CHAPTER 8. MAKING MEANING: THE PRINCIPLES OF SOCIAL SEMIOTICS
In the last 10 years, students of social behavior have begun to construct a new theoretical synthesis. It provides a radically different way of looking at how human beings make sense of and to one another: how we make meaning. This new theory is still incomplete, but I believe that it is the foundation on which the social science of the future will be built. Because it is so new, and still unfinished, it doesn't yet have a commonly accepted name. Following the lead of the social linguist Michael Halliday and others, I will call it social semiotics (Halliday, 1978; Lemke, 1984, 1987a, 1987b, 1989; Threadgold, in press).
THE GENEALOGY OF SOCIAL SEMIOTICS
Social semiotics is a synthesis of several modern approaches to the study of social meaning and social action. One of them, obviously, is semiotics itself: the study of our social resources for communicating meanings. Historically, semiotics (also called semiology) was invented as part of the effort to find a scientific basis for linguistics (de Saussure, 1915; Bakhtin-Voloshinov, 1929; Hjelmslev, 1943). Semiotics is the study of all systems of signs and symbols (including gestures, pictures, even hairstyles) and how we use them to communicate meanings. Linguistics covers the one special case of language and so is part of semiotics.
The name social semiotics is meant to distinguish the new synthetic theory from more traditional approaches to semiotics (e.g., Peirce, 1908/1958; Eco, 1976), which we can call formal semiotics. Formal semiotics is mainly interested in the systematic study of the systems of signs themselves. Social semiotics includes formal semiotics and goes on to ask how people use signs to construct the life of a community.
Social semiotics is not new in trying to unite the study of human behavior, especially meaning-making behavior (talking, writing, reasoning, drawing, gesturing, etc.), with the study of society. There is a long tradition of doing this in cultural anthropology and ethnography. While many anthropologists and ethnographers have taken the role of language too much for granted, they have still made great contributions to the study of symbols and symbolic actions. One of the founders of modern anthropology, Bronislav Malinowski (1923, 1935), also formulated several of the principles of social linguistics incorporated in modern social semiotic theory. And social semiotics also builds directly on the work of the modern anthropologist, Gregory Bateson (1972).
Linguistics itself has often been concerned with language as a tool of social action, especially in the European traditions of functional linguistics (Propp, 1928; Bakhtin-Voloshinov, 1929; Bakhtin, 1935, 1953; Jakobson, 1971; and many members of the Prague School, see Garvin, 1964). A branch of this tradition took root in the United States and influenced early American anthropology to investigate the relations of language and culture among Native Americans (Sapir, 1921; Boas, 1922; Whorf, 1956). Another branch of functional linguistics flourished in England (Firth, 1957) and led to the work of Michael Halliday (e.g., 1961, 1975, 1978, 1985a), whose theory of linguistic meaning is generalized by social semiotics into a theory of meaningful social action.
There is one essential piece still missing from this synthesis: a theory of society as a whole. Cultural anthropology has usually been more interested in describing the similarities and differences between societies than in explaining them in the way that physics tries to explain material processes. Traditionally this job has fallen to sociology, but it too has found it easier to describe and compare than to explain. Recently a minor branch of sociology, known as ethnomethodology, picked up the ethnographers' interest in everyday life, and looked at processes of social action in our own society (Goffman, 1959, 1974, 1981; Garfinkel, 1967; Garfinkel & Sacks, 1971). This approach, too, is incorporated in social semiotics, but it still lacks any theory of society in the large. It is still only a microsociology; what we need is macrosociology.
The problem with most general theories of society is that they are written from the point of view of the dominant groups in the society. They tend to be elaborate rhetorics that really only repeat commonly accepted rationalizations for the way things are. There is, of course, one famous exception: Marx's political sociology.
In the United States, mainstream sociology has largely ignored Marx's social theory because of its political implications. But in Europe, Asia, Latin America, and even Canada important parts of Marx's theory have been reworked and incorporated into a number of different modern social theories. Most of these theories have a few basic principles in common, however different they may be from Marx's original theory of over a century ago, or from each other (cf. Gramsci, 1935; Althusser, 1971; Habermas, 1972; Bourdieu, 1972). The basic theory today can be called critical sociology because it explains social processes without assuming that the way they are is the way they have to be, or the way they ought to be. Social semiotics modifies critical sociology considerably in the process of joining it to other essential elements of the synthesis, hopefully improving it in the process.
There are two other pieces in the puzzle that are worth mentioning here. Ethnomethodology is the application in social science of some philosophical approaches known collectively as phenomenology(Husserl, 1960, 1965; Schutz, 1932; Merleau-Ponty, 1945). These perspectives also make important contributions to social semiotics' view of the relation of social action to the human individual (the so-called "problem of the subject" cf. Lemke, 1988c). Finally, social semiotics makes use of the insights of Michel Foucault (1969, 1976), the French historian and social theorist, who has analyzed the relations between how we talk about the world and how we act and are acted upon in it. He connects discourse and the technologies of action and control to the larger patterns of belief and power in society.
All these approaches to the study of language, symbol, symbolic action, and human culture have influenced and been influenced by a major philosophical change in Western culture itself. The great battle between theology and science from the Renaissance to the Enlightenment had led philosophers to try to reconcile Idealist theories, which held that Truths existed in an absolute sense (like God), with Materialist theories, for which only Things existed absolutely. Eventually concern with these metaphysical issues gave way to interest in epistemological questions: how can people know, or find out, what truths or things really exist? In science, these became the familiar questions of scientific method. The dominant belief for most of the 18th and 19th centuries was that we discover absolute truths by systematically studying things and by making theories that correspond to our observations.
This answer, however, has not held up very well in the 20th century. In science, and in all other fields, it seems, we do not so much "discover truths" as we construct meanings. We devise useful ways of talking about things and processes, and useful systems of technical action (e.g., measurements, technologies). We construct systems of meanings by using language, mathematics, diagrams, and techniques. They are our social tools, and they differ from one social community to another. This is the view of Social Constructionism in philosophy (cf. Wittgenstein, 1949; Rorty, 1979), and it is also the view of social semiotics.
I want now to give a systematic overview of the principles of social semiotics. A complete version would take another whole book, so this outline will have to leave out many of the details. They can be found in the various books and articles in the References that are cited in this chapter. Social semiotics presents a way of looking at human behavior and human society that makes a lot of sense once you get used to it. But it has a way of talking about these subjects that can seem a little strange at first. Having read the rest of this book, however, you should find that much of what I will be saying in this last chapter will sound at least a little familiar.
ACTION, CONTEXT, AND MEANING
Social semiotics is basically a theory of how people make meaning. It asks how we make sense of and to one another and how we make sense of the world. It concerns itself with everything people do that is socially meaningful in a community: talking, writing, drawing pictures and diagrams, gesturing, dancing, dressing, sculpting, building -- in effect, everything. But it looks at everything from a particular point of view. Social semiotics tries to answer these questions:
How does does the performance of any particular socially meaningful action make sense to the members of a community?
How do people interpret it?
What are its parts and how are they related to each other?
What alternatives could have been done in its place, and how would their meanings have differed?
When do people make this particular meaning? Engage in this particular action?
How does the meaning change in different circumstances or contexts? How do people feel about the action and its meaning?
What larger social patterns does the action belong to?
How does it tend to recreate or change the basic patterns of the society?
The basic assumption of social semiotics is that meanings are made. This is a change in the semantics of the term meaning. It is misleading to say, as people often do, that something has meaning, as if the meaning was somehow built-in. A word, or a diagram, or a gesture does not have meaning. A meaning has to be made for it, by someone, according to some set of conventions for making sense of words, diagrams, or gestures.
