CHAPTER 1
Conceptual Framework
If âdiscourseâ is nothing less than language itself, and âdiscourse analysisâ attempts to admit a broad range of research to the analysis of language, then it is by nature interdisciplinary.
âTannen, 1990, p. 10
Discourse is a difficult concept, largely because there are so many conflicting and overlapping definitions formulated from various theoretical and disciplinary standpoints.
âFairclough, 1992, p. 3
Discourse is language-in-the-world, and discourse analysis is thus the study of language in all its located complexity and glory. Although discourse and discourse analysis have been variously conceptualized, these are the definitions I will adopt here.
Given these broad definitions, studies in rhetoric/composition of how people use language to operate in and on their worldly environments are located squarely within the field of discourse analysis. At the same time, rhetorical studies at least has maintained an existence, historically speaking, somewhat apart from the more linguistically oriented traditions of discourse analysis. It has therefore developed theoretical assumptions and methods that do not always comport comfortably withâand which in some cases may even opposeâtheir counterparts in the linguistics-influenced traditions. To give but one general example: The formerâs emphasis on rhetorical effectiveness seems epistemologically opposed to the latterâs relativist stance that all language is by nature equally suited to its contexts of use and culture.1
As an interdisciplinary endeavor, discourse analysis will always contain within it the potential for internal discoherence and fragmentation based on such differences in âprofessional visionâ (Goodwin, 1994)âthis is not a unique feature of the relationship between rhetorical and linguistic discourse analysis. It does suggest, however, that the introduction of concepts central to a study such as this one, which attempts to bring the two traditions closer together, may be a necessary place to begin. In order therefore to establish grounds for clear communication and understanding in this complex, âpluralisticâ endeavor (Kirsch, 1992), this chapter is devoted to defining major theoretical concepts, or âconventions for construing realityâ (Bizzell, 1982), underlying the research on which this book is based. As will be seen, the definitions themselves often aim to integrate theory in rhetoric/composition and linguistics, or to use earlier attempts at integration as points of departure. Similar but less involved integrative efforts are also made in chapter 3 in regard to the research methodologies used in this study, and in chapter 6 at the level of empirical results. Both of the latter two attempts, howeverâas well as much of the rest of the bookâare predicated on the concepts established first in this chapter.
DISCOURSE AND DISCOURSE ANALYSIS
Having defined discourse broadly as language-in-the-world, and discourse analysis as its study, I will now specify these concepts more closely. Given the complex worldliness of discourse, however, discourse analysis will be treated not as a discrete science with precise definitions and a rational internal structure, but as a process of âentering something messy with messy toolsâprobably the only way of doing itâ (Galtung, 1988, cited in Phillipson, 1992, p. 3).
Fairclough (1992) distinguishes among three levels of discourse and its analysis:
Any discursive âeventââŠis seen as being simultaneously a piece of text, an instance of discursive practice, and an instance of social practice. The âtextâ dimension attends to language analysis of texts. The âdiscursive practiceâ dimensionâŠspecifies the nature of the processes of text production and interpretation The âsocial practiceâ dimension attends to issues of concern in social analysis such as the institutional and organizational circumstances of the discursive event and how that shapes the nature of the discursive practice, and the constitutive/constructive effects of discourse. (p. 4)
This conceptualization provides a useful heuristic for moving toward an integrative definition of discourse for analytical purposes. Faircloughâs level of âtextââof discourse as more-or-less pure languageâis what has historically been studied by linguists: Such definitions of discourse as âlanguage above the level of the sentence,â although less popular than formerly, are even now current in linguistically oriented discourse analysis (Pennycook, 1994; Tannen, 1990). Faircloughâs discourse as âdiscursive practice,â on the other hand, is substantially congruent with the traditional concerns of rhetoric, perhaps even much of the ânew rhetoricâ into the 1980s. Bitzerâs (1968) classic analysis of the rhetorical situation, for example, would seem to be located squarely in this tradition, as wouldâthough perhaps less squarelyâMillerâs (1984) concept of genre as social action.
Faircloughâs third categoryâdiscourse as social practiceâis the one that has received special attention across academic fields over the past 15 years. Its root formulation is, of course, that of Foucault, although his ideas have been widely modified and adapted. When Jarratt (1994, p. 1) defines rhetoric as âa set of inquiries around the uses of language, the institutional and power structures creating and created by historically specific discourse worlds,â I take her to be describing, in Faircloughâs terms, the analysis of discourse as social practice.
