TD   Introduction
ways of thinking or feeling, ways of manipulating objects or tools, ways of using non-linguistic symbol systems, etc.), we can dispense with this device. It will just clutter up the text and the point is now made. Throughout this book I will simply use the phrase "discourse analysis," but will mean by this phrase analyses that deal with both "little d" discourse and "big D" Discourse.
2    Discourses and social languages
2.1 Building things through language
Language has a magical property: when we speak or write we craft what we have to say to fit the situation or context in which we are communicating. But, at the same time, how we speak or write creates that very situation or context. It seems, then, that we fit our language to a situation or context that our language, in tum, helped to create in the first place.
This is rather like the "chicken and egg" question: Which comes first? The situation we're in (e.g. a committee meeting)? Or the language we use (our committee ways of talking and interacting)? Is this a "committee meeting" because we are speaking and acting this way, or are we speaking and acting this way because this is a committee meeting? After all, if we did not speak and act in certain ways, committees could not exist; but then, if institutions, committees, and committee meetings didn't already exist, speaking and acting this way would be nonsense. The answer here is that this magical property is real and language and institutions "boot ,A strap" each other into existence in a reciprocal process through time.
Another way to look at the matter is this: we always actively use spoken and written language to create or build the world of activities (e.g. committee meetings) and institutions (committees) around us. However, thanks to the workings of history and culture, we often do this in more or less routine ways. These routines make activities and institutions, like committees and committee meetings, seem to (and, in that sense, actually) exist apart from language and action in the here and now. None the less, these activities and institutions have to be continuously and actively rebuilt in the here and now. This is what accounts for change, transformation, and the power of language-in-action in the world.
We continually and actively build and rebuild our worlds not just through language, but through language used in tandem with actions, interactions, non-linguistic symbol systems, objects, tools, technologies, and distinctive ways of thinking, valuing, feeling, and believing. Sometimes what we build is quite similar to what we have built before; sometimes it is not. But language-in-action is always and everywhere an active building process.
12    Discourses and social languages
Whenever we speak or write, we always and simultaneously construct or build six things or six areas of "reality":
1      Tlie meaning and value of aspects of the material world: I enter a plain, square room, and speak and act in a certain way (e.g. like someone about to run a meeting), and, low and behold, where I sit becomes the "front" of the room.
2     Activities: We talk and act in one way and we are engaged in formally opening a committee meeting; we talk and act in another way and we are engaged in "chit-chat" before the official start of the meeting.
3     Identities and relationships: I talk and act in one way one moment and I am speaking and acting as "chair" of the committee; the next moment I speak and talk in a different way and I am speaking and acting as one peer/colleague speaking to another.
4     Politics (the distribution of social goods): I talk and act in such a way that a visibly angry male in a committee meeting (perhaps it's me!) is "standing his ground on principle," but a visibly angry female is "hysterical."
5      Connections: I talk and act so as to make what I am saying here and now in this committee meeting about whether we should admit more minority students connected to or relevant to (or, on the other hand, not connected to or relevant to) what I said last week about my fears of losing my job given the new government's rum to the right.
6     Semiotics (what and how different symbol systems and different forms of knowledge "count"): I talk and act so as to make the knowledge and language of lawyers relevant (privileged), or not, over "everyday language" or over "non-lawyerly academic language" in our committee discussion of facilitating the admission of more minority students.
In Chapter 5 I will elaborate these "building tasks" and their relevance for discourse analysis. But in the next three chapters, I want to develop several "tools of inquiry" (ways of looking at the world of talk and interaction) that will help us study how these building tasks are carried out and with what social and political consequences. The tools of inquiry I will introduce in this chapter are primarily relevant to how we (together with others) build identities and activities and recognize the identities and activities that are being built around us. However, the tools of inquiry introduced here are most certainly caught up with all the other buüding tasks above, as well, as we will see progressively in this book. The tools to be discussed in this chapter are:
(a)   "Situated identities," that is, different identities or social positions we enact and recognize in different settings.
(b)   "Social languages," that is, different styles of language that we use to enact and recognize different identities in different settings; different
Discourses and social languages    13
social languages also allow us to engage in all the other building tasks above (in different ways, building different sorts of things).
(c)   "Discourses" with a capital "D," that is, different ways in which we humans integrate language with non-language "stuff," such as different ways of thinking, acting, interacting, valuing, feeling, believing, and using symbols, tools, and objects in the right places and at the right times so as to enact and recognize different identities and activities, give the material world certain meanings, distribute social goods in a certain way, make certain sorts of meaningful connections in our experience, and privilege certain symbol systems and ways of knowing over others (i.e. carry out all the building tasks above).
(d)   "Conversations" with a capital "C," that is, long-running and important themes or motifs that have been the focus of a variety of different texts and interactions (in different social languages and Discourses) through a significant stretch of time and across an array of institutions.
2.2 Wlios and zvhats
When you speak or write anything, you use the resources of English to project yourself as a certain kind of person, a different kind in different circumstances. You also project yourself as engaged in a certain kind of activity, a different kind in different circumstances. If I have no idea who you are and what you are doing, then I cannot make sense of what you have said, written, or done.
You project a different identity at a formal dinner party than you do at the family dinner table. And, though these are both dinner, they are none the less different activities. The fact that people have differential access to different identities and activities, connected to different sorts of status and social goods, is a root source of inequality in society. Intervening in such matters can be a contribution to social justice. Since different identities and activities are enacted in and through language, the study of language is integrally connected to matters of equity and justice.
An oral or written "utterance" has meaning, then, only if and when it communicates a who and a what (Wieder and Pratt 1990a). What I mean by a "who" is a socially-situated identity, the "kind of person" one is seeking to be and enact here and now. What I mean by a "what" is a socially-situated activity that the utterance helps to constitute.
Lots of interesting complications can set in when we think about identity enacted in and through language. Whos can be multiple and they need not always be people. The President's Press Secretary can issue an utterance that is, in fact, authored by a speech writer and authorized (and even claimed) by the President. In this case, the utterance communicates a sort of overlapping and compound who. The Press Secretary, even if she is directly quoting the speech writer, must inflect the remark with her own
voice. In turn, the speech writer is both "mimicking" the President's "voice" and creating an identity for him.
Not just individuals, but also institutions, through the "anonymous" texts and products they circulate, can author or issue "utterances." For example, we will see below that the warning on an aspirin bottle actually communicates multiple whos. An utterance can be authored, authorized by, or issued by a group or a single individual.
Finally, we can point out that whos and whats are not really discrete and separable. You are who you are partly through what you are doing and what you are doing is partly recognized for what it is by who is doing it. So it is better, in fact, to say that utterances communicate an integrated, though often multiple or "heteroglossic," who-doing-what.
2.3 "Real Indians"
Though I have focused on language, it is important to see that making visible and recognizable who we are and what we are doing always requires more than language. It requires, as well, that we act, think, value, and interact in ways that together with language render who we are and what we are doing recognizable to others (and ourselves). In fact, to be a particular who and to pull off a particular what requires that we act, value, interact, and use language in sync with or in coordination with other people and with various objects ("props") in appropriate locations and at appropriate times.
To see this wider notion of language as integrated with "other stuff" (other people, objects, values, times and places), we will briefly consider Wieder and Pratt's (1990a, b) fascinating work on how Native Americans (from a variety of different groups, though no claim is made that the following is true of all Native American groups) recognize each other as "really Indian." Wieder and Pratt point out that real Indians "refer to persons who are 'really Indian' in just those words with regularity and standardization" (1990a: 48). Wieder and Pratt's work will also make clear how the identities (the whos) we take on are flexibly negotiated in actual contexts of practice.