Different people make different meanings for the same word, the same diagram, the same gesture. The same person may make different meanings for something at different times, depending on circumstances and past experience. The most important differences are differences in the conventions for how to make a meaning in a particular context. People from different communities, including different groups within one larger community, tend to have different ways of making meaning. We can only make sense of and to one another to the extent that we share the same ways of making meanings. We must belong to the same, similar, or overlapping communities to do this. To the extent that we do share meanings, we have become members of the same social group, at least partially.
We do not, of course, just make sense of, or for, words and gestures. We make sense with words and gestures, that is, we use them to make socially recognizable meanings, to perform socially meaningful actions. We use them to communicate information, to make requests and offers, to praise and blame, to insult, joke, and pray. The ways in which we use them are characteristic of the communities we belong to.
We do not use just words and gestures in this way, we also use lines and spaces (in diagrams), notes and rests (in music), mathematical symbols (in formulas), steps (in dance), fabrics and colors (in dressing), supports and spaces (in architecture), moves and plays (in games), and every other form of symbol and action. We speak meaningfully, draw meaningfully, compose and choreograph meaningfully, dress and move meaningfully, build and play meaningfully by deploying the resources our community gives us (words, lines, notes, steps, moves), according to patterns that make sense to others in our community. We use those same patterns to make sense of the actions of others.
Every community has its own meaning-making (i.e. semiotic) practices. These are the ways in which its members perform actions that are meaningful in the community. They are patterns of action that are repeated many times. Everything you do that makes sense of, or with, a word, an object, or an action follows one of the semiotic practices of your community. It is a semiotic practice of your community (or a combination of semiotic practices). Semiotic practices are actions that make sense in a community. The form of action may be speech, gesture, drawing, building, or even washing the dishes. Semiotic practices are the fundamental elements of social semiotic theory.
What makes an action a semiotic practice? This is another way of asking how an action becomes meaningful in a community, how we make sense of an action (or event). Social semiotics begins with an answer to this basic question. Fundamentally, every action is made meaningful by placing it in some larger context. In fact, we place every action or event in many contexts in order to make it meaningful. The meaning we make for an action or event consists of the relations we construct between it and its contexts. Making meaning is the process of connecting things to contexts. We make actions and events meaningful by contextualizing them. The most important of all semiotic practices are these contextualizing practices. Social semiotics analyzes the kinds of contexts in which we place things, and the kinds of relationships we construct between them and these contexts.
Consider, for example, the meaning of a word said by someone. Take the word "scientific" used by the student who complained about his teacher, saying, "Why can't he explain science in a scientific way?" (see Chapter 5). What are the contexts we connect that word to in order to make it mean something? First of all, there is the context of the whole utterance: What words preceded and followed it? Then there is the context of the situation in which it was said: Who said it to whom? What was the event or activity that was going on at the time? What were the relations of the participants to each other? There are also the wider contexts of the community: Under what circumstances is that word ordinarily used in this community? What alternative words might have been used in its place? Is use of the word in this way typical of a particular social group?
What we can say about contextualizing a word holds just as well for a gesture, a diagram, or any action. Social semiotics identifies and names these different sorts of contexts. The wholes in which any action (or thing, event, word) is placed as a part are its syntagmatic contexts. The most common sorts of syntagmatic contexts are sequences of actions that are themselves built up of shorter sequences and in turn belong to longer sequences.
The activity structures of the classroom, like Triadic Dialogue, are syntagmatic contexts of this kind. A word is taken to be part of a Teacher Question, or a Student Answer, as well as part of a (complete or incomplete) sentence. Sentences and paragraphs in writing, moves and exchanges in dialogue, plays and innings in baseball are all larger contexts of this kind. There are also syntagmatic contexts in spatial arrangement that are not necessarily sequential: details and figures in a painting, place settings and table arrangements in dining, and so on. Ultimately, however, all semiotic structures can be analyzed as activity structures, because they are all the product of semiotic practices, the results of sequences of social actions (writing, drawing, painting, arranging a dinner table, etc.).
In addition to syntagmatic contexts, there are also paradigmatic contexts. These are the contexts of "what might have been." They consist of other words, or actions, that might have taken the place of the one that occured, in the same syntagmatic context. In the same sentence, what other words could have been used? At the same point in the game, what other plays might have been made? For the same detail in the painting, what other colors could have been used? What matters here is the relation in meaning of what was to what could have been. How would we make sense of something differently if a word or color had been different?
Finally, there are what I can call indexical contexts. These are social contexts that are usually associated in some way with an action in a particular community. The action and the context "index" or point to one another with a certain degree of probability. For example, there are some actions which index or point to the teacher as the person who is performing them (e.g., Admonitions, Evaluations), and others which are more usually associated with students (e.g., Calling Out, Requests for the Pass). The meaning of an action does depend on who performs the action, who speaks the words or makes the gesture. It depends, specifically, on the type of person: teacher or student, doctor or patient, lawyer or client, policeman or suspect, parent or child, and so on. Most actions can be done by anyone, but the meaning will be very different if it is done by someone of the "wrong" type.
What the right type of person to say or do something is, or what the right situation or circumstances are to do it, depends in turn on wider social contexts. They index and are indexed by one another. If parents and children speak as equals at the dinner table, we are more likely (in America, in the 1980s) to be in a liberal middle-class home than in a conservative, working-class one.
Who speaks how to whom, when, points to social group, to culture, to historical period. And vice versa. What, after all, does it mean to be middle-class, except that you do certain things and say certain things in certain ways under particular conditions? Indexical definitions of social context and the social meaning of actions apply not just to social group, class, and period. They apply as well to gender, nationality, ethnic group, occupation, age, and all significant social categories -- to all the ways in which we divide society according to differenes in patterns of behavior.
Notice, by the way, that what matters in indexical contextualization is the pattern of actions, the systematic relations of actions to each other and to categories or other sorts of "context." Young and old, male and female, rich and poor, Black and White, artist and scientist do not behave completely differently in our society. They perform many of the same actions, but they perform them in slightly different contexts, with differences in background and circumstances, and so with different social meanings. They read, write, reason, labor, and play according to different patterns. These patterns of action of different social groups within a single society are not just different: They have systematic relationships to one another. We will come back to these issues a little later (e.g., see the discussion of "heteroglossia" and "heteropraxia" below).
There is one very important special case of indexical contextualization: thematic contexts. Everything we say can be made sense of by hearing it in relation to other things we have heard on other occasions that use the same thematic pattern. And just as different social groups behave differently in similar circumstances, their members also identify themselves (i.e. index their social group) by how they talk differently about a subject. Scientists and artists, teachers and students, managers and workers talk about many subjects in characteristically different ways. They construct (or reconstruct) in their speech, writing, and reasoning the different thematic patterns that index their social group.
Contextualization is a very powerful notion. From it a complete theory of meaning and social relations can be built, if we take into account that it is actually the pattern of probabilities that an action will index and be indexed by some social category or type of context which distinguishes (and relates) one social group or community from (and to) another. A social group can be indexed by what actions index what contexts in that group. As you can see, this can get pretty complicated, but the mathematical theory of redundancy provides a way to keep straight what is indexing or contextualizing what. The complex patterns of relationships that result can be called "metaredundancies" or "metacontextualizations" (see Lemke, 1984, pp. 33-44).
PRACTICES, PROCESSES, AND COMMUNITIES
An action that makes a socially recognizable meaning in a community is a semiotic practice. We need to look at these meaning-making practices in two ways. First, they are actions which make sense in the community. Second, they include the actions by which we make sense of other actions (and, by analogy, make sense also of events and things). As we have just seen, a social group or community can be defined indexically by the typical (probable) patterns of action of different types of people under different circumstances in that community. In this sense, a community is not composed of people per se, but of people acting. It is made up, not of individual biological organisms, but of interconnected life-processes. It is the patterns of those processes that define a community and tell us how it is similar to and different from other communities, how it keeps itself going, and perhaps even how it may change in the future.
