Discourse analysis in applied linguistics: key concepts and theoretical frameworks

ELT Pub

Knowledge hub · Concept guide

1. What counts as discourse?

A university seminar, a journal article, a feedback conversation, and a messaging thread all involve more than individual sentences. Contributions relate to earlier contributions, invoke shared expectations, and position people in relation to one another. Discourse analysis investigates how such meaning-making is organised and how it connects with the activities and social settings in which it occurs.

The field is sometimes introduced as the study of language beyond the sentence. That definition is useful but incomplete. A one-word response can be analytically important because of its position in an exchange, while a lengthy text can be studied for the social relationships it constructs. Discourse analysis is a family of approaches, not one universally agreed procedure. The Routledge handbook illustrates the breadth of methods and educational or institutional applications.

In applied linguistics, the focus might be how teachers allocate participation, how writers establish a research contribution, how multilingual speakers manage understanding, or how policy texts represent language learners. The relevant data might be recorded talk, written documents, images, screen recordings, or combinations of these. The task is to show how observable details support an explanation, rather than simply attaching a theoretical label to a topic.

Discourse analysis overlaps with pragmatics. Pragmatics often foregrounds situated interpretation and action; discourse approaches may foreground textual organisation, sequential interaction, social practice, or ideology. The boundary is neither absolute nor a division between short and long texts. A researcher can combine perspectives, provided the combination is theoretically coherent and the evidence can support the resulting claims.

2. Key concepts for reading texts and interaction

Cohesion and coherence

Cohesion concerns linguistic connections across a text: reference, conjunction, lexical relations, ellipsis, and substitution. In “The interview was postponed. This delayed the analysis,” the demonstrative expression connects the second sentence with the earlier event. Repetition of related vocabulary can also maintain a topic. These devices make connections available, but their presence alone does not guarantee an effective text.

Coherence concerns how an audience makes sense of a text or interaction as a meaningful whole. It relies on topic development, communicative purpose, contextual knowledge, and expectations about the activity. “The road was flooded. We held the class online” can be coherent without an explicit causal conjunction. Conversely, adding “therefore” between unrelated claims does not establish a defensible argument. In writing instruction, this distinction helps explain why a paragraph can contain many linking expressions yet remain difficult to follow.

Genre and communicative purpose

A genre is a recognisable way of accomplishing communicative work in a community or setting. Research article introductions, grant applications, consultation openings, and complaint emails involve patterned expectations, although those expectations are variable and negotiable. A genre is not just a format or a fixed sequence of headings. Its organisation relates to what participants are trying to accomplish and to the institutional circumstances of that work.

In academic writing, move analysis examines segments that perform rhetorical functions, such as establishing a research territory, identifying an unresolved problem, or presenting the current study. A move may span several sentences, and one sentence may perform more than one function. John Swales’s Genre Analysis is a foundational contribution to this tradition. Treat such models as tools for investigating variation, rather than templates every successful text must reproduce.

Register, stance, and identity

Register refers to variation associated with situations of use. In systemic functional descriptions, field concerns the activity and subject matter, tenor the relationships between participants, and mode the role and organisation of language in the situation. This is broader than labelling a text “formal” or “informal.” A tutorial message and a published abstract can differ in their goals, degree of shared context, interactional possibilities, and relationships.

Stance concerns how speakers or writers express commitment, evaluation, and alignment. “The findings demonstrate” and “the findings may suggest” position a claim differently, but the effect depends on discipline, evidence, and surrounding argument. Identity is likewise approached through communicative practices: who is positioned as expert, novice, legitimate speaker, or decision-maker? Analysts should connect such interpretations with repeated patterns and contextual evidence rather than equate a single pronoun or modal verb with a stable identity.

Intertextuality and recontextualisation

Texts frequently incorporate or respond to other texts. Quotation, citation, paraphrase, reported speech, and familiar institutional formulations are forms of intertextual connection. Recontextualisation concerns what happens when material moves into a new setting: a research finding becomes a press release, an assessment policy becomes a classroom instruction, or a student’s words become an example in a report.

Such movement changes what is foregrounded, omitted, and authorised. An analyst can compare versions, identify whose voice is represented, and examine how certainty or responsibility changes. This is particularly relevant to public communication about language education, where nuanced research claims can become simplified prescriptions.

3. Major approaches and their commitments

Conversation analysis: action in sequence

Conversation analysis examines how participants organise social action through the sequential placement and design of turns. Turn-taking, adjacency pairs, repair, and sequence expansion are central concerns. A question makes an answer relevant; a request makes acceptance or refusal relevant. The details of timing, overlap, intonation, and embodied conduct may matter, depending on the question and data.