The term "real Indian" is, of course, an "insiders' term." The fact that it is used by some Native Americans in enacting their own identity work does not license non-Native Americans to use the term. Thus, though it may clutter the text, I will below always place the term "real Indian" in scare quotes to make clear that I am talking about the term and not claiming that I have the "right" to actually use it of anyone. In any case, however I might use it, it certainly would do different work than it does for the Native Americans we will discuss below. Finally, let me say that I am not discussing Native Americans here because I think they are "esoteric." In fact, I am using this example, because I think it is a clear and dramatic example of what we all do all the time, though in different ways.
1—'tOl-ĽtH i3E,i3   HUW   OUHHt   J-M-H.XIAI4-XĽÍ?         Xui
The problem of "recognition and being recognized" is very consequential and problematic for Native Americans. While in order to be considered a "real Indian," one must be able to make some claims to kinship with others who are recognized as "real Indians," this by no means settles the matter. People with such (biological) ties can fail to get recognized as a "real Indian," and people of mixed kinship can be so recognized.
Being a "real Indian" is not something one can simply be. Rather, it is something that one becomes in and through the doing of it, that is, in carrying out the actual performance itself. Though one must have certain kinship ties to get in the "game," beyond-this entry criterion, there is no being (once and for all) a "real Indian," rather there is only doing being-or-becoming-a-"real-Indian." If one does not continue to "practice" being a "real Indian," one ceases to be one.
Finally, doing being-and-becoming-a-"real-Indian" is not something that one can do all by oneself. It requires the participation of others. One cannot be a "real Indian" unless one appropriately recognizes "real Indians" and gets recognized by others as a "real Indian" in the practices of doing bemg-and-becoming-a-"real-uidian." Being a "real Indian" also requires appropriate accompanying objects (props), times, and places.
There are a multitude of ways one can do being-and-becoming-a-"real-Indian." Some of these are (following Wieder and Pratt 1990a): "Real Indians" prefer to avoid conversation with strangers, Native American or otherwise. They cannot be related to one another as "mere acquaintances," as some "non-Indians" might put it. So, for "real Indians," any conversation they do have with a stranger who may turn out to be a "real Indian" will, in the discovery of the other's "Indianness," establish substantial obligations between the conversational partners just through the mutual acknowledgment that they are "Indians" and that they are now no longer strangers to one another.
In their search for the other's "real Indianness" and in their display of their own "Indianness," "real Indians" frequently engage in a distinctive form of verbal sparring. By correctly responding to and correctly engaging in this sparring, which "Indians" call "razzing," each participant further establishes cultural competency in the eyes of the other.
"Real Indians" manage face-to-face relations with others in such a way that they appear to be in agreement with them (or, at least, they do not overtly disagree); they are modest and "fit in." They show accord and harmony and are reserved about their own interests, skills, attainments, and positions. "Real Indians" understand that they should not elevate themselves over other "real Indians." And they understand that the complex system of obligations they have to kin and other "real Indians" takes priority over those contractual obligations and pursuit of self-interest that some "non-Indians" prize so highly.
"Real Indians" must be competent in "doing their part" in participating in conversations that begin with the participants exchanging greetings and
16 j Discourses and social languages
other amenities and then lapsing into extended periods of silence. They must know that neither they nor the others have an obligation to speak -that silence on the part of all conversants is permissible.
When they are among "Indians," "real Indians" must also be able to perform in the roles of "student" and "teacher" and be able to recognize the behaviors appropriate to these roles. These roles are brought into play exclusively when the appropriate occasion arises for transmitting cultural knowledge (i.e. tilings pertinent to being a "real Indian"). Although many "non-Indians" find it proper to ask questions of someone who is instructing them, "Indians" regard questions in such a situation as being inattentive, rude, insolent, and so forth. The person who has taken the role of "student" shows attentiveness by avoiding eye contact and by being silent. The teaching situation, then, as a witnessed monologue, lacks the dialogical features that characterize some Western instruction.
While the above sort of information gives us something of the flavor of what sorts of things one must do and say to get recognized as a "real Indian," such information can lead to a bad mistake. It can sound as if the above features are necessary and sufficient criteria for doing being-and-becoming-a-"real-Indiari." But this is not true.
These features are not a test that can be or ever is administered all at once, and once and for all, to determine who is or is not a "real Indian." Rather, the circumstances under which these features are employed by "Indians" emerge over the course of a developing history among groups of people. They are employed always in the context of actual situations, and at different times in the life history of groups of people. The ways in which the judgment, "He (or she) is (or is not) a 'real Indian'," is embedded within situations that motivate it make such judgments intrinsically provisional. Those now recognized can spoil their acceptance or have it spoiled and those not now accepted can have another chance, even when others are reluctant to extend it.
The same tiling applies, in fact, in regard to many other social identities, not just being "a real Indian." There are no once and for all tests for who is a "real" feminist, gang member, patriot, humanist, cutting-edge scientist, "yuppie," or "regular" at the local bar. These matters are settled provisionally and continuously, in practice, as part and parcel of shared histories and on-going activities. When I was young, my community certainly had (very rigid) tests through which we continually, always provisionally, and sometimes contentiously, displayed and recognized who was and was not a "real Catholic" (versus being a "Catholic in name only" or being a non- Catholic). That community, and those tests, have, over the least several decades, changed radically, however much we then viewed them as static and eternal.
Different social identities (different whos) may seriously conflict with one another. For instance, Scollon and Scollon (1981) point out that for the Native Americans they studied (Athabaskans in Canada and the U.S.),
Discourses and social languages    Y7
writing essays, a practice common in school, can constitute a crisis in iden- ' rity. To produce an essay requires the Athabaskan to produce a major self-display, which is appropriate to Athabaskans only when a person is in a position of dominance in relation to the audience (in the case of school, the teacher, not the student).
Furthermore, in essayist prose, the audience and the author are "fictionalized" (not really me and you, but decontextualized and rather generic readers and writers) and the text is decontextualized from specific social networks and relationships. Where the relationship of the communicants is decontextualized and unknown, Athabaskans prefer silence.
The paradox of prose for Athabaskans, the Scollons point out, is that if it is communication between known author and audience it is contextualized and compatible with Athabaskan values, but not good essayist prose. To the extent that it becomes decontextualized and thus good essayist prose, it becomes uncharacteristic of Athabaskans to seek to communicate. What is required to do and be an Athabaskan is in large part mutually exclusive with what it is required to do and be a writer of school-based essayist prose. This doesn't mean Athabaskans cannot do both (remember, we are all multiple), it simply means that they may face very real conflicts in terms of values and identity. And, as the Scollons point out, many other groups of people have similar or related "identity issues" with essayist literacy.
2.4 Discourses (with a big "D")
I want to argue that the problem of "recognition and being recognized" is very consequential, not only for Native Americans, but for all of us all the time. And, as we saw above, making visible and recognizable who we are and what we are doing always involves a great deal more than "just language." It involves actmg-interacting-thinldng-valuing-talking-(some-times writing-reading) in the "appropriate way" with the "appropriate" props at the "appropriate" times in the "appropriate" places.
Such socially accepted associations among ways of using language, of thinking, valuing, acting, and interacting, in the "right" places and at the "right" times with the "right" objects (associations that can be used to identify oneself as a member of a socially meaningful group or "social network"), I will refer to as "Discourses," with a capital "D" (Gee 1990b, 1992, 1996; see also Bourdieu 1990b; Foucault 1985). I will reserve the word "discourse," with a little "á," to mean language-in-use or stretches of language (like conversations or stories). "BigD" Discourses are always language plus "other stuff." There are innumerable Discourses in any modem, technological, urban-based society: for example, (enacting) being something as general as a type of African-American or Anglo-Australian or something as specific as being a type of modern British young second-generation affluent Sikh woman. Being a type of middle-class American, factory worker, or
i8    Discourses and social langiiages
executive, doctor or hospital patient, teacher, administrator, or student, student of physics or of literature, member of a club or street gang, regular at the local bar, or, as we saw earlier, "real Indian" are all Discourses.
