Search

The past few decades have seen the rapid development of topic modeling. So far, research has been more concerned with determining the ideal number of topics or meaningful topic clustering words than with applying topic modeling techniques to evaluate linguistic theories. This study proposes the Structural Topic Model (STM)-led framework to facilitate the interpretation of topic modeling results and standardize text analysis. STM encompasses various model training mechanisms, thereby requiring systematic designs to properly combine language studies. “Structural” in STM refers to the inclusion of metadata structure. Unlike the corpus-based keyness approach, STM can capture contextual cues and meta-information for the interpretation of topical results. Besides, STM can make crosscorpora comparisons via topical contrast, a challenging task for corpusdriven related models such as the Biterm Topic Model (BTM). Stylistic variations in song lyrics are taken as an illustration to show how to use the suggested framework to delve into the linguistic theory proposed by Pennebaker (2013). The topical model and iterable model in the proposed paradigm can clarify how pronouns affect style distinction. We believe the proposed STM-led framework can shed light on text analysis by conducting a reproducible cross-corpora comparison on short texts. [ABSTRACT FROM AUTHOR]

Corpus linguistic methods can now be easily employed in a wide range of studies within sub-disciplines of linguistics and well beyond. In a two-part paper, Gries (2022a, 2022b) challenges some of the most widely used 'association measures' of what many might feel to be powerful aspects of text patterning: collocation and key words. While the additional association measure offers some new possibilities, this paper highlights the strong influence of another frequency parameter on odds ratio and Gries's suggested association measure, and questions the applicability of his cautions for many different kinds of corpus research. Nevertheless, having been inspired to look at different aspects of association and dispersion more carefully, the author presents some new visualizations which were designed to communicate some of the important lessons to be learned from Gries's papers, especially for learners and teachers using corpus tools in Second Language classrooms. [ABSTRACT FROM AUTHOR]

Keyness is a commonly used method in corpus linguistics and is assumed to identify key items that are characteristic of 1 corpus when compared to another. This paper puts this assumption to the test by comparing case study corpora in the fields of genetic, immunological and psychiatric biomedical association studies, using what we refer to as a 'K-FLUX' analysis to produce a set of key items. Experts from within these fields are asked to evaluate the extent to which identified key items are characteristic of their discipline. The paper concludes that less than 50% of the items identified by the method are rated as highly characteristic by experts and that this ranges between types of association study. Further, there is difficulty in reaching a consensus over what is deemed to be 'characteristic', thus posing a challenge to the ultimate aim of the keyness method. The paper demonstrates the value of supporting corpus linguistic studies with expert assessments to evaluate whether (and which) items can be said to be indicative of a particular field. [ABSTRACT FROM AUTHOR]

Throughout the social sciences, there has been growing pressure to present effect sizes when publishing empirical data (see American Psychological Association, 2001; Parsons & Nelson, 2004). While it seems indisputable that for the majority of quantitative research foci, effect size is an essential element of statistical analysis, this paper argues that specifically for key word analysis in corpus linguistics, the means of reporting effect size must depend on the level of the unit of study of each investigation (single text, collection or large corpus). After exploring some main criticisms of the log-likelihood measure, this paper unpacks the parameters of different measures for keyness and how they might address underlying concerns. It maintains that for the exploration of foregrounded/deviant/salient/marked features in text, the use of log-likelihood scores to rank the results is still fit for purpose and coupled with Bayes Factors is a solid approach for key word analyses. [ABSTRACT FROM AUTHOR]

This paper attempts to shed some light on the importance of adjectives in the linguistic characterisation of tourism discourse in English in general and in adventure tourism in particular as well as to prove how significant the difference in usage is compared to the general language. It seeks to understand the role that adjectives play in this specific subdomain and to contribute to the linguistic characterisation of tourism discourse in this respect. It also aims to confirm or reject previous assumptions regarding the use, and frequency of use, of adjectives and adjectival patterns in this specialised domain and, in general, to promote the study of adjectivisation in domain-specific discourses. To do so, it proposes a corpus-based study that measures the keyness of adjectives in promotional texts of the adventure tourism domain in English by comparing their usage in the compiled corpus to the two most relevant reference corpora of English (coca and the bnc). [ABSTRACT FROM AUTHOR]

