This study provides a bibliometric review of research in corpus linguistics, mapping out its development from 1996 to 2021. CiteSpace 6.1.R2 was utilised, based on a dataset of 7,553 studies and 126,113 cited references retrieved from the Web of Science (wos) core collection. The co-citation analysis identified the major research themes and the landmark publications in this field. Based on Shneider's (2009) four-stage evolutionary model and the timeline generated, this study found that corpus linguistics has experienced three major developmental stages: the conceptualising stage, the maturing stage, and the flourishing stage. The analysis also revealed potential themes in the near future, including language acquisition, lexical items, and rhetorical structures. The findings provide a systematic and structured understanding of the core topics, seminal works, as well as the developmental trajectory of the field of corpus linguistics, and highlight the increasing significance of corpus-based research in advancing the comprehension of linguistic phenomena.
The Cognitive Linguistic approach is an essential development in Critical Discourse Analysis in that it provides a conceptual framework for analysing discourse representation at both the levels of text and embodiment. In alignment with this belief, this research aims to merge Cognitive Linguistics, Corpus Linguistics, and Critical Discourse Analysis, with a case study of critically analysing the ELEMENTAL metaphorical markers in discourses of social unrest constructed by six US-based, English-language online news media outlets, namely Fox News, CNN, New York Times, The Wall Street Journal, Washington Post and USA Today. In particular, it employs a corpus-assisted discourse analysis focussing on lexical choice, frequency and grammatical patterns to investigate discourses of social unrest. We argue that Critical Discourse Analysis can benefit tremendously from Corpus Linguistics research design to ward off the subjective (under/over) and cherry-picking interpretation nature of discourse texts. Thus, the suggested research design may shed light on the metaphorical framing operations of ELEMENTAL metaphorical markers in social unrest. The findings demonstrate how the use of metaphors varies amongst media sources, reflecting ideological disparities. Whilst police are deputised as stabilising forces using fire and wind metaphors, right-leaning media outlets such as Fox News and USA Today use water and earth metaphors to portray protests as chaotic threats. This contrast shows how liberal media outlets promote protests as a response to systemic injustice, whilst conservative media delegitimises them by focussing on societal disorder.
This paper describes the creation of the 14.6 million word Pakistani Local English Corpus (PakLocCorpus). The corpus consists of representative non-native English used in Pakistan from the fields of literature, media and academia. The corpus is available through an online portal, with concordance searches enabled, and this paper highlights the potential of the corpus for English language users, language policy makers, English language teachers, learners, and curriculum developers. The paper highlights the challenges faced during the processes of corpus building which may help corpus developers interested in purpose-built corpus building. The significance of the complex linguistic phenomenon of Urduised words that call for manual annotation is also highlighted.
Scientific writing has become increasingly less readable over time. However, few studies have examined diachronic changes in readability by comparing academic abstracts with full texts. Taking the journal Science as a specific case, our research examined diachronic changes in the readability of both abstracts and full texts of research articles published between 1997 and 2021, and explored how linguistic features may have influenced these changes in readability. The findings of this study show that in Science research articles, abstracts have evolved to be less readable than full texts over time. An increased presence of certain linguistic features may have implicitly contributed to the observed decline in readability, which is further examined and analysed in this study. Finally, this study discusses the implications of its findings for academic researchers and the broader scientific community.
Media representation is crucial in shaping public perception of social issues such as homelessness. This is particularly the case for those who lack direct contact with the issue, since news coverage can provide the primary means for such people to learn about these matters. Whilst some discourse-based research has looked at the representation of homelessness in partisan news sources, this study offers the first such analysis of bbc news uk – a non-partisan, and the most widely consumed, news provider in the UK. Drawing on a specialised corpus comprising 358 online news articles published in 2022 and 2023 (totalling 157,854 words), we conducted a keyword-driven framing analysis, closely reading 100 concordances of each keyword obtained by comparing this corpus against the BE21 corpus. Our findings demonstrate that the bbc predominantly foregrounds responses to and experiences of homelessness rather than its causes, mirroring a long-standing trend amongst partisan news media. However, our analysis also identified housing and polycrisis discourses that could help readers to understand the issue from structural perspectives, as well as a ‘nuisance’ discourse that may stigmatise people experiencing homelessness. Notably, political actors in the central/devolved governments are far less salient than local councils and charities in the corpus.
