Messages behind the Unheard Sounds: Crossing the Word Boundaries through Songs
摘要
The knowledge and application of reduced forms in word recognition is a decisive factor for English learners to attain a higher language proficiency. Listening comprehension for non-native English speakers involves aspects of diction, syntax, grammar and content. Language in the spoken form is more challenging in that learners have to discern sounds that tend to cross the word boundaries for fluency. Therefore, the ability to comprehend and identify correctly what is said is the key to understanding spoken English. To fully comprehend a naturally pronounced sentence well, ESL/EFL learners need a basic knowledge of reduced forms including C-C linking (elision), C-V linking, /h/-deletion, contraction, palatalization, flapping, and the like. An appropriate approach for students to improve listening comprehension is through songs (Crawford, 2006). The present study is a language experiment practiced on two homogeneous groups of sophomore English majors who aim at passing high-intermediate GEPT in the junior year. Both groups were given the intermediate GEPT listening test to establish their initial listening proficiency. The experimental group then underwent a six-session (240 minutes) training course, in which the afore-mentioned six reduced forms were taught through twelve songs in Hit Parade Listening (Kumai & Timson, 2003, 2010). The techniques and knowledge of reduced forms were explicitly taught through lectures and dictation exercises of song lyrics. Two forms of a dictation test of 17 digitally recorded sentences of similar reliability coefficient, containing the six targeted connected speech patterns, were developed as the pretest and the posttest, respectively. The results showed the subjects of the experimental group made overall significant improvement in recognizing spoken words than the control group after receiving the explicit connected speech instruction using the Hit Parade Listening Class CD and Song CD. Regarding the error types, in the pre-instructi