The key to Discourses is "recognition." If you put language, action, interaction, values, beliefs, symbols, objects, tools, and places together in such a way that others recognize you as a particular type of who (identity) engaged in a particular type of what (activity) here and now, then you have pulled off a Discourse (and thereby continued it through history, if only for a while longer). Whatever you have done must be similar enough to other performances to be recognizable. However, if it is different enough from what has gone before, but still recognizable, it can simultaneously change and transform Discourses. If it is not recognizable, then you're not "in" the Discourse.
Discourses are always embedded in a medley of social institutions, and often involve various "props" like books and magazines of various sorts, laboratories, classrooms, buildings of various sorts, various technologies, and a myriad of other objects from sewing needles (for sewing circles) through birds (for bird watchers) to basketball courts and basketballs (for basketball players). Think of all the words, symbols, deeds, objects, clothes, and tools you need to coordinate in the right way at the right time and place to "pull off" (or recognize someone as) being a cutting-edge particle physicist or a Los Angeles Latino street gang member or a sensitive lťgh-culture humanist (of old).
It is sometimes helpful to think about social and political issues as if it is not just us humans who are talking and interacting with each other, but rather, the Discourses we represent and enact, and for which we are "carriers." The Discourses we enact existed before each of us came on the scene and most of them will exist long after we have left the scene. Discourses, through our words and deeds, carry on conversations with each other through history, and, in doing so, form human history.
Think, for instance, of the long-running and ever-changing "conversation" in the U.S. and Canada between the Discourses of "being an Indian" and "being an Anglo" or of the different, but equally long-running "conversation" in New Zealand between "being a Maori" and "being an Anglo" (or, for that matter, think of the long-running conversation between "being a British Anglo" and "being an American Anglo"). Think of the long-running and ever-changing "conversation" between creationists and biologists. Think of the long-running and ever-changing "conversation" in Los Angeles between African-American teenage gang members and the L.A. police (some of whom, for instance, are leading experts, even academically speaking, on the "grammar" of gang graffiti, which varies significantly, by the way, between African-American gangs and Latino gangs), mtriguingly, we humans are very often unaware of the history of these conversations, and thus, in a deep sense, not fully aware of what we mean when we act and talk.
Discourses and.social languages    19
When we discussed being a "real Indian," we argued that "knowing how" to be a "real Indian" rests on one's being able to "be in sync with other 'real Indians'" and with objects (e.g. the material items of the culture) in the appropriate times and places. Recent studies of science suggest much the same thing is true for scientists.
For example, these studies argue the physics experimental physicists "know" is, in large part, not in their heads. Rather, it is spread out (distributed), inscribed in (and often trapped in) apparatus, symbolic systems, books, papers, and journals, institutions, habits of bodies, routines of practice, and other people (Latour 1987; Traweek 1988). Each domain of practice, each scientific Discourse - for example, a specific area within physics or biology - attunes actions, expressions, objects, and people (the scientists themselves) so that they become "workable" in relation to each other (Knorr Cetina 1992). They are "in sync."
Just as there were verbal and non-verbal ways to be a "real Indian," there are ways to be a "real experimental physicist." They are both (being an experimental physicist or being a "real Indian") ways with words, feelings, values, beliefs, emotions, people, actions, things, tools, and places that allow us to display and recognize characteristic whos doing characteristic whats. They are both, then, Discourses.
The scientist's "know how" is the ability to coordinate and be coordinated by constellations of expressions, actions, objects, and people. In a sense, the scientist is both an actor (coordinating other people and various things, tools, technologies, and symbol systems) and a patient (being coordinated by other people and various things, tools, technologies, and symbol systems). Scientists become agent-patients "in sync with," "linked with," "in association with," "in coordination with," however we want to put it, other "actants" (adapting a term from Callon and Latour 1992), such as particular forms of language, other people, objects (e.g. scientific equipment, atoms, molecules, or birds), places (e.g. labs or fields), and nonverbal practices.
In the end a Discourse is a "dance" that exists in the abstract as a coordinated partem of words, deeds, values, beliefs, symbols, tools, objects, times, and places and in the here and now as a performance that is recognizable as just such a coordination. Like a dance, the performance here and now is never exactly the same. It all comes down, often, to what the "masters of the dance" will allow to be recognized or will be forced to recognize as a possible instantiation of the dance.
2.5 Discourses are not "units" with clear boundaries
Tire notion of Discourses will be important throughout this book. It is important, therefore, to make some points clear to avoid some common misunderstandings. Imagine I freeze a moment of thought, talk, action, or interaction for you, in the way in which a projector can freeze a piece
20    Discourses and social languages
of film. To make sense of that moment, you have to recognize the identifies and activities involved in it. Perhaps, for this frozen moment you can't do so, so you move the film back and forward enough until you can make such a recognition judgment.
"Oh, now I see," you say, "it's a 'real Indian' razzing another 'real Indian'," or "it's a radical feminist berating a male for a crass male remark" or "it's a laboratory physicist orienting colleagues to a graph" or "it's a first-grader in Ms. X's class starting a sharing time story." Perhaps, if you now move the film backwards and forwards a bit more, you will change your judgments a little, a lot, or not at all.
Perhaps, you aren't sure. You and I even argue about the matter. You say that "It's a skinhead sending mtirnidating glances to a passing adult on the street" and I say, "No, it's just a wanna-be trying to act tough." You say, "It's a modern classroom teacher leading a discussion" and I say, "No, it's a traditional teacher giving a lecture in the guise of a series of known-answer questions."
This is what I call "recognition work." People engage in such work when they try to make visible to others (and to themselves, as well) who they are and what they are doing. People engage in such work when they try to recognize others for who they are and what they are doing. People engage in such work within interactions, moment by moment. They engage in such work when they reflect on their interactions later. They engage in such work, as well, when they try to understand human interaction as researchers, practitioners, theoreticians, or interventionists.
Sometimes such recognition work is conscious, sometimes it is not. Sometimes people have labels they can articulate for the mhos and whats they recognize, sometimes they don't. Sometimes they fight over the labels, sometimes they don't. And the labels change over time.
Thanks to the fact that we humans engage, inside and outside interactions, in recognition work, Discourses exist in the world. For example, there is a way of being a kindergarten student in Ms. X's class with its associated activities and ways with words, deeds, and tilings. Ms. X, her students, her classroom, with its objects and artifacts, and characteristic activities, are all in the Discourse she and her students create. These same people and things, of course, can be in other Discourses as well.
Recognition work and Discourses out in the world go hand-in-hand. Ms. X and her students engage in recognition work, for example, a certain sort of sharing time story isn't recognized as "acceptable" in this class, another type is. That recognition work creates a Discourse, that is, ways with words, actions, beliefs, emotions, values, interactions, people, objects, tools, and technologies that come to constitute "being and doing a student in Ms. X's class." In turn, this Discourse renders recognition work possible and meaningful. It's another "chicken and egg" question, then: Which comes first, recognition work or Discourses? Neither. They are reflexively related, such that each creates the other.
Discourses and social languages    21
Discourses have no discrete boundaries because people are always, in history, creating new Discourses, changing old ones, and contesting and pushing the boundaries of Discourses. You, an African-American male, speak and act here and now in an attempt to get recognized as a "new capitalist manager coaching a project team." If you get recognized as such, then your performance is in the Discourse of new capitalist management. If you don't, it isn't.
If your performance has been influenced, intentionally or not, by another one of your Discourses (say, your membership in the Discourse of doing and being a jazz fan or your membership in a certain version of African-American culture as a Discourse), and it gets recognized in the new capitalist management Discourse, then you have just, at least for here and now, "infected" one Discourse with another and widened what "counts" in the new capitalist management Discourse. You pushed the boundaries. In another time and place they may get narrowed.
