General Language Research Articles

Low- to mid-level high school first-year EFL learners’ growth language mindset, grit, burnout, and engagement: Using serial mediation models to explore their relationships

English-as-a-foreign-language (EFL) learning is a journey that rewards those with positive psychological qualities. This study investigated the relationships among low- to mid-level high school EFL learners' growth language mindset (GLMS), grit, burnout, and engagement, to gain insights into how learner beliefs, personality traits, and emotions affect EFL learning. A total of 868 first-year students from four schools located away from China's central cities participated in the questionnaire survey. Their data were analyzed using structural equation modelling (SEM). The serial mediation models show, among others, that 1) GLMS predicted L2 burnout directly and via L2 grit indirectly and influenced L2 engagement indirectly via L2 grit and L2 burnout, with second language aptitude beliefs (L2B) having more effects than general language intelligence beliefs (GLB), 2) in general, POE (perseverance of effort) had stronger effects than COI (consistency of interest), and 3) L2 burnout showed negative associations with GLMS, L2 grit, and L2 engagement and worked as an important mediator contributing to indirect relationships leading to L2 engagement. The theoretical and pedagogical implications of these findings are discussed. We especially call for support given to learners at low- to mid-level senior high schools to build their positive psychology.

Evolution and advancements in deep learning models for Natural Language Processing

This paper provides a comprehensive review of the evolution and advancements in deep learning models for Natural Language Processing (NLP). It explores the transition from statistical models to neural networks, highlighting the paradigm shift towards data-driven methodologies and the implications for NLP tasks. The emergence of neural network architectures, such as Recurrent Neural Networks (RNNs) and transformer-based models like BERT and GPT, has revolutionized language understanding and generation. Furthermore, the integration of deep learning in traditional NLP tasks, such as part-of-speech tagging and named entity recognition, has led to significant improvements in accuracy and efficiency. The paper also discusses the quantitative analysis of deep learning models, including performance metrics, computational efficiency, and mathematical modeling of language tasks. Case studies and applications, including sentiment analysis, machine translation, and automated content generation, exemplify the transformative impact of deep learning in NLP.

An Evaluation of Khanmigo, a Generative AI Tool, as a Computer-Assisted Language Learning App

The recent advancement in technology has attracted learners’ attention worldwide to Generative Artificial Intelligence (GenAI) for educational purposes. While GenAI has shown promising results for general language purposes (Godwin-Jones, 2023; Xiao & Zhi, 2023), the potential of GenAI for language learning has not been fully explored. This paper, therefore, endeavors to decipher the potential of a GenAI app, Khanmigo, as a language learning tool, specifically for learning French. The app was analyzed by the researcher through her interactions of about 17.5 hours using Chapelle’s (2001) Evaluation Framework for discerning the task appropriateness of a given Computer-Assisted Language Learning (CALL) tool. While the app does not show robust performance in all the six criteria suggested for evaluation, it still holds some promise.

A Wolf in Sheep’s Clothing? Critical Discourse Analysis of Five Online Automated Paraphrasing Sites

Research on academic integrity used to focus more on student character and behaviour. Now this research includes wider viewing of this issue as a current teaching and learning challenge which requires pedagogical intervention. It is now the responsibility of staff and institutions to treat the creation of a learning environment supporting academic integrity as a teaching and learning priority. Plagiarism by simply copying other people’s work is a well-known misconduct which undermines academic integrity; moreover, technological developments have evolved plagiarism to include the generation and copying of computer-generated text. Automated paraphrasing tool (APT) websites have become increasingly common, offering students machine-generated rephrased text that students input from their own or others’ writing. These developments present a creeping erosion of academic integrity under the guise of legitimate academic assistance. This also has implications for arrival of large language model (LLM) generative AI tools. In accessing these sites, students must discern what is a legitimate use of the tool and what may constitute breaching academic integrity. This study critically analysed the text from five online paraphrasing websites to examine the discourses used to legitimise and encourage APT use in both appropriate and inappropriate ways. We conceptualised these competing discourses using Sheep and Wolf metaphors. In addition, we offer a metaphor of the Educator as a Shepherd to become aware of APT website claims and assist students to develop critical language awareness when exposed to these sites. Educators can assist students with this through knowledge of how these sites use language to entice users to circumvent learning.

