Generally speaking, the research of the VOICE project is informed by perspectives of three kinds: conceptual, descriptive and methodological.
Research in the VOICE project is conceptualized along the same lines as two other areas of language study, namely corpus linguistics and research into language variation. The difference is that the VOICE researchers were concerned with data which these areas have not hitherto focused on.
Corpus linguistics has demonstrated how effective computer-aided analysis can be in the detailed description of actually-occurring language behaviour. It has, however, been mainly concerned with native-speaker usage. The aim of VOICE was to extend the scope of such analysis to actually-occurring ELF usage.
Language variation and change is a field of research which sociolinguistics and historical linguistics have been concerned with for generations. In this sense, there is nothing original about VOICE. The factors which affect the variable use of ELF are the same factors which are relevant to the understanding of how languages always vary and eventually change according to different circumstances. What is surprising is that the phenomenon of ELF has so far not been given the same scholarly attention as other kinds of linguistic variation. VOICE seeks to remedy this situation.
The overall purpose of VOICE is to provide an empirical basis for a sound and in-depth description of spoken ELF.
The primary research interests of the VOICE team lie in the identification of lexicogrammatical features of ELF, an area in which hardly any description has been undertaken so far. The mark-up conventions developed for VOICE are designed to support research into this aspect of ELF.
Although only initial observations about the lexicogrammar of ELF talk have emerged so far, these suggest that various features which have usually been regarded as learner errors are produced regularly by ELF users from many different first language backgrounds without posing an obstacle to communicative success. A number of these features are listed on a reference page about ELF in the Oxford Advanced Learner's Dictionary (7th edition).
The members of the VOICE team have necessarily been involved in developing a new methodology appropriate for dealing with ELF.
Methodological challenges faced in the process of compiling a corpus of spoken ELF have included, among other things, theoretical questions such as how to delimit 'English as a lingua franca', i.e. which criteria to use for the identification and selection of data for the corpus, but also more practical issues such as the relationship between the immediate reality of spoken language and its representation on paper and screen. Concerning the latter, Breiteneder, Pitzl, Majewski & Klimpfinger (2006) highlight that the question of what a transcript should actually look like and which decisions have been made prior to and during the process of transcribing is crucial for all subsequent analyses. For instance, as VOICE is concerned with a use of English that has hitherto not been described, let alone codified, a new set of spelling conventions was designed to render the diversity of ELF speech in a standardized way, but dissociated from a purely British or American standard (cf. the VOICE spelling conventions). The challenge for transcribers of ELF talk is a considerable one because of the temptation to note down what is correct or appropriate in native-speaker terms, rather than what is actually said. To make the difficult task of transcribing spoken ELF easier, the VOICE team have been working on software which supports the processes of data transcription and checking. This has resulted in the creation of the custom-made software VoiceScribe released for free-of-charge access in September 2007.