Chapter 5 describes the process of constructing the MI 1984–2014 Corpus, from compiling the sampling frame (e.g. search terms, dates covered) to the compilation procedure for the illness and year subcorpora. In particular, a detailed discussion of the interpretative status of search terms is provided. Practical issues related to compiling corpora such as cleaning the data are discussed. Furthermore, the problems that the interdisciplinary nature of corpus construction poses for the researcher are outlined.

We find out what scientists know about a topic by searching the scientific literature. Literature searches range from a preliminary search to find out what we know about a general area to a specific search on a precise topic. As we explore a research topic, we focus our searches to identify the main open questions, the hypotheses proposed and the support for them, potential model systems and methods, and the experts in the field. Broad background reading is also fundamental preparation for a study because no study goes as planned and we may need to identify new research questions as we progress. I begin this chapter with sources of information we have available, then describe how we identify search terms and assess the quality of the literature we find, I explain the importance of reading broadly and how to choose what to read, and end with how we keep up with the literature

Chapter 4 contains extensive references to exemplifications of the voice quality categories described in Chapters 1 and 2. Distinctive voice quality settings of actors, singers, media announcers, politicians, and other personalities are cited with specific references to be used as search terms to locate video and audio material online. A separate ‘Multimedia References’ section is included at the end of the text to facilitate searching. Long-term voice qualities are essentially extralinguistic, but they can also alternate among phrases of speech as paralinguistic ‘registers’. Many laryngeal postures, and their auditory output, are employed linguistically as syllabic registers or segmental units to signal contrastive lexical meaning. This is a result of the positioning of the laryngeal articulator at the beginning of the shaping of the sound stream, where background elements are easily patterned behind oral articulations. Extensive references to linguistic examples of lower-vocal-tract consonantal strictures and tonal register effects are presented within the text notes and are included among the video and audio materials online.