Having used Faircloughâs categorization to bootstrap us into the discussion, let me now complexify it. His categories are obviously mere analytical conveniences; there is no such thing, for example, as âpure languageââall language is language in context, so to speak. Likewise, Faircloughâs intermediate levelâthe production and reception of language/ textsâis predicated on their imminent, immediate, or trace material existence (Faircloughâs âlanguage/textâ level) and their discursive social histories (Faircloughâs âso cial practiceâ level), and nontrivially involves these phenomena at every turn. Similarly, although pioneers like Foucault were less than cognizant of the importance of microlevel linguistic detail in realizing and modifying macrolevel social action and organization, discourse as social practice presupposes dynamic activity at all levels of discourse, as later theorists have shown (e.g., Bazerman, 1994, chap. 12; Berkenkotter & Huckin, 1995, chap. 1; Fairclough, 1992; Gee, 1990; Giddens, 1984; Stubbs, 1996).
A theory of discourse that usefully complements and expands on Faircloughâs is that of Gee.2 For Gee (1990), discourse is always part of a specific form of life, or a âDiscourse,â which he defines as:
a socially accepted association among ways of using language, of thinking, feeling, believing, valuing, and of acting that can be used to identify oneself as a member of a socially meaningful group or âsocial network,â or to signal (that one is playing) a socially meaningful ârole.â (p. 143)3
Geeâs definition is useful in that it clarifies the place of language in social life, wherein language enjoys equal status with a variety of other sociocognitive practices but is in no sense privileged over or to be studied apart from them. Rather, it functions as an organic part of a larger sociocognitive wholeâa Discourse.
Together with those of Fairclough, Geeâs ideas provide an âopportunity spaceâ for the form of discourse analysis I try to develop in this study. Viewing all concrete instances of language as part of one or more larger âsaying-writing-doing-being-valuing-believingâ combinations (Gee, 1990, p. 142)âcombinations that are furthermore always historically constituted (p. 145)âlicenses the use of non-language focused historical resources and knowledge as an integral part of a discourse/analytic framework. Thus, the study of discourse canâindeed mustâbe undertaken by studying social practices (including historical ones) that may have little to do with language per se.
With rare exceptions,4 linguistically oriented discourse analysts have avoided granting any status to historical concerns in their research, perhaps due to Saussureâs foundational separation of synchronic and diachronic perspectives on language, and his subsequent banishment of the latter (D. Atkinson, 1994, 1995).5 Even studies in composition/rhetoric have tended to treat history more as a variably powerful contextual explanans than as an integral part of the object under analysis (Bazerman, 1988, chap. 1, note 1). Geeâs decentering of language from a privileged position, howeverâa move made less explicitly also by Faircloughâallows history in as a more equal partner.
My own version of discourse analysis, described in more detail in chapter 3, attempts to capture activity at all three of the levels described by Fairclough (1992), and to treat them integratively. It does so gradually, howeverâhistorical social practice, rhetoric, and language are first treated more or less independently (in chapters 2, 4, and 5, respectively), and only afterwards do I attempt to bring them together (in chapter 6). This is partly for analytical convenience, and partly because my understanding of the relationships between these different aspects of discourseâand the very fact that they can be encompassed in a larger integrated conception of discourse/Discourse at allâevolved substantially in the course of writing this book. Such an approach makes for a less-than-homogeneous presentation of materialâand no doubt for some internal inconsistenciesâbut is, I believe, a more accurate record of the thinking that has gone into this volume.6
The four main concepts described in the remainder of this chapterâdiscourse community, discourse conventions, genre, and registerâserve further to undergird the integrative model of discourse analysis I have introduced here.
DISCOURSE COMMUNITY
Geeâs Discourses represent complex combinations of cognitive but intersubjective knowledge and activity with external, socially purposive activity and practice.7 The much-discussed concept of âdiscourse community,â on the other hand, has focused largely on the outward manifestations and consequences of the intersubjective knowledge systems accounted for partly in Discourses.
Since the discourse community concept has been widely discussed and debated (e.g., Bizzell, 1982, 1992; Cooper, 1989; Freed & Broadhead, 1987; J.Harris, 1989; Killingsworth, 1992; C.R.Miller, 1993; J.Porter, 1986; Rafoth, 1988, 1990; Swales, 1990, 1993), I will limit my discussion here to two or three major points focusing on the single most widely cited definition of âdiscourse communityââthat of Swales (1990, chap. 2). This definition has been criticized for emphasizing the foundational, consensual, and realist nature of discourse communities. In my own view, Swalesâ six definitional criteria are useful as identifying features of mature discourse communities, but they have an after-the-fact status as mostly the public results or consequences of the strong social and communal motivations and mechanisms that lead, in the first place, to the constitution and maintenance of such groups.