Its emphasis is on participants’ displayed orientations. Rather than assume that a pause expresses anxiety, an analyst examines how the pause figures in the surrounding interaction and how others respond. Repair concerns trouble in speaking, hearing, or understanding; it is broader than grammatical correction. A teacher’s “Sorry?” may invite repetition, clarification, or a revised answer, and the next turn helps establish which action participants treat as relevant.

In language education, this approach can reveal how learning opportunities arise through interaction and how competence is jointly accomplished. The Cambridge chapter on conversation analysis in L2 teaching and learning discusses this connection. Conversation analysis is not simply any qualitative study using interview transcripts: its analytic commitments concern how interaction is organised, including the interview itself.

Systemic functional linguistics: language as a resource for meaning

Systemic functional linguistics, associated with Halliday and developed by many scholars, treats language as a system of meaning-making choices. Three broad metafunctions concern construing experience and logical relations, enacting interpersonal relationships, and organising messages. The linguistic description connects these functions with grammatical resources rather than treating grammar only as a list of rules for correctness.

An analyst might examine who is represented as acting, what processes are selected, how certainty is expressed, or which information becomes the point of departure of a clause. Compare “The committee rejected the appeal” with “The appeal was rejected.” The second formulation leaves the actor unexpressed, but that observation does not by itself establish deliberate concealment. Interpretation depends on what the wider text and situation make relevant. Halliday and Matthiessen’s reference grammar provides a detailed account of the framework.

Genre analysis: recurring purposes and rhetorical organisation

Genre analysis investigates how recurrent communicative purposes are organised within disciplinary or professional communities. In English for Academic Purposes, it can connect close reading of texts with interviews, institutional documents, and knowledge of publication practices. Analysts examine typical moves, their realisations, their ordering, and the variations associated with different contexts.

For a study of literature reviews, counting citations might identify a pattern, but it cannot explain how writers use those citations to establish agreement, identify a limitation, or create a space for their own contribution. A useful analysis attends to both rhetorical function and linguistic expression. It also recognises that disciplinary conventions change and that writers may strategically depart from them.

Critical discourse studies: language, power, and social practice

Critical discourse approaches ask how discourse participates in sustaining, negotiating, or challenging social inequalities. They examine relationships between textual choices and broader institutions, histories, and ideologies. Fairclough’s work, for example, connects detailed attention to language with analysis of social practices. His Critical Discourse Analysis brings together work developing this perspective.

In applied linguistics, a question might concern how a policy represents multilingual students: as deficient recipients of support, as measurable performance units, or as knowledgeable participants. An analyst could inspect labels, agency, evaluation, presuppositions, and whose perspectives are included. The critical orientation must be supported by an explicit argument. A single passive clause is not proof of oppression, and an institutional intention should not be inferred without appropriate evidence. The handbook of critical discourse studies offers a broader view of the field’s methodological diversity.

Corpus-assisted and multimodal approaches

Corpus-assisted analysis uses collections of texts to identify patterns such as frequency, collocation, keywords, and recurring phraseology. Concordance lines allow researchers to return from a numerical pattern to examples in context. Corpus design matters: a collection of published abstracts cannot represent everything researchers write, and apparent differences may reflect topic or sampling rather than the social category under investigation. Frequency is evidence of recurrence, not an automatic measure of importance.

Multimodal analysis examines how language works with images, gesture, gaze, layout, sound, and other resources. A teacher’s verbal instruction may depend on a pointing gesture, while a digital learning page may establish a reading path through size, colour, spacing, and links. Researchers need to justify which modes they record and how they analyse their coordination. The Paltridge introduction places these concerns within the wider landscape of discourse analysis.

4. A worked classroom example

A sequential analysis observes how the teacher’s follow-up makes elaboration relevant and how the student supplies it. The final turn reformulates the answer and opens a further activity. The exchange can be related to the familiar initiation–response–follow-up pattern, but that label alone misses the specific work performed by the follow-up. Not every third turn simply evaluates an answer as right or wrong.

A functional analysis could examine how the modal “might” and the student’s “maybe” position explanations as tentative. A genre perspective could ask how this classroom activity prepares students to participate in methodological evaluation. A critical analysis could investigate how rights to ask, reformulate, and evaluate are distributed across a wider collection of classroom episodes.

The transcript cannot establish whether a particular student learned a concept, how all classroom participation is organised, or whether the teacher’s intentions matched the analyst’s account. Those claims require further data. Because the example omits timing, intonation, and embodied conduct, it also cannot support fine-grained claims about turn transitions or gaze. The point is to compare questions, not to suggest that every approach can be applied fully to the same minimal extract.

