International Journal of Corpus Linguistics - Volume 27, Issue 4, 2022

This paper applies a new approach to the identification of discourses, based on Multiple Correspondence Analysis (MCA), to the study of discourse variation over time. The MCA approach to keywords deals with a major issue with the use of keywords to identify discourses: the allocation of individual keywords to multiple discourses. Yet, as this paper demonstrates, the approach also allows us to observe variation in the prevalence of discourses over time. The MCA approach to keywords allows the allocation of individual texts to multiple discourses based on patterns of keyword co-occurrence. Metadata in the corpus data analysed (here, UK newspaper articles about Islam) can then be used to map those discourses over time, resulting in a clear view of how the discourses vary relative to one another as time progresses. The paper argues that the drivers for these fluctuations are language external; the real-world events reported on in the newspapers.

This paper demonstrates the value of studying co-occurrence ‘quads’ – constellations of four non-adjacent lemmas that consistently co-occur across spans of up to 100 tokens – for understanding discursive change. We map meaning onto quads as ‘discursive concepts’, which encompass encyclopaedic semantics, pragmatics, and context. We investigate a high-frequency quad with high co-occurrence strength in EEBO-TCP: world-heaven-earth-power. We conduct semantic and pragmatic analysis to generate hypotheses regarding discursive change. The quad’s components are semantically underspecified; thus, although the quad indicates a discursive concept, each instantiation of the quad is variable, contingent, and dependent upon context and pragmatic processes for interpretation. We observe how the vague lexemes that constitute building blocks of religious discourse are employed to generate new, timely secular discourses; and we argue that semantic underspecification is the site and source of discursive change. Indeed, the volatile, unstable nature of the component lexical meanings renders them indispensable to early modern debate.

This paper examines the utility of metaphor as an investigative tool in “long-distance” corpora and discourse studies. I show that metaphor is both important for understanding discourses and useful for diachronic analysis because it allows us to abstract out above the purely lexical level, enabling comparison across contexts where the same concept could be lexicalised differently. The case-study is concerned with the oft-discussed metaphor of migrants are water in the UK-based Times newspaper from 1800–2018 and the conventionalisation and evaluative patterns are presented. The findings confirm that the water metaphor has an extensive discourse history regarding how migration is represented in the UK press, but also that evaluations may differ significantly. The paper shows how metaphor can provide a way to find discourse evaluations and framings across different time periods. The use of second-order collocates illustrates how corpus tools can help re-contextualise data to ensure interpretation heeds contemporary framings.

The ways in which politicians have discussed who, what, and where was considered “uncivilized’” across the past two centuries gives an insight into how speakers in a position of authority classified and constructed the world around them, and how those in power in Britain see the country and themselves. This article uses the Hansard Corpus 1803–2003 of speeches in the UK Parliament alongside data from the Historical Thesaurus of English to analyse diachronic variation in usage of words for persons, places and practices considered uncivil. It proposes new methods and offers quantitative data to describe the period’s shift in political attitudes towards not just the so-called “uncivil” but also the country as a whole.

This paper tracks stylistic variation in the use of two roughly synonymous suffixes, the Romance -ity and the native -ness, during the Early Modern English period. We seek to verify from a statistical viewpoint the claims of Rodríguez-Puente (2020), who reports on a decrease of -ness in favour of -ity in registers representative of the speech-written and formal-informal continua at that time. To this end, we develop new methods of statistical and visual analysis that enable diachronic comparisons of competing processes across subcorpora, building upon an earlier method by Säily and Suomela (2009). Our results confirm that -ity gained ground first in written registers and then spread towards speech-related registers, and we are able to time this change more accurately thanks to a novel periodisation. We also provide strong statistical support indicating that the proportion of -ity was significantly higher in legal registers than in other registers.

This article introduces Corpus PalaeoHibernicum (CorPH), a corpus currently consisting of 78 texts in Early Irish (c. 7th–10th cent.) created by the ERC-funded Chronologicon Hibernicum (ChronHib) project by bringing together pre-existing lexical and syntactic databases and adding further crucial texts from the period. In addition to being annotated for POS, morphological and syntactic information, another layer of annotation has been developed for CorPH – ‘Variation Tagging’, i.e. a tagset that numerically encodes synchronic language variation during the Early Irish period, thus allowing for much improved research on the chronological variation among the material. Another new pillar of studying linguistic variation is Bayesian Language Variation Analysis (BLaVA), in order to address the challenge that “not-so-big data” poses to statistical corpus methods. Instead of reflecting feature frequencies, BLaVA models language variation as probabilities of variation.