Language Data Sources Research Articles

Transcending cultural barriers in Thai mythology: Unveiling the intercultural dynamics through the “Home Sweet Home” video game’s translation

This study examines the translation strategies employed for culture-specific terms from Thai to English in episodes 1 and 2 of the video game "Home Sweet Home" (2017–2023). Applying a thematic approach and textual analysis, the source language data were categorized into six groups: ghosts, humans, ritual and religious acts, abstract religious concepts, items, and places. The findings indicate that the most frequently used strategy was absolute universalization (41.5%), followed by linguistic (non-cultural) translation (27.7%), limited universalization (16.2%), deletion (11.5%), orthographic adaptation (2.3%), and intratextual gloss (0.8%). However, no instances of repetition, extratextual gloss, synonymy, or autonomous creation strategies were identified. The discussion explores why different strategies were used for various phrasal tokens. The study also highlights five major topics concerning video game translation: universalization versus preservation, translatability and untranslatability, language limitations, cultural barriers in the target language, and the game's future translations. It can be concluded that a comprehensive understanding of the language and culture used in the game and its translation is critical for the game's success and effectiveness for both players and creators. Furthermore, in the digital era, where postmodern imaginations and reinterpretation of Thai folk beliefs and superstitions can boost the economy, numerous opportunities for creative soft powers are always available.

Documenting Geographically and Contextually Diverse Language Data Sources

Documenting Geographically and Contextually Diverse Language Data Sources

Effectiveness of student academic presentations: What do we learn from presentation guidebooks?

This research is based on a survey of recently published presentation guidebooks designed for both general academic readership as well as scientific and technical communities. The primary objectives were 1) to examine how Anglophone academic culture views effectiveness of academic presentations and, respectively, determine the shared core principles which contribute to it; 2) to identify additional explicitly stated aspects of effective presentations, both general and/or area-specific; 3) to get a better sense of the generalizability of the language recommendations in the guidebooks by looking into whether they were derived from large language data sources or were based primarily on the authors’ subjective language observations.Overall, the survey revealed that Anglophone academic culture has a consistent view of several core principles on which effective and successful academic presentations should be built. The guidebooks also offered a shared set of additional characteristics, associated with both the preparation and delivery stage of effective presentations. However, the majority of the surveyed guidebooks lacked in the depth and breadth of their treatment of the language use in presentations. The findings of the survey can be beneficial to instructors and novice presenters alike.

CL-NERIL: A Cross-Lingual Model for NER in Indian Languages (Student Abstract)

Developing Named Entity Recognition (NER) systems for Indian languages has been a long-standing challenge, mainly owing to the requirement of a large amount of annotated clean training instances. This paper proposes an end-to-end framework for NER for Indian languages in a low-resource setting by exploiting parallel corpora of English and Indian languages and an English NER dataset. The proposed framework includes an annotation projection method that combines word alignment score and NER tag prediction confidence score on source language (English) data to generate weakly labeled data in a target Indian language. We employ a variant of the Teacher-Student model and optimize it jointly on the pseudo labels of the Teacher model and predictions on the generated weakly labeled data. We also present manually annotated test sets for three Indian languages: Hindi, Bengali, and Gujarati. We evaluate the performance of the proposed framework on the test sets of the three Indian languages. Empirical results show a minimum 10% performance improvement compared to the zero-shot transfer learning model on all languages. This indicates that weakly labeled data generated using the proposed annotation projection method in target Indian languages can complement well-annotated source language data to enhance performance. Our code is publicly available at https://github.com/aksh555/CL-NERIL.

Quantifying Changes in the Language Used Around Mental Health on Twitter Over 10 Years: Observational Study.

