Abstract Lee

Methodological issues in building a corpus of doctor-patient dialogues annotated for metaphor.

In this paper we will discuss our experiences in building a corpus of doctor-patient dialogues and annotating the corpus for use in metaphor research. We will concentrate on the methodological issues of building and annotating the corpus but also present some recent research findings about the range and type of metaphor employed. Our corpus was created from the transcription of consultations between doctors and patients in a rheumatology outpatients clinic and our long-term research goal is to discover how doctors and patients conceptualize and describe rheumatoid conditions such as arthritis and related symptoms. We accept the cognitive linguistic claim that metaphors are essentially conceptual rather than linguistic and therefore hope that studying how metaphors are used in this domain will provide insight into how doctors and patients conceptualize notions of illness, recovery etc.

Semino et al. (2004) identify several methodological issues which need to be addressed with corpus-based studies of metaphor. In addition, building a corpus of doctor patient dialogues presents particular problems since the researcher must be careful not to intrude or otherwise influence the dialogue while still ensuring that both the doctor and patient give their full consent that the dialogue is recorded. Furthermore the recorded data must be anonymised to protect patient confidentiality. We will therefore present a research methodology for building such a corpus while not influencing the data either during collection, anonymisation or annotation as far as possible and describe a protocol for annotating metaphor in this corpus which has resulted in high interannotator agreement in terms of kappa scores.

Our investigation shows that metaphor is very common in the language of this type of medical consultation. There are two reasons for this. First, metaphor is often used to either describe or substitute for medical terminology which may be unfamiliar to the patient. Secondly, there is often a need to describe abstract aspects of rheumatoid conditions in concrete terms. For example, different types of pain are described using different conceptual metaphors. There is also often an emotional aspect to the use of metaphor for instance where the patient uses metaphorical language to communicate emotional judgements.

References

Deignan, A. (2005). “Metaphor and Corpus Linguistics”. Amsterdam: John Benjamins.

Low, G. (1999) “Validating metaphor research projects”, in Cameron, L. and Low, G. (eds.) Researching and Applying Metaphor, Cambridge: Cambridge University Press.

Semino, E., Heywood, J. and Short, M. (2004) ‘Methodological problems in the analysis of a corpus of conversations about cancer’, Journal of Pragmatics, 36:7, 1271-1294