Doing Discourse Analysis

Methods for Studying Action in Talk and Text

Linda A. Wood, Rolf O. Kroger

• Publisher: SAGE Publications, Inc
• Year: 2000
• ISBN:
  9781506320113
• Edition: 1
• Topic: Social Sciences
• Subtopic: Social Science Research & Methodology
• Index: Social Sciences

Providing both the practical steps for doing discourse analysis and the theoretical justifications for those steps, this book shows students how the social world revolves around talk and text. The authors draw on conversation analysis, critical discourse analysis, pragmatics, and the discursive approach developed in social psychology. The book presents actual examples, covers data collection methods and strategies for data analysis, and addresses issues of reliability and validity.

PART I

THEORETICAL BACKGROUND

1

Language, Discourse, and Discourse Analysis

Discourse analysis is a perspective on social life that contains both methodological and conceptual elements. Discourse analysis involves ways of thinking about discourse (theoretical and metatheoretical elements) and ways of treating discourse as data (methodological elements). Discourse analysis is thus not simply an alternative to conventional methodologies; it is an alternative to the perspectives in which those methodologies are embedded. Discourse analysis entails more than a shift in methodology from a general, abstracted, quantitative to a particularized, detailed, qualitative approach. It involves a number of assumptions that are important in their own right and also as a foundation for doing discourse-analytic research. Our purpose in this chapter is to cover these basic principles. We note that there are actually multiple versions of the discourse-analytic perspective. For our present purposes, we present one particular version, which is our take on the version originally developed by Potter and Wetherell (1987). The Potter and Wetherell version is one of the better known approaches; furthermore, it is accessible, and its central principles are common to a number of versions.

Multiple perspectives on discourse mean that there are multiple definitions of discourse and of what counts as discourse (e.g., spoken language, written language, language use above the level of the sentence, etc.). What we want to stress here is an approach to definition that views discourse not just as an object, but as a way of treating language. In general, we find congenial the definition offered by Potter (1997):

[Discourse analysis] has an analytic commitment to studying discourse as texts and talk in social practices. That is, the focus is not on language as an abstract entity such as a lexicon and set of grammatical rules (in linguistics), a system of differences (in structuralism), a set of rules for transforming statements (in Foucauldian genealogies). Instead, it is the medium for interaction; analysis of discourse becomes, then, analysis of what people do. (p. 146)

In other words, language is taken to be not simply a tool for description and a medium of communication (the conventional view), but as social practice, as a way of doing things. It is a central and constitutive feature of social life. The major assumption of discourse analysis is that the phenomena of interest in social and psychological research are constituted in and through discourse. “Discourse theorists maintain that talk is constitutive of the realities within which we live, rather than expressive of an earlier, discourse-independent reality” (Sampson, 1993, p. 1221). Talk creates the social world in a continuous, ongoing way; it does not simply reflect what is assumed to be already there.

The discursive perspective does not deny physical reality. As Wetherell and Potter (1992) note, a mountain is no less real for being constituted discursively as either a volcanic eruption or as a result of tectonic shifts. You still die if your plane crashes into it. They add, “However, material reality is no less discursive for being able to get into the way of planes. How those deaths are understood … and what caused them is constituted through our systems of discourse” (p. 65). As Sampson (1993) puts it:

The very objects [and events] of our world are constituted as such in and through discourse. There is no meaning to reality behind discourses that discourse represents; in the representation lies the constitution of what we come to accept as the real. (p. 1222)

These ideas require elaboration. To grasp the nature of discourse in the sense that we use the term means confronting a number of assumptions, assumptions that are usually taken for granted by both social scientists and laypersons. This view of discourse entails (at least) three major shifts from conventional orientations: (a) from a distinction between talk (discourse) and action to an emphasis on talk as action, (b) from a view of talk (discourse) as a route to internal or external events or entities to an emphasis on talk as the event of interest, and (c) from a view of variability as an anomalous feature of action to an appreciation of variability both within and between people (Potter & Wetherell, 1987).