BackgroundMental health challenges are thought to affect approximately 10% of the global population each year, with many of those affected going untreated because of the stigma and limited access to services. As social media lowers the barrier for joining difficult conversations and finding supportive groups, Twitter is an open source of language data describing the changing experience of a stigmatized group.ObjectiveBy measuring changes in the conversation around mental health on Twitter, we aim to quantify the hypothesized increase in discussions and awareness of the topic as well as the corresponding reduction in stigma around mental health.MethodsWe explored trends in words and phrases related to mental health through a collection of 1-, 2-, and 3-grams parsed from a data stream of approximately 10% of all English tweets from 2010 to 2021. We examined temporal dynamics of mental health language and measured levels of positivity of the messages. Finally, we used the ratio of original tweets to retweets to quantify the fraction of appearances of mental health language that was due to social amplification.ResultsWe found that the popularity of the phrase mental health increased by nearly two orders of magnitude between 2012 and 2018. We observed that mentions of mental health spiked annually and reliably because of mental health awareness campaigns as well as unpredictably in response to mass shootings, celebrities dying by suicide, and popular fictional television stories portraying suicide. We found that the level of positivity of messages containing mental health, while stable through the growth period, has declined recently. Finally, we observed that since 2015, mentions of mental health have become increasingly due to retweets, suggesting that the stigma associated with the discussion of mental health on Twitter has diminished with time.ConclusionsThese results provide useful texture regarding the growing conversation around mental health on Twitter and suggest that more awareness and acceptance has been brought to the topic compared with past years.

In Search of State and Trait Emotion Markers in Mobile-Sensed Language: Field Study

BackgroundEmotions and mood are important for overall well-being. Therefore, the search for continuous, effortless emotion prediction methods is an important field of study. Mobile sensing provides a promising tool and can capture one of the most telling signs of emotion: language.ObjectiveThe aim of this study is to examine the separate and combined predictive value of mobile-sensed language data sources for detecting both momentary emotional experience as well as global individual differences in emotional traits and depression.MethodsIn a 2-week experience sampling method study, we collected self-reported emotion ratings and voice recordings 10 times a day, continuous keyboard activity, and trait depression severity. We correlated state and trait emotions and depression and language, distinguishing between speech content (spoken words), speech form (voice acoustics), writing content (written words), and writing form (typing dynamics). We also investigated how well these features predicted state and trait emotions using cross-validation to select features and a hold-out set for validation.ResultsOverall, the reported emotions and mobile-sensed language demonstrated weak correlations. The most significant correlations were found between speech content and state emotions and between speech form and state emotions, ranging up to 0.25. Speech content provided the best predictions for state emotions. None of the trait emotion–language correlations remained significant after correction. Among the emotions studied, valence and happiness displayed the most significant correlations and the highest predictive performance.ConclusionsAlthough using mobile-sensed language as an emotion marker shows some promise, correlations and predictive R2 values are low.

ЛИЧНОСТЬ ДРУГОГО В АКАДЕМИЧЕСКОМ ДИСКУРСЕ

The paper focuses on the semantic role of Other in academic discourse. Drawing on ideas expressed by M.M. Bakhtin and Yu.M. Lotman, the authors see Other as a key actor of written communication in the academic community and aim to describe the specifics of verbalizing Other’s presence in academic discourse as compared to fiction genres. The empirical basis of the study is constituted by a corpus of research papers written by Russian linguists and published over the period of 1991 through 2012 as part of the Logical Analysis of Language series edited by N.D. Arutyunova. In the initial stage of the research the corpus was used as a source of language data featuring contexts indicative of omnipresent Other. The second stage consisted in semantic analysis of language samples followed by cognitive interpretation and relied both on present-day linguistic theories and recent research in psychology. Results include a typology of Other in academic discourse and a description of relevant lexical and syntactic units systemically functioning as explicit and implicit markers of Other in Russian written academic communication. Addressing the figure of Other in academic discourse offers insights on the cognitive background of academic communication and brings to light key pragmatic intentions shaping interaction within the academic community

Coarse Alignment of Topic and Sentiment: A Unified Model for Cross-Lingual Sentiment Classification.

Cross-lingual sentiment classification (CLSC) aims to leverage rich-labeled resources in the source language to improve prediction models of a resource-scarce domain in the target language. Existing feature representation learning-based approaches try to minimize the difference of latent features between different domains by exact alignment, which is achieved by either one-to-one topic alignment or matrix projection. Exact alignment, however, restricts the representation flexibility and further degrades the model performances on CLSC tasks if the distribution difference between two language domains is large. On the other hand, most previous studies proposed document-level models or ignored sentiment polarities of topics that might lead to insufficient learning of latent features. To solve the abovementioned problems, we propose a coarse alignment mechanism to enhance the model's representation by a group-to-group topic alignment into an aspect-level fine-grained model. First, we propose an unsupervised aspect, opinion, and sentiment unification model (AOS), which trimodels aspects, opinions, and sentiments of reviews from different domains and helps capture more accurate latent feature representation by a coarse alignment mechanism. To further boost AOS, we propose ps-AOS, a partial supervised AOS model, in which labeled source language data help minimize the difference of feature representations between two language domains with the help of logistics regression. Finally, an expectation-maximization framework with Gibbs sampling is then proposed to optimize our model. Extensive experiments on various multilingual product review data sets show that ps-AOS significantly outperforms various kinds of state-of-the-art baselines.

