Discourse can be seen as the things we use in communication. That is: language, forms of writings, images and objects, our bodies and so forth. Each of these items - let us call them discursive (or semiotic) resources - can be studied in its own right as a latent object. Language, for instance, will thus be studied in linguistics. But such studies have little to do with discourse. Language (to stick to the example) changes into discourse as soon as we look at language-in-action, as something that is effectively used in communication.
When we study discourse, consequently, we will need to take into account vastly more factors than when we study the resources in their own right, latent objects. In discourse analysis, for instance, we will always look at:
- Context: where, when and with which means was the communication conducted? Can we describe, in detail, the communicative event, and can this event be situated into broader categories such as genres (think of e.g. a talkshow or a quarrel between people).
- Intertextuality: what were the sources from which the resources used drew their actual meaning? What are the histories of use (and abuse) of those resources? Do certain words, for instance, carry a slang or jargon meaning that can be used in this particular context? And do certain accents in a language, for instance, send signals as to identity, tone, intention to the participants?
- Participants: who performed the communication? Were they equal or unequal participants (e.g. teacher-student or doctor-patient), and how did existing social relationships (e.g. between participants of a different gender, race, class, ethnicity of sociolinguistic group) get enacted? Did they remain stable or were they challenged, modified, altered?
- Actions: in communication, we rarely do just one thing. A "conversation", for instance, can be seen as one big action, but within that conversation, there will be question-answer parts, small narratives, people performing parallel actions (e.g. looking at their smartphones) and so forth. So what exactly did people do in communication?
These four factors combined show the difference between e.g. linguistics and discourse analysis: whereas in linguistics we would primarily be interested in what remains stable in language, in discourse analysis we would be interested in looking at how language gets continually adjusted to the circumstances of communication, shaped by these four parameters. Communicative resources are, in actual communication, twisted and bent in the directions defined by the actions. Thus, a conversation can start in a friendly tone but change into an angry one, a quarrel, and that change in general direction can change the meanings of words and other resources used there. This effect of context on text, in which meanings become adjusted to what they must contextually convey to interlocutors, is known as indexicality, and it is at the heart of discourse analysis.