Investigating the presentation of speech, writing and thought in spoken British English : A corpus-based approach

In this paper we describe the Lancaster Speech, Writing and Thought Presentation (SW&TP) Spoken Corpus. We have constructed this corpus to investigate the ways in which speakers present speech, thought and writing in contemporary spoken British English, with the associated aim of comparing our findings with the patterns revealed by the previous Lancaster corpus-based investigation of SW&TP in written texts. We describe the structure of the corpus and the archives from which its composite texts are taken. These are the spoken section of the British National Corpus, and archives currently housed in the Centre for North West Regional Studies (CNWRS) at Lancaster University. We discuss the decisions that we made concerning the selection of suitable extracts from the archives, the re-transcription that was necessary in order to use the original CNWRS archive texts in our corpus, and the problems associated with the original archived transcripts. Having described the sources of our corpus, we move on to consider issues surrounding the mark-up of our data with TEI-conformant SGML, and the problems associated with capturing in electronic form the CNWRS archive material. We then explain the tagging format we adopted in annotating our data for Speech, Writing and Thought Presentation and discuss how this was developed from the earlier version used for tagging written texts. We also discuss some preliminary analyses which point towards fruitful future lines of investigation