An action, if it is socially meaningful in a community, can be talked about as a semiotic practice. In those terms it has semiotic relations to other practices in the community, relations of indexing and being indexed, contextualizing and being contextualized. It has syntagmatic relations to the larger wholes in which it is placed, and paradigmatic relations to the alternatives that could have stood in its place. But an action can also be talked about as a material, physical (and usually biological and ecological) process, as well. As such, it has other sorts of relationships to other actions: relationships of exchange of matter, energy, and information (entropy).
A community is not just a system of semiotic practices: it is also a dynamic, physical, biological system. To survive it must regulate itself internally and interact externally. It needs to cycle matter and energy through itself: It needs food and resources, and it needs to get rid of wastes. It also needs to maintain a relatively stable, useful environment, and to be prepared to change and adapt to that environment when it cannot control it. Most such systems also develop from immature to mature forms, reproduce imperfectly, and evolve (Lemke, 1984, pp. 27-30, 104-112; and Lemke, in preparation).
A social community, therefore, is both a dynamic, open material system of physical and biological processes, and it is a dynamic, open semiotic system of meaningful actions and meaning-making practices. Every semiotic practice is simultaneously a material process. And every material process we know is assigned a social meaning by how we talk about it and how we act with respect to it. Because social practices are also social processes, they have material relations to one another that may not already be recognized in the social system of meanings. This makes social change both possible and inevitable (see below).
In social semiotics "things" are not fundamental. An object or entity of any kind is always analyzed as a social construction, that is, as the product of social practices/material processes that make it something meaningful in a community. When you see a pencil, your perception combines biological processes and semiotic practices to "see" something that you have been taught to regard as an object, with size, shape, color, and a name, a description, a value, and a set of typical uses. Your community endows the object with meaning, and every meaningful thing you do with the object is guided by its meanings in your community.
This applies to people as well. An individual biological organism is socially constructed in much the same way that any other object is, but in our society we combine (and often confuse) this notion of an organism with the very different notion of social individual. Social individuals are known, recognized, and identified by how they act. In the case of people, this also has a lot to do with how they look. An individual is assigned a biological sex and a social gender (and, as in the case of transsexuals, these two don't have to be the same). How do we tell that a person is male or female? Masculine or feminine? Straight or gay? Handsome or ugly? We do it by employing specific semiotic practices of our community, whether they are scientific ones or commonsense ones.
Our community also has its specific ways of connecting the individual at one moment in time to "the same" individual earlier or later. We apply particular criteria to construct this social continuity of personalities, just as we do to construct the temporal continuity of organisms and other objects. In our own community, for instance, we insist on an exact "one body, one person" correspondence and regard multiple personalities in a single body as abnormal, even if each of the personalities is normal by itself. We do not construct personal continuity from generation to generation (cf. reborn spirits, reincarnation), and we are divided over whether to construct any continuity of individuality at all beyond the moment of death. Other cultures disagree with us on these matters and on many of the details of constructing the social meaning of persons, places, and things. (See Lemke, 1988c.)
We construct social "subjects" as well as social "objects." Subjects act and live; they make meaning. And yet they are also made meaningful. Social semiotics has a long way yet to go in analyzing such things as human emotion and the social construction of personality and individuality. There is more to an individual than the sum of all the social groups and social categories he or she can be assigned to in a community. But the larger social patterns of the community tend to depend more on what social role or type an individual represents than on their uniqueness, because those patterns are patterns of relationships among groups and categories. The patterns are themselves social constructions, but they are not made up by individuals. They are the result of history, of many individual actions that have tended to recreate and change these patterns over long periods of time.
SOCIAL SEMIOTICS VS. MENTALISM
Where in all this grand picture is the "mind"? Somehow it hasn't seemed necessary to use this word, or others that go with it, like intention, cognition, thought. The language of mentalism, which is basic to the recent revival of cognitive psychology, assumes that there is an autonomous domain of phenomena between the biological and the social. I do not believe that this is so, and social semiotics rejects mentalism completely.
After a long decline, mentalism was revived just a few decades ago because of the failure of behaviorist psychology to explain how people learned and used language (Chomsky, 1959). It was then rapidly generalized, under the name cognitive science, to include computer models of many forms of semiotic activity: problem solving, writing, learning behavior, and so on. Much of the research that has been done in these areas is extremely valuable, but the basic theoretical language used remains hampered by the limitations of mentalism.
Mentalism ignores both the biological and the social, but the problems it addresses are precisely the problems of the relations of the biological to the social. Cognitive science originally sought a close relationship with neurobiology, but it has long since gone its own way instead. Mentalist models treat cognitive processes as isolated phenomena that happen within a single mind isolated from others, as mental processes isolated from social processes.
Mentalism hides from social reality behind an assumption of universality: that all human minds work in the same way. But the work that cognitive science seeks to describe is the work of engaging in semiotic practices, and those practices are crucially different in different cultures and different social groups. Most cognitive science research describes a small number of social practices of middle and upper-middle class Americans. In many cases, apart from its use of the language of mentalism (which it does not actually need), it does this very well. There is no reason it should not, since the data of cognitive science is not "mental" data at all: It is data in the form of language and social behavior, records and descriptions of semiotic practices.
But by ignoring the social, mentalism hides, not just from the reality of social differences, but from the problem of social values. Cognitive science is fond of identifying "expert" ways of solving problems, writing, reading, learning, and so on. It then describes these and implies that everyone should learn to re-program themselves to do things in these "expert" ways (or else rely on computers that do). These are the narrow values of technocrats and efficiency experts. They take the preferences of one small but powerful group in society (white, male, upper-middle class, with North European cultural values) and project them as intrinsically superior. Since cognitive science acts as if all minds basically worked the same way, it is saying that the way this one group does things is the way the Mind, all minds, work best. They have safely insulated their "science" from cultural diversity and social conflicts of values and interests.
Mentalism also largely ignores linguistics and semiotics, despite the fact that it was revived by a well-known American linguist (Chomsky, 1959). Most cognitivist accounts of language use pay little or no attention to the semantics of functional linguistics, but try instead to invent an autonomous cognitive logic (lingua mentis, the language of the mind). But what we call "thought" is itself constructed through the medium of language and other semiotic resources (depiction, action structures), and the semantics of natural language, far richer than any cognitive logic, is all that's needed to account for what people do with language. I have put "thought" in quotation marks, because, so far as social semiotics is concerned, there is no separate phenomenon to bear this name.
What we call thinking is simply material processes which enact the meaning-making practices of a community: the use of language and other semiotic resources. This is not to say that there are not individual differences in what we say or write or reason: most texts and many actions are unique, but every meaningful text and action largely conforms to recognizable social patterns.
Social semiotics identifies those patterns and analyzes individual behavior in relation to them. Mentalism does not have a theoretical language for doing this, though in practice, given the nature of its data and methods of analysis, much cognitive research does describe parts of these patterns very well, even though it misunderstands and misidentifies their social nature. For more detailed critiques of mentalism, see Geertz (1983, Chapter 7), Lemke (1989), and Thibault (In press, Chapter 2).
SEMIOTIC RESOURCES AND FORMATIONS
Semiotics describes social action in terms of semiotic resources and semiotic formations. A semiotic resource system, such as language, is a system of possible ways of meaning. Information about it tells us what one can say in the language, and how to say it. In general, a semiotic resource system matches the kinds of meanings you can make (semantic functions) with the actions (such as words) needed to make those meanings in a particular community. This is a generalization of Halliday's model of language as "meaning potential," a semantic resource (Halliday, 1978).
