
During public health crises, rapid dissemination of accurate information in the languages spoken by vulnerable populations is a matter of linguistic justice and, in many cases, a determinant of survival. Although machine translation (MT) systems constitute a scalable solution for multilingual crisis communication, low-resource languages such as West African varieties of Pidgin remain effectively excluded from these technologies owing to the absence of standardized orthography, terminological resources, and parallel corpora. This study examines the obstacles developers would face when integrating COVID-19 terminology into MT systems for Pidgin English. Drawing on a corpus of over 20,000 Cameroon Pidgin English tokens scraped from Facebook at the height of the pandemic and analysed with Sketch Engine, we investigate how speakers rendered key public health concepts, identify the predominant translation strategies, and document the resulting challenges for terminology compilation and downstream MT integration. The analysis focuses on orthographic variation, domain specificity, annotation consistency, preprocessing requirements, and ethical considerations surrounding community consent and representation. The study highlights linguistic and technical barriers preventing Pidgin-speaking communities from benefiting from automated crisis-translation tools and underscores the urgent need for community-driven terminological resources as a prerequisite for inclusive MT development in low-resource settings. For time of public health crisis, e dey very important make correct information reach vulnerable people dem for di language weh dem di speak, because dis na matter of language justice and fit even save life. Although machine translation (MT) system fit help spread information for plenty language, low-resource language dem like Cameroon Pidgin English still no di benefit because standard writing, terminology resources, and parallel corpora no plenty. Dis study examine di challenge dem weh developer go face if dem want put COVID-19 terminology inside MT system for Cameroon Pidgin English. Wi use corpus of over 20,000 Cameroon Pidgin English tokens weh wi collect from Facebook during di COVID-19 pandemic and analyse with Sketch Engine. Wi investigate how speakers express key public health concepts, identify di main translation strategies, and document di challenge dem for terminology development and MT integration. Di analysis focus on orthographic variation, domain-specific terminology, annotation consistency, text preprocessing, and ethical issues like community consent and representation. Di findings show say dis language and technical challenge dem still di prevent Pidgin-speaking community dem from benefiting from automated crisis-translation tools and show say community-driven terminology resources na key for build inclusive MT systems.
Chinese contracted sentences are often viewed as a strategy to enhance processing efficiency, yet quantitative evidence for their syntactic and cognitive features remains limited. This study examines whether their dependency distance distributions (DDDs) follow Zipfian regularities and whether their mean dependency distances (MDDs) reflect reduced cognitive load. Using dependency grammar, three corpora of contracted sentences (parallel, causal, adversative) and a reference corpus were annotated for dependency syntax. Results show that (i) the DDDs of both contracted and reference sentences follow Zipf's law, supporting the principle of dependency distance minimization; and (ii) contracted sentences exhibit significantly shorter MDDs, suggesting lower cognitive load. Comparison with full-form counterparts further confirms that contraction reduces syntactic complexity and processing effort. These findings provide quantitative evidence that Chinese contracted sentences conform to universal dependency distance patterns while optimizing cognitive efficiency.
Researching media texts such as advertising has evolved from studying language as a linking, communal experience constituting a speech community to a mass individualized experience with the fragmenting effects of digital communication and the mainstreaming of large language models. This has involved a shift from the medium being regarded as the site of the text through an increasingly stronger focus on mediatization to the current situation where the one technology can be the site, source, and tool for collecting and analysing the data. Disruptions in the processes of production have led to changes in the types of data, ways of collecting, and ways of analysing. A key challenge of mass individualization and the interactional experience offered by large language models is the ever-changing nature of the texts produced and the difficulty of capturing the fragmented, individual experience with these texts.
This study investigates how nonexperts conceptualize linguistics and what they most want to know about language. We recruited 812 English-speaking participants from Canada, the United States, and the United Kingdom via Prolific to complete four open-ended surveys (N = 406 on perceived professional foci; N = 406 on desired research topics). All responses were segmented into 2,684 statements and analyzed with unsupervised natural language processing: sentence embeddings (all-MiniLM-L6-v2) followed by k-means clustering, yielding 29 topic clusters. Results show that respondents broadly recognize core subfields (e.g., syntax, phonetics/phonology, morphology, semantics, historical linguistics) and also reproduce familiar misconceptions (e.g., linguists as translators; linguists as language or literature teachers). Public curiosity, however, concentrates on applied and experiential themes - language learning and learnability, bilingualism, dialect differences, and the origins, spread, and loss of languages. Topics linking variation, cognition, and development (sociolinguistics, acquisition, psycho-/neurolinguistics) form shared ground. We discuss how these high-interest domains can serve as entry points for communicating foundational linguistic concepts.
This special issue examines the use of dependency-annotated corpora across different languages for quantitative typological research. The contributions explore how large-scale resources enable new analyses of morphological and syntactic complexity, word order variation, language contact, and language evolution, as well as challenges in the annotation of specific phenomena. Together, these contributions demonstrate the potential of dependency grammar-based resources for advancing crosslinguistic research.