In an article published in this journal, Partington (2014) addresses the criticism often made against corpus linguistics that it is apparently unable to cope with absences. He convincingly argues that corpus linguistics is better suited to account for absences than has been claimed. I resume the debate by discussing a type of absence not fully addressed in Partington (2014) which I have termed 'creative absences'. With a focus on corpus stylistics, I consider the way in which the author Henry Green dispenses with a compulsory element in the grammatical structure of Standard English, i.e. the determiner (mainly, the definite article). By means of a manual analysis as well as two corpus stylistic analyses (keyness and text-type analysis) of the novel Living (Green 1929), I explore the effects of such an unorthodox use and argue, alongside Partington (2014), for the usefulness of corpus approaches to account for at least certain types of absences. [ABSTRACT FROM AUTHOR]

Aim. The aim of this paper is to present and exemplify a number of basic uses of corpus-based text analysis tools that can supplement and provide additional insight for an otherwise qualitative analysis of a text. I attempt to show that nowadays certain corpus tools are easily accessible to any researcher and can be used to enrich the results of studies concerned with texts. Methods. This paper comprises the basics of corpus building, the main types of data that can be drawn from a simple corpus and a detailed description of four methods that can aid text analysis: wordlists, concordances, dispersion plots, and keywords. Each of those four methods is thoroughly described, including a number of examples of its applications and indicates its possible limitations. Results. The examples provided suggest that even performing a very simple corpus analysis of a text might unveil certain trends and phenomena not noticeable through the classic qualitative text analysis methods (e.g. close reading). The paper argues that corpus research can hence work as an extension of a quantitative analysis (or be its starting point) by examining themes and keywords present in a given text and enrich the results of a qualitative study with a fresh perspective. Finally, the paper claims that basic corpus analysis can, in fact, be successfully employed by researchers who do not have any prior experience with statistics or corpora. [ABSTRACT FROM AUTHOR]

Keyword analysis is used in a range of sub-disciplines of applied linguistics from genre analyses to critically-oriented studies for different purposes ranging from producing a general characterization of a genre to identifying text-specific ideological issues. This study compares the use of log-likelihood (LL), a probability statistic, and odds ratio (OR), an effect size statistic, for keyword identification and argues that the two methods produce different keywords applicable to research focusing on different purposes. Through two case studies, keyword analyses of advance fee scams against the British National Corpus and research articles in applied linguistics against research articles from other academic disciplines, we show that both the LL and OR keywords concern the aboutness of the corpus, but differ in their specificity and pervasiveness through the corpus. LL highlights words which are relatively common in general use serving genre purposes, whereas OR highlights more specialized words serving critically-oriented purposes. Methodological and practical contributions to keyword analysis are discussed. [ABSTRACT FROM AUTHOR]

This paper looks at a corpus of British and US poetry, uncovering phraseological units which, through their frequency, are indicators of key concepts. Multi-word-units (MWUs) have been discussed extensively with reference to corpus-based research, for example by Sinclair (1996) [2004], Biber and Conrad (1999), or, referred to as formulaicity by Wray (2002); O'Keefe et al. (2007), Greaves and Warren (2010) and Pace-Sigge (2015) describe MWUs preferred in different spoken and written genres. So far, however, there has been very little research in how far MWUs appear in the genre of poetry. A commonly held view is that poetry by definition should not be yielding patterns - it subverts every pattern (linguistically speaking) that it can. Through focus on the main themes surfacing in multiword units, this research looks at usages found in poetic texts in-depth and compares sets of words found with their occurrence patterns in prose literature. Key issues will be highlighted through a number of theme-based case studies, looking at themes of world and sky. Results show that there are common clusters found in poetry and prose corpora: itis depth of usage that marks their divergence. [ABSTRACT FROM AUTHOR]