A semiotic formation, on the other hand, is an actual pattern of meaningful action, using semiotic resources, that is repeatedly performed and recognized in a community. Activity structures and thematic patterns (more properly called thematic formations) are examples of semiotic formations. A community deploys its semiotic resources in certain habitual ways, and these are its semiotic formations. A formation is a sort of "institutionalized" way of talking, or gesturing, or behaving. Semiotic resource systems tell us what you can meaningfully do or say in a community; semiotic formations describe what repeated does get done and said.
Resource systems and formations are interdependent. Ultimately, a resource system, like the English language, is an abstraction from the uses of English in a community. Those regular, repeatable, habitual uses are formations (or parts of formations). When the formations change in systematic ways in relation to one another, the language resources have changed, too. But semiotic resources, because they are defined at a higher level of abstraction than formations, change very slowly. New thematic formations in science or politics can appear very quickly, but the grammar and semantics of English changes much more slowly.
There are many different resource systems in addition to language. There is the system of Depiction: the conventions of drawing, painting, diagrams, and so on (cf. O'Toole, in press). As a resource system, Depiction defines what can be pictured and how, what options are available to us. Pictorial formations include typical kinds of depiction: bar graphs, outline drawings, stick-figures, portraits, still lifes. Gesture and movement give us further semiotic resources, and common gesture-routines (shaking hands, waving good-bye) and movement patterns (sauntering, jogging, waltzing) are examples of formations that deploy these resources.
There are, of course, many more semiotic resource systems: Music (van Leeuwen, in press), Architecture (Preziosi, 1983), Dress and Grooming (Barthes, 1983), Cooking and Dining (Douglas, 1984), and so on. Each has its own typical formations: sonatas and concertos, villas and skyscrapers, tuxedos and sarongs, roasts and puddings. The most general semiotic resource system is that of social action itself, and its formations are the activity structures of a community, from factoring polynomials to washing the dishes, from writing sonnets to playing a game of tennis.
Social actions, including speaking particular words or writing them, do not make meaning simply by repeating the patterns of common semiotic formations in a community. As we have seen, they also make meaning (or we make meaning of and with them) by how they are used in a particular situation. It matters when we say or do something, where we do, and with or to whom (i.e., what type of person). You can dance, and if you follow the conventions of dancing for your time and place, others will recognize that that is what you are doing (and not having a fit, for example). If you enact a recognizable formation, they may see that you are waltzing. But the meaning of your waltzing then and there will depend on the situation: Is there audible music or not? are you dancing alone or with a partner? Is there a dance floor? are you in a place and situation where dancing is normal? Are you dancing with someone of your own or the oppposite sex? Is your partner unusually young, or old? Is your partner related to you socially in a way that makes the dancing specially significant? Is this kind of dancing normal for your social group?
A record of social action, whether it is a piece of writing, a transcription of a tape recording, a film or video, or just an account of some events, is a semiotic text. The actual events constitute a semiotic performance, and any material artifact that results can be called a semiotic production. When there is no particular reason to use these distinctions, and especially when the resource system of language plays an important role, I will just use the word text to refer to the actual, concrete, particular enactment of some semiotic practices, or a record or product of the performance. Semiotic texts are the basic data of social semiotics. They are the stuff of lived social life, so far as it is available for systematic study.
When we engage in a semiotic performance, that is, when we do any meaningful social action, we deploy semiotic resources strategically. We can never stray too far from some recognizable pattern of our community (semiotic formation), or our actions will become meaningless for others. But within those broad limits, we have considerable freedom to make an enormous variety of subtly different meanings through what we do and how we do it. That is why we need to take so many different sorts of contexts into account in analyzing how a particular "text" of actions means what it does. And we can never forget that it means different things (or nothing) to those whose social meaning-making practices are significantly different from our own.
In analyzing any action-text, including a purely verbal one, there are two perspectives we need to use. One is the dynamic perspective. This is the point of view of someone witnessing the events as they were actually happening. At each point in the sequence of action (or words), we have a different sense of what is being meant or done now, what things mean up to this present point, and what is likely to happen next. When the next action occurs, all that could change: What we thought just happened could, retrospectively, have turned out to be or mean something quite different than it seemed to as it happened. It is even possible that our whole sense of what was going on up to that point might have to be revised. And certainly our expectations about what may happen next can be radically changed. In real life the unexpected happens. It art and literature it often does, too.
There is a second perspective which is also important: the synoptic perspective. This is the viewpoint of someone who stands outside of time, after the whole sequence of actions has taken place, with the complete text of what happened in hand. This is the usual perspective of researchers, and it is quite useful, but it must be complemented by the usual perspective of participants, the dynamic perspective. We need to see how strategically (and tactically?) the events unfolded; not just how things turned out in the end. This is especially true when there are surprises, or when the events unfold over a fairly long period of time.
From now on I am mostly going to talk about semiotic formations, rather than semiotic resource systems. But first I want to identify the general semiotic functions that all semiotic resource systems provide the means to do. These are again generalizations from Halliday's model of the most basic semantic functions of language (which he calls its metafunctions; see Halliday, 1978, especially Chapter 2).
First, a semiotic resource system enables us to make representations. Basically this means that we can perform one action to represent the meaning of another. This is possible because contextualizing practices associate one action with another, allowing them to index each other. In representation, an action constructed with the semiotic resources of one system indexes the meaning of an action usually constructed with another. We can index the act of climbing a tree with a sentence, with a picture, with mime, with music, and possibly with other resource systems as well. In most communities, language is probably the most versatile semiotic system in this respect (cf. the ideational or experiential function in Halliday). Other semiotic systems tend to be specialized to represent only some kinds of actions (or objects, events, processes).
The second, closely related function is the ability to make relations or connections between actions, objects, events, processes. There are many kinds of relations obviously: spatial, temporal, sequential, structural, causal, behavioral, possessive, attributive, equative, conjunctive, disjunctive; relations of means, manner, condition, similarity, and so on. (Cf. Halliday's logical function.)
The third and fourth functions are also closely related (cf. Halliday's interpersonal function, with its distinction between Mood and Modality; Halliday, 1985a, pp. 68-94, 332-346). On the one hand there is the ability to interact, to constitute a dialogue, either explicit (represented) or implicit: to make a move to which there are possible responses, and to convey meaning through the relation of move and response that goes beyond what either can mean alone. And then there is the ability to establish an orientation towards one's action and its meaning, a point of view: favorable or unfavorable, serious or joking, literal or metaphoric, committed or uncommitted, tentative or definite, and so forth. Of these, probably the most important for social analysis is®MDBO¯ evaluative orientation. There are a great number of ways in which we indicate whether we approve or disapprove of things, associate ourselves with them or dissociate ourselves from them.
Finally there is the organizational function. A semiotic resource system must provide the means to bind together actions into coherent wholes, to fashion activity structures and other types of formations, and to distinguish between actions randomly strung together and those that are organized according to a pattern into a larger whole. (Cf. the discussion of the textual function and of textual cohesion in Halliday & Hasan, 1976).
ACTIVITY STRUCTURES, GENRES, AND THEMATIC FORMATIONS
An activity structure is a socially recognizable sequence of actions. Actually, it is a little more abstract than that. The action sequence itself is the result of enacting an activity structure. The same activity structure can be realized in many ways, by many actual sequences of actions. Think of all the sequences of actions that can constitute a lesson, or even an exchange in Triadic Dialogue. What all these sequences have in common is their activity structure. It is a structure in the sense that it has parts, each a functionally defined action type (e.g., Teacher Question, Student Challenge), and that these functional elements have specific relationships to one another (e.g., Teacher Evaluation to Student Answer), including restrictions on the order in which they can meaningfully occur.