Refining the Assessment Literacy Competence among K12 EFL Teachers

Language assessment literacy as a relatively new term denotes teachers’ understanding of the assessment construct both theoretically and practically. Given the importance of assessment in the teaching and learning process, the main purpose of the current study is to investigate the general and skill-based language assessment literacy levels of EFL teachers working at K12 schools in Türkiye and to analyse their needs both in pre-service and in-service years. Considering the purpose of the study, a mixed-method research design was employed. At the initial phase of the study, the quantitative data were collected by means of the Language Assessment Knowledge Scale (LAKS), and 272 EFL teachers working at K12 schools in Türkiye participated. Descriptive and inferential analyses were run, and results showed that the participants’ mean score in LAKS was 31.59 out of 60. The participants demonstrated the highest level of knowledge in assessing reading, while assessing listening was identified as the area where teachers had the least expertise. In the second phase, semi-structured interviews were conducted with volunteers to learn more about the teachers’ opinions on language assessment knowledge and their needs. Qualitative findings complemented the findings from the questionnaire by revealing details about the reasons of the participants’ scores in LAK scale.

SEMANTIC DYNAMICS OF BUSINESS IN STYLISTICS OF SOCIAL COMMUNICATION

The article is devoted to the linguistic and stylistic motivation of the "business" nature in the time and space of culture. The research idea regarding the need for professional analysis of the evolution of business is formulated and argued: from the primary meaning of the word to the modern one, which reflects the values and priorities of the development of the modern world, verbalized as a result of active processes of literary language development. The functioning of "business" concept in heterogeneous communicative fragments — in the structure of word combinations and fixed expressions — is studied, which explains the nature and scope of its application and the effective expediency of its use in multilingual contexts. The status of business in the language consciousness of society and the general state of business activity are considered through a diachronic comparison of the semantics of the term "business" in general language dictionaries of different editions. Therefore, attention is focused on common and professional definitions of the concept of business, their features and differences. For the dynamics of the objectivity of stylistic studies of business communication, it is determined that it is relevant to characterize the nature of business as a phenomenon, its strategies and tactics, which appear as systematized communicative intentions of speakers and determine effective models of their verbal interaction in this area. At this stage, we consider such tasks to be promising for further research, since business linguistics aspectology is projected onto the system of current issues that determine the vectors of development of national linguistics. Taking into account the development of multimedia technologies, which today actively affect all spheres of public communication, determine the dynamic nature of their functional and stylistic adaptation to the priorities and values of the mass audience, the task of the research perspective of creating and implementing an effective functional model of business communication, in particular, in the media space, was formulated.

Prompt response variation in learner corpus research: Implications for data interpretation

While general first language corpora are composed of samples from various naturalistic sources (e.g., websites, books), language samples in most written learner corpora (LC) are texts produced in response to prompts. In this context, LC users need to develop a clear awareness of the affordances and limitations of specific prompts and how responses to said prompts may affect the investigation of their intended object(s) of study. Through an analysis of the presence/absence of specific Spanish verb tenses in texts written in response to two supposedly narrative prompts in a Spanish LC (COWS-L2H; Yamada et al., 2020), this article illustrates the impact of inter- and intra-prompt response variation on LC data interpretation. Based on this evidence, we caution against rapid assumptions about text content based solely on the superficial phrasing of LC writing prompts. Instead, we recommend that LC users perform in-depth quantitative and qualitative analyses of learners’ samples written in response to each prompt they aim to include in their study prior to running statistical models on those data.

