Article Preview
TopIntroduction
The language of evaluation has gained considerable interest in the field of language and discourse studies (e.g., Hunston, 2011; Martin & White, 2005; Thompson & Hunston, 2000). According to Thompson and Hunston (2000), evaluation can be broadly defined as a cover term “for the expression of the speaker or writer’s attitude or stance towards, viewpoint on, or feelings about the entities or propositions” (p.5), which overlaps to varying degrees with the related constructs such as stance (Biber, 2006), metadiscourse (Hyland, 2005), and appraisal (Martin & White, 2005), to name just a few. While academic writing is traditionally characterized as objective and impersonal, research on evaluative language in academic discourse has indicated that evaluation plays a crucial part in helping writers to construct knowledge and interact with readers (Hyland, 2005). The importance of evaluation has been highlighted in a variety of academic genres, such as research articles (e.g., Hyland, 2005), master’s and doctoral theses (e.g., Geng & Wharton, 2016; Hyland, 2004), and undergraduate essays (Lee, 2015; Wu 2007).
Although it is well recognized that evaluative language poses difficulties for analysts because evaluation is often context-dependent and implicitly expressed (Hunston, 2011), the recent development of corpus linguistic methods has assisted the identification and analysis of evaluative language. The work on “pattern grammar” (Francis et al.,1998; Hunston & Francis, 2000), for example, finds that particular syntactic structures tend to co-occur with a restricted set of lexical items and vice versa. This co-selection of lexis and syntactic units are further explored in corpus-driven studies on multi-word sequences such as lexical bundles or phrase-frames (Biber, 2009). Lexical bundles refer to continuous multi-word sequences which are recurrent in a discourse regardless of their idiomaticity and structural status (Biber et al., 1999). Phrase-frames (p-frames henceforth), on the other hand, refers to discontinuous multi-word sequences with one variable slot (Biber, 2009). The most frequent lexical bundles in a corpus are often associated with particular p-frames. For example, when moving from lexical bundles to extended units of phraseologies, Römer (2010) found that the five-word lexical bundles it would be interesting to, it would be useful to, and it would be nice to are all realizations of the p-frame it would be * to, and the filler words like interesting, useful, nice, better, possible, and helpful are the top-ranking variants for this p-frame in a 3.5-million-word academic book review corpus.
Although the research emphasis was not always placed on evaluative expressions, the studies on multi-word sequences have consistently showed, among other findings, that evaluation is a key function of formulaic expressions and phraseologies (Biber, 2009; Römer, 2008, 2010). In what follows, we provide a brief overview of previous research on evaluative patterns and p-frames associated with adjectives.