Abstract: Logographic Chinese and alphabetic English exhibit a sharp contrast in both phonology and orthography. Although about eighty-five percent of present-day Chinese characters are phonetic compounds containing a phonetic component that may give information about the pronunciation of the compound, estimates of the validity of this information reveal that only thirty-eight percent of phonetic components are helpful for the pronunciation of whole characters (Perfetti & Tan, 1999). However, the pronunciation of most English words is known largely predictable based upon grapheme-to-phoneme conventions. Given the huge difference, what is the cognitive and neural-anatomical mechanism when those bilinguals, whose native language is Chinese and who learn English as a second language (L2), are performing a phonological task on English words has attracted much attention in cognitive science. This paper was to find out the possible approaches adopted by those bilinguals in L2 phonological process and to locate the brain areas activated in the process as well.
Twenty-nine Chinese post-graduates (13 females and 16 males; aged 23 to 24), who were late bilinguals presumably with high proficiency in English, participated in the experiment. They were asked to judge the rhymes of real-word pairs, which were of high frequency, regular grapheme-phoneme correspondence and high lexicality, and meanwhile, they were given some pronounceable pseudoword pairs and control-letter pairs (consonant-letter strings) to judge too. Functional magnetic resonance imaging (fMRI) was employed to scan their brain areas activated in the process, and E-prime was used to record their behavioral performance. The fMRI experiment was done in 6 sessions, each of which contained 3 blocks respectively corresponding to rhyme decision of read-word pairs, pronounceable pseudoword pairs, and font size decision of control-letter pairs.
The results obtained from behavioral measurement and fMRI analysis showed that these late bilinguals were at approximately the same speed in judging the real-word pairs and pseudoword pairs (MD=3.27; p=0.932>0.05), but much faster in judging the control-letter pairs (MD=214.95; P<0.01); that their brain areas activated in L2 phonological process involved the left dorsal lateral frontal system, the left ventral prefrontal system, the dorsal aspect of left inferior parietal system, and the bilateral ventral occipitotemporal system where the phonological process of their mother tongue works; and that the brain areas activated as a whole were strongly left-lateralized , but the areas in occipital lobe (Asymmetric Index: -0.10) and cerebellum (Asymmetric Index: -0.03) were slightly right-lateralized.
In a word, the late proficient bilinguals were found to employ the same brain areas in phonological process for both their mother tongue and L2, and they appeared to be taking an assembled approach rather than an addressed approach to L2 phonological process.

Key words: fMRI, second language acquisition, phonology