This study investigated the phonetic features of the English diphthongs /ei/ and /ou/ pronounced by Japanese elementary school students in comparison with the pronunciation of native English speakers and Japanese vowels. Pronouncing diphthongs is difficult for Japanese learners as they do not exist in the Japanese sound system. However, very few previous studies on the phonetics of young Japanese EFL learners have examined diphthongs. This study recorded two male native speakers of General American English (NESs) and nine Japanese third-grade students (JSs), with no experience of English learning before school lessons, pronouncing each English alphabet letter. The students were recorded after receiving six lessons on the English alphabet. Their recorded productions of the letters "A, H, J, K, and O" were extracted and acoustically analyzed using Praat. The values of the first and second formant (F1 and F2) frequencies were measured at every temporal point of 25% of the vowel duration of the pronunciations of the NESs and JSs for comparison. Six out of nine JSs pronounced these English alphabet letters in the way they were pronounced in Japanese daily conversation, and their Japanese pronunciation was also digitally recorded and acoustically analyzed in the same way as their English productions.
All the young Japanese learners substituted the Japanese long vowels /e:/ and /o:/ or the vowel sequences /ei/ and /ou/ for the English diphthongs /ei/ and /ou/. The diphthong /ei/ pronounced by the NESs and the JSs and the Japanese vowel sequence /ei/ shared the direction of temporal formant changes: F1 decreased and F2 increased; thus, the ratio of F2 to F1 (formant ratio) was increased throughout the vowel portion. However, the degree of changes in the English productions of the JSs was greater than in those of the NESs, and this was also observed in the Japanese /ei/: the mean value of the difference in the formant ratio between the 25% and the 75% temporal points was 1.06 for /ei/ pronounced by NESs, whereas those of the JSs and the Japanese /ei/ were 1.31 and 1.75, respectively. This indicates that for JSs, the tongue was raised from the mid position of /e/ to the higher position in the latter part of the vowel for the English and Japanese pronunciations of /ei/ than for the NESs' diphthong. As for the diphthong /ou/, the productions of NESs decreased in both F1 and F2 in the latter half of the vowel, while F1 decreased and F2 increased in the latter part of the English and Japanese productions of JSs. This means that the back of the tongue was raised and moved backward in the latter part of the vowel /ou/ for the NESs but that it was raised and moved forward for both English and Japanese pronunciations for the JSs. The formants throughout the JSs' English diphthongs substituted by the Japanese long vowels did not change so greatly as those throughout the NESs' diphthongs or the Japanese vowel sequences.
A larger number of JSs (62.2%) substituted the Japanese long vowels than the vowel sequences for the English diphthongs. However, the substitution rates of the long vowels and the vowel sequences for the English diphthongs by the young Japanese learners differed across the alphabet. All the JSs pronounced the Japanese /ei/ for H, but only one student pronounced the vowel sequences /ei/ and /ou/ for the letters A and O, respectively; eight JSs (88.9%) preferred the long vowels for these alphabets. In contrast, the long vowel /e:/ was substituted for the English diphthong /ei/ by five JSs (55.6%) for J and seven JSs (77.8%) for K.
The findings revealed that even after six lessons of pronunciation and identification training for each alphabet letter, the diphthongs pronounced by the young Japanese learners were affected by the negative transfer of Japanese vowels. Sound fossilization occurs in the early stages of English learning without adequate instructions, constant training, or feedback of English pronunciation. Teaching the pronunciation of each alphabet letter can be an effective way for young learners to learn English sounds because half of English vowels and consonants appear in the alphabet when pronounced. From this perspective, the authors developed a teacher's manual on teaching and pronouncing the English alphabet. It was designed for Japanese elementary school teachers, most of whom have no basic knowledge of English sounds and insufficient training in teaching and pronouncing English sounds. It included a QR code of the motion pictures in which an English native speaker pronounced each letter with a Japanese explanation of how to pronounce. This manual will help teachers learn the correct pronunciation of the English alphabet and provide them with teaching points and the basic knowledge of English sounds.
