Closed-class keywords and corpus-driven discourse analysis

Abstract

Keywords belonging to closed grammatical classes (i.e. conjunctions, determiners, prepositions and pronouns) are often perceived as useful indicators of the characteristic style of a particular text or corpus, but as being of less interest to researchers interested in its semantic properties. The aim of this chapter is to propose, contrary to this mainstream view, that closed-class keywords can form a valid and even preferable basis for empirical linguistic research into specialized discourses, “discourses” being defined here as constellations of meanings and values associated with specific communities or institutions. The argument is illustrated with practical examples drawn from a keywords analysis of a 3-million-word corpus of academic journal articles representing the academic disciplinary discourse of history.

Candidate Knowledge? Exploring epistemic claims in scientific writing: a corpus-driven approach

“…Hyland 1998), which, whilst providing excellent empirically-based descriptions of known epistemic structures, are unlikely to contribute to the discovery of additional or unknown epistemic devices. In this paper I will argue, in agreement with Groom (2007;2010) that the answer to this impasse is to explore corpus-driven methods of analysis in order to uncover new or unexpected epistemic devices in English. Through a close analysis of four clusters, I demonstrate that it is possible to discover a number of additional strategies for nuancing claims, which are not typically mentioned in seemingly exhaustive studies such as Hyland (1998).…”

“…Whilst the identification of hedging devices has proven to be a very useful and successful enterprise within applied linguistics, it has been argued that the study of these devices has become concentrated onto a small group of the 'usual suspects' (Groom, 2007;2010;Plappert, 2012) of words and structures that are known to have an epistemic effect in a claim or proposition. As such linguistic markers of modality such as modal verbs (eg: may, might, can, could), modal adjectives (eg: possibly, probably) and n-grams identified as functioning as hedges (such as it is possible that and it is likely that) often form the starting place for analysis of the linguistic aspects of epistemology.…”

Using Antconc: A Corpus-Based Tool, to Investigate and Analyse the Keywords in Dickens’ Novel ‘A Tale of Two Cities’.

“…In keyword analysis, it is generally considered by discourse analysts that open-class keywords represent the aboutness of a specific test/texts or corpus, while closed-class words represent the style of a specific test/texts or corpus (Groom,2010;Scott, and Tribble, 2006). However, analysing the meaning of a corpus depends on the sequences of words not on the form of these words (Groom, 2010). In addition, Sinclair (1991 cited in Groom, 2010: 62) mentioned that "[m]ost everyday words do not have an independent meaning, or meanings, but are components of a rich repertoire of multi-word patterns that make up text".…”

“…There is a slight distinction between closed-class and open-class keywords (Scott, and Tribble, 2006). Therefore, Groom (2010) argued that this distinction is regarded as a reason in skipping the analysis of closed-class keywords. It is possible to say that this issue is not a rule of thumb; it significantly depends on the text/texts or corpus.…”

“…In a type of text, for instance, one can find the use of functional words (closed-class) is a representative of the style as well as the aboutness of this text but in other texts s/he may find these words represent just the style and not the aboutness. However, closed-class keywords would not be treated as the same as open-class keywords in analysing the aboutness, open-class keywords are always the prime indicators of aboutness (Groom, 2010;Baker, 2006).…”