Analyzing the Style of Writing of Russian Press on Russian-Turkish War of 1877-1878 for Serbia

Media language has always attracted the attention of linguists, particularly applied linguists and sociolinguists. There are four practical and principled reasons for this interest. First, the media provide an easily accessible source of language data for research and teaching purposes. Second, the media are important linguistic institutions. Their output makes up a large proportion of the language that people hear and read every day. Media usage reflects and shapes both language use and attitudes in a speech community. For second language learners, the media may function as the primary—or even the sole—source of native-speaker models. Third, the ways in which the media use language are interesting linguistically in their own right; these include how different dialects and languages are used in advertising, how tabloid newspapers use language in a projection of their assumed readers' speech, or how radio personalities use language—and only language–to construct their own images and their relationships to an unseen, unknown audience. Fourth, the media are important social institutions. They are crucial presenters of culture, politics, and social life, shaping as well as reflecting how these are formed and expressed. Media ‘discourse’ is important both for what it reveals about a society and for what it contributes to the character of society. This study tried to analyze the style of writing of Russian Press on Russian-Turkish War of 1877-1878 for Serbia. The article deals with the problem of gaining independence of the principality of Serbia in the Russian-Turkish war 1877-1878. The article considers the events with the Serbian participation in the liberation war through the pages of the Russian press. It describes all the vicissitudes of gaining Serbia’s independence from the Ottoman rule.

From a Smoking Gun to Spent Fuel: Principled Subsampling Methods for Building Big Language Data Corpora from Monitor Corpora

With the influence of Big Data culture on qualitative data collection, acquisition, and processing, it is becoming increasingly important that social scientists understand the complexity underlying data collection and the resulting models and analyses. Systematic approaches for creating computationally tractable models need to be employed in order to create representative, specialized reference corpora subsampled from Big Language Data sources. Even more importantly, any such method must be tested and vetted for its reproducibility and consistency in generating a representative model of a particular population in question. This article considers and tests one such method for Big Language Data downsampling of digitally accessible language data to determine both how to operationalize this form of corpus model creation, as well as testing whether the method is reproducible. Using the U.S. Nuclear Regulatory Commission’s public documentation database as a test source, the sampling method’s procedure was evaluated to assess variation in the rate of which documents were deemed fit for inclusion or exclusion from the corpus across four iterations. After performing multiple sampling iterations, the approach pioneered by the Tobacco Documents Corpus creators was deemed to be reproducible and valid using a two-proportion z-test at a 99% confidence interval at each stage of the evaluation process–leading to a final mean rejection ratio of 23.5875 and variance of 0.891 for the documents sampled and evaluated for inclusion into the final text-based model. The findings of this study indicate that such a principled sampling method is viable, thus necessitating the need for an approach for creating language-based models that account for extralinguistic factors and linguistic characteristics of documents.

Study of Large Data Resources for Multilingual Training and System Porting

This study investigates the behavior of a feature extraction neural network model trained on a large amount of single language data (“source language”) on a set of under-resourced target languages. The coverage of the source language acoustic space was changed in two ways: (1) by changing the amount of training data and (2) by altering the level of detail of acoustic units (by changing the triphone clustering). We observe the effect of these changes on the performance on target language in two scenarios: (1) the source-language NNs were used directly, (2) NNs were first ported to target language.The results show that increasing coverage as well as level of detail on the source language improves the target language system performance in both scenarios. For the first one, both source language characteristic have about the same effect. For the second scenario, the amount of data in source language is more important than the level of detail.The possibility to include large data into multilingual training set was also investigated. Our experiments point out possible risk of over-weighting the NNs towards the source language with large data. This degrades the performance on part of the target languages, compared to the setting where the amounts of data per language are balanced.