BASIC ASSUMPTIONS ABOUT DISCOURSE

Language Is Action

A major source of the view that language is action is Austin’s (1962) theory of speech acts. Austin pointed out a feature of language that is known implicitly by all language users: Utterances not only have a certain “meaning” (i.e., they refer to states, persons, events, etc.), they also have force, that is, they are not only about things, they also do things. In other words, talk (and language use more generally) is action. Specifically, utterances can be considered in terms of three features: (a) their locutionary or referential meaning (what they are about), (b) their illocutionary force (what the speaker does with them), and (c) their perlocutionary force (their effects on the hearer). For example, if we say to someone, “You stole the money,” we are not only describing an event—and the statement could thus be judged as true or false—we are also accusing the hearer (who may be hurt by the accusation, become angry, etc.). The emphasis in discourse analysis is on what talk is doing and achieving.

Further, we need to recognize that talk has multiple functions. For example, the speaker of the above statement might also be said to be insulting the hearer, constructing the identity of the hearer as a liar, and so on. Finally, what a person might be doing with talk depends on the other persons involved, the circumstances, and so on. These factors are captured in Austin’s notion of felicity conditions, that is, the conditions that must obtain for an utterance to have force as a certain speech act (accusing, etc.). For example, an imperative statement can only be an order if uttered by someone with authority over the hearer. In sum, the discourse analyst will look at discourse for what is being done, not primarily for what it is about.

To appreciate the wider importance of language, language must be seen as action, but this is not always easy to do. In part, this is because both social science and everyday discourse make an unfortunate distinction between talk and action. Both systems privilege action and tend to downgrade talk. There are numerous examples of this tendency, for example, the dismissive reference to utterances as “just talk” or expressions such as “We’ve talked enough—it’s time for action” and more recently, “He can talk the talk, but can he walk the walk?” Nonetheless, people do recognize that talk has a performative quality, that we can do things with words. Let us look more closely at a frequently used expression to emphasize the importance of the distinctions drawn in Austin’s speech-act theory. The statement “actions speak louder than words” appears to make a distinction between talk and action, giving primary importance to actions. But the verb that carries the notion of importance and the idea of action is “speak”; the expression appears simultaneously to deny and recognize the importance and action nature of language. The seeming paradox can be resolved by distinguishing between two senses of talk: talk as words, and talk as what people are doing with words. Similarly, the utterance “talk doesn’t count” would reflexively render itself pointless unless we interpret talk as referring to words rather than to what can be done with them.

What are speakers doing with their utterances? We can say first of all that they are recognizing the performative quality of discourse. But if we consider such utterances in context, we can identify other functions. For example, a newspaper article reported that China would observe the 50th anniversary of the Universal Declaration of Human Rights while continuing to crack down on dissidents (Cernetig, 1998). The headline read, “China’s words on human rights louder than its actions.” The claim that China supports human rights in words but not in actions not only serves as an accusation that China is behaving badly; it marks that behavior as particularly egregious. China cannot reasonably offer an account that it is unaware of the Declaration and what it involves. By combining reports of China’s observation of the anniversary with reports of the crackdown, the journalist presents China’s actions as a mockery, not simply a neglect, of human rights. Or consider the headline of a report on Canadian chartered banks: “Banking and Poor People: Talk is Cheap” (see Philp, 1998). The article also includes a quotation from the report: “When it comes to serving the needs of poor people, their record is lots of talk and not much action” (quoted in Philp). In this case, the emphasis on talk versus action serves as an accusation that the banks have failed to keep their commitment to open their doors to low-income people. In this case, we have talk in the form of a promise; clearly, such talk is not inherently empty, or we would pay it little attention. This talk serves to commit the speaker to a future action. In this case, the action is not forthcoming. But what is the nature of that action? The phrase “open their doors” does not appear to refer to talk, but it is also not used literally. That is, no one is suggesting that poor people are not allowed to come into the bank. Rather, they are not permitted to engage in the sorts of conversation and exchange of texts (checks, signature cards, account agreements, etc.) that would give them access to the banks’ services.