An activity structure must also be completable, in the sense that you can get to the end of the structure and have a sense of closure to the activity. Activity structures are repeatable in a community, and most of them are repeated frequently, even though any particular action sequence may never be exactly repeated again. That is why is it useful to define activity structures in this abstract way.
In addition to the activity structures of the classroom (Do Now's, Going-Over-Homework, Student-Teacher Debates, Admonition Sequences, etc.), there are recognizable activity structures in every aspect of human life. Every routine and ritual, every activity for which we can specify rules or procedures, every activity that is endlessly repeated with minor variations can be described as an activity structure. Large, complex activity structures can be analyzed as being composed of smaller, simpler ones. Dining-in-a-Restaurant is a major activity structure, and within it we could identify Getting-the-Check as a smaller one. Washing-the-Dishes (hopefully not in the restaurant) is an activity structure, and so are Going-to-the-Movies, Making-a-Phone-Call, Writing-a-Letter, Painting-a-Portrait, Taking-a-Photo, Buying-Stamps, Telling-a-Story, Changing-Diapers, Mowing-the-Lawn, Doing-your-Taxes, Washing-a-Car, Measuring-Blood-Pressure, and so on.
Activity structures can be everyday or special. Many of them in our society are technical, learned and performed by specialists: medical lab procedures (even when automated), scientific measuring techniques, astronomical observations, drawing weather maps, compiling economic statistics, cataloguing new books in a library, fitting new pipes, taking inventory, writing an annual report. Some of these involve the use of language or mathematics; others deploy only other sorts of semiotic resources (including the basic resource of meaningful action itself).
Some human activities may not be regular enough, not similar enough from one performance to another, to be described as activity structures: writing a book, taking a vacation, committing a murder. On the other hand, there are undoubtedly more specialized versions of such activities that are structured in a more or less repeatable way. Part of the job of social and cultural analysis is to identify the regularities in human action.
An activity structure can be interrupted and resumed later; it need not be enacted continuously from start to finish (except in special cases where continuity is an essential relation between its elements). This is possible because the basic relationship between elements in an activity structure is usually not simply that one comes (immediately) after the other. The relationship is functional. An Answer can come long after its Question and still be recognized as the Answer to the Question.
A synonym for activity structure might be "action genre," by analogy with the terms "speech genre" (Bakhtin, 1953) and genre itself, which usually refers to a written genre. The notion of genre is a useful one when it is understood in relation to activity structure. Originally, a genre meant a literary genre: short story, one-act play, epic poem, sonnet, limerick, novel. It referred to certain standard types of literary product that were repeatedly produced in European society. The notion is sometimes extended to other arts: The genres of painting include still-lifes and crucifixions, those of music include concertos and sonatas. But a sonnet or a sonata is not an activity structure. The activity of®MDBO¯ Writing-a-Sonnet or Composing-a-Sonata, however, is. A sonnet is a semiotic production resulting from a performance of the activity structure of Writing-a-Sonnet. The sonnet is also a semiotic text, a sort of indirect and transformed record of the activity of writing (it is usually highly edited, not an accurate record of the actual writing process).
This interpretation of the traditional notion of genre does not always work. A novel is probably not the product of a definable activity structure because writing novels does not seem to result in the degree of similarity in functional structure from one novel to another that is found in other genres like limericks or haiku. Some kinds of novels (formula mysteries, for example) may be specialized and regular enough that writing them is describable as an activity structure. The Novel is not a genre in social semiotics, and Writing-a-Novel is not an activity structure. It is a recognizable activity type, however, because there are semiotic practices in our community by which we can usually decide whether somebody is "writing a novel" or not. Similarly, the Novel is a semiotic text-type because we can decide whether something counts as a novel or not.
Speech genres tend to be less literary and more typical of everyday life. They are also usually integral parts of activity structures that deploy other sorts of action in addition to speech. Formal, extemporaneous speaking is somewhat out of style today, but Toasts and Eulogies, spontaneous Limericks, and even Playing the Dozens in Black English are obvious examples. So, less formally, is the language of buying and selling (Ventola, 1987), or the language of the classroom, or the courtroom (e.g., Walter, 1988). It is not clear whether Casual Conversation is truly an activity structure, or just a recognizable activity type (Eggins & Slade, 1987).
It is particularly important to be clear that activity structures are not rigid formulas for speaking, writing, or doing. Only a very few sorts of ritualized or automatic activities are performed in a mechanical way. Because activity structures are defined at a relatively high level of abstraction from actual performances, they leave a lot of room for dynamic variability (and creativity) in their enactment. To write a sonnet is not usually a mechanical process, but however the parts of the sonnet are produced, they do conform to certain regularities (or else the result is not a sonnet). An activity structure always gives a predominantly synoptic view of human action: it tells what the dynamic performance amounted to in the end, not how it was produced.
We can make a more dynamic account of human action by looking at the moment-to-moment strategies and options within a performance, but the categories we use in doing so are always synoptic ones. This "slippage" between the dynamics of performance and the synoptic perspective of functional analysis is an aspect of the more fundamental incommensurability of analysis in terms of material processes and analysis in terms of semiotic practices. This slippage makes it possible for semiotic systems to change (Lemke, 1984, pp. 143-146).
Probably the most important and neglected kinds of genres are nonliterary written genres (Kress, 1982; Martin, 1985b; Lemke 1988a). From Book Reports to Book Reviews, Persuasive Essays to Editorials, Lab Reports to Research Papers, both as students and as mature writers we need to master many specialized activity structures for writing a variety of genres. Very few of us will write novels, plays, or short stories, but most careers and lives make use of the skills of writing these nonliterary genres. Those genres through which power is exercised in our society particularly need to be analyzed and taught. Synoptically, we need to know what the parts of each genre are, how they are functionally defined and recognized, what their functional relations are to each other, and what possible orderings they can meaningfully have. Dynamically, we need to know at least some of the ways in which each part can be realized, right down to the level of phrase- and sentence- types.
We all know that you can start a fable with "Once upon a time, there lived ...." How many other ways can you start one, and still have the opening be recognized as probably the start of a fable? (Cf. Hasan, 1984b; Halliday & Hasan, 1985, pp. 63-69). How do you start a Book Report? Or a Legal Petition? They too have their rules and conventions and ought to be taught systematically in the curriculum.
The functional elements of a particular genre or activity structure are specific to that one genre or activity. A Teacher Question in Triadic Dialogue (where the teacher knows the answer) does not perform the same function as a Teacher Question in True Dialogue (where the teacher does not), even if we happen to give them the same name. On the other hand, the basic semantic relation of Question to Answer is the same in all forms of dialogue, although at a slightly more abstract level. The Question-Answer pattern, like the Problem-Solution, Cause-Consequence, or Generalization-Example patterns are widely used across many different genres and language-using activities. One can think of these as "mini-genres" that are used to fill the functional slots of the major genres. They are certainly structures in their own right, though typically they only have two or three functional elements of their own. It is useful, however, to recognize that they are more abstract than genres or activity structures proper, and that because of this fact, they can be used in different genres, while retaining their own basic structure. I call them rhetorical structures (cf. Mann & Thompson 1983, 1987).
A piece of written or spoken language, that is, a text, can be analyzed at three sequential levels: as the product of an activity structure which results in a definite sequence of functionally defined and related elements, as a sequence of rhetorical structures that realize these elements and relations, and as a structure of grammatical constructions and words that realize the elements of the rhetorical structures. The scheme of analysis looks something like this:
GRAMMATICAL CONSTRUCTIONS AND WORDS
THEMATIC FORMATIONS, DISCOURSE, AND TEXT
Analyzing the uses of language in a community in terms of activity structures, genres, and rhetorical formations helps us to identify syntagmatic contexts. Whatever is said or written is always a part of some functional element in an activity, and it will have syntagmatic relations to other elements in these larger wholes. A phrase in a Teacher Question is part of the whole question, part of the complete exchange (with relations to Preparation, Answer, Evaluation, and Elaboration), and part of the overall activity of the episode, the lesson, the unit, and the course.