5. Applications in teaching and academic literacy

For writing instruction, discourse analysis directs attention to how a text does its work. Students can compare introductions from their own discipline, identify how a problem is established, and examine where claims are qualified. They can then test those observations against a contrasting example. This teaches rhetorical judgement rather than a universal set of sentence starters.

For spoken interaction, recordings can help teachers notice who obtains turns, which questions receive extended responses, and how misunderstandings are handled. A useful task is to compare an exchange that closes discussion with one that invites explanation. The goal is to make the relationship between interactional design and participation visible, while recognising that the same practice can have different effects in different activities.

In multilingual education, discourse analysis can reveal resources that narrow correctness-based accounts overlook: translation, gesture, shared writing, reformulation, and collaborative completion. It can also expose how labels such as “weak learner” or “non-participating student” become established in institutional talk. These labels deserve investigation rather than automatic acceptance as neutral descriptions.

Assessment can include genre awareness, coherence, responsiveness, and audience design alongside sentence-level accuracy. Criteria should explain how these dimensions relate to the communicative task. A strong written argument and effective real-time discussion require overlapping but different resources, and neither can be evaluated adequately through a single count of grammatical errors or connective words.

6. A practical research workflow

Define the question and unit of analysis. Specify whether the study examines clauses, rhetorical moves, turns, sequences, whole documents, or a social practice spanning several texts. A broad interest in “identity” needs an account of what observations would support a claim about identity in the selected setting.

Build a defensible dataset. Explain inclusion criteria, time period, collection procedures, and relevant contextual differences. For recordings, capture the modes needed for the question and document the conditions of recording. For textual corpora, consider genre, topic, publication stage, authorship, and document provenance. Larger datasets do not compensate for a mismatch between the question and the sample.

Preserve context and make decisions traceable. Keep original files, document transcription conventions, and retain enough surrounding material to assess interpretations. Develop a coding guide where categorisation is part of the method. Explain ambiguous cases and inspect counterexamples. Reliability procedures should suit the approach: agreement statistics may support a coding study, while sequential analysis requires a transparent account of participant orientations and the organisation of the collection.

Connect pattern with explanation. Present extracts that demonstrate the argument, not merely illustrate a conclusion already announced. Where counts are used, explain denominators and comparability. Where interpretations rely on institutional context, identify the evidence establishing that context. Distinguish a textual possibility from an observed audience effect.

Handle ethics as part of analysis. Consent, anonymisation, and secure storage matter, particularly in classrooms where participation occurs within unequal relationships. Publicly visible online material is not automatically free of ethical concerns. Searchable quotations, distinctive events, faces, and voices may identify participants. Where translation is involved, explain choices that affect interpretation and avoid presenting a translated extract as if it preserved every feature of the original.

7. Common pitfalls and stronger alternatives

A frequent problem is announcing a framework and then offering thematic summaries with little attention to discourse. The stronger alternative is to demonstrate how specific forms, positions, or patterns contribute to the phenomenon being claimed. Another problem is selecting only striking examples. Build a collection, explain how cases were selected, and discuss examples that complicate the preferred account.

Frameworks also should not be treated as interchangeable toolboxes. Conversation analysis and critical discourse approaches differ in how they establish relevant context and explain social order. Combining them requires an explicit rationale; adding terminology from several traditions does not automatically produce triangulation. Choose an approach because it can answer the question, and describe the limits of what the resulting analysis shows.

Discourse analysis offers applied linguistics a way to connect close attention to language with questions about learning, participation, expertise, and institutional life. Its contribution depends on the clarity of that connection. For the interpretation of utterances and the management of social actions, see the companion guide to pragmatics.

References and further reading

• Paltridge, B. (2021). Discourse analysis: An introduction (3rd ed.). Bloomsbury Academic. Publisher page.
• Swales, J. M. (1990). Genre analysis: English in academic and research settings. Cambridge University Press. Cambridge University Press Bookshop.
• Halliday, M. A. K., & Matthiessen, C. M. I. M. (2014). Halliday’s introduction to functional grammar (4th ed.). Routledge. Publisher page.
• Fairclough, N. (2010). Critical discourse analysis: The critical study of language (2nd ed.). Routledge. Publisher page.
• Handford, M., & Gee, J. P. (Eds.). (2023). The Routledge handbook of discourse analysis (2nd ed.). Routledge. Publisher page.
• Flowerdew, J., & Richardson, J. E. (Eds.). (2018). The Routledge handbook of critical discourse studies. Routledge. Publisher page.