We can extend our analyses if we consider phrases such as “Actions speak louder than words” and “Talk is cheap” as idiomatic expressions. Drew and Holt (1988) have shown that idiomatic formulations are difficult to challenge, both because of their generality, that is, their independence of the specific details of the particular situation or person to which they are applied, and because they invoke and constitute the taken-for-granted knowledge shared by all competent members of the culture. Because they are difficult to resist, they are rhetorically effective.

These samples of discourse analysis are by no means complete; we demonstrate later how to provide a much more thorough analysis and how to warrant that analysis. Our main goal here has been to illustrate that discourse analysts are not uninterested in content but that their aim is to go beyond content to see how it is used flexibly to achieve particular functions and effects. The role of content requires some elaboration, because people sometimes take the action emphasis of discourse analysis to mean a lack of concern for content. On the contrary, discourse analysis can be said to take content more seriously than do conventional approaches in which the focus is on the psychological processes ostensibly underlying content (Potter & Wetherell, 1995a, p. 82). The misunderstanding also reflects a narrow view of content, whereas the conceptualization of content in discourse analysis is very broad. Thus, discourse analysts do consider the referential and propositional content of discourse, although not in terms of its truth value. The analysis of how language is used also involves the identification of content at a number of different levels: lexical, pragmatic, discourse, and social. For example, the utterance “You’re not going out again tonight, are you?” might be a reprimand, an initiation (of a conversational exchange), and a threat to the face of the other. All of these involve the performative force or function of the talk and might all be said to be the content of the discourse. In some instances, a word might be semantically empty in that it has no representative or referential meaning. But it can function as a speech act. For example, in the instance above, if “Well?” were to follow the failure of the other person to respond to the “question,” it might serve as a challenge. Thus, there is no reason to restrict the term content to the referential aspect of utterances.

Another comment we hear quite frequently is that discourse analysis is not concerned with content, only with “style.” Style is sometimes referred to as the how versus the what (content) of language use. But the two cannot be separated. Style is how we (both users of language and discourse analysts) get from one type of content to another; doing something in a certain way makes it into another sort of thing. For example, the linguistic or intonational form of an utterance as a question versus a statement has implications for the content of the utterance as a request rather than as a command. This does not mean that pragmatic content is always discernible from semantic content and style; for example, a hint is by definition not obvious from the semantic or propositional content of an utterance, regardless of how it is uttered. The recipient and the analyst must also consider the context in which it is uttered. There is a further complication. Style is sometimes thought of only in relation to the nonverbal aspects of discourse. But verbal features, including content, may also contribute to style. For example, the use of formal language (e.g., Latinate, nominal forms) produces a very different style from the use of informal language (as is apparent to anyone who has listened to the remarks to the media by a police officer at a crime scene). In sum, just as form or style is part of content, content is part of style. As soon as we move beyond the level of content as usually understood, we cannot separate content and style. “One cannot properly analyse content without simultaneously analysing form, because contents are always necessarily realised in forms, and different contents entail different forms and vice versa. In brief, form is a part of content” (Fairclough, 1992a, p. 194).

Function

The concept of function is central to discourse analysis. It requires some elaboration, in part because, like discourse and text, the term is used in a number of ways. For example, function may be construed in terms of relationships between units (e.g., a linguistic form and the other elements of the system in which it is used). The iexmfunction as we use it refers to action, or to what it is that people are doing in and with their talk and text (or more casually, what the talk or text is doing). Action may refer to anything that can be done using words, from connecting words and phrases to speech acts, including the communication of ideas, attitudes, meaning, and so on. Particular units may have characteristic or expected functions, although the interpretation must always be confirmed in the particular instance. For example, we expect that and will usually be used to connect two words or clauses, and it usually does. In some cases, there may be agreement about function, for example, that the function of a phrase such as ‘Tm not prejudiced, but” is to disclaim. The discourse in question may have no characteristic function in that its action varies from context to context. For example, “That’s not a good idea” may be used to evaluate in one context, but to warn in another. Function as action may entail the idea of a specific consequence, as in “persuading,” which requires not only activity on the part of a speaker, but a particular effect on the hearer.