A first point to make regarding Swalesâ definition is that the weakness mentioned just above is partly addressed by Geeâs notion of Discourses, because, fully developed, the latter foregrounds the underlying motivations for and socialization/apprenticeship practices by which people are actually inducted into sociocultural groups. An additional element of Geeâs notion not mentioned so farâthat humans are normally participants in multiple such Discourses, some of which will be mutually inconsistent with or opposed to one another in terms of their inherent systems of values and practicesâalso suggests how dissonance, fragmentation, heterogeneity, and change can be natural attributes of discourse communities, and the individuals who make them up.8
Second, Swales (1993) himself has backed away from some of the more realist implications of his earlier definition, preferring what he calls a âvirtual existenceâ for discourse communities of the sort theorized by Giddens (e.g., 1979; cf. Berkenkotter & Huckin, 1995, chap. 1; Stubbs, 1996, chap. 3) for social structures and systems in general. As is widely recognized, Giddensâ structuration theory is important here because it suggests a mechanism by which ongoing social change can operate in conjunction with relatively stable (or metastable) social systems and structures. Bakhtinâs (1981) âcentripedal versus centrifugalâ tendencies in language use and development, Foucaultâs (1972) juxtaposition of âinclination versus institutionâ in discourse, and Bourdieuâs (1977) social fields versus individual habitus in social life are other dialectical conceptualizations of the relationship between microlevel social practices and macrolevel social structures.
An alternative way of conceptualizing expert communitiesâas âcommunities of practiceâ with their multiple, historically evolved âactivity systemsââhas been proposed by students of Vygotskian psychology and situated cognition (e.g., Brown, Collins, & Duguid, 1989; Chaiklin & Lave, 1993; Lave & Wenger, 1991; Rogoff, 1990; cf. Berkenkotter & Huckin, 1995, chap. 1). Such concepts would seem to capture some of the multiplex variety of interwoven human actions and activities that characterize disciplinary networks. That is, rather than seeing such communities as necessarily organized around texts per se, we can profitably view texts as part of larger, integrated systems around which sociodisciplinary activity is organized and through which it is accomplished. As with Geeâs notion of âDiscourse(s),â language/text is thereby decentered and thus able to be viewed in a holistic ecological relationship with other social practices.9
At a different conceptual level, the pragmatic notion of discourse conventions also helps to explain the dynamics underlying the establishment, survival, and change inherent in discourse communities.
DISCOURSE CONVENTIONS
In Lewis (1969), a convention is defined as an institutionalized solution to a recurring coordination problem. A coordination problem, in turn, is a class of interpersonal situations in which a mutually beneficial activity, in order to be performed, demands the coordinated efforts of those involved. The idea here, then, is that where social action of any type will bring greater benefit to individuals than they could achieve acting separately, and where conditions for such action recur, this coordinated action will tend to become regularized, or conventionalized. Such conventions can be found in virtually all domains of human social activityâwritten discourse conventions are simply conventionalized solutions to recurring coordination problems that are addressed through written communication (D. Atkinson, 1991).
Written discourse conventions, from this perspective, exist for the primary purpose of ensuring smooth and efficent communication among discourse community members. Although linguists commonly view all language as conventional in this sense, some of the most concrete examples of conventions beyond the basic units of language are certain collocationally restricted grammatical features, formulaic phrases, and technical vocabulary that are highly discourse-community specific. Legal formulas, for example, such as in good faith or by or on behalf of, or technical medical expressions like human blood group B and acute myocardial infarction, can be seen as providing the tools necessary to communicate about events either particular to, or viewed from the particular standpoints of, the communities that use them (Goodwin, 1994). At a more abstract level, principles of textual organization such as the conventionalized three-part ârhetorical moveâ sequence of experimental research article introductions (Swales, 1981, 1990)âor the highly conventionalized Introduction-Methods-Results-Discussion (IMRD) experimental research-report format (e.g., Bazerman, 1985)âhave been shown to provide templates by means of which both scientific reading and writing can be performed with maximal efficiency.
At the most abstract level of discourse conventions are what Bizzell (1982) has called âconventions for construing reality.â These are basically the socially normative folk theories that Gee (1990, 1992) has characterized as partly making up Discourses, and which likewise in part constitute the communal knowledge bases of discourse communities. For example, a dominant shared assumption in the scientific community is that of âscientific objectivityâ (e.g., Daston, 1992); that is, that by following a certain prescribed set of methods in the study of nature, enough analytical rigor can be attained to effectively neutralize fallible human judgments. Although such guiding conventions are substantially nonlinguistic, they have significant rhetoricallinguistic parallels; thus, such features as the âscientific passiveâ (Ding, in press; Halliday & Martin, 1993, chap. 3), and heavily sanctioned restrictions like those on lexical self-reference (i.e., the use of âIâ or âweâ), are powerful linguistic indexes of the convention of scientific objectivity.
But perhaps the most significant aspect of discourse conventions is that they âlook both ways,â or are multifunctional. That is, although they are often directly represented in language and rhetoric, discourse conventions serve larger social and cognitive functions as well. Cognitively, they represent the input to and output of the schematic patterns or mental models that are believed to organize thought and memory (e.g., Adams & Collins, 1979; Holland & Quinn, 1987). Socially, besides serving as tools for efficient ingroup co...