Within the field of structural linguistics, Harris defined discourse as ‘stretches longer than one utterance’ (1951: 11), which are analyzed at the discourse level. He referred to discourse analysis as the study of ‘stretches of speech longer than one sentence’ (1952: 30). He noted that utterances occurred in a ‘conversational exchange’ (1951: 14), using the term utterance to differentiate it from sentence. However, the notion of social action did not play a central role in Harris’ model of discourse. Finally, in his socio-semiotic theory of language, Halliday (1978) alluded to discourse in his concept of text, which he defined as ‘instances of linguistic interaction in which people actually engage: whatever is said’ (1978: 108). Halliday used the concept of ‘text’ to refer to a ‘supersentence,’ namely, a linguistic unit bigger than a sentence (p. 109).
Although ‘discourse analysis’ has been approached from different perspectives, there is a general consensus regarding the elements that characterize discourse. According to Stubbs, discourse analysis refers to ‘the organization of language above the sentence or above the clause, and therefore to study larger linguistic units, such as conversational exchanges or written texts’ (1983: 1). Brown and Yule stated that discourse analysis should not be limited to the analysis of descriptive linguistic forms disconnected from social interaction; instead, it should situate ‘the analysis of language in use’ (1983: 1). And after a meticulous analysis of the conceptualization of discourse from diverse approaches, Schiffrin concluded that discourse refers to ‘utterances as units of language production (whether spoken or written) that are inherently contextualized’ (1994: 41).
The term discourse analysis has been used as an umbrella term to examine stretches of talk or text. There are different types of spoken interaction: natural conversation, narratives, and institutional interactions in formal (e.g. doctor-patient interactions, court room discourse, job interviews) and non-formal settings (e.g. service encounters in small shops, department stores, information centers, or travel agencies). Other researchers take a narrow methodological perspective to analyze conversation in everyday and institutional talk.
The type of discourse selected, a family dinner, a court hearing, or a service encounter, for example, is often analyzed on the basis of at least one of the following characteristics: the level of formality/informality, the situation or speech event, symmetric or asymmetric relations, goals and statuses of participants, the rights and responsibilities of the participants, and the social setting in which socialization takes place.