You can get several of your Discourses recognized all at once. You (thinking of one of my esteemed colleagues at a university where I previously worked) "pull off" being here and now, in a class or meeting, for example, "a British, twice-migrant, globally oriented, traditional and modern, fashionable, female, Sikh, American Professor of cultural studies and feminist postmodern anthropology" by weaving strands of your multiple Discourses together. If this sort of thing gets enacted and recognized enough, by enough people, then it will become not multiple strands of multiple Discourses interleaved, but a single Discourse whose hybridity may ultimately be forgotten. The point is not how we "count" Discourses; the point is the performance, negotiation, and recognition work that goes into creating, sustaining, and transforming them, and the role of language (always with other things) in this process.
Let me make several other brief, but important points about Discourses:
1     Discourses can split into two or more Discourses. For example, medieval "natural philosophy" eventually split into philosophy, physics and other sciences.
2     Two or more Discourses can meld together. For example, after the movie Colors came out some years ago, mixed Latino, African-American, and white gangs emerged. Prior to that, Latinos, African-Americans, and whites had quite separate ways of being and doing gangs, as they still do in the case of segregated gangs.
3     It can be problematic whether a Discourse today is or is not the same as ■ one in the past. For example, modern medicine bears little resemblance
to medicine before the nineteenth century, but perhaps enough to draw some important parallels for some purposes, though not for others.
4     New Discourses emerge and old ones die all the time. For example, in Palmdale, California (a desert community outside Los Angeles), and I assume other places as well, an anti-racist skinhead Discourse is
dying because people, including the police, tend to confuse its members with a quite separate, but similar looking, racist Neo-Nazi skinhead Discourse.
5     Discourses are always defined in relationships of complicity and contestation with other Discourses, and so they change when other Discourses in a society emerge or die. For example, the emergence of a "new male" Discourse in the 1970s (ways of doing and being a "new male") happened in response to various gender-based Discourses (e.g: various sorts of feminism) and class-based Discourses (the baby-boom middle class was too big for all young males to stay in it, so those who "made it" needed to mark their difference from those who did not), and, in turn, changed the meanings and actions of these other Discourses.
6     Discourses need, by no means, be "grand" or large scale. I used to eat regularly at a restaurant with a long bar. Among the regulars, there were two different Discourses at opposite ends of the bar, that is, ways, of being and doing that end of the bar. One involved young men and women and a lot of male-dominated sexual bantering; the other involved older people and lots of hard luck stories. The restaurant assigned different bartenders to each end (always a young female at the young end) and many of the bartenders could fully articulate the Discourse at their end of the bar and their role in it.
7     Discourses can be hybrids of other Discourses. For example, the school yards of many urban middle and high schools are places where teenagers of different ethnic groups come together and engage in what I have elsewhere called a "borderland Discourse" of doing and being urban teenager peers (Gee 1996), when they cannot safely go into each other's neighborhoods and when they each have their own neighborhood peer-based Discourses. The borderland Discourse is quite manifestly a mixture of the various neighborhood peer Discourses, with some emergent properties of its own.
8     There are limitless Discourses and no way to count them, both because new ones, even quite non-grand ones, can always emerge and because boundaries are always contestable.
One way to think about the role of Discourses is this: Imagine you have a giant map. Each Discourse is represented on the map like a country, but with movable boundaries that you can slide around a bit. You place the map on top of any language, action, or interaction you participate in or want to think about. You move the boundaries of the Discourse areas on the map around in negotiation with others or as your reflections change.
The map gives you a way to understand what you are seeing in relationship to the full set of Discourses in an institution (maybe it is just a map of all the Discourses in a given community, business, school, or university) or the society as a whole (if it's a map of the whole society), at least as far as
Discourses and social languages    23
you know it. Wherever on the map you line up the current thought, action, interaction, or language, it is immediately placed in relation to all the other countries (Discourses) on the map (though "fuzzily," since you can move the boundaries around or others can try to make you do so).
Such a map is a Discourse grid against which you understand your own and others' thought, language, action, and interaction. It is an ever changing map with which you can engage in recognition work. It is, as it exists across people and social groups, both the origin and the product of the reality of actual Discourses in the world, aligning and disaligning themselves with each other through history.
Understanding is always relative to the whole grid ör map. The complex relationships among Discourses, which we can imagine as intricate crisscrossing lines connecting the various Discourse-areas on the map in complex positive and negative ways, define and demarcate individual Discourses. Your own Discourse grid is the limit of your understanding, and it is the fundamental job of education to give people bigger and better Discourse maps, ones that reflect the working of Discourses throughout society, the world, and history in relationship to each other and to the learner.
So Discourses are out in the world and history as coordinations ("a dance") of people, places, times, actions, interactions, verbal and nonverbal expression, symbols, things, tools, and technologies that betoken certain identities and associated activities. Thus, they are material realities. But Discourses also exist as the work we do to get people and things recognized in certain ways and not others, and they exist as maps that constitute our understandings. They are, then, social practices and mental entities, as well as material realities.
2.6 A heteroglossic aspirin bottle
I want now to return to how whos and whats are communicated in language (keeping in mind that language alone is rarely enough and is always put together with "other stuff" to pull off a Discourse). It is time, then, to turn to examples in order to make my points about whos-doing-whats more concrete. Consider, then, the warning on my aspirin bottle (Gee 1996), reprinted below (italics and capitals are on the warning):
Warnings: Children and teenagers should not use this medication for chicken pox or flu symptoms before a doctor is consulted about Reye Syndrome, a rare but serious illness reported to be associated with aspirin. Keep this and all drugs out of the reach of children. In case of accidental overdose, seek professional assistance or contact a poison control center immediately. As with any drug, if you are pregnant or nursing a baby, seek the advice of a health professional before using this product, rr is especially
IMPORTANT NOT TO USE ASPIRIN DURING THE LAST 3 MONTHS OF PREGNANCY
24    Discourses and social languages
UNLESS SPECIFICALLY DIRECTED TO DO SO BY A DOCTOR BECAUSE TT MAY CAUSE PROBLEMS TN THE UNBORN CHILD OR COMPLICATIONS DURING DELIVERY. See
carton for arthritis use+ and Important Notice.
My interpretation of this text is that there are two who-doing-whats in this warning, and they are interleaved. The first is made up of the following sentences:
Children and teenagers should not use this medication for chicken pox or flu symptoms before a doctor is consulted about Reye Syndrome, a rare but serious illness reported to be associated with aspirin. It is especially important not to use aspirin during the last 3 months of pregnancy unless specifically directed to do so by a doctor because it may cause problems in the unborn child or complications during delivery.
Here things are referred to quite specifically ("children or teenagers," "this medication," "chicken pox," "flu," "Reye Syndrome," "aspirin," "last 3 months," "unborn child," "delivery")/ doctors are called "doctor," and matters are treated emphatically (italics, capitals, "should not," "rare but serious," "especially important," "specifically directed").
The second who-doing-what is made up of the following sentences, placed in the middle of the other two:
Keep this and all drugs out of the reach of children. In case of accidental overdose, seek professional assistance or contact a poison control center immediately. As with any drug, if you are pregnant or nursing a baby, seek the advice of a health professional before using this product.
Here things are referred to more generally and genericaUy ("this and all drugs," "any drug," and "this product," rather than "this medication" and "aspirin"; "children" rather than "children and teenagers," "pregnant" rather than "last 3 months of pregnancy"), doctors are not mentioned, rather the health profession is referred to more generally ("professional assistance," "poison control center," "health professional"), and matters are treated less stridently with the exception of that "immediately" (small print, "keep out of reach," "accidental overdose," "seek . . . assistance," "seek advice," rather than "should not" and "important not to use").
