Semantic shifts have been explored via a range of methods (Allan & Robinson 2012). Typically, semantic shifts were usually noted or described with methods such as a literature review or dictionary checking (e.g. Blank & Koch, 1999; Stockwell & Minkova, 2001; Williams, 1976), which are very labour-intensive and time-consuming methods. Other more recently developed methods involve sociolinguistic interviews (Robinson, 2012; Sandow & Robinson, 2018). However, with the development of large-sized corpora and computational semantics, diachronic semantic shifts have started to be captured in a data-driven way (Kutuzov et al., 2018). Recently, the word embeddings technique (Mikolov et al., 2013) has been proven to be a promising tool for the tracking of semantic shifts (e.g. Hamilton, Leskovec & Jurafsky, 2016a, 2016b; Kulkarni et al., 2015; Kutuzov et al., 2017). For example, Hamilton et al. (2016b) exemplified how to use the technique to capture the subjectification process of the word ‘actually’ during the 20th century.

English-medium instruction (EMI) has been perceived as a key strategy through which universities, propelled by academic, political, social and economic motives, respond to the influence of globalisation (Altbach & Knight, 2007). This has been fuelled by the fact that English, defined as the global common language, is needed to create the knowledge base in global tertiary education (Fishman, 2000). In the process, English has become the universal second language of advanced education (Brumfit, 2004), due to the value attached to the language in present times and the advantage of using the language in the existing global language order (Zhang, 2017). These motivations have contributed to the global phenomenon of English being the medium of instruction (MOI), and higher education has been the venue where EMI could be implemented more consistently (Dearden, 2014). This has resulted in the generation of a growing body of work on how universities plan their language policies (Liddicoat, 2016).

In the Longman Grammar of Spoken and Written English (Biber et al., 1999), a new category is identified in the grammar of the English phrase. In conversational data, the most frequent forms cited as examples of this category are or something, and everything, and things and and stuff, which are described as ‘coordination tags’ by Biber et al. (1999: 115–16). This label has not been widely adopted, but the linguistic category it describes has clearly become established as part of modern English. The term ‘general extender’ (Overstreet, 1999) is now commonly used to refer to this category: ‘“general” because they are nonspecific and “extender” because they extend otherwise complete utterances’ (1999: 3). There are two subcategories: adjunctive general extenders, beginning with and, and disjunctive general extenders, beginning with or. In casual conversation, general extenders are typically phrase- or clause-final, consisting of and/or plus a vague noun (stuff/things) or a pronoun (something/everything), with an optional comparative phrase (like that/this). In everyday spoken British English, the phrase and (all) that is also extremely common. In written and formal spoken English, forms with quite different structures, such as et cetera, and so on, and so forth, and or so are more typically used to fulfill related functions. All of these forms are grammatically optional and fall within the more general category of pragmatic markers, along with you know, I mean, like and sort of, ‘expressions which may have little obvious propositional meaning but which oil the wheels of conversational social interaction’ (Beeching, 2016: 1).