In relation to these activity structures, we can also place the phrase in some of its paradigmatic contexts: What else might the teacher have done at this moment? How else could he or she have asked the question? But there are still other equally important contexts in which we place the phrase to make sense of its full meaning. The phrase may repeat the words used by the teacher or by a student at an earlier point in the lesson, or it may use the wording of a problem or discussion in the textbook. The wording may serve to remind students of something they learned earlier in the year, or even in a different course or outside school. It may even disturb students because it sounds too colloquial or "unscientific."
If the question was about a particular topic in science, then the odds are that there are only a few ways in which the teacher could have asked it. There are only a few, because all the meaningful ways of asking it have to use the same basic semantic relationships among key concepts or terms. The teacher can ask: "What kind of wave is sound?" or "Sound is what type of wave motion?" or "Is sound a longitudinal wave or a transverse wave?" However the question is asked, it has to express the semantic relationship of classification and it has to refer to what is to be classified (sound), and in what category the classification is to be made (waves). It also has to imply, or state, that there is more than one type (the Classifiers) to choose from. The question must fit this thematic pattern of semantic relationships. That pattern is a small piece of a semiotic formation, a recognizable way of talking about this topic in a particular community. Such a thematic formation is another important context for the meaning of the question, or any words used in phrasing it.
Thematic contextualization is the process of placing anything said or written in the context of some larger, familiar thematic pattern of semantic relationships. Because there is often more than one way to express the parts of a thematic pattern in words, the pattern itself has to be defined at a slightly more abstract level than that used to describe wordings. A scientific "concept" (unfortunately, a mentalist term) can always be expressed in different words: sound can be expressed as sound, sound wave, acoustic vibration, pulse, and so on at different points in the same text (or from one text to another). The element of a thematic pattern which can be expressed in all these ways is called a thematic item (to avoid mentalist terms like concept or idea). The web of semantic relationships among different thematic items form the thematic pattern or thematic formation of the topic.
We can use words to construct the semantic relationships of a thematic formation in many different ways, even using very different genres or activity structures in doing so (e.g., poems, essays, debates). Thematic formations are what all the different texts that talk about the same topic in the same ways (semantically) have in common.
The thematic meaning of a word or phrase is the meaning we make for it by placing it in the context of a particular thematic formation. If we are talking about money, bank and deposit will probably be contextualized by their semantic relations to cashier and withdrawal in one particular thematic formation. If we are talking about a river, we will construct meanings for them according to a different pattern that relates them to current and sediment. Saying the words bank or deposit are merely ways of using language to make meanings; the meaning depends on the thematic pattern. A meaning is made by our thematic contextualization of a word, by which pattern we place it in. In many cases, we can place it in more than one pattern. What is true of a word, is true also of phrases, clauses, sentences, and even whole texts.
While it is basically true that thematic formation, genre structure, and particular choice of wording are all in principle independent of one another, in practice these different features of a text index one another and the community in which they are used. We do not see the thematics of science constructed in the genres of poetry as often as we see them in the genres of the textbook or research paper. We do not hear the word choice "pickled rope" as often as we do "salt bridge" or "conduction pathway," when the genre is a textbook and the thematics is that of chemistry. If we do find it, the activity structure is more likely to be informal talk about science than formal writing.
The stylistic norms of science, discussed in Chapter 5, are social conventions (metaredundancy relations, as in Lemke, 1984, pp. 33-44), that link thematics, genre, and wording to make some combinations more likely than others in a particular community. Those patterns of combinations in turn index the community itself, distinguishing one social group from another.
A particular, recognizable combination of thematics, genre, and stylistic choices of rhetorical strategies and word-choices can be called a discourse formation. The actual texts of such a formation make up a very specific text-type. The discourse formation is an important indexical context for the meaning of every part of every text of its text-type. Every Shakespearean sonnet about love is a potentially relevant context for making sense of every other one. Every research article about the theory of superconductivity is potentially a relevant context in which to interpret every other one.
INTERTEXTUALITY AND TEXT SEMANTICS
Social semiotics begins with texts and other records of human behavior or products of human activity. It does not begin with activity structures, genres, or thematic formations. Semiotic formations are abstractions from texts: They are the common patterns shared by many similar texts. They describe how we make meaning by placing actions and words in some contexts rather than others, connecting them to certain other actions and certain other words. Every community or social group has its own characteristic ways of making meaning, its own ways of contextualizing and connecting, its own activity structures, genres, and thematic formations.
When we participate in an activity, read a text, or make sense of talk and other forms of socially meaningful action, we connect words or events up in familiar patterns. They may be words and events in the same text or action sequence, or they may be words and events from different texts or times. This is the principle of general intertextuality (Lemke, 1985a, in press-b): Everything makes sense only against the background of other things like it.
The intertexts of a text are all the other texts that we use to make sense of it. Some of them are texts that share the same thematic pattern (cothematic texts). Others belong to another element in the same activity structure (coactional texts), or have the same genre structure (cogeneric texts). A poem and a textbook passage, both about evolution, may be cothematic. A speech by a defense lawyer and the text of a letter entered in evidence in the same trial may not necessarily be cothematic, but they are coactional. Any two limericks are cogeneric.
The text-connecting practices of a community are an important part of its ways of making meaning. We can make meanings through the relations between two texts that cannot be made within any single text.
Real texts and sequences of actions are not pure, ideal types. They do not necessarily stick to one genre or one thematic formation. Many texts mix different thematic formations, often creatively. Some texts also mix different genres. Semiotic analysis is not a straightjacket, it is a systematic expression of how all of us make sense of texts and events, including the irreducible ambiguities and multiple meanings we find.
A real text will have many thematic strands, many thematic formations which it links together to make its arguments. At each point in a text, there will be one or a few thematic formations that will normally be used to interpret the text there. In addition to its generic or activity structure, a text is also organized by the ways in which these thematic formations run through the text. Like a piece of music, there are various themes that appear, disappear, return, are transformed and linked to other themes (see Lemke, 1988d, in press, b). This sort of thematic analysis of the meaning of a text is part of general text semantics.
Semiotic productions in general, of course, use resources other than (or in addition to) language. A thematic formation does not have to be expressed only through the medium of language. Very often when speaking we use gestures or objects around us, rather than words, to stand for thematic items or express relationships. In writing we can also use pictures for these purposes, and various multimedia productions (which will become more common with the use of computer systems) can express thematic formations through the deployment of many kinds of semiotic resources. For a discussion of codeployment of different semiotic resource systems, see Lemke (1987a).
Just as communities have specific ways of connecting texts to one another, thematically or actionally, so they also construct characteristic patterns of relationships among activity structures and thematic formations themselves. Some activity structures are considered to be rival ways of doing things, or are characteristic of social groups that may be in conflict with one another. Certain thematic formations are also set up as being opposed to one another, as being contradictory or representing conflicting ways of talking about a subject. In order to understand the importance of these relationships in analyzing the social practices of a community, there is one more aspect of semiotic formations we need to discuss.
So far we have talked mainly about how formations express three of the five fundamental semiotic functions: representation, relation, and organization. To the extent that we have emphasized the importance of dialogue, we have also touched on a fourth: interaction. But the fifth function, orientation, is the one that enables us to communicate our attitude or stance toward what we are saying or doing, and especially our evaluation of it: whether we value it positively or negatively. The resources of language and social action enable us to sneer and ridicule as well as admire and promote, to dissociate ourselves from an action or way of speaking as well as to take it for our own.
