as fesh, picture as piture or picature, toothbrush as toothbush, freeze as feeze or fareeze, drowsy as dowsy, and concrete as concarete. Moreover, because Chinese syllables usually do not end with consonants except /n/, /H/, or /r/, some participants were inclined to delete English final consonants or to add a vowel to them. Hence, cool was mispronounced as coo, weight as weighter, stood as stoo, cute as cue, jade as jada, tight as tigh or tighter, pearl as pearlo, and wood as woo. As for consonant digraphs, ph, sh, and th, similarity, confusion, and substitution problems were found. As previous studies claimed, the error of the th sound appears due to the absence of its corresponding sound in participants’ mother tongue (Chang, 1994; Ing, 1997; Chu, 1997; Li, 1998). Since there are no /L/ and /M/ sounds in Mandarin Chinese, a
majority of the participants were found to substitute /s/ for /L/ and /l/, /d/, or /z/ for /M/.
Thus, thank becomes sank, weather becomes weler or weder, toothbrush becomes toosbrush, and clothe becomes close. With regard to the ph and sh sounds, owing to the similarity between the written letters and the phonetic symbols, some participants didn’t distinguish that they are consonant digraphs that represent sounds different from their component letters p or s, and sounded out the word phone as pone, sulphur as sulpur, fresh as fres, and toothbrush as toothbrus. As to mistaking shape as chape
and toothbrush as toothbruch, the error indicated that the participants confused sh with ch (Chu, 1997). About other errors of consonants no matter in word-initial, word-middle, or word-final positions, articulation errors were commonly found. The participants mispronounced hard g as /d/ or /dI/ in the word wagon, j as /z/ or /dz/ in jade, m as /n/ in game and fume, n as /r/ in phone and yawn, v as /f/, /b/, /m/, /w/ in village and vine, and w as /b/ in wood. It seems that the participants still did not have enough phonemic awareness to revise their own articulation.
Similarly, participants had articulation errors towards vowels. Though only five letters, a, e, i, o, and u, belong to vowels in English, the participants did not make fewer errors towards vowels than consonants due to the complicated qualities of vowels.
With respect to single vowel letters, the major errors resulted from inaccurate articulation of phonemes. Thus, long a was mispronounced as /G/, /$/, /A/, short a as /0/, /G/, /e/, long e as /G/, /e/, /a0/, short e as /0/, /e/, /$/, long i as /0/, /o/, /e/, short i as /G/, /i/, long u as /$/, /i/, /o/, /u/, and short u as /a?/ or /A/, some of which were categorized as “errors of confusing vowel letters with phonetic symbols” in Chu’s (1997) study. For example, the letter a was wrongly corresponded to the phoneme /A/, e to /e/, i to /i/, and u to /u/. As for substituting /i/ for /0/ and mixing long a, short a, and short e, it seemed that dull sensitivity to the articulation of phonemes or
fossilization in articulation caused the inaccuracy.
Regarding vowel digraphs, different from Lin’s (2001) report that subjects in her experiment were apt to pronounce the aw by substituting /a?/ for /C/, the participants in this study either mistook the word saw as say or read it as /so/ and yawn as /jon/. It seemed that they knew the letters aw corresponded to the phoneme /C/, but
inaccurately articulated the phoneme as /o/. Likewise, some participants failed to pronounce ay in layout well as /e/, but just read it as ['lGa?t]. A similar situation was
also found when referring to the ea sound pattern as in the words weather and
treadmill, in which ea was mispronounced as /e/, not /G/. Again, dull sensitivity to the articulation of phonemes or fossilization in articulation seemed to a factor of the inaccuracy. Besides, several cases pronounced them as ['wiMQ] and ['tridm0l]. This error resulted from their wrong concept of regarding the words as if they were regular ones, one of “regularization errors” (Chu, 1997:33). As for the oo sound pattern, similar to Lin’s (2001) study, a majority of wrong cases pronounced it as /o/ rather than the correct /u/ or /?/. Thus, cool became [kol], stood [stod], wood [wod], and toothbrush ['toLbrKN]. It was likely that the participants failed to recognize oo as a phoneme unit and confusion arose between the digraph and the phonetic symbol /o/
similar to what mentioned above in some cases of single vowel letters. As to other mispronunciation of these vowel digraphs or others, it seemed to suggest that some participants had limited knowledge of vowel digraphs and failed to pronounce them correctly.
