Earticle

The purpose of this study is to explore an alternative model for the TOEIC speaking test based on language ‘Complexity’, ‘Accuracy’ and ‘Fluency’ (CAF) by drawing comparisons between task-based CAF (Model 1) and communicative adequacy-based CAF (Model 2). Thirty-two college students participated in the study and carried out CAF based English speaking tests. The results of Pearson analysis revealed that Model 2 has a higher correlation with the TOEIC speaking scores than Model 1. Based on factor analysis, Model 2 was restructured, to create a new format, NCAF (New CAF). Under the NCAF model, ‘Complexity’ has three sub-items: sentence length, sentence elaboration, and lexical density; ‘Accuracy’ has error accuracy 1, and error accuracy 2; and ‘Fluency’ includes pause distribution, speed fluency, and pause frequency. To assess the relationship between the NCAF model and TOEIC speaking scores, a regression analysis was performed. The results showed that NCAF could account for 76.1% of the TOEIC speaking scores. Pedagogical implications and discussions are suggested.

The Metacognitive Awareness Listening Questionnaire(MALQ), developed and validated by Vandergrift et al. (2006), has become an important tool in research on the role of metacognition in learning foreign or second language listening. Despite its wide use, studies about the measurement characteristics of the questionnaire are rare and little is known about its measurement properties especially among Korean learners. In this paper, the psychometric properties of the MALQ were investigated in a study of 1000 college students by applying exploratory structural equation modeling. The results were then compared to those in the developers’ validation paper. While the current findings overall support the 5-dimensional structure of the MALQ, many discrepancies and problems were also identified, including: a low model fit, a high incidence of factor crossloads, the low total variance explained by the model, the possibility of missing some key factors, and the lack of explicit modeling for factor interdependence. These results are discussed in regard to their psychometric implications for the study of metacognition using the MALQ and some steps for improving the measure are suggested.

The purpose of this study is to explain that Korean learners of English are likely to pronounce English word-initial voiced stops as tensificated sounds, attributing this phenomenon to perceptual similarity and articulatory ease. As voiced stops do not exist in the word-initial position in the Korean language, if English word-initial voiced stops are realized as voiceless tensificated ones, they are sometimes mistaken for different sounds. Therefore, this research examines whether the tensification of English word-initial voiced stops by Korean learners is the effect of their lack of awareness of a difference between English phonemes and the constraints imposed by Korean loanwords. I explore the relevance of English voiced stop and the tensification observed in Korean loanwords, considering constraints associated with ease of articulation and perceptual similarity using Optimality Theory. Furthermore, I analyze these using native English speakers' phonetic sounds as input, rather than focusing on English phonemes, since this study is focused on phonetic perception, not on the underlying features of English voiced stops.

This study explored the perceptions of Korean university students studying English as a foreign language (EFL) after using the AI chatbot tool ChatGPT in an advanced writing class for one semester. A questionnaire was used to collect data from students about three learner variables (being comfortable in completing a task, being motivated to complete a task, and being confident that they can complete a task). The questionnaire also asked students if they felt their learning experience was improved by using AI tools (as opposed to classes taught only using traditional methods) and if they thought a blended method using both AI tools and traditional methods was preferable to one using strictly one or the other. The results showed that students were much more comfortable writing in English with the assistance of AI writing tools, that they felt less anxiety while completing tasks with the assistance of AI writing tools, and that they felt more confident when writing in English with the assistance of AI writing tools. They also indicated that using AI writing tools in the classroom was preferable to classes taught only in the traditional method, and that classes using a blended method were preferable to those taught using strictly AI tools or traditional methods alone.

In the traditionally male-dominated maritime industry, it is important to improve gender representation in line with the International Maritime Organization's commitment to SDG 5 on gender equality. In this study, we use corpus-based analysis to examine gender-biased and gender-inclusive language by comparing a corpus of maritime web news(MWN) to a corpus of general English news on the web(NOW). Our findings show that while inclusive terms such as "seafarer" are increasing in use, terminology inclusive to women remains underrepresented, although when used it is more often associated with leadership roles. Traditional male pronouns still dominate, emphasizing the ongoing challenges of achieving linguistic gender parity. We suggest the need for policy changes, targeted training and continued efforts to promote gender-inclusive language and true gender parity in the maritime industry.

This study investigates the semantic variations of four synonymous verbs within the BNC and COCA. By employing a combination of corpus analysis and advanced network analysis through tools like Wmatrix 5 and NetMiner 4.0, the study identifies both semantic differences and similarities between British and American English. The research reveals that, although these verbs share common semantic domains, there are notable distinctions in their usage that reflect deeper cultural and linguistic differences between the UK and the US. This study underscores the critical importance of context in language usage, demonstrating significant variations in the connotation, frequency, and function of synonymous verbs across regional dialects. By bridging static dictionary definitions with the dynamic nature of real-world language use, this research enhances the understanding of English synonymy and strongly advocates for expanded comparative linguistic studies. This approach ultimately aims to deepen comprehension of linguistic intricacies, thereby promoting more effective communication and improved language education across various English-speaking communities.

This research seeks to provide a genuine explanation of the prosodic licensing of wh-in-situ analysis in Korean. First, it is demonstrated that Korean wh-words are underspecified and possess two distinct features: definite [+wh] and indefinite [-wh]. In certain licensing contexts, wh-words can be ambiguous, having either a definite [+wh] or indefinite [-wh] value. This ambiguity determines whether the sentence functions as a wh-question (WHQ) or a yes-no question (YNQ). The selected value determines the prosodic structure associated with the corresponding sentence. This research aims to naturally describe the interaction between syntax and PF by deploying a Labeling Algorithm that encodes wh-features in the syntax.

This paper investigates verb-object idioms in Korean and their impact on the theory of verb structure. Two novel findings are presented from the results of this study. First, examining the patterns of psych idioms in Korean, I propose that Experiencers are not introduced by Voice. Instead, I suggest that Experiencers are internal arguments, based on the inalienable relationship exhibited by the body part Theme and its Possessor, which is commonly entailed in psych idioms. Second, I offer new evidence that the domain of verb-object idioms cannot exceed the first verbalizing phrase. In deriving this conclusion, I make use of Jung(2024)’s classification of verbs suffixed with -eci into two types: root-selecting and verb-selecting. I then demonstrate that only the former class of -eci verbs form an idiom with a Theme argument. The failure of verb-selecting -eci to participate in idiom formation reveals that the domain of idiosyncratic expressions must be restricted to the first verbalizing layer.