Perception And Production Of English /l/-/r/ And /l/-/n/ By Chinese EFL Learners

Researchers at home and abroad have investigated a lot into second language acquisition such as semantics and syntax(Roger 2006; Lihong Lin 2012), but still less in both areas of speech perception and production. Currently, with more and more researchers paying attention to the effect of the dialect on the second language learning, this paper studied the perception and production of English phonemic contrasts of /l/- /r/ and /l/- /n/ by Xinhua EFL learners from Hunan Province and Suihua EFL learners from Hei longjing province and aimed to disclose L1's effect on L2 phonetic acquisition to testify Flege(1987/1995)'s Hypothesis 1and Hypothesis 3 of Speech Learning Model. This study will discuss following four questions:1. To what extent do Xinhua and Suihua EFL learners perceive /l/- /r/ and /l/- /n/ correctly?2. To what extent do they produce /l/- /r/ and /l/- /n/ correctly?3. which contrast of /l/-/r/ or /l/-/n/ is better acquired ?4. How does the perception correlate with the production?In this comparing study, informants are high-school, grade-2 students with 26 from Xinhua and 26 from Suihua. They are all English as Foreign language learners and never leave their hometown, with Xinhua dialect as the former's daily language and Suihua dialect as the latter's daily language. The informants in this study are in good health and have normal intelligence, hearing and verbal abilities. The experiments include perception and production task. For the perception task, nonce words are employed to rule out the influence of memory or familiarity of real words. Target phones of each contrast are inserted in word-initial,word-medial and word-final positions with the neighboring vowels confined to one of the three / i/, /?/ and /u/, and all of the materials in this task are recorded by a native American English speaker and the recording and perception task are all conducted with the help of Senhiseer HD600 micro-earphone. Each nonce word is recorded twice and separately saved into a sound file. The informants listen to the recordings played in a random order on computer and choose a sound from AB choice which consists of a minimal pair. Two sets of stimulus materials are used in the production task, with one set of real words containing these two contrasting sounds and the other set of minimal pairs. All of the words are real words and selected from their high-school textbooks. The first set of stimulus materials are used in delayed repetition technique which were like mini-dialogues eliciting target phones in the carrier sentences. The second stimulus materials are minimal pairs written on note-cards and read by the informants. All of the productions were transcribed through Praat and assessed by two Americans, thus generating the intra-reliability of 0.88 and inter-reliability of 0.85 via SPSS19.0.The major findings are:1) When /l/-/r/ phonemic contrast is in word-initial and word-medial positions, XH and Suihua informants have a similar perceptual accuracy without significant differences.However, when this contrast is in word-final position, XH informants' perceptual mean correct percentage is apparently higher than SH informants'; In terms of /l/-/n/ phonemic contrast, when it is in word-initial and word-medial positions, XH informants' perceptual results are much worse than those of SH informants but when it's in word-final position, two groups of informants have a very high mean correct percentage in perception and it shows no difference.2) Similar to what has been found in the perceptual experiment, XH and SH informants show no obvious difference of productive accuracy when /l/-/r/ contrast is in word-initial and word-medial positions. While, SH informants produce much worse than the XH informants in dealing with the word-final /l/-/r/ phonemic contrast; In terms of /l/-/n/ phonemic contrast,when it's in word-initial and word-medial positions XH informants' production mean correct percentage is much lower than SH informants' But, there is no obvious difference when dealing with word-final l/-/n/ contrast, both XH and SH informants produce very well.3) For XH informants, word-initial and word-medial /l/-/n/ phonemic contrast poses the greatest problem for them to perceive and produce, while /l/-/r/ phonemic contrast is relatively easy; As for SH informants, due to the existence of word-final retroflex which is very similar to English word-final /r/, SH informants tend to make mistakes by assimilating the two contrasting sounds into the native phone, which is hard for them to discriminate.4) The two groups of informants' perception is strongly correlated with their production,the findings show that their mean correct percentage of perception is higher than that of production and those who can successfully perceive all these two English phonemic contrasts are very likely to produce them accurately.By doing this comparing study, we can find Xinhua EFL informants have a difficulty in discriminating two different phonemes in L2 which belong to the same phoneme in L1. Suihua EFL informants have the greatest problem in discerning the difference between the phone in L2 and the closest sound in L1, which testifies Flege(1995)'s H3 in his SLM; Moreover, both two groups present an obvious difference when dealing with the contrast in word-initial,word-medial and word-final positions, which is in support of the H1 of his model that the sounds in the L1 and L2 are related perceptually to one another at a position-sensitive allophonic level, rather than at a more abstract phonemic level. Besides, the study also lends support for the proposition of SLM that the production accuracy of L2 is based on the perception accuracy of L2.