Concerning diphthongs, several participants tended to pronounce the oi sound pattern as /C/ rather than /C0/ in join and toilet, ow as /K/ or /o/ in down and /C/ in drowsy. Like what Lin’s (2001) study suggested, the participants “tended to substitute two vowel sequences with a monophthong” (Lin, 2001:101). Again, it seemed that some participants did not have thorough understanding about diphthongs, and neither could they pronounce phonemes accurately.
With regard to r-controlled vowels, participants’ common errors were found in the sound patterns of ear as in learn and pearl and ur as in sulphur. The former sound pattern was commonly sounded out as /Gr/ or /0r/, and the latter as /Ar/ or /Cr/. It is not surprising that the participants mispronounced the ear sequence in pearl as /Gr/ or /0r/, but not /"/. This might result from the analogy from familiar words such as pear, wear,
ear, year, dear, hear, and near. Namely, these words had become their base-word vocabulary, and they were apt to pronounce the sequence of ear in pearl as /Gr/ or /0r/.
Regarding schwa, theoretically speaking, unstressed syllables are usually
reduced to the vowel /0/ or /D/. However, several participants paid no attention to this and mispronounced the schwa a as /o/ or /C/ in the word allusion.
Concerning common cases, the mispronunciation of sion as /dIDn/ in television, allusion and tion as /IDn/ in mention seemed to originate from the participants’
confusion about the articulation of /I/, /dI/, and /N/. This suggests that participants need to have keen hearing to differentiate these similar sounds from each other.
Another error was that five participants mispronounced culture as cultry. Maybe the word country occurred to their minds when they were producing culture. As to the ex sequence in the words exercise and exchange, it was commonly simplified as ek or es, which suggested that the participants were inclined to omit one consonant in the consonant cluster /ks/ due to the absence of consonant clusters in their mother tongue.
Finally, the interesting phenomenon that the participants were inclined to replace one word with another word visually similar to it follows to be discussed. Chu
(1997:32) classified such errors as “visually similar errors” in her report, and the wrong pronunciation of the participants in this study fitted in her first two categories of this type of error: (1) Initial and more letters shared with the target word:
pronouncing weight as what or white, doctor as daughter, house as home, join as joy or June, saw as say, shape as ship or shop, (2) Initial and final letters shared with the target word: pronouncing thank as think, fresh as fish, new as now, exercise as excuse, and weather as winter or waiter. Furthermore, some other visually similar errors were related to the target word in meaning, termed as “meaning-preserving errors” by Baron (1979, cited in Chu, 1997:32). These errors shared not only many common letters but also some of the meaning with the target word. Examples of such errors
include Sunday as Saturday, new as news, thank as thanks, sleep as sleeping, and learn as learning. One possibility about the occurrence of the last three errors might be related to the arrangement of English textbooks in junior high school. In Vol. 1, students learned thanks earlier than thank and learned verbs and their present participle forms simultaneously. It seemed that the participants read these words according to their intuition or the very first idea entering their minds, but not based upon the grapheme-phoneme correspondence relationships. What is more, all these visually similar errors appeared in the pre-test except for only one word, shape, in the post-test. Perhaps the reflection of one interviewee with progress had a good
illustration on this: he depended on his intuition and impressions to sound out the words in the pre-test because those words had been learned in junior high school, but applied phonics rules to word pronunciation in the post-test owing to the novelty of words. This suggests that the participants’ intuition and impressions without thinking too much might cause visually similar errors in word pronunciation.
To summarize, as far as the participants’ major errors in this study were
concerned, different ways of letter combination might result in different difficulties in their word pronunciation. Besides, they had deletion, insertion, substitution, and confusion problems along with articulation, regularization errors and visually similar errors towards pronouncing consonants, vowels, and whole words.
4.3.4 Discussion of Participants’ Awareness in Memorizing Vocabulary Prior to