The life cycle of large language models in education: A framework for understanding sources of bias

AbstractLarge language models (LLMs) are increasingly adopted in educational contexts to provide personalized support to students and teachers. The unprecedented capacity of LLM‐based applications to understand and generate natural language can potentially improve instructional effectiveness and learning outcomes, but the integration of LLMs in education technology has renewed concerns over algorithmic bias, which may exacerbate educational inequalities. Building on prior work that mapped the traditional machine learning life cycle, we provide a framework of the LLM life cycle from the initial development of LLMs to customizing pre‐trained models for various applications in educational settings. We explain each step in the LLM life cycle and identify potential sources of bias that may arise in the context of education. We discuss why current measures of bias from traditional machine learning fail to transfer to LLM‐generated text (eg, tutoring conversations) because text encodings are high‐dimensional, there can be multiple correct responses, and tailoring responses may be pedagogically desirable rather than unfair. The proposed framework clarifies the complex nature of bias in LLM applications and provides practical guidance for their evaluation to promote educational equity. Practitioner notesWhat is already known about this topic The life cycle of traditional machine learning (ML) applications which focus on predicting labels is well understood. Biases are known to enter in traditional ML applications at various points in the life cycle, and methods to measure and mitigate these biases have been developed and tested. Large language models (LLMs) and other forms of generative artificial intelligence (GenAI) are increasingly adopted in education technologies (EdTech), but current evaluation approaches are not specific to the domain of education. What this paper adds A holistic perspective of the LLM life cycle with domain‐specific examples in education to highlight opportunities and challenges for incorporating natural language understanding (NLU) and natural language generation (NLG) into EdTech. Potential sources of bias are identified in each step of the LLM life cycle and discussed in the context of education. A framework for understanding where to expect potential harms of LLMs for students, teachers, and other users of GenAI technology in education, which can guide approaches to bias measurement and mitigation. Implications for practice and/or policy Education practitioners and policymakers should be aware that biases can originate from a multitude of steps in the LLM life cycle, and the life cycle perspective offers them a heuristic for asking technology developers to explain each step to assess the risk of bias. Measuring the biases of systems that use LLMs in education is more complex than with traditional ML, in large part because the evaluation of natural language generation is highly context‐dependent (eg, what counts as good feedback on an assignment varies). EdTech developers can play an important role in collecting and curating datasets for the evaluation and benchmarking of LLM applications moving forward.

Developing neutral linguistic resources for the implementation of an automatic transformational analyzer

Bakground The linguistic pursuit of describing natural languages stands as a commendable scientific endeavor, regardless of immediate software application prospects. It transcends mere documentation of possible sentences to establish connections between sentences derived from transformations. Methods Amid the dominance of Large Language Models (LLMs) in research and technology, which offer intriguing advancements in text generation, the approaches presented in this article confront challenges like opacity, limited human intervention, and adaptation difficulties inherent in LLMs. The alternative or complementary approaches highlighted here focus on the theoretical and methodological challenges of describing linguistic transformations and are firmly rooted in the field of linguistics, the science of language. We propose two solutions to address the problem of language transformations: (i) the procedural approach, which involves representing each transformation with a transducer, and (ii) the declarative method, which entails capturing all potential transformations in a single neutral grammar. Results These approaches simplify the generation of complex sentences from elementary ones and vice versa. Conclusion This work has benefited from research exchanges within the Multi3Generation COST Action (CA18231), and the resources produced can contribute to enhancing any language generation system.

Effects of transcranial direct current stimulation on different cognitive domains in Alzheimer's disease: a meta-study.

Numerous studies have investigated the potential effects of transcranial direct current stimulation (tDCS) on improving symptoms related to Alzheimer's disease (AD). However, these studies have produced inconsistent results, leading to a need for further investigation. A comprehensive search was conducted, including articles published from the initial availability date to 5 April 2024. The extracted study data were analyzed using STATA 12.0 software. The standard mean difference (SMD) and a 95% confidence interval (CI) were calculated to assess the effects of tDCS. A total of 18 studies assessing the effects of tDCS on AD were included in the study. The study revealed that tDCS has an immediate positive impact on general cognitive, executive, language, and visuospatial function. However, the study did not observe any other significant effect of tDCS treatment on improvements in brain function, including long-term effects on general cognitive, attention, language, and memory function, as well as immediate effects on attention and memory function. In conclusion, the study suggests that tDCS may be a promising intervention for improving the cognitive function of patients with AD. However, given the complex and multifactorial nature of AD, further well-designed studies with larger sample sizes are necessary to clarify the effectiveness of tDCS and determine the optimal combination of tDCS parameters.