These two who-doing-whats "feel" different. The are authorized and issued by different "voices" to different purposes and effects. The first speaks with a lawyerly voice responding to specific court cases; the second speaks with the official voice of a caring, but authoritatively knowledgeable company trying to avoid anyone thinking that aspirin in particular is a potentially harmful drug. Of course, this second who-doing-what
Discourses and social languages    25
partly contradicts the first. By the way, the second who-doing-what on the aspirin bottle used to be the only warning on the bottle (with the order of the sentences a bit different).
This warning, like all utterances, reflects the company it has kept, or, to put the matter another way, it reflects a history that has given rise to it. In this case, presumably, the new sterner, more direct who-doing-what was added to the more general and avuncular one because the company got sued over things like Reye Syndrome.
The warning on the aspirin bottle is heteroglossic. That is, it is "double-voiced," since it interleaves two different whos-doing-whats together. Of course, in different cases, this sort of interleaving could be much more intricate, with the two (or more) whos-doing-whats more fully integrated, and harder to tease apart.
2.7 Social languages
There is another term that it is useful in place of the cumbersome phrase "who-doing-what," at least as far as the language aspects of "who-doing-whats" are concerned (remembering that language is caught up with "other stuff" in Discourses). This term is "social language" (Gee 1996: ch. 4; Bakhtin 1986). Each of the who-doing-whats we saw on the aspirin bottle is linguistically expressed in different "social languages." All languages, like English or French, are composed of many (a great many) different social languages. Social languages are what we learn and what we speak.
Keep in mind that "social languages" and "Discourses" are terms for different things. I will use the term "social languages" to talk about the role of language in Discourses. But as I said above, Discourses always involve more than language. They always involve coordinating language with ways of acting, interacting, valuing, believing, feeling, and with bodies, clothes, non-linguistic symbols, objects, tools, technologies, times, and places.
Let me give a couple of examples of social languages at work, beyond the example of the two different social languages in the warning on the aspirin bottle, examples I have used over the years as particularly clear instances of different social languages (e.g. Gee 1996). Consider, for instance, the following case of an upper-middle-class, Anglo-American young woman named "Jane," in her twenties, who was attending one of my courses on language and communication. The course was discussing different social languages and, during the discussion, Jane claimed that she herself did not use different social languages in different contexts, but rather, was consistent from context to context. In fact, to do otherwise, she said, would be "hypocritical," a failure to "be oneself."
In order to support her claim that she did not switch her style of speaking in different contexts and for different conversational partners, Jane decided
26   Discourses and social languages
to record herself talking to her parents and to her boyfriend. In both cases, she decided to discuss a story the class had discussed earlier, so as to be sure that, in both contexts, she was talking about the same thing.
In the story, a character named Abigail wants to get across a river to see her true love, Gregory. A river boat captain (Roger) says he will take her only if she consents to sleep with him. In desperation to see Gregory, Abigail agrees to do so. But when she arrives and tells Gregory what she has done, he disowns her and sends her away. There is more to the story, but this is enough for our purposes here. Students in my class had been asked to rank order the characters in the story from the most offensive to the least.
In explaining to her parents why she thought Gregory was the worst (least moral) character in the story, the young woman said the following:
Well, when I thought about it, I don't know, it seemed to me that Gregory should be the most offensive. He showed no understanding for Abigail, when she told him what she was forced to do. He was callous. He was hypocritical, in the sense that he professed to love her, then acted like that.
Earlier, in her discussion with her boyfriend, in an informal setting, she had also explained why she thought Gregory was the worst character. In this context she said:
What an ass that guy was, you know, her boyfriend. I should hope, if I ever did that to see you, you would shoot the guy. He uses her and he says he loves her. Roger never lies, you know what I mean?
It was clear - even to Jane - that she had used two very different forms of language. The differences between Jane's two social languages are everywhere apparent in the two texts.
To her parents, she carefully hedges her claims ("I don't know," "it seemed to me"); to her boyfriend, she makes her claims straight out. To her boyfriend, she uses terms like "ass" and "guy," while to her parents she uses more formal terms like "offensive," "understanding," "callous," "hypocritical" and "professed." She also uses more formal sentence structure to her parents ("it seemed to me that . . . ," "He showed no understanding for Abigail, when . . . ," "He was hypocritical in the sense that...") than she does to her boyfriend (".. . that guy, you know, her boyfriend," "Roger never lies, you know what I mean?"). ■
Jane repeatedly addresses her boyfriend as "you," thereby noting his social involvement as a listener, but does not directly address her parents in this way. In talking to her boyfriend, she leaves several points to be inferred, points that she spells out more explicitly to her parents (e.g. her boyfriend must infer that Gregory is being accused of being a hypocrite
Discourses and social languages   27
from the information that though Roger is bad, at least he does not lie, which Gregory did in claiming to love Abigail).
All in all, Jane appears to use more "school-like" language to her parents. Her language to them requires less inferencing on their part and distances them as listeners from social and emotional involvement with what she is saying, while stressing, perhaps, their cognitive involvement and their judgment of her and her "intelligence." Her language to her boyfriend, on the other hand, stresses social and affective involvement, solidarity, and co-participation in meaning making.
This young woman is making visible and recognizable two different versions of who she is and what she is doing. In one case she is "a dutiful and intelligent daughter having dinner with her proud parents" and in the other case she is "a girl friend being intimate with her boyfriend." Of course, I should add, that while people like Jane may talk at dinner this way to their parents, not all people do; there are other identities one can take on for one's parents, other social languages one can speak to them. And, indeed, there may well be others that Jane would use to her parents in different settings.
Let me give one more example of social languages at work, an example taken from Greg Myers' work (1990). Biologists, and other scientists, write differently in professional journals than they do in popular science magazines. These two different ways of writing do different things and display different identities. The popular science article is not merely a "translation" or "simplification" of the professional article.
To see this, consider the two extracts below, the first from a professional journal, the second from a popular science magazine, both written by the same biologist on the same topic (Myers 1990: 150):
Experiments show that Heliconius butterflies are less likely to ovipost on host plants that possess eggs or egg-like structures. These egg-mimics are an unambiguous example of a plant trait evolved in response to a host-restricted group of insect herbivores.
(Professional journal)
Heliconins butterflies lay their eggs on Passiflora vines. In defense the vines seem to have evolved fake eggs that make it look to the butterflies as if eggs have already been laid on them.
(Popular science)
The first extract, from a professional scientific journal, is about the conceptual structure of a specific theory within the scientific discipline of biology. The subject of the initial sentence is "experiments," a methodological tool in natural science. The subject of the next sentence is "these egg-mimics": note how plant-parts are named, not in tenns of the plant itself, but in terms of the role they play in a particular theory of natural selection and
28    Discourses and social languages
evolution, namely "coevolution" of predator and prey (that is, the theory that predator and prey evolve together by shaping each other). Note also, in this regard, the earlier "host plants" in the preceding sentence, rather than the "vines" of the popular passage.
In the second sentence, the butterflies are referred to as "a host-restricted group of insect herbivores," which points simultaneously to an aspect of scientific methodology (like "experiments" did) and to the logic of a theory (like "egg-mimics" did). Any scientist arguing for the theory of co-evolution faces the difficulty of demonstrating a causal connection between a particular plant characteristic and a particular predator when most plants have so many different sorts of animals attacking them. A central methodological technique to overcome this problem is to study plant groups (like Passiflora vines) that are preyed on by only one or a few predators (in this case, Heliconius butterflies). "Host-restricted group of insect herbivores," then, refers to both the relationship between plant and insect that is at the heart of the theory of coevolution and to the methodological technique of picking plants and insects that are restricted to each other so as to "control" for other sorts of interactions.
The first passage, then, is concerned with scientific methodology and a particular theoretical perspective on evolution. On the other hand, the second extract, from a popular science magazine, is not about methodology and theory, but about animals in nature. The butterflies are the subject of the first sentence and the vine is the subject of the second. Further, the butterflies and the vine are labeled as such, not in terms of their role in a particular theory.
