Empirical work on linguistic description
Speakers in ELF encounters normally come from different linguacultural backgrounds, and are likely to display varying levels of competence in English. They are expected to have had different experiences with the language, having learnt it formally in the education system or informally under different circumstances in different parts of the world. In all this, the influence of the local context and the domain in which they function are likely to manifest themselves in various localisations at the different linguistic levels.
Descriptive research of ELF communication has been particularly productive in the past 15 years, especially since the creation of a number of corpora of spoken data (VOICE, ELFA and ACE, as well as some individual and small-scale corpus projects), and increasingly more work on written data too. In this section, for reasons of space, I will provide only a succinct summary of work on the nature of ELF that is relevant for structuring and situating the contributions in this volume. This will include the main findings of research in pronunciation, pragmatics and intercultural aspects. Research in lexico-grammar is not reviewed here as papers in this collection do not involve this aspect in the teaching practices explored.
Pronunciation was the first area of linguistic description to be empirically researched. Jenkins' seminal work (2000) explored intelligibility in ELF spoken communication and the kind of accommodation processes speakers engage in. Her data showed the speakers' ability to accommodate to more or less 'nativelike' speech in order to enhance intelligibility. The findings also cast light on the core aspects of pronunciation that are essential for intelligibility - that is, all the consonants (apart from the dental fricatives), consonant deletion in initial clusters, vowel length distinctions and nuclear stress. As such, teaching implications require that practitioners focus more on the pronunciation items that are core and are found to enhance intelligibility, rather than on the entire pronunciation inventory. This research also has fundamental implications for assessment - the ELF pronunciation influenced by speakers' linguistic repertoires can be considered as legitimate rather than as pronunciation errors (cf. Deterding 2013; Schaller-Schwaner, this volume; Walker 2010).
Research in pragmatics has been more extensive, but has provided similar results in terms of recurrent uses of ELF rather than random learner errors. Numerous pragmatic studies have focused on understanding/non-understanding in the attempt to identify those aspects or expressions that facilitate the solution of understanding problems (see Cogo and Dewey 2012; Kaur 2009; Mauranen 2006; Pitzl 2005). Research has shown that misunderstanding issues are less frequent than might be expected and that interlocutors tend to preempt or signal possible issues in a problematic exchange. The focus, therefore, has been on the strategies used for dealing with pre-empting, addressing or resolving issues in communication, such as the use of repetition, paraphrasing or co-construction of idiomatic expressions (Cogo 2010; Kaur 2012; Pitzl 2009; Seidlhofer 2009).
Another aspect of pragmatic investigation is the repertoire of multilingual practices that is creatively co-constructed and flexibly integrated (Hulmbauer 2011; Kalocsai 2013; Pitzl 2012; Vettorel 2014) in ELF communication. This implies strategies involving both code-switching and trans-languaging (Cogo 2012b) for meaning making, expressing a specific orientation to the talk (playful, engaged, irritated etc.) and expressing cultural and identity functions.