A comparative study of critical thinking implementation in face-to-face and online classrooms: Medical students’ critical thinking and reading comprehension

Background: The enhancement of critical thinking (CT) is an essential goal in the educational strategic framework of medical universities across the globe. Therefore, it is of utmost importance to evaluate the competence in CT abilities as well as the inclination of medical students to actively participate in CT practices. Moreover, educational decision-makers believe that granting medical students the opportunity to utilize technology can enhance their academic performance and stimulate their CT ability level. Hence, this research aimed to compare the influence of implementing a CT framework in traditional classroom settings and online platforms on the CT and reading comprehension of medical students. Methods: Over four months, two groups of 60 medical students (30 students in each group) were exposed to two environments of face-to-face and online English language instruction. The current quasi-experimental study was conducted at a general English language classroom, situated in Islamic Azad University, Kerman branch. Pre-tests and post-tests of CT and reading comprehension were used to measure learners’ improvements in CT and reading comprehension levels before and after the project. Results: The results of this study showed some improvements in the CT and reading comprehension abilities of both face-to-face and online groups (P<0.01). However, the improvements were particularly noticeable in the online group regarding the reading comprehension (M=17.98) and CT (M=107.462) variables compared to the face-to-face group (M=11.39) and (M=51.438). Conclusion: Medical education experts strongly believe that developing and nurturing CT skills among medical students is crucial for improving the overall quality of medical education. In today’s globalized era, the importance of CT cannot be overstated, as it equips students with essential skills needed to navigate complex challenges. Integrating CT into online learning environments is a testament to the progressive nature of medical education in the modern world.

Detecting LLM-assisted writing in scientific communication: Are we there yet?

Abstract Large Language Models (LLMs), exemplified by ChatGPT, have significantly reshaped text generation, particularly in the realm of writing assistance. While ethical considerations underscore the importance of transparently acknowledging LLM use, especially in scientific communication, genuine acknowledgment remains infrequent. A potential avenue to encourage accurate acknowledging of LLM-assisted writing involves employing automated detectors. Our evaluation of four cutting-edge LLM-generated text detectors reveals their suboptimal performance compared to a simple ad-hoc detector designed to identify abrupt writing style changes around the time of LLM proliferation. We contend that the development of specialized detectors exclusively dedicated to LLM-assisted writing detection is necessary. Such detectors could play a crucial role in fostering more authentic recognition of LLM involvement in scientific communication, addressing the current challenges in acknowledgment practices.

COGNITIVE-MATRIX ANALYSIS AS A TOOL FOR EXPLORING CULTURAL IDENTITY OF A FICTION AUTHOR

The integration of artificial intelligence (AI) technologies into the educational system in general and foreign language teaching in particular makes it possible to significantly enhance students’ foreign language practice and create conditions for more effective formation of foreign language communicative competence components. At the same time, the gradual implementation of AI tools in the educational process, along with its obvious advantages, may cause teachers’ concerns related to the gradual exclusion of the teacher from the learning process and their complete replacement by AI. As a modern paradigm of education, the authors suggest the triad “learner - teacher - artificial intelligence”, within which the interaction of learners with AI is built into traditional teaching methods, enhancing the educational process with new additional opportunities for learners’ foreign language practice. For objective reasons, little time and attention is often allocated to teaching learners to write in a foreign language. AI, which has significant language didactic potential, is able to take over the teacher’s functions of evaluating students’ written work and providing them with feedback in the form of an AI-generated version of an essay and/or recommendations for the revision of written work. In this paper, the authors propose 11 stages of a method for teaching learners to write creative works in a foreign language in the triad “learner - teacher - artificial intelligence”. Special attention in the method is given to the following aspects, which are new for teaching practice: a) compliance of students with the norms of authorship ethics when receiving various types of feedback from AI; b) the need to discuss the results of students’ interaction with AI during classroom sessions. At the same time, the authors note that the teacher is not excluded from the process of teaching learners a foreign language and evaluating their written work. The traditional method of teaching students’ written speech is enriched with additional stages of interaction with artificial intelligence, creating more conditions for students to develop the necessary foreign language speech skills.