The second passage is a story about the straggles of insects and plants that are transparently open to the trained gaze of the scientist. Furthermore, the plant and insect become "intentional" actors in the drama: the plants act in their own "defense" and things "look" a certain way to the insects, they are "deceived" by appearances as humans sometimes are.
These two examples replicate in the present what, in fact, is an historical difference. In the history of biology, the scientist's relationship with nature gradually changed from telling stories about direct observations of nature to carrying out complex experiments to test complex theories (Bazerman 1989). Myers (1990) argues that professional science is now concerned with the expert "management of uncertainty and complexity" and popular science with the general assurance that the world is knowable by and directly accessible to experts.
The need to "manage uncertainty" was created, in part, by the fact that mounting "observations" of nature led scientists not to consensus, but to growing disagreement as to how to describe and explain such observations (Shapin and Schaffer 1985). This problem led, in turn, to the need to convince the public that such uncertainty did not damage the scientist's claim to professional expertise or the ultimate "knowability" of the world.
Discourses and social languages    29
This example lets us see, then, not just that ways with words are connected to different whos (here the experimenter/theoretician versus the careful observer of nature) and whats (the professional contribution to science and the popularization of it), but that they are always acquired within and licensed by specific social and historically shaped practices representing the values and interests of distinctive groups of people.
So, it is clear now, I hope, that in using language what is at stake are whos-doing-whats. But, you cannot be any old who you want to. You cannot engage in any old what you want to. That is to say that whos and whats are creations in history and change in history, as we have just seen in the examples from biology.
2.8 Two grammars
Each social language has its own distinctive grammar. However, two different sorts of grammars are important to social languages, only one of which we ever think to study formally in school. One grammar is the traditional set of units like nouns, verbs, inflections, phrases and clauses. These are real enough, though quite inadequately described in traditional school grammars. Let's call this "grammar one."
The other - less studied, but more important — grammar is the "rules" by which grammatical units like nouns and verbs, phrases and clauses, are used to create patterns which signal or "index" characteristic whos-doing-whats-within-Discourses. That is, we speakers and writers design our oral or written utterances to have patterns in them in virtue of which interpreters can attribute situated identities and specific activities to us and our utterances. We will call this "grammar two."
These patterns, I hasten to add, are not fancy devices of postmodern social science. They have been named in linguistics for a long time. Linguists call them "collocational patterns." This means that various sorts of grammatical devices "co-locate" with each other. The patterns I am trying to name here are "co-relations" (correlations) among many grammatical devices, from different "levels" of grammar one. These correlations, in turn, also co-relate to (coordinate with) other non-language "stuff" to constitute (for historical, i.e. conventional reasons) whos-doing-whats-within-Discourses.
For example, in Jane's utterance to her boyfriend, "What an ass that guy was, you know, her boyfriend," note how informal terms like "ass" and "guy," the vague reference "that guy," the informal parenthetical device "you know," and the informal syntactic device of "right dislocation" (i.e. letting the phrase "her boyfriend" hang out at the end of the sentence) all pattern together to signal that this utterance is in an informal social language used to achieve solidarity.
30    Discouises and social languages
The situation here is much like choosing clothes that go together in such a way that they communicate that we are engaged in a certain activity or are taking up a certain style connected to such activities. For example, consider how thongs, bathing suit, tank top, shades, and sun hat "co-locate" together to "signal" to us things like outdoor and water activities and the situated identities we take up in such situations.
2.9 Grammar and conversations
Let me give you another example of grammar one being used to create grammar two, that is, to create co-locational patterns in virtue of which we recognize a specific social language and its concomitant social identities and activities. Consider the sentence below (adapted from Halliday and Martin 1993: 77):
1   Lung cancer death rates are clearly associated with an increase in smoking.
A whole bevy of linguistic features mark this sentence as part of a distinctive academic social language (though without more connected text we can't actually tell exactly which one). Some of these are: a heavy subject ("lung cancer death rates"), deverbal nouns ("increase," "smoking";), a complex compound noun ("lung cancer death rates"), a "low transitive" relational predicate ("are associated with"), passive or passive-like voice ("are associated"), the absence of agency (no mention of who does the associating), an abstract noun ("rates"), and an assertive modifier to the verb ("clearly").
No single grammatical feature marks the social language of this sentence. Rather, all these features (and a great many more if we took a larger stretch of text, including many discourse-level features) form a distinctive configuration (a correlation or, better, co-relation) that marks the social language. This co-relational (co-locational) pattern is part of the grammar of this social language (in the sense of "grammar two").
I hasten to point out that the configuration of features that mark a social language are too complex and too situated in the specific context they are helping to create (after all, there is no such thing as a "general social science context") to be open to much generalized and rote learning. Linguistic relationships like these do not exist, and are not learned, outside the distinctive social practices (whats) of which they are an integral part. They are part and parcel of the very "voice" or "identity" (lohos) of people who speak and write and think and act and value and live that way (e.g. as a social scientist) for a given time and place. To learn such relationships is part of what it means to leam to recognize the very social context one is in (and helping to create). This is not to say there is no role here for overt instruction (there is). It is only to say that there is no way we can
Discourses and social languages   31
leave out immersion in situated practices if we want to teach people new social languages.
It is sometimes said that what distinguishes "informal" social languages like the one Jane used to her boyfriend from more "formal" ones characteristic of literacy and "literate talk," like the social language Jane used to her parents, or the smoking example on p. 30, is that, in the "infonnal" case, "context" determines meaning and you just have to have been there to understand what was being said. In the more "formal" cases, it is held that the words and sentences mean in a more explicit, less contextual way. In fact, it is sometimes said that such language is "decontextualized." Some people in education claim that what many minority and lower socioeconomic children who do not succeed in school fail to know is how to use such "decontextualized language."
AU this is seriously in error, and in ways that not only mislead us, but actually damage some people (e.g. the children just referred to). Consider sentence 1 again. This sentence is no more explicit than informal language. It is no less contextualized. It is simply inexplicit and contextualized in a different way.
Though we tend to think of writing, at least academic writing, as clear, unambiguous, and explicit in comparison to speech, sentence 1, in fact, has at least 112 different meanings! What is odder still is that anyone reading sentence 1 (at least anyone reading this book) hits on only one of these meanings (or but one of a select few) without any overt awareness that the other 111 meanings are perfectly possible.
There are theories in psycholinguistics that claim that what happens in a case like sentence 1 is that we unconsciously consider all 112 possible meanings and rule out all but one, but we do this so fast and so below the level of consciousness that we are completely unaware of it. Be that as it may, how can sentence 1 have so many meanings and why do we all, none the less, hit on one and, in fact, exactly the same one?
This fact is due to the grammar (in the grammar one sense) of the sentence. The subject of sentence 1 ("Lung cancer death rates") is a "nom-inalization" made up of a compound noun. Nominalizations are like trash compactors: they allow one to take a lot of information - indeed, a whole sentence's worth of information - and compact it into a compound word or a phrase. One can then insert this compacted information into another sentence (thereby making bigger and bigger sentences). The trouble is this: once one has made the compacted item (the nominalization), it is hard to tell what information exactly went into it. Just like the compacted trash in the trash compactor, you can't always tell exactly what's in it.
"Lung cancer death rates" could be a compaction of any of the following more expanded pieces of information:
2a   [lung cancer] [death rates] = rates (number) of people dying from lung cancer = how many people die from lung cancer
32    Discourses and social languages
2b   [lung cancer] [death rates] = rates (speed) of people dying from lung
cancer = how quickly people die from lung cancer 2c   [lung] [cancer death] [rates] = rates (number) of lungs dying from
cancer = how many lungs die from cancer 2d   [lung] [cancer death] [rates] = rates (speed) of lungs dying from cancer
= how quickly lungs die from cancer
The first two meanings (2a/b) parse the phrase "lung cancer death rates" as "lung-cancer (a disease) death-rates," that is "death-rates from lung-cancer," where "rates" can mean number of people dying or the speed of their death from the disease. The second two meanings (2c/d) parse the phrase "lung cancer death rates" as "lung cancer-death-rates," that is "cancer-death-rates for lungs," where, once again, "rates" can mean number of (this time) lungs dying from cancer or the speed with which they are dying from cancer. This way of parsing the phrase is analogous to the most obvious reading of "pet cancer death rates" (i.e. "cancer-death-rates for pets," that is, how many/how fast pets are dying from cancer). Of course, everyone reading this paper interpreted "lung cancer death rates" to be a compaction of 2a. Our question is, why?