Measuring the number of foreign words used daily in conversation or in general language discourse is a challenge to linguists. This is because all instances of a language's use cannot be collected and measured reliably for the occurrence of specific language elements. This is in part because the overall size of a language population, that is, all words and expressions used in a language cannot be known (Kilgarriff & Grefenstette, 2003). They are dynamic in nature. There is, however, value in the enquiry of determining how much of foreign words exist in any given discourse. This paper proposes an approach of measuring the amount of foreign terms in a language. It uses a 13-million-word corpus to measure the amount of English in a Setswana corpus. The data is analysed using Oxford Wordsmith Tools software. In the study, we use keyword analysis. The different components of the corpus are measured to extract terms which are typical or descriptive of a specific text. The key terms are then listed collectively on the basis of their keyness measures. The English words are then counted in the extracted list at 100 word intervals and then tabulated. An average is subsequently computed to characterise the amount of English words found in the keyword list. One of the key findings of the study is that about 25% of Batswana's language use is English. This is attributed to Batswana's bilingualism and the Botswana education system's systematic promotion of English dating back to the late 70s. The article recommends that for the preservation and promotion of Setswana, the language must be used beyond language classes in social functional domains such as medicine, law and farming. [ABSTRACT FROM AUTHOR]

This study details a comparative, corpus-based discourse analysis of corpora containing educational documents distributed to parents and guardians of K-12 children in public schools in the United States (US). The exploratory local corpus (n=152,934) contains parent-directed educational documents collected from four public schools in a city located in the south-eastern US with an unusually high percentage of foreign-born residents. The comparison corpus (n=147,796) contains parent-directed documents collected from a sampling of K-12 schools across the US. Following Baker et al. (2008), keyness and collocations were utilised as central theoretical notions and tools of analysis, in addition to a lexical sophistication comparison, in order to investigate text simplification across corpora. Results show that while the first corpus used labels for students that were superficially inclusive, English language learners themselves were discursively represented as outsiders facing barriers to inclusion that native-English speaking monolingual students do not face. Furthermore, the first corpus revealed an emphasis on identifying and categorising language learners so as to provide them with immediate services, while the non-geographically specific corpus focussed more on the long-term development of learners and on preparation for post-secondary education. We discuss the implications for language policy in public education and for policies related to K-12 school-to-home correspondence. [ABSTRACT FROM AUTHOR]

This study describes some characteristics of maritime institutional legal texts in terms of corpus methodology. We self-built two study corpora: a public maritime institutional corpus and a private maritime institutional corpus. The differences between the two corpora can be distinguished by identifying typical linguistic features from the keyness aspects of key words, key clusters, and key semantic domains. Specific words and phrases in complementary distribution are offered to distinguish public maritime legal characteristics from private maritime legal characteristics by comparing the self-built specialized corpora with the more general British National Corpus (BNC informative genre). Linguistic features are discussed from the view of keyness, thus enabling non-legal practitioners as well as non-specialist readers to discover and describe underlying parameters that best depict the differences between legal registers or genres. [ABSTRACT FROM AUTHOR]

This exploratory study is concerned with the extent to which corpus-driven analyses of two generically very similar corpora, namely television coverage of the South African general election of 2009 by a state-owned public broadcasting channel (SABC3) and by a commercial channel (e.tv), can nevertheless reveal interesting differences between them. The focus is mostly on those differences that relate to the channels' political stances, but stylistic contrasts in the coverage are also considered, some of which prove to be of indirect political relevance. Central to the research approach here is the notion of 'keyness' (a statistically significant higher density of an item in one corpus when compared to another), and this study takes its cue from Rayson (2008) by using his Wmatrix corpus analysis and comparison tool to examine differences not only at word level, but also at part-of-speech and semantic domain levels. [ABSTRACT FROM AUTHOR]