We can distinguish types of function in several ways. For example, functions may be linguistic (e.g., to connect two phrases), pragmatic (to convey nonsemantic meaning, e.g., the use of well to mark the insufficiency of a previous utterance), or social (e.g., to save face). Functions may be specific or global (e.g., requesting vs. self-presentation; see Potter & Wetherell, 1987, p. 33). Functions may also be described in terms of the type of domain to which they are relevant, for example, legal or religious functions. As we noted above, there is also a connection between content and function in that the former is often described in terms of the latter, particularly when function involves meaning. For example, a request may be viewed as both the content of an utterance and its pragmatic function.

We add a brief comment on structure. Most discourse-analytic approaches can be described as structural in that they do not treat specific elements of discourse in isolation from other elements, but rather attend to their combinations and arrangement. Structure may be sequential or hierarchical, within or across utterances (e.g., in examinations of turn taking). The emphasis, however, is not simply on the arrangement of elements, but on structure as an active process, that is, on structure in the sense of construction.

Discourse as Focus or Topic, Not as Route or Resource

The emphasis on discourse as action and as constitutive of phenomena entails a shift from the usual focus of interest in the phenomena to which the discourse refers to a focus on the discourse itself. Ethnomethodologists (e.g., Garfinkel, 1967) view this as a shift from using features of talk to explain behavior (talk as resource) to a focus on the features of talk as the behavior to be explained (talk as topic). For many, this is a difficult move to make. It means that the emphasis shifts away from a concern with talk as a description of or route to what people “really” think. This is not because it is impossible to gain access to the internal in any direct way, nor is it simply a behaviorist escape hatch (Potter & Wetherell, 1987). But it is an essential move if we accept the proposition that mind (vs. brain) is constituted discursively (rather than existing as an entity prior to and independent of language) and that the separation of thought and talk (and the privileging of one or the other as more “real”) is therefore problematic (Harré & Gillett, 1994). Similarly, the argument from discursive metatheory, that events in the world are constructed discursively, means a shift in concern from what “really” happened to how those events are discursively constructed in the social world. The discursive orientation constitutes a counterpoint to the current emphasis on cognition and perception (on things “in there,” under the skull, in the mind) rather than on events in the world, including talk as action (see Edwards, 1997, for an extensive discussion of the problems with the prevailing cognitive approach).

The idea of looking at talk or discourse as a topic for research in its own right can be problematic for a different reason than that discussed previously. It is sometimes taken to mean a focus on language as language, that is, in terms of such features as phonology, syntax, or semantics. Some discourse analysts do attend to such features for specific purposes. But generally speaking, the topic for discourse analysts is more properly framed not as language or talk, but in terms of the phenomena that are constructed discursively (e.g., racism, abuse), that is, in terms of what people are doing with words. So, discourse analysts generally do not focus on words as such, as linguistic objects. Nor do they focus on the referential function of words. (We take up the issue of topic in more detail in Chapter 2.)

As a brief illustration of these views about talk, consider the response in the following exchange (from a telephone conversation between Wood and a 75-year-old family friend): “Would you like to come for tea on Sunday?” “I’m sure I can make it.” Our interest is not in the literal truth of the friend’s statement, in whether the woman is really sure. Both analyst and inviter will be more concerned with what the woman is doing with the words, that is, accepting the invitation. But utterances almost invariably have multiple functions. For example, the phrase “I’m sure” is not required as part of the acceptance of an invitation, so we need to ask what it is doing there. It suggests (and simultaneously discounts) the possibility that the woman might not be able to accept, perhaps because she has a conflicting social engagement or because she has health problems. Both possibilities can serve in different ways to construct the identity of the woman as an active social...

Table of contents

• Cover Page
• Title page
• Copyright
• Contents
• Acknowledgments
• Introduction
• PART I: THEORETICAL BACKGROUND
• PART II: METHOD
• PART III: ANALYSIS
• PART IV: EVALUATION AND REPORTING
• Appendix A: Transcript Notation
• Appendix B: Selected Varieties of Discourse Analysis
• Glossary: Terms for Discourse Analysis
• References
• Index
• About the Authors