Consider now the verbal phrase "are clearly associated with" in sentence 1. Such rather "colorless" relational predicates are typical of certain social languages. Such verbal expressions are ambiguous in two respects. First, we cannot tell whether "associated with" indicates a relationship of causation or just correlation. Thus, does sentence 1 say that one thing causes another (e.g. smoking causes cancer) or just that one thing is correlated with another (smoking and cancer are found together, but, perhaps, something else causes both of them)? Second, even if we take "associated with" to mean cause, we still cannot tell what causes what. You and I may know, in fact, that smoking causes cancer, but sentence 1 can perfectly mean that lung cancer death rates lead to increased smoking. "Perhaps," as Halliday remarks, "people are so upset by fear of lung cancer that they need to smoke more in order to calm their nerves" (Halliday and Martin 1993: 77-8). It is even possible that the writer did not want to commit to a choice between cause and correlate, or to a choice between smoking causing cancer or fear of cancer causing smoking. This gives us at least the following meaning possibilities for the verbal phrase "are clearly associated with";
3a   cause
3b   caused by
3c   correlated with
3d  writer does not want to commit herself                       ,      ■ ■
Now, let's finish with the phrase "increased smoking." This is another nominalization, compacting information. Does it mean "people smoke more" (smokers are increasing the amount they smoke), or "more people
Discourses and social languages    33
smoke" (new smokers are being added to the list of smokers), or is it a combination of the two, meaning "more people smoke more"?
We can also ask, in regard to the death rates and the increased smoking taken together, if the people who are increasing their smoking (whether old smokers or new ones) are the people who are dying from lung cancer, or whether other people dying as well (e.g. people who don't smoke, but, perhaps, are "associated with" smokers). Finally, we can ask of the sentence as a whole, whether it is representing a "real" situation ({'because more people are smoking more people are dying") or just a hypothetical one ("if more people were to smoke we know more people would die")? This gives us at least seven more meaning possibilities:
4a   increased smoking = people smoke more
4b   increased smoking = more people smoke
4c   increased smoking = more people smoke more
4d   the same people are smoking and dying
4e   the people smoking and dying are not all the same
4f   the situation being talked about is real (because)
4g   the situation being talked about is hypothetical {if)
We now have considered four possible meanings for the subject ("lung cancer death rates"), four possible meanings for the verbal phrase ("are clearly associated with") and seven possibilities for the complement ("increased smoking"). Like an old-fashioned Chinese menu, you can take one from list A and another from list B and yet another from list C and get a specific combination of meanings. This gives us four times four times seven possibilities, that is, 112 different possible meanings.
All of these meanings are perfectly allowed by the grammar of sentence 1 in the "grammar one" sense of grammar. And, in fact, there are other possibilities I have not discussed, e.g. taking "rates" to mean "monetary costs" or "lung cancer death rates" to be the rates at which lung cancer is dying. And yet - here's our mystery again - everyone reading this paper in a micro second hit on just one of these many meanings and the same one (or, at worst, considered a very few of the possibilities). Why?
The answer to the mystery I am discussing here may be perfectly obvious to you, but I want to suggest that, none the less, it is important for how we view language and language learning. We all hit on only one (and the same one) of the 112 meanings because we have all been part of - we have all been privy to - the ongoing discussion or conversation iri our society about smoking, disease, tobacco companies, contested research findings, warnings on cartons, ads that entice teens to smoke, and so on and so forth.
Given this conversation as background, sentence 1 has one meaning. Without that conversation - with only the grammar of English in one's head - the sentence has more than 112 meanings. Obviously, however important grammar is, the conversation is more important. It leaves
34   Discourses arid social languages
open one meaning (or a small number of possibilities, like allowing that sentence 1 also covers people getting lung cancer from secondary smoke).
A more technical way to put this point is this: meaning is not merely a matter of decoding grammar, it is also, and more importantly, a matter of knowing which of the many inferences that one can draw from an utterance are relevant (Sperber and Wilson 1986). And relevance is a matter deeply tied to context, point of view, and culture. One knows what counts for a given group of people at a given time and place as "relevant" by having been privy to certain "conversations" those people have heretofore had. If there had been a major conversation about environmentally induced lung cancer in a nervous society, then sentence 1 could perfectly well have been taken to mean that the prevalence of lung cancer is causing many more people to turn to smoking to calm their nerves (2a + 3a + 4b).
So, we have concluded, we speak and write not in English alone, but in specific social languages. The utterances of these social languages have meaning - or, at least, the meanings they are taken to have - thanks to being embedded in specific social conversations. Though I have established these points in regard to a single sentence (sentence 1), I take them to be generally true.
To teach someone the meaning of sentence 1 - or any sentence for that matter - is to embed them in the conversational sea in which sentence 1 swims. To teach someone the sort of social language in which sentences like sentence 1 occur is to embed them in the conversations that have recruited (and which, in turn, continually reproduce) that social language.
2.10 Big "C" Conversations: Conversation among Discourses
Now it is time to become clearer about what I mean by "conversation." The word "conversation," as I am using it here, can be misleading. We tend to think of conversations as "just words." But the sorts of conversations I am talking about involve a lot more than words; they involve, in fact, Discourses. It is better, perhaps, to call them "Conversations" with a "big C," since they are better viewed as (historic) conversations between and among Discourses, not just among individual people. Think, for instance, as we mentioned above, of the long-running, historic Conversation between biology and creationism, or between the Los Angeles police department and Latino street gangs.
More than people, and more than language, are involved in Conversations. They involve, as well, at least the following three non-verbal things:
1     controversy, that is, "sides" we can identify as constituting a debate (Billig 1987);
2     values and ways of thinking connected to the debate; and
Discourses and social languages   35
3     the "symbolic" value of objects and institutions that are what we might call non-verbal participants in the Conversation (Latour 1987).
Let me give you an example of what I am trying to get at here. It is fashionable today for businesses to announce (in "mission statements") their "core values" in an attempt to create a particular company "culture" (Collins and Porras 1994, examples below are from pp. 68-9). For instance, the announced core values of Johnson & Johnson, a large pharmaceutical company, include "The company exists to alleviate pain and disease" and "Individual opportunity and reward based on merit," as well as several others.
One might wonder, then, what the core values of a cigarette company might be. Given the Conversations that most of us are familiar with -about the U.S. and its history in this case, as well as about smoking - we can almost predict what they will be. For example, the espoused core values of Philip Monis, a large company which sells cigarettes among a great many other products, include "The right to personal freedom of choice (to smoke, to buy whatever one wants) is worth defending," "Winning - being the best and beating others," and "Encouraging individual initiative," as well as (in a statement similar to one of Johnson & Johnson's statements) "Opportunity to achieve based on merit, not gender, race, or class."
We all readily connect Philip Morris's core value statements to themes of American individualism and freedom. Note how the values of "individual initiative" and "reward for merit," which are part of the core values of both Johnson & Johnson and Philip Morris, take on a different coloring in the two cases. In the first case, they take on a humanistic coloring and in the other the coloring of "every man for himself." This coloring is the effect of our knowledge of the two sides to the "smoking Conversation" in which, we all know, individual freedom is pitted against social responsibility.