Everything we say, write, or do carries with it an evaluative orientation. It is "colored" by ways of saying or doing that indicate our attitude and stance toward what we do (Lemke, 1988b, in press-a, in press-e). We can speak or act with reluctance or enthusiasm, disapproval or endorsement. We can express our stance explicitly in so many words, or more subtly in tone, body language, facial expression, pacing, or choice of words. In a community where the attitudes of different groups can be taken for granted, we need only index a group by one of its views or characteristics to project its attitudes toward anything else.
When used in a text or discourse formation, a thematic formation is always given, in a particular community or social group, not only a set of semantic relations to other formations, but also a set of value relations (or axiological relations). Formations are regarded as good or bad, right or wrong, appropriate or inappropriate. This is part of the means by which they are made to oppose each other, complement each other, or directly ally with each other (Lemke, in press-b). Since social groups are indexed by (and identify closely with) the formations they use and their stances toward them, these kinds of relationships between different ways of speaking and acting play an important role in the social dynamics of a community. They form the community's systems of heteroglossia and heteropraxia. These notions are essential to the ways in which social semiotics analyzes a community.
HETEROGLOSSIA AND DISJUNCTION
One of the most important social facts about how people use language (and all the other systems of semiotic resources) is that people from different communities deploy these resources differently. Every different social group and category makes different meanings. They have different activity structures, different thematic formations, and different probabilities for which formations and activities will be used by whom and when (i.e. under what circumstances, in what contexts). Not just people from what we recognize as different countries or cultures, but people from different groups within a single social community, and even people who simply belong to significantly socially different categories in the community act and speak differently. People who differ in age, sex, social class, religion, occupation, and political views, for example, whether they form distinct communities or not, talk about many subjects differently and act differently in many situations.
From the way a person talks about a subject, especially if it is a controversial one (e.g., abortion, Gay rights, feminism, legalization of drugs, euthanasia, military spending, nuclear reactors and weapons, multinational corporations, etc.) you can tell a lot about the social groups and categories they belong to or have been influenced by. But even in much more everyday matters, like choice of newspaper or preference in clothing, people from different social groups behave differently and talk differently (Bourdieu, 1979/1984).
Different social groups and categories correspond to different patterns of combination of social practices. From those differences arise differences of interests and differences in values. Included in these differences are differences in tastes and preferences, differences in opinions and beliefs, and differences in propensities for actions of various kinds in various situations.
If we were to take any topic (e.g., see Lemke, 1988b on Gay rights) in any community, we could identify a certain relatively small number of basic ways of talking about that topic (i.e., different thematic formations). Nearly everything said or written by anybody could be analyzed as some combination of elements from these formations, and would vary systematically with social group and social interests. But in addition to this, we would find that each group also had its opinions of the views of the other groups.
For example, on the issue of Gay rights, there is the Moral Majority's view of homosexuality as "willful and sinful," calling the secular scientific view (that it is a normal sexual variation) "blasphemous," while civil libertarians opppose the Moral Majority's positions for violating the separation of church and state, and mainstream religious denominations denounce their methods of interpreting the Christian bible. Each social group has its own way of discussing the subject and its own stance toward the others. There are about a dozen different positions on this issue, and their innumerable combinations (Lemke, 1988b).
From any social position in the community, we can construct a picture of how the different possible viewpoints on an issue are talked of as being allied, opposed, complementary, and so on. This is the system of heteroglossia for that topic, from that social position, in that community. The term heteroglossia derives from the work of Bakhtin (1935, pp. 262-300), who first emphasized the importance of social differences in the use of language for analyzing society.
While heteroglossia (diversity of ways of speaking) covers the differences of thematic formations and the different value orientations of each group to those formations, we need a more general term to cover the similar sorts of differences in forms of social action other than just speaking. For this I use the term heteropraxia (diversity of ways of acting). A complete description of the system of heteropraxia (or even heteroglossia) for any community would be very complex indeed. But what these concepts give us is an important perspective: They remind us that not only do different social groups and categories speak and act differently, but that these differences form a system. That is, they are not just differences: They are differences that are systematically related to each other. The ways in which different groups speak about a subject, or tend to act, are part of the dynamics of their overall social relations to one another: their conflicts and alliances in matters of economic and political interests and power (Lemke, in press-d).
Within the system of heteroglossia, any two thematic formations which people regard as being more or less "about the same thing" (see Lemke, in press-b) have two different sorts of relations to one another. One kind, obviously is their thematic-semantic relations: how do they construct similar or different semantic relations among similar or different thematic items. The other kind is their axiological relations: What value orientations do they take toward each other? These two are closely interrelated because each social discourse voice (thematic-formation-plus-value-orientation, as spoken by a particular group) tends to reconstruct the thematics of each of the others in the process of approving or disapproving of them (cf. Lemke, 1988b, in press-a). It is thus discourse voices, rather than abstract thematic formations alone, which constitute the basic elements of a system of heteroglossia.
Discourse is a mode of social action. It is not just language, but language-in-use in a community. A discourse voice does not just speak in a particular way about a topic (thematics), even with a value orientation to other voices (axiological stance). It is always doing some social work in the community, it is always "up to something." Very often, as we have seen in Chapter 5, the use of a particular way of talking (e.g., using the stylistic norms of formal science, or promoting the technocratic view of scientific objectivity) tends to promote certain social interests at the expense of others, and help people from some social groups while hindering others, even when we are not aware of this, and even when we would not want to do it if we were aware. This phenomenon is often called the ideological use of language.
In social semiotics, an ideology is a discourse voice that systematically promotes the social interests of a privileged or powerful group while at the same time disguising the fact that it is doing so. It is important to note that a thematic formation is not necessarily in and of itself ideological. It is the way it is used in the community that determines its ideological force. The same thematic formation can often be used by several groups whose interests are in fundamental conflict.
SOCIAL SEMIOTICS AND SOCIAL CHANGE
The brief sketch of principles in this chapter has taken us from a theory of how people act meaningfully in a community to a model of the structure and conflicts among factions within that community. At the same time, social semiotics recognizes that communities change and that their dynamics depend on human actions seen both as semiotic practices and as material processes. The ultimate challenge for a social science based on the principles of social semiotics is to tell us what we can predict about the future of a community and what we cannot.
It is probably not possible for any individual or group to directly control the future of a social system. The processes which organize, sustain, and modify large, complex communities operate on a vastly different scale than do their parts, whether groups or individuals. Communities live on a different time scale than we do, and the processes which change them are processes operating between groups, between discourse voices, and most generally within and between systems of heteropraxia.
Communities also change because of rapid or gradual responses to slippages or mismatches between the complex systems of relations among their semiotic practices and the equally complex but very different systems of relations among the material processes which sustain them (and which include the semiotic practices themselves). Individuals and groups can no more control the future of a community in the long term than a cell or organ in our body can control our lives. One cancerous cell or defective organ can perhaps end our life, but no single mutant cell or specialized organ, even the brain, determines the overall course of human development and individual response to an environment which the organism itself partly controls.
The tools of social semiotics provide the means to refine our present, rather hazy models of social systems in the large. Even critical sociologies only give us a starting point. Notions like heteroglossia enable us to relate individual actions to wider social processes by way of intermediate constructs such as activity structures, thematic formations, and discourse voices. We can look at the ways in which the activity structures of a community converge differently in different types of individuals, creating and subdividing significant social categories. We can redefine and refine notions like social class, gender, and life stage. We can reconnect individual and groups interests with their systems of values and patterns of action. But, most importantly, we can now begin to analyze the dynamic relationships between two patterns: the pattern of social relations among actions viewed as semiotic practices and the pattern of biophysical interactions among those same actions viewed as material processes.
