Discourse analysis (DA) is underpinned by a social constructionist orientation to knowledge. Social constructionism rests on the philosophical assumptions that multiple versions of the world are legitimate; that texts are open to multiple readings; and that language is non-representational. As social constructionism is relativistic, the status of 'evidence' generated by DA is questionable from more traditional research perspectives. On a common-sense level, people obviously construct meaning in relation to their lives. Thus, DA can help us to examine constructions of meaning in relation to nursing care. Equally the discourse analyst constructs one possible meaning in relation to a phenomenon that may compete with other versions. Multiplicity does not necessarily entail anarchy and competing versions prevent authoritarianism and loss of freedom. However, judgements have to be made about competing versions, for example, by assessing the level of 'facticity', or referring to the ethics embedded in the cultural context. In this paper, Bob White discusses DA as a form of qualitative research that offers promise for nursing research. Subsequent papers will examine the methodology and methods of DA and its application to nursing research.