On Representation Learning-based Methods for Effective, Efficient, and Scalable Code Retrieval

Code retrieval consists of finding relevant code snippets regarding a programmer’s query — an increasingly important task due to software ubiquity. Although significant progress has been made, there is still a need for refinement and improvement in current code retrieval solutions. Besides effective, we argue that solutions for the code retrieval task should be efficient and scalable to ever-increasing large code repositories. This paper introduces xCoFormer, the first representation learning-based model for the code retrieval task in the literature, which meets all these requirements. Our contributions include: (i) an interactive tag-training method that efficiently places a query closer to its relevant codes in the embedding space and (ii) use of a specialized loss function (N-pair) better suited to the code retrieval task. To evaluate our proposal, we conducted experiments with different versions of xCoFormer against several state-of-the-art deep learning models in datasets with different properties. In our experiments, xCoFormer produced the best cost-effectiveness tradeoff — it is as effective as our closest competitor while being thousands of times faster at search time. xCoFormer’s attention mechanism also (partially) meets the desirable requirement of explainability by exposing the “reasoning” of the model’s predictions. Finally, when used as a “proxy” method, fine-tuned with domain-specific code language models, such as Unixcoder and CodeBERT, our proposal achieves gains of more than 33% in effectiveness when compared to the use of general purpose language models such as BERT and RoBERTa.

THE IMPACT OF THE ETYMOLOGICAL STUDY OF THE CONSTITUENTS OF THE CONCEPT LANGUAGE POLICY ON THE FORMATION OF A CONCEPTUAL METAPHOR

The article is dedicated to the study of the semantic features of the constituent lexemes language and policy in the English worldview, which are analysed as components of the concept LANGUAGE POLICY, as well as the peculiarities of their influence on the formation of aconceptual metaphor. It has been established that the word combination language policy is rarely found in general language dictionaries, so it is appropriate to study its lexemes-constituents by their origin in order to find out how their cognitive components interact with each other. 17 and 22 definitions of the lexeme-constituents language and policy have been analysed, respectively. It is noted that the cognitive component a course of action of the lexeme-constituent policy explains the meaning of policy as directed actions on this or that issue. Cognitive components body of words; body of methods of combination of words and used by a nation, people, or race of the lexemes-constituent language explain language as a grouping of words, methods of their combination used by people. It has been established that the expediency of such an analysis consists in a detailed study of the verbalizers of the nominative field of the concept LANGUAGE POLICY. This requires a comprehensive analysis of the concept, taking into account qualitative and quantitative studies. Several conceptual metaphors are identified such as ECOLOGY, LANGUAGE PLANNING IS AN ORGAN, LANGUAGE IS STATUS, LANGUAGE IS FIGHT,MULTILINGUALISM IS MOSAIC, MULTILINGUALISM IS CONSTELLATION and LANGUAGE IS A FLOWER.It was noted that the etymological analysis of the combination of lexemes-constituents narrowed the meaning of lexemes in such a combination, which affected the LANGUAGE POLICY conceptual metaphor.

Spontaneous Generation of Unconventional Language and Its Link with Grammatical Performance in Chinese Adults With and Without ASD.

This study investigated the generation of unconventional language in the spontaneous speech of Chinese adults with autism spectrum disorder (ASD), and how it was related to their grammatical performance, when compared to neurotypical (NT) controls. Twenty Cantonese-speaking adults with ASD and 20 NT controls completed three interview tasks in the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2), and their spontaneous speech was recorded and transcribed. Utterances containing unconventional language (neologisms, idiosyncratic phrases, and pedantic language), morphosyntactic errors, mean length of utterance (MLU), and mazes were computed. The ASD group produced more neologisms, idiosyncratic phrases, and pedantic language than the NT group and their grammatical difficulties were shown in shorter MLU but not morphosyntactic errors. Mazes were more frequent in the ASD than the NT group. While the use of unconventional language increased with MLU in the NT group, it correlated positively with mazes in the ASD group. Generation of unconventional language, particularly pedantic language, in Cantonese-speaking NT adults is linked to more advanced grammar, while it appears to be a common speech characteristic among autistic speakers regardless of individual grammatical performance.