Note, then, here how values, beliefs, and objects play a role in the sorts of Conversations I am talking about. We know that in this Conversation some people will hold values and beliefs consistent with expressions about individualism, freedom, the "American way," and so forth, while others will express values and beliefs consistent with the rights of others, social responsibility, and protecting people from harm, even harm caused by their own desires. In turn, these two value and belief orientations can be historically tied to much wider dichotomies centering around beliefs about the responsibilities and the role of governments.
Furthermore, within this Conversation, an object like a cigarette or an institution like a tobacco company, or the act of smoking itself, takes on meanings - symbolic values - within the Conversation, but dichotomous meanings. Smoking can be seen as an addiction, an expression of freedom,
36    Discourses and social languages
or a lack of caring about others. The point is that those familiar with the Conversation know, just as they can select the meaning of sentence 1 out of 112 possibilities, the possible meanings of cigarettes, tobacco companies, and smoking.
When we teach literature or physics, or anything else, for that matter, we index a multiple, but specific world of Conversations, though it is no easy matter in classrooms to get most of these Conversations going "for real." When we teach language - whether this be French, English as Second Language, composition, basic skills, literacy, or what have you - we face in the purest and hardest form the question of what Conversation or Conversations make words and phrases meaningful and relevant here and now.
The themes and values that enter into Conversations circulate in a multitude of texts and media and have done so in the past. They are the products of historic meaning making within Discourses. Of course, people today often know these themes and values without knowing the historical events that helped create or sustain them in the past and pass them down to us today.
For example, throughout the latter half of the nineteenth century in Massachusetts, courts were asked to return escaped slaves to their Southern "owners" (von Frank 1998). These court battles, and the accompanying controversies in newspapers and public meetings, engaged two distinctive Discourses among several others (for example, several Discourses connected to Black churches and to Massachusetts' significant nineteenth-century population of free Black people, some of them professionals, such as ministers, doctors, and lawyers - note that it is hard to know what to call these people, they were of African descent, born in the U.S., of all different colors, but were not full citizens).
One Discourse, connected to people like Emerson and Thoreau, championed freedom, personal responsibility, and morality as constituting a "higher law" than the law of states, the federal government, or the courts. They argued and fought, not only to not return the slaves, but to disobey the court and the federal officials seeking to enforce its mandate. The other Discourse, heavily associated with nationally-oriented political and business elites, championed the rule of law at the expense of either the slave's freedom or one's own personal conscience.
These two Discourses were, by no means, just "statements" and "beliefs." There were, for example, distinctive ways, in mind, body, and social practice, to mark oneself in nineteenth-century Massachusetts as a "Transcendentalist" (i.e. a follower of Emerson and his colleagues) and to engage in social activities seen as part and parcel of this identity.
Many people today have no knowledge of the debates over escaped slaves in Massachusetts and nationally in the nineteenth century (though these debates, of course, helped lead to the Civil War). However, these debates sustained, transformed, and handed down themes and values
Discourses and social languages    37
that are quite recognizable as parts of ongoing Conversations in the mid-twentieth century (e.g. in the Civil Rights Movement) and today.
Of course, I must hasten to add, again, that a number of other important Discourses played a significant role in the escaped slave cases in Massachusetts. Blacks were part of some integrated Discourses, as well as their own distinctive Discourses. Furthermore, all these Discourses interacted with each other, in complex relations of alliance and contestation, with some important overlaps between Discourses (e.g. between the Transcendental-ists and John Brown's distinctive and violent Discourse in regard to slavery and abolition).
Because people are often unaware of historical clashes among Discourses, it is often easier to study Conversations, rather than Discourses directly, though it is always important and interesting to uncover the historical antecedents of today's Conversations. Conversations are the precipitates of what we will call, in subsequent chapters, "situated meanings" and "cultural models" as these have circulated with and across Discourses in history. .
The way in which I have used the term "Conversation" here is a use that is sometimes covered in other work by the term "discourse." People who use the term "discourse" this way mean something like this: the range of things that count as "appropriately" "sayable" and "meaning-able," in terms of (oral or written) words, symbols, images, and things, at a given time and place, or within a given institution, set of institutions, or society, in regard to a given topic or theme (e.g. schools, women's health, smoking, children, prisons, etc.). Such a use of the term "discourse" or "Conversation" (the term I will use) concentrates on themes and topics as they are "appropriately" "discussible" within and across Discourses at a particular time in history, across a particular historical period, within a given institution or set of them, or within a particular society or across several of them (Foucault 1985).
2.11 Social languages and Discourses as tools of inquiry
hi this chapter, I have treated the terms "social languages," "Discourses," and "Conversations" realistically. That is, I have spoken about them as things that exist in the mind and in the world. And indeed, this is, I believe, both true and the easiest way to grasp what they mean and how and why they are significant for discourse analysis.
But it is important to realize that, in the end, these terms are ultimately our ways as theoreticians and analysts of talking about, and, thus, constructing and construing the world. And it is in this guise that I am primarily interested in them. They are "tools of inquiry." "Social languages," "Discourses," and "Conversations" are "thinking devices" that guide us to ask certain sorts of questions. Faced with a piece of oral or written language, we ask the following sorts of questions:
38   Discourses and social languages
•    What social languages are involved? What sorts of "grammar two" patterns indicate this? Are different social languages mixed? How so?
■ What socially situated identities and activities do these social languages enact?
•    What Discourse or Discourses are involved? How is "stuff" other than language " ("mind stuff" and "emotional stuff" and "world stuff" and "interactional stuff" and non-language symbol systems, etc.) relevant in indicating socially situated identities and activities?
•    What sort of performance and recognition work (negotiations and struggles) has gone on in interactions over this language? What are the actual or possible social, institutional, and political consequences of this work?
•    In considering this language, what sorts of relationships among different Discourses are involved (institutionally, in society, or historically)? How are different Discourses aligned or in contention here?
•    What Conversations are relevant to understanding this language and to what Conversations does it contribute (institutionally, in society, or historically)?
Note: The term "Discourse" (with a big "D") is meant to cover important aspects of what others have called: discourses (Foucault 1966, 1969, 1973, 1977, 1978, 1980, 1984, 1985); communities of practice (Lave and Wenger 1991); cultural communities (Clark 1996); discourse communities (Berken-kotter and Huckin 1995; Miller 1984); distributed knowledge or distributed systems (Hutchins 1995; Lave 1988); thought collectives (Fleck 1979); practices (Barton and Hamilton 1998; Bourdieu 1977,1985,1990a, b; Heidegger 1962); cultures (Geertz 1973,1983); activity systems (Engestrom 1987,1990; Leont'ev 1981; Wertsch 1998); actor-actant networks (Callon and Latour 1992; Latour 1987); and (one interpretation of) "forms of life" (Wittgenstein 1958). Discourses, for me, crucially involve:
•    situated identities;
•    ways of performing and recognizing characteristic identities and activities;
•    ways of coordinating and getting coordinated by other people, things, tools, technologies, symbol systems, places, and times;
•     characteristic ways of actmg-mteracting-feeling-emoting-valuing-gesturmg-posturing-dressmg-tiiMdng-beHeving-laiowmg-speaking-listening (and, in some Discourses, reading-and-writing, as well).
A given Discourse can involve multiple identities (e.g. a teacher, Ms. X , and her kindergarten students take on different situated identities, and different, but related, ones in diverse activities within the "Ms. X-and-her-students classroom Discourse," provided that Ms. X has, in fact, created a coherent Discourse in and around her classroom). Some people dislike
Discourses and social languages   39
the term "situated identity" and prefer, instead, something like "(social) position" or "subjectivity" (they tend to reserve the term "identity" for a sense of self that is relatively continuous and "fixed" over time). I use the term "identity" (or, to be specific, "socially-situated identity") for the multiple identities we take on in different practices and contexts and would use the term "core identity" for whatever continuous and relatively "fixed" sense of self underlies our contextually shifting multiple identities.