A community is simultaneously a social system and a dynamic, open biophysical system (Lemke, 1984, pp. 104-112; in press-d). As a material system, it belongs to a special and important class of systems: those which maintain their existence by continual dynamic interaction with their environments. Like open flames, hurricanes, biological cells, and independent organisms, communities have particular nonequilibrium thermodynamic properties that lead to internal self-regulation, exchange of matter, energy, and information among subsystems and with the environment, and a developmental pathway characteristic of their type or species. Along that pathway they undergo discontinuous changes in their internal organization, build up internal complexity, and export entropic disorder to their environments. They pass through successive stages of ascendance, maturity, and senescence in which they tend to respond differently to the challenges of a changing environment (cf. Salthe, 1985, in press).
Even more than this, communities are systems whose types can evolve, because the material base of their semiotic practices can preserve information, accomodate variability, and transmit information to future communities of the same or a successor type (Lemke, in preparation). In this way communities are very much like biological organisms, though in other respects they are a quite different kind of system. The community maintains itself, follows the developmental course specific to its type, undergoes at the same time an individuation unique to its history, and contributes to an evolutionary lineage of communities over longer periods of time.
A complete discussion of the developmental-evolutionary model of social change would require several more chapters or another book (See Textual Politics, 1995). Here it may be enough to point out that the developmental forces in the lifetime of a community, or even of one of its components, in affecting the dynamics of the material processes of the system, also play a role in the changes in our meaning-making practices: in the life-histories of activity structures, thematic formations, and discourse voices as social phenomena. By correctly defining the various constituent subsystems of a community, we can in principle distinguish predictable, type-specific developmental changes from individuating community-specific changes and inherently unpredictable evolutionary changes. Across various time scales and communities there will be a spectrum from (probabilistically) predictable to unpredictable responses to internal and external forces.
SOCIAL SEMIOTICS AND POLITICAL RESPONSIBILITY: MAKING TROUBLE
I want to conclude this chapter by projecting from the model of social systems sketched here a few conclusions about our own responsibility for social change. Social systems are neither objects nor machines, and we are inside and part of them, not outside and independent of them. We cannot deceive ourselves that we can guide and control them with the help of social science the same way we manipulate our conventional technologies on the basis of physical and biological science. We have a better chance of "managing" even the ecosystems of which we are also a part than we do our own social systems, though we would do better with ecosystems if we at least tried to imagine a viewpoint larger than that of our own interests. In the case of social systems, even our viewpoints are necessarily contained within the system: we cannot imagine a viewpoint outside our own system of heteroglossia (in trying to do so we merely enlarge that system, remaining inside it).
Every attempt to construct an overview of the whole of a social system, encompassing all the interests and differences represented within it, will always necessarily be constructed from the perspective of some single social position within the system, and therefore from the viewpoint of a particular set of social interests and values. Neither traditional liberal politics nor the newer technocratic ideology of rational management (see Lemke, in press-a) want to face this issue.
The traditional view is that we can rely on a political class (lawyers, politicians, their advisors) to represent among themselves the interests of social groups and categories to which they themselves do not belong and whose patterns of social practices (i.e., lifeways) they do not share. The newer technocratic position is that complex problems can be solved by expertise without considering conflicts of values and interests or the diversity of viewpoints in society. The traditional political class and the new technocrats claim to be representing, in the one case, everyone's interests, and in the other, a purely objective viewpoint, but in neither case do they admit that they must in the long run represent the only interests they can: their own.
No individual, group, or class, can represent -- or even imagine -- a complete, representative, or positionless view of the whole of a social system. In the case of the schools, no principal or administrator can claim to be able to make decisions that truly take into account the interests of teachers, students, parents, and the public in all their real diversity. Neither, for society as a whole, can politicians or even elected legislatures as presently constituted do so. And certainly no class of professional experts or managers, even educational researchers or social scientists, can legitimately claim to be able to make "objective" decisions.
Society itself, as a whole, is the only complete representation of the social system. Ideally, decisions should be made, after issues and interests are articulated by each social constituency in its own terms, by the whole of the community. Short of that, any representative system of decision making, whether in schools, corporations, or government bodies, needs to seek representation by members of diverse social categories, and means to articulate the interests of these constituencies.
Any such system of decision making would result in much more open conflict of views than now exists. In our present system of social ventriloquism, a relatively homogeneous few speak for a narrow range of their own interests in the name of all. This minimizes the appearance of conflict, but the reality of conflict remains. Denied a forum and political power, the full diversity of social interests, values, and viewpoints creates a dangerous social instability. This instability arises from the long-term contradictions between decisions which reflect the special interests of the decision makers and the full spectrum of actual social interests in the community.
Let us not romanticize the social upheavals and political revolutions which result. They are as cruel and devastating as any war; they bring about enormous real suffering, and they can wreck the complex and technologically vulnerable economic base of a modern society. Neither those who benefit from the status quo, nor those who seek to change it radically should be eager for a violent confrontation of conflicting social interests. Just as national wars are pointless for modern nations in the nuclear age, so are violent revolutions and civil wars too potentially devastating to consider in complex, technologically vulnerable societies. In both cases there would be little left for the victors to govern. Far better to institutionalize these conflicts, and all future conflicts of interest, in the political process itself and strain the political system to its limits, rather than destroy the social system itself.
I am speaking here of changes in institutions: in activity structures in politics and decision-making, in the range of discourse voices to be heard on every issue, and in the distribution of power among social groups and categories. These changes do not have to begin with the national government, and are not likely to. They can begin with individual institutions: schools, social organizations, businesses, advocacy groups, local governments, and so on. Pressure for changes in the distribution of power in major corporations or government bodies will require that people first experience new ways of making decisions on a smaller scale.
We need to build confidence in our ability to articulate viewpoints that have never had the opportunity to speak for themselves effectively. We need to learn that we can cope with far more open conflict of interests and values than we have in the past. We need to see that better decisions are made when all interests are truly represented, and that no single viewpoint can ever be either comprehensive or objective. We need to grow jealous of our right to speak for ourselves.
New social practices can spread in a community when they meet needs that may never have been adequately articulated before. In any social system there are some meanings that are never made, some connections that are never put into words, some actions that are not yet part of the social repertory because to make, say, or do them would be to change that system against the interests of those who regulate our social practices (see on the "system of disjunctions" in Lemke, 1984, pp. 131-150, 1985a). But such regulation, and the ideologies which rationalize it, only exacerbate the contradictions between what people do and what it is possible to say. This is a crucial source of the slippages between social action analyzed as semiotic practice and as material process.
We always do more than we recognize that we do. There are interests arising from patterns of social action that do not yet have social meaning or recognition. There is more potential meaning to every act than what its definition as a semiotic practice recognizes at any one point in history. Some significant social patterns are always "presemiotic" in this sense. The social power of the status quo is especially vulnerable to changes which articulate and make recognizable actions that express previously unarticulable interests. It is here that social semiotics can be a guide to effective political action.
What an individual or a group does only matters to the social system as a whole to the extent that it spreads and leads to changes in the relations among groups in the system. That is most likely to happen when what we do operates in the critical zones of vulnerability of the status quo: those places where its power is maintained by the absence of alternatives rather than by force.
It is in those areas where there seems to be no alternative to the status quo that any real alternative can begin to upset it. Constructing those alternatives means finding new ways to talk about problems that matter to people, ways that engage their real interests and their truest values. Constructing them often means adopting new points of view, fashioning new ways of talking about education and learning, about politics and decision making. In this, social semiotics offers far more than a new way of talking science, it offers what is inevitably a new way of making trouble, and hopefully a new way of making sense.
For References, see J.L. Lemke, Talking Science (Ablex, 1990).