Unsupervised evolution of protein and antibody complexes with a structure-informed language model.

Large language models trained on sequence information alone can learn high-level principles of protein design. However, beyond sequence, the three-dimensional structures of proteins determine their specific function, activity, and evolvability. Here, we show that a general protein language model augmented with protein structure backbone coordinates can guide evolution for diverse proteins without the need to model individual functional tasks. We also demonstrate that ESM-IF1, which was only trained on single-chain structures, can be extended to engineer protein complexes. Using this approach, we screened about 30 variants of two therapeutic clinical antibodies used to treat severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. We achieved up to 25-fold improvement in neutralization and 37-fold improvement in affinity against antibody-escaped viral variants of concern BQ.1.1 and XBB.1.5, respectively. These findings highlight the advantage of integrating structural information to identify efficient protein evolution trajectories without requiring any task-specific training data.

Investigating the increase of violent speech in Incel communities with human-guided GPT-4 prompt iteration

This study investigates the prevalence of violent language on incels.is. It evaluates GPT models (GPT-3.5 and GPT-4) for content analysis in social sciences, focusing on the impact of varying prompts and batch sizes on coding quality for the detection of violent speech. We scraped over 6.9M posts from incels.is and categorized a random sample into non-violent, explicitly violent, and implicitly violent content. Two human coders annotated 3, 028 posts, which we used to tune and evaluate GPT-3.5 and GPT-4 models across different prompts and batch sizes regarding coding reliability. The best-performing GPT-4 model annotated an additional 45, 611 posts for further analysis. We find that 21.91% of the posts on the forum contain some form of violent language. Within the overall forum, 18.12% of posts include explicit violence, while 3.79% feature implicit violence. Our results show a significant rise in violent speech on incels.is, both at the community and individual level. This trend is particularly pronounced among users with an active posting behavior that lasts for several hours up to one month. While the use of targeted violent language decreases, general violent language increases. Additionally, mentions of self-harm decline, especially for users who have been active on the site for over 2.5 years. We find substantial agreement between both human coders (κ = 0.65), while the best GPT-4 model yields good agreement with both human coders (κ = 0.54 for Human A and κ = 0.62 for Human B). Overall, this research offers effective ways to pinpoint violent language on a large scale, helping with content moderation and facilitating further research into causal mechanisms and potential mitigations of violent expression and online radicalization in communities like incels.is.

Reform and Development of English Teaching in Colleges and Universities with the Integration of ChatGPT Technology

As an artificial intelligence-powered language generation model based on natural language processing, ChatGPT, like other artificial intelligence technologies, possesses super real-time dialogue capabilities and is considered an innovative and intelligent auxiliary teaching tool in the field of language learning. This article first analyzes the current state of development of artificial intelligence technologies such as ChatGPT. Secondly, it examines the challenges encountered in the reform and development of English teaching in colleges and universities in China. Lastly, it summarizes the applications and recommendations for using ChatGPT technology in college-level English teaching.

Evaluating Generative Artificial Intelligence on Multilingual Sentiment Analysis

This study shows that natural language processing (NLP), large language models (LLM), and generative Artificial Intelligence are becoming the most important. LLM is considered essential for many NLP tasks, including question-answering, classification, interpretation, and writing. New LLMs must be able to analyze and create documents in different languages as they know how to train training materials in different languages such as ChatGPT, BLOOMZ, and others. Considering how frequently LLMs are used, it is critical to assess the effectiveness in multi-lingual environments. In a zero-shot context, the present generative models are still useful in producing text in Indian languages. On the other hand, generative models routinely outperform human quality-based evaluation when it comes to English language generation and Indian languages. LLMs are not meant to be used in a zero-shot manner in downstream applications due to poor generating performance.