We investigated the possibility of whether there is a post-vocalic ‘r’ in phonological representations of Korean L2 English speakers and the extent to which they exploit their knowledge of mapping graphemes onto phonemes within each correspondence between orthography and phonemes. First, the results obtained in the pseudohomophone task showed that R-items were responded to with higher accuracy and shorter RT than Non-R items. It suggests that there is no post-vocalic ‘r’ in phonological representations of Korean L2 speakers unlike Australian native English speakers and that Korean L2 learners are truly non-rhotic speakers. Another striking finding is that accuracy and RT for visual lexical access varied depending on the transparency between orthography and its corresponding phonemes. This indicates that Korean L2 speakers’ knowledge about the association of graphemes and phonemes varies depending on each type of correspondence between spellings and phonemes. Finally, it was found that the frequency of the base words also affected the retrieval of words along with the orthographic depth in grapheme-to-phoneme correspondences.

This study investigates how the notion of feature geometry can be effectively incorporated into the instantiation of general constraint definition in Optimality Theory. To this end, we examine three types of palatalization observed across languages: the fronting of velar, the change of place within coronals, and the addition of secondary palatalization. For the first two cases, unlike previous work by Clements and Hume (1995) and Lahiri and Evers (1991), we propose that both of those processes involve the spreading of the whole [coronal] node only of the triggering vowel onto C-place node of the coronal consonant. Furthermore, unlike Iverson (1993), we also distinguish two types of coronalization from the secondary articulation as in Acadian French. Following Clements and Hume (1995), we assume that the secondary articulation is the result of spreading the coronal node onto the V-place of the consonant with the primary place intact. This pattern opens the way to put another featural node [C-place] into use in the OT constraints. This study confirms a possibility to take advantage of feature geometry in specifying the substance of constraints.

This study investigated the family size effect on English word processing via visual lexical decision task with three different groups of speakers, i.e., L1 English speakers, Korean L2 English learners, and English bilinguals. For English simple nouns, verbs, and adjectives, we examined the effects of the type count of morphologically related members and the surface base-frequency on lexical processing. First, results showed that the family size effect emerged in Korean L2 learners, but it was mostly inhibitory. To be specific, words with a large family size elicited slower response latencies than those with a small family size. However, the facilitatory effect arose for bilinguals and native speakers of English. Second, it was exhibited that high-frequency base words were recognized more quickly than low-frequency counterparts, confirming that token frequency as well as type frequency codetermines their recognition latency. These findings suggest that L1, bilinguals, and L2 learners' mental lexicons are organized by morphologically related words along with surface frequency although their effect size differs depending on the amount of language exposure. Finally, building on the results in the current experimental study, we propose a formal account for the processing advantages of words with a large family size under a psycholinguistic model and processing constraints.

This study probes whether relative frequency is a potential factor which can determine Korean L2 speakers' access to morphologically complex words and whether their processing pattern varies according to affix type and participants' L2 proficiency. A processing experiment was conducted whereby Korean L2 speakers rated English affixed words as more complex and measured their percentages and reaction time. Our results exhibited that subjects judged as more complex the forms where the whole derived words were less frequent than their embedded base forms and that the words which were more frequent than their base forms were judged as less complex in a robust manner like L1 speakers. Further, such a trend did not vary according to their proficiency level and affix type. These findings provide supportive evidence that relative frequency facilitates L2 speakers' processing of morphologically complex words. Additionally, they imply that they access L2 derived words through decomposition to constituents such as base forms, supporting dual-route model. Finally, the relative frequency might emerge from an earlier stage of L2 development, coupled with the possibility that affixed words are represented along with frequency of the whole words and their constituents.

This study examines whether formulaic sequences might be stored and retrieved as wholes in a holistic fashion rather than analytically. Another goal is to see whether the processing of L2 sequences is affected by L2 proficiency, grammaticality and length of expressions. Grammaticality judgment task was conducted and data was collected from Korean L2 English learners of low- and intermediate-level of proficiency. First, results showed that formulaic sequences were responded more rapidly and accurately than nonformulaic sequences. This provides additional support for the claim that formulaic sequences are stored and processed as single lexical units even in L2 speakers' mental lexicon. Second, the formulaicity effect was constant at both levels, indicating that formulaic sequences are acquired and processed from the early stage of L2 development. Finally, it was found that grammatical sequences were processed more quickly and accurately than ungrammatical ones and that shorter expressions were judged faster and more accurately than longer ones, implying more efficient processing and less memory load for shorter and grammatical expressions. Building on the results in the present study, we propose a formal account for the quicker and accurate processing of formulaic sequences with reference to psycholinguistic model and processing constraints.

This study examines the production of English word-final codas by Korean speakers. First, we investigate how Korean learners phonetically realize English word final codas. The second question explores what factors exert an influence on production type. Results show that Korean speakers' production most favors the released type and then prefers the release and vowel epenthesis type and the unreleased type. This finding indicates that Korean learners are making progress toward targetlike production while retaining the transfer from Korean syllable structures. Finally, their production type significantly differs by the number of codas, grammatical condition, preceding vowel, and following linguistic environment.