Meta-analysis of receptive and expressive language skills in autism spectrum disorder

Clinical anecdotes suggest that children with autism spectrum disorder (ASD) often show an atypical language profile in which expressive language exceeds receptive language competency. However, the few studies to directly explore this language profile have yielded inconsistent findings. This meta-analysis examined 74 studies that reported the receptive and expressive language performances of children and youth with ASD. Four potential predictors (age, language domain, source of language data, method of ASD diagnosis) were separately analyzed for their contribution to the relative receptive and expressive language impairment in ASD. Contrary to popular belief, the current meta-analyses found no evidence that an expressive advantage is common in ASD. Overall, children and youth with ASD showed equally impaired receptive and expressive language skills, both falling roughly 1.5 SD below peers with typical development. No discrepancies were found in receptive and expressive language across developmental stages, cognitive abilities, vocabulary, global language skills, caregiver report measures, clinician-administered measures, mixed method measures, or method of ASD diagnosis. Although some individual children with ASD may have an expressive-better-than-receptive language profile, this profile is not common enough to be a useful marker of ASD.

A Novel Online Encyclopedia-Oriented Approach for Large-Scale Knowledge Base Construction

In the process of constructing large-scale knowledge base, manual-based construction approach lacks efficiency as well as flexibility. Therefore, automatically extracting of massive knowledge from online encyclopedia has attracted attention from an increasing number of scholars. Current research is mainly focused on the extracting of data from English online encyclopedia, whereas research about knowledge extraction from Chinese or other language data sources is rare. For such reason, the present paper proposes an automatic construction scheme for large-scale knowledge base based on Chinese online Encyclopedia. (i)In the first phase of the scheme, self-expanded learning is performed on the semantic relations between subjects and objects among the knowledge triples. (ii)In the second phase, semantic relations between the marked attributes and their entities is predicted using Conditional Random Fields (CRFs) and Support vector machine (SVM) classifier. A large-scale knowledge base is automatically constructed based on the scheme, and the experiment results indicate that the scheme possesses feasibility and effectiveness.

E-SovTox: An Online Database of the Main Publicly-available Sources of Toxicity Data concerning REACH-relevant Chemicals Published in the Russian Language

A new open-access online database, E-SovTox, is presented. E-SovTox provides toxicological data for substances relevant to the EU Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) system, from publicly-available Russian language data sources. The database contains information selected mainly from scientific journals published during the Soviet Union era. The main information source for this database - the journal, Gigiena Truda i Professional'nye Zabolevania [Industrial Hygiene and Occupational Diseases], published between 1957 and 1992 - features acute, but also chronic, toxicity data for numerous industrial chemicals, e.g. for rats, mice, guinea-pigs and rabbits. The main goal of the abovementioned toxicity studies was to derive the maximum allowable concentration limits for industrial chemicals in the occupational health settings of the former Soviet Union. Thus, articles featured in the database include mostly data on LD50 values, skin and eye irritation, skin sensitisation and cumulative properties. Currently, the E-SovTox database contains toxicity data selected from more than 500 papers covering more than 600 chemicals. The user is provided with the main toxicity information, as well as abstracts of these papers in Russian and in English (given as provided in the original publication). The search engine allows cross-searching of the database by the name or CAS number of the compound, and the author of the paper. The E-SovTox database can be used as a decision-support tool by researchers and regulators for the hazard assessment of chemical substances.

Barriers to kidney transplants in Indonesia: a literature review

People living with chronic kidney disease will require renal dialysis or a kidney transplant to maintain life. Although Indonesia has a developing healthcare industry, Indonesia's kidney transplant rates are lower than comparable nations. To explore the healthcare literature to identify barriers to kidney transplants in particular in relation to Indonesia. Healthcare databases were searched (CINAHL, Medline, EBSCOhostEJS, Blackwell Synergy, Web of Science, PubMed, Google Scholar and Proquest 5000) using the search terms: transplant, kidney disease, renal, dialysis, haemodialysis, Indonesia and nursing. The search was limited to English and Indonesian language data sources from 1997 to 2007. Reference lists of salient academic articles were hand searched. The results of our search identified six articles that met our criteria. Costs are the major barrier to kidney transplant in Indonesia, followed by cultural beliefs, perception of the law, lack of information and lack of infrastructure. In addition, kidney disease prevention strategies are required. There are many complex socio-economic, geographical, legal, cultural and religious factors that contribute to low kidney transplant rates in Indonesia. Although an increase in transplantation rates will require strategies from various agencies, healthcare professionals, including nurses, can play a role in overcoming some barriers. Community education programmes, improving their own education levels and by increasing empowerment in nursing we may contribute to improved kidney transplant rates in Indonesia.

Sodobna dvojezična leksikografija

The article deals with some practical issues in modern bilingual lexicography. Its themes include language resources necessary for the work of a modern lexicographer, basic methodology of lexicographic work, computer format of the dictionary and the final result, a dictionary database. Among the language resources, the existing dictionaries are highlighted together with language corpora as indispensable source of language data for modern lexicography. Methodology is presented by using some practical examples from the Slovenian-English language pair. As the standard computer format of modern dictionaries, SGML/XML format is pointed up as the one allowing to control the structure of the dictionary and because of its software independence. The final result of the lexicographer’s work is the computer database in SGML/XML format that contains information of various levels, with the printed dictionary only one of them.

Cross-language use of acoustic information for automatic speech recognition

Techniques are investigated that use acoustic information from existing source language databases to implement automatic speech recognition (ASR) systems for new target languages. The assumption is that the amount of target language data available is too little for the training of a robust ASR system. Strategies for cross-language use of acoustic information are evaluated which include (i) training on pooled source and target language data, (ii) adapting source language models using target language data, (iii) adapting models trained on pooled source and target language using target language data only and (iv) transforming source language data to augment target language data for model training. These strategies are allied with Bayesian and transformation-based techniques to present a framework for cross-language reuse of acoustic information. Experiments are performed for a large number of approaches from the framework, using relatively large amounts of English speech data from either a separate database or from the same-database as smaller amounts of Afrikaans speech data to improve the performance of an Afrikaans speech recogniser. Results indicate that a significant reduction in word error rate is achievable (between 14% and 48% for experiments), depending on the amount of target language data available.

Language and the Media

Media language has always attracted the attention of linguists, particularly applied linguists and sociolinguists. There are four practical and principled reasons for this interest. First, the media provide an easily accessible source of language data for research and teaching purposes. Second, the media are important linguistic institutions. Their output makes up a large proportion of the language that people hear and read every day. Media usage reflects and shapes both language use and attitudes in a speech community. For second language learners, the media may function as the primary—or even the sole—source of native-speaker models. Third, the ways in which the media use language are interesting linguistically in their own right; these include how different dialects and languages are used in advertising, how tabloid newspapers use language in a projection of their assumed readers' speech, or how radio personalities use language—and only language–to construct their own images and their relationships to an unseen, unknown audience. Fourth, the media are important social institutions. They are crucial presenters of culture, politics, and social life, shaping as well as reflecting how these are formed and expressed. Media ‘discourse’ is important both for what it reveals about a society and for what it contributes to the character of society.

The Use of the Cantigas de amigo for a History of Portuguese Language Course

Introduction. This article presents a contextsage or passages of the cantigas for examples. In based framework for teaching the history of this way linguistic analysis proceeds cyclically Portuguese which incorporates passages from over the same text and, as a result, the student the medieval cantigas tie amigo to illustrate becomes thoroughly familiar with an actual mediachronic linguistic processes in the evolution dievab corpus. In this way the cantiga passages of the language. The context-based method not only provide examples of each diachronic means that actual passages of attested language principle, they also put it in a meaningful context. are employed to exemplify diachronic processes After carrying out the analysis of a medieval text, in phonology, lexicon and grammar, wherein the students then examine subsequent linguistic delinguistic elements of the sample text are interrevetopments which lead up to the modern form of tated as part of an actual communicative event.i the language. As before, cantiga passages proProminent educators have maintained for two vide a source of language data in context from decades that the acquisition of language skills which the student traces forward each Gaticianis enhanced by the use of communication-oriPortuguese phoneme, morpheme, or other gramented and natural language materials in the aticat element to its reflex in the modern classroom.2 It is the opinion of the authors that language and, in doing so, teams to identify the instruction i t e istory of language can also relevant diachronic processes. benefit from a more contextbased approach. This The advantages to such an approach are: stands in contrast to the methodology which (1) The use of the cantigas tie amigo facititraditional history of language courses often tates the beaming of linguistic elements in much employ. Not infrequently, traditional text books the same way that communication-oriented tanmake use of lists of vocabulary items to illustrate guage skills programs do, that is, by placing the the linguistic processes which have taken place material to be teamed in as real a language in the language. Students in such a course deal context as possible. In this case, a poetic work with lists of words in isolation long before they serves as a means of s-elating sets of Latin, see contextuatized samples of the language. Gatician-Portuguese and Modern Portuguese The method proposed here